Development of an index of landscape development intensity for predicting the ecological condition of aquatic and small ...

MISSING IMAGE

Material Information

Title:
Development of an index of landscape development intensity for predicting the ecological condition of aquatic and small isolated palustrine wetland systems in Florida
Physical Description:
Mixed Material
Language:
English
Creator:
Vivas, Manuel Benjamin
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla
Publication Date:

Subjects

Subjects / Keywords:
Ecological, Florida, lakes, land, landscape, LDI, streams, water, wetlands
Dissertations, Academic -- Environmental Engineering Sciences -- UF
Environmental Engineering Sciences thesis, Ph. D
Genre:
government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Abstract:
ABSTRACT: Freshwater ecosystems, while vital components of landscapes and essential for human well-being, are increasingly threatened by the ever-escalating intensity of human development. In this study, the main objective was to analyze the influence of human development intensity on the ecological condition and water quality of isolated forested wetlands, streams, and lakes in Florida. An index of Landscape Development Intensity (LDI), derived from the non-renewable areal empower density of land use, was used as a measure of the human disturbance gradient against which the ecological condition and water quality of 118 isolated palustrine forested wetlands, 69 streams, and 54 lakes were analyzed at different landscape scales. Landscape pattern metrics were also calculated for study ecosystems and tested for relationships with indicators of ecosystem condition and water quality. Overall, the LDI had the greatest predictive ability for bioindicators of ecological condition in wetlands and streams, explaining up to 30% and 27% of variability, respectively. The LDI was a significant factor in explaining the variability of water quality variables only for streams. Changes in landscape scale (grain and extent) had small effects on the LDI. Differences in LDI scores were noticeable when developed lands were added with increasing area.
Abstract:
The use of distance-weighting functions provided little enhancement of the predictive power of the LDI; distance-weighted LDIs did increase by 7% the predictive power for bioindicators for streams. Landscape pattern metrics explained up to 44% of variability in bioindicators of wetland condition, and 22% and 42% for stream and lakes. They also accounted for up to 60% and 39% in the variance of water quality variables for streams and lakes, respectively. When included with the LDI in multiple regressions they increased by 25% the amount of variance explained in bioindicators of wetland condition and 52% for lakes. In general, the LDI had higher predictive power for bioindicators of ecosystem condition than for chemical constituents of the ecosystems studied, which are more variable with season, time of day, and hydrologic conditions. The LDI may have greater correlation with bioindicators because they may be more integrative of anthropogenic impacts and have higher correlation with ecological condition.
Thesis:
Thesis (Ph. D.)--University of Florida, 2007.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Manuel Benjamin Vivas.
General Note:
Title from title page of source document.
General Note:
Document formatted into pages; contains 341 pages.
General Note:
Includes vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 003904717
oclc - 660156469
System ID:
AA00003991:00001


This item is only available as the following downloads:


Full Text

PAGE 1

DEVELOPMENT OF AN INDEX OF LANDS CAPE DEVELOPMENT INTENSITY FOR PREDICTING THE ECOLOGICAL CONDITION OF AQUATIC AND SMALL ISOLATED PALUSTRINE WETLAND SYSTEMS IN FLORIDA By MANUEL BENJAMIN VIVAS A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2007 1

PAGE 2

2007 Manuel Benjamin Vivas 2

PAGE 3

To Sarita and Sofa Manuela. 3

PAGE 4

ACKNOWLEDGMENTS I would like to acknowledge th e invaluable and unconditional support provided by my advisor, Dr. Mark T. Brown, at the Howard T. Odum Center for Environmental Policy at the University of Florida. Dr. Brown was unfailing in his encouragement, advice, and vast repository of knowledge provided to me dur ing research and writing. Dr. Cl ay L. Montague, Dr. Michael W. Binford, and Dr. Clyde F. Kiker offered thoughtful advice and input during the process. Additional acknowledgment is due to the resear ch group from the H. T. Odum Center for Wetlands, especially Kelly Reiss, Jim Surdic k, Mike Murray-Hudson, and Charles Lane. The data collection phase, as well as so me of the ideas that led to this dissertation was a team effort. Dr. Larry Winner and Dr. Bhramar Mukherjee from the University of Floridas Department of Statistics provided advice on the statistical analyses used. I would like to thank the Tropical Conservation and Development Program, Center for Latin American Studies, University of Florida and the Office of Gradua te Minority Programs, Research and Graduate Programs, University of Florida for financial support. I will always be grateful to Dr. Mark T. Brown for his fina ncial support when funds were running short. Finally, I would like to thank my wife Sa ra Vivas for her encouragement, support, patience, and unconditional love. Her editorial co mments were also invalu able to this work. I hope I can make up for the playground time I missed with my daughter Sofia when I was working. I love them both. 4

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........9 LIST OF FIGURES.......................................................................................................................16 LIST OF ACRONYMS.................................................................................................................19 ABSTRACT...................................................................................................................................20 CHAPTER 1 INTRODUCTION................................................................................................................. .22 Statement of the Problem....................................................................................................... .22 Plan of Study...........................................................................................................................25 Concepts and Approaches to Understanding Human Impacts on Freshwater Systems.........27 Emergy and Empow er Density........................................................................................27 Human Disturbance Gradient..........................................................................................29 Landscape Pattern............................................................................................................29 Importance of Scale.........................................................................................................30 Watershed as the Unit of Analysis..................................................................................31 Previous Studies......................................................................................................................32 Emergy Studies................................................................................................................3 2 Watersheds studies using emergy............................................................................33 Landscapes and land use..........................................................................................35 Measures of the intensity of landscape development...............................................36 Landscape Pattern: Quantific ation and Application........................................................38 Selection of pattern metrics......................................................................................39 Landscape influences on ecosystem condition.........................................................40 Landscape scale influences on fr eshwater ecosystem condition..............................42 Study Area: Florida, an Overview..........................................................................................45 Physical and Ecological Aspects.....................................................................................45 Land Use..........................................................................................................................48 Freshwater Ecosystem Degradation................................................................................49 Scale of Investigation......................................................................................................... ....50 2 METHODS...................................................................................................................... .......54 Site Selection..........................................................................................................................54 Isolated Forested Wetlands..............................................................................................54 Streams........................................................................................................................ ....55 Lakes................................................................................................................................56 Data Sources...........................................................................................................................57 5

PAGE 6

Land Use / Land Cover Data and Land Use Classification Systems..............................57 Land Use Intensity Classification....................................................................................58 Water Chemistry Data.....................................................................................................59 Isolated forested wetlands........................................................................................59 Streams.....................................................................................................................60 Lakes........................................................................................................................62 Biological Data................................................................................................................ 62 Isolated forested wetlands........................................................................................62 Streams.....................................................................................................................63 Lakes........................................................................................................................64 Scales of Analysis............................................................................................................. ......65 Isolated Forested Wetlands..............................................................................................65 Streams........................................................................................................................ ....66 Lakes................................................................................................................................68 Landscape Indices.............................................................................................................. .....68 Landscape Development Intensity Index........................................................................69 Method of calculation...............................................................................................69 Scale dependency.....................................................................................................70 Landscape Pattern Metrics..............................................................................................71 Metric selection........................................................................................................71 Landscape metric calculation...................................................................................73 Data Analysis..........................................................................................................................74 Study Sites, Water Chemistry Variab les, and Biological Variables...............................74 Landscape Development Intensity Index........................................................................74 Description and behavior.........................................................................................74 Relationship between the LDI and ecosystem condition.........................................75 Landscape Pattern Metrics..............................................................................................76 Description and selection.........................................................................................76 Influence of landscape pattern on ecosystem condition...........................................77 The LDI, Landscape Pattern, and Ecosystem Condition.................................................78 3 RESULTS...................................................................................................................... .......104 Land Use/Land Cover Composition of the Freshwater Systems..........................................104 Isolated Forested Wetlands............................................................................................104 Streams and Lakes.........................................................................................................106 Description of the Landscape Development Intensity Index................................................107 Scale Dependence: Grain Size.......................................................................................107 Isolated forested wetlands......................................................................................107 Streams...................................................................................................................108 Scale Dependence: Spatial Extent.................................................................................110 Isolated forested wetlands......................................................................................110 Streams...................................................................................................................111 Lakes......................................................................................................................112 Relationship between Land Use Intensity and Ecosystem Condition...........................113 Isolated forested wetland condition........................................................................113 Stream condition....................................................................................................117 6

PAGE 7

Lake condition........................................................................................................121 Landscape Pattern Metrics....................................................................................................12 2 Isolated Forested Wetlands............................................................................................122 Metric selection: grain size....................................................................................122 Regression analysis................................................................................................127 Streams........................................................................................................................ ..129 Metric selection: grain size....................................................................................129 Metric selection: spatial extent...............................................................................132 Regression analysis: grain size...............................................................................134 Regression analysis : spatial extent.........................................................................135 Lakes..............................................................................................................................138 Metric selection: grain size....................................................................................138 Metric selection: spatial extent...............................................................................140 Regression analysis: grain size...............................................................................143 Regression analysis : spatial extent.........................................................................144 Land Use Intensity, Landscape Pattern, and Ecosystem Condition.....................................145 Isolated Forested Wetlands............................................................................................145 Streams........................................................................................................................ ..146 Lakes..............................................................................................................................148 4 DISCUSSION................................................................................................................... ....228 Summary...............................................................................................................................228 Spatial Properties of the LDI................................................................................................22 9 Effects of Changing Grain Size on the LDI..................................................................229 Effects of Changes in Extent on the LDI.......................................................................231 Distance Weighting Factors..........................................................................................232 Land Use Intensity and Ecosystem Condition......................................................................233 Biological Indices..........................................................................................................233 Gradients of Change and Thresholds............................................................................235 Water Quality................................................................................................................237 Correlations with Changes in Grain Size......................................................................240 Biological indices...................................................................................................240 Water quality..........................................................................................................241 Correlations with Changes in Extent.............................................................................242 Biological indices...................................................................................................242 Water quality..........................................................................................................244 Land Use Intensity and Distance-Weighting.................................................................245 Landscape Pattern and Ecosystem Condition.......................................................................247 Correlations with Changes in Scale...............................................................................247 Water quality..........................................................................................................247 Biological indicators..............................................................................................252 Landscape Pattern, Land Use Intensity, and Ecosystem Condition.....................................256 Limitations and Further Research.........................................................................................259 Conclusions...........................................................................................................................260 7

PAGE 8

APPENDIX A ENERGY CIRCUIT LANGUAGE......................................................................................263 B LAND USE/LAND COVER CLASSIFICATION SYSTEM..............................................264 C SUMMARY OF EMERGY EVAL UATIONS FOR LAND USES.....................................269 D WATER CHEMISTRY DATA FOR THE SAMPLE FRESHWATER SYSTEMS...........271 E WETLAND CONDITION INDEX......................................................................................275 F STREAM CONDITION INDEX.........................................................................................277 G LAKE CONDITION INDEX...............................................................................................279 H MFWORKS SCRIPTS.........................................................................................................281 I LAND USE/LAND COVER SURROUNDING THE ISOLATED FORESTED WETLANDS........................................................................................................................283 J LAND USE/LAND COVER SURROUNDING STREAMS..............................................286 K LAND USE/LAND COVER SURROUNDING LAKES....................................................288 L LDI SCORES FOR THE ISOL ATED FORESTED WETLANDS.....................................290 M LDI SCORES FOR STREAMS...........................................................................................299 N DESCRIPTIVE STATISTICS FOR LANDSCAPE PATTERN METRICS.......................305 O LANDSCAPE INDICES AND INDICATO RS OF ECOSYSTEM CONDITION.............315 LIST OF REFERENCES.............................................................................................................323 BIOGRAPHICAL SKETCH.......................................................................................................341 8

PAGE 9

LIST OF TABLES Table page 2-1 Surrounding land use and date of sampli ng for 118 isolated forested wetlands in Florida................................................................................................................................79 2-2 STORET sampling station numbers, drainage basins, Hydrologic Unit Codes (HUC), and bioregions for 69 study streams in Florida..................................................................82 2-3 STORET station numbers, lake name, Hydrologic Unit Codes (HUC), and ecoregions for 54 study lakes in Florida............................................................................84 2-4 Summary information of the land use data used................................................................86 2-5 Land uses and definitions.................................................................................................. .88 2-6 Intensity of human development clas sification and non-renewable and purchased areal empower density fo r selected land uses....................................................................90 2-7 Description of landscape pattern metrics selected for this study.......................................92 3-1 Percent values for the land use/land cover (LU/LC) for a priori defined 200-meter buffer areas for the sample of isol ated forested wetlands (n = 118)................................150 3-2 Summary statistics of the non-renewable and purchased areal empower density (E+14 sej/ha/yr) for a priori defined buffer area classes of the isolated forested wetlands...........................................................................................................................150 3-3 Summary statistics on the size and the land use/land c over (LU/LC) composition for the drainage areas for the sample st reams (n = 69) and lakes (n = 54)............................150 3-4 Summary statistics of the non-renewable and purchased areal empower density (E+14 sej/ha/yr) for the sample streams and lakes..........................................................151 3-5 Spearman correlation between the three fo rms of the LDI calculated for the sample isolated forested wetlands at three different spatial extents.............................................151 3-6 Spearman correlations between the three forms of the LDI calculated for the sample streams at three different spatial extents..........................................................................151 3-7 Spearman correlations between the three forms of the LDI calculated for the sample lakes at three different spatial extents..............................................................................152 3-8 Simple linear regression values (r2) for regressions between the LDI and the water chemistry variables measured at three la ndscape extents for the sample isolated forested wetlands.............................................................................................................152 9

PAGE 10

3-9 Simple linear regression results (r2) showing the proportion of total variance in each of five water chemistry variables explaine d by the LDI calculated at eight different grain sizes (meters on a side) for the isolated forested wetlands.....................................153 3-10 Simple linear regression values (r2) for regressions between the LDI and the WCI measured at three spatial extents for th e sample isolated forested wetlands...................153 3-11 Simple linear regression results (r2) showing the proportion of total variance in each of the three WCIs explained by the LDI in its three forms calculated at different grain sizes (meters on a side) for the sample isolated forested wetlands.........................154 3-12 Coefficients of determination (r2) for simple linear regressions between the three forms of the LDI and the water chemistr y variables and the Water Condition Index (WQI) measured at three spatial extents for the sample streams.....................................155 3-13 Simple linear regressions (r2) showing the proportion of tota l variance in each of five water chemistry variables and the WQI e xplained by the LDI calculated at six different grain sizes (meter s on a side) for streams.........................................................156 3-14 Simple linear regression values (r2) for regressions between the three forms of the LDI and the SCI measured at three spa tial extents for the sample streams.....................156 3-15 Simple linear regression results (r2) showing the proportion of total variance in the SCI explained by the LDI in its three form s calculated at six di fferent grain sizes (meters on a side) for the sample streams........................................................................157 3-16 Coefficients of determination (r2) for simple linear regressions between the three forms of the LDI and the water chemistry variables and the Lake Condition Index (LCI) measured at three spatial extents for the sample lakes..........................................157 3-17 Simple linear regression results (r2) showing the proportion of total variance in each of five water chemistry variables and the LCI explained by the LDI in its three forms calculated at six different grain si zes (meters on a side) for lakes...................................158 3-18 Pearsons correlations betw een landscape pattern metrics cal culated at four different grain sizes (meters on a side) for the isolated forested wetlands.....................................159 3-19 Eigenvalues and variance explained by the first seven axes for the principal components analysis at four different grai n sizes (meters on a si de) for the isolated forested wetlands.............................................................................................................161 3-20 The 5 x 5-meter grain size for isolated forested wetland buffers: principal component matrices showing pattern me tric factor loadings.............................................................161 3-21 The 10 x 10-meter grain size for isolated forested wetland buffers: principal component matrices showing patte rn metric factor loadings..........................................162 10

PAGE 11

3-22 The 20 x 20-meter grain size for isolated forested wetland buffers: principal component matrices showing patte rn metric factor loadings..........................................162 3-23 The 30 x 30-meter grain size for isolated forested wetland buffers: principal component matrices showing patte rn metric factor loadings..........................................163 3-24 Coefficients of determination, probabili ties, and regression equations for multiple regressions between indicators of ecos ystem condition and significant components resulting from the PCA of landscape pattern metrics at four grain sizes for the sample isolated forested wetlands....................................................................................163 3-25 Pearsons correlations betw een landscape pattern metrics cal culated at four different grain sizes (meters on a side ) for the sample streams......................................................166 3-26 Eigenvalues and variance explained by the first six axes for the principal components analysis at four different grain sizes (meters on a side ) for stream watersheds...............168 3-27 The 20 x 20-m grain size for stream wa tersheds: principal component matrices showing pattern metric factor loadings............................................................................168 3-28 The 50 x 50-m grain size for stream wa tersheds: principal component matrices showing pattern metric factor loadings............................................................................169 3-29 The 80 x 80-m grain size for stream wa tersheds: principal component matrices showing pattern metric factor loadings............................................................................169 3-30 The 110 x 110-m grain size for stream wa tersheds: principal component matrices showing pattern metric factor loadings............................................................................170 3-31 Pearsons correlations betw een landscape pattern metrics calculated at three spatial extents for the sample streams.........................................................................................171 3-32 Eigenvalues and variance explained by the first six axes for the principal components analysis at three different spatial extents for streams......................................................173 3-33 The 100-meter spatial extent for stream s: principal component matrices showing pattern metric factor loadings..........................................................................................173 3-34 The 400-meter spatial extent for stream s: principal component matrices showing pattern metric factor loadings..........................................................................................174 3-35 The watershed spatial extent for stream s: principal component matrices showing pattern metric factor loadings..........................................................................................174 3-36 Coefficients of determination, probabili ties, and regression equations for multiple regressions between indicators of ecos ystem condition and significant components resulting from the PCA of landscape pattern metrics at four grain sizes for the sample streams.................................................................................................................175 11

PAGE 12

3-37 Coefficients of determination, probabili ties, and regression equations for multiple regressions between indicators of ecos ystem condition and significant components resulting from the PCA of landscape pattern metrics at three spatial extents for the sample streams.................................................................................................................177 3-38 Pearsons correlations betw een landscape pattern metrics cal culated at four different grain sizes (meters on a side ) for the sample lakes..........................................................179 3-39 Eigenvalues and variance explained by the first six axes for the principal components analysis at four different grain sizes (meters on a side) for lakes watersheds.................181 3-40 The 20 x 20-m grain size for lake waters heds: principal component matrices showing pattern metric factor loadings..........................................................................................181 3-41 The 40 x 40-m grain size for lake waters heds: principal component matrices showing pattern metric factor loadings..........................................................................................182 3-42 The 60 x 60-m grain size for lake waters heds: principal component matrices showing pattern metric factor loadings..........................................................................................182 3-43 The 80 x 80-m grain size for lake waters heds: principal component matrices showing pattern metric factor loadings..........................................................................................183 3-44 Pearsons correlations betw een landscape pattern metrics calculated at three spatial extents for the sample lakes.............................................................................................184 3-45 Eigenvalues and variance explained by the first six axes for the principal components analysis at three different spatial extents for lakes..........................................................186 3-46 The 100-m spatial extent for lakes: prin cipal component matrices showing pattern metric factor loadings......................................................................................................186 3-47 The 400-m spatial extent for lakes: prin cipal component matrices showing pattern metric factor loadings......................................................................................................187 3-48 The watershed spatial extent for lake s: principal component matrices showing pattern metric factor loadings..........................................................................................187 3-49 Coefficients of determination, probabili ties, and regression equations for multiple regressions between indicators of ecos ystem condition and significant components resulting from the PCA of landscape pattern metrics at four grain sizes for the sample lakes................................................................................................................... ..188 3-50 Coefficients of determination, probabili ties, and regression equations for multiple regressions between indicators of ecos ystem condition and significant components resulting from the PCA of landscape pattern metrics at three spatial extents for the sample lakes................................................................................................................... ..190 12

PAGE 13

3-51 Multiple regression models at four grain sizes for the sample isolated forested wetlands: coefficients of determination, pr obabilities, and change in the amount of variability ( R2) in indicators of ecosystem condition....................................................191 3-52 Multiple regression models at four grain sizes for the sample streams: coefficients of determination, probabilities, and cha nge in the amount of variability ( R2) in indicators of ecosystem condition....................................................................................193 3-53 Multiple regression models at three spatial extents for the sample streams: coefficients of determinati on, probabilities, and change in the amount of variability ( R2) in indicators of ecosystem condition.....................................................................195 3-54 Multiple regression models at four grain sizes for the sample lakes: coefficients of determination, probabilities, and cha nge in the amount of variability ( R2) in indicators of ecosystem condition....................................................................................197 3-55 Multiple regression models at three spatial extents for the sample lakes: coefficients of determination, probabi lities, and change in th e amount of variability ( R2) in indicators of ecosystem condition....................................................................................198 A-1 Primary symbols of the energy circuit diagramming.......................................................263 B-1 FLUCCS categories and correspond ing land use intensity classes.................................264 C-1 Non-renewable and purchased empower de nsity for urban land uses according to M. T. Brown..........................................................................................................................269 C-2 Non-renewable and purchased empower de nsity for urban land uses according to N. Parker...............................................................................................................................269 C-3 Non-renewable and purchased empower de nsity for agricultura l land uses according to S. Brandt-Williams......................................................................................................270 D-1 Water chemistry variables considered fo r 75 sample isolated forested wetlands............271 D-2 Water chemistry variables consid ered for 47 STORET stream station...........................272 D-3 Water chemistry variables consid ered for 54 STORET lake stations..............................273 E-1 Metric composition of the WCI incl uding diatoms WCI, macrophytes WCI, and macroinvertebrates WCI..................................................................................................275 E-2 WCI scores for 118 wetlands based on three assemblages including diatoms, macrophytes, and macroinvertebrates..............................................................................275 F-1 Macroinvertebrate metric composition of the SCI defined by Barbour and colleagues..277 F-2 Macroinvertebrate metric compos ition of the SCI defined by S. Fore............................277 13

PAGE 14

F-3 SCI scores for 69 streams for the macroinvertebrate assemblage...................................278 G-1 Macroinvertebrate metric composition of the LCI..........................................................279 G-2 LCI scores for 54 lakes for th e macroinvertebrate assemblage.......................................279 I-1 Characteristics of the land use/land cove r surrounding the isolated forested wetlands...283 J-1 Characteristics of the land use/la nd cover surrounding the sample streams....................286 K-1 Characteristics of the land use/land cover surrounding the sample lakes........................288 L-1 LDI scores calculated for eight different grain sizes (units: meters on a side) and based on the area occupied by each la nd use type in the landscape unit.........................290 L-2 LDI scores calculated for eight different spatial resolutions (units in meters) and assuming that the effect of development in tensity on the landscape decreases linearly with distance....................................................................................................................293 L-3 LDI scores calculated for eight different spatial resolutions (units in meters) and assuming that the effect of development intensity on the landscape decreases in inverse-square with distance............................................................................................296 M-1 LDI scores calculated for six different gr ain sizes (units: meters on a side) and based on the area occupied by each land use type in the drainage basin unit............................299 M-2 LDI scores calculated for six different grain sizes (units: meters on a side) and assuming that the effect of development in tensity on the landscape decreases linearly with distance....................................................................................................................301 M-3 LDI scores calculated for six different grain sizes (units: meters on a side) and assuming that the effect of development intensity on the landscape decreases in inverse square with distance............................................................................................303 N-1 Isolated forested wetlands (n = 51) : summary statistics and transformation information for landscape pattern metrics calculated at different grain sizes.................305 N-2 Streams (n = 68): summary statistics and transformation information for landscape pattern metrics calculated at four different grain sizes....................................................307 N-3 Streams (n = 63): summary statistics and transformation information for landscape pattern metrics calculated at th ree different spatial extents.............................................309 N-4 Lakes (n = 48): summary statistics an d transformation information for landscape pattern metrics calculated at four different grain sizes....................................................311 N-5 Lakes (n =44): summary statistics an d transformation information for landscape pattern metrics calculated at different spatial extents......................................................313 14

PAGE 15

O-1 Multiple regression models at four grain sizes for the sample isolated forested wetlands: coefficients of determination, pr obabilities, and change in the amount of variability ( R2) in indicators of ecosystems condition..................................................315 O-2 Multiple regression models at four grain sizes for the sample streams: coefficients of determination, probabilities, and cha nge in the amount of variability ( R2) in indicators of ecosystems condition..................................................................................317 O-3 Multiple regression models at three spatial extents for the sample streams: coefficients of determinati on, probabilities, and change in the amount of variability ( R2) in indicators of ecosystems condition....................................................................319 O-4 Multiple regression models at four grain sizes for the sample lakes: coefficients of determination, probabilities, and cha nge in the amount of variability ( R2) in indicators of ecosystems condition..................................................................................320 O-5 Multiple regression models at three spatial extents for the sample lakes: coefficients of determination, probabi lities, and change in th e amount of variability ( R2) in indicators of ecosystems condition..................................................................................322 15

PAGE 16

LIST OF FIGURES Figure page 1-1 Systems diagram of the impacts of ag ricultural lands on freshwater ecosystems.............51 1-2 Systems diagram of a Florida cypress dome......................................................................52 1-3 Systems diagram of a stream section in Florida................................................................52 1-4 Systems diagram of a Florida lake.....................................................................................53 2-1 Study site location by ecoreg ions of 118 isolated forest ed wetlands in Florida................94 2-2 Study site location by ecoregions of 69 streams in Florida...............................................95 2-3 Study site location of 54 lakes in Florida...........................................................................96 2-4 Major land use patches delineated from an aerial photo within a 200 meter buffer area surrounding a study wetland......................................................................................97 2-5 Landscape surrounding an isolated forested wetland in Florida........................................98 2-6 Land use types for a Florida stream watershed..................................................................99 2-7 Spatial scales of analysis for isolated forested wetlands.................................................100 2-8 Spatial scales of analysis for streams...............................................................................101 2-9 Flow chart showing the main steps followe d for the delineation of drainage basins......101 2-10 Graphic representation of the steps followe d in the delineation of the area draining to a water quality sampling site............................................................................................102 2-11 Spatial scales of analysis for lakes...................................................................................103 3-1 Mean LDI scores for a subsampl e of isolated forested wetlands.....................................199 3-2 Scalograms showing the effect of changi ng the grain size on the LDI for six isolated forested wetlands.............................................................................................................200 3-3 Landscapes surrounding a subsample of st udy isolated forested wetlands shown at three different grain sizes.................................................................................................201 3-4 Mean LDI scores for a 15 stream drainage basins...........................................................204 3-5 Scalograms showing the effect of changi ng the grain size on the LDI for six streams...205 16

PAGE 17

3-6 Comparison among LDI values calculated for isolated forested wetlands for three extents..............................................................................................................................206 3-7 Matrix plot of the rela tionship between pairs of LDI scores calculated for three landscape extents for the is olated forested wetlands.......................................................207 3-8 Comparison between LDI scores calcula ted for buffer areas of 100 and 400 meters from streams and for the entire drainage basin................................................................208 3-9 Matrix plot of the rela tionship between pairs of LDI scores calculated for three landscape extents for streams...........................................................................................209 3-10 Comparison among LDI scores calculated for buffer areas of 100 and 400 meters from lakes and for the entire drainage basin....................................................................210 3-11 Matrix plot of the rela tionship between pairs of LDI scores calculated for three landscape extents for lakes...............................................................................................211 3-12 Variability in DO for the isolated forested wetlands explained by the LDI calculated at the 200-meter buffer.....................................................................................................212 3-13 Variability in SC for the isolated forest ed wetlands explained by the LDI calculated at the 200-meter buffer.....................................................................................................212 3-14 Variability in TP for the isolated forest ed wetlands explained by the LDI calculated for a spatial extent of 20 meters.......................................................................................213 3-15 Regression results at seve ral spatial scales showing how much of the variability (measured in r2 values) in the water chemistry va riables for the isolated forested wetlands was explained by the LDI.................................................................................214 3-16 Variability in the macrophyte WCI explai ned by the LDI calculated for a buffer of 20-meters surrounding the sample isolated forested wetlands........................................215 3-17 Variability in the macroinvertebrate WCI explained by the LDI calculated for a buffer of 20-meters surrounding the isolated forested wetlands......................................215 3-18 Variability in the diatom WCI explai ned by the LDI calculated for a buffer of 100meters surrounding the isolated forested wetlands..........................................................216 3-19 Regression results at seve ral landscape grains showing how much of the variability (measured in r2 values) in the WCI for the isolat ed forested wetlands was explained by the LDI........................................................................................................................217 3-20 Variability in the macrophyte WCI explaine d by the LDI calculated at a grain size of 5 x 5 meters................................................................................................................... ...218 17

PAGE 18

3-21 Variability in the macroinvertebrate WCI explained by th e LDI calculated at a grain size of 50 x 50 meters......................................................................................................218 3-22 Variability in the diatom WCI explained by the LDI calculated at a grain size of 30 x 30 meters...................................................................................................................... ....219 3-23 Variability in the concentration of DO for the sample streams explained by the LDI calculated for the watershed scale....................................................................................219 3-24 Variability in NO3-N for the sample streams explained by the LDI calculated for the watershed scale................................................................................................................ 220 3-25 Variability in TN for the sample stre ams explained by the LDI calculated for the watershed scale................................................................................................................ 220 3-26 Variability in the WQI scores for the sa mple streams explained by the LDI calculated for the watershed scale.....................................................................................................221 3-27 Regression results at seve ral spatial scales showing how much of the variability (measured in r2 values) in the water chemistry variables and the WQI for streams was explained by the LDI................................................................................................222 3-28 Variability in the concentration of DO for the sample streams explained by the LDI calculated at a grain size of 170 x170 meters..................................................................223 3-29 Variability in the concentration of NO3-N for the sample streams explained by the LDI calculated at a grain size of 170 x 170 meters..........................................................223 3-30 Variability in TN for streams explained by the LDI calculated at a grain size of 170 x 170 meters..................................................................................................................... ...224 3-31 Variability in the WQI scores for stream s explained by the LDI calculated at a grain size of 170 x 170 meters..................................................................................................224 3-32 Variability in the SCI_1 for streams e xplained by the LDI calculated for the 100meter buffer......................................................................................................................225 3-33 Variability in the SCI_2 for streams e xplained by the LDI calculated for the 100meter buffer......................................................................................................................225 3-34 Regression results at seve ral spatial scales showing how much of the variability (measured in r2 values) in the SCI for str eams was explained by the LDI......................226 3-35 Variability in the SCI_1 for streams explai ned by the LDI calculated at a grain size of 20 x 20 meters.............................................................................................................. 226 3-36 Variability in the SCI_2 for the stream s explained by the LDI calculated at a grain size of 20 x 20 meters......................................................................................................227 18

PAGE 19

LIST OF ACRONYMS ANOVA Analysis of variance BASINS Better Assessment Science In tegrating Point and Nonpoint Sources CWA Clean Water Act DD Decimal degrees DEM Digital elevation model DO Dissolved oxygen (units: mg/L) DOQQ Digital Orthogra phic Quarter Quad DTM Digital terrain model ED Edge density (units: meters/hectare) FDEP Florida Department of Environmental Protection FDOT Florida Department of Transportation FLUCCS Florida Land Use and Cove r and Forms Classification System FWCI Florida Wetland Condition Index (unit less) GIS Geographic Information System HUC Hydrologic Unit Code IBI Index of Biological Integrity IJI Interspersion and Juxtap osition Index (units: %) LCI Lake Condition Index (unit less) LDI Landscape Development Intensity Index (unit less) LDI-ILD LDI-inverse linear distance (unit less) LDI-ISD LDI-inverse square distance (unit less) LDI-PLU LDI-proportion of land use (unit less) LU/LC Land use/land cover NAPP National Aerial Photography Program NED National Elevation Dataset NO2/NO3 nitrate/nitrite-nitrogen (units: mg/L) NTU Nephelometric turbidity units PCA Principal Components Analysis PD Patch density (units: # of patches /100 hectares) PR Patch richness (unit less) PRD Patch richness density (units: # /100 hectares) SC Specific conductance (units: mhos/cm) SCI Stream Condition Index (unit less) SHDI Shannon Diversity Index (unit less) SHEI Shannon Evenness Index (unit less) TKN Total Kjeldahl nitrogen (units: mg/L) TN Total nitrogen (units: mg/L) TP Total phosphorus (units: mg/L) USDA U.S. Department of Agriculture USGS U.S. Geological Survey USEPA U.S. Environmental Protection Agency VIF Variance Inflation Factor WCI Wetland Condition Index (unit less) WMD Water Management District WQI Water Quality Index (unit less) 19

PAGE 20

Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy DEVELOPMENT OF AN INDEX OF LANDS CAPE DEVELOPMENT INTENSITY FOR PREDICTING THE ECOLOGICAL CONDITION OF AQUATIC AND SMALL ISOLATED PALUSTRINE WETLAND SYSTEMS IN FLORIDA By Manuel Benjamin Vivas August 2007 Chair: Mark T. Brown Major Department: Environmental Engineering Sciences Freshwater ecosystems, while vital compone nts of landscapes and essential for human well-being, are increasingly thr eatened by the ever-escalating intensity of human development. In this study, the main objective was to analyze the influence of human development intensity on the ecological condition and water quality of isol ated forested wetlands, streams, and lakes in Florida. An index of Landscape Development In tensity (LDI), derived from the non-renewable areal empower density of land use, was used as a measure of the human disturbance gradient against which the ecological condition and water quality of 118 isolated palustrine forested wetlands, 69 streams, and 54 lakes were analy zed at different lands cape scales. Landscape pattern metrics were also calculated for study ec osystems and tested for relationships with indicators of ecosystem condition and water quality. Overall, the LDI had the greatest predictive ability for bioindicators of ecological condition in wetlands and streams, explai ning up to 30% and 27% of variab ility, respectively. The LDI was a significant factor in explaini ng the variability of water quality variables only for streams. Changes in landscape scale (grain and extent) had small effects on the LDI. Differences in LDI scores were noticeable when developed lands were added w ith increasing area. The use of 20

PAGE 21

distance-weighting functions provi ded little enhancement of the predictive power of the LDI; distance-weighted LDIs did incr ease by 7% the predictive power for bioindicators for streams. Landscape pattern metrics explained up to 44% of variability in bioindicators of wetland condition, and 22% and 42% for stream and lakes. They also accounted for up to 60% and 39% in the variance of water quality variables for streams and lakes, respectively. When included with the LDI in multiple regressions they increase d by 25% the amount of variance explained in bioindicators of wetland condition and 52% for lakes. In general, the LDI had higher predictive pow er for bioindicators of ecosystem condition than for chemical constituents of the ecosystems studied, which are more variable with season, time of day, and hydrologic conditions. The LDI may have greater correlation with bioindicators because they may be more integrative of anthro pogenic impacts and have higher correlation with ecological condition. 21

PAGE 22

CHAPTER 1 INTRODUCTION Freshwater ecosystems are essential component s of many landscapes. They are also home to myriad species and are vital to humans for the countless products and services that they provide. Despite their value, freshwater system s are increasingly modified by peoples use of resources, causing changes in their ecological condition and threateni ng societal well-being. Thus, investigating how human behavior in th e landscape affects freshwater ecosystems is a matter of critical importance (Naiman et al. 1995; Baron et al. 2003; Allan 2004). To this end, this research was designed to develop a quantit ative understanding of the influence of human development intensity on the ecologi cal condition of streams, lakes, and small isolated palustrine forested wetlands in Florida. Statement of the Problem The effect of land use/land cover changes on freshwater ecosystems has been recognized for some time (Likens et al. 1970; Haynes 1975; Omernik, 1977; Karr and Schlosser 1978; Peterjohn and Correl 1984; Allan and Flecker 1993). Human-related activities such as deforestation, silviculture, agricultural intensification, urbani zation, and drainage of flooded areas may all negatively affect freshwater systems (Carpenter et al. 1996; Carpenter et al. 1998; Giller and Malmqvist 1998). The problem may be exacerbated as human population increases, the remaining natural lands are converted to other uses, and human activities in already transformed lands are further intensifie d to respond to development needs. Population growth and land development are the main forces driving changes in Floridas landscapes (Reynolds 1999). Currentl y, Florida is the fourth most -populated state in the United States, with more than 17 million people; it is projected that by 2030 th e states population will total 28.7 million (U.S. Census Bureau 2005). By 1997 the rate of rural land loss in the state was 22

PAGE 23

about 60,000 hectares per year, and it is proj ected that from 2000 to 2020 an additional 53,000 hectares per year will be converted from ru ral to urban uses (Re ynolds 1999). These trends suggest that Floridas freshwater systems will be facing increased human pressures and the need to better understand how the states streams, lakes and isolated forested wetlands may be affected by land development. Wetlands, streams, and lakes integrate all ener gy flows occurring in th eir drainage basins, with water acting as the main element for materials transport and defining the cumulative impacts of human activities in the lands surr ounding them (Naiman et al. 1995; Wear et al. 1998). Energy inputs into freshwater systems from the surrounding landscape that are different from their more natural flows may result in gradua l differences in internal structure, processes, and eventually system organization. Thus, fres hwater systems may experience water quality modifications and changes in aquatic ecological communities as energy inputs are altered due to land development. The assessment of how human landscape-level activities affect aquatic systems requires the use of indicators that describe landscape attributes (ONeil et al. 1997; Gergel et al. 2002). Land use intensity is an attribute of landscapes that might be relate d to the condition of freshwater systems, since as the intensity of human development of landscapes increases, the greater the potential for ecological degrad ation (Brown 2003b; Brown and Vivas 2005). Land use intensity can be quantified based on the amount of energy used by humans in their development activities. Emergy, the energy that was used to make a product or service and expressed in units of one type of energy (Odum 1996), allows quantifying natural and economic flows in meaningful common units and developing metrics that describe land use intensity which 23

PAGE 24

can then be related to water quality and biotic variables. This in turn may allow for new insights in to the assessment of the condition of Floridas aquatic systems. Different indices or metrics of landscape patt ern, which describe la ndscape heterogeneity in terms of its composition (presence, rela tive abundance and diversity of patches) and configuration (spatial distributi on of patches) have also been proposed (ONeil et al. 1997; Cifaldi et al. 2004) and to some extent tested (J ohnson et al. 1997; Griffith et al. 2002) to analyze the relationship between human uses of landscap es and the condition of freshwater systems. Despite the lack of conclusive ev idence that would suggest a set of metrics that can best be used to describe this relationship, patt ern metrics are valuable landscape indicators that can help in understanding how humans impact freshw ater systems (Turner et al. 2001). Human influences on freshwater systems occur at multiple spatial scales ranging from the local scale, where for example the removal of riparian forest s may result in increased local sediment inputs to water bodies, altered water h eat flows due to shade removal, changes in woody-debris inputs, and shifts in the compos ition of aquatic biological communities (Jones et al. 1999; Naiman and Dcamps 1997; Poole and Berman 2001); to the watershed scale, where land use regulates hydrological budgets and differen ces in the amount of sediments and nutrients transported by surface runoff (Soranno et al. 1996; Allan 2004). Levin (1992) has argued that the description of the variab ility and predictability of the envi ronment is only meaningful when the multiple scales that may be of importance to an organism or process are considered. Landscape scale is described in terms of its grain (i.e., spatial resolution) and extent (i.e., size of the study area) (Wiens 1989; Turner et al. 2001). Multiple studies (Richards et al. 1996; Allan et al. 1997; Gergel et al. 1999) have documented the importan ce of scale in determin ing the influence of landscape attributes on the ecological condition of freshwater syst ems. However, there are mixed 24

PAGE 25

conclusions about the scales at which land use most affects freshwater systems, suggesting that this matter still requires further investigation (Turner et al 2001). The influence of human activ ities on freshwater ecosystems may also depend on the distance between landscape features. For exam ple, material transport originating from agricultural lands can be attenu ated with increasing distance fr om freshwater systems (Soranno et al. 1996) and may be significan tly reduced by the pres ence of riparian forests in drainage basins where developed lands are common (Co rrell et al. 1992; Naiman and Dcamps 1997). Since land use proximity is believed to be impor tant in how human activities affect freshwater systems, emphasizing the value of nearby lands over that of more distant lands in land-water interaction studies may prove to be valuable (O Neil et al. 1997). Severa l studies of land-water interface (Comeleo et al. 1996; King et al. 2005) ha ve considered distance-weighting of land use; however, the benefits of this approach as well as the best way of weighting land use remains unclear (King et al. 2005). This dissertation explores how landscape deve lopment intensity might affect the condition of isolated forested wetlands, streams, and la kes in Florida. It investigates properties of indicators of landscape development intensity with changes in landscape scale. It statistically relates landscape development intensity and land scape pattern with indicators of ecosystem conditions and water quality at different spatial scales, and c onsidering distance-weighting of land use. The question of how landscape patt ern indicators may complement landscape development intensity indicators in their abil ity to predict ecological condition is also considered. Plan of Study How human activities affect aquatic ecosystems is a central aspect of land-water interface studies. The assessment of this re lationship requires measurable indicators that describe human 25

PAGE 26

behavior in the landscape and that will allow an understanding of how it contributes to the ecological impoverishment of fres hwater ecosystems. This investigation approaches this need through a landscape analysis that uses concepts and tools of environmental accounting using emergy. Accordingly, the three main objective of this work are the following: Describe spatial properties of indicators of landscape developmen t intensity with changes in landscape scale (i.e., grain and extent) and distance. Determine statistically the ability of landscape development intensity indicators to predict the ecological condition of streams, lakes, and sma ll isolated forested wetlands at different landscape scales and considering land use distance. Develop statistical models to predict the ecological condition and water quality of streams, lakes, and small isolated forested wetlands from landscape pattern and determine the spatial scale at which relationships are most relevant. Assess how landscape pattern can complement th e ability of landscape development intensity to predict the ecological conditi on of streams, lakes, and small isolated forested wetlands. To accomplish the first objective emergy analys is enabled the evaluation of the spatial non-renewable energy characteristics (non-renewa ble areal empower dens ity) of different land uses in Florida. These in turn allowed quan tifying the non-renewable ar eal empower density of the drainage basins or hydrol ogical contribution areas of the systems under investigation, which were determined using Geographic Information Systems (GIS) methods. Finally, an index of Landscape Development Intensity (LDI) was calc ulated for each drainage basin. The LDI was calculated at various landscape grains and exte nts to assess landscape-scale effects on the index. The landscape grain was varied systematically considering the spatial characteristics of the original land use data used for each system studied. The landscape extent over which the LDI was calculated was varied from artificially defi ned hydrological contributi ng areas of different sizes to entire drainage basins. Two distance-weighting algorithms were used to test the effect of distance on the LDI. 26

PAGE 27

The second and third objectives were achieve d by statistically relating the landscape indicators to indicators of eco logical condition and wate r quality variables for isolated forested wetlands, lakes, and streams. The LDI was used as a measure of the human disturbance gradient against which the ecological condition and water quality of freshwater systems was analyzed using simple linear regression. The spatial dist ribution and patterns of human development in drainage basins were determined following concep ts and guidelines from the field of landscape ecology and within the GIS environment. Mul tiple regression analys is allowed developing statistical models to investigat e how much of the variation in indicators of ecosystem condition and water quality variables was explained by the landscape pattern metrics. Relationships were explored considering various grain sizes and spatial extents to assess scale effects on the predictive power of both types of landscape indices. For the fourth objective pattern metrics were used as additional independent variables in multiple regression analysis to explore how much of the remaining vari ance initially accounted for by the LDI could be explained when the tw o types of landscape indicators were used together. A discussion of the value of landscape indicators for predicting ecological condition of freshwater systems fallowed. The results from this dissertation are intended to be a contribution to the development and testing of methods fo r the assessment and monitoring of freshwater systems in Florida. Concepts and Approaches to Understanding Human Impacts on Freshwater Systems Emergy and Empower Density Emergy is a measure of the available energy th at was used to make a product or process and that has been corrected for different qualities. The solar emergy of a product is the emergy of the product expressed in equivalent solar energy that is required to generate it (Odum 1996). The 27

PAGE 28

units of emergy are emjoules (for emergy joules) and the units of solar em ergy are solar emjoules (abbreviated sej). A flow of emergy is empower (measured in sola r emjoule per time); when it is applied in a unit area it is referred to as em power density (areal empower dens ity) and can be interpreted as a measure of work per area per time (units: sej/ ha-yr) (Odum 1996). An area with high energy use, such as a city, will have a higher areal empower density than areas using less energy, such as rural areas. Emergy flows are organized hierarchically into spatial patterns with emergy flows per area more concentrated in hierarchical centers such as cities (Brown 1980; Odum 1996). Based on this observation, Brown and Vivas (2005) suggested that the impacts of human activities might be related spatially to the intensity of energy us e and that the areal empow er density might serve as a measure of the level of human-induced impacts on ecological systems. Consider, for example, a land-use practice such as agriculture (Figure 1-1); it is highly probable that as agriculture is intensified by the increased use of fertilizers, pesticid es, heavy machinery, and fossil fuels, among other non-renewable inputs as shown in the right side of Figure 1-1, the chances of degrading nearby freshwater systems (on the left in Figure 1-1) with higher loads of nutrients and sediments in runoff water will be higher over time. As the use of the land intensifies, there is greater potential for envir onmental impacts. These impacts can be assessed spatially using the empower densities of land use as a measure of the degree of the impacts of human activities. When multiple land uses or land scapes with different levels of development intensity are considered, descri bed, and organized based on their empower densities, the areal empower density can be use as a metric that describes the human disturbance gradient. 28

PAGE 29

Human Disturbance Gradient Environmental variation is ordered in space and spatial environmental patterns determine the structure and function of ecological systems. Like natural environm ental gradients, land development gradients provide a means to ex plain spatially or predict human impacts on ecological systems (McDonnell and Pickett 1990). Since humans have influenced, altered, or transformed almost every ecosystem around the globe, with most landscapes today falling between a gradient of complete ly natural to highly developed areas (Brown and Vivas 2005), the human-induced impacts on the landscape can be organized along a di sturbance gradient depending on the intensity of the human ac tivities within it (McDonnell and Pickett 1990; McMahon and Cuffney 2000; Brown and Vivas 2005) The human disturbance gradient, which can be described and quantified using landscape emergy-based indices, was used to assess the impact on freshwater ecosystems that result fr om human activities within their watersheds. Landscape Pattern Landscape pattern refers to the spatial com position and configuration of the elements present in a landscape. Landscape pattern results from the interaction of multiple abiotic and biotic factors and the way in which humans use the land (Turner et al. 2001). The quantification of landscape pattern is a necessary step in order to describe the interactions between spatial pattern and ecological processes (Gustafson 1998; Turner et al. 2001; Turner 2005). The relationship between spatial patter n and ecological processes is th e fundamental idea that defines the science of landscape ecology (Turner et al. 2001). Many landscape measures or metrics have been developed that desc ribe how landscapes are structured (ONeill et al. 1988; McGariga l and Marks 1995; ONeill et al. 1999) and constitute a direct way for establishing the re lationship between pattern and process (Gustafson 1998; Turner et al. 2001). Landscape metrics pr ovide information about the composition and 29

PAGE 30

configuration of landscapes (McGarigal and Marks 1995). Composition metrics provide descriptions of the presence, relative abundance, and diversity of patches in the landscape, while configuration metrics represent the spatial distribution of patches in the landscape. The concepts and analytical tools used in landscape ecology are also applicable to the study of freshwater systems (Wiens 2002). Thus metrics of landscape pattern allow quantifying anthropogenic disturbance and provide a means to establishing links between human actions in watersheds and their impact on wetlands, streams, and lakes. Importance of Scale The review of the recent literature on land-water interactions reveals that human impacts on freshwater systems occur at multiple spatial scales. It is also clear from this review that determining the scale(s) that are appropriate for establishing predictive relationships between human actions at the landscape level and thei r effects on freshwater systems is an ongoing challenge. In this dissertation the relationship be tween the intensity and pattern of development and the ecological condition of fr eshwater systems was assessed at several scales in order to determine the scales at which such relationships are best observed. Scale has been defined as the spatial or tempor al dimensions of an object and is described by the grain and extent (Turner et al. 2001). Grain refers to the finest spatial resolution of observation of a phenomenon, while extent is the total area of st udy (Wiens 1989; Turner et al. 2001). Wiens (1989) clarifies that, defined this wa y, the grain of an investigation is different from the way MacArthur and Levins (1964) cons idered grain as being a function of how animals exploit resource patchiness in environm ents (page 387). Throug hout this study grain and extent will be used following the definitio ns by Wiens (1989) and Turner et al. (2001). 30

PAGE 31

Watershed as the Unit of Analysis A watershed can be defined as the area of la nd draining to a specifi c point or area on a stream, lake, or wetland1. Watersheds are usually considered to be defined landscape units since their boundaries can be relatively easily defi ned based on topographi c and hydrologic factors (Hunsaker and Levine 1995; NRC 1999). This charac teristic, and their inte grative properties of environmental process including human impacts, makes watersheds functional systems for research in many different scientific fiel ds including geomorphology, hydrology, stream ecology, landscape ecology, and ecosystem management (USEPA 1996; Allan et al. 1997; NRC 1999; Turner et al. 2001). Floridas surface waters have been divided in to 52 major basins or hydrologic units based on the countrys nationwide surface hydrologic f eatures system, which was developed by the United States Geological Survey. The system is hierarchical and allows th e definition of more detailed hydrologic units. As a result, Floridas surface waters have been further subdivided into about 3,830 sub-basins or waters heds. This watershed subdivision has allowed the state to manage its water resources on the basis of hydrologi c units despite the comp lexity of the states surface and subsurface water flows. Through th e development of surface and groundwater monitoring programs and the assessme nt of the impacts of point a nd non-point source discharges on surface waters within Floridas watersheds, the state fulfills the requirements that were established under Sections 305(b) and 303(d) of the Federal Clean Water Act, and accomplishes the goals of the 1999 Florida Watershed Restoration Act (FDEP 2004). 1 Allan (1995) mentions that in American us age the terms drainage basin, catchment, and watershed are synonymous, even though catchment s may be used to refer to small basins. Accordingly, these three terms were used in terchangeably throughout this dissertation. 31

PAGE 32

Emergy has often been used to evaluate the wo rk of nature and the work of humans in generating products and services in watershed syst ems, to describe and assess their patterns of development, and to propose management alte rnatives (Odum 1996). Turner et al. (2001) has pointed out that the study of watersheds overlaps with that of landscape ecology, either when different watersheds are compared or when th e internal characteristics of a watershed are explored. Accordingly, concepts and tools that are commonly us ed in environmental accounting using emergy and landscape ecology were used in th is study in the assessment of the relationship of patterns of watershed development and their ef fect on isolated forested wetlands, streams, and lakes in Florida. Previous Studies Although concerns over the effect of land use/land cover on freshwater systems has existed for some time only recently, with the development of GIS tec hnology, remotely sensed imagery, and advances in landscape ecology, has there been a greater attemp t to quantify these relationships. In this section, a re view of the literature related to the main research objectives of this dissertation is presented. Emergy studies of watersheds, landscapes, and land use are summarized herein. Some of the concepts and results from these studies are the basis for the calculation of and index of Landscape Developm ent Intensity (LDI) whose development and applications are also considered. Finally, studi es that relate landscape pattern metrics to ecosystems condition at different scales and for different freshwater systems are reviewed. Emergy Studies Emergy accounting has enabled relating econom ic development with environmental change for a great variety of products and pro cesses around the world. Most of this work is summarized in Odum (1996), and more recently was published in a series of folios by the H. T. Odum Center for Environmental Policy of the University of Florida (Odum 2000; Odum et al. 32

PAGE 33

2000a; Brown and Bardi 2001; Brandt-Williams 2001; Kangas 2002), and in the proceedings of the biennial Emergy Synthesis Research Conf erences initiated in 1999 (Brown 2000; Brown 2003a; Brown 2005). The scientific basis of the emergy methodology is desc ribed in great detail in Odum (1994). Within this work, studies on watersheds, landscapes, and land use are of particular interest for this research. Watersheds studies using emergy The study of watersheds using emergy began mo re than 20 years ago. These studies have described properties of watersheds, their patterns of development, and have been used as the basis for management purposes. Diamond (1984) and Odum et al. (1987) evaluated the properties of stream orders in the Mississipp i River Basin based on their environmental and economic empower. These studies revealed that the geopotential energy fluxes were greatest at intermediate to high order leve ls while the delta and floodplains presented the greatest emergy since different energies (i.e., watershed, coast, and trade) converged in these areas. Romitelli (1997) evaluated the work done by water energies in six watersheds in southeastern Brazil and in the Coweeta basin in North Carolina, United Stat es. The chemical potential and geopotential energies of water were found to be coupled ; the geopotential energy, maximized at middle elevations, allowed the dispersal of nutrients and sediments in the floodplains. Transformities2 of chemical potential energy were found to increase downstream. Research aimed at analyzing the patterns of development in watersheds include the study by Odum et al. (1986) in the Amazon Basin, where economic development and ecological organization were found to be hi erarchical and organized to ma ximize available energies from within and without the system Howington (1999) looked at th e spatial organization of the 2 Transformity is defined as the emergy of one type required to make a unit of energy of another type (Odum 1996, page 289). 33

PAGE 34

Catatumbo River basin from an emergy perspect ive. Spatial emergy patterns were used to describe how resources were used and the patterns of development in this basin which is shared between two countries, Colombia and Venezuela. Total renewable empower density increases downstream while nonrenewable empower densit y increased upstream. Since the middle to upper basin was more developed, the author suggeste d that growth in this area should be done in such a way that its impact on water quality is minimized, thus improving the ability of downstream wetlands to buffer development imp acts. Studying the Cach e River watershed in northeastern Arkansas, United States, Odum et al. (1998b) found that environmental contributions within the system accounted for about half of the watersheds wealth (measured in emergy units) while the other half was from i nputs purchased outside the system. The Cache River watershed, mostly an agricu ltural area based on indigenous so ils and waters, proved to be a net emergy exporter. In Florida, Brandt-Williams (1999) used measures of materials, energy and emergy, to study the links between two lakes and their respective waters heds. Lakes were found to be areas of high emergy concentration due to the flow of energy a nd materials from their watersheds. Simulation models determined that the influence of the watersheds on the lakes increased with higher use of nonrenewable inputs within the basin. Also in Florida, Parker (1998) found that the empower density increase d downstream in the St. Marks watershed and was highest in the central region of the basin, wh ere nonrenewable inputs such as electricity and fuel dominated. Intermediate empower densities characterized the lower end of the basin where the landscape changes from urbanto rural-dominated land uses. Tilley (1999) estimated the benefits provided by three forested watersheds in the Southern Appalachians, United States, based on the emergy required to develop and maintain services and products. The benefits of several combinations of economic investment in recreation and 34

PAGE 35

timbering were evaluated using empower. Max imum empower (optimum intensity of forest development) was found at intermedia te levels of economic investment. Landscapes and land use Studies aimed at evaluating landscapes us ing emergy are originally found in Brown (1980), which studied the relations hip between energy flows and the hierarchical organization of urban and regional systems in Florida to test th e hypothesis of energy control of landscapes. The results suggested that landscap es are organized hierarchi cally based on their quality measurements obtained using embodied energy (e mergy). In this study, the power density and volume of structure associated w ith 11 urban land uses in Florid a were also calculated. Power density, as well as volume of structure, showed a strong correlation with increasing complexity of land uses. Whitfield (1994) used emergy to analyze the organization of urban systems based on land use patterns in Jacksonville, Florida. There was a high ratio of purchased resources to renewable resources indicating the high depend ency of the city on outside re sources. The land use pattern of Jacksonville largely determined the type and am ount of resources used. Alternative land use patterns suggested by the author were found to reduce resource use and to reduce the impact on the contribution of natural systems. In this study emergy evaluations for 13 different urban land uses and three agricultural land uses were developed. Lambert (1999) created a spatial model of the distribution of energy flows and storages in Alachua County, Florida, and used it to anal yze spatial patterns of energy transformation hierarchy in relation to spatial patterns of urba n landscapes. Maps of transformities showed that areas with low transformities are more dispersed and organize d surrounding centrally-located areas with higher transformities. The results s uggested that urban landscapes tend to develop spatial patterns that can be described in te rms of an energy tran sformation hierarchy. 35

PAGE 36

Brandt-Williams (1999) developed emergy evaluati ons for different land uses within lake watersheds in Florida. Key emergy-based ratios re levant to agriculture (transformity, emergy per mass, and empower density) were calculated. Bo ggess (1994) related the sp atial distribution of empower to phosphorus runoff from di fferent land uses in an agri cultural watershed of the north Okeechobee basin in Florida. Maximum phosphorus loads occurred at intermediate empower densities (agricultural lands uses) while em power increased downstream. Other production processes of interest that we re evaluated using emergy incl ude natural systems (Orell 1998; Bardi and Brown 2001), forestry (Christianson 1984; Doherty 1995; Odum et al. 2000b), mining (Kangas 1983; Odum 1996; Kangas 2002), aquacu lture (Odum and Arding 1991; Brown et al. 1992; Ortega et al. 2000) and transporta tion (Odum and Odum 1987; McGrane 1994). Measures of the intensity of landscape development Brown and Vivas (2005) calculated an index of landscape development intensity (LDI) using land use data and measures of the intens ity of development derived from energy use per unit area. The LDI was developed to estimate the potential impacts from human-dominated activities on ecological systems within watersheds of different sizes in Florida. The LDI was proposed as a measure of the human disturbance gradient. Previously, Parker (1998) used preliminar y versions of the LDI based on physical and emergy measurements to correlate them to model results from a spatial pollutant model for total phosphorus (TP) for subbasins of the St. Marks Watershed in Northern Florida. The LDIs showed a good association with the TP loads above background levels, particularly an imperviousness LDI and the empower density LDIs. Th is study showed that despite the fact that predicting TP loads at low development intensit ies is difficult; at higher levels of human development the LDI in its various forms may be a good predictor of nutrients accumulation that 36

PAGE 37

can result from more intense human activities. Based on these re search findings, Parker (1998) suggested that development intensity indices could be useful for policy purposes. Cohen et al. (2004) used a preliminary vers ion of the LDI calculated by Brown and Vivas (2005) as a measure against which a floristi c quality assessment i ndex (FQAI) could be compared and to provide evidence of its importa nce in the assessment of the ecological condition of small isolated herbaceous we tland systems. The FQAI is a nu merically-based expert opinion index used to associate plant species to a partic ular habitat or to determine their tolerance to varying disturbance intensity. Strong associa tions between the LDI and the FQAI provided evidence of the relevance of the floristic index for biological assessment studies and of the LDI as a measure of the human disturbance gradient. Using the LDI, Lane (2003) developed three in dices as quantitative me asures of biological integrity using metrics for diatoms, macrophytes, and macroinvertebrates fo r isolated herbaceous depressional wetlands in Florid a. Similarly, Reiss (2004) deve loped a Wetland Condition Index (WCI) for the same groups of organisms for isolat ed forested wetlands in Florida; and Reiss and Brown (2005) developed a preliminary Florid a Wetland Condition Index (FWCI) for forested strand and floodplain wetlands. In all three cases these indices allowe d the comparison of changes in the composition of bi ological communities of wetland s along gradients of landscape development intensity. Fore (2004, 2005) used m odified versions of the LDI to assess the biological conditions of streams and lakes in Florida, respectively. Surdick (2005) analyzed how human land uses of varying intensi ties surrounding isolated forested wetlands in Florida affected the species composition of birds and amphibians. Amphibians that were obligatory ephemeral pond breeders decreased with increasing land use intensity. For birds, insectivor es, bark gleaners, canopy gleaners, territorial species, ground 37

PAGE 38

nesters, and cavity nesters also decreased with increasing land use inte nsity, while omnivores, herbivores, ground gleaners, canopy nesters, and exotic species increased with increasing land use intensity. The differences between species composition in less de veloped landscapes and highly developed landscapes were significant, fo llowing a gradient of increasing dissimilarity from undeveloped lands to silviculture, agricu lture, and urban land uses, respectively. Surdick (2005) pointed out the relevance of the LDI fo r ecological studies involving changes along a disturbance gradient. Since the LD I is GIS-based, it is more time and cost-efficient than indices that rely heavily on field data. Additionally, the LDI provides a quantitative measure of the human disturbance gradient compared to most indices used in studies involving measures of land use intensity which tend to be qualitative in nature. Mack (2006) tested the robustness of the LD I as a wetland condition assessment procedure using a large reference wetland data set in Ohio. The LDI was significantly correlated with the Ohio Rapid Assessment Method for Wetlands (ORA M), an independent measure of the human disturbance gradient. The LDI was also correlat ed with Ohios Vegetation Index of Biotic Integrity (VIBI), a multi-metric index of wetland integrity. The most significant relationships were found between the LDI and metrics from emergent wetlands, followed by forested wetlands, and shrub wetlands. Mack (2006) emphas ized the robustness of the LDI as a measure of the human disturbance gradient given its theoretical foundations and quantitative nature. Landscape Pattern: Quantification and Application Metrics of landscape patter n allow the quantification of composition and spatial configuration of humandomi nated land uses in watershe ds and provide quantitative measurements of human-induced impacts. They are useful in watershed management because they can be used to assess changes in water qua lity (USEPA 1994; ONeill et al. 1997; Gergel et al. 2002; Griffith 2002; Turner 2005). Such recogn ition has led to the development of research 38

PAGE 39

directed towards understanding the behavior and relevance of metr ics in watershed studies, to explore relationships between land use/land cover and water chemistry and between land use/land cover and the condition of freshwater biological communities, and to determine the spatial scale that best de scribes such relationships. Selection of pattern metrics The quantification of landscape pattern has received considerable attention due to the widely accepted relationship between spatial pattern and ecological proc esses. The question of what the minimum number and types of measurem ents are that appropri ately describe overall landscape status is of special interest. Since one metric alone cannot cap ture all the relevant aspects of pattern, multi-metric measurements are required. To help in the identification of landscape indicators or metrics for monitoring landscapes in te rms of land cover and land use patterns, Riitters et al (1995) have suggested th e use of measurements th at provide information on six dimensions or orthogonal factors that can be represented by six independent metrics: average perimeter-area ratio, contagion, standardi zed patch size, patchperimeter-area scaling, number of attribute classes, and large-patch density-area scaling. Turn er et al. (2001) have recommended the use of a minimum of five metr ics that are independent from each other when used to describe a landscape. In the assessment of watershed integrity, USEPA (1994) and ONeill et al. (1997) have recommended the use of landscape indicators of patterns that provide information on at least three dimensions: spat ial composition, shape complexity, and spatial configuration. Several authors (C ain et al. 1997; Cifald i et al. 2004; Kearns et al. 2005) have made efforts to identify a set of core metrics that can be generally used as pattern measurements in watershed and ecological condition of aquatic ecosystem studies Factor analysis and principal component analyses; which account for most of the variance among the original variables (Johnson and Gage 1997; Dytham 1999), have been frequently used for such a purpose. Even 39

PAGE 40

though similar sets of metrics have been propos ed, their behavior across different landscape remains uncertain and requires further investigation (Cifaldi et al. 2004). Landscape influences on ecosystem condition Studies of the influence of land use on su rface waters for different ecosystems using landscape pattern metrics can be divided into two related groups ; those that study changes in water chemistry, and those that focus on the change s in the characteristic s of aquatic biological communities. Water chemistry Water chemistry studies have empha sized the study of the relationship between land use and excess nutrients, partic ularly nitrogen and phosphorus. Hunsaker and Levine (1995) used the proportion of seven land use types, dominance, and contagion as landscape metrics in an effort to develop methods to characterize landscape attributes that influence water quality at different scales in the Wabash River in Illinois. Their results showed that landscape metrics were effective predicto rs of phosphorus and nitrogen loads. Landscape metrics (eight variables combined) accounted for 71 to 85% of the variance in phosphorus loads, and 42 to 53% of the variance for nitrogen loads acr oss a range of watershe d sizes. In a study of multiple watersheds in the Mid-Atlantic region, Jones et al. (2001) used metrics of land use composition (fraction occupied by different land us es) to explain the high amounts of variation in nutrient and sediment loads to streams. Landscape metrics consistently explained the variability of nitrogen (65 to 85%) and dissolved phosphorus (73%) inputs to streams. Crosbie and ChowFraser (1999) showed that the water chemistr y of 22 marshes in the Great Lakes basin was significantly affected by land use in their resp ective watersheds. Usi ng principal components analysis they were able to show that the c ontent of inorganic solid s and phosphorus in the sediments and the ionic strength of the water positively correlated with percent agricultural land. However, soluble reactive phosphorus and nitrate nitrogen concen tration in the water did not 40

PAGE 41

correlate with land use. Soranno et al. (1996) de veloped a simple model to account for spatial pattern in topography and land use using GIS. La nd use scenarios were used to explain annual phosphorus loads in Lake Mendota, Wisconsin, whic h is a nutrient-rich lake in a watershed dominated by agricultural and urban lands. M odeling results showed that phosphorus loading was greatest when natural vegetation was convert ed to agricultural or urban land uses. Studies that have found evidence of the relationship be tween landscape metrics and water chemistry for nitrogen and phosphorus include Osborne and Wille y (1988), Johnes et al. (1996), Johnson et al. (1997), Jones et al. (2001), Gri ffith et al. (2002), Brett et al (2005), and Uuemaa et al. (2005); for suspended sediments Johnson et al. (1997), Cr osbie and ChowFraser (1999), Jones et al. (2001), Sponseller et al. (2001) and Houlahan and Findlay (2 004); for conductivity Hunsaker and Levine (1995), Crosbie and Ch owFraser (1999), Sponseller et al. (2001), and Griffith et al. (2002); for alkalinity Johnson et al. (1997) and Sponseller et al (2001); for dissolved organic carbon (Gergel et al. 1999); a nd Uuemaa et al. (2005) for biological and chemical oxygen demand. Biological integrity The use of landscape metrics to assess changes in the characteristics of aquatic biological communities ha s also been investigated. Roth et al. (1996) studied the effect of land use/land cover on the biological integrity of stream ecosystems in River Raisin in Michigan. They evaluated the c ondition of streams using an Index of Biological Integrity (IBI) and a habitat index (HI). The IBI was developed based on 10 metrics of fish collected from the streams. The HI was calculated based on nine metr ics that measure in-stream and bank variables. The results showed that the stream biotic integr ity and habitat quality co rrelated negatively with the percent of agricultural land and positively with the extent of wetlands and forests. For the same watershed, Lammert and Allan (1999) examined fish and macroinvertebrate assemblages to 41

PAGE 42

relate the overall biotic condition of streams to patterns of land use. They found that land use immediate to the streams was a good predictor of biotic condition. They used proportions of agricultural and forested land as the metrics of landscape patt ern. Sponseller et al. (2001) have suggested that differences in the structure of the macroinvert ebrate assemblages in nine headwater basins in the southern Appalachian region could be explaine d by land cover patterns. Two land use classes, forested and non-forested land, and their proportions, were used in the analysis. Wang et al. (2001) analyzed the relati onship of the amount and spatial pattern of land cover with stream fish communities in 47 small wa tersheds in Wisconsin. Their results suggested that the increase in urbanization had negatively affected stream habitat and fish communities. Again, land use proportion was the preferred lands cape metric. In a study of 20 watersheds in West Virginia, Snyder et al. (2003) found that urban land use negatively affected the biological integrity of streams. Biological integrity was qua ntified using a fish-based IBI. No meaningful relationships were found between th e proportion of agricultural lands and the IBI. Similar studies that included fish and macroinver tebrates are those by Richards et al. (1996), Allan et al. (1997), Mensing et al. (1998), Galatowits ch et al. (1999), Gri ffith et al. (2002), Pess et al. (2002), Townsend et al. (2003), and Wang et al (2003). Other studies have considered amphibians (Galatowitsch et al. 1999; Knutson et al. 1999; Guerry and Hunter 2002), birds (Miller et al. 1997; Galatowitsch et al. 1999; Austin et al. 2001) and plants (Miller et al. 1997; Galatowitsch et al. 2000). Landscape scale influences on freshwater ecosystem condition Since landscapes and watersheds are comple x systems, their study requires a multi-scale approach to fully understand and manage them (USEPA 1994; Allan and Johnson 1997; Hay et al. 2001). The importance of landscape scale in de termining land use-water relationships has led to questions regarding how responses vary when relationships are anal yzed using different 42

PAGE 43

spatial resolutions, and how the sp atial extent at which land use pa ttern can best explain changes in the condition of freshwater ecosystems. These studies have at times shown contrasting results and at other times complementary observations. Grain size Hunsaker and Levine (1995) develo ped methods to characterize landscape attributes that influence water quality at various scal es. They compared land use data from the Wabash River basin in Illinois a nd the Lake Ray Roberts watershed in Texas with very different spatial resolutions; the Illinois data were courser. Specifically they looked at different areas of influence, the entire watershed, hydrological active areas, and equidistant corridors of 200 meters and 400 meters around streams to model the infl uence of the landscape on water quality. They were able to show that for the Illinois basin wa ter quality could be accurately predicted from land use information for the entire watershed, while land use near streams was not a critical factor in modeling water quality. These result s contradicted to some extent those of the Texas site, whose results showed that all land use areas were impor tant to modeling. In Michigan, several studies have looked at the landscape influence on stream biotic integrity at different scales in the River Raisin watershed. Based on Frisse ll et al. (1986), the hierarchical framework for stream habitat classification (habitat-reach-segment-subcatchment-b asin), Allan et al. (1997) and Roth et al. (1996) have suggested that the influence of land use on stream integrity is scale-dependent. These studies showed that the sub-catchment s cale was the best predic tor of stream biotic condition when analyzing land use at a regional scale. A similar study conducted by Lammert and Allan (1999) used a finer scale measuremen t for the same watershed, which presented different results. In this case, habitat and immediate land use were better predictors of stream biotic integrity than land us e for the entire catchment. However, this was only true for 43

PAGE 44

macroinvertebrates. The authors argued that m easurements at the habitat level may not be appropriate in explaining the variability in fish assemblages at that scale. Spatial extent Studies that considered different spatial extents have also shown varying results. Wang et al. (2001) found in a study of 47 small watersheds in southeastern Wisconsin that land use within 50-meter buffers had more influence on fish assemblages than land use within buffers of 100-meters and greater; and mo re influence for land use data within a 1.6-km radius upstream from sampling sites than for land use data within greater radiuses upstream. Snyder et al. (2003) found that la nd use patterns for the whole watershed were more strongly associated with fish assemblages than riparian land use patterns in a West Virginia study; Mensing et al. (1998) obtained si milar results in a study of ripari an wetlands in Minnesota. Wang et al. (2003) concluded from a study of 79 watersheds in Grea t Lakes region that the relative influence of land use among other variables at th e reach scale on fish assemblages may be more important in non-developed areas, while land use at the watershed scale may be a more important variable with increasing modifi cation of the landscape by humans. The findings by Sponseller et al. (2001) for macroinvertebrates in streams of the Appalachian region seem to support the hypothesis that local or near-stream land use ha s a strong influence on the structure of the assemblage. They quantify land cover at five spat ial scales: the entire catchment, the riparian corridor, and three buffers extending 200, 1000, and 2000 meters upstream of sampling reaches. Landscape pattern within the 200-meter buffe r presented the best relationship with macroinvertebrate metrics. Mensing et al. (1998) results for macroinver tebrates also provided evidence in this direction. In New Zealand, To wnsend et al. (2003) found that the proportion of pasture land within the riparian zone of the Taieri River accounted for most of the variation in macroinvertebrate assemblages than for the entire watershed. 44

PAGE 45

In the analysis of the influence of land use on changes in the water chemistry, results by Johnson et al. (1997) showed that total phosphorus and total suspended solids were much better explained by the land use within a 100-meter buffer for streams of the Saginaw Bay Catchment in central Michigan than for the entire subcat chment. Total nitrogen, nitr ate, orthophosphate, and alkalinity were equally explained by land use at the two scales. For a watershed in southeastern Michigan, Allan et al. (1997) f ound that nutrient and sediment i nputs were influenced more by the land use measured at broader scales than at mo re local scales. In a study that included a total of 73 wetlands in Ontario, Canada, Houlahan and Findlay (2004) found that the effect that the land use surrounding wetlands has on water and se diments nutrients could extend for relatively large distances after considering a series of land buffers at different inte rvals that range up to 5000 meters from the wetlands edges. They f ound that water phosphorus and nitrogen levels correlated strongly with forest cover at 2,250 meters from the we tland edge. Gergel et al. (1999) reported that the proportion of wetlands in the watershed was a better pr edictor of dissolved organic carbon (DOC) than in 200-me ter riparian buffers for rivers in Wisconsin. For lakes, the proportion of wetlands within 50-meter buffers se emed to explain more of the DOC than the proportion of wetlands for entire watersheds. Study Area: Florida, an Overview Physical and Ecological Aspects Florida is an ecologically and climatically dive rse region located in th e southeastern tip of the United States. The state has approximately 14 million hectares in land area, two-thirds of which occur as a long peninsula running in a nort h-south direction and with a humid subtropical climate. The remaining third of Floridas land su rface, know as the Panhandle, is located in the northwest portion of the state and has a more temperate climate. Florida has a relatively flat terrain with a mean elevation of approximately 30 meters above sea level. Floridas climate is 45

PAGE 46

generally warm and humid, with rains that vary seasonally averaging between 1.2 to 1.5 meters per year (Chen and Gerber 1990). Northern Florida is within th e southern temperate zone and is home to both deciduous and evergreen hardwood forests, including pine trees in the uplands and bottom hardwoods in the alluvial plains (Odum et al. 1998a). The Ok efenokee Swamp, a headwater wetland, is an important feature of the landscap e to the northeast. Pine flatw oods are also common in Central Florida where individual stands may comprise very large areas intermixed with other less extended forest systems and wetlands including isolated cypress heads, bayheads, wet prairies, and marshes (Abrahamson and Hartnett 1990). The southern tip of the peninsula, although highly modified by development, still contains tropical ly-influenced hammock forests. However, the most outstanding natural feature of Floridas southern landscape is the presence of a complex system of wetlands that include the Everglades, wet prairies, sawgrass and tree islands, and the Big Cypress headwater wetland (Odum et al. 1998a). Wetlands (marshes and swamps) are a common f eature of Floridas landscapes. The largest freshwater mash area is the Everglades in Sout h Florida, while other freshwater mashes are unevenly distributed throughout the state. Unlike marshes, swamps are widely distributed throughout the state and constitute a diverse set of systems that can be found in river floodplains, in the margins of lakes, or as isolated systems in the form of strands or ponds (cypress domes). A cypress dome is illustrated in Fi gure 1-3 in the form of a systems diagram which shows the main components and processes of a forested wetland in Florida. The main inflows into the cypress dome are sunlight, wind, and wa ter (rain, runoff, and groundwater) as shown to the left of the diagram. Rainwater is the main water inflow and source of nutrients. Oxygen and carbon dioxide are brought into the wetland from the air. R unoff from the surroundings is limited since the 46

PAGE 47

drainage area of cypress domes is usually sma ll. Wetland trees, herbaceo us vegetation, and algae are the main producers in the system. Organic ma tter, mostly produced within the system, may accumulate in large quantities as part of the wetlands soil. Consumers include bottom-dwelling macroinvertebrate and amphibians, as well as birds and mammals which are usually occasional visitors. Florida is home to more than 1,700 streams and rivers, which comprise approximately 80,000 kilometers (km) in length (Nordlie 1990; Fernald and Purdum 1998). There is great variation among Floridas rivers and streams in te rms of their physical ch aracteristic and their chemical and biological features. The state also has over 300 springs, most of which are artesian (Nordlie 1990). Figure 1-4 is a system diagram of the main components and processes of a section of a Florida stream. Main inflows in the system are light, wind, and water in the form or rain, runoff, and stream-inflow as shown to the le ft in the diagram. Light drives photosynthetic production as it is used by phytoplankton, submerged plants, and emergent pl ants when present. Carbon dioxide required for photosynt hesis is brought into the syst em from the air. Runoff from the surrounding lands bring into the system organi c matter, nutrients, and sediments. Nutrients are absorbed by producers and microbes. Organic matter accumulates as detritus, a portion of which is decomposed by microbes. Some organic matter and the sediments flow downstream. Filter feeders, bottom feeders, and fish are among the stream/river consumers. The state is also home to around 7,800 lakes that are over one-half of a hectare in size and which cover at least 9,720 square kilometers (km2), or 6 percent, of Floridas surface area (Brenner et al. 1990; Fernald and Purdum 1998). Th e lakes are distributed unevenly in the state with more than half located in the central sandy ridge system (Brenner et al. 1990). However, the largest lake in the state, Lake Okeechobee, is located in the South. Floridas lakes are very 47

PAGE 48

diverse in terms of their physical chemical, and biological featur es. Figure 1-5 depicts a typical lake in Florida. Main inflows in the system ar e light, wind, and rain. Othe r waters inputs are in the form of runoff and groundwater. Photosynthetic production is driven by light and may be limited depending on the lakes depth. Phytoplankton, submerged plants and algae, and emergent plants are among the main producers. Carbon di oxide and oxygen are brought into the lakes water by the wind. Runoff from the surrounding lands brings organic matter and nutrients into the system. Nutrients are absorbed by producer s and microbes while organic matter accumulates as detritus, a portion of which is decomposed my microbes. Organic matter and sediments tend to accumulate in the lakes bottom. In the absenc e of a surface water outflow, which is a common aspect of most Florida lakes, the hydrologic c onnections with the outer systems are limited to groundwater exchange and evaporation. Zooplankt on, bottom feeders, fish, and birds are among the lakes consumers. Estuarine systems are also common in Florida. In the northern part of the state they dominated by salt marsh communities, while in th e southern half mangroves and sea grasses are the predominant ecosystems. The southern-most por tion of the state contains the Florida Keys, which are separated from the peninsula by the gra ss flats of Florida Bay. Coral reefs are present along the Keys. Land Use Humans have lived in Florida for more than 10,000 years. The earliest Floridians were hunter-gathers and had little impact on the state s landscapes (Ewel 1990a). With the adoption of agriculture by native Floridians some 800 years ago, the features of the states landscapes began to change first by the use of fire, and more recently by deforestation, intensification of agriculture, dewatering and canal ization of aquatic systems, and urbanization (Ewel 1990a; Reynolds 1999). 48

PAGE 49

Currently, about 31% percent of Floridas land area is in upl and forests (Wear 2002). It is expected that this figure will decrease in the future due to development pressures in the state. It is estimated that approximately 2.5 mil lion hectares of forest lands have been lost in Florida since the 1930s; and land development greatly contributed to this change (FDACS 2005). Wetlands (marshes and swamps), which used to occupy more than half of the states land area, today account for 30% percent of Floridas land area with 2.2 million hectares covered by forested wetlands or swamps (Ainslie 2002). Although the rate of loss of we tlands to other land uses has decreased compared to the period between the 19 50s to the 1970s when it reached its peak, wetlands will continue to be lost in Florid a due to agriculture, ur ban development, and silviculture. Forested wetlands will be affected particularly by urbanization and conversion to other wetland types (Ainslie 2002). Agricultural lands cover 30% of Floridas land area (USDA 2005). Although between 1945 and 1974 th e area in agriculture increas ed steadily, since then it has remained about the same. Floridas urban land uses account for approximately 12 percent of the land area (Wear 2002). Freshwater Ecosystem Degradation Overall, the main sources for the impairment of the states surface waters are urban and agricultural runoff, domestic and industrial wastewaters, and hydrol ogic alterations (Paulic et al. 1996). The presence of high levels of nutrients, hi gh loads of organic matter that may result in low concentrations of dissolved oxygen, si ltation, habitat degradation, and bacterial contamination are the main problems leading to stream degradation (Paulic et al. 1996). Problems for lakes have resulted mostly from nutrient enrichment, the presence of toxins and metals, acidification, and habitat degradation (Bre nner et al. 1990, Paulic et al. 1996). Most of the problems that are facing lakes and streams ar e also common in the forested wetlands of the state. Physical disturbance or modification has been of special concern because of the 49

PAGE 50

50 characteristics of these systems. Since forested wetlands are abundant with trees, almost every wetland in the state has experienced some type of logging since the late 1800s (Ewel 1990b). Although these forests may regene rate after logging has occurre d, this regeneration may not happen without some degree of shif t in their community composition. Scale of Investigation Floridas hydrologic units subdiv ision of watersheds defined the area of investigation for streams and lakes in the assessment of human imp acts on the quality of su rface waters. Selected watersheds were distributed throughout the state excluding the Ev erglades and Florida Keys. In addition, a sample of 118 isolated forested we tlands, also distributed throughout the state, was used to complement the analysis of human landscape scale impacts on freshwater systems.

PAGE 51

51 Figure 1-1. Systems diagram of the impacts of agricultural lands on freshwater ecosyst ems. B = biomass, Spp. = species, Sed. = sediments, N & P = nitrogen and phosphorus, Tox. = toxins, and O. M. = organic matter (modified from Brown and Vivas 2005). The symbols used are explained in Appendix A-1. Agri cultural lands are represente d with a box symbol provided that the elements included (and their inte ractions) are not limited to the fields or crops but to a variety of components.

PAGE 52

Figure 1-2. Systems diagram of a Florida cypress dome. Figure 1-3. Systems diagram of a stream section in Florida. 52

PAGE 53

Figure 1-4. Systems diagram of a Florida lake. 53

PAGE 54

CHAPTER 2 METHODS The Landscape Development Intensity Index (LDI) and landscape pa ttern indices were calculated for watersheds and areas of influe nce of different sizes surrounding 118 isolated forested wetlands, 69 streams, and 54 lake s in Florida. The indices were used to assess the influence of human development on these aquatic ecosystems. This chapter describes site selection, data sources, scales of analysis, and selection and calcula tion of landscape indi ces. The statistical analyses that were used are also described. Site Selection Isolated Forested Wetlands The isolated forested wetlands included in th is study were selected following the criteria set by Reiss (2004) for the development of sets of biological indicators to assess th e ecological integrity of isolated forested wetlands in Flor ida. With a nearly equal spatial distribution of wetlands within each of the four Florida wetla nd ecoregions (Panhandle, North, Central, and South) defined by Lane (2000), a total of 118 isol ated forested wetlands were identified through field surveys and with the aid of aerial photogra phy. The isolated forested wetlands varied in size from 0.07 to 2.1 hectares (mean = 0.68; SD = 0.44). Their locations are pres ented in Figure 2-1. Isolated forested wetlands were selected a priori as belonging to one of three different landscapes: isolated forested wetlands surrounde d mainly by undeveloped lands (n = 37), which are hereafter referred to as refere nce isolated forested wetlands; isolated forested wetlands within an agricultural landscape (n = 40), which are hereafte r referred to as agricultural isolated forested wetlands; and isolated forested wetlands within an urban landscape (n = 41 ), which are hereafter referred to as urban isolated fo rested wetlands. The reference isol ated forested wetlands were generally located on conservation lands including state and national parks and forests, county 54

PAGE 55

and city lands, and private cons ervation tracts. Agricultural isolated forested wetlands were usually surrounded by cattle pasture, row crops, tree crops, and silvicultural land uses. Urban isolated forested wetlands were surrounded by a variety of residential, commercial, industrial, recreational, and public land uses; many of the urban isolated forest ed wetlands were believed to have previously belonged to an agricultura l landscape (Reiss 2004). Tabl e 2-1 provides general information about each wetland, including the sampling date and surrounding land use. Streams Site selection was made from the stream samp le used in the devel opment of the stream condition index (SCI) for Florid a and initially reported by Bar bour et al. (1996b). The SCI was developed using seven descriptors, or metrics, of stream benthic macroinvertebrates that are altered with increased human disturbance (Barbo ur et al. 1996a; Barbour et al. 1996b). The SCI was recalibrated by Fore (2004) and new streams were included in the analysis with a final total sample of 223 streams. The streams that were se lected in the development of the original SCI were distributed homogenously in Florida into three bioregions (Panhandle, Peninsula, and Northeast) in order to control for the biological variance that w ould otherwise occur (Barbour et al. 1996b). Bioregions resulted fr om the aggregation of Floridas stream ecoregions that were developed by Griffith et al. (1994) and did not consider the Evergl ades since the streams in this area have been subject to significant hydrologic alterations (Barbour et al. 1996b). Site selection was made cons idering only those streams for which the macroinvertebrate data were collected between 1993 and 1995. The stream sample size was further narrowed based on the availability of water qual ity data for each stream and for the same period of time using EPAs Water Quality Storage and Retrieval (STORET) database (available at www.epa.gov/storpubl/legacy/gateway.htm ) and the data used in the surface water quality assessment that was developed for Florida in 1996 by Paulic et al. ( 1996) (also known as the 55

PAGE 56

305[b] Report). Since not every site had water quality data associ ated with it, and to increase sample size, the selection of water quality data was made increasing the sampling time range for each station (or site) 1year. Fina lly, sampling stations used as tests sites in the development of the SCI that were associated with point sour ces of pollution were removed from the stream sample. As a result, a total of 69 streams were selected in Florid a. Their locations are shown in Figure 2-2 as STORET sampling stations. Tabl e 2-2 provides information on each stream, including the STORET sampling station number, the drainage basin name, and the United States Geological Survey (USGS) Hydrologic Unit Code (HUC) that identifies the hydrologic unit to which each stream belongs, and the corresponding stream bioregion. Lakes Lakes were selected using a subset of the la kes currently used by FD EP in the development of a biological index for Florida lakes which was initially reported by Gerritsen et al. (2000). The development of this index originally included six metrics in three alternative lake condition indexes (LCI) that were based on biological data collected between 1993 and 1997 for 206 lakes within 36 of the 47 Florida lake re gions that are identified by Gr iffith et al. ( 1997). In addition, the lakes were classified into five categories ba sed on three independent f actors: water color, pH, and ecoregions. This classification allowed the id entification of Floridas lakes by types: acidclear lakes of ecoregion 65 (Southe ast Plains), acid-clear lakes of ecoregion 75 (Atlantic Coastal Plain), acid-colored lakes, alkalin e-clear lakes, and alkaline-colored lakes. The lakes were also associated to Level 3 ecoregi ons as defined by Omernik (1987). The development of the LCI is an ongoing process led by FDEP; thus far 500 potential lakes have b een identified to examine the relationships between macroinvertebrate commun ities, water quality variables, and land use within lake basins (FDEP 2005). 56

PAGE 57

Lake selection for this dissertation was made from a subset of 102 lake sampling stations provided by FDEP (R. Frydenborg 2005, Environmental Assessment Section, Bureau of Laboratories, personal communicatio n). Only those lakes for whic h the macroinvertebrate data were collected by the same agency between the years 1993 and 1995 were considered. The availability of water quality data for each lake and for the same period of time was determined using the data provided by FDEP and using th e STORET database. Since not every site had water quality data associated with it, the fina l selection of sites was made increasing the sampling time range 1year in order to increase sample size. As a result, 54 lakes were selected and their location is shown in Figure 2-3 as STORET sampling stations. Table 2-3 provides general information about each lake, includi ng the lakes name, STORET sampling station number, and ecoregion. Data Sources Land Use / Land Cover Data and Land Use Classification Systems The required land-use data of the surrounding landscapes of the isolated forested wetlands were delineated using color-inf rared orthorectified images (D igital Orthographic Quarter Quad [DOQQ]) for the year 19 99, with a 1-meter resolution and 3.7 5 x 3.75-minutes in extent. The onscreen digitizing was displayed on a computer screen at a scale of approximately 1:5,000. The spatial data were updated and verified in the field during May-September of 2001 and MayOctober of 2002. Figure 2-4 illustra tes the land-use data delineated for a wetland within an urban landscape. The data were comparable to biological a nd water chemistry data collected during the same dates and were used by Reiss (2004) in developing the Wetland C ondition Index (WCI), a quantitative measure of the biolog ical integrity of isolated forested wetlands in Florida. The land-use data for the drainage areas for streams and lakes were obtained from FDEPs geodata directory, which is available at www.dep.state.fl.us/gis The data were originally 57

PAGE 58

developed by the five Water Management District s (WMDs) in the State of Florida for the years 1995, 1999, and 2000. The 1995 datasets were devel oped by each WMD via photo interpretation to capture land use/cover information from DO QQs. The National Aerial Photography Program (NAPP), a multiple-agency program coordinated by the USGS that provides panchromatic and color infrared aerial photos of the United States at 1:40,000 (USGS 2006), recorded the images from 1993 through 1995. The 1999 and 2000 datasets were developed from the 1995 datasets and were updated using DOQQs that were obtained by NAPP during 1999 and 2000. Table 2-4 summarizes the information related to the land use datasets that were developed by each of Floridas five WMDs that are used in this study. The land use/cover features developed by th e WMDs are categorized according to the Florida Land Use and Cover and Forms Classi fication System (FLUCCS). The FLUCCS was developed by the Florida Department of Trans portation (FDOT), following guidelines set by the USGS, and is a hierarchical system of categorie s with four levels ra nging from general to specific. Level I consists of ni ne classes (Urban and Built-up, Agriculture, Rangeland, Upland Forest, Water, Wetlands, Barren Land, Transportation/Communication/Ut ilities, and Special Classifications) (FDOT 1999) which are further subdivided into finer deta il (Levels II through IV) and increase in resolution with each level. Land Use Intensity Classification Using the general structure of the FLUCCS and th e definitions set for each of the land uses included in this classification system, and based on the energy flow characteristics of each land use type, a classification scheme of land use was developed according to the intensity of human development. The land use categories for the in tensity of human development are defined in Table 2-5. This classification system of land use intensity consists of 25 categories and is limited to one level that was matched with the classes of Levels II and III of the FLUCCS. The matched 58

PAGE 59

classes are presented in Appendix B. The choice of land use intensity categories adheres to the specific research needs for th is dissertation, and complement s the research done by Brown (1980), Harper (1994), Parker (199 8), and Brandt-Williams (1999). Non-renewable and purchased energies we re the primary source for quantifying the intensity of human activity, which in emergy calculations is e xpressed as emergy use per unit area per time or areal empower density (Odum 1996). The non -renewable and purchased areal empower density was calculated as average values for land use categories from previous studies by Brown (1980), Whitfield (1994), Doherty (19 95), Parker (1998), and Brandt-Williams (1999; 2001). In these previous studies, th e energy consumption data were collected from actual billing records and from the literature, and the data were averaged on a per unit area basis for different land use types. A summary of the emergy evalua tions used is presented in Appendix C. The resulting areal empower densities for the 25 la nd use categories or the intensity of human development are presented in Table 2-6. Most of Florida landscapes will fall somewhere within the provided scale of values when they are de scribed in terms of the use of non-renewable and purchased energies. The land use data for the surrounding landscapes (buffer areas) of the sample isolated forested wetlands, as well as the drainage ba sins of the sample streams and lakes, were reclassified using the land use intensity classes and values from Table 2-6. Figures 2-5 and 2-6 shows results of this reclassification process. Water Chemistry Data Isolated forested wetlands The water chemistry data for the isolated forested wetlands were collected between 2001 and 2002 by a research team from the Center for Wetlands at the University of Florida. Water samples were collected for a total of 75 isolated forested wetlands and were taken from the 59

PAGE 60

deepest pool of each wetland when a minimum of 10 centimeters of standing water was present throughout at least half of the wetland. Wa ter samples were collected only once. Water samples were analyzed by the FDEP Ce ntral Chemistry Laboratory in Tallahassee, Florida. The water chemistry measures measured by FDEP were color, turbidity, pH, specific conductance (SC), ammonia-nitroge n, nitrate/nitrite-nitrogen (NO2/NO3), total Kjeldahl nitrogen (TKN), and total phosphorus (TP). Dissolved oxygen (DO) and water temperature measurements were taken onsite using a YSI-55 Dissolved O xygen hand meter. Among the variables measured, only turbidity (n = 75), SC (n = 33), TP (n = 75), and DO (n = 71) were used in this study to explore the relationship between the changes in land use and the potential impact on the ecological condition of these systems. Total nitrogen (TN, n = 62) was also included as a water chemistry variable and was calculated as the sum of NO2/NO3 and TKN. DO, TN, and TP were as given concentrations (mg/L), turbidity as ne phelometrics turbidity units (NTU) and SC was measured as micromhos per centimeter ( mhos/cm). All of these water chemistry variables have been reported to be relevant for watershed-land use water quality studies and used to assess the ecosystem condition of freshwater systems (O sborne and Wiley 1988; Johnson et al. 1997; Tufford et al. 1998; Tufford et al. 2003; H oulahan and Findlay 2004; Brett et al. 2005). Information on the availability of water chemistry data for the sample isolated forested wetlands is provided in Appendix D (Table D-1). Streams The water chemistry data for streams used in this study were obtained primarily from the data used in developing the Water Quality I ndex (WQI) to assess the condition of Floridas streams in 1996, and included as part of the 1996 305( b) Report. This is a biennial water quality report submitted by the State of Florida to the EPA in which the status of the water quality in the state is detailed, as required by Section 305(b) of the Clean Water Act (CWA) (Paulic et al. 60

PAGE 61

1996). The WQI is the arithmetic mean of six water quality categories (clarity, DO, oxygendemanding substances, bacteria, nutr ients, and biological diversity) and uses data obtained from EPAs STORET database. The WQI is reported w ith values ranging from 0 to 99. High index scores are indicative of water quality impairment. Paulic and collaborators (1996) used 26 wate r quality measurements from STORET water chemical data for two different periods of tim e, 1980 to 1989 and 1990 to 1995 to define the six water quality categories that compose the WQI. In some cases, more than one measurement was reported for water quality variables for each ST ORET station. Thus, the annual average was calculated for the data in these cases and then use d. The criteria that were followed in calculating an annual mean dictated that a station had to be sampled at least once during the colder months and once during the warmer months for a gi ven year. For the study performed for this dissertation, only five water quality measurem ents were considered (i.e. turbidity, DO, NO3/NO2-N, TN, and TP) in addition to the WQI to explore the relationship between human development, quantified through the landscape indices, and their potential impact on the water quality of Florida streams. Only data reporte d for the period between 1992 and 1996 were used, with the exception of three si tes for which data reported for 1990-1994 were also included. This exception was made to increase the sample size, since the water chemistry data were not available for all the streams that were initially considered. The water quality data used from the 305(b) Report were complemented with data ob tained directly from the STORET database (available at www.epa.gov/storpubl/legacy/gateway.htm ) for sampling stations not considered in the development of the WQI. These stations included mainly those sampled during 1996. Data selection and usage was made following the same cr iteria set by Paulic et al. (1996), as well as that which was reported as collected by FDEP. TN was calculated from STORET data as the 61

PAGE 62

sum of nitrate/nitrite nitrogen a nd TKN. Water chemistry data were available for a total of 47 streams, and the water chemistry measurements were reported as concentrations. Appendix D (Table D-2) presents a summary of the information on the water chemistry variables considered for streams in this study. Lakes The water quality data for lakes used in th is study were collected by FDEP between 1993 and 2000. Unlike the data for streams, many lakes were sampled only once; therefore, the data were not discarded and were considered for furt her analysis. For cases in which there was more than one measurement for a given year, the annual average was calculated for each variable. The water quality data, which corresponded to th e years 1996-2000 were provided by FDEP (R. Frydenborg 2005, Environmental Assessment Sec tion, Bureau of Laboratories personal communication), and are available by accessing EP As STORET database. Five water chemistry variables were considered: ammonia nitrogen, NO3/NO2-N, TKN, TN, and TP. TN was calculated as the sum of NO3/NO2-N and TKN, and all water chemistry measurements were reported as concentrations. Appendix D (Table D3) presents a summary of the water quality variables used for th e lakes in this study. Biological Data Isolated forested wetlands The individual index for each of the three biol ogical assemblages used in the development of the Florida WCI by Reiss (2004 ) was used in the assessment of the effect of land use on the condition of wetlands. The WCI is a multimetric inde x that quantifies the bi ological integrity of pond-cypress wetlands. The index resulted from the combination of 19 metrics (7 diatom metrics, 6 macrophyte metrics, and 6 macroinverteb rate metrics) that measure changes in the 62

PAGE 63

biological community composition of wetlands along a gradient of human disturbance. The metric composition of each WCI is presented in Appendix E. Potential scores for the diatoms WCI ranged from 0-70, while the WCIs for the macrophytes and the macroinvertebrates ranged from 0-60. In all three cases, higher values represent the isolated forested wetlands that are surrounde d by low-intensity land uses (Reiss 2004). The WCI scores for each wetland are also presented in Appendix E. It should be noted that Reiss (2004) used the LDI that was reported by Brown and Vi vas (2005) as the measure of the human disturbance gradient in the developmen t of the WCI. This LDI was an earlier version of the area-based LDI that is us ed in this dissertation (a descri ption of this LDI is presented later). However, the WCI was still considered usef ul for testing the predictive power of the new form of the LDI. Streams The biological data used for streams are summar ized in the SCI that was first developed by Barbour et al. (1996b) for Florida. The SCI is an index that quantifies the biologi cal integrity of streams and is used as a biomonitoring tool to assess the effectiven ess of non-point source pollution control in the state. The original SCI aggregated seven metrics representing characteristics of bottom-dwelling macroinvertebrat es that are expected to change along a human disturbance gradient. Least-impaired streams were defined as sites that we re wadeable (first to third order), showed minimum signs of disturbance, and were completely within subecoregions. These sites were defined as reference sites and serve to differentiate between least-impaired and impaired streams (Barbour et al. 1996b). SCI scor es ranged between 7, which refer to a very poor biological condition, and 33, which relate to a very good biological condition. The SCI was calibrated for summer and wint er biological index periods. 63

PAGE 64

The original version of the SCI was recently refined by Fore (2004). This newest version of the SCI aggregates 10 metrics that also represent characteristics of bottom-dwelling macroinvertebrates that are expected to change along a human disturbance gradient. The human disturbance gradient was devel oped by combining and scoring measures of hydrologic condition, habitat condition, and water quality measured as the concentration of ammonia nitrogen. It also includes an earlier form of the LD I (area-weighted). The values are reported on a scale from 0 to 100 with low values associated with most disturbed streams, and high values corresponding to the least disturbed sites. Add itionally, the SCI is independent of watershed size and geographic region, and shows a small vari ability between seasonal and annual changes (Fore 2004). Both versions of the SCI were used in this dissertation and in thos e cases in which more than one SCI score was reported for a given si te (one score for each of multiple sampling seasons), the average was used (arithmetic mean ). The metric composition of both versions of the SCI and the SCI values for each sampling stat ion used are presented in Appendix F. The SCI data used in this dissertat ion were provided by FDEP (R. Frydenborg 2005, Environmental Assessment Section, Bureau of Laboratories, personal communication) and were developed using field data for benthic macroinverteb rates collected by FDEP between 1992 and 1995. Lakes The Lake Condition Index (LCI) for Florida wa s used to assess changes in the biotic community composition of lakes due to the infl uence of land use surroun ding these bodies of water. The LCI was developed by Gerritsen et al. (2000) and quantifies the bi ological integrity of lakes. The index aggregates six metrics that represent characteristics of bottom-dwelling macroinvertebrates that are expected to cha nge along the human distur bance gradient. Index values range from 0 to 100, with low values as sociated with most-dis turbed lakes and high values corresponding to leas t-impaired lakes. 64

PAGE 65

The LCI scores for lakes used in this di ssertation were also provided by FDEP (R. Frydenborg 2005, Environmental Assessment Secti on, Bureau of Laboratories, personal communication). The lake data were collected by FDEP staff between 1996 and 2000. The LCIs metric composition and the LCI score fo r each lake is presented in Appendix G. Scales of Analysis The boundaries of hydrological active areas were overlain on land use maps to quantify the total area of each land use class within each boundary contributing or affecting each surface water system. For the isolated forested wetlands, equidistant areas or buffe rs were used rather than a strict drainage area to determine the co ntributing effect of different land uses on the wetlands. For streams and lakes, two drainage basins were used: 1) basins delineated using elevation data and a geographic information system (GIS) and 2) equidistant buffers. Isolated Forested Wetlands The land uses associated with the isolated forested wetlands were considered based on the landscape surrounding each site at 20, 100, and 200 me ters from its borders. The use of buffer areas instead of strict drainage areas was considered appropriat e due to the lack of detailed topographic data that would al low modeling the hydrological cont ributing areas to these small systems, most of which are located in the relative ly flat terrain of many Florida landscapes. Land use buffer areas for each wetland were obtaine d using GIS (ArcView 3.2, ESRI -1999) by first delineating buffer contours that were dr awn using a specified distance (20, 100, and 200 meters) from the study isolated forested wetlands and then by using the buffers to extract the land use data from the previously delineated land use within the surrounding landscape of each wetland. The original wetland boundary was deli neated from DOQQs for 1999. To ensure the accuracy of the delineation, boundaries for the isol ated forested wetlands were verified on the ground and corrected when required. In all cases, the geographic location of each sample wetland 65

PAGE 66

was obtained using a Global Positioning System (Garmin III GPS) unit. The extracted land use was reclassified based on the land use intensity classification system. Figure 2-7 depicts the three spatial scales used. Streams The land uses associated with streams were cons idered at three different spatial scales or hydrological active areas: th e total drainage area or total wate rshed for each stream, a riparian zone or equidistant areas of influence of 400 meters around str eams, and a riparian zone or equidistant area of influence of 100 meters around stream. Figure 2-8 illustrates the three spatial scales used. The drainage areas for streams, as well as the stream networks, were determined using the Better Assessment Scie nce Integrating Point and N onpoint Sources 3.0 (BASINS 3.0) environmental analysis system. The BASINS computer program wa s developed by the Office of Water of the EPA to support environmental and eco logical studies, and decision-making at the watershed level (EPA 2001). The BASINS asse ssment tools are integrated into ArcView 3.2. The boundaries of the 400and 100-meter buffers we re calculated using equidistant widths of 400 meters and 100 meters on both sides of the st reams networks that were obtained from the drainage basins modeling. The delineation of drainage basins and the stream networks using BASINS required the use of a digital terrain model (DTM), a grid map that masks the DTM, and a pre-digitized stream network, as summarized in Figure 2-9. The DTM provides the topographic information required in many watershed models (Sole and Valanza no 1996). The National El evation Dataset (NED) was used as the preferred DTM. The NED is a 30-meter raster-based dataset produced by the USGS, and assembled from quadrangle-based (7 .5-minute) 10and 30-meter digital elevation models (DEM). The NED significantly reduces the pre-processing steps for spatial analysis that is usually required when using DEMs. In additio n, the NED is available for large areas in only 66

PAGE 67

one file, thereby avoiding the need to use multiple files of digital elevation data when used in large-scale watershed analysis (Gesch et al 2002). The DTM for each drainage basin was masked using a state-wide watershed boundari es coverage developed by FDEP between 1994 and 1997. FDEP subdivided the state into appr oximately 4,400 watersheds using EPAs River Rich File 3 (RF3) and the USGS Hydrological Units (HUCs) (Paulic et al. 1996). The predigitized state-wide streams netw orks used were obtained from FDEPs geodata database, which is available at http://www.dep.state.fl.us/gis/. The final calculation of the drainage basin boundary was made using a stream outlet that corresponded to the location of a STORET water quality sampling station that was previously identified and with know n biological and water chemistry data. The output for a modeled drai nage basin is presen ted in Figure 2-10. Even though the NED was the best publicly elev ation data available at the time that this study was undertaken, the accuracy of the NE D for watershed delineation stream network definition was not quantitatively tested. It has been noted (Kost et al. 2002) that the NED may limit the quality of certain hydrologi c procedures. To reduce potent ial errors, the delineation of watersheds and stream channels was done using the automatic delineation method available in BASINS. The number of sub-watersheds and stream channels to be delineated was determined in all cases using the threshold ar ea, or critical source area, su ggested by BASINS. By using an automatic delineation method over a manual delin eation method, both available in BASINS, the chances for mistakes in the delineation are re duced as long as the DTM used is accurate (Oksanen and Sarjakoski 2005). Additionally, BASINS removes non-draining zones (sinks), thereby improving the accuracy in watershed delineation stream network definition. Once the drainage basins and stream networ k were delineated, the land use within the drainage basins was selected from Floridas fi ve WMDs land use datasets using GIS. The land 67

PAGE 68

use within a 100-meter and 400-meter distance of the modeled streams was captured using buffers in a similar way as described for the isolat ed forested wetlands. In all cases, the area of the different land uses was recalculated once the land use data were obtained. Lakes For lakes, three spatial scales of analysis we re also considered: the total drainage area or total watershed for each lake, an area of influe nce or equidistant buffer of 400 meters, and an area of influence or equidistant buffer of 100 mete rs. Figure 2-11 illustrate s the three scales of work. The drainage areas for each lake were determined using a hydrological extension for ArcView 3.2 and the steps followed were similar to those described previously for streams. The DTM used was the NED. For each lake, the DTM was cut (clipped) using a corresponding watershed boundary from FDEPs state-wide wa tershed boundaries covera ge. After clipping the DTM, sinks were removed and each resulting un it was smoothed and flow directions were calculated. Finally, using the resulting drainage di rection map and grid coverage for each lake studied, the drainage area or ca tchment was derived draining the uphi ll area to the lake. After the drainage basins for each lake were delineated, the land use within each catchment was selected from Floridas five WMD land use datasets us ing GIS and following the same procedure as described above for the streams. Landscape Indices To assess the impact of human activities on the condition of a quatic systems, two different sets of landscape indices were used: the LD I and landscape pattern metrics. The following paragraphs provide a description of how the indices were selected and calculated, and how they were tested to assess their behavior wh en computed at different grain sizes. 68

PAGE 69

Landscape Development Intensity Index Method of calculation Three measurements of the landscape devel opment intensity (LDI) or level of human activity in the landscape based on energy flow and expressed as emergy (sej), were used. The LDI was calculated based on the developed or no nrenewable empower density of the landscape which is the emergy per area per ti me and its units are sej/ha/yr, and can be interpreted as a measure of the areal work per time. This quantit ative measure of the inte nsity of human use of landscapes permits the scaling of the intensity of activity based on non-renewable energy used, which is common to all human-dominated landscapes (Brown and Vivas 2005). The LDI for each watershed or area of influe nce was calculated using land use data and existing emergy evaluations for diffe rent land uses in Florida. Th e quantification of the metrics was done using the GIS software MFwo rks Version 3.0 (Keigan Systems, Inc. 2002) and a spreadsheet (Microsoft Office Excel 2003). The total metric value for the landscape (drainage basin, area of influence, or other land use pr oportion) to the receiving surface water system was calculated based on the log scale of the ratio of nonrenewable empower density to the average state renewable empower density, which is the baseline (Brown and Vivas 2006). The equation used for calculation of the LDI is the following: LDI. = 10 log (empPDTotal/emPDRef) Equation 2-1 where LDI is the landscape development intensity index; empPDTotal is the areal empower density for a landscape unit (wetland, stream, lake) including the background environment; and emPDref is the areal empower density of backgr ound environment which is equal to 1.81E 15 sej/ha/yr or the empower density of rain in Florida calculated after Odum et al. (1998a). The areal empower density for each lands cape unit is calculate d based on the nonrenewable and purchased empower densities of the various land use types present within the 69

PAGE 70

landscape unit. The non-renewable energies may include electricity, fuels, fertilizers, pesticides, water (both public water supply and irrigation), and human labor Human services were not included in the calculation of the indices for ag ricultural land uses. Refer to Table 2-6 for the non-renewable and purchased areal empower densities of the various land use types considered. The LDI was calculated in three different ways for each area of influence and for entire watersheds. First, the index was calculated ba sed on the proportion of each land use type and their development intensity (non renewable and pur chased areal empower density), regardless of the distance of each land use type from the ta rget aquatic system. Second, it was calculated assuming that the effect of deve lopment intensity on the landscap e unit decreases linearly with distance from the target aquatic system. Third, th e metric was calculated a ssuming that the effect of development intensity on the la ndscape unit decreases in invers e-square with distance from the target aquatic system. For simplicity, during the rest of this document these metrics will be referred to as to as the LDI-pr oportion of land use (LDI-PLU), th e LDI-inverse linear distance (LDI-ILD), and the LDI-inverse square distan ce (LDI-ISD), respectively. The specific steps (scripts) followed for each calculation of the non renewable and purchased areal empower density for each landscape using the GIS are presented in Appendix H. Scale dependency To investigate the scale depende ncy of the LDI, it was calculated for each landscape using different grain sizes and keeping the extent constant. For isolated forested wetlands, grain (pixel) size was systematically changed from 5 meters to 70 meters for a total of eight different pixel sizes (5, 10, 20, 30, 40, 50, 60, and 70 meters). Each time the rescaling of the grain size began with the 5 x 5-meter cell size, which correspond s to the approximate minimum mapping area of the original land use data, following the same procedure used by Wu et al. (2002). The higher end of the range corresponds to the maximum gr ain size after which less than 30 pixels will 70

PAGE 71

result for a 200-meter buffer. The 200-meter buffe r was the preferred exte nt, due to the process of pixel aggregation that results from the increase in grain size. For streams, grain (pixel) size was systematically changed from 20 meters to 1 70 meters using the same rescaling rule applied for the isolated forested wetlands, with distance in tervals of 30 meters for a total of six different grain sizes (20, 50, 80, 110, 140, and 170 meters). The lower end of this range represents the approximate minimum mapping area of the original land use data. The higher end corresponds to the maximum grain size after which less than 30 pixels will result for the smallest watershed included in the stream sample studied. The preferre d extent used was the drainage basin or total watershed range. All rescaling was done in MFworks Version 3.0 using the most frequent occurring value method, a widely used procedure in ecological and spatial data analysis to aggregate categorical data (Wu 2004). In this rescaling method the value of a pixel in the new map is determined based on the patch type with the most pixels within a moving window in the original map; to break a tie the highest valu e is used by default in MFworks Version 3.0. Landscape Pattern Metrics Landscape pattern metrics have been suggested as appropriate lands cape indices to assess the impact of land use on surface waters (EPA 1994; ONeill et al. 1997 ; Liu and Cameron 2001; Gergel et al. 2002; Cifaldi et al. 2004) for watersheds of differe nt sizes (Cifaldi et al. 2004; Kearns et al. 2005). It is recommended that metrics select ion for quantifying landscape pattern should at least consider metrics that will effectiv ely respond to research needs, that the behavior of the metrics be known, and that redundanc y among metrics should be avoided whenever possible (Gustafson 1998; Turner et al. 2001). Metric selection Preliminary metric selection was done through literature review of studies that have previously linked landscape pattern with ecosystem condition and water quality variables. A total 71

PAGE 72

of 17 metrics were identified and are described in Table 2-7. Metric identification was done regardless of the type of freshwater system investigated. Most studies seem to be direct ed to studying land-water interac tions in rivers and streams, with apparently less work having been done for lakes and wetlands. To analyze the relationship of landscape pattern and indicators of ecosystem condition and water qualit y variables, the same set of 17 metrics was initially used for lakes, streams, and isolated forested wetlands. For isolated forested wetlands, pattern metrics were calculat ed only for the 200-meter buffer areas since these systems and their surrounding landscapes are small and bias in calculating pattern metrics may result when calculated for the 100and 20-me ter buffer areas. Following ONeill et al. (1996) and Turner et al. (2001)s reco mmendation that the extent for the study landscape should be at least two times larger than the landscape patches in order to avoid a bias in metric measurements, a total of 56 of 118 wetland lands capes (or 200-m buffer areas) were selected for metrics calculations. For streams, 68 out of 69 sites were included in watershed analyses, and 64 sites out of 69 sites were included for metrics calculation in buffer areas of 400 and 100 meters. For lakes 48 of 54 sites were included for metrics calculation for all three spatial extents. The effects on landscape pattern metrics when changing spatial scale is well documented and many pattern metrics are sensitive to change s in both grain and extent (Turner et al. 1989, Griffith et al. 2000; Wu et al. 2000; Wu et al 2002; Shen et al. 2004; Wu 2004; Uuemaa et al. 2005). As a result, investigating ch anges in the behavior of lands cape pattern metrics as a result of variations in spatia l scale was of secondary interest in this dissertation. However, landscape pattern metrics were calculated at four grain sizes for each freshwater system: for isolated forested wetlands at 5 x 5, 10 x 10, 20 x 20, and 30 x 30 meters; for streams at 20 x 20, 50 x 50, 80 x 80, and 110 x 110 meters; and for lakes at 20 x 20, 40 x 40, 60 x 60, and 80 x 80 meters. 72

PAGE 73

Only four scales were considered to minimize the number of one-cell patc hes formed due to cell aggregation when rescaling since multiple once-cell patches may result in bias metric calculation (McGarigal and Marks 1995). A dditionally, working with a re duced number of grain sizes helped to preserve to some extent the accur acy of representation of the original data To evaluate the spatial extent at which patter n metrics best explain variations in ecosystem condition and water quality variables for streams and lakes, th e grain size has held constant and metrics were calculated for the whole watershe d and for 400and 100-meter buffers. The grain size used in the spatial extent analysis was the approximate minimum mapping area. Landscape metric calculation Landscape metrics were calculated for each wa tershed and buffer area using the spatial analysis program Fragstats version 3.3 (McGarig al et al. 2002). Fragst ats quantifies landscape structure and has been used widely in landscap e ecology research, includi ng research performed for watersheds and aquatic ecosystems. Fragstats computes three groups of metrics: patch, type or class, and landscape metrics. To calculate the metrics, ArcGri ds were exported into Fragstats after converting vector coverages of watersheds and buffer areas into raster (5 x 5-meter cell size) coverages for isolated fore sted wetlands, and raster (20 x 20-meter cell size) coverages for stream and lakes. The chosen cell sizes co rresponded to the approximate minimum mapping units of the land use data for the three systems under investigation. To obta in data at different grain sizes, the grids were rescaled using a ne arest neighbor method alwa ys starting from the smallest cell size considered. Although this method is used for resampling spatial data it can also be used for rescaling since it effectively increases the grain size of a map (Li and Wu 2004). Since the rescaling of raster data may result in disjoint patc hes due to the aggregation or division of patches, the rescaled data were visually ex amined to ensure accurate representation of the original data as suggested by McGarigal et al. (2002). All metrics were calculated in batch file 73

PAGE 74

format using an 8-cell rule in a standard window mode with two output files: class metrics and landscape metrics. Data Analysis In this section, the methods used to descri be and investigate the behavior of the LDI calculated in three different ways at different spatial scales ar e described. The process followed to identify a set of uncorrelated landscape pattern metrics that could be used to assess land useecological condition interactions is also considered. To investig ate the relationship between land use and ecosystem condition using the LDI and landscape pattern indices regression analysis was used. The different te sts used are explained. Study Sites, Water Chemistry Vari ables, and Biological Variables Descriptive statistics were used to typify the systems under investigation. Watershed characteristics for each set of systems assessed (isolated forested wetlands, streams, and lakes) were described based on their sizes and land us e composition. Since isolated forested wetlands were not analyzed based on strict drainage areas, the 200-meter buffer was used to describe the surrounding landscape of each wetland. Watersheds were also described based on thei r intensity of development measured as the sum of the areal nonrenewable and purchased empower density (sej/ha/yr). Comparative statistics were used to describe differences in the intensity of human activity between reference, agricultural, and urban is olated forested wetlands. Landscape Development Intensity Index Description and behavior LDI values were compared for different grain sizes at which they were calculated for a subset of isolated forested wetlands buffers (n = 15) and streams watersheds (n = 15) with LDI 74

PAGE 75

scores that were considered re presentative of the whole range of LDI values for each group of systems. The degree of dispersion of the LDI va lues for each site was analyzed visually and calculating the standard deviati on from the mean LDI values. Scalograms were then used to emphasize differences among low, intermediate and high LDI scores for the landscapes investigated. Scalograms have frequently been us ed to investigate the behavior of landscape indices due to changes in both grain size and exte nt (Turner et al. 1989, Wu et al. 2000; Wu et al. 2002; Shen et al. 2004; Wu 2004; Uuemaa et al. 2005). To test the behavior of the LDI with change s in spatial extent, the Kruskal-Wallis test was used. This test is the non-parametric form of the one-way analysis of variance (Dytham 1999). The Wilcoxons signed ranks test, the non-par ametric equivalent of the paired t-test (Dytham 1999), was used to disc ern differences among medians of the LDI calculated in its three forms: LDI-PLU, LDI-ILD, and LDI-ISD. Both non-parametric tests were run using Minitab (Version 14.1, Minitab Release 14 Statistical Software). Spearmans rank order correlation, the non-para metric measure of correlation (Dytham 1999), was used to assess the degree of associ ation between the LDI calculated in the three different ways. This analysis determined the degree to which each form of the LDI contains redundant information. Correlations were run using Minitab. Relationship between the LDI and ecosystem condition Simple linear regression analysis was used to explore the relationship between the LDI and indicators of ecosystem condition. Simple linear re gression also allowed the determination of the grain size at which the LDI was more strongly related to measures of ecosystem condition. Regression analysis has often been used in lands cape ecology studies to quantify the explanatory power of landscape variables at different spatia l scales (Pearson 1993; Pearson and Turner 1995; 75

PAGE 76

Tufford et al. 1998; Gergel et al. 1999, Houl ahan et al. 2006). Regression results were graphically represented using scalograms by plotting the coeffi cients of determination (r2) against the LDI measured at increasing scales. Re sidual plots allowed to visually determining for inequality of variances (Dytham 1999; Minitab 2003). Water chemistry variables were log10 transformed when required. All regr essions were run using Minitab. Landscape Pattern Metrics Description and selection The process followed to reduce the 17 landscape pattern metrics to a smaller set of uncorrelated variables that can be used as independent measures to assess the relationship between land use and ecosystem condition was sim ilar to that used by Cifaldi et al. (2004) and Kearns et al. (2005). Using Kearns et al. (2005) rationales, firs t, descriptive statistics (mean, standard deviation, minimum and maximum values) were used to characterize metrics and to determine if any metrics presented values that wo uld suggest error in metr ic calculation. Metrics with high standard deviation (i.e ., greater than the mean) were c onsidered indicative of possible errors in metric calculation, and metrics with ve ry low standard deviation were considered as unable to discriminate among landscapes. Minimu m and maximum values of metrics helped to determine whether calculated values were outside the range of values in accordance with definitions of the metrics. Any of these condi tions would result in the removal of metrics. Second, the remaining metrics were tested fo r redundancy using Pearsons product-moment correlation. In deciding which metrics should be removed from further analysis, the criteria for data reduction set by Riitters et al. (1995) for P earsons correlations grea ter than 0.90 were used. Before correlation analysis could be performed, metrics were tested for normality using the Anderson-Darling normality test. Me trics with a p-value of less than 0.05 were considered to be unlikely normally distributed (Dytham 1999). After testing for normality, metrics that were not 76

PAGE 77

normally distributed were transformed (square root, logarithmic, and arcsine square-root transformations). Principal Components Analysis (PCA) was used to further reduce the nu mber of metrics to a set that explained most of the variation in the data. PCA is an ordination (multivariate) technique that graphically summarizes complex relationships among variables, and reduces the number of variables to a set of compound axes that represent most of the information contained in the original set of variables (ter Braak 1995; McCune and Grace 2002). The resulting axes can then be used as independent variables in ANOVA or regression analysis (Johnson and Gage 1997). PCA has frequently been used to reduce th e number of landscape patte rn variables to a set of uncorrelated metrics based on a set of orthog onal (uncorrelated) axes (Riitters et al. 1995; Cain et al. 1997; Johnson et al. 1997; Griffith et al. 2000; Cifaldi et al. 2004; Kearns et al. 2005). Variables considered for further analysis were dete rmined using the general rule that an axis with eigenvalues of less than one is considered not significant (Riitters et al 1995; Kearns et al. 2005). PCA was run using PC-ORD version 4.41 for Windows (McCune and Mefford 1999). Influence of landscape pattern on ecosystem condition Multiple regression analysis was performed using the water chemistry and biological indicators of ecosystem condition as dependent variables and the PCAs resulting axes as independent variables. Multiple regressions determined the best prediction of a dependent variable by using more than one independent va riable simultaneously. The prediction is defined as a linear equation (Sokal a nd Rohlf 1981; Dytham 1999). To assess the adequacy of the regression models, residuals were examined. Residual analysis allows testing for the normal distribution of the residuals, an assumption of regression anal ysis, and establishes that the relationship between the variables is linear (Dytham 1999; Minitab 2003). Variance Inflation Factors (VIF) were calcul ated to confirm the linear independen ce of the predictor variables. The 77

PAGE 78

VIF is the factor by which the standardized unexplained variance is inflated due to multicollinearity among the predictor variable s (Sokal and Rohlf 1981; Minitab 2003). The greater the correlation among the predictor variables, the larger the VIF. As a general rule, a VIF > 5 is signal of poor regression coefficients estimation (Montgomery a nd Peck 1982; Berk 2004). Multiple regression models and significance tests were estimated using Minitab. The LDI, Landscape Pattern, and Ecosystem Condition To test the effect of landscape pattern on the relationship be tween the LDI and indicators of ecosystem condition, multiple regression analysis was used. Multiple regression analysis is an extension of simple regression and allows using more than one predictor variable to estimate values of response variable s (Sokal and Rohlf 1981; Dyth am 1999). This technique was considered useful because it helped to investig ate how the LDI and the la ndscape pattern metrics performed together as predictor variables. To assess the adequacy of the regression models, residuals and VIFs were examined. All tests were run using Minitab. 78

PAGE 79

Table 2-1. Surrounding land use and date of sa mpling for 118 isolated forested wetlands in Florida (source: Reiss 2004). Site Code* Sample Date Surrounding La nd Use** Latitude (DD) Longitude (DD) SA1 6/5/01 Cattle & Crops 26.45639 -81.62487 SA2 6/6/01 Citrus 26.65347 -81.49650 SA3 6/27/01 Cattle 26.87612 -80.21323 SA4 7/30/01 Row Crops 26.26321 -81.41238 SA5 7/31/01 Cattle & Crops 26.28800 -81.22514 SA6 9/5/01 Cattle 26.75042 -81.35194 SA7 7/31/02 Woodland 26.69303 -81.64284 SA8 7/31/02 County Park 26.72535 -81.65383 SA9 8/1/02 Cattle 26.69032 -81.58596 SR1 6/28/01 County Park 26.95371 -80.18218 SR2 7/3/01 State Park 27.00919 -80.14564 SR3 7/24/01 State Reserve 26.86817 -80.41380 SR4 8/1/01 National Park 25.98619 -81.24261 SR5 8/21/01 State Preserve 26.10439 -81.34650 SR6 9/18/01 NWR 26.39318 -80.24319 SR7 7/15/02 County Park 26.73059 -80.25719 SR8 7/17/02 County Airport 26.86345 -80.23689 SR9 7/24/02 County park 26.72182 -80.25839 SU1 6/6/01 Residential & Golf 26.58418 -81.82126 SU2 6/29/01 School Campus 26.70849 -80.20811 SU3 7/4/01 Residential 26.82646 -80.15216 SU4 8/22/01 Residential 26.32488 -81.77125 SU5 8/23/01 Industrial 26.38015 -81.79557 SU6 9/30/01 Industrial 26.31050 -81.78732 SU7 7/16/02 Comercial 26.65032 -80.21265 SU8 7/16/02 Com. & Residential 26.73417 -80.11917 SU9 7/23/02 Residential 26.62327 -80.20051 SU10 7/30/02 Roads & Canals 26.77605 -81.35576 CA1 5/23/01 Row Crops 29.48414 -81.44358 CA2 5/30/01 Cattle 28.04368 -81.03569 CA3 6/7/01 Pullet Farm 28.24835 -82.09155 CA4 6/21/01 Cattle 27.81016 -80.53548 CA5 7/10/01 Cattle 28.06925 -81.42633 CA6 7/23/01 Citrus 27.53684 -80.64084 CA7 7/3/02 Silviculture & Cattle 28.47059 -82.11763 CA8 7/19/02 Dairy farm 28.14343 -82.22698 CA9 7/24/02 Citrus 27.43523 -80.64827 CR1 5/30/01 Conservation Tract 28.03325 -81.01988 CR2 6/14/01 Conservation Tract 28.08000 -81.40000 CR3 6/20/01 State Park 27.84150 -80.56868 CR4 8/10/01 WMD 28.39459 -81.97103 79

PAGE 80

Table 2-1. Continued. Site Code* Sample Date Surrounding La nd Use** Latitude (DD) Longitude (DD) CR5 8/13/01 State Park 28.77954 -81.39938 CR6 8/15/01 State Forest 27.78911 -81.46244 CR7 5/30/02 City Owned 28.11667 -82.34595 CR8 7/2/02 State Forest 28.48059 -82.00019 CR9 7/11/02 State Preserve 28.50238 -80.98923 CR10 10/9/02 State Park 27.22257 -82.29842 CR11 10/9/02 State Park 27.23897 -82.34602 CU1 5/31/01 University Campus 28.60519 -81.19494 CU2 6/15/01 Residential 28.45969 -81.26527 CU3 7/16/01 Commercial 28.07078 -82.50734 CU4 8/14/01 Roadside 28.46612 -81.27323 CU5 9/11/01 Roadside 28.32963 -81.53056 CU6 9/12/01 Golf Course 28.31473 -81.54881 CU7 5/30/02 City Owned 28.10109 -82.38816 CU8 7/1/02 Industrial 29.47656 -81.26731 CU9 7/8/02 Commercial 28.42875 -81.47282 CU10 8/7/02 Park 27.88246 -82.27471 CU11 8/8/02 Park 29.56134 -81.21028 NA1 5/21/01 Cattle 29.74165 -82.26923 NA2 6/4/01 Cattle 29.94500 -81.72512 NA3 6/19/01 Silviculture 30.30300 -82.41403 NA4 7/20/01 Row Crops 29.79279 -81.50204 NA5 7/27/01 Cattle 29.59070 -82.72842 NA6 7/31/01 Silv., Cat., Row Crops 30.02212 -83.09392 NA7 5/22/02 Row Crops 29.79495 -82.41924 NA8 5/21/02 Silviculture 29.67920 -81.73589 NA9 6/10/02 Silviculture 29.47003 -83.09615 NA10 7/12/02 Silviculture 29.81470 -81.83806 NA11 7/24/02 Cattle 29.01557 -82.38937 NA12 7/26/02 Cattle & Crops 29.80023 -82.41417 NR1 5/26/01 University Land 29.76662 -82.20515 NR2 6/18/01 City Park 29.66101 -82.27516 NR3 7/10/01 State Forest 29.13109 -82.62290 NR4 7/11/01 WMD 30.47229 -81.49923 NR5 8/6/01 Military 30.01672 -82.01836 NR6 8/21/01 State Park 28.63427 -82.57047 NR7 5/28/02 State Park 30.16982 -81.93639 NR8 8/5/02 State Park 30.17871 -81.93942 NR9 8/29/02 State Forest 29.28337 -82.61729 NU1 5/22/01 Road Side 29.73240 -82.38854 NU2 6/11/01 Residential & Golf 28.73751 -82.54058 80

PAGE 81

Table 2-1. Continued. Site Code* Sample Date Surrounding La nd Use** Latitude (DD) Longitude (DD) NU3 6/26/01 Residential 29.72551 -82.35376 NU4 6/27/01 Residential 30.17120 -82.66393 NU5 6/28/01 Residential 30.23957 -81.52941 NU6 8/1/01 Residential & Instit. 30.20305 -81.76350 NU7 5/15/02 Residential & Comm. 29.67127 -82.32561 NU8 6/3/02 Residential & Golf 30.11208 -81.62379 NU9 6/12/02 Industrial 30.20991 -82.64868 NU10 7/29/02 Residential & Instit. 30.40546 -81.72289 PA1 5/24/01 Cattle 30.46537 -82.70119 PA2 5/29/01 Cattle 30.50303 -83.12786 PA3 7/3/01 Crops & Turf Grass 30.77000 -87.14000 PA4 7/2/01 Row Crops 30.97707 -87.49655 PA5 8/8/01 Cattle 30.61916 -85.74185 PA6 8/9/01 Cattle 30.78991 -85.88736 PA7 6/5/02 Cattle 30.58314 -83.72990 PA8 8/8/02 Silviculture 29.95437 -84.59852 PA9 8/13/02 Row Crops 30.78313 -84.95956 PA10 8/14/02 Silviculture 30.83094 -86.96921 PR1 6/15/01 National Forest 29.95488 -84.99321 PR2 7/3/01 WMD 30.47204 -87.07998 PR3 7/4/01 Military 30.42565 -86.75117 PR4 8/9/01 State Forest 30.40339 -85.88346 PR5 8/10/01 State Forest 30.35190 -86.17112 PR6 8/18/01 National Forest 30.26229 -84.82198 PR7 6/4/02 Conservation Tract 30.67379 -84.22337 PR8 8/7/02 NWR 30.04041 -84.44025 PU1 6/14/01 Residential 30.44130 -84.32875 PU2 7/5/01 Residential 30.41486 -86.79694 PU3 8/17/01 Residential & Comm. 30.21191 -85.64720 PU4 8/17/01 Residential Park 30.78597 -85.68024 PU5 9/28/01 Commercial& Silv. 30.76760 -85.68549 PU6 9/29/01 Commercial 30.19020 -85.77982 PU7 6/18/02 Resid. & Orchard 30.45497 -87.32888 PU8 6/19/02 Industrial & Silv. 30.19108 -85.68608 PU9 6/20/02 Residential 29.93822 -85.39423 PU10 7/25/02 Instituti onal 30.48783 -84.27853 *Site Codes correspond to the ecoregion (Lane 2000), land use category, and sample order: S = south, C =central, N = north and P = panhandle; R = reference, A = agriculture and U = urban. **Surrounding Land Use abbreviations: NWR = Nati onal Wildlife Refuge; WMD = Water Management District; Resid. = Residential; Cat. = Cattle; Comm. = Commercial; Instit. = Institutional; Silv. = Silviculture. 81

PAGE 82

Table 2-2. STORET sampling station numbers, drainage basins, Hydrologic Unit Codes (HUC), and bioregions for 69 study streams in Florida. STORET station # Drainage Basin HUC Bioregion 19010042 Calkins Creek 03070204 Northeast 19010099 Pigeon Creek 03070204 Northeast 19020027 Alligator Creek 03070205 Northeast 20010454 Juniper Creek 03080101 Peninsula 20010455 Blackwater Creek 03080101 Peninsula 20020004 Little Orange Creek 03080102 Peninsula 20020012 Oklawaha River 03080102 Peninsula 20020317 Silver River 03080102 Peninsula 20020404 Orange Creek 03080102 Peninsula 20020424 Oklawaha River 03080102 Peninsula 20030263 Rowell Creek 03080103 Northeast 20030264 Sal Taylor Creek 03080103 Northeast 20030265 Sal Taylor Creek 03080103 Northeast 20030340 Rowell Creek 03080103 Northeast 20030341 Yellow Water Creek 03080103 Northeast 20030342 Yellow Water Creek 03080103 Northeast 20030419 Black Creek 03080103 Northeast 20030437 North Fork Black Creek 03080103 Northeast 20030549 Yellow Water Creek 03080103 Northeast 20030550 Rowell Creek 03080103 Northeast 21010018 South Falling Creek 03110201 Northeast 21010032 Hamilton Rocky Creek 03110201 Northeast 22020010 Quincy Creek 03120003 Panhandle West 22020049 Mule Creek 03120003 Panhandle West 22020062 Oklawaha Creek 03120003 Panhandle West 22020077 Unnamed Branch 03120001 Panhandle West 22020093 Quincy Creek 03120003 Panhandle West 22030062 McBride Slough 03120001 Panhandle East 22030064 Central Drainage Ditch 03120001 Panhandle East 23010464 Withlacooche River 03100208 Peninsula 24010002 Manatee River Ab Dam 03100202 Peninsula 24020134 Fishhawk Creek 03100204 Peninsula 24030013 Hillsborough River 03100205 Peninsula 24030044 Hillsborough River 03100205 Peninsula 24030142 Hillsborough River 03100205 Peninsula 82

PAGE 83

Table 2-2. Continued. STORET station # Drainage Basin HUC Bioregion 25020014 Oak Creek 03100101 Peninsula 25020111 Horse Creek 03100101 Peninsula 26010029 Arbuckle Creek 03090101 Peninsula 26010430 Parker Bay Drain 03080101 Peninsula 26010593 Fisheating Creek 03090103 Peninsula 26010972 Reedy Creek 03090101 Peninsula 26011019 Livingston Creek 03090101 Peninsula 26011020 Lake Weohyakapka 03090101 Peninsula 28010223 Jonathan Dickinson 3090202 Peninsula 28010224 South Indian River 03090202 Peninsula 28010232 North St. Lucie 03090202 Peninsula 28010239 Tidal St. Lucie 03090202 Peninsula 28010608 Tidal St. Lucie 03090202 Peninsula 28020147 West Caloosahatchee 03090205 Peninsula 28020148 Tidal Caloosahatchee 03090205 Peninsula 28020221 Telegraph Swamp 03090205 Peninsula 28020232 Tidal Caloosahatchee 03090205 Peninsula 28020233 Tidal Caloosahatchee 03090205 Peninsula 28020234 Estero Bay 03090204 Everglades 31010050 Crooked Creek 03130011 Panhandle West 31010051 Sweetwater Creek 03130011 Panhandle West 31020037 Bridge Creek 03130012 Panhandle West 31020038 Waddells Mill Creek 03130012 Panhandle West 31020040 Ten Mile Creek 03130012 Panhandle West 32010021 Alaqua Creek 03140102 Panhandle West 32020030 Camp Branch 03140203 Panhandle West 32020063 Little Crooked Creek 03140203 Panhandle West 32030023 Ecofina Creek 03140101 Panhandle West 32030024 So Fk Little Bear Creek 03140101 Panhandle West 33010054 McDavid Creek 03140106 Panhandle West 33010065 Rest Area Run 03140106 Panhandle West 33010068 Eleven Mile Creek 03140106 Panhandle West 33040014 Big Horse Creek 03140103 Panhandle West 33040015 Pine Log Creek 03140103 Panhandle West 83

PAGE 84

Table 2-3. STORET station numbers, lake name, Hydrologic Unit Codes (HUC), and ecoregions for 54 study lakes in Florida. STORET station # Lake name HUC Ecoregion 20010048 Lake Orienta 3080101 Central Florida Ridges and Uplands 20010110 Sawgrass Lake 3080101 Eastern Florida Flatwoods 20010222 Lake Kathryn 3080101 Central Florida Ridges and Uplands 20010299 Lake Kilarney 3080101 Central Florida Ridges and Uplands 20010311 Lake Underhill 3090101 Central Florida Ridges and Uplands 20010334 Lake Fairview 3080101 Central Florida Ridges and Uplands 20010336 Lake Ivanhoe 3080101 Central Florida Ridges and Uplands 20010337 Lake Minnehaha 3080101 Central Florida Ridges and Uplands 20020014 Hammond Lake 3080102 Central Florida Ridges and Uplands 20020015 Dixie Lake 3080102 Central Florida Ridges and Uplands 20020062 Lake Jumper 3080102 Central Florida Ridges and Uplands 20020064 South Twin Lake 3080102 Central Florida Ridges and Uplands 20020065 Lake Gibson 3080102 Central Florida Ridges and Uplands 20020066 Lake Umatilla 3080102 Central Florida Ridges and Uplands 20030417 Georges Lake 3080103 Eastern Florida Flatwoods 20030438 Lake Johnson 3080103 Central Florida Ridges and Uplands 23010434 Lake Rousseau West 3100208 Gulf Coast Flatwoods 23010435 Lake Rousseau East 3100208 Gulf Coast Flatwoods 25010079 Lake Webb 3100103 Southwestern Florida Flatwoods 25020552 Sunshine Lake 3100101 Southwestern Florida Flatwoods 25020554 Lake Zappa 3100101 Southwestern Florida Flatwoods 26010032 Alligator Lake 3090101 Eastern Florida Flatwoods 26010037 Lake Lizzie 3090101 Eastern Florida Flatwoods 26010039 Trout Lake 3090101 Eastern Florida Flatwoods 26010040 Brick Lake 3090101 Eastern Florida Flatwoods 26010105 Lake Porter 3090101 Central Florida Ridges and Uplands 26010116 Fish lake 3090101 Eastern Florida Flatwoods 26010303 Lake Persimmon 3090101 Central Florida Ridges and Uplands 26010304 Lake Clay 3090101 Central Florida Ridges and Uplands 26010325 Lake Adelaide 3090101 Central Florida Ridges and Uplands 26010326 Lake Little Bonnet 3090101 Central Florida Ridges and Uplands 26010327 Lake Trout 3090101 Central Florida Ridges and Uplands 26010331 Lake Huntley 3090101 Central Florida Ridges and Uplands 26010526 Lake Jackson 3090101 Central Florida Ridges and Uplands 26010528 Lake Jackson 3090101 Central Florida Ridges and Uplands 84

PAGE 85

Table 2-3. Continued. STORET station # Lake name HUC Ecoregion 26010531 Lake Jackson 3090101 Central Florida Ridges and Uplands 26010556 Lake Sebring 3090101 Central Florida Ridges and Uplands 26010585 Lake Carrie 3090101 Central Florida Ridges and Uplands 26010591 Dinner Lake 3090101 Central Florida Ridges and Uplands 26010605 Lake Viola 3090101 Central Florida Ridges and Uplands 26010644 Lake Josephine East 3090101 Central Florida Ridges and Uplands 26010645 Lake JosephineMid 3090101 Central Florida Ridges and Uplands 26010646 Lake JosephineWest 3090101 Central Florida Ridges and Uplands 26010647 Lake Rachard 3090101 Central Florida Ridges and Uplands 26010648 Lake Denton 3090101 Central Florida Ridges and Uplands 28020242 Crystal Lake 3100103 Southwestern Florida Flatwoods 28030068 Lake Avalon 3090204 Big Cypress 32010038 Sand Hammock Pond 3140203 Dougherty/Marianna Plains 32020113 Juniper Lake 3140203 Dougherty/Marianna Plains 32030081 Martin Lake 3140101 Gulf Coast Flatwoods 33010064 Crescent Lake 3140107 Gulf Coast Flatwoods 33020097 Lake Stone 3140305 Southern Pine Plains and Hills 33020098 Cotton Lake 3140305 Southern Pine Plains and Hills 33030057 Bear Lake 3140104 Southern Pine Plains and Hills 85

PAGE 86

Table 2-4. Summary information of the land use data used (source: Florida Department of Environmental Protection, www.dep.state.fl.us/gis ). Northwest Florida Water Management District Date of Data 1995 Data Layer Name Land Use North West Florida Description Northwest Florida Water Ma nagement District Land Use, Cover, and Forms Classification System Type Polygon Scale 1:24,000 Source Material National Aerial Photogr aphy Program color-infrared imaging Scale of Source Material 1:40,000 Date of Source Material 1994/1995 Suwannee River Water Management District Date of Data 1995 Data Layer Name sr_landuse95 Description Florida Land Use, Cove r, and Forms Classification System Type Polygon Scale 1:40,000 Source Material National Aerial Photogr aphy Program color-infrared imaging Scale of Source Material 1:40,000 Date of Source Material 1994/1995 St. Johns River Water Management District Date of Data 1995 Data Layer Name landcover95 Description FLUCC Land Use / Land Cover Type Polygon Scale 1:40,000 Source Material National Aerial Photogr aphy Program color-infrared imaging Scale of Source Material 1:40,000 Date of Source Material 1993/1994/1995 Date of Data 2000 Description Land cover and land use in the St. Johns River Water Management District Type Polygon Scale 1:12,000 Source Material National Aerial Photogr aphy Program color-infrared imaging Scale of Source Material 1:12,000 and 1:24,000 Date of Source Material 1999/2000 Southwest Florida Water Management District Date of Data 1995 Data Layer Name Southwest Florida Wate r Management District Land Use / Land Cover 1994-1995 Description FLUCC Land Use / Land Cover Type Polygon Scale 1:24,000 Date of Source Material 1994/1995 86

PAGE 87

Table 2-4. Continued. Southwest Florida Water Management District Date of Data 1999 Data Layer Name SWFWMD 1999 Land Use Description SWFWMD 1999 la nd use/cover features categorized according to the Florida Land Use, Cover, a nd Forms Classification System Type Polygon Scale 1:12,000 Source Material 1995 land use updated via 1999 DOQQs Scale of Source Material 1:12,000 Date of Source Material 1999/2000 Date of Data 1995 Data Layer Name sf_landuse95 Description South Florida Water Mana gement District Land use/cover 19941995 Type Polygon Scale 1:24,000 Source Material National Aerial Photogr aphy Program color-infrared imaging Scale of Source Material 1:40,000 Date of Source Material 1994/1995/1996 87

PAGE 88

Table 2-5. Land uses and definitions. Land Use Definition Natural Land / Open Water Open water, upland, or wetland with low manipulations. Pine Plantation Land devoted to the growth of mostly pine trees with different stocking densities. Rangeland Native rangeland and woodland pasture with presence of livestock. Pasture Areas where the natural vegetati on has been altered by drainage, irrigation, etc., for the grazing of domestic animals. Does not include livestock. Low Intensity Pasture Areas where the natural vegetati on has been altered by drainage, irrigation, etc., for the grazing of domestic animals with a density of less than 1.0 animal/ha. High Intensity Pasture Areas where the natu ral vegetation has been altered by drainage, irrigation, etc., for the grazing of domestic animals with a density of more than 1.0 animals/ha. Tree Crops Areas devoted to the production of tree crops such as citrus groves, fruit orchards, vine yards, and other groves. Row crops Areas devoted to the pro duction of all types of vegetables usually grown in rows, whether producing or not. High Intensity Agriculture Dair y farms and large-scale cattle feed lots, chicken farms, and hog farms. Low Intensity Open Space / Recreational Areas of natural vegetation in ci ties maintained as nature parks, and undeveloped land that may be occupied by low impacted natural vegetation in an agricultural or urban landscape. Medium Intensity Open Space/Recreational Areas with grassy lawns in urban landscapes including recreational land such as playgr ounds, ball fields, and swimming beaches. Also applies to land that has been cleared and prepared for construction and/or developmen t, dirt roads, barren land, and open areas surrounding paved roads and power lines. Includes human-created water bodies (ret ention ponds, canals, reservoirs, etc). High-intensity Open Space / Recreational Applies to stadiums not associated with institutions such as schools and universities, golf c ourses, and racetracks (horse, dog, car). Low-density Single Family Residential Areas that are predominantly residential units with a density less than 5 units/ha. Medium Intensity Single Family Residential Areas that are predominantly residential units with a density between 5 and 12 units/ha. High Intensity Single Family Residential Areas that are predominantly residential units with a density of more than 12 units/ha. Low-intensity Multi-family Residential Areas that are predominantly multi-family residential units such as condominiums and apartment buildings up to 2 stories. High-intensity Multi-family Residential Areas that are predominantly multi-family residential units such as condominiums and apartment buildings with 3 or more stories. 88

PAGE 89

Table 2-5. Continued. Land Use Definition Low Intensity Commercial Commercial strip. High-intensity Commercial Comm ercial mall with associated storage buildings and parking lots, hotels, convention cente rs, and theme parks. Institutional Schools, universitie s, religious, military, medical and professional facilities, and government buildings. Industrial Land uses include manufacturing, assembly or processing of materials/products and associat ed buildings and grounds. Also includes extractive areas and mining operations, water supply plants, and solid waste disposal. Low Intensity Transportation Paved road w ith no more than 2 lanes, and railroads. High-intensity Transportation Paved road with more than 2 lanes, airports, railroad terminals, b us and truck terminals, port facilit ies, and auto parking facilities when not directly related to other land uses. Low-intensity Central Business District Central business districts with an average of 2 stories. High-intensity Central Business District Central business districts with an average of more than 2 stories. 89

PAGE 90

Table 2-6. Intensity of human development clas sification and non-renewable and purchased areal empower density for selected land uses. Land Use Intensity Category Non-Renewable & Purchased (LI=Low Intensity; MI=Middle Inte nsity; Areal Empower Density Notes HI=High Intensity) (E14 sej/ha/yr) 1 Natural Land / Open Water 0.00 2 LI-Open Space / Recreational 2.77 3 Rangeland 3.20 4 Pine Plantation 5.10 5 Pasture (improved) 19.53 6 LI-Pasture 36.90 7 HI-Pasture 51.52 8 Tree Crops 65.39 9 MI-Open Space / Recreational 67.35 10 Row crops 117.11 11 HI-Agriculture 195.00 12 LI-Single Family Residential 1077.00 13 HI-Open Space / Recreational 1230.00 14 MI-Single Family Residential 2461.50 15 LI-Transportation 3080.00 16 HI-Single Family Residential 3729.50 17 LI-Commercial 3758.00 18 Institutional 4042.20 19 HI-Transportation 5020.00 20 Industrial 5210.60 21 LI-Multifamily Residential 7391.50 22 HI-Commercial 12661.00 23 HI-Multifamily Residential 12825.00 24 LI-Central Business District 16150.30 25 HI-Central Business District 29401.30 Notes 1 Non-renewable and purchased empower density for natural systems. 0.00 2 Average of empower densities of 1, 3 and 4 Assumed 3 Based on 18 acres/animal 0.14 cow/ ha/yr After Tanner & Bradley (1992) Empower density to support one steer: 22.85 E14 sej/J Brandt-Williams (2002) = Natural Land (1) + 3.2 E 14 sej/ha/yr 3.20 E14 sej/J 4 After Doherty (1995) 5 After Brandt-Williams (2002) 6 Based 3.25 acres/animal 0. 76 cow/ha/yr After Pate (1999) Empower density to support one steer: 22.85 E14 sej/J Brandt-Williams (2002) = Improved Pasture (5) + 17.37 E14 sej/ha/yr 36.90 E14 sej/J 7 Based on 1.4 brood cow/ha/yr (1.75 acres/animal) After Pate (1999) Empower density to support one steer: 22.85 E14 sej/J Brandt-Williams (2002) = Improved Pasture (5) + 31.99 E14 sej/ha/yr 51.52 E14 sej/J 8 Brandt-Williams (2002) 9 Average of empower densities of 5, 8, and 10. Assumed after Odum (1994) 90

PAGE 91

Table 2-6. Continued. Notes 10 Average of empower densities for 12 row crops. After Brandt-Williams (2002) 11 Based on 0.79 dairy cows /acre/yr 1.95 cow/ha/yr After USDA (2002) Empower density to support one cow: 100.00 E14 sej/J Brandt-Williams (2002) 195.00 E14 sej/J 12 Based on Parker (1998) and Brown (1980) 13 Based on the emergy evaluation for a golf course After Behrend (2000) 14 Based on Parker (1998) and Brown (1980). Includes mobile home medium density. 15 Based on Parker (1998) 16 Based on Brown (1980). Includes mobile home high density. 17 Based on Parker (1998) and Brown (1980) 18 After Brown (1980) 19 Based on Parker (1998) 20 Based on Parker (1998) and Brown (1980) 21 Based on Parker (1998) and Brown (1980) 22 Based on Parker (1998) and Brown (1980) 23 After Brown (1980) 24 After Brown (1980) 25 After Brown (1980) 91

PAGE 92

Table 2-7. Description of landscape pattern metrics selected for this study. Acronyms, descriptions, and units accordi ng to McGarigal et al. (2002). Landscape Pattern Metric Acronym Description Units References Class level metrics Percent agriculture PLAND_Ag PLAND equals the percentage of the landscape comprised of agricultural land. % Allan et al. 1997; Jones et al. 2001; Cifaldi et al. 2004; Kearns et al. 2005 Percent urban PLAND_Urb PLAND equals the percentage of the landscape comprised of urban land. % Allan et al. 1997; Jones et al. 2001; Cifaldi et al. 2004; Kearns et al. 2005 Percent forests PLAND_For PL AND equals the percentage of the landscape comprised of forested land. % Allan et al. 1997; Jones et al. 2001; Cifaldi et al. 2004; Kearns et al. 2005 Percent wetlands PLAND_Wet PL AND equals the percentage of the landscape comprised of wetland land. % Allan et al. 1997; Jones et al. 2001; Cifaldi et al. 2004; Kearns et al. 2005 Landscape level metrics Patch richness PR PR equals the number of different patch types present within the landscape boundary. None (PR 1, without limit) Griffith et al. 2002 Patch richness density PRD PRD equals the number of different patch types present within the landscape boundary divided by total landscape area (m2), multiplied by 10,000 and 100 (to convert to 100 hectares). #/100 ha (PRD > 0, without limit) Uuemaa et al. 2005 Patch density PD Number of patches of the landscape divided by total landscape area (m2), multiplied by 10,000 and 100 (to convert to 100 hectares). #patches/100ha (PD > 0, constrained by cell size) Johnson et al. 1997; Griffith et al. 2002; Cifaldi et al. 2004; Kearns et al. 2005; Uuemaa et al. 2005 Mean patch size AREA_MN MN equals the sum, across all patches in the landscape, of the corresponding patch metric values1, divided by the total number of patches. Ha (AREA > 0, without limit) Miller et al. 1997 Patch size coefficient of variation AREA_CV CV equals the standard deviation divided by the mean, multiplied by 100 to convert to a percentage, for the corresponding patch metric1. % Cifaldi et al. 2004 Edge density ED ED equals the sum of the lengths (m) of all edge segments in the landscape, divided by the total landscape area (m2), multiplied by 10,000 (to convert to hectares). m/ha (ED 0, without limit) Cifaldi et al. 2004; Uuemaa et al. 2005 92

PAGE 93

Table 2-7. Continued. Landscape Pattern Metric Acronym Description Units References Interspersion and juxtaposition index IJI IJI measures the extent to which patch types are interspersed (not necessarily dispersed); higher values result from landscapes in which the patch types are well interspersed (i.e., equally adjacent to each other), whereas lower values characterize landscapes in which the patch types are poorly interspersed. % Griffith et al. 2002, Cifaldi et al. 2004; Kearns et al. 2005 Mean shape index SHAPE_MN MN equals the sum, across all patches in the landscape, of the corresponding patch metric values2, divided by the total number of patches. None (SHAPE 1, without limit) Kearns et al. 2005; Uuemaa et al. 2005 Shannons diversity index SHDI SHDI equals minus the sum, across all patch types, of the proportional abundance of each patch type multiplied by that proportion. None (SHDI 0, without limit) Miller et al. 1997; Uuemaa et al. 2005 Dominance (calculated as its complement evenness, where evenness = 1dominance) SHEI Shannons evenness index is expressed such that an even distribution of area among patch types results in maximum evenness. None (0 SHEI 1) USEPA 1994; Hunsaker and Levine 1995 Contangion CONTAG Contagion measures both patch type interspersion (i.e., the intermixing of units of different patch types) as well as patch dispersion (i.e., the spatial distribution of a patch type). % USEPA 1994; Hunsaker and Levine 1995; Miller et al. 1997; Griffith et al. 2002; Kearns et al. 2005; Uuemaa et al. 2005 Fractal dimension (calculated as the mean patch fractal dimension) FRAC_MN FRAC_MN equals the sum of 2 times the logarithm of patch perimeter (m) divided by the logarithm of patch area of patch (m2) for each patch in the landscape, divided by the number of patches. An indicator of shape complexity. None (1 FRAC_MN 2) USEPA 1994; Liu and Cameron 2001 Mean Euclidian nearest neighbor distance ENN_MN MN equals the sum, across all patches in the landscape, of the corresponding patch metric values3, divided by the total number of patches. Meters (ENN > 0, without limit.) Uuemaa et al. 2005 1 AREA equals the area (m2) of the patch, divided by 10,000 (to convert to hectares). 2 SHAPE equals patch perimeter (given in numbe r of cell surfaces) divided by the minimum perimeter (given in number of cell surfaces) possi ble for a maximally compact patch (in a square raster format) of the corresponding patch area. 3 ENN equals the distance (m) to the nearest neighboring patch of the same type, based on shortest edge-to-edge distance. 93

PAGE 94

Figure 2-1. Study site locati on by ecoregions of 118 isolated forested wetlands in Florida. 94

PAGE 95

# S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S Sample Streams (n=69) SCI BioregionsEverglades Northeast Panhandle Peninsula050 100 Kilometers N Figure 2-2. Study site location by ecoregions of 69 streams in Florida. Streams locations indicated as STORET sampling stations. 95

PAGE 96

# S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S # S Sample Lakes (n=54) Florida Ecoregions0 40 80 Kilometers N Figure 2-3. Study site locati on of 54 lakes in Florida. 96

PAGE 97

Land use bounderies 0100 meters NStudy wetland (CU3) Figure 2-4. Major land use patches delineated from an aerial photo within a 200 meter buffer area surrounding a study wetland (CU3) within an urban landscape. 97

PAGE 98

Residential Highway Recreational Study wetland (CU3) Commercial Upland forest Paved Road (2 lanes) Lake Open Space Land Use Type 5020 3750 3729.5 3080 1230 67.35 2.77 NR Empower Density (E+14 sej/ha/yr) Study wetland (CU3) 0100 meters N (a) (b) Figure 2-5. Landscape surrounding an isolated fore sted wetland in Florida showing (a) general land use categories and (b) land use base d on the intensity of human activities measured as non-renewable empower density. 98

PAGE 99

(a) (b) Figure 2-6. Land use types for a Florida stream watershed based on (a) the FLUCCS (refer to Appendix B for codes descriptions) and (b) the intensity of human activities measured as non-renewable empower density. 99

PAGE 100

#20-meter Buffer 200-meter Buffer 100-meter Buffer Study Wetland #20-meter Buffer Landscape Matrix 020 40Meters N Figure 2-7. Spatial scales of analysis for the is olated forested wetlands. Equidistant areas of 20, 100, and 200 meters were used to determin e the contributing effect of land use on isolated forested wetlands. 100

PAGE 101

(a) (b) (c) Figure 2-8. Spatial scales of analysis for stream s: (a) total drainage area or total watershed, (b) equidistant area of influence of 400 meters around the stream, and (c) equidistant area of influence of 100 meters around the stream. DTM MASK (Basin Boundary) PRE DIGITIZED STREAM MODELED STREAM NETWORK STORET SAMPLING STATION Outlet Definition MODELED DRAINAGE BASIN Figure 2-9. Flow chart showing the main steps followed for the de lineation of drainage basins using US EPAs BASINS 3.0 software. 101

PAGE 102

(c) (b) N #N 012Kilometers 26 30 31 35 36 39 40 44 45 48 49 53 54 57 58 62 63 66 67 71 72 75 76 80 81 84 85 89 90 94(a) Figure 2-10. Graphic representation of the step s followed in the delineation of the area draining to a water quality sampling site (STORET 22020093). (a) DTM of the main drainage basin (units = meters); (b) basins mask w ith a pre-digitized str eam network; and (c) modeled drainage boundary with modeled st ream network and ou tlet (water quality sampling station). 102

PAGE 103

(a) (b) (c) Figure 2-11. Spatial scales of analysis for lakes: (a) total drainage area or total watershed, (b) equidistant area of influence of 400 meters ar ound the lake, and (c) equidistant area of influence of 100 meters around the lake. 103

PAGE 104

CHAPTER 3 RESULTS This chapter presents the results of the landscape analysis performed on isolated forested wetlands, streams, and lakes in Florida to study the relationship between human land use intensity (LDI), landscape pattern metrics, and indicators of ecosystem condition water quality variables for these freshwater systems. The LDI was correlated to water quality variables and biological indicators for each freshwater system at multiple landscape scales. The effects of spatial scale on the LDI and its predictive power are detailed herein. For landscape pattern metrics, the selection process of a set of metrics that best de scribes the surrounding landscape of the systems studied is shown. The metrics selected were also correlated to water quality variables and biological indicators at multiple landscape scales. Finally, the effect of landscape pattern on the relationship between the LD I and indicators of ecosyst em condition and water quality variables that resulted after using the LDI together with the pattern metrics in multiple regression analysis is presented. Land Use/Land Cover Composition of the Freshwater Systems Isolated Forested Wetlands Each isolated forested wetland buffer area was a priori classified as reference, agricultural, or urban through inspection of aerial photography (DOQQs ). The result of the a priori classification was that 37 isolated forested wetla nds had buffer areas classified as reference, 40 had agricultural buffer areas, and 41 had urban buffer areas. Table 3-1 summarizes the land use characteristics of the a priori buffer area classes. The land us e/land cover composition for buffer areas for each of the 118 isolated forested wetlands is presented in Appendix I. Reference isolated forested wetlands were primarily su rrounded by forests, which accounted for almost 73% of the land use/land cover within a 200meter buffer distance from the study sites. 104

PAGE 105

Approximately 25% of the 200-meter buffer areas were covered with wetlands other than the site under investigation. Other land use/land covers accounted for less than 3% of the surrounding landscapes. Agricultural isolated forested wetlands were imbedded within landscapes surrounded mostly by agricultural crops, pasture lands, and rangelands. These land uses accounted for approximately 66% of the land surrounding the isolat ed forested wetlands within this category. Forests represented 21% of the land use/land cove r; and wetlands were present in 28 sites, covering about 8% of the landsca pes. Sixty-four percent (64%) of the landscapes surrounding the urban sites were occupied by urban land uses. Fore sted areas and wetlands were less prevalent in urban sites than in agricultural sites, occupyi ng only 18% and 4% of the surrounding landscapes, respectively. Lands devoted to transportation uses were more representative of urban lands, accounting for about 10% of the bu ffer areas in urban landscapes. The non-renewable and purchased areal empower density was calcula ted within the 200meter buffer surrounding each wetland. Table 3-2 presents summary statistics for the areal empower density for the a priori defined buffer areas. Although the surrounding landscapes of the reference sites consisted mostly of natura l lands, a high variability in the areal empower density for this wetland category was obser ved. A maximum value of 227.7 E+14 sej/ha/yr suggests that there were sites among the referenc e isolated forested we tlands that included highintensity land uses in their surrounding landscap es. Summary statistics for the agricultural isolated forested wetlands show that on average these sites were w ithin the range of what may be a characteristic value for a landscape that consists mostly of agricultural lands. The minimum non-renewable and purchased areal empower dens ity value corresponds to lands planted with pine trees that have an areal empower density similar to that of natural lands. The maximum nonrenewable and purchased areal empower densit y value reflects the development intensity 105

PAGE 106

characteristics of land uses with high energy dema nds such as dairy farms or chicken farms. On average, urban isolated forested wetlands had non-renewable and purchased areal empower densities that were two orders of magnitude hi gher than the reference a nd agricultural isolated forested wetlands. The range of non-renewable a nd purchased areal empower values at the lower end describes sites with transiti onal lands that are less developed, while the higher range of areal empower values describes sites with intense us e of energy such as commercial malls and highrise residential areas. Streams and Lakes Table 3-3 presents summary statistics for th e drainage area size and land use composition for 69 streams and 54 lakes. Information on the total area and the land use/land cover composition for each drainage area is included in Appendices I and J for streams and lakes, respectively. The drainage areas for streams varied considerably in size with the largest drainage basin three orders of magnitude larger th an the smallest one. The land use/land cover composition of the drainage basins was very di verse and represented a wide range of land use types with landscapes varying from almost comple tely forested (site S56; see Appendix J) to completely urbanized (site S29). For lakes, drainage basins also varied consid erably in size, however, on average these were much smaller than the streams watersheds. The smallest lake dr ainage basin was just over 8 hectares and the largest was more than 3,500 hectares. Although the land use/land cover composition of the drainage basins was diverse, urban lands were most common in the landscapes surrounding the sample lakes. The land use composition of the drainage basins varied from completely covered with natural lands (site L53; see Appendix K) to entirely urban (site L20). 106

PAGE 107

Mean values the standard deviation of the non-renewable and purchased areal empower density for the drainage areas of the sample stre ams and lakes calculated in three different ways are shown in Table 3-4. There was a large vari ability among the streams in terms of their development intensity. Results showed that none of the watersheds were completely occupied by natural lands or waters, and that at least some of the resources used were either non-renewable or purchased. At the other end of the development gr adient, the resource use within the watershed with the highest non-renewable and the purchased areal empower density value was characteristic of a highl y urbanized landscape. There was also a large variab ility among the lakes based on their development intensity. Note that the average areal density was higher for the lakes than for the streams. Lakes watersheds were more urbanized than those of the streams. The average resource use among all lake watersheds was equivalent to a landscape with a resource use comparable to low-intensity family residential lands. Description of the Landscape Development Intensity Index The LDI was calculated for varying grain sizes and spatial extents of the sample isolated forested wetlands, streams, and lakes, to test the effect of spatial grain size and spatial extent on which the LDI is calculated. Scale Dependence: Grain Size Isolated forested wetlands The LDI was calculated for eight different grain sizes (5 x 5, 10 x 10, 20 x 20, 30 x 30, 40 x 40, 50 x 50, 60 x 60, and 70 x 70 meters) and usi ng the 200-meter buffer area surrounding the sample of isolated forested wetlands as the sp atial extent for analysis. LDI scores for 15 representative isolated forested wetlands selected from the tota l set of 118 wetlands are shown in Figure 3-1. The LDI scores for the total wetland sample are presente d in Appendix L. In general, 107

PAGE 108

changing the grain size had a modest effect on the LDI scores. The highest variability was among intermediate LDI scores that corresponded to the most heterogeneous landscapes or landscapes composed of a combination of natural, agricultural, and/or urban land uses. The site that presented the highest variability among the total sample of isolated forested wetlands (n = 118) was SU10 (mean = 7.39, SD = 3.49 for the LDI-ILD). The surrounding landscape for this site was in transition between natural lands and urban (i.e., housing) that were developed with streets already constructed but without housing units. LDI scores varied from high values to lower values as the LDI was calculated with incr easing grain size (refer to Appendix L). It was also observed that the LDI scores based on area only were consistently higher than the LDI scores based on a decrease in distance from the study isolated forested wetlands. Individual landscape scalograms presented in Figure 3-2 provide detailed information on the effect of changing grain size on the LDI fo r six representative isolated forested wetland buffers with low, medium, and high LDI values. Tr ends in the variation of the LDI scores for small wetland buffers suggest that the effect of grain size on the LDI may be important for landscapes with intermediate LDI scores. For th ese landscapes, the effect of cell aggregation may remove small patches with high development in tensity that when presen t at fine grain sizes may have a strong influence on the landscape LDI sc ore. This situation can be appreciated in Figure 3-3, where the effect of changing grain size for the six wetland buffers is shown. Of particular interest is Figure 3-3(c) wetland PA1, where the di sappearance of the low-intensity transportation patch beyond a grai n size of 40 x 40 meters had a st rong effect on the LDI score of the wetland buffer when the LDI was calcu lated based on area only (LDI-PLU). Streams For streams, the LDI was calculated for six different grain sizes (20 x 20, 50 x 50, 80 x 80, 110 x 110, 140 x 140, and 170 x 170 meters) and using the total drainage area as the spatial 108

PAGE 109

extent for analysis. Figure 3-4 shows LDI scores for 15 representative streams watersheds selected from the total set of 69 streams. The LDI scores for the total samp le of drainage basins for streams are presented in Appendix M. Changi ng the grain size had a small effect on the LDI scores. This can be attributed to the fact that for stream waters heds urban patches were not rare. The disappearance of some urban patches with ce ll aggregation had less effect on the LDI score for streams than the effect observed for the isolated forested wetlands, since for the former some urban lands tended to remain with changes in sc ale allowing for less variability in LDI scores. The highest variability was among intermediate LDI scores that correspond to the most heterogeneous landscapes or landscapes composed of a combination of natural, agricultural, and/or urban land uses. The site that presented the highest variability am ong the total sample of streams was site S42 (mean = 9.30, SD = 2.59, for the LDI-PLU; see Appendix M). Site S26 in Figure 3-4 was unique among the sample streams since it had a high LDI score and a relatively high standard deviation. A bigger spread in the LDI values was reported for sites with intermediate LDI scores that were calculated based on the proportion occupied by each land use type. Of interest in Figure 3-4 is that there appears to be a difference in the trend among the scores for the three LDI forms. While the LD I-PLU and the LDI-ILD scores tend to increase linearly with no large departures for the different sites, the LDI-ISD presents more variability suggesting the effect of the flood plain (n atural buffers) on some of the sites. Individual drainage basin scalograms presente d in Figure 3-5 provide detailed information on the effect of changing grain size on the LDI fo r six of the streams watersheds that were considered. Trends in the varia tion of the LDI scores for each LDI form suggest that, despite some small variations, the emergy-based index va ries very little with changes in grain size between 20 x 20and 170 x 170-meters for stream watersheds. 109

PAGE 110

Scale Dependence: Spatial Extent Isolated forested wetlands To test how the LDI varies with changes in spatial extent, th e grain size was held constant at 5 x 5 meters and the LDI in its three forms was calculated for buffer areas of 20, 100, and 200 meters around the study isolated forested wetlan ds. The Kruskal-Wallis test initially implied that there were no significant differences among the LD I scores for the three extents (n = 118; LDIPLU, H = 4.80, p = 0.091; LDI-ILD, H = 5.35, p = 0.069; LDI-ISD, H = 5.02, p = 0.081). However, a comparison of LDI values with changes in extent based on a priori classes, as shown in Figure 3-6, suggested that there were differe nces among LDI values with increasing extent, especially among reference sites and urban sites. Th ese differences can be attributed to the fact that there was an increase in the number of landuse classes included in the wetland buffers with increasing scale. LDI values tended to be higher w ith changes in extent as more developed lands were included in the buffers. The Kruskal-Wallis test confirmed the significance of the differences in LDI scores with increasing extent among reference isolated forested wetlands (n = 38; LDI-PLU, H = 13.79, p = 0.001; LDI-ILD, H = 13.42, p = 0.001; LDI-ISD, H = 12.33, p = 0.002) and among urban isolated forested wetl ands (n = 41; LDI-PLU, H = 8.88, p = 0.012; LDIILD, H = 12.32, p = 0.002; LDI-ISD, H = 12.87, p = 0.002). Table 3-5 and Figure 3-7 show the extent of the association between each form of the LDI calculated for the three different extents. In a ll cases there was a very strong positive correlation among the three forms of the LDI; the weakes t association was found between the LDI-PLU calculated for the 200-meter buffer area and the LDI-ISD calculated for the 20-meter buffer area (r = 0.86, p < 0.001). The matrix plot of the rela tionship between pairs of LDI scores for different extents confirmed that LD I scores tended to be higher for the largest extent that, as was suggested previously, presented more developed la nds than the smallest extent considered (20110

PAGE 111

meter buffer). Note that differences among the 100-meter buffer and the 200-meter buffer were minimal, suggesting that it might not be nece ssary to consider the land beyond 100 meters in order to account for most of the lands that mi ght influence small isolated forested wetland systems. Streams The grain size for streams was held constant at 20 x 20 meters and the LDI in its three forms was calculated for buffer areas of 100 and 400 meters from the sample streams and for the entire drainage basin. The Kruskal-Wallis test initially implied that there were no significant differences among the LDI-PLU scores for th e three extents (n = 69; H = 5.14, p = 0.077). However, differences among the LDI-ILD scores with changes in scale were significantly different (n = 69; H = 6.65, p = 0.036), as were the differences among the LDI-ISD scores (n = 69; H = 6.97, p = 0.031), suggesting that the LDI is scale-dependent. When LDI values were disaggregated into low (n = 17), intermediate (n = 35), and high (n = 17) development intensity classes as shown in Figure 3-8, differences am ong LDI values with increasing extent were suggested for the LDI-PLU at the low and interm ediate ranges of values. These results also suggest that LDI values tend to increase as more developed lands are included in larger landscape areas. The Kruskal-Wallis test confir med the significance of the differences among streams within the low LDI range of values (n =69; LDI-PLU, H = 11.97, p = 0.003; LDI-ILD, H = 13.48, p = 0.001; LDI-ISD, H = 12.45, p = 0.002) and among streams within the intermediate LDI range of values (LDI-PLU, H = 13.90, p = 0.001; LDI-ILD, H = 17.72, p < 0.001; LDI-ISD, H = 18.72, p < 0.001). The fact that there were no significant differences among sites with high LDI values could be attributed to the large extent overlap among these sites, most of which tended to have small watersheds. 111

PAGE 112

Correlations between each form of the LDI calculated for th e three different extents are shown in Table 3-6 and Figure 3-9. In all case s there was a very strong positive correlation among all forms of the emergy-based index; the weakest association was found between the LDI-PLU calculated for the watershed scale a nd the LDI-ISD calculated for the 100-meter buffer area scale (r = 0.84, p < 0.001). The sm all variability among pairs of LDI scores for landscapes surrounding streams observed in Figure 3-9 sugges ts that patterns of development remained relatively similar with changes in extent. Lakes The grain size for the lakes was also held cons tant at 20 x 20 meters and the LDI in its three forms was calculated for buffer areas of 100 and 400 meters from the sample lakes and for the entire drainage basin The Kruskal-Wallis test suggested that there were no significant differences among the LDI-PLU scores for the three extents (n = 54; H = 1.36, p = 0.507), and that there were no significant differences among the LDI-ILD scores (n = 54; H = 2.87, p = 0.238) or the LDI-ISD scores (n = 54; H = 5.23, p = 0.073). When LDI values were compared by disaggregating them into low (n = 13), intermed iate (n = 28), and hi gh (n = 13) development intensity classes as shown in Figure 3-10, statis tically significant differences (Kruskal-Wallis test) among LDI values with increasing extent we re only observed for the LDI-ILD at the low (H = 7.04, p = 0.03) and intermediate ranges of valu es (H =8.42, p = 0.015), and for the LDI-ISD at the low (H = 8.82, p = 0.012), intermediate (H =14.02, p = 0.001), and high ranges of values (H = 7.62, p = 0.02). Small differences among LDI groups fo r lake basins were due to the fact that the lands within the basins were more devel oped and no new land use t ypes that will result in significant changes in LDI scores we re added with increasing extent. Table 3-7 provides information on the correla tions between the three forms of the LDI calculated at three different spatial extents for the sample lakes. A strong positive relationship for 112

PAGE 113

all pairs of LDI forms was re ported. The weakest association was found between the LDI-PLU calculated for the watershed scale and the LDI-ISD calculated for the 100-meter buffer area scale (r = 0.78, p < 0.001). Figure 3-11 shows the relations hip between pairs of LDI scores for the different extents for the sample lakes. Note that in Figure 3-11 it can be observed that patterns of development of the lakes watersheds tended to be different with changes in extent with developed lands, becoming more common with increasing scale. Relationship between Land Use In tensity and Ecosystem Condition Isolated forested wetland condition Water quality: spatial extent. Simple linear regression was us ed to estimate the level of association between the LDI and fi ve water chemistry variables at different spatial extents while holding the grain size constant at 5 x 5-meters. All depe ndent variables were log10 transformed to satisfy requirements of regression analysis. Resu lts for the regressions are presented in Table 38. The highest variability in the DO was explai ned by the LDI-PLU at the 200-meter scale. The regression had a very poor fit (r2 = 0.07), but the overall re lationship was significant (F1,69 = 5.01, p = 0.028). Figure 3-12 shows the relationshi p between the LDI-PLU and the isolated forested wetlands DO (log10 transformed) at the 200-meter sp atial extent. Regression results among the three LDI forms were very small at all scales. The LDI-PLU explained slightly more of the va riability in the waters specific conductance (SC) at the 200-meter scale (r2 = 0.21, F1,31 = 7.98, p = 0.008) with an in crease in the strength of the association between the two va riables with increasing scale. Regression results for the LDIPLU were also slightly higher than for the other two forms of the LDI at all three scales. The relationship between the LDI-PLU and isol ated forested wetlands water SC (log10 transformed) at the 200-meter scale is shown in Figure 3-13. Significant results for TP resulted only when TP was regressed against the LDI-PLU with the strongest association established at the 20-meter 113

PAGE 114

scale. The regression had a very poor fit (r2 = 0.065), but the overall re lationship was significant (F1,73 = 5.07, p = 0.027). The relationship between th e LDI-PLU and isolated forested wetlands water TP (log10 transformed) at the 20-meter scale is shown in Figure 3-14. None of the regressions between the LDI and TN or betw een the LDI and the water turbidity were significant. The analysis of residuals suggested that the mode l for SC satisfied the requirements of regression. The regression model for DO showed some level of heteroscedasticity. For TP, the residuals were not normally distributed and unequal variances were observed. Water quality: grain size. The proportion of the variance in water chemistry variables explained by the LDI measured at different grain sizes was tested by calculating regression coefficients (r2) for each scale independently and holding the spatial extent constant at the 200meter scale (see Figure 3-15). The significance of the regression results is shown in Table 3-9. More of the variability in the water DO was ex plained by the LDI-PLU almost equally at all grain sizes. The relationship between DO and th e LDI-PLU was shown previously in Figure 312. For the water SC, more of the variability was explained by the LDI-PLU at the 20 x 20meter grain size (r2 = 0.208, F1,31 = 8.14, p = 0.008). Minimal differe nces were observed among all of the scales considered for the LDI-PLU. The relationship between the specific conductance and the LDI-PLU was shown previously in Figure 3-13. The LDI was not significantly associated to the concentration of TN, TP, or the isolated fore sted wetlands water turbidity at any scale. Biological indicators: spatial extent Simple linear regression allowed estimating the degree of association between the LDI and th e WCI for macrophytes, macroinvertebrates, and diatoms at different spatial extents and holdi ng the grain size constant at 5 x 5 meters. 114

PAGE 115

Regression results are presented in Table 3-10. The LDI-PLU explai ned more of the variability of the macrophyte WCI at the 20-meter buffer scale (r2 = 0.30, F1,116 = 48.78, p < 0.001), with the strength of the rela tionship decreasing with increasing scale for the LDI-PLU and the LDIILD. Figure 3-16 shows the relationship between the LDI-PLU and the macrophyte WCI at the 20-meter scale. The plot between the LDI-PLU and the macrophyte WCI shows a decrease in WCI scores with increasing development intensity. However, two distinct sets of data points grouped at approximate LDI values of 5 to10 and 15 to 23 for isolated forested wetlands with low scores in the macrophyte WCI are apparent. For the macroinvertebrate WCI more of the variance was explained by the LDI-PLU at the 20-meter scale (r2 = 0.24, F1,77 = 24.17, p < 0.001), and a decreas e in the strength of the association with increasing scale was also reported for the LDI-PLU and the LDI-ILD. The relationship between the LDI-PLU and macroinvertebrate WCI at the 20-meter scale is shown in Figure 3-17. The plot shows differences between the WCI scores for isolated forested wetlands with surrounding landscapes with low LDI-PLU values and wetland landscapes with intermediate and high LDI-PLU values. For the diatom WCI the str ongest relationship was found with LDI-PLU at the 100-meter scale (r2 = 0.24, F1,48 = 15.32, p < 0.001). For diatoms, the pr oportion of the va riance explained by the LDI was higher at broader scales for all forms of the LDI. Figure 3-18 shows the relationships between the LDI-PLU and diatom WCI at the 100-meter scale. The plot between the LDI-PLU and the diatom WCI showed higher WCI scores for sites with low LDI-PLU than for sites with middle and high LDI-PLU. However, among the isolated forested wetlands with middle and high LDI-PLU two distinct sets of da ta points grouped at approximate LDI values of 5 to10 and at 15 to 25 were observed. 115

PAGE 116

The analysis of residuals (nor mal probability plots, residual s versus variables plots, and plots of residuals versus the f itted values) suggested that the relationship between the LDI and the macrophyte WCI was non-linear and that the da ta were not normally distributed. For the macroinvertebrate WCI and the diatom WCI residual plots sugge sted a linear relationship between the variables and a norma l distribution of the data. Howe ver, the regression model for the diatom WCI showed some level of heteroscedasticity. Biological indicators: grain size The variability of the WCI for macrophytes, macroinvertebrates, and diatoms explained by the LDI measured at different grain sizes while holding the spatial extent consta nt at the 200-meter scale, was tested by calculating regression coefficients (r2) for each scale independently (see Figure 3-19 and Table 3-11). The proportion of the variance in the macrophytes WCI explained by the LDI decreased as the grain size was increased, with more of the variability explaine d by the LDI at the 5 x 5-meter grain size. The LDI-PLU accounted for more of the variability in the WCI than the other two forms of the index. Figure 3-20 shows the relationship between th e macrophyte WCI and the LDI at the most significant scale (r2 = 0.243, F1,116 = 37.15, p < 0.001). The macrophyte WCI decreased with increasing development intensity. More of the va riability that was explained by the LDI for the macroinvertebrate WCI occurred at the 50 x 50-meter cell resolution. At this grain size the LDIILD explained most of the proportion of the vari ance. However, the difference with the other two forms of the LDI was very small. The relations hip between the macroinvertebrate WCI and the LDI at the most significant sc ale is shown in Figure 3-21 (r2 = 0.198, F1,77 = 18.95, p < 0.001). The macroinvertebrate WCI decreased with increasing development intensity. For the diatom WCI, more of the vari ability was explained by the LDI-PLU at the 30 x 30-meter cell resolution, although differences in the 5 x 5-meter to 30 x 30-meter cell size range were minimal. Beyond 116

PAGE 117

the 30 x 30-meter cell size, the proportion of th e total variance explained by the LDI tended to decrease with increasing grain size. The relations hip between the diatom WCI and the LDI at the most significant scale is shown in Figure 3-22 (r2 = 0.231, F1,48 = 14.44, p < 0.001). Stream condition Water quality: spatial extent Simple linear regression was used to asses the relationship between the LDI and five water chemistry variab les and the WQI at diffe rent spatial extents, holding the grain size constant at 20 x 20 meters. When needed, the dependent variables were log-transformed to satisfy regres sion analysis requirements. Regr ession results are presented in Table 3-12. More of the variabil ity in the concentration of DO was explained equally by the LDI-ILD and the LDI-ISD at the watershed scale. Regressions had a fair fit (r2 = 0.41) and were highly significant (LDI-ILD: F1,35 = 24.03, p < 0.001; LDI-ISD: F1,35 = 24.22, p < 0.001, respectively). Figure 3-23 shows the relationships between the LDI-ISD and the DO at the watershed scale. The concentration of DO decreased with increasing development intensity. The LDI-PLU explained more of the va riability in the concentration of NO3-N and TN at the watershed scale (NO3-N: r2 = 0.14, F1,44 = 7.13, p = 0.011; TN: r2 = 0.17, F1,45 = 9.21, p = 0.004). For NO3-N, the strength of the association with the LDI decreased with decreasing scale and was lowest with the LDI-ISD. The relationships between the LDI-PLU and the NO3-N at the watershed scale are shown in Figure 3-24. An increase in NO3-N with increasing development intensity is suggested, despite the weak association between th e variables. For TN, regression results showed that for the LDI-PLU were only s lightly higher than for the other two forms of the LDI at all three scales. Figure 3-25 presents the relationships between the LDI-PLU and TN at the watershed scale. An increase in TN with increasing development intensity can be observed despite the scatter in the data. 117

PAGE 118

None of the regressions between the LDI and TP were statisti cally significant except for the LDI-ISD at the 100-meter scale, where a poor fit was reported (r2 = 0.08, F1,45 = 4.09, p = 0.049). Significant results for the relationship between the WQI a nd the three forms of the LDI were observed at all scales. The strongest a ssociation was found when the WQI was regressed against the LDI-ISD at the watershed scale (r2 = 0.33; F1,35 = 17.38, p < 0.001). The strength of the association between the two variables tended to slightly decrease with increasing scale. The relationship between the LDI-ISD and the WQI for streams is shown in Figure 3-26. The plot shows what appears to be a linear relationship between the two variables with the WQI increasing with increasing development intensit y. However, a relatively high scatter at low development intensity values was apparent. None of the regressions be tween the LDI and the streams water turbidity were significant. The anal ysis of residuals for each significant regression model showed that all models seemed adequate. Some level of heteroscedasticity was observed for the TP and the WQI models. Water quality: grain size The proportion of the variance in five water chemistry variables and the WQI explained by the LDI measured at differe nt grain size was tested by calculating regression coefficients (r2) for each scale independently and holding the spatial extent constant at the watershed scale (Figure 3-27). The significance of the regression results are shown in Table 3-13. More of the variability in the water DO was explained by the LDI-ISD at the 170 x 170-meter grain size with the LDI-ILD explaining almost the same amount of the variation in the concentration of DO. Figure 3-28 shows the relationship between DO and the LDI-ISD at the most statistically significant scale (r2 = 0.452, F1,35 = 28.86, p < 0.001). The plot shows that the concentration of DO decrea sed with increasing development intensity. 118

PAGE 119

More of the variance in the concentration of NO3-N was explained by the LDI-PLU at the 170 x 170-meter grain size. The differences between the regressions for the LDI-PLU and the regressions for the LDI-ISD were relatively large, with the LDI-ISD explaining almost 50% less of the total variance explained by the LDI-PLU at all scales measured. Figure 3-29 shows the relationship between th e concentration of NO3-N and the LDI-PLU at the most statistically significant scale (r2 = 0.147, F1,44 = 7.61, p = 0.008). The concentration of NO3-N tended to increase with increasing development intensity. For TN, most of the variati on in its concentration was also explained by the LDI-PLU at the 170 x 170-meter grain size. The three forms of the explained almost the same amount of variability in the concentration of TN at finer grain sizes. Some variation in the predictive power of the LD I was seen at the broader grain sizes where the LDI-ILD and the LDI-ISD explained less of the tota l variance in the concentration of TN than the LDI-PLU. Figure 3-30 shows the relationship be tween the concentration of TN and the LDIPLU at the most statistically significant scale (r2 = 0.232, F1,45 = 13.163, p = 0.001). The concentration of TN increased with increasing development intensity. The LDI-ISD at the 170 x 170-meter grain size was the best predictor of the variability in the WQI scores. The LDI calculated at coarser grain sizes explained slightly more of the variance than when it was calculated at finer grain sizes. At all scales, the LDI-ISD explained approximately 10% and 5% more of the total variance in the WQI than the LDI-ILD and the LDI-PLU, respectively. Figure 3-31 shows the relationship between the WQI and the LDI-ISD at the most statistically significant scale (r2 = 0.364, F1,35 = 20.04, p < 0.001). Higher scores for the WCI were reported at higher valu es of development intensity. The relationship between the LDI and TP and between the LDI and the streams water turbidity were not statistically significant at any grain size. 119

PAGE 120

Stream condition index: spatial extent The degree of associa tion between the LDI and the SCI at different spatial extents was estimate d using simple linear regression and holding the grain size constant at 20 x 20 me ters. The regression results are presented in Table 3-14. Among the three forms of the LDI, the LDI-ISD explained more of the variability for the SCI_1 scores at the 100-meter buffer scale (r2 = 0.27, F1,66 = 24.11, p < 0.001; ) as well as the variability for the SCI_2 scores at the same spatial extent (r2 = 0.26, F1,67 = 23.71, p < 0.001). However, differences with the proportion of the variance in the SCI explai ned at other scales were small, particularly at the watershed scale. At all scales, the LDI-ISD expl ained more of the variability for the SCI than the other forms of the LDI. Figure 3-32 shows the relationships between the LDI-ISD and the SCI_1 at the 100-meter scale. The plot between the LDI-ISD and the SCI _1 shows a decrease in the SCI scores with increasing development intensity, especially beyond an LDI score of 10. The plot of residuals versus the fitted values showed that the relatio nship between the SCI_1 and the LDI-ISD was not linear. Additionally, the residuals were not normally distributed. The relationships between the LDI-ISD and the SCI_2 at the 100-meter scale are shown in Figure 3-33. The plot of residuals versus the fitted values also suggested a non-li near relationship between the SCI_2 and the LDIISD. However, the residuals were normally distributed. Stream condition index: grain size The proportion of the variance in the SCI explained by the LDI measured at different grain sizes wa s tested by calculating regression coefficients (r2) for each scale independently and holding the spatia l extent constant at the watershed scale (see Figure 3-34 and Table 3-15). Changes in the grain size had minimal effect on the amount of the variability in the SCI_1 explained by the LDI. Among all forms of the LDI, the LDI-ISD calculated at the 20 x 20-meter grain size expl ained more of the variance in the SC_1 (r2 = 0.228, 120

PAGE 121

F1,66 = 19.54, p < 0.001). The relationship between these two variables is presented in Figure 3-35. Similarly, changes in the grain size had minima l effect on the amount of the variability in the SCI_2 explained by the LDI. Once more, th e LDI-ISD calculated at the 20 x 20-meter grain size explained more of the variance in the SCI_2 (r2 = 0.252, F1,67 = 22.53, p < 0.001). Figure 3-36 shows the relationship betw een the SCI_2 and the LDI-ISD. Lake condition Simple linear regression models indicated that the LDI was not significantly related at any of the lake condition variables tested when relationships were analyzed for different landscape extents (Tables 3-16). The LDI only explained close to 4% of the variance in the concentration of ammonia-N (F1,52 = 1.99, p = 0.164, for the LDI-ILD at the watershed scale). For the concentration of NO3/NO2-N, only 2% of the variance was accounted for by the LDI (F1.52 = 1.14, p = 0.29, for the LDI-ISD at the 100-meter scale). The concentrations of TKN and TN were also poorly correlated with the LDI. Only less than 2% of the variation in the c oncentrations of both variables were explained, with the highest regression reported for the LDI-ISD at the 400meter scale in both cases (F1.51 = 0.60, p = 0.441; and F1,51 = 0.81, p = 0.374; respectively). Similarly, TP was poorly associated with the LDI (r2 = 0.053, F1.51 = 2.85, p = 0.097, for the LDI-ILD at the watershed scale). The LCI was also poorly associated, with the LDI-PLU explaining only 1% of the variance in th e index scores at the watershed scale (F1,51 = 0.67, p = 0.417). Similarly, the LDI was not significantly related at any of the lake condi tion variables tested when different landscape grains were consider ed (Table 3-17). For the concentration of ammonia-N and among the different forms of th e LDI, the LDI-ILD only explained slightly more than 4% of the variance (F1,52 = 2.33, p = 0.133, for the LDI-ILD for the 40 x 40-m scale). For the concentration of NO3/NO2-N, less than 1% of the vari ance was explained by the LDI 121

PAGE 122

(F1.52 = 0.38, p = 0.538, for the LDI-ILD at the 20 x 20-meter scale). Approximately 2% of the variation in the concentration of TKN was accounted for by the LDI (F1.51 = 0.76, p = 0.338, for the 40 x 40-m scale). A similar result was reporte d for the association between the concentration of TN and the LDI (r2 = 0.022, F1.51 = 1.17, p = 0.285). For TP, the LDI-PLU accounted for more of the variation in the con centration of TP, explaining 5% at the 80 x 80-meter scale (F1.51 = 3.14, p = 0.082). Finally, only 2% of the variation in the LCI was explained by the LDI-PLU at the 40 x 40-meter scale (F1,51 = 1.08, p = 0.305). Landscape Pattern Metrics Isolated Forested Wetlands Metric selection: grain size Pattern metrics were calculated for landscapes surrounding isolated forested wetlands by exporting ArcGrids into Frag stats with 5 x 5-, 10 x 10-, 20 x 20-, and 30 x 30-meter cell resolutions. None of the metrics showed unusual values that would sugge st that any of these could not be considered for further analysis. De scriptive statistics on each metric calculated for the scales considered is provided in Appendi x N, Table N-1, as well as information on the transformations used to obtain a normal dist ribution in the metrics scores. The land use composition metrics (PLAND_Urb, PLAND_Ag, PLAND_For, and PLAND_Wet) showed high variability; however, this behavior was considered normal given the fact that zero scores were common among the metrics due to the a priori selection of the sample isolated forested wetlands. All metrics showed some level of variability (SD > 0) making them useful to discriminate between landscapes. However, the metric IJI wa s reported as undefined for one site (NA7) where the landscape measured had less than three patc h types. IJI is based on patch adjacencies and does not report a score for landscapes with less than three patches (McGarigal et al. 2002). In addition, transformation normality was not achieved in any case and a very significant statistic 122

PAGE 123

for the Anderson-Darling test was reported (p < 0.005). As a result, IJI was removed from further analysis. Correlation analysis to test for redundancy among the remain ing landscape pattern metrics and to reduce the number of metrics used in mu ltivariate statistics tests are shown in Table 3-18. The metrics PD and AREA_MN were highly correlated at the 5 x 5-meter (r = .00, p < 0.001), 10 x 10-meter (r = .00, p < 0.001), 20 x 20-meter (p = .97, p < 0.001), and 30 x 30-meter (p = .97, p < 0.001) cell sizes. At the landscape leve l these metrics are both dependent on the number of patches and the total landscape area, making them highly redundant (McGarigal et al. 2002). The metrics SHAPE_MN and FRAC_MN were also highly correlated at all the four scales considered (r = 0.92, p < 0.001; r = 0.94, p < 0.001; r = 0.95, p < 0.001; r = 0.94, p < 0.001, with increasing grain size, respectiv ely). Both SHAPE_MN and FRAC_MN are measurements of patch shape complexity and ar e based on perimeter-to-area ratio relationships (McGarigal et al. 2002). As a re sult, they may convey the same information about the complexity of patch forms in the landscape. Similarly, PR a nd PRD were highly correlated regardless of the grain size at which the metric s were calculated (r = 0.91, p < 0.001; r = 0.93, p < 0.001; r = 0.93, p < 0.001; r = 0.93, p <0.005, with increasing grain si ze, respectively). PR and PRD are highly redundant when calculated for landscapes that are similar in area and when the maximum number of patch classes is a constant (McGarigal and Mark s 1995). Pearsons correlations between the metrics CONTAG and SHEI were also very significant when calculated at the 5 x 5meter (r = .95, < 0.001) and 10 x 10-meter (r = .9003, p < 0.001) cell sizes. The strength of the relationship decreased slightly with increases in grain size (r = .84, p < 0.001; and r = 0.86, p < 0.001 at the 20 x 20-meter and 30 x 30-meter grain sizes, respectively). 123

PAGE 124

Following Riitters et al. (1995)s measure of Pearson correlations of |r| > 0.9 for the reduction of metrics, AREA_MN, FRAC_MN, and PR were eliminated from further analysis for all of the scales considered. Based on the same criterion SHEI was eliminated from the 5 x 5meter and the 10 x 10-meter grain size set of va riables. AREA_MN was chosen for elimination over PD since first-order statistics provide information of limite d value about the size of patches. More meaningful information about the variabil ity of landscapes based on patch sizes can be obtained from the metric AREA_CV (a metric not among the candidates for elimination), whose scores can be better explained when considered along with the results for PD (McGarigal and Marks 1995). FRAC_MN was eliminated because it may present problems of interpretation (McGarigal and Marks 1995). PR measures the number of patch types present in the landscape while PRD measures richness on a per area basis, allowing for comparisons among landscapes (McGarigal and Marks 1995). As a result, the latt er was preferred. SHEI was not considered for further analysis at the 5 x 5-meter and 10 x 10meter grain sizes since two other measures of patch diversity were already included (PRD and SHDI). Additionally, CONTAG was the only metric left that would aid in describing the is olated forested wetlands landscapes based on patch adjacencies; thus, it was preferred over SHEI. Principal components analysis of the landscape patte rn metrics showed varying results for each scale considered. Eigenvalues and the pr oportion of the variance explained by each component are presented in Table 3-19. For the 5 x 5-meter grain size the first five components explained close to 81% of the variation in the 12 landscape variables. The total percent of the variance explained by the same number of component s for the rest of the grain sizes was similar to the 5 x 5-meter dataset, with the 10 x 10-me ter and 20 x 20-meter scales being almost equal and the 30 x 30-meter scale slightly higher ( 82.4%). However, the number of significant 124

PAGE 125

components (eigenvalue > 1) was of six for th e 10 x 10-, 20 x 20-, and 30 x 30-meter scales. These components explained 89.21%, 89.22%, and 90.93 % of the variance in the landscape variables, respectively. The factor loadings for the first six eigenvect ors for each grain size are presented in Tables 3-20 through 3-23. Components were labeled to represent the metric that loaded highest with each axis for each grain size. For example, component 1 for the 5 x 5-meter grain size was labeled URB since the metric PLAND_Urb loaded highest with this component. Component 2 was labeled HETER since the metric PD loaded highest with this component and describes aspects of patch type heterogeneity in the la ndscape. The metric SHAPE_MN loaded highest with component 3, so it was labeled SHAPE. Com ponents 4 and 5 were labeled to describe that the percent of agricultural land (AG) and the percent of forest land (FOR) were most important in these axes, respectively. Fina lly, in component 6 the metric that quantifie s the nearest neighbor distance among patches of the same type (ENN_MN) loaded highest and was labeled DIST. Other metrics not represented in PC matr ix for the 5 x 5-meter grain size include those that describe the diversity of path types; when these metrics correlated highest with any of the components, they were labeled DIVERS (e.g., co mponent 1 at the 20 x 20-meter grain size). The proportion of urban land had the highest lo ading (negative) with the first component for both the 5 x 5-meter and 10 x 10-meter scales while the proportion of isolated forested wetlands had a positive but slightly smaller loading for both datasets as well. In addition, at both scales there was a negative correl ation between the metrics that describe the diversity of patches in the landscape (PRD and SHDI) and this com ponent. This distribution can be attributed to changes along a gradient of land uses rangi ng from highly urbanized wetland landscapes to landscapes with no urban lands present. For the 20 x 20-meter and 30 x 30-meter scales, the 125

PAGE 126

proportion of urban lands and wetla nds were not as important in the first component. Instead, the diversity metrics SHDI and SHEI presented the hi ghest loading (negative) with this axis. The third diversity metric, PRD, also was negatively correlated with this component. Additionally, the metric AREA_CV which measures the variability of patch size relative to the mean patch size in the landscape presented a positive correlat ion with this component for the 20 x 20-meter grain size. The second component was dominated by the pa tch density (PD) and edge density (ED) variables. These metrics were approximately equally important for all scales and had a positive correlation with this compone nt, except for the 30 x 30-meter grain size. The metric SHAPE_MN had the highest negati ve correlation with the third component at the 5 x 5-meter grain size. However, this metric showed a low loadi ng with this axis for the rest of the grain sizes as a consequence of the aggregation of pixels with increasing scale, which resulted in the simplification of patch forms. Th e metric CONTAG also showed a negative correlation with this component at the 5 x 5-meter grain size as well as for the 10 x 10-meter and 20 x 20-meter grain sizes where the loading was even higher. At the 30 x 30-meter grain size CONTAG also had a negative correlation with this ax is although the strength of the relationship was lower than for the other scales. Additionally, with increasi ng scale the class metrics PLAND_Urb and PLAND_Wet increased their correlation with this axis. The former had a negative correlation with this component while the la tter had a positive co rrelation with this component. In summary, this component can be described as a patch interspersion/dispersion grad ient with landscapes with a few large, contiguous patches at one end of the axis and landscapes with more dispersed patches at the other end of the axis with a separation of urban sites form the agricultural and reference sites. 126

PAGE 127

Metrics with the highest loadi ng with the fourth component for the 5 x 5-meter grain size included PLAND_Ag and CONTAG, with both metr ics presenting a negativ e correlation with this axis. For the 10 x 10-meter grain size, SHAPE_MN had the highest (positive) loading with this component. For the 20 x 20-meter and 30 x 30-meter grain sizes, PLAND_Ag again had the highest loading with this component; however, the trend of th e relationship with this axis was negative. When PLAND_Ag was dominant, PLAND_ For correlated inversely but moderately with this axis. The fifth component summarizes the spatial patterns in forest cover (PLAND_For) at the 5 x 5-meter and 10 x 10-meter grain sizes ; at the 20 x 20-meter and 30 x 30-meter grain sizes ENN-MN and SHAPE_MN were dominant, with PLAND_For decreasing in importance with increasing scale. Finally, the sixth component was highly correlated to PLAND_For at the 20 x 20and 30 x 30-meter grain sizes, while for the 10 x 10-meter grain size ENN-MN had a high positive correlation with this axis. Regression analysis Significant components that resu lted from the PCA of landscape pattern metrics were used as independent variables in multiple regression anal ysis to test for relationships with the WCI for macrophytes, macroinvertebrates, and diatoms, and with water chemistry variables. These regression results are presented in Table 3-24. Significant relationships were found only for TP among the water chemistry variable s considered. The strongest relationship was reported at the 30 x 30-meter grain size (R2 = 0.45, F6,25 = 5.18, p < 0.001) with very similar results for the 10 x10-meter (R2 = 0.44, F6,25 = 4.99, p = 0.002) and 20 x 20-meter (R2 = 0.43, F6,25 = 4.82, p = 0.002) scales. For the 30 x 30-meter grain size, components 4 and 5 were the only independent variables significantly related to TP (p = 0.005 and p = 0.001, respectively). For the 20 x 20meter grain size, the same compon ents were the only predictors si gnificantly associated to TP (p = 0.002 and p = 0.003, respectively). For the 10 x 10-me ter scale the p-values for the estimated 127

PAGE 128

coefficients of components 5 and 6 (p = 0.01 a nd p = 0.001, respectively) indicated that these were the only independent vari ables significantly related to TP. At the 5 x 5-meter scale the relationship between significant components resulting from the PCA of landscape pattern metrics and TP was not significant. The analysis of residuals (normal plots of residuals, residuals versus variables plots, and plots of residuals versus the fitted values) for the regression model at the 30 x 30-meter scale suggested that the relationship between the LDI and TP was fairly linear and that the residuals were normally distributed. Ho wever, the residuals presented some level of heteroscedasticity. Variance Inflation Factors (VIF) were reported as being < 5 for all components, which indicated a good regression coefficient estimation. Among the WCIs, significant relationships were found for the macrophytes, with the highest variance explained at the 10 x 10-meter grain size (R2 = 0.44, F6,44 = 7.49, p < 0.001). Differences between the four grain sizes were small: 5 x 5-meter (R2 = 0.42, F5,45 = 7.45, p < 0.001), 20 x 20-meter (R2 = 0.39, F6,44 = 6.41, p < 0.001), and 30 x 30-meter (R2 = 0.42, F6,44 = 6.97, p < 0.001). Components that were significan tly related (p < 0.05) to the macrophyte WCI varied between scales. At the 5 x 5-meter scal e components 1 and 4 were the only independent variables significantly related to the macrophyte WCI, with estimated coefficients with p-values of 0.023 and < 0.001, respectively. For the 10 x 10-meter scale components 4 and 5 were significantly associated with the macrophyte WC I, with p < 0.001 in both cases. Components 3 and 4 were also the only axes significantly related to the macrophyte WCI at the 20 x 20-meter scale, with p-values for the estimated coeffi cients of 0.015 and < 0.001, respectively. At the 30 x 30-meter level three components (2, 3, and 4) expl ained most of the variation in the macrophyte WCI (p = 0.034, p = 0.009, and p < 0.001, respectively). For the macroinvertebrate WCI, regressions were significant only at the 20 x 20-meter grain size (R2 = 0.26, F6,24 = 2.71, p = 128

PAGE 129

0.038). Component 1 was the only predictor signifi cantly associated with the response variable (p = 0.003). No significant rela tionships were found between the components from the PCA of landscape pattern metrics and the diatom WCI. The analysis of residuals for the macrophyte WCI regression model at the most statistically significant scale suggested that the relationship between the PCA components a nd the macrophyte WCI was linear a nd that the residuals were normally distributed. Additionally, variances were fairly homogeneous. The residual plots also suggested a linear relationship be tween the variables and a normal distribution of the data for the macroinvertebrate WCI. The regression model for the macroinvertebrate WCI also showed fairly homogeneous variances. For all cases th e VIFs were reported as being < 5. Streams Metric selection: grain size Pattern metrics were calculated for drainage basins by exporting ArcGrids into Fragstats with 20 x 20-, 50 x 50-, 80 x 80-, and 110 x 110 -meter cell resolutions. Descriptive statistics on each metric calculated for each scale considered, as well as information on the transformations used to obtain a normal distribution in the metric s scores are provided in Appendix N, Table N-2. The metric FRAC_MN was removed from further anal ysis since it was not able to differentiate between different landscapes. At all scales the standard devi ation for this metric was 0.01. Redundancy was tested for among the remaining landscape pattern metrics using Pearsons correlation analysis. Table 3-25 shows the correlation results of all pairs of metrics considered. The metrics PD and AREA_MN were highly correl ated at all the scales considered (r = -0.96, p < 0.001, for all four grain sizes). The metrics CO NTAG and SHEI were also highly correlated at all the four scales considered (r = -0.98, p < 0.001; r = -0.93, p < 0.001; r = -0.91, p < 0.001; r = 0.92, p < 0.001, with increasing grain size, respecti vely). Following Riitters et al. (1996)s measure of Pearson correlations of |r| > 0.9 for the reduction of me trics, AREA_MN and 129

PAGE 130

CONTAG were eliminated from fu rther analysis for all the s cales considered. AREA_MN was chosen over PD for elimination since first-orde r statistics provide information of limited value about the size of patches. More meaningful info rmation about the variabil ity of landscapes based on patch sizes can be obtained from the metr ic AREA_CV (a metric not among the candidates for elimination), whose scores can be better explai ned when considered together with the results for PD (McGarigal and Marks 1995). The excl usion of CONTAG was quite arbitrary. Since SHEI measures the evenness component of landscap e diversity, an aspect that is not captured by the PR or SHDI (the other two metrics that qua ntify diversity of the la ndscape level), it was preferred over CONTAG. Principal components analysis of the landscape patte rn metrics calculated for different scales showed similar results. Eigenvalues and the proportion of the variance explained by each component are presented in Table 3-26. The total percent of the variance explained by the six components was very similar among all grain sizes and explained between 90% and 92% of the variation in the 14 landscape variables. The results for the first four components with an eigenvalue of greater than one and worthy of interpretation were also very similar. These four components together explained approximately 80% of the variation in the landscape variables. The factor loadings for these four eigenvectors for each grain size are pr esented in Tables 3-27 through 3-30. Components were labeled to represent the metric that loaded highest with each axis for each grain size. For example, for the 20 x 20-meter grain size component 1 was labeled DIVERS1 since the metric SHEI (Shannon Evenness Index) loaded highest with this component. Component 2 was labeled DIVERS2 since the metric PR (patch richness) loaded highest with this component. The metric PLAND_Wet loaded highest with component 3, so it was labeled 130

PAGE 131

WET. In component 4 the metric that quantifie s the nearest neighbor distance among patches of the same type (ENN_MN) loaded highest and was labeled DIST. The same reasoning was used to label the components at other grain sizes. At all scales, the firs t component tended to be dominate d by the metrics that describe the diversity of patches in the landscape, particularly by SHEI which emphasizes the evenness component of diversity. The metric that quantif ies patch interspersion and juxtaposition (IJI) correlated fairly with axis 1, especially at the 20 x 20and 50 x 50 meter grain sizes. At the 80 x 80and 110 x 110-meter scales AREA_CV and ED were more important that IJI. The second component was also influenced by me trics that measure as pects relatively to the diversity of patches in the landscape. In this component the richness of patches was emphasized rather that their even distribution in space. The metrics PR and SHEI correlated highest with this component, par ticularly at the 80 x 80and 110 x 110-meter scales. The metric ENN_MN, which measures the distance of a patch type to its nearest ne ighboring patch of the same type, was also fairly correlated with this ax is at all grain sizes with its influence tending to decrease with increasing scale. The proportion of land use under forests (PLAND_For) was also an important variable in this axis particul arly at the 80 x 80a nd 110 x 110-meter scales. Component 3 was clearly summarized by a gradie nt of urban to wetland types. The metrics PLAND_Urb and PLAND_Wet correlated negatively w ith each other at all grain sizes and fairly with this component. Also, the im portance of spatial patterns in agricultural and forested land covers tended to increase with the increasing scale. However, a shift in the direction of the association between the metrics and axis 3 was noticed between the finer scales (20 x 20 and 50 x 50 meters) and the broader scales (80 x 80 and 110 x 110 meters). The metric with the highest loading with the forth component for the 20 x 20-meter grain size was ENN_MN which 131

PAGE 132

correlated negatively with this axis at this scale. This metric became less important with increasing scale while the metric SHAPE_MN tende d to load higher with increasing scale. This axis was also influenced by the metrics AREA_CV and PLAND_Ag. Metric selection: spatial extent In addition to the watershed scale, landscape pattern metrics were calculated for stream buffers of 100 and 400 meters, keeping the grai n size constant at 20 x 20 meters. Descriptive statistics on each metric calculated for 63 sample st reams for the three scales considered, as well as information on the transformations used to obtai n a normal distribution in the metrics scores is provided in Appendix N, Table N-3. The metric FRAC_MN had a standard deviation of 0.01 at all scales. As a result, this metric was remove d from further analysis since FRAC_MN was not able to differentiate between different landscapes. Pearsons correlations were used to test for the redundancy among the remaining metrics. The results of all pairs of metrics are shown in Table 3-31. The metrics PD and AREA_MN were highly correlated at all scales considered (r = -0.98, p < 0.001; r = -0.99, p < 0.001; r = -0.96, p < 0.001, with increasing extent respectively). The metrics C ONTAG and SHEI were also highly correlated at a ll the three scales (r = -0.93, p <0.001; r = -0.96, p < 0.001; r = -0.98, p < 0.001, with increasing extent, respectively). Followi ng Riitters et al. (1996)s measure of Pearson correlations of |r| > 0.9 for the reduction of metrics, AREA _MN and CONTAG were eliminated from further analysis for all the scales considered. The results of the principal components analys is of the landscape pattern metrics showed varying results with changes in spatial exte nt. Table 3-32 shows the eigenvalues and the proportion of the variance explained by each com ponent for each scale considered. The total percent of the variance explaine d by the six components was sim ilar among all spatial extents and explained approximately 88% to 91% of the va riation in the remaining landscape variables. 132

PAGE 133

The results for the first four components (eigenvalue > 1) were also fairly similar, explaining 79% of the variance in the landscape metrics at the 400-meter and watershed scales, and 75% at the 100-meter scale. The factor load ings for the first four eigenvectors for each extent are presented in Tables 3-33 through 3-35. Components we re labeled to represent the metric that loaded highest with e ach axis for each spatial extent. The first component was influenced by the metric s that describe the diversity of patches in the landscape, particularly by S HDI which loaded fairly and nega tively at all scales and with SHEI increasing its influence with increasi ng scale. The metric that quantifies patch interspersion and juxtapos ition (IJI) also loaded fair ly and negatively with this component at all scales. The metric PLAND_For also had some influence and correlated positively with this axis, tending to decrease with increasing scale. The second component was dominated by th e metrics PRD and AREA_CV at the 100meter scale. Both metrics loaded equally on this component and were inversely correlated with each other. The PD metric also had some influen ce on this component at this scale. The metrics PRD, AREA_CV, and PD also were important at the 400-meter scale; however, ED appeared as the metric that loaded highest and had a positive associa tion with axis 2 at this scale. ED also loaded highest and positively with axis 2 at the watershed scale; a change in the magnitude of the loadings was observed. Also at th e watershed scale, diversity metr ics related to the richness of patches (PR and PRD) also loaded fairly (n egatively) with this component. The metrics PLAND_For and PLAND_Ag also had some influen ce on this axis and correlated inversely with each other. Component 3 was influenced by the proportion of wetlands (PLA ND_Wet), particularly at the 100and 400-meter scale, although there was a ch ange in the magnitude of the loadings with 133

PAGE 134

changes in scale. At the finer sc ale there was also some influence of the ED and PD variables, which had a negative correlation wi th this component. The influence of these metrics on this axis decreased with increasing scale. The patch area coefficient of variation (AREA_CV) also correlated fairly especially at the 400-m and watershed scales. At the watershed scale the metric ENN_MN correlated highest with this axis. This component can be summarized as a gradient of wetland spatial patterns ranging from landscapes with high presence of wetlands to landscapes where land types other than wetlands were more common, with the presence of isolated patches of the same type at the watershed scale. The fourth component showed differences am ong the metrics that loaded highest with changes in scale. At the 100-meter extent the metrics ED, SHAPE_MN, and IJI correlated highest with axis 4; although IJ I had a positive association with the axis while the other two metrics presented a negative asso ciation with the axis For the 400-meter and watershed scales the proportion of agricultural lands (PLAND_Ag) and the proportion of wetlands (PLAND_Wet) had the strongest influence on the axis, with the metrics correlating inversel y with each other. A gradient of landscape ranging fr om agriculturalto wetlands-dominated landscapes seemed apparent at broader scales. Regression analysis: grain size Multiple regression analysis was used to explore the relationships between significant components that resulted from the PCA of landscape pattern metrics and water chemistry variables for streams and the SCI. These results are presented in Tabl e 3-36. A significant relationship was found at all scales when the la ndscape pattern variable s were related to DO, with the strongest association at the 110 x110-meter grain size (R2 = 0.22, F4,31 = 3.42, p = 0.020). Component 2 explained more of the variab ility in the dependent variable (estimated coefficient with p-value of 0.032). For NO3-N, significant but weak relationships were found 134

PAGE 135

only for the 50 x 50and 110 x 110-meter grain sizes. At these scales the amount of the variability that the landscape pa ttern variables were able to account for was the same (R2 = 0.13, F4,40 = 2.62, p = 0.049). There was a good association at all scales between th e landscape pattern variables and TN. The strongest relationship was found at the 20 x 20-meter grain size (R2 = 0.60, F4,41 = 18.14, p < 0.001), with components 1, 2, and 4 significantly related to TN (estimated coefficients with p-values of < 0.001, 0.002, and 0.007, respectively). A decrea se in the strength of the association between va riables was observed with in creasing scale. A significant relationship between the landscape pattern variables and TP was al so found for every grain size considered, with the strongest asso ciation at the 20 x 20-meter scale (R2 = 0.42, F4,41 = 8.96, p < 0.001). The p-value for the estimated coefficien t of component 4 was 0.001, indicating that it was the only component significantly related to TP. The strength of the association between variables decreased with increasing grain size. There were no significant relationships between the independent variables and the water turbidity and the WQI. Residual analysis (normal probability plots, residuals versus variables pots, and plots of residuals versus the fitted values) suggested that all of the relationships between the independent and dependent variables were fairly linear and that variables satisfied the requirements of normality. Some level of heteroscedasticity was observed in the regression models for DO, TN, and TP. There were no significant relationships found between the la ndscape pattern variables and the SCI_1 at any of the grain sizes, or between the landscape pattern variables and the SCI_2. Regression analysis: spatial extent The relationships between significant components that resulted from the PCA of landscape pattern metrics and the water chemistry variables for streams and the SCI are summarized in Table 3-37. A significant relations hip was reported between water turbidity and the landscape pattern variables for the 100-meter scale only (R2 = 0.24, F4,29 = 3.66, p = 0.016). However, the 135

PAGE 136

analysis of residual plots show ed that the level of associa tion between variables was highly influenced by one observation (outli er), which may help to explain the differences reported in the regression results between scales. Estimated coefficients for components 1 and 2 were significantly associated to the dependent vari able (p = 0.004 and p = 0.02, respectively). For DO, a significant relationship with the landscape pattern variable s was found at all scales, with the strongest association f ound at the 100-meter scale (R2 = 0.34, F4,30 = 5.40, p = 0.002). Components 1 and 3 were significa ntly related to DO (estimated coefficients with p-value of 0.006 for both cases). The amount of the variance in the concentration for DO explained by the independent variables showed a tendency to decrease with increasing scale. For NO3-N, a significant relationship with the landscape pattern variables wa s found for the watershed scale only (R2 = 0.24, F4,37 = 4.20, p = 0.007). Component 1 wa s significantly related to NO3-N (estimated coefficient with p-values of 0.016). The strongest relationship with the landscape pattern variables for TN was also found at the watershed scale (R2 = 0.59, F4,38 = 16.35, p < 0.001), with components 1, 2, and 3 significantly relate d (estimated coefficients with p-values of < 0.001, <0.001, and 0.033, respectively). The strength of the association between variables tended to decrease with decreasing scale. Regressions at all scales were st atistically significant. TP was also significantly related to the landscape pattern variable s at all scales. The strongest relationship was also reported for the watershed scale (R2 = 0.44, F4,41 = 9.32, p < 0.001). Components 2, 3, and 4 were significantly related to TP (estimated coefficients with p-values of 0.008, 0.002, and 0.043, respectively) at this scale. The results showed th at the amount of variability in the concentration of TP explained by the independe nt variables tended to decrease with decreasing scale. A significant relationship between the independent variables and the WQI was found at the 400-meter and watershed scales Among these, the strongest association was 136

PAGE 137

found at the 400-meter scale (R2 = 0.29, F4,30 = 4.46, p = 0.006). Component 4 was the only axis significantly related to the WQI with an estimated coefficient with a p-value of 0.004. The analysis of residuals for the most significant fi ndings suggested a linear relationship between the response and predictor variables. All regression models satisfied the requirements of normality. Some level of heteroscedasticity was observed for the DO, NO3-N, TP, TN, and WQI models. Unequal variances among residuals were more evident for the TN and the WQI models. A significant relationship between the landsca pe pattern variables and the SCI_1 was found only for the 100-meter scale (R2 = 0.17, F4,57 = 4.18, p = 0.005). The analysis of residual plots seemed to suggest that the regression mode l was not adequate to explain the relationship between variables; the normal plot of residua ls showed that the re siduals were not normal distributed and the residuals agains t the fits plot showed that the residuals did not have a constant variance. For the SCI_2, a significant relationship with the land scape variables was found at the 100-meter and 400-meter scales. The most significant model was the one for the 100-meter scale (R2 = 0.22, F4,63 = 5.41, p = 0.001). According to the analysis of residual plots, this regression model seemed more adequate to explain the re lationship between variab les. Plots showed a normal distribution of the residual s and a fairly constant varian ce. A linear relationship between the variables was suggested in the plot of th e standardized residuals and the standardized predicted values. For the SCI_1, component 4 was the only axis significan tly related to this variable with an estimated coefficient with a p-value of 0.001; components 1 and 3 were significantly associated with the SCI_2 (estimated coefficients with p-values of 0.002 and 0.003, respectively). 137

PAGE 138

Lakes Metric selection: grain size Pattern metrics were calculated for lake draina ge basins (n = 48) e xporting ArcGrids into Fragstats with 20 x 20-, 40 x 40-, 60 x 60-, and 80 x 80-meter cell grain sizes. Descriptive statistics and information on the transformations in scores for each metric calculated for each scale considered are provided in Appendix N, Table N-4. None of the metrics showed unusual values that would suggest that any of these could not be considered for further analysis. However, the metric FRAC_MN was not able to differentiate between di fferent landscapes, and showed a standard deviation of 0.01 for all gr ain sizes. As a result, this metric was not considered for further analysis. Correlation analysis was used to test for redundancy among the remaining landscape pattern metrics. Table 3-38 shows the correlation results of all pa irs of metrics considered. The metrics PD and AREA_MN were highly correlated at all scales considered (r = -1.00, p < 0.001; r = -0.96, p < 0.001; r = -0.96, p < 0.001; r = -1.00, p < 0.001, with increasing scale, respectively). The metrics CONTAG and SHEI were also highly correlated at all of the scales considered (r = -0.98, p < 0.001; r = -0.96, p < 0.001; r = -0.951, p < 0.001; r = -0.93, p < 0.001, with increasing grain size, respectively). T hus, the metrics AREA_MN and CONTAG were eliminated from further analysis. The results of the principal components an alysis, including the eigenvalues and the proportion of the variance explained by the firs t six components for each grain size considered, are presented in Table 3-39. The first five components (eigenvalue > 1) explained between 81.4% (40 x 40-meter scale) and 83.7% (20 x 20-met er scale) of the variance in the landscape pattern metrics. The factor loadings for the fist five components are shown in Tables 3-40 through 3-43. Components were labe led to represent the metric th at loaded highest with each 138

PAGE 139

axis for each grain size. For example, for the 20 x 20-meter grain size component 1 was labeled DIVERS1 since the metric SHEI (Shannon Evenness Index) loaded highest with this component. Component 2 was labeled DIVERS2 since the metr ic PRD (path richness density) loaded highest with this component. The metrics PLAND_Urb and AREA_CV correlated highest and equally with component 3, so it was labeled URB/ SIZE. In components 4 and 5 the metrics PLAND_Wet and PLAND_Ag loaded highest in each component and were labeled WET and AG, respectively. A similar reasoning was used to label the components at other grain sizes. The first component was influenced by the metric s that describe the diversity of patches in the landscape, particularly by the metrics SHDI a nd SHEI, which presented a negative load at all scales. The metric ED also loaded fairly and nega tively with this component at all scales. At the 20 x 20and 60 x 60-meter scales the metric PD also had a fair correlation with component 1. The metric IJI also had some influence on axis 1 although it loaded poorly at the 80 x80-meter scale. This component can be summarized as a patch diversity gradient ranging from landscapes with diverse land uses evenly distributed in sp ace to landscapes dominated by a few land uses. For the second component PRD correlated highest with this axis at the finer to medium scales, with less influence at 80 x 80-meter grain size. The metric PLAND_Wet was also fairly correlated with this axis, particularly at the 20 x 20and 60 x 60-meter scales. The metric PD also loaded fairly with this component, with it s importance tending to in crease with increasing scale. In summary, this component is characterized as a diversity /heterogeneity gradient ranging from landscapes dominated by a small number of patch types, among which wetlands were dominant, to more diverse, patchy landscapes. The metrics that correlated highest with the third component included PLAND_Urb and PLAND_For, which were negatively correlated with each other. However, the direction of the 139

PAGE 140

relationship of these metrics wi th component 3 shifted with ch anges in scale. The metric ENN_MN also had a fair load with this axis at the 40 x 40-, 60 x 60-, and 80 x 80-meter scales, increasing its influence with increasing scale. At the 20 x 20-meters grain size the metrics AREA_CV and PR correlated fairly and negati vely with axis 3. This component can be summarized as a development gr adient ranging from landscapes with a few large forested patches to landscapes with a high diversity of patches and with urban lands well represented. The fourth component summarizes spatial patterns in agricultural land cover. The metric PLAND_Ag correlated highly with this component, particularly at broader grain sizes. At the finest scale considered, the metr ic PLAND_Wet had a fair loading with this axis and correlated negatively with PLAND_Ag. For the fifth co mponent, the metrics PLAND_Ag and PLAND_For had a fair correlation with this axis, especially at the finer scales considered. At the broader scales AREA_CV correlated hi ghest with this component. Metric selection: spatial extent In addition to the watershed scale, landscape pattern metrics were cal culated for landscape buffers of 100 and 400 meters surrounding lakes (n = 44) and keeping the grain size constant at 20 x 20 meters. Descriptive statistics and inform ation on the transformations used to obtain a normal distribution in scores for each metric calcul ated for each scale cons idered are provided in Appendix N, Table N-5. Among the 17 metrics, only FRAC_MN was excluded from further analysis since it was not able to differentiate between the different landscapes showing (SD = 0.02 for the 100-m buffer and 400-m buffer scales and SD = 0.01 for the watershed scale). The analysis for establishing redundancy among metrics showed very similar results to those presented for the selection of metrics base d on changes in grain size. Pearsons correlation results of all pairs of metrics are presente d in Table 3-44. The metrics PD and AREA_MN showed a very high correlation (r = -0.92, p < 0.001; r = -0.99, p < 0.001; r = -0.97, p < 0.001, 140

PAGE 141

with increasing scale, respectively). The metric s CONTAG and SHEI were also highly correlated at all of the spatial extent s considered (r = -0.95, p < 0.001; r = -0.98, p < 0.001; r = -0.98, p < 0.001; with increasing scale, respectively). As a result, the metrics AREA_MN and CONTAG were eliminated from further analysis. Results of the principal components analysis including the eigenvalues and the proportion of the variance explained by the first six com ponents for each spatial extent considered, are presented in Table 3-45. The first five compone nts (eigenvalue > 1) explained between 80.0% (100-meter scale) and 81.5% (watershed scale) of the variance in the landscape pattern metrics. The factor loadings for the first five com ponents are shown in Tables 3-46 through 3-48. Components were labeled to represent the metric that loaded highest with each axis for each spatial extent using a similar reasoning to the one used to label the components in the landscape grain analysis. The first component correlated highest with the metrics SHDI and SHEI, which loaded negatively at all of the scales with this axis. Th e metric PRD had a fair negative loading with this component. Its influence decreased with increas ing scale and PR became more important. The metrics PD and ED also correlated to some extent with component 1, with their influence tending to decrease with increasing scale. The metric IJI also loaded fairly and negatively with this component, especially at coarser scales. This co mponent was a diversity gradient ranging from landscapes with numerous patches relatively evenly distributed in space and with a variety of land use types represented, to landscapes dominated by a small number of patches usually belonging to one or few land use types. The second component also correlated highest with metrics that provide information on patch diversity. The metric PR loaded positively with component 2 with its influence tending to 141

PAGE 142

decrease with increasing scale. PRD loaded negatively with component 2 and its influence tended to increase with increasing scale. AREA _CV was positively correlated with component 2 at all scales. Similarly, ENN_MN had a positive correlation with componen t 2, especially at the broader scales. This component summarizes a di versity gradient contra sting patch diversity quantified based on the total number of patches and quantified relative to the total landscape area. Landscapes ranged from sites with numerous patches of various sizes and with a variety of land use types represented to landscapes dominated by a small number of rather homogenous patches sizes belonging to a few land use types. Component 3 summarized a development gradient where metrics that described the proportion of land use types in th e landscape correlated highest w ith this component. The metric PLAND_Urb correlated fairly with component 3 at all scales a lthough it showed a shift in the direction of the correlation (pos itive) at the 400-meter buffer s cale. The metric PLAND_Wet also loaded fairly with component 3 at all scales and was inversel y correlated with PLAND_Urb. The metric PLAND _For loaded fairly and negatively with component 3 at the watershed scale. The metrics ENN_MN and IJI also had a fairly positive level of associati on to this axis at the finest scale. Their influence on component 3 decreased with increasing scale. The metric SHAPE_NM correlated highest an d negatively with component 4 at the 100meter buffer scale. The influence of this metric on axis 4 tended to decrease with increasing scale. The metric PD also loaded fairly with this component with the highest correlation (positive) reported at the 400-meter scale. At th e watershed scale AREA_C V loaded fairly with component 4. The fifth component also seemed to summarize a development gradient in which PLAND_For had a high negative corr elation with component 5 at the 100-meter scale. At the 400-meter and watershed scales PLAND_Ag had a fairly high correlation with this axis, 142

PAGE 143

although a shift in the directi on of the relationship was obser ved. At the 400-meter scale, PLAND_For loaded fairly and positively with com ponent 5 but loaded poorly with this axis at the watershed scale. At this same scale PLAND_ Urb had some effect on axis 5 and correlated inversely with PLAND_Ag. Regression analysis: grain size The relationships between significant components that resulted from the PCA of landscape pattern metrics and water chemistry variables for lakes and the LCI were explored using multiple regression analysis. For each grain considered th e spatial extent was held constant at the watershed scale. Regression resu lts are presented in Table 3-49. A significant relationship was found at all scales when the land scape pattern variables were rela ted to TN, with the strongest association established at th e 20 x 20-meters grain size (R2 = 0.29, F5,41 = 4.80, p = 0.002). Components 1 and 5 were significant ly related to TN (estimated co efficients with p-values of 0.014 and < 0.001, respectively). A significant but weak relationship be tween the landscape pattern metrics and TP was also found but only for the 20 x 20-meter grain size (R2 = 0.17, F5,41 = 2.93, p = 0.024). The p-values for the estimated co efficients of compon ents 2 and 4 were 0.016 and 0.012, indicating that they we re significantly related to TP For the LCI, significant relationships with the landscap e variables were found at all gr ain sizes with the strongest association reported at the 40 x 40-meter scale (R2 = 0.39, F5,41 = 6.86, p < 0.001). Components 1 (estimated coefficient with p < 0.001) and 5 (estimated coefficient with p = 0.013) were significantly associated to the LCI. Residual analys is showed that residuals for all models were normally distributed and that the most significant relationships between the above variables were fairly linear. However, for the TP model the variance among residuals di d not appear constant. 143

PAGE 144

Regression analysis: spatial extent Table 3-50 summarizes the regression results of multiple regression analysis used to test for relationships between significan t components that resulted from the PCA of landscape pattern metrics measured at three different spatial extents and the water chemistry variables for lakes and the LCI. For this analysis the grain size was held constant at the 20 x 20meter grain size. When the landscape pattern metrics were related to TKN, a significant relationship was reported for the 400-meter buffer scale (R2 = 0.25, F5,37 = 3.83, p = 0.007). Among all components only component 5 was significantly related to TKN (es timated coefficient with p value of 0.037). The analysis of residuals showed th at the residuals did not follow a normal distribution. In addition, the variance among residuals was not constant an d a non-linear pattern was observed from the plot of residuals versus fits. There was a sign ificant relationship between the landscape pattern metrics and TN at the 400-meter buffer and wa tershed scales. The str ongest association was reported for the 400-meter buffer scale (R2 = 0.39, F5,37 = 6.38, p < 0.001) with components 1, 3, and 5 explaining most of the variation in the con centration of TN (estimated coefficients with p values of 0.004, 0.012, and 0.001, respectively). The an alysis of residuals revealed that the residuals did follow a normal distribution. Ho wever, the variance among residuals was not constant. When TP was regressed against the landscape variables, a significant relationship was found only for the 400-meter buffer scale (R2 = 0.21, F5,37 = 3.25, p = 0.015). Component 3 explained most of the variation in the concentration of TP (estimat ed coefficients with p value of 0.001). The analysis of residuals confirmed that the residuals followed a normal distribution. The plot of residuals versus fits showed that the ra ndom variation of the residuals increased as the fitted values increased, an indicative of non-constant variance. The strongest association between the LCI and the landscape vari ables was found for the 400-meter buffer scale. The model had a 144

PAGE 145

fair fit (R2 = 0.42) and was statis tically significant (F5,41 = 7.01, p < 0.001). The estimated coefficients for component 1 (p = 0.001) a nd component 2 (p < 0.001) showed that these components explained more of the variability in the LCI scores. According to the residuals analysis the residuals had a nor mal distribution and the variance among residuals did not appear constant. Regression models for the LCI at the 1 00-meter buffer and watershed scales were also statistically significant. Land Use Intensity, Landscape Pa ttern, and Ecosystem Condition Multiple factor models that included both the LDI and pattern metrics as independent factors were use to analyze the predictive power the landscape variables had when used together to explain the variability in water quality vari ables and biological indicators. Changes in the proportion of the total variability of the response variables expl ained by the landscape variables were interpreted as the added predictive power to the LDI that resulted after using the LDI together with significant compone nts that resulted from the PC A of landscape pattern metrics. Results are presented for all three freshwater systems studied. Isolated Forested Wetlands For isolated forested wetlands, relationships were assessed using only the data for the 200meter buffer, and at four grain sizes which corresponded to the spat ial scales at which the pattern metrics were calculated. Regression results for si gnificant associations for a sample of isolated forested wetlands are shown in Table 3-51. Amo ng the water chemistry variables considered, an increase in the amount of the variability explai ned was reported only for TP, with more of the variance of the concentration of TP explaine d at the 30 x 30-m grain size when the LDI-PLU was used with the landscape pattern metrics (R2 = 0.43, F7,24 = 4.31, p = 0.03; R2 = 0.39). Among the biological indicators of wetland condition, when the LDI-ISD and the landscape metric pattern variable s were used together, up to an additional 25% of the total 145

PAGE 146

variance in the macrophytes WCI scores was explained at the 10 x 10-meter grain size (R2 = 0.44, F7,43 = 6.53, p < 0.001). For the macroinvertebrate WCI, up to an additional 17% of the total variance in the macroinvertebrate WCI scor es was explained at the 30 x 30-meter grain size when the LDI-ISD and the landscape metric pattern variables were used together as independent variables (R2 = 0.34, F7,23 = 3.18, p = 0.017). There were no significant relationships between the diatom WCI and the landscape indices. Adding the landscape pattern metrics to the LDI resulted in a decrease in the amount of the variance explained by these variables (see Table O-1 in Appendix O). Streams Significant results for multiple regressions are shown in Tables 3-52 and 3-53 for landscape variables measured with changes in gr ain size and with change s in spatial extent, respectively (all other regression results are pr esented in Appendix O). For TN, and additional 46% of the total variance in th e concentration of TN was expl ained when the pattern metrics were used together with the LDI-PLU at the 5 x 5-m grain size (R2 = 0.63, F5,40 = 16.61, p < 0.001). For TP, and additional 37% of the total variance in th e concentration of TP was explained when the pattern metrics were used together with the LDI-ISD also at the 5 x 5-m grain size (R2 = 0.40, F5,40 = 7.05, p < 0.001). In both cases, the differences among the amount of the variance accounted for by the landscape variab les was minimal when the different forms of the LDI were used. When the landscape pattern metrics were used together with the LDI to explain the variability in the WQ I, an additional 11% to 13 % of the total variance was explained with changes in grain size. The most signifi cant relationship was re ported at the 110 x 110-m scale when the LDI-ISD was used in comb ination with the landscape variables (R2 = 0.46, F5,30 = 6.99, p < 0.001; R2 = 0.12). 146

PAGE 147

For changes in spatial extent, regression re sults showed that when the LDI and the landscape pattern variables were used together in multiple regression analysis these variables allowed a significant prediction of the variation in the streams water SC but only at the 100meter buffer scale. More of the SC was explai ned by the factor model that included the LDI-ISD (R2 = 0.29, F5,28 = 3.67, p = 0.011; R2 = 0.28). For DO, more of the variance was explained by landscape variables at the watershed scal e level when the LDI-PLU was used (R2 = 0.44, F5,29 = 6.25, p = 0.004; R2 = 0.04). Models with the LDI and the landscape pattern metrics were better predictors of TN concentrations at all scales. More of the remaining variance was explained at the watershed scale when the LDI-PLU was used (R2 = 0.62, F5,37 = 14.54, p < 0.001; R2 = 0.45) with minimal differences among models with different LDI forms. The variability in the concentration of TP was equally explained by the landscape variables at the 400-meter buffer scale and at the watershed scale (R2 = 0.42) with minimal diff erences among models with different forms of the LDI. The LDI and the land scape pattern variables explained an additional 35% to 41% of the total varian ce in TP depending on the form of the LDI considered. The landscape pattern metrics were also important add itional factors in explaining together with the LDI the remaining variance for the WQI. The largest change was reported for the 400-meter buffer scale for the LDI-PLU ( R2 = 0.24). Nevertheless, the st rongest association between variables was established for the LD I-ISD at the same spatial extent (R2 = 0.49, F5,29 = 7.65, p < 0.001; R2 = 0.20). When the LDI and the landscape pattern metrics were included as independent variables in multiple regression models, at all grain sizes ther e was a decrease in the am ount of the variability in the SCI explained by these va riables compared to the amount of the variance explained by the LDI alone. The most significant relationship betw een the SCI_1 and the landscape variables was 147

PAGE 148

found for the 80 x 80-meter grain size for the regression model that included the LDI-ISD (R2 = 0.10, F5,60 = 2.45, p = 0.044; R2 = -0.12). For the SCI_2, the most significant relationship was found at the 20 x 20-meter grain size for the model that included the LDI-ISD (R2 = 0.15, F5,61 = 3.37, p = 0.009; R2 = -0.10). For changes in spatial extent, there was also a decrease in the amount of the variability in the SCI explained by these variab les compared to the amount of the variance explained by the LDI alone. More of the varian ce in the SCI_1 scores was accounted for by the landscape variables at the 100-meter buffer scale where when the LDI-ISD was used (R2 = 0.19, F5,56 = 3.81, p = 0.005; R2 = -0.03). For the SCI_2, the strongest association between variables was reported at the 100-meter buffer scale for the model that included the LDI-ISD (R2 = 0.31, F5,57 = 5.05, p = 0.001, R2 = -0.01). For both the SCI_1 and the SCI_2, differences in the amount of the variance of the SCI explained by the indepe ndent variables were very small when the different forms of the LDI were used. Lakes The significant results for multiple regressions that included the LDI and the landscape pattern metric variables to explore how much of the variation in water quality variables and indicators of ecosystem condition was explained by these variables at di fferent grain sizes and different spatial extents are shown in Tabl es 3-54 and 3-55, respect ively (non-significant regression results are presented in Appendix O) A significant increase in the amount of the variability explained in the c oncentration of TN was observed at all scales considered. The largest change occurred at the 20 x 20-meter scale ( R2 = 0.34) for the model that included the LDI-PLU (R2 = 0.34, F6,40 = 4.97, p = 0.001). For TP, the relationship with the landscape variables was only signif icant at the 20 x 20 meter scale with more of the variance explained by the model that included the LDI-PLU (R2 = 0.17. F6,40 = 2.61, p < 0.031; R2 = 0.15). For the 148

PAGE 149

LCI, the largest change in the unexplained vari ance of LCI scores occurr ed at the 40 x 40-meter scale with a R2 of 0.48 when the LDI-PLU was used as an independent variable. The landscape indicators explained 50 % of the total variance in the LCI scores (F6,40 = 8.76, p < 0.001). For changes in spatial extent, regression results showed that multiple factor models that included the LDI jointly with the landscape pa ttern metrics allowed explaining 24% of the variance in TKN at the 400-meter buffer scale when the LDI-PLU was used as one of the independent variables (F6,36 = 3.20, p = 0.013; R2 = 0.24). Very similar results were reported when the other forms of the LDI were used. For NO3/NO2-N, the multiple regression model that included the LDI-PLU for the watershed scale wa s the only model that reported a significant relationship between the landscape variables a nd TKN; however, the relationship was weak (R2 = 0.17, F6,37 = 2.51, p = 0.038; R2 = 0.17). The models that incl uded the LDI-PLU and the LDIILD at the 400-meter scale showed the strongest associations between the landscape variables and TN (R2 = 0.40, p < 0.001, R2 = 0.40, in both cases). For TP the relationship with the landscape variables was only significant at the 4 00-meter scale. Among the different LDI forms, the model that included the LDI-ILD explained more of the variance in TP (R2 = 0.23, F6,36 = 3.09, p < 0.015; R2 = 0.23). Significant relationships between the LCI and the landscape variables were observed at all sp atial extents considere d. Among these, more of the variance in the LCI was explained by the multiple factor model that included the LDI-PLU at the watershed scale (R2 = 0.52 F6,36 = 8.72, p < 0.001; R2 = 0.52). 149

PAGE 150

Table 3-1. Percent values for the land use/land cover (LU/LC) for a priori defined 200-meter buffer areas for the sample of isolated fo rested wetlands (n = 118). The number of sites for each a priori category that included each LU/LC class in their surrounding landscape is shown. LU/LC classes were defined according to Level 1 of the FLUCCS. A priori Classes Urban Agriculture Rangeland Forest Water Wetland Transportationa Reference % of total buffer 0.5 0.0 0.0 72.8 0.1 24.5 2.2 # of sites 2 0.0 0.0 36 3 31 26 Agricultural % of total buffer 0.5 61.9 4.5 21.2 2.0 8.2 1.7 # of sites 5 32 5 21 25 28 19 Urban % of total buffer 63.9 1.5 0.0 17.5 2.7 4.1 10.5 # of sites 41 5 0.0 32 23 16 39 a Includes access, dirt, and paved roads. Table 3-2. Summary statistics of the non-renewa ble and purchased areal empower density (E+14 sej/ha/yr) for a priori defined buffer area classes of the isolated forested wetlands. The non-renewable areal empower density was calculated based on the proportion occupied by each land use type within a 200-meter buffer of the sample wetlands. n Mean SD () Minimum Maximum Reference 37 24.1 54.5 0.0 227.7 Agricultural 40 61.3 57.6 3.3 260.9 Urban 41 2239.2 1563.9 288.8 8164.4 Table 3-3. Summary statistics on the size and th e land use/land cover (LU/LC) composition for the drainage areas for the sample streams (n = 69) and lakes (n = 54). Streams Lakes Mean SD () Mean SD () Size (ha) 11,371.5 14,740.0 468.2 643.7 LU/LC a (%) Urban 12.97 18.20 49.21 31.60 Agriculture 19.55 20.28 27.23 19.59 Rangeland 5.09 6.71 11.47 14.44 Forest 44.79 25.68 18.29 19.45 Water 1.65 2.83 3.06 4.46 Wetland 16.73 10.34 13.82 14.46 Barren landb 0.35 1.01 0.17 0.22 Transportation 1.63 2.86 5.88 11.83 a LU/LC categories defined according to Level 1 of the FLUCCS classification scheme. b Areas of bare soil or rock (FDOT 1999). 150

PAGE 151

Table 3-4. Summary statistics of the non-renewa ble and purchased areal empower density (E+14 sej/ha/yr) for the sample streams and lakesa. Streams Lakes EmpDen-PLU 432.67 765.01 1662.98 1668.6 EmpDen-ILD 218.89 347.46 1072.56 1163.35 EmpDen-ISD 135.25 207.07 813.99 888.53 a The areal empower density was calculated ba sed on the proportion occupied by each land use type (EmpDen-PLU), and assuming that the eff ect of development intensity on the landscape decreased linearly with distance (EmpDen-ILD), a nd in inverse square with distance (EmpDenISD). All calculations were made for the total drainage basin. Table 3-5. Spearman correlation between the three forms of the LDI calculated for the sample isolated forested wetlands at three diffe rent spatial extents (buffer areas of 20, 100, and 200 meters surrounding wetlands). 20-meter 100-meter 200-meter LDI PLU ILD ISD PLU ILD ISD PLU ILD ILD 0.99 20-m ISD 0.99 0.99 PLU 0.94 0.93 0.91 ILD 0.95 0.94 0.93 0.99 100-m ISD 0.96 0.95 0.94 0.98 1.00 PLU 0.89 0.88 0.86 0.96 0.95 0.94 ILD 0.90 0.89 0.88 0.97 0.96 0.96 1.00 200-m ISD 0.92 0.92 0.90 0.99 0.98 0.97 0.99 0.99 Table 3-6. Spearman correlations between the thr ee forms of the LDI calculated for the sample streams at three different spatial extents (100 meters, 400 meters, and the total watershed). 100-meter 400-meter Watershed LDI PLU ILD ISD PLU ILD ISD PLU ILD ILD 0.97 100-m ISD 0.93 0.98 PLU 0.98 0.96 0.91 ILD 0.96 0.97 0.95 0.98 400-m ISD 0.93 0.96 0.96 0.95 0.99 PLU 0.94 0.90 0.84 0.96 0.93 0.90 ILD 0.93 0.92 0.88 0.95 0.95 0.93 0.99 Watershed ISD 0.91 0.92 0.90 0.93 0.95 0.96 0.95 0.99 151

PAGE 152

Table 3-7. Spearman correlations between the thr ee forms of the LDI calculated for the sample lakes at three different spatial extent s (100 meters, 400 meters, and the total watershed). 100-meter 400-meter Watershed LDI PLU ILD ISD PLU ILD ISD PLU ILD ILD 0.90 100-m ISD 0.82 0.92 400-m PLU 0.99 0.87 0.79 ILD 0.93 0.99 0.91 0.91 ISD 0.90 0.97 0.94 0.87 0.96 PLU 0.98 0.86 0.78 1.00 0.90 0.86 ILD 0.92 0.94 0.85 0.91 0.96 0.92 0.90 Watershed ISD 0.91 0.97 0.92 0.88 0.97 0.99 0.87 0.93 Table 3-8. Simple linear regression values (r2) for regressions between the LDI and the water chemistry variables measured at three la ndscape extents for the sample isolated forested wetlands ( -level of 0.05). 20-meter buffer 100-meter buffer 200-meter buffer LDI-PLU r2 p r2 p r2 p Log10(DOa) 0.05 0.051 0.06 0.040 0.07 0.028 Log10(SCb) 0.12 0.053 0.15 0.024 0.21 0.008 Log10(TNc) 0.04 0.096 0.03 0.165 0.02 0.215 Log10(TPd) 0.07 0.027 0.06 0.043 0.04 0.071 Log10(Turbe) 0.03 0.164 0.04 0.078 0.03 0.111 LDI-ILD Log10(DO) 0.04 0.100 0.05 0.063 0.06 0.046 Log10(SC) 0.12 0.052 0.13 0.043 0.18 0.014 Log10(TN) 0.04 0.080 0.03 0.139 0.03 0.155 Log10(TP) 0.05 0.060 0.05 0.059 0.04 0.094 Log10(Turb) 0.02 0.250 0.04 0.106 0.03 0.141 LDI-ISD Log10(DO) 0.03 0.130 0.04 0.079 0.05 0.062 Log10(SC) 0.12 0.051 0.11 0.056 0.16 0.023 Log10(TN) 0.04 0.075 0.04 0.110 0.03 0.129 Log10(TP) 0.04 0.092 0.04 0.079 0.04 0.104 Log10(Turb) 0.01 0.362 0.03 0.138 0.03 0.156a Dissolved oxygen; b Specific conductance; c Total nitrogen; d Total phosphorus; e Turbidity 152

PAGE 153

Table 3-9. Simple linear regression results (r2) showing the proportion of total variance in each of five water chemistry variables explaine d by the LDI calculated at eight different grain sizes (meters on a side) for the sample isolated forested wetlands (* = p < 0.05, ** = p < 0.01). 5-m 10-m 20-m 30-m 40-m 50-m 60-m 70-m LDI-PLU Log10(DO) 0.069* 0.067* 0.066* 0.064* 0.067* 0.065* 0.060* 0.068* Log10(SC) 0.205** 0.207** 0.208** 0.204** 0.205** 0.206** 0.200** 0.193* Log10(TN) 0.019 0.020 0.019 0.018 0.019 0.019 0.018 0.025 Log10(TP) 0.044 0.042 0.042 0.043 0.044 0.042 0.042 0.038 Log10(Turba) 0.034 0.034 0.035 0.035 0.033 0.032 0.033 0.027 LDI-ILD Log10(DO) 0.057* 0.057* 0.056* 0.053 0.067* 0.064* 0.059* 0.058* Log10(SC) 0.179* 0.182* 0.181* 0.173* 0.177* 0.173* 0.142* 0.154* Log10(TN) 0.026 0.027 0.028 0.024 0.029 0.030 0.027 0.032 Log10(TP) 0.038 0.037 0.036 0.039 0.040 0.039 0.043 0.042 Log10(Turba) 0.030 0.029 0.030 0.031 0.025 0.020 0.030 0.031 LDI-ISD Log10(DO) 0.050 0.050 0.048 0.047 0.061* 0.059* 0.047 0.052 Log10(SC) 0.155* 0.159* 0.157* 0.151* 0.153* 0.158* 0.125* 0.136* Log10(TN) 0.031 0.032 0.033 0.030 0.034 0.035 0.030 0.035 Log10(TP) 0.036 0.035 0.034 0.036 0.037 0.037 0.039 0.039 Log10(Turba) 0.027 0.027 0.028 0.028 0.02 0.016 0.029 0.025 a Turbidity. Table 3-10. Simple linear regression values (r2) for regressions between the LDI and the WCI measured at three spatial extents for the sample isolated forested wetlands. 20-meter buffer 100-meter buffer 200-meter buffer LDI-PLU r2 p r2 p r2 p Macrophyte WCI 0.30 <0.001 0.27 <0.001 0.24 <0.001 Macroinvertebrate WCI 0.24 <0.001 0.23 <0.001 0.18 <0.001 Diatom WCI 0.19 0.002 0.24 <0.001 0.23 <0.001 LDI-ILD Macrophyte WCI 0.23 <0.001 0.24 <0.001 0.21 <0.001 Macroinvertebrate WCI 0.20 <0.001 0.20 <0.001 0.18 <0.001 Diatom WCI 0.15 0.006 0.20 0.001 0.20 0.001 LDI-ISD Macrophyte WCI 0.19 <0.001 0.21 <0.001 0.19 <0.001 Macroinvertebrate WCI 0.18 <0.001 0.19 <0.001 0.17 <0.001 Diatom WCI 0.12 0.016 0.17 0.003 0.19 0.002 153

PAGE 154

Table 3-11. Simple linear regression results (r2) showing the proportion of total variance in each of the three WCIs explained by the LDI in its three forms calculated at different grain sizes (meters on a side) for the sample isol ated forested wetlands. All results were significant (p < 0.01). 5-m 10-m 20-m 30-m 40-m 50-m 60-m 70-m LDI-PLU Macrophyte WCI 0.243 0.239 0.237 0.236 0.229 0.229 0.214 0.215 Macroinvertebrate WCI 0.183 0.180 0.174 0.179 0.179 0.188 0.164 0.179 Diatom WCI 0.228 0.229 0.228 0.231 0.225 0.225 0.219 0.213 LDI-ILD Macrophyte WCI 0.209 0.207 0.207 0.205 0.203 0.202 0.193 0.183 Macroinvertebrate WCI 0.179 0.177 0.172 0.171 0.183 0.198 0.163 0.172 Diatom WCI 0.203 0.205 0.206 0.208 0.208 0.196 0.187 0.205 LDI-ISD Macrophyte WCI 0.190 0.189 0.190 0.191 0.188 0.184 0.174 0.174 Macroinvertebrate WCI 0.172 0.172 0.170 0.168 0.182 0.191 0.156 0.177 Diatom WCI 0.186 0.189 0.192 0.195 0.193 0.186 0.164 0.189 154

PAGE 155

Table 3-12. Coefficients of determination (r2) for simple linear regressions between the three forms of the LDI and the water chemistr y variables and the Water Condition Index (WQI) measured at three spatial extents for the sample streams ( -level of 0.05). 100-meter buffer 400-meter buffer Watershed LDI-PLU r2 p r2 p r2 p Log10(Turba) 0.04 0.299 0.02 0.405 0.03 0.334 DOb 0.40 <0.001 0.34 <0.001 0.40 <0.001 Log10(NO3-Nc) 0.07 0.080 0.09 0.040 0.14 0.011 Log10(TNd) 0.13 0.011 0.10 0.030 0.17 0.004 Log10(TPe) 0.02 0.350 0.01 0.490 0.01 0.491 WQIf 0.21 0.004 0.21 0.004 0.22 0.003 LDI-ILD Log10(Turb) 0.001 0.886 0.002 0.829 0.01 0.564 DO 0.33 <0.001 0.31 <0.001 0.41 <0.001 Log10(NO3-N) 0.04 0.188 0.07 0.084 0.10 0.031 Log10(TN) 0.11 0.021 0.09 0.043 0.16 0.005 Log10(TP) 0.05 0.139 0.02 0.353 0.02 0.335 WQI 0.26 0.001 0.25 0.002 0.28 <0.001 LDI-ISD Log10(Turb) 0.007 0.654 0.003 0.765 0.001 0.845 DO 0.31 <0.001 0.31 <0.001 0.41 <0.001 Log10(NO3-N) 0.01 0.446 0.04 0.168 0.08 0.061 Log10(TN) 0.11 0.025 0.09 0.039 0.16 0.005 Log10(TP) 0.08 0.049 0.03 0.241 0.03 0.247 WQI 0.32 <0.001 0.29 0.001 0.33 <0.001a Turbidity, b Dissolved oxygen, c Nitrate nitrogen, d Total nitrogen, e Total phosphorus, f Water Quality Index. 155

PAGE 156

Table 3-13. Simple linear regressions (r2) showing the proportion of tota l variance in each of five water chemistry variables and the WQI e xplained by the LDI calculated at six different grain sizes (meters on a side) for streams (* = p < 0.05, ** = p < 0.01). 20-m 50-m 80-m 110-m 140-m 170-m LDI-PLU Log10(Turb) 0.030 0.029 0.030 0.033 0.035 0.045 DO 0.403** 0.395** 0.398** 0.427** 0.439** 0.444** Log10(NO3-N) 0.139* 0.140* 0.142* 0.138* 0.141* 0.147** Log10(TN) 0.170** 0.167** 0.168** 0.196** 0.210** 0.232** Log10(TP) 0.011 0.010 0.011 0.015 0.017 0.027 WQI 0.222** 0.220** 0.223** 0.225** 0.238** 0.255** LDI-ILD Log10(Turb) 0.011 0.010 0.011 0.013 0.014 0.019 DO 0.407** 0.402** 0.407** 0.438** 0.448** 0.451** Log10(NO3-N) 0.101* 0.100* 0.101* 0.099* 0.100* 0.106* Log10(TN) 0.164** 0.160** 0.163** 0.189** 0.194** 0.209** Log10(TP) 0.021 0.019 0.020 0.024 0.027 0.036 WQI 0.279** 0.277** 0.275** 0.285** 0.301** 0.313** LDI-ISD Log10(Turb) 0.001 0.001 0.002 0.002 0.003 0.004 DO 0.409** 0.405** 0.409** 0.442** 0.449** 0.452** Log10(NO3-N) 0.078 0.077 0.078 0.076 0.077 0.081 Log10(TN) 0.159** 0.156** 0.157** 0.186** 0.184** 0.197** Log10(TP) 0.030 0.027 0.029 0.033 0.036 0.045 WQI 0.332** 0.330** 0.328** 0.340** 0.354** 0.364** Table 3-14. Simple linear regression values (r2) for regressions between the three forms of the LDI and the SCI measured at three spatial extents for the sample streams. 100-meter buffer 400-meter buffer Watershed LDI-PLU r2 p r2 p r2 p SCI_1a 0.20 <0.001 0.17 0.001 0.17 0.001 SCI_2b 0.24 <0.001 0.19 <0.001 0.19 <0.001 LDI-ILD SCI_1 0.24 <0.001 0.20 <0.001 0.21 <0.001 SCI_2 0.26 <0.001 0.22 <0.001 0.23 <0.001 LDI-ISD SCI_1 0.27 <0.001 0.22 <0.001 0.23 <0.001 SCI_2 0.26 <0.001 0.22 <0.001 0.25 <0.001aSCI defined by Bar bour et al. (1996b); bSCI defined by Fore (2004). 156

PAGE 157

Table 3-15. Simple linear regression results (r2) showing the proportion of total variance in the SCI explained by the LDI in its three form s calculated at six di fferent grain sizes (meters on a side) for the sample streams. A ll regressions were significant at p < 0.01. 20-m 50-m 80-m 110-m 140-m 170-m LDI-PLU SCI_1a 0.167 0.162 0.162 0.163 0.167 0.167 SCI_2b 0.185 0.178 0.177 0.179 0.181 0.178 LDI-ILD SCI_1 0.205 0.198 0.200 0.200 0.203 0.203 SCI_2 0.228 0.219 0.222 0.221 0.222 0.221 LDI-ISD SCI_1 0.228 0.221 0.223 0.224 0.227 0.227 SCI_2 0.252 0.242 0.245 0.244 0.246 0.247aSCI defined by Bar bour et al. (1996b); bSCI defined by Fore (2004). Table 3-16. Coefficients of determination (r2) for simple linear regressions between the three forms of the LDI and the water chemistry variables and the Lake Condition Index (LCI) measured at three spatial extents for the sample lakes ( -level of 0.05). 100-meter buffer 400-meter buffer Watershed LDI-PLU r2 p r2 p r2 p Log10(Ammonia-N) 0.007 0.547 0.024 0.267 0.014 0.387 Log10(NO3/NO2-N) 0.007 0.535 0.003 0.703 0.003 0.712 Log10(TKN) 0.011 0.445 0.008 0.532 <0.001 0.996 Log10(TN) 0.005 0.603 0.016 0.366 0.002 0.714 Log10(TP) 0.007 0.553 0.025 0.255 0.028 0.229 LCI <0.001 0.968 <0.001 0.957 0.013 0.417 LDI-ILD r2 p r2 p r2 p Log10(Ammonia-N) 0.005 0.615 0.020 0.303 0.037 0.164 Log10(NO3/NO2-N) 0.016 0.366 <0.001 0.938 0.007 0.538 Log10(TKN) 0.010 0.469 0.010 0.471 0.005 0.608 Log10(TN) 0.003 0.714 0.015 0.383 0.013 0.421 Log10(TP) 0.002 0.723 0.017 0.355 0.053 0.097 LCI 0.001 0.873 <0.001 0.948 0.002 0.746 LDI-ISD r2 p r2 p r2 p Log10(Ammonia-N) 0.004 0.650 0.020 0.312 0.037 0.163 Log10(NO3/NO2-N) 0.022 0.290 <0.001 0.899 0.004 0.644 Log10(TKN) 0.008 0.516 0.012 0.441 0.007 0.544 Log10(TN) 0.001 0.820 0.016 0.374 0.014 0.403 Log10(TP) 0.001 0.820 0.013 0.414 0.043 0.136 LCI 0.002 0.751 0.004 0.666 <0.001 0.888 157

PAGE 158

158 Table 3-17. Simple linear regression results (r2) showing the proportion of total variance in each of five water chemistry variables and the LCI explained by the LDI in its three forms calculated at six different grain sizes (meters on a side) for the sample lakes ( -level of 0.05). 20-m 40-m 60-m 80-m 100-m 120-m LDI-PLU Log10(Ammonia-N) 0.014 0.017 0.017 0.034 0.012 0.010 Log10(NO3/NO2-N) 0.003 0.001 0.002 0.003 0.002 0.002 Log10(TKN) <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 Log10(TN) 0.002 0.001 0.002 0.004 0.002 0.003 Log10(TP) 0.028 0.024 0.027 0.058 0.023 0.022 LCI 0.013 0.021 0.013 0.009 0.011 0.011 LDI-ILD Log10(Ammonia-N) 0.037 0.043 0.036 0.032 0.033 0.029 Log10(NO3/NO2-N) 0.007 0.001 0.006 0.006 0.005 0.005 Log10(TKN) 0.005 0.015 0.005 0.006 0.006 0.005 Log10(TN) 0.013 0.022 0.013 0.013 0.013 0.013 Log10(TP) 0.053 0.021 0.051 0.049 0.044 0.042 LCI 0.002 0.001 0.002 0.002 0.001 0.001 LDI-ISD Log10(Ammonia-N) 0.037 0.035 0.035 0.033 0.032 0.028 Log10(NO3/NO2-N) 0.004 0.003 0.003 0.003 0.002 0.002 Log10(TKN) 0.007 0.008 0.007 0.008 0.008 0.008 Log10(TN) 0.014 0.013 0.013 0.013 0.014 0.014 Log10(TP) 0.043 0.044 0.041 0.040 0.033 0.032 LCI <0.001 <0.001 <0.001 <0.001 <0.001 <0.001

PAGE 159

Table 3-18. Pearsons correlations between landscape pattern metrics calculated at four different grain sizes (meters on a side ) for the isolated forested wetlands. Some of th e metrics descriptors (acronyms) of the co lumns have been renamed for convenience. Metric % Urba % Agb % Forc % Wetd PD ED A_MNe A_CVf SHAPg FRACh ENNi CONTj PR PRD SHDI 5-m PLAND_Ag -0.65 PLAND_For -0.24 -0.28 PLAND_Wet -0.73 0.21 0.08 PD -0.06 -0.15 0.21 0.18 ED -0.01 -0.23 0.10 0.16 0.78 AREA_MN 0.06 0.15 -0.21 -0.18 -1.00 -0.78 AREA_CV -0.46 0.29 0.07 0.40 0.49 0.35 -0.49 SHAPE_MN 0.09 -0.10 -0.04 -0.11 -0.02 0.46 0.02 0.15 FRAC_MN -0.02 -0.02 0.01 -0.04 0.13 0.46 -0.13 0.33 0.92 ENN_MN -0.17 0.08 0.11 0.07 0.22 0.21 -0.22 0.05 0.02 0.05 CONTAG 0.03 0.10 0.06 -0.22 -0.05 -0.11 0.05 0.34 0.22 0.28 -0.15 PR 0.38 -0.11 -0.01 -0.42 0.39 0.15 -0.39 -0.05 -0.20 -0.15 0.16 0.10 PRD 0.33 -0.10 0.02 -0.32 0.46 0.20 -0.46 -0.06 -0.20 -0.14 0.16 0.00 0.91 SHDI 0.35 -0.16 -0.03 -0.31 0.26 0.07 -0.26 -0.31 -0.33 -0.34 0.19 -0.45 0.83 0.79 SHEI 0.07 -0.07 -0.10 0.08 -0.11 -0.13 0.11 -0.46 -0.36 -0.42 0.11 -0.95159 0.05 0.12 0.59 10-m PLAND_Ag -0.65 PLAND_For -0.24 -0.28 PLAND_Wet -0.73 0.21 0.08 PD -0.06 -0.16 0.21 0.13 ED 0.01 -0.23 0.07 0.15 0.75 AREA_MN 0.06 0.16 -0.21 -0.13 -1.00 -0.75 AREA_CV -0.27 0.11 0.09 0.22 0.61 0.39 -0.61 SHAPE_MN 0.26 -0.12 -0.19 -0.19 -0.27 0.31 0.27 -0.18 FRAC_MN 0.21 -0.06 -0.18 -0.17 -0.17 0.32 0.17 -0.04 0.94 ENN_MN -0.18 0.12 0.12 0.06 -0.05 0.08 0.05 -0.10 0.11 0.08 CONTAG 0.07 0.09 0.05 -0.28 -0.11 -0.27 0.11 0.37 -0.05 -0.02 -0.11 PR 0.38 -0.11 -0.02 -0.42 0.27 0.19 -0.27 -0.16 -0.06 -0.13 0.12 0.10 PRD 0.32 -0.10 0.01 -0.32 0.37 0.23 -0.37 -0.14 -0.13 -0.15 0.01 0.04 0.93 SHDI 0.35 -0.16 -0.03 -0.31 0.16 0.14 -0.16 -0.44 -0.07 -0.16 0.11 -0.37 0.85 0.80 SHEI 0.06 -0.06 -0.10 0.09 -0.12 -0.06 0.12 -0.60 -0.06 -0.12 0.09 -0.91 0.11 0.14 0.59

PAGE 160

160Table 3-18. Continued. Metric % Urba % Agb % Forc % Wetd PD ED A_MNe A_CVf SHAPg FRACh ENNi CONTj PR PRD SHDI 20-m PLAND_Ag -0.65 PLAND_For -0.23 -0.29 PLAND_Wet -0.73 0.21 0.07 PD -0.05 -0.12 0.03 0.19 ED 0.00 -0.22 0.06 0.16 0.82 AREA_MN -0.01 0.15 -0.07 -0.07 -0.97 -0.81 AREA_CV -0.17 0.10 -0.02 0.09 0.66 0.48 -0.67 SHAPE_MN 0.05 -0.08 0.01 -0.10 -0.41 0.08 0.40 -0.29 FRAC_MN 0.15 -0.18 0.08 -0.17 -0.33 0.13 0.29 -0.27 0.95 ENN_MN -0.14 0.13 0.03 0.02 0.20 0.21 -0.26 0.24 0.08 0.12 CONTAG 0.11 0.11 0.00 -0.32 -0.22 -0.38 0.17 0.30 -0.12 -0.10 -0.06 PR 0.38 -0.09 -0.06 -0.42 0.18 0.21 -0.28 -0.14 -0.04 0.09 0.14 0.11 PRD 0.34 -0.09 -0.03 -0.33 0.24 0.23 -0.32 -0.16 -0.13 0.00 0.08 0.04 0.93 0-m SHDI 0.36 -0.16 -0.04 -0.32 0.07 0.17 -0.15 -0.45 0.02 0.12 0.03 -0.30 0.87 0.84 SHEI 0.08 -0.08 -0.09 0.09 -0.13 -0.01 0.16 -0.67 0.10 0.09 -0.06 -0.84 0.15 0.20 0.59 3 PLAND_Ag -0.65 PLAND_For -0.24 -0.28 PLAND_Wet -0.71 0.20 0.07 PD -0.05 -0.17 0.00 0.24 ED -0.03 -0.21 0.08 0.16 0.86 AREA_MN 0.01 0.17 -0.06 -0.12 -0.97 -0.85 AREA_CV -0.19 0.11 -0.05 0.13 0.46 0.25 -0.49 SHAPE_MN 0.07 -0.06 0.00 -0.12 -0.37 0.06 0.38 -0.36 FRAC_MN 0.18 -0.14 0.02 -0.21 -0.29 0.13 0.29 -0.45 0.94 ENN_MN -0.14 0.17 0.08 -0.05 0.11 0.13 -0.14 0.28 0.03 0.07 CONTAG 0.13 0.08 0.02 -0.31 -0.32 -0.52 0.24 0.44 -0.26 -0.35 0.02 PR 0.30 -0.06 0.00 -0.38 0.35 0.31 -0.43 -0.09 -0.17 -0.03 0.12 0.04 PRD 0.28 -0.06 0.01 -0.31 0.42 0.31 -0.48 -0.10 -0.31 -0.16 0.01 0.02 0.93 SHDI 0.30 -0.13 -0.02 -0.30 0.31 0.37 -0.35 -0.43 -0.02 0.15 -0.04 -0.36 0.86 0.82 SHEI 0.05 -0.06 -0.09 0.09 0.06 0.23 0.00 -0.73 0.25 0.36 -0.15 -0.86 0.20 0.20 0.62 a % Urb = PLAND_Urb; b % Ag = PLAND_Ag; c % For = PLAND_For; d % Wet = PLAND_WET; e A_MN = Area_MN; f A_CV = AREA_CV; g SHAP = SHAPE_MN; h FRAC = FRAC_MN; i ENN = ENN_MN; j CONT = CONTAG.

PAGE 161

Table 3-19. Eigenvalues and vari ance explained by the first seven axes for the principal components analysis at four different gr ain sizes (meters on a side) for isolated forested wetlands. Component 1 2 3 4 5 6 7 5-m Eigenvalue 2.93 2.53 1.83 1.30 1.12 0.95 0.56 % of variance explained 24.43 21.11 15.26 10.80 9.36 7.89 4.68 Cumulative % of variance 24.43 45.54 60.80 71.59 80.96 88.85 93.53 10-m Eigenvalue 2.84 2.42 1.71 1.47 1.21 1.06 0.59 % of variance explained 23.68 20.17 14.23 12.24 10.08 8.81 4.94 Cumulative % of variance 23.68 43.85 58.08 70.32 80.40 89.21 94.15 20-m Eigenvalue 3.11 2.59 2.25 1.41 1.15 1.08 0.71 % of variance explained 23.90 19.93 17.32 10.88 8.88 8.33 5.47 Cumulative % of variance 23.90 43.82 61.14 72.02 80.89 89.22 94.69 30-m Eigenvalue 3.31 2.55 2.31 1.41 1.14 1.11 0.61 % of Variance explained 25.49 19.60 17.75 10.85 8.75 8.50 4.67 Cumulative % of variance 25.49 45.09 62.83 73.68 82.43 90.93 95.59 Table 3-20. The 5 x 5-meter grain size for isolated forested wetland buffers: principal component matrices showing pattern metric factor loadings. Metrics URB HETER SHAPE AG FOR DIST PLAND_Urb -0.49 -0.10 -0.32 -0.01 -0.08 -0.11 PLAND_Ag 0.31 -0.08 0.32 -0.49 -0.20 0.15 PLAND_For 0.05 0.18 -0.03 0.38 0.71 0.28 PLAND_Wet 0.41 0.16 0.26 0.28 -0.06 -0.26 PD -0.02 0.58 -0.04 -0.07 0.07 -0.21 ED 0.02 0.52 -0.27 0.13 -0.27 -0.08 AREA_CV 0.36 0.31 -0.12 -0.31 0.10 -0.17 SHAPE_MN 0.10 0.06 -0.55 0.06 -0.39 0.27 ENN_MN 0.00 0.26 0.19 0.05 -0.16 0.80 CONTAG 0.14 -0.09 -0.41 -0.50 0.40 0.14 PRD -0.37 0.31 0.14 -0.39 0.13 0.00 SHDI -0.44 0.21 0.33 -0.11 -0.04 -0.01 161

PAGE 162

Table 3-21. The 10 x 10-meter grain size for is olated forested wetland buffers: principal component matrices showing patte rn metric factor loadings. Metrics URB HETER CONTAG SHAPE FOR ENN PLAND_Urb -0.50 -0.09 -0.31 0.12 0.02 -0.09 PLAND_Ag 0.32 -0.13 0.24 -0.26 -0.54 0.16 PLAND_For 0.07 0.20 0.00 -0.13 0.73 0.29 PLAND_Wet 0.41 0.16 0.31 0.12 0.08 -0.25 PD -0.01 0.62 -0.09 -0.06 -0.08 -0.06 ED -0.07 0.53 0.00 0.40 -0.17 0.07 AREA_CV 0.29 0.37 -0.40 -0.05 -0.18 0.08 SHAPE_MN -0.13 -0.13 -0.07 0.65 -0.21 0.27 ENN_MN 0.03 0.02 0.35 0.11 0.06 0.75 CONTAG 0.09 -0.12 -0.55 -0.31 -0.10 0.39 PRD -0.40 0.25 0.12 -0.38 -0.20 0.13 SHDI -0.45 0.14 0.36 -0.23 -0.09 0.00 Table 3-22. The 20 x 20-meter grain size for is olated forested wetland buffers: principal component matrices showing patte rn metric factor loadings Metrics DIVERS HETER CONTAG AG SHAPE FOR PLAND_Urb -0.34 0.04 -0.44 -0.19 0.03 -0.30 PLAND_Ag 0.22 -0.15 0.22 0.62 -0.17 -0.01 PLAND_For 0.05 0.04 0.04 -0.41 0.08 0.81 PLAND_Wet 0.25 0.07 0.48 -0.03 0.18 0.04 PD 0.15 0.58 -0.03 0.02 0.15 -0.07 ED 0.05 0.56 0.05 -0.18 -0.16 -0.12 AREA_CV 0.42 0.31 -0.22 0.01 -0.05 -0.13 SHAPE_MN -0.13 -0.16 0.10 -0.33 -0.67 -0.09 ENN_MN 0.09 0.20 0.04 0.14 -0.65 0.17 CONTAG 0.22 -0.23 -0.50 0.17 -0.03 0.21 PRD -0.32 0.27 -0.19 0.38 -0.02 0.30 SHDI -0.47 0.21 0.00 0.27 -0.03 0.20 SHEI -0.41 0.05 0.42 0.02 0.09 -0.10 162

PAGE 163

Table 3-23. The 30 x 30-meter grain size for isolat ed forested wetland buffers the wetland buffer: principal component matrices showi ng pattern metric factor loadings Metrics DIVERS HETER URB/WET AG SHAPE FOR PLAND_Urb -0.19 0.32 -0.43 -0.21 -0.11 0.26 PLAND_Ag 0.18 -0.11 0.25 0.64 -0.20 -0.06 PLAND_For 0.01 -0.07 0.02 -0.37 0.35 -0.75 PLAND_Wet 0.08 -0.34 0.43 -0.04 0.29 0.12 PD -0.24 -0.52 -0.16 -0.09 0.01 0.16 ED -0.32 -0.42 -0.03 -0.25 -0.19 0.07 AREA_CV 0.28 -0.42 -0.26 -0.09 -0.16 0.15 SHAPE_MN -0.07 0.27 0.26 -0.30 -0.48 -0.07 ENN_MN 0.07 -0.19 -0.04 0.03 -0.65 -0.44 CONTAG 0.39 0.10 -0.40 0.13 0.09 -0.11 PRD -0.33 -0.09 -0.35 0.37 0.11 -0.22 SHDI -0.47 0.03 -0.14 0.28 0.02 -0.18 SHEI -0.44 0.12 0.32 0.10 0.01 0.03 Table 3-24. Coefficients of determination, proba bilities, and regression equations for multiple regressions between indicators of ecos ystem condition and significant components resulting from the PCA of landscape pattern me trics at four grain sizes for the sample isolated forested wetlands ( level of 0.05). Component s that were significantly related to the dependent vari able (p < 0.05) are indicated with an asterisk in the regression equation Variable (Y) n R2 (adj) p Regression equation 5-meter Water chemistry Log10(DO) 29 0.01 0.399 Y = 0.171 0.003( URB) + 0.045(HETER ) + 0.043(SHAPE) + 0.064(AG) + 0.043(FOR) Log10(SC) 17 0.15 0.247 Y = 2. 18 0.043(URB) + 0.021 (HETER) + 0.121(SHAPE) + 0.024(AG) 0.106(FOR) Log10(TN) 32 0.01 0.389 Y = 0. 226 + 0.02(URB) 0.043( HETER) + 0.037(SHAPE) 0.015(AG) 0.075(FOR) Log10(TP) 32 0.16 0.091 Y = 0.773 + 0.0 03(URB) 0.06(HETER ) + 0.081(SHAPE) 0.243(AG) 0.097(FOR) Log10(Turb) 32 0.00 0.637 Y = 0.581 0.107( URB) 0.047(HETER) + 0.006(SHAPE) + 0.011(AG) 0.008(FOR) WCI Macrophyte 51 0.42 <0.001 Y = 24.9 + 2.10(URB)* + 1.70(HETER) 2.01(SHAPE) + 7.29(AG)* 0.14(FOR) Macroinvertebrate 31 0.17 0.086 Y = 24.8 + 3.17(URB) + 0.77(HETER) 2.65(SHAPE) + 0.85(AG) 0.04(FOR) Diatom 21 0.12 0.229 Y = 35.9 + 1.88(URB) + 1.15(HETER) + 2.34(SHAPE) + 6.48(AG) + 1.83(FOR) 163

PAGE 164

Table 3-24. Continued. Variable (Y) n R2 (adj) p Regression equation 10-meter Water chemistry Log10(DO) 29 0.00 0.591 Y = 0.157 + 0.081(CONTAG) + 0.018(SHAPE) + 0.067(FOR) 0.017(DIST) 0.013(URB) + 0.015(HETER) Log10(SC) 17 0.05 0.402 Y = 2. 20 0.042(URB) + 0.013 (HETER) + 0.123(CONTAG) 0.019(SHAPE) 0.062(FOR) 0.009(DIST) Log10(TN) 32 0.07 0.262 Y = 0.221 + 0.018(URB) 0.045(HETER) + 0.047(CONTAG) + 0.035(SHAPE) 0.055(FOR) + 0.044(DIST) Log10(TP) 32 0.44 0.002 Y = 0.755 + 0.036(URB) 0.054(HETER) + 0.025(CONTAG) 0.089(SHAPE) 0.192(FOR)* + 0.318(DIST)* Log10(Turb) 32 0.00 0.79 Y = 0.584 0.098(UR B) 0.046(HETER) 0.012(CONTAG) + 0.007(SHAPE) + 0.023(FOR) + 0.019(DIST) WCI Macrophyte 51 0.44 <0.001 Y = 24.9 + 1.61(URB) + 1.51(HETER) + 0.96(CONTAG) + 5.94(SHAPE)* + 5.38(FOR)* + 0.42(DIST) Macroinvertebrate 31 0.16 0.119 Y = 24.8 + 2.88(URB) + 1.17(HETER) 2.05(CONTAG) + 2.35(SHAPE) + 0.24(FOR) + 0.78(DIST) Diatom 21 0.06 0.363 Y = 36.0 + 1.82(URB) + 0.84(HETER) + 3.91(CONTAG) + 0.63(SHAPE) + 4.77(FOR) 4.37(DIST) 20-meter Water chemistry Log10(DO) 29 0.01 0.531 Y = 0.171 0.02(DIVERS) + 0.026(HETER) + 0.056(CONTAG) 0.054(AG) 0.033(SHAPE) + 0.051(FOR) Log10(SC) 17 0.07 0.379 Y = 2.18 0.065(DIVERS) + 0.024(HETER) + 0.088(CONTAG) + 0.061(AG) 0.005(SHAPE) 0.078(FOR) Log10(TN) 32 0.02 0.379 Y = 0.225 0.007(DIVERS) 0.029(HETER) + 0.05(CONTAG) + 0.041(AG) 0.072(SHAPE) 0.049(FOR) Log10(TP) 32 0.43 0.002 Y = 0.725 + 0.042(DIVERS) 0.051(HETER) 0.041(CONTAG) + 0.245(AG)* 0.231(SHAPE)* + 0.075(FOR) Log10(Turb) 32 0.00 0.741 Y = 0.618 0.012(DIVERS) + 0.019(HETER) 0.111(CONTAG) + 0.038(AG) 0.027(SHAPE) + 0.055(FOR) WCI Macrophyte 51 0.39 <0.001 Y = 24.9 + 1.26(DIVERS) + 1.27 (HETER) + 2.62(CONTAG)* 6.91(AG)* 1.01(SHAPE) + 0.65(FOR) Macroinvertebrate 31 0.26 0.038 Y = 24.9 + 4.25(DIVERS)* + 1.48(HETER) 0.79(CONTAG) 0.62(AG) 0.07(SHAPE) + 0.38(FOR) Diatom 21 0.09 0.312 Y = 35.7 + 1.44 (DIVERS) + 1.26(HETER ) + 3.91(CONTAG) 3.90(AG) + 1.99(SHAPE) + 4.03(FOR) 164

PAGE 165

165 Table 3-24. Continued. Variable (Y) n R2 (adj) p R egression equation 30-meter Water chemistry Log10(DO) 29 0.01 0.591 Y = 0.157 0.013(DIVERS) + 0.015(HETER) + 0.081(URB/WET) + 0.018(AG) + 0.067(SHAPE) 0.017(FOR) Log10(SC) 17 0.08 0.369 Y = 2.19 0.0552(DIVERS) 0.0146(HETER) + 0.0416(URB/WET) + 0.103(AG) 0.003(SHAPE) + 0.065(FOR) Log10(TN) 32 0.17 0.094 Y = 0.225 0.021( DIVERS) + 0.038(HETER) + 0.079(URB/WET) + 0.046(AG) 0.097(SHAPE) + 0.016(FOR) Log10(TP) 32 0.45 0.001 Y = 0.745 + 0.026(DIVERS) + 0.039(HETER) + 0.009(URB/WET) + 0.206(AG)* 0.261(SHAPE)* 0.132(FOR) Log10(Turb) 32 0.01 0.592 Y = 0.607 0.035(DIVERS) + 0.086(HETER) 0.047(URB/WET) 0.010(AG) 0.079(SHAPE) 0.061(FOR) WCI Macrophyte 51 0.42 <0.001 Y = 24.9 0.493(DIVERS) 2.10(HETER)* + 2.76(URB/WET)* 6.82(AG)* + 1.53(SHAPE) 0.02(FOR) Macroinvertebrate 31 0.16 0. 119 Y = 24.4 + 1.65(DIVERS) 1.70(HETER) + 1.07(URB/WET) 2.10(AG) 1.73(SHAPE) + 0.53(FOR) Diatom 21 0.13 0.251 Y = 35.5 0.12(DIVERS) 2. 77(HETER) + 3.68(URB/WET) 2.83(AG) + 4.95(SHAPE) 3.02(FOR)

PAGE 166

Table 3-25. Pearsons correlations between landscape pattern metrics calculated at four different grain sizes (meters on a side ) for the sample streams. Some of the metric s descriptors (acronyms) of the column s have been renamed for convenience. Metric % Urba %Agb %Forc % Wetd PD ED A_MNe A_CVf SHAPg ENNh CONTi IJI PR PRD SHDI 20-m PLAND_Ag -0.15 PLAND_For -0.41 -0.66 PLAND_Wet -0.33 -0.09 -0.15 PD 0.04 -0.03 -0.22 0.21 ED 0.05 -0.26 0.06 0.10 0.83 AREA_MN -0.02 0.03 0.25 -0.29 -0.96 -0.83 AREA_CV -0.35 0.33 -0.13 0.26 -0.06 -0.46 0.10 SHAPE_MN -0.31 -0.15 0.54 -0.28 -0.64 -0.31 0.71 -0.16 ENN_MN -0.36 0.29 -0.11 0.11 -0.38 -0.46 0.35 0.15 0.22 CONTAG -0.18 0.04 0.37 -0.18 -0.60 -0.62 0.62 0.46 0.36 0.04 IJI 0.41 0.19 -0.59 -0.05 0.30 0.07 -0.29 -0.13 -0.47 0.05 -0.67 PR -0.10 0.28 -0.38 0.26 -0.09 -0.35 0.03 0.40 -0.19 0.69 -0.07 0.23 PRD 0.44 -0.25 -0.03 -0.15 0.34 0.53 -0.36 -0.72 -0.28 -0.23 -0.36 -0.03 -0.40 SHDI 0.01 0.18 -0.51 0.35 0.25 0.05 -0.31 0.03 -0.33 0.52 -0.67 0.60 0.77 -0.10 SHEI 0.26 0.02 -0.48 0.18 0.48 0.45 -0.50 -0.42 -0.38 0.07 -0.98166 0.76 0.20 0.33 0.76 50-m PLAND_Ag -0.15 PLAND_For -0.41 -0.66 PLAND_Wet -0.33 -0.09 -0.15 PD 0.03 -0.05 -0.18 0.19 ED 0.07 -0.33 0.13 0.07 0.82 AREA_MN -0.01 0.06 0.20 -0.27 -0.96 -0.81 AREA_CV -0.39 0.37 -0.08 0.19 -0.20 -0.54 0.19 SHAPE_MN -0.19 -0.21 0.51 -0.30 -0.65 -0.26 0.73 -0.18 ENN_MN -0.41 0.31 -0.14 0.18 -0.29 -0.47 0.28 0.28 0.13 CONTAG -0.12 0.13 0.24 -0.17 -0.67 -0.70 0.66 0.57 0.27 0.11 IJI 0.40 0.21 -0.64 -0.02 0.36 0.15 -0.35 -0.30 -0.40 0.11 -0.70 PR -0.11 0.29 -0.38 0.25 -0.10 -0.44 0.05 0.46 -0.25 0.69 0.07 0.26 PRD 0.45 -0.25 -0.03 -0.16 0.34 0.56 -0.34 -0.76 -0.13 -0.24 -0.37 0.24 -0.43 SHDI -0.01 0.17 -0.49 0.35 0.25 -0.03 -0.29 0.02 -0.35 0.54 -0.54 0.66 0.77 -0.13 SHEI 0.26 -0.02 -0.44 0.17 0.50 0.45 -0.50 -0.52 -0.31 0.05 -0.93 0.84 0.15 0.34 0.72

PAGE 167

167Table 3-25. Continued. Metric % Urba % Agb % Forc % Wetd PD ED A_MNe A_CVf SHAPg ENNh CONTi IJI PR PRD SHDI 80-m PLAND_Ag -0.15 PLAND_For -0.40 -0.66 PLAND_Wet -0.33 -0.09 -0.15 PD 0.03 -0.10 -0.13 0.15 ED 0.11 -0.35 0.13 0.07 0.84 AREA_MN 0.00 0.11 0.15 -0.23 -0.96 -0.83 AREA_CV -0.42 0.34 -0.03 0.20 -0.31 -0.61 0.31 SHAPE_MN -0.27 -0.22 0.53 -0.18 -0.55 -0.22 0.62 -0.07 ENN_MN -0.47 0.34 -0.16 0.24 -0.21 -0.47 0.18 0.37 0.04 CONTAG -0.16 0.19 0.17 -0.14 -0.70 -0.75 0.70 0.66 0.23 0.21 IJI 0.56 0.17 -0.63 -0.06 0.34 0.19 -0.33 -0.41 -0.47 -0.08 -0.62 PR -0.13 0.29 -0.38 0.25 -0.10 -0.44 0.06 0.47 -0.24 0.72 0.18 0.16 PRD 0.44 -0.25 -0.03 -0.15 0.39 0.59 -0.38 -0.78 -0.17 -0.27 -0.43 0.31 -0.42 SHDI -0.02 0.17 -0.49 0.35 0.27 -0.02 -0.30 0.01 -0.32 0.56 -0.43 0.54 0.78 -0.12 SHEI 0.33 -0.05 -0.42 0.15 0.53 0.51 -0.54 -0.61 -0.29 -0.07 -0.91 0.82 0.07 0.40 0.64 110-m PLAND_Ag -0.15 PLAND_For -0.40 -0.66 PLAND_Wet -0.30 -0.05 -0.18 PD 0.10 -0.12 -0.14 0.14 ED 0.18 -0.36 0.12 0.04 0.85 AREA_MN -0.07 0.12 0.17 -0.23 -0.96 -0.83 AREA_CV -0.42 0.31 0.01 0.17 -0.46 -0.67 0.42 SHAPE_MN -0.28 -0.24 0.56 -0.16 -0.58 -0.25 0.66 0.01 ENN_MN -0.48 0.37 -0.19 0.33 -0.21 -0.48 0.17 0.44 -0.04 CONTAG -0.18 0.19 0.17 -0.18 -0.76 -0.76 0.75 0.70 0.27 0.27 IJI 0.41 0.20 -0.62 0.14 0.42 0.19 -0.41 -0.39 -0.51 0.04 -0.65 PR -0.13 0.29 -0.38 0.35 -0.15 -0.49 0.09 0.48 -0.28 0.69 0.22 0.22 PRD 0.45 -0.23 -0.04 -0.13 0.48 0.64 -0.45 -0.76 -0.22 -0.28 -0.43 0.27 -0.44 SHDI -0.02 0.17 -0.50 0.48 0.25 -0.07 -0.31 -0.01 -0.40 0.53 -0.40 0.64 0.77 -0.14 SHEI 0.35 -0.06 -0.41 0.23 0.60 0.54 -0.60 -0.67 -0.35 -0.14 -0.92 0.83 0.02 0.42 0.60 a % Urb = PLAND_Urb; b % Ag = PLAND_Ag; c % For = PLAND_For; d % Wet = PLAND_WET; e A_MN = Area_MN; f A_CV = AREA_CV; g SHAP = SHAPE_MN; h ENN = ENN_MN; i CONT = CONTAG.

PAGE 168

Table 3-26. Eigenvalues and vari ance explained by the first six axes for the principal components analysis at four different gr ain sizes (meters on a side) for stream watersheds. Component 1 2 3 4 5 6 20-m Eigenvalue 4.08 3.61 1.81 1.57 0.93 0.68 % of variance explained 29.13 25.78 12.95 11.20 6.64 4.84 Cumulative % of variance 29.13 54.91 67.85 79.06 85.69 90.54 50-m Eigenvalue 4.10 3.83 1.79 1.52 0.89 0.67 % of variance explained 29.25 27.35 12.80 10.86 6.34 4.78 Cumulative % of variance 29.25 56.60 79.39 80.25 86.61 91.39 80-m Eigenvalue 4.34 3.73 1.83 1.28 0.88 0.69 % of variance explained 31.11 26.61 13.05 9.11 6.31 4.95 Cumulative % of variance 31.11 57.71 70.76 79.87 86.18 91.12 110-m Eigenvalue 4.58 3.92 1.69 1.09 0.83 0.65 % of Variance explained 32.69 28.00 12.05 7.80 5.92 4.65 Cumulative % of variance 32.69 60.68 72.74 80.53 86.45 91.10 Table 3-27. The 20 x 20-m grain size for stream watersheds: principa l component matrices showing pattern metric factor loadings. Metrics DIVERS1 DIVERS2 WET DIST PLAND_Urb -0.20 0.18 0.46 -0.21 PLAND_Ag -0.09 -0.28 0.14 -0.37 PLAND_For 0.35 0.21 -0.19 0.33 PLAND_Wet -0.11 -0.15 -0.50 0.11 PD -0.33 0.18 -0.40 -0.13 ED -0.23 0.35 -0.33 0.10 AREA_CV 0.11 -0.34 -0.28 -0.40 SHAPE_MN 0.36 0.00 0.19 0.35 ENN_MN 0.01 -0.38 0.12 0.41 IJI -0.39 -0.07 0.24 -0.02 PR -0.14 -0.43 -0.02 0.17 PRD -0.19 0.36 0.15 0.13 SHDI -0.35 -0.29 -0.03 0.32 SHEI -0.43 0.02 0.04 0.28 168

PAGE 169

Table 3-28. The 50 x 50-m grain size for stream watersheds: principa l component matrices showing pattern metric factor loadings. Metrics DIVERS1 DIVERS2 WET SHAPE PLAND_Urb -0.21 -0.15 0.48 -0.15 PLAND_Ag -0.03 0.29 0.20 -0.38 PLAND_For 0.30 -0.26 -0.26 0.31 PLAND_Wet -0.08 0.16 -0.49 0.07 PD -0.35 -0.11 -0.39 -0.22 ED -0.27 -0.32 -0.34 -0.01 AREA_CV 0.23 0.33 -0.18 -0.36 SHAPE_MN 0.30 -0.11 0.22 0.45 ENN_MN 0.04 0.37 0.01 0.39 IJI -0.41 0.12 0.22 0.08 PR -0.07 0.44 -0.02 0.15 PRD -0.24 -0.32 0.13 0.10 SHDI -0.31 0.33 -0.07 0.30 SHEI -0.44 0.02 0.01 0.27 Table 3-29. The 80 x 80-m grain size for stream watersheds: principa l component matrices showing pattern metric factor loadings. Metrics DIVERS1 DIVERS2 URB SHAPE PLAND_Urb -0.28 -0.01 -0.48 0.08 PLAND_Ag 0.10 0.29 -0.29 -0.33 PLAND_For 0.14 -0.39 0.31 0.19 PLAND_Wet 0.03 0.18 0.46 -0.09 PD -0.33 0.06 0.36 -0.35 ED -0.36 -0.15 0.34 -0.17 AREA_CV 0.37 0.17 0.02 -0.35 SHAPE_MN 0.21 -0.26 0.00 0.52 ENN_MN 0.22 0.32 0.18 0.29 IJI -0.34 0.26 -0.22 0.13 PR 0.14 0.43 0.07 0.18 PRD -0.35 -0.15 -0.03 0.11 SHDI -0.10 0.44 0.20 0.30 SHEI -0.38 0.20 0.10 0.26 169

PAGE 170

170 Table 3-30. The 110 x 110-m grain size for stream watersheds: principal component matrices showing pattern metric factor loadings. Metrics DIVERS1 DIVERS2 URB SHAPE PLAND_Urb -0.26 -0.03 -0.48 -0.29 PLAND_Ag 0.09 0.28 -0.37 0.41 PLAND_For 0.14 -0.37 0.38 -0.12 PLAND_Wet 0.00 0.24 0.47 -0.02 PD -0.36 0.04 0.27 0.41 ED -0.37 -0.17 0.27 0.26 AREA_CV 0.38 0.16 -0.01 0.27 SHAPE_MN 0.22 -0.26 0.09 -0.47 ENN_MN 0.20 0.33 0.20 -0.02 IJI -0.31 0.29 -0.13 -0.20 PR 0.14 0.42 0.07 -0.16 PRD -0.35 -0.15 -0.03 0.03 SHDI -0.11 0.44 0.21 -0.27 SHEI -0.39 0.17 0.10 -0.26

PAGE 171

Table 3-31. Pearsons correlations between landscape pattern metrics calculated at three spatial extents fo r the sample streams Some of the metrics descriptors (acronyms) of the columns have been renamed for convenience. Metric % Urba % Agb % Forc % Wetd PD ED A_MNe A_CVf SHAPg ENNh CONTi IJI PR PRD SHDI 100-m buffer PLAND_Ag 0.12 PLAND_For -0.31 -0.62 PLAND_Wet -0.36 -0.16 -0.32 PD 0.08 -0.18 0.17 0.15 ED -0.12 -0.33 0.29 0.19 0.68 AREA_MN -0.02 0.24 -0.22 -0.19 -0.98 .66 -0 AREA_CV -0.32 0.01 -0.14 0.42 -0.04 0.04 0.04 SHAPE_MN -0.20 -0.08 0.00 0.04 -0.53 0.23 0.56 0.19 ENN_MN 0.33 0.42 -0.58 0.21 -0.30 -0.40 0.32 0.18 0.04 CONTAG -0.20 0.16 0.03 0.07 -0.40 -0.36 0.42 0.43 0.22 0.32 IJI 0.29 0.20 -0.46 0.08 0.20 -0.27 -0.20 -0.09 -0.55 0.27 -0.51 PR 0.37 0.36 -0.52 0.24 -0.12 -0.24 0.15 0.24 0.01 0.84 0.30 0.23 PRD 0.56 0.06 -0.03 -0.28 0.33 0.09 -0.29 -0.66 -0.34 -0.01 -0.29 0.17 0.01 SHDI 0.56 0.31 -0.53 0.13 0.15 -0.11 -0.14 -0.21 -0.30 0.61 -0.35 0.58 0.72 0.36 SHEI 0.37 0.02 -0.24 -0.10 0.24 0.06 -0.26 -0.48 -0.33 -0.06 -0.93171 0.67 -0.06 0.38 0.59 400-m buffer PLAND_Ag 0.02 PLAND_For -0.38 -0.74 PLAND_Wet -0.23 -0.13 -0.25 PD 0.18 0.03 -0.18 0.25 ED -0.04 -0.21 0.21 0.14 0.81 AREA_MN -0.14 -0.01 0.16 -0.28 -0.99 .81 -0 AREA_CV -0.19 0.26 -0.35 0.40 0.06 -0.20 -0.06 SHAPE_MN -0.31 -0.19 0.38 -0.13 -0.63 -0.14 0.65 -0.22 ENN_MN 0.44 0.27 -0.48 0.13 -0.21 -0.40 0.20 0.19 -0.09 CONTAG -0.32 0.10 0.18 -0.21 -0.51 -0.48 0.51 0.42 0.31 0.01 IJI 0.48 0.25 -0.59 0.24 0.38 0.07 -0.37 -0.06 -0.51 0.37 -0.72 PR 0.39 0.32 -0.57 0.29 0.07 -0.21 -0.07 0.45 -0.23 0.80 0.04 0.38 PRD 0.50 -0.04 -0.04 -0.22 0.36 0.23 -0.37 -0.51 -0.38 0.08 -0.35 0.31 -0.05 SHDI 0.55 0.19 -0.58 0.36 0.34 0.11 -0.35 0.01 -0.36 0.63 -0.64 0.77 0.72 0.24 SHEI 0.45 0.01 -0.34 0.23 0.40 0.31 -0.40 -0.39 -0.31 0.21 -0.96 0.81 0.16 0.40 0.79

PAGE 172

172Table 3-31. Continued. Metric % Urba % Agb % Forc % Wetd PD ED A_MNe A_CVf SHAPg ENNh CONTi IJI PR PRD SHDI Watershed PLAND_Ag -0.04 PLAND_For -0.41 -0.72 PLAND_Wet -0.22 -0.20 -0.17 PD 0.09 -0.01 -0.28 0.31 ED -0.03 -0.25 0.06 0.19 0.83 AREA_MN -0.04 0.01 0.30 -0.38 -0.96 .82 -0 AREA_CV -0.15 0.31 -0.24 0.21 -0.02 -0.43 0.06 SHAPE_MN -0.22 -0.20 0.55 -0.41 -0.66 -0.27 0.73 -0.30 ENN_MN 0.28 0.24 -0.28 -0.15 -0.45 -0.48 0.40 -0.10 0.07 CONTAG -0.35 0.05 0.40 -0.21 -0.60 -0.61 0.61 0.46 0.36 0.02 IJI 0.47 0.23 -0.58 -0.01 0.30 0.03 -0.28 -0.04 -0.44 0.24 -0.68 PR 0.30 0.24 -0.50 0.18 -0.06 -0.31 -0.01 0.31 -0.31 0.67 -0.11 0.34 PRD 0.21 -0.21 -0.09 -0.04 0.36 0.54 -0.38 -0.66 -0.19 0.02 -0.37 0.08 -0.29 SHDI 0.45 0.13 -0.62 0.28 0.29 0.10 -0.35 -0.07 -0.44 0.46 -0.72 0.70 0.76 0.02 SHEI 0.43 0.02 -0.50 0.20 0.48 0.44 -0.50 -0.41 -0.38 0.15 -0.98 0.76 0.26 0.32 0.82 a % Urb = PLAND_Urb; b % Ag = PLAND_Ag; c % For = PLAND_For; d % Wet = PLAND_WET; e A_MN = Area_MN; f A_CV = AREA_CV; g SHAP = SHAPE_MN; h ENN = ENN_MN; i CONT = CONTAG.

PAGE 173

Table 3-32. Eigenvalues and vari ance explained by the first six axes for the principal components analysis at three different spatial extents for streams. Component 1 2 3 4 5 6 100-meter buffer Eigenvalue 4.21 3.06 1.91 1.32 0.94 0.86 % of variance explained 30.05 21.87 13.62 9.39 6.71 6.17 Cumulative % of variance 30.05 51.92 65.54 74.93 81.64 87.82 400-meter buffer Eigenvalue 4.84 2.91 1.97 1.35 0.96 0.69 % of variance explained 34.56 20.75 14.05 9.62 6.83 4.94 Cumulative % of variance 34.56 55.31 69.36 78.98 85.81 90.75 Watershed Eigenvalue 4.42 3.22 2.14 1.25 0.94 0.81 % of Variance explained 31.59 23.00 15.30 8.89 6.71 5.76 Cumulative % of variance 31.59 54.60 69.90 78.79 85.49 91.25 Table 3-33. The 100-meter spatial extent for stre ams: principal component matrices showing pattern metric factor loadings. Metric DIVERS SIZE WET HETER PLAND_Urb -0.32 0.17 0.18 -0.34 PLAND_Ag -0.25 -0.18 0.18 0.10 PLAND_For 0.35 0.23 0.08 -0.06 PLAND_Wet -0.01 -0.22 -0.55 0.03 PD -0.01 0.35 -0.49 -0.11 ED 0.18 0.21 -0.41 -0.45 AREA_CV 0.10 -0.38 -0.34 0.08 SHAPE_MN 0.19 -0.28 0.16 -0.43 ENN_MN -0.34 -0.32 0.01 -0.18 IJI -0.34 0.12 -0.18 0.44 PR -0.33 -0.28 -0.11 -0.30 PRD -0.20 0.38 0.14 -0.29 SHDI -0.44 0.04 -0.14 -0.16 SHEI -0.26 0.33 -0.07 0.18 173

PAGE 174

Table 3-34. The 400-meter spatial extent for stre ams: principal component matrices showing pattern metric factor loadings. Metric DIVERS HETER WET AG PLAND_Urb -0.29 0.09 -0.36 -0.05 PLAND_Ag -0.17 -0.25 0.01 -0.56 PLAND_For 0.34 0.23 -0.06 0.27 PLAND_Wet -0.14 -0.08 0.47 0.45 PD -0.21 0.35 0.40 -0.19 ED -0.03 0.43 0.33 0.03 AREA_CV -0.05 -0.38 0.44 -0.11 SHAPE_MN 0.27 -0.13 -0.20 0.36 ENN_MN -0.28 -0.30 -0.23 0.20 IJI -0.39 0.10 -0.03 0.04 PR -0.31 -0.31 0.02 0.15 PRD -0.17 0.35 -0.28 -0.19 SHDI -0.42 0.00 -0.03 0.27 SHEI -0.33 0.27 -0.08 0.24 Table 3-35. The watershed spatial extent for streams: principal component matrices showing pattern metric factor loadings. Metric DIVERS HETER WET AG PLAND_Urb -0.24 0.12 -0.29 0.05 PLAND_Ag -0.06 0.34 0.03 0.57 PLAND_For 0.30 -0.36 -0.07 -0.23 PLAND_Wet -0.12 -0.01 0.40 -0.49 PD -0.33 -0.18 0.37 0.13 ED -0.25 -0.39 0.20 0.04 AREA_CV 0.09 0.33 0.38 0.15 SHAPE_MN 0.32 -0.09 -0.32 -0.12 ENN_MN -0.05 0.27 -0.46 -0.14 IJI -0.35 0.22 -0.10 -0.05 PR -0.26 -0.35 -0.21 0.26 PRD -0.23 -0.37 -0.23 0.20 SHDI -0.36 0.23 -0.09 -0.36 SHEI -0.42 0.02 -0.10 -0.24 174

PAGE 175

Table 3-36. Coefficients of determination, proba bilities, and regression equations for multiple regressions between indicators of ecos ystem condition and significant components resulting from the PCA of landscape pattern me trics at four grain sizes for the sample streams ( level of 0.05). Components that were significantly related to the dependent variable (p < 0.05) are indicated with an as terisk in the regression equation. Variable (Y) n R2 (adj) p Regression equation 20-meter Water chemistry Log10(Turbidity) 32 0.00 0.664 Y = 0.339 + 0.0228(DIVERS1) 0.0468(DIVERS2) 0.0662(WET) + 0.0367(DIST) DO 36 0.18 0.033 Y = 6.33 + 0.265(DIVERS1) + 0.183(DIVERS2) 0.397(WET) + 0.309(DIST) Log10(NO3-N) 45 0.12 0.055 Y = 0.952 0.124(DIVERS1)* + 0.0303(DIVERS2) + 0.106(WET) 0.0998(DIST) Log10(TN) 46 0.60 <0.001 Y = 0.0806 0.0621(DIVERS1)* 0.0608(DIVERS2)* 0.036(WET) 0.0614(DIST)* Log10(TP) 46 0.42 <0.001 Y = 1.28 0.0553(DIVERS1) 0.0776(DIVERS2) 0.0992(WET) 0.193(DIST)* WQI 36 0.11 0.105 Y = 35.0 0.343(DIVERS1) 0.74(DIVERS2) + 0.93(WET) 3.09(DIST)* SCI SCI_1 66 0.01 0.354 Y = 28.5 + 0.372(DIVERS1) + 0.157(DIVERS2) 0.120(WET) + 0.235(DIST) SCI_2 67 0.03 0.205 Y = 61.8 + 1.89(DIVERS1) + 1.03(DIVERS2) 3.03(WET) + 0.71(DIST) 50-meter Water chemistry Log10(Turbidity) 32 0.00 0.769 Y = 0.36 + 0.0278(DIVERS1) + 0.0405(DIVERS2) 0.0572(WET) + 0.0156(SHAPE) DO 36 0.20 0.028 Y = 6.37 + 0.196(DIVERS1) 0.248(DIVERS2) 0.466(WET) + 0.224(SHAPE) Log10(NO3-N) 45 0.13 0.049 Y = 0.958 0.123(DIVERS1)* 0.009(DIVERS2) + 0.118(WET) 0.101(SHAPE) Log10(TN) 46 0.57 <0.001 Y= 0.0874 0.0502(DIVERS1)* + 0.074(DIVERS2)* 0.137(WET) 0.0578(SHAPE)* Log10(TP) 46 0.41 <0.001 Y = 1.29 0.0359(DIVERS1) + 0.0913(DIVERS2)* 0.0421(WET) 0.212(SHAPE)* WQI 36 0.14 0.068 Y = 34.5 + 0.188(DIVERS1) 1.06(DIVERS2) + 1.95(WET) 2.87(SHAPE) SCI SCI_1 66 0.01 0.342 Y = 28.5 + 0.356(DIVERS1) 0.188(DIVERS2) 0.081(WET) + 0.27(SHAPE) SCI_2 67 0.02 0.272 Y = 61.8 + 1.68(DIVERS1) 1.23(DIVERS2) 2.7(WET)* + 0.42(SHAPE) 175

PAGE 176

Table 3-36. Continued. Variable (Y) n R2 (adj) p Regression equation 80-meter Water chemistry Log10(Turbidity) 32 0.00 0.735 Y = 0.343 + 0.0447(DIVERS1) + 0.0186(DIVERS2) + 0.0652(URB) 0.0109(SHAPE) DO 36 0.19 0.032 Y = 6.38 + 0.076(DIVERS1) 0.339(DIVERS2)* + 0.483(URB) 0.024(SHAPE) Log10(NO3-N) 45 0.12 0.057 Y = 0.962 0.106(DIVERS1) + 0.0557(DIVERS2) 0.124(URB) 0.0764(SHAPE) Log10(TN) 46 0.54 <0.001 Y = 0.0898 0.0037(DIVERS1)* + 0.0934(DIVERS2)* + 0.001(URB) 0.0573(SHAPE)* Log10(TP) 46 0.36 <0.001 Y = 1.29 0.0249(DIVERS1) + 0.108(DIVERS2)* 0.0136(URB) 0.208(SHAPE)* WQI 36 0.11 0.106 Y = 34.6 + 0.64(DIVERS1) + 0.904(DIVERS2) 2.51(URB) 2.91(SHAPE) SCI SCI_1 66 0.01 0.337 Y = 28.5 + 0.205(DIVERS1) 0.361(DIVERS2) + 0.086(URB) + 0.265(SHAPE) SCI_2 67 0.02 0.288 Y = 61.8 + 1.13(DIVERS1) 2.02(DIVERS2) + 2.39(URB) 0.42(SHAPE) 110-meter Water chemistry Log10(Turbidity) 32 0.00 0.701 Y = 0.343 + 0.0382(DIVERS1) + 0.0224(DIVERS2) + 0.0575(URB) + 0.0403(SHAPE) DO 36 0.22 0.020 Y = 6.42 + 0.039(DIVERS1) 0.326(DIVERS2)* + 0.505(URB) + 0.210(SHAPE) Log10(NO3-N) 45 0.13 0.049 Y = 0.962 0.0983(DIVERS1) + 0.0441(DIVERS2) 0.152(URB) + 0.0666(SHAPE) Log10(TN) 46 0.52 <0.001 Y = 0.0947 0.0044(DIVERS1) + 0.0953(DIVERS2)* 0.0122(URB) + 0.043(SHAPE) Log10(TP) 46 0.32 <0.001 Y = 1.30 + 0.0318(DIVERS1) + 0.118(DIVERS2)* 0.044(URB) + 0.176(SHAPE)* WQI 36 0.14 0.066 Y = 34.6 + 0.68(DIVERS1) + 0.883(DIVERS2) 3.18(URB) 1.28(SHAPE) SCI SCI_1 66 0.01 0.376 Y = 28.5 + 0.177(DIVERS1) 0.330(DIVERS2) + 0.190(URB) 0.266(SHAPE) SCI_2 67 0.02 0.298 Y = 61.8 + 1.06(DIVERS1) 1.78(DIVERS2) + 2.76(URB) + 0.76(SHAPE) 176

PAGE 177

Table 3-37. Coefficients of determination, proba bilities, and regression equations for multiple regressions between indicators of ecos ystem condition and significant components resulting from the PCA of landscape pattern metrics at three spatial extents for the sample streams ( level of 0.05). Components that we re significantly related to the dependent variable (p < 0.05) are indicated with an as terisk in the regression equation. Variable (Y) n R2 (adj) p Regression equation 100-meter Water chemistry Log10(Turbidity) 34 0.24 0.016 Y = 0.354 + 0.0954(DIVERS)* 0.0909(SIZE)* 0.0416(WET) 0.0577(HETER) DO 35 0.34 0.002 Y = 6.16 + 0.44(DIVERS)* 0.051(SIZE) 0.512(WET)* + 0.14(HETER) Log10(NO3-N) 41 0.00 0.549 Y = 0.998 0.0399(DIVERS) 0.0475(SIZE) 0.0378(WET) 0.0891(HETER) Log10(TN) 43 0.35 <0.001 Y = 0.0744 0.076(DIVERS)* 0.0328(SIZE) + 0.0136(WET) 0.0259(HETER) Log10(TP) 43 0.18 0.021 Y = 1.32 0.0652(DIVERS) 0.112(SIZE) 0.0715(WET) + 0.0583(HETER) WQI 35 0.13 0.086 Y = 35.3 0.37(DIVERS) 1.33(SIZE) + 0.25(WET)* + 0.06(HETER) SCI SCI_1 62 0.17 0.005 Y = 28.4 + 0.658(DIVERS)* 0.32(SIZE) 0.561(WET) + 0.175(HETER) SCI_2 63 0.22 0.001 Y = 61.7 + 3.5(DIVERS)* 1.43(SIZE) 4.87(WET)* + 1.20(HETER) 400-meter Water chemistry Log10(Turbidity) 34 0.00 0.713 Y = 0.405 + 0.0421(DIVERS) 0.0105(HETER) + 0.0311(WET) + 0.0011(AG) DO 35 0.22 0.020 Y = 6.30 + 0.294(DIVERS)* + 0.252(HETER) + 0.236(WET) + 0.342(AG) Log10(NO3-N) 42 0.08 0.139 Y = 0.931 0.101(DIVERS)* + 0.0312(HETER) 0.0225(WET) 0.122(AG) Log10(TN) 43 0.53 <0.001 Y= 0.0730 0.0710(DIVERS)* 0.0409(HETER)* + 0.0361(WET) 0.0447(AG) Log10(TP) 43 0.41 <0.001 Y = 1.29 0.0565(DIVERS) 0.0874(HETER)* + 0.148(WET)* 0.150(AG)* WQI 35 0.29 0.006 Y = 35.6 0.496(DIVERS) 1.60(HETER) 0.84(WET) 4.38(AG)* SCI SCI_1 62 0.08 0.077 Y = 28.4 + 0.484(DIVERS)* + 0.002(HETER) + 0.405(WET) + 0.405(AG) SCI_2 63 0.11 0.027 Y = 61.7 + 2.32(DIVERS)* + 0.2.32(HETER) + 3.71(WET)* + 2.91(AG) 177

PAGE 178

178 Table 3-37. Continued. Variable (Y) n R2 (adj) p Regression equation Watershed Water chemistry Log10(Turbidity) 34 0.00 0.893 Y = 0.430 + 0.0158(DIVERS) 0.0274(HETER) + 0.0036(WET) + 0.0017(AG) DO 35 0.18 0.043 Y = 6.40 + 0.267(DIVERS) 0.336(HETER)* + 0.146(WET) 0.259(AG) Log10(NO3-N) 42 0.24 0.007 Y = 0.968 0.133(DIVERS)* 0.0202(HETER) 0.0698(WET) 0.251(AG) Log10(TN) 43 0.59 <0.001 Y = -0.0738 0.0687(DIVERS)* 0.0618(HETER)* + 0.0451(WET)* + 0.0346(AG) Log10(TP) 43 0.44 <0.001 Y = 1.29 0.0398(DIVERS) + 0.104(HETER)* + 0.165(WET)* + 0.1294(AG)* WQI 35 0.17 0.046 Y = 35.2 0.712(DIVERS) +1 .26(HETER) 0.11(WET) + 3.96(AG)* SCI SCI_1 62 0.06 0.123 Y = 28.4 + 0.429(DIVERS)* 0.202(HETER) + 0.145(WET) 0.513(AG) SCI_2 63 0.07 0.083 Y = 61.7 + 2.17(DIVERS) 1.50(HETER) + 2.56(WET) 2.58(AG)

PAGE 179

Table 3-38. Pearsons correlations between landscape pattern metrics calculated at four different grain sizes (meters on a side ) for the sample lakes. Some of the metrics descriptors (acronyms) of the columns have been renamed for convenience. Metric % Urba % Agb % Forc % Wetd PD ED A_MNe A_CVf SHAPg ENNh CONTi IJI PR PRD SHDI 20-m PLAND_Ag 01 0.708 .3 5 PLAND_For .0 PLAND_Wet -0.69 -0.02 0.35 PD 0.10 0.00 0.02 -0.03 ED -0.09 -0.05 0.21 0.16 0.56 AREA_MN -0.10 0.00 -0.02 0.03 -1.00 6 -0 .5 AREA_CV -0.04 -0.12 -0.23 0.39 -0.04 -0.11 0.04 SHAPE_MN -0.33 0.04 0.38 0.07 -0.31 0.29 0.31 -0.26 ENN_MN 0.17 0.21 -0.08 -0.16 -0.45 -0.29 0.45 -0.03 -0.01 CONTAG 0.32 -0.30 -0.21 -0.23 -0.43 -0.68 0.43 0.28 -0.12 0.09 IJI -0.09 0.32 0.06 -0.06 0.27 0.29 -0.27 -0.13 0.01 0.23 -0.56 PR 0.21 -0.03 0.06 0.11 0.02 0.32 -0.02 0.38 -0.01 0.47 -0.15 0.28 PRD 0.19 0.07 0.01 -0.45 0.53 0.11 -0.53 -0.68 -0.11 -0.25 -0.12 0.10 -0.46 SHDI -0.04 0.19 0.15 0.21 0.23 0.56 -0.23 0.08 0.01 0.33 -0.72 0.54 0.78 -0.25 SHEI -0.30 0.33 0.18 0.22 0.36 0.55 -0.36 -0.25 0.05 0.04 -0.98179 01 0.707 0.58 0.23 0.07 0.78 40-m PLAND_Ag .3 5 PLAND_For .0 PLAND_Wet -0.67 -0.05 0.35 PD 0.10 0.06 0.02 -0.14 ED -0.11 -0.04 0.24 0.19 0.53 AREA_MN -0.14 0.00 -0.06 0.15 -0.96 2 -0 .5 AREA_CV 0.06 -0.20 -0.24 0.28 -0.18 -0.08 0.15 SHAPE_MN -0.46 0.05 0.38 0.32 -0.52 0.14 0.55 -0.03 ENN_MN 0.19 0.03 -0.08 -0.02 -0.37 -0.18 0.36 0.18 0.01 CONTAG 0.31 -0.27 -0.24 -0.21 -0.49 -0.71 0.45 0.35 -0.02 0.08 IJI -0.04 0.25 0.05 -0.06 0.31 0.31 -0.37 -0.09 -0.12 0.22 -0.52 PR 0.21 -0.05 0.05 0.12 -0.03 0.36 -0.01 0.44 0.07 0.60 -0.11 0.26 PRD 0.16 0.12 0.02 -0.44 0.61 0.05 -0.63 -0.65 -0.45 -0.39 -0.14 0.14 -0.46 SHDI -0.03 0.16 0.15 0.21 0.22 0.61 -0.22 0.07 0.03 0.46 -0.68 0.50 0.78 -0.25 SHEI -0.27 0.31 0.19 0.20 0.39 0.56 -0.35 -0.33 -0.03 0.10 -0.96 0.54 0.20 0.08 0.76

PAGE 180

180Table 3-38. Continued. Metric % Urba % Agb % Forc % Wetd PD ED A_MNe A_CVf SHAPg ENNh CONTi IJI PR PRD SHDI 6 0-m PLAND_Ag 00 088 .3 .5 PLAND_For -0 .0 PLAND_Wet -0.69 -0.02 0.36 PD 0.06 0.14 0.03 -0.19 ED -0.17 -0.04 0.27 0.26 0.45 AREA_MN -0.09 -0.11 -0.05 0.20 -0.96 6 -0 .4 AREA_CV 0.04 -0.13 -0.26 0.21 -0.28 -0.06 0.21 SHAPE_MN -0.50 -0.05 0.40 0.50 -0.64 0.05 0.67 -0.04 ENN_MN 0.30 -0.02 -0.09 -0.10 -0.24 -0.03 0.25 0.19 -0.10 CONTAG 0.27 -0.21 -0.24 -0.20 -0.54 -0.70 0.49 0.42 0.03 0.01 IJI 0.09 0.16 -0.05 -0.17 0.50 0.30 -0.55 -0.10 -0.46 0.21 -0.51 PR 0.15 0.04 0.05 0.13 -0.03 0.40 -0.04 0.48 -0.09 0.61 -0.08 0.23 PRD 0.15 0.13 0.02 -0.44 0.64 -0.02 -0.64 -0.64 -0.49 -0.37 -0.18 0.27 -0.45 SHDI -0.04 0.18 0.16 0.22 0.23 0.63 -0.26 0.08 -0.07 0.51 -0.63 0.45 0.79 -0.23 SHEI -0.21 0.26 0.19 0.19 0.43 0.54 -0.39 -0.40 -0.05 0.17 -0.95 01 077 0.51 0.17 0.13 0.72 80-m PLAND_Ag .3 .5 PLAND_For -0 .0 PLAND_Wet -0.66 -0.05 0.32 PD 0.06 0.10 0.16 -0.19 ED -0.24 -0.02 0.34 0.28 0.42 AREA_MN -0.06 -0.10 -0.16 0.19 -1.00 2 -0 .4 AREA_CV 0.02 -0.11 -0.20 0.26 -0.25 -0.06 0.25 SHAPE_MN -0.53 -0.12 0.32 0.52 -0.49 0.32 0.49 0.00 ENN_MN 0.19 0.12 -0.12 -0.03 -0.35 -0.08 0.35 0.21 -0.03 CONTAG 0.27 -0.30 -0.21 -0.17 -0.38 -0.62 0.38 0.36 -0.14 0.00 IJI 0.10 0.22 -0.12 -0.10 0.47 0.34 -0.47 -0.02 -0.30 0.23 -0.44 PR 0.14 0.03 0.08 0.13 -0.02 0.36 0.02 0.47 0.02 0.60 -0.06 0.27 PRD 0.15 0.10 0.04 -0.42 0.69 -0.03 -0.69 -0.64 -0.48 -0.42 -0.09 0.18 -0.42 SHDI -0.07 0.19 0.18 0.21 0.19 0.62 -0.19 0.10 0.16 0.48 -0.60 0.45 0.79 -0.23 SHEI -0.24 0.28 0.20 0.18 0.34 0.57 -0.34 -0.36 0.18 0.12 -0.93 0.43 0.18 0.09 0.73 a % Urb = PLAND_Urb; b % Ag = PLAND_Ag; c % For = PLAND_For; d % Wet = PLAND_WET; e A_MN = Area_MN; f A_CV = AREA_CV; g SHAP = SHAPE_MN; h ENN = ENN_MN; i CONT = CONTAG.

PAGE 181

Table 3-39. Eigenvalues and vari ance explained by the first six axes for the principal components analysis at four different grain sizes (meters on a side) for lakes watersheds. Component 1 2 3 4 5 6 20-m Eigenvalue 3.722 2.898 2.193 1.601 1.309 0.69 % of variance explained 26.587 20.702 15.666 11.434 9.351 4.927 Cumulative % of variance 26.587 47.289 62.954 74.388 83.739 88.666 40-m Eigenvalue 3.451 3.014 2.478 1.412 1.044 0.757 % of variance explained 24.649 21.53 17.699 10.084 7.456 5.409 Cumulative % of variance 24.649 46.179 63.878 73.962 81.419 86.828 60-m Eigenvalue 3.458 3.249 2.65 1.147 1.077 0.677 % of variance explained 24.703 23.207 18.928 8.193 7.694 4.838 Cumulative % of variance 24.703 47.909 66.837 75.031 82.725 87.563 80 x 80-m Eigenvalue 3.564 3.045 2.586 1.25 1.035 0.799 % of variance explained 25.456 21.748 18.47 8.925 7.391 5.708 Cumulative % of variance 25.456 47.204 65.674 74.6 81.99 87.698 Table 3-40. The 20 x 20-m grain size for lake wate rsheds: principal component matrices showing pattern metric factor loadings. Metric DIVERS1 DIVERS2 URB/SIZE WET AG PLAND_Urb 0.10 -0.33 -0.43 0.23 0.29 PLAND_Ag -0.13 0.04 0.15 0.34 -0.59 PLAND_For -0.14 0.20 0.40 -0.17 0.27 PLAND_Wet -0.10 0.40 0.10 -0.44 -0.18 PD -0.36 -0.32 -0.07 -0.34 -0.04 ED -0.41 0.01 0.02 -0.11 0.35 AREA_CV 0.07 0.27 -0.43 -0.34 -0.18 SHAPE_MN -0.01 0.21 0.38 0.19 0.42 ENN_MN 0.36 0.32 0.07 0.34 0.04 IJI -0.34 0.01 -0.02 0.33 -0.19 PR -0.21 0.26 -0.41 0.17 0.25 PRD -0.10 -0.50 0.26 0.01 0.00 SHDI -0.41 0.22 -0.21 0.22 0.07 SHEI -0.43 0.09 0.09 0.17 -0.15 181

PAGE 182

Table 3-41. The 40 x 40-m grain size for lake wate rsheds: principal component matrices showing pattern metric factor loadings. Metric DIVERS1 DIVERS2 URB AG SIZE PLAND_Urb 0.17 0.10 0.53 0.09 0.26 PLAND_Ag -0.11 0.13 -0.11 -0.57 -0.45 PLAND_For -0.19 -0.03 -0.40 0.09 0.42 PLAND_Wet -0.23 -0.25 -0.32 0.22 -0.32 PD -0.11 0.47 0.04 0.34 -0.19 ED -0.36 0.19 -0.06 0.39 0.13 AREA_CV -0.05 -0.35 0.23 0.33 -0.46 SHAPE_MN -0.12 -0.30 -0.33 -0.12 0.33 ENN_MN -0.19 -0.22 0.34 -0.37 0.20 IJI -0.30 0.22 0.11 -0.22 -0.09 PR -0.36 -0.18 0.34 0.09 0.13 PRD 0.16 0.50 -0.07 -0.04 0.12 SHDI -0.50 0.04 0.18 -0.02 0.06 SHEI -0.41 0.25 -0.07 -0.14 -0.06 Table 3-42. The 60 x 60-m grain size for lake wate rsheds: principal component matrices showing pattern metric factor loadings. Metric DIVERS1 DIVERS2 URB AG SIZE PLAND_Urb 0.01 -0.27 0.45 -0.19 0.24 PLAND_Ag -0.13 -0.03 -0.12 0.79 -0.26 PLAND_For -0.08 0.20 -0.37 -0.21 0.29 PLAND_Wet -0.03 0.40 -0.25 -0.07 -0.26 PD -0.33 -0.32 -0.14 -0.21 -0.30 ED -0.38 0.12 -0.14 -0.39 -0.10 AREA_CV 0.08 0.26 0.36 -0.11 -0.56 SHAPE_MN 0.19 0.39 -0.24 0.06 0.28 ENN_MN -0.18 0.16 0.39 0.21 0.43 IJI -0.38 -0.16 0.06 0.08 -0.07 PR -0.30 0.27 0.33 -0.06 -0.03 PRD -0.08 -0.47 -0.22 -0.03 0.06 SHDI -0.47 0.21 0.10 0.04 0.09 SHEI -0.43 0.03 -0.19 0.15 0.18 182

PAGE 183

183 Table 3-43. The 80 x 80-m grain size for lake wate rsheds: principal component matrices showing pattern metric factor loadings. Metric DIVERS HETER URB AG SIZE PLAND_Urb 0.24 -0.09 -0.45 0.24 -0.31 PLAND_Ag -0.09 -0.15 -0.01 -0.73 0.34 PLAND_For -0.21 -0.03 0.36 0.18 -0.05 PLAND_Wet -0.30 0.22 0.26 0.05 0.29 PD 0.00 -0.50 0.06 0.27 0.29 ED -0.37 -0.22 0.12 0.32 -0.04 AREA_CV -0.11 0.31 -0.27 0.23 0.57 SHAPE_MN -0.26 0.28 0.31 -0.02 -0.32 ENN_MN -0.21 0.11 -0.41 -0.27 -0.26 IJI -0.17 -0.35 -0.22 -0.05 0.21 PR -0.36 0.03 -0.37 0.17 0.03 PRD 0.25 -0.44 0.15 0.01 -0.02 SHDI -0.46 -0.17 -0.19 0.01 -0.13 SHEI -0.34 -0.31 0.08 -0.19 -0.25

PAGE 184

Table 3-44. Pearsons correlations between landscape pattern metrics calculated at three spatial extents fo r the sample lakes. Some of the metrics descriptors (acronyms) of the co lumns have been renamed for convenience. Metric % Urba % Agb % Forc % Wetd PD ED A_MNe A_CVf SHAPg ENNh CONTi IJI PR PRD SHDI 100-m buffer PLAND_Ag -0.27 PLAND_For -0.44 0.11 PLAND_Wet -0.53 -0.03 -0.09 PD -0.29 0.17 0.17 0.25 ED -0.40 0.28 0.14 0.18 0.59 AREA_MN 0.19 -0.18 -0.21 -0.08 -0.92 -0.58 AREA_CV -0.06 -0.45 -0.08 0.29 0.09 -0.25 -0.06 SHAPE_MN -0.13 0.13 -0.15 0.09 -0.51 0.32 0.51 -0.37 ENN_MN 0.35 -0.13 -0.15 0.00 -0.34 -0.40 0.26 0.15 -0.06 CONTAG 0.36 -0.51 -0.06 -0.15 -0.51 -0.77 0.49 0.55 -0.25 0.32 IJI 0.27 0.18 0.09 -0.21 0.23 0.23 -0.23 -0.27 -0.13 0.15 -0.28 PR 0.15 -0.18 0.04 0.19 0.18 0.07 -0.27 0.47 -0.18 0.46 0.17 0.19 PRD 0.07 0.31 0.23 -0.39 0.36 0.45 -0.40 -0.72 0.02 -0.34 -0.51 0.37 -0.21 SHDI -0.03 0.22 0.11 0.14 0.45 0.45 -0.52 -0.13 -0.06 0.29 -0.56 0.39 0.65 0.26 SHEI -0.23 0.51 0.03 0.09 0.44 0.58 -0.43 -0.58 0.16 -0.13 -0.95184 0.31 -0.07 0.49 0.69 400-m buffer PLAND_Ag -0.31 PLAND_For -0.56 -0.08 PLAND_Wet -0.57 -0.02 0.16 PD -0.15 0.13 0.11 0.13 ED -0.17 0.04 0.22 0.11 0.60 AREA_MN 0.16 -0.14 -0.13 -0.14 -0.99 -0.62 AREA_CV 0.08 -0.23 -0.31 0.37 0.04 -0.13 -0.05 SHAPE_MN -0.14 0.00 0.14 0.05 -0.37 0.29 0.34 -0.07 ENN_MN 0.28 0.03 -0.16 0.03 -0.37 -0.10 0.36 0.09 0.24 CONTAG 0.42 -0.43 -0.24 -0.13 -0.48 -0.65 0.50 0.38 -0.15 -0.03 IJI -0.15 0.40 0.16 -0.02 0.22 0.32 -0.26 -0.17 0.20 0.30 -0.58 PR 0.22 -0.10 0.05 0.20 0.06 0.35 -0.09 0.32 0.07 0.53 -0.11 0.34 PRD 0.06 0.13 0.11 -0.43 0.39 0.17 -0.39 -0.66 -0.22 -0.35 -0.26 0.12 -0.43 SHDI -0.11 0.23 0.17 0.19 0.32 0.57 -0.34 -0.07 0.07 0.41 -0.74 0.59 0.73 -0.10 SHEI -0.38 0.46 0.20 0.12 0.40 0.51 -0.41 -0.38 0.09 0.12 -0.98 0.59 0.15 0.22 0.78

PAGE 185

185Table 3-44. Continued. Metric % Urba % Agb % Forc % Wetd PD ED A_MNe A_CVf SHAPg ENNh CONTi IJI PR PRD SHDI Watershed PLAND_Ag -0.28 PLAND_For -0.62 -0.09 PLAND_Wet -0.66 -0.05 0.36 PD 0.01 0.11 0.02 0.06 ED -0.14 0.00 0.23 0.19 0.54 AREA_MN 0.01 -0.09 -0.07 -0.08 -0.97 -0.54 AREA_CV 0.02 -0.08 -0.27 0.33 -0.07 -0.17 0.10 SHAPE_MN -0.29 -0.05 0.39 0.07 -0.31 0.33 0.29 -0.25 ENN_MN 0.18 0.13 -0.07 -0.15 -0.41 -0.25 0.41 0.05 -0.07 CONTAG 0.36 -0.36 -0.21 -0.23 -0.42 -0.67 0.43 0.31 -0.15 0.06 IJI -0.19 0.35 0.12 0.05 0.26 0.33 -0.31 -0.13 -0.01 0.22 -0.59 PR 0.17 -0.03 0.08 0.14 -0.01 0.29 0.03 0.35 0.01 0.56 -0.15 0.32 PRD 0.11 0.07 0.08 -0.34 0.50 0.12 -0.56 -0.70 -0.13 -0.29 -0.14 0.11 -0.48 SHDI -0.10 0.22 0.16 0.23 0.22 0.54 -0.20 0.04 0.04 0.41 -0.72 0.58 0.78 -0.25 SHEI -0.33 0.39 0.17 0.22 0.34 0.55 -0.34 -0.28 0.07 0.08 -0.98 0.61 0.22 0.08 0.78 a % Urb = PLAND_Urb; b % Ag = PLAND_Ag; c % For = PLAND_For; d % Wet = PLAND_WET; e A_MN = Area_MN; f A_CV = AREA_CV; g SHAP = SHAPE_MN; h ENN = ENN_MN; i CONT = CONTAG.

PAGE 186

Table 3-45. Eigenvalues and vari ance explained by the first six axes for the principal components analysis at three diffe rent spatial extents for lakes. Component 1 2 3 4 5 6 100-meter buffer Eigenvalue 3.83 2.487 2.215 1.586 1.069 0.877 % of variance explained 27.359 17.764 15.822 11.331 7.632 6.264 Cumulative % of variance 27.359 45.123 60.945 72.276 79.908 86.172 400-meter buffer Eigenvalue 3.676 2.736 1.953 1.626 1.249 0.878 % of variance explained 26.256 19.544 13.952 11.613 8.92 6.268 Cumulative % of variance 26.256 45.8 59.751 71.365 80.284 86.552 Watershed Eigenvalue 3.604 2.683 2.202 1.576 1.346 0.841 % of variance explained 25.742 19.162 15.729 11.258 9.616 6.004 Cumulative % of variance 25.742 44.904 60.633 71.89 81.507 87.511 Table 3-46. The 100-m spatial extent for lakes: principal component matrices showing pattern metric factor loadings. Metric DIVERS1 DIVERS2 URB SHAPE FOR PLAND_Urb 0.17 -0.07 0.55 0.06 0.31 PLAND_Ag -0.30 -0.11 -0.02 -0.14 -0.21 PLAND_For -0.14 0.04 -0.14 0.32 -0.76 PLAND_Wet 0.00 0.30 -0.41 -0.30 0.12 PD -0.30 0.32 -0.14 0.36 0.30 ED -0.40 0.10 -0.17 -0.10 0.21 AREA_CV 0.32 0.39 -0.19 0.14 0.09 SHAPE_MN -0.06 -0.27 -0.09 -0.61 -0.04 ENN_MN 0.18 0.24 0.36 -0.31 -0.29 IJI -0.22 0.10 0.39 0.10 -0.03 PR 0.04 0.53 0.19 -0.12 -0.12 PRD -0.38 -0.20 0.19 0.23 0.04 SHDI -0.30 0.40 0.22 -0.19 -0.07 SHEI -0.44 0.04 0.06 -0.21 0.09 186

PAGE 187

Table 3-47. The 400-m spatial extent for lakes: principal component matrices showing pattern metric factor loadings. Metric DIVERS1 DIVERS2 WET HETER AG PLAND_Urb 0.21 0.14 -0.54 0.27 0.15 PLAND_Ag -0.22 -0.13 -0.08 -0.21 -0.63 PLAND_For -0.19 -0.14 0.30 -0.23 0.41 PLAND_Wet -0.12 0.21 0.55 0.02 -0.16 PD -0.26 -0.22 0.11 0.56 -0.02 ED -0.37 -0.04 0.04 0.22 0.38 AREA_CV 0.13 0.39 0.26 0.32 -0.16 SHAPE_MN -0.10 0.15 0.01 -0.50 0.36 ENN_MN -0.10 0.41 -0.31 -0.21 -0.07 IJI -0.38 0.04 -0.20 -0.12 -0.14 PR -0.23 0.44 -0.11 0.20 0.17 PRD -0.07 -0.50 -0.25 0.09 0.10 SHDI -0.45 0.20 -0.12 0.10 0.01 SHEI -0.46 -0.10 -0.06 -0.06 -0.16 Table 3-48. The watershed spatial extent for la kes: principal compone nt matrices showing pattern metric factor loadings. Metric DIVERS1 DIVERS2 URB HETER AG PLAND_Urb 0.23 0.13 0.49 -0.02 -0.35 PLAND_Ag -0.18 -0.06 0.14 0.12 0.63 PLAND_For -0.21 -0.15 -0.40 0.21 -0.13 PLAND_Wet -0.21 0.08 -0.45 -0.32 0.13 PD -0.19 -0.32 0.22 -0.49 -0.08 ED -0.36 -0.18 0.02 -0.17 -0.42 AREA_CV 0.05 0.42 -0.12 -0.48 0.09 SHAPE_MN -0.11 -0.07 -0.33 0.41 -0.33 ENN_MN -0.09 0.39 0.20 0.40 0.09 IJI -0.37 -0.02 0.21 0.09 0.19 PR -0.28 0.41 0.13 -0.01 -0.27 PRD 0.04 -0.51 0.25 0.08 -0.01 SHDI -0.47 0.19 0.16 0.01 -0.10 SHEI -0.45 -0.11 0.12 0.03 0.14 187

PAGE 188

Table 3-49. Coefficients of determination, proba bilities, and regression equations for multiple regressions between indicators of ecos ystem condition and significant components resulting from the PCA of landscape pattern me trics at four grain sizes for the sample lakes ( level of 0.05). Components that were significantly related to the dependent variable (p < 0.05) are indi cated with an asterisk in the regression equation. Variable (Y) n R2 (adj) p Regression equation 20-meter Water chemistry Log10(Ammonia-N) 48 0.00 0.507 Y = 1.66 + 0.0183(DIVERS1) + 0.0139(DIVERS2) + 0.0408(URB/SIZE) 0.0240(WET) 0.0312(AG) Log10(NO3/NO2-N) 48 0.01 0.400 Y = 1.73 0.0154(DIVERS1) + 0.0352(DIVERS2) 0.0513(URB/SIZE) + 0.0454(WET) 0.166(AG) Log10(TKN) 47 0.10 0.093 Y = 0.188 + 0.0244(DIVERS1) + 0.0079(DIVERS2) 0.0017(URB/SIZE) 0.0414(WET) 0.0552(AG) Log10(TN) 47 0.29 0.002 Y = 0.109 + 0.0323(DIVERS1)* + 0.0022(DIVERS2) + 0.0098(URB/SIZE) 0.0161(WET) 0.0875(AG)* Log10(TP) 47 0.17 0.024 Y = 1.51 + 0.0241(DIVERS1) + 0.0619(DIVERS2)* 0.0051(URB/SIZE) 0.0871(WET)* 0.0095(AG) LCI 47 0.38 <0.001 Y = 44.1 3.08(DIVERS1)* + 4.09(DIVERS2)* 1.77(URB/SIZE) + 5.13(WET)* + 3.42(AG) 40-meter Water chemistry Log10(Ammonia-N) 48 0.00 0.565 Y = 1.66 + 0.0157(DIVERS1) 0.0116(DIVERS2) 0.0478(URB) 0.0044(AG) 0.0040(SIZE) Log10(NO3/NO2-N) 48 0.00 0.518 Y = 1.73 0.0290(DIVERS1) 0.0286(DIVERS2) + 0.0340(URB) 0.0921 AG_1 0.142(SIZE) Log10(TKN) 47 0.06 0.199 Y = 0.188 + 0.0204(DIVERS1) 0.0180(DIVERS2) 0.0168(URB) + 0.0010(AG) 0.0635(SIZE) Log10(TN) 47 0.24 0.006 Y = 0.110 + 0.0307(DIVERS1)* 0.0155(DIVERS2) 0.0240(URB) 0.0348(AG) 0.0732(SIZE)* Log10(TP) 47 0.11 0.081 Y = 1.51 0.0035(DIVERS1) 0.0649(DIVERS2)* 0.0371(URB) + 0.0502(AG) 0.0193(SIZE) LCI 47 0.39 <0.001 Y = 44.2 5.02(DIVERS1)* 1.78(DIVERS2) + 2.22(URB) 2.34 AG_1 + 5.14(SIZE)* 60-meter Water chemistry Log10(Ammonia-N) 48 0.00 0.417 Y = 1.66 + 0.028(DIVERS1) + 0.0137(DIVERS2) 0.0431(URB) + 0.0135(AG) + 0.0186(SIZE) Log10(NO3/NO2-N) 48 0.06 0.184 Y = 1.73 0.0128(DIVERS1) + 0.0129(DIVERS2) + 0.0397(URB) + 0.16(AG) 0.183(SIZE) Log10(TKN) 47 0.01 0.404 Y = 0.189 + 0.0304(DIVERS1) + 0.0162(DIVERS2) 0.0124(URB) + 0.0235(AG) 0.0042(SIZE) Log10(TN) 47 0.24 0.005 Y = 0.110 + 0.0406(DIVERS1)* + 0.0053(DIVERS2) 0.0192(URB) + 0.0675(AG)* 0.0246(SIZE) Log10(TP) 47 0.12 0.070 Y = 1.51 + 0.0342(DIVERS1) + 0.0673(DIVERS2) 0.0204(URB) 0.0281(AG)* 0.0024(SIZE) LCI 47 0.34 <0.001 Y = 44.5 3.76(DIVERS1)* + 3.37(DIVERS2)* + 2.27(URB) + 0.77(AG) + 4.97(SIZE)* 188

PAGE 189

Table 3-49. Continued. Variable (Y) n R2 (adj) p Regression equation 80-meter Water chemistry Log10(Ammonia-N) 48 0.01 0.386 Y = 1.66 + 0.0058(DIVERS) + 0.0227(HETER) + 0.0461(URB) 0.0327(AG) 0.0270(SIZE) Log10(NO3/NO2-N) 48 0.13 0.052 Y = 1.73 0.0196(DIVERS) + 0.0027(HETER) 0.0850(URB) 0.133(AG) + 0.235(SIZE)* Log10(TKN) 47 0.06 0.173 Y = 0.190 + 0.011(DIVE RS) + 0.0343(HETER) + 0.0227(URB) 0.0395(AG) 0.0125(SIZE) Log10(TN) 47 0.28 0.002 Y = 0.111 + 0. 0219(DIVERS) + 0.0290(HETER)* + 0.0162(URB) 0.0824(AG)* + 0.0211(SIZE) Log10(TP) 47 0.13 0.059 Y = 1.51 0.0237(DIVERS) + 0.0758(HETER)* + 0.0286(URB) + 0.0084(AG) 0.0118(SIZE) LCI 47 0.28 0.002 Y = 44.4 4.64(DIVERS)* 0.32(HETER) 2.42(URB) + 1.02(AG) 4.21(SIZE) 189

PAGE 190

Table 3-50. Coefficients of determination, proba bilities, and regression equations for multiple regressions between indicators of ecos ystem condition and significant components resulting from the PCA of landscape pattern metrics at three spatial extents for the sample lakes ( level of 0.05). Components that were significantly related to the dependent variable (p < 0.05) are indicated with an as terisk in the regression equation. Variable (Y) n R2 (adj) p Regression equation 100-meter Water chemistry Log10(Ammonia-N) 44 0.00 0.498 Y = 1.67 0.009(DIVERS1) + 0.0148(DIVERS2) 0.0475(URB) + 0.0388(SHAPE) 0.0251(FOR) Log10(NO3/NO2-N) 44 0.06 0.212 Y = 1.80 0.0051(DIVERS1) + 0.149(DIVERS2) + 0.0301(URB) + 0.0486(SHAPE) + 0.0012(FOR) Log10(TKN) 43 0.00 0.575 Y = 0.182 + 0.006(DIVERS1) 0.0232(DIVERS2) 0.028(URB) 0.0025(SHAPE) + 0.0276(FOR) Log10(TN) 43 0.00 0.533 Y = 0.126 + 0.0066(DIVERS1) 0.0101(DIVERS2) 0.0247(URB) + 0.0141(SHAPE) + 0.0349(FOR) Log10(TP) 43 0.08 0.152 Y = 1.50 + 0.0373(DIVERS1) + 0.0405(DIVERS2) 0.0664(URB)* + 0.0003(SHAPE) + 0.0139(FOR) LCI 43 0.26 0.005 Y = 45.6 + 0.33(DIVERS1) + 4.42(DIVERS2) + 1.24(URB) 5.45(SHAPE) 3.25(FOR) 400-meter Water chemistry Log10(Ammonia-N) 44 0.06 0.210 Y = 1.67 + 0.012(DIVERS1) 0.0281(DIVERS2) + 0.0632(WET)* 0.0155(HETER) 0.0511(AG) Log10(NO3/NO2-N) 44 0.02 0.347 Y = 1.80 0.053(DIVERS1) + 0.0777(DIVERS2) + 0.0144(WET) + 0.0592(HETER) 0.113(AG) Log10(TKN) 43 0.25 0.007 Y = 0.178 + 0.0372(DIVERS1)* 0.0224(DIVERS2) + 0.0423(WET)* + 0.0097(HETER) 0.0622(AG)* Log10(TN) 43 0.39 <0.001 Y= 0.122 + 0.0352(DIVERS1)* 0.0165(DIVERS2) + 0.0417(WET)* + 0.018(HETER) 0.0732 AG)* Log10(TP) 43 0.21 0.016 Y = 1.50 + 0.0234(DIVERS1) + 0.0187(DIVERS2) + 0.115(WET)* + 0.0342(HETER) 0.0132(AG) LCI 43 0.42 <0.001 Y = 45.3 4.28(DIVERS1)* + 5.35(DIVERS2)* 1.27(WET) 3.07(HETER) + 0.38(AG) Watershed Water chemistry Log10(Ammonia-N) 44 0.08 0.149 Y = 1.67 + 0.0127(DIVERS1) 0.0008(DIVERS2) 0.0716(WET)* 0.0017(HETER) + 0.0545(AG) Log10(NO3/NO2-N) 44 0.06 0.222 Y = 1.80 0.0456(DIVERS1) + 0.084(DIVERS2) + 0.0202(WET) 0.0894(HETER) + 0.121(AG) Log10(TKN) 43 0.12 0.077 Y = 0.179 + 0.0302(DIVERS1) + 0.0019(DIVERS2) 0.0224(WET) 0.0273(HETER) + 0.0578(AG)* Log10(TN) 43 0.29 0.003 Y = 0.122 + 0.0289(DIVERS1)* + 0.0064(DIVERS2) 0.0176(WET) 0.0433(HETER)* + 0.0676 (AG)* Log10(TP) 43 0.14 0.063 Y = 1.50 + 0.004(DIVERS1) + 0.0322(DIVERS2) 0.0859(WET)* 0.0532(HETER) + 0.0193(AG) LCI 43 0.41 <0.001 Y = 45.3 4.69(DIVERS1)* + 4.2(DIVERS2)* + 0.39(WET) + 4.47(H ETER)* 1.12(AG) 190

PAGE 191

Table 3-51. Multiple regression models at four gr ain sizes for the sample isolated forested wetlands: coefficients of determination, pr obabilities, and change in the amount of variability ( R2) in indicators of ecosystem condition ( level of 0.05). R2 is the added predictive power to the LDI that resu lted after using the LDI together with significant components that re sulted from the PCA of la ndscape pattern metrics. Dependent Variable Independent Variables R2 (adj) p R2 5 x 5-meter WCI Macrophytes LDI-PLU; URB; HETER; SHAPE; AG; FOR 0.41 <0.001 0.17 LDI-ILD; URB; HETER; SHAPE; AG; FOR 0.41 <0.001 0.20 LDI-ISD; URB; HETER; SHAPE; AG; FOR 0.41 <0.001 0.22 Macroinvertebrates LDI-ILD; URB; HETER; SHAPE; AG; FOR 0.25 0.040 0.07 LDI-ISD; URB; HETER; SHAPE; AG; FOR 0.26 0.035 0.09 10 x 10-meter Water chemistry Log10(TP) LDI-PLU; URB; HETER; CONTAG; AG; FOR; ENN 0.41 0.004 0.37 LDI-ILD; URB; HETER; CONTAG; AG; FOR; ENN 0.41 0.004 0.37 LDI-ISD; URB; HETER; CONTAG; AG; FOR; ENN 0.41 0.004 0.38 WCI Macrophytes LDI-PLU; URB; HETER; CONTAG; AG; FOR; ENN 0.43 <0.001 0.19 LDI-ILD; URB; HETER; CONTAG; AG; FOR; ENN 0.43 <0.001 0.23 LDI-ISD; URB; HETER; CONTAG; AG; FOR; ENN 0.44 <0.001 0.25 Macroinvertebrates LDI-ISD; URB; HETER; CONTAG; AG; FOR; ENN 0.26 0.048 0.08 20 x 20-meter Water chemistry Log10(TP) LDI-PLU; DIVERS; HETER; CONTAG; AG; SHAPE; FOR 0.40 0.005 0.36 LDI-ILD; DIVERS; HETER; CONTAG; AG; SHAPE; FOR 0.40 0.005 0.37 LDI-ISD; DIVERS; HETER; CONTAG; AG; SHAPE; FOR 0.40 0.005 0.37 WCI Macrophytes LDI-PLU; DIVERS; HETER; CONTAG; AG; SHAPE; FOR 0.38 <0.001 0.15 LDI-ILD; DIVERS; HETER; CONTAG; AG; SHAPE; FOR 0.39 <0.001 0.18 LDI-ISD; DIVERS; HETER; CONTAG; AG; SHAPE; FOR 0.40 <0.001 0.17 Macroinvertebrates LDI-PLU; DIVERS; HETER; CONTAG; AG; SHAPE; FOR 0.28 0.036 0.11 LDI-ILD; DIVERS; HETER; CONTAG; AG; SHAPE; FOR 0.31 0.026 0.13 LDI-ISD; DIVERS; HETER; CONTAG; AG; SHAPE; FOR 0.33 0.019 0.16 30 x 30-meter Water chemistry Log10(TP) LDI-PLU; DIVERS; HETER; URB/WET; AG; SHAPE; FOR 0.43 0.003 0.39 LDI-ILD; DIVERS; HETER; URB/WET; AG; SHAPE; FOR 0.43 0.003 0.39 LDI-ISD; DIVERS; HETER; URB/WET; AG; SHAPE; FOR 0.43 0.003 0.39 191

PAGE 192

Table 3-51. Continued. Dependent Variable Independent Variables R2 (adj) p R2 WCI Macrophytes LDI-PLU; DIVERS; HETER; URB/WET; AG; SHAPE; FOR 0.41 <0.001 0.17 LDI-ILD; DIVERS; HETER; URB/WET; AG; SHAPE; FOR 0.41 <0.001 0.21 LDI-ISD; DIVERS; HETER; URB/WET; AG; SHAPE; FOR 0.42 <0.001 0.23 Macroinvertebrates LDI-PLU; DI VERS; HETER; URB/W ET; AG; SHAPE; FOR 0.28 0.035 0.10 LDI-ILD; DIVERS; HETER; URB/WET; AG; SHAPE; FOR 0.32 0.021 0.15 LDI-ISD; DIVERS; HETER; URB/WET; AG; SHAPE; FOR; 0.34 0.017 0.17 192

PAGE 193

Table 3-52. Multiple regression models at four grain sizes for the sample streams: coefficients of determination, probabilities, and cha nge in the amount of variability ( R2) in indicators of ecosystem condition ( level of 0.05). R2 is the added predictive power to the LDI that resulted after us ing the LDI together with significant components that resulted from the PCA of landscape pattern metrics. Dependent Variable Independent Variables R2 (adj) p R2 20 x 20-meter Water chemistry DO LDI-PLU; DIVERS1; DIVERS2; WET; DIST 0.43 <0.001 0.02 LDI-ILD; DIVERS1; DIVERS2; WET; DIST 0.40 0.001 0.00 LDI-ISD; DIVERS1; DIVERS2; WET; DIST 0.38 0.001 -0.03 Log10(TN) LDI-PLU; DIVERS1; DIVERS2; WET; DIST 0.63 <0.001 0.46 LDI-ILD; DIVERS1; DIVERS2; WET; DIST 0.63 <0.001 0.47 LDI-ISD; DIVERS1; DIVERS2; WET; DIST 0.63 <0.001 0.47 Log10(TP) LDI-PLU; DIVERS1; DIVERS2; WET; DIST 0.40 <0.001 0.39 LDI-ILD; DIVERS1; DIVERS2; WET; DIST 0.40 <0.001 0.38 LDI-ISD; DIVERS1; DIVERS2; WET; DIST 0.40 <0.001 0.37 WQI LDI-PLU; DIVERS1; DIVERS2; WET; DIST 0.34 0.003 0.12 LDI-ILD; DIVERS1; DIVERS2; WET; DIST 0.40 0.001 0.12 LDI-ISD; DIVERS1; DIVERS2; WET; DIST 0.44 <0.001 0.11 SCI SC_2 LDI-ILD; DIVERS1; DIVERS2; WET; DIST 0.13 0.018 -0.10 LDI-ISD; DIVERS1; DIVERS2; WET; DIST 0.15 0.009 -0.10 50 x 50-meter Water chemistry DO LDI-PLU; DIVERS1; DIVERS2; WET; SHAPE 0.42 0.001 0.02 LDI-ILD; DIVERS1; DIVERS2; WET; SHAPE 0.40 0.001 0.00 LDI-ISD; DIVERS1; DIVERS2; WET; SHAPE 0.38 0.001 -0.02 Log10(TN) LDI-PLU; DIVERS1; DIVERS2; WET; SHAPE 0.59 <0.001 0.43 LDI-ILD; DIVERS1; DIVERS2; WET; SHAPE 0.60 <0.001 0.44 LDI-ISD; DIVERS1; DIVERS2; WET; SHAPE 0.59 <0.001 0.44 Log10(TP) LDI-PLU; DIVERS1; DIVERS2; WET; SHAPE 0.39 <0.001 0.38 LDI-ILD; DIVERS1; DIVERS2; WET; SHAPE 0.39 <0.001 0.37 LDI-ISD; DIVERS1; DIVERS2; WET; SHAPE 0.38 0.001 0.36 WQI LDI-PLU; DIVERS1; DIVERS2; WET; SHAPE 0.35 0.002 0.13 LDI-ILD; DIVERS1; DIVERS2; WET; SHAPE 0.41 0.001 0.13 LDI-ISD; DIVERS1; DIVERS2; WET; SHAPE 0.44 <0.001 0.11 SCI SC_2 LDI-ILD; DIVERS1; DIVERS2; WET; SHAPE 0.12 0.026 -0.10 LDI-ISD; DIVERS1; DIVERS2; WET; SHAPE 0.14 0.015 -0.10 193

PAGE 194

Table 3-52. Continued. Dependent Variable Independent Variables R2 (adj) p R2 80 x 80-meter Water chemistry DO LDI-PLU; DIVERS1; DIVERS2; URB; SHAPE 0.41 0.001 0.01 LDI-ILD; DIVERS1; DIVERS2; URB; SHAPE 0.40 0.001 -0.01 LDI-ISD; DIVERS1; DIVERS2; URB; SHAPE 0.39 0.001 -0.02 Log10(TN) LDI-PLU; DIVERS1; DIVERS2; URB; SHAPE 0.57 <0.001 0.40 LDI-ILD; DIVERS1; DIVERS2; URB; SHAPE 0.57 <0.001 0.40 LDI-ISD; DIVERS1; DIVERS2; URB; SHAPE 0.56 <0.001 0.41 Log10(TP) LDI-PLU; DIVERS1; DIVERS2; URB; SHAPE 0.34 <0.001 0.33 LDI-ILD; DIVERS1; DIVERS2; URB; SHAPE 0.34 <0.001 0.32 LDI-ISD; DIVERS1; DIVERS2; URB; SHAPE 0.34 <0.001 0.31 WQI LDI-PLU; DIVERS1; DIVERS2; URB; SHAPE 0.35 0.003 0.13 LDI-ILD; DIVERS1; DIVERS2; URB; SHAPE 0.41 0.001 0.13 LDI-ISD; DIVERS1; DIVERS2; URB; SHAPE 0.45 <0.001 0.12 SCI SC_1 LDI-ISD; DIVERS1; DIVERS2; URB; SHAPE 0.10 0.044 -0.12 SC_2 LDI-ILD; DIVERS1; DIVERS2; URB; SHAPE 0.12 0.024 -0.10 LDI-ISD; DIVERS1; DIVERS2; URB; SHAPE 0.14 0.013 -0.10 110 x 110-meters Water chemistry DO LDI-PLU; DIVERS1; DIVERS2; URB; SHAPE 0.44 <0.001 0.01 LDI-ILD; DIVERS1; DIVERS2; URB; SHAPE 0.42 0.001 -0.02 LDI-ISD; DIVERS1; DIVERS2; URB; SHAPE 0.41 0.001 -0.04 Log10(TN) LDI-PLU; DIVERS1; DIVERS2; URB; SHAPE 0.55 <0.001 0.35 LDI-ILD; DIVERS1; DIVERS2; URB; SHAPE 0.55 <0.001 0.36 LDI-ISD; DIVERS1; DIVERS2; URB; SHAPE 0.55 <0.001 0.36 Log10(TP) LDI-PLU; DIVERS1; DIVERS2; URB; SHAPE 0.30 0.001 0.29 LDI-ILD; DIVERS1; DIVERS2; URB; SHAPE 0.30 0.001 0.28 LDI-ISD; DIVERS1; DIVERS2; URB; SHAPE 0.31 0.001 0.27 WQI LDI-PLU; DIVERS1; DIVERS2; URB; SHAPE 0.35 0.002 0.13 LDI-ILD; DIVERS1; DIVERS2; URB; SHAPE 0.41 0.001 0.13 LDI-ISD; DIVERS1; DIVERS2; URB; SHAPE 0.46 <0.001 0.12 SCI SC_1 LDI-ISD; DIVERS1; DIVERS2; URB; SHAPE 0.10 0.046 -0.13 SC_2 LDI-ILD; DIVERS1; DIVERS2; URB; SHAPE 0.12 0.024 -0.10 LDI-ISD; DIVERS1; DIVERS2; URB; SHAPE 0.14 0.013 -0.10 194

PAGE 195

Table 3-53. Multiple regression models at thr ee spatial extents for the sample streams: coefficients of determinati on, probabilities, and change in the amount of variability ( R2) in indicators of ecosystem condition ( level of 0.05). R2 is the added predictive power to the LDI that resulted after using the LDI together with significant components that resulted from the PCA of landscape pattern metrics. Dependent Variable Independent Variables R2 (adj) p R2 100-meter buffer Water chemistry Log10(Turbidity) LDI-PLU; DIVERS; SIZE; WET; HETER 0.22 0.034 0.18 LDI-ILD; DIVERS; SIZE; WET; HETER 0.25 0.020 0.25 LDI-ISD; DIVERS; SIZE; WET; HETER 0.29 0.011 0.28 DO LDI-PLU; DIVERS; SIZE; WET; HETER 0.41 0.001 0.01 LDI-ILD; DIVERS; SIZE; WET; HETER 0.39 0.010 0.06 LDI-ISD; DIVERS; SIZE; WET; HETER 0.40 0.001 0.09 Log10(TN) LDI-PLU; DIVERS; SIZE; WET; HETER 0.33 0.001 0.20 LDI-ILD; DIVERS; SIZE; WET; HETER 0.33 0.001 0.22 LDI-ISD; DIVERS; SIZE; WET; HETER 0.33 0.001 0.22 Log10(TP) LDI-PLU; DIVERS; SIZE; WET; HETER 0.18 0.027 0.16 LDI-ILD; DIVERS; SIZE; WET; HETER 0.21 0.018 0.16 LDI-ISD; DIVERS; SIZE; WET; HETER 0.22 0.012 0.14 WQI LDI-PLU; DIVERS; SIZE; WET; HETER 0.35 0.003 0.14 LDI-ILD; DIVERS; SIZE; WET; HETER 0.41 0.001 0.15 LDI-ISD; DIVERS; SIZE; WET; HETER 0.48 <0.001 0.16 SCI SCI_1 LDI-PLU; DIVERS; SIZE; WET; HETER 0.17 0.009 -0.04 LDI-ILD; DIVERS; SIZE; WET; HETER 0.18 0.006 -0.06 LDI-ISD; DIVERS; SIZE; WET; HETER 0.19 0.005 -0.03 SCI_2 LDI-PLU; DIVERS; SIZE; WET; HETER 0.23 0.001 -0.01 LDI-ILD; DIVERS; SIZE; WET; HETER 0.24 0.001 -0.02 LDI-ISD; DIVERS; SIZE; WET; HETER 0.25 0.001 -0.01 400-meter buffer Water chemistry DO LDI-PLU; DIVERS; HETER; WET; AG 0.32 0.006 -0.02 LDI-ILD; DIVERS; HETER; WET; AG 0.27 0.012 -0.04 LDI-ISD; DIVERS; HETER; WET; AG 0.27 0.014 -0.04 Log10(TN) LDI-PLU; DIVERS; HETER; WET; AG 0.53 <0.001 0.43 LDI-ILD; DIVERS; HETER; WET; AG 0.52 <0.001 0.43 LDI-ISD; DIVERS; HETER; WET; AG 0.52 <0.001 0.43 Log10(TP) LDI-PLU; DIVERS; HETER; WET; AG 0.40 <0.001 0.39 LDI-ILD; DIVERS; HETER; WET; AG 0.41 <0.001 0.39 LDI-ISD; DIVERS; HETER; WET; AG 0.42 <0.001 0.39 195

PAGE 196

Table 3-53. Continued. Dependent Variable Independent Variables R2 (adj) p R2 WQI LDI-PLU; DIVERS; HETER; WET; AG 0.45 <0.001 0.24 LDI-ILD; DIVERS; HETER; WET; AG 0.47 <0.001 0.22 LDI-ISD; DIVERS; HETER; WET; AG 0.49 <0.001 0.20 SCI SCI_1 LDI-ISD; DIVERS; HETER; WET; AG 0.11 0.040 -0.11 SCI_2 LDI-PLU; DIVERS; HETER; WET; AG 0.13 0.026 -0.07 LDI-ILD; DIVERS; HETER; WET; AG 0.14 0.017 -0.08 LDI-ISD; DIVERS; HETER; WET; AG 0.15 0.014 -0.07 Watershed Water chemistry DO LDI-PLU; DIVERS; HETER; WET; AG 0.44 0.004 0.04 LDI-ILD; DIVERS; HETER; WET; AG 0.42 0.005 0.01 LDI-ISD; DIVERS; HETER; WET; AG 0.39 0.006 -0.02 Log10(TN) LDI-PLU; DIVERS; HETER; WET; AG 0.62 <0.001 0.45 LDI-ILD; DIVERS; HETER; WET; AG 0.62 <0.001 0.46 LDI-ISD; DIVERS; HETER; WET; AG 0.61 <0.001 0.45 Log10(TP) LDI-PLU; DIVERS; HETER; WET; AG 0.42 <0.001 0.41 LDI-ILD; DIVERS; HETER; WET; AG 0.42 <0.001 0.35 LDI-ISD; DIVERS; HETER; WET; AG 0.42 <0.001 0.39 WQI LDI-PLU; DIVERS; HETER; WET; AG 0.32 0.006 0.10 LDI-ILD; DIVERS; HETER; WET; AG 0.37 0.002 0.09 LDI-ISD; DIVERS; HETER; WET; AG 0.42 0.001 0.09 SCI SC_1 LDI-ISD; DIVERS; HETER; WET; AG 0.10 0.045 -0.13 SC_2 LDI-ILD; DIVERS; HETER; WET; AG 0.13 0.020 -0.10 LDI-ISD; DIVERS; HETER; WET; AG 0.15 0.012 -0.10 196

PAGE 197

Table 3-54. Multiple regression models at four grain sizes for the sample lakes: coefficients of determination, probabilities, and cha nge in the amount of variability ( R2) in indicators of ecosystem condition ( level of 0.05). R2 is the added predictive power to the LDI that resulted after us ing the LDI together with significant components that resulted from the PCA of landscape pattern metrics. Dependent Variable Independent Variables R2 (adj) p R2 20 x 20-meter Water chemistry Log10(TN) LDI-PLU; DIVERS1; DIVERS2; URB/SIZE; WET; AG 0.34 0.001 0.34 LDI-ILD; DIVERS1; DIVERS2; URB/SIZE; WET; AG 0.29 0.003 0.27 LDI-ISD; DIVERS1; DIVERS2; URB/SIZE; WET; AG 0.28 0.003 0.27 Log10(TP) LDI-PLU; DIVERS1; DIVERS2; URB/SIZE; WET; AG 0.17 0.031 0.15 LDI-ILD; DIVERS1; DIVERS2; URB/SIZE; WET; AG 0.15 0.046 0.10 LDI-ISD; DIVERS1; DIVERS2; URB/SIZE; WET; AG 0.15 0.045 0.11 LCI LDI-PLU; DIVERS1; DIVERS2; URB/SIZE; WET; AG 0.44 <0.001 0.43 LDI-ILD; DIVERS1; DIVERS2; URB/SIZE; WET; AG 0.41 <0.001 0.40 LDI-ISD; DIVERS1; DIVERS2; URB/SIZE; WET; AG 0.39 <0.001 0.39 40 x 40-meter Water chemistry Log10(TN) LDI-PLU; DIVERS1; DIVERS2; URB; AG; SIZE 0.32 0.001 0.32 LDI-ILD; DIVERS1; DIVERS2; URB; AG; SIZE 0.22 0.012 0.20 LDI-ISD; DIVERS1; DIVERS2; URB; AG; SIZE 0.24 0.009 0.22 LCI LDI-PLU; DIVERS1; DIVERS2; URB; AG; SIZE 0.50 <0.001 0.48 LDI-ILD; DIVERS1; DIVERS2; URB; AG; SIZE 0.41 <0.001 0.41 LDI-ISD; DIVERS1; DIVERS2; URB; AG; SIZE 0.42 <0.001 0.42 60 x 60-meter Water chemistry Log10(TN) LDI-PLU; DIVERS1; DIVERS2; URB; AG; SIZE 0.30 0.002 0.30 LDI-ILD; DIVERS1; DIVERS2; URB; AG; SIZE 0.27 0.005 0.27 LDI-ISD; DIVERS1; DIVERS2; URB; AG; SIZE 0.27 0.004 0.27 LCI LDI-PLU; DIVERS1; DIVERS2; URB; AG; SIZE 0.39 <0.001 0.39 LDI-ILD; DIVERS1; DIVERS2; URB; AG; SIZE 0.37 <0.001 0.37 LDI-ISD; DIVERS1; DIVERS2; URB; AG; SIZE 0.29 <0.001 0.29 80 x 80-meter Water chemistry Log10(TN) LDI-PLU; DIVERS1; HETER; URB; AG; SIZE 0.30 0.002 0.30 LDI-ILD; DIVERS1; HETER; URB; AG; SIZE 0.27 0.004 0.27 LDI-ISD; DIVERS1; HETER; URB; AG; SIZE 0.27 0.004 0.27 LCI LDI-PLU; DIVERS1; HETER; URB; AG; SIZE 0.32 0.001 0.32 LDI-ILD; DIVERS1; HETER; URB; AG; SIZE 0.30 0.002 0.30 LDI-ISD; DIVERS1; HETER; URB; AG; SIZE 0.29 0.003 0.29 197

PAGE 198

Table 3-55. Multiple regression models at three spa tial extents for the sample lakes: coefficients of determination, probabi lities, and change in th e amount of variability ( R2) in indicators of ecosystem condition ( level of 0.05). R2 is the added predictive power to the LDI that resulted after us ing the LDI together with significant components that resulted from the PCA of landscape pattern metrics. Dependent Variable Independent Variables R2 (adj) p R2 100-meter buffer LCI LDI-PLU; DIVERS1; DIVERS2; URB; SHAPE; FOR 0.27 0.007 0.27 LDI-ILD; DIVERS1; DIVERS2; URB; SHAPE; FOR 0.28 0.005 0.28 LDI-ISD; DIVERS1; DIVERS2; URB; SHAPE; FOR 0.30 0.004 0.30 400-meter buffer Water chemistry Log10(TKN) LDI-PLU; DIVERS1; DIVERS2; WET; HETER; AG 0.24 0.013 0.24 LDI-ILD; DIVERS1; DIVERS2; WET; HETER; AG 0.24 0.014 0.24 LDI-ISD; DIVERS1; DIVERS2; WET; HETER; AG 0.23 0.015 0.23 Log10(TN) LDI-PLU; DIVERS1; DIVERS2; WET; HETER; AG 0.40 <0.001 0.40 LDI-ILD; DIVERS1; DIVERS2; WET; HETER; AG 0.40 <0.001 0.40 LDI-ISD; DIVERS1; DIVERS2; WET; HETER; AG 0.38 0.001 0.38 Log10(TP) LDI-PLU; DIVERS1; DIVERS2; WET; HETER; AG 0.20 0.026 0.20 LDI-ILD; DIVERS1; DIVERS2; WET; HETER; AG 0.23 0.015 0.23 LDI-ISD; DIVERS1; DIVERS2; WET; HETER; AG 0.22 0.017 0.22 LCI LDI-PLU; DIVERS1; DIVERS2; WET; HETER; AG 0.43 <0.001 0.43 LDI-ILD; DIVERS1; DIVERS2; WET; HETER; AG 0.44 <0.001 0.44 LDI-ISD; DIVERS1; DIVERS2; WET; HETER; AG 0.41 <0.001 0.41 Watershed Water chemistry Log10(NO3/NO2-N) LDI-PLU; DIVERS1; DIVERS2; URB; HETER; AG 0.17 0.038 0.17 Log10(TN) LDI-PLU; DIVERS1; DIVERS2; URB; HETER; AG 0.37 0.001 0.37 LDI-ILD; DIVERS1; DIVERS2; URB; HETER; AG 0.28 0.005 0.28 LDI-ISD; DIVERS1; DIVERS2; URB; HETER; AG 0.23 0.006 0.23 LCI LDI-PLU; DIVERS1; DIVERS2; URB; HETER; AG 0.52 <0.001 0.52 LDI-ILD; DIVERS1; DIVERS2; URB; HETER; AG 0.46 <0.001 0.46 LDI-ISD; DIVERS1; DIVERS2; URB; HETER; AG 0.44 <0.001 0.44 198

PAGE 199

0 3 6 9 12 15 18 21 24 27 30CR4 SA 7 N A 1 0 N A 9 PA 1 N A 5 PA 9 PA 4 SR9 N U 1 N U 8 N U 6 SU 5 N U 9 CU 9LDI LDI-PLU LDI-ILD LDI-ISD Figure 3-1. Mean LDI scores for a subsample of isolated forested wetlands (n = 15). For each site LDI scores are shown calculated in three different ways: based on land use proportions only (LDI-PLU), and considering a linear decrease with distance (PLUILD) and an inverse square decrease with distance (LDI-ISD). E rror bars indicate the variance of LDI scores across grain sizes Wetlands are identified by site codes. 199

PAGE 200

SA7 0 0.3 0.6 0.9 1.2 010203040506070 Grain size (m)LDI NA10 0.0 0.4 0.8 1.2 1.6 010203040506070 Grai n s ize (m)LDI (a) PA1 0 3 5 8 10 010203040506070 Grain size (m)LDI PA4 4 7 9 12 14 010203040506070 Grain size (m)LDI (b) SU5 14 17 19 22 24 010203040506070 Grain size (m)LDI NU9 15 18 20 23 25 010203040506070 Grain size (m)LDI (c) Figure 3-2. Scalograms showing the effect of changing the grain size on the LDI for a subsample of six isolated forested wetlands representing wetland buffers with (a) low LDI scores, (b) intermediate LDI scores, and (c) high LDI scores. ( LDI-PLU; LDI-ILD; LDI-ISD). The site code for each wetland is shown. 200

PAGE 201

SA7 5 x 5-m grain size 40 x 40-m grain size 70 x 70-m grain size 0100Meters N NR Empower Density (E+14 sej/ha/yr) and Land Use Intensity Class 2.77 LI-Open Space / Recreational 3.20 Rangeland 67.35 MI-Open Space / Recreationa l 2.77 LI-Open Space / Recreational 3.20 Rangeland 67.35 MIOp ecreation 3.20 Rangeland-en Space / Ra l (a) NA10 5 x 5-m grain size 40 x 40-m grain size 70 x 70-m grain size 0100Meters N NR Empower Density (E+14 seja/yr) and Land Use Intensity Class /h 0.00 5.10 0.00 Natural land 2.77 Natural land 2.77 LI-Open Space / Recreational 5.10 Pine PlantationPine Plantation LIOpen Space / Recreational 5.10 Pin e Plantation 67.35 MI-Open Space / Recreationa l (b) Figure 3-3. Landscapes surrounding a subsample o gina eters. Landscapes were aggregated using the most frequently occurring value method. Sites: (a) SA7; (b) NA10; (c) PA1; (d) PA4; SU 5; and (NU9). Table 2-1 explains site codification. Light to dark red denotes increasing non-renewable empower density. f study isolated forested wetlands shown at l landscape), 40 x 40, and 70 x 70 m three different grain sizes: 5 x 5 (ori 201

PAGE 202

PA1 5 x 5-m grain size 40 x 40-m grain size 70 x 70-m grain size 090Meters N NR Empower Density (E+14 sej/ha/yr) and Land Use Intensity Class 3.20 Rangeland 51.52 HI-Pasture 67.35 MI-Open Space / Recreational 3.20 Rangeland 51.52 HI-Pasture 67.35 MI-Open Space / Recreational 3080.00 LI-Transportation 3.20 Rangeland 51.52 HI-Pasture 67.35 MI-Open Space / Recreational 3080.00 LI-Transportation(c) PA4 5 x 5-m grain 40 x 40-m grain size size 70 x 70-m grain size 0100Meters N NR Empower Density (E+14 sej/ha/yr) and Land Use Intensity Class 67.35 MI-Open Space / Recreational 117.11 Row Crops 1077.00 LI-Single Family Residential 3080.00 LI-Transportation (d) Figure 3-3. Continued. 67.35 MI-Open Space / Recreational 117.11 Row Crops 1077.00 LI-Single Family Residential 67.35 MI-Open Space / Recreational 117.11 Row Crops 1077.00 LI-Single Family Residential 202

PAGE 203

SU5 5 x 5-m grain size 40 x 40-m grain size 70 x 70-m grain size 0100Meters N NR Empower Density (E+14 sej/ha/yr) and Land Use Intensity Class 2.77 LI-Open Space / Recreational 67.35 MI-Open Space / Recreational 3080.00 LI-Transportation 5210.60 Industrial 2.77 LI-Open Space / Recreational 67.35 MI-Open Space / Recreational 3080.00 LI-Transportation 5210.60 Industrial 67.35 MI-Open Space / Recreational 3080.00 LI-Transportation 5210.60 Industrial (e) NU9 5 x 5-m grain size 40 x 40-m grain size 70 x 70-m grain size 0100Meters N NR Empower Density (E+14 sej/ha/yr) and Land Use Intensity Class 2.77 LI-Open Space / Recreational 67.35 MI-Open Space / Recreation al 3080.00 LI-Transportation 5020.00 HI-Transportation 5210.60 Industrial (f) Figure 3-3. Continued. 2.77 LIpace / Recreational -Open S 67.35 MI-Open Space / Recreational 5020.00 HI-Transportation 5210.60 Industrial2.77 LI-Open Space / Recreational 67.35 MI-Open Space / Recreationa l 5210.60 Industrial 203

PAGE 204

0 3 6 9 12 15 18 21 S1S22S4S69S21S9S5S44S14S2S8S23S13S53S26LDI LDI-PLU LDI-ILD LDI-ISD Figure 3-4. Mean LDI scores for a subsample of 15 stream drainage basins. For each site, LDI scores are shown calculated in three diffe rent ways: based on land use proportions only (LDI-PLU), and considering a linear decrease with distance (PLU-ILD) and an inverse square decrease with distance (LD I-ISD). Error bars indicate the variance of LDI scores across grain sizes. Stream draina ge basins are identif ied by site codes. 204

PAGE 205

S1 0.0 0.3 0.6 0.9 1 2 205080110140170 Grain size (m)LDI S2 2 0.0 0.4 0.8 1.2 1 6 205080110140170 Grin s aize (m)LDI (a) S21 3 4 5 7 8 205080110140170 Grain size (m)LDI S2 7 9 10 12 13 205080110140170 Grain size (m)LDI (b) S53 14 16 17 19 20 205080110140170 Grn s S13 10 12 14 16 18 205080110140170 Grain size (m)LDIaiize (m)LDI (c) Figure 3-5. Scalograms showing the effect of changing the grain size on the LDI for a subsample of six streams representing stream watersheds with (a) low LDI scores, (b) intermediate LDI scores, and (c) high LDI scores. ( LDI-PLU; LDI-ILD; LDI-ISD). The site code for each stream is shown. 205

PAGE 206

LDI-PLU 200-m 100-m 20-m UrbAg Ref UrbAg Ref UrbAg Ref 30 25 20 15 10 5 0 (a) 25 20LDI-ILD 15 10 5 0 UrbAg Ref UrbAg Ref Urb Ag Ref 20m 100m 200m (b) Figure 3-6. Comparison among LDI values calculat ed for isolated forested wetlands for three extents (20, 100, and 200 meters from wetlands edge). LDIs were calculated based on (a) the proportion occupied by each land use type in the la ndscape unit LDIPLU; and assuming that the effect of development intensity d ecreased (b) linearly with distance LDI-ILD, and (c) in inve rse-square with distance LDI-ISD. Groups within extent categories show wetlands distribution based on a priori land use categories: Ref = reference, Ag = agricu ltural, and Urb = urban. Differences among extent categories were not significant (Kru skal-Wallis test, p < 0.05). Boxes delimit the 25th and 75th percentiles and solid lines indica te the median. Whiskers extend to the lowest and highest data values a nd asterisks show unusual observations. 206

PAGE 207

LDI-ISD 200-m 100-m 20-m UrbAg Ref UrbAg Ref Urb Ag Ref 20 15 10 5 0 (c) Figure 3-6. Continued. LDI-PLU: 20-m LDI-ILD: 20-m LDI-ISD: 20-m LDI-PLU: 100-m LDI-ILD: 100-m LDI-ISD: 100-m LDI-PLU: 200-m LDI-ILD: 200-m LDI-ISD: 200-m 20 10 0 20 10 0 20 10 0 20 10 0 20 10 0 20 10 0 20 10 0 2010 0 20 10 0 2010 0 2010 0 2010 0 2010 0 2010 0 2010 0 20 10 0 Figure 3-7. Matrix plot of the relationship between pairs of LD I scores calculated for three landscape extents (20-, 100-, and 200-meter) for the sample of isolated forested wetlands. LDIs were calculated based on landuse type proportions in the landscape unit (LDI-PLU); and based on a linear (L LD) and inverse square (LDI-ISD) decrease with distance of development intensity. DI-I 207

PAGE 208

LDI-PLU 25 20 15 10 5 0 Watershed(a) 400-m(a) (a) 100-m High Med Low High Med Low High Med Low (a) LDI-ILD Watershed(b) 400-m(ab) 100-m(a) High Med Low 20 15 10 5 0 H d igh Me Low High Med Low (b) Figure 3-8. Comparison between LDI scores calculate d for buffer areas of 100 and 400 meters ediate 50% (Med ), and the higher 25% (High) of the data. Extent categories with similar letters were not significantly different (Kruskal-Wallis test, p < 0.05). Boxes delimit the 25th and 75th percentiles and solid lines indicate the median. Whiskers extend to the lowest a nd highest data values and asterisks show unusual observations. from the sample streams and for the entire drainage basin. LDIs were calculated based on (a) the proportion occupied by each land use type in the landscape unit LDI-PLU, and assuming that the effect of development intensity on the landscape decreased (b) linearly with distance LD I-ILD, and (c) in inverse-square with distance LDI-ISD. Groups within extent category show streams distribution for the first 25% (Low), the interm 208

PAGE 209

LDI-ISD Watershed(b) 400-m(ab) 100-m(a) High Med Low High Med Low High Med Low 20 15 10 5 0 (c) Figure 3-8. Continued. LDI-PLU: 100-m LDI-ILD: 100-m LDI-ISD: 100-m LDI-PLU: 400-m LDI-ILD: 400-m LDI-ISD: 400-m LDI-PLU: Wash LDI-ILD: Wash LDI-ISD: Wash 20 10 0 20 10 0 20 10 0 20 10 0 20 10 0 20 10 0 20 10 0 2010 0 20 10 0 2010 0 20 10 0 2010 0 2010 0 20 10 0 2010 0 2010 0 Figure 3-9. Matrix plot of the relationship between pairs of LD I scores calculated for three landscape extents (100-m, 400-m, and watershe d) for the sample of streams. LDIs were calculated based on landuse type proportions in th e landscape unit (LDI-PLU); and based on a linear (LDI-ILD) and inverse square (LDI-ISD) decrease with distance of development intensity. 209

PAGE 210

LDI-PLU Watershed 400-m 100-m High Med Low High Med Low High Med Low 25 20 15 10 5 0 (a) LDI-ILD 25 20 5 0 15 10 Watershed 400-m 100-m High Med Low High Med Low High Med Low (b) Figure 3-10. Comparison among LDI scores calc ulated for buffer areas of 100 and 400 meters from the sample lakes and for the entire dr ainage basin. LDI valu es were calculated based on (a) the proportion occupied by each land use type in the landscape unit LDI-PLU, and assuming that the effect of development intensity on the landscape decreased (b) linearly with distance LD I-ILD, and (c) in inverse-square with distance LDI-ISD. Groups within extent category show lakes distribution for the first 25% (Low), the intermediate 50% (Med ), and the higher 25% (High) of the data. There were no significant di fferences among extent categor ies (Kruskal-Wallis test, p < 0.05). Boxes delimit the 25th and 75th percentiles and soli d lines indicate the median. Whiskers extend to the lowest a nd highest data values and asterisks show unusual observations. 210

PAGE 211

LDI-ISD Watershed 400-m 100-m High Med Low High Med Low High Med Low 25 20 15 10 5 0 (c) Figure 3-10. Continued. LDI-PLU: 100-m LDI-ILD: 100-m LDI-ISD: 100-m LDI-PLU: 400-m LDI-ILD: 400-m LDI:ISD: 400-m LDI-PLU: Wash LDI-ILD: Wash LDI-ISD: Wash 20 10 0 20 10 0 20 10 0 20 10 0 20 10 0 20 10 0 20 10 0 20 10 0 20 10 0 2010 0 20 10 0 2010 0 2010 0 2010 0 2010 0 2010 0 Figure 3-11. Matrix plot of th e relationship between pairs of LDI scores calculated for three landscape extents (100-m, 400-m, and watershed) for the sample of lakes. LDIs were calculated based on landuse type proportions in the landscape unit (LDI-PLU); and based on a linear (LDI-ILD) and inverse square (LDI-ISD) decrease with distance of development intensity. 211

PAGE 212

LDI-PLUlog10(DO) 30 25 20 15 10 5 0 1.00 0.75 0.50 0.25 0.00 -0.25 -0.50 Reference Agricultural Urban Figure 3-12. Variability in DO for the sample isolated forested wetlands explained by the LDI calculated at the 200-meter buffer. Log10(DO) = 0.338 0.0079(LDI-PLU); r2 = 0.069, p = 0.028. Sample wetlands are designated by a priori land use category: reference, agricultural, or urban. LDI-PLUlog10(SC) 25 20 15 10 5 0 2.75 2.50 2.25 2.00 1.75 1.50 Reference Agricultural Urban Figure 3-13. Variability in SC for the sample isolated forested wetlands explained by the LDI calculated at the 200-meter buffer. Log10(SC) = 1.83 + 0.0194(LDI-PLU); r2 = 0.205, p = 0.008. Sample wetlands are designated by a priori land use category: reference, agricultural, or urban. 212

PAGE 213

LDI-PLUlog10(TP) 25 1.0 0.5 0.0 -0.5 -1.0 -1.5 -2.0 20 15 10 5 0 Urban Figure 3-14. Variability in TP for the sample isolated forested wetlands explained by the LDI calculated for a spatial ex tent of 20 meters. Log10(TP) = 1.06 + 0.0165(LDI-PLU); r2 = 0.065, p = 0.027. Sample wetlands are designated by a priori land use category: reference, agricultural, or urban. Reference Agricultural 213

PAGE 214

0.03 0.04 0.05 020406080R20.06 0.07 Grain size (meters on a side) 0.10 0.13 0.16 02R0.19 0.22 0406080 Grain size (meters on a side)2 (a) (b) 0.01 0.02 0.03 0.0 4 020406080 Grain size (meters on a side) 0.030 0.035 0.040 0.045 0.050R2020406080 Grain size (meters on a side) (c) R2 (d) 0.01 0.02 0.03 0.04 0204060 Grain size (meters on a side)R2 80(e) Figure 3-15. Regression results at several spat ial scales showing how mu ch of the variability (measured in r2 values) in the water chemistry variables for the sample isolated forested wetlands was explained by the LDI: (a) log10DO, (b) log10SC, (c) log10TN, (d) log10TP, and (e) log10Turbidity. ( LDI-PLU; LDI-ILD; LDI-ISD). 214

PAGE 215

LDI-PLUMacrophyte WCI 60 50 40 30 20 10 0 20 15 10 5 0 25 Figure 3-16. Variability in the macrophyte WCI explained by the LDI calculated for a buffer of 20-meters surrounding the sample isolat ed forested wetlands. Macrophyte WCI = 40.6 1.25(LDI-PLU); r2 = 0.296, p < 0.001. Sample wetlands are designated by a priori land use category: referen ce, agricultural, or urban. Reference Agricultural Urban 60 50 40 10 LDI-PLUMacroinverte WCI brate 30 20 25 20 15 10 5 0 0 Reference Agricultural Urban Figure 3-17. Variability in th e macroinvertebrate WCI explained by the LDI calculated for a buffer of 20-meters surrounding the sample isolated forested wetlands. Macroinvertebrate WCI = 33.2 0.70(LDI-PLU); r2 = 0.239, p < 0.001. Sample wetlands are designated by a priori land use category: reference, agricultural, or urban. 215

PAGE 216

LDI-PLUDiatom WCI 25 20 15 10 5 0 70 60 50 40 30 20 10 0 Figure 3-18. Variability in th e diatom WCI explained by the LD I calculated for a buffer of 100meters surrounding the sample isolated forested wetlands. Diatom WCI = 51.8 1.01(LDI-PLU); r2 = 0.242, p < 0.001. Sample wetlands are designated by a priori land use category: reference, agricultural, or urban. Reference Agricultural Urban 216

PAGE 217

217 0.14 0.17 0.20 0.23 0.26 010203040506070 Grain size (meters on a side)R2 0.14 0.16 0.18 0.20 0.22 010203040506070Grain size (meters on a side)R2 (a) (b) (c) Figure 3-19. Regression results at several landscape grains show ing how much of the variability (measured in r2 values) in the WCI for the sample isolated forested wetlands was explained by the LDI: (a) macrophyte WCI, (b) macroinvertebrate WCI, and (c) diatom WCI. ( LDI-PLU; LDI-ILD; LDI-ISD). 0.14 0.23 0.26 010203040506070 0.17 0.20 Grainsize(metersonaside) R2

PAGE 218

LDI-PLUMacrophyte WCI 30 25 20 15 10 5 0 60 50 40 30 20 10 0 Figure 3-20. Variability in the macrophyte WCI explained by the LDI calculated at a grain size of 5 x 5 meters. Macrophyt e WCI = 41.1 1.04(LDI-PLU); r2 = 0.243, p < 0.001. Sample isolated forested wetlands are designated by a priori Reference Agricultural Urban land use category: reference, agricultural, or urban. Macroinvertebrate WCI 25 20 15 10 5 0 60 50 40 30 20 0 10 LDI-ILD Reference Agricultural Urban Figure 3-21. Variability in th e macroinvertebrate WCI explained by the LDI calculated at a grain size of 50 x 50 meters. Macroinv ertebrate WCI = 32.9 0.692(LDI-ILD); r2 = 0.198, p < 0.001. Sample isolated forest ed wetlands are designated by a priori land use category: reference, agricultural, or urban. 218

PAGE 219

LDI-PLUDiatom WCI 25 20 15 10 5 0 70 60 50 40 30 20 10 0 Figure 3-22. Variability in the diatom WCI explained by the LDI calculated at a grain size of 30 x 30 meters. Diatom WCI = 52.2 0.994 (LDI-PLU); r2 = 0.231, p < 0.001. Sample Reference Agricultural Urban isolated forested wetlands are designated by a priori land use category: reference, agricultural, or urban. LDI-ISDDO (mg/l) 1816141210 86420 9 8 7 6 5 4 3 2 1 Figure 3-23. Variability in the concentration of DO for the sample streams explained by the LDI calculated for the watershed s cale. DO = 7.85 0.296(LDI-ISD); r2 = 0.409, p < 0.001. 219

PAGE 220

LDI-LPUlog10(NO3) 20 15 10 5 0 0.0 -0.5 -1.0 -1.5 -2.0 Figure 3-24. Variability in NO3-N for the sample streams explained by the LDI calculated fo the watershed scale. Log10(NO3-N) = 1.46 + 0.0467 (LDI-PLU); r2 = 0.139, p = 0.011. r LDI-LPUlog10(TN) 20 15 10 5 0 0.4 0.2 0.0 -0.2 -0.4 -0.6 Figure 3-25. Variability in TN for the sample streams explained by the LDI calculated for the watershed scale. Log10(TN) = 0.330 + 0.0242(LDI-PLU); r2 = 0.17, p = 0.004. 220

PAGE 221

70WQI 1816 14 1210 86420 60 50 40 30 20 10 LDI-ISD Figure 3-26. Variability in the WQI scores for the sample streams explained by the LDI calculated for the watershed s cale. WQI = 26.7 + 1.60(LDI-ISD); r2 = 0.332, p < 0.001. 221

PAGE 222

0.00 0.01 0.02 0.03 0.04 0.05 0306090120150180 Grain size (meters on a side)R2 0.38 0.40 0.41 0.43 0.44 0.46 0306090120150180 Grain size (meters on a side)R2 (a) (b) 0.05 0.07 0.09 0.11 0.13 0.15 0306090120150180 Grain size (meters on a side)R2 0.15 0.17 0.19 0.21 0.23 0.25 0306090120150180 Grain size (meters on a side)R2 (c) (d) 0.00 0.01 0.02 0.03 0.04 0.05 0306090120150180 Grain size (meters on a side)R2 0.20 0.24 0.28 0.32 0.36 0.40 0306090120150180 Grain size (meters on a side)R2 (e) (f) Figure 3-27. Regression results at several spat ial scales showing how mu ch of the variability (measured in r2 values) in the water chemistry variables and the WQI for the sample streams was explained by the LDI: (a) Turbidity, (b) DO, (c) NO3-N, (d) TN (e) TP, and (f) WQI. ( LDI-PLU; LDI-ILD; LDI-ISD). 222

PAGE 223

LDI-ISDDOmg/l) ( 18161412 10 86420 9 8 5 1 7 6 4 3 2 Figure 3-28. Variability in the concentration of DO for the sample streams explained by th calculated at a grain size of 170 x170 meters. DO = 7.80 0.298(LDI-ISD); r2 = 0.452, p < 0.001. e LDI LDI-PLUlog10(NO3) 20 15 10 5 0 0.0 -0.5 -1.0 -1.5 -2.0 Figure 3-29. Variability in the concentration of NO3-N for the sample streams explained by the LDI calculated at a grain size of 170 x 170 meters. Log10(NO3-N) = 1.43 + 0.0447(LDI-PLU); r2 = 0.147, p = 0.008. 223

PAGE 224

0.4 LDI-PLUl og10(TN) 0.2 0.0 -0.2 -0.4 -0.6 5 0 20 15 10 Figure 3-30. Variability in TN for the sample streams explained by the LDI calculated at a grain size of 170 x 170 meters. Log (TN) = 0.345 + 0.0264(LDI-PLU); r2 = 0.232, p = 100.001. 70 60 LDIISDWQI 50 10 40 30 20 1816 420 14 1210 86 Figur calculated at a grain size of 170 x 170 meters. WQI = 27.0 + 1.61(LDI-ISD); r2 = e 3-31. Variability in the WQI scores for the sample streams explained by the LDI 0.364, p < 0.001. 224

PAGE 225

35 LDI-ISDSCI _1 20 15 10 5 0 30 25 20 15 10 Figure 3-32. Variability in the SCI_1 for the samp le streams explained by the LDI calculated the 100-meter buffer. SC_1 = 30.6 0.450(LDI-ISD); r = 0.268, p < 0.001. for 2 LDI-ISDSC I_2 20 15 10 5 0 90 60 30 0 80 70 50 40 20 10 Figure 3-33. Variability in the SCI_2 for the samp le streams explained by the LDI calculated for the 100-meter buffer. SCI_2 = 73.2 2.34(LDI-ISD); r2 = 0.261, p < 0.001. 225

PAGE 226

0.12 0.15 0.21 0.24 Grain size (meters on a side) 0.18 0306090120150180R20.15 0.18 0.24 0.27 Grain size (meters on a side)R (a) (b) Figure 3-34. Regre0.21 03060901201501802ssion results at several spat ial scales showing how mu ch of the variability 2(measured in r values) in the SCI for the sample streams was explained by the LDI: (a) SC_1, and (b) SCI_2. ( LDI-PLU; LDI-ILD; LDI-ISD). LDI-ISD 35 25SC I _1 30 15 20 20 15 10 5 0 10 Figure 3-35. Variability in the SCI_1 for the samp le streams explained by the LDI calculated at2 a grain size of 20 x 20 meters. SC_1 = 31.2 0.409(LDI-ISD); r = 0.228, p < 0.001. 226

PAGE 227

LDI-ISDSCI2 90 70 60 40 30 10 0 20 15 10 5 0 80 50 20 Figure 3-36. Variability in the SCI_2 for the samp le streams explained by the LDI calculated at a grain size of 20 x 20 meters. SC_2 = 77.1 2.29(LDI-ISD); r = 0.252, p < 0.001. 2 227

PAGE 228

CHAPTER 4 land development may result in changes in the ecological condition of freshwater ecosystem DISCUSSION Summary Land use change may be the greatest threat to freshwater ecosystems worldwide. Since s and the degradation of resources vital to hum in the landscape affects freshwater ecosystems is a matter of critical importance. The problem development are the main driving forces of change in the landscape (Reynolds 1999). This quantitative measure of Landscape Development In tensity (LDI) calculated based on the use of human activity. The objectives of this study were to (1) investigate spatial properties of statistically relate landscape pattern with indicators of and (4) evaluate how landscape pattern indicators may complement the landscape development intensity indicators in their ability to predict ecological condition. A ccordingly, the major findings of this investigation are summarized as follows: Changes in landscape grain and extent had minor effects on the LDI. Grain size affected LDI forested wetlands). The effect of landscape gr ain on the LDI tended to become less important Differences in LDI scores with changes in landscape scale were more noticeable at low and an well-being, understanding how human behavior may be most significant in places like Flor ida where rapid population growth and land dissertation evaluated how landscape-scale development inte nsity affects the ecological condition and water quality of isolated forest ed wetlands, streams, and lakes in Florida. A non-renewable energy use and several measures of landscape pattern were used as indicators of indicators of landscape development intensity w ith changes in landscape scale and distance; (2) test the ability of indicators landscape development intensity to predict ecosystem condition at different landscape scales and c onsidering land use distance; (3) ecosystem conditions and water quality at diffe rent landscape scales; scores, especially in the middle ranges of th e LDI and for smaller landscapes (isolated with increasing landscape area. LDI scores tende d to increase with increasing spatial extent. 228

PAGE 229

intermediate LDI ranges where differences in LDI scores are more likely to occur as condition. The LDI calculated in closest proximity to the sample isolated forested wetlands respectively. There was a lack of correlati on between the LDI and indicators of lake Florida lakes. There were minor relationships between the LDI and water quality variables. For isolated correlations. Larger data samples for mu developed lands were added with increasing area. In general, the LDI had a greater predictive ab ility for biological-based indices of ecological and streams had the strongest pr edictive power, explaining up to 30% and 27% of variability, condition, which may be attributed to the sp atial variability and complex hydrology of forested wetlands and lakes small water quality sample size seemed to be a factor for poor ltiple years allowed for stronger associations for streams. Methods for calculating the LDI at th e watershed scale and for coarser grain sizes ha power of the LDI. Only for streams, LDIs based on distanceweighting explained up to 7% forested wetlands, and about 22% and 42% of the variability in biologic al indices of stream e r t Spatial Properties of the LDI Effec d the strongest predictive power. The use of a distance-weighting functions pr ovided little enhancement of the predictive more of the variance of biological indicators. Landscape pattern metrics were fairly associat ed to biological indi cators of ecological condition and water quality for the freshwater ec osystems studied. Landscape pattern indices explained up to 44% of variability in biological indicators of ecological condition in isolated and lake condition, respectively. Pattern metrics were important factors in explaining th variability of water quality variables, especi ally for streams and lakes, which accounted for up to 60% and 39% in the variance in water quality, respectively. In general, using pattern indice s with the LDI had a moderate effect on predictive power. Fo isolated forested wetlands, the stronges t influences were for the macrophyte and macroinvertebrate WCIs with an additional 25% and 17% of the total variance explained a the most significant scales, respectively. For la kes, landscape pattern wa s an important factor in determining lake condition; together, the landscape indices explained up to 52% of the total variance in the LCI. Combining the LDI and the landscape patter n variables was useful in assessing the influence of human development on water quality, especially for streams. ts of Changing Grain Size on the LDI Landscape indices allow the quant ification of landscape attrib utes and provide a direct means to assess how landscapes may affect ecological systems. Their interpretation and usefulness is even more relevant when the sensi tivity of landscape indices to changes in scale is 229

PAGE 230

well comprehended. In this study a Landscape Development Intensity Index (LDI) was developed using land use data surrounding various freshwaters systems, and was systematically investigated as to how the index varied with changes in landscape grain and extent. Important spatial properties of the LDI we re described through this work, enhancing the usefulness of th LDI as a measure of land use intensity The effect of cell aggrega tion (increase in landscape grain), which may result in the elimination of patch types, had more influence on the LDI score of the smaller landscapes (isolated forested wetla e s nds) than on larger landscapes (streams). More variab ng rough 3-5). g e out the nce or nds, s ed wetlands ca lculating the LDI at a grain size of 5 x 5 meter grain ility in LDI scores was observed when the LDI was quantified for landscapes surroundi isolated forested wetlands than for landscapes surrounding streams (see Figures 3-1 th LDI scores were also more variable among lands capes that were more heterogeneous in their land use composition (mix of lowto high-intens ity land uses). These results suggest that the grain size at which the LDI is calculated influences the result. It also suggests that the intensity of land use measured at different spatial scales may not be comparable. That the LDI is scale dependent is a logical outcome provided that th e LDI is derived usin land-use data, which when aggregated into coarser grain data suffers from the elimination of th less common land use types (Turne r et al. 1989). Despite this, tw o important findings ab properties of the LDI resulted from this study. First, the LDI is very sensitive to the prese absence of urban land-use types in the landscape Since the development intensity of urban la measured as non-renewable areal empower density, is at least an order of magnitude greater than other land uses classes like agri cultural and natural lands, the pr esence of even a small area of urban lands may result in a high LDI score for a given landscape that otherwise may appear les developed. Second, for the isolated forest s may be most appropriate (refer to Figure 3-2). The calculation of the LDI at coarser 230

PAGE 231

sizes may result in the elimination of land use pa tches that have strong influence on the cond of these systems. In doing so, the causes of eco system degradation rela ted to land use may be obscured. For streams, the behavior of the LDI wh en calculated at different grain sizes suggests that differences in LDI scores are minimal for th e finest scales considered (Figure 3-5, Appendix M) and that calculating the LDI at grain sizes between 20 x 20to 80 x 80-meters will preserve land use patches that have strong influence on th e condition of these systems, especially for small watersheds. It must be noted that in this study only one of several aggregation me thods for spatial data was used. The data were rescaled based on the mo st frequently occurring cell value, which may eliminate patch types with aggregation. This change may have had an effect on the ition LDI score for differ ions, tigation an development patterns of watersheds using the LD I and how the land use within them may affect ent grain sizes. The use of other aggregation methods may lead to different conclus since different aggregation met hods may produce different results when the spatial data are rescaled (Bian and Butler 1999; Tu rner et al. 2001; Wu 2004). Effects of Changes in Extent on the LDI The LDI is also sensitive to changes in landscape extent (see Fi gures 3-6 through 3-11). Considering that a change in spatial extent while holding the grain consta nt usually results in greater spatial heterogeneity as new patch type s are added to the land scape under inves (Wiens 1989), the increase in extent most likely resulted in the addition of urban lands and increase in LDI scores; thus, the largest diffe rences among LDI scores were observed between landscapes that presented low LDI scores at small extents and high LDI scores at broad extents. Since the results of this study suggested that the spatial extent over which the LDI is calculated influences the outcome, the ability to extrapolate development intensity values across different extents is limited. The implications of these results are important when analyzing the 231

PAGE 232

freshwater ecosystems. For largel y urbanized watersheds, which we re more characteristic of lakes analyzed in this study, the difference among extents may not be relevant as LDI scores tend to be similar regardless of changes in spatial exte nt. This might also be the case for watersheds where developed lands are found in the near vicin ity of the freshwaters systems studied as once included in the smaller extents, the influence of urban lands on the LDI score for the broad the er exten more in nnot n ated to land-use distance were also described. ng was effective at emphasizing the value of the development intensity of nearb ts may still be large because of the high non-renewable empower dens ity of these type of lands. When this is the case, considering just the smaller spatial extents to characterize the development intensity of a landscape surrounding a freshwater system may be sufficient as the variability in LDI scores tends to be small w ith increasing extent. Where landscapes are heterogeneous in their land use composition and undeveloped lands tend to be more common the near vicinity of the freshwater systems studied the effect of spatial extent on the LDI ca be ignored. This was the case for the landscape surrounding the streams an alyzed in this study, which tended to be surrounded by natural lands (floodplain effect) with their influence on the LDI decreasing with increasing extent. It is importa nt to point out the drai nage basins for streams were larger compared to those of the sample lakes, allowing for less extent overlap and more variability in land use types. This suggests that the size of the drai nage basin is al so important i determining LDI scores. However, the effect of watershed size on the LDI remains a matter for future research, as it was not qua ntitatively analyzed in this study. Distance-Weighting Factors Several spatial aspects of the LDI rel Distance-weighti y lands over that of more distant lands, partic ularly at the intermediate range of LDI values where landscapes were more heterogeneous in term s of the different lands use types present and where the presence of urban lands may have th e most influence on LDI scores. Another noted 232

PAGE 233

effect of distance-weighting on the LDI at the intermediate range of LDI values was that the LDI was less variable with changes in landscape scale because distance-weighting allowe the effect that methods of cell aggregation may have on LDI scores and by reducing the influence of developed lands that are added when the area of analysis is increased. The LDI in its three forms was also highly correlated (see Tables 3-5, 3-6, and 3-7 and Figures 3-7, 3-9, and 3-11), suggesting that all forms of the LDI provided very similar information about the land use intensity when calculated at the same landscape extent. Accordingly, either form of the LDI can be used to describe the intens ity of human development of the lands surrounding freshwater systems. However, the LDI calculated using distanceweighted factors allows disti nguishing between patterns of de velopment within buffers and whole watersheds, which can also be highly corr elated, especially fo d reducing r small watersheds where the ex of nds, ith n of aquatic ecosystems, biological indicators ar e believed to be useful in determining the tent overlap can be considerable. Possible applications of emphasizing the land uses within buffers over entire watersheds include giving mo re value to the presence of natural lands for their role in mitigating impacts or weighting more heavily the presence of contiguous patches developed lands along water bodies to give emphasis to their pot ential negative effect on their ecological condition and water quality. Land Use Intensity and Ecosystem Condition Biological Indices The LDI served as a measure to predict the condition of isolated forested wetlands and streams in spite of the inherent complexity of these systems. For the isolated forested wetla the LDI was significantly relate d to the macrophyte WCI, the macroinvertebrate WCI, and the diatom WCI (see Tables 3-10 and 3-11). Similarl y, the LDI was significantly associated w both forms of the SCI (see Tables 3-14 and 3-15) In the assessment of the ecological conditio 233

PAGE 234

condition of freshwater ecosystems since biologic al communities are affected by a wide range environmental factors and integrate the effect of human disturbances over time (Karr and C 1999; Karr and Chu 2000; Adams 2002). Accordingl y, the significant relationship between the LDI and measures of biotic cond ition for the isolated forested wetlands and streams seem to indicate that the LDI is an effective predictor of the various human factors that may of hu affect the condi argued mong Tangen et al. 20 variables may have a stronger effect on the biol ogical composition of Floridas streams than the tion of freshwater systems as reflected by the response of the biotic components of the systems investigated to land use intensity. Despite the significant association between the biological indices test ed and the LDI, the relationships reported were not strong. This sugge sts that factors other than land use intensity were in part responsible for the variability in the biological indices test ed. Reiss (2004) reported that given the wide latitudinal and longitudinal range of Florida there were differences in the biological composition of the asse mblages used in developing th e WCI for Florida. Reiss that the development of WCIs for specific re gions may improve the state-wide WCI since sources of environmental variability that ma y account for some of the differences found a the biological composition of the three assembla ges of organisms used in developing the WCI may be reduced. Other studies have also reported that biological indicators used to assess the level of wetland degradation by hum an activities can be highly in fluenced by natural factors in determining the composition of wetlands biologi cal communities (Wilcox et al. 2002; 03). Biological differences among regions for Florida s streams have been attributed mostly to topography, water velocity, and water chemistry ch aracteristics (Barbour et al. 1996a; Barbour 1996b). The weak correlation between the LDI and the state-wide SCI suggests that such natural 234

PAGE 235

intensity of human development. Natural factors have also be en reported to have a stronger influence on some streams biological communities than anthropogenic landscape variables (see d Wang et al. 2003). As suggested for the isolated e development of biologica l indices for the assessment of the ecological condi nthe sen et al. akes ining the condi Is for example Arbuckle and Downing 2002, an forested wetlands, th tion of streams for individual bioregions may result in the reduc tion of sources of no anthropogenic variability, and will possibly allow for stronger predictions of the effect of land use intensity on streams condition in Florida. For lakes, the LDI and the LCI correlated ve ry poorly (see Tables 3-16 and 3-17). The reasons for this poor correlation may be rela ted to the complexity of the hydrological characteristics of Floridas lakes, most of which are classified as seepage lakes. In developing LCI, Gerritsen et al. (2000) investigated the rela tionship between the biological condition of lakes and the proportion of land us es within their watersheds repr esented as equidistant buffers from the lakeshores. Poor relationships were found among the variables te sted. Gerrit (2000) pointed out the difficulty in defining the c ontributing drainage basi ns in Floridas l due to their complex groundwater connections, and s uggested that this fact or could be in part responsible for the lack of association between changes in land use and lake condition. Additional factors may also be more important than human influence in determ tion of the states lakes. The classification of Floridas lakes propos ed by Griffith et al. (1997) included 47 lake regions based on soil and sediment types, lake origin, water chemistry, and hydrology. Perhaps, an analysis based on indi vidual lake regions may result in the reduction of some of the variability of th e LCI due to factors other than those of anthropogenic origin Gradients of Change and Thresholds The relationship between the LDI and the indi ces of biological integrity for isolated forested wetlands indicated non-linear patterns. The relationshi p between the LDI and the WC 235

PAGE 236

showed clear differences in WCI scores among th e isolated forested we tlands with low LDI values and isolated forested wetlands with inte rmediate to high LDI values (see Figures 3 17, and 3-18). WCI scores for isolated forested wetlands with intermedia te to hi were approximately within the sam -16, 3gh LDI values e range, suggesting that there were minimal differences in the croinvertebrate, a nd diatom community compositions among the sample isolated forest Is es) s), d e may ck (2005) in explaining the variability observed in the streng urban macrophyte, ma ed wetlands. Reiss (2004, 2006) noted, after analyzing the relationships between the WC and a slightly different version of the LDI, that there may be a convergence of species among some of the isolated forested wetlands with in termediate LDI scores (mostly agricultural sit and isolated forested wetlands with high LDI sc ores (mostly urban isolated forested wetland despite the fact that the human influences were different for each type of wetland. The LDI incorporates land use intensity estimates from multiple land use types as measured by the use of non-renewable energies in each land use type. Energy flows such as electricity or energy storages, such as constructi on materials (buildings), help to define the differences between the energy usages of an urba n land from other land uses such as agriculture. These types of flows make urban lands more ener gy-intense than areas wi th other types of lan uses. However, the use of high energy-intense re sources like electricity in the landscap have a smaller effect on ecological systems than other less intense energy flows. This observation was also made by Surdi th of the relationship be tween land use intensity, as meas ured through the LDI, and the avian and amphibian species composition of isolat ed forested wetlands. The position of the isolated forested wetlands along the gradient of human disturbance observed in this study was the result of the higher energy us age in the landscapes surrounding these wetlands. However, the response of the biological assemblages analyzed to anthropogenic activi ties seemed to be the 236

PAGE 237

result of an energy usage that was related to less energy-intense lands, such as those that are characteristic of agricultural lands. Testing the LDI with other wetland data for different regions e whole range of the developmen t gradient may provide to the nature of the relationship betw een land use intensity and ecological condi shold t ng in which sites are selected to represent th additional insight in tion, as well as the effect of highly in tense energy flow on biological communities. For streams, the relationships between variables were non-line ar, and suggested a thre behavior with sites reporting low SCI values approximately mid-way on the LDI scale (see Figures 3-36 and 3-36). Species responses to environmental gradients are usually non-linear (McCune and Grace 2002), as are responses of stream condition to gradients of human disturbance (Allan 2004). Accordingly, results fro m this dissertation seemed to indicate tha streams may remain in relatively good conditi on until their drainage basins become highly developed. Only until these levels of developm ent are reached will their biological communities show evidence of degradation. Howe ver, other studies have reporte d a decline in the condition of biological communities of streams with relatively low levels of urban development (Wang et al. 2000; Paul and Meyer 2001), and for agricultural landscapes (W ang 1997; Fitzpatrick et al. 2001) It appears that the relationship between la nd use intensity and streams condition seems too complex to be explained solely based on one th reshold, or to be able to differentiate amo types of developed lands that may be more comp atible for preserving healthy streams. Despite this complexity, this study demonstrates that streams biological communities are adversely affected by increasing land development. Water Quality With the exception of some water chemistry variables for streams, the water quality of the freshwater systems studied were generally poorly associated with the development intensity of their surrounding lands (see Tables 3-8, 3-9, 3-12 and 3-13). The assessment of the condition of 237

PAGE 238

aquatic ecosystems based on a chemical criterion has been recognized by some authors as being of limited use, since it assumes that ecologi cal degradation is th e result of chemical contamination alone (Barbour et al. 1996a; Ka rr and Chu 1999; USEPA 2002b). In addition to chemical-based degradation, human actions can cause the alteration of hydrological, physical and biological factors that control the struct ure a nd functionality of aquatic ecosystems (Karr 1999) n the suite I can for er t rocesses (Barbour et al. 1996a; Brnmark and Hannon 1998; USEPA 2002b). Results from this study Since the LDI represents the combined act ions of several human actions that may be agents of ecosystem degradation (air and water pollutants, physical damage, changes i of environmental conditions such as groundwater levels and increased flooding) (Brown and Vivas 2005), the poor correlation between the water chemistry variables te sted and the LD be partially attributed to the limitation of wate r chemistry variables in integrating the multiple human factors that may affect th e condition of freshwater systems. The water chemistry data available for streams consisted of a larger number of samples for multiple years than the water chemistry data fo r lakes and isolated forested wetlands, many of which were based on one grab sample. It will seem that the larger water quality data sets streams allowed for a better integration of the cumulative effect of human actions than the wat quality data sets for lakes and isolated forested wetlands. Poor correlation between the LDI and water quality for lakes and isolated forested wetla nds could be partially e xplained by the fact tha water quality assessments based on grab samples may only reflect a temporary condition since the unusual concentration of chem icals in the water may be the result of a one-time pollution event instead of the effect of stressors ove r time (USEPA 2002b). Furthermore, the water chemical characteristics of most aquatic systems may change quickly as many of the chemical constituents present in the water column ar e altered by biological and physical p 238

PAGE 239

suggest that large sampling efforts may be requ ired to account for the different anthropogenic factor he ucture rainages were not representative of the flows that would determ ts to the sheds s that may contribute to determine the chemical composition of the water column of freshwater systems. This constitutes an important limitation to using a chem ical approach in t evaluation of the ecological condition of aquatic ecosystems, as extended sampling can be timeconsuming and costly, especially for studies at a regional level (Smith et al. 1997). Other factors may also have contributed to the lack of correlation be tween the intensity of land use and the water quality variables for the systems under investigation. For example, Devito et al. (2000) noted that in ar eas with a complex hydrology and geology, the landscape variables alone may not be good predictors of the variability in water quality of lakes as a consequence of human disturbance. In Florida, most of the la kes are seepage lakes with groundwater-dominating water flows, with their chemical composition clos ely related to the stat es geological str (Canfield and Hoyer 1988; Brenner et al. 1990 ) and largely determining nutrient budgets (Deevey 1988). Accordingly, it coul d be that contributing areas defined in this study for the sample lakes based on surface d ine the water chemistry composition of these systems. This possible situation poin need to analyze the relationship between huma n disturbance and lake condition based on the distribution of lakes by ecoregi ons or other land category that will allow the grouping of lakes based on their ecological simila rities and controlling for s ources of natural variability. Watershed size may also affect the water chemistry composition of freshwater systems. For example, low concentrations of nutrients ar e more common for lakes with smaller drainage basins (Brnmark and Hannon 1998; McDowell et al 2004). Watershed size was not evaluated as a source of variability in this research; however, it was observe d that the sizes of water 239

PAGE 240

for the sample lakes and streams were very variab le (see Table 3-3). Futu re research should te for potential dependence of lake condition on watershed size. Correlations with Changes in Grain Size Biological indices This study revealed t st hat the effect of landscape grain on the relationships between biolog as 11 the ns 1989), the ty of human deve lopment is independent of landscape grain, at s in contrast to wh at was observed for the isolat ical indicators of ecosyste m condition and the LDI was small. For the isolated forested wetlands, slightly more of the variability in the macrophyte WCI and the diatom WCI w explained at the finer and intermediate grain si zes (5 x 5-meter to 30 x 30-meter) (see Table 3and Figure 3-19). For the macroinve rtebrate WCI, a slightly la rger proportion of the total variance was explained at a coarser grain size (50 x 50-meter scal e). Considering that as grain becomes larger the spatial variance in the whole landscape study decreases (Wie these results seem to suggest that macrophytes an d diatoms are more affected by more local or fine-grained variations in driving energies associated with human development than macroinvertebrates. Perhaps this is due to the fact that macrophytes and diatoms have no mobility or depend on physical factors for moveme nt and finer-scale events may affect them more directly. On the other hand, macroinver tebrates can move away from sources of disturbance and may be affected more by stressors that are more widespread. For streams, the effect of grain size on the relationship between land use intensity and SCI was minimal (see Table 3-15 and Figure 3-34). It appears that th e response of stream macroinvertebrates to the intensi least within the range of grain sizes tested. This i ed forested wetlands, suggesting that the scale-dependency of the relationship between land use intensity and ecosystem condition differs for different freshwaters systems, and limiting the establishment of generalizations about system response to land use intensity with changes in 240

PAGE 241

landscape grain. Differences in the response obser ved for the two types of ecosystems to the influence of land use intensity could be attrib uted to differences in patterns of landscape development as well as to differences in the structure and composition of their macroinvertebrates communities. Water quality This study also showed that the effect of landscape grain on the relationships between water quality variables and the LDI was small. Fo r the isolated forested wetlands, the LDI was only significantly related to DO and SC (Table 3-9 and Figure 3-15). Since information at different grain sizes did not repr esent major differences in predictability (higher or lower r2 values), it would appear that to predict the im pact of human development on the water chemistry composition for small wetlands systems with small hydrological contributing areas, this relationship could be assessed at a ny scale within the range of grai ns sizes tested here. However, since the correlation between the wate r quality variables a nd the LDI were weak and only statist the because of the widespread nature of the proble m related to land use and non-point sources of ically significant for two variables, conclu sive statements on how land use intensity affects the condition of isolated forest ed wetlands as viewed by their water quality are not possible Perhaps using a larger dataset of water chemistry for wetlands, consisting of multiple measurements for each wetland sampled, might provide additional information about this relationship with changes in scale. For streams, significant associations were found between the LDI and DO, NO3-N, TN, and the WQI (see Table 3-13 and Figure 3-27). In all cases, the LDI e xplained more of the variability in the response values at the coarser grain size (170 x 170 meters), suggesting that effect of land use intensity on the water chemistr y composition of streams is highest when spatial variance is reduced. These results have an importa nt implication for studies at a regional level 241

PAGE 242

pollution. Additionally, since regional-scale land-u se data are usually coarse-grained, they may be more suitable than fined-grained data for studying this phenomenon. Correlations with Changes in Extent Biological indices Landscape extent was an important factor in determining the response of different biological communities ape development tween la nd use intensity and the community composition and macroinvertebrates for the is olated forested wetlands was reported for the small uman s is e isolated nd n ing ds of the sample isolated forested wetlands to landsc intensity. The strongest relationship be of macrophytes est extent tested (20-meter buffers) (see Table 3-10, Figure 3-16, and Figure 3-17). It appears that macrophytes and macroinvertebrate s are more likely to be affected by h activities within the immediate vicinity of the isolated fore sted wetlands, suggesting that maintaining the natural lands ar ound the first 20-meters surrounding is olated forested wetland critical for securing the ecological condition of these systems. However, it also seems that human activities on larger scales also have the potential to alter the ec osystem condition of th forested wetlands, as implied by the strongest relationship found between land use intensity a the diatom WCI for the 100-meter buffer scale (see Table 3-10 and Figure 3-18). These findings, as well as those reported for the response of the WCIs to land use intensity with changes i landscape grain, suggest that the assessment of ecosystem condition requires a multiple-scale approach since, as it has been noted, one scal e of analysis alone may only provide a limited understanding of the effects of human activities on the integrity of ecological systems (K 1993; Karr 1994; Allan 2004). Additionally, suppor t is provided for Reisss (2004) proposition of the use of a multi-metric multi-assemblage approach over a single-assemblage approach for determining the ecological condition of isolated forested wetlands Although this study seems to indicate that the conditio n of these wetlands systems can be assessed considering only the lan 242

PAGE 243

nearest to the study wetlands (within 100 meters), it will still be of interest to investigate how t lands beyond 200 meters might affe ct the condition of the isolated forested wetlands. Also, us elevation data with fine resolution when they becomes available for the entire state will allo delineating more precisely the hydrologic contri buting areas to these small systems, perhaps allowing for new insights into this relationship. For streams, the two forms of the SCI for Flor ida were most significantly associated with the LDI at the 100-meter buffer scale (see Ta ble 3-14). This suggests that changes in development intensity in the he ing w lands immediatel y surrounding the sample streams may be enough to alte et DI the nces to ment r r their ecological conditi on. It also points to the importa nce of riparian buffers for maintaining healthy streams. Other studies have also documented that local lands may have a stronger effect on stream macroinvertebrate asse mblages than more distant lands (Sponseller al. 2001; Townsend et al. 2004). However, small di fferences in the relatio nship between the L and the SCI_2 at the 100-meter buffer scale and the results for the watershed scale imply that both local and watershed-wide development may affect the community composition of macroinvertebrates in streams. Morley and Karr (2002) have suggested th at this might be case, while other studies have re ported that macroinvertebrate communities are more impacted by land use at the watershed scal e (Richards et al. 1996; Wang et al. 1997). Despite differe among studies that seem to indicate that there is no single best landscape at which the effect of human development on stream macroinvertebrates shoul d be analyzed and that efforts directed reduce the impact of land use intensity on stream s should consider multiple-scale manage strategies, these study indicates that for Florida streams the effect of land use intensity in thei nearest lands have the strongest effect on their macroinvertebrate communities. 243

PAGE 244

Water quality Differences were revealed on how landscape development intensity affects the water chemistry characteristics of the sample isolated forested wetlands with changes in landscape extent. For the isolated forested wetlands, si gnificant relationships were found between the LDI and DO, SC, and TP when changes in the spatial extent were considered (see Table 3-8). Th LDI explained more of the variance of the concen tration of DO at the broadest scale (200-meter buffer). Similarly, for SC more of the variability was explained at the 200-meter buffer scale. For TP, the LDI explained more proportions of the va riance at the smallest extent (20-meter buffer scale). These results initiall e y sugge st that the effect of land use in tensity on the isolated forested wetla ariable emistry s distance on the suggested that large areas with a large coverage of natura l lands surrounding wetlands are required to prevent the impact of human deve lopment on their water quality. The results nds water quality operate acr oss scales, varying depending on the water chemistry v considered. However, as was mentioned before when referring to the results from this dissertation regarding the relatio nship between the LDI and water quality, the use of a larger dataset for water chemistry variable for the isolat ed forested wetlands would have been useful to confirm these initial findings. When an effort was done to compare the re sults discussed here to those of similar investigations, only one study was found that has analyzed the in fluence of human development of adjacent lands on the water chemistry composition of wetlands. Houlahan and Findlay (2004) reported that the effect of the surrounding land uses on the composition of the water ch (nitrogen and phosphorus) of a sample of wetlands in Ontario, Canada, may extend to distance of up to 4,000 meters from the wetlands edge. Acco rdingly, it appears that the optimum or extent for most accurately predicting the variab ility of the water chemistry is dependent wetland type and varies for each variable c onsidered. Houlahan and Findlay (2004) also 244

PAGE 245

provided herein seem to indicate that for small isolated wetland systems, and specifically for phosphorus concentrations, only a narrow strip of na tural lands adjacent to the wetlands may be adequ more e n e n land use proportions (see Tables 314 and 3-15), it will seem that the effect of distance of land ate. However, it will still be useful to inve stigate beyond 200 meters in order to gain understanding of the influence of human activiti es on the condition of the isolated forested wetlands at larger scales. For Floridas streams the variability in the concentrations of DO, NO3-N, TN, and for th WQI were best predicted by the LDI at the wate rshed scale (see Table 3-12). This is in agreement with other studies that have reported that the influen ce of land use on the chemical condition of streams is best explained at the wa tershed scale (Hunsaker and Levine 1995; Scott et al. 2002; Strayer et al. 2003) Together these results sugge st that changes induced by watershed-scale development are more important in determining the water chemistry composition of streams than nearby lands and that studies aiming at analyzing the cumulative impacts of land use intensity on streams based on water quality assessments should be focused on the watershed scale. Yet, other studies have shown that the lands near est streams may have a stronger influence on the chemical composition of these systems (Richards et al. 1996; Johnso et al. 1997; Tufford et al. 1998). Provided that th e water quality data used in this dissertation appears representative of a wide variety of factors that may dete rmine the water chemistry of th sample streams as well as their cumulative im pact, and that the variability of water quality variables was consistently pred icted by the LDI at the watershed scale, this constitutes an important finding of this study. Land Use Intensity and Distance-Weighting Considering that the LDI calculated based on di stance-weights explained up to 7% more of the variance in stream condition as measure by the SCI than the LDI calculated based only o 245

PAGE 246

use is only a critical factor in determining th e effect of land use in tensity on the ecolog condition of these systems. On the other hand, th e effect of distance of land use had a limited effect on the ecosystem condition of the isolated forested wetlands as th e distance-weighted L less of the variability in the biological indices for these systems than the LDI based only o use (see Tab ical DI n land les 3-10 and 3-11). Distance-weighti ng was also of limited use when relating the LDI t variance use ture despite being s e er quality variables and the WCIs than distance-we ighing land use intensity measures. On the other o water quality variables. The distance-weighted LDI only explained more of the in DO and the WQI for streams (Table 3-12). Other significant relationships between land intensity and water quality variables for streams and isolated forested wetlands were stronger when the LDI-PLU was used. Accordingly, it ap pears that the distance of land use may have different effects on the chemical composition of different freshwater systems and depending on the response variable. Differences in the effect of land use intensity ba sed on distance-weighting for wetlands and streams also seemed to suggest th at watershed size may play a role in the effect of land use on ecosystem condition. The use of land-use distance weighing has receiv ed little attention in the litera suggested by ONeill et al. (1997) as a wa y to further refine landscape indicators of ecological integrity. Strayer et al. (2003) have reported that the spatial arrangement of land use i more important than non-spatially explicit landsca pe measures in small watersheds for predicting ecosystems condition, and King et al. (2005) has provided similar evidence after using distanceweighting of land cover for watersheds of different sizes to assess the cond ition of streams in th Coastal Plain of Maryland. However, the results of this dissertation s uggest the opposite since measures of land use intensity of the hydrological contributing ar eas for the sample isolated forested wetlands based on land us e proportions alone consistently correlated stronger with wat 246

PAGE 247

hand, for streams and their larger watersheds, th e response to the effect of distance-weighting was v seful for r en e on because they a nt ers ariable; suggesting that at broader scales the dist ance-weighted LDI may be more u assessing the water chemistry composition of freshwater systems Contrasting results between this and other research call for further investigation of the be nefits and limitations of the distance-weighting approach in freshwat er condition assessment studies. Landscape Pattern and Ecosystem Condition Correlations with Changes in Scale Water quality Isolated forested wetlands Landscape pattern was a poor pr edictor of water quality fo the isolated forested wetlands. Pattern me trics were only significantly related to the concentrations of TP (see Table 3-24). Agricult ural land uses as well as the distance betwe patches of the same type were metrics that were significantly associated to TP at the coarsest grain size tested (30 x 30 meters). The lack of correlation between landscape pattern metrics and water quality variables could be attributed once more to th e fact that water chemistry measurements were based on a single water sample rather than the inadequacy of landscap pattern metrics to predict water quality. The disadvantages of basing water quality assessment grab samples have already been discussed. Ho wever, the limitations of pattern metrics to accurately represent landscape patte rn also needs to be emphasized Some pattern metrics may give erroneous measurements at very small extents or some cannot be computed at all re dependent on the number of patch types present (McGarigal and Marks 1995; McGarigal et al. 2002). This certa inly represents a limitation of pattern metrics. In this study, testing the relationship between la ndscape pattern metrics and water quality variables of isolated forested wetlands with changes in landscape exte nt was not possible since only the largest exte could be considered as few patches tend to rema in at spatial extents smaller than 200 met 247

PAGE 248

surrounding the isolated forested wetlands, th ereby leaving the quest ion about how landsc pattern influences the condition of small isolat ed forested wetland with changes in landscape extent wide open. However, at small spatial extents simple measurements such as land u proportions may be useful in predicting how land use intensity affects the water quality of small wetland systems. Evidence in this direction ha s been provided by Lane and Brown (2006), who effectively used landscape measures of percent agriculture, percent urban, percent natural lands for 100meter buffers surrounding small herbaceous wetlands in Florida to assess the ecologica condition of these systems. Streams. For streams, landscape pattern metrics were useful in predicti ng changes in wat quality related to land use. Landscape pattern metrics were significantly associated to DO, NO3N, TN, TP, and the WQI when relationships were assessed at different landscape scales (see Tables 3-36 and 3-37). The water chemistry co mposition of streams varied along a land use diversity gradient ranging from a landscape w ith few land-use types among which forests were common, to landscapes richer in land use types where agriculture was more prev quality conditions were usually better for forested watersheds than for watersheds where agriculture lands were more common. The rela tionship between landscape pattern metrics and ape se l er alent. Water water the s chemistry varied with grain size and fo r each variable consid ered. Additionally, the relative importance of landscape pa ttern factors was also variable among variables. Based on results presented herein, it appe ars that the influen ce of land use on the water quality operates across multiple landscape grains, depending on th e water chemistry variable of interest. Accordingly, it seems that the assessment of the ecosystem c ondition of freshwater system using a chemical criterion will be incomplete if it relies only on one or few variables and if the effect of landscape grain is not considered, as different water chemistry variables are affected 248

PAGE 249

differently by land use and at different spatia l scales. Very few studi es have reported on the relationship discussed here. When this has been th e case, the use of differe nt grain sizes has been the result of the availability of spatial data at different resolution scales to compare different landscapes (Hunsaker and Levine 19 95), rather than a systematic e ffort to assess how changes in grain size determines the predictive power of landscape pattern metrics in assessing stream condition. Including the analysis of the effect of change sin lands cape grain in land-water stu for streams is a necessary step for the understand ing of how landscape pattern influences stream condition. Not doing so may lead to erroneou dies s conclusion about the scales at which water quality is con ds sion determines the oxygen concentrations in stream trolled by land use. With changes in spatial extent, the landscap e pattern metrics explained more of the variability in DO at the 100-meter buffer scale. St reams with adjacent lands with high presence of forests and wetlands presented higher concentr ations of DO than streams where adjacent lan were more fragmented and presented higher proporti ons of agricultural and urban lands. This is consistent with previous studies that have reported that the concentration of DO in streams decreases with increasing lands cape disturbance levels (Young and Huryn 1999; Mulholland et al. 2005). However, there seem to be no previous studies that considered the spatial scales at which land use has the greatest influence on DO in streams. Mulholland et al. (2005) suggested that changes in DO in streams should be anal yzed at the watershe d scale since stream metabolism, the process that together with di ffu s (Allan 1995; Young and Huryn 1999), is affected by watershed-wide variables. However, since differences in vegetation in adjacent lands (reduction in forest cover) can influence DO concentrations in streams (Findlay et al. 2001), land in the vicinity of streams may 249

PAGE 250

also play an important role in regulating DO concentrations as suggested by the result presented here. For nutrients, the landscape pattern metrics explained more of th e variability at the watershed scale. Previous studies have also es tablished that inorganic nitrogen (Johnson et al 1997; Sponseller et al. 2001; Gr iffith 2002) and TN (Hunsaker and Levine 1995; Jones et al. 2001; Sponseller et al. 2001) are more related to watershed-scale land us e patterns. Howev other investigations have also reported that the relationship between landscape pattern and nitrogen is best observed at more local scales (Johnson et al 1997; Tufford et al. 1998). Yet, Brett et al. (2005) found no significant relationships between land development and the concentratio of inorganic nitrogen at the wa tershed scale. Previous studies have also reported that TP in streams is controlled by human activities at the watershed scale (Hunsaker and Levine 1995; Brett et al. 2005). However, more local scales se em also important (Tufford et al. 1998; Joh et al. 1997) and others (Sponseller et al. 2001; Tufford et al. 2003) have found no relationship at all. The lack of agreement among studies in relating nutrients (nitroge n and phosphorus) to land use with changes in landscape extent may be attributed to the complexity of the phenomeno under investigation. Physical, chemical, and biological variables related to land use make it difficult to find patterns that allow the consistent description of nutrien t pathways over range of landscapes (Townsend and Riley 1999; McDowell et a er, n nson n a wide l. 2004 ). Differences among studie of s might also have to do with differences in classification systems used and the number of land use/land cover classes consid ered. Despite these differences a nd limitations, the results from this dissertation are a contribu tion to establishing a relationship between landscape pattern and water quality for Floridas streams. Landscape pattern metrics were able to explain up to 60% the variability of TN and up to 44% of the va riability of TP, providi ng evidence of the link 250

PAGE 251

between land use and stream condition. Add itionally, since in this study the land-use classification system used was the FLUCCS, which is widely used by state agencies and resea institutions in ecological assessment studies, uncertai rch nties regarding the usefulness the land use/la an with ween ter quality based solely on landscape variables. Despite these results, testing the ab ld be to furthe nd cover classification system and the number of land use classes considered were minimized. Lakes Landscape pattern metric s were useful in determining differences in the concentrations of nutrient in lakes with watersheds with va rying patterns of development. Landscape pattern metrics were significantly associ ated with TN and TP with changes in grain size. Relationships were best ex plained at the 20 x 20-meter scale in both cases (see Table 3-49). For different landscape extents, la ndscape pattern metrics were significantly associated to TKN, TN, and TP with all relationships were best ex plained for the 400-meter buffer (see Table 3-50). Lakes within watersheds with higher diversity of patch types (a mix of natural, agricultural, and urban lands) tended to have lower concentrations of nutrients than lakes within watersheds where patch diversity was low and where urba n or agricultural lands were common. Considering that excess phosphorus and nitrogen that result from agricultural and urb land use is the most common cause of concern re garding the eutrophicati on of aquatic systems, landscape pattern metrics proved to be effective in linking human activities in the landscape the condition of Florida lakes. Th e usefulness of pattern metrics is even more relevant since the hydrological complexity of Florida lakes makes it difficult to establish a relationship bet human disturbance and wa ility of landscape pattern metrics with larger water quality datasets wou r evidence the usefulness of pattern metrics to predict the water quality of Florida lakes. 251

PAGE 252

Biological indicators Isolated forested wetlands Landscape pattern metrics were fair predictors of the ecolo as nd an et al. 2006). Urban land use has also been reported to affect vegetation comm t cs was macroinvertebrates have been suggested as pote ntial indicators of eco logical integrity in gical condition of the isolated forested wetlands as measured by the macrophyte WCI and macroinvertebrate WCI (see Table 3-24). For the macrophyte WCI, more of the variability w explained at the 10 x 10-meter resolution scale, w ith the complexity of patch types shapes a forest lands being the most important explanat ory factors. In both cas es the proportion of agricultural land use was negativel y correlated with the patch type s shapes and the percent of forests, suggesting that among th e land use types present in the different landscapes, agriculture may be a determining factor for wetland impair ment. These results are consistent with other studies that have suggest ed that agricultural land use is an important factor in determining the level of impairment of wetland systems when related to different assemblages of wetland vegetation (Crosbie and Chow-Fra ser 1999; Galatowitsch et al. 2000; Lopez et al. 2002) as well as forest cover (Houlah unities in wetlands (Galatowitsch et al. 2000; Lopez et al. 2002); however, urban lands were not found to be an importa nt factor in this research Of interest was the fact that the influence of land development on the macrophytes WCI was most notable at a fine landscape grain when pattern metrics were used as an indicator of human activity. A similar result was reported when the LDI was used, providing support to the idea that because macrophytes canno escape the direct effect of distance, finer-s cale events may affect them more directly For the macroinvertebrate WCI, the relations hip between the landscape pattern metri only significant at the 20 x 20-meter resolution scale. Urban land use was the most important defining factor for wetland impairment Few studies have reported on the influence of landscape scale development on wetlands macroinvertebrate communities, despite that fact that aquatic 252

PAGE 253

wetlands (Adamus 1996; USEPA 2002a). Mensi ng et al. (1998) f ound no indic landscape scale disturbances, when measured along a disturbance gradient that included urb lands, affec ation that an ted aquatic macroinvertebrate commu nities in riparian wetlands of northern tempe te how land development affects the community of macro wet meadows (herbaceous communities) was most affected by human activities at less than 500 meters from the wetlands. Houlahan et al. (2006) found that the rate landscapes. Other studies have also re ported on the lack of correlation between land use and macroinvertebrates in wetlands (Steinma n et al. 2003; Tangen et al. 2003). Yet, Euliss and Mushet (1999) found that agri cultural lands had a significant influence on macroinvertebra assemblages for wetlands in the Prairie Pothol e region. A disagreement between the results on the grain size at which the influence of land development on the macroinvertebrate is most significant, seemed apparent between the pattern metrics and the LDI. According to the latter, the results indicated that the influence of la nd development on the macroinvertebrate was best observed at coarser grain sizes; for the former an intermediate scale seemed most important. Unfortunately pattern metrics could not be calc ulated beyond a grain size of 30 x 30 meters, thus limiting the possibility of comparison of results. Th erefore, multiple spatial scales might need to be considered when assessing invertebrates for isolated forested wetl ands as the evidence provided herein is not conclusive about the scale at wh ich macroinvertebrates are most affected by human disturbances. Previous studies have also repor ted, as is reported in this dissertation, on the importance of the landscape extent at which adjacent land us e may have the largest influence on wetlands. Mensing et al. (1998) found that for riparian wetlands, herbaceous vegetation was more affected by disturbance at local scales (less than 500 meters), while sh rub communities responded to disturbance at intermediate sc ales (500 and 1,000 meters). Gala towitsch et al. (2000) also observed that the vegetation of 253

PAGE 254

highest effect of adjacent land pecies was within 250 to 300 meter n udy s CI ir nd di atom assemblages when measured within 100 meter pattern I us e to the species richness on forest s s from temperate wetlands. For macroinvert ebrates, Mensing et al. (1998) found that these responded to disturbance at local scales (less than 500 meters). Although the effect of huma activities in the landscap e on the sample isolated forested wetl ands was not assessed in this st beyond 200 meters from the wetlands edge, the results presented herein together with the result from other studies suggest that in the asse ssment of wetland condition using vegetation and macroinvertebrate as indicators, the effects of human disturbances on wetla nds can be effectively assessed at relatively small spatial extents. For diatoms, the lack of correlation between landscape pattern metrics and the diatom W implies that the response of diatoms to human di sturbance may be controll ed at spatial scales different to the one that was c onsidered here. For example, Lane and Brown (2006) found a fa association between land use proportions a s from isolated depressional marshes in Fl orida. For streams, Pan et al. (1999) suggested that the correspondence between diatoms and land scape pattern can extend to the watershed scale. Testing the relationship between landsca pe pattern and condition diatom assemblages in wetlands at multiple scales may provide additional information about the spatial scales at which land development affect diatom communities. However, the limitation of using landscape metrics at very fine extents will always exist. Perhaps simple land use proportions may suffice to analyze this relationship at very small extents. Streams. No significant relationships between the landscape pattern variables and the SC for Florida were reported for different grain si zes (see Table 3-36). The lack of association between variables can be attributed to the fact that variations in grain size were analyzed only at the watershed scale, which was al so statistically not significantly related to the SCI. Landscape 254

PAGE 255

patter able s 03; portion have have 996) h er land use at the catchments scale, rather than at local scales (100meter t However, the results from this work indic ridas streams the response of macro es n variables were more strongly related to the SCI at the 100-meter spatial scale (see T 3-37). Accordingly, human activitie s in the landscape seem to have greater significance in determining the composition of macroinvertebrate a ssemblages at smaller spatial extents. This i consistent with other studies (Lammert and Al lan 1999; Sponseller et al. 2001; Roy et al. 20 Townsend et al. 2003; Townsend et al. 2004) and with the results reported in this work when the LDI was used as measure of human activity. Ho wever, differences also exist between this investigation and what has been reported elsewher e. Although in this inve stigation the pro of urban lands was not a very important factor in determining the variability the SCI, it did some influence on the composition of the macroi nvertebrates assemblages. Urban lands been reported to have more influence on stream macroinvertebrates at the watershed scale than at more local scales (Morley and Karr 2002; Roy et al 2003). Additionally, Richards et al. (1 suggested that all toget s buffers), is more important in determin ing the condition of stream assemblages including macroinvertebrates Similarities and differences between this work and other studies suggest tha land development affects the c ondition of streams at both loca l and basin-wide scales, and further studies of this relationship should incorporate analyses at multiple landscape extents. ate that for Flo invertebrates to human disturbance is controlled locally. Lakes Landscape pattern metrics were a fair pr edictor of the ecological condition of lak as measured by the LCI (see Table 3-49). More of the variability in lake condition was explained at finer resolution scales (20 x 20 meters and 40 x 40 meters). When different spatial extents were analyzed, more of the variability in la ke condition was explaine d at the 400-meter scale (see Table 3-50). Results showed that the diversity of patch types had a positive influence on 255

PAGE 256

macroinvertebrate communities. Higher levels of impairment of the macroinvertebrate communities of lakes were associa ted with low diversity of patch types where large urban or agricu n rban lands, was associated with less impaired conditions. Lewis hed ot s atterns of al owever, more research is ne eded in order to produce more conclusive statem logical condition and water quality vari ables initially explained by th e LDI alone, was additionally ltural patches were common in the landscape. Although macroinvertebrates have been identif ied as potential indicators of lake conditio (USEPA 1998), few studies have reported on the influence of land use on macroinvertebrates in lakes and none seem to have considered the sc ale(s) at which such a relationship is best predicted. Blocksom et al. (2002) found that macroinvertebrate assemblages were correlated to land use, with increased levels of impairment in lakes more strongly associated with u while the proportion of forests in lakes watershe ds et al. (2001) also provided evidence on the relationship between land use at the waters scale and the condition of lakes based on macroinvertebrate data. The fact that the LCI was n significantly related to th e LDI, as reported earlier in this di ssertation, adds more uncertainty a to how human disturbance impacts lake condition. Of in terest is to note that landscape p development for lakes and streams were different, with the former presenting higher levels of urbanization and less buffering by natural lands. Such differences may help to explain the contrast observed among the spatia l extents at which land use s eems to control the ecologic condition of these aquatic systems. Altogether, the finding of this work as well as the little evidence available from other studies initially indi cates that land use may affect lake condition at multiple scales; h ents on this matter. Land Use Intensity, Landscape Pa ttern, and Ecosystem Condition When the LDI and pattern metrics were used together in multiple regression analysis as predictor variables, different propo rtions of the remaining variance of the indicators of eco 256

PAGE 257

accounted for. The results varied for each depend ent variable and for each type of freshwater system investigated. The effect of the pattern metrics on the rela tionship between the LDI and the water quality variab by assessment tool of wetla ape n ealed that combining the LDI and the landscape pattern variab ned at indicators of les for the isolated forested wetlands was limited. More of the variability in the water quality measurements was only explained for TP concentrations (see Table 3-51). Stronger relationships were reported for the macrophyte WCI and the macroinvertebrate WCI. These results emphasize the need to focus on the use of biological indicators of ecosystem condition over water chemistry indicators. However, since landscape pattern metrics can be constrained the scale of analysis for very fine extents, th e LDI alone may be a better nds ecosystem condition than it would be when used together with measures of landsc pattern. At fine extents simple metrics such as land use proportions may be useful when used together with the LDI. However, care must be used since land use pr oportions and the LDI ca be highly correlated. Multiple regression analysis rev les was useful in assessing the influe nce of human development on the water quality variables for streams. Stronger relationships we re observed for DO, TN, TP, and the WQI (see Tables 3-52 and 3-53). That the combined predictiv e power of the two types of landscape pattern metrics was more noticeable for streams can be at tributed once again to the fact that the water chemistry data for these systems were more robust in terms of the number of independent samples included, and perhaps allowing for a better characterization of the streams water chemistry composition. Of interest was the fact that the landscape pattern metrics did not help to improve the relationship between the LDI and th e SCI. Although the landscape metrics explai a relatively small proportion of the variance in the SCI, it still can be argu ed th 257

PAGE 258

stream condition based on assemblages o are useful since macroinvertebrates communities can be negatively impacted at high levels of development intensity. he analysis of both landscape intensity and landscape pattern variables showed that when sed together in multiple factor models, slightly better relationships with indicators of ecosystem condition were observed for lake s (see Table 3-54 and 3-55). Pr ovided that the LDI alone was not significantly related to any of the water qua lity variables for lakes, the relative contribution of landscape pattern metrics in explaining the vari ability in the water quality variables for lakes suggest that the combined effect of both types of landscape indices may be useful in ecosystem condition assessment efforts for this type of systems. However, problems with collinearity were observed in the factor models for TN and TP. This was attributed to the fact that the LDI was highly correlated to the proportion of urban land use in the surr ounding landscapes to the lakes. Thus, testing for high correlation between the LDI and proportions of land use before using these metrics together is desirable. The LDI accounted for a proportion of the vari ability of some of the indicators of ecosystems condition and water quality tested in this study, while the la ndscape pattern metrics explained an additional proportion of the variability of these same variables and in some cases of other variables not explained by the LDI. This suggests that although both groups of landscape indices are measures of human activ ity in the landscape, different attributes of the land use are being quantified. Therefore, both types of lands cape indicators are complementary and can be used together to improve the ability to pred ict how landscape-level human disturbances may impact freshwaters systems. f macroinvertebrates T u 258

PAGE 259

Limitation Research A major limitation to this study was the small sa mple size of the streams and lakes studied sm ater chem istry measurem d tlaakes. A larger sample size of streams and lakes would have allowed including orenge ba ent inten ved the ccucy o r sh n, and oul pos h It n analysis y ecregir ework w s latinshiul ribed. A large mistry easurements fo or some of the varia to the multiple environmental factors that determine the water hemtry o he n ter chemis ges were ot alyzr e nmental ariaility i cal e structure of biological communities in fres hwater systems (Cattaneo and Prairie 1995; Cyr et al. 2004). By e mporal disconnections also existed between the land use da I and the ata ed ai ndition. The date the ere developed did ot aayse ater ch ng. For the sample olad fod ata we re developed otographs and efforts ereadero s not th ms and kesincela were datasets generated by others and corresponded to a prior eriod, and covered large ar eas throughout Florida. s and Further as well as the all sample size of w ents for the isolated foreste we nds nd la m drai a sins at differe nt levels of developm sity that could have impro a ra f the elationship of human devel opment and fre water ecosystem conditio w d sibly ave helped to iden tify critical thresholds. would also have allowed a b o ons o bioregions, providing a spatia l fram ithin which aspects of thi re o p co d have been more clearly desc r sample size of water che m r the isolated forested wetlands and lakes cou ld have permitted accounting f bility related c is comp sition of these systems. T seaso al or yearly variati ons in both wa try a nd biological assembla n na ed fo any of the systems studied, ev en though th temporal as pect of enviro v b is an mportant factor in determining the chemi composition as well as th doing so, aspects of variability of ecosys tem condition could have been explained Som te ta used to calculate the LD d us s ind cators of ecosystem co land use da ta w n lw agre with the date of the biologi cal and w emistry sampli is te reste wetlands, the land use d using aerial ph w m to g und truth the data. However, th is wa e case for the sample strea la s the nd use data used p 259

PAGE 260

The analysis of the landscape scales (grain and extent) at which the effect of human development was described was constrained by so me spatial attribu vestin.in si the extent of the drainage bach system. The olateforestetla imite d to the use of buffers as surrogates for real ydrolical cibutsolat ed forested wetlands can be p r watersheds and ubjecto hydica other freshwater systems not included in buffers. For loridas lakest of which are seepage sy stems, the identificati ontributing rainag basins lim bsurface exchanges that may trongllu lake es e limitations are technically difficult to vercoe, thenstitles of the comple xity of the systemated and call for the eed toconsidesewhen interpreting the relationship nd use and cosysm condition. Finally, the interpretation of landscape pattern indices was ch is difficulty has een reognizy othGriffith et al. 2002; Li and Wu 2004). Perhaps using simple nd a feltested m ve simp lified the understanding of how landscape pattern fluenes ecoem condition and can complement the LDI as pr iables of impact. Conclusions Changes in the ecological condition and water qu ality of isolated forested wetlands were nked ndinte th e landscape developmsity index (LDI). LDas evelope ac tivity based on a d at is erived no ng landscape and was used as a easur of hu-ind ecological systems. Accord ed as n effe lase-b that allows predicting the condition of freshwater stems, while the areal empower density of la nd use allows describing patterns of landscape tes of the systems under in gatio Gra ze was limited by asins for e is d ed w nds analyses were l h og ontr ing areas. I a rt of large s t rolog l exchanges with F s, mo on of the c d e s wa ited to topographic flows, ignoring su s y inf ence condition. Although th o m y co ute examp s investig n er th variables between la e te allenging. Th b c ed b er authors ( a ew w letrics would ha in c syst edictor var li to la use nsity, as measured by ent inten The I w d d as an index of human evelopment intensity measure th d from n-renewa ble energy use in the surroundi m e man uced impact s on ingly, the LDI may be us a ctive nd u ased assessment index sy 260

PAGE 261

development which may aid in land use planning directed to minimizing anthropogenic influences on ecological systems. Te stud the the LDI is scale-dependent, aryingwith ces i ain and extent. Change s ed DI scres, esally aller landscapes; while changes landape et we at low and intermediate L nces LDIcoresore likely to occur as deve loped lands were added with increasing area. The ndersndinghe s ges in spatial sca onsidation utur the LDI. Geater elati DI and biol ogical indicato l ommuities inte huma n disturbances an e effective dicatrs of ecological condition. The limited co rrelation between ater quality aentionhe li chemical criteri on to assess the condition of reshwsysems. Although a significant level of association was found between the LDI and ome wter chstry am onditin of fate ge sa equire to act for h sition f fresater ms n ct Athoug ov r s redictowr of ec n at the sc ales tested was sm ue ale atich t ent and ecological condition can e bestss ultiple s h y of spatial propert ies of the LDI revealed that v hang n both la ndscape gr in landscape grain affect L o peci in its middle ranges of values and for sm in sc xten re more noticeable DI ranges where differe in s are m u ta of t ensitivity of the LDI to chan le is an important c er for f e app lications of r corr ons between the L rs s uggest that biologica c n may grate a wide range of d may be mor in o the LDI and w draw tt to t mitations of using a f ater t s a emi variables, as in the case of the sample stre s; in order to assess the c o reshw r systems using a chemi cal criterion, lar mpling efforts may be r d coun the environmental factors that determine t e water chemical compo o hw syste and may only provide a partial understandi g of how humans may impa aquatic system s. l h the erall effect of changes in landscape scale (g ain and extent) on the LDI p ive p e osystem conditio all, there is no one uniq spatial sc wh he relationship between human developm icators analyzed at m b asse ed. Rather, multiple ind patial scales should be 261

PAGE 262

considered when analyzing this relationship. Despite this, the land nearest to the isolated forested wetlands and streams studied seemed to have the greatest effect on as esy e respo cal indicators to land use inten Tseistan provided little e nhan ower f the Lespe shwater sy stems with very sma er, istanc-weig bec ainage sizes increa istanc-weig ma een the effect of land use intensity within buffers nd whle waeds, r highlighting the im s ithin oodpl and ecological condit Lndscaatte air predicto rs of the condition and water quality for olateestd wetla rences among t est xplaind by tnds metrics and the LDI indicate tha e dicesan be tog ive power of landscape indices for assessing e conition shw Foridareaspopulation will contin ue to dema use ith th inevi con ining natural lan ban ses. Idfyig patte nt that w ill minimize the im water ystemais imperative. The LDI and mtrics of landscape pattern are analytical tools that an aidhe ssme andscapes, a n der to nsuree pernce stream tate. Amore complete g as to how the LDI can be us of reshwters sys could include an analysis th at considers a larg l asporal varia grity their ecological condition, sugg ted b th nse of biologi sity. he u of d ce-weighti ng functions cement of the predictive p o DI, cially for fre ll dr ainage areas; howev d e hting ame more important as dr sed. At broader extents d e hting y allow distin ction betw a o tersh which m ay be useful fo portance of natural land w fl ains their positive influence on ion. a pe p rn metrics were f is d for e nds, streams, and lakes. Diffe he vari ables that were b e e he la cape pattern t bo th types of landscap in c used ether to enhance the predict th d of fre ater systems. l s inc ing human nd intensification in land w e table version of some of the rema ds to agricu ltural and ur u enti n rns of developme pact on the states fresh s s rem n e c in t asse nt and management of enti re l ecessary approach in or e th siste of healt hy isolated forested wetlands s, and lakes throughout the s understandin ed to predict the condition f a stem er systems sample size as wel tem tions in water ch emistry and biotic inte measures. 262

PAGE 263

APPENDIX A CIRCUIT LANGUAGE le A Pr symbols of the energy circuit diagramming. ENERGY Tab -1. imarySymbol Name Description Sys tem nda ystem being diagramm ross boundary indicate inflo of em. the system bou ry Defines the s ed. Lines that c ws and outflows the syst gy in Ener circui A pathway with a flow proportiona t the storage or source upstream l to the quantity urce ing function or outside sourcgy forces according to a prtrolled So delivering A forc e of ener ogram con from outside. Flow li rce Outside source of energy with a floternally mited sou w that is ex controlled. rage rtment of energy storage ystem ntity as the balance ofnd Sto tank storing a qua A compa within the s inflows a outflows. nsor (small square box on st ols some other fl supply the main energy for it. Se storage tank contr The sensor orage) suggests the ow but does not Producer Unit that collects and transforms low-quality energy under the control of high-quality flows. Consumer Unit that transforms energy quality, stores it, and feeds it back autocatalytically to improve inflow. Box Miscellaneous symbol to use for whatever unit or function is needed. Heat sink Dispersion of potential energy into heat that accompanies all real transformation processes and storages. Dispersed energy is no longer available to the system. 263

PAGE 264

APPENDIX B LAND USE/LAND COVER CLASSIFICATION SYSTEM Table B-1. FLUCCS categories and corre sponding land use in tensity classes. de FLUCCS Co L Land Use Intensity Description I=M it 1 FLUCCS Description (LI=Low Intensity, M Intensity, HI=H iddle y) igh Intens I LII LIII 00 Urban and Built-up LI-Single Famtial Fixed Single Family Units LI-Single Famtial anchettes Mobile Units LI-Single Famtial anchettes Mixed Units LI-Single Famtial 9 Low Density Under Construction LI-Single Famtial Density MI-Single Famntia sidential Fixed S ts MI-Single Fantia Medium Density Residential Mobile Home Units MI-Single Family Residential 123 Medium Density Residential Mixed Units MI-Single Family Residential LI-Multifamily Residential 13 ixed Units HI-Singleesiden Low Rise HI-Multsident igh Density Residential Mixed Units < and mobile home units> HI-Singeside 139 High Density Under Construction HI-Single Famiesid mercial and Services LI-Commercial rvices HI-Commercial 1411 Shopping Center HI-Commercial 142 Wholesale Sales and Services LI-Commercial 143 Professional Services LI-Commercial 144 Cultural and Entertainment Institutional 145 Tourist Services LI-Commercial 193 Urban Land in Transition Without Positive Indicators of Intended Activity MI-Open Space / Recreational 194 Other Open Land MI-Open Space / Recreational 200 Agriculture 110 Residential, Low Density dwelling units per acre> ily Residen 111 Low Density Residential Fixed Units Family ily Residen 112 Low Density Residential Mob e Units ily Residen 113 Low Density R Fixed and mobile anchettes ily Residen 114 R 115 R ily Residen ily Residen 116 R ily Residen 11 ily Residen 120 Residential, Medium dwelling units per ac e ily Reside l 121 Medium Density Re ingle Family Uni mily Reside l 122 129 Medium Density Under Construction MI-Single Family Residential 130 Residential, High Dens ity 1 High Density Residential F Single Family per acre> ential 132 High Density Residentia Units ifamily Re l 134 M ee s 135 H ifamily Re ial Fixed le Family R ntial ly R ential 140 Com 141 Retail Sales and Se 264

PAGE 265

Table -1. CCCBS Cone tinued. FLU od LI LII Land Use Intensity Description (LI=Low Intensity, MI=Middle Intensity, HI=High Intensity) LI-Pasture LIII FLUCCS Description 210 Pastures a nd Fi elds 300 400 211 Im proved Pa ved stur Pastures es HI-Pasture 212 Unimpro LI-Pasture LI-Pasture Row Crops Row Crops Tree Crops Tree Crops Tree Crops Tree Crops 3 Tree Crops Feeding Operations HI-Agriculture HI-Agriculture HI-Agriculture 3 Swine Feg Oati HI-Agriculture Nurseries and Vineyards Tree Crops Tree Crops Tree Crops 3 s Tree Crops 244 Vineyards Tree Crops 245 Floriculture Tree Crops Tree Crops Tree Crops 251 Horse Farms HI-Agriculture HI-Agriculture HI-Agriculture 4 Aqulture HI-Agriculture 259 Other Specialty Farms HI-Agriculture Rangeland Rangeland 262 Old Field nd nd Shrub and Brushland nd 321 Palmetto Prairies ngeland nd nd nd Upland Forests 41 tural Land / Open Water tural Land / Open Water 412 Longleaf Pineric Oak or Longleaf Sandhitural Land / Open Water 413 Sand Pine or Sand Pine Scrub tural Land / Open Water 414 Pine Mesic Oak tural Land / Open Water 419 g Plantation Woodlands e Plantation d Hwoorests tural Land / Open Water tural Land / Open Water 220 2 2 2 13 14 15 W Ro Fi ood w C eld lan ro Cro d Pa ps ps stures 230 2 2 22 21 22 Tr Fr O ee C uit ther rop Orc Grove s G hard s rov s es 240 2 2 23 31 32 Ca Po ttle ultr Fe y F edin eed edin g O ing per Op e per atio rat ns ions ons 2 2 24 41 42 Tr So O ee N d F rnam ur arm en serie s tal N s urserie 25 2 46 Ti Sp mbe eci r N alty urs Far ery ms 0 2 2 25 52 53 D K airie enne uac s ls 260 2 O Fa ther llow Op C en L ropla an nd ds 61 Rangela Rangela Rangela Ra Rangela Rangela Rangela Na Na Na Na Na Pin Na Na 310 320 3 00 Ra H nge erba lan ceo d us 3 3 22 29 Co O M ast ther ixed al S Sh Ra cru rubs nge b an lan 330 d Br d ush 0 4 U Pi plan ne F d C lat oni woo fero ds o e-X us F r M ore esi sts c Fl 11 atwood s ll H U O untin plan ak S 420 4 ard hill od F 21 and 265

PAGE 266

Table B-1. Continued. FLUCCS Code FLUCCS Description Land Use Intensity Descripti LI LII LIII Intensity, HI=High Intensity) 422 Brazilian P on (LI=Low Intensity, MI=Middle epper Natural Land / Open Water Hickory Natural Land / Open Water Melaleuca Natural Land / Open Water 25 Temperate Hardwood ral Land / Open Water 26 T s tural Land / Open Water 27 La tural Land / Open Water 28 Cagem tural Land / Open Water 29 Wax Myrtle Ww tural Land / Open Water 0 Ud woore tural Land / Open Water 31 Beh Mno tural Land / Open Water 32 Oacru tural Land / Open Water Wern Eerglades oods tural Land / Open Water 4 Hawoo Coixed tural Land / Open Water 35 Dead Tree tural Land / Open Water Aune tural Land / Open Water 8 Md Hawo tural Land / Open Water 39 Maritimem tural Land / Open Water 0 Tree Planion e Plantation 41 Pine Planion e Plantation 2 Hawoolans e Plantation 3 Foener e Plantation 4 Exrimel Tlots e Plantation 5 Seed Tree Plan e Plantation Wr 0 Streams and Waterways tural Land / Open Water 0 Lakes tural Land / Open Water 1 Lakes Lar T00 Acres (2 Hectares) tural Land / Open Water 22 Lakes Lar T00 Acres (4ect, b Less Than 500 tural Land / Open Water 3 Lakes Lar T0 Acs (4 Hctar, but Less Than 100 tural Land / Open Water 24 Las Lehres (4 hectares) Which e Doinant Feat tural Land / Open Water 0 Reservoirs MI-Open Space / Recreational 531 Reservoirs Larger Than 500 Acres (202 Hectares) MI-Ope n Space / Recreational 532 Reservoirs Larger Th an 100 Acres (40 Hectares), but Less Than 500 Acres MI-Open Space / Recreational 533 Reservoirs Larger Than 10 Acres (4 Hectares), but Less Than 100 Acres MI-Open Space / Recreational 534 Reservoirs less than 10 Acres (4 Hectares) which are dominant features MI-Open Space / Recreational 540 Bays and Estuaries Natural Land / Open Water 541 Embayments Opening Directly into the Gulf of Mexico or the Atlantic Ocean Natural Land / Open Water 542 Embayments Not Openin g Directly into the Gulf of Mexico or the Atlantic Ocean Natural Land / Open Water 550 Major Springs Natural Land / Open Water 423 Oak Pine 424 4 4 4 4 4 Natu Na Na Na Na Na ropical Hardwood ive O abb plan k Pal Hard 43 illo od F sts 4 ec ag lia Na 4 k S b Na 433 est v Hardw Na 43 rd d nifer M Na 4 s Na 437 stralian Pi Na 43 ixe rd ods Na 4 Ha mock Na 44 tat s Pin 4 tat s Pin 44 rd d P tation Pin 44 rest R ge ation Pin 44 pe nta ree P Pin 44 tations Pin 500 ate 51 Na 52 Na 52 ger han 5 02 Na 5 ger han 1 0 H ares) ut Acres Na 52 ger han 1 re e e) s Acres Na 5 ke ss T an 10 Ac ar m ures Na 53 266

PAGE 267

Table B-1. Continued. FLUCCS Code LI LII LIII FLUCCS Description Land Use Intensity Description (LI=Low Intensity, MI=Middle Intensity, HI=High Intensity) 560 Slough Waters Natural Land / Open Water 600 Wetlands 610 Wetland Hardwood Forests Natural Land / Open Water 611 Bay Swamps Natural Land / Open Water 612 Mangrove Swamps Natural Land / Open Water 613 Gum Swamps Natural Land / Open Water 614 Shrub Swamps Natural Land / Open Water 615 Bottomland Hardwood Forest Natural Land / Open Water 616 Inland Ponds and Sloughs Natural Land / Open Water 617 Mixed Wetland Hardwoods Natural Land / Open Water 620 Wetland Coniferous Forests Natural Land / Open Water 621 Cypress Natural Land / Open Water 622 Wet Flatwoods Natural Land / Open Water 623 Atlantic White Cedar Natural Land / Open Water 624 Cypress Pine Cabbage Palm Natural Land / Open Water 630 R Wetland Mixed Forest Natural Land / Open Water 631 Hydric Hammock Natural Land / Open Water 632 Tidal Swamp Natural Land / Open Water 640 Vegetated Non-forested Wetlands Natural Land / Open Water 641 Freshwater Marshes Natural Land / Open Water 642 Salt marshes Natural Land / Open Water 643 Wet Prairies Natural Land / Open Water 644 Emergent Aquatic Vegetation Natural Land / Open Water 645 Submergent Aquatic Vegetation Natural Land / Open Water 6451 Hydrilla Natural Land / Open Water 646 Mixed Scrub-Shrub Wetland Natural Land / Open Water 650 Non-vegetated Natural Land / Open Water 651 Salt Barrens Natural Land / Open Water 652 Intertidal Areas Natural Land / Open Water 653 Inland Shores/Ephemeral Ponds Natural Land / Open Water 654 Oyster Bars Natural Land / Open Water 660 Cut over Wetlands Natural Land / Open Water 700 Barren Land 710 Beaches Natural Land / Open Water 720 Sand Other Than Beaches Natural Land / Open Water 730 Exposed Rock Natural Land / Open Water 731 Exposed Rock with Marsh Grasses Natural Land / Open Water 740 Disturbed Lands MI-Open Space / Recreational 741 Rural Land in Tr ansition Without Positive Indicators of Intended Activity MI-Open Space / Recreational 742 Borrow Areas MI-O pen Space / Recreational 743 Spoil Areas MI-O pen Space / Recreational 744 Fill Areas MI-Open Space / Recreational 745 Burned Areas MI-O pen Space / Recreational 800 Transportation, Communication and Utilities 810 Transportation LI-Transportation 267

PAGE 268

Table B-1. Continued. FLUCCS Code LI LII LIII FLUCCS Description Land Use Intensity Description (LI=Low Intensity, MI=Middle Intensity, HI=High Intensity) 811 Airports HI-Transportation 812 Railroads LI-Transportation 813 Bus and Truck Terminals HI-Transportation 814 Roads and Highways HI-Transportation 815 Port Facilitie s HI-Transportation 816 Canals and Locks HI-Transportation 817 Oil, Water, or Gas Long Distance Transmission Line HI-Transportation 818 Auto Parking Facilities (Highw ay Rest Areas) HI-Transportation 819 Transportation Facilities Unde r Construction HI -Transportation 820 Communications Industrial 821 Transmission Towers Industrial 822 Communication Facilities Industrial 829 Communication Facilities Under Construction Industrial 830 Utilities Industrial 831 Electrical Power Facilities Industrial 832 Electrical Power Tran smission Lines Industrial 833 Water Supply Plants Industrial 834 Sewage Treatment Industrial 835 Solid Waste Disposal Industrial 839 Utilities Under Construction Industrial 900 Special Classifications Natural Land / Open Water 910 Vegetative Natural Land / Open Water 911 Sea Grass Natural Land / Open Water 268

PAGE 269

APPENDIX C SUMMARY OF EMERGY EVALUATIONS FOR LAND USES Table C-1. Non-renewable and purchased empower density for urban land uses according to M. T. Browna (units: E14 sej/ha/yr). Land use Fuel Goods & services Total Single family residential (low-density) 290.33 1013.26 1303.59 Single family residential (med-density) 467.34 1685.33 2152.67 Single family residential (high-density) 454.93 1914.86 2369.80 Multi-family residential (low rise: 2 stories) 1943.81 7349.26 9293.07 Multi-family residential (high rise: 4 stories) 2568.31 10256.70 12825.02 Mobile home medium density 500.43 2247.79 2748.22 Mobile home high density 949.16 4137.84 5086.99 Commercial strip 2812.32 1823.87 4636.20 Commercial mall 13565.32 8486.60 22051.91 Industrial 3143.18 2266.40 5409.58 Central business district (2 stories) 9843.13 6307.05 16150.17 Central business district (4 stories) 17866.52 11534.66 29401.17 Universities (Institutiona l) 1207.64 2828.87 4036.51a Brown (1980). Table C-2. Non-renewable and purchased empower density for urban land uses according to N. Parkera (units: E14 sej/ha/yr). Land use Earth loss Electricity Fuel Total Single family residential (low-density) 5.83 820.04 17.47843.34 Single family residential (med-density) 5.832139.02 40.802185.65 Multi-family residential (low rise) 5.835335.20 143.605484.63 Commercial strip 8.742494.70 379.002882.44 Commercial mall 8.742884.96 379.003272.70 Industrial 5.834626.84 379.005011.67 Highway 5.83 4075.504081.33aParker (1998). 269

PAGE 270

270Table C-3. Non-renewable and purchased empower density for agricultural land uses according to S. Brandt-Williamsa (units: E13 sej/ha/yr). Land use Net Fuel Electricity Labor Lime Potash Pesticides Phosphate Nitrogen Services Subtotalb Totalc Topsoil Loss Potatoes 95.29 194.21 36.68 61.00 94.92 30.06 86.94 145.82 192.38 435.41 937.30 1372.70 Sugarcane 95.29 62.57 0.00 6.00 0.00 27.45 4.94 38.97 0.00 453.60 235.22 688.82 Tomatoes 0.78 817.36 0.00 381.00 552.72 25.76 400.68 170.13 192.38 1199.42 2540.81 3740.23 Watermelon 95.29 229.66 0.00 178.00 0.00 13.76 95.51 97.22 115.94 287.53 825.37 1112.90 Green beans 95.29 215.15 44.22 28.00 94.92 12.90 30.74 73.00 96.19 512.08 690.41 1202.49 Lettuce 95.29 291.14 0.00 172.00 0.00 34.35 111.64 97.22 192.38 451.84 994.01 1445.84 Cucumber 95.29 242.96 0.00 285.00 94.92 27.45 123.48 155.37 192.38 410.76 1216.85 1627.61 Cotton 1020.14 107.58 8.48 40.00 94.92 13.76 12.52 58.42 77.04 123.08 1432.85 1555.92 Cabbage 95.29 193.20 36.68 91.00 94.92 34.35 16.63 170.13 192.38 121.31 924.57 1045.88 Cornd 526.89 90.01 21.10 6.00 62.66 20.70 4.26 77.95 231.03 132.75 1040.59 1173.34 Corne 302.68 138.31 0.00 113.00 0.00 25.76 27.97 145.82 192.38 212.50 945.93 1158.43 Pepper 95.29 690.64 20.13 728.00 0.00 31.83 330.12 194.75 178.24 577.80 2269.00 2846.80 Oranges 7.85 221.01 12.58 120.00 40.32 43.55 45.11 41.58 121.95 121.36 653.94 775.31 Pasturef 0.78 27.31 5.95 2.00 62.69 6.71 0.00 27.27 62.63 6.13 195.34 201.47 Beefg 1.00 133.00 0.00 37.00 93.00 13.00 27.00 28.00 125.00 136.00 457.00 593.00 Milkh 95.00 194.00 135.00 57.00 156.00 28.00 6.00 124.00 205.00 1177.00 1000.00 2177.00 a Brandt-Williams (2001) bWithout services; cWith services. dGrain; eSweet; fBahia; g2 steers/ha; hPer cow/yr.

PAGE 271

APPENDIX D WATER CHEMISTRY DATA FOR THE SAMPLE FRESHWATER SYSTEMS Table D-1. Water chemistry variables considered for 75 sample isol ated forested wetlands (x = variable measured). Site code Turbidity Specific conductance Total nitrogen Total phosphorus Dissolved oxygen Site code Turbidity Specific conductance Total nitrogen Total phosphorus Dissolved oxygen SA2 x x x x x CR3 x x x x x SA3 x x x x x CR4 x x x x x SA4 x x x CR5 x x x x SA5 x x x x CR6 x x x x SA6 x x x x CR8 x x x x SA7 x x x x CR9 x x x SA8 x x x x CR10 x x x x SA9 x x x CR11 x x x x SR1 x x x x x CU1 x x x x x SR2 x x x x x CU3 x x x x x SR3 x x x x x CU5 x x x x x SR4 x x x x CU6 x x x x x SR5 x x x x CU7 x x x x SR7 x x x x CU8 x x x SR8 x x x CU9 x x x SR9 x x x x CU10 x x x SU1 x x x x x NA6 x x x x SU2 x x x x x NA10 x x x SU3 x x x x x NA11 x x x SU4 x x x x NR2 x x x x x SU5 x x x x NR3 x x x x x SU6 x x x x x NR4 x x x x x SU7 x x x x NR6 x x x x SU8 x x x NR8 x x x x CA2 x x x x NR9 x x x CA3 x x x x x NU2 x x x x x CA4 x x x x x NU4 x x x x x CA5 x x x x x NU5 x x x x x CA7 x x x NU6 x x x x CA8 x x x x NU10 x x x CR1 x x x x PA2 x x x x x PA3 x x x x x PR8 x x x x PA5 x x x x x PU1 x x x x x PA6 x x x x x PU3 x x x x x PR4 x x x x x PU4 x x x x PR5 x x x x x PU6 x x x x PR6 x x x x PU10 x x PR7 x x x x 271

PAGE 272

Table D-2. Water chemistry variables considered for 47 STORET stream stations (x = data available). STORET Station # Data Source Period sampled # of samples Turbidity DOa NO3 b TNc TPd WQIe 19010099 305(b) report 93-94 18 x x x x x x 20010454 305(b) report 93-95 8 x x x x x x 20010455 305(b) report 93-94 6 x x x x x x 20020004 305(b) report 94-95 5 x x x x x x 20020012 305(b) report 90-95 30 x x x x x x 20020317 305(b) report 90-94 10 x x x x x x 20020404 305(b) report 93-94 7 x x x x x x 20030342 STORET 95-96 2 x x x 20030419 STORET 94-95 2 x x x 20030437 STORET 94-96 3 x x x 21010018 305(b) report 94-94 8 x x x x x x 22020049 305(b) report 93-94 8 x x x x x x 22020062 STORET 94-96 4 x x x 22030062 305(b) report 93-93 6 x x x x x x 23010464 305(b) report 93-95 7 x x x x x x 24010002 305(b) report 93-94 6 x x x x x x 24020134 305(b) report 94-95 3 x x x x x x 24030013 305(b) report 93-95 7 x x x x x x 24030044 305(b) report 93-94 6 x x x x x x 25020014 305(b) report 93-95 7 x x x x x x 25020111 305(b) report 90-94 32 x x x x x x 26010029 305(b) report 92-94 4 x x x x x x 26010430 STORET 95-96 7 x x x 26010593 305(b) report 94-94 4 x x x x x x 26010972 STORET 94-95 2 x x x 26011019 305(b) report 93-94 6 x x x x x x 26011020 305(b) report 93-94 6 x x x x x x 28010223 305(b) report 93-95 344 x x x x x x 28010224 305(b) report 93-95 11 x x x x x x 28010239 305(b) report 94-95 3 x x x x x x 28010608 305(b) report 93-95 365 x x x x x x 28020147 STORET 94-96 2 x x x 28020148 305(b) report 92-94 7 x x x x x x 28020221 305(b) report 93-94 6 x x x x x x 31010050 305(b) report 93-94 6 x x x x x x 31010051 305(b) report 93-94 5 x x x x x x 31020038 305(b) report 92-94 10 x x x x x x 31020040 305(b) report 93-94 9 x x x x x x 32010021 305(b) report 92-94 9 x x x x x x 32020063 305(b) report 93-94 6 x x x x x x 32030023 305(b) report 93-94 8 x x x x x x 32030024 305(b) report 93-94 9 x x x x x x 33010054 STORET 92-95 8 x x x 33010065 STORET 95-96 2 x x x 33010068 STORET 95-96 2 x x x 33040014 305(b) report 93-94 7 x x x x x x 33040015 305(b) report 93-94 7 x x x x x x aDO = Dissolved oxygen; bNO3 = Nitrate nitrogen; cTN = Total nitrogen, calculated from STORET data; dTP = Total phosphorus; eWQI = Water Quality Index. 272

PAGE 273

Table D-3. Water chemistry variables considered for 54 STORET lake stations. All data provided by FDEP (R. Frydenborg 2005, Envi ronmental Assessment Section, Bureau of Laboratories personal communica tion) (x = data available). STORET Station # Period sample d # of samples Ammonia Na NO3/NO2 b TKNc TNd TPe 20010048 2000 1 x x x x x 20010110 1998 1 x x x x x 20010222 1998 1 x x x x x 20010299 1997 1 x x x x x 20010311 1998 1 x x x x x 20010334 1998 1 x x x x x 20010336 1998 1 x x x x x 20010337 1998 1 x x x x x 20020014 1996-1997 2 x x x x x 20020015 1996-1998 2 x x x x x 20020062 1997 1 x x x x x 20020064 1998-2000 2 x x x x x 20020065 1998-2000 3 x x 20020066 1998 1 x x x x x 20030417 1998 1 x x x x x 20030438 1998 1 x x x x x 23010434 1998 1 x x x x x 23010435 1997 1 x x x x x 25010079 1997 2 x x x x x 25020552 1998 2 x x x x x 25020554 1999 2 x x x x x 26010032 1996-1998 2 x x x x x 26010037 1998 1 x x x x x 26010039 1998 1 x x x x x 26010040 1997 1 x x x x x 26010105 2000 1 x x x x x 26010116 1997 1 x x x x x 26010303 1998 2 x x x x x 26010304 1998 2 x x x x x 26010325 1999 2 x x x x x 26010326 1999 2 x x x x x 26010327 2000 2 x x x x x 26010331 1998 2 x x x x x 26010526 2000 2 x x x x x 26010528 2000 2 x x x x x 26010531 2000 2 x x x x x 26010556 1999 2 x x x x x 26010585 1999 2 x x x x x 26010591 1999 2 x x x x x 26010605 1998 2 x x x x x 26010644 1999 2 x x x x x 26010645 1997 2 x x x x x 26010646 1997 2 x x x x x 26010647 1997 2 x x x x x 26010648 1997 2 x x x x x 28020242 1998 2 x x x x x 28030068 1997 2 x x x x x 273

PAGE 274

Table D-3. Continued. STORET Station # Period sample d # of samples Ammonia Na NO3/NO2 b TKNc TNd TPe 32010038 1997-1998 2 x x x x x 32020113 1997-1999 2 x x x x x 32030081 1998-1999 2 x x x x x 33010064 1997-1998 2 x x x x x 33020097 1998-1999 2 x x x x x 33020098 1998-1999 2 x x x x x 33030057 1996-1998 3 x x x x x aAmmonia N = Ammonia nitrogen; b NO3/NO2 = Nitrite-nitrate nitrogen; cTKN = Total Kjeldahl nitrogen; dTN = Total nitrogen; eTP = Total phosphorus. 274

PAGE 275

APPENDIX E WETLAND CONDITION INDEX Table E-1. Metric composition of the WCI in cluding diatoms WCI, macrophytes WCI, and macroinvertebrates WCI (Source: Reiss 2004). Diatoms Macrophytes Macroinvertebrates % Tolerant Indicator Species % Tolerant Indicator Species % Tolerant Indicator Species % Sensitive Indicator Species % Sensitive Indi cator Species % Sensitive Indicator Species Pollution class 1a Modified FQIf Florida Indexg Nitrogen class 3b % Exotic species % Mollusca Saprobity class 4c % Native perennial % Noteridae pH class 3d % Wetland status species % Scrapers Dissolved oxygen class 1e a Very tolerant to pollution; b Need periodically elevated concentration of organically bound nitrogen; c inhabit aquatic environments w ith an oxygen saturation between 1025% and a biological oxygen demand of approximately 13-22 mg/L; d Mainly occurring at pH values close to 7; e requiring continuously high dissolved oxygen concentrations near 100%; f Modified Floristic Quality Index; g Weighted sum of intolerant taxa, which are classed as 1 (least to lerant) or 2 (intolerant). Table E-2.WCI scores for 118 wetlands based on three assemblages including diatoms, macrophytes, and macroinverteb rates (Source: Reiss 2004). Site code Diatom WCI Macrophyte WCI Macroinvertebrate WCI Site code Diatom WCI Macrophyte WCI Macroinvertebrate WCI PA1 30.4 CA1 8.9 PA2 38.1 11.9 30.1 CA2 10.6 0.7 19.4 PA3 34.9 8.3 25.1 CA3 7.9 7.1 20.0 PA4 12.6 CA4 56.9 38.8 31.3 PA5 51.1 6.5 21.2 CA5 43.6 26.9 7.2 PA6 28.2 7.7 12.9 CA6 22.7 7.1 21.1 PA7 17.7 CA7 9.8 32.1 PA8 50.6 CA8 37.7 31.3 PA9 12.1 CA9 11.8 22.6 PA10 41.7 CR1 51.0 PR1 61.1 55.9 37.6 CR2 49.9 PR2 50.5 CR3 57.7 47.6 33.9 PR3 49.5 CR4 57.8 51.2 48.9 PR4 64.5 51.2 40.0 CR5 43.8 43.5 29.7 PR5 58.0 53.6 30.0 CR6 65.5 54.6 50.4 PR6 63.9 58.4 34.4 CR7 51.7 PR7 34.8 26.5 CR8 54.3 28.8 PR8 53.6 40.7 CR9 49.4 34.4 PU1 6.2 CR10 53.5 45.0 275

PAGE 276

Table E-2. Continued. Site code Diatom WCI Macrophyte WCI Macroinvertebrate WCI Site code Diatom WCI Macrophyte WCI Macroinvertebrate WCI PU2 31.5 CR11 59.0 49.5 PU3 33.1 31.0 35.7 CU1 61.1 42.9 40.6 PU4 10.5 4.0 21.6 CU2 10.0 PU5 22.1 CU3 28.5 13.5 22.3 PU6 16.5 CU4 21.4 PU7 24.1 CU5 21.5 22.3 17.8 PU8 33.6 CU6 15.1 41.5 23.4 PU9 48.8 CU7 20.7 10.6 PU10 9.2 30.2 CU8 21.1 10.1 NA1 0.0 CU9 28.3 28.3 NA2 3.0 CU10 38.3 32.3 NA3 56.4 CU11 21.3 34.1 NA4 33.8 16.3 10.4 SA1 0.7 NA5 2.9 SA2 34.1 9.4 15.0 NA6 56.3 18.8 16.9 SA3 47.9 23.1 28.6 NA7 37.0 SA4 15.8 11.3 9.1 NA8 46.0 SA5 46.3 18.9 19.0 NA9 37.3 SA6 31.9 3.7 19.0 NA10 51.5 30.0 SA7 30.8 29.8 NA11 32.6 28.7 SA8 34.5 11.0 NA12 8.0 SA9 29.8 17.9 NR1 52.0 SR1 66.8 54.1 33.4 NR2 65.8 34.8 30.0 SR2 68.9 50.8 46.6 NR3 66.8 58.2 52.8 SR3 51.6 51.2 26.4 NR4 58.3 42.2 39.8 SR4 43.7 57.9 28.2 NR5 52.3 SR5 39.4 49.8 38.4 NR6 57.9 55.0 48.6 SR6 41.0 51.8 39.0 NR7 52.3 SR7 49.9 42.7 NR8 58.4 30.0 SR8 47.5 57.0 NR9 56.7 33.0 SR9 50.1 43.4 NU1 35.2 SU1 17.2 17.8 22.1 NU2 24.1 23.7 15.5 SU2 46.2 20.3 15.2 NU3 25.6 SU3 31.7 42.6 35.4 NU4 54.5 35.1 31.0 SU4 42.3 21.8 18.9 NU5 60.0 40.1 24.0 SU5 38.9 23.9 5.3 NU6 48.8 20.7 28.0 SU6 46.1 28.1 21.0 NU7 11.8 SU7 12.5 9.1 NU8 38.6 SU8 2.7 23.3 NU9 37.5 SU9 20.4 32.3 NU10 17.2 23.0 SU10 11.7 276

PAGE 277

APPENDIX F STREAM CONDITION INDEX Table F-1. Macroinvertebrate metric composition of the SCI defined by Barbour and colleaguesa. Core Metrics Description Total taxa Measures the overall variety of macroinvertebrates. EPT taxa Sum of the number of taxa of the orders Ephemeroptera (mayflies), Plecoptera (st oneflies), and Trichoptera (caddisflies). Chironomidae taxa Number of unique taxa of chiromids (midges). % Dominant taxon Relative abundance of the most abundant taxon. % Diptera Relative abundance of indivi duals classed as dipterans (true fly larvae). Florida index Weighted sum of intole rant taxa, classified as 1 (least tolerant) and 2 (intolerant). Fl orida index = 2 X class 1 taxa + 1 X class 2 taxa. % Filterers Relative abundance of the sample that filters suspended detritus. aBarbour et al (1996b). Table F-2. Macroinvertebrate metric co mposition of the SCI defined by S. Forea. Core Metrics Description Total taxa Measures the overall variety of macroinvertebrates. Ephemeroptera taxa Number of un ique taxa found within the order Ephemeroptera (mayflies). Trichoptera taxa Number of unique taxa found within the order Trichoptera (caddisflies). % Filterers Relative abundance of the f ilterer functional feeding group. Long-lived taxa Number of unique taxa which requires more than a year to complete their life cycles. Clinger taxa Number of unique ta xa that attaches to substrates. % Dominance Relative abundance of the most abundant taxon. % Tanytarsini Relative abundance of the Tanytarsini tribe of the Chironimid (midges) family. Sensitive taxa Number of unique taxa sensitive to human disturbance. % Very tolerant Relative abundance of taxa very tolerant to human disturbance. aFore (2004) 277

PAGE 278

Table F-3. SCI scores for 69 streams for the macroinvertebrate asse mblage. (Source: R. Frydenborg 2005, FDEP, Environmental Assessment Section, Bureau of Laboratories, personal communication). Site code STORET station # # of Samples SCI_1a Score SCI_2b Score Site code STORET station # # of Samples SCI_1 Score SCI_2 Score S1 19010042 3 29 77 S36 25020014 6 28 58 S2 19010099 5 29 73 S37 25020111 5 31 82 S3 19020027 1 29 70 S38 26010029 3 21 27 S4 20010454 5 27 60 S39 26010430 3 28 63 S5 20010455 5 29 55 S40 26010593 5 32 75 S6 20020004 4 32 94 S41 26010972 2 30 80 S7 20020012 7 30 69 S42 26011019 5 29 70 S8 20020317 5 28 57 S43 26011020 5 30 75 S9 20020404 6 31 89 S44 28010223 7 28 51 S10 20020424 1 19 10 S45 28010224 8 31 50 S11 20030263 1 29 60 S46 28010232 1 19 15 S12 20030264 1 29 65 S47 28010239 5 29 59 S13 20030265 1 25 25 S48 28010608 8 27 38 S14 20030340 1 29 75 S48 28020147 2 32 70 S15 20030341 1 29 65 S50 28020148 5 31 78 S16 20030342 1 29 40 S51 28020221 6 31 81 S17 20030419 2 28 65 S52 28020232 1 29 45 S18 20030437 3 28 65 S53 28020233 1 25 20 S19 20030549 1 29 85 S54 28020234 1 60 S20 20030550 1 21 35 S55 31010050 7 32 79 S21 21010018 3 25 62 S56 31010051 8 32 84 S22 21010032 2 26 40 S57 31020037 1 27 50 S23 22020010 1 31 50 S58 31020038 1 25 45 S24 22020049 7 29 62 S59 31020040 6 31 71 S25 22020062 5 26 54 S60 32010021 6 31 75 S26 22020077 1 29 70 S61 32020030 2 16 8 S27 22020093 1 31 55 S62 32020063 6 31 73 S28 22030062 4 31 59 S63 32030023 6 31 75 S29 22030064 1 13 5 S64 32030024 5 33 75 S30 23010464 7 28 54 S65 33010054 6 32 81 S31 24010002 5 30 84 S66 33010065 1 33 85 S32 24020134 4 31 74 S67 33010068 1 29 65 S33 24030013 7 28 57 S68 33040014 6 31 73 S34 24030044 5 28 65 S69 33040015 6 31 83 S35 24030142 1 29 90 aSCI defined by Barbour et al. (1996b). bSCI defined by Fore (2004). 278

PAGE 279

APPENDIX G LAKE CONDITION INDEX Table G-1. Macroinvertebrate metric compositi on of the LCI (source: Gerritsen et al. 2000). Core Metrics Description Total taxa Measures the overall variety of macroinvertebrates. EOT taxa Sum of the number of ta xa of the orders Ephemeroptera (mayflies), Odonata (dragonf lies and damselflies), and Trichoptera (caddisflies). % EOT Relative abundance of indivi duals classed as mayflies, dragonflies and damselflies, and caddisflies. Hulbert Index Macroinvertebrate component of the Hulberts Lake Condition Index. Shannon-Wiener diversity Measure of the general di versity and composition of macroinvertebrates (considers both richness and evenness). % Diptera Relative abundance of indivi duals classed as dipterans (true fly larvae). Table G-2. LCI scores for 54 lakes for the macr oinvertebrate assemblage (source: R. Frydenborg 2005, FDEP, Environmental Assessment Secti on, Bureau of Laboratories, personal communication). STORET station # # of Samples LCI score STORET station # # of Samples LCI score 20010048 26010303 2 13.94 20010110 1 34.39 26010304 2 52.60 20010222 1 24.89 26010325 2 66.65 20010299 1 42.43 26010326 2 25.07 20010311 1 18.48 26010327 2 46.97 20010334 1 37.34 26010331 2 66.09 20010336 1 22.99 26010526 2 75.57 20010337 1 11.74 26010528 2 72.34 20020014 1 49.18 26010531 2 69.21 20020015 1 35.91 26010556 2 61.81 20020062 1 21.49 26010585 2 53.42 20020064 2 65.36 26010591 2 31.89 20020065 3 54.34 26010605 2 55.56 20020066 1 36.54 26010644 2 38.13 20030417 1 87.96 26010645 1 38.15 20030438 1 57.64 26010646 1 41.38 23010434 1 51.66 26010647 2 24.25 279

PAGE 280

Table G-2. Continued. STORET station # # of Samples LCI score STORET station # # of Samples LCI score 23010435 1 54.74 26010648 2 36.27 25010079 2 58.47 28020242 2 26.27 25020552 2 9.68 28030068 2 58.95 25020554 2 43.85 32010038 2 30.28 26010032 1 78.53 32020113 2 49.06 26010037 1 44.33 32030081 2 43.81 26010039 1 44.32 33010064 2 26.67 26010040 1 16.29 33020097 2 37.57 26010105 1 37.71 33020098 2 64.30 26010116 1 25.64 33030057 2 35.56 280

PAGE 281

APPENDIX H MFWORKS SCRIPTS Script 1: Land use proportion LDI (regardl ess of distance from study aquatic system). Recode_map= Recode Empower_map Assigning 1 To 0...15000; Total_area_map= Measure Recode _map Hectares Ignore VOID; LU_area_map= Measure Empower_map Hectares; Fraction_map= LU_area_map/Total_area_map; LU_empower_map= Fraction_map*Empower_map; Cell_measure_map= Measure LU_empower_map Cells; New_fraction_map= (LU_empower_map/Cell_measure_map)*100000000; Trunk_map= Trunc(New_fraction_map); New_trunc_map= Trunc(Total_area_map); Total_score_map= Score New_trunc _map By Trunc_map Total; Float_total_score_map= Float(Total_score_map); Aw_final_value_map= Float_total_score_map/1000000; Save Aw_final_value_map; Script 2: Inverse linear distan ce LDI (linear decrease with dist ance from study aquatic system). Mask_map= (Empower_map*0)+1; Recode_mask_map= Recode Mask_map Assigning 9999999 To VOID CarryOver; Spread_map= Spread Seed_map To 100000 In Recode_mask_map; Mask_all_map= Recode Mask_map Assigning 1 to 100000000.00000 CarryOver; Trunc_mask_all_map= Trunc(Mask_all_map); Float_spread_map= Float(Spread_map); Max_spread_value_map= Score Trunk_mask_all_map By Float_spread_map Maximum; Inverse_distance_map= (Spread_mapMax_spread_value_map)*-1; Norm_distance_map= Inverse_distance_map/Max_spread_value_map; Ldw_empower_map= Norm_distance_map*Empower_map; Trunc_mask_map= Trunc(Mask_map); Ldw_final_value_map= Score Trunk_mask_map By Ldw_empower_map Average; Save Ldw_final_value_map; Script 3: Inverse square distan ce LDI (square decrease with distance from study aquatic system). Mask_map= (Empower_map*0)+1; Recode_maskmap= Recode Mask_map Assigning 9999999 To VOID CarryOver; Spread_map= Spread Seed_map To 100000 In Recode_mask; Mask_all_map= Recode Mask_map Assigning 1 to 100000000.00000 CarryOver; Trunk_mask_all_map= Trunc(Mask_all_map); Float_spread_map= Float(Spread_map); Max_spread_value_map= Score Trunk_mask_all_map By Float_spread_map Maximum; Inverse_distance_map= (Spread_mapMax_spread_value_map)*-1; Square_distance_map= Distance_map*Distance_map; 281

PAGE 282

Norm_distance_map= Square_distance_map/ (Max_spread_value_map*Max_spread_value_map); Sqdw_empower_map= Norm_distance_map* Empower_map; Trunk_mask_map= Trunc(Mask_map); Sqdw_final_value_map= Score Trunk_mask_map By Sqdw_empower_map; Save Sqdw_final_value_map; 282

PAGE 283

APPENDIX I LAND USE/LAND COVER SURROUNDING THE ISOLATED FORESTED WETLANDS Table I-1. Characteristics of the land use/land co ver surrounding the isolated forested wetlands (n = 118). Categories defined according to Le vel 1 of the FLUCCS classification scheme. Blank spaces denote a value of zero. Site Code % Urban % Agriculture % Rangeland % Forest % Water % Wetland % Transportation CA1 94.6 1.7 3.7 CA2 86.9 1.6 11.5 CA3 84.1 0.4 1.0 12.5 2.0 CA4 84.7 15.3 CA5 98.6 1.4 CA6 94.1 2.1 3.8 CA7 51.9 16.5 27.8 3.8 CA8 0.8 72.3 18.6 8.3 CA9 97.7 2.3 CR1 89.8 10.2 CR10 98.1 1.9 CR11 12.0 0.3 87.6 CR2 97.3 2.7 CR3 58.7 41.3 CR4 76.8 20.4 2.8 CR5 95.8 2.0 2.2 CR6 63.0 37.0 CR7 67.0 29.3 3.6 CR8 92.2 4.3 3.5 CR9 89.7 3.1 CU1 16.8 62.6 0.5 20.1 CU10 96.4 3.6 CU11 96.2 2.1 1.7 CU2 78.4 21.6 CU3 75.0 11.3 0.8 12.9 CU4 52.4 6.7 24.5 0.3 16.1 CU5 40.2 2.4 7.1 50.2 CU6 86.1 9.2 1.7 2.9 CU7 71.9 15.7 5.2 7.3 CU8 23.8 56.3 2.9 17.0 CU9 56.6 3.5 2.7 37.2 NA1 87.9 0.7 7.4 3.9 NA10 91.2 0.5 8.3 NA11 86.4 2.2 0.5 9.2 1.7 283

PAGE 284

Table I-1. Continued. Site Code % Urban % Agriculture % Rangeland % Forest % Water % Wetland % Transportation NA12 64.5 15.1 0.2 17.0 3.2 NA2 93.7 3.8 2.6 NA3 70.4 26.6 2.9 NA4 88.0 2.6 3.9 5.5 NA5 86.3 2.6 4.4 6.7 NA6 79.9 3.7 8.8 4.1 3.6 NA7 59.1 40.9 NA8 99.5 0.5 NA9 85.1 1.3 9.1 4.5 NR1 97.6 1.0 1.4 NR2 89.7 7.2 3.1 NR3 82.3 17.7 NR4 81.3 16.1 2.5 NR5 96.5 3.5 NR6 84.5 9.8 5.7 NR7 97.7 1.9 0.4 NR8 98.6 1.4 NR9 93.1 5.2 1.7 NU1 26.3 9.7 49.0 1.7 2.3 11.0 NU10 87.0 7.0 0.4 5.5 NU2 73.3 22.3 4.4 NU3 39.7 53.1 1.0 6.2 NU4 51.2 33.2 2.4 3.0 10.1 NU5 69.9 12.2 1.8 16.1 NU6 54.3 40.0 5.6 NU7 76.9 15.6 1.5 6.0 NU8 58.8 26.2 8.4 6.5 NU9 77.1 17.5 5.4 PA1 49.7 48.9 1.5 PA10 97.4 2.6 PA2 0.4 71.2 15.2 10.7 2.5 PA3 14.5 83.3 0.8 1.4 PA4 2.2 90.7 1.5 5.6 PA5 66.2 2.7 31.1 PA6 0.3 62.9 25.2 3.1 4.3 4.2 PA7 80.1 7.0 0.9 12.0 PA8 98.2 1.8 PA9 43.2 53.8 3.1 PR1 64.4 33.1 2.5 PR2 7.7 0.4 91.8 PR3 89.7 7.2 3.1 PR4 2.9 90.7 6.5 PR5 77.2 18.6 4.2 PR6 81.3 13.9 4.9 284

PAGE 285

Table I-1. Continued. Site Code % Urban % Agriculture % Rangeland % Forest % Water % Wetland % Transportation PR7 98.5 1.5 PR8 84.5 14.2 1.3 PU1 77.2 12.8 10.0 PU10 91.4 3.4 5.2 PU2 62.3 2.4 18.5 7.6 9.3 PU3 84.3 5.7 10.1 PU4 79.3 16.0 4.7 PU5 40.3 32.7 1.1 17.7 8.2 PU6 25.1 36.2 0.7 28.5 9.4 PU7 68.7 20.0 11.3 PU8 87.8 5.9 6.3 PU9 83.0 14.1 2.9 SA1 96.1 3.9 SA2 76.0 5.4 18.7 SA3 60.4 15.5 24.1 SA4 77.0 0.1 10.9 11.2 0.7 SA5 76.6 14.7 8.7 SA6 93.5 3.7 0.9 1.8 SA7 98.2 1.8 SA8 11.7 84.3 3.9 SA9 99.9 0.1 SR1 77.9 22.1 SR2 54.2 45.8 SR3 78.7 20.7 0.6 SR4 13.4 85.2 1.4 SR5 14.6 78.4 7.0 SR6 100.0 SR7 69.4 29.1 1.6 SR8 56.9 36.4 6.7 SR9 14.5 66.8 15.1 3.6 SU1 51.3 25.6 5.3 17.8 SU10 77.7 3.3 1.7 4.0 13.3 SU2 48.8 24.8 12.5 13.9 SU3 17.6 4.2 50.5 27.7 SU4 85.9 1.3 12.7 SU5 53.2 43.1 0.5 3.3 SU6 46.9 21.1 22.0 7.0 3.1 SU7 67.5 2.8 20.3 9.4 SU8 84.1 2.4 13.4 SU9 77.6 16.8 3.7 2.0 285

PAGE 286

APPENDIX J LAND USE/LAND COVER SURROUNDING STREAMS Table J-1. Characteristics of the land use/land cover surrounding the sample streams (n = 69). Categories defined according to Level 1 of the FLUCCS classification scheme. Blank spaces denote a value of zero. Site code % Urban % Agriculture % Rangeland % Forest % Water % Wetland % Barren land % Transportation S1 0.1 62.0 37.8 S2 6.1 5.9 10.6 56.9 0.3 16.7 0.4 3.1 S3 6.7 12.4 1.8 51.5 0.4 27.1 0.1 S4 0.2 0.3 90.3 0.4 8.7 0.1 S5 10.0 12.7 4.3 39.5 7.7 25.2 0.6 0.1 S6 10.9 3.3 0.8 49.9 11.2 23.0 0.6 0.2 S7 9.3 9.3 2.2 49.3 6.5 23.0 0.2 0.2 S8 12.6 4.3 1.9 66.4 0.8 13.1 0.9 S9 3.4 16.3 1.7 33.1 10.2 34.8 0.1 0.3 S10 17.9 11.8 3.2 30.0 3.0 33.3 0.4 0.4 S11 9.4 0.4 2.5 59.2 1.1 22.9 4.4 S12 2.5 4.4 7.1 50.5 0.2 20.7 0.1 14.6 S13 2.5 3.6 5.9 54.7 0.1 20.8 0.1 12.3 S14 4.0 0.4 0.5 63.6 0.7 29.4 1.5 S15 5.9 2.3 1.8 67.0 0.3 20.7 2.0 S16 5.5 2.6 3.3 62.2 0.3 20.3 5.9 S17 10.8 8.1 60.9 4.3 15.3 0.3 0.3 S18 7.1 2.3 2.3 63.9 1.9 19.6 0.1 2.8 S19 5.4 2.4 1.7 68.4 0.3 19.6 2.2 S20 9.8 0.4 1.7 60.0 1.2 22.8 4.1 S21 2.1 1.9 1.1 71.0 0.1 22.8 0.2 0.7 S22 0.2 1.7 0.1 56.2 0.1 41.9 S23 11.2 17.0 0.7 65.3 1.4 3.4 0.1 0.8 S24 2.6 0.7 85.6 10.8 0.1 S25 0.4 4.8 85.6 9.1 S26 39.7 47.3 13.0 S27 10.5 20.0 0.4 63.3 1.6 3.4 0.1 0.7 S28 28.8 4.2 3.2 56.7 5.1 0.2 1.7 S29 99.4 0.6 S30 6.9 30.2 13.3 16.2 0.5 32.4 0.1 0.5 286

PAGE 287

Table J-1. Continued. Site code % Urban % Agriculture % Rangeland % Forest % Water % Wetland % Barren land % Transportation S31 1.4 27.3 41.7 18.2 0.3 10.7 0.4 S32 16.9 51.6 4.5 12.6 1.9 12.4 0.1 S33 25.3 25.4 14.3 10.6 0.8 21.9 1.7 S34 23.4 36.1 8.7 9.7 1.0 19.9 1.3 S35 17.1 21.3 18.4 12.5 0.8 27.4 2.5 S36 68.8 6.2 7.2 0.2 17.7 S37 0.5 48.6 16.2 15.9 0.1 18.6 0.1 S38 45.3 24.0 2.8 8.5 7.7 11.4 0.2 S39 2.0 75.9 4.8 4.7 1.8 10.6 0.2 S40 0.9 42.2 11.9 20.7 0.3 23.9 0.1 S41 9.8 36.6 0.4 7.7 2.5 39.4 0.9 2.6 S42 6.1 41.1 8.5 12.1 14.6 17.2 0.4 S43 10.3 53.1 1.6 14.5 6.4 13.7 0.3 S44 6.5 4.6 1.5 51.8 0.7 32.5 2.3 S45 70.0 0.6 0.6 21.4 0.9 6.6 S46 10.2 65.6 0.3 16.3 2.1 2.8 2.7 S47 4.3 68.5 2.8 13.0 1.7 7.5 0.7 1.4 S48 5.1 63.4 3.0 16.2 1.6 8.7 0.7 1.3 S48 27.2 22.3 6.1 21.6 0.5 21.5 0.8 S50 65.7 4.7 5.9 9.1 2.0 8.3 3.2 1.1 S51 0.1 21.0 3.3 49.6 25.8 0.1 S52 13.2 25.5 9.4 27.9 0.1 21.3 1.8 0.7 S53 45.9 14.1 3.6 17.5 0.9 9.0 4.1 5.0 S54 14.6 24.5 0.8 14.8 0.9 37.0 6.2 1.2 S55 2.2 2.1 4.8 89.7 0.3 0.9 S56 0.4 1.6 3.3 93.4 0.8 0.6 S57 1.8 19.6 20.6 49.8 0.3 7.8 0.1 S58 3.4 38.4 0.8 41.6 0.5 14.4 0.2 0.7 S59 3.0 7.7 17.2 55.9 0.6 15.0 0.5 S60 6.3 2.9 4.0 79.5 0.5 5.3 0.1 1.5 S61 41.9 30.7 2.3 20.4 0.2 4.4 S62 76.8 23.1 S63 12.4 3.1 12.7 58.4 2.9 9.5 1.1 S64 0.1 0.2 0.1 91.4 7.0 1.3 S65 2.1 23.7 0.3 61.2 0.2 12.0 0.4 S66 19.3 4.0 56.6 0.5 13.9 5.6 S67 25.4 12.8 0.7 53.8 1.3 3.7 2.4 S68 1.3 19.2 1.9 64.5 0.8 11.6 0.6 S69 1.3 26.4 4.1 51.9 1.3 14.8 0.1 287

PAGE 288

APPENDIX K LAND USE/LAND COVER SURROUNDING LAKES Table K-1. Characteristics of the land use/land cover surrounding the sample lakes (n = 54). Categories defined according to Level 1 of the FLUCCS classification scheme. Blank spaces denote a value of zero. Site code % Urban % Agriculture % Rangeland % Forest % Water % Wetland % Barren land % Transportation L1 99.83 0.17 L2 32.73 3.44 17.49 46.33 L3 86.99 2.88 0.34 1.04 8.75 L4 93.37 0.69 5.94 L5 30.32 2.72 0.28 1.98 64.70 L6 87.91 0.10 0.75 0.80 2.81 7.64 L7 88.02 1.40 10.58 L8 89.95 0.12 6.60 3.34 L9 6.44 43.50 29.92 20.14 L10 0.84 24.93 48.32 25.90 L11 9.88 21.77 17.52 50.83 L12 45.44 22.10 10.78 21.69 L13 10.82 58.02 2.75 24.11 4.29 L14 59.54 17.04 8.51 5.21 7.70 2.01 L15 32.48 3.57 45.13 0.08 18.05 0.08 0.61 L16 6.06 22.48 46.54 4.34 19.97 0.62 L17 37.35 24.37 0.25 9.86 2.57 25.59 0.02 L18 37.35 24.37 0.25 9.86 2.57 25.59 0.02 L19 0.40 0.70 39.55 31.84 0.00 22.89 0.03 4.59 L20 100.00 L21 90.73 3.31 1.05 3.56 L22 20.80 38.22 0.64 9.04 8.14 22.71 0.11 0.34 L23 15.42 21.21 39.56 0.22 23.08 0.49 L24 10.48 27.56 16.43 0.58 44.84 0.10 L25 40.44 5.89 0.14 53.53 L26 96.13 0.97 2.90 L27 52.63 33.04 0.42 1.42 5.99 6.50 L28 35.19 54.48 3.93 6.40 L29 63.19 20.36 4.73 3.46 2.95 3.56 1.75 L30 64.94 32.90 0.48 0.83 0.85 L31 13.86 67.51 6.13 8.85 3.65 L32 44.99 25.09 14.58 1.66 13.19 0.48 288

PAGE 289

Table K-1. Continued. Site code % Urban % Agriculture % Rangeland % Forest % Water % Wetland % Barren land % Transportation L33 36.48 35.52 1.53 18.77 7.71 L34 67.31 10.51 0.68 5.59 10.25 3.08 2.58 L35 67.31 10.51 0.68 5.59 10.25 3.08 2.58 L36 67.31 10.51 0.68 5.59 10.25 3.08 2.58 L37 62.67 11.03 2.20 12.68 2.19 6.40 2.84 L38 98.42 0.55 1.04 L39 47.27 42.00 0.90 4.97 4.34 0.53 L40 50.79 49.21 L41 30.33 25.81 0.87 25.01 1.39 13.90 2.69 L42 49.96 6.08 6.97 7.02 0.20 29.78 L43 26.24 17.99 23.98 17.35 0.95 13.49 L44 42.04 44.28 13.68 L45 18.73 73.35 7.69 0.23 L46 83.83 16.17 L47 75.21 0.54 2.86 12.26 6.80 2.33 L48 2.35 53.33 29.25 15.06 L49 29.16 26.06 35.64 0.38 8.19 0.57 L50 77.35 5.93 0.44 1.96 14.31 L51 87.22 2.01 4.98 1.74 4.06 L52 11.48 7.10 78.11 0.41 2.32 0.58 L53 61.96 5.38 32.67 L54 3.98 75.10 20.83 0.09 289

PAGE 290

APPENDIX L LDI SCORES FOR THE ISOL ATED FORESTED WETLANDS Table L-1. LDI scores calculated for eight differe nt grain sizes (units: meters on a side) and based on the area occupied by each la nd use type in the landscape unit. Site Code 5-m 10-m 20-m 30-m 40-m 50-m 60-m 70-m Mean SD SA1 8.67 8.69 8.69 8.72 8.73 8.73 8.73 8.73 8.71 0.02 SA2 6.00 6.04 6.03 6.06 5.99 5.96 6.12 6.03 6.03 0.05 SA3 5.88 5.95 5.89 5.91 5.88 5.94 5.59 6.05 5.89 0.13 SA4 8.08 8.11 8.07 8.06 8.10 8.06 7.92 8.03 8.05 0.06 SA5 5.08 5.13 5.12 5.06 4.99 5.13 5.02 5.04 5.07 0.05 SA6 5.68 5.69 5.65 5.67 5.67 5.66 5.67 5.61 5.66 0.02 SA7 0.81 0.86 0.87 0.89 0.81 0.70 0.71 0.71 0.80 0.08 SA8 8.99 10.37 9.38 5.50 8.72 8.44 8.06 9.56 8.63 1.45 SA9 5.27 5.28 5.26 5.26 5.31 5.28 5.29 5.29 5.28 0.02 SR1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 SR2 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 SR3 0.00 0.00 0.00 0.00 0.01 0.00 0.00 0.00 0.00 0.00 SR4 0.20 0.33 0.14 0.08 0.00 0.43 0.00 0.00 0.15 0.16 SR5 7.90 9.36 9.72 9.66 7.68 7.93 6.72 7.76 8.34 1.10 SR6 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 SR7 0.26 0.25 0.08 0.09 0.32 0.49 0.68 0.49 0.33 0.21 SR8 5.47 6.19 6.74 4.84 6.40 1.39 6.79 2.27 5.01 2.08 SR9 9.93 10.23 10.76 10.84 11.72 11.94 12.11 12.94 11.31 1.03 SU1 18.58 18.69 18.69 18.53 18.44 18.73 18.13 18.72 18.56 0.20 SU10 14.16 15.10 14.10 12.72 8.39 9.98 8.58 9.09 11.52 2.79 SU2 22.63 22.82 22.85 22.47 22.34 22.25 22.97 22.83 22.65 0.27 SU3 15.88 16.55 16.61 15.37 15.56 15.60 15.63 14.99 15.77 0.56 SU4 12.29 12.74 12.66 12.71 12.77 12.53 11.57 12.74 12.50 0.41 SU5 21.98 22.16 22.01 22.06 22.07 22.23 21.95 22.55 22.13 0.19 SU6 19.74 19.87 19.77 19.73 19.65 19.51 19.20 19.97 19.68 0.24 SU7 24.81 24.95 25.03 25.31 24.89 24.95 24.97 25.04 24.99 0.15 SU8 22.12 22.18 22.17 22.08 21.90 22.25 22.03 21.77 22.06 0.16 SU9 19.89 20.01 20.02 19.65 19.74 19.99 19.70 19.55 19.82 0.18 CA1 8.57 8.60 8.59 8.56 8.60 8.65 8.67 8.57 8.60 0.04 CA2 5.49 5.52 5.50 5.50 5.50 5.49 5.58 5.50 5.51 0.03 CA3 6.76 6.88 6.85 6.86 6.79 6.69 7.06 6.74 6.83 0.11 CA4 4.40 4.40 4.38 4.35 4.39 4.38 4.37 4.50 4.40 0.04 CA5 5.86 5.86 5.85 5.85 5.85 5.85 5.85 5.85 5.85 0.00 CA6 6.52 6.53 6.52 6.52 6.52 6.52 6.55 6.53 6.53 0.01 290

PAGE 291

Table L-1. Continued. Site Code 5-m 10-m 20-m 30-m 40-m 50-m 60-m 70-m Mean SD CA7 4.77 4.87 4.86 4.88 4.75 5.11 5.04 5.50 4.97 0.25 CA8 11.33 12.26 11.36 9.62 8.84 7.48 9.39 10.72 10.12 1.57 CA9 6.42 6.43 6.41 6.37 6.39 6.38 6.42 6.39 6.40 0.02 CR1 0.00 0.48 0.46 0.47 0.48 0.47 0.48 0.48 0.42 0.17 CR10 8.34 9.04 8.82 7.97 6.13 8.93 8.85 7.40 8.18 1.00 CR11 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 CR2 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 CR3 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 CR4 0.44 0.51 0.47 0.33 0.43 0.46 0.63 0.86 0.52 0.16 CR5 0.62 0.78 0.72 0.63 0.70 0.22 0.33 0.45 0.56 0.20 CR6 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 CR7 0.49 0.61 0.59 0.48 0.51 0.39 0.80 0.74 0.58 0.14 CR8 0.02 0.03 0.01 0.00 0.01 0.02 0.03 0.04 0.02 0.01 CR9 1.39 1.51 1.59 1.89 1.81 2.22 2.38 2.18 1.87 0.36 CU1 17.00 17.35 17.31 16.33 16.39 16.15 16.26 15.62 16.55 0.61 CU10 19.74 19.90 19.81 19.64 19.89 19.83 19.78 20.08 19.83 0.13 CU11 19.39 19.56 19.54 19.52 19.56 19.41 19.16 19.41 19.44 0.14 CU2 20.24 20.31 20.33 20.49 20.36 20.47 20.39 20.24 20.35 0.09 CU3 22.33 22.38 22.20 22.22 22.09 22.18 22.22 22.00 22.20 0.12 CU4 22.21 22.35 22.38 22.48 22.26 22.53 22.24 22.81 22.41 0.20 CU5 24.71 24.78 24.73 24.70 24.84 24.62 24.61 24.57 24.70 0.09 CU6 23.98 24.03 24.00 23.90 23.94 23.59 23.78 24.34 23.95 0.21 CU7 18.90 19.00 18.71 18.67 18.42 18.33 18.26 18.76 18.63 0.27 CU8 19.56 19.74 19.56 19.64 19.48 19.13 19.02 19.69 19.48 0.26 CU9 26.55 26.63 26.58 26.63 26.46 26.55 26.46 26.71 26.57 0.09 NA1 8.73 9.32 8.52 6.14 6.07 6.12 6.08 6.11 7.14 1.44 NA10 1.45 1.48 1.49 0.32 1.06 1.06 1.08 1.07 1.13 0.38 NA11 0.90 1.01 0.76 0.62 0.62 0.62 0.62 0.62 0.72 0.15 NA12 10.10 10.75 10.91 9.73 10.17 9.29 6.40 6.78 9.27 1.73 NA2 5.96 5.97 5.93 5.91 5.87 5.91 5.94 5.89 5.92 0.03 NA3 0.95 0.96 0.96 0.96 0.97 0.97 0.97 0.96 0.96 0.01 NA4 11.88 12.00 11.67 12.13 10.86 11.69 7.83 7.80 10.73 1.84 NA5 5.61 4.76 4.72 4.66 4.68 4.62 4.64 4.70 4.80 0.33 NA6 6.32 6.34 6.28 6.20 6.22 6.15 6.29 6.13 6.24 0.08 NA7 5.09 5.12 5.10 4.99 4.91 4.91 5.02 4.79 4.99 0.12 NA8 1.08 1.08 1.08 1.08 1.08 1.08 1.08 1.08 1.08 0.00 NA9 1.54 1.75 1.29 1.06 1.01 1.02 1.24 1.04 1.24 0.27 NR1 0.21 0.26 0.00 0.00 0.15 0.25 0.34 0.00 0.15 0.14 NR2 0.21 0.35 0.12 0.16 0.34 0.30 0.38 1.59 0.43 0.48 NR3 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 NR4 0.10 0.15 0.06 0.01 0.14 0.02 0.00 0.04 0.07 0.06 NR5 0.59 0.67 0.61 0.54 0.29 0.00 0.33 0.43 0.43 0.22 291

PAGE 292

Table L-1. Continued. Site Code 5-m 10-m 20-m 30-m 40-m 50-m 60-m 70-m Mean SD NR6 0.10 0.15 0.06 0.01 0.14 0.02 0.04 0.04 0.07 0.05 NR7 0.72 0.70 0.67 0.68 0.70 0.68 0.65 0.73 0.69 0.03 NR8 0.01 0.01 0.00 0.01 0.02 0.04 0.05 0.05 0.02 0.02 NR9 0.01 0.01 0.00 0.00 0.00 0.01 0.01 0.00 0.00 0.01 NU1 16.26 16.63 16.55 16.29 16.33 15.54 15.26 16.05 16.11 0.48 NU10 23.25 23.24 23.16 23.09 22.91 23.42 23.14 23.00 23.15 0.16 NU2 20.99 21.08 20.96 20.86 20.68 20.60 21.05 21.03 20.91 0.18 NU3 19.44 19.47 19.56 19.36 19.30 18.79 19.51 18.75 19.27 0.32 NU4 15.68 15.79 16.15 15.54 15.89 15.90 15.32 14.97 15.65 0.37 NU5 20.56 20.72 20.71 20.81 20.65 20.46 20.51 20.72 20.64 0.12 NU6 19.43 19.51 19.62 19.46 18.95 19.12 19.11 19.29 19.31 0.23 NU7 21.12 21.27 21.35 21.27 21.08 21.35 21.34 21.38 21.27 0.11 NU8 18.00 18.15 18.13 17.99 18.15 17.88 18.23 18.42 18.12 0.17 NU9 22.96 23.04 23.02 23.04 22.86 23.12 22.93 22.68 22.96 0.14 PA1 3.88 4.58 5.42 5.55 5.90 2.07 2.31 2.30 4.17 1.87 PA10 1.07 1.08 1.08 1.08 1.08 1.08 1.08 1.08 1.08 0.00 PA2 6.48 6.52 6.52 6.56 6.45 6.57 6.44 6.68 6.53 0.08 PA3 8.91 8.96 8.96 8.93 8.96 8.85 8.85 8.80 8.90 0.06 PA4 11.24 11.84 11.78 10.86 9.84 9.02 9.11 9.27 10.37 1.20 PA5 4.82 4.88 4.87 4.88 4.90 4.87 4.70 4.87 4.85 0.06 PA6 8.20 8.47 8.55 8.71 7.80 7.83 6.92 3.92 7.55 1.57 PA7 5.30 5.33 5.30 5.26 5.40 5.29 5.32 5.15 5.29 0.07 PA8 1.24 1.33 1.18 1.03 1.01 1.03 1.02 1.02 1.11 0.12 PA9 5.93 6.00 5.78 5.83 5.75 5.46 5.45 5.66 5.73 0.20 PR1 0.02 0.02 0.01 0.00 0.02 0.02 0.04 0.05 0.02 0.02 PR2 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 PR3 0.50 0.63 0.58 0.31 0.14 0.22 0.59 0.46 0.43 0.18 PR4 9.12 9.37 9.09 9.44 9.40 9.73 9.65 8.24 9.25 0.47 PR5 0.03 0.04 0.01 0.01 0.02 0.03 0.05 0.07 0.03 0.02 PR6 1.10 1.21 1.19 1.21 1.50 1.04 0.82 1.56 1.20 0.24 PR7 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 PR8 0.01 0.02 0.00 0.00 0.01 0.02 0.01 0.04 0.01 0.01 PU1 21.33 21.61 21.56 21.61 21.44 21.34 21.56 20.83 21.41 0.26 PU10 21.33 21.46 21.30 21.18 21.12 21.15 20.94 21.15 21.20 0.16 PU2 19.54 19.67 19.67 19.58 19.47 19.25 19.47 19.29 19.49 0.16 PU3 21.90 21.96 22.01 21.87 21.74 22.08 21.79 21.91 21.91 0.11 PU4 16.53 16.88 16.71 16.40 15.94 15.65 15.34 15.86 16.16 0.55 PU5 19.59 19.73 19.62 19.72 19.70 19.16 19.48 18.70 19.46 0.36 PU6 19.55 19.79 19.49 19.64 19.98 19.24 19.22 19.72 19.58 0.26 PU7 17.64 17.75 17.79 17.70 17.80 17.87 17.57 17.97 17.76 0.13 PU8 22.61 22.77 22.74 22.71 23.14 22.81 23.66 23.80 23.03 0.46 PU9 16.15 16.45 16.06 16.26 16.08 16.20 16.74 16.49 16.30 0.24 292

PAGE 293

Table L-2. LDI scores calculated for eight differe nt spatial resolutions (units in meters) and assuming that the effect of development in tensity on the landscape decreases linearly with distance. Site code 5-m 10-m 20-m 30-m 40-m 50-m 60-m 70-m Mean SD SA1 5.63 5.65 5.78 5.98 6.17 6.02 5.19 4.99 5.68 0.41 SA2 3.31 3.36 3.31 3.30 3.46 3.63 2.73 3.14 3.28 0.26 SA3 3.23 3.14 3.29 3.48 3.61 3.32 2.57 2.64 3.16 0.37 SA4 5.47 5.51 5.49 5.49 5.10 5.29 4.70 4.83 5.24 0.32 SA5 2.54 2.69 2.69 2.56 2.98 2.44 1.88 1.74 2.44 0.42 SA6 3.37 3.42 3.40 3.42 3.62 2.95 2.84 3.65 3.33 0.29 SA7 0.31 0.30 0.34 0.30 0.35 0.27 0.24 0.30 0.30 0.03 SA8 6.36 7.68 6.99 1.77 5.87 6.72 0.15 6.72 5.28 2.75 SA9 2.97 3.01 3.02 3.09 3.24 5.28 2.46 2.42 3.19 0.90 SR1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 SR2 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 SR3 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 SR4 0.02 0.05 0.02 0.00 0.00 0.00 0.00 0.00 0.01 0.02 SR5 4.60 5.64 6.24 6.70 3.19 2.37 0.00 4.24 4.12 2.22 SR6 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 SR7 0.11 0.09 0.02 0.01 0.03 0.06 0.12 0.17 0.08 0.05 SR8 0.96 1.09 1.87 0.67 2.67 0.34 0.10 0.66 1.04 0.85 SR9 4.66 4.76 5.51 5.93 7.07 6.29 5.48 6.43 5.76 0.83 SU1 14.92 15.09 15.21 14.99 14.72 14.98 13.10 15.77 14.85 0.77 SU10 10.78 11.75 10.75 8.64 3.36 5.24 5.31 3.29 7.39 3.49 SU2 18.35 18.48 18.85 18.54 19.18 18.48 17.39 17.32 18.32 0.65 SU3 10.10 11.00 10.73 8.95 10.06 9.21 8.05 7.09 9.40 1.35 SU4 7.29 7.97 7.84 8.01 8.54 7.22 4.98 8.71 7.57 1.17 SU5 18.32 18.45 18.27 18.52 18.68 18.49 18.89 17.45 18.38 0.42 SU6 15.60 15.48 15.51 15.88 15.51 15.17 13.42 16.62 15.40 0.91 SU7 20.40 20.65 20.75 20.77 20.85 22.03 18.76 19.50 20.46 0.97 SU8 18.49 18.57 18.51 18.99 18.63 18.57 19.01 18.10 18.61 0.29 SU9 16.40 16.46 16.77 16.39 17.14 16.34 16.81 16.87 16.65 0.29 CA1 5.58 5.61 5.64 5.38 5.43 6.02 5.99 4.67 5.54 0.42 CA2 3.08 3.11 3.08 3.38 3.25 2.78 3.35 2.44 3.06 0.31 CA3 3.62 3.67 3.81 3.88 4.25 3.22 3.21 3.97 3.70 0.36 CA4 2.43 2.49 2.52 2.32 2.41 2.43 2.65 2.84 2.51 0.16 CA5 3.49 3.51 3.48 3.80 3.79 3.69 2.96 3.76 3.56 0.28 CA6 4.02 4.00 4.17 4.10 4.39 4.09 3.66 4.53 4.12 0.26 CA7 2.97 3.02 2.97 3.16 2.92 2.83 2.76 3.35 3.00 0.18 CA8 8.58 9.47 8.56 6.47 6.75 0.00 3.31 4.16 5.91 3.22 CA9 4.06 4.11 4.14 4.08 4.44 4.40 3.79 4.11 4.14 0.21 CR1 0.00 0.20 0.19 0.20 0.20 0.22 0.15 0.22 0.17 0.07 CR10 5.92 6.52 5.93 5.89 1.18 1.96 0.01 3.98 3.92 2.54 CR11 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 293

PAGE 294

Table L-2. Continued. Site Code 5-m 10-m 20-m 30-m 40-m 50-m 60-m 70-m Mean SD CR2 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 CR3 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 CR4 0.19 0.22 0.19 0.05 0.03 0.06 0.00 0.31 0.13 0.11 CR5 0.20 0.25 0.25 0.19 0.15 0.00 0.00 0.16 0.15 0.10 CR6 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 CR7 0.18 0.22 0.22 0.19 0.00 0.00 0.00 0.26 0.14 0.11 CR8 0.01 0.14 0.05 0.00 0.00 0.01 0.01 0.02 0.03 0.05 CR9 0.37 0.41 0.40 0.54 0.21 0.30 0.37 0.00 0.33 0.16 CU1 12.46 13.09 12.82 11.28 11.69 9.01 10.88 10.65 11.48 1.34 CU10 14.99 15.28 14.88 14.91 15.59 15.78 13.13 15.54 15.01 0.83 CU11 14.47 14.75 14.91 14.27 15.49 15.30 12.61 13.65 14.43 0.94 CU2 15.61 15.51 15.72 16.52 16.08 16.55 14.51 15.88 15.80 0.65 CU3 19.04 19.21 19.05 19.15 19.20 18.71 18.21 19.60 19.02 0.41 CU4 18.07 18.41 18.46 19.27 18.68 17.68 17.14 18.83 18.32 0.68 CU5 20.83 20.97 20.99 20.69 20.59 20.65 19.77 21.22 20.71 0.44 CU6 20.60 20.67 20.95 20.78 20.67 19.84 20.10 21.62 20.65 0.54 CU7 15.69 15.63 15.59 15.52 15.48 15.34 14.20 16.28 15.47 0.58 CU8 16.51 16.85 16.83 16.48 16.17 16.55 15.53 15.42 16.29 0.55 CU9 22.86 23.02 22.93 23.61 23.15 23.66 21.52 23.75 23.06 0.71 NA1 4.97 5.43 5.00 3.54 3.36 3.96 2.81 3.70 4.10 0.93 NA10 0.70 0.74 0.75 0.53 0.49 0.45 0.40 0.56 0.58 0.14 NA11 0.42 0.49 0.33 0.29 0.26 0.24 0.23 0.28 0.32 0.09 NA12 7.32 7.86 8.37 6.56 7.58 6.27 3.19 3.54 6.34 1.96 NA2 3.51 3.61 3.57 3.50 3.84 3.51 3.89 2.99 3.55 0.27 NA3 0.43 0.44 0.45 0.43 0.48 0.41 0.34 0.41 0.42 0.04 NA4 8.87 8.95 8.48 9.14 7.83 8.64 3.97 4.81 7.59 2.02 NA5 3.27 2.68 2.64 2.68 2.74 2.34 2.05 2.31 2.59 0.37 NA6 3.97 3.97 3.93 4.13 4.17 4.28 3.60 4.09 4.02 0.21 NA7 3.19 3.29 3.32 3.34 3.31 3.07 2.83 3.05 3.18 0.18 NA8 0.48 0.48 0.48 0.48 0.53 0.47 0.37 0.45 0.47 0.05 NA9 0.77 0.88 0.59 0.46 0.52 0.46 0.36 0.51 0.57 0.17 NR1 0.11 0.14 0.00 0.00 0.00 0.25 0.34 0.00 0.11 0.13 NR2 0.05 0.09 0.03 0.03 0.06 0.11 0.01 0.49 0.11 0.16 NR3 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 NR4 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 NR5 0.22 0.24 0.27 0.22 0.09 0.00 0.00 0.15 0.15 0.11 NR6 0.04 0.05 0.02 0.00 0.00 0.00 0.01 0.02 0.02 0.02 NR7 0.37 0.35 0.33 0.33 0.32 0.29 0.24 0.42 0.33 0.05 NR8 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 NR9 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 NU1 12.47 12.91 12.61 13.02 12.61 11.55 11.40 10.94 12.19 0.78 NU10 20.40 20.51 20.65 20.64 20.30 21.56 21.22 20.70 20.75 0.43 294

PAGE 295

Table L-2. Continued. Site Code 5-m 10-m 20-m 30-m 40-m 50-m 60-m 70-m Mean SD NU2 17.08 17.22 17.06 17.06 16.46 15.90 15.82 17.00 16.70 0.56 NU3 16.60 16.70 16.80 16.83 16.56 16.42 16.59 16.18 16.59 0.21 NU4 11.54 11.65 11.94 11.90 10.99 11.71 10.58 10.24 11.32 0.64 NU5 16.68 16.75 16.79 16.98 17.29 17.12 15.72 17.61 16.87 0.56 NU6 15.85 15.86 16.29 16.00 14.41 14.62 14.36 15.42 15.35 0.77 NU7 16.59 16.82 16.89 16.87 15.86 15.62 14.95 17.76 16.42 0.89 NU8 13.67 13.83 13.90 13.92 13.83 13.05 12.63 15.26 13.76 0.77 NU9 19.30 19.48 19.61 19.27 19.91 19.96 18.39 19.40 19.41 0.49 PA1 1.24 1.44 1.87 1.57 0.90 0.76 0.78 0.74 1.16 0.43 PA10 0.45 0.48 0.47 0.47 0.50 0.41 0.52 0.52 0.48 0.04 PA2 3.71 3.90 3.90 3.95 3.64 3.44 4.08 3.29 3.74 0.27 PA3 6.10 6.13 6.32 6.40 5.82 6.50 5.56 6.50 6.17 0.34 PA4 8.61 9.31 9.39 8.89 7.55 6.68 5.99 6.95 7.92 1.30 PA5 2.71 2.82 2.88 2.65 2.97 2.72 2.28 2.45 2.69 0.23 PA6 4.21 4.46 4.23 4.78 4.05 3.24 2.13 2.56 3.71 0.95 PA7 3.21 3.24 3.24 3.21 3.06 2.85 2.82 3.34 3.12 0.19 PA8 0.57 0.63 0.53 0.47 0.47 0.52 0.36 0.46 0.50 0.08 PA9 4.34 4.41 4.24 4.43 4.38 4.37 3.68 4.39 4.28 0.25 PR1 0.01 0.01 0.00 0.00 0.01 0.01 0.01 0.03 0.01 0.01 PR2 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 PR3 0.22 0.27 0.29 0.11 0.00 0.00 0.00 0.00 0.11 0.13 PR4 4.46 4.53 4.29 5.32 2.60 3.85 0.19 0.40 3.21 1.95 PR5 0.01 0.02 0.00 0.00 0.00 0.01 0.02 0.04 0.01 0.01 PR6 0.57 0.64 0.66 0.64 0.86 0.39 0.16 0.79 0.59 0.22 PR7 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 PR8 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.00 PU1 17.49 17.84 17.75 18.52 17.77 17.75 18.42 17.38 17.86 0.41 PU10 17.83 18.01 17.77 17.93 17.77 16.58 16.05 17.76 17.46 0.73 PU2 16.14 16.34 16.53 16.21 16.16 15.98 15.25 14.84 15.93 0.58 PU3 17.99 18.12 18.43 18.04 18.16 18.33 18.56 18.81 18.31 0.28 PU4 13.35 13.70 13.79 13.38 12.78 11.62 10.54 13.17 12.79 1.14 PU5 15.66 15.89 15.79 16.14 15.86 16.11 16.19 14.36 15.75 0.59 PU6 16.14 16.35 16.46 15.93 16.81 16.26 15.53 16.35 16.23 0.38 PU7 13.00 13.19 13.24 13.27 13.98 13.78 10.72 14.24 13.18 1.08 PU8 17.41 17.90 17.55 18.18 18.77 19.53 17.43 17.61 18.05 0.75 PU9 11.58 11.96 11.67 11.52 10.77 10.99 10.08 11.75 11.29 0.63 295

PAGE 296

Table L-3. LDI scores calculated for eight differe nt spatial resolutions (units in meters) and assuming that the effect of development intensity on the landscape decreases in inverse-square with distance. Site code 5-m 10-m 20-m 30-m 40-m 50-m 60-m 70-m Mean SD SA1 4.05 4.13 4.28 4.55 4.71 4.69 3.99 4.18 4.32 0.29 SA2 2.06 2.11 2.06 2.19 2.21 2.39 1.79 2.09 2.11 0.17 SA3 1.99 1.94 2.11 2.34 2.44 2.07 1.70 1.97 2.07 0.23 SA4 4.03 4.10 4.12 4.23 3.88 4.09 3.62 4.07 4.02 0.19 SA5 1.49 1.61 1.61 1.55 1.90 1.53 1.12 1.19 1.50 0.25 SA6 2.26 2.33 2.36 2.44 2.58 2.04 2.03 2.65 2.34 0.23 SA7 0.17 0.16 0.19 0.18 0.19 0.16 0.15 0.18 0.17 0.01 SA8 4.87 6.08 5.93 0.68 4.93 6.05 0.05 6.72 4.41 2.58 SA9 1.92 1.98 2.02 2.13 2.21 1.81 1.67 1.88 1.95 0.17 SR1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 SR2 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 SR3 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 SR4 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 SR5 2.59 3.29 3.96 4.54 1.62 0.82 0.00 1.91 2.34 1.55 SR6 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 SR7 0.05 0.04 0.00 0.00 0.01 0.02 0.04 0.06 0.03 0.02 SR8 0.13 0.15 0.36 0.12 0.93 0.12 0.04 0.23 0.26 0.29 SR9 1.91 1.98 2.50 2.94 3.95 3.65 2.66 3.29 2.86 0.74 SU1 12.44 12.64 12.93 12.83 12.48 12.78 11.04 13.58 12.59 0.72 SU10 8.61 9.56 8.66 6.64 2.64 3.24 3.44 2.57 5.67 3.01 SU2 15.65 15.85 16.26 16.25 16.82 15.77 15.04 15.84 15.93 0.52 SU3 6.14 7.15 6.87 5.30 6.09 6.05 4.48 3.78 5.73 1.15 SU4 4.26 4.89 4.81 5.12 5.69 4.61 2.93 5.81 4.76 0.91 SU5 15.66 16.04 15.96 16.29 16.46 16.11 16.61 16.06 16.15 0.30 SU6 12.64 12.59 12.62 13.39 12.56 12.06 10.39 14.08 12.54 1.07 SU7 17.37 17.71 17.87 18.18 17.93 19.76 15.87 16.79 17.69 1.13 SU8 16.15 16.28 16.24 16.91 16.46 16.48 16.97 16.17 16.46 0.32 SU9 14.22 14.35 14.75 14.56 15.40 14.35 14.93 15.01 14.70 0.40 CA1 3.94 4.00 4.08 3.90 4.02 4.46 4.43 3.77 4.08 0.25 CA2 1.97 2.01 2.01 2.28 2.14 1.80 2.15 1.86 2.03 0.16 CA3 2.24 2.30 2.41 2.54 2.85 2.03 2.03 2.58 2.37 0.28 CA4 1.55 1.62 1.67 1.53 1.81 1.66 1.78 2.08 1.71 0.18 CA5 2.35 2.39 2.39 2.71 2.71 2.58 2.13 2.79 2.51 0.23 CA6 2.80 2.82 3.01 3.03 3.25 3.06 2.75 3.56 3.03 0.27 CA7 2.14 2.18 2.17 2.40 2.11 2.13 2.15 2.47 2.22 0.14 CA8 7.15 8.06 7.39 5.23 6.67 0.00 3.31 4.16 5.25 2.69 CA9 2.87 2.93 3.03 3.03 3.37 3.32 2.92 3.14 3.08 0.19 CR1 0.00 0.11 0.10 0.12 0.11 0.13 0.09 0.14 0.10 0.04 CR10 4.48 5.06 4.61 4.54 0.27 0.59 0.01 1.77 2.66 2.21 CR11 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 296

PAGE 297

Table L-3. Continued. Site Code 5-m 10-m 20-m 30-m 40-m 50-m 60-m 70-m Mean SD CR2 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 CR3 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 CR4 0.10 0.12 0.11 0.01 0.01 0.01 0.00 0.10 0.06 0.05 CR5 0.07 0.09 0.09 0.06 0.05 0.00 0.00 0.05 0.05 0.04 CR6 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 CR7 0.08 0.10 0.10 0.11 0.00 0.00 0.00 0.09 0.06 0.05 CR8 0.09 0.01 0.00 0.00 0.00 0.01 0.01 0.02 0.02 0.03 CR9 0.11 0.12 0.12 0.18 0.04 0.08 0.13 0.00 0.10 0.05 CU1 9.66 10.34 9.95 8.01 8.17 5.64 9.01 7.41 8.53 1.55 CU10 11.71 12.08 11.50 11.85 12.50 12.74 9.96 12.98 11.92 0.94 CU11 10.83 11.19 11.53 10.67 12.30 12.12 9.09 9.74 10.93 1.11 CU2 12.56 12.51 12.82 14.00 13.24 13.76 12.03 12.98 12.99 0.66 CU3 16.86 17.09 16.97 17.22 17.21 16.79 16.38 18.07 17.07 0.49 CU4 15.63 16.01 16.16 17.10 16.53 15.65 15.37 16.84 16.16 0.62 CU5 18.18 18.40 18.54 18.24 18.15 18.29 17.64 19.07 18.31 0.40 CU6 18.45 18.62 19.04 18.93 18.91 17.88 18.35 20.11 18.79 0.66 CU7 13.48 13.45 13.50 13.55 13.32 13.25 12.28 14.70 13.44 0.66 CU8 14.68 15.09 15.24 14.86 14.66 14.98 14.53 13.46 14.69 0.55 CU9 20.36 20.61 20.57 21.47 20.88 21.56 19.27 21.90 20.83 0.83 NA1 2.92 3.20 3.06 2.37 2.27 2.74 1.92 2.59 2.63 0.43 NA10 0.43 0.47 0.48 0.32 0.32 0.28 0.26 0.37 0.37 0.08 NA11 0.24 0.29 0.20 0.19 0.17 0.15 0.15 0.18 0.20 0.05 NA12 5.78 6.30 6.86 4.91 6.10 4.47 2.24 2.94 4.95 1.65 NA2 2.36 2.46 2.46 2.46 2.78 2.54 2.77 2.40 2.53 0.16 NA3 0.25 0.27 0.27 0.27 0.30 0.26 0.22 0.25 0.26 0.02 NA4 6.95 7.00 6.48 7.38 5.77 6.67 2.71 3.38 5.79 1.77 NA5 2.18 1.76 1.76 1.84 1.83 1.59 1.41 1.53 1.74 0.24 NA6 2.75 2.77 2.78 2.97 3.00 3.20 2.66 3.12 2.91 0.19 NA7 2.25 2.37 2.42 2.51 2.46 2.26 2.07 2.31 2.33 0.14 NA8 0.27 0.28 0.29 0.29 0.32 0.29 0.23 0.27 0.28 0.03 NA9 0.48 0.55 0.35 0.29 0.33 0.27 0.23 0.33 0.35 0.11 NR1 0.07 0.09 0.00 0.00 0.00 0.25 0.34 0.00 0.09 0.13 NR2 0.02 0.03 0.01 0.01 0.01 0.05 0.00 0.17 0.04 0.06 NR3 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 NR4 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 NR5 0.09 0.10 0.13 0.09 0.03 0.00 0.00 0.05 0.06 0.05 NR6 0.02 0.02 0.01 0.00 0.00 0.00 0.01 0.02 0.01 0.01 NR7 0.24 0.22 0.22 0.22 0.21 0.19 0.16 0.29 0.22 0.04 NR8 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 NR9 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 NU1 10.00 10.48 10.23 10.71 10.10 8.98 8.69 9.21 9.80 0.74 NU10 18.43 18.65 18.98 19.06 18.70 20.21 19.93 19.09 19.13 0.63 297

PAGE 298

Table L-3. Continued. Site Code 5-m 10-m 20-m 30-m 40-m 50-m 60-m 70-m Mean SD NU2 14.24 14.44 14.30 14.49 13.88 13.12 13.01 14.47 13.99 0.60 NU3 14.65 14.85 14.96 15.20 14.55 14.82 15.06 14.32 14.80 0.28 NU4 8.92 9.03 9.33 9.49 8.67 9.67 8.67 7.99 8.97 0.54 NU5 14.45 14.58 14.59 14.96 15.22 15.05 14.07 15.81 14.84 0.54 NU6 13.58 13.64 14.14 13.98 11.98 12.26 12.17 14.29 13.25 0.96 NU7 13.59 13.90 14.00 14.28 13.12 12.76 12.34 15.49 13.68 0.98 NU8 10.58 10.77 10.97 11.21 11.04 10.19 9.69 13.06 10.94 0.99 NU9 16.87 17.12 17.39 17.07 17.83 18.03 16.38 17.23 17.24 0.52 PA1 0.60 0.66 0.77 0.68 0.55 0.44 0.52 0.54 0.59 0.10 PA10 0.26 0.28 0.27 0.28 0.29 0.24 0.30 0.33 0.28 0.03 PA2 2.44 2.63 2.67 2.75 2.56 2.34 2.76 2.59 2.59 0.15 PA3 4.64 4.73 4.96 5.12 4.62 5.13 4.51 5.38 4.89 0.31 PA4 7.12 7.84 7.96 7.85 6.29 5.22 4.74 5.55 6.57 1.29 PA5 1.81 1.90 2.01 1.84 2.05 1.91 1.65 1.71 1.86 0.14 PA6 2.26 2.42 2.27 2.72 2.23 1.89 1.53 1.90 2.15 0.37 PA7 2.18 2.23 2.27 2.29 2.19 2.03 2.08 2.47 2.22 0.14 PA8 0.34 0.39 0.32 0.29 0.28 0.32 0.23 0.29 0.31 0.05 PA9 3.35 3.43 3.31 3.60 3.52 3.61 2.87 3.59 3.41 0.25 PR1 0.01 0.01 0.00 0.00 0.01 0.01 0.01 0.02 0.01 0.01 PR2 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 PR3 0.11 0.14 0.16 0.05 0.00 0.00 0.00 0.00 0.06 0.07 PR4 1.92 2.00 1.75 2.72 0.74 1.38 0.09 0.23 1.35 0.92 PR5 0.01 0.01 0.00 0.00 0.00 0.01 0.01 0.02 0.01 0.01 PR6 0.37 0.42 0.46 0.43 0.65 0.22 0.10 0.56 0.40 0.18 PR7 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 PR8 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 PU1 15.18 15.63 15.63 16.46 15.50 15.48 16.29 15.45 15.70 0.44 PU10 15.56 15.79 15.58 15.97 15.51 14.18 14.04 15.92 15.32 0.77 PU2 14.00 14.26 14.61 14.37 14.34 14.26 13.56 13.67 14.13 0.36 PU3 15.34 15.55 16.04 15.64 15.93 16.10 16.31 16.80 15.96 0.46 PU4 11.30 11.68 11.93 11.59 10.78 9.57 8.36 11.29 10.81 1.23 PU5 13.08 13.39 13.43 13.78 13.43 13.95 13.88 11.63 13.32 0.74 PU6 13.99 14.20 14.46 14.02 14.84 14.64 14.20 14.84 14.40 0.35 PU7 9.93 10.21 10.34 10.50 11.30 10.89 7.99 11.66 10.35 1.11 PU8 14.00 14.60 14.19 14.95 15.59 16.97 14.64 14.44 14.92 0.96 PU9 8.46 8.86 8.78 8.66 7.92 8.28 7.21 9.13 8.41 0.61 298

PAGE 299

APPENDIX M LDI SCORES FOR STREAMS Table M-1. LDI scores calculated for six different grai n sizes (units: meters on a side) and based on the area occupied by each land use type in the drainage basin unit. Site code 20-m 50-m 80-m 110-m 140-m 170-m Mean SD S1 0.76 0.79 0.92 0.93 0.80 0.99 0.86 0.09 S2 11.52 12.09 12.19 12.09 11.88 11.70 11.91 0.26 S3 9.80 10.23 10.31 10.49 10.48 10.02 10.22 0.27 S4 2.01 2.14 1.93 1.67 1.62 1.50 1.81 0.25 S5 8.79 9.00 9.09 8.80 8.71 8.67 8.84 0.17 S6 10.13 10.42 10.27 10.40 10.11 10.92 10.37 0.30 S7 9.91 10.20 10.26 10.20 10.19 10.40 10.19 0.16 S8 12.90 13.10 13.30 13.07 12.92 13.16 13.07 0.15 S9 7.14 7.43 7.44 7.27 7.04 7.00 7.22 0.19 S10 12.21 12.35 12.40 12.28 12.31 12.23 12.30 0.07 S11 14.34 14.67 14.77 14.46 14.44 14.33 14.50 0.18 S12 16.74 16.83 16.92 16.94 16.85 16.86 16.85 0.07 S13 16.07 16.26 16.30 16.22 16.05 16.38 16.21 0.13 S14 10.93 11.41 11.50 11.21 11.09 11.27 11.24 0.21 S15 11.50 11.90 11.94 12.02 11.86 12.03 11.87 0.20 S16 14.13 14.33 14.33 14.17 14.08 14.05 14.18 0.12 S17 13.38 13.53 13.53 13.44 13.45 13.56 13.48 0.07 S18 12.59 12.83 12.81 12.61 12.46 12.37 12.61 0.18 S19 11.92 12.35 12.37 12.48 12.35 12.46 12.32 0.21 S20 14.30 14.67 14.77 14.53 14.53 14.41 14.53 0.17 S21 6.77 7.17 7.01 6.53 6.02 5.83 6.55 0.54 S22 1.08 1.15 1.12 1.15 1.10 1.09 1.12 0.03 S23 14.29 14.59 14.56 14.25 14.12 14.35 14.36 0.18 S24 2.13 2.46 3.20 1.53 1.19 0.96 1.91 0.85 S25 3.77 3.88 3.86 3.85 4.03 4.07 3.91 0.12 S26 19.00 19.47 19.78 19.78 19.54 19.75 19.55 0.30 S27 13.93 14.26 14.30 14.05 13.89 14.18 14.10 0.17 S28 15.54 15.71 15.72 15.46 15.34 15.27 15.51 0.19 S29 25.03 25.06 24.99 25.08 25.09 25.14 25.06 0.05 S30 11.00 11.26 11.31 11.18 11.05 11.25 11.17 0.12 S31 8.40 8.75 9.12 8.36 8.16 8.33 8.52 0.35 299

PAGE 300

Table M-1. Continued. Site code 20-m 50-m 80-m 110-m 140-m 170-m Mean SD S32 16.69 16.93 17.03 17.06 17.08 17.15 16.99 0.16 S33 16.64 16.80 16.85 16.74 16.71 16.71 16.74 0.07 S34 15.97 16.18 16.23 16.13 16.13 16.18 16.14 0.09 S35 14.70 14.94 14.91 14.89 14.87 14.94 14.87 0.09 S36 4.24 4.35 4.37 4.40 4.39 4.42 4.36 0.06 S37 4.92 5.28 5.35 5.36 5.38 5.50 5.30 0.20 S38 20.01 20.03 20.02 20.01 20.00 19.98 20.01 0.02 S39 6.68 6.84 6.80 6.61 6.64 6.66 6.70 0.09 S40 4.66 4.87 4.88 4.68 4.56 4.45 4.68 0.17 S41 12.48 12.83 12.96 12.87 12.86 12.89 12.82 0.17 S42 10.17 10.38 10.47 10.32 10.30 10.30 10.32 0.10 S43 12.05 12.19 12.19 12.17 11.74 12.08 12.07 0.17 S44 9.56 9.86 10.06 10.13 10.21 9.98 9.96 0.23 S45 16.76 16.87 16.95 16.94 16.91 16.97 16.90 0.08 S46 14.31 14.52 14.58 14.47 14.20 14.17 14.37 0.17 S47 10.44 10.72 10.80 10.71 10.68 10.64 10.66 0.12 S48 10.53 10.84 10.94 10.87 10.85 10.86 10.81 0.14 S48 8.24 8.44 8.22 8.95 7.84 8.46 8.36 0.37 S50 16.33 16.44 16.47 16.46 16.35 16.33 16.40 0.07 S51 3.06 3.19 3.18 3.17 2.99 2.94 3.09 0.11 S52 12.90 13.07 13.08 13.01 12.87 12.83 12.96 0.11 S53 18.09 18.35 18.32 18.34 18.30 18.24 18.27 0.10 S54 12.88 13.13 13.16 13.28 13.44 13.37 13.21 0.20 S55 4.57 4.73 4.78 4.40 4.71 4.47 4.61 0.15 S56 5.41 5.85 5.70 3.32 2.28 2.10 4.11 1.75 S57 6.55 7.01 7.06 7.08 6.04 5.33 6.51 0.71 S58 9.43 9.67 9.63 9.37 9.17 8.77 9.34 0.33 S59 6.56 6.92 6.26 5.52 5.34 5.06 5.94 0.74 S60 11.30 11.66 11.69 11.27 11.07 10.28 11.21 0.51 S61 20.28 20.43 20.60 20.40 20.22 20.14 20.35 0.16 S62 0.90 1.12 1.35 0.78 0.78 0.80 0.95 0.23 S63 8.89 9.22 9.01 8.27 7.91 7.30 8.43 0.74 S64 6.93 7.16 7.09 6.81 5.82 2.95 6.12 1.63 S65 6.98 7.29 7.53 6.97 6.59 6.29 6.94 0.45 S66 16.68 17.04 17.10 17.22 17.27 17.05 17.06 0.21 S67 17.83 18.15 18.15 18.13 18.08 17.85 18.03 0.15 S68 6.58 7.03 7.22 6.05 5.31 5.30 6.25 0.83 S69 5.59 6.05 6.26 5.90 5.83 5.86 5.91 0.22 300

PAGE 301

Table M-2. LDI scores calculated for six different grain sizes (units: meters on a side) and assuming that the effect of development in tensity on the landscape decreases linearly with distance. Site code 20-m 50-m 80-m 110-m 140-m 170-m Mean SD S1 0.41 0.43 0.54 0.53 0.43 0.62 0.49 0.08 S2 9.61 10.11 10.17 10.00 9.73 9.39 9.84 0.31 S3 8.84 9.30 9.36 9.53 9.56 9.02 9.27 0.28 S4 0.71 0.74 0.69 0.62 0.57 0.54 0.65 0.08 S5 6.54 6.69 6.78 6.68 6.56 6.58 6.64 0.09 S6 6.31 6.56 6.49 6.49 6.22 6.33 6.40 0.13 S7 6.98 7.22 7.24 7.07 6.64 7.12 7.05 0.22 S8 9.52 9.66 9.87 9.68 9.59 9.73 9.67 0.12 S9 4.10 4.30 4.28 4.13 3.96 3.90 4.11 0.16 S10 10.46 10.58 10.63 10.51 10.55 10.48 10.54 0.06 S11 12.40 12.66 12.84 12.49 12.72 12.44 12.59 0.18 S12 14.87 14.94 14.98 15.02 14.97 15.00 14.96 0.05 S13 13.16 13.28 13.27 13.26 13.05 13.42 13.24 0.12 S14 7.99 8.53 8.47 8.45 8.22 8.82 8.41 0.28 S15 8.43 8.73 8.81 8.90 8.79 8.89 8.76 0.17 S16 11.43 11.58 11.59 11.44 11.39 11.36 11.46 0.10 S17 9.87 10.01 9.97 9.88 9.89 9.96 9.93 0.06 S18 9.45 9.68 9.63 9.46 9.29 9.24 9.46 0.18 S19 9.59 9.99 10.03 10.12 10.02 10.25 10.00 0.22 S20 12.67 12.99 13.16 12.93 13.14 12.92 12.97 0.18 S21 5.43 5.77 5.58 5.30 4.91 4.69 5.28 0.41 S22 0.49 0.52 0.51 0.51 0.50 0.50 0.50 0.01 S23 12.43 12.69 12.65 12.43 12.34 12.52 12.51 0.14 S24 1.06 1.19 1.55 0.82 0.67 0.54 0.97 0.37 S25 3.33 3.43 3.42 3.40 3.61 3.67 3.48 0.13 S26 14.03 14.45 14.64 15.26 15.55 15.28 14.87 0.58 S27 12.15 12.45 12.54 12.32 12.16 12.51 12.36 0.17 S28 10.83 11.00 11.04 10.68 10.69 10.47 10.79 0.22 S29 21.24 21.20 21.28 20.97 21.27 21.31 21.21 0.12 S30 6.37 6.65 6.63 6.62 6.58 6.73 6.60 0.12 S31 6.43 6.68 7.01 6.32 6.15 6.31 6.48 0.31 S32 13.69 13.94 14.02 14.02 14.08 14.08 13.97 0.15 S33 14.97 15.11 15.16 15.01 15.00 15.00 15.04 0.08 S34 12.68 12.86 12.91 12.81 12.79 12.85 12.82 0.08 S35 12.15 12.36 12.30 12.31 12.24 12.30 12.28 0.07 S36 2.43 2.47 2.50 2.50 2.50 2.53 2.49 0.04 S37 2.97 3.13 3.09 3.05 3.02 3.06 3.05 0.06 301

PAGE 302

Table M-2. Continued. Site code 20-m 50-m 80-m 110-m 140-m 170-m Mean SD S38 17.99 18.01 18.00 17.99 17.98 17.96 17.99 0.02 S39 4.80 4.97 4.93 4.69 4.67 4.69 4.79 0.13 S40 2.93 3.04 3.04 2.88 2.81 2.76 2.91 0.12 S41 10.36 10.73 10.83 10.72 10.71 10.59 10.66 0.16 S42 7.14 7.31 7.40 7.25 7.21 7.22 7.25 0.09 S43 7.77 7.86 7.85 7.90 7.56 7.83 7.79 0.12 S44 6.54 6.80 7.00 7.10 7.04 6.88 6.89 0.20 S45 13.45 13.58 13.68 13.63 13.61 13.58 13.59 0.08 S46 12.57 12.78 12.87 12.80 12.53 12.53 12.68 0.16 S47 8.60 8.87 8.98 8.89 8.88 8.84 8.84 0.13 S48 8.86 9.15 9.28 9.19 9.19 9.23 9.15 0.15 S48 5.87 5.97 5.85 6.32 6.08 6.83 6.15 0.37 S50 13.96 14.08 14.09 14.07 13.96 13.92 14.01 0.07 S51 1.29 1.37 1.31 1.28 1.21 1.21 1.28 0.06 S52 10.82 11.02 11.04 10.94 10.79 10.74 10.89 0.13 S53 16.53 16.81 16.79 16.77 16.80 16.75 16.74 0.10 S54 11.75 12.00 12.03 12.17 12.29 12.25 12.08 0.20 S55 2.36 2.47 2.49 2.27 2.43 2.31 2.39 0.09 S56 3.37 3.71 3.21 1.96 1.64 1.46 2.56 0.98 S57 3.58 3.96 3.92 3.94 3.22 3.02 3.61 0.40 S58 5.18 5.37 5.27 5.06 4.81 4.30 5.00 0.39 S59 4.56 4.85 4.37 3.90 3.75 3.65 4.18 0.48 S60 8.24 8.49 8.51 8.06 7.96 7.87 8.19 0.27 S61 17.53 17.66 17.85 17.70 17.57 17.32 17.61 0.18 S62 0.52 0.73 0.95 0.39 0.38 0.40 0.56 0.23 S63 6.53 6.82 6.63 5.87 5.44 4.88 6.03 0.76 S64 4.24 4.43 4.42 4.03 3.31 1.45 3.65 1.15 S65 5.55 5.84 6.10 5.41 5.08 4.80 5.46 0.48 S66 13.42 13.82 13.81 13.84 13.98 13.67 13.76 0.19 S67 14.06 14.43 14.38 14.46 14.28 14.27 14.31 0.15 S68 4.10 4.44 4.63 3.09 2.57 2.41 3.54 0.97 S69 2.65 2.90 2.99 2.88 2.85 2.62 2.82 0.15 302

PAGE 303

Table M-3. LDI scores calculated for six different grain sizes (units: meters on a side) and assuming that the effect of development intensity on the landscape decreases in inverse square with distance. Site code 20-m 50-m 80-m 110-m 140-m 170-m Mean SD S1 0.28 0.29 0.39 0.36 0.29 0.47 0.35 0.07 S2 8.36 8.83 8.88 8.68 8.44 8.01 8.53 0.33 S3 8.09 8.57 8.61 8.76 8.83 8.23 8.52 0.29 S4 0.28 0.29 0.28 0.26 0.24 0.24 0.26 0.02 S5 5.43 5.57 5.65 5.59 5.47 5.51 5.54 0.08 S6 3.95 4.19 4.12 4.10 3.76 3.89 4.00 0.16 S7 5.72 5.95 5.96 5.80 5.33 5.88 5.77 0.23 S8 6.67 6.78 6.96 6.83 6.79 6.84 6.81 0.10 S9 2.40 2.54 2.51 2.41 2.30 2.27 2.40 0.11 S10 9.00 9.11 9.14 9.04 9.09 9.03 9.07 0.06 S11 11.17 11.39 11.61 11.24 11.58 11.17 11.36 0.20 S12 13.35 13.42 13.45 13.52 13.49 13.50 13.46 0.06 S13 10.67 10.77 10.74 10.78 10.52 10.95 10.74 0.14 S14 6.58 7.09 7.04 7.11 6.82 7.54 7.03 0.32 S15 6.23 6.46 6.57 6.68 6.59 6.66 6.53 0.17 S16 9.17 9.30 9.31 9.18 9.18 9.14 9.21 0.07 S17 6.86 6.98 6.91 6.83 6.84 6.86 6.88 0.06 S18 7.01 7.21 7.16 7.01 6.87 6.82 7.01 0.16 S19 7.81 8.19 8.26 8.35 8.27 8.51 8.23 0.24 S20 11.74 12.03 12.22 12.01 12.29 12.00 12.05 0.19 S21 4.46 4.76 4.55 4.43 4.12 3.90 4.37 0.31 S22 0.25 0.26 0.26 0.26 0.26 0.26 0.26 0.00 S23 11.14 11.38 11.32 11.20 11.14 11.26 11.24 0.10 S24 0.55 0.61 0.75 0.48 0.42 0.36 0.53 0.14 S25 3.01 3.11 3.12 3.08 3.32 3.40 3.17 0.15 S26 10.20 10.67 10.91 12.27 12.45 12.17 11.45 0.96 S27 11.06 11.36 11.46 11.31 11.16 11.50 11.31 0.17 S28 7.76 7.93 7.98 7.59 7.71 7.41 7.73 0.21 S29 18.70 18.75 18.89 18.58 19.07 19.28 18.88 0.26 S30 3.94 4.22 4.08 4.22 4.19 4.28 4.16 0.12 S31 4.73 4.94 5.22 4.62 4.49 4.63 4.77 0.26 S32 10.99 11.24 11.31 11.28 11.38 11.32 11.25 0.14 S33 13.77 13.90 13.96 13.78 13.77 13.77 13.83 0.08 S34 10.45 10.63 10.67 10.58 10.55 10.61 10.58 0.08 S35 10.68 10.88 10.83 10.83 10.75 10.82 10.80 0.07 S36 1.52 1.55 1.57 1.58 1.58 1.60 1.57 0.03 S37 2.21 2.34 2.30 2.28 2.27 2.29 2.28 0.04 303

PAGE 304

Table M-3. Continued. Site code 20-m 50-m 80-m 110-m 140-m 170-m Mean SD S38 16.28 16.28 16.27 16.26 16.26 16.23 16.26 0.02 S39 3.57 3.73 3.69 3.45 3.40 3.42 3.54 0.14 S40 1.92 2.00 2.00 1.88 1.84 1.80 1.91 0.08 S41 8.76 9.13 9.21 9.10 9.07 8.88 9.03 0.17 S42 4.86 5.00 5.08 4.95 4.89 4.91 4.95 0.08 S43 4.87 4.94 4.94 5.01 4.81 4.98 4.93 0.07 S44 4.08 4.27 4.46 4.57 4.41 4.31 4.35 0.17 S45 11.08 11.24 11.35 11.28 11.31 11.25 11.25 0.09 S46 11.18 11.40 11.52 11.47 11.22 11.26 11.34 0.14 S47 7.53 7.79 7.94 7.85 7.86 7.82 7.80 0.14 S48 7.86 8.15 8.30 8.20 8.21 8.25 8.16 0.15 S48 4.44 4.53 4.48 4.76 5.18 6.00 4.90 0.61 S50 11.93 12.06 12.06 12.03 11.91 11.89 11.98 0.08 S51 0.63 0.69 0.63 0.60 0.58 0.59 0.62 0.04 S52 9.08 9.29 9.33 9.20 9.04 8.99 9.15 0.14 S53 15.33 15.63 15.62 15.59 15.66 15.63 15.58 0.12 S54 10.82 11.08 11.10 11.27 11.36 11.34 11.16 0.20 S55 1.26 1.34 1.37 1.24 1.31 1.28 1.30 0.05 S56 2.37 2.67 2.22 1.31 1.30 1.11 1.83 0.67 S57 1.94 2.22 2.14 2.20 1.75 1.75 2.00 0.22 S58 2.75 2.88 2.79 2.65 2.45 2.05 2.60 0.30 S59 3.58 3.83 3.45 3.07 2.98 2.88 3.30 0.38 S60 6.05 6.26 6.27 5.83 5.77 5.72 5.98 0.24 S61 15.16 15.30 15.49 15.41 15.33 14.98 15.28 0.18 S62 0.35 0.54 0.75 0.22 0.22 0.23 0.39 0.22 S63 4.76 5.01 4.85 4.11 3.65 3.22 4.27 0.73 S64 2.56 2.72 2.74 2.42 2.01 0.88 2.22 0.71 S65 4.62 4.89 5.13 4.36 4.03 3.78 4.47 0.51 S66 10.86 11.27 11.22 11.22 11.37 11.06 11.17 0.18 S67 11.98 12.43 12.34 12.49 12.25 12.41 12.32 0.18 S68 2.88 3.18 3.35 1.90 1.55 1.28 2.36 0.89 S69 1.55 1.68 1.73 1.72 1.78 1.37 1.64 0.15 304

PAGE 305

APPENDIX N DESCRIPTIVE STATISTICS FOR LANDSCAPE PATTERN METRICS Table N-1. Isolated forested wetlands (n = 51): summary statistics and transformation information for landscape pattern metrics calc ulated at different grain sizes (meters on a side). Summary statistics expressed in untransformed values (refer to Table 2-7 for units and metrics names). Grain size Metric (acronym) Mean SD Min Max Transformation ADa statistics p 5-m PLAND_Urb 47.03 34.300.0096.38 arcsine sqrt 2.83<0.005 PLAND_Ag 13.71 27.590.0087.93 arcsine sqrt 10.17<0.005 PLAND_For 18.22 20.340.0085.08 arcsine sqrt 0.850.027 PLAND_Wet 10.15 17.140.0085.17 arcsine sqrt 3.41<0.005 PD 70.29 28.7022.42151.71 log10 0.300.575 ED 313.94 64.49207.40468.66 sqrt 0.540.155 AREA_MN 1.69 0.790.664.46 log10 0.300.575 AREA_CV 119.44 35.5735.73207.92 0.710.062 SHAPE_MN 1.77 0.241.372.28 sqrt 0.610.109 FRAC_MN 1.12 0.031.071.19 0.490.218 ENN_MN 47.70 34.790.00145.10 sqrt 0.320.517 CONTAG 56.42 6.6042.1771.99 0.210.842 IJIa 67.88 10.8131.0095.56 arcsine sqrt 4.27<0.005 PR 6.51 2.112.0012.00 sqrt 1.060.008 PRD 35.59 11.2611.4562.79 0.150.956 SHDI 1.36 0.370.552.26 0.320.528 SHEI 0.75 0.120.400.98 0.480.222 10-m PLAND_Urb 47.04 34.330.0096.67 arcsine sqrt 2.80<0.005 PLAND_Ag 13.69 27.550.0087.68 arcsine sqrt 10.18<0.005 PLAND_For 18.21 20.320.0085.18 arcsine sqrt 0.840.028 PLAND_Wet 10.11 17.130.0085.20 arcsine sqrt 3.41<0.005 PD 73.35 31.2122.88146.08 log10 0.170.927 ED 303.98 59.58203.03446.28 sqrt 0.540.156 AREA_MN 1.64 0.770.684.37 log10 0.170.927 AREA_CV 137.06 62.7167.25333.68 log10 0.900.020 SHAPE_MN 1.63 0.211.222.24 0.520.184 FRAC_MN 1.10 0.021.041.16 0.380.384 ENN_MN 56.38 31.370.00133.31 sqrt 0.600.116 CONTAG 52.39 7.1534.6868.67 0.280.618 IJIa 68.00 10.8329.5995.23 arcsin sqrt 4.01<0.005 PR 6.51 2.112.0012.00 sqrt 1.060.008 PRD 35.59 11.2911.4462.73 0.140.971 SHDI 1.36 0.370.552.25 0.330.506 305

PAGE 306

Table N-1. Continued. Grain size Metric (acronym) Mean SD Min Max Transformation ADa statistics p SHEI 0.75 0.120.400.98 0.390.363 20-m PLAND_Urb 47.02 34.280.0096.85arcsine sqrt 2.85<0.005 PLAND_Ag 13.72 27.600.0087.38arcsine sqrt 10.13<0.005 PLAND_For 18.26 20.410.0086.18arcsine sqrt 0.850.027 PLAND_Wet 10.18 17.160.0085.50arcsine sqrt 3.44<0.005 PD 72.30 30.1723.04151.84 0.730.055 ED 276.65 47.28205.62395.02 0.710.061 AREA_MN 1.67 0.800.664.34log10 0.410.339 AREA_CV 146.88 65.9248.30390.72log10 0.180.912 SHAPE_MN 1.38 0.121.131.70 0.430.295 FRAC_MN 1.07 0.021.031.11 0.380.384 ENN_MN 74.31 37.890.00221.02sqrt 3.16<0.005 CONTAG 46.76 7.9425.1962.64 0.360.446 IJIb 69.41 11.6330.3995.64arcsine sqrt 3.76<0.005 PR 6.35 2.112.0012.00sqrt 1.010.010 PRD 34.64 10.9411.5262.50 0.280.636 SHDI 1.35 0.370.492.23 0.310.548 SHEI 0.75 0.120.450.98 0.460.248 30-m PLAND_Urb 46.92 34.560.0095.21arcsine sqrt 2.65<0.005 PLAND_Ag 13.77 27.570.0085.71arcsine sqrt 10.12<0.005 PLAND_For 18.27 20.370.0084.02arcsine sqrt 0.850.027 PLAND_Wet 10.11 17.320.0085.22arcsine sqrt 2.54<0.005 PD 63.23 25.1822.45136.66 0.600.116 ED 257.26 40.51187.91353.04 0.710.059 AREA_MN 1.85 0.800.734.46log10 0.210.865 AREA_CV 133.38 50.6256.30327.78log10 0.270.664 SHAPE_MN 1.30 0.101.121.69 0.460.253 FRAC_MN 1.06 0.021.021.11 0.420.313 ENN_MN 101.87 57.270.00369.93sqrt 3.05<0.005 CONTAG 42.77 9.2818.2066.16 0.280.645 IJIa 69.87 11.5125.0295.78arcsine sqrt 3.76<0.005 PR 6.31 2.042.0012.00 0.620.099 PRD 34.43 10.9511.2263.19 0.240.780 SHDI 0.66 0.130.280.87 0.330.504 SHEI 0.75 0.130.390.97 0.560.145a The Anderson-Darling test was used to determin e whether the metrics scores were normally distributed; if p < 0.05 the data did not follow a normal distribution. 306

PAGE 307

Table N-2. Streams (n = 68): summary statistic s and transformation information for landscape pattern metrics calculated at four differe nt grain sizes (meter s on a side). Summary statistics are expressed in untransformed values (refer to Table 2-7 for units and metrics names). Grain size Metric (acronym) Mean SD Min Max Transformation ADa statistic p 20-m PLAND_Urb 12.97 18.200.0099.44arcsine sqrt 8.35 <0.005 PLAND_Ag 18.11 20.170.0075.90arcsine sqrt 4.25 <0.005 PLAND_For 43.44 25.590.0093.39arcsine sqrt 1.58 <0.005 PLAND_Wet 16.87 10.340.0041.88 0.480.222 PD 7.75 3.461.9216.85 0.470.235 ED 84.30 20.0834.18137.14 0.200.868 AREA_MN 16.58 10.055.9352.21log10 1.000.012 AREA_CV 505.92 291.07104.361656.25log10 0.210.861 SHAPE_MN 1.61 0.161.332.27sqrt 1.65<0.005 FRAC_MN 1.08 0.011.051.11 0.690.067 ENN_MN 421.99 152.700.00816.68sqrt 1.39<0.005 CONTAG 63.67 5.5350.4379.63 0.970.842 IJI 59.76 8.1132.1671.64arcsine sqrt 0.810.035 PR 32.53 16.245.0070.00sqrt 0.410.340 PRD 0.96 1.540.077.74log10 0.560.141 SHDI 2.06 0.500.803.13 0.280.528 SHEI 0.61 0.100.350.83 0.630.222 50-m PLAND_Urb 12.97 18.220.0099.60 arcsine sqrt 8.40<0.005 PLAND_Ag 18.11 20.180.0075.90 arcsine sqrt 4.25<0.005 PLAND_For 43.45 25.610.0093.48 arcsine sqrt 1.58<0.005 PLAND_Wet 16.87 10.320.0041.59 0.480.222 PD 7.68 3.142.2719.48 0.400.360 ED 78.74 17.7735.03131.82 0.350.456 AREA_MN 15.82 8.275.1344.10 log10 0.980.013 AREA_CV 566.46 321.10104.031689.04 log10 0.370.418 SHAPE_MN 1.44 0.121.271.85 sqrt 1.46<0.005 FRAC_MN 1.06 0.011.041.09 0.410.343 ENN_MN 438.96 141.130.00797.28 sqrt 1.93<0.005 CONTAG 58.06 6.1042.1972.30 0.230.802 IJI 59.73 8.2630.4372.06 arcsin sqrt 0.690.068 PR 32.21 16.115.0070.00 sqrt 0.410.330 PRD 0.93 1.470.078.03 log10 0.590.118 SHDI 2.07 0.490.803.13 0.320.522 SHEI 0.62 0.090.360.83 0.230.790 307

PAGE 308

Table N-2. Continued. Grain size Metric (acronym) Mean SD Min Max Transformation ADa statistic P 80-m PLAND_Urb 12.97 18.22 0.00 99.60 arcsine sqrt 8.59<0.005 PLAND_Ag 18.11 20.18 0.00 75.90 arcsine sqrt 4.26<0.005 PLAND_For 43.45 25.61 0.00 93.48 arcsine sqrt 1.590.027 PLAND_Wet 16.87 10.32 0.00 41.59 0.480.222 PD 6.91 2.68 2.34 17.65 0.460.259 ED 72.55 16.29 32.81 119.83 0.400.357 AREA_MN 16.97 7.69 5.67 42.65 log10 0.710.061 AREA_CV 559.4 313.9 81.38 1582.3 log10 0.450.272 SHAPE_MN 1.34 0.09 1.20 1.67 sqrt 1.87<0.005 FRAC_MN 1.05 0.01 1.04 1.07 1.13<0.005 ENN_MN 511.9 155.0 0.00 870.48 sqrt 1.89<0.005 CONTAG 54.17 6.64 36.63 69.11 0.210.855 IJI 60.28 8.46 30.83 83.07 arcsine sqrt 0.630.095 PR 31.72 16.21 4.00 69.00 sqrt 0.440.276 PRD 0.89 1.32 0.07 6.89 log10 0.680.075 SHDI 2.07 0.49 0.82 2.13 0.320.523 SHEI 0.62 0.09 0.37 0.85 0.200.875 110-m PLAND_Urb 13.02 18.27 0.00 100.00 arcsine sqrt 8.60<0.005 PLAND_Ag 18.18 20.19 0.00 75.42 arcsine sqrt 4.21<0.005 PLAND_For 43.44 25.64 0.00 93.43 arcsine sqrt 1.640.027 PLAND_Wet 16.30 10.20 0.00 41.89 0.390.336 PD 5.86 2.09 1.98 14.35 0.380.393 ED 65.81 14.60 30.23 108.13 0.650.085 AREA_MN 19.66 8.59 6.97 50.39 log10 0.970.014 AREA_CV 530.1 297.0 78.50 1796.7 log10 0.520.181 SHAPE_MN 1.28 0.07 1.18 1.59 sqrt 2.21<0.005 FRAC_MN 1.04 0.01 1.03 1.06 1.11<0.005 ENN_MN 607.7 173.1 0.00 1022.3 sqrt 2.16<0.005 CONTAG 51.44 7.34 32.42 68.14 0.210.852 IJI 60.37 8.35 29.38 74.01 arcsine sqrt 0.590.118 PR 31.04 16.25 4.00 69.00 sqrt 0.460.256 PRD 0.85 1.23 0.07 6.48 log10 0.670.075 SHDI 2.07 0.49 0.80 3.12 0.330.508 SHEI 0.63 0.10 0.37 0.86 0.260.716aThe Anderson-Darling test was used to determin e whether the metrics scores were normally distributed; if p < 0.05 the data did not follow a normal distribution. 308

PAGE 309

Table N-3. Streams (n = 63): summary statistic s and transformation information for landscape pattern metrics calculated at three different spatial extents (units in meters). Summary statistics are expressed in untransformed values (refer to Table 2-7 for units and metrics names). Buffer size Metric (acronym) Mean SD Min Max Transform ADa statistic p 100-m PLAND_Urb 8.43 13.630.0559.28arcsine sqrt 3.14<0.005 PLAND_Ag 13.75 18.820.3276.66arcsine sqrt 2.42<0.005 PLAND_For 37.44 23.803.8494.81arcsine sqrt 0.580.128 PLAND_Wet 34.83 16.480.4369.15 0.350.463 PD 24.15 6.2910.3336.14 0.250.736 ED 213.20 23.85166.44263.37 0.620.104 AREA_MN 4.49 1.492.779.68log10 0.980.013 AREA_CV 322.50 150.8895.44729.32log10 0.200.868 SHAPE_MN 1.62 0.141.382.17 0.650.084 FRAC_MN 1.09 0.011.071.12 0.480.231 ENN_MN 449.96 220.82103.08952.56sqrt 0.320.517 CONTAG 59.91 5.7045.0971.98arcsine sqrt 0.860.025 IJI 56.43 8.4321.8874.31arcsine sqrt 1.55<0.005 PR 21.10 8.495.0039 0.480.223 PRD 3.45 3.890.6527.64log10 0.450.261 SHDI 1.89 0.420.802.87 0.200.881 SHEI 0.64 0.100.450.89 0.630.098 400-m PLAND_Urb 8.87 13.30 0.0064.17arcsine sqrt 2.03<0.005 PLAND_Ag 15.70 19.67 0.0077.10arcsine sqrt 1.67<0.005 PLAND_For 44.28 25.02 4.7193.90arcsine sqrt 1.24<0.005 PLAND_Wet 22.50 12.52 1.2249.33 0.640.090 PD 10.07 3.72 3.8321.34sqrt 0.670.075 ED 106.96 16.87 64.68142.25 0.330.504 AREA_MN 11.53 5.09 4.6926.12log10 1.14<0.005 AREA_CV 345.77 142.51 116.12720.12log10 0.490.216 SHAPE_MN 1.61 0.14 1.402.10 0.720.057 FRAC_MN 1.08 0.01 1.061.11 0.200.876 ENN_MN 472.16 180.22 219.39946.48sqrt 0.940.016 CONTAG 61.83 5.18 48.7175.03 0.350.466 IJI 59.36 7.60 29.7471.76arcsin sqrt 0.990.012 PR 26.08 10.41 8.0050.00sqrt 0.380.386 PRD 1.27 1.48 0.2510.08log10 0.790.039 SHDI 2.02 0.42 0.862.94 0.240.772 SHEI 0.63 0.09 0.410.86 0.360.430 309

PAGE 310

Table N-3. Continued. Buffer size Metric (acronym) Mean SD Min Max Transform ADa statistic p Washb PLAND_Urb 12.97 18.220.0099.60arcsine sqrt 5.44<0.005 PLAND_Ag 18.11 20.180.0075.90arcsine sqrt 4.03<0.005 PLAND_For 43.45 25.610.0093.48arcsine sqrt 1.53<0.005 PLAND_Wet 16.87 10.320.0041.59 0.440.277 PD 6.91 2.682.3417.65 0.420.316 ED 72.55 16.2932.81119.83 0.310.541 AREA_MN 16.97 7.695.6742.65log10 1.27<0.005 AREA_CV 559.4 313.981.381582.3log10 0.160.943 SHAPE_MN 1.34 0.091.201.67sqrt 1.75<0.005 FRAC_MN 1.05 0.011.041.07 0.780.040 ENN_MN 511.9 155.00.00870.48sqrt 1.10<0.005 CONTAG 54.17 6.6436.6369.11 0.150.963 IJI 60.28 8.4630.8383.07arcsine sqrt 0.700.039 PR 31.72 16.214.0069.00sqrt 0.460.254 PRD 0.89 1.320.076.89log10 0.390.378 SHDI 2.07 0.490.822.13 0.590.122 SHEI 0.62 0.090.370.85 0.290.614a The Anderson-Darling test was used to determin e whether the metrics scores were normally distributed; if p < 0.05 the data did not follow a normal distribution. bWash = watershed. 310

PAGE 311

Table N-4. Lakes (n = 48): summary statistics and transformation information for landscape pattern metrics calculated at four different grain sizes (meters on as side). Summary statistics expressed in untransformed values (refer to Table 2-7 for units and metrics names). Grain size Metric (acronym) Mean SD Min Max Transformation ADa statistic p 20-m PLAND_Urb 45.60 30.23 0.40 99.83 arcsine sqrt 0.23 0.801 PLAND_Ag 26.40 19.29 0.10 73.35 arcsine sqrt 1.82 <0.005 PLAND_For 17.92 18.61 0.12 78.11 arcsine sqrt 0.88 0.022 PLAND_Wet 13.85 14.51 0.23 53.53 arcsine sqrt 0.96 0.014 PD 20.05 9.05 5.87 39.40 log10 0.32 0.520 ED 74.90 20.41 29.31 127.08 0.26 0.697 AREA_MN 6.24 3.26 2.54 17.03 log10 0.32 0.520 AREA_CV 225.76 76.04 91.54 413.90 0.55 0.153 SHAPE_MN 1.54 0.13 1.26 1.81 0.52 0.182 FRAC_MN 1.08 0.01 1.04 1.12 0.28 0.619 ENN_MN 299.29 167.33 85.32 852.98 sqrt 0.37 0.422 CONTAG 58.19 6.88 45.74 74.39 0.35 0.458 IJI 65.38 7.14 43.16 75.22 arcsine sqrt 1.59 <0.005 PR 12.10 4.82 5.00 24.00 sqrt 0.99 0.012 PRD 6.11 4.37 0.90 21.77 log10 0.69 0.067 SHDI 1.67 0.42 0.88 2.33 sqrt 1.33 0.002 SHEI 0.69 0.11 0.41 0.88 0.56 0.139 40-m PLAND_Urb 46.13 30.35 0.39 99.89 arcsine sqrt 0.24 0.756 PLAND_Ag 25.88 19.76 0.18 73.21 arcsine sqrt 1.80 <0.005 PLAND_For 17.85 18.59 0.15 77.75 arcsine sqrt 0.90 0.020 PLAND_Wet 13.74 14.47 0.21 53.45 arcsine sqrt 1.02 0.010 PD 18.44 8.48 4.47 43.60 0.54 0.157 ED 68.03 18.63 28.42 110.47 0.19 0.890 AREA_MN 6.85 3.81 2.29 22.36 log10 0.33 0.506 AREA_CV 214.37 68.61 104.07 378.14 0.56 0.139 SHAPE_MN 1.43 0.11 1.27 1.70 0.54 0.161 FRAC_MN 1.06 0.01 1.04 1.09 0.38 0.399 ENN_MN 335.29 162.12 108.08 924.10 sqrt 0.86 0.026 CONTAG 53.29 7.79 37.60 71.56 0.26 0.715 IJI 65.87 7.09 45.87 82.17 arcsine sqrt 1.15 0.005 PR 11.94 4.81 5.00 24.00 sqrt 0.84 0.028 PRD 5.99 4.17 0.89 18.94 log10 0.68 0.071 SHDI 1.67 0.42 0.88 2.33 sqrt 1.35 0.001 SHEI 0.69 0.11 0.41 0.89 0.48 0.228 60-m PLAND_Urb 45.52 30.02 0.43 100.00 arcsine sqrt 0.22 0.833 PLAND_Ag 26.46 19.26 0.10 73.78 arcsine sqrt 1.84 <0.005 PLAND_For 18.42 18.64 0.90 78.40 arcsine sqrt 0.93 0.017 PLAND_Wet 13.88 14.40 0.35 53.59 arcsine sqrt 0.87 0.024 311

PAGE 312

Table N-4. Continued. Grain size Metric (acronym) Mean SD Min Max Transformation ADa statistic p PD 17.21 7.93 4.46 41.40 0.65 0.086 ED 61.01 16.05 27.50 95.91 0.32 0.533 AREA_MN 7.28 3.95 2.42 22.42 log10 0.29 0.591 AREA_CV 205.93 65.69 110.05 346.18 log10 0.43 0.303 SHAPE_MN 1.34 0.10 1.12 1.61 0.26 0.708 FRAC_MN 1.05 0.01 1.03 1.08 0.26 0.682 ENN_MN 369.24 152.63 153.45 977.69 sqrt 0.41 0.338 CONTAG 49.15 8.04 33.16 69.63 0.25 0.722 IJI 67.35 6.20 48.97 81.31 arcsine sqrt 1.21 0.003 PR 11.58 4.88 4.00 24.00 sqrt 0.71 0.060 PRD 5.75 3.99 0.89 18.73 log10 0.67 0.074 SHDI 1.67 0.42 0.88 2.33 sqrt 1.32 0.002 SHEI 0.70 0.11 0.41 0.93 0.39 0.380 80-m PLAND_Urb 45.59 30.32 0.38 100.00 arcsine sqrt 0.21 0.860 PLAND_Ag 27.17 18.95 0.72 72.05 arcsine sqrt 1.81 <0.005 PLAND_For 18.59 18.71 0.69 78.65 arcsine sqrt 0.90 0.020 PLAND_Wet 14.78 14.42 1.02 53.05 arcsine sqrt 0.93 0.017 PD 15.98 7.23 4.57 37.01 log10 0.37 0.420 ED 54.64 16.10 24.55 107.92 0.36 0.427 AREA_MN 7.71 3.95 2.70 21.90 log10 0.37 0.420 AREA_CV 195.12 64.58 67.00 357.05 0.59 0.119 SHAPE_MN 1.27 0.09 1.06 1.54 sqrt 0.87 0.023 FRAC_MN 1.04 0.01 1.01 1.07 0.59 0.118 ENN_MN 413.09 161.34 178.48 949.38 0.68 0.072 CONTAG 46.83 8.59 29.60 68.85 0.23 0.784 IJI 67.42 6.18 51.58 79.14 0.59 0.118 PR 11.42 4.81 4.00 24.00 sqrt 0.79 0.038 PRD 5.71 3.97 0.79 17.69 log10 0.58 0.123 SHDI 1.67 0.42 0.86 2.33 sqrt 1.26 0.002 SHEI 0.70 0.11 0.41 0.95 0.23 0.790 a The Anderson-Darling test was used to determin e whether the metrics scores were normally distributed; if p < 0.05 the data did not follow a normal distribution. 312

PAGE 313

Table N-5. Lakes (n =44): summary statistics and transformation information for landscape pattern metrics calculated at different sp atial extents (units in meters). Summary statistics expressed in untransformed values (refer to Table 2-7 for units and metrics names). Buffer size Metric (acronym) Mean SD Min Max Transformation ADa statistic p 100-m PLAND_Urb 45.62 28.89 0.66 99.78 arcsine sqrt 0.393 0.363 PLAND_Ag 22.53 17.41 0.41 55.12 arcsine sqrt 3.320 <0.001 PLAND_For 16.69 18.48 0.14 80.98 arcsine sqrt 1.868 <0.001 PLAND_Wet 17.41 16.47 0.13 64.83 arcsine sqrt 0.619 0.100 PD 51.00 23.05 15.82 122.03 sqrt 0.616 0.102 ED 85.10 40.54 15.58 205.18 0.592 0.118 AREA_MN 2.37 1.07 0.82 6.32 0.717 0.057 AREA_CV 168.37 63.23 72.00 383.34 log10 0.308 0.547 SHAPE_MN 1.61 0.21 1.36 2.21 sqrt 1.382 0.001 FRAC_MN 1.09 0.02 1.05 1.15 0.543 0.154 ENN_MN 232.14 132.89 0.00 597.82 sqrt 0.287 0.607 CONTAG 55.35 11.01 30.09 83.00 0.228 0.800 IJI 59.50 10.36 17.62 76.86 0.709 0.060 PR 8.27 3.69 3.00 19.00 sqrt 0.472 0.233 PRD 16.91 9.16 2.85 35.55 0.375 0.399 SHDI 1.40 0.41 0.40 2.08 0.498 0.200 SHEI 0.70 0.16 0.25 0.93 0.637 0.091 400-m PLAND_Urb 45.62 28.89 0.66 99.78 0.574 0.128 PLAND_Ag 22.53 17.41 0.41 55.12 arcsine sqrt 1.831 <0.001 PLAND_For 16.69 18.48 0.14 80.98 arcsine sqrt 0.931 0.017 PLAND_Wet 17.41 16.47 0.13 64.83 arcsine sqrt 0.611 0.105 PD 24.19 9.99 8.41 47.28 sqrt 0.413 0.325 ED 78.26 21.96 24.12 126.64 0.163 0.940 AREA_MN 4.92 2.19 2.12 11.88 log10 0.306 0.553 AREA_CV 201.05 62.06 98.00 419.74 0.584 0.121 SHAPE_MN 1.54 0.14 1.34 1.91 0.617 0.102 FRAC_MN 1.08 0.02 1.05 1.13 0.675 0.073 ENN_MN 297.03 167.28 91.24 769.25 sqrt 0.477 0.226 CONTAG 57.50 7.23 44.52 74.44 0.333 0.502 IJI 63.13 8.43 32.17 73.59 arcsine sqrt 1.410 0.001 PR 11.11 3.87 5.00 20.00 sqrt 0.643 0.088 PRD 7.90 4.82 1.59 26.01 log10 0.514 0.183 SHDI 1.63 0.39 0.74 2.27 0.521 0.176 SHEI 0.69 0.12 0.42 0.89 0.339 0.486 Washb PLAND_Urb 45.00 30.01 0.40 99.83 0.728 0.053 PLAND_Ag 23.85 17.68 0.10 67.51 arcsine sqrt 1.786 <0.001 PLAND_For 18.90 19.06 0.12 78.11 arcsine sqrt 0.693 0.066 PLAND_Wet 13.43 13.35 0.48 50.83 arcsine sqrt 0.780 0.039 313

PAGE 314

Table N-5. Continued. Buffer size Metric (acronym) Mean SD Min Max Transformation ADa statistic p PD 20.34 9.27 5.87 39.40 0.693 0.065 ED 75.18 20.91 29.31 127.08 0.249 0.732 AREA_MN 6.20 3.33 2.54 17.03 log10 0.326 0.511 AREA_CV 225.83 76.05 91.54 413.90 0.675 0.073 SHAPE_MN 1.54 0.13 1.26 1.81 0.654 0.082 FRAC_MN 1.08 0.01 1.04 1.12 0.361 0.431 ENN_MN 292.07 164.71 85.32 852.98 sqrt 0.354 0.449 CONTAG 57.98 7.11 45.74 74.39 0.428 0.299 IJI 65.17 7.27 43.16 75.22 arcsine sqrt 1.622 <0.001 PR 12.05 4.88 5.00 24.00 sqrt 0.927 0.017 PRD 6.11 4.55 0.90 21.77 log10 0.574 0.128 SHDI 1.67 0.43 0.88 2.33 sqrt 1.346 0.002 SHEI 0.69 0.12 0.41 0.88 0.723 0.055 a The Anderson-Darling test was used to determin e whether the metrics scores were normally distributed; if p < 0.05 the data did not follow a normal distribution. bWash = Watershed. 314

PAGE 315

APPENDIX O LANDSCAPE INDICES AND INDICATO RS OF ECOSYSTEM CONDITION Table O-1. Multiple regression models at four gr ain sizes for the sample isolated forested wetlands: coefficients of determination, pr obabilities, and change in the amount of variability ( R2) in indicators of ecosystems condition ( level of 0.05). Dependent Variable Independent Variables R2 (adj) p R2 5 x 5-meters Water chemistry Log10(DO) LDI-PLU; URB; HETER; SHAP E; AG; FOR 0.00 0.539 -0.07 LDI-ILD; URB; HETER; SHAPE; AG; FOR 0.00 0.534 -0.06 LDI-ISD; URB; HETER; SHAPE; AG; FOR 0.00 0.526 -0.05 Log10(SC) LDI-PLU; URB; HETER; SHAP E; AG; FOR 0.09 0.355 -0.12 LDI-ILD; URB; HETER; SHAPE; AG; FOR 0.08 0.364 -0.10 LDI-ISD; URB; HETER; SHAPE; AG; FOR 0.08 0.368 -0.08 Log10(TN) LDI-PLU; URB; HETER; SHAP E; AG; FOR 0.00 0.521 -0.02 LDI-ILD; URB; HETER; SHAPE; AG; FOR 0.00 0.523 -0.03 LDI-ISD; URB; HETER; SHAPE; AG; FOR 0.00 0.519 -0.03 Log10(TP) LDI-PLU; URB; HETER; SHAPE; AG; FOR 0.16 0.104 0.12 LDI-ILD; URB; HETER; SHAPE; AG; FOR 0.16 0.111 0.12 LDI-ISD; URB; HETER; SHAPE; AG; FOR 0.15 0.124 0.11 Log10(Turbidity) LDI-PLU; URB; HETER; SH APE; AG; FOR 0.00 0.640 -0.03 LDI-ILD; URB; HETER; SHAPE; AG; FOR 0.00 0.683 -0.03 LDI-ISD; URB; HETER; SHAPE; AG; FOR 0.00 0.717 -0.03 WCI Macroinvertebrates LDI-PLU; URB; HETER; SHAPE; AG; FOR 0.23 0.050 0.05 Diatoms LDI-PLU; URB; HETER; SHAPE; AG; FOR 0.07 0.332 -0.15 LDI-ILD; URB; HETER; SHAPE; AG; FOR 0.08 0.321 -0.12 LDI-ISD; URB; HETER; SHAPE; AG; FOR 0.08 0.318 -1.78 10 x 10-meters Water chemistry Log10(DO) LDI-PLU; URB; HETER; CONTAG ; AG; FOR; ENN 0.00 0.704 -0.07 LDI-ILD; URB; HETER; CONTAG; AG; FOR; ENN 0. 00 0.712 -0.06 LDI-ISD; URB; HETER; CONTAG; AG; FOR; ENN 0. 00 0.716 -0.05 Log10(SC) LDI-PLU; URB; HETER; CONTAG ; AG; FOR; ENN 0.03 0.447 -0.17 LDI-ILD; URB; HETER; CONTAG; AG; FOR; ENN 0. 01 0.477 -0.17 LDI-ISD; URB; HETER; CONTAG; AG; FOR; ENN 0. 00 0.488 -0.16 Log10(TN) LDI-PLU; URB; HETER; CONT AG; AG; FOR; ENN 0.03 0.369 0.01 LDI-ILD; URB; HETER; CONTAG ; AG; FOR; ENN 0.03 0.371 0.00 LDI-ISD; URB; HETER; CONTAG ; AG; FOR; ENN 0.03 0.373 0.00 Log10(Turbidity) LDI-PLU; URB; HETER; C ONTAG; AG; FOR; ENN 0.00 0.782 -0.03 LDI-ILD; URB; HETER; CONTAG; AG; FOR; ENN 0. 00 0.807 -0.03 LDI-ISD; URB; HETER; CONTAG; AG; FOR; ENN 0. 00 0.830 -0.03 WCI Macroinvertebrates LDI-PLU; URB; HETER; CONTAG; AG; FOR; ENN 0.21 0.077 0.03 LDI-ILD; URB; HETER; CONTAG ; AG; FOR; ENN 0.24 0.060 0.06 Diatoms LDI-PLU; URB; HETER; C ONTAG; AG; FOR; ENN 0.01 0.448 -0.22 LDI-ILD; URB; HETER; CONTAG; AG; FOR; ENN 0. 03 0.429 -0.18 315

PAGE 316

Table O-1. Continued. Dependent Variable Independent Variables R2 (adj) p R2 LDI-ISD; URB; HETER; CONTAG; AG; FOR; ENN 0. 03 0.431 -0.16 20 x 20-meters Water chemistry Log10(DO) LDI-PLU; DIVERS; HETER; CONT AG; AG; SHAPE; FOR 0.00 0.643 -0.07 LDI-ILD; DIVERS; HETER; CONTAG; AG; SHAPE; FOR 0.00 0.624 -0.06 LDI-ISD; DIVERS; HETER; CONTAG; AG; SHAPE; FOR 0.00 0.604 -0.05 Log10(SC) LDI-PLU; DIVERS; HETER; CONT AG; AG; SHAPE; FOR 0.01 0.473 -0.15 LDI-ILD; DIVERS; HETER; CONTAG; AG; SHAPE; FOR 0.01 0.481 -0.17 LDI-ISD; DIVERS; HETER; CONTAG; AG; SHAPE; FOR 0.00 0.485 -0.16 Log10(TN) LDI-PLU; DIVERS; HETER; CONTAG; AG; SHAPE; FOR 0.00 0.491 -0.02 LDI-PLU; DIVERS; HETER; CONTAG; AG; SHAPE; FOR 0.00 0.502 -0.03 LDI-ILD; DIVERS; HETER; CONTAG; AG; SHAPE; FOR 0.00 0.507 -0.03 Log10(Turbidity) LDI-PLU; DIVERS; HETER; C ONTAG; AG; SHAPE; FOR 0.00 0.826 -0.04 LDI-ILD; DIVERS; HETER; CONTAG; AG; SHAPE; FOR 0.00 0.840 -0.03 LDI-ISD; DIVERS; HETER; CONTAG; AG; SHAPE; FOR 0.00 0.839 -0.03 WCI Diatoms LDI-PLU; DIVERS; HETER; CONTAG; AG; SHAPE; FOR 0.02 0.444 -0.21 LDI-ILD; DIVERS; HETER; CONTAG; AG; SHAPE; FOR 0.02 0.435 -0.18 LDI-ISD; DIVERS; HETER; CONTAG; AG; SHAPE; FOR 0.03 0.431 -0.17 30 x 30-meters Water chemistry Log10(DO) LDI-PLU; DIVERS; HETER; URB/WET; AG; SHAPE; FOR 0.00 0.727 -0.06 LDI-ILD; DIVERS; HETER; URB/WET; AG; SHAPE; FOR 0.00 0.697 -0.05 LDI-ISD; DIVERS; HETER; URB/WET; AG; SHAPE; FOR 0.00 0.672 -0.05 Log10(SC) LDI-PLU; DIVERS; HETER; URB/WET; AG; SHAPE; FOR 0.03 0.454 -0.18 LDI-ILD; DIVERS; HETER; URB/WET; AG; SHAPE; FOR 0.02 0.459 -0.15 LDI-ISD; DIVERS; HETER; URB/WET; AG; SHAPE; FOR 0.03 0.454 -0.12 Log10(TN) LDI-PLU; DIVERS; HETER; URB/WET; AG; SHAPE; FOR 0.14 0.150 0.12 LDI-PLU; DIVERS; HETER; URB/WET; AG; SHAPE; FOR 0.14 0.155 0.11 DIVERS; HETER; URB/WET; AG; SHAPE; FOR; LDI-ILD 0.14 0.156 0.11 Log10(Turbidity) LDI-PLU; DIVERS; HETER; URB/WET; AG; SHAPE; FOR 0.00 0.665 -0.04 LDI-ILD; DIVERS; HETER; URB/WET; AG; SHAPE; FOR 0.00 0.679 -0.03 LDI-ISD; DIVERS; HETER; URB/WET; AG; SHAPE; FOR 0.00 0.689 -0.03 WCI Diatoms LDI-PLU; DIVERS; HETER; URB/WET; AG; SHAPE; FOR 0.07 0.369 -0.17 LDI-ILD; DIVERS; HETER; URB/WET; AG; SHAPE; FOR 0.07 0.364 -0.14 LDI-ISD; DIVERS; HETER; URB/WET; AG; SHAPE; FOR 0.07 0.360 -0.13 316

PAGE 317

Table O-2. Multiple regression models at four grain sizes for the sample streams: coefficients of determination, probabilities, and cha nge in the amount of variability ( R2) in indicators of ecosystems condition ( level of 0.05). Dependent Variable Independent Variables R2 (adj) p R2 20 x 20-meters Water chemistry Log10(Turbidity) LDI-PLU; DIVERS1; DIVE RS2; WET; DIST 0.00 0.798 -0.03 LDI-ILD; DIVERS1; DIVERS2; WET; DIST 0.00 0.753 -0.01 LDI-ISD; DIVERS1; DIVERS2; WET; DIST 0.00 0.643 0.00 Log10(NO3-N) LDI-PLU; DIVERS1; DIVERS2; WET; DIST 0.12 0.079 -0.02 LDI-ILD; DIVERS1; DIVERS2; WET; DIST 0.10 0.098 0.00 LDI-ISD; DIVERS1; DIVERS2; WET; DIST 0.10 0.103 0.02 SCI SC_1 LDI-PLU; DIVERS1; DIVERS2; WET; DIST 0.04 0.204 -0.13 LDI-ILD; DIVERS1; DIVERS2; WET; DIST 0.07 0.086 -0.13 LDI-ISD; DIVERS1; DIVERS2; WET; DIST 0.09 0.051 -0.13 SC_2 LDI-PLU; DIVERS1; DIVERS2; WET; DIST 0.09 0.058 -0.10 50 x 50-meters Water chemistry Log10 (Turbidity) LDI-PLU; DIVERS1; DIVERS2; WET; SHAPE 0.00 0.871 -0.03 LDI-ILD; DIVERS1; DIVERS2; WET; SHAPE 0.00 0.861 -0.01 LDI-ISD; DIVERS1; DIVERS2; WET; SHAPE 0.00 0.787 0.00 Log10(NO3-N) LDI-PLU; DIVERS1; DIVERS2; WET; SHAPE 0.13 0.066 -0.01 LDI-ILD; DIVERS1; DIVERS2; WET; SHAPE 0.11 0.085 0.01 LDI-ISD; DIVERS1; DIVERS2; WET; SHAPE 0.11 0.092 0.03 SCI SC_1 LDI-PLU; DIVERS1; DIVERS2; WET; SHAPE 0.04 0.082 -0.12 LDI-ILD; DIVERS1; DIVERS2; WET; SHAPE 0.08 0.066 -0.12 LDI-ISD; DIVERS1; DIVERS2; WET; SHAPE 0.10 0.050 -0.13 SC_2 LDI-PLU; DIVERS1; DIVERS2; WET; SHAPE 0.08 0.064 -0.10 80 x 80-meters Water chemistry Log10(Turbidity) LDI-PLU; DIVERS1; DIVERS2; URB; SHAPE 0.00 0.852 -0.30 LDI-ILD; DIVERS1; DIVERS2; URB; SHAPE 0.00 0.824 -0.01 LDI-ISD; DIVERS1; DIVERS2; URB; SHAPE 0.00 0.731 0.00 Log10(NO3-N) LDI-PLU; DIVERS1; DIVERS2; URB; SHAPE 0.13 0.061 -0.01 LDI-ILD; DIVERS1; DIVERS2; URB; SHAPE 0.11 0.089 0.01 LDI-ISD; DIVERS1; DIVERS2; URB; SHAPE 0.10 0.101 0.02 SCI SC_1 LDI-PLU; DIVERS1; DIVERS2; URB; SHAPE 0.04 0.072 -0.12 LDI-ILD; DIVERS1; DIVERS2; URB; SHAPE 0.08 0.058 -0.12 SC_2 LDI-PLU; DIVERS1; DIVERS2; URB; SHAPE 0.08 0.081 -0.10 110 x 110-meters Water chemistry Log10(Turbidity) LDI-PLU; DIVERS1; DIVERS2; URB; SHAPE 0.00 0.813 -0.03 LDI-ILD; DIVERS1; DIVERS2; URB; SHAPE 0.00 0.810 -0.01 LDI-ISD; DIVERS1; DIVERS2; URB; SHAPE 0.00 0.720 0.00 317

PAGE 318

Table O-2. Continued. Dependent Variable Independent Variables R2 (adj) p R2 Log10(NO3-N) LDI-PLU; DIVERS1; DIVERS2; URB; SHAPE 0.13 0.061 -0.01 LDI-ILD; DIVERS1; DIVERS2; URB; SHAPE 0.12 0.081 0.02 LDI-ISD; DIVERS1; DIVERS2; URB; SHAPE 0.11 0.090 0.03 SCI SC_1 LDI-PLU; DIVERS1; DIVERS2; URB; SHAPE 0.04 0.208 -0.13 LDI-ILD; DIVERS1; DIVERS2; URB; SHAPE 0.08 0.080 -0.12 SC_2 LDI-PLU; DIVERS1; DIVERS2; URB; SHAPE 0.08 0.080 -0.10 318

PAGE 319

Table O-3. Multiple regression models at thr ee spatial extents for the sample streams: coefficients of determinati on, probabilities, and change in the amount of variability ( R2) in indicators of ecosystems condition ( level of 0.05). Dependent Variable Independent Variables R2 (adj) p R2 100-meter Water chemistry Log10(NO3-N) LDI-PLU; DIVERS; SIZE; WET; HETER 0. 01 0.390 -0.06 LDI-ILD; DIVERS; SIZE; WET; HETER 0.00 0.650 -0.04 LDI-ISD; DIVERS; SIZE; WET; HETER 0.00 0.687 -0.01 400-meter Water chemistry Log10(Turbidity) LDI-PLU; DIVERS; HETER; WET; AG 0.00 0.795 -0.02 LDI-ILD; DIVERS; HETER; WET; AG 0.00 0.535 0.00 LDI-ISD; DIVERS; HETER; WET; AG 0.04 0.315 0.03 Log10(NO3-N) LDI-PLU; DIVERS; HETER; WET; AG 0.08 0.164 -0.01 LDI-ILD; DIVERS; HETER; WET; AG 0.06 0.202 -0.01 LDI-ISD; DIVERS; HETER; WET; AG 0.05 0.229 0.01 SCI SC_1 LDI-PLU; DIVERS; HETER; WET; AG 0.07 0.102 -0.10 LDI-ILD; DIVERS; HETER; WET; AG 0.09 0.060 -0.11 Watershed Water chemistry Log10(Turbidity) LDI-PLU; DIVERS; HETER; WET; AG 0.00 0.951 -0.03 LDI-ILD; DIVERS; HETER; WET; AG 0.00 0.949 -0.01 LDI-ISD; DIVERS; HETER; WET; AG 0.00 0.904 0.00 Log10(NO3-N) LDI-PLU; DIVERS; HETER; WET; AG 0.22 0.052 0.08 LDI-ILD; DIVERS; HETER; WET; AG 0.22 0.067 0.12 LDI-ISD; DIVERS; HETER; WET; AG 0.22 0.070 0.14 SCI SC_1 LDI-PLU; DIVERS; HETER; WET; AG 0.05 0.136 -0.12 LDI-ILD; DIVERS; HETER; WET; AG 0.08 0.070 -0.13 SC_2 LDI-PLU; DIVERS; HETER; WET; AG 0.09 0.053 -0.10 319

PAGE 320

Table O-4. Multiple regression models at four grai n sizes for the sample lakes: coefficients of determination, probabilities, and cha nge in the amount of variability ( R2) in indicators of ecosystems condition ( level of 0.05). Dependent Variable Independent Variables R2 (adj) p R2 20 x 20-meters Water chemistry Log10(Ammonia-N) LDI-PLU; DIVERS1; DIVERS2; URB/SIZE; WET; AG 0.00 0.443 -0.01 LDI-ILD; DIVERS1; DIVERS2; URB/SIZE; WET; AG 0.00 0.643 -0.04 LDI-ISD; DIVERS1; DIVERS2; URB/SIZE; WET; AG 0.00 0.642 -0.04 Log10(NO3/NO2-N) LDI-PLU; DIVERS1; DIVERS2; URB/SIZE; WET; AG 0.07 0.182 0.06 LDI-ILD; DIVERS1; DIVERS2; URB/SIZE; WET; AG 0.01 0.416 0.00 LDI-ISD; DIVERS1; DIVERS2; URB/SIZE; WET; AG 0.02 0.359 0.01 Log10(TKN) LDI-PLU; DIVERS1; DIVERS2; URB/SIZE; WET; AG 0.12 0.077 0.12 LDI-ILD; DIVERS1; DIVERS2; URB/SIZE; WET; AG 0.08 0.156 0.07 LDI-ISD; DIVERS1; DIVERS2; URB/SIZE; WET; AG 0.08 0.153 0.07 40 x 40-meters Water chemistry Log10(Ammonia-N) LDI-PLU; DIVERS1; DIVERS2; URB; AG; SIZE 0.00 0.650 -0.02 LDI-ILD; DIVERS1; DIVERS2; URB; AG; SIZE 0.00 0.601 -0.04 LDI-ISD; DIVERS1; DIVERS2; URB; AG; SIZE 0.00 0.311 -0.04 Log10(NO3/NO2-N) LDI-PLU; DIVERS1; DIVERS2; URB; AG; SIZE 0.07 0.176 0.07 LDI-ILD; DIVERS1; DIVERS2; URB; AG; SIZE 0.00 0.436 0.00 LDI-ISD; DIVERS1; DIVERS2; URB; AG; SIZE 0.00 0.460 0.00 Log10(TKN) LDI-PLU; DIVERS1; DIVERS2; URB; AG; SIZE 0.06 0.194 0.06 LDI-ILD; DIVERS1; DIVERS2; URB; AG; SIZE 0.06 0.214 0.04 LDI-ISD; DIVERS1; DIVERS2; URB; AG; SIZE 0.04 0.273 0.03 60 x 60-meters Water chemistry Log10(Ammonia-N) LDI-PLU; DIVERS1; DIVERS2; URB; AG; SIZE 0.00 0.439 0.00 LDI-ILD; DIVERS1; DIVERS2; URB; AG; SIZE 0.00 0.550 0.00 LDI-ISD; DIVERS1; DIVERS2; URB; AG; SIZE 0.00 0.482 0.00 Log10(NO3/NO2-N) LDI-PLU; DIVERS1; DIVERS2; URB; AG; SIZE 0.13 0.063 0.13 LDI-ILD; DIVERS1; DIVERS2; URB; AG; SIZE 0.07 0.165 0.07 LDI-ISD; DIVERS1; DIVERS2; URB; AG; SIZE 0.13 0.071 0.13 Log10(TKN) LDI-PLU; DIVERS1; DIVERS2; URB; AG; SIZE 0.00 0.456 0.00 LDI-ILD; DIVERS1; DIVERS2; URB; AG; SIZE 0.00 0.535 0.00 LDI-ISD; DIVERS1; DIVERS2; URB; AG; SIZE 0.04 0.257 0.04 Log10(TP) LDI-PLU; DIVERS1; DIVERS2; URB; AG; SIZE 0.10 0.114 0.10 LDI-ILD; DIVERS1; DIVERS2; URB; AG; SIZE 0.10 0.117 0.10 LDI-ISD; DIVERS1; DIVERS2; URB; AG; SIZE 0.11 0.090 0.11 80 x 80-meters Water chemistry Log10(Ammonia-N) LDI-PLU; DIVERS1; HETER; URB; AG; SIZE 0.00 0.517 0.00 LDI-ILD; DIVERS1; HETER; URB; AG; SIZE 0.00 0.516 0.00 LDI-ISD; DIVERS1; HETER; URB; AG; SIZE 0.00 0.227 0.00 Log10(NO3/NO2-N) LDI-PLU; DIVERS1; HETER; URB; AG; SIZE 0.14 0.058 0.14 LDI-ILD; DIVERS1; HETER; URB; AG; SIZE 0.13 0.710 0.13 LDI-ISD; DIVERS1; HETER; URB; AG; SIZE 0.14 0.059 0.14 320

PAGE 321

Table O-4. Continued. Dependent Variable Independent Variables R2 (adj) p R2 Log10(TKN) LDI-PLU; DIVERS1; HETER; URB; AG; SIZE 0.07 0.175 0.07 LDI-ILD; DIVERS1; HETER; URB; AG; SIZE 0.04 0.267 0.04 LDI-ISD; DIVERS1; HETER; URB; AG; SIZE 0.04 0.125 0.04 Log10(TP) LDI-PLU; DIVERS1; HETER; URB; AG; SIZE 0.12 0.085 0.12 LDI-ILD; DIVERS1; HETER; URB; AG; SIZE 0.12 0.084 0.12 LDI-ISD; DIVERS1; HETER; URB; AG; SIZE 0.11 0.095 0.11 321

PAGE 322

Table O-5. Multiple regression models at three spa tial extents for the sample lakes: coefficients of determination, probabi lities, and change in th e amount of variability ( R2) in indicators of ecosystems condition ( level of 0.05). Dependent Variable Independent Variables R2 (adj) p R2 100-meter Water chemistry Log10(Ammonia-N) LDI-PLU; DIVERS1; DIVERS2; URB; SHAPE; FOR 0.00 0.596 0.00 LDI-ILD; DIVERS1; DIVERS2; URB; SHAPE; FOR 0.00 0.578 0.00 LDI-ISD; DIVERS1; DIVERS2; URB; SHAPE; FOR 0.00 0.576 0.00 Log10(NO3/NO2-N) LDI-PLU; DIVERS1; DIVERS2; URB; SHAPE; FOR 0.08 0.174 0.08 LDI-ILD; DIVERS1; DIVERS2; URB; SHAPE; FOR 0.11 0.108 0.11 LDI-ISD; DIVERS1; DIVERS2; URB; SHAPE; FOR 0.13 0.080 0.13 Log10(TKN) LDI-PLU; DIVERS1; DIVERS2; URB; SHAPE; FOR 0.00 0.554 0.00 LDI-ILD; DIVERS1; DIVERS2; URB; SHAPE; FOR 0.00 0.604 0.00 LDI-ISD; DIVERS1; DIVERS2; URB; SHAPE; FOR 0.00 0.660 0.00 Log10(TN) LDI-PLU; DIVERS1; DIVERS2; URB; SHAPE; FOR 0.00 0.647 0.00 LDI-ILD; DIVERS1; DIVERS2; URB; SHAPE; FOR 0.00 0.669 0.00 LDI-ISD; DIVERS1; DIVERS2; URB; SHAPE; FOR 0.00 0.655 0.00 Log10(TP) LDI-PLU; DIVERS1; DIVERS2; URB; SHAPE; FOR 0.07 0.189 0.07 LDI-ILD; DIVERS1; DIVERS2; URB; SHAPE; FOR 0.08 0.169 0.08 LDI-ISD; DIVERS1; DIVERS2; URB; SHAPE; FOR 0.09 0.159 0.09 400-meter Water chemistry Log10(Ammonia-N) LDI-PLU; DIVERS1; DIVERS2; WET; HETER; AG 0.03 0.315 0.03 LDI-ILD; DIVERS1; DIVERS2; WET; HETER; AG 0.03 0.308 0.03 LDI-ISD; DIVERS1; DIVERS2; WET; HETER; AG 0.03 0.311 0.03 Log10(NO3/NO2-N) LDI-PLU; DIVERS1; DIVERS2; WET; HETER; AG 0.00 0.454 0.00 LDI-ILD; DIVERS1; DIVERS2; URB; HETER; AG 0.01 0.392 0.01 LDI-ISD; DIVERS1; DIVERS2; URB; HETER; AG 0.01 0.399 0.01 Watershed Water chemistry Log10(Ammonia-N) LDI-PLU; DIVERS1; DIVERS2; URB; HETER; AG 0.10 0.120 0.10 LDI-ILD; DIVERS1; DIVERS2; URB; HETER; AG 0.06 0.220 0.06 LDI-ISD DIVERS1; DIVERS2; URB; HETER; AG 0.06 0.227 0.06 Log10(NO3/NO2-N) LDI-ILD; DIVERS1; DIVERS2; URB; HETER; AG 0.06 0.219 0.06 LDI-ISD DIVERS1; DIVERS2; URB; HETER; AG 0.07 0.197 0.07 Log10(TKN) LDI-PLU; DIVERS1; DIVERS2; URB; HETER; AG 0.13 0.091 0.13 LDI-ILD; DIVERS1; DIVERS2; URB; HETER; AG 0.10 0.132 0.10 LDI-ISD DIVERS1; DIVERS2; URB; HETER; AG 0.10 0.125 0.10 Log10(TP) LDI-PLU; DIVERS1; DIVERS2; URB; HETER; AG 0.13 0.083 0.13 LDI-ILD; DIVERS1; DIVERS2; URB; HETER; AG 0.11 0.111 0.11 LDI-ISD DIVERS1; DIVERS2; URB; HETER; AG 0.12 0.106 0.12 322

PAGE 323

LIST OF REFERENCES Abrahamson, W. G. and D. C. Hartnett. 1990. Pine flatwoods and dry prairies. p. 103. In R. L. Meyers and J. J. Ewel (eds.) Ecosystems of Florida. University of Central Florida Press, Orlando, FL, USA. Adams, S. M. 2002. Biological indicators of aquatic ecosystem stress: introduction and overview. p. 1. In S. M. Adams (ed.) Biological indi cators of aquatic ecosystem stress. American Fisheries Society, Bethesda, MD, USA. Adamus, P. R. 1996. Bioindicators for assessi ng ecological integrity of prairie wetlands. EPA600-R-96/082. National Health and Envi ronmental Effects Laboratory, Western Ecology Division, United States Environmenta l Protection Agency. Corvallis, OR, USA. Ainslie, W. B. 2002. Chapter 20: forested wetlands. p. 479. In D. N. Wear and J. G. Greis (eds.) Southern forest resource assessment Gen. Tech Rep. SRS-53. U.S. Department of Agriculture, Forest Service, Southern Research Station, Asheville, NC, USA. Allan, J. D. 1995. Stream ecology: structure and function of running waters. Kluwer Academic Publishers, Dordrecht, The Netherlands. Allan, J. D. 2004. Landscapes and riverscapes: the influence of land use on stream ecosystems. Annual Review of Ecology, Evolution, and Systematics 35:257. Allan, J. D., D. L. Erickson, and J. Fay. 1997. The influence of catchment land use on stream integrity across multiple spatial scales. Freshwater Biology 37:149. Allan, J. D. and A. S. Flecker. 1993. Biodiversit y conservation in running waters. BioScience 43:32. Allan, J. D. and L. Johnson. 1997. Catchment-scale analysis of aquatic eco systems. Freshwater Biology 37:107. Arbuckle, K. E. and J. A. Downing. 2002. Fres hwater mussel abundance and species richness: GIS relationships with watershed land use a nd geology. Canadian Journal of Fisheries and Aquatic Sciences 59:310. Austin, J. E., T. K. Buhl, G. R. Guntenspergen, W. Norling, and H. T. Sklebar. 2001. Duck populations as indicators of landscape c ondition in the Prairie Pothole Region. Environmental Monitoring and Assessment 69:29. Barbour, M. T., J. Gerristen, G. E. Griffith, R. Frydenborg, E. McCarron, J. S. White, and M. L. Bastian. 1996a. A framework for biological criteria for Florida streams using benthic macroinvertebrates. Journal of the Nort h American Benthological Society 15:185. 323

PAGE 324

Barbour, M. T., J. Gerristen, and J. S. Wh ite. 1996b. Development of the Stream Condition Index (SCI) for Florida. A Report to the Flor ida Department of Environmental Protection, Stormwater and Nonpoint Source Management Section. Tetra Tech, Inc. Owing Mills, MD, USA. Bardi, E. and M. T. Brown. 2001. Emergy evaluati on of ecosystems: a basis for environmental decision making. p. 81. In M. T. Brown (ed.) Emergy synthesis: theory and applications of the emergy methodolo gy. The Center for Environmental Policy, Department of Environmental Engineering Scie nces, University of Florida, Gainesville, FL, USA. Baron, J. S., N. L. Poff, P. L. Angermeier, C. N. Dahm, P. H. Gleick, N. G. Hairston, R. B. Jackson, C. A. Johnston, B. D. Richter, and A. D. Steinman. 2003. Sustaining healthy freshwater ecosystems. Issues in Ecology, Nu mber 10, Ecological Society of America, Washington, DC, USA. Berk, R. A. 2004. Regression analysis: a c onstructive critique. Advanced Quantitative Techniques in the Social Sciences Series 11. Sage Publications, Thousand Oaks, CA, USA. Bian, L. and R. Butler. 1999. Comparing effects of aggregation methods on statistical and spatial properties of simulated spatial data. Photogr ammetric Engineering and Remote Sensing 65:73. Blocksom, K. A., J. P. Kurtenbach, D. J. Kl emm, F. A. Fulk, and S. M. Cormier. 2002. Development and evaluation of the lake macroi nvertebrate integrity index (LMII) for New Jersey lakes and reservoirs. Enviro nmental Monitoring and Assessment 77:311. Boggess C. F. 1994. Biogeoeconomics of phosphorus in a Florida watershed. Ph.D. Dissertation, University of Florida, Gainesville, FL, USA. Brandt-Williams, S. 1999. Evaluation of waters hed control of two central Florida lakes: Newnans Lake and Lake Weir. Ph.D. Dissertatio n, University of Florida, Gainesville, FL, USA. Brandt-Williams, S. 2001. Handbook of emergy evaluation. Folio #4, Emergy of Florida agriculture. Center for Environmental Po licy, Environmental Engineering Sciences, University of Florida, Gainesville, FL, USA. Brenner, M., M. W. Binford, and E. S. Deeway. 1990. Lakes. p. 364. In R. L. Meyers and J. J. Ewel (eds.) Ecosystems of Florida. Univ ersity of Central Florida Press, Orlando, FL. USA. Brett, M. T., G. B. Arhonditsis, S. E. Mueller, D. M. Hartley, J. D. Frodge, and D. E. Funke. 2005. Non-point source impacts on stream nutrient concentrations along a forest to urban gradient. Environmental Management 35:330. Brnmark, C. and L. A. Hansson. 1998. The bi ology of lakes and ponds. Oxford University Press, New York, NY, USA. 324

PAGE 325

Brown, M. T. 1980. Energy basis for hierarchie s in urban and regional landscapes. Ph. D. Dissertation, University of Fl orida, Gainesville, FL, USA. Brown, M. T. (ed.). 2000. Emergy synthesis: theory and applications of the emergy methodology. Proceedings of the First Biennial Emergy Analysis Research Conference, Gainesville, FL, September 2001. The Center fo r Environmental Policy, University of Florida, Gainesville, FL, USA. Brown, M. T. (ed.). 2003a. Emergy synthesis 2: theory and applica tions of the emergy methodology. Proceedings of the Second Bienni al Emergy Conference, Gainesville, FL, December 2003. The Center for Environmental Po licy, University of Florida, Gainesville, FL, USA. Brown, M. T. 2003b. Spatial modeling and environmental loading. p. 35. In M. T. Brown (ed.) Emergy synthesis 2: theory and applic ations of the emergy methodology. Proceedings of the Second Biennial Emergy Analysis Rese arch Conference, Gainesville, FL, December 2003. The Center for Environmental Policy, Univer sity of Florida, Gainesville, FL, USA. Brown, M. T. (ed.). 2005. Emergy synthesis 3: theory and applications of the emergy methodology. Proceedings of the 3rd Biennial Emergy Conference, Gainesville, FL, November 2005. The Center for Environmental Policy, University of Florida, Gainesville, FL, USA. Brown, M. T. and E. Bardi. 2001. Handbook of emergy evaluation. Folio #3, Emergy of ecosystems. Center for Environmental Po licy, Environmental Engineering Sciences, University of Florida, Gainesville. FL, USA. Brown, M. T., P. Green, A. Gonzalez, and J. Vanegas. 1992. Emergy analysis perspectives, public policy options, and development guidelines for the coastal zone of Nayarit, Mexico. Report to The Cousteau Society and the Government of Nayarit, Mexico. Volume 2: Emergy analysis and public policy portions. Center of Wetlands and Water Resources, University of Florida, Gainesville, FL, USA Brown, M. T. and M. B. Vivas. 2005. Landscape Development Intensity Index. Environmental Monitoring and Assessment 101:289. Brown, M. T. and M. B. Vivas. 2006. A new La ndscape Development Intensity Index based on areal empower density. Paper presented at Em ergy Synthesis: Theory, Applications and Methodologies, Fourth Bienni al Emergy Research Conference, Gainesville, FL, USA. Cain, D. H., K. Riitters, and K. Orvis. 1997. A multi-scale analysis of landscape statistics. Landscape Ecology 12:199. Canfield, D. E. and M. V. Hoyer. 1988. Regiona l geology and the chemi cal and trophic state characteristics of Florida lakes. La ke and Research Management 1:21. 325

PAGE 326

Carpenter, S. R., N. F. Caraco, D. L. Correll, R. W. Howarth, A. N. Sharpley, and V. H. Smith. 1998. Nonpoint pollution of surface waters with phosphorus and nitrogen. Ecological Applications 8:559. Carpenter, S., T. Frost, L. Persson, M. Po wer, and D. Soto. 1996. Freshwater ecosystems: linkages of complexity and processes. p. 299. In H. A. Mooney (ed.) Functional roles of biodiversity: a global perspective. J ohn Wiley & Sons, Inc., New York, NY, USA. Cattaneo, A. and Y. T. Prairie. 1995. Temporal va riability in the chemical characteristics along the Riviere de lAnchigan: how many samples ar e necessary to describe stream chemistry? Canadian Journal of Fisherie s and Aquatic Sciences 52:828. Chen, E. and J. F. Gerber. 1990. Climate. p. 11-34. In R. L. Meyers and J. J. Ewel (eds.) Ecosystems of Florida. University of Central Florida Press, Orlando, FL, USA. Christianson, R. A. 1984. Energy perspectives on a tropical forest/plantation system at Jari, Brazil. Masters Thesis, University of Florida, Gainesville, FL, USA. Cifaldi, R. L., J. D. Allan, J. D. Duh, and D. G. Brown. 2004. Spatial patterns in land cover of exurbanizing watersheds in southeastern Michigan. Landscape and Urban Planning 66:107. Cohen, M. J., S. Carstenn, and C. R. Lane. 2004. Floristic quality indices for biotic assessment of depressional marsh condition in Florida. Ecological Applications 14:784. Comeleo, R. L., J. F. Paul, P. V. August, C. Baker, S. S. Hale, and R. W. Latimer. 1996. Relationships between watershed stressors a nd sediment contamination in Chesapeake Bay estuaries. Landscape Ecology 11:307. Correll, D. L., T. E. Jordan, and D. E. Weller. 1992. Nutrient flux in a landscape: effects of coastal land use and terre strial community mosaic on nutrien t transport to coastal waters. Estuaries 15:431. Crosbie, B. and P. Chow-Fraser. 1999. Percenta ge land use in the watershed determines the water and sediment quality of 22 marshes in the Great Lake Basin. Canadian Journal of Fisheries and Aquatic Sciences 56:1781. Cyr, H., P. J. Dillon, and J. E. Parker. 2004. The te mporal scaling of enviro nmental variability in rivers and lakes. p. 201-213. In L. Seuront and P. G. Strutton (eds.) Handbook of scaling methods in aquatic ecology: measurements, analysis, simulation. CRC Press, Boca Raton, FL, USA. Devito, K. J., I. F. Creed, R. L. Rothwell, and E. E. Prepas. 2000. Landscape controls on phosphorus loading to boreal lakes: implications for the potential impacts of forest harvesting. Canadian Journal of Fish eries and Aquatic Sciences 57:1977. Dewey, E. S. 1988. Estimation of downward le akage from Florida lakes. Limnology and Oceanography 33:1308. 326

PAGE 327

Diamond, C. 1984. Energy basis for the regional or ganization of the Mississippi River Basin. Masters Thesis, University of Florida, Gainesville, FL, USA. Doherty, S. J. 1995. Emergy evaluations of and li mits to forest production. Ph.D. Dissertation, University of Florida, Gainesville, FL, USA. Dytham, C. 1999. Choosing and using statistics: a biologists guide. Blackwell Science, Oxford, UK. Euliss, N. H. Jr. and D. M. Mushet. 1999. Infl uence of agriculture on aquatic invertebrate communities of temporary wetlands in the Prai rie Pothole Region of North Dakota, USA. Wetlands 19:578. Ewel. J. J. 1990a. Introduction. p. 3. In R. L. Meyers and J. J. Ewel (eds.) Ecosystems of Florida. University of Central Florida Press, Orlando, FL, USA. Ewel, K. C. 1990b. Swamps. p. 281. In R. L. Meyers and J. J. Ewel (eds.) Ecosystems of Florida. University of Central Florida Press, Orlando, FL, USA. Fernald, E. A. and E. D. Purdum (eds.). 1998. Wa ter Resources Atlas of Florida. Institute of Science and Public Affairs, Florida State University, Tallahassee, FL, USA. FDACS (Florida Department of Agriculture a nd Consumer Services). 2005. Floridas forest resource plan: setting the course for 2030. Florida Division of Forestry, Florida Department of Agriculture and Consumer Serv ices, Tallahassee, FL, USA. Available at www.fl-dof.com Accessed on 06/2006. FDEP (Florida Department of Environmenta l Protection). 2004. Inte grated water quality assessment for Florida: 2004 305(b) Report and 303(d) List Update. Bureau of Watershed Management, Division of Water Resour ce Management, Florida Department of Environmental Protection, Ta llahassee, FL, USA. 190 p. FDEP (Florida Department of Environmen tal Protection ). 2005. Bioassessment: current projects. Florida Department of Envir onmental Protection, Tallahassee, FL, USA. Available at http://www.dep.state.fl.us/water/bioassess/currproj.htm Accessed on 10/2005. FDOT (Florida Department of Transportation). 1999. Florida Land Use and Cover and Forms Classification System: handbook. Third Editi on. Thematic Mapping Section, Surveying and Mapping Office, Tallahassee, FL, USA. Findlay, S., J. M. Quinn, C. W. Hickey, G. Burrell, and M. Downes. 2001. Effects of land use and riparian flowpath on delivery of dissolved organic carbon to streams. Limnology and Oceanography 46:345. Fitzpatrick, F. A., B. C. Scudder, B. N. Lenz and D. J. Sullivan. 2001. Effects of multi-scale environmental characteristics on agricultural stre am biota in eastern Wisconsin. Journal of the American Water Res ource Association 37:1489. 327

PAGE 328

Fore, L. S. 2004. Development and testing of biomonitoring tools for macroinvertebrates in Florida streams. Statistical Design, Seat tle, Washington. A report for the Florida Department of Environmental Pr otection, Tallahassee, FL, USA. Fore, L. S. 2005. Assessing the biological condition of Florida lakes: development of the Lake Vegetation Index (LVI). Statistical Design, Se attle, Washington. A re port for the Florida Department of Environmental Prot ection, Tallahassee, FL, USA. Frissel, C. A., W. J. Liss, C. E. Warren, and M. D. Hurley. 1986. A hierarchical framework for stream habitat classification: viewing streams in a watershed context. Environmental Management 10:199. Galatowitsch, S. M., D. C. Whited, R. Lehtin en, J. Husveth, and K. Schik. 2000. The vegetation of wet meadows in relation to their land use. Environmenta l Monitoring and Assessment 60:121. Galatowitsch, S. M., D. C. Whited, and J. R. Tester. 1999. Development of community metrics to evaluate recovery of Minnesota wetlands Journal of Aquatic Ecosystem Stress and Recovery 6:217. Gergel, S. E., M. G. Turner, and T. K. Kratz. 1999. Dissolved organic carbon as an indicator of the scale of watershed influence on lakes a nd rivers. Ecological Applications 9:1377. Gergel, S. E., M. G. Turner, J. M. Miller, J. M. Melack, and E. H. Stanley. 2002. Landscape indicators of human impacts on riveri ne systems. Aquatic Sciences 64:118. Gerritsen, J, B. Jessup, E. W. Leppo, and J. White. 2000. Development of the Lake Condition Index for Florida. Tetra Tech, Inc. and Flor ida Department of Environmental Protection, Owing Mills, MD, USA. Gesch, D., M. Oimoen, S. Greenlee, C. Nels on, M. Steuck, and D. Tyler. 2002. The National Elevation Dataset. Photogrammetric Engineering and Remote Sensing 68:5. Giller, P. S. and B. Malmqvist. 1998. The biolo gy of streams and rivers. Oxford University Press, New York, NY, USA. Griffith, G. E., D. E. Canfield, Jr., C. A. Horsburgh, J. M. Omernik, and S. H. Azevedo. 1997. Florida lake regions. Environmental Res earch Laboratory, Office of Research and Development, U.S. Environmental Protection Agency, Corvallis, OR, USA. Griffith, G. E., J. M. Omernik, C. M. Rohm and S. M. Pierson. 1994. Florida Regionalization Project. Environmental Research Laboratory, Office of Research and Development, U.S. Environmental Protection Agency, Corvallis, OR, USA. Griffith, J. A. 2002. Geographic techniques and recent applications of remote sensing to landscape-water quality studies. Wa ter, Air, and Soil Pollution 138:181. 328

PAGE 329

Griffith, J. A., E. A. Martinko, and K. P. Pri ce. 2000. Landscape structure analysis of Kansas at three scales. Landscape and Urban Planning 52:45. Griffith, J. A., E. A. Martinko, J. L. Whistler, and K. P. Price. 2002. Preliminary comparison of landscape pattern-Normalized Difference Ve getation Index (NDVI) relationships to Central Plain stream conditions. Journa l of Environmental Quality 31:846. Guerry, A. D. and M. L. Hunter. 2002. Amphibian distribution in a land scape of forests and agriculture: an examination of landscape composition and co nfiguration. Conservation Biology 16:745. Gustafson, E. J. 1998. Quantifying landscape spatial pattern: what is the state of the art? Ecosystems 1:143. Harper, H. H. 1994. Stormwater loading rate parameters for Central and South Florida. Environmental Research & Design, Inc. Hay, G. J., D. J. Marceau, P. Dube, and A. Bouchard. 2001. A multiscale framework for landscape analysis: object-specific anal ysis and upscaling. Landscape Ecology 16:471 490. Haynes, H. B. N. 1975. The stream and its valle y. Verhandlugen der Internationalen Vereinigung fr Theoretische und Angewandte Limnologie 19:1. Houlahan, J. E. and C. S. Findlay. 2004. Estima ting the critical distance at which adjacent land-use degrades wetland water and se diment quality. Landscape Ecology 19:677. Houlahan, J. E., P. A. Keddy, K. Makkay, and C. S. Findlay. 2006. The effects of adjacent land use on wetlands species richness and community composition. Wetlands 26:79. Howington, T. M. 1999. A spatial analysis of emergy of an interna tionally shared drainage basin and the implications for policy decisions. P h.D. Dissertation, Univ ersity of Florida, Gainesville, FL, USA. Hunsaker, C. T. and D. A. Levine. 1995. Hierarchi cal approaches to the study of water quality in rivers. BioScience 45:193. Johnes, P., B. Moss, and G. Phillips. 1996. The determination of total nitrogen and total phosphorus concentrations in freshwater from land use, stock headage and population data: testing of a model for use in conservation and water quality management. Freshwater Biology 36:451. Johnson, L. B. and S. H. Gage. 1997. Landscape approaches to the analysis of aquatic ecosystems. Freshwater Biology 37:113. Johnson, L. B., C. Richards, G. E. Host, and J. W. Arthur. 1997. Landscape influences on water chemistry in Midwestern stream ecosystems. Freshwater Biology 37:193. 329

PAGE 330

Jones, E. B. D., G. S. Helfman, J. O. Harper, and P. V. Bolstad. 1999. Effects of riparian forest removal on fish assemblages in southern Appalachian streams. Conservation Biology 13:1454-1465. Jones, K. B., A. C. Neale, M. S. Nash, R. D. Van Remortel, J. D. Wickha m, K. H. Riitters, and R. O. Neill. 2001. Predicting nutrient and sediment loading to streams from landscape metrics: a multiple watershed study form the United States Mid-Atlantic Region. Landscape Ecology 16:301. Kangas, P. C. 1983. Landforms: succession and recl amation. Ph.D. Dissertation, University of Florida, Gainesville, FL, USA. Kangas, P. C. 2002. Handbook of emergy evaluati on. Folio #5, Emergy of landforms. Center for Environmental Policy, Environmental Engine ering Sciences, University of Florida, Gainesville. FL, USA. Karr, J. R. 1994. Landscapes and manageme nt of ecological integrity. p. 229-251. In K. C. Kim and R. D. Weaver (eds.) Bi odiversity and landscapes: a para dox of humanity. Cambridge University Press, Cambridge, UK. Karr, J. R. 1999. Defining and measuring ri ver health. Freshwater Biology 41:221-234. Karr, J. R. and E. W. Chu. 1999. Restoring life in running waters. Island Press, Washington, DC, USA. Karr, J. R. and E. W. Chu. 2000. Sustaining liv ing rivers. Hydrobiolog ia 422/423:1. Karr, J. R. and I. J. Schlosser. 1978. Water resources and the land-water interface. Science 201:229. Kearns, F. R., N. M. Kelly, J. L. Carter, and V. H. Resh. 2005. A method for the use of landscape metrics in freshwater research a nd management. Landscape Ecology 20:113. King, A. W. 1993. Considerations of scale and hierarchy. p. 19-45. In S. Woodley, J. M. Kay, and G. Francis (eds.) Ecological integrity and the manageme nt of ecosystems. St. Lucie Press, Delray Beach, FL, USA. King, R. S., M. E. Baker, D. F. Whigham, D. E. Weller, T. E. Jordan, P. F. Kazyak, and M. K. Hurd. 2005. Spatial considerations for lin king watershed land cover to ecological indicators in streams. Ec ological Applications 15:137. Kost, J., K. Verdin, B. Worstell, and G. Kell y. 2002. Methods and tools for the development of hydrologically conditioned elevation data and derivatives for nationa l applications, 2nd Annual Conference on Watershed Modeli ng, Las Vegas, NV. Available at http://edna.usgs.gov/. Accessed on 07/2007. 330

PAGE 331

Knutson, M. G., J. R. Sauer, D. A. Olsen, M. J Mossman, L. M. Hemesath, and M. J. Lannoo. 1999. Effects of landscape composition and wetland fragmentation on frog and toad abundance and species richness in Iowa a nd Wisconsin, U.S.A. Conservation Biology 13:1437. Lambert, J. D. 1999. A spatial emergy model fo r Alachua County, Florida. Ph.D. Dissertation, University of Florida, Gainesville, FL, USA. Lammert, M. and J. D. Allan. 1999. Assessing biotic integrity of streams: effects of scale in measuring the influence of land use/c over and habitat structure on fish and macroinvertebrates. Environmental Management 23:257. Lane, C. R. 2000. Proposed ecological regions fo r freshwater wetlands of Florida. Masters Thesis, University of Florid a, Gainesville, FL, USA. Lane, C. R. 2003. Biological indicators of wetland condition for isolated depressional herbaceous wetlands in Florida. Ph.D. Di ssertation, University of Florida, Gainesville, FL, USA. Lane, C. R. and M. T. Brown. 2006. Energy-based land use predictors of proximal factors and benthic diatom composition in Florida freshwat er marshes. Environmental Monitoring and Assessment 117:433. Levin, S. A. 1992. The problem of pattern and scale in ecology. Ecology 73:1943. Lewis, P. A., D. J. Klemm, and W. T. Thoe ny. 2001. Perspectives on use of a multimetric lake bioassessment integrity index using benthic macroinvertebrates. Northeastern Naturalist 8:233. Li, H. and J. Wu. 2004. Use and misuse of landscape indices. Landscape Ecology 19:389. Likens, G. E., F. H. Bormann, N. M. Johnson, D. W. Fisher, and R. S. Pierce. 1970. Effects of forest cutting and herbicide treatment on nutrient budgets in the Hubbard Brook watershed-ecosystem. Ecological Monographs 40:23. Liu, A. J. and G. N. Cameron. 2001. Analysis of landscape patterns in coastal wetlands of Galveston Bay, TX, USA. Landscape Ecology 16:581. Lopez, R. D., C. B. Davis, and M. S. Fennessy. 2002. Ecological relationships between landscape change and plant guilds in de pressional wetlands. Landscape Ecology 00:1. MacArthur, R. H. and R. Levins. 1964. Co mpetition, habitat selection, and character displacement in a patchy environment. Proceed ings of the National Academy of Sciences of the USA 51:1207. Mack, J. J. 2006. Landscape as a predictor of we tland condition: an evaluation of the landscape development index (LDI) with a large reference wetland dataset from Ohio. Environmental Monitoring and Assessment 120:221-241. 331

PAGE 332

McCune, B. and J. B. Grace. 2002. Analysis of ecological communities. MjM Software Design, Gleneden Beach, OR, USA. McCune, B. and M. J. Mefford. 1999. PC-ORD: multiv ariate analysis of ecological data. Version 4.41, MjM Software, Gleneden Beach, OR, USA. McDonnell, M. J. and S. T. A. Pickett. 1990. Ec osystem structure and function along urban-rural gradients: an unexploited opport unity for ecology. Ecology 71:1232. McDowell, R. W., B. J. F. Biggs, A. N., Aharpley, and L. Nguyen. 2004. Connecting phosphorus loss from agricultural landscapes to surf ace water quality. Chemistry and Ecology 20:1. McGarigal, K., S. A. Cushman, M. C. Neel, and E. Ene. 2002. FRAGSTATS: Spatial Pattern Analysis Program for Categorical Maps. Co mputer software program produced by the authors at the University of Massachusetts, Amherst, MA, USA. Available at www.umass.edu/landeco/researc h/fragstats/fr agstats.html Accessed on 09/2005. McGarigal, K. and B. J. Marks. 1995. FRAGST ATS: spatial pattern analysis program for quantifying landscape structure. Genera l Technical Report PNW-GTR-351. Pacific Northwest Research Station, Forest Service, Department of Agriculture, Portland, OR, USA. McGrane, G. 1994. An emergy evaluation of pers onal transportation alternatives. Masters Thesis, University of Florid a, Gainesville, FL, USA. McMahon, G. and T. F. Cuffney. 2000. Quantifying urban intensity in dr ainage basins fro assessing stream ecological condition. Journal of the American Water Resource Association 36:1247. Mensing, D. M., S. M. Galatowitsch, and J. R. Tester. 1998. Anthropogenic effects on the biodiversity of riparian wetlands of a nor thern temperate landscape. Journal of Environmental Management 53:349. Miller, J. N., R. P. Brooks, and M. J. Croonqui st. 1997. Effects of landscape pattern on biotic communities. Landscape Ecology 12:137. Minitab Release 14 Statistical Software. 2003. St ate College. PA, USA. Available at www.minitab.com Montgomery, D. C. and E. A. Peck. 1982. Introduc tion to linear regression analysis. John Wiley & Sons, Inc., New York, NY, USA. Morley, S. H. and J. R. Karr. 2002. Assessing and restoring the health of urban streams in the Puget Sound Basin. Conservation Biology 16:1498. Mulholland, P. J., J. N. Houser, and K. O. Maloney. 2005. Stream diurnal dissolved oxygen profiles as indicators of in-stream metabolism and disturbance effects: Fort Benning as a case study. Ecological Indicators 5:243. 332

PAGE 333

Naiman, R. J. and H. Decamps. 1997. The ecology of interfaces: riparian zones. Annual Review of Ecology and Systematics 28:621-658. Naiman, R. J., J. J. Magnuson, D. M. McKni ght, and J. A. Stanford. 1995. The freshwater imperative: a research agenda. Island Press, Washington, DC, USA. NRC (National Research Council). 1999. New stra tegies for America's watersheds. Committee on Watershed Management, Water Scie nce and Technology Board, Commission on Geosciences, Environment, and Resources, Na tional Research Council. National Academy Press, Washington, DC, USA. Nordlie, F. G. 1990. Rivers and springs. p. 392. In R. L. Meyers and J. J. Ewel (eds.) Ecosystems of Florida. University of Central Florida Press, Orlando, FL, USA. Odum, H. T. 1994. Ecological and general sy stems: an introduction to systems ecology. University Press of Colorado, Niwot, CO, USA. Odum, H. T. 1996. Environmental accounting: emergy and environmental decision-making. John Wiley & Sons, Inc., New York, NY, USA. Odum, H. T. 2000. Handbook of emergy evaluation. Folio #2, Emergy of Global Processes. Center for Environmental Policy, Environmen tal Engineering Sciences, University of Florida. Gainesville, FL, USA. Odum H. T. and J. E. Arding. 1991. Emergy analysis of shrimp mariculture in Ecuador. Working Paper. Prepared for the Coastal Resour ce Center, University of Rhode Island. Environmental Engineering Sciences and Cent er for Wetlands, University of Florida, Gainesville, FL, USA. Odum, H. T., M. T. Brown, and S. Bra ndt-Williams. 2000a. Handbook of emergy evaluation. Folio #1, Introduction and Global Budget. Center for Environmental Policy, Environmental Engineering Sciences, Univer sity of Florida. Gainesville, FL, USA. Odum, H. T., M. T. Brown, and R. A. Ch ristianson. 1986. Energy systems overview of the Amazon basin. Report to the Cousteau So ciety. CFW Publication #86-1, Center for Wetlands, University of Florid a, Gainesville, FL, USA. Odum, H. T., C. Diamond, and M. T. Brown. 1987. Energy systems overview of the Mississippi River Basin. Report to the Cousteau So ciety, CFW Publication #87-1. Center for Wetlands, University of Florid a, Gainesville, FL, USA. Odum, H. T., S. J. Doherty, F. N. Scatena, and P. A. Kharecha. 2000b. Emergy evaluation of reforestation alternatives in Puer to Rico. Forest Science 46:521. Odum, H. T. and E. C. Odum (eds.). 1987. Ecol ogy and economy: emergy analysis and public policy in Texas. Lyndon B. Johnson School of Public Affairs, Policy Research Project Report Number 78, The University of Texas, Austin, TX, USA. 333

PAGE 334

Odum, H. T., E. C. Odum, and M. T. Brown. 1998a Environment and society in Florida. Lewis Publishers, Boca Raton, FL, USA. Odum, H. T., S. Romitelli, and R. Tighe. 1998b. Evaluation overview of the Cache River and the Black Swamp in Arkansas. Final Report on Contract #DACW39-94-K-0300 Energy Systems Perspectives for Cumulative Impact s Assessment between Waterways Experiment Station, U.S. Dept. of the Army, Vicksburg, Mi ss. and University of Florida. Center for Environmental Policy, Environmental Engine ering Sciences, University of Florida, Gainesville, FL, USA. Oksanen, J. and T. Sarjakoski. 2005. Error propag ation analysis of DEM-based drainage basin delineation. International Journal of Remote Sensing 26:3085-3102. Omernik, J. M. 1977. Nonpoint source-stream nutrient level relationships : a nation wide study. EPA-600/3-77-105. Corvallis Environmental Resear ch Laboratory, Office of Research and Development, United States Environmental Protection Agency, Corvallis, OR, USA. Omernik, J. M. 1987. Ecoregions of the contermi nous United States. Annals of the Association of American Geographers 77:118. O'Neill, R. V., C. T. Hunsaker, K. B. Jones, K. H. Riitters, J. D. Wickham, P. M. Schwartz, I. A. Goodman, B. L. Jackson, and W. S. Bailla rgeon. 1997. Monitoring environmental quality at the landscape level. BioScience 47:513. O'Neill, R. V., C. T. Hunsaker, S. P. Timms, B. L. Jackson, K. B. Jones, K. H. Riitters, and J. D. Wickham. 1996. Scale problems in reporting la ndscape pattern at the regional scale. Landscape Ecology 11:169. ONeill, R. V., J. R. Krummel, R. H. Gardner, G. Sugihara, B. Jackson, D. L. DeAngelis, B. T. Milne, M. G. Turner, B. Zygmunt, S. Christensen, V. H. Dale, and R. L. Graham. 1988. Indices of landscape patter n. Landscape Ecology 1:153. O'Neill, R. V., K. H. Riitters, J. D. Wickham, and K. B. Jones. 1999. Landscape pattern metrics and regional assessment. Ecosystem Health 5:225. Orrell, J. J. 1998. Cross scale comparison of plant production and dive rsity. Masters Thesis, University of Florida, Gainesville, FL, USA. Ortega, E., J. F. Queiroz, C. E. Boyd, and J. M. G. Ferraz. 2000. Emergy analysis of channel catfish farming in Alabama, USA. p. 171-180. In M. T. Brown (ed.) Emergy synthesis: theory and applications of the emergy methodology. The Cent er for Environmental Policy, Department of Environmental Engineering Scie nces, University of Florida, Gainesville, FL, USA. Osborne, L. L. and M. J. Willey. 1988. Empirical relationships between land use/land cover and stream water quality in an agricultural wate rshed. Journal of Environmental Management 26:9. 334

PAGE 335

Pan, Y., R. J. Stevenson, B. H. Hill, P. R. Kaufmann, and A. T. Herlihy. 1999. Spatial patterns and ecological determinants of benthic algal assemblages in Mid-Atlantic streams, USA. Journal of Phycology 35:460. Parker, N. M. 1998. Spatial models of tota l phosphorus loading and landscape development intensity in a North Florida watershed. Masters Thesis, University of Florida, Gainesville, FL, USA. Pate, F. 1999. Bahiagrass is Floridas most dependable pasture forage. Available at http://rcrecona.ifas.ufl.edu/PRFAR14.html Accessed on 10/2001. Paul, M. J. and J. L. Meyer. 2001. Streams in the urban landscape. Annual Review of Ecology and Systematics 32:333. Paulic, M., J. Hand, and L. Lord. 1996. 1996 water qua lity assessment of the State of Florida: Section 305(b) Main Report. Bureau of Wa ter resources Protection, Division of Water Facilities, Florida Departme nt of Environmental Protec tion, Tallahassee, FL, USA. Pearson, S. M. 1993. The spatial extent and rela tive influence of landscape-level factors on wintering bird populations Landscape Ecology 8:3. Pearson, S. M. and M. G. Turn er. 1995. Winter habitat use by large ungulates following fire in northern Yellowstone National Par k. Ecological Applications 5:744. Pess, G. R., D. R. Montgomery, E. A. Steel, R. E. Bilby, B. E. Feist, and H. M. Greenberg. 2002. Landscape characteristics, land use, and coho salmon ( Oncorhynchus kisutch) abundance, Snohomish River, Wash., U.S.A. Canadian J ournal of Fisheries and Aquatic Sciences 59:613. Peterjohn, W. T. and D. L. Correll. 1984. Nutrient dynamics in an agricultural watershed: observations on the role of a riparian forest. Ecology 65:1466. Poole, G. C. and C. H. Berman. 2001. An eco logical perspective on in-stream temperature: natural heat dynamics and mechanisms of human-caused thermal degradation. Environmental Management 27:787-802. Reiss, K. C. 2004. Developing biolog ical indicators for isolated forested wetlands in Florida. Ph.D. Dissertation, University of Fl orida, Gainesville, FL, USA. Reiss, K. C. 2006. Florida Wetland Condition In dex for depressional forested wetlands. Ecological Indicators 6:337. Reiss, K. C. and M. T. Brown. 2005. The Flor ida Wetland Condition Index (FWCI): preliminary development of biological indicators for fo rested strand and floodplain wetlands. Report submitted to the Florida Department of E nvironmental Protection under contract #WM683. Howard T. Odum Center for Wetlands, Univer sity of Florida, Gainesville, FL, USA. 335

PAGE 336

Reynolds, J. E. 1999. Urban land conversion and competition for rural land use. SP 99-15. University of Florida, Institute of Food and Agricultural Sciences, Gainesville, FL, USA. Richards, C., L. B. Johnson, and G. E. Host. 1996. Landscape-scale influences on stream habitats and biota. Canadian Journal of Fi sheries and Aquatic Sciences 53:295. Riitters, K. H., R. V. ONeill, C. T. Hunsaker, J. O. Wickham, D. H. Yankee, S. P. Timmins, K. B. Jones, and B. L. Jackson. 1995. A factor analysis of landscape pattern and structure metrics. Landscape Ecology 10:23. Romitelli, M. S. 1997. Energy analysis of watersheds. Ph.D. Dissertation, University of Florida, Gainesville, FL, USA. Roth, N. E., J. D. Allan, and D. L. Ericks on. 1996. Landscape influences on stream biotic integrity assessed at multiple scales. Landscape Ecology 11:141-156. Roy, A. H., A. D. Rosemond, M. J. Paul, D. S. Leigh, and J. B. Wallace. 2003. Stream macroinvertebrate response to catchment urbanization (Georgia, USA). Freshwater Biology 48:329. Scott, M. C., G. S. Helfman, M. E. McTamma ny, E. F. Benfield, and P. V. Bolstad. 2002. Multiscale influences on physical and chem ical stream conditions across Blue Ridge landscapes. Journal of the American Water Resource Association 38:1379. Shen, W., G. D. Jenerette, J. Wu, and R. H. Gardner. 2004. Evaluating empirical scaling relations of pattern metrics with simulated landscapes. Ecography 27:459. Smith, R. A., G. E. Schwarz, and R. B. Alexa nder. 1997. Regional interpretation of water-quality monitoring data. Water Resources Research 33:2781. Snyder, C. D., J. A. Young, R. Villella, and D. P. Lemari. 2003. Influences of upland and riparian land use patterns on stream biotic integrity. Landscape Ecology 18:647. Sokal, R. R. and F. J. Rohlf. 1981. Biometry: the principals and practice of statistics in biological research. Second Edition. W. H. Freema n and Company, New York, NY, USA. Sole, A. and A. Valanzano. 1996. Di gital terrain modeling. p. 175-194. In V. P. Singh and M. Fiorentino (eds.) Geographical information systems in hydrology. Kluwer Academic Publishers, Dordrecht, The Netherlands. Soranno, P. A., S. L. Hubler, S. R. Carpenter, and R. C. Lathrop. 1996. Phosphorus loads to surface waters: a simple model to account for spatial pattern of land use. Ecological Applications 6:865. Sponseller, R. A., E. F. Benfield, and H. M. Valett. 2001. Relationships between land use, spatial scale and stream macroinvertebrate communities. Freshwater Biology 46:1409. 336

PAGE 337

337 Steinman, A. D., J. Conklin, P. J. Bohlen, a nd D. G. Uzarski. 2003. Influence of cattle grazing and pasture land use on macroinvertebrate comm unities in freshwater wetlands. Wetlands 23:877–889. Strayer, D. L., R. E. Beighley, L. C. Thom pson, S. Brooks, C. Nilsson, G. Pinay, and R. J. Naiman. 2003. Effects of land cover on stream ecosystems: roles of empirical models and scaling issues. Ecosystems 6:407–423. Surdick, A. J. 2005. Amphibian and avian species composition of forested depressional wetlands and circumjacent habitat: the influence of la nd use type and intensity. Ph.D. Dissertation, University of Florida, Gainesville, FL, USA. Tangen, B. A., M. G. Butler, and M. J. Ell. 2003. Weak correspondence between macroinvertebrate assemblages and land use in Prairie Pothole Region wetlands, USA. Wetlands 23:104–115. Tanner, G. W. and S. B. Bran tley. 1992. Woodland livestock grazing. In M. Duryea, W. Hubbard, D. McGrath, and C. Marcus (eds.) Florida’s Forest Stewardship Program. Cooperative Extension Services, Institute of F ood and Agriculture Sciences, University of Florida, Gainesville, FL, USA. Available at http://www.sfrc.ufl.edu/Ex tension/pubtxt/c1020g.htm Accessed on 07/2001. ter Braak, C. J. F. 1995. Ordination. p. 91–173. In R. H. G. Jongman, C. J. F. ter Braak, and O. F. R. van Tongeren (eds.) Data analysis in community and landscape ecology. Cambridge University Press, Cambridge, UK. Tilley, D. R. 1999. Emergy basis of forest systems. Ph.D. Dissertation, University of Florida, Gainesville, FL, USA. Townsend, C. R., S. Doledec, R. Norris, K. P eacock, and C. J. Arbuckle. 2003. The influence of scale and geography on relati onships between stream co mmunity composition and landscape variables: description and prediction. Freshwater Biology 48:768–785. Townsend, C. R., B. J. Downes, K. Peacock, and C. J. Arbuckle. 2004. Scale and the detection of land-use effects on morphology, vegetation and macroinvertebrate communities of grassland streams. Freshwater Biology 49:448–462. Townsend, C. R. and R. H. Riley. 1999. Assessment of river health: accoun ting for perturbation pathways in physical and ecological space. Freshwater Biology 41:393–405. Tufford, D. L., H. N. McKellar, and J. R. Hussey. 1998. In-stream nonpoint source nutrient prediction with land-use proximity and seas onality. Journal of Environmental Quality 27:100–111. Tufford, D. L., C. L. Samarghitan, H. N. McKella r, D. E. Porter, and J. R. Hussey. 2003. Impacts of urbanization on nutrient concen trations in small southeastern coastal streams. Journal of the American Water Res ource Association 39:301–312.

PAGE 338

338 Turner, M. G. 1989. Landscape ecology: the effect of pattern on process. Annual Review of Ecology and Systematics 20:171–197. Turner, M. G. 2005. Landscape ecology: what is th e state of the science? Annual review of Ecology, Evolution, and Systematics 36:319–344. Turner, M. G., R. H. Gardner, and R. V. O’Neill. 2001. Landscape ecology in theory and practice: patterns and process. Springer-V erlag New York, Inc., New York, NY, USA. Turner, M. G., R. V. O’Neill, R. H. Gardner, and B. T. Milne. 1989. Effects of changing spatial scale on the analysis of landscap e pattern. Landscape Ecology 3:153–162. United States Census Bureau. 2005. Population projections. Population Projections Branch, Population Division, U.S. Census Burea u, Washington DC, USA. Available at http://www.census.gov Accessed on 06/2006. USDA (United States Department of Agricult ure). 2002. Census of Agriculture State Data. USDA, National Agricultural Statistics Serv ice, Washington DC, USA. Available at www.nass.usda.gov Accessed on 08/2005. USDA (United States Department of Agriculture) 2005. State fact sheets: Fl orida. United States Department of Agriculture, Wash ington DC, USA. Available at http://www.nass.usda.gov. Accessed on 06/2006. USEPA (United States Environmental Protec tion Agency). 1994. Landscape monitoring and assessment research plan. EPA 620/R-94/009. Office of Research and Development, Washington DC, USA. USEPA (United States Environmental Prot ection Agency). 1996. Watershed approach framework. EPA840-S-96-001. Office of Water, Washington DC, USA. USEPA (United States Environmental Prot ection Agency). 1998. Lake and reservoir bioassessment and biocriteria: EPA 841B-98-007 Technical Guidance Document. Washington, DC, USA. Available at http://www.epa.gov/owow/m onitoring/tech/lakes.html Accessed on 10/2006. USEPA (United States Environmental Protec tion Agency). 2001. Better assessment science integrating point and nonpoint sources: BASINS 3.0 User’s ma nual. EPA-823-B-01-001. Office of Water, Washington, DC, USA. USEPA (United States Environmental Protec tion Agency). 2002a. Methods for evaluating wetland condition: developing an invertebrate index of biological integrity for wetlands.EPA-833-R-02-019. Office of Water, Washington, DC, USA. USEPA (United States Environmental Prot ection Agency). 2002b. Methods for evaluating wetland condition: introduction to wetland biological assessment. EPA-822-R-02-014. Office of Water, U.S. Environmental Protection Agency, Washington, DC, USA.

PAGE 339

339 USGS (United States Geological Survey). 2006. National Aerial Photogr aphy Program (NAPP). U.S. Geological Survey, Washi ngton, DC, USA. Available at http://edc.usgs.gov/guides/napp.html Accessed on 05/2006. Uuemaa, E., J. Roosaare, and U. Mander. 2005. S cale dependence of landscape metrics and their indicatory value for nutrient and organic matter losses form catchments. Ecological Indicators 5:350–369. Wang, L., J. Lyons, and P. Kanehl. 2001. Impacts of urbanization on stream habitat and fish across multiple spatial scales. Environmental Management 28:255–266. Wang, L., J. Lyons, P. Kanehl, R. Bannerma n, and E. Emmons. 2000. Watershed urbanization and changes in fish communities in Southeas tern Wisconsin streams. Journal of the American Water Resource Association 36:1173-1189. Wang, L., J. Lyons, P. Kanehl, and R. Gatti. 1997. Influences of watershed land use on habitat quality and biotic integrity in Wi sconsin streams. Fisheries 22:6–12. Wang, L., J. Lyons, P. Rasmussen, P. Seelbach, T. Simon, M. Wiley, P. Kanehl, E. Baker, S. Niemela, and P. M. Stewart. 2003. Watershed, r each, and riparian influences on stream fish assemblages in the Northern Lakes and Fore st Ecoregion, U.S.A. Canadian Journal of Fisheries and Aquatic Sciences 60:461–505. Wear, D. N. 2002. Chapter 6: land use. p. 153-173. In D. N. Wear and J. G. Greis (eds.) Southern forest resource assessment. Gen. Tech Rep. SRS-53. U.S. Department of Agriculture, Forest Service, Southern Resear ch Station, Asheville, NC, USA. Wear, N. W., M. G. Turner, and R. J. Naim an. 1998. Land cover along urban-rural gradients: implications for water quality. Ecological Applications 8:619–630. Whitfield, D. F. 1994. Emergy basis for urban land us e patterns in Jacksonville, Florida. Ph.D. Dissertation, University of Fl orida, Gainesville, FL, USA. Wiens, J. A. 1989. Spatial scaling in ecology. Functional Ecology 3:385–397. Wiens, J. A. 2002. Riverine landscapes: taking landscape ecology into the water. Freshwater Biology 47:501–515. Wilcox, D. A., J. E. Meeker, P. L. Hudson, B. J. Armitage, M. G. Black, and D. G. Uzarski. 2002. Hydrologic variability and th e application of index of bi otic integrity metrics to wetlands: a Great Lakes ev aluation. Wetlands 22:588–615. Wu, J. 2004. Effects of changing scale on la ndscape pattern analysis : scaling relations. Landscape Ecology 19:125–138. Wu, J., D. E. Jelinski, M. Luck, and P. T. Tueller. 2000. Multiscale analysis of landscape heterogeneity: scale variance and pattern metr ics. Geographic Information Sciences 6:6– 19.

PAGE 340

340 Wu, J., W. Shen, W. Sun, and P. T. Tueller. 2002. Empirical patterns of the effects of changing scale on landscape metrics. Landscape Ecology 17:71–782. Young, R. G. and A. D. Huryn. 1999. Effects of land-use on stream metabolism and organic matter turnover. Ecologi cal Applications 9:1359–1376.

PAGE 341

341 2 4 BBIOGRAPHICAL SKETCH Manuel Benjamin Vivas was born in Bogot, Colombia. He graduated from Universidad de los Andes in 1988 after having earned a degree in biology. He earned a Ma ster of Arts degree from the Center for Latin American Studies/Tro pical Conservation Development Program at the University of Florida in 1998. In January 2000, Be njamin enrolled in the Ph.D. program in Systems Ecology through the Center for Wetlands, Department of Environmental Engineering Sciences, University of Florida, under the tutelage of Dr. Mark T. Brown. During his professional career in the fiel ds of natural resource management and conservation science, Benjamin has served in various positions of project management and consultancy in environmental non-governmental or ganizations. He is well-versed in ecoregional planning and national parks management, and has wo rked in close collabora tion with a variety of stakeholders to define strategies for conserva tion and natural resources management in South America. Benjamin is married to Sara Vivas, and is the proud father of Sofa Manuela. He enjoys cycling, fishing, Latin American lit erature, and classical music.