Citation
Characterization of the Impacts of the Built Environment on the Hydrologic Cycle for Life Cycle Assessment

Material Information

Title:
Characterization of the Impacts of the Built Environment on the Hydrologic Cycle for Life Cycle Assessment
Creator:
Castro-Raventos, R
Place of Publication:
[Gainesville, Fla.]
Florida
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (228 p.)

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Design, Construction, and Planning
Design, Construction and Planning
Committee Chair:
RIES,ROBERT
Committee Co-Chair:
KIBERT,CHARLES JOSEPH
Committee Members:
FRANK,KATHRYN I
JAWITZ,JAMES W
Graduation Date:
8/9/2014

Subjects

Subjects / Keywords:
Basins ( jstor )
Correlation coefficients ( jstor )
Groundwater ( jstor )
Land cover ( jstor )
Population density ( jstor )
River basins ( jstor )
Stream flow ( jstor )
Surface water ( jstor )
Water balance ( jstor )
Watersheds ( jstor )
Design, Construction and Planning -- Dissertations, Academic -- UF
impact-assessment -- land-cover -- life-cycle -- urban-ecology -- water-balance
Pasco County ( local )
Genre:
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Design, Construction, and Planning thesis, Ph.D.

Notes

Abstract:
The goal of this research is to develop a method to characterize the impacts of the built environment on freshwater resources in life cycle studies. The focus of this methodology is the impacts brought about by changes in land cover on the hydrological cycle. In order to calculate an impact indicator it is necessary to understand how changes in land cover can influence how water flows within a catchment. To this end, the water balance of a sample of drainage basins in the South Eastern Coastal Plains ecoregion is analyzed on an annual basis using the Budyko framework. The Budyko model provides a catchment coefficient that represents properties of the drainage basin and its effect on the partitioning of rainfall into evapotranspiration and run-off. The annually calculated catchment coefficient was studied to detect a trend in time and found two diverging trends regardless of the increase in urban development. On one hand, an increase in the catchment coefficient could be partially explained through the cross-correlation of surface water withdrawals at the county level. Nevertheless, further research is necessary to explain the sensitivity of changes in streamflow from water transfers and/or climate variability on the optimization procedure used to calculate the catchment coefficient. In contrast, a reduced sample of basins presented the expected run-off behavior and was used to explore the applicability of the proposed methodology. The resulting panel data regression models included random effects and morphometric parameters that allowed capturing the heterogeneity of the sample. Also, the models represented statistically significant relationships between the catchment coefficient and land cover classes and population parameters. This facilitated the estimation of annual stream-flow based on given urban growth expectations and climate scenarios. Finally, the impact of the built environment on freshwater resources was characterized as the proportion of the run-off ratio of a developed basin to a baseline condition. The methodology allowed assessment of the relative impact of land cover change on the hydrological cycle and could assist planners in the development of sustainable communities. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (Ph.D.)--University of Florida, 2014.
Local:
Adviser: RIES,ROBERT.
Local:
Co-adviser: KIBERT,CHARLES JOSEPH.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2015-02-28
Statement of Responsibility:
by R Castro-Raventos.

Record Information

Source Institution:
UFRGP
Rights Management:
Copyright Castro-Raventos, R. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
2/28/2015
Resource Identifier:
968131537 ( OCLC )
Classification:
LD1780 2014 ( lcc )

Downloads

This item has the following downloads:


Full Text

PAGE 1

CHARACTERIZATION OF THE IMPACTS OF THE BUILT ENVIRONMENT ON THE HYDROLOGIC CYCLE FOR LIFE CYCLE ASSESSMENT By RODRIGO CASTRO RAVENT"S A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2014

PAGE 2

© 2014 Rodrigo Castro Raventós

PAGE 3

To my father

PAGE 4

4 ACKNOWLEDGMENTS The process behind this dissertation has benefited from exchanges with many individuals. I would like to start by thanking my committee members, whose dedication to this project is greatly appreciated : Dr. Robert J. Ries for his patient ment orship and for sharing the joy of research , Dr. James Jawitz for helping me discover a passion for the field of hydrology , Dr. Kibert and Dr. Frank for their valuable insights on the applicability of this project to multiple spatial scales of the built environment . In addition, Dr. Zwick and Dr. Fik whose respective courses on spatial analysis and statistics were fundamental in helping me frame a new personal perspective on the built environment. Also, William Elizabeth the Shimberg Center for Housing Studies by providing assistance with census data and technical advice . Similarly, I would like to thank David Keelling and Ícaru Alzuru for their help in obtaining and preparing the climate data sets. In additi on, this work was made possible by the financial support provided by the University of Florida A lumni F und and the M.E. Rinker Sr., Sch ool of Construction Management. Moreover, I would like to recognize the influence on this work by Prof. Mirady Sebastiani whose course on strategic environmental impact assessment at Universidad Simón Bolívar planted the seed for this endeavor many years back. T he learning experience behind this project demanded an imp ortant emotional commitment. I would like to start by tha nk ing Katherine Castro for walking part of this road with me. Also, I appreciate the encouragement my mother and brother provided all along. Special thanks to Dr. Torres Antonini and Roberto Koenecke for their care. And l ast but not least, thanks to Charlotte Skov for her support and sense of humor .

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 8 LIST OF FIGURES ................................ ................................ ................................ ........ 11 LIST OF ABBREVIATIONS ................................ ................................ ........................... 16 ABSTRACT ................................ ................................ ................................ ................... 17 CHAPTER 1 INTRODUC TION ................................ ................................ ................................ .... 19 Aim and Objectives of Research ................................ ................................ ............. 19 Background and Significance ................................ ................................ ................. 19 Statement of Problem and Research Questions ................................ ..................... 22 ................. 24 r Across Catchments? ................................ ................................ ................................ . 24 Rationale of Study ................................ ................................ ................................ .. 24 Summary ................................ ................................ ................................ ................ 26 2 CHARACTERIZATION OF THE IMPACT OF THE BUILT ENVIRONMENT ON WATER RESOURCES IN LIFE CYCLE ASSESSMENT ................................ ........ 28 Background ................................ ................................ ................................ ............. 28 Methodological Discussion: Water in LCA ................................ .............................. 30 From Water Inventory Analysis to Impact Modeling ................................ ......... 30 Impacts from Product Systems ................................ ................................ ......... 32 The Built Environment, a Site Dependent Assessment ................................ .... 35 Challenges to LCIA of Water Resources ................................ .......................... 37 Modeling Approach ................................ ................................ ................................ . 40 Goal and Scope Definition ................................ ................................ ................ 40 Life Cycle Inventory Analysis ................................ ................................ ............ 40 Impact Assessment ................................ ................................ .......................... 42 Classification ................................ ................................ .............................. 42 Characterization (model) ................................ ................................ ............ 42 Cause effect chain ................................ ................................ ..................... 42 System Elements ................................ ................................ ............................. 43 Hydrological Model ................................ ................................ ........................... 44 Spatial scale ................................ ................................ ............................... 45 Temporal scale ................................ ................................ .......................... 46 Summary ................................ ................................ ................................ ................ 46

PAGE 6

6 3 IMPACT ASSESSMENT OF LAND COVER CHANGE ON STREAMFLOW WITHIN A WATER ENERGY BALANCE FRAMEWORK ................................ ....... 52 Background ................................ ................................ ................................ ............. 52 Literature Review ................................ ................................ ................................ .... 53 Representation of Catchment Behavior ................................ ............................ 53 Hypothesis ................................ ................................ ................................ ........ 56 Methodological Fra mework ................................ ................................ ..................... 56 Basic Statistical Model and Assumptions ................................ ......................... 56 Dependent variable ................................ ................................ .................... 58 Independent variables ................................ ................................ ................ 58 Land c over classification ................................ ................................ ............ 60 Land cover variable normalization ................................ ............................. 61 Population density ................................ ................................ ...................... 61 Morphometric parameters ................................ ................................ .......... 62 Regression Procedure ................................ ................................ ...................... 63 An alytical scenarios ................................ ................................ ................... 64 Calculation of catchment coefficient ................................ ........................... 64 Trend Analysis ................................ ................................ ................................ .. 65 Case Study Pre Selection ................................ ................................ ................ 65 Results and Discussion ................................ ................................ ........................... 66 Case Study Pre Selection Assessment ................................ ............................ 66 Exploratory analysis ................................ ................................ ................... 67 Trend analysis ................................ ................................ ............................ 68 Alternative Hypothesis ................................ ................................ ...................... 7 0 Alternative hypothesis testing: withdrawals ................................ ................ 71 Alternative hypothesis testing: discharges ................................ ................. 73 Alternative hypothesis testing: conclusions ................................ ................ 76 Case Study Sample ................................ ................................ .......................... 76 Regression Analysis ................................ ................................ ......................... 79 Conclusions ................................ ................................ ................................ ............ 82 4 LIFE CYCLE IMPACT ASSESSMENT CHARACTERIZATION OF CHANGES IN LAND COVER ON WATER RESOURCES: METHOD APPLICATION .............. 91 Background ................................ ................................ ................................ ............. 91 Impact Ranges and Baseline Condition ................................ ........................... 91 Impact Indicator ................................ ................................ ................................ 93 Differentiat ion from water stress indices ................................ .................... 93 Climate variability ................................ ................................ ....................... 94 Methodological Approach ................................ ................................ ....................... 94 Retrospective Assessment ................................ ................................ ............... 95 Prospective Assessment ................................ ................................ .................. 96 Conclu sions ................................ ................................ ................................ ............ 98

PAGE 7

7 5 CONCLUSIONS AND FUTURE RESEARCH ................................ ....................... 104 Discussion ................................ ................................ ................................ ............ 104 Challenges in the Application of the Budyko Hypothesis ................................ 105 Challenges in the Calculation of the Impact Factor ................................ ........ 106 Conclusions ................................ ................................ ................................ .......... 106 Future Research ................................ ................................ ................................ ... 107 APPENDIX A WATER BALANCE CALCULATION METHODS ................................ .................. 109 B INDEPENDENT VARIABLES CALCULATION ................................ ..................... 125 C CASE STUDIES SUMMARY ................................ ................................ ................ 127 D CASE STUDY ANALYSIS SUPPORT INFORMATION ................................ ........ 169 E STATISTICAL MODEL OUTPUT ................................ ................................ .......... 198 F IMPACT ASSESSMENT SCENARIOS ................................ ................................ . 208 G LIST OF REFERENCES ................................ ................................ ....................... 215 H BIOGRAPHICAL SKETCH ................................ ................................ ................... 228

PAGE 8

8 LIST OF TABLES Table page 2 1 Water use characterization factors within Abiotic Resources in terms of their availability for future generations ................................ ................................ ........ 47 2 2 Land use characterization factors related to water use: land occupation (area*time), and land transformation (changing quality per area unit) ................ 47 2 3 Characteristics of watershed management units ................................ ................ 48 3 1 Data Time Steps ................................ ................................ ................................ . 83 3 2 Hydr ologic data requirements for water balance optimization ............................ 83 3 3 Hypothetical Influence on the water budget by land cover type ......................... 83 3 4 Trend analysis of the catchment coefficient of the nine basins sub sample at different temporal ranges ................................ ................................ .................... 84 3 5 Potential sources of streamflow alterations at the nine basins in the sub sample ................................ ................................ ................................ ................ 85 3 6 Summary of regression models for the four (4) basins sample ........................... 86 4 1 Impact ranges ................................ ................................ ................................ ..... 99 4 2 Land cover class ratios and population for the Charlie Creek near Ga rdner basin ................................ ................................ ................................ ................... 99 4 3 Land cover class ratios and population for the Anclote River near Elfers basin . 99 4 4 Prospective urban growth scenarios for the Anclote River near Elfers basin, 2010 2030 ................................ ................................ ................................ ........ 100 A 1 Data for Matlab scripts ................................ ................................ ...................... 114 C 1 Time series of land cover types and population density for St. Marys River b asin (USGS 02231000) ................................ ................................ .................. 128 C 2 Time series of land cover types and population density for Ft. Drum Creek basin (USGS 02231342) ................................ ................................ .................. 130 C 3 Time series of land cover types and population density for Gee Creek basin (USGS 02234400) ................................ ................................ ............................ 132 C 4 Time series of land cover types and population density for Wekiva River basin (USGS 02235000) ................................ ................................ .................. 134

PAGE 9

9 C 5 Time series of land cover types and population density for South Black Fork Creek basin (USGS 02245500) ................................ ................................ ........ 136 C 6 Time series of land cover types and population density for Tomoka River basin (USGS 02247510) ................................ ................................ .................. 138 C 7 Time series of land cover types and population density for Shingle Creek basin (USGS 02263800) ................................ ................................ .................. 140 C 8 Time series of land cover types and population density for Davenport Creek basin (USGS 02266480) ................................ ................................ .................. 142 C 9 Time series of land cover types and population density for Catfish Creek basin (USGS 02267000) ................................ ................................ .................. 144 C 10 Time series of land cover types and population density for Charlie Creek basin (USGS 02296500) ................................ ................................ .................. 146 C 11 Time series of land cover types and population density for Joshua Creek basin (USGS 02297100) ................................ ................................ .................. 148 C 12 Time series of land cover types and population density for Prairie Creek basin (USGS 02298123) ................................ ................................ .................. 150 C 13 Time series of land cover types and popul ation density for Bullfrog Creek basin (USGS 02300700) ................................ ................................ .................. 152 C 14 Time series of land cover types and population density for Trout Creek basin (USGS 02303350), ................................ ................................ ........................... 154 C 15 Time series of land cover types and population density for Brooker Creek basin (USGS 023073 59) ................................ ................................ .................. 156 C 16 Time series of land cover types and population density for South Branch Anclote River basin (USGS 02309848) ................................ ............................ 158 C 17 Time series of land cover types and population density for Anclote River basin (USGS 02310000) ................................ ................................ .................. 160 C 18 Time series of land cover types and population density for Steinhatchee River basin (USGS 02324000) ................................ ................................ ......... 162 C 19 Time series of land cover types and population density for Sopchoppy River basin (USGS 02327100) ................................ ................................ .................. 164 C 20 List of potential case studies with flow records and remarks ............................ 166 D 1 Climate type and meteorological ranges ................................ .......................... 169

PAGE 10

10 D 2 Non linear regressions of Q/P=f(aridity index, catchment coefficient) .............. 169 D 3 Land Cover Classification ................................ ................................ ................. 170 D 4 List of pre selected case studies ................................ ................................ ...... 171 D 5 Morphometric variables for the 19 ba sins sample ................................ ............ 172 D 6 Summary of descriptive statistics for two case study samples ......................... 172 D 7 Trend analysis of 10 year average discharge (Q_10) and 10 year average precipitation (P_10) and land cover types (LC it ) and population density (PD t ), 1980 2010 ................................ ................................ ................................ ........ 173 D 8 Trend analysis of 10 year average discharge (Q_10) and 10 year average precipitation (P_10) and land cover types (LC it ) and population density (PD t ), 1992 2010 ................................ ................................ ................................ ........ 175 D 9 List of springs per basin ................................ ................................ .................... 176 D 10 List of domestic wastewater facilities per basin ................................ ................ 177 D 11 List of industrial wastewater facilities per basin ................................ ................ 178 D 12 Cross correlation analysis between the catchment coefficient and surface water withdrawals at the county level ................................ ............................... 179 F 1 Retrospective assessment of Charlie Creek basin ................................ ........... 208 F 2 Prospective scenario #1 of Anclote River basin (1992 2010, 2010 2030) ........ 209 F 3 Prospective scenario #1 of Anclote River basin (2010 2030), and mean aridity index plus confidence interval ................................ ................................ 210 F 4 Prospective scenario #1 of Anclote River basin (2010 2030), and mean aridity index minus confidence interval ................................ ............................. 211 F 5 Prospective scenario #2 of Anclote River basin (2010 2030), and mean aridity index ................................ ................................ ................................ ...... 212 F 6 Prospective scenario #2 of Anclote River basin (2010 2030), and mean aridity index plus confidence interval ................................ ................................ 213 F 7 Prospective scenario #2 of Anclote River basin (2010 2030), and mean aridity index minus confidence interval ................................ ............................. 214

PAGE 11

11 LIST OF FIGURES Figure page 1 1 Total freshwater withdrawals in million gallons per day and % of total water withdrawals in Florida by category in 2010 (Marella 2014) ................................ . 27 2 1 Schematic diagram of the phases in an LCA (ISO 14040 1997) ........................ 48 2 2 Life cycle impact assessment based on ISO 14042 (Baumann and Tillman 2004). ................................ ................................ ................................ ................. 49 2 3 Schematic presentation of LCIA modeling (Adapted from Finnveden et al. 2009) ................................ ................................ ................................ .................. 49 2 4 Life cycle inventory water flow model (Adapted from Chhabra 2011) ................. 50 2 5 Hypothetical relationship of variables to catchment storages ............................. 50 2 6 Cause effect chain showing the relationship between land cover classes and morphometric variables on the catchment coefficient resulting run off ratio ....... 51 3 1 Relationship between the ratio of mean annual evapotranspiration to precipitation (E/P) as a function of the aridity index (E 0 /P) for different values ................................ ................................ ................................ ... 87 3 2 The ratio of run off to precipitation at different catchment coefficients ( ) and a consta ................................ ................................ ............ 87 3 3 Historical freshwater withdrawals by water source in Florida, 1950 2010 (Marella 2014) ................................ ................................ ................................ .... 88 3 4 Scatterplot of the relationship between UBG and FWW for sub sample of 9 basins (1992 2010) ................................ ................................ ............................. 88 3 5 Scatterplot of the relationship between UBG and FWW for the selected sample of four (4) basins (1992 2010) ................................ ................................ 89 3 6 Geographical distribution of basin case studies ................................ .................. 89 3 7 Linear regression plot for pooled time series, catchment coefficient ( _10)=f(UBG) ................................ ................................ ................................ ... 90 3 8 Scatterplot of the catchment coefficient ( ) for the four (4) case studies (1992 2010) ................................ ................................ ................................ ........ 90 4 1 Partitioning of precipitation according to ISA r anges (Source: Livingston and McCarron 1992) ................................ ................................ ................................ 100

PAGE 12

12 4 2 Presumptive standards of protection for ecological flows (Richter et al. 201 1) . 1 01 4 3 Charlie Creek basin near Gardner 2011 land cover map with Anderson level II land cover classification ................................ ................................ ................. 101 4 4 Model analysis for Charlie Creek basin near Gardner, 1992 2010. .................. 102 4 5 Anclote River basin near Elfers 2011 land cover map with Anderson level II land cover classification ................................ ................................ .................... 102 4 6 Model analysis for Anclote River basin near Elfers, 1992 2010 and 2010 2030. ................................ ................................ ................................ ................ 103 A 1 Water balance calculation algorithm ................................ ................................ . 115 C 1 Catchment coefficient for St. Marys River basin (USGS 02231000), 1950 2010 ................................ ................................ ................................ ................. 129 C 2 Catchment coefficient for Ft. Drum Creek basin (USGS 02231342), 1980 2010 ................................ ................................ ................................ ................. 131 C 3 Catchment coefficient for Gee Creek basin (USGS 02234400), 1980 2010 ..... 133 C 4 Catchment coefficient for Wekiva River basin (USGS 02235000), 1950 2010 . 135 C 5 Catchment coefficient for South Black Fork Creek basin (USGS 02245500), 1950 2010 ................................ ................................ ................................ ........ 137 C 6 Catchment coefficient for Tomoka River basin (USGS 02247510), 1970 2010 139 C 7 Catchment coefficient for Shingle Creek basin (USGS 02263800), 1950 2010 141 C 8 Catchment coefficient for Davenport Creek basin (USGS 02266480), 1970 2010 ................................ ................................ ................................ ................. 143 C 9 Catchment coefficient for Catfish Creek basin (USGS 02267000), 1950 2010 145 C 10 Catchment coefficient for Charlie Creek basin (USGS 02296500), 1950 2010 147 C 11 Catchment coefficient for Joshua Creek basin (USGS 02297100), 1950 2010 149 C 12 Catchment coefficient for Prairie Creek basin (USGS 02298123), 1980 2010 . 151 C 13 Catchment coefficient for Bullfrog Creek basin (USGS 02300700), 1980 2010 153 C 14 Catchment coefficient for Trout Creek basin (USGS 02303350), 1980 2010 ... 155 C 15 Catchment coefficient for Brooker Creek basin (USGS 02307359), 1951 2010 ................................ ................................ ................................ ................. 157

PAGE 13

13 C 16 Catchment coefficient for South Branch Anclote River basin (USGS 02309848), 1971 2010 ................................ ................................ ..................... 159 C 17 Catchment coefficient for Anclote River basin (USGS 02310000), 1950 2010 . 161 C 18 Catchment coefficient for Steinhatchee River basin (USGS 02324000), 1950 2010 ................................ ................................ ................................ ................. 163 C 19 Cat chment coefficient for Sopchoppy River basin (USGS 02327100), 1970 2010 ................................ ................................ ................................ ................. 165 C 20 Geographical distribution of pre selected case st udy catchments. ................... 168 D 1 Time series of catchment coefficient ( t ) for the pre selected sample ............. 179 D 2 Geographical distribution of nine (9) basins sub sample for long term analysis (1950 2010) of trends in the catchment coefficient. ............................ 180 D 3 Linear regression plot for pooled time series, catchment coefficient ( _10)=f(UBG) differentiated by case study basin ................................ ........... 180 D 4 Surface water withdrawals and the catchment coefficient for St. Marys River basin (USGS 02231000), 1960 2010 ................................ ............................... 181 D 5 Surface water withdrawals and the catchment coefficient for Wekiva River basin (USGS 02235000), 1960 2010 ................................ ............................... 181 D 6 Surface water withdrawals and the catchment coefficient for South Black Fork Creek basin (USGS 02245500), 1960 2010 ................................ ............. 182 D 7 Surface water withdrawals and the catchment coefficient for Catfish Creek basin (USGS 02267000), 1960 2010 ................................ ............................... 182 D 8 Surface water withdrawals and the catchment coefficient for Charlie Creek basin (USGS 02296500), 1960 2010 ................................ ............................... 183 D 9 Surface water withdrawals and the catchment coefficient for Joshua Creek basin (USGS 02297100), 1960 2010 ................................ ............................... 183 D 10 Surface water withdrawals and the catchment coefficient for Brooker Creek basin (USGS 02307359), 1960 2010 ................................ ............................... 184 D 11 Surface water withdrawals and the catchment coefficient for Anclote River basin (USGS 02310000), 1960 2010 ................................ ............................... 184 D 12 Surface water withdrawals and the catchment coefficient for Steinhatchee River basin (USGS 02324000), 1960 2010 ................................ ...................... 185

PAGE 14

14 D 13 Surface water withdrawals and the catchment coefficient for Shingle Creek basin (USGS 02263800), 1965 2010 ................................ ............................... 185 D 14 Surface water withdrawals and the catchment coefficient for Davenport Creek basin (USGS 02266480), 1965 2010 ................................ ..................... 186 D 15 Surface water withdrawals and the catchment coefficient for Prairie Creek basin (USGS 02298123), 1965 2010 ................................ ............................... 186 D 16 Surface water withdrawals and the catchment coefficient for Trout Creek basin (USGS 02303350), 1965 2010 ................................ ............................... 187 D 17 Cross correlation plot of the catchment coefficient and county level surface water withdrawals with upper and lower confidence limits for St. Marys River basin (USGS 02231000) in Baker County. ................................ ....................... 187 D 18 Cross correlation plot of the catchment coefficient and county level surface water withdrawals with upper and lower confi dence limits for Wekiva River basin (USGS 02235000). A) Seminole County; B) Orange County. ................. 188 D 19 Cross correlation plot of the catchment coefficient and county level surface water withdrawals with upper and lower confidence limits for South Black Fork Creek basin (USGS 02245500) in Clay County. ................................ ....... 189 D 20 Cross correlation plot of the catchment coefficient and county level surface water withdrawals with upper and lower confidence limits for Catfish Creek basin (USGS 02267000) in Polk County. ................................ ......................... 189 D 21 Cross correlation plot of the catchment coefficient and county level surface water withdrawals with upper and lower confidence limits for Charlie Creek basin (USGS 02296500) in Hardee County. ................................ ..................... 190 D 22 Cross correlation plot of the catchment coefficient and county le vel surface water withdrawals with upper and lower confidence limits for Joshua Creek basin (USGS 02297100) in Desoto County. ................................ ..................... 190 D 23 Cross correlation plot of the catchment coefficient and county level surface water withdrawals with upper and lower confidence limits for Brooker Creek basin (USGS 02307359) in Hillsborough County. ................................ ............. 191 D 24 Cross correlation plot of the catchment coefficient and county level surface water withdrawals with upper and lower confidence limits for Anclote River basin (USGS 02310000) in Pasco County. ................................ ...................... 191 D 25 Cross correlation plot of the catchment coefficient and county level surface water withdrawals with upper and lower confidence limits for Steinhatchee River basin (USGS 02324000) in Lafayatte County. ................................ ........ 192

PAGE 15

15 D 26 Cross correlation plo t of the catchment coefficient and county level surface water withdrawals with upper and lower confidence limits for Shingle Creek basin (USGS 02263800) in Orange County. ................................ .................... 192 D 27 Cross correlation plot of the catchment coefficient and county level surface water withdrawals with upper and lower confidence limits for Davenport Creek basin (USGS 02266480). A) Osceola County; B) Polk C ounty. ............. 193 D 28 Cross correlation plot of the catchment coefficient and county level surface water withdrawals with upper and low er confidence limits for Prairie Creek basin (USGS 02298123) in Desoto County. ................................ ..................... 194 D 29 Cross correlation plot of the catchment coefficient and county level surface water withdrawals with upper and lower confidence limits for Trout Creek basin (USGS 02303350). A) Hillsborough County; B) Pasco County. .............. 195 D 30 Cross correlation plot of the catchment coefficient and county level surface water withdrawals with upper and lower confidence limits for Wekiva River basin (USGS 02235000). A) Orange County; B) Seminole County. ................. 196 D 31 Cross correlation plot of the catchment coefficient and county level surface water withdrawals with upper and l ower confidence limits for Anclote River basin (USGS 02310000) in Pasco County. ................................ ...................... 197 D 32 Cross correlation plot of the catchment c oefficient and county level surface water withdrawals with upper and lower confidence limits for Steinhatchee River basin (USGS 02324000) in Lafayatte County. ................................ ........ 197

PAGE 16

16 LIST OF ABBREVIATIONS ADR Annual Data Report BGD Billion gallons per day CBG Census block group CF Characterization factor GAGES II Geospatial Attributes of Gages for Evaluating Streamflow , version 2 GOES Geostationary Operational Environmental Satellite LCA Life cycle assessment LCIA Life cycle impact assessment LCI Life cycle inventory MGD Million gallons per day MRLC Multi resolution land characteristics consortium NLCD National land cover database NAWQA National Water Quality Assessment Program SBA Sustainable boundary approach USGS United States Geological Survey

PAGE 17

17 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy CHARACTERIZATION OF THE IMPACTS OF THE BUILT ENVIRONMENT ON THE HYDROLOGIC CYCLE FOR LIFE CYCLE ASSESSMENT By Rodrigo Castro Raventós August 2014 Chair: Robert J. Ries Major: De sign, Construction and Planning The goal of this research is to develop a method to characterize the impacts of the built environment on freshwater resources in life cycle studies. The focus of this methodology is the impacts brought about by changes in land cover on the hydrologic al cycle. In order to calculate an impact indicator it is necessary to understand how changes in land cover can influence how water flows within a catchment. To this end, the water balance of a sample of drainage basin s in the South Eastern Coast al Plains ecoregion is analyzed on an annual basis using the Budyko framework. The Budyko model provides a catchment coefficient that represents properties of the drainage basin and its effect on the partitioning of rainfall into evapotranspiration and run off. T he annually calculated catchment coefficient was stu died to detect a trend in time and found two diverging trends regardless of the increase in urban development. On one hand, an increase in the catchment coefficient could be partially explained thr ough the cross correlation of surface water withdrawals at the county level. Nevertheless, further research is necessary to explain the sensitivity of changes in streamflow from water transfers and/or climate variability on the optimization procedure

PAGE 18

18 used to calculate the catchment coefficient. In contrast, a reduced sample of basins presented the expected run off behavior and was used to explore the applicability of the proposed methodology. The resulting panel data regression models included random effect s and morphometric parameters that allowed capturing the heterogeneity of the sample. Also, the models represented statistically significant relationships between the catchment coefficient and land cover classes and population parameters. This facilitated the estimation of annual stream flow based on given urban growth expectations and climate scenarios. Finally , the impact of the built environment on freshwater resources was characterized as the proportion of the run off ratio of a developed b asin to a bas eline condition . Th e methodology allowed assess ment of the relative impact of land cover change on the hydrologic al cycle and could assist planners in the development of sustainable communities .

PAGE 19

19 CHAPTER 1 INTRODUCTION Aim and Objectives of Research The aim of this research project is to explore the impacts of the built environment on the hydrological cycle at different spatial scales within a life cycle perspective. The specific objectives of the study are : Develop a me thod to characterize the impacts of the built environment through changes in land cover on freshwater resources in life cycle assessment studies. Define representative baseline conditions of the water mass balance at a regional level. Define a significant land cover change condition for water balance alterations. D efine impact thresholds based on distance to target approach of catchment behavior . Background and Significance The built environment impacts fre shwater resources through the alterations to the h ydrological cycle brought about by exploitation and changes in land cover. W ater withdrawals limit the availability of the resource to satisfy human demand and ecosystem services. Changes in land cover impact the hydrologic cycle by altering the rainfall r unoff process. Both processes are interrelated and their proper management would help achieve the sustainability of human society. For instance, e nsuring that cities have an adequate supply of water is increasingly important as human populations continue t o concentrate in urban areas. Globally, urban population as a percentage of total population has expanded from 30% in 1950 to over 50% as of 2010 (UN 2012). In the United States (US), a similar trend has been recorded, as more than 80% of the US population now live in urban areas, compared to 64% in 1950 (UN 2012). Correspondingly, freshw ater demands in the US

PAGE 20

20 public supply sector have increased steadily with population growth, from 14 billion gallons per day (BGD) in 1950 to 44.2 BGD in 2005 . Similarly, self supplied domestic freshwater withdrawals ha ve increased from 2.1 BGD in 1950 to 3.8 BGD in 2005 (Huston et al. 2005 ; Kenny et al. 2009 ). In addition, saline and fresh water demand for thermoelectric generation has increased from 40 BGD in 1950 to 201 B GD in 2005, making the the rmoelectric power generation sector the largest water withdrawal in the US. In Florida for 2010, freshwater withdrawals for domestic, industrial, commercial and energy use were approximately 3,850 million gallons per day (MGD), cl ose to 60% of state wide freshwater withdrawals. Figure 1 1 shows a breakdown of freshwater withdrawals for Florida. The overall growth in demand can be attributed to demographic and economic trends as reflected in household and commercial floor space, and climatic fluctuations (Kenny et al. 2009 ; USEIA 2012 ). As a consequence, r apidly growing urban demands are straining local and regional water supplies and concerns over urban water scarcity in the US are becoming more p rominent (Levin et al. 2002). Howeve r, the impact of the built environment on freshwater resources extends beyond water use to include the influence of land use on the hydrologic cycle , the process by which water flows through ecosystems replenishing water bodies and renewing groundwater . L a nd use impacts the hydrologic cycle by altering land cover which changes the dynamics of rainfall run off generation and flow . In addition , land use generates pollutants that are carried to water bodies as they are washed out by stormwater . Therefore, t he environmental consequences of land cover change are increasingly recognized as potentially undermining the capacity of ecosystems to provide critical services, such as freshwater for irrigation, industry, and domestic

PAGE 21

21 consumption (Foley et al. 2005). La nd cover change alters the hydrologic cycle at local and regional levels and conditions the availability of surface water for appropriation (Vörösmarty et a l. 2000). Life cycle assessment (LCA) is a standardized method for measuring the environmental impac t of any product or service by accounting for its life cycle material and energy flows. Consequently, in 2009 the Society of Environmental Toxicology and Chemistry (SETAC) launched a Work Area Interest G roup (WAIG) to develop standards for the assessment o f water use (Humbert 2013) . A recent methodological review by Kounina et al. (2013) shows that the cause effect chain between land use and water resources has not been addressed. Application of LCA to the built environment dates back to the 1970's (Stein 1 977) for assessing the environmental performance of a major sector of the economy with a long service life and high resource use during operation (Cole and Rousseau 1992, Cole and Kernan 1996, Adalberth 1997, Horvath and Hendrickson 1998, Mahdavi and Ries 1998, Hendrickson and Horvath 2000, Ries and Mahdavi 2001). The application of LCA to the building industry has principally been in four different ways: using LCA for the analysis of building materials and component combinations, the construction process, building operations, and end of life. Ortiz, et a l. (2007) found that most LCA studies in the built environment focus on energy consumption. While considerable attention has been paid to energy in the LCA of buildings, relatively little has been quantified in terms of water use and water consumption specific to the built environment. For instance, Arpke and Hutzler (2006) analyzed operational domestic water use for different building types using energy requirements for water supply and

PAGE 22

22 waste treatment. Lund in et al. (2002), Lundie et al. (2004), El Sayed et al. (2010), and Borghi et al. (2013) define the system boundaries for waterworks and wastewater treatment systems in isolation from the natural hydrologic system. Moreover, the authors focus on the impacts of effluents and emissions on environmental categories without acknowledging the impact of water use on the drainage basins mass balance , such as, changes in streamflow . Therefore, there is an opportunity to develop a comprehensive und erstanding of the impact of the built envi ronment on the hydrologic cycle . We propose to develop a characterization model for the impact assessment of land cover changes from the built environment on the hydrologic cycle. The model framework would assist i n the evaluation of alternative engineered solutions, the development of adaptation and mitigation strategies, and the management of water resources. Statement of Problem an d Research Questions Historically, the typical practice for impact assessment of water as a resource in life cycle assessment was calculating an inventory of water use reported on a mass or volume basis. This practice has changed as impact oriented water footprinting methods h ave been formulated (Berger and Finkbeiner 2013, Kounina et al. 2013). Recent work has examined the relative scarcity of water resources in regions around the world and compared that to withdrawals (Pfister et a l. 2009) using the water stress concept as th e basis of the impact calculation (UN 1997). Other recent work that reviewed water resource impact models concluded that virtual water and the water stress index are promising approaches for estimating water resource use and their impact on resource deplet ion and ecosystem requirements (Milà i Canals et a l. 2009). These and others

PAGE 23

23 recognize the need to differentiate between consumptive and non consumptive use of water for life cycle impact assessment. Furthermore, current methodological advances have suppor ted the detailed development of inventories of water use and its spatial differentiation but have mostly focused on the associated impacts of water use on human health and ecosystem quality (Kounina et a l. 2013). This is consistent with the perception in l ife cycle assessment that an impact oriented water footprinting approach is preferable to a volumetric footprinting one consumption depend on local scarcity, the type of watercourse used, water quality, the time of withdrawal, Finkbeiner 2013, 87). However, the interrelationship between land use and water resources has not been explicitly acknowledged as the area of protection that represents resource depletion treats water use and land use separately (Finnveden et a l. 2009). However, Heuvelmans et al. (2005a, 2005b) showed the impact of land use on hydrologic behavior by defining the drainage basin as its system boundaries, focusing on changes in vegetation and proposing a new impact category: 'regional water balance'. This method does not elaborate a characterization factor by land use type but shows the gap in our understanding of the environmental impact of the built environment as it represents a land use tr ansformation and occupation that has an effect on the hydrologic cycle. We propose an impact assessment modeling approach for ground and surface water resources that can be used to evaluate land use and the built environment. Pre development water flows ar e altered in several ways when development occurs. For

PAGE 24

24 example, water extraction for both consumptive and non consumptive uses generally increases with development. Imperviousness is altered by building and infrastructure construction, and stormwater struc tures are typically put in place to manage stormwater flows. These changes affect the quantity of water and the flow regime in an ecosystem compared to a pre determined state. In other words, the alterations of land cover and land use impact the water mass balance. The research questions that are addressed in the study are: How D oes E ach C W ater B alance C hange O ver T ime? as shown in the run off ratio ? Is the impact on the water balance proportional at different spatial levels ? What is the marginal rate of change between urban land cover and run off ratio? Is it feasible to find a regional or global marginal rate of change? How would the impact from changes of land cover at a par cel level differ from changes in land cover at a basin level? How D oes E ach C W ater B alance C hange D iffer A cross C atchments? Are there other variables in addition to land cover that have an impact on the Is the impact of la nd cover changes on the water balance dependent on the Rationale of Study This study seeks to expand the understanding of the life cycle studies of the built environment and its impact on freshwater resources by characterizing hydrological conditions associated with land cover variables within adequate spatial scales and system boundaries. This would allow quantifying the impacts on the hydrological cycle of land cover alterations from urbanization or other land m anagement criteria.

PAGE 25

25 This work starts from the premise that the built environment can be analyzed from two complementary perspectives: p roduct system and building system . The first is part of a supply chain network of products with varying spatial outreach, similar to the agriculture industry. As a consequence, t he assembly of building products aggregate s the environmental impact from the different locations w h ere the goods are extracted and processed for use . On the other hand, the buildi ng system has a sit e specific impact which for our purposes manifests itself on the hydrologic cycle through alterations to the water mass balance . For this study, the conceptual hydrological framework is defined by the Budyko hypothesis, which states that the long term equi librium in the water mass balance depends on the interaction between water supply through precipitation and the atmospheric water demand represented by potential evapotranspiration. This modeling approach was selected based on its parsimonious formulation and data requirements, which includes the use of a catchment coefficient to account for the physical characteristics of the basin as the unit of analysis. Furthermore , this projects plans to use the watershed as the unit of analysis within an ecosystem reg ion framework . This approach is part of current research on the trends, causes and impacts of land use and land cover change across the conterminous United States (Omernik 1987; USEPA 1999; Loveland et al. 2002). It is expected that land cover cannot solel y determine the complex process behind run off generation, and other variables would play a role in this process. However, by confining the analysis to areas that share similar ecosystem characteristics, such as climate, soils, and vegetation, the signal p rovided by changes in land cover can be more easily identified.

PAGE 26

26 The expected result is a better understandi ng on the role of land cover on the water balance that could inform planners on the impacts and tradeoffs of land cover mixes at a regional level. In the case of the built environment, this could frame the proportional presence of urban areas to other land cover classes inside watersheds and define watershed specific impacts. Consequently, urbanization can be studied within hydrological thresholds of r ainfall runoff behavior. Th is study proposes the following hypotheses to explain the relationship between land cover and catchment behavior: 1. If there is a non linear inverse relationship between the catchment coefficient and runoff, then, an increase of a land cover class that promotes run off should decrease the catchment coefficient to reflect the alteration in the water balance. Alternatively, 2. If the calculation of the catchment coefficient depends on a given runoff ratio under soil moisture storage eq uilibrium , then water transfers in the basin, either in the form of discharge or withdrawals, could alter the water balance and consequently the coefficient. Summary This project proposes to address a current gap in life cycle assessment methodology : the inclusion of impacts of changes in land cover on the hydrologic cycle. The method will explore the application of hydrologic modeling of drainage basins for the assessment of changes in streamflow as they correlate to changes in land cover . The aim is to better understand and communicate the impact of land use on freshwater resources in life cycle assessment studies. This would help inform decision makers on the trade offs of strategies that have different land use implications and potentially inform l and use planning.

PAGE 27

27 Figure 1 1. Total freshwater withdrawals in million gallons per day and % of total water withdrawals in Florida by category in 2010 ( Marella 2014) Public supply , 2,250.85 , 35% Domestic self supplied , 213.84 , 3% Commercial industrial mining self supplied , 378.35 , 6% Agricultural self supplied , 2,551.10 , 40% Recreational landscape irrigation , 391.93 , 6% Power generation , 613.19 , 10%

PAGE 28

28 CHAPTER 2 CHARACTERIZATION OF THE IMPACT OF THE BUILT ENVIRONMENT ON WATER RESOURCES IN LIFE CYCLE ASSESSMENT Background The life cycle assessment (LCA) framework provides a methodology to study the environmental impacts associated with a product or process from creation to decommission ing . The LCA procedure follows four concept ual phases as shown in Figure 2 1 : Goal and scope definition; Inventory analysis; Impact assessment; Interpretation. deals with defining the object of analysis and the system boundaries. Also, it gives policy makers the flexibi lity to address particular issues in the analysis. The relevant energy and mass balance. The last two LCA phases fall within what has been defined as a life cycle impact assessment (LCIA) model, an The aim of an LCIA is to translate the environmental loads or changes defined in the inventory analysis into environmental impacts, thus, facilitating the understanding of the e ffects of those loads and the comparison of selected options. This is performed by classifying the inventory according to the type of impact and by characterizing the relative contributions to each cause effect category based on the quantities of emissions and resources used and their equivalencies. Finally, the i nterpretation phase provides the opportunity to improve the process under study to ameliorate its impacts .

PAGE 29

29 As a result of the technical workshop on LCIA sponsored by SETAC in 1992, Figure 2 2 shows the three phases of LCIA: classification, characterization and valuation. These stages have evolved into the international standard ISO 14042, which requires an LCIA to have three core or mandatory elements: impact category definition, classification and characterization. These make up what is called the LCIA profile. In addition, there are four optional elements: normalization, grouping, weighting and data quality analysis (Baumann and Tillman 2004). These elements are shown in Figure 2 3 and are describe d as follows: 1. Impact category definition: Impacts are grouped into broad groups, such as, resource use, human health and ecological consequences. As such, this step defines the direction of the cause effect chain to a particular end point. These categories can be further sub divided depending on the scope of the study and/or data requirements. These categories are also called areas of protection. 2. Classification: In this process the inventory is grouped according the selected impact categories. In case an el ement can impact more than a single category, care must be taken to avoid double counting, and only direct effects should be consider. 3. Characterization: To facilitate the comparison of effects or understand their aggregate behavior, this process uses equivalency factors to transform the inventory quantities into comparable impacts. This is based on physico chemical characteristics of the inventory items. In the literature, the term characterization factor (CF) and category indicator are interchangeable. The application of carbon equivalent values to different greenhouse gases is an example of this step. 4. Normalization: The characterization 2004, 141). 5. Grouping: Sorting the CFs into sub sets for better communication of results. 6. Weighting: This process determines the rel ative importance of the environmental impacts based on quantitative or qualitative criteria. Ethical and ideological values influence the selection and application of this process, thus making it very controversial.

PAGE 30

30 7. Data quality analysis: The application o f techniques to deal with the significance, sensitivity and uncertainty of the results. Methodological D iscussion: W ater in LCA T he LCA community has moved from volumetric to impact oriented water footprints to better characterize the environmental impacts of water use (Berger and Finkbeiner 201 3 ). This has led to the development of a plethora of inventory and impact assessment methods as summarized in Kounina et al. (2013). Nonetheless, it is important to rev iew the evolution in the treatment of water in LCA to comprehend the role of this project within the existing body of knowledge. From Water Inventory Analysis to Impact Modeling Two concepts that have helped understand and communicate water impacts are vi rtual water (Allan 1993, 1994, 2003) and water footprint (Hoekstra et al. 2011) . The first represents the amount of soil water that it takes to produce a gricultural good s , and explains how countries can mitigate their water needs through trade 1 . The second expands on this idea by including the water used for manufactur ing products , and makes its spatio temporal use explicit by showing the volumes of water consumed b y sources at different time frame s (Hoekstra et al. 20 11 ). However, to understand the impact of water abstraction it is necessary to relate the volume of withdrawals to the available volume of the resource. In addition, other impacts must be considered, such as the impacts of land use cha nge on the hydrologic cycle, as these affect the natural ren ewability of the resource. 1 A the amount of the resource necessary to produce it.

PAGE 31

31 T he ratio of consumption to reserves or withdrawals to availability is one approach that has been used to assess the level of resource depletion (Lindfors et al. 1995). This has led to the development of impact ranges , such as th ose found in the water stress index 2 (UN 1997; WWAP 2006) and has help ed forecast the environmental outlook worldwide based on current trends (OECD 2008 , 2012 ; WWAP 2012). T his type of analysis uses modeling techniques that incorporate hydrological and wat er use components to determine the water balance in global river basins . One such example is the Water Global Assessment Prognosis or WaterGAP model (Döll et al. 2003; Alcamo et al. 2003a, 2003b). WaterGAP provides a comprehensive view of water use impact through a system that combines systematic water use and hydrological modeling at the river basin macro scale . It highlights the importance of understanding resource flows and defining appropriate system boundaries for water stress analysis. The use of riv er basins as a spatial unit is based on water mass balance analysis 3 . T his approach creates a link between resource use and pollution that is useful for water governance (OECD 2008). Its global scope matches the potentially global scope of LCA. However, it s spatial resolut ion is 0.5 degree x 0.5 degree 4 grid cells that aggregate consumptive use and the hydrologic partitioning of precipitation into run off and infiltration to calculate the ratio of 2 The categories of water stress are as follows: areas with 0 10% have low water stress; from 10 20% moderate water stress; from 20 40% represents medium high water stress; and over 40% is considered to i ndicate severe water scarcity (UN 1997). 3 The water mass balance analysis or water budget is based on the concept of equating the system inputs outle t and accounting for losses to the atmosphere through evapotranspiration. 4 The length of a degree of latitude ranges from 110.574 km at 0 degrees to 111.694 km at 90 degrees. The length of a degree of longitude ranges from 111.320 km at 0 degrees to 0 km at 90 degrees. A projected cell of 1 degree x 1 degree at the Equator would be approximately 110 km x 110 km. However, at 45 degrees latitude, the cell would have 110 km x 79 km in area.

PAGE 32

32 withdrawal to availability per cell. Because each cell repre sents approximately 6,000 sq. km. or 1.5 million acres, it is difficult to evaluate the impact of urban patterns of development without land include other types of environmental impact linked to changes in water availability that are necessary for a comprehensive impact assessment of water resources. Impacts from P roduct S ystems Life cycle impact assessment (LCIA) methods provide different indicators to understand the impacts in quantity and qual ity of in stream 5 and off stream 6 water use for production (Owens 2001), by user category (Boulay et al. 2011), and to explore technological options (Stewart and Weidema 2005) relevant to water scarcity. This has led to the development of a comprehensive f ramework for off stream use 7 that distinguishes from consumptive and degradative uses (Bayart et al. 2010; Ridoutt and Pfister 2012). LCA inventory analysis overlaps with water footprint in the quantification of resource use. However, LCIA goes further by modeling or assuming potential consequences. For example, by assessing the damage from domestic water scarcity through infectious diseases associated with the resource appropriation for the environmental and human health endpoints (Motoshita et al. 2011). Heuvelmans et al . ( 2005 a, 2005b ) regional water balance category to incorporate the impact s of land use and their corresponding flows as 5 6 1, 39) 7 W ater used off stream can return to that same system with a lower quality (i.e. degradative use) or represent a net loss to the system (i.e. consumptive use) through evapotranspiration or its transfer by product integration or discharge into anoth er watershed.

PAGE 33

33 affected by changes in land cover and withdrawal rates . The characterization model scale hydrologic model 8 to provide values for flood and drought risks based on stream flow statistics. In addition, the authors propose the us indicator to quantify resource depletion based on precipitation and water consumption. Although, this approach links both land use and water resources that had been treated separately in LCA, this work did not operationa lize the results into an assessment framework. In addition, Milà i Canals et al. (2009) criti ci ze d the extensive data and modeling requirements suggesting that it would be unfeasible for incorporation in practice. Nevertheless, Milà i Canals et al. (2009) recognized the role of land use on bio geographical conditions such as biodiversity, biotic production potential and ecological soil quality as they relate to freshwater availability and health in the . Van Ek e t al. (2002) have proposed the use of mid point indicators for groundwater levels and soil moisture content to assess the impact of groundwater extraction. These indicators are linked into an end gh an eco hydrological model 9 . The goal is to understand the environmental stress brought by desiccation on ground water dependent ecosystems. Similarly, Van Zelm et al. (2011) provide d the characterization factors (CF) addressing the impact of groundwater withdrawals on terrestrial vegetation richness for the 8 http://swatmodel.tamu.edu/ 9 DEMNAT (Dose Effect Model Nature Terrestrial) is a model developed in the Netherlands for the evaluation of dehydration or dessication of wetlands on plant life. Accessed on July 5 th , 2012: http://www.helpdeskwa ter.nl/onderwerpen/applicaties modellen/water_en_ruimte/demnat/literatuur/@21481/demnat 2_1_latest/

PAGE 34

34 Netherlands b y defining the alteration in the number of plant species in a given area as a result of groundwater withdrawals. The modeling effort was produced with U.S. el. Also, Saad et al. (2011) recognizes the role of land use in regulating erosion, freshwater partitioning into flows and storages, and water purification through the physicochemical and mechanical properties of ecosystems. These parameters are evaluated through a multivariate analysis of variance (MANOVA) model to understand the statistical significance of its results and the interactions among the selected variables. Although the model aims for site generic values, the characterization factors show signi ficant variability that calls for better spatial parameter differentiation. In other words, the parameters are site dependent. Work by Pfister et al. (2009) has tried to provide a holistic understanding of the environmental impacts of freshwater consumpti on through a modified water to availability (WTA*) ratio that accounts for monthly and annual variation of stress brought about by changes in precipitation. The spatial distribution of the water mass balance is provided by the WaterGAP model ( Döll et al . 2 003; Alcamo et al . 2003a, 2003b ) . The characterization factor is used to calculate the water stress index (WSI), assess impacts on human health, ecosystem quality and resource depletion. These damage factors are then integrated into the Eco indicator 99 10 f ramework and its endpoints. In addition, Pfister et al. (2011) uses the concept of land water trade off to frame the analysis of agricultural production globally. This assessment is based on a ratio of RED (relevant for environmental deficiency) water use to a proposed land stress index. 10 indicator th , 2012: http://www.pre sustainability.com/content/eco indic ator 99/

PAGE 35

35 The relevance of the analysis is in looking at land use changes as trade offs between different water requirements that are site specific. Tables 2 1 and 2 2 summarize the characterization factors (CFs) proposed by differen t authors for water and land use impact assessment. Each CF is classified under its level of site dependence. However, the actual geographical area of the spatial level is not as clear. As a testimony to the difficulty of providing appropriate data, some C Fs were calculated at a regional level, others at a local level. For instance, although Pfister et al. (2009) recognized the regionalization of hydrologic responses, the authors worked with national water inventories. As noted earlier, the statistical anal ysis of site generic CFs indicates that greater spatial differentiation is needed (Saad et al. 2011). Consequently, the approaches of Heuvelmans et al. (2005a, 2005b), Van Ek et al. (2002) and Van Zelm et al. (2011) demand a level of hydrological modeling with a higher spatial resolution than the one provided by a global model, such as WaterGAP2. At a glance, there seems to be a tradeoff between data requirements and modeling capabilities, but in the case of th e built environment, unless building products are the unit of analysis, the assessment of the building system in itself requires a site dependent approach. The Built E nvironment , a S ite D ependent A ssessment The definition of the system boundaries is cruci al for the understanding the impacts of the built environment on freshwater resources. For instance, Kenway et al. (2011) advocate for the analysis of cities as systems isolated from their surrounding environment because this facilitates the evaluation of urban performance and the comparison across metropolitan areas. The authors propose a n urban water mass balance limited to the urban footprint and maps input output resource flows . Thus, the

PAGE 36

36 city is seen as a water storage device, and its performance is ev aluated based on the volume that flows through it and/or remains inside. On the other hand, Kennedy et al. (2007) expanded the definition of the urban system and studied the relationship between a settlement and its groundwater system. According to the aut hors, urban growth promotes local aquifer contamination and pushes for the development of peri urban groundwater wells. Also, infiltration from irrigation or pipe leaks may increase the water table under the urban footprint above its pre development level, potentially causing flooding or damage to infrastructure. Other problems linked to overexploitation of groundwater are salt water intrusion, and land subsidence. However , this leaves the problem of understanding the environmental impacts of urbanization o n freshwater resources. Looking at urban areas in isolation misses the extent of the impact of land use change on the hydrologic cycle. The current satellite image based global land cover model with highest representation performance, MODIS 500m, shows tha t approximately less than 1% of the total land area is occupied by urban clusters with more than 100,000 inhabitants (Potere et al. 2009; Schneider et al. 2009). However, 43% of the total land surface has been transformed by humans for agricultural and urb an landscapes with the potential of reaching an ecological threshold where irreversible system changes might occur at a global scale (Barnosky et al. 2012). This makes it difficult to look at the impacts of the built environment in isolation from other lan d uses in a region or watershed. In addition , how can low density development patterns such as sprawl be analyzed? O r , what is the impact of single buildings on their hydrological context. These kinds of questions highlight the importance of providing an

PAGE 37

37 a lternative to the global water scarcity analysis provided by models such as WaterGAP2 that are limited in their spatial resolution due to their coupling with climate models. Challenges to LCIA of Water Resources There are two types of LCA based on how the impacts are quantified: an a ccounting and a change oriented analysis. The first adds the impacts to a single product system , much like the assembly of construction products to a single building. The second approach looks at the effects of a building or product system on a condition of interest , an impact category . Therefore, a major challenge for the inve ntory analysis lies in determining what impacts are considered and how to characterize them . Th e approach taken here is to treat th e impacts from land c over changes as a change oriented LCA and use a distance to target methodology. In order to measure the impact of a certain effect, it is necessary to describe or define a baseline condition for comparison. The hydrologic cycle is the natural process by wh ich water resources flow and move through the biosphere. Water falls from the atmosphere in the form of precipitation and can be separated into distinct components once it reaches the land cover. Water infiltrates the ground to recharge groundwater storage s, but once the soil saturates, it flows as runoff to feed surface water bodies. In this process, water is used by vegetation and is transpired, adding to the vapor created from insolation. These components are the partitions that make the water budget or mass balance. As a consequence, i n the selected model the baseline condition refers to the way the water budget is calculated based solely on climate and a defined land cover mix. However, the definition of the baseline and the analysis of the impact of ch anges in land cover on the water budget depend on the spatial and temporal scales of analysis. Klemes (1978) warns that hydrological processes are complex and the

PAGE 38

38 relevancy of physical variables depends on the spatial scale of the study. Moreover, the effe ct of those variables does not translate linearly across space. Similarly, the selection of temporal scale depends on the aim of the analysis. For instance, run off generation is usually stud ied at an hourly or daily time step. However, the water balance o f a basin depends on the assumption of an equilibrium condition that may take years to establish and therefore is best analyzed at an annual time step. In addition , the definition of system boundary for hydrological studies is a major challenge, illustrate d in the considerations surrounding the myth of a sustainable groundwater yield (Bredehoeft 2002) . the virgin recharge before development is important in determining the magnitude of sustainable development resources as a metric of sustainable use because the response of an aquifer to withdrawals cannot be accounted for simply through a mass balance approach, and will depend on its boundary conditions and its system properties (Alley et al. 1999; Bredehoeft 2002; Sophocleus 2000; Theis 1940). Alley et al. (1999) go further when is of limited value in determining the groundwater systems under predevelopment conditions are in a long term equilibrium of recharge and discharge, and changes in the syst em balance require adjustments that cannot be predicted based on the initial water budget. This only indirectly provides the amount of perennial discharge available that can be captured. In addition, groundwater systems serve both as reservoirs and distrib ution systems connected to surface water

PAGE 39

39 bodies, such as lakes or rivers. Changes to the flow system after development can take three forms: increases in the recharge or decreases in the discharge, removal of water stored in the system, or a combination of duration of withdrawals are important to understanding how the aquifer system responds, and that could mean changes in the direct ion of flow as surface water bodies flow to aquifers to balance water that is being withdrawn. Consequently, a system modeling perspective is fundamental in the development of water resource indicators. Outside of the LCA community, it is understood that p atterns of land development modify the hydrologic connectivity within drainage basins (Yang et al. 2011). These changes come in part due to alterations in chemical and hydraulic soil properties (Scalenghe and Marsan 2009) and by physically covering the gro und with buildings and pavement. An increase in impervious areas translates into surges in runoff that alter stream flows over time and impair water quality, exacerbate flooding, and also reduce aquifer recharge by limiting infiltration. In addition, reduc tion in soil water aggravates drought conditions. As a consequence, the planning community considers the ratio of impervious surfaces to total area as an impact indicator with a threshold for the evaluation of land development (Arnold and Gibbons 1996; Mog len and Kim 2007). Studies such as Haase and Nuissl (200 7 ) support the link between surface sealing, spatial patterns of development such as sprawl, and hydrological alterations. These cause effect chains need to be translated into LCIA.

PAGE 40

40 Modeling Approach Goal and Scope Definition The goal of this project is to understand the impact of changes of land cover on the hydrological cycle as quantified in the water balance. To this end, it is necessary to define a baseline condition by a particular mix of land co vers within a watershed . Th is s mass balance and allow the correlation of its behavior in time based on how the mix of land covers change. The model should be capable of forecasting catchment behavior by unders tanding the effect of urban areas in relationship to other land cover classes in the generation of water balance. Life Cycle Inventory Analysis This include s an accounting of water flows based on land cover and other characteristics in this work . Kenway et al. (2011) isolate the urban system from its surrounding environment because the authors want to understand the water exchange between the city and its surrounding environment and facilitate the accounting of these flows . On the other hand, Kennedy (2012) considers impact of the urban system on the aquifer without clearly defining the boundaries of the groundwater system . This project builds upon the work of Chhabra (2011), who developed an urban water flow model (see Figure 2 6) that includes water use a nd rainfall runoff generation pathways. This model is a conceptual derivation from earlier hydrological models that separate water flow into storage compartments, such as the explicit soil moisture accounting ( ESMA ) rainfall runoff model from onnell (1965) also known as the Stanford Watershed model .

PAGE 41

41 The LCI water flow model assumes a mass balance approach where the long term equilibrium in a basin is given by equation 2 1, where P stands for precipitation, Q for streamflow, E for evapotranspira ( 2 1 ) (2 2) (2 3) Human intervention is seen in the streamflow and changes in storage compartments as precipitation and evapotranspiration depend largely on climatic variables that are not directly under the control of society. Streamflow (Q) in equation 2 2 results from ba se flow (D AQ ) and the net discharge (D N ) from the system, which includes run off (D P ), stormwater runoff and wastewater discharge (D U ) minus surface water withdrawals (W S ). It is important to clarify that runoff is differentiated between that generated in urban areas (i.e. stormwater) and that resulting from rainfall in other land cover classes so as to facilitate the understanding of water flows in the urban system later on. In addition, changes in land use and land cover produce change s in water storages . This is represented in equation 2 3 as the result of recharge from precipitation or infiltration (R P ), urban recharge from leakage in water supply and sewage conveyance systems (R U ), groundwater withdrawals (W AQ ), baseflow (D AQ ), and inflows (GW I ) an d outflows (GW O ) from the groundwater systems that cannot be accounted for using the drainage basin system boundaries. The myth of sustainable groundwater yield is based on the misunderstanding that GW I and GW O flows can be ignored, which makes it difficul t to describe the impact of changes in storage as groundwater would flow from outside the system to replenish

PAGE 42

42 withdrawals but at rates based on the geomorphology of the system . This represents a major challenge as it could mean impact shifting as the effec ts of withdrawals could be perceived outside the basin under analysis. Moreover, this could increase the data required to model the process in LCA studies. Therefore, for this project the impact will be evaluated at the streamflow level, where surface runoff and baseflow merge, and concentrate on the impact brought about by changes in land cover. Impact Asses s ment Classification This involves the definition of water flows, natural and man made for the inventory analysis so as to relate through the cause effect model. Characterization (model) The unit of analysis or functional unit for this project is the proportion of land cover classes to total basin area. This wil l help relate the presence of a land cover classes against each other . Cause effect chain The proposed causal relationship between land cover and the water balance is illustrated in Figure 2 8. The diagram shows the cause effect chain within the ecosystem region framework . Land development activity alters the distribution of the land cover clas ses that make up a catchment. The impact of the changes in land cover depends on the morphometric variables or physical characteristics of the catchment. The goal is to develop a runoff ratio that can be determined by a catchment coefficient (see Chapter 3 ) estimated with hydrological model .

PAGE 43

43 System Elements The selected geographic framework for this analysis is the ecosystem region or ecoregion as defined by ecosystems or relationships among organisms and understanding of the impacts brought about by the built environment (Drummond and Loveland, 2010; Napton et al. 2010; Sleeter et al. 2011) . Within this framework, the watershed is the representative unit of analysis (M agnusson 2001; Omernik 2003) which will allow the measurement of impacts based on changes on the water mass balance, or the process of partitioning rainfall into runoff and evapotranspiration . The built environment is represented by the urban land cover, w hich has different ranges of impervious surface area (ISA) based on its population density . As naturally vegetated land cover is substituted for other land uses, more impervious surface area is built to connect and support infrastructure systems, and water use increases. Figure 2 7 shows the hypothetical relationships between ISA and water use to the surface and groundwater storage compartments. Increases in ISA augment runoff but decrease the corresponding evapotranspiration and infiltration. Meanwhile, runoff and infiltration processes increase surface and groundwater, while evapotranspiration reduces both compartments. On the other hand, water use increases withdrawals and, discharge and injection. The former could increase surface and groundwater storages while the latter reduces them. In addition, increases in ISA produce hydrologic alterati ons to stream flow that have an impact on ecosystem health , and studies have classified the impact thresholds based on percentages of ISA and the configuration of the stormwater system (Roy and Shuster 2009) . These thresholds can include physical and chemi cal parameters such as

PAGE 44

44 channel stability (Booth and Jackson 1997), geomorphic response patterns (Cianfrani et al. 2006), water chemistry (Walsh et al 2005) and alterations to summer baseflows (Finkenbine et al. 2000) . Also, alterations to biological parame ters have been detected that include changes in algal biomass (Walsh et al. 2005) and decline in macroinvertebrate diversity (Stepenuck et al. 2002) to name a few. For more information visit the US EPA Causal Analysis / Diagnosis Decision Information Syste m (CADDIS) website ( http://www.epa.gov/caddis/ssr_urb_is4.html ). The variables that will be studied in this research project are primarily land cover classes such as urban and forested. The intermediate variables are precipitation and evapotranspiration as these will determine the water balance of a given climate zone and would help characterize the spatial variability of the model. Dependent variables: Water availability in the form of the run off ratio: streamflow to precipitation (water mass balance). In dependent variables: The ratios of land cover classes to total system area (ex. urban area (URB), population density as a proxy for urban areas, and morphometric variables to explicitly describe the physical characteristics of the basins. Intervening varia bles: (affect the cause effect chain): Precipitation, atmospheric demands (evapotranspiration), and catchment morphometric variables. Hydrologic al Model The relationship between the se variables is based on a water mass balance . This approach determines to a certain extent , the characteristics of the spatial and temporal analysis. The hydrologic al model ling approach for this study is based on the Budyko framework, a n empirical relationship between precipitation (P) and evapotranspiration (E) that allows for the calculation of runoff (Q) as the difference

PAGE 45

45 between precipitation and evapotranspiration . E quation 2 1 follows the work of Fu (1981) as presented by Zhang et al. (2004) . An important contribution of this model is the use of a catchment coefficient ( ) that represents the physical characteristics that drive the rainfall runoff process. This is a lumped model because it treats the basin as a single unit and does not differentiate in its elements or their spatial distribution. This approach was preferr ed to other more detailed model s based on its parsimony data requirements , and assumptions. Also, the selected formulation of the Budyko framework provides a deterministic solution but can include output variability based on the probability distribution fu nction of the input variables. The model is explained in detail in Chapter 3. ( 2 1) Spatial scale The spatial context of this study is the ecosystem regions proposed by the U.S. Environmental Protection Agency as a framewor k for ecosystem management (Omernik et al . 2011 ; Wiken et al 2011 ). Each region is defined by common geographical characteristics, such as biota, topography and climate. Within this framework, the watershed is the system of analysis because it facilitates the measurement of the water mass balance in time through the estimation of precipitation and the validation of its partitioning by gauged streamflow . I t has been found that there is a range of spatial area where ISA has a noticeable an effect on streamflow . Randolph (2004, 257) has shown that in drainage basins with an area of 26 to 260 square kilometers (10 100 square miles) , the study will include representative drainage basin s with in that range . Table 2 3 includes

PAGE 46

46 the tiered spatial hierarchy for the use of watersheds as management units. As a cautionary note, it is important to remember that hydrologic unit levels organize basins into units with similar areas. However, only approximately 45% of hydrologic units are true drainage basins (Omernik et al. 2003, 2011). Temporal scale The assumptions in the Budyko framework require average annual data of climate variables . The period of study will depend on the availability of land cover and climate data, but should expand a representative horizon to detect the impact of land cover on the hydrological cycle. Summary The proposed characterization model of the impacts of the built environment on freshwater resources depends on the cal culation of the annual water mass balance within a watershed. The Budyko hypothesis provides a conceptual framework for the calculation based on the limits of water supply by precipitation and water demand expressed in the evapotranspiration potential. The model will build upon the category be based upon the ratio of land cover classes to total basin area. In conclusion, t he aim of the study is the definition of a LCA mid point model of the impacts of the built environment on freshwater resources based on the definition of an ecoregional characterization baseline and life cycle impact assessment thresholds of streamflow availability as a distance to target baseline .

PAGE 47

47 Table 2 1. Water use characterization factors within Abiotic Resources in terms of their availability for future generations Category Indicator Site dependent attribute Site generic attribute Reference Water availability and scarcity No Yes Bauer and Zapp (SETAC 2005) Dynamic water reserve life (DWRL) Yes No Heuvelmans et al. 2005 a Water stress index (water deprivation) Yes ? Pfister et al. 2009 Relevant for Environmental Deficiency (RED) water Yes No Pfister et al. 2011 Consumptive water use (CWU) Degradative water use (DWU) Yes ? Ridoutt and Pfister 2012 Table 2 2. Land use characterization factors related to water use : land occupation (area*time), and land transformation (changing quality per area unit) Category Indicator Site dependent attribute Site generic attribute Reference Groundwater extraction (dessication) Yes No Van Ek et al. 2002 Regional water balance Yes No Heuvelmans et al. 2005 a Groundwater function and run off regulation Yes No Pennington et al. 200 4 Erosion regulation potential (ERP) Freshwater regulation potential (FWRP) Water Purification Potential (WPP) Yes Yes Saad et al. 2011 Biodiversity Biotic production potential Ecological soil quality Freshwater depletion (FD) Freshwater Ecosystem Impact (FEI) ? ? Milà i Canals et al. 2007, 2009

PAGE 48

48 Table 2 3. Characteristics of watershed management units HU level HUC Watershed management unit (HU) Typical area (square km) Influence of impervious surfaces Sample Management Catchment < 1.3 Very strong Practices and site design 6 th 14 digit Subwatershed 2.6 to 26 Strong Stream classification and management 5 th 11 digit Watershed 26 to 260 Moderate Watershed based zoning 4 th 8 digit Sub basin (cataloging unit) 260 2600 Weak Basin planning 3 rd 6 digit Basin (accounting unit) 2,600 26,000 Very weak Basin planning 2 nd 4 digit Subregion 1 st 2 digit Region Note: adapted from Randolph (2004). Figure 2 1 . Schematic diagram of the phases in an LCA (ISO 14040 1997)

PAGE 49

49 Figure 2 2 . Life cycle impact assessment based on ISO 14042 (Baumann and Tillman 2004). Figure 2 3 . Schematic presentation of LCIA modeling (Adapted from Finnveden et al . 2009)

PAGE 50

50 Figure 2 4 . Life cycle inventory water flow model ( Adapted from Chhabra 2011) Figure 2 5 . Hypothetical relationship of v ariables to c atchment s torages GW I GW O

PAGE 51

51 Figure 2 6 . Cause effect chain showing the r elationship between land cover classes and morphometric variables on the catchment coefficient resulting run off ratio

PAGE 52

52 CHAPTER 3 IMPACT ASSESSMENT OF LAND COVER CHANGE ON STREAMFLOW WITHIN A WATER ENERGY BALANCE FRAMEWORK Background The purpose of this study is to assess the impacts of land cover changes on streamflow. The process by which precipitation partition s i nto run off and evapotranspiration within a drainage basin is referred to here as catchment behavior . This follows the assumption that the basin as a system is under a long term equilibrium where infiltration and base flow are in balance and can be ignored . This equilibrium , and deviations from this reference behavior can be attributed to impacts in the rainfall runoff generation process from changes in land cover . T he first step in this study involves the hydrologic modeling of selected drainage basins under the Budyko framework which describes catchment behavior as equilibrium between water supply from precipitation and water demand from evapotranspiration and resul ting streamflow. This process is also known as a water energy balance framework and uses a catchment coefficient in its empirical formulation to define the physical attributes of the basin . Secondly, a trend analysis is performed to detect if there are sta tistically significant changes in the catchment behavior during the study period , and if those changes can be correlated to changes in land cover. The exploratory hypothesis is : f a basin exhibits changes in its land cover, the impact of that change on t he water mass balance will be reflected in the catchment coefficient. This chapter describes the corresponding research design to test this hypothesis. The methodology is based on the regression analysis of the changes of different land

PAGE 53

53 cover classes on catchment behavior. The sample is drawn from basins found in the South Eastern Coastal Plains ecoregion (level III). Literature Review Representation of C atchment B ehavior The Budyko hypothesis states that the rate of evapotranspiration is the result of t he equilibrium between atmospheric water supply (precipitation) and water demand represented by potential evapotranspiration. (1964) and Budyko (1958) have proposed formulae for the mean annual water energy balance. However these formulations are unable to capture the role of catchment characteristics on the mass balance. Roderick and Farquhar (2008) identified two theoretical threads that have given shape to the water energy balance for a drainage basin and their cor responding empirical formula e that have introduced landscape characteristics through a dimensionless parameter . The first thread has taken the generalized form given in equation 3 1 by Choudhury (1999; Bagrov 1953; Mezentsev 1955; Turc 1954; Pike 1964; Mi lly and Dunne 2002). The alternate formulation was proposed by Fu (1981) and presented by Zhang et al. (2004) in equation 3 2. (3 1) ( 3 2 ) Both equations have been found to be almost mathematically equivalent by Yang et al. (2008) and . The relationship between water demand and supply can be described graphically in Figure 3 1 .

PAGE 54

54 Milly (1994) sought to establish the control of soil wat er storage and climate seasonality in the water energy balance. Zhang et al. (2004) approximated the value of the catchment parameter to the major vegetated land cover (i.e. forested and grass land ) of several basins, the authors found it difficult to defi ne the parameter a priori . Roderick and Farquhar (20 11 ) delve d in to the meaning of the catchment properties parameter or coefficient to relate variations in run off and climatic conditions for the water balance formulation proposed by Choudhury (1999) . For instance, the aridity index as a ratio of potential evapotranspiration to precipitation was used by Ponce et al. (2000) to classif y climate types as shown in Table D 1 . Moreover, Roderick and Farquhar (20 11 ) differentiate between properties that will rema in nearly unchanged in time, such as geologic and topographic properties, from properties that can change easily, as vegetated land cover. In addition, the authors precipitation intensity or changes in the spatial distribution and/or seasonal timing of P and E 0 Donohue et al. (20 07 ) considered the role of vegetation through evapotranspiration as a major driver in the mass balance as well as soil properties and topography. Oudin et al. (2008) explored the inclusion of differentiated v egetation land cover classes in existing water balance models with improved results on small and wet basins (<10,000 sq. km). However, the authors commented that improvements in model performance were not consistent across the sample . Consequently, there seems to be a general agreement that a better understanding of the catchment parameter could lead to the improvement of hydrological model performance.

PAGE 55

55 For the purposes of this investigation the formul a tion proposed by Fu (1981) and described by Zhang et al. (2004) in equation 3 2 is selected to study the effects of land cover i n the mass balance model. The catchment coefficient has a range of [1, ]. The selection of this approach was based on the poten tial use of this formula at inter annual and monthly time steps according to Zhang et al. (2008). In addition , the evapotranspiration ratio (E/P) calculated with equation 3 2 is translated into a relationship that is more meaningful to describe the impact of the built environment when presented as a ratio of runoff to precipitation (Q/P) , also known as the runoff ratio which provides a measure of the availability of surface water to satisfy the needs of the ecosystem, including the built environment . This r elationship is derived from the steady state water mass balance equation 3 3 , where P stands for precipitation, Q represents , considered negligible given long term equilib rium . Unfortunately, the literature is not specific in what can be considered long term in the hydrologic balance of a basin. Therefore, this study will calculate the catchment coefficient using the moving average of the climate variables, precipitation a nd potential evapotranspiration, at different time steps , namely 2, 3, 4, 5 and 10 year moving averages. ( 3 3 ) relationship between Q/P and the catchment coe represents the most common ratio of the selected meteorological variables for Florida (sub humid climate type) according to Ponce et al. (2000) is shown in Figure 3 2.

PAGE 56

56 Figure 3 2 shows that Q/P can be understood as a power function of the , i.e. aridity indices , can be evaluated as described in Table D 2 . T he relationship between Q/P and the catchment coefficient can be described with s ingular power functions. Understanding this behavior can be valuable in the definition of thresholds of catchment behavior. The exploratory hypothesis is further specified to include the use of the Budyko framework for the hydrological modeling. Hypothes i s If there is a non linear inverse relationship between the catchment coefficient and runoff, then, a proportional increase within a basin of a land cover class that promotes run off should decrease the catchment coefficient to reflect the alteration in the water balance. Methodological Framework The study explore s if the changes in the water balance in time as represented by the catchment coefficient ( ) are correlated to changes in land cover. T he impact of y, soil characteristics, and morphometric variables are considered in the analysis. . To that end, a model that combines l ongitudinal and cross sectional stud ies is proposed for the evaluat ion of sample basins in the South Eastern Coastal Plains of Florida (ecoregion level III) . A panel data regression or pooled time series analysis model is used . Basic Statistical Model and Assumptions The model seeks to explain the process rainfall partitioning in each catchment at two levels: first, as the process relat es to the changes composition in time , and second, how do changes in land cover have an impact across

PAGE 57

57 different catchments. To this end, a general panel data regression model a s described by equation 3 4 is proposed for the an alysis of the effects of land cover change on the catchment coefficient. (3 4 ) Where n = represents the cross sections , i.e. sample basins; t = are the time intervals , i.e. , annual; and k = are the explanatory variables used . The use of a panel data model presents important challenges that include the violations to the assumptions that support linear models. For instance , the inclusion of stochastic and non stochastic variables in the model result in non zero expected error mean and non constant variance (Sayrs 1989). This leads to contamination by correlation of the error terms between different cross sections at the same time periods or/and at different time periods. As a consequence, Sayrs (1989) p resents four models that could deal with these challenges based on different assumptions about the error term : a constant coefficients model, a least squares dummy variable (LSDV), an errors components model, and the structural equations model . The constan t coefficients model assumes that the regression coefficients are the same for all the cross sections. In addition, the model must not present auto regression and non constant variance ( heteroscedasticity ) of the residuals, as the former will produce biase d estimates while the latter will lead to inefficient estimates. This model theoretical assumptions. The LSDV model assumes the problems brought by non constant variance and heteroscedasticity by introducing a fixed effect value that represents sample uniqueness and using a dummy variable to condition the variance.

PAGE 58

58 T he random effects model assumes random error in time and space and not unique to either one to produce unbiased and efficient estimators. The advantage of the random effects over the fixed effects model is its universality subject to a representative sample; the fixed effects model is unique to its sample. However, the random effects model requires a stronger theor etical base to avoid meaningless estimators, a problem that could also affect the other models but to a lesser exten t . Finally, the structure equation model is based on a system of equations to deal with endogenous lagged variables and the problems associa ted with correlated errors. For the purpose of this project, the preferred method of analysis is the random effects model , as the objective is to develop a regression equation to explain the effects of land cover classes and catchment morphometric variable s on the catchment coefficient for a sample of basins in South East ern Coastal Plains ecoregion. Nevertheless, OLS and fixed effects models will be considered as alternatives representations of the processes under analysis. These three methods are available in the LIMDEP 10S software package under panel data models. Dependent variable The c atchment coefficient represents the physical characteristics that control the partitioning of the rainfall runoff process . I t is calculated by optimizing the wate r energy balance using equation 3 2 subject to annual streamflow records. This process is explained in detail in section . Independent v ariables The initial hypothesis assumes the relationship between land cover cl asses and the catchment coefficient. In addition, population density is another variable that represent s land occupation and modification. The regression analysis seeks to identify

PAGE 59

59 the correlation between changes in these variables and the catchment coeffi cient. Moreover, the influence of these variables on the catchment coefficient could depend on other characteristics that do not change in time such as the morphometric variables. The list of independent variables used in this study is presented as follows explained in greater detail in the following sections. Land cover classes: This variable uses the Anderson Level 1 Class codes as follows: Open Water (WAT): Groups all areas of open water with less than 25% vegetation or soil cover. Urban areas ( URB): Aggregates areas with different percentages of impervious surface areas, and includes residential, commercial, and industrial land uses. Barren areas (BAR): Classifies areas of bedrock or accumulated earthen material that have less than 15% vegetatio n. Forested areas (FOR): Groups areas with greater than 20% vegetation cover and dominated by trees taller than 5 meters. Grassland areas (GRA): Represents areas dominated by herbaceous vegetation and shrubs less than 5 meters tall. Planted/cultivated (AGR ): Groups cultivated crops and pasture/hay. Wetlands (WET): Groups woody wetlands and herbaceous wetlands. Population Density (PD): This parameter represents people per unit area. It is calculated from census block groups using 2010 census block group boun daries. The spatial boundaries for evaluation are the drainage basin area and urban area.

PAGE 60

60 Morphometric parameters: This is a select group of representative physical attributes of the basin. Th e s e variables are time invariant Land cover classification In or der to understand the effects of changes in land cover on the catchment coefficient , a geographic information system (GIS) is used to visualize and compute changes in land use/land cover in time within the boundaries of the selected basin s . The land use/la nd cover classification system employed is based on the modified Anderson Level 1 class codes for Florida (Anderson et al. 1976; FDOT 1999). T he following land cover variables were selected from the U.S. Geological Survey Enhanced Historical Land Use and Land Cover Data for the State of Florida 1970s (Price et al. 2006), National Land Cover Database products NLCD 1992 retrofit (Fry et al. 2009), NLCD 2001 (Homer et al. 2007), NLCD 2006 (Fry et al. 2011), and NLCD 2011 (Jin et al. 2013). Table D 3 summarizes the equivalency between land cover classes used in this study between the 1992 and 2001 2006 databases. Furthermore, land use/land cover maps for the State of Florida have been published approximately every 5 years since 1990 by its five water management districts. Each district covers only the area within its jurisdiction, which does not necessarily coincide with the county boundaries. These maps are the product of digitizing a e r i al representations with graphic al scales between 1:8,000 and 1:12 ,000. T he water management districts started implementing a quality assurance program for their land cover products in 1995. The maps that pre date 1995 could add uncertainty into any land cover comparison study. This is partly the reason that this project selected the NLCD products as they are the more consistent land cover products for temporal studies. Nonetheless, the time scale is limited to 1992 2011 by the products available.

PAGE 61

61 For this reason, the state wide 1974 land cover map created by the Univers ity of Florida GeoPlan Center , which uses the USGS Enhanced Historical La nd Use and Land Cover Data Sets was used . The analysis of catchments outside Florida required the direct download of data from the USGS website (a ccessed on May 13 th , 2014: http://water.usgs.gov/GIS/dsdl/ds240/index.html ) . The original scale of these map s is 1:100,000 and 1:250,000, and not only its methodological differences but its resolution adds uncertainty to the cur rent study of land cover changes. However, the possibility of extending the temporal timeframe for this study makes it an acceptable trade off. Land cover variable normalization In order to facilitate inter catchment comparison, the selected variables ( LC i ) are normalized by dividing the area of each land cover variable (A i ) by the basin area (A B ), i ( 3 5 ) Population density Population density is proposed as a proxy of urban development . The variable is calculated by dividing the population in the basin (POP B ) by the area in the basin (A B ). The population of the basin is the weighted sum of population per census block group (POP B Gi ) contained within the basin. The weights are the ratio of the area of the block group contained within the basin (A i BGi ). The calculation is given by equation 3 4. The parameter is calculated for years 1980, 1990, 2000 and 2010 based on the availability of census data with the same block group boundaries for year 2010 (Geolytics 2013). Further details on the calculation procedures are found in Appendix B.

PAGE 62

62 (3 6 ) Similarly, the urban population density (PD U ) was calculated as a ratio of population to urban area (A URB ) per basin. This was done to account for the differences in urban use intensity and impervious surfaces. Morphometric parameters The variables or predictors to include in the regression analysis are selected fro m a pre Geospatial Attributes of Gages for Evaluating Streamflow dataset (Falcone et al. 2011). This dataset provides geospatial data and classifications for 9,322 stream gages maintained by the USGS . These variables were selected based on their relative importance to catchment behavior ( S plinter et al. 2011; Costa 1987 ; Patton and Baker 1976 ; Chow 1964 ) and are defined as follows: Drainage area (Area): Watershed drainage area in Sq. Km. Base flow index (BFI): Base Flow Index (BFI), The BFI is a ratio of base flow to total streamflow, expressed as a percentage and ranging from 0 to 100. Base flow is the sustained, slowly varying component of streamflow, usually attributed to ground water discharge t o a stream. Basin compactness (Compact): Watershed compactness ratio, = area/perimeter^2 * 100; higher number = more compact shape. Drainage density (DrainDen): Stream density, km of streams per watershed Sq. Km., from NHD 100k streams. Permeability, avera ge (P ermAve ): Average permeability (inches/hour).

PAGE 63

63 Relief ratio, mean (RR_mean): Dimensionless elevation relief ratio, calculated as (mean elevation minimum elevation)/(maximum elevation minimum elevation). Slope percentage (S lope ): Mean watershed slo pe, percent. This parameter is d erived from 100 m resolution National Elevation Dataset, so slope values may differ from those calculated from data of other resolutions. Circularity ratio (CircRatio): Ratio of basin area to the area of a circle with the sa me perimeter as the basin. This variable was calculated independently as it is not found amongst the Gages II dataset. Regression Procedure For the evaluation of land cover to catchment coefficient over time, a hypothesis of how the land cover types will i nfluence run off is generated . Figure 3 2 shows the hypothetical non off to (E 0 /P) as given by the Budyko framework . In ot her words, as the catchment coefficient increases, run off decreases. Therefore, this reasoning should guide the selection and arrangement of the land cover variable s for this study. Urban areas with varying percentages of i mpervious surface areas (ISA) wi ll be studied . It is expected that run off will be positively correlated to ISA , and thus, to urban areas . Similarly, barren and grassland areas will produce more run off to stream flow than forest, wetlands and/or open water areas. These last three types of land cover reduce the transform ation of precipitation into run off either because the forest canopy intercepts rainfall and promotes evapotranspiration or wetlands and open water areas serve as storage areas for precipitation and run off. This reduces the contribution from these land types to s tream flow. As a consequence, t wo aggregate groups of land cover

PAGE 64

64 types can be proposed based on their contribut ion to run off and stream flow: urban barren grassland ( U BG), and forest wetland water (FWW). On the other hand , t he urban areas (URB) and agricu ltural lands (AG) represent unique cases. Urban development creates impervious surfaces that increase run off but also withdraw water from surface or groundwater storages, thus altering streamflow. C rops do not normally have such large canopies for rainfal l interception but their intensive water use and configuration could make them a land cover type that is structured to transfer water through evapotranspiration via crop growth. Table 3 1 summarizes the hypothetical influence on run off generation by land cover classes . As a starting point, AG, UBG and FWW land cover groups will be used in the regression analysis. Also, URB and population density will also be explored in the regression analysis in order to get a more nuanced understanding of the impacts of the built environment. Analytical scenarios Two scenarios were analyzed based on the available land cover data (see Table 3 2) . The first included the linear interpolation of land cover data from aerial photographs from 1974 with the satellite data from 19 92, 2001, 2006 and 2011. The second scenario sought to reduce the uncertainty associate with mixing land cover products and limited its period of anal ysis between 1992 through 2010. 1. 1980 2010: Land cover (NLCD & USGS) datasets & population density 2. 1992 2010: Land cover (NLCD) datasets & population density Calculation of catchment coefficient The catchment coefficient ( ) is calculated by finding the maximum optimal solution to the ratio of actual evapotranspiration to precipitation in equation 3 2 by

PAGE 65

65 changing the values of , and assum ing long term equilibrium with no changes in storage as represented in equation 3 3 . 3 2 Generalized Reduced Gradient (GRG) Nonlinear solvi ng method. The data requirements are listed in Table 3 3 . The optimal solution was sought on an annual basis over 2,3,4,5 and 10 years of average climate data, namely potential evapotranspiration (E 0 ), precipitation (P) and run off (Q). This also means tha t the regression analysis requires an additional number of years of flow records equal to the selected number of years included in the moving average. For instance, to calculate the catchment coefficient for scenario 1980 2010, at least 40 years of flow re cords are necessary: 31 years for the analytical scenario and 9 additional for the calculation of the initial climate data average. A detailed description of the extraction and processing of climate data is presented in Appendix A. Trend A nalysis The annua l time series of the catchment coefficient, land cover types, and the process helps identify which catchments to include in the regression analysis. Case S tudy P re S election The calculation of the catchment coefficient requires at least 10 years of climate data to precede the actual calculation of the coefficient. Therefore, for the first analytical scenario which includes land cover data from 1980 to 2010, the basins require at least 40 years of flow records that span from 1970 to 2010. For the second analytical scenario (1992 2010), a minimum of 30 years (1980 2010) of flow records is necessary. In addition, these basins should not have dams that could alter their stream flo w by regulating flows and/or withdrawing surface water. Also, this sample of basins will be in

PAGE 66

66 the same level III ecoregion , namely the South Eastern Coastal Plains and in Florida ) to control the influence of patterns and composition of biotic and abiotic phenomena on catchment behavior. The sample of case studies was selected from the GAGES II dataset (Falcone et al. 2011). The selection was based on available flow records and a provision that these basins will be located in the South East ern Coastal Plai ns level III ecoregion and would not have dam structures that could alter natural streamflow. The flow record had to include at least 40 years of continuous records from 1970 to 2010. In addition, the potential case studies were further screened for commen ts on the presence of hydrological alterations by the USGS National Water Quality Assessment Program (NAWQA) personnel in the Annual Data Report (ADR) used in the GAGES II dataset. Results and Discussion This section explains the rationale in the case study selection, the exploratory analysis of the selected basins, and the results of the regression analysis. Case Study Pre Selection A ssessment Th e pre selection process found 2 6 catchments that comply with the minimal 40 years of f low records and absence of dam structure s . Fourteen of those catchments do not show the presence of hydrologic modifications according to NAWQA , such as stream flow diversions by canals , and are used in the exploratory analysis. In addition, five basins wi th similar characteristics and a minimum of 30 years of continuous flow records were added to expand the sample size. Table D 4 contains the list of the twenty one pre selected basins with their flow records correlation coeff icient for the catchment coefficient and UBG in respect to time for the two analytical scenarios . A discussion of the initial findings follows.

PAGE 67

67 Exploratory analysis The sample of basins w as analyzed to identify changes in the catchment coefficient in time and their correlation with changes in land cover. A detailed analysis of each basin was performed and is included in Appendix C. This exploration consisted in performing a trend analysis to test the initial hypothesis of a negative correlation between urban land cover and the catchment coefficient. The calculation of the catchment coefficient us ed different water balance annual moving averages for 2, 3, 4, 5 and 10 years. The 10 year moving average proved the most stable f or the calculation of the catchment coefficient as can be seen in Figures C 1 through C 19. Also, the areas of the land cover classes were extracted from the selected maps and the ratios of land cover class to basin area are included with the annual value of the catchment coefficient and population density in Tables C 1 through C 19 for each of the basins. In addition, to account for the possible effect of morphometric differences in the basins on the water balance , the time invariant parameters for each ca tchment are presented in Table D 5 . After the calculation of the catchment coefficient, two basins were excluded from the sample. Orange Creek at Orange Springs b asin ( USGS 02243000 ) has an incomplete record of streamflow between 1971 and 1975 which result s in a reduced series of the coefficient from 1985 through 2010. Also, the optimization procedure did not produce viable results for Fenholloway River basin near Foley ( USGS 02324400 ) . Th e s e catchments are not included in Appendix C. The next step in the a nalysis was the identification of outliers as presented in Figure 3 4 , which shows the catchment coefficient for the selected nineteen (19) basins. A clear outlier in the graph is the South Branch Anclote River basin near Odessa

PAGE 68

68 (USGS 02309848 ) , with values in its catchment coefficient that range between 6 and 18 versus the sample median of 3.5 . Furthermore, the sample has an excess kurtosis of 13.31 and a skewness of 3.18 which signals that the variance could result from extreme deviations . However, w hen South Branch Anclote river basin is excluded, excess kurtosis becomes 0.04 and the skewness 0.38. Thus, the exclusion of the outlier approximates the distribution of the catchment coefficient to a normal distribution. Table D 6 summarizes the descriptive statistics calculated by MS Excel for the two samples. The resulting sample has eighteen (18) basins for trend analysis. Trend analysis The scatter plots of the catchment coefficient time series attached in Appendix C showed varying degrees of regression line fit across the sample . The Pearson correlation coefficient (parametric test) was used to explore the relationship proposed in the hypothesis between the catchment coefficient and UBG by analyzing the correlation between t he trends of the two parameters in time. According to the hypothesis, the parameters should have opposite trends ; for example, a positive trend in UBG should be accompanied by a negative trend in the catchment coefficient . Table D 7 and D 8 show the result s of the Pearson correlation coefficient between time (Year), the catchment coefficient ( _10) , the normalized land cover classes , and population density for the eighteen (18) basins in the scenarios of 1980 2010 and 1992 2010 respectively. The results are not clear across the two scenarios. For the period of 1980 2010, six (6) basins presented the expected behavior of opposite significant trends in the catchment coefficient and UBG, and eight (8) basins showed significant trends of both

PAGE 69

69 para meters moving in the same direction. For the period of 1992 2010, ten (10) basins showed the expected behavior while three (3) showed the opposite behavior. Moreover, the trend in the catchment coefficient changed direction significantly within the same ba sin between the two scenarios. This led to explor ing the potential phenomen a affecting the behavior of th e catchment coefficient in time. A sub group of nine basins with flow records greater than 50 years was selected for the analysis . T he statistical anal ysis of the time series of catchment coefficients were studied at five time periods and are shown in Table 3 4 : the complete timeframe of analysis between 195 9 and 2010, which is then split into time periods 1959 1980 and 1980 2010. Additionally , a closer look is provided for the 1980 1992 and 1992 2010 scenarios . Between 1959 and 2010 , seven of the nine basins exhibit significant positive trends in the catchment coefficient. This means that according to the catchment coefficient, there has been a significa nt decline in streamflow. However, since 1980 records show a significant positive trend in UBG . This evidence is contrary to our hypothesis of an inverse relationship between the catchment coefficient and UBG. However, the trend is more nuanced once the ti meframe is divided into 1959 1980 and 1980 2010. Of the previous seven basins, only three have significant negative trends, two do not exhibit a significant trend and two have significant positive trends. For the period of 1980 2010, the trends have change d significantly in the opposite direction . Consequently, the catchment coefficient does not seem correlated to the ratio of aggregate urban, barren and grassland land cover to basin area. This provides an

PAGE 70

70 opportunity for an alternative explanation to the c auses behind the changes in the catchment coefficient. Alternative H ypothesis The changes in the catchment coefficient shown in tables D 7 and D 8 do not seem to be correlated with the changes in land cover for the periods of analysis. Therefore, there cou ld be cause s other than land cover changes for the catchment behavior in the selected sample. In order to propose an alternative hypothesis, it is important to recapitulate what is understood in this framework. T he catchment coefficient describes the rainf all runoff process conditioned to the demands of water from evapotranspiration assuming a steady state in long term groundwater storages . Also , the catchment coefficient has been associated with the physical attributes of the basin , such as vegetation cove r (Rod e rick and Farquhar 2011; Zhang 2004, 2008) . The working hypothesis assumes the effect of land cover classes on runoff generation (see Table 3 1 ) . Furthermore, Dietrich points to the evidence of a steady population growth in the State of Florida (1978 ; US Census 2011) which translates into a positive urbanization trend supported by the trend in population density as shown in tables D 7 and D 8 . Figure 3 4 shows the scatterplot of the relationship between UBG and FWW for the period of 1992 2010 . For this sub sample of nine basins, six of the basins show a substitution of FWW for UBG, while the remaining three basins seem to be made up of mostly AG with little change in this period. This could indicate that if change in land cover from FWW to UBG w as the predominant process that determines the behavior of these catchments, then run off should increase in time according to the hypothetical relationships described in Table 3 1 .

PAGE 71

71 From the sub sample of nine (9) basins with long term data (1959 2010) pre sented in Table 3 4, at least four basins show the expected inverse relationship between the catchment coefficient and UBG between 1980 and 2010 : Charlie Creek (USGS 02296500) , Joshua Creek (USGS 02297100 ) , Brooker Creek (USGS 02307359 ) and Anclote River ( USGS 02310000 ) . However, four of the remaining five basins show a positive trend in the coefficient for the period of 19 80 2010 : St. Marys River (USGS 02231000 ) , South Black Fork Creek (USGS 02245500 ) , Catfish Creek (USGS 02267000 ) and Steinhatchee River ( USGS 02324000 ) , while Wekiva River (USGS 02235000 ) do es not show a significant trend in either period . Therefore, although it was hypothesized that the substitution of FWW f or UBG would increase run off, this is not the case for the whole sub sample. Furthermore, considering that the signal from land cover substitution is clear and significant, this leaves the possibility that an alteration of the water balance as a result of water withdrawals an d/or discharges could have an imp act on the catchment coefficient as shown in a study by Istanbu l l u oglu et al . (201 2 ) for a group of catchments in the Sand Hills of Nebraska. Therefore, the impacts associated with water transfers could be incorporated into an alternative hypothesis as follows: I f the calculation of the catchment coefficient depends on a given runoff ratio under mass balance steady state , then water transfers in the basin, either in the form of discharge or withdrawals, could alter the wate r balance and consequently the coefficient. Alternative hypothesis testing : withdrawals Surface water withdrawals provided 52% to 68% of freshwater in Florida between 1950 and 1975 ( Marella 2008) . Figure 3 5 shows how by 1980, surface water withdrawals ha ve stabilized for the whole State of Florida while groundwater

PAGE 72

72 withdrawals increased. This fact lends credibility to the alternative hypothesis that surface water withdrawals could change the calculation of the catchment coefficient by reducing the amount of observed stream flow. This hypothesis is evaluated for the sub sample of basins by comparing surface water withdrawals per county to the catchment coefficient (see Figures D 4 through D 1 6 ) . The cross correlation coefficient was calculated with SPSS ver sion 22 for a lag range between 10 and 10 periods. Surface water withdrawals per county are available every 5 year s ( Marella 2008, 2014; USGS 2014b ) and were interpolated linearly to perform the analysis. Considering that the catchment coefficient is estimated with the annual average of climate data from the previous 10 years, a cross correlation should be positively significant with a negative lag number for surface water withdrawals in order to prove the lagged correlation. In other words, as surface water withdrawals increase leading to a decrease in available streamflow, the catchment coefficient should increase. Also, a decrease in withdrawals will result in an increase in available streamflow and a red uction in the catchment coefficient. Appendix D includes the bar plots of the cross correlation analysis at different lag values with their upper and lower confident limits in Figures D 17 through D 2 9 . Also the impact of groundwater withdrawals was analyz ed on the basins that had spring vents. Spring flow data was not available at the NWIS site, so a similar cross correlation analysis was performed between groundwater withdrawals at a county level and the catchment coefficient. The bar plots are shown in F igures D 3 0 , D 3 1 and D 32 . Table D 12 summarizes the significant cross correlation coefficient and its corresponding lag number between surface withdrawals and the catchment coefficient.

PAGE 73

73 The results show that only one (1) basin with the long term flow re cord has the expected cross correlation. This basin is Steinhatchee River (USGS 02324000) basin with a positive cross correlation of 0.429 at 7 lag period. The same basin has a spring vent and the analysis was replicated with groundwater withdrawals but s howed no significant results. Figure D 12 supports the statistical evidence by showing the displaced effect from surface water withdrawals and the trend in the catchment coefficient. While the withdrawals peaked in 1990, the coefficient did so in 1995. How ever, as withdrawals decreased the coefficient did the same. And, there is evidence of a positive significant cross correlation between the coefficient and withdrawals seven years before the coefficient. This could indicate that as withdrawals decrease, th e catchment coefficient decreases to account for the increase of available streamflow. This left the remaining eight basins without a plausible explanation to their behavior, requiring further inquiry. Alternative hypothesis testing: discharges Alternative ly, other way to analyze the changes in streamflow was by considering discharges from domestic and industrial wastewater outfalls. Table 3 6 shows permitted wastewater outfall capacity in million gallons per day and in mm by dividing volume by basin area. However, this information alone is not enough to understand the impact by wastewater discharges and water withdrawals, so a brief explanation of practices that could have impacted the catchments is presented as follows. From the group of basins that showed the expected behavior, only the Anclote River basin (USGS 02310000 ) has an outfall , which has an annual domestic wastewater treatment capacity of 26.75 mgd. The treated volume is reused for irrigation and/or discharge d through rapid rate infil tration basins or RRIBs (FDEP 20 11 ).

PAGE 74

74 Consequently, this volume of approximate 209 mm/yr is not directly discharged into streamflow. On the other hand, surface water withdrawals showed two peaks in abstraction: one in 1975 and the other in 1995. However, there was a significant increase in the catchment coefficient from 1960 until 1995 2000 (see Table 3 15); the trend changes direction coincides with the implementation of the Pasco County Master Reuse System (PCMRS) which implemented the use of reclaimed water (RCW) for irrigation, provided rapid rate infiltration basins and created infrastructure to store RCW during wet weather for later reuse. In addition, the decline in the catchment coefficient agrees with the reduction in surface water withdrawals ( FDEP 2006a) . Nevertheless, there is not a significant lagged cross correlation between the coefficient and withdrawals (see Table D 1 2 ). In addition , three othe r basins have wastewater facility outfalls within their boundaries that could increase streamflow and consequ en tly , decrease the catchment coefficient . These basins are St. Marys River (USGS 02231000 ) , Wekiva River (USGS 02235000 ) and South Black Fork Creek (USGS 02245500 ) ; the first and last have shown significant positive trends in the catchment coe fficient bet ween 1980 and 2010. In the Wekiva River basin , a mining operation uses a heavy mineral wastewater system with a capacity of 25 mgd to treat stormwater and dredge water. The system uses settling ponds and discharges into wetlands (FDEP 2006b). The catchment coefficient started an increasing trend in 1990 , matching a period of surface water withdrawals and the ex pansion of the mining operation . However, the positive trend in the coefficient continued after the withdrawal period s but coincid ed with the wastewa ter treatment discharges from the mining operation. In addition, t he boundary of the Wekiva

PAGE 75

75 River b asin overlap s Orange and Seminole counties and Figure D 2 shows the surface water withdrawals from both counties. As stated in the previous section, there is not a significant cross correlation between the behavior of the catchment coefficient and withdrawals . Therefore, the trend in UBG does not seem to be strong enough to control the runoff generation process . In th e South Black Fork Creek b asin the discharge of domestic wastewater is approximately 0.9 mg d or 4mm as compared to the average annual streamflow of 330 mm. Figure D 6 shows that w hile the two peaks in withdrawals seem to support the increase in the catchme nt coefficient, the decrease in withdrawals between 1990 and 2005 do not seem to have a similar effect on the catchment coefficient . T he behavior does not present a cross correlation between the coefficient and withdrawals. The opposite apparent relationsh ip between surface water withdrawal and the catchment coefficient seems to take place in the Catfish Creek basin (USGS 02267000 ) where withdrawals decline as the coefficient increases. Nonetheless, the withdrawals are much larger than in other basins studi ed so far, ranging between 50 and 300 mgd , or 400 to 2,500 mm . This surpasses the average annual streamflow of 170 mm. In addition, there is no evidence of wastewater facilities discharges and/or a groundwater discharge from a spring vent to alter streamfl ow. Also, while the distribution between land cover classes does not seem to change greatly, the increase in UBG is significant for the period of 1980 2010. No significant cross correlation was detected. There fore, surface water withdrawals could be respon sible for the increase in the catchment only because their magnitude exceeds average streamflow. Further investigation is warranted.

PAGE 76

76 Alternative hypothesis testing: conclusions In conclusion, the visual analysis of the relationship between surface water wi thdrawals at a county level and its impact on streamflow do not have strong support statistically: only the Steinhatchee River basin s howed the expected positive cross correlation between the coefficient and surface water withdrawals. The difficulty in defining where the withdrawals take place gives pause to make a definite statement . Also, the water management practices which include wastewater reuse per county show shown by this analysi s. Therefore , water transfers remain the best explanation available for the behavior of the basins that have re duced their streamflow in time, and have shown an increas e in the catchment coefficient. Nevertheless, the basins that do not seem to be affected by changes in the direction of the trend in their catchment coefficient as far back as the 1950s are Joshua Creek and Anclote River. These basins do not appear to be affected by withdrawals and/or discharges. In addition, both ba sins have significant decrease in the coefficient and a significant increase in UBG . These characteristics ma ke them appropriate candidates for the study . Case Study S ample For the selection of the case study sample two modifications to the selection crite ria were implemented. First, only the analytical scenario for the period of 1992 2010 will be further explored. The reason is that some inconsistencies were detected in the some basins regarding the classification of land cover classes which impacts their trends. The land cover products between 1992 and 2011 are considered more consistent for their use in this regression. Second, in order to expand the sample size,

PAGE 77

77 the same methodology was applied to the sub set of ten (10) basins that show the expected tre nd for the period between 1992 and 2010 . It is assumed that samples that have a positive trend or do not show a trend in the catchment coefficient show significant decreases in the run off ratio that cannot be explained by the hypothetical assumptions made about the land cover classes. From this group, four basins are candidates for the regression based on their negative trend in the catchment coefficient and the positive trend in UBG , as well as, the lack of statistical evidence that could show a cross cor relation between the catchment coefficient and surface water withdrawals . These basins are Shingle Creek (USGS 02263800 ) , Davenport Creek (USGS 02266480 ) , Prairie Creek (USGS 02298123 ) , and Trout Creek (USGS 02303350 ) . These basins were evaluated to detect the influence of withdrawals (see Figure D 13, D 14, D 15 and D 16) and/or discharges (see Figure D 26, D 27, D 28 and D 29) to clear their suitabili ty for the regression analysis. Shingle Creek b asin shows a decreasing trend in surface water withdrawals from 1970 through 1980 that seems to match the decrease in the catchment coefficient. This support s the correlation between withdrawals and streamflow. However, from 1985 until 2005 the catchment coefficient se ems to be stable at around 3.5 while water withdrawals continue to decrease. It is only after 2005 that the coefficient starts decreasing again. Therefore, it is possible that there is another parameter maintaining the run off ratio . Nevertheless, a statis tical correlation between the changes in land cover and the coefficient cannot be established. On the other hand, in Davenport Creek basin the relationship between withdrawals and the coefficient is clearer. As surface water withdrawals in Polk C ounty

PAGE 78

78 star t decreasing in 1980 so does the coefficient. This could signal a return of previous streamflow patterns which are higher than the resulting balance after withdrawals. This trend is briefly interrupted when withdrawals increase in 1985 and it is reflected in the coefficient. Afterwards, both trends continue matching their behavior. The positive cross correlation coefficient at a negative lag interval supports this analysis (see Table D 12). The last two basins have shorter catchment coefficient records whic h limits the interpretation of a potential longer term trend. In Prairie Creek basin it is difficult to find a relations hip between the two parameters . S urface water withdrawals start decreasing in 1990, and there is a slight decrease in the catchment coef ficient from 3 to 2.5. However, once water withdrawals begin increas ing in 2000 , the coefficient remains unchanged. Nevertheless, the cross correlation coefficient supports the significant relationship between withdrawals and the catchment coefficient. Fin ally, Trout Creek basin shows a stable trend in withdrawals until 1990 , when withdrawals increase from 60 mgd to 80 mgd. This has an echo in the increase of the coefficient in 1995. However when the coefficient starts decreasing after 1995, withdrawals are still increasing and peak in 2005. I t is difficult to explain why the c oefficient did not continue increasing after 1995 when withdrawals kept on increasing . Also, there is not a significant cross correlation between withdrawals and the catchment coeffici ent. T his analysis reduces the possibility of using any of these basins as cases for the regression analysis. It also shows that this approach to seeking a correspondence between withdrawals and the catchment coefficient is not always straight forward and it

PAGE 79

79 is filled with uncertainty in the interpretation of the relationship between the processes in the basin . As a result of this screening process, f our ( 4 ) basins were retained for the regression analysis : Shingle Creek, Joshua Creek, Trout Creek and Anclote River . The period of analysis ranges from 1992 until 2010 , and includes 19 years of data per basin. The panel data contains 76 observatio ns. Figure 3 6 shows the geographical distribution of the sample. The sample primarily has a substitution of FW W for UBG with the exception of Joshua Creek which has low ratios for both UBG and FWW and shows little change between the two (see Figure 3 6). In addition, the scatterplot of the relationship between UBG ratio and the catchment coefficient (see Figure 3 7) show s negative relationship between the two parameters through the slope of the trend lines . However, the basins as a group do not show a clear relationship between the trends in the ratio of UBG and the catchment coefficient. However, the change from o ne run off generation process to another is clear as shown in Figure 3 8. With the exception of Shingle Creek (USGS 02263800), the observations of the remaining three (3) basins point to a change in the catchment coefficient. Nevertheless, t h e sample size limits the explanatory power of the regression analysis but provides the opportunity to complete the presentation of the proposed methodology . Regression A nalys i s The linear regression analysis was performed using LIMDEP 10S for panel data. Three models we re tested: ordinary least squares, fixed effects and random effects. All these models follow equation 3 7 which represents a pooled time series which combines cross sect ional and longitudinal analysis.

PAGE 80

80 (3 7 ) nt LC knt are the land cover class ratio , P D knt is basin or urban area population density, MV kn represents the morphometric variable ; periods; k K are the explanatory variables. Several regression models were explored and are included in Appendix E. The summary of the results are presented in Table 3 7 , and are discuss as follows . F or the purposes of studying the impacts of the built environment on catchment behavior, the preferred regression models included urban ratio (URB) or UBG, population (POP) and urban population density (PDU) parameters, and other land cover variables (AG an d FWW). The analyzed models can be organized into two groups based on whether time invariant parameters , i.e. , morphometric variables , are included . The models that only included changes in land cover or population density showed poor fits with either low or negative adjusted R squares. This supports the fir st impression provided by Figure 3 7 where as a group , there seems to be a low correspondence between UBG and the catchment coefficient within the sample , but when each basin is studied separately there is a clear negative correlation between the two parameters. The fixed effects models agree in the relationship of the land cover and population density variables and the catchment coefficient by making the heterogeneity of the sample explicit through un ob servable fixed , i.e. , time invariant effects and result in high adjusted R squared models. Nevertheless, the explanatory power of the fixed effects is limited exclusively to the sample that generated it.

PAGE 81

81 The second group of models included the morphometric variables, which would serve as fixed effects. Consequently, the fixed effects models could not be generated . Mean while , the random effects models could become significant by using the morphometric variables as the intercept that could distinguish each ca se from each other , and thus, account for the heterogeneity . The selected models show high R squares in the random effects models that are corroborated with similar fit values in the ordinary least square s models. The latter are stricter model representati ons that serve to l end credibility to the results. The following random effect model s support the hypothetical relationships proposed in this project : The model proposed by equation 3 8 shows how the catchment coefficient has a positive relationship with A G and FWW, and urban population density (PDU). The first two parameters would hypothetically reduce run off generation . It is difficult to interpret the role of population density as it was implied that population density serves as a proxy for urban areas and impervious surfaces, both of which should have a negative relationship with the catchment coefficient. (3 8) On the other hand, equation 3 9 presents a mo del that separates urban areas and population. Both parameters have a negative relationship with the catchment coefficient and argue for an increase in the run off ratio associated with urbanization. This model could be employed in analyzing the relationsh ip between urban growth and catchment behavior. (3 9)

PAGE 82

82 Conclusions The land cover variables cannot solely explain changes in the catchment coefficient for the selected sample. A challenge in applying the Budyko framework to catchments that have significant interven tions by the built environment is the water transfers, either in the form of wastewater discharge or surface water withdrawals. The change in the water balance alters the calculation of the catchment coefficient, and can mislead the analysis and forecasting of the runoff generation process. The application of the fixed effects model showed that for t he selected sample it is possible to explain a large portion of the in group variation by an unobserved fixed effects variable. This supported the inclusion of time invariant variables to explain the large portion of the heterogeneity in the sample. The re sulting panel data model which mixes land cover variables and basin morphometric variables shows an improve d fit. Its applicability for forecasting and impact analysis will be studied in the next chapter. Nevertheless, f urther research is required to understand if there are cross basin characteristics that could be added into the regression equation, the existence of thresholds in changes in behavior, and if it possible to include water use in the form of changes to the w ater budget such as irrigation and/or water harvesting.

PAGE 83

83 Table 3 1 . Data Time Steps Data 1970 1975 1980 1985 1990 1995 2000 2005 2010 Hydrologic variables (catchment coefficient, ) X X X X X X X X X Land cover classes ( USGS; NLCD) 1974 1992 2001 2006 2011 Population density (census block group) X X X X Table 3 2 . Hydrologic data requirements for water balance optimization Variable Calculation Method Data Requirements Source of Data Evapotranspiration potential, (E 0 ) Priestley Taylor (1972) formula adjusted for monthly time steps (Jensen et al. 1990) Monthly cumulative precipitation (P) and minimum and maximum temperature (Tmin; Tmax), and mean basin elevation Maurer et al. 20 02 Evapotranspiration potential, (E 0 ) calibration E 0 Dataset calculated using Geostationary Operational Meteorological Satellite Imagery from 1995 2010 U.S. Geological Survey Florida Evapotranspiration Network Jacobs et al. 2008 Mecikalski et al. 2011 Stream flow Discharge, (Q) Stream flow gauge station r ecords Monthly stream flow in cubic feet per second (cfs) U.S. Geological Survey National Water Information System Falcone et al. 2011 Catchment coefficient, ( ) Optimization of equation 1 method) calibrated with gauged stream flow and constrained by equation 1 3. Monthly gauged stream flow or discharge (Q), precipitation (P), and potential evapotranspiration (E 0 ) N/A Actual Evapotranspiration, (E) Zhang et al. 2004) Evapotranspiration (E 0 ), precipitation (P), catchment coefficient ( ) N/A Table 3 3 . Hypothetical Influence on the water budget by land cover type LC Code Description VAR code Run off (Q) Evapo transpiration (E) Withdrawal (WD) Change in 1 Open Water WAT + n/a n/a 2 Urban URB + + 3 Barren BAR + n/a n/a 4 Forest FOR + n/a n/a 5 Grass /Shrub GRA + n/a n/a 6 Agriculture AGR + + + /+ 7 Wetlands WET + n/a n/a

PAGE 84

84 Table 3 4 . Trend analysis of the catchment coefficient of the nine basins sub sample at different temporal ranges Temporal range Variable Correlation coefficient 1958 2010 1958 1980 1980 2010 1980 1992 1992 2010 Year Pearson Correlation 1 1 1 1 1 Sig. (2 tailed) N 52 22 31 13 19 _02231000 Pearson Correlation .257 .752 ** .839 ** .444 .902 ** Sig. (2 tailed) .066 .000 .000 .128 .000 N 52 22 31 13 19 _02235000 Pearson Correlation .479 ** .783 ** .448 * .925 ** .361 Sig. (2 tailed) .000 .000 .011 .000 .129 N 52 22 31 13 19 _02245500 Pearson Correlation .717 ** .167 .470 ** .075 .310 Sig. (2 tailed) .000 .459 .008 .806 .196 N 52 22 31 13 19 _02267000 Pearson Correlation .878 ** .084 .876 ** .916 ** .675 ** Sig. (2 tailed) .000 .711 .000 .000 .002 N 52 22 31 13 19 _02296500 Pearson Correlation .461 ** .559 ** .404 * .804 ** .929 ** Sig. (2 tailed) .001 .008 .024 .001 .000 N 51 21 31 13 19 _02297100 Pearson Correlation .438 ** .277 .856 ** .321 .863 ** Sig. (2 tailed) .001 .224 .000 .285 .000 N 51 21 31 13 19 _02307359 Pearson Correlation .661 ** .960 ** .430 * .065 .686 ** Sig. (2 tailed) .000 .000 .016 .832 .001 N 51 21 31 13 19 _02310000 Pearson Correlation .630 ** .838 ** .503 ** .358 .663 ** Sig. (2 tailed) .000 .000 .004 .229 .002 N 52 22 31 13 19 _02324000 Pearson Correlation .714 ** .706 ** .594 ** .280 .111 Sig. (2 tailed) .000 .000 .000 .353 .651 N 51 21 31 13 19 **. Correlation is significant at the 0.01 level (2 tailed). *. Correlation is significant at the 0.05 level (2 tailed).

PAGE 85

85 Table 3 5 . Potential sources of streamflow a lterations at the nine basins in the sub sample Gage_ID Flow record , years Area (Sq Km) Spring vent WAFR permitted capacity outfall by source WFR outfall (mm) Q streamflo w (mm) _10 (1992 2010) UBG (1992 2010) _10 (1980 2010) UBG (1980 2010) 02231000 52 1748.4 DW = 1.3 mgd; IW=5x5 mgd 21 28 0 .902 ** .940 ** .839 ** .983 ** 02235000 52 449.0 yes DW = 2.9+12.5 mgd 47 570 .952 ** .952 ** 02245500 52 348.4 DW = 0.9 mgd 4 330 .984 ** .470 ** .992 ** 02267000 52 168.6 170 .675 ** .988 ** .876 ** .913 ** 02296500 51 886.4 240 .929 ** .881 ** .404 * .495 ** 02297100 51 350.4 290 .863 ** .920 ** .856 ** .751 ** 02307359 51 80.0 170 .686 ** .960** .430 * .987 ** 02310000 52 177.6 yes DW = 26.75 mgd 208 260 .663 ** .995** .503 ** .796 ** 02324000 51 791.0 yes 305 .974 ** .594 ** .991 ** Notes: DW, domestic wastewater outfall; IW, industrial wastewater outfall; WAFR, Wastewater Facility Regulation

PAGE 86

86 Table 3 6 . Summary of regression models for the four (4) basins sample Ordinary Least Squares (OLS) Fixed effects Random effects Variable Coefficient Model 1992 2010, test Adj. R squared Coefficient Adj. R squared Coefficient R squared UBG 0.544(*) F[1, 74]= 0.032 0.0131 3.82 0.779 1.45 0.281 Constant 3.15 n/a n/a UBG 0.495 F[2, 73]= 0.194 0.022 2.76 0.778 8.31 1.12 PDB 4.10E 04 (*) 5.54E 03 Constant 3.23 n/a n/a UBG 7.95 F[3, 72]= 1.81 0.031 7.99 0.801 5.34 3.38 PDB 3.75E 03 1.27E 03 2.85E 03 AG 2.73 12.7 5.24 Constant 5.77 n/a n/a URB 0.019(*) F[1, 74]= 3.85E 3 1.30E 02 4.10 0.756 2.02 0.607 Constant 3.17 n/a n/a URB 3.34 F[2, 73]= 26.3 0.402 3.03 0.762 5.53 2.34 POP 9.36E 06 4.49E 06 6.40E 03 Constant 3.55 n/a n/a URB 2.72 F[4, 71]= 62.1 0.765 2.59 0.775 POP 5.07E 06 6.05E 06 Compact 1.93 2.11 Slope 6.77 7.11 Constant 0.302 AG 0.030(*) F[5, 70]= 68.8 0.819 0.438 0.830 FWW 0.126(*) 0.673(*) PDU 1.30E 03 1.16E 03 Compact 2.6 2.07 Slope 2.82 3.08 Constant 0.469 n/a AG 0.739(*) F[5, 70]= 69.9 0.821 5.98 0.767 FWW 1.02(*) 3.98 PDU 1.03 2.20E 04 RR_mea n 3.04 6.17 CircRatio 77.6 106 Constant 3.79 n/a

PAGE 87

87 Figure 3 1. Relationship between the ratio of mean annual evapotranspiration to precipitation (E/P) as a function of the aridity index (E 0 /P) for different values Figure 3 2. The ratio of run off to precipitation at different catchment coefficients ( ) and a con 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 Evapotranspiration ratio (E/P) Aridity Index, (E 0 /P) Boundary Sub-humid (Lower Bound) Sub-humid (Upper Bound) y = 0.9548x 1.166 R² = 0.9986 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 1 2 3 4 5 Q/P catchment coefficient, Q/P=f( , =1) Q/P=f( Power (Q/P=f(

PAGE 88

88 Figure 3 3 . Historical freshwater withdrawals by water source in Florida, 1950 2010 ( Marella 2014) Figure 3 4 . Scatterplot of the relationship between UBG and FWW for sub sample of 9 basins (19 92 2010) 0.0 1,000.0 2,000.0 3,000.0 4,000.0 5,000.0 6,000.0 7,000.0 8,000.0 9,000.0 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 Water withdrawals, million gallons per day (MGD) Year Surface Ground Total 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 0.2 0.4 0.6 0.8 1 UBG FWW 02231000 02235000 02245500 02267000 02296500 02297100 02307359 02310000 02324000

PAGE 89

89 Figure 3 5 . Scatterplot of the relationship between UBG and FWW for the selected sample of four (4) basins (1992 2010) Figure 3 6 . Geographical distribution of basin case studies 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.2 0.4 0.6 0.8 1 UBG FWW 02263800 02297100 02303350 02310000

PAGE 90

90 Figure 3 7 . Linear regression plot for pooled time series, catchment coefficient ( _10)=f(UBG) Figure 3 8 . Scatterplot of the catchment coefficient ( ) for the four (4) case studies (1992 2010) y = 2.835x + 5.4034 R² = 0.3736 y = 8.2396x + 6.0879 R² = 0.494 y = 3.0657x + 3.6831 R² = 0.5338 y = 48.89x + 7.333 R² = 0.8478 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Catchment coefficient, Land cover ratio UBG_02263800 UBG_02310000 UBG_02303350 UBG_02297100 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 0.00 0.50 1.00 1.50 2.00 2.50 3.00 Evapotranspiration ratio, E/P Aridity Index, (E 0 /P) Boundary Sub-humid (Lower Bound) Sub-humid (Upper Bound) 02263800 02297100 02303350 02310000

PAGE 91

91 CHAPTER 4 LIFE CYCLE IMPACT ASSESSMENT CHARACTERIZATION OF CHANGES IN LAND COVER ON WATER RESOURCES : METHOD APPLICATION Background Chhabra (2011) proposed a characterization factor to account for the changes to surface water , i.e. , lakes or rivers , and groundwater storage compartments from water use and changes in land cover. The inventory flow model proposed by Chhabra (201 1 ) combine d water use flows in the built environment with four scenarios of urban built up based on density of impervious surface areas as presented by Livingston and McCarron (1992; USEPA 1993) and illustrated in Figure 4 1 . The resulting impact indicator was the p roduct of a distance to target measure between a developed state and a hypothetical pre development or natural vegetated land cover state for surface discharge and aquifer recharge. The main drawback of the simplified approach to rainfall runoff generation shown in Livingston and McCarron (1992) is that there is no support for the calculation of these ranges available. Also, basin morphometric variables, ecosystem characteristics and changes in climate are not accounted for. So, it is difficult to estimate regional specific changes on catchment behavior associated to land cover changes and climate variability . This limits its potential to project climate change scenarios. Therefore, the characterization model presented in the previous chapter is expected to facilitate the development of regional specific impact factors. Impact Ranges and Baseline Condition Th e translat ion of the impacts of changes in land cover over time on streamflow into mid point LCA indicators uses impact thresholds based upon the Sustain ability Boundary Approach (SBA) proposed by Richter (2009). This approach establishes a

PAGE 92

92 baseline of the un exploited and unregulated daily streamflow , also known as the ecological flow. Furthermore, Richter et al. (2011) proposed presumptive standards of pr of baseline conditions . Figure 4 2 shows a high level of protection range between 10% and 10% of ecological flow , which represent s minimal changes to the riverine ecosystem. A range of +/ 11 20% of ecological flow is considered to have a moderate level of protection as changes in structure and ecosystem functions begin to be perceived . For alterations to ecological flows greater than +/ 20% , the riverine ecosys tem is considered at risk . For instance, the minimum flows and levels proposed by the U.S. EPA and in the Florida Statute 373.042 would most likely consider th at alterations to ecological flows greater than +/ 20% are to be avoided. Similarly, the proposed LCA impact indic ator compares a n expected run off ratio based on land cover conditions to a baseline run off ratio . This baseline condition could represent ecological flow conditions or an already impacted state that is under evaluation. In the latter case, care must be e xercise d when making statements of ecological impact only if the ecological flow condition is known. Otherwise, the assessment sh ould be limited to express changes in cumulative flow. Moreover, although the use of the SBA approach is similar to the LCA dis tance to target method , it is important to distinguish that ecological flow s are measured on a daily time step to assess streamflow behavior while the proposed LCA impact factor uses the annual average behavior . This difference in the temporal resolution could miss the signal from seasonal variability in streamflo w as it will only account for the magnitude of change on a cumulative annual volume base. In addition, as future climate

PAGE 93

93 change impacts predict more extreme occurrences of precipitation and drough t conditions (Kundzewicz and Döll 2009) , the resulting consequences to flow could be missed within the annual aggregate volume. Nevertheless, the proposed method allows for a n indicative and parsimonious assessment of the water balance . In addition, the me thod allows for retrospective and prospective analyses by compar ing selected scenario s to baseline conditions . Impact Indicator The impact indicator defined in equation 4 2 represents alterations in the run off ratio given by relationship between run off ( Q) and precipitation (P). The impact factor shows the change between th e expected run off ratio at ti run off ratio subject to mean climate c onditio ns . ( 4 1) ( 4 2) Table 4 1 includes a presumptive maximum value of +/ 20% natural flow alteration before the change in flow could have an ecological impact ( Richter 2009; Richter et al. 2011; Hoekstra et al . 2012). Differentiation from water stress indices If the impact indicator is linked to the water stress or scarcity indices (i.e. the ratio of withdrawal to availability), w e could get counter intuitive results as increases in

PAGE 94

94 environmental outcome. Climate variability In the calculations of the run off ratio with equation 4 1 , the mean aridity index for the period of analysis is used with its confidence interval at 95% level. This provides three values for the run off ratio: low, mean and high values. Conseq uently, the impact factor includes a similar range. For prospective analysis, estimated values of the aridity index could be used based on probabilistic climate scenarios, and thus, account for the non stationarity of the climate system (Milly et al. 2008) . Methodological A pproach In Chapter 3, several regression models were calculated based on a sample of four (4) basins. The regression models combined different land cover variables, population density and morphometric parameters to estimate the catchment coefficient . The model fit assessment is summarized in Table 3 7 . The proposed impact assessment methodology selects the best regression model according to the objective of analysis. For instance, the retrospective analysis seeks to understand if changes i n land cover a responsible for the historical changes in streamflow for a selected basin, and thus, limits the model selection to those that include the particular land cover variables of interest. On the other hand, the prospective assessment seeks to eva luate two potential urban area scenarios with different population densities. Therefore, the main variables of interest for the regression model are urban areas and population . The assessment process includes the definition of baseline conditions, the calc ulation of expected streamflow using equation 4 3 and the catchment coefficient

PAGE 95

95 from the regression equation . The impact indicator is the ratio of the expected to baseline run off ratios , which serves to estimat e changes in the streamflow annual volume. ( 4 3 ) Retrospective Asses s ment Charlie Creek near Gardner (Florida) was subject of a retrospective assessment. This basin is identified with the USGS gage number 02296500, and has a drainage area of 855 square kilometers ( 330 square miles) . The basin is located west of Sebring Avon Park (Highlands County), and its area falls within the boundaries of Hardee, Highlands and Desoto counties (see Figure 4 3 ). The land cover class ratios from 1992 until 2011 and its population es timates for 1990, 2000 and 2010 are included in T able 4 2. The regression equation used to calculate the expected catchment coefficient is equation 4 4 , which included the ratios of agricultural area (AG) and the aggregate of forest, water and wetlands are as (FWW), urban population density (PDU), the relief ratio (RR) and the circularity ratio (CR) a s its explanatory variable s . This equation is a random effects model with an R squared of 0.77 , and a Nash Sutcliffe efficiency coefficient of 0.11 when use to evaluate a basin from the sample that generated it . The observed values of the catchment coefficient are the result of . The comparisons between observed and expected catchment coefficient values are summarized in Table D 1 . Figure 4 4 shows the comparison between the observed and expected values of the catchment coefficient. (4 4 )

PAGE 96

96 The observed catchment coefficient shows a decreasing trend in time between 1992 and 2010. The trendline has an R squared of 0.86. However, the expected catchment coefficient based on equation 4 3 does not show such a trend because the land cover class rat ios have not change d significantly in time (see Table 4 2 ) . Therefore, the trend in the catchment coefficient is a result of factors other than changes in land cover. Moreover, the decrease in the observed catchment coefficient means that there has been an increase in the run off ratio. Consequently, t he impact factor for year 2005 was calculated based on the expected streamflow value based on the calculated catchment coefficient for 2005 and the observed streamflow value for the same year ; both run off rat ios use the same precipitation value . This impact factor is equal to 1.13, which means that year 2005 shows an increase of 13% in streamflow based on the existing land cover mix. In addition, if the study required assess ing the change in streamflow for yea r 2005 as compared to year 1992, the impact factor will be a ratio of the expected streamflow value s to its corresponding observed precipitation at the two temporal points . This provides an impact factor of 1. 21 or an increase of 21 % in streamflow in 2005 from expected streamflow in1992. Prospective A ssessment This basin is Anclote River near Elfers (Florida), identified by the USGS gage number 02310000. This basin has a drainage area of 188 square kilometers (72.5 square miles), and it is located north of the Tampa St. Petersburg urbanized area (see Figure 4 5 ). The basin falls within the boundaries of Pasco County. The land cover class ratios from 1992 until 2011 and population estimates for 1990, 2000 and 2010 are included in T able 4 3 . The regression equ ation used to calculate the catchment coefficient is equation 4 4 .

PAGE 97

97 (4 4) In this regression equation, the catchment coefficient depends on the ratio of urban area to basin area (URB), basin population (POP), basin compactness (C ompact ) and the average slope (S lope ) of the basin. This equation is a random effects model with an R squared of 0.78. The observed values of the catchment coefficient are the n. For the prospective analysis, two scenarios are evaluated based on the medium population projections for Pasco County between 2010 and 2030 (BEBR 2010) . The population projections and the estimated increase in urban areas based on pre defined population densities for both scenarios are included in Table 4 4 . substituting AG land for urban land while the FWW ratio remains unchanged. In addition, both scenarios use the mean aridity index for th e period 1992 2010 with a confidence interval at 95% level (0.94 +/ 0.01). T he impact factor is represented as a mean value and a range that includes the corresponding run off ratio values for the low, mean and high aridity indices. Figure 4 6 shows the co mparison between the observed and expected values of the catchment coefficient. In scenario 1, by maintaining current urban population density (PDU) at 1,025 inhabitants per square kilometer, the projected population growth for 2030 will expand urban areas from a ratio of 0.18 to 0.26 of total basin area . This could translate in an impact factor of 1.05 or a range between 1.049 and 1.052 from 2010 streamflow levels, an approximate increase in streamflow of 5% by 2030. On the other hand, scenario 2 assumes a projected decrease in urban population density from 1,025 in 2010 to 625 inhabitants per square kilometer at the same time that the expected population growth

PAGE 98

98 is taking place . This means an increase in the URB ratio from 0.18 to 0.42. This could result in an impact factor of 1.19 or a range between 1.187 and 1.200 for 2030 from 2010 streamflow levels, an approximate increase of 19% in streamflow by 2030. Moreover, by 2027, the basin agricultural land will have been spent and the FWW ratio would have to dec rease from 0.17 to 0.11 absorbing the expansion of urban areas . Conclusions The impact indicator represents changes in the annual water budget , specifically run off generation. The indicator is indicative of changes in the magnitude of stream flow but cann ot substitute for the impacts of stream flow variability on shorter time interval s . The proposed calculation method based on the Budyko hypothesis includes climate variability through the aridity index and could be used in retrospective and prospective analysis of the impact of land cover and population changes on catchment behavior.

PAGE 99

99 Table 4 1 . Impact ranges Value Impact factor description Environmental flow protection standard (Richter et al. 2011) Water stress (UN 1997) >1.20 Ecological risk, high flooding Ecological risk, upper range 1. 11 1.20 Ecological risk, moderate flooding Moderate level of protection 1.00 1.10 Ecological risk, low flooding High level of protection 1.00 Ecological baseline Ecological flow (b aseline ) None 0.90 0.99 Ecological risk, low depletion High level of protection 0 10%; low 0.80 0.89 Ecological risk, moderate depletion Moderate level of protection 10 20%; moderate 0.0 0 0.79 Ecological risk, high depletion Ecological risk, lower range 20 40%; medium high >40%; high Table 4 2. Land cover class ratios and population for the Charli e Creek near Gardner basin Variable Code Period Land Cover Class 1992 2001 2006 2011 Urban URB 0.0610 0.0579 0.0580 0.0580 Barren BAR 0.0024 0.0026 0.0027 0.0009 Grassland/Shrub GRA 0.0309 0.0620 0.0620 0.0620 Forest FOR 0.0225 0.0212 0.0210 0.0210 Open Water WAT 0.0011 0.0008 0.0010 0.0015 Wetlands WET 0.3162 0.2686 0.2686 0.2682 Agriculture AGR 0.5658 0.5869 0.5867 0.5884 Period 1990 2000 2010 Population POP 12,201 17,830 16,377 Table 4 3 . Land cover class ratios and population for the Anclote River near Elfers basin Variable Code Period Land Cover Class 1992 2001 2006 2011 Urban URB 0.1339 0.1328 0.1707 0.1855 Barren BAR 0.0024 0.0039 0.0130 0.0073 Grassland/Shrub GRA 0.0996 0.1590 0.1555 0.1593 Forest FOR 0.1853 0.1117 0.1082 0.1046 Open Water WAT 0.0163 0.0192 0.0199 0.0204 Wetlands WET 0.3556 0.3569 0.3439 0.3408 Agriculture AGR 0.2069 0.2165 0.1889 0.1820 Period 1990 2000 2010 Population POP 7,821 11,560 33,305

PAGE 100

100 Table 4 4 . Prospective urban growth scenarios for the Anclote River near Elfers basin , 2010 2030 Period 2010 2015 2020 2025 2030 M edium projection 440,300 479,100 527,800 574,700 619,000 Population increase 1.00 1.09 1.20 1.31 1.41 Basin Population 33,305 36,240 39,924 43,471 46,822 URB, scenario 1 0.18 0.20 0.22 0.24 0.26 URB 2030, density (hab./sq. km.) 1 , 025 1 , 025 URB, scenario 2 0.18 0.22 0.27 0.34 0.42 URB 2030, density (hab./sq. km.) 1 , 0 25 625 Figure 4 1 . Partitioning of precipitation according to ISA ranges (Source: Livingston and McCarron 1992)

PAGE 101

101 Figure 4 2 . Presumptive standards of protection for ecological flows (Richter et al. 2011) Figure 4 3 . Charlie Creek basin near Gardner 2011 land cover map with Anderson level II land cover classification

PAGE 102

102 Figure 4 4 . Model analysis for Charlie Creek basin near Gardner , 1992 2010 . Figure 4 5 . Anclote River basin near Elfers 2011 land cover map with Anderson level II land cover classification y = 0.0813x + 166.54 R² = 0.8635 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 1990 1995 2000 2005 2010 right: catchment coefficient ( ), left: ratio of UBG area to basin area Years Exp_ UBG FWW Linear (

PAGE 103

103 Figure 4 6 . Model analysis for Anclote River basin near Elfers, 1992 2010 and 2010 2030. 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 5.50 6.00 1990 1995 2000 2005 2010 2015 2020 2025 2030 left: catchment coefficient ( ), right: ratio of URB area to basin area Years Obs_ Exp_ SCN 01 SCN 02 URB URB_SCN-01 URB_SCN-02

PAGE 104

104 CHAPTER 5 CONCLU SIONS AND FUTURE RESEARCH Discussion The proposed approach has sought to model the impacts of the built environment on the water balance by concentrating i n changes on the run off ratio that result from alterations to land cover. To this end, a formulation of the Budyko hypothesis was used to capture the properties of the drainage basin and its effect on the partitioning of rainfall into evapotranspiration and run off through a catchment coefficient . The methodology was applied to a sample of drainage basin s in the South Eastern Coa stal Plains ecoregion based on the premise that other parameters such as vegetation types and hydro geomorphology could be controlled in the modelin g. T he annually calculated catchment coefficient for the sample showed diverging trends regardless of the in crease in urban development questioning the first hypothesis that proposed an inverse relationship between urbanization and the catchment coefficient. As a consequence, a nine (9) basin sub set with available long term streamflow data (59 60 years) of the original sample was used to explore the presence of long term cycles and found a general agreement in the upward trend in the coefficient with the period of increase surface withdrawals for the State of Florida . However, an increase in the catchment coeffi cient could only be partially explained through its cross correlation with surface water withdrawals at the county level. In addition, the investigation found extensive wastewater treatment and reuse capacity in place in some of the basins under analysis . This led to the proposal of an alternative hypothesis that could account for the impact of changes in the water balance on the catchment coefficient. These presumed changes in the water balance made difficult the definition

PAGE 105

105 of baseline conditi ons as well as find a corre lation between land cover classes and the catchment coefficient. Nevertheless, further research is necessary to explain the sensitivity of changes in streamflow from water transfers and/or climate variability on the optimization procedure used to calculate the catchment coefficient. In contrast, a reduced sample of basins presented the expected run off behavior and was used to explore the applicability of the proposed methodology. The resulting panel data regression models include d random effects and morphometric parameters that allowed capturing the heterogeneity of the sample. Also, the models represented statistically significant relationships between the catchment coefficient, land cover classes , and population parameters. This facilitated the estimation of annual stream flow based on given urban growth expectations and climate scenarios. Finally , the impact of the built environment on freshwater resources was characterized as the proportion of the run off ratio of a developed b asin to a baseline condition. Th e methodology allowed assessment of the relative impact of land cover change on the hydrologic al cycle and could assist planners in the development of sustainable communities . Challenges in the Application of the Budyko Hypothesis Alterations to the water mass balance through water transfers impact the calculation of the catchment coefficient. This violates the equilibrium state equation. Moreover, the impact from groundwater withdrawals was not fully explored but conside ring the previous statement it is possible to infer that alterations of the long term equilibrium in soil moisture storage could also have an impact. Moreover, the ten year moving average used in the optimization of the Budyko formulation probably does not capture the complete extent of the basin hydrological memory and climate variability . In

PAGE 106

106 addition, the ten year average misses the presence of longer term climate cycles such as the Atlantic Multi decadal Oscillation. This process is estimated to have a c ycle of 60 to 80 years and has a significant impact on precipitation in Florida (Enfield et al. 2001). Challenges in the Calculation of the Impact Factor The impact factor is limited to the analysis of the annual water balance. The ecological risks attribu ted to changes in streamflow are studied at a daily time intervals. This makes the impact factor more suitable for the understanding of water availability than the ecological impacts of the built environment. Nevertheless, the approximation is indicative o f potential impacts and a modeling trade off considering the data requirements of a modeling effort at a shorter time intervals. In addition, the impact factor could include confidence intervals based on the probability distribution of the aridity index. H owever, the Budyko equation establishes a non linear relationship between the run off ratio through the evapotranspiration ratio and the aridity index . Consequently, the impact factor could have an impact range based on the aridity index but not a confiden ce interval. Conclusions There is evidence that water transfers could alter the calculation of the catchment coefficient by altering the water balance, violating the equilibrium state assumption in the long term rainfall runoff generation process . This lim its the applicability of the Budyko formulation unless the changes in storage can be incorporated into it. However, this does not discard the possibility that there is a correlation between land cover classes and the catchment coefficient. Nevertheless, th e spatial resolution of the land cover data misses important characteristics of the built environment. The aggregation of urban areas into a single category regardless of percentage impervious areas misses

PAGE 107

107 information that could help evaluate mitigating st rategies such as the implementation of low impact development. This does not mean that this information could not be included by a better characterization of the urban area ratio. In conclusion , the impact of the built environment on freshwater resources c an be characterized as a ratio of the catchment coefficient of a developed basin to a baseline coefficient. Thus, it is possible to measure the effects of land cover alteration on changes in the relative water balance through the implementation of the Budy ko hypothesis . This methodology could assist in the assessment of the relative impact of land cover change and the built environment on the hydrologic cycle and assist planners in the development of sustainable communities. Further research will incorporat e the impacts of water withdrawals from surface and groundwater storages, and extend the methodology to cover different spatial levels that would allow analyzing building units within a regional perspective through better characterization of the land cover variables . Future Research As a consequence of this study, several future venues of research are proposed. First, e xplore the relevance of the ecoregions in the model by conducting studies across other ecoregions and compare the results to find commonalities and differences in the variables that could influence the catchment coefficient . Second, investigate the impact of streamflow elasticity to climate variability in the model in order to incorporate in the forec asting process the s ensitivity to climate scenarios. Third, d evelop water use model based on system dynamics to couple with the hydrological model through the water compartments: surface and groundwater storages. In this way, the impact analysis of the wat er mass balance can differentiate between water flows from the natural

PAGE 108

108 hydrological cycle and water flows from the systems used to supply the built environment and other land uses. Fourth, study the impact of land cover fragmentation and distribution on th e catchment coefficient. Fifth, p erform the study including different urban densit ies and their spatial distribution in the catc hment in relation to the outlet. And finally, characterize different urban stormwater management strategies represented by total impervious surfaces and effective impervious surfaces with their corresponding impact on run off generation.

PAGE 109

109 APPENDIX A WATER BALANCE CALCULATION METHODS Summary The proposed implementation of the Budyko framework for the determination of the basin annua equation (A 1). The water balance is represented by equation A 2, and assumes long term equilibrium where groundwater storage changes become negligible : (A 1) (A 2) Where P represents the annual cumulative precipitation or rainfall depth in the basin in millimeters (mm); Q is the annual cumulative stream flow or basin discharge in rainfall depth equivalent; E is evapotranspiration, and it is the product of the optimal solution for the catchment coefficient ( equation as described in Chapter 3. E 0 stands for potential evapotranspiration and it is calculated with the Priestley Taylor equation . The calculation of these parameters is described below. Stream fl ow (Q) Monthly statistics for catchment discharge were downloaded from the USGS Data is in cubic feet per second. Average m onthly d ischarge rate was transformed into monthly total discharge volume, a nd th en, into annual discharge volume by summing the monthly values from January through December (i.e. calendar year) . Finally, total annual discharge volume was converted into rainfall depth equivalent by di viding volume by catchment area in SI units. Data do wnload accessed on May 3, 2014:

PAGE 110

110 http://waterdata.usgs.gov/nwis/inventory?search_criteria=search_site_no&submitted_for m=introduction Precipitation (P) and temperature Monthly cumulative rainfall depth (mm) and, monthly minimum and monthly maximum temperature in degrees Celsius were obtained from the gridded meteorological data (1949 2010) at 1/8 degree spatial resolution (Maurer et al. 2002). Data download accessed on May 10, 2014: http://www.engr.scu.edu/~emaurer/gridded_obs/index_gridded_obs.html The calculation process involved the spatial projection o f the dataset coordinate points using ArcGIS 10.2, and the creation of the corresponding Thiessen polygons Thiessen polygons to calculate the area weighted influence of ea ch data point in the annual rainfall depth (P) in mm is an input in equation A 1. The monthly average temperature parameters were used in the calculation of potential evapotr anspiration. Potential evapotranspiration (E 0 ) The calculation of potential evapotranspiration for the selected case studies use d equation A 3 formulated by P riestley and Taylor (1972) and the methodology presented by Mecikalski et al. (2011) with supp lemented methods from Allen et al. 1998 for the estimat ion of parameters that were provided by satellite data from the Geostationary Operational Environmental Satellite (GOES) system . The Priestly Taylor formula was selected for this study because of its f ew data requirements and good fit for the South Eastern U.S. (Lu et al. 2005). The following abridged version includes the assumptions made for the calculation process.

PAGE 111

111 (A 3) Where, kg 1 w is the density of water (1000 kg m 3 Taylor coefficient assumed to be 1.26, and the slope of the saturated vapor pressure ( ) and the psychrometric constant ( ) are calculated using equations A 4 and A 5 respectively. Also, Rn stands for net radiation (MJ m 2 d 1 ) and G is the soil heat flux (MJ m 2 d 1 ). For daily calculations G is considered negligible. (A 4) In A 4, e s is the saturated vapor pressure (kPa), and T min is the minimum daily tempera ture in degrees Celsius ( C). For A 5, c p is the specific heat of moist air (1.013 kJ kg 1 C 1 ), P is atmospheric pressure (101.3 kPa), is the ratio of molecular weight of water vapor to dry air. Also, T stands for mean temperature in degrees Celsiu s for equation A 7. (A 5) Saturated vapor pressure was calculated using equation A 6, the latent heat of vaporization was estimated using equation A 7, and net radiation (Rn) was calculated according to the method proposed by Mecik alski et al. (2011). (A 6) (A 7) Net radiation method (Rn) The calculation of net radiation (R n ) uses the four component methodology recommended by Mecikalski et al. (2011), which uses incoming solar radiation (R s ), lu ) and downwelling (R ld ) longwave radiation as

PAGE 112

112 shown in equation A 8 . Radiation values are in MJ m 2 day 1 , surface emissivity s =0.97) s et al. (2008) . (A 8 ) Each component is calculated as presented in the following sections: Incoming solar radiation ( Rs ) The incoming solar radiation was estimated using the Hargreaves and Samani (1982) formulation and calibrated using the coefficients (k Rs ) proposed by Martinez and Thepadia (2010) and, Thepadia and Martinez (2012). (A 9 ) (A 10) Where k Rs stands for an empirical calibration coefficient calculated using equation A 10 , R a is the extraterrestrial incoming radiation estimated using equation A 11 , and TD is the difference between maximum daily temperature (T max ) and minimum daily temperature (T min ) . Equation A 10 uses the average values of mean temperat ure and temperature difference . (A 11) Where Gsc is the solar constant (0.0820 MJ m 2min 1), dr is the inverse relative distance Earth Sun, s is solar declination in radians. Upwelling longwave radiation (Rlu) The calculation uses the following equation: (A 1 2 )

PAGE 113

113 s Boltzmann constant ( 4.903e 9 MJ m 2 day 1 K 4 ) and T s is surface temperature or average daily air temperature in degrees Kelvin . Downwelling longwave radiation (R ld ) The estimation of this parameter is organized into two steps: 1) estimation of clear sky radiation, and 2) the correction for cloud cover. The equations employed are: (A 13) Where e a is the mean actual atmospheric vapor pressure in millibars (see equation A 14) Boltzmann constant ( 4.903e 9 MJ m 2 day 1 K 4 ) and T is average daily air temperature i n degrees Kelvin, and the Florida specific values for a 1 and a 2 of 0.575 and 0.054 respectively. (A 14) The actual vapor pressure was calculated using equations 14 and 48 from Allen et al. (1998) which assumes minimum temperature equals dew temperature in degrees Celsius. The correction for cloud cover (equation A 16) uses the estimated fractional cloud cover (see equation A 15), where R s is the estimated incoming solar radiation at the surface and R so is the e stimated clear sky radiation (MJ m 2 day 1) and elevation (z). (A 15) (A 16) (A 17)

PAGE 114

114 Water Balance Calculation Algorithm The water mass balance calculation procedures were sep arated into stand alone Matlab R2013 scripts to facilitate debugging. These scripts are grouped within a single script to provide the annual cumulative water depth for precipitation, evapotranspiration and stream flow. Figure A 1 shows the algorithm of the se procedures and the following section presents the transcription of each script. Matlab scripts instructions Figure A 1 shows the nested relationship of the scripts used in the calculation of the water mass balance. Prefix es stand for parameter calculation at annual and monthly time steps respectively. The scripts use as call data found in a matlab file named parameters. The file is located in each case study folder. These parameters are summarized in Table A 1. Table A 1. Data for Matlab scripts Variable Name Description Basin_area Drainage area in square kilometers Latdeg Basin centroid latitude in degrees angle M _cfsQ Monthly average stream flow (Q) in cubic feet per second M_days Days per month M_prcp Monthly average rain fall (PRCP) in millimeters M_tmax Maximum monthly temperature in degrees Celsius M_tmin Minimum monthly temperature in degrees Celsius M_tmean Mean monthly temperature in degrees Celsius Z Elevation in meters

PAGE 115

115 Figure A 1. Water balance calculation algorithm Matlab scripts a _calc_WB %this script calculates the water mass balance of a basin on an annual %basis a_PET=a_calc_PET; a_Q=a_calc_Q; a_PRCP=a_calc_PRCP; a_PRCP=a_calc_PRCP %this function script will calculate the annual PCPT from a dataset with monthly values A_PRCP A_P ET A_ Q INPUT DATA M _P ET M A_P ET M A_ RN M A_ Rld M A_ Rns M A_ Rlu Rs Rs Rso aVP kRs Ra OUTPUT DATA A_ WB

PAGE 116

116 load('basin_inputdata', 'm_prcp'); a_PRCP=zeros(length(m_prcp)/12,1); jan=1; for a=1:length(m_prcp)/12 dec=jan+11; a_PRCP(a)=sum(m_prcp(jan:dec)); jan=dec+1; end end a_Q=a_calc_Q %This function script calculates the annual stream flow from monthly averages load('basin_inputdata', 'm_cfsQ', 'm_days', 'basin_area'); %Converts 'monthly average discharge' from cubic feet per second into cubic meters per second %(English units to SI units) ma_m3Qsec=zeros(length(m_ cfsQ),1); for i=1:length(m_cfsQ) ma_m3Qsec(i)=m_cfsQ(i)./35.315; end %calculates cumulative monthly discharge (m_m3Q) in m3 ma_m3Q=zeros(length(m_cfsQ),1); %empty array for cumulative daily discharge m_m3Q=zeros(length(m_cfsQ),1); %discharge in cubic feet per second for i=1:length(m_cfsQ) ma_m3Q(i)=ma_m3Qsec(i).*(60*60*24); %60sec/min in 60min/hr in 24hrs/day m_m3Q(i)=ma_m3Q(i).*m_days(i); %days is a vector with the days per month for the study period end %calculates streamflow/basin area (depth?) m_Q=zeros(length(m_cfsQ),1); for i=1:length(m_cfsQ) m_Q(i)=m_m3Q(i)./(basin_area*10^3); %sqkm to sqm multiplies area by 10^6; meters to mm multiplies area by 1000 end %calculate the annual cumulative streamflow discharge (a_Q) per unit area %(depth in mm) a_Q=zeros(length(m_cfsQ)/12,1); jan=1; for a=1:length(m_cfsQ)/12 dec=jan+11; a_Q(a)=sum(m_Q(jan:dec)); jan=dec+1; end end

PAGE 117

117 a_PET=a_calc_PET %this function script calculates the cumulative annual PET from monthly values load('b asin_inputdata', 'm_tmean'); m_PET=m_calc_PET; %calculate the annual cumulative sum of PET by the sum of 1 to 12 months into an annual total %annual average!!!! a_PET=zeros(length(m_tmean)/12,1); jan=1; for a=1:length(m_tmean)/12 dec=jan+11; a_PET(a)=sum(m_PET(jan:dec)); jan=dec+1; end end m_PET = m_calc_PET %calculates the cumulative sum of daily PET per month %each month is assumed to have 30.4 days based on the calculation for %julian days used in the ma_calc_Rn script ma_PET=m a_calc_PET; load('basin_inputdata', 'm_tmean'); m_PET=zeros(length(m_tmean),1); for i=1:length(m_tmean) m_PET(i)=ma_PET(i).*30.4166; end end ma_PET=ma_calc_PET %This script will calculate the PET using the Priestley Taylor equation %Unless otherwise specified, the variables are daily values %see explanation of methodology in FAO: http://www.fao.org/docrep/x0490e/x0490e07.htm #chapter 3 load('basin_inputdata','m_tmax','m_tmin','m_tmean','z'); %latent heat of vaporization (MJ/kg) lhv=zeros(length(m_tmean),1); for i=1:length(m_tmean) lhv(i)=2.501 0.002361*m_tmean(i); end %slope of vapor pressure; equation 13 in FAO chapter 3

PAGE 118

118 %the saturated vapor pressure (sVP) is calculated at 'tmean' for the slope of %saturation vapour pressu re sVP_tmean=zeros(length(m_tmean),1); for i=1:length(m_tmean) sVP_tmean(i)=0.6108*exp(17.27*m_tmean(i)/(m_tmean(i)+237.3)); end slope_sVP=zeros(length(m_tmean),1); for i=1:length(m_tmean) slope_sVP(i)=4098*(sVP_tmean(i))/((m_tmean(i)+237.3)^2); end %soil heat flux density (only for monthly calculations); units in MJ m 2 day 1 %g(i)=0; %on daily calculations "g" is ignored %see http://www.fao.org/docrep/x0490e/x0490e07.htm#radiation for a detailed %explanation of soil heat flux calculations %g=zeros(length(m_tmean),1); %for i=1:length(m_tmean) % if i==1 % g(i)=0.14*(m_tmean(i+1) m_tmean(i)); % elseif i==numel(m_tmean) % g(i)=0.14*(m_tmean(i) m_tmean(i 1)); % else % g(i)=0.07*(m_tmean(i+1) m_tmean(i 1)); % end %end %Calculates psychrometric constant (psy) P=101.3*((293 0.0065*z)/293)^5.26; %atmospheric pressure in kPa psy=zeros(length(m_tmean),1); for i=1:length(m_tmean) cp=1.013/(10^3); %MJ/kgC; specific heat at constant pressure mol=0.622; %ratio molecular weight of water vapor to dry air psy(i)=(cp*P)/(mol.*lhv(i)); end %daily PET using the Priestly Taylor formula PTc=1.26; %Priestley Taylor calibration constant for humid or wet areas; other values are possible (i.e. f or arid areas PTc~1.70) %Calls for ma_RN calculation ma_RN=ma_calc_RN; ma_PET=zeros(length(m_tmean),1); %this equation ignores g(i) as an input because the calculation of PET is on daily basis for i=1:length(m_tmean) ma_PET(i)=(1./lhv(i)).*PTc.*slope_sVP(i).*ma_RN(i)./((slope_sVP(i)+psy(i))); end end

PAGE 119

119 ma_RN = ma_calc_RN %this script calculates net solar radiation (RN) as the difference between the incoming net shortwave radiation (Rns) and net longwave radiation (Rnl= Rld Rlu) % It calls on three functions: ma_Rns and ma_Rnl %Calls function to calculate mid month net shortwave radiation using the %Hargreaves formula and a given albedo ma_Rns=ma_calc_Rns; %Calls functions to calculate mid month longwave radiation c omponents ma_Rld=ma_calc_Rld; ma_Rlu=ma_calc_Rlu; %Calculates RN as the difference between Rns and Rnl %FAO chapter 3, eq. 40 load('basin_inputdata', 'm_tmean'); num_months=length(m_tmean); ems=0.97; %surface emissivity ma_RN=zeros(num_months,1); for i=1:num_months ma_RN(i)=ma_Rns(i)+(ems.*ma_Rld(i)) ma_Rlu(i); end end ma_Rns = ma_calc_Rns %This script calculates solar radiation from air temperature differences on %daily periods using the Hargreaves formula in FAO 56 chapter 3 Eq. 50 %(Rs) in MJ m^ 2 day^ 1 ma_Rs=ma_calc_Rs; load('basin_inputdata', 'm_tmean'); num_months=length(m_tmean); %Calculates mid month net shortwave radiation ma_Rns=zeros(num_months,1); for i=1:num_months ma_Rns(i)=ma_Rs(i).*(1 0.141); %Rns=Rs*(1 As); As=0.141, based on Jacobs et al (2008, 58) GOES report. %Note: 0.32 is Earth's average albedo end end ma_Rs = ma_calc_Rs %This script calculates solar radiation from air temperature differences on %daily periods using the Hargreaves formula in FAO 56 chapter 3 Eq. 50 %(Rs) in MJ m^ 2 day^ 1 %NOTE: script requires work in the input request; default set for inland

PAGE 120

120 %basin kRs %location=input('Is the basin located inland? Y/N [Y]: ','s'); %if location=='Y' % kRs=0.16;%adjustment coefficient for 'interior' locations %else % kRs=0.19; %adjustment coefficient for 'coastal' locations %end %kRs=0.16; load('basin_inputdata', 'm_tmax', 'm_tmin'); num_m onths=length(m_tmax); %temperature difference of monthly averages of temperature in degrees Celsius m_tdif=zeros(num_months,1); for i=1:num_months m_tdif(i)=m_tmax(i) m_tmin(i); end %Calculates the solar radiation using the Hargreaves formula %d_Rs_H in MJ m^ 2 day^ 1 ma_Ra=ma_calc_Ra; kRs=calc_kRs; %this functions estimates kRs based on equation 11 Thepadia and Martinez (2012) ma_Rs=zeros(num_months,1); for i=1:num_months ma_Rs(i)=kRs.*ma_Ra(i).*sqrt(m_tdif(i)); end end kRs=calc_kRs %this script calculates the empirical coefficient for the Hargreaves %formula and follows equation 11 in Thepadia & Martinez 2012 load ('basin_inputdata', 'm_tmax', 'm_tmin'); num_months=length(m_tmax); m_tdif=zeros(num_months,1); for i=1:num_months m_tdif(i)=m_tmax(i) m_tmin(i); end avg_td=sum(m_tdif)/(length(m_tdif)); m_tmean=zeros(num_months,1); for i=1:num_months m_tmean(i)=(m_tmax(i)+m_tmin(i))./2; end

PAGE 121

121 avg_tmean=sum(m_tmean)/(length(m_tmean)); kRs=0.0305*(avg_tmean/avg_td)+0.0976; end end ma_Ra=ma_calc_Ra %this script calculates the extraterrestrial radiation for daily periods %at the middle of the month %(Ra) in MJ m^ 2 day^ 1 load('basin_inputdata','latdeg','m_tmean'); latrad=latdeg*pi/180; lat=latrad; %solar constant in MJ m^ 2 min^ 1; the d_Ra formula includes the multiplication by 24 hrs/day * 60 min/hrs to obtain the solar radiation Gsc=0.0820; %calculates the Julian day at mid month num_months=length(m_tmean); %number of months in the study %ex. 492 months/12 months year = 41 years months=1:12; %creates horizontal 1x12 matrix year=months'; %creates vertical 12X1 matrix m=repmat(year,num_months/12,1); %creates a vertical matrix 'num_months'x1. It repeats (492/12)times 1:12 (months) jd=zeros(num_months,1); for i=1:num_months jd(i)=round(30.4*m(i) 15); %30.4375=[(3*365)+366]/48 months end %calculates the solar declination at mid month %uses equation (24) from http://www.fao.org/docrep/x0490e/x0490e07.htm#solar radiation %sd(i) is the vector that contains the daily solar declination in radians % sd=zeros(num_months,1); for i=1:num_months sd(i)=23.45*(pi/180).*sin(2*(pi/365).*jd(i) 1.39); %0.409=23.45*pi/180 end %calculates sunset hour angle (ws) in radians ws=zeros(n um_months,1); for i=1:length(sd) ws(i)=acos( tan(sd(i)).*tan(lat)); end %calculates inverse relative distance Earth Sun (dr); unitless? dr=zeros(num_months,1); for i=1:num_months dr(i)=1+0.033.*cos((2*pi/365).*jd(i));

PAGE 122

122 end %calculates extraterrestrial radiation (Ra) in MJ m^ 2 day^ 1 ma_Ra=zeros(num_months,1); for i=1:length(sd) ma_Ra(i)=(24*60*Gsc/pi).*dr(i).*((ws(i).*sin(sd(i)).*sin(lat))+(cos(sd(i)).*c os(lat).*sin(ws(i)))); end %Note: the previous equation had this form %w=0.2618; %angular velocity in rad/hour, or w=15 degrees/hour %Rs=(2*w0)*((tsun(i)*sin(sd(i))*sin(lat))+(cos(sd(i))*cos(lat)*sin(w*tsun(i)) /w)); end end ma_Rlu=ma_calc_Rlu %this script calculates the upwelling longwave radiation Rlu %Background information can be found in section 4.3.2 of Jacobs et al (2008) SATELLITE BASED SOLAR %RADIATION, NET RADIATION, AND POTENTIAL AND REFERENCE EVAPOTRANSPIRATION ESTIMATES OVER FLORIDA %This is a technical report downloaded from the USGS website %3) Upwel ling longwave radiation Rlu %Stefan Boltzmann constant = 5.670 ×10^ 8 W m^ 2 K^ 4 %Rlu=ems*sbc*(m_tmean_K).^4; (.^) elementwise POWER function load('basin_inputdata', 'm_tmean'); ems=0.97; %surface emissivity SBc=4.903e 9; %Stefan Boltzmann constant in MJ m^ 2 day^ 1 K^ 4 m_tmean_K=zeros(length(m_tmean),1); %this creates a monthly mean temperature matrix in degrees Kelvin for i=1:length(m_tmean_K) m_tmean_K(i)=m_tmean(i)+273.15; end ma_Rlu=zeros(length(m_tmean_K),1); for n=1:lengt h(m_tmean_K) ma_Rlu(n)=ems*SBc.*(m_tmean_K(n)).^4; end end ma_Rld=ma_calc_Rld %this script calculates the downwelling longwave radiation Rld

PAGE 123

123 %Downwelling longwave radiation requires two steps; 1) estimate the clear sky radiation and 2) correct for c loud cover. load('basin_inputdata', 'm_tmean'); %1) estimate clear sky ratiation %Rldc=a1+a2*sqrt(ea)*SBc*(tmean_K)^4; ea: vapor pressure, and SBc: Stefan Boltzmann %constant %ea=?; %mean actual vapor pressure SBc=4.903e 9; %Stefan Boltzmann constant in MJ m^ 2 day^ 1 K^ 4 a1=0.575; %Sellers (1965) estimated variable a1=0.605 a2=0.054; %Sellers (1965) estimated variable a2=0.048 m_tmean_K=zeros(length(m_tmean),1); %this creates a monthly mean temperature matrix in degrees Kelvin for i=1:length(m_tmean_K) m_tmean_K(i)=m_tmean(i)+273.15; end %actual vapor pressure in mb derived from Tdew=Tmin using FAO 56 eq. 48 ma_aVP = ma_calc_aVP; % 1) clear sky longwave radiation ma_Rldc=zeros(length(m_tmean_K),1); for n=1:length(m_tmean_K) ma_Rldc(n)=(a1+a2.*sqrt(ma_aVP(n))).*SBc.*(m_tmean_K(n))^4; end %2) method to correct for cloudy conditions (Crawford and Duchon 1999) %estimate fractional cloud cover (c) %incoming solar radiation (ma_Rs) at the surface calculated at site ma_Rs=ma_calc_Rs; %Rso=1; %theoretical clear sky downward solar radiation calculated using the method described in Allen et al 1998 (FAO 56) ma_Rso=ma_calc_Rso; %c=1 Rs/Rso; %method to calculate the fractional cloud cover used in the %GOES report ma_c =zeros(length(ma_Rs),1); for n=1:length(ma_Rs) ma_c(n)=1 (ma_Rs(n)./ma_Rso(n)); end ma_Rld=zeros(length(ma_Rldc),1); for x=1:length(ma_Rldc) ma_Rld(x)=ma_Rldc(x).*(1 ma_c(x))+(ma_c(x).*SBc.*(m_tmean_K(x))^4); end end

PAGE 124

124 ma_Rso = ma_calc_Rso %This script calculates the clear sky solar radiation (Rso) % it uses FAO chapter 3 eq. 39 % MJ m^ 2 day^ 1 load('basin_inputdata', 'm_tmean'); ma_Ra=ma_calc_Ra; num_months=length(m_tmean); ma_Rso=zeros(num_months,1); for i=1:num_months ma_Rso(i)=ma_Ra(i).*0.75; %the sum of the Angstrom non calibrated parameters equals 0.75 end end ma_aVP = ma_calc_aVP %this scripts calculates the 'actual' vapor pressure in mb (millibars) % this procedure estimates aVP without humidity data as shown in FAO % chapter 3, eq. 14 and eq. 48 % this method assumes that Tmin=Tdew, and that data is in degrees Celsius; it is recommended that this be % checked for the selected region of analysis load ('basin_inputdata', 'm_tmin') num_months=leng th(m_tmin); ma_aVP_kPa=zeros(num_months,1); %mean 'actual' vapor pressure in kPa for i=1:num_months ma_aVP_kPa(i)=0.6108.*exp((17.27.*m_tmin(i))./(m_tmin(i)+237.3)); end ma_aVP=zeros(num_months,1); %mean 'actual' vapor pressure in mb for i=1:num_months ma_aVP(i)=ma_aVP_kPa(i).*10; end end

PAGE 125

125 APPENDIX B INDEPENDENT VARIABLES CALCULATION Land cover calculation m ethods The 1974 cover calculations were based on shapefil es made available by the USGS in their historical archives and pre processed by the University of Florida GeoPlan Center and made available in the Florida Geographic Data Library (FGDL) (accessed on July 21, 2014: http://www.fgdl.org/download/index.html ). The land cover products for 1992 , 2001, 2006 and 2011 were available from the Multi Resolution Land Characteristics Consortium (MRLC) in their N ational L and C over D atabase (NLCD) and accessed on July 21, 2014: http://www.mrlc.gov/about.php . The calculation procedures were carried out using ArcGIS 10.2 . The process consisted in extracting the land cover data from the available raster files using the selected basin shapefile boundaries. The land cover data was a ggregated into Anderson level I classes. Population density calculation method To calculate the area weight of a census block group the following procedures boundaries. Ca lculate area of the basin contain within the block group and calculate the total area of the block group by adding land and water areas (units are in square meters). Then, calculate the ratio of each portion of the census block group contained within the b asin and the total area of its corresponding block group area. Estimated population equals the contribution of each census bloc k group population to the basin area weight of the census block group that the corresponding census blocks group total population . Finally, total population is divided by basin area to derive population density.

PAGE 126

126 Time invariant parameters The selected morphometric variables were extracted from the GAGES II dataset with the exception of the circularity ratio, which was calculated by e quating the basin area to the area of a circle with equal perimeter .

PAGE 127

127 APPENDIX C CASE STUDIES SUMMARY The case studies were extracted from the Gages II dataset. The selected basins are located in the State of Florida and within the South East Coa stal P lain Ecoregion (Wiken et al. 2011) . Each case is identified by its stream flow gage number provided by the U.S. Geological Survey. The basins included have at least 30 consecutive years of stream flow records. For further understanding of the catchments behavior, the remarks from the Water Report were included into the screening process. Only basins with un regulated flows were used in this project , which were assumed to be those without any remarks . Table C 20 presents the complete list of potential case studies.

PAGE 128

128 USGS 02231000 , S t . M arys R iver near M acClenny Description : Latitude 30°21'31", Longitude 82°04'54" NAD27 Baker County, Florida, Hydrologic Unit 03070204 Drainage area: 700 square miles Table C 1 . Time series of l and cover types and population density for St. Marys River basin (USGS 02231000 ) YEAR _10 AG URB FOR WAT WET BAR GRA PD 1980 3.68 0.0324 0.0218 0.6029 0.0065 0.3107 0.0007 0.0249 8.26 1981 3.57 0.0318 0.0232 0.5949 0.0065 0.3138 0.0006 0.0291 8.44 1982 3.57 0.0312 0.0246 0.5868 0.0065 0.3170 0.0006 0.0332 8.61 1983 3.85 0.0307 0.0260 0.5787 0.0065 0.3202 0.0006 0.0374 8.79 1984 3.70 0.0301 0.0274 0.5706 0.0064 0.3234 0.0006 0.0415 8.97 1985 3.71 0.0295 0.0288 0.5625 0.0064 0.3266 0.0006 0.0457 9.15 1986 3.44 0.0289 0.0302 0.5545 0.0064 0.3297 0.0005 0.0498 9.32 1987 3.37 0.0283 0.0316 0.5464 0.0063 0.3329 0.0005 0.0540 9.50 1988 3.57 0.0277 0.0329 0.5383 0.0063 0.3361 0.0005 0.0581 9.68 1989 3.54 0.0271 0.0343 0.5302 0.0063 0.3393 0.0005 0.0623 9.85 1990 3.47 0.0265 0.0357 0.5222 0.0063 0.3424 0.0004 0.0664 10.03 1991 3.49 0.0259 0.0371 0.5141 0.0062 0.3456 0.0004 0.0706 10.24 1992 3.62 0.0253 0.0385 0.5060 0.0062 0.3488 0.0004 0.0747 10.46 1993 3.64 0.0255 0.0396 0.4921 0.0062 0.3499 0.0004 0.0861 10.67 1994 3.77 0.0258 0.0408 0.4783 0.0062 0.3510 0.0005 0.0974 10.88 1995 3.84 0.0260 0.0419 0.4644 0.0062 0.3521 0.0005 0.1088 11.10 1996 3.93 0.0263 0.0430 0.4505 0.0062 0.3532 0.0006 0.1202 11.31 1997 4.10 0.0265 0.0442 0.4367 0.0061 0.3542 0.0006 0.1315 11.52 1998 3.75 0.0268 0.0453 0.4228 0.0061 0.3553 0.0007 0.1429 11.73 1999 3.76 0.0270 0.0464 0.4089 0.0061 0.3564 0.0007 0.1543 11.95 2000 3.97 0.0273 0.0476 0.3951 0.0061 0.3575 0.0008 0.1656 12.16 2001 4.07 0.0275 0.0487 0.3812 0.0061 0.3586 0.0008 0.1770 12.43 2002 4.35 0.0275 0.0488 0.3797 0.0061 0.3574 0.0008 0.1797 12.69 2003 4.10 0.0275 0.0489 0.3781 0.0061 0.3562 0.0008 0.1824 12.96 2004 4.20 0.0275 0.0491 0.3766 0.0062 0.3549 0.0007 0.1850 13.22 2005 4.22 0.0275 0.0492 0.3750 0.0062 0.3537 0.0007 0.1877 13.49 2006 4.24 0.0275 0.0493 0.3735 0.0062 0.3525 0.0007 0.1904 13.76 2007 4.41 0.0273 0.0495 0.3710 0.0062 0.3544 0.0008 0.1909 14.02 2008 4.84 0.0272 0.0497 0.3685 0.0062 0.3562 0.0009 0.1914 14.29 2009 5.00 0.0270 0.0498 0.3659 0.0063 0.3581 0.0009 0.1920 14.55 2010 4.94 0.0269 0.0500 0.3634 0.0063 0.3599 0.0010 0.1925 14.82

PAGE 129

1 29 Figure C 1 . Catchment coefficient for St. Marys River basin (USGS 02231000 ) , 1950 2010 y = 0.0063x 8.5473 R² = 0.0653 1.00 2.00 3.00 4.00 5.00 6.00 7.00 1950 1960 1970 1980 1990 2000 2010 Catchment coefficient, Year (t) Linear (

PAGE 130

130 USG S 02231 342, F t. Drum C reek at S unshine S t. P ky. Near F t. D rum Description : Latitude 27°34'06", Longitude 80°47'47" NAD27 Okeechobee County, Florida, Hydrologic Unit 03080101 Drainage area: 52.6 square miles Table C 2 . Time series of land cover types and population density for Ft. Drum Creek basin (USGS 02231 342) YEAR _10 AG URB FOR WAT WET BAR GRA PD 1980 0.5970 0.0192 0.0509 0.0129 0.1895 0.0000 0.1275 1.34 1981 0.5848 0.0222 0.0502 0.0123 0.2089 0.0000 0.1188 1.37 1982 0.5726 0.0253 0.0496 0.0116 0.2283 0.0000 0.1100 1.41 1983 0.5605 0.0284 0.0489 0.0110 0.2478 0.0001 0.1013 1.44 1984 0.5483 0.0314 0.0482 0.0104 0.2672 0.0001 0.0925 1.48 1985 0.5361 0.0345 0.0476 0.0097 0.2866 0.0001 0.0838 1.51 1986 0.5239 0.0375 0.0469 0.0091 0.3060 0.0001 0.0750 1.54 1987 2.78 0.5118 0.0406 0.0462 0.0085 0.3254 0.0001 0.0663 1.58 1988 2.91 0.4996 0.0437 0.0456 0.0078 0.3448 0.0001 0.0575 1.61 1989 3.07 0.4874 0.0467 0.0449 0.0072 0.3643 0.0001 0.0488 1.65 1990 3.04 0.4752 0.0498 0.0442 0.0066 0.3837 0.0001 0.0400 1.68 1991 3.18 0.4631 0.0528 0.0436 0.0059 0.4031 0.0001 0.0313 1.80 1992 3.29 0.4509 0.0559 0.0429 0.0053 0.4225 0.0001 0.0225 1.91 1993 3.11 0.4524 0.0563 0.0421 0.0050 0.4171 0.0001 0.0270 2.03 1994 3.14 0.4540 0.0567 0.0414 0.0048 0.4117 0.0001 0.0314 2.15 1995 3.04 0.4555 0.0570 0.0406 0.0045 0.4063 0.0002 0.0359 2.27 1996 3.15 0.4571 0.0574 0.0398 0.0043 0.4009 0.0002 0.0404 2.38 1997 3.17 0.4586 0.0578 0.0391 0.0040 0.3955 0.0002 0.0448 2.50 1998 3.12 0.4602 0.0582 0.0383 0.0038 0.3901 0.0002 0.0493 2.62 1999 2.98 0.4617 0.0585 0.0375 0.0035 0.3847 0.0003 0.0538 2.73 2000 2.87 0.4633 0.0589 0.0368 0.0033 0.3793 0.0003 0.0582 2.85 2001 2.70 0.4648 0.0593 0.0360 0.0030 0.3739 0.0003 0.0627 2.95 2002 2.66 0.4648 0.0593 0.0360 0.0030 0.3739 0.0003 0.0627 3.05 2003 2.74 0.4648 0.0593 0.0360 0.0030 0.3739 0.0003 0.0627 3.15 2004 2.85 0.4648 0.0593 0.0360 0.0030 0.3739 0.0003 0.0627 3.25 2005 2.99 0.4648 0.0593 0.0360 0.0030 0.3739 0.0003 0.0627 3.36 2006 2.85 0.4648 0.0593 0.0360 0.0030 0.3739 0.0003 0.0627 3.46 2007 2.85 0.4648 0.0593 0.0360 0.0029 0.3740 0.0003 0.0627 3.56 2008 2.86 0.4648 0.0593 0.0360 0.0028 0.3741 0.0003 0.0627 3.66 2009 2.96 0.4648 0.0592 0.0360 0.0028 0.3743 0.0003 0.0626 3.76 2010 3.12 0.4648 0.0592 0.0360 0.0027 0.3744 0.0003 0.0626 3.86

PAGE 131

131 Figure C 2 . Catchment coefficient for Ft. Drum Creek basin (USGS 02231 342 ) , 19 8 0 2010 y = 0.0085x + 19.866 R² = 0.1253 1.00 2.00 3.00 4.00 5.00 6.00 7.00 1980 1985 1990 1995 2000 2005 2010 Catchment coefficient, Year (t) Linear (

PAGE 132

132 USG S 02234400, G ee Creek near L ongwood Description : Latitude 28°42'14", Longitude 81°17'27" NAD27 Seminole County, Florida, Hydrologic Unit 03080101 Drainage area: 12.8 square miles Table C 3. Time series of land cover types and population density for Gee Creek basin (USGS 0223 4400) YEAR _10 AG URB FOR WAT WET BAR GRA PD 1980 0.0453 0.5277 0.1341 0.0567 0.1579 0.0768 0.0012 564.53 1981 0.0434 0.5389 0.1255 0.0562 0.1639 0.0705 0.0014 597.35 1982 0.0414 0.5500 0.1169 0.0558 0.1700 0.0641 0.0016 630.17 1983 0.0395 0.5612 0.1083 0.0554 0.1760 0.0577 0.0019 662.98 1984 0.0376 0.5723 0.0996 0.0550 0.1820 0.0513 0.0021 695.80 1985 0.0356 0.5835 0.0910 0.0546 0.1881 0.0449 0.0023 728.62 1986 0.0337 0.5946 0.0824 0.0541 0.1941 0.0386 0.0025 761.44 1987 0.0317 0.6058 0.0738 0.0537 0.2002 0.0322 0.0027 794.26 1988 0.0298 0.6169 0.0652 0.0533 0.2062 0.0258 0.0029 827.07 1989 0.0278 0.6281 0.0566 0.0529 0.2123 0.0194 0.0031 859.89 1990 0.0259 0.6392 0.0479 0.0524 0.2183 0.0131 0.0033 892.71 1991 0.0239 0.6504 0.0393 0.0520 0.2244 0.0067 0.0035 896.51 1992 0.0220 0.6615 0.0307 0.0516 0.2304 0.0003 0.0037 900.31 1993 0.0196 0.6674 0.0321 0.0511 0.2248 0.0007 0.0049 904.12 1994 0.0171 0.6732 0.0335 0.0506 0.2193 0.0011 0.0062 907.92 1995 3.23 0.0147 0.6791 0.0349 0.0501 0.2137 0.0015 0.0074 911.72 1996 3.27 0.0122 0.6849 0.0363 0.0496 0.2081 0.0019 0.0087 915.52 1997 3.49 0.0098 0.6908 0.0377 0.0491 0.2026 0.0022 0.0099 919.32 1998 3.27 0.0073 0.6966 0.0391 0.0486 0.1970 0.0026 0.0112 923.13 1999 3.32 0.0049 0.7025 0.0405 0.0481 0.1914 0.0030 0.0124 926.93 2000 3.25 0.0024 0.7083 0.0419 0.0476 0.1859 0.0034 0.0137 930.73 2001 3.27 0.0000 0.7142 0.0433 0.0471 0.1803 0.0038 0.0149 934.76 2002 3.40 0.0000 0.7142 0.0433 0.0471 0.1796 0.0038 0.0149 938.79 2003 3.24 0.0000 0.7142 0.0433 0.0471 0.1789 0.0038 0.0149 942.82 2004 3.33 0.0000 0.7142 0.0433 0.0471 0.1781 0.0038 0.0149 946.85 2005 3.43 0.0000 0.7142 0.0433 0.0471 0.1774 0.0038 0.0149 950.88 2006 3.37 0.0000 0.7142 0.0433 0.0471 0.1767 0.0038 0.0149 954.90 2007 3.24 0.0000 0.7154 0.0432 0.0471 0.1756 0.0034 0.0152 958.93 2008 3.39 0.0000 0.7166 0.0431 0.0471 0.1746 0.0031 0.0155 962.96 2009 3.33 0.0000 0.7177 0.0431 0.0470 0.1735 0.0027 0.0159 966.99 2010 3.29 0.0000 0.7189 0.0430 0.0470 0.1725 0.0024 0.0162 971.02

PAGE 133

133 Figure C 3. Catchment coefficient for Gee Creek basin (USGS 0223 4400 ) , 19 8 0 2010 y = 0.0023x 1.2049 R² = 0.0197 1.00 2.00 3.00 4.00 5.00 6.00 7.00 1980 1985 1990 1995 2000 2005 2010 Catchment coefficient, Year (t) Linear (

PAGE 134

134 USG S 02235000 , W ekiva River near Sanford Description : Latitude 28°48'54", Longitude 81°25'10" NAD27 Seminole County, Florida, Hydrologic Unit 03080101 Drainage area: 189 square miles Datum of gage: 4.96 feet above NGVD29 Table C 4 . Time series of land cover types and population de nsity for Wekiva River basin (USGS 0223 5 000 ) YEAR _10 AG URB FOR WAT WET BAR GRA PD 1980 1.77 0.2754 0.2999 0.1761 0.0325 0.1857 0.0110 0.0194 351.20 1981 1.82 0.2633 0.3118 0.1712 0.0321 0.1894 0.0102 0.0221 367.46 1982 1.87 0.2512 0.3237 0.1663 0.0317 0.1930 0.0093 0.0248 383.72 1983 1.90 0.2392 0.3357 0.1614 0.0313 0.1966 0.0085 0.0275 399.97 1984 1.90 0.2271 0.3476 0.1564 0.0309 0.2002 0.0076 0.0302 416.23 1985 1.90 0.2150 0.3595 0.1515 0.0305 0.2038 0.0068 0.0329 432.49 1986 1.87 0.2029 0.3714 0.1466 0.0301 0.2075 0.0059 0.0356 448.75 1987 1.90 0.1909 0.3833 0.1417 0.0297 0.2111 0.0051 0.0383 465.01 1988 1.96 0.1788 0.3952 0.1368 0.0293 0.2147 0.0042 0.0410 481.26 1989 1.97 0.1667 0.4072 0.1319 0.0289 0.2183 0.0034 0.0437 497.52 1990 1.97 0.1546 0.4191 0.1269 0.0285 0.2220 0.0025 0.0464 513.78 1991 2.03 0.1426 0.4310 0.1220 0.0281 0.2256 0.0017 0.0491 525.28 1992 1.99 0.1305 0.4429 0.1171 0.0277 0.2292 0.0008 0.0518 536.78 1993 1.94 0.1238 0.4502 0.1193 0.0273 0.2247 0.0009 0.0538 548.28 1994 2.01 0.1171 0.4576 0.1215 0.0269 0.2203 0.0010 0.0557 559.78 1995 2.00 0.1104 0.4649 0.1237 0.0265 0.2158 0.0011 0.0577 571.28 1996 2.02 0.1037 0.4722 0.1259 0.0261 0.2113 0.0012 0.0596 582.77 1997 2.09 0.0970 0.4796 0.1280 0.0258 0.2069 0.0012 0.0616 594.27 1998 1.97 0.0903 0.4869 0.1302 0.0254 0.2024 0.0013 0.0635 605.77 1999 1.98 0.0836 0.4942 0.1324 0.0250 0.1979 0.0014 0.0655 617.27 2000 1.95 0.0769 0.5016 0.1346 0.0246 0.1935 0.0015 0.0674 628.77 2001 1.93 0.0702 0.5089 0.1368 0.0242 0.1890 0.0016 0.0694 633.52 2002 1.95 0.0693 0.5105 0.1363 0.0247 0.1879 0.0021 0.0693 638.28 2003 1.93 0.0684 0.5121 0.1358 0.0253 0.1869 0.0025 0.0691 643.03 2004 1.93 0.0675 0.5136 0.1352 0.0258 0.1858 0.0030 0.0690 647.79 2005 1.97 0.0666 0.5152 0.1347 0.0264 0.1848 0.0034 0.0688 652.54 2006 1.91 0.0657 0.5168 0.1342 0.0269 0.1837 0.0039 0.0687 657.29 2007 1.86 0.0650 0.5194 0.1337 0.0269 0.1827 0.0036 0.0687 662.05 2008 1.99 0.0642 0.5221 0.1333 0.0268 0.1816 0.0033 0.0687 666.80 2009 1.97 0.0635 0.5247 0.1328 0.0268 0.1806 0.0030 0.0686 671.56 2010 2.01 0.0627 0.5273 0.1324 0.0267 0.1796 0.0027 0.0686 676.31

PAGE 135

135 Figure C 4 . Catchment coefficient for Wekiva River basin (USGS 0223 5 000 ) , 1950 2010 y = 0.0039x 5.9309 R² = 0.2282 1.00 2.00 3.00 4.00 5.00 6.00 7.00 1950 1960 1970 1980 1990 2000 2010 Catchment coefficient, Year (t) Linear (

PAGE 136

136 USG S 02245500, S outh Black Fork Creek near Penney Farms Description : Latitude 29°58'45", Longitude 81°51'08" NAD27 Clay County, Florida, Hydrologic Unit 03080103 Drainage area: 134 square miles Datum of gage: 9.82 feet above NGVD29 Table C 5 . Time series of land cover types and population density for South Black Fork Creek basin (USGS 022 455 00 ) YEAR _10 AG URB FOR WAT WET BAR GRA PD 1980 3.47 0.0252 0.0315 0.7389 0.0078 0.1158 0.0071 0.0738 5.48 1981 3.53 0.0249 0.0339 0.7258 0.0079 0.1175 0.0071 0.0830 5.81 1982 3.55 0.0246 0.0363 0.7127 0.0080 0.1192 0.0071 0.0921 6.15 1983 3.97 0.0243 0.0387 0.6997 0.0082 0.1209 0.0072 0.1013 6.48 1984 3.64 0.0240 0.0410 0.6866 0.0083 0.1226 0.0072 0.1104 6.82 1985 3.81 0.0237 0.0434 0.6735 0.0084 0.1243 0.0072 0.1196 7.15 1986 3.69 0.0234 0.0458 0.6604 0.0085 0.1260 0.0072 0.1287 7.48 1987 3.50 0.0231 0.0482 0.6473 0.0087 0.1277 0.0073 0.1379 7.82 1988 3.40 0.0228 0.0506 0.6342 0.0088 0.1294 0.0073 0.1470 8.15 1989 3.34 0.0225 0.0530 0.6211 0.0089 0.1311 0.0073 0.1562 8.49 1990 3.42 0.0222 0.0554 0.6080 0.0090 0.1328 0.0073 0.1653 8.82 1991 3.75 0.0219 0.0578 0.5949 0.0092 0.1345 0.0074 0.1745 9.14 1992 3.94 0.0216 0.0602 0.5818 0.0093 0.1362 0.0074 0.1836 9.47 1993 3.87 0.0210 0.0612 0.5703 0.0092 0.1371 0.0076 0.1936 9.79 1994 4.21 0.0204 0.0622 0.5589 0.0091 0.1381 0.0078 0.2037 10.11 1995 3.90 0.0198 0.0631 0.5474 0.0089 0.1390 0.0080 0.2137 10.44 1996 3.97 0.0192 0.0641 0.5360 0.0088 0.1399 0.0082 0.2237 10.76 1997 3.86 0.0186 0.0651 0.5245 0.0087 0.1409 0.0085 0.2338 11.08 1998 3.62 0.0180 0.0661 0.5131 0.0086 0.1418 0.0087 0.2438 11.40 1999 3.69 0.0174 0.0670 0.5016 0.0084 0.1427 0.0089 0.2538 11.73 2000 3.72 0.0168 0.0680 0.4902 0.0083 0.1437 0.0091 0.2639 12.05 2001 3.59 0.0162 0.0690 0.4787 0.0082 0.1446 0.0093 0.2739 11.90 2002 3.58 0.0162 0.0690 0.4757 0.0081 0.1445 0.0095 0.2768 11.76 2003 3.45 0.0162 0.0690 0.4727 0.0081 0.1445 0.0098 0.2796 11.61 2004 3.37 0.0162 0.0690 0.4698 0.0080 0.1444 0.0100 0.2825 11.46 2005 3.64 0.0162 0.0690 0.4668 0.0080 0.1444 0.0103 0.2853 11.32 2006 3.54 0.0162 0.0690 0.4638 0.0079 0.1443 0.0105 0.2882 11.17 2007 4.28 0.0162 0.0690 0.4542 0.0079 0.1442 0.0108 0.2977 11.02 2008 4.72 0.0161 0.0690 0.4446 0.0079 0.1442 0.0110 0.3071 10.87 2009 4.59 0.0161 0.0690 0.4350 0.0079 0.1441 0.0113 0.3166 10.73 2010 4.52 0.0160 0.0690 0.4255 0.0079 0.1440 0.0116 0.3261 10.58

PAGE 137

137 Figure C 5 . Catchment coefficient for South Black Fork Creek basin (USGS 022 455 00 ) , 1950 2010 y = 0.0205x 37.13 R² = 0.5159 1.00 2.00 3.00 4.00 5.00 6.00 7.00 1950 1960 1970 1980 1990 2000 2010 Catchment coefficient, Year (t) Linear (

PAGE 138

138 USG S 02247510, T omoka River near Holly Hill Description : Latitude 29°13'02", Longitude 81°06'32" NAD27 Volusia County, Florida, Hydrologic Unit 03080201 Drainage area: 76.8 square miles Table C 6 . Time series of land cover types and population density for Tomoka Ri ver basin (USGS 022 4751 0 ) YEAR _10 AG URB FOR WAT WET BAR GRA PD 1980 3.16 0.0201 0.1232 0.4249 0.0103 0.3947 0.0033 0.0235 33.84 1981 3.21 0.0202 0.1294 0.4205 0.0103 0.3905 0.0032 0.0260 37.33 1982 3.37 0.0203 0.1355 0.4162 0.0104 0.3862 0.0030 0.0285 40.83 1983 3.63 0.0204 0.1416 0.4118 0.0104 0.3820 0.0028 0.0310 44.32 1984 3.48 0.0205 0.1477 0.4074 0.0104 0.3778 0.0026 0.0335 47.82 1985 3.38 0.0206 0.1538 0.4031 0.0105 0.3735 0.0024 0.0360 51.31 1986 3.52 0.0207 0.1600 0.3987 0.0105 0.3693 0.0023 0.0385 54.80 1987 3.64 0.0208 0.1661 0.3944 0.0106 0.3651 0.0021 0.0410 58.30 1988 3.80 0.0210 0.1722 0.3900 0.0106 0.3608 0.0019 0.0434 61.79 1989 3.96 0.0211 0.1783 0.3857 0.0107 0.3566 0.0017 0.0459 65.29 1990 3.94 0.0212 0.1845 0.3813 0.0107 0.3524 0.0016 0.0484 68.78 1991 3.96 0.0213 0.1906 0.3770 0.0108 0.3481 0.0014 0.0509 71.25 1992 3.88 0.0214 0.1967 0.3726 0.0108 0.3439 0.0012 0.0534 73.72 1993 3.70 0.0216 0.1978 0.3566 0.0111 0.3493 0.0016 0.0619 76.19 1994 4.53 0.0219 0.1989 0.3407 0.0114 0.3547 0.0020 0.0704 78.66 1995 4.61 0.0221 0.2000 0.3247 0.0118 0.3601 0.0024 0.0789 81.13 1996 4.33 0.0224 0.2011 0.3087 0.0121 0.3655 0.0028 0.0874 83.59 1997 4.44 0.0226 0.2022 0.2928 0.0124 0.3709 0.0031 0.0958 86.06 1998 3.82 0.0229 0.2033 0.2768 0.0127 0.3763 0.0035 0.1043 88.53 1999 3.83 0.0231 0.2044 0.2608 0.0131 0.3817 0.0039 0.1128 91.00 2000 3.83 0.0234 0.2055 0.2449 0.0134 0.3871 0.0043 0.1213 93.47 2001 3.62 0.0236 0.2066 0.2289 0.0137 0.3925 0.0047 0.1298 93.99 2002 3.73 0.0235 0.2074 0.2277 0.0136 0.3924 0.0048 0.1303 94.51 2003 3.60 0.0235 0.2083 0.2265 0.0135 0.3924 0.0048 0.1309 95.03 2004 3.06 0.0234 0.2091 0.2252 0.0135 0.3923 0.0049 0.1314 95.55 2005 3.01 0.0234 0.2100 0.2240 0.0134 0.3923 0.0049 0.1320 96.07 2006 2.96 0.0233 0.2108 0.2228 0.0133 0.3922 0.0050 0.1325 96.58 2007 2.95 0.0232 0.2125 0.2192 0.0133 0.3918 0.0050 0.1350 97.10 2008 3.21 0.0230 0.2142 0.2156 0.0132 0.3914 0.0049 0.1375 97.62 2009 3.10 0.0229 0.2159 0.2120 0.0132 0.3911 0.0049 0.1400 98.14 2010 3.10 0.0228 0.2176 0.2084 0.0131 0.3907 0.0048 0.1426 98.66

PAGE 139

139 Figure C 6 . Catchment coefficient for Tomoka River basin (USGS 022 4751 0 ) , 19 7 0 2010 y = 0.0069x + 17.336 R² = 0.019 1.00 2.00 3.00 4.00 5.00 6.00 7.00 1970 1975 1980 1985 1990 1995 2000 2005 2010 Catchment coefficient, _t Year (t) Linear (

PAGE 140

140 USG S 02263800, S hingle Creek at airport near K issimee Description : Latitude 28°18'14", Longitude 81°27'04" NAD27 Osceola County, Florida, Hydrologic Unit 03090101 Drainage area: 89.2 square miles Datum of gage: 60.66 feet above NGVD29 Table C 7 . Time series of land cover types and population density for Shingle Creek basin (USGS 02263800) YEAR _10 AG URB FOR WAT WET BAR GRA PD 1980 3.80 0.2287 0.3401 0.0786 0.0590 0.1439 0.0155 0.1342 322.35 1981 3.79 0.2193 0.3573 0.0751 0.0592 0.1508 0.0142 0.1242 344.12 1982 3.69 0.2099 0.3744 0.0716 0.0593 0.1576 0.0130 0.1143 365.89 1983 3.70 0.2005 0.3915 0.0681 0.0594 0.1645 0.0118 0.1044 387.66 1984 3.64 0.1910 0.4086 0.0645 0.0595 0.1713 0.0106 0.0944 409.43 1985 3.53 0.1816 0.4257 0.0610 0.0596 0.1781 0.0094 0.0845 431.21 1986 3.44 0.1722 0.4429 0.0575 0.0598 0.1850 0.0081 0.0745 452.98 1987 3.46 0.1628 0.4600 0.0540 0.0599 0.1918 0.0069 0.0646 474.75 1988 3.55 0.1534 0.4771 0.0505 0.0600 0.1986 0.0057 0.0547 496.52 1989 3.56 0.1440 0.4942 0.0470 0.0601 0.2055 0.0045 0.0447 518.29 1990 3.56 0.1346 0.5114 0.0435 0.0603 0.2123 0.0032 0.0348 540.06 1991 3.52 0.1252 0.5285 0.0400 0.0604 0.2192 0.0020 0.0248 563.30 1992 3.62 0.1158 0.5456 0.0365 0.0605 0.2260 0.0008 0.0149 586.53 1993 3.46 0.1060 0.5567 0.0358 0.0609 0.2207 0.0010 0.0190 609.77 1994 3.69 0.0961 0.5678 0.0351 0.0612 0.2154 0.0013 0.0232 633.00 1995 3.63 0.0863 0.5788 0.0344 0.0616 0.2101 0.0015 0.0273 656.24 1996 3.81 0.0765 0.5899 0.0337 0.0620 0.2048 0.0017 0.0315 679.47 1997 4.06 0.0666 0.6010 0.0329 0.0623 0.1995 0.0020 0.0356 702.71 1998 3.64 0.0568 0.6121 0.0322 0.0627 0.1942 0.0022 0.0398 725.94 1999 3.64 0.0470 0.6231 0.0315 0.0631 0.1889 0.0024 0.0439 749.18 2000 3.63 0.0371 0.6342 0.0308 0.0634 0.1836 0.0027 0.0481 772.41 2001 3.52 0.0273 0.6453 0.0301 0.0638 0.1783 0.0029 0.0522 789.03 2002 3.49 0.0264 0.6491 0.0300 0.0644 0.1753 0.0034 0.0514 805.65 2003 3.55 0.0254 0.6528 0.0298 0.0649 0.1724 0.0039 0.0506 822.27 2004 3.39 0.0245 0.6566 0.0297 0.0655 0.1694 0.0044 0.0498 838.89 2005 3.54 0.0235 0.6603 0.0295 0.0660 0.1665 0.0049 0.0490 855.52 2006 3.38 0.0226 0.6641 0.0294 0.0666 0.1635 0.0054 0.0482 872.14 2007 3.09 0.0214 0.6671 0.0293 0.0669 0.1610 0.0062 0.0478 888.76 2008 3.25 0.0203 0.6700 0.0292 0.0672 0.1585 0.0071 0.0475 905.38 2009 3.03 0.0191 0.6730 0.0292 0.0675 0.1561 0.0079 0.0471 922.00 2010 3.06 0.0180 0.6759 0.0291 0.0678 0.1536 0.0088 0.0468 938.62

PAGE 141

141 Figure C 7 . Catchment coefficient for Shingle Creek basin (USGS 02263800 ) , 1950 2010 y = 0.0399x + 83.286 R² = 0.6629 1.00 2.00 3.00 4.00 5.00 6.00 7.00 1950 1960 1970 1980 1990 2000 2010 Catchment coefficient, Year (t) Linear (

PAGE 142

142 USG S 02266480, D avenport Creek near L oughman Description : Latitude 28°16'15", Longitude 81°35'28" NAD27 Osceola County, Florida, Hydrologic Unit 03090101 Drainage area: 23.0 square miles Datum of gage: 77.69 feet above NGVD29 Table C 8 . Time series of land cover types and population density for Davenport Creek basi n (USGS 022 6648 0 ) YEAR _10 AG URB FOR WAT WET BAR GRA PD 1980 4.77 0.5164 0.0653 0.0206 0.0264 0.3199 0.0075 0.0437 2.04 1981 4.81 0.5037 0.0735 0.0240 0.0249 0.3224 0.0079 0.0435 2.19 1982 4.56 0.4910 0.0817 0.0275 0.0233 0.3249 0.0083 0.0432 2.35 1983 4.30 0.4783 0.0899 0.0309 0.0218 0.3275 0.0087 0.0430 2.50 1984 3.92 0.4656 0.0981 0.0343 0.0202 0.3300 0.0090 0.0428 2.66 1985 3.77 0.4529 0.1063 0.0378 0.0186 0.3325 0.0094 0.0425 2.81 1986 3.56 0.4402 0.1145 0.0412 0.0171 0.3350 0.0098 0.0423 2.96 1987 3.60 0.4275 0.1227 0.0446 0.0155 0.3375 0.0102 0.0420 3.12 1988 3.59 0.4147 0.1308 0.0481 0.0139 0.3400 0.0106 0.0418 3.27 1989 3.51 0.4020 0.1390 0.0515 0.0124 0.3425 0.0110 0.0415 3.43 1990 3.56 0.3893 0.1472 0.0549 0.0108 0.3450 0.0113 0.0413 3.58 1991 3.42 0.3766 0.1554 0.0584 0.0093 0.3475 0.0117 0.0410 4.27 1992 3.61 0.3639 0.1636 0.0618 0.0077 0.3500 0.0121 0.0408 4.97 1993 3.42 0.3575 0.1674 0.0598 0.0073 0.3519 0.0120 0.0440 5.66 1994 3.89 0.3511 0.1712 0.0578 0.0070 0.3538 0.0118 0.0473 6.36 1995 3.72 0.3447 0.1750 0.0558 0.0066 0.3556 0.0117 0.0505 7.05 1996 3.86 0.3383 0.1788 0.0538 0.0062 0.3575 0.0116 0.0537 7.74 1997 3.81 0.3320 0.1825 0.0519 0.0059 0.3594 0.0114 0.0570 8.44 1998 3.42 0.3256 0.1863 0.0499 0.0055 0.3613 0.0113 0.0602 9.13 1999 3.32 0.3192 0.1901 0.0479 0.0051 0.3631 0.0112 0.0634 9.83 2000 3.24 0.3128 0.1939 0.0459 0.0048 0.3650 0.0110 0.0667 10.52 2001 3.34 0.3064 0.1977 0.0439 0.0044 0.3669 0.0109 0.0699 12.66 2002 3.39 0.2958 0.2042 0.0429 0.0044 0.3657 0.0166 0.0704 14.79 2003 3.13 0.2852 0.2107 0.0418 0.0044 0.3646 0.0224 0.0710 16.93 2004 2.93 0.2746 0.2172 0.0408 0.0044 0.3634 0.0281 0.0715 19.06 2005 2.94 0.2640 0.2237 0.0397 0.0044 0.3623 0.0339 0.0721 21.20 2006 2.83 0.2534 0.2302 0.0387 0.0044 0.3611 0.0396 0.0726 23.34 2007 2.80 0.2453 0.2438 0.0357 0.0048 0.3585 0.0380 0.0738 25.47 2008 3.11 0.2372 0.2574 0.0327 0.0052 0.3559 0.0365 0.0750 27.61 2009 3.10 0.2291 0.2711 0.0298 0.0056 0.3533 0.0349 0.0763 29.74 2010 3.07 0.2210 0.2847 0.0268 0.0060 0.3507 0.0334 0.0775 31.88

PAGE 143

143 Figure C 8 . Catchment coefficient for Davenport Creek basin (USGS 022 6648 0 ) , 19 7 0 2010 y = 0.0484x + 100.13 R² = 0.7406 1.00 2.00 3.00 4.00 5.00 6.00 7.00 1970 1975 1980 1985 1990 1995 2000 2005 2010 Catchment coefficient, Year (t) Linear (

PAGE 144

144 USG S 02267000, C atfish Creek near Lake Wales Description : Latitude 27°57'40", Longitude 81°29'48" NAD27 Polk County, Florida, Hydrologic Unit 03090101 Drainage area: 58.9 square miles Datum of gage: 72.70 feet above NGVD29 Table C 9 . Time series of land cover types and population density for Catfish Creek basin (USGS 022 67 000 ) YEAR _10 AG URB FOR W AT WET BAR GRA PD 1980 3.75 0.4814 0.1021 0.1289 0.1137 0.0990 0.0193 0.0557 44.42 1981 3.72 0.4756 0.1105 0.1201 0.1134 0.1033 0.0182 0.0589 46.42 1982 4.02 0.4698 0.1189 0.1114 0.1132 0.1076 0.0171 0.0621 48.42 1983 3.99 0.4640 0.1273 0.1027 0.1130 0.1120 0.0160 0.0653 50.42 1984 3.81 0.4581 0.1356 0.0940 0.1127 0.1163 0.0149 0.0685 52.42 1985 3.96 0.4523 0.1440 0.0853 0.1125 0.1206 0.0138 0.0717 54.43 1986 4.05 0.4465 0.1524 0.0765 0.1122 0.1249 0.0127 0.0749 56.43 1987 4.13 0.4407 0.1608 0.0678 0.1120 0.1292 0.0116 0.0781 58.43 1988 4.21 0.4349 0.1692 0.0591 0.1118 0.1335 0.0104 0.0812 60.43 1989 4.24 0.4291 0.1776 0.0504 0.1115 0.1379 0.0093 0.0844 62.43 1990 4.48 0.4232 0.1859 0.0416 0.1113 0.1422 0.0082 0.0876 64.43 1991 4.84 0.4174 0.1943 0.0329 0.1110 0.1465 0.0071 0.0908 65.23 1992 4.76 0.4116 0.2027 0.0242 0.1108 0.1508 0.0060 0.0940 66.03 1993 4.50 0.4123 0.1996 0.0247 0.1110 0.1475 0.0063 0.0987 66.83 1994 5.00 0.4129 0.1966 0.0251 0.1112 0.1441 0.0066 0.1035 67.63 1995 5.02 0.4136 0.1935 0.0256 0.1115 0.1408 0.0069 0.1082 68.43 1996 4.92 0.4142 0.1905 0.0261 0.1117 0.1375 0.0072 0.1129 69.22 1997 4.95 0.4149 0.1874 0.0265 0.1119 0.1341 0.0075 0.1177 70.02 1998 4.56 0.4155 0.1844 0.0270 0.1121 0.1308 0.0078 0.1224 70.82 1999 4.54 0.4162 0.1813 0.0275 0.1124 0.1275 0.0081 0.1271 71.62 2000 4.31 0.4168 0.1783 0.0279 0.1126 0.1241 0.0084 0.1319 72.42 2001 4.52 0.4175 0.1752 0.0284 0.1128 0.1208 0.0087 0.1366 74.79 2002 4.64 0.4163 0.1759 0.0284 0.1131 0.1207 0.0092 0.1364 77.16 2003 4.84 0.4151 0.1766 0.0284 0.1134 0.1205 0.0098 0.1362 79.54 2004 5.36 0.4139 0.1774 0.0283 0.1138 0.1204 0.0103 0.1359 81.91 2005 5.51 0.4127 0.1781 0.0283 0.1141 0.1202 0.0109 0.1357 84.28 2006 5.22 0.4115 0.1788 0.0283 0.1144 0.1201 0.0114 0.1355 86.65 2007 4.90 0.4102 0.1796 0.0283 0.1147 0.1201 0.0115 0.1356 89.02 2008 5.73 0.4089 0.1803 0.0283 0.1150 0.1200 0.0117 0.1357 91.40 2009 5.76 0.4077 0.1811 0.0283 0.1153 0.1200 0.0118 0.1359 93.77 2010 6.14 0.4064 0.1818 0.0283 0.1156 0.1199 0.0120 0.1360 96.14

PAGE 145

145 Figure C 9 . Catchment coefficient for Catfish Creek basin (USGS 022 67 000 ) , 1950 2010 y = 0.0401x 75.359 R² = 0.7708 1.00 2.00 3.00 4.00 5.00 6.00 7.00 1950 1960 1970 1980 1990 2000 2010 Catchment coefficient, Year (t) Linear (

PAGE 146

146 USGS 02296500, C harlie Creek near Gardner Description : Latitude 27°22'29", Longitude 81°47'48" NAD27 Hardee County, Florida, Hydrologic Unit 03100101 Drainage area: 330 square miles Datum of gage: 21.66 feet above NGVD29 Table C 10 . Time series of land cover t ypes and population density for Charlie Creek basin (USGS 02296500) YEAR _10 AG URB FOR WAT WET BAR GRA PD 1980 3.75 0.5725 0.0211 0.0300 0.0019 0.2205 0.0074 0.1464 10.15 1981 3.81 0.5720 0.0245 0.0294 0.0018 0.2285 0.0070 0.1367 10.51 1982 3.58 0.5714 0.0278 0.0288 0.0018 0.2365 0.0066 0.1271 10.87 1983 3.57 0.5709 0.0311 0.0282 0.0017 0.2445 0.0062 0.1175 11.23 1984 3.77 0.5703 0.0344 0.0275 0.0016 0.2524 0.0057 0.1079 11.59 1985 3.79 0.5697 0.0377 0.0269 0.0016 0.2604 0.0053 0.0983 11.96 1986 4.02 0.5692 0.0411 0.0263 0.0015 0.2684 0.0049 0.0886 12.32 1987 3.92 0.5686 0.0444 0.0256 0.0014 0.2763 0.0045 0.0790 12.68 1988 3.92 0.5680 0.0477 0.0250 0.0014 0.2843 0.0041 0.0694 13.04 1989 3.91 0.5675 0.0510 0.0244 0.0013 0.2923 0.0037 0.0598 13.40 1990 3.96 0.5669 0.0544 0.0238 0.0012 0.3003 0.0032 0.0501 13.76 1991 4.04 0.5664 0.0577 0.0231 0.0012 0.3082 0.0028 0.0405 14.40 1992 4.36 0.5658 0.0610 0.0225 0.0011 0.3162 0.0024 0.0309 15.03 1993 4.54 0.5681 0.0607 0.0224 0.0011 0.3109 0.0024 0.0344 15.67 1994 4.40 0.5705 0.0603 0.0222 0.0010 0.3056 0.0024 0.0378 16.30 1995 4.28 0.5728 0.0600 0.0221 0.0010 0.3003 0.0025 0.0413 16.94 1996 4.08 0.5752 0.0596 0.0219 0.0010 0.2950 0.0025 0.0447 17.58 1997 4.30 0.5775 0.0593 0.0218 0.0009 0.2898 0.0025 0.0482 18.21 1998 4.25 0.5799 0.0589 0.0216 0.0009 0.2845 0.0025 0.0516 18.85 1999 4.27 0.5822 0.0586 0.0215 0.0009 0.2792 0.0026 0.0551 19.48 2000 4.08 0.5846 0.0582 0.0213 0.0008 0.2739 0.0026 0.0585 20.12 2001 3.88 0.5869 0.0579 0.0212 0.0008 0.2686 0.0026 0.0620 19.96 2002 3.79 0.5869 0.0579 0.0212 0.0008 0.2686 0.0026 0.0620 19.79 2003 3.46 0.5868 0.0579 0.0211 0.0009 0.2686 0.0026 0.0620 19.63 2004 3.35 0.5868 0.0580 0.0211 0.0009 0.2686 0.0027 0.0620 19.46 2005 3.28 0.5867 0.0580 0.0210 0.0010 0.2686 0.0027 0.0620 19.30 2006 3.21 0.5867 0.0580 0.0210 0.0010 0.2686 0.0027 0.0620 19.14 2007 3.11 0.5870 0.0580 0.0210 0.0011 0.2685 0.0023 0.0620 18.97 2008 3.30 0.5874 0.0580 0.0210 0.0012 0.2684 0.0020 0.0620 18.81 2009 3.26 0.5877 0.0580 0.0210 0.0013 0.2684 0.0016 0.0620 18.64 2010 3.43 0.5881 0.0580 0.0210 0.0014 0.2683 0.0013 0.0620 18.48

PAGE 147

147 Figure C 10 . Catchment coefficient for Charlie Creek basin (USGS 02296500) , 1950 2010 y = 0.0135x 23.203 R² = 0.2126 1.00 2.00 3.00 4.00 5.00 6.00 7.00 1950 1960 1970 1980 1990 2000 2010 Catchment coefficient, Year (t) Linear (

PAGE 148

148 USG S 02297100, J oshua Creek at Nocatee Description : Latitude 27°09'59", Longitude 81°52'47" NAD27 De Soto County, Florida, Hydrologic Unit 03100101 Drainage area: 132 square miles Datum of gage: 3.94 feet above NGVD29 Table C 11 . Time series of land cover types and population density for Joshua Creek basin (USGS 022 971 00 ) YEAR _10 AG URB FOR WAT WET BAR GRA PD 1980 3.48 0.7055 0.0268 0.0116 0.0043 0.0909 0.0006 0.1604 9.52 1981 3.56 0.7078 0.0293 0.0113 0.0040 0.0985 0.0006 0.1486 9.77 1982 3.58 0.7101 0.0317 0.0110 0.0037 0.1061 0.0006 0.1369 10.03 1983 3.57 0.7124 0.0342 0.0107 0.0034 0.1137 0.0006 0.1252 10.28 1984 3.75 0.7147 0.0366 0.0104 0.0031 0.1212 0.0005 0.1134 10.54 1985 3.80 0.7170 0.0391 0.0101 0.0028 0.1288 0.0005 0.1017 10.79 1986 4.04 0.7193 0.0415 0.0098 0.0025 0.1364 0.0005 0.0899 11.04 1987 4.01 0.7216 0.0440 0.0095 0.0022 0.1439 0.0005 0.0782 11.30 1988 3.95 0.7239 0.0464 0.0092 0.0020 0.1515 0.0005 0.0665 11.55 1989 3.70 0.7262 0.0489 0.0089 0.0017 0.1591 0.0005 0.0547 11.81 1990 3.74 0.7285 0.0513 0.0086 0.0014 0.1667 0.0005 0.0430 12.06 1991 3.67 0.7308 0.0538 0.0083 0.0011 0.1742 0.0005 0.0312 12.55 1992 3.57 0.7331 0.0562 0.0080 0.0008 0.1818 0.0005 0.0195 13.04 1993 3.66 0.7344 0.0560 0.0076 0.0008 0.1796 0.0005 0.0211 13.53 1994 3.55 0.7356 0.0558 0.0071 0.0008 0.1773 0.0005 0.0228 14.02 1995 3.44 0.7369 0.0556 0.0067 0.0008 0.1751 0.0005 0.0244 14.51 1996 3.14 0.7381 0.0554 0.0062 0.0008 0.1729 0.0005 0.0261 14.99 1997 3.15 0.7394 0.0552 0.0058 0.0007 0.1706 0.0005 0.0277 15.48 1998 3.09 0.7406 0.0550 0.0053 0.0007 0.1684 0.0005 0.0294 15.97 1999 3.15 0.7419 0.0548 0.0049 0.0007 0.1662 0.0005 0.0310 16.46 2000 3.10 0.7431 0.0546 0.0044 0.0007 0.1639 0.0005 0.0327 16.95 2001 3.14 0.7444 0.0544 0.0040 0.0007 0.1617 0.0005 0.0343 17.65 2002 3.14 0.7443 0.0545 0.0040 0.0007 0.1617 0.0005 0.0343 18.35 2003 2.94 0.7441 0.0547 0.0040 0.0007 0.1617 0.0005 0.0343 19.06 2004 2.83 0.7440 0.0548 0.0040 0.0006 0.1617 0.0005 0.0343 19.76 2005 2.89 0.7438 0.0550 0.0040 0.0006 0.1617 0.0005 0.0343 20.46 2006 2.89 0.7437 0.0551 0.0040 0.0006 0.1617 0.0005 0.0343 21.16 2007 2.80 0.7436 0.0551 0.0040 0.0007 0.1617 0.0006 0.0343 21.86 2008 2.95 0.7435 0.0551 0.0040 0.0007 0.1617 0.0006 0.0343 22.57 2009 2.97 0.7433 0.0552 0.0040 0.0008 0.1616 0.0007 0.0343 23.27 2010 2.97 0.7432 0.0552 0.0040 0.0008 0.1616 0.0007 0.0343 23.97

PAGE 149

149 Figure C 11 . Catchment coefficient for Joshua Creek basin (USGS 022 971 00 ) , 1950 2010 y = 0.0088x + 20.933 R² = 0.1917 1.00 2.00 3.00 4.00 5.00 6.00 7.00 1950 1960 1970 1980 1990 2000 2010 Catchment coefficient, Year (t) Linear (

PAGE 150

150 USG S 02298123, P rairie Creek near Fort Ogden Description : Latitude 27°03'06", Longitude 81°47'05" NAD27 De Soto County, Florida, Hydrologic Unit 03100101 Drainage area: 233 square miles Datum of gage: 25.00 feet above NGVD29 Table C 1 2. Time series of land cover types and population density for Prairie Creek basin (USGS 022 98123) YEAR _10 AG URB FOR WAT WET BAR GRA PD 1980 0.4370 0.0111 0.0029 0.0033 0.1713 0.0002 0.3735 3.60 1981 0.4477 0.0111 0.0032 0.0031 0.1799 0.0002 0.3540 3.70 1982 0.4583 0.0111 0.0035 0.0030 0.1885 0.0002 0.3346 3.81 1983 0.4690 0.0111 0.0039 0.0029 0.1972 0.0003 0.3152 3.91 1984 0.4797 0.0111 0.0042 0.0028 0.2058 0.0003 0.2958 4.02 1985 0.4903 0.0111 0.0045 0.0027 0.2144 0.0003 0.2764 4.12 1986 0.5010 0.0111 0.0048 0.0025 0.2230 0.0003 0.2569 4.22 1987 2.96 0.5117 0.0111 0.0051 0.0024 0.2316 0.0004 0.2375 4.33 1988 2.96 0.5223 0.0111 0.0054 0.0023 0.2402 0.0004 0.2181 4.43 1989 3.01 0.5330 0.0111 0.0058 0.0022 0.2488 0.0004 0.1987 4.54 1990 3.04 0.5437 0.0111 0.0061 0.0020 0.2574 0.0004 0.1792 4.64 1991 2.91 0.5543 0.0111 0.0064 0.0019 0.2660 0.0005 0.1598 4.84 1992 3.02 0.5650 0.0111 0.0067 0.0018 0.2746 0.0005 0.1404 5.04 1993 3.13 0.5691 0.0111 0.0065 0.0017 0.2718 0.0005 0.1394 5.24 1994 3.17 0.5731 0.0111 0.0063 0.0017 0.2690 0.0005 0.1384 5.44 1995 2.89 0.5772 0.0111 0.0060 0.0016 0.2663 0.0005 0.1374 5.65 1996 2.77 0.5812 0.0111 0.0058 0.0015 0.2635 0.0005 0.1364 5.85 1997 2.81 0.5853 0.0112 0.0056 0.0015 0.2607 0.0006 0.1354 6.05 1998 2.68 0.5893 0.0112 0.0054 0.0014 0.2579 0.0006 0.1344 6.25 1999 2.67 0.5934 0.0112 0.0051 0.0013 0.2552 0.0006 0.1334 6.45 2000 2.63 0.5974 0.0112 0.0049 0.0013 0.2524 0.0006 0.1324 6.65 2001 2.68 0.6015 0.0112 0.0047 0.0012 0.2496 0.0006 0.1314 6.61 2002 2.59 0.6015 0.0112 0.0047 0.0012 0.2496 0.0006 0.1314 6.56 2003 2.47 0.6015 0.0112 0.0047 0.0012 0.2496 0.0006 0.1314 6.52 2004 2.44 0.6015 0.0112 0.0047 0.0012 0.2496 0.0006 0.1314 6.47 2005 2.52 0.6015 0.0112 0.0047 0.0012 0.2496 0.0006 0.1314 6.43 2006 2.51 0.6015 0.0112 0.0047 0.0012 0.2496 0.0006 0.1314 6.39 2007 2.41 0.6012 0.0112 0.0047 0.0013 0.2496 0.0008 0.1313 6.34 2008 2.55 0.6009 0.0112 0.0047 0.0015 0.2496 0.0009 0.1313 6.30 2009 2.54 0.6005 0.0112 0.0047 0.0016 0.2497 0.0011 0.1312 6.25 2010 2.58 0.6002 0.0112 0.0047 0.0018 0.2497 0.0012 0.1312 6.21

PAGE 151

151 Figure C 1 2. Catchment coefficient for Prairie Creek basin (USGS 022 98123 ) , 19 8 0 2010 y = 0.0289x + 60.517 R² = 0.7696 1.00 2.00 3.00 4.00 5.00 6.00 7.00 1980 1985 1990 1995 2000 2005 2010 Catchment coefficient, Year (t) Linear (

PAGE 152

152 USG S 02300700, B ullfrog Creek near Wimauma Description : Latitude 27°47'30", Longitude 82°21'08" NAD27 Hillsborough County, Florida, Hydrologic Unit 03100206 Drainage area: 29.1 square miles Datum of gage: 0.88 feet above NAVD88 Table C 13 . Time series of land cover types and population density for Bullfrog Creek basin (USGS 023 007 00 ) YEAR _10 AG URB FOR WAT WET BAR GRA PD 1980 0.5443 0.0592 0.0843 0.0034 0.1270 0.0009 0.1829 34.79 1981 0.5425 0.0648 0.0807 0.0033 0.1395 0.0011 0.1700 37.65 1982 0.5406 0.0703 0.0771 0.0033 0.1519 0.0012 0.1572 40.51 1983 0.5388 0.0759 0.0735 0.0033 0.1644 0.0014 0.1443 43.36 1984 0.5370 0.0815 0.0698 0.0033 0.1768 0.0016 0.1314 46.22 1985 0.5351 0.0870 0.0662 0.0033 0.1893 0.0017 0.1186 49.08 1986 0.5333 0.0926 0.0626 0.0032 0.2017 0.0019 0.1057 51.94 1987 2.19 0.5314 0.0981 0.0590 0.0032 0.2142 0.0020 0.0929 54.80 1988 2.08 0.5296 0.1037 0.0554 0.0032 0.2266 0.0022 0.0800 57.65 1989 2.01 0.5277 0.1092 0.0518 0.0032 0.2391 0.0023 0.0672 60.51 1990 2.01 0.5259 0.1148 0.0482 0.0031 0.2515 0.0025 0.0543 63.37 1991 1.98 0.5240 0.1203 0.0446 0.0031 0.2640 0.0026 0.0415 64.67 1992 1.94 0.5222 0.1259 0.0410 0.0031 0.2764 0.0028 0.0286 65.97 1993 2.07 0.5184 0.1291 0.0401 0.0033 0.2756 0.0029 0.0319 67.26 1994 1.92 0.5146 0.1323 0.0392 0.0035 0.2748 0.0030 0.0353 68.56 1995 1.90 0.5109 0.1355 0.0383 0.0038 0.2741 0.0030 0.0386 69.86 1996 1.83 0.5071 0.1387 0.0374 0.0040 0.2733 0.0031 0.0420 71.16 1997 1.82 0.5033 0.1418 0.0365 0.0042 0.2725 0.0032 0.0453 72.46 1998 1.84 0.4995 0.1450 0.0356 0.0044 0.2717 0.0033 0.0487 73.75 1999 1.91 0.4958 0.1482 0.0347 0.0047 0.2710 0.0033 0.0520 75.05 2000 1.87 0.4920 0.1514 0.0338 0.0049 0.2702 0.0034 0.0554 76.35 2001 1.87 0.4882 0.1546 0.0329 0.0051 0.2694 0.0035 0.0587 97.11 2002 1.88 0.4882 0.1546 0.0329 0.0051 0.2669 0.0035 0.0587 117.87 2003 1.79 0.4882 0.1546 0.0329 0.0051 0.2644 0.0035 0.0587 138.63 2004 1.80 0.4882 0.1546 0.0329 0.0051 0.2620 0.0035 0.0587 159.39 2005 1.86 0.4882 0.1546 0.0329 0.0051 0.2595 0.0035 0.0587 180.16 2006 1.87 0.4882 0.1546 0.0329 0.0051 0.2570 0.0035 0.0587 200.92 2007 1.92 0.4776 0.1639 0.0328 0.0064 0.2548 0.0036 0.0608 221.68 2008 2.05 0.4670 0.1733 0.0327 0.0077 0.2527 0.0038 0.0628 242.44 2009 1.96 0.4564 0.1826 0.0326 0.0091 0.2505 0.0039 0.0649 263.20 2010 2.04 0.4458 0.1920 0.0325 0.0104 0.2484 0.0041 0.0669 283.96

PAGE 153

153 Figure C 13 . Catchment coefficient for Bullfrog Creek basin (USGS 023 007 00 ) , 19 8 0 2010 y = 0.0056x + 13.206 R² = 0.1546 1.00 2.00 3.00 4.00 5.00 6.00 7.00 1980 1985 1990 1995 2000 2005 2010 Catchment coefficient, Year (t) Linear (

PAGE 154

154 USG S 02303350, T rout Creek near Sulphur Springs Description : Latitude 28°08'05", Longitude 82°21'43" NAD83 Hillsborough County, Florida, Hydrologic Unit 03100205 Drainage area: 23.0 square miles Datum of gage: 0.75 feet above NAVD88 Table C 14. Time series of land cover types and population density for Trout Creek basin (USGS 02 30335 0 ) , YEAR _10 AG URB FOR WAT WET BAR GRA PD 1980 0.1619 0.0894 0.0312 0.0037 0.1900 0.0081 0.5123 37.33 1981 0.1678 0.0965 0.0329 0.0036 0.2127 0.0075 0.4760 40.79 1982 0.1738 0.1037 0.0345 0.0034 0.2354 0.0068 0.4397 44.24 1983 0.1797 0.1109 0.0361 0.0032 0.2581 0.0061 0.4034 47.70 1984 2.58 0.1856 0.1181 0.0377 0.0030 0.2808 0.0054 0.3671 51.15 1985 2.64 0.1916 0.1253 0.0393 0.0028 0.3035 0.0047 0.3308 54.61 1986 2.51 0.1975 0.1324 0.0410 0.0027 0.3262 0.0041 0.2945 58.06 1987 2.27 0.2034 0.1396 0.0426 0.0025 0.3489 0.0034 0.2582 61.52 1988 2.26 0.2094 0.1468 0.0442 0.0023 0.3715 0.0027 0.2220 64.97 1989 2.38 0.2153 0.1540 0.0458 0.0021 0.3942 0.0020 0.1857 68.43 1990 2.39 0.2212 0.1611 0.0475 0.0020 0.4169 0.0014 0.1494 71.88 1991 2.36 0.2272 0.1683 0.0491 0.0018 0.4396 0.0007 0.1131 86.70 1992 2.60 0.2331 0.1755 0.0507 0.0016 0.4623 0.0000 0.0768 101.51 1993 2.79 0.2215 0.1964 0.0498 0.0018 0.4598 0.0005 0.0783 116.33 1994 2.87 0.2098 0.2172 0.0488 0.0021 0.4574 0.0011 0.0798 131.14 1995 2.75 0.1982 0.2381 0.0479 0.0023 0.4549 0.0016 0.0813 145.96 1996 2.92 0.1865 0.2590 0.0470 0.0025 0.4524 0.0022 0.0828 160.77 1997 2.84 0.1749 0.2798 0.0460 0.0028 0.4500 0.0027 0.0843 175.59 1998 2.46 0.1632 0.3007 0.0451 0.0030 0.4475 0.0033 0.0858 190.40 1999 2.55 0.1516 0.3216 0.0442 0.0032 0.4450 0.0038 0.0873 205.22 2000 2.47 0.1399 0.3424 0.0432 0.0035 0.4426 0.0044 0.0888 220.03 2001 2.55 0.1283 0.3633 0.0423 0.0037 0.4401 0.0049 0.0903 248.79 2002 2.28 0.1283 0.3633 0.0423 0.0037 0.4255 0.0049 0.0903 277.55 2003 2.09 0.1283 0.3633 0.0423 0.0037 0.4110 0.0049 0.0903 306.30 2004 1.93 0.1283 0.3633 0.0423 0.0037 0.3964 0.0049 0.0903 335.06 2005 1.99 0.1283 0.3633 0.0423 0.0037 0.3819 0.0049 0.0903 363.82 2006 1.94 0.1283 0.3633 0.0423 0.0037 0.3673 0.0049 0.0903 392.58 2007 2.03 0.1245 0.3763 0.0421 0.0043 0.3604 0.0048 0.0877 421.34 2008 2.31 0.1207 0.3893 0.0419 0.0049 0.3536 0.0048 0.0851 450.09 2009 2.38 0.1168 0.4023 0.0416 0.0054 0.3467 0.0047 0.0824 478.85 2010 2.37 0.1130 0.4153 0.0414 0.0060 0.3399 0.0047 0.0798 507.61

PAGE 155

155 Figure C 14 . Catchment coefficient for Trout Creek basin (USGS 02 30335 0 ) , 19 8 0 2010 y = 0.0157x + 33.744 R² = 0.1985 1.00 2.00 3.00 4.00 5.00 6.00 7.00 1980 1985 1990 1995 2000 2005 2010 Catchment coefficient, Year (t) Linear (

PAGE 156

156 USG S 02307359, B rooker Creek near Tarpon Springs Description : Latitude 28°05'45", Longitude 82°41'15" NAD27 Pinellas County, Florida, Hydrologic Unit 03100206 Drainage area: 30.0 square miles Datum of gage: 0.00 feet above NGVD29 Table C 15 . Time series of land cover types and population density for Brooker Creek basin (USGS 02 307359) YEAR _10 AG URB FOR WAT WET BAR GRA PD 1980 5.22 0.4077 0.1480 0.0474 0.0800 0.2321 0.0002 0.0845 28.89 1981 5.10 0.3856 0.1530 0.0550 0.0790 0.2453 0.0003 0.0818 34.98 1982 5.05 0.3635 0.1579 0.0625 0.0781 0.2584 0.0003 0.0792 41.08 1983 5.04 0.3414 0.1629 0.0701 0.0772 0.2715 0.0004 0.0766 47.17 1984 5.31 0.3193 0.1679 0.0776 0.0762 0.2846 0.0004 0.0740 53.27 1985 5.34 0.2972 0.1728 0.0852 0.0753 0.2977 0.0004 0.0714 59.36 1986 5.40 0.2751 0.1778 0.0927 0.0743 0.3109 0.0005 0.0687 65.45 1987 5.31 0.2530 0.1827 0.1003 0.0734 0.3240 0.0005 0.0661 71.55 1988 5.22 0.2309 0.1877 0.1078 0.0725 0.3371 0.0005 0.0635 77.64 1989 5.14 0.2088 0.1926 0.1154 0.0715 0.3502 0.0006 0.0609 83.74 1990 4.97 0.1867 0.1976 0.1229 0.0706 0.3634 0.0006 0.0582 89.83 1991 5.12 0.1646 0.2025 0.1305 0.0696 0.3765 0.0007 0.0556 96.19 1992 5.31 0.1425 0.2075 0.1380 0.0687 0.3896 0.0007 0.0530 102.54 1993 5.02 0.1429 0.2093 0.1354 0.0687 0.3863 0.0007 0.0567 108.90 1994 5.24 0.1433 0.2111 0.1328 0.0687 0.3829 0.0007 0.0604 115.25 1995 5.44 0.1438 0.2130 0.1301 0.0688 0.3796 0.0007 0.0641 121.61 1996 5.57 0.1442 0.2148 0.1275 0.0688 0.3763 0.0007 0.0678 127.97 1997 6.11 0.1446 0.2166 0.1249 0.0688 0.3729 0.0007 0.0714 134.32 1998 5.55 0.1450 0.2184 0.1223 0.0688 0.3696 0.0007 0.0751 140.68 1999 5.58 0.1455 0.2203 0.1196 0.0689 0.3663 0.0007 0.0788 147.03 2000 5.49 0.1459 0.2221 0.1170 0.0689 0.3629 0.0007 0.0825 153.39 2001 5.56 0.1463 0.2239 0.1144 0.0689 0.3596 0.0007 0.0862 156.79 2002 5.54 0.1455 0.2263 0.1142 0.0690 0.3584 0.0011 0.0855 160.18 2003 4.80 0.1446 0.2288 0.1140 0.0691 0.3572 0.0015 0.0849 163.58 2004 4.64 0.1438 0.2312 0.1137 0.0692 0.3559 0.0020 0.0842 166.97 2005 4.31 0.1429 0.2337 0.1135 0.0693 0.3547 0.0024 0.0836 170.37 2006 4.21 0.1421 0.2361 0.1133 0.0694 0.3535 0.0028 0.0829 173.77 2007 4.00 0.1419 0.2365 0.1132 0.0694 0.3533 0.0028 0.0829 177.16 2008 4.56 0.1417 0.2370 0.1132 0.0694 0.3531 0.0028 0.0829 180.56 2009 4.83 0.1416 0.2374 0.1131 0.0693 0.3530 0.0028 0.0828 183.95 2010 4.53 0.1414 0.2379 0.1131 0.0693 0.3528 0.0028 0.0828 187.35

PAGE 157

157 Figure C 15 . Catchment coefficient for Brooker Creek basin (USGS 02 307359 ) , 195 1 2010 y = 0.0328x 60.49 R² = 0.4373 1.00 2.00 3.00 4.00 5.00 6.00 7.00 1950 1960 1970 1980 1990 2000 2010 Catchment coefficient, Year (t) Linear (

PAGE 158

158 USG S 02309848, S outh Branch Anclote River near Odessa Description : Latitude 28°11'07.6", Longitude 82°33'12.1" NAD83 Pasco County, Florida, Hydrologic Unit 03100207 Drainage area: 17.1 square miles Datum of gage: 46.22 feet above NGVD29 Table C 16 . Time series of land cover types and population density for South Branch Anclote River basin (USGS 0230 9848) YEAR _10 AG URB FOR WAT WET BAR GRA PD 1980 9.38 0.4031 0.1913 0.0315 0.0446 0.1672 0.0001 0.1613 72.87 1981 9.44 0.3791 0.2053 0.0367 0.0441 0.1839 0.0001 0.1498 78.29 1982 10.32 0.3552 0.2194 0.0420 0.0436 0.2005 0.0001 0.1384 83.72 1983 11.85 0.3312 0.2335 0.0472 0.0432 0.2172 0.0002 0.1269 89.14 1984 12.57 0.3072 0.2475 0.0524 0.0427 0.2339 0.0002 0.1154 94.56 1985 13.36 0.2833 0.2616 0.0577 0.0422 0.2505 0.0002 0.1040 99.99 1986 14.69 0.2593 0.2756 0.0629 0.0417 0.2672 0.0002 0.0925 105.41 1987 15.41 0.2353 0.2897 0.0682 0.0413 0.2839 0.0002 0.0811 110.83 1988 18.19 0.2114 0.3038 0.0734 0.0408 0.3005 0.0002 0.0696 116.25 1989 14.31 0.1874 0.3178 0.0787 0.0403 0.3172 0.0003 0.0582 121.68 1990 12.88 0.1634 0.3319 0.0839 0.0398 0.3339 0.0003 0.0467 127.10 1991 13.30 0.1395 0.3459 0.0892 0.0394 0.3505 0.0003 0.0353 134.16 1992 12.43 0.1155 0.3600 0.0944 0.0389 0.3672 0.0003 0.0238 141.22 1993 10.21 0.1136 0.3620 0.0918 0.0400 0.3618 0.0017 0.0302 148.27 1994 10.67 0.1117 0.3640 0.0892 0.0412 0.3563 0.0031 0.0365 155.33 1995 11.20 0.1098 0.3660 0.0866 0.0423 0.3509 0.0046 0.0429 162.39 1996 11.78 0.1079 0.3680 0.0840 0.0435 0.3454 0.0060 0.0492 169.45 1997 13.36 0.1061 0.3700 0.0815 0.0446 0.3400 0.0074 0.0556 176.51 1998 9.08 0.1042 0.3720 0.0789 0.0458 0.3345 0.0088 0.0619 183.56 1999 9.07 0.1023 0.3740 0.0763 0.0469 0.3291 0.0103 0.0683 190.62 2000 8.76 0.1004 0.3760 0.0737 0.0481 0.3236 0.0117 0.0746 197.68 2001 8.97 0.0985 0.3780 0.0711 0.0492 0.3182 0.0131 0.0810 214.86 2002 8.69 0.0985 0.3780 0.0711 0.0492 0.3164 0.0131 0.0810 232.04 2003 7.73 0.0985 0.3780 0.0711 0.0492 0.3146 0.0131 0.0810 249.21 2004 7.94 0.0985 0.3780 0.0711 0.0492 0.3127 0.0131 0.0810 266.39 2005 7.57 0.0985 0.3780 0.0711 0.0492 0.3109 0.0131 0.0810 283.57 2006 7.08 0.0985 0.3780 0.0711 0.0492 0.3091 0.0131 0.0810 300.75 2007 6.27 0.0937 0.3860 0.0708 0.0496 0.3067 0.0123 0.0809 317.93 2008 7.53 0.0889 0.3940 0.0704 0.0500 0.3043 0.0115 0.0808 335.10 2009 8.29 0.0842 0.4021 0.0701 0.0504 0.3019 0.0108 0.0806 352.28 2010 8.37 0.0794 0.4101 0.0697 0.0508 0.2995 0.0100 0.0805 369.46

PAGE 159

159 Figure C 16. Catchment coefficient for South Branch Anclote River basin (USGS 0230 9848 ) , 19 71 2010 y = 0.2071x + 423.87 R² = 0.4395 6.00 8.00 10.00 12.00 14.00 16.00 18.00 1980 1985 1990 1995 2000 2005 2010 Catchment coefficient, Year (t) Linear (

PAGE 160

160 USG S 02310000, A nclote River near Elfers Description : Latitude 28°12'50", Longitude 82°40'00" NAD27 Pasco County, Florida, Hydrologic Unit 03100207 Drainage area: 72.5 square miles Datum of gage: 0.00 feet above NGVD29 Table C 17 . Time series of land cover types and population density for A nclote River basin (USGS 02310000) YEAR _10 AG URB FOR WAT WET BAR GRA PD 1980 3.73 0.3691 0.0739 0.1351 0.0168 0.2027 0.0008 0.2016 22.82 1981 3.96 0.3556 0.0789 0.1393 0.0168 0.2154 0.0009 0.1931 24.94 1982 3.72 0.3421 0.0839 0.1435 0.0167 0.2282 0.0011 0.1846 27.06 1983 3.80 0.3286 0.0889 0.1477 0.0167 0.2409 0.0012 0.1761 29.19 1984 3.96 0.3150 0.0939 0.1518 0.0167 0.2536 0.0013 0.1676 31.31 1985 4.00 0.3015 0.0989 0.1560 0.0166 0.2664 0.0015 0.1591 33.43 1986 3.94 0.2880 0.1039 0.1602 0.0166 0.2791 0.0016 0.1506 35.55 1987 3.70 0.2745 0.1089 0.1644 0.0165 0.2919 0.0017 0.1421 37.67 1988 3.62 0.2610 0.1139 0.1686 0.0165 0.3046 0.0019 0.1336 39.80 1989 3.83 0.2475 0.1189 0.1728 0.0164 0.3174 0.0020 0.1251 41.92 1990 3.67 0.2339 0.1239 0.1769 0.0164 0.3301 0.0021 0.1166 44.04 1991 3.58 0.2204 0.1289 0.1811 0.0163 0.3429 0.0023 0.1081 46.15 1992 3.84 0.2069 0.1339 0.1853 0.0163 0.3556 0.0024 0.0996 48.25 1993 3.65 0.2080 0.1338 0.1771 0.0166 0.3557 0.0026 0.1062 50.36 1994 3.87 0.2090 0.1337 0.1689 0.0169 0.3559 0.0027 0.1128 52.46 1995 3.90 0.2101 0.1335 0.1608 0.0173 0.3560 0.0029 0.1194 54.57 1996 4.18 0.2112 0.1334 0.1526 0.0176 0.3562 0.0031 0.1260 56.67 1997 4.47 0.2122 0.1333 0.1444 0.0179 0.3563 0.0032 0.1326 58.78 1998 3.84 0.2133 0.1332 0.1362 0.0182 0.3565 0.0034 0.1392 60.88 1999 3.86 0.2144 0.1330 0.1281 0.0186 0.3566 0.0036 0.1458 62.99 2000 3.90 0.2154 0.1329 0.1199 0.0189 0.3568 0.0037 0.1524 65.09 2001 4.13 0.2165 0.1328 0.1117 0.0192 0.3569 0.0039 0.1590 77.33 2002 4.00 0.2110 0.1404 0.1110 0.0193 0.3543 0.0057 0.1583 89.58 2003 3.33 0.2055 0.1480 0.1103 0.0195 0.3517 0.0075 0.1576 101.82 2004 2.96 0.1999 0.1555 0.1096 0.0196 0.3491 0.0094 0.1569 114.07 2005 3.00 0.1944 0.1631 0.1089 0.0198 0.3465 0.0112 0.1562 126.31 2006 3.01 0.1889 0.1707 0.1082 0.0199 0.3439 0.0130 0.1555 138.55 2007 2.99 0.1875 0.1737 0.1075 0.0200 0.3433 0.0119 0.1563 150.80 2008 3.44 0.1861 0.1766 0.1068 0.0201 0.3427 0.0107 0.1570 163.04 2009 3.51 0.1848 0.1796 0.1060 0.0202 0.3420 0.0096 0.1578 175.29 2010 3.24 0.1834 0.1825 0.1053 0.0203 0.3414 0.0084 0.1585 187.53

PAGE 161

161 Figure C 17 . Catchment coefficient for A nclote River basin (USGS 02310000 ) , 1950 2010 y = 0.023x 42.387 R² = 0.3958 1.00 2.00 3.00 4.00 5.00 6.00 7.00 1950 1960 1970 1980 1990 2000 2010 Catchment coefficient, Year (t) Linear (

PAGE 162

162 USG S 02324000, S teinhatchee River near Cross City Description : Latitude 29°47'11", Longitude 83°19'18" NAD27 Taylor County, Florida, Hydrologic Unit 03110102 Drainage area: 3 50.00 square miles Datum of gage: 7.84 feet above NGVD29 Table C 18 . Time series of land cover types and population density for Steinhatchee River basin (USGS 023 24 000 ) YEAR _10 AG URB FOR WAT WET BAR GRA PD 1980 3.84 0.0068 0.0070 0.4663 0.0002 0.4993 0.0000 0.0203 1.63 1981 4.15 0.0066 0.0082 0.4611 0.0002 0.5002 0.0000 0.0237 1.67 1982 4.33 0.0065 0.0093 0.4559 0.0002 0.5010 0.0000 0.0270 1.70 1983 4.67 0.0064 0.0105 0.4507 0.0002 0.5019 0.0001 0.0304 1.74 1984 4.25 0.0063 0.0117 0.4454 0.0001 0.5027 0.0001 0.0337 1.78 1985 4.20 0.0062 0.0128 0.4402 0.0001 0.5036 0.0001 0.0370 1.82 1986 4.14 0.0060 0.0140 0.4350 0.0001 0.5044 0.0001 0.0404 1.85 1987 4.10 0.0059 0.0152 0.4297 0.0001 0.5053 0.0001 0.0437 1.89 1988 4.56 0.0058 0.0163 0.4245 0.0001 0.5061 0.0001 0.0470 1.93 1989 4.31 0.0057 0.0175 0.4193 0.0001 0.5070 0.0001 0.0504 1.96 1990 4.09 0.0055 0.0187 0.4141 0.0000 0.5078 0.0001 0.0537 2.00 1991 4.13 0.0054 0.0198 0.4088 0.0000 0.5087 0.0001 0.0571 2.04 1992 4.63 0.0053 0.0210 0.4036 0.0000 0.5095 0.0001 0.0604 2.07 1993 4.59 0.0048 0.0228 0.3889 0.0000 0.5153 0.0001 0.0679 2.11 1994 5.18 0.0043 0.0247 0.3742 0.0000 0.5212 0.0001 0.0754 2.14 1995 5.46 0.0039 0.0265 0.3595 0.0000 0.5270 0.0001 0.0829 2.18 1996 5.90 0.0034 0.0283 0.3448 0.0000 0.5328 0.0001 0.0904 2.21 1997 7.19 0.0029 0.0302 0.3302 0.0000 0.5387 0.0001 0.0979 2.25 1998 5.99 0.0024 0.0320 0.3155 0.0000 0.5445 0.0001 0.1054 2.28 1999 5.96 0.0020 0.0338 0.3008 0.0000 0.5503 0.0001 0.1129 2.32 2000 6.31 0.0015 0.0357 0.2861 0.0000 0.5562 0.0001 0.1204 2.35 2001 5.30 0.0010 0.0375 0.2714 0.0000 0.5620 0.0001 0.1279 2.33 2002 5.43 0.0010 0.0375 0.2675 0.0000 0.5616 0.0001 0.1323 2.31 2003 5.12 0.0010 0.0375 0.2635 0.0000 0.5612 0.0001 0.1367 2.29 2004 5.19 0.0010 0.0375 0.2596 0.0000 0.5607 0.0001 0.1410 2.27 2005 4.93 0.0010 0.0375 0.2556 0.0000 0.5603 0.0001 0.1454 2.25 2006 4.67 0.0010 0.0375 0.2517 0.0000 0.5599 0.0001 0.1498 2.22 2007 4.54 0.0009 0.0375 0.2499 0.0000 0.5596 0.0005 0.1516 2.20 2008 5.48 0.0009 0.0375 0.2481 0.0000 0.5592 0.0008 0.1535 2.18 2009 5.82 0.0008 0.0375 0.2464 0.0000 0.5589 0.0012 0.1553 2.16 2010 5.32 0.0008 0.0375 0.2446 0.0000 0.5585 0.0015 0.1572 2.14

PAGE 163

163 Figure C 18 . Catchment coefficient for Steinhatchee River basin (USGS 023 24 000 ) , 1950 2010 y = 0.0397x 74.261 R² = 0.5091 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 1950 1960 1970 1980 1990 2000 2010 Catchment coefficient, Year (t) Linear (

PAGE 164

164 USG S 02327100, S opchoppy River near Sopchoppy Description : Latitude 30°07'45", Longitude 84°29'40" NAD27 Wakulla County, Florida, Hydrologic Unit 03120003 Drainage area: 102.00 square miles Datum o f gage: 0.00 feet above NGVD29 Table C 19 . Time series of land cover types and population density for Sopchoppy River basin (USGS 023 27 100 ) YEAR _10 AG URB FOR WAT WET BAR GRA PD 1980 2.49 0.0013 0.0062 0.5347 0.0000 0.4538 0.0000 0.0040 1.81 1981 2.45 0.0015 0.0072 0.5170 0.0000 0.4696 0.0000 0.0047 1.85 1982 2.40 0.0017 0.0083 0.4993 0.0000 0.4854 0.0000 0.0053 1.89 1983 2.46 0.0019 0.0093 0.4816 0.0000 0.5012 0.0000 0.0060 1.93 1984 2.30 0.0021 0.0103 0.4638 0.0000 0.5170 0.0000 0.0067 1.97 1985 2.30 0.0023 0.0114 0.4461 0.0000 0.5328 0.0000 0.0073 2.01 1986 2.39 0.0025 0.0124 0.4284 0.0000 0.5486 0.0000 0.0080 2.05 1987 2.30 0.0027 0.0134 0.4107 0.0000 0.5644 0.0000 0.0087 2.09 1988 2.26 0.0030 0.0145 0.3930 0.0000 0.5802 0.0000 0.0093 2.13 1989 2.32 0.0032 0.0155 0.3753 0.0000 0.5960 0.0000 0.0100 2.17 1990 2.31 0.0034 0.0165 0.3575 0.0000 0.6118 0.0000 0.0107 2.21 1991 2.18 0.0036 0.0176 0.3398 0.0000 0.6276 0.0000 0.0113 2.35 1992 2.27 0.0038 0.0186 0.3221 0.0000 0.6434 0.0000 0.0120 2.49 1993 2.29 0.0034 0.0185 0.3227 0.0000 0.6432 0.0000 0.0121 2.63 1994 2.37 0.0030 0.0185 0.3233 0.0000 0.6430 0.0000 0.0122 2.77 1995 2.35 0.0025 0.0184 0.3238 0.0000 0.6428 0.0000 0.0123 2.91 1996 2.38 0.0021 0.0184 0.3244 0.0000 0.6426 0.0000 0.0124 3.04 1997 2.51 0.0017 0.0183 0.3250 0.0000 0.6423 0.0000 0.0126 3.18 1998 2.38 0.0013 0.0183 0.3256 0.0000 0.6421 0.0000 0.0127 3.32 1999 2.36 0.0008 0.0182 0.3261 0.0000 0.6419 0.0000 0.0128 3.46 2000 2.31 0.0004 0.0182 0.3267 0.0000 0.6417 0.0000 0.0129 3.60 2001 2.36 0.0000 0.0181 0.3273 0.0000 0.6415 0.0000 0.0130 3.65 2002 2.21 0.0000 0.0181 0.3272 0.0000 0.6415 0.0000 0.0130 3.71 2003 2.11 0.0000 0.0181 0.3272 0.0000 0.6415 0.0000 0.0131 3.76 2004 2.14 0.0000 0.0181 0.3271 0.0000 0.6416 0.0000 0.0131 3.81 2005 2.19 0.0000 0.0181 0.3271 0.0000 0.6416 0.0000 0.0132 3.87 2006 2.15 0.0000 0.0181 0.3270 0.0000 0.6416 0.0000 0.0132 3.92 2007 2.12 0.0000 0.0181 0.3269 0.0000 0.6416 0.0000 0.0133 3.97 2008 2.38 0.0000 0.0181 0.3268 0.0000 0.6416 0.0000 0.0134 4.02 2009 2.35 0.0000 0.0181 0.3267 0.0000 0.6416 0.0000 0.0136 4.08 2010 2.37 0.0000 0.0181 0.3266 0.0000 0.6416 0.0000 0.0137 4.13

PAGE 165

165 Figure C 19 . Catchment coefficient for Sopchoppy River basin (USGS 023 27 100 ) , 19 7 0 2010 y = 0.0058x + 13.895 R² = 0.2411 1.00 2.00 3.00 4.00 5.00 6.00 7.00 1970 1975 1980 1985 1990 1995 2000 2005 2010 Catchment coefficient Year Linear (

PAGE 166

166 Table C 20. List of potential case studies with flow records and remarks Gage ID Area (Sq Km) Flow record (years, 1950 2009) Flow record (years, 1990 2009) WR report remarks 022 31000 1748.4 60 20 022 31342 111.6 32 20 022 32200 57.1 53 20 Since October 1970 flow regulated to some extent following the construction of Jane Green Reservoir; levees were constructed and an interconnecting canal was dug joining the watershed areas of Taylor, Pennywash, Cox, and Wolf Creeks. 022 35000 449.0 60 20 022 35200 297.7 26 20 022 43000 3071.7 53 20 022 45500 348.4 60 20 022 34324 85.1 35 20 Some regulation by retention ponds in urban areas upstream from station. 022 34384 46.3 28 20 Since about 1980, some regulation by retention ponds in headwaters. 022 34400 42.7 31 20 022 34990 110.3 34 20 Flow includes occasional pumpage from Cranes Roost basin. 022 91500 513.8 27 20 022 96500 886.4 59 20 022 97100 350.4 59 20 022 98123 540.7 37 20 022 36500 146.9 51 20 Some interconnection at high stages with Little Creek and Withlacoochee River basin. 022 47510 145.0 45 20 022 48000 93.5 58 20 Some diversions for irrigation above station. 022 62900 216.9 50 20 Some diversion to ground water through drainage wells in lakes upstream from station. 022 63800 274.8 51 20 022 63869 8.5 33 20 Flow regulated by automatic gates in control structure 15. 022 64100 113.0 43 20 Since October 1968, flow regulated by automatic gates upstream and since December 1970, by control structure S 11. Natural flow of stream affected by canals and control structures above station which divert an undetermined amount of water into the Reedy C 022 66205 27.0 23 20 Flow regulated by operation of structure 411. At high stages interconnection exists between Reedy Creek, Whittenhorse Creek, and Boggy Creek. 0 2324000 791.0 59 20 022 66300 257.8 43 20 Natural flow of stream affected by several canals, levees, and control structures. 022 66480 43.6 40 20

PAGE 167

167 Table C 20. Continued. Gage_ID Area (Sq Km) Flow record (years, 1950 2009) Flow record (years, 1990 2009) WR report remarks 0 2312000 1372.4 60 20 High water diversion above station into Hillsborough River basin through Withlacochee Hillsborou gh Overflow near Richland 022 67000 168.6 60 20 022 91580 75.1 22 20 No flow many days during the water year. 0 2300700 73.8 34 20 0 2301750 35.8 25 20 0 2311500 923.8 26 20 High water diversion above station into Hillsborough River basin 0 2301900 22.4 45 20 Some diversion at times by pumpage for irrigation. 0 2301990 238.9 26 20 022 70500 982.2 60 20 Records include small diversions into Lake Arbuckle from Lake Weohyakapka 0 2303350 42.0 35 20 022 31600 641.3 56 20 Since April 1990, flow regulated to some extent by flood control lift gates 0 2306500 13.6 58 20 Flow affected by regulation of control structures upstream from station. 0 2306647 41.8 24 20 0 2307359 80.0 59 20 0 2309848 43.6 39 20 0 2310000 177.6 60 20 0 2324400 176.3 54 20 0 2310947 649.6 42 20 Some interconnection with Gator Creek and some diversions to the north may exist during periods of extreme high water. 0 2327100 271.1 45 20

PAGE 168

168 Figure C 20 . Geographical distribution of pre selected case study catchments.

PAGE 169

169 APPENDIX D CASE STUDY ANALYSIS SUPPORT INFORMATION Table D 1 . Climate type and meteorological ranges Climate type Mean annual Evapotranspiration (Eo, mm) Mean annual precipitation (P, mm) Humid 1200 = Eo Sub humid Semi arid Arid Source: Arora ( 2002 ) ; Ponce et al. ( 2000) Table D 2 . Non linear regressions of Q/P=f(aridity index, catchment coefficient) Regression equation R square 0.375 Q/P = 0.8828 0.242 0.8090 0.75 Q/P = 0.8818 0.718 0.9686 2.00 Q/P = 1.4143 2.831 0.9794 5.00 Q/P = 2.4298 5.269 0.9640 12.00 Q/P = 3.9821 7.597 0.9587

PAGE 170

170 Table D 3 . Land Cover Classification Class \ Value Anderson Level 1 Class Codes and Descriptions (NLCD 1992) Class \ Value Anderson Level 2 Class Codes and Descriptions (NLCD 2001/2006/2011) 0 No Data Water 1 Open Water 11 Open Water 12 Perennial Ice/Snow Developed 2 Urban 21 Developed, Open Space 22 Developed, Low Intensity 23 Developed, Medium Intensity 24 Developed High Intensity Barren 3 Barren 31 Barren Land (Rock/Sand/Clay) Forest 4 Forest 41 Deciduous Forest 42 Evergreen Forest 43 Mixed Forest Shrubland/Grassland 5 Grassland/Shrub 51 Dwarf Scrub 52 Shrub/Scrub Herbaceous 71 Grassland/Herbaceous 72 Sedge/Herbaceous 73 Lichens 74 Moss Planted/Cultivated 6 Agriculture 81 Pasture/Hay 82 Cultivated Crops Wetlands 7 Wetlands 90 Woody Wetlands 95 Emergent Herbaceous Wetlands Other 8 Ice/Snow See 12 See Perennial Ice/Snow

PAGE 171

171 Table D 4 . List of pre selected case studies Gage_ID Flow record (years, 1949 2009) Area (Sq Km) _10 (1992 2010) UBG (1992 2010) _10 (1980 2010) UBG (1980 2010) 02231000 60 1748.4 .902 ** .940 ** .839 ** .983 ** 02231342 32 111.6 .534 * .893 ** 02234400 31 42.7 .912 ** .979 ** 02235000 60 449.0 .952 ** .952 ** 02243000 54 3071.7 ? ? ? ? 02245500 60 348.4 .984 ** .470 ** .992 ** 02247510 40 145.0 .832 ** .957 ** .987 ** 02263800 51 274.8 .765 ** .953 ** .596 ** .984 ** 02266480 45 43.6 .825 ** .988 ** .850 ** .994 ** 02267000 60 168.6 .675 ** .988 ** .876 ** .913 ** 02296500 59 886.4 .929 ** .881 ** .404 * .495 ** 02297100 59 350.4 .863 ** .920 ** .856 ** .751 ** 02298123 34 540.7 .856 ** .879 ** .876 ** .852 ** 02300700 35 73.8 .970 ** 02303350 37 42.0 .748 ** .940 ** .449 * 02307359 59 80.0 .686 ** .960** .430 * .987 ** 02309848 39 43.6 .806** .956** .663 ** .979 ** 02310000 60 177.6 .663 ** .995** .503 ** .796 ** 02324000 59 791.0 .974 ** .594 ** .991 ** 02324400 53 176.3 ? ? ? ? 02327100 45 271.1 .987 ** .439 * .856 **

PAGE 172

172 Table D 5 . Morphometric variables for the 19 basins sample Gage_ID Area Drain Den Compac t Perm A ve RR mean S lope Circ Ratio Bas Relief 02231000 1748.4 0.2161 1.56 7.45 0.552 0.64 0.0312 55.00 02231342 111.6 1.4154 2.21 11.68 0.666 0.01 0.0281 12.00 02234400 42.7 0.3420 1.51 12.58 0.515 0.36 0.0193 25.00 02235000 449.0 0.1846 1.74 12.36 0.392 0.94 0.0335 54.00 02245500 348.4 0.5285 1.75 10.36 0.423 0.85 0.0352 68.00 02247510 145.0 0.2305 1.47 10.83 0.577 0.03 0.0187 14.00 02263800 274.8 0.3620 1.68 11.92 0.282 0.40 0.0332 38.00 02266480 43.6 0.0786 1.52 13.04 0.259 0.87 0.0194 44.00 02267000 168.6 0.0024 1.23 13.31 0.189 0.85 0.0157 68.00 02296500 886.4 0.7397 1.35 10.85 0.401 0.12 0.0174 48.00 02297100 350.4 1.1477 1.43 10.66 0.648 0.04 0.0181 24.00 02298123 540.7 1.1922 1.48 10.75 0.584 0.00 0.0189 19.00 02300700 73.8 0.5518 2.53 11.27 0.458 0.28 0.0503 37.00 02303350 42.0 0.1506 1.35 9.45 0.384 0.07 0.0171 22.00 02307359 80.0 0.3126 2.86 12.20 0.510 0.11 0.0600 17.00 02309848 43.6 0.0599 3.89 11.70 0.529 0.03 0.0757 8.00 02310000 177.6 0.4269 1.85 11.08 0.572 0.04 0.0365 22.00 02324000 791.0 0.2143 2.06 8.73 0.397 0.74 0.0411 44.00 02327100 271.1 0.2062 1.32 9.47 0.494 0.75 0.0267 43.00 Table D 6 . Summary of d escriptive statistics for two case study samples _10 CS_19 CS_18 Mean 3.88 3.47 Standard Error 0.09 0.04 Median 3.56 3.51 Mode #N/A #N/A Standard Deviation 2.03 0.99 Sample Variance 4.14 0.99 Kurtosis 13.31 (0.04) Skewness 3.18 0.38 Range 16.42 5.42 Minimum 1.77 1.77 Maximum 18.19 7.19 Sum 2,130.41 1,799.69 Count 549.00 518.00 Confidence Level (95.0%) 0.17 0.09

PAGE 173

173 Table D 7 . Trend analysis of 10 year average discharge (Q_10) and 10 year average precipitation (P_10) and land cover types (LC it ) and population density (PD t ), 1980 2010 Case study Parametric test YEAR _10 AG UBG FWW PD CS_02231000 Pearson Correlation 1 .839 ** .590 ** .983 ** .981 ** .997 ** Sig. (2 tailed) .000 .000 .000 .000 .000 N 31 31 31 31 31 31 CS_02231342 Pearson Correlation 1 .356 .773 ** .008 .605 ** .986 ** Sig. (2 tailed) .088 .000 .967 .000 .000 N 31 24 31 31 31 31 CS_02234400 Pearson Correlation 1 .143 .974 ** .979 ** .987 ** .884 ** Sig. (2 tailed) .596 .000 .000 .000 .000 N 31 16 31 31 31 31 CS_02235000 Pearson Correlation 1 .361 .932 ** .952 ** .983 ** .977 ** Sig. (2 tailed) .129 .000 .000 .000 .000 N 31 19 19 19 19 19 CS_02245500 Pearson Correlation 1 .470 ** .977 ** .992 ** .992 ** .893 ** Sig. (2 tailed) .008 .000 .000 .000 .000 N 31 31 31 31 31 31 CS_02247510 Pearson Correlation 1 .231 .935 ** .987 ** .987 ** .967 ** Sig. (2 tailed) .211 .000 .000 .000 .000 N 31 31 31 31 31 31 CS_02263800 Pearson Correlation 1 .596 ** .978 ** .984 ** .696 ** .998 ** Sig. (2 tailed) .000 .000 .000 .000 .000 N 31 31 31 31 31 31 CS_02266480 Pearson Correlation 1 .850 ** .993 ** .994 ** .404 * .921 ** Sig. (2 tailed) .000 .000 .000 .024 .000 N 31 31 31 31 31 31 CS_02267000 Pearson Correlation 1 .876 ** .850 ** .913 ** .943 ** .988 ** Sig. (2 tailed) .000 .000 .000 .000 .000 N 31 31 31 31 31 31 CS_02296500 Pearson Correlation 1 .404 * .864 ** .495 ** .127 .932 ** Sig. (2 tailed) .024 .000 .005 .494 .000 N 31 31 31 31 31 31 CS_02297100 Pearson Correlation 1 .856 ** .939 ** .751 ** .575 ** .984 ** Sig. (2 tailed) .000 .000 .000 .001 .000 N 31 31 31 31 31 31 CS_02298123 Pearson Correlation 1 .876 ** .926 ** .852 ** .613 ** .939 ** Sig. (2 tailed) .000 .000 .000 .000 .000 N 31 24 31 31 31 31 CS_02300700 Pearson Correlation 1 .401 .975 ** .295 .635 ** .870 ** Sig. (2 tailed) .052 .000 .107 .000 .000 N 31 24 31 31 31 31

PAGE 174

174 Table D 7 . Continued Case study Parametric test YEAR _10 AG UBG FWW PD CS_02303350 Pearson Correlation 1 .449 * .711 ** .042 .495 ** .953 ** Sig. (2 tailed) .019 .000 .822 .005 .000 N 31 27 31 31 31 31 CS_02307359 Pearson Correlation 1 .430 * .835 ** .987 ** .623 ** .994 ** Sig. (2 tailed) .016 .000 .000 .000 .000 N 31 31 31 31 31 31 CS_02309848 Pearson Correlation 1 .663 ** .876 ** .979 ** .582 ** .968 ** Sig. (2 tailed) .000 .000 .000 .001 .000 N 31 31 31 31 31 31 CS_02310000 Pearson Correlation 1 .503 ** .890 ** .796 ** .392 * .913 ** Sig. (2 tailed) .004 .000 .000 .029 .000 N 31 31 31 31 31 31 CS_02324000 Pearson Correlation 1 .594 ** .969 ** .991 ** .992 ** .861 ** Sig. (2 tailed) .000 .000 .000 .000 .000 N 31 31 31 31 31 31 CS_02327100 Pearson Correlation 1 .439 * .696 ** .856 ** .747 ** .985 ** Sig. (2 tailed) .013 .000 .000 .000 .000 N 31 31 31 31 31 31

PAGE 175

175 Table D 8 . Trend analysis of 10 year average discharge (Q_10) and 10 year average precipitation (P_10) and land cover types (LC it ) and population density (PD t ), 1992 2010 Case study Parametric test YEAR _10 AG UBG FWW PD CS_02231000 Pearson Correlation 1 .902 ** .775 ** .940 ** .938 ** .999 ** Sig. (2 tailed) .000 .000 .000 .000 .000 N 19 19 19 19 19 19 CS_02231342 Pearson Correlation 1 .534 * .894 ** .893 ** .893 ** .999 ** Sig. (2 tailed) .019 .000 .000 .000 .000 N 19 19 19 19 19 19 CS_02234400 Pearson Correlation 1 .143 .894 ** .912 ** .946 ** 1.000 ** Sig. (2 tailed) .596 .000 .000 .000 .000 N 19 16 19 19 19 19 CS_02235000 Pearson Correlation 1 .361 .932 ** .952 ** .983 ** .977 ** Sig. (2 tailed) .129 .000 .000 .000 .000 N 19 19 19 19 19 19 CS_02245500 Pearson Correlation 1 .310 .903 ** .984 ** .986 ** .402 Sig. (2 tailed) .196 .000 .000 .000 .088 N 19 19 19 19 19 19 CS_02247510 Pearson Correlation 1 .832 ** .695 ** .957 ** .954 ** .941 ** Sig. (2 tailed) .000 .001 .000 .000 .000 N 19 19 19 19 19 19 CS_02263800 Pearson Correlation 1 .765 ** .926 ** .953 ** .981 ** .996 ** Sig. (2 tailed) .000 .000 .000 .000 .000 N 19 19 19 19 19 19 CS_02266480 Pearson Correlation 1 .825 ** .995 ** .988 ** .904 ** .974 ** Sig. (2 tailed) .000 .000 .000 .000 .000 N 19 19 19 19 19 19 CS_02267000 Pearson Correlation 1 .675 ** .512 * .988 ** .864 ** .975 ** Sig. (2 tailed) .002 .025 .000 .000 .000 N 19 19 19 19 19 19 CS_02296500 Pearson Correlation 1 .929 ** .908 ** .881 ** .893 ** .659 ** Sig. (2 tailed) .000 .000 .000 .000 .002 N 19 19 19 19 19 19 CS_02297100 Pearson Correlation 1 .863 ** .850 ** .920 ** .893 ** .996 ** Sig. (2 tailed) .000 .000 .000 .000 .000 N 19 19 19 19 19 19 CS_02298123 Pearson Correlation 1 .856 ** .883 ** .879 ** .887 ** .699 ** Sig. (2 tailed) .000 .000 .000 .000 .001 N 19 19 19 19 19 19 CS_02300700 Pearson Correlation 1 .129 .949 ** .970 ** .995 ** .934 ** Sig. (2 tailed) .599 .000 .000 .000 .000 N 19 19 19 19 19 19

PAGE 176

176 Table D 8 . Continued. Case study Parametric test YEAR _10 AG UBG FWW PD CS_02303350 Pearson Correlation 1 .748 ** .930 ** .940 ** .965 ** .989 ** Sig. (2 tailed) .000 .000 .000 .000 .000 N 19 19 19 19 19 19 CS_02307359 Pearson Correlation 1 .686 ** .420 .960 ** .939 ** .988 ** Sig. (2 tailed) .001 .073 .000 .000 .000 N 19 19 19 19 19 19 CS_02309848 Pearson Correlation 1 .806** .955** .956** .964** .982** Sig. (2 tailed) 0 0 0 0 0 N 19 19 19 19 19 19 CS_02310000 Pearson Correlation 1 .663 ** .795 ** .995 ** .966 ** .954 ** Sig. (2 tailed) .002 .000 .000 .000 .000 N 19 19 19 19 19 19 CS_02324000 Pearson Correlation 1 .111 .909 ** .974 ** .976 ** .183 Sig. (2 tailed) .651 .000 .000 .000 .452 N 19 19 19 19 19 19 CS_02327100 Pearson Correlation 1 .327 .894 ** .987 ** .818 ** .973 ** Sig. (2 tailed) .172 .000 .000 .000 .000 N 19 19 19 19 19 19 Note: (**) Correlation is significant at the 0.01 level (2 tailed). (*) Correlation is significant at the 0.05 level (2 tailed). Table D 9 . List of springs per basin Gage_ID WBID Data Source Spring Name Station ID County Sprint type Magnitude 02324000 3573A USGS Iron Spring 294940083182800 Lafayette Spring 3 3573Z DEP FGS Steinhatchee Spring DEP FGS 384 Lafayette Spring, vent 4 3581 DEP FGS Unnamed Spring, 2953400831438 DEP FGS 428 Lafayette Spring, vent Unknown 02310000 1440F USGS Seven Spring 281251082395700 Pasco Spring None 02235000 2993 SJRWMD Witherington Spring 73691 Orange Spring 3 2967 DEP FGS Rock Spring DEP FGS 330 Orange Spring, vent 2 2967 DEP FGS Sulfur Spring DEP FGS 385 Orange Spring, vent 4 2956 SJRWMD Barrel Spring 73623 Orange Spring 3 2956C DEP FGS Wekiwa Spring DEP FGS 450 Orange Spring, vent 2 2987Z DEP FGS Sanlando Spring DEP FGS 346 Seminole Spring, vent 2 2987X DEP FGS Starbuck Spring DEP FGS 382 Seminole Spring, vent 2 2956Z DEP FGS Miami Spring DEP FGS 256 Seminole Spring, vent 3 2987X USGS Pegasus Spring 284147081232400 Seminole Spring 3 2987Y SJRWMD Palm Spring 73664 Seminole Spring 3 2987 DEP FGS Ginger Ale Spring DEP FGS 138 Seminole Spring, vent 5 2967 SJRWMD Tram Spring 79004 Orange Spring Unknown

PAGE 177

177 Table D 1 0 . List of domestic wastewater facilities per basin Gage ID Facility ID Site Name Treatment Process Summary Permitted Capacity (MGD) DW Class Effective Date 02231000 FL0040495 D 001 Activated Sludge W/Effluent To Sprayfield W/Overland Flow 1.3 IIC 1/16/2014 02235000 FL0036251 D 001 3 Ext Aeration Trains, Nutrient Removal, 2 ABW Filters, w/Eff to Reuse, Perc Ponds or Sweetwater Creek 2.9 IB 3/31/2011 02235000 FL0033251 D 001 Existing 12.5 MGD AADF activated sludge domestic WWTF; 12.5 IA 8/27/2012 02245500 FL0022853 D 001 An SBR; Consisting Of One Mechanical Grit Screen, One (9) Inch Parshall Flume, One Flow Splitter, Four Sbr Basins, Two Chlorine Contact Chambers, One Dechlorination; Two Belt Thickeners & One Aerobic Digester. 0.9 IIIC 12/5/2012 02303350 FL0039896 D 002 2 Stage Ext. Aeration, Nitrification/Denite "Rabco" Process, Golf Course Reuse, UV for stream discharge 0.4 IC 3/18/2014 02303350 FL0039896 D 001 2 Stage Ext. Aeration, Nitrification/Denite "Rabco" Process, Golf Course Reuse, UV for stream discharge 0.4 IC 3/18/2014 02310000 FL0127272 D 002 Master reuse system 26.75 10/25/2012

PAGE 178

178 Table D 1 1 . List of industrial wastewater facilities per basin Gage ID Facility ID Site Name Treatment Process Summary Capacity (MGD) Effective Date 02247510 FL0037877 D 001 Settling Discharge To Tomoka River Via Spruce Creek 0.11 2/1/2011 02231342 FL0174696 D 001 Anaerobic Digestion In Unlined Impoundment 0.56 2/11/2014 02231000 FL0435490 D 004 5 million gallon per day (MGD) maximum daily flow heavy mineral mining wastewater treatment system. Process water from the dredge mining operation will be collected in humate settling ponds that will be constructed as the mining progresses. Stormwat er fr 5 5/17/2011 FL0435490 D 003 5 million gallon per day (MGD) maximum daily flow heavy mineral mining wastewater treatment system. Process water from the dredge mining operation will be collected in humate settling ponds that will be constructed as the mining progresses. Stormwater fr 5 5/17/2011 FL0435490 D 001 5 million gallon per day (MGD) maximum daily flow heavy mineral mining wastewater treatment system. Process water from the dredge mining operation will be collected in humate settling ponds that will be constructed as the mining progresses. Stormwater fr 5 5/17/2011 FL0435490 D 002 5 million gallon per day (MGD) maximum daily flow heavy mineral mining wastewater treatment system. Process water from the dredge mining operation will be collected in humate settling ponds that will be constructed as the mining progresses. Stormwater fr 5 5/17/2011 FL0435490 D 005 5 million gallon per day (MGD) maximum daily flow heavy mineral mining wastewater treatment system. Process water from the d redge mining operation will be collected in humate settling ponds that will be constructed as the mining progresses. Stormwater fr 5 5/17/2011 02263800 FL0622591 D 001 The Aquatica treatment and disposal system (4.1 acre Pond) receives the following waste streams 1) ¿ stormwater runoff, imported from a portion of the International Drive and a large parking lot on the east side of the Aquatica park; 2) , the Commerson 0.673 10/19/201 0 FL0622648 D 002 The Discovery Cove treatment and disposal system receives the following waste streams from the Sea World attraction north of Discovery Cove (and north of the Central Florida Parkway, see DEP Exhibit number 6): 1. Non contact stormwater runoff; 2. Produc 0.861 10/19/201 0 FL0629332 D 004 Wastewater tre atment and disposal 0.73 10/19/201 0 FL0629332 D 003 Wastewater treatment and disposal 0.73 10/19/201 0

PAGE 179

179 Table D 1 2 . Cross correlation analysis between the catchment coefficient and surface water withdrawals at the county level Gage_ID cross correlation coefficient (surface water withdrawals) lag number cross correlation coefficient (groundwater withdrawals) Lag numbe r 02231000 n/a 02235000 0.334 2 n/a 02245500 n/a 02267000 n/a 02296500 n/a 02297100 0.436 4 02307359 0.422 2 02310000 n/a n/a 02324000 0.429 7 n/a Additional basins: 02263800 n/a 02266480 0.518 1 02298123 0.563 4 02303350 0.481 1 Figure D 1 . Time series of catchment coefficient ( t ) for the pre selected sample 0 2 4 6 8 10 12 14 16 18 20 1980 1985 1990 1995 2000 2005 2010 Catchment coefficient, _10 Year

PAGE 180

180 Figure D 2. Geographical distribution of nine (9) basins sub sample for long term analysis (1950 2010) of trends in the catchment coefficient. Figure D 3. Linear regression plot for pooled time series, catchment coefficient ( _10)=f(UBG) differentiated by case stu dy basin 1.00 2.00 3.00 4.00 5.00 6.00 7.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 Catchment coefficient, UBG 02231342 02235000 02263800 02266480 02296500 02297100 02298123 02303350 02307359 02310000

PAGE 181

181 Figure D 4 . Surface water withdrawals and the catchment coefficient for St. Marys River basin (USGS 02231000 ) , 1960 2010 Figure D 5 . Surface water withdrawals and the catchment coefficient for Wekiva River basin (USGS 02235000 ) , 1960 2010 0.50 1.00 1.50 2.00 2.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 5.50 6.00 1960 1970 1980 1990 2000 2010 surface freshwater withdrawals, million gallons per day (MGD) catchment coefficient, Year 02231000 surface water (Baker county) 20.00 40.00 60.00 80.00 100.00 120.00 140.00 160.00 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 5.50 6.00 1960 1970 1980 1990 2000 2010 surface freshwater withdrawals, million gallons per day (MGD) catchment coefficient, Year 02235000 surface water (Orange county) surface water (Seminole county)

PAGE 182

182 Figure D 6 . Surface water withdrawals and the catchment coefficient for South Black Fork Creek basin (USGS 02245500 ) , 1960 2010 Figure D 7 . Surface water withdrawals and the catchment coefficient for Catfish Creek basin (USGS 02267000 ) , 1960 2010 1.00 2.00 3.00 4.00 5.00 6.00 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 5.50 6.00 1960 1970 1980 1990 2000 2010 surface freshwater withdrawals, million gallons per day (MGD) Catchment coefficient, Year 02245500 surface water (Clay county) 50.00 100.00 150.00 200.00 250.00 300.00 350.00 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 5.50 6.00 1960 1970 1980 1990 2000 2010 surface freshwater withdrawals, million gallons per day (MGD) catchment coefficient, Year 02267000 surface water (Polk county)

PAGE 183

183 Fig ure D 8 . Surface water withdrawals and the catchment coefficient for Charlie Creek basin (USGS 02296500) , 1960 2010 Figure D 9 . Surface water withdrawals and the catchment coefficient for Joshua Creek basin (USGS 02297100 ) , 1960 2010 1.00 2.00 3.00 4.00 5.00 6.00 7.00 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 5.50 6.00 1960 1970 1980 1990 2000 2010 surface freshwater withdrawals, million gallons per day (MGD) catchment coefficient, Year 02296500 surface water (Hardee county) 5.00 10.00 15.00 20.00 25.00 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 5.50 6.00 1960 1970 1980 1990 2000 2010 surface freshwater withdrawals, million gallons per day (MGD) Catchment coefficient, Year 02297100 surface water (Desoto county)

PAGE 184

184 Figure D 10 . Sur face water withdrawals and the catchment coefficient for Brooker Creek basin (USGS 02307359 ) , 1960 2010 Figure D 11 . Surface water withdrawals and the catchment coefficient for Anclote River basin (USGS 02310000 ) , 1960 2010 20.00 40.00 60.00 80.00 100.00 120.00 140.00 160.00 180.00 200.00 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 5.50 6.00 1960 1970 1980 1990 2000 2010 surface freshwater withdrawals, million gallons per day (MGD) catchment coefficient, Year 02307359 surface water (Hillsborough county) 2.00 4.00 6.00 8.00 10.00 12.00 14.00 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 5.50 6.00 1960 1970 1980 1990 2000 2010 surface freshwater withdrawals, million gallons per day (MGD) catchment coefficient, Year 02310000 surface water (Pasco county)

PAGE 185

185 Figure D 1 2 . Surface water withdrawals and the catchment coefficient for Steinhatchee River basin (USGS 02324000 ) , 1960 2010 Figure D 1 3 . Surface water withdrawals and the catchment coefficient for Shingle Creek basin (USGS 02263800 ) , 1965 2010 0.10 0.20 0.30 0.40 0.50 0.60 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 1960 1970 1980 1990 2000 2010 surface freshwater withdrawals, million gallons per day (MGD) catchment coefficient, Year 02324000 surface water (Lafayatte county) 20.00 40.00 60.00 80.00 100.00 120.00 140.00 160.00 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 5.50 6.00 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 surface freshwater withdrawals, million gallons per day (MGD) catchment coefficient, 02263800 surface water (Orange county)

PAGE 186

186 Figure D 1 4 . Surface water withdrawals and the catchment coefficient for Davenport Creek basin (USGS 02266480 ) , 1965 2010 Figure D 1 5 . Surface water withdrawals and the catchment coefficient for Prairie Creek basin (USGS 02298123 ) , 1965 2010 50.00 100.00 150.00 200.00 250.00 300.00 350.00 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 5.50 6.00 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 surface freshwater withdrawals, million gallons per day (MGD) catchment coefficient, 02266480 surface water (Polk county) surface water (Osceola county) y = 0.0289x + 60.53 R² = 0.7676 5.00 10.00 15.00 20.00 25.00 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 5.50 6.00 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 surface freshwater withdrawals, million gallons per day (MGD) catchment coefficient, 02298123 surface water (Desoto county) Linear (02298123)

PAGE 187

187 Figure D 1 6 . Surfac e water withdrawals and the catchment coefficient for Trout Creek basin (USGS 02303350 ) , 1965 2010 Figure D 1 7 . Cross correlation plot of the catchment coefficient and county level surface water withdrawals with upper and lower confidence limits for St. Marys River basin (USGS 02231000 ) in Baker County . y = 0.0158x + 34.015 R² = 0.2017 20.00 40.00 60.00 80.00 100.00 120.00 140.00 160.00 180.00 200.00 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 5.50 6.00 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 surface freshwater withdrawals, million gallons per day (MGD) catchment coefficient, 02303350 surface water (Pasco county) surface water (Hillsborough county) Linear (02303350)

PAGE 188

188 A B Figure D 18 . Cross correlation plot of the catchment coefficient and county level surface water withdrawals with upper and lower confidence limits for Wekiva River basin (USGS 02235000 ) . A) Seminole County; B) Orange County.

PAGE 189

189 Figure D 19 . Cross correlation plot of the catchment coefficient and county level surface water withdrawals with upper and lower confidence limits for South Black Fork Creek basin (USGS 02245500 ) in Clay County . Figure D 2 0 . Cross correlation plot of the catchment coefficient and county level surface water withdrawals with upper and lower confidence limits for Catfish Creek basin (USGS 02267000 ) in Polk County .

PAGE 190

190 Figure D 2 1 . Cross correlation plot of the catchment coefficient and county level surface water withdrawals with upper and lower confidence limits for Charlie Creek basin (USGS 02296500) in Hardee County . Figure D 2 2 . Cross correlation plot of the catchment coefficient and county level surface water withdrawals with upper and lower confidence limits for Joshua Creek basin (USGS 02297100 ) in Desoto County .

PAGE 191

191 Figure D 2 3 . Cross correlation plot of the catchment coeffic ient and county level surface water withdrawals with upper and lower confidence limits for Brooker Creek basin (USGS 02307359 ) in Hillsborough County . Figure D 2 4 . Cross correlation plot of the catchment coefficient and county level surface water withdr awals with upper and lower confidence limits for Anclote River basin (USGS 02310000 ) in Pasco County .

PAGE 192

192 Figure D 2 5 . Cross correlation plot of the catchment coefficient and county level surface water withdrawals with upper and lower confidence limits for S teinhatchee River basin (USGS 02324000 ) in Lafayatte County . Figure D 2 6 . Cross correlation plot of the catchment coefficient and county level surface water withdrawals with upper and lower confidence limits for Shingle Creek basin (USGS 02 2638 00 ) in Orange County .

PAGE 193

193 A B Figure D 27 . Cross correlation plot of the catchment coefficient and county level surface water withdrawals with upper and lower confidence limits for Davenport Creek basin (USGS 02 266 4 8 0 ) . A) Osceola County; B) Polk County.

PAGE 194

194 Figure D 28 . Cross correlation plot of the catchment coefficient and county level surface water withdrawals with upper and lower confidence limits for Prairie Creek basin (USGS 02 298123 ) in Desoto County .

PAGE 195

195 A B F igure D 29 . Cross correlation plot of the catchment coefficient and county level surface water withdrawals with upper and lower confidence limits for Trout Creek basin (USGS 0230 335 0 ) . A) Hillsborough County; B) Pasco County.

PAGE 196

196 A B F igure D 3 0 . Cross correla tion plot of the catchment coefficient and county level surface water withdrawals with upper and lower confidence limits for Wekiva River basin (USGS 02 235000 ) . A) Orange County; B) Seminole County.

PAGE 197

197 F igure D 3 1 . Cross correlation plot of the catchment c oefficient and county level surface water withdrawals with upper and lower confidence limits for Anclote River basin (USGS 023 1000 0 ) in Pasco County . F igure D 3 2 . Cross correlation plot of the catchment coefficient and county level surface water withdrawals with upper and lower confidence limits for Steinhatchee River basin (USGS 023 2400 0 ) in Lafayatte County .

PAGE 198

198 APPENDIX E STATISTICAL MODEL OUTPUT The following panel data models were produced with LIMDEP 10, student version ( http://www.limdep.com/ ). The sample group consists of the four ( 4 ) selected basins described in Chapter 3. The data records cover 19 years from 1992 unt il 2010, representing 76 pooled observations. Model = f(UBG) ----------------------------------------------------------------------------Ordinary least squares regression ............ LHS=PI_10 Mean = 3.17199 Standard deviation = .57353 ---------No. of observations = 76 DegFreedom Mean square Regression Sum of Squares = .105964E 01 1 .01060 Residual Sum of Squares = 24.6594 74 .33323 Total Sum of Squares = 24.6700 75 .32893 ---------Standard error of e = .57726 Root MSE .56962 Fit R squared = .00043 R bar squared .01308 Model test F[ 1, 74] = .03180 Prob F > F* .85896 [High values of LM favor FEM/REM over base model] Baltagi Li form of LM Statistic = 295.00601 [= BP if balanced panel] Moulton/Randolph form:SLM N[0,1] = 31.42404 B P test Chi squared [ 1] = 295.00601 Prob C2 > C2* = .00000 -------------------------------------------------Panel Data Analysis of PI_10 [ONE way] Unconditional ANOVA (No regressors) Source Variation Deg. Free. Mean Square Between 16.731 3 5.5768 Residual 7.9395 72 .11027 Total 24.670 75 .32893 -------+ -------------------------------------------------------------------| Standard Prob. 95% Confidence PI_10| Coefficient Error t |t|>T* Interval -------+ -------------------------------------------------------------------UBG| .05437 .30492 .18 .8590 .5 5319 .66194 Constant| 3.15208*** .12984 24.28 .0000 2.89337 3.41078 -------+ -------------------------------------------------------------------***, **, * ==> Significance at 1%, 5%, 10% level. Model was estimated on Jul 15, 2014 at 07:16:46 PM --------------------------------------------------------------------------------------------------------------------------------------------------------LSDV least squares with fixed effects .... LHS=PI_10 Mean = 3.17199 Standard deviation = .57353 ---------No. of observations = 76 DegFreedom Mean square Regression Sum of Squares = 19.5165 4 4.87912 Residual Sum of Squares = 5.15351 71 .07258 Total Sum of Squares = 24.6700 75 .32893 ---------Standard error of e = .26942 Root MSE .26040 Fit R squared = .79110 R bar squared .77933 Estd. Autocorrelation of e(i,t) = .688789 -------------------------------------------------Panel:Groups Empty 0, Valid data 4 Smallest 19, Largest 19 Average group size i n panel 19.00 Variances Effects a(i) Residuals e(i,t) 1.286963 .072585 Std.Devs. 1.134444 .269415 Rho squared: Residual variation due to ai .946611 Within groups variation in PI_10 .79395D+01 R squared based on within group variation .350899 Between group variation in PI_10 .16731D+02 -------+ -------------------------------------------------------------------| Standard Prob. 9 5% Confidence PI_10| Coefficient Error t |t|>T* Interval -------+ -------------------------------------------------------------------UBG| 3.82183*** .61689 6.20 .0000 5.05073 2.59292 -------+ ------------------------------------------------------------------

PAGE 199

199 ***, **, * ==> Significance at 1%, 5%, 10% level. Model was estimated on Jul 15, 2014 at 07:16:46 PM ----------------------------------------------------------------------------------------------------------------------------------------------Test Statistics for the Fixed Effects Regression Model -------------------------------------------------------------------Model Log Likelihood Sum of Squares R squared (1) Constant term only 65.08376 24.66997 .00000 (2) Group effects only 22.00154 7.93945 .67817 (3) X variables only 65.06743 24.65937 .00043 (4) X and group effects 5.57920 5.15351 .79110 -------------------------------------------------------------------Hypothesis Tests Likelihood Ratio Test F Tests Chi squared d.f. Prob F num denom P value (2) vs (1) 86.16 3 .0000 50.57 3 72 .00000 (3) vs (1) .03 1 .8566 .03 1 74 .85896 (4) vs (1) 119.01 4 .0000 67.22 4 71 .00000 (4) vs (2) 32.84 1 .0000 38.38 1 71 .00000 (4) vs (3) 118.98 3 .0000 89.58 3 71 .00000 -----------------------------------------------------------------------------------------------------------------------------------------------Random Effects Model: v(i,t) = e(i,t) + u(i) Estimates: Var[e] = .072585 Var[u] = .260650 Corr[v(i,t),v(i,s)] = .782182 Sum of Squares 562.175 R squared .281122 Variances co mputed using OLS and LSDV with d.f. ----------------------------------------------Lagrange Multiplier Test vs. Model (3) = 295.01 [ 1 degrees of freedom, prob. value = .000000] [High values of LM favor FEM/REM over CR model] Fixed vs. Random Effects (Hausman) = 143.80 [ 1 degrees of freedom, prob. value = .000000] [High (low) values of H favor F.E.(R.E.) model] -------+ -------------------------------------------------------------------| Standard Prob. 9 5% Confidence PI_10| Coefficient Error z |z|>Z* Interval -------+ -------------------------------------------------------------------UBG| 1.44648*** .43306 3.34 .0008 .59769 2.29527 -------+ ------------------------------------------------------------------***, **, * ==> Significance at 1%, 5%, 10% level. Model was estimated on Jul 15, 2014 at 07:16:46 PM ----------------------------------------------------------------------------Model = f(UBG, PDB) ----------------------------------------------------------------------------Ordinary least squares regression ............ LHS=PI_10 Mean = 3.17199 Standard deviation = .57353 ---------No. of observations = 76 DegFreedom Mean square Regression Sum of Squares = .130516 2 .06526 Residual Sum of Squares = 24.5395 73 .33616 Total Sum of Squares = 24.6700 75 .32893 ---------Standard error of e = .57979 Root MSE .56823 Fit R squared = .00529 R bar squared .02196 Model test F[ 2, 73] = .19413 Prob F > F* .82397 [High values of LM favor FEM/REM over base model] Baltagi Li form of LM Statistic = 285.70264 [= BP if balanced panel] Moulton/Randolph form:SLM N[0,1] = 45.95764 B P test Chi squared [ 1] = 285.70264 Prob C2 > C2* = .00000 -------------------------------------------------Panel Data Analysis of PI_10 [ONE way] Unconditional ANOVA (No regressors) Source Variation Deg. Free. Mean Square Between 16.731 3 5.5768 Residual 7.9395 72 .11027 Total 24.670 75 .32893 -------+ -------------------------------------------------------------------| Standard Prob. 95% Confidence PI_10| Coefficient Error t |t|>T* Interval -------+ -------------------------------------------------------------------UBG| .49476 .96907 .51 .6112 2.42611 1.43659 PD| .00041 .000 68 .60 .5522 .00096 .00177 Constant| 3.23396*** .18921 17.09 .0000 2.85687 3.61106 -------+ -------------------------------------------------------------------***, **, * ==> Significance at 1%, 5%, 10% level.

PAGE 200

200 Model was es timated on Jul 15, 2014 at 07:24:33 PM --------------------------------------------------------------------------------------------------------------------------------------------------------LSDV least squares with fixed effects .... LHS=PI_10 Mean = 3.17199 Standard deviation = .57353 ---------No. of observations = 76 DegFreedom Mean square Regression Sum of Squares = 19.5614 5 3.91228 Residual Sum of Squares = 5.10858 70 .07298 Total Sum of Squares = 24.6700 75 .32893 ---------Standard error of e = .27015 Root MSE .25926 Fit R squared = .79292 R bar squared .77813 Estd. Autocorrelation of e(i,t) = .689495 -------------------------------------------------Panel:Groups Empty 0, Valid data 4 Smallest 19, Largest 19 Average group size in panel 19.00 Variances Effects a(i) Residuals e(i,t) 1.229472 .072980 Std.Devs. 1.108816 .270147 Rho squared: Residual variation due to ai .943967 W ithin groups variation in PI_10 .79395D+01 R squared based on within group variation .356558 Between group variation in PI_10 .16731D+02 -------+ -------------------------------------------------------------------| Standard Prob. 95% Confidence PI_10| Coefficient Error t |t|>T* Interval -------+ -------------------------------------------------------------------UBG| 2.75810* 1.49013 1.85 .0682 5.727 24 .21104 PD| .00068 .00087 .78 .4352 .00241 .00105 -------+ -------------------------------------------------------------------***, **, * ==> Significance at 1%, 5%, 10% level. Model was estimated on Jul 15, 2014 at 0 7:24:33 PM -----------------------------------------------------------------------------------------------------------------------------------------------Test Statistics for the Fixed Effects Regression Model ------------------------------------------------------------------Model Log Likelihood Sum of Squares R squared (1) Constant term only 65.08376 24.66997 .00000 (2) Group effects only 22.00154 7.93945 .67817 (3) X variables only 64.88219 24.53945 .00529 (4) X and group effects 5.24644 5.10858 .79292 -------------------------------------------------------------------Hypothesis Tests Likelihood Ratio Tes t F Tests Chi squared d.f. Prob F num denom P value (2) vs (1) 86.16 3 .0000 50.57 3 72 .00000 (3) vs (1) .40 2 .8174 .19 2 73 .82397 (4) vs (1) 119.67 5 .00 00 53.61 5 70 .00000 (4) vs (2) 33.51 2 .0000 19.39 2 70 .00000 (4) vs (3) 119.27 3 .0000 88.75 3 70 .00000 ----------------------------------------------------------------------------------------------------------------------------------------------Random Effects Model: v(i,t) = e(i,t) + u(i) Estimates: Var[e] = .072980 Var[u] = .263177 Corr[v(i,t),v(i,s)] = .782900 Sum of Squares 283.137 R squared 1.120487 Variances computed using OLS and LSDV with d.f. ----------------------------------------------Lagrange Multiplier Test vs. Model (3) = 285.70 [ 1 degrees of freedo m, prob. value = .000000] [High values of LM favor FEM/REM over CR model] Fixed vs. Random Effects (Hausman) = 95.85 [ 2 degrees of freedom, prob. value = .000000] [High (low) values of H favor F.E.(R.E.) model] -------+ -------------------------------------------------------------------| Standard Prob. 95% Confidence PI_10| Coefficient Error z |z|>Z* Interval -------+ ------------------------------------------------------------------UBG| 8.31135*** .95753 8.68 .0000 6.43462 10.18808 PD| .00554*** .00069 8.06 .0000 .00689 .00419 -------+ ------------------------------------------------------------------***, **, * ==> Significance at 1%, 5%, 10% level. Model was estimated on Jul 15, 2014 at 07:24:33 PM ----------------------------------------------------------------------------Model = f(UBG, PDB, AG) ----------------------------------------------------------------------------

PAGE 201

201 Ordinary least squares regression ............ LHS=PI_10 Mean = 3.17199 Standard deviation = .57353 ---------No. of observations = 76 DegFreedom Mean square Regression Sum of Squares = 1.72844 3 .57615 Residual Sum of Squares = 22.9415 72 .31863 Total Sum of Squares = 24 .6700 75 .32893 ---------Standard error of e = .56448 Root MSE .54942 Fit R squared = .07006 R bar squared .03131 Model test F[ 3, 72] = 1.80818 Prob F > F* .15334 [High values of LM favor FEM/REM over base model] Baltagi Li form of LM Statistic = 247.42070 [= BP if balanced panel] Moulton/Randolph form:SLM N[0,1] = 46.00507 B P test Chi squared [ 1] = 247.42070 Prob C2 > C2* = .00000 -------------------------------------------------Panel Data Analysis of PI_10 [ONE way] Unconditional ANOVA (No regressors) Source Variation Deg. Free. Mean Square Between 16.731 3 5.576 8 Residual 7.9395 72 .11027 Total 24.670 75 .32893 -------+ -------------------------------------------------------------------| Standard Prob. 95% Confidence PI_10| Coefficient Error t |t|>T* Interval -------+ -------------------------------------------------------------------AG| 2.73355** 1.22066 2.24 .0282 5.16689 .30021 UBG| 7.95060** 3.46048 2.30 .0245 14.84895 1.05225 PD| .00375** .00163 2.29 .0247 .00049 .00700 Constant| 5.77448*** 1.14932 5.02 .0000 3.48335 8.06560 -------+ ------------------------------------------------------------------***, **, * ==> Significance at 1%, 5%, 10% level. Model was estimated on Jul 15, 2014 at 08:30:16 PM -------------------------------------------------------------------------------------------------------------------------------------------------------LSDV least squares with fixed effects .... LHS=PI_10 Mean = 3.17199 Standard deviation = .57353 ---------No. of observations = 76 DegFr eedom Mean square Regression Sum of Squares = 20.1586 6 3.35976 Residual Sum of Squares = 4.51141 69 .06538 Total Sum of Squares = 24.6700 75 .32893 --------Standard error of e = .25570 Root MSE .24364 Fit R squared = .81713 R bar squared .80123 Estd. Autocorrelation of e(i,t) = .647161 -------------------------------------------------Panel :Groups Empty 0, Valid data 4 Smallest 19, Largest 19 Average group size in panel 19.00 Variances Effects a(i) Residuals e(i,t) 5.438767 .065383 Std.Devs. 2.332116 .255701 Rho squared: Residual variation due to ai .988121 Within groups variation in PI_10 .79395D+01 R squared based on within group variation .431773 Between group variation in PI_10 .16731D+02 ------+ -------------------------------------------------------------------| Standard Prob. 95% Confidence PI_10| Coefficient Error t |t|>T* Interval -------+ ------------------------------------------------------------------AG| 12.6950*** 4.20066 3.02 .0035 21.0669 4.3231 UBG| 7.99986*** 2.23554 3.58 .0006 12.45529 3.54443 PD| .00127 .00084 1.51 .1366 .00296 .00041 -------+ -------------------------------------------------------------------***, **, * ==> Significance at 1%, 5%, 10% level. Model was estimated on Jul 15, 2014 at 08:30:16 PM ----------------------------------------------------------------------------------------------------------------------------------------------Test Statistics for the Fixed Effects Regression Model -------------------------------------------------------------------Model Log Likelihood Sum o f Squares R squared (1) Constant term only 65.08376 24.66997 .00000 (2) Group effects only 22.00154 7.93945 .67817 (3) X variables only 62.32352 22.94153 .07006 (4) X and group effects .52264 4.51141 .81713 -------------------------------------------------------------------Hypothesis Tests Likelihood Ratio Test F Tests Chi squared d.f. Prob F num denom P value (2) vs (1) 86.16 3 .0000 50.57 3 72 .00000 (3) vs (1) 5.52 3 .1374 1.81 3 72 .15334 (4) vs (1) 129.12 6 .0000 51.39 6 69 .00000

PAGE 202

202 (4) vs (2) 42.96 3 .0000 17.48 3 69 .00000 (4) vs (3) 123.60 3 .0000 93.96 3 69 .00000 -----------------------------------------------------------------------------------------------------------------------------------------------Random Effects Mo del: v(i,t) = e(i,t) + u(i) Estimates: Var[e] = .065383 Var[u] = .253250 Corr[v(i,t),v(i,s)] = .794802 Sum of Squares 130.688 R squared 3.377668 Variances computed using OLS and LSDV with d.f. ----------------------------------------------Lagrange Multiplier Test vs. Model (3) = 247.42 [ 1 degrees of freedom, prob. value = .000000] [High values of LM favor FEM/REM over CR model] Fixed vs. Random Effects (Hausman) = 61.95 [ 3 degrees of freedom, prob. value = .000000] [High (low) values of H favor F.E.(R.E.) model] -------+ -------------------------------------------------------------------| Standard Prob. 95% Confidence PI_10| Coefficient Error z |z|>Z* Interval -------+ -------------------------------------------------------------------AG| 5.23820*** .62790 8.34 .0000 4.00753 6.46887 UBG| 5.34148*** .97751 5.46 .0000 3.42560 7.25736 PD| .00285*** .00072 3.94 .0001 .00427 .00143 -------+ -------------------------------------------------------------------***, **, * ==> Significance a t 1%, 5%, 10% level. Model was estimated on Jul 15, 2014 at 08:30:16 PM --------------------------------------------------------------------------------------------------------------------------------------------------------Ordinary least squares regression ............ LHS=PI_10 Mean = 3.17199 Standard deviation = .57353 ---------No. of observations = 76 DegFreedom Mean square Regression Sum of Squares = .128209E 02 1 .00128 Residual Sum of Squares = 24.6687 74 .33336 Total Sum of Squares = 24.6700 75 .32893 ---------Standard error of e = .57737 Root MSE .56973 Fit R squared = .00005 R bar squared .01346 Model test F[ 1, 74] = .00385 Prob F > F* .95072 [High values of LM favor FEM/REM over base model] Baltagi Li form of LM Statistic = 297.41334 [= BP if balanced panel] Moulton/Randolph form:SLM N[0,1] = 31.78425 B P test Chi squared [ 1] = 297.41334 Prob C2 > C2* = .00000 -------------------------------------------------Panel Data Analysis of PI_10 [ONE way] Unconditional ANOVA (No regressors) Source Variation Deg. Free. Mean Square Between 16.731 3 5.5768 Residual 7.9395 72 .11027 Total 24.670 75 .328 93 -------+ -------------------------------------------------------------------| Standard Prob. 95% Confidence PI_10| Coefficient Error t |t|>T* Interval -------+ -------------------------------------------------------------------URB| .01850 .29838 .06 .9507 .57604 .61305 Constant| 3.16666*** .10850 29.19 .0000 2.95048 3.38285 -------+ ------------------------------------------------------------------***, **, * ==> Significance at 1%, 5%, 10% level. Model was estimated on Jul 15, 2014 at 07:15:53 PM -------------------------------------------------------------------------------------------------------------------------------------------------------LSDV least squares with fixed effects .... LHS=PI_10 Mean = 3.17199 Standard deviation = .57353 ---------No. of observations = 76 D egFreedom Mean square Regression Sum of Squares = 18.9824 4 4.74560 Residual Sum of Squares = 5.68759 71 .08011 Total Sum of Squares = 24.6700 75 .32893 --------Standard error of e = .28303 Root MSE .27356 Fit R squared = .76945 R bar squared .75646 Estd. Autocorrelation of e(i,t) = .716445 -------------------------------------------------P anel:Groups Empty 0, Valid data 4 Smallest 19, Largest 19 Average group size in panel 19.00 Variances Effects a(i) Residuals e(i,t)

PAGE 203

203 1.447463 .080107 Std.Devs. 1.203105 .283032 Rho squared: Residual variation due to ai .947559 Within groups variation in PI_10 .79395D+01 R squared based on within group variation .283630 Between group variation in PI_10 .16731D+02 ------+ -------------------------------------------------------------------| Standard Prob. 95% Confidence PI_10| Coefficient Error t |t|>T* Interval -------+ ------------------------------------------------------------------URB| 4.10047*** .77339 5.30 .0000 5.64114 2.55980 -------+ -------------------------------------------------------------------***, **, * ==> Significance at 1%, 5%, 10% level. Model was e stimated on Jul 15, 2014 at 07:15:53 PM -----------------------------------------------------------------------------------------------------------------------------------------------Test Statistics for the Fixed Effects Regression Model ------------------------------------------------------------------Model Log Likelihood Sum of Squares R squared (1) Constant term only 65.08376 24.66997 .00000 (2) Group effects only 22.00154 7.93945 .67817 (3) X variables only 65.08178 24.66869 .00005 (4) X and group effects 9.32632 5.68759 .76945 -------------------------------------------------------------------Hypothesis Tests Likelihood Ratio Test F Tests Chi squared d.f. Prob F num denom P value (2) vs (1) 86.16 3 .0000 50.57 3 72 .00000 (3) vs (1) .00 1 .9499 .00 1 74 .95072 (4) vs (1) 111.51 4 .0000 59.24 4 71 .00000 (4) vs (2) 25.35 1 .0000 28.11 1 71 .00000 (4) vs (3) 111.51 3 .0000 78.98 3 71 .00000 ----------------------------------------------------------------------------------------------------------------------------------------------Random Effects Model: v(i,t) = e(i,t) + u(i) Estimates: Var[e] = .080107 Var[u] = .253254 Corr[v(i ,t),v(i,s)] = .759699 Sum of Squares 549.768 R squared .606731 Variances computed using OLS and LSDV with d.f. ----------------------------------------------Lagrange Multiplier Test vs. Model (3) = 297.41 [ 1 degrees of freedom, prob. value = .000000] [High values of LM favor FEM/REM over CR model] Fixed vs. Random Effects (Hausman) = 114.22 [ 1 degrees of freedom, prob. value = .000000] [High (low) values of H favor F.E.(R.E.) model] -------+ -------------------------------------------------------------------| Standard Prob. 95% Confidence PI_10| Coefficient Error z |z|>Z* Interval -------+ -------------------------------------------------------------------URB| 2.01800*** .51998 3.88 .0001 .99886 3.03714 -------+ -------------------------------------------------------------------***, **, * ==> Significan ce at 1%, 5%, 10% level. Model was estimated on Jul 15, 2014 at 07:15:53 PM ----------------------------------------------------------------------------Model = f(URB, POP) ----------------------------------------------------------------------------O rdinary least squares regression ............ LHS=PI_10 Mean = 3.17199 Standard deviation = .57353 ---------No. of observations = 76 DegFreedom Mean square Regression Sum of Squares = 10.3226 2 5.16129 Residual Sum of Squares = 14.3474 73 .19654 Total Sum of Squares = 24.6700 75 .32893 ---------Standard error of e = .44333 Root MSE .43449 Fit R squared = .41843 R bar squared .40249 Model test F[ 2, 73] = 26.26079 Prob F > F* .00000 [High values of LM favor FEM/REM over base model] Baltagi Li form of LM Statistic = 88.72261 [= BP if balanced panel] Moulton/Randolph form:SLM N[0,1] = 26.94779 B P test Chi squared [ 1] = 88.72261 Prob C2 > C2* = .00000 -------------------------------------------------Panel Data Analysis of PI_10 [ONE way] Unconditional ANOVA (No regressors) Source Variation Deg. Free. Mean Square Between 16.731 3 5.5768

PAGE 204

204 Residual 7.9395 72 .11027 Total 24.670 75 .32893 -------+ -------------------------------------------------------------------| Standard Prob. 95% Confidence PI_10| Coefficient Error t |t|>T* Interval -------+ ------------------------------------------------------------------URB| 3.33713*** .51663 6.46 .0000 4.36678 2.30748 POP| .93602D 05*** .1292D 05 7.25 .0000 .67860D 05 .11935D 04 Constant| 3.55181*** .09882 35.94 .0000 3.35487 3.74875 -------+ -------------------------------------------------------------------nnnnn.D xx or D+xx => multiply by 10 to xx or +xx. ***, **, * ==> Significance at 1%, 5%, 10% level. Model was estimated on Jul 15, 2014 at 07:26:03 PM --------------------------------------------------------------------------------------------------------------------------------------------------------LSDV least squares with fixed effects .... LHS=PI_10 Mean = 3.17199 Standard deviation = .57353 ---------No. of observations = 76 DegFreedom Mean square Regression Sum of Squares = 19.1947 5 3.83894 Residual Sum of Squares = 5.47529 70 .07822 Total Sum of Squares = 24.6700 75 .32893 ---------Standard error of e = .27968 Root MSE .26841 Fit R squared = .77806 R bar squared .76221 Estd. Autocorrelation of e(i,t) = .718329 -------------------------------------------------Panel:Groups Empty 0, Valid data 4 Smallest 19, Largest 19 Average group size in panel 19.00 Variances Effects a(i) Residuals e(i,t) 1.977354 .078218 Std.Devs. 1.406184 .279675 Rho squared: Residual variation due to ai .961948 Within groups variation in PI_10 .793 95D+01 R squared based on within group variation .310370 Between group variation in PI_10 .16731D+02 -------+ -------------------------------------------------------------------| Standard Prob. 95% Confide nce PI_10| Coefficient Error t |t|>T* Interval -------+ -------------------------------------------------------------------URB| 3.02991*** 1.00314 3.02 .0035 5.02871 1.03111 POP| .44899D 05 .2725D 05 1.65 .1037 .99202D 05 .94040D 06 -------+ -------------------------------------------------------------------nnnnn.D xx or D+xx => multiply by 10 to xx or +xx. ***, **, * ==> Significance at 1%, 5%, 10% level. Model was estimated on Jul 15, 2014 at 07:26:03 PM -----------------------------------------------------------------------------------------------------------------------------------------------Test Statistics for the Fixed Effects Regression Mo del -------------------------------------------------------------------Model Log Likelihood Sum of Squares R squared (1) Constant term only 65.08376 24.66997 .00000 (2) Group effects only 22.00154 7.93945 .67817 (3) X variables only 44.48706 14.34740 .41843 (4) X and group effects 7.88075 5.47529 .77806 -------------------------------------------------------------------Hypoth esis Tests Likelihood Ratio Test F Tests Chi squared d.f. Prob F num denom P value (2) vs (1) 86.16 3 .0000 50.57 3 72 .00000 (3) vs (1) 41.19 2 .0000 26.26 2 73 .00000 (4) vs (1) 114.41 5 .0000 49.08 5 70 .00000 (4) vs (2) 28.24 2 .0000 15.75 2 70 .00000 (4) vs (3) 73.21 3 .0000 37.81 3 70 .00000 -----------------------------------------------------------------------------------------------------------------------------------------------Random Effects Model: v(i,t) = e(i,t) + u(i) Estimates: Var[e] = .078218 V ar[u] = .118321 Corr[v(i,t),v(i,s)] = .602023 Sum of Squares 379.229 R squared 2.340197 Variances computed using OLS and LSDV with d.f. ---------------------------------------------Lagrange Multiplier Test vs. Model (3) = 88.72 [ 1 degrees of freedom, prob. value = .000000] [High values of LM favor FEM/REM over CR model] Fixed vs. Random Effects (Hausman) = 165.25 [ 2 degrees of freedom, prob. value = .000000] [ High (low) values of H favor F.E.(R.E.) model] -------+ -------------------------------------------------------------------

PAGE 205

205 | Standard Prob. 95% Confidence PI_10| Coefficient Error z |z|>Z* Interval -------+ -------------------------------------------------------------------URB| 5.53129*** .70880 7.80 .0000 4.14207 6.92051 POP| .63977D 05*** .2222D 05 2.88 .0040 .10753D 04 .20419D 05 -------+ ------------------------------------------------------------------nnnnn.D xx or D+xx => multiply by 10 to xx or +xx. ***, **, * ==> Significance at 1%, 5%, 10% level. Model was estimated on Jul 15, 2014 at 07:26:03 PM --------------------------------------------------------------------------------------------------------------------------------------------------------Ordinary least squares regression ............ LHS=PI_10 Mean = 3.17199 Standard deviation = .57353 ---------No. of observations = 76 DegFreedom Mean square Regression Sum of Squares = 19.1861 4 4.79653 Residual Sum of Squares = 5.48383 71 .07724 Total Sum of Squares = 24.6700 75 .32893 ---------Standard error of e = .27792 Root MSE .26862 Fit R squared = .77771 R bar squared .76519 Model test F[ 4, 71] = 62.10143 Prob F > F* .00000 [High values of LM favor FEM/REM over base model] Baltagi Li form of LM Statistic = 2.09506 [= BP if balanced panel] B P test Chi squared [ 1] = 2.09506 Prob C2 > C2* = .14778 -------------------------------------------------Panel Data Analysis of PI_10 [ONE way] Unconditional ANOVA (No regressors) Source Variation Deg. Free . Mean Square Between 16.731 3 5.5768 Residual 7.9395 72 .11027 Total 24.670 75 .32893 -------+ -------------------------------------------------------------------| Standard Prob. 95% Confidence PI_10| Coefficient Error t |t|>T* Interval -------+ -------------------------------------------------------------------URB| 2.72108*** .36316 7.49 .0000 3.44521 1.99695 POP| .50774D 05** .2053D 05 2.47 .0158 .91712D 05 .98354D 06 COMPACT| 1.93394*** .18842 10.26 .0000 1.55824 2.30963 SLOPE| 6.77454*** 1.17600 5.76 .0000 4.42966 9.11942 Constant| .30188 .31888 .95 .3470 .33395 .93770 -------+ -------------------------------------------------------------------nnnnn.D xx or D+xx => multiply by 10 to xx or +xx. ***, **, * ==> Significance at 1%, 5%, 10% level. Model was estimated on Jul 15, 2014 at 07:54:16 PM --------------------------------------------------------------------------------------------------------------------------------------------------------Random Effects Model: v(i,t) = e(i,t) + u(i) Estimates: Var[e] = .072043 Var[u] = .000112 Corr[v(i,t),v(i,s)] = .001559 Sum of Squares 5.55306 R squared .774934 Variances computed using no d.f. corrections. ----------------------------------------------Lagrange Multiplier Test vs. Model (3) = 2.10 [ 1 degrees of freedom, prob. value = .147776] [High values of LM favor FEM/REM over CR model] -------+ ------------------------------------------------------------------| Standard Prob. 95% Confidence PI_10| Coefficient Error z |z|>Z* Interval -------+ ------------------------------------------------------------------URB| 2.58829*** .32711 7.91 .0000 3.22942 1.94717 POP| .60556D 05*** .1709D 05 3.54 .0004 .94055D 05 .27058D 05 COMPACT| 2.10889*** .03673 57.42 .0000 2.03690 2.18087 SLOPE| 7.1 1060*** 1.08244 6.57 .0000 4.98905 9.23215 -------+ -------------------------------------------------------------------nnnnn.D xx or D+xx => multiply by 10 to xx or +xx. ***, **, * ==> Significance at 1%, 5%, 10% level. Model was estimat ed on Jul 15, 2014 at 07:54:17 PM ----------------------------------------------------------------------------AG, FWW, PDU, COMPACT , SLOPE) ----------------------------------------------------------------------------Ordinary least squa res regression ............ LHS=PI_10 Mean = 3.17199 Standard deviation = .57353

PAGE 206

206 ---------No. of observations = 76 DegFreedom Mean square Regression Sum of Squares = 20.4992 5 4.09983 Residual Sum of Squares = 4.17080 70 .05958 Total Sum of Squares = 24.6700 75 .32893 ---------Standard error of e = .24410 Root MSE .23426 Fit R squared = .83094 R bar squared .81886 Model test F[ 5, 70] = 68.80899 Prob F > F* .00000 [High values of LM favor FEM/REM over base model] Baltagi Li form of LM Statistic = 2.10770 [= BP if balanced panel] Moulton/Randolph form:SLM N[0,1] = .86721 B P test Chi squared [ 1] = 2.10770 Prob C2 > C2* = .14656 -------------------------------------------------Panel Data Analysis of PI_10 [ ONE way] Unconditional ANOVA (No regressors) Source Variation Deg. Free. Mean Square Between 16.731 3 5.5768 Residual 7.9395 72 .11027 Total 24.670 75 .32893 ------+ -------------------------------------------------------------------| Standard Prob. 95% Confidence PI_10| Coefficient Error t |t|>T* Interval -------+ ------------------------------------------------------------------AG| .03065 .98662 .03 .9753 1.93710 1.99839 FWW| .12629 1.37776 .09 .9272 2.62157 2.87415 PDU| .00130*** .00036 3.64 .0005 .00201 .00059 COMPACT| 2.05667*** .16334 12.59 .0000 1.73089 2.38244 SLOPE| 2.81979*** .76244 3.70 .0004 1.29916 4.34042 Constant| .46883 1.0853 6 .43 .6671 1.69585 2.63352 -------+ -------------------------------------------------------------------***, **, * ==> Significance at 1%, 5%, 10% level. Model was estimated on Jul 15, 2014 at 08:00:35 PM -------------------------------------------------------------------------------------------------------------------------------------------------------Random Effects Model: v(i,t) = e(i,t) + u(i) Estimates: Var[e] = .053445 Var[u] = .001434 Corr[v(i,t),v(i,s)] = .026137 Sum of Squares 4.18211 R squared .830478 Variances computed using no d.f. corrections. ----------------------------------------------Lagrange Multi plier Test vs. Model (3) = 2.11 [ 1 degrees of freedom, prob. value = .146559] [High values of LM favor FEM/REM over CR model] -------+ -------------------------------------------------------------------| Standard P rob. 95% Confidence PI_10| Coefficient Error z |z|>Z* Interval -------+ -------------------------------------------------------------------AG| .43834* .24161 1.81 .0696 .03521 .91189 FWW| .67270 .46244 1.45 .1458 .23366 1.57905 PDU| .00116*** .00014 8.31 .0000 .00143 .00089 COMPACT| 2.06843*** .18503 11.18 .0000 1.70577 2.43109 SLOPE| 3.08447*** .4396 8 7.02 .0000 2.22271 3.94624 -------+ -------------------------------------------------------------------***, **, * ==> Significance at 1%, 5%, 10% level. Model was estimated on Jul 15, 2014 at 08:00:37 PM -------------------------------------------------------------------------------------------------------------------------------------------------------Ordinary least squares regression ............ LHS=PI_10 Mean = 3.17199 Standard deviation = .57353 ---------No. of observations = 76 DegFreedom Mean square Regression Sum of Squares = 20.5578 5 4.11155 Residual Sum of Squares = 4.11220 70 .05875 Total Sum of Squares = 24.6700 75 .32893 ---------Standard error of e = .24238 Root MSE .23261 Fit R squared = .83331 R bar square d .82141 Model test F[ 5, 70] = 69.98901 Prob F > F* .00000 [High values of LM favor FEM/REM over base model] Baltagi Li form of LM Statistic = 2.10828 [= BP if balanced panel] B P test Chi squared [ 1] = 2.10828 Prob C2 > C2* = .14650 -------------------------------------------------Panel Data Analysis of PI_10 [ONE way] Unconditional ANOVA (No regressors) Source Variation Deg. Free. Mean Square Between 16.731 3 5.5768 Residual 7.9395 72 .11027

PAGE 207

207 Total 24.670 75 .32893 -------+ -------------------------------------------------------------------| Standard Pro b. 95% Confidence PI_10| Coefficient Error t |t|>T* Interval -------+ -------------------------------------------------------------------AG| .73918 1.12457 .66 .5131 1.50371 2.98206 FWW| 1.02884 1.04161 .99 .3267 3.10627 1.04859 PDU| .00142*** .00033 4.35 .0000 .00208 .00077 RR_MEAN| 3.03861*** .97223 3.13 .0026 4.97765 1.09956 CIRCRATI| 77.5778*** 7.63479 10.16 .0000 62.3506 92.8049 Constant| 3.79144*** .72113 5.26 .0000 2.35318 5.22969 -------+ -------------------------------------------------------------------***, **, * ==> Significance at 1%, 5%, 10% level. Model was estimated on J ul 15, 2014 at 08:48:02 PM --------------------------------------------------------------------------------------------------------------------------------------------------------Random Effects Model: v(i,t) = e(i,t) + u(i) Estimates: Var[e] = .053445 Var[u] = .000663 Corr[v(i,t),v(i,s)] = .012260 Sum of Squares 5.74002 R squared .767474 Variances computed using no d.f. corrections. ----------------------------------------------Lagrange Multiplier Test vs. Model (3) = 2.11 [ 1 degrees of freedom, prob. value = .146504] [High values of LM favor FEM/REM over CR model] -------+ ------------------------------------------------------------------| Standard Prob. 95% Confidence PI_10| Coefficient Error z |z|>Z* Interval -------+ -------------------------------------------------------------------AG| 5.97900 *** .48806 12.25 .0000 5.02242 6.93557 FWW| 3.98392*** .39504 10.08 .0000 3.20965 4.75819 PDU| .00022*** .8599D 04 2.60 .0092 .00006 .00039 RR_MEAN| 6.17482*** .73896 8.36 .0000 7.62316 4.72649 CIRCRATI| 106.222*** 5.35725 19.83 .0000 95.722 116.722 -------+ -------------------------------------------------------------------nnnnn.D xx or D+xx => multiply by 10 to xx or +xx. ***, **, * ==> Significance at 1%, 5%, 10% level. Model was estimated on Jul 15, 2014 at 08:48:04 PM ----------------------------------------------------------------------------

PAGE 208

208 APPENDIX F IMPACT ASSESSMENT SCENARIOS Retrospective Assessment Table F 1. Retrospective assessment of Charlie Creek basin Year Obs_ Exp_ Obs_Q Exp_Q 1992 3.62 3.75 0.14 290.83 279.91 10.92 1993 3.46 3.71 0.26 274.68 253.31 21.37 1994 3.69 3.67 0.02 295.16 296.42 1.26 1995 3.63 3.64 0.01 300.80 300.28 0.52 1996 3.81 3.60 0.21 307.46 324.36 16.91 1997 4.06 3.56 0.50 305.38 343.14 37.77 1998 3.64 3.52 0.12 310.72 321.03 10.31 1999 3.64 3.48 0.17 318.89 333.41 14.52 2000 3.63 3.44 0.19 309.72 326.82 17.11 2001 3.52 3.40 0.12 306.46 317.16 10.70 2002 3.49 3.39 0.11 332.92 343.21 10.29 2003 3.55 3.37 0.18 355.85 373.03 17.18 2004 3.39 3.36 0.03 358.87 362.21 3.34 2005 3.54 3.34 0.19 375.68 394.10 18.42 2006 3.38 3.33 0.05 356.95 362.20 5.25 2007 3.09 3.32 0.22 349.98 325.52 24.46 2008 3.25 3.30 0.06 354.44 348.64 5.80 2009 3.03 3.29 0.26 378.54 348.94 29.60 2010 3.06 3.28 0.22 402.28 376.95 25.33

PAGE 209

209 Prospective Assessment Table F 2. Prospective scenario #1 of Anclote River basin (1992 2010, 2010 2030) Y ear Compact S lope POP URB Exp_ P , mm/yr Eo/P 1/ Exp_ Q 1992 1.85 0.04 8,569 0.13 3.82 1,313 0.97 0.26 3.82 275 1993 1.85 0.04 8,943 0.13 3.82 1,266 1.01 0.26 3.82 246 1994 1.85 0.04 9,317 0.13 3.82 1,329 0.96 0.26 3.82 285 1995 1.85 0.04 9,691 0.13 3.82 1,328 0.96 0.26 3.82 286 1996 1.85 0.04 10,064 0.13 3.82 1,358 0.94 0.26 3.82 307 1997 1.85 0.04 10,438 0.13 3.82 1,383 0.92 0.26 3.82 322 1998 1.85 0.04 10,812 0.13 3.82 1,348 0.95 0.26 3.82 297 1999 1.85 0.04 11,186 0.13 3.82 1,361 0.94 0.26 3.82 305 2000 1.85 0.04 11,560 0.13 3.82 1,347 0.95 0.26 3.82 295 2001 1.85 0.04 13,735 0.13 3.82 1,328 0.97 0.26 3.82 282 2002 1.85 0.04 15,909 0.14 3.80 1,368 0.94 0.26 3.80 308 2003 1.85 0.04 18,084 0.15 3.78 1,409 0.91 0.26 3.78 337 2004 1.85 0.04 20,258 0.16 3.76 1,392 0.93 0.27 3.76 327 2005 1.85 0.04 22,433 0.16 3.74 1,434 0.90 0.27 3.74 358 2006 1.85 0.04 24,607 0.17 3.72 1,388 0.93 0.27 3.72 327 2007 1.85 0.04 26,782 0.17 3.71 1,330 0.97 0.27 3.71 290 2008 1.85 0.04 28,956 0.18 3.71 1,359 0.94 0.27 3.71 312 2009 1.85 0.04 31,131 0.18 3.70 1,357 0.94 0.27 3.70 312 2010 1.85 0.04 33,305 0.18 3.69 1,391 0.91 0.27 3.69 339 2011 1.85 0.04 33,892 0.1862 3.68 1,357 0.94 0.27 3.68 313 2012 1.85 0.04 34,479 0.1894 3.67 1,357 0.94 0.27 3.67 314 2013 1.85 0.04 35,066 0.1926 3.66 1,357 0.94 0.27 3.66 314 2014 1.85 0.04 35,653 0.1958 3.66 1,357 0.94 0.27 3.66 315 2015 1.85 0.04 36,240 0.1991 3.65 1,357 0.94 0.27 3.65 316 2016 1.85 0.04 36,977 0.2031 3.64 1,357 0.94 0.27 3.64 317 2017 1.85 0.04 37,713 0.2071 3.63 1,357 0.94 0.28 3.63 317 2018 1.85 0.04 38,450 0.2112 3.62 1,357 0.94 0.28 3.62 318 2019 1.85 0.04 39,187 0.2152 3.61 1,357 0.94 0.28 3.61 319 2020 1.85 0.04 39,924 0.2193 3.59 1,357 0.94 0.28 3.59 320 2021 1.85 0.04 40,633 0.2232 3.58 1,357 0.94 0.28 3.58 321 2022 1.85 0.04 41,343 0.2271 3.57 1,357 0.94 0.28 3.57 322 2023 1.85 0.04 42,052 0.2310 3.56 1,357 0.94 0.28 3.56 323 2024 1.85 0.04 42,762 0.2349 3.55 1,357 0.94 0.28 3.55 324 2025 1.85 0.04 43,471 0.2388 3.54 1,357 0.94 0.28 3.54 324 2026 1.85 0.04 44,141 0.2425 3.53 1,357 0.94 0.28 3.53 325 2027 1.85 0.04 44,812 0.2461 3.53 1,357 0.94 0.28 3.53 326 2028 1.85 0.04 45,482 0.2498 3.52 1,357 0.94 0.28 3.52 327 2029 1.85 0.04 46,152 0.2535 3.51 1,357 0.94 0.29 3.51 328 2030 1.85 0.04 46,822 0.2572 3.50 1,357 0.94 0.29 3.50 329

PAGE 210

210 Table F 3 . Prospective scenario #1 of Anclote River basin (2010 2030 ), and mean aridity index plus confidence interval Y ear Compact S lope POP URB Exp_ P, mm/yr Eo/P , 1/ Exp_ Q 2011 1.85 0.04 33,892 0.1862 3.68 1,357 0.96 0.27 3.68 305 2012 1.85 0.04 34,479 0.1894 3.67 1,357 0.96 0.27 3.67 306 2013 1.85 0.04 35,066 0.1926 3.66 1,357 0.96 0.27 3.66 306 2014 1.85 0.04 35,653 0.1958 3.66 1,357 0.96 0.27 3.66 307 2015 1.85 0.04 36,240 0.1991 3.65 1,357 0.96 0.27 3.65 308 2016 1.85 0.04 36,977 0.2031 3.64 1,357 0.96 0.27 3.64 309 2017 1.85 0.04 37,713 0.2071 3.63 1,357 0.96 0.28 3.63 309 2018 1.85 0.04 38,450 0.2112 3.62 1,357 0.96 0.28 3.62 310 2019 1.85 0.04 39,187 0.2152 3.61 1,357 0.96 0.28 3.61 311 2020 1.85 0.04 39,924 0.2193 3.59 1,357 0.96 0.28 3.59 312 2021 1.85 0.04 40,633 0.2232 3.58 1,357 0.96 0.28 3.58 313 2022 1.85 0.04 41,343 0.2271 3.57 1,357 0.96 0.28 3.57 314 2023 1.85 0.04 42,052 0.2310 3.56 1,357 0.96 0.28 3.56 315 2024 1.85 0.04 42,762 0.2349 3.55 1,357 0.96 0.28 3.55 316 2025 1.85 0.04 43,471 0.2388 3.54 1,357 0.96 0.28 3.54 316 2026 1.85 0.04 44,141 0.2425 3.53 1,357 0.96 0.28 3.53 317 2027 1.85 0.04 44,812 0.2461 3.53 1,357 0.96 0.28 3.53 318 2028 1.85 0.04 45,482 0.2498 3.52 1,357 0.96 0.28 3.52 319 2029 1.85 0.04 46,152 0.2535 3.51 1,357 0.96 0.29 3.51 320 2030 1.85 0.04 46,822 0.2572 3.50 1,357 0.96 0.29 3.50 321

PAGE 211

211 Table F 4. Prospective scenario #1 of Anclote River basin (2010 2030), and mean aridity index minus confidence interval Y ear Compact S lope POP URB Exp_ P, mm/yr Eo/P 1/ Exp_ Q 2011 1.85 0.04 33,892 0.1862 3.68 1,357 0.93 0.27 3.68 320 2012 1.85 0.04 34,479 0.1894 3.67 1,357 0.93 0.27 3.67 321 2013 1.85 0.04 35,066 0.1926 3.66 1,357 0.93 0.27 3.66 322 2014 1.85 0.04 35,653 0.1958 3.66 1,357 0.93 0.27 3.66 322 2015 1.85 0.04 36,240 0.1991 3.65 1,357 0.93 0.27 3.65 323 2016 1.85 0.04 36,977 0.2031 3.64 1,357 0.93 0.27 3.64 324 2017 1.85 0.04 37,713 0.2071 3.63 1,357 0.93 0.28 3.63 325 2018 1.85 0.04 38,450 0.2112 3.62 1,357 0.93 0.28 3.62 326 2019 1.85 0.04 39,187 0.2152 3.61 1,357 0.93 0.28 3.61 327 2020 1.85 0.04 39,924 0.2193 3.59 1,357 0.93 0.28 3.59 327 2021 1.85 0.04 40,633 0.2232 3.58 1,357 0.93 0.28 3.58 328 2022 1.85 0.04 41,343 0.2271 3.57 1,357 0.93 0.28 3.57 329 2023 1.85 0.04 42,052 0.2310 3.56 1,357 0.93 0.28 3.56 330 2024 1.85 0.04 42,762 0.2349 3.55 1,357 0.93 0.28 3.55 331 2025 1.85 0.04 43,471 0.2388 3.54 1,357 0.93 0.28 3.54 332 2026 1.85 0.04 44,141 0.2425 3.53 1,357 0.93 0.28 3.53 333 2027 1.85 0.04 44,812 0.2461 3.53 1,357 0.93 0.28 3.53 333 2028 1.85 0.04 45,482 0.2498 3.52 1,357 0.93 0.28 3.52 334 2029 1.85 0.04 46,152 0.2535 3.51 1,357 0.93 0.29 3.51 335 2030 1.85 0.04 46,822 0.2572 3.50 1,357 0.93 0.29 3.50 336

PAGE 212

212 Table F 5. Prospective scenario #2 of Anclote River basin (2010 2030), and mean aridity index Year Compact Slope POP URB Exp_ P, mm/yr Eo/P 1/ Exp_ Q 2011 1.85 0.04 33,892 0.1862 3.68 1,357 0.94 0.27 3.68 313 2012 1.85 0.04 34,479 0.1934 3.66 1,357 0.94 0.27 3.66 314 2013 1.85 0.04 35,066 0.2009 3.64 1,357 0.94 0.27 3.64 316 2014 1.85 0.04 35,653 0.2087 3.62 1,357 0.94 0.28 3.62 318 2015 1.85 0.04 36,240 0.2169 3.60 1,357 0.94 0.28 3.60 320 2016 1.85 0.04 36,977 0.2263 3.58 1,357 0.94 0.28 3.58 322 2017 1.85 0.04 37,713 0.2363 3.55 1,357 0.94 0.28 3.55 324 2018 1.85 0.04 38,450 0.2467 3.52 1,357 0.94 0.28 3.52 326 2019 1.85 0.04 39,187 0.2576 3.50 1,357 0.94 0.29 3.50 329 2020 1.85 0.04 39,924 0.2690 3.47 1,357 0.94 0.29 3.47 332 2021 1.85 0.04 40,633 0.2809 3.44 1,357 0.94 0.29 3.44 334 2022 1.85 0.04 41,343 0.2934 3.40 1,357 0.94 0.29 3.40 337 2023 1.85 0.04 42,052 0.3065 3.37 1,357 0.94 0.30 3.37 341 2024 1.85 0.04 42,762 0.3204 3.33 1,357 0.94 0.30 3.33 344 2025 1.85 0.04 43,471 0.3351 3.29 1,357 0.94 0.30 3.29 348 2026 1.85 0.04 44,141 0.3504 3.26 1,357 0.94 0.31 3.26 352 2027 1.85 0.04 44,812 0.3666 3.21 1,357 0.94 0.31 3.21 357 2028 1.85 0.04 45,482 0.3838 3.17 1,357 0.94 0.32 3.17 362 2029 1.85 0.04 46,152 0.4022 3.12 1,357 0.94 0.32 3.12 367 2030 1.85 0.04 46,822 0.4218 3.07 1,357 0.94 0.33 3.07 373

PAGE 213

213 Table F 6. Prospective scenario #2 of Anclote River basin (2010 2030), and mean aridity index plus confidence interval year Compact S lope POP URB Exp_ P mm/yr Eo/P 1/ Exp_ Q 2011 1.85 0.04 33,892 0.1862 3.68 1,357 0.96 0.27 3.68 305 2012 1.85 0.04 34,479 0.1934 3.66 1,357 0.96 0.27 3.66 306 2013 1.85 0.04 35,066 0.2009 3.64 1,357 0.96 0.27 3.64 308 2014 1.85 0.04 35,653 0.2087 3.62 1,357 0.96 0.28 3.62 310 2015 1.85 0.04 36,240 0.2169 3.60 1,357 0.96 0.28 3.60 312 2016 1.85 0.04 36,977 0.2263 3.58 1,357 0.96 0.28 3.58 314 2017 1.85 0.04 37,713 0.2363 3.55 1,357 0.96 0.28 3.55 316 2018 1.85 0.04 38,450 0.2467 3.52 1,357 0.96 0.28 3.52 318 2019 1.85 0.04 39,187 0.2576 3.50 1,357 0.96 0.29 3.50 321 2020 1.85 0.04 39,924 0.2690 3.47 1,357 0.96 0.29 3.47 324 2021 1.85 0.04 40,633 0.2809 3.44 1,357 0.96 0.29 3.44 327 2022 1.85 0.04 41,343 0.2934 3.40 1,357 0.96 0.29 3.40 330 2023 1.85 0.04 42,052 0.3065 3.37 1,357 0.96 0.30 3.37 333 2024 1.85 0.04 42,762 0.3204 3.33 1,357 0.96 0.30 3.33 337 2025 1.85 0.04 43,471 0.3351 3.29 1,357 0.96 0.30 3.29 341 2026 1.85 0.04 44,141 0.3504 3.26 1,357 0.96 0.31 3.26 345 2027 1.85 0.04 44,812 0.3666 3.21 1,357 0.96 0.31 3.21 349 2028 1.85 0.04 45,482 0.3838 3.17 1,357 0.96 0.32 3.17 354 2029 1.85 0.04 46,152 0.4022 3.12 1,357 0.96 0.32 3.12 360 2030 1.85 0.04 46,822 0.4218 3.07 1,357 0.96 0.33 3.07 366

PAGE 214

21 4 Table F 7. Prospective scenario #2 of Anclote River basin (2010 2030), and mean aridity index minus confidence interval Year Compact Slope POP URB Exp_ P, mm/yr Eo/P 1/ Exp_ Q 2011 1.85 0.04 33,892 0.1862 3.68 1,357 0.93 0.27 3.68 320 2012 1.85 0.04 34,479 0.1934 3.66 1,357 0.93 0.27 3.66 322 2013 1.85 0.04 35,066 0.2009 3.64 1,357 0.93 0.27 3.64 323 2014 1.85 0.04 35,653 0.2087 3.62 1,357 0.93 0.28 3.62 325 2015 1.85 0.04 36,240 0.2169 3.60 1,357 0.93 0.28 3.60 327 2016 1.85 0.04 36,977 0.2263 3.58 1,357 0.93 0.28 3.58 329 2017 1.85 0.04 37,713 0.2363 3.55 1,357 0.93 0.28 3.55 331 2018 1.85 0.04 38,450 0.2467 3.52 1,357 0.93 0.28 3.52 334 2019 1.85 0.04 39,187 0.2576 3.50 1,357 0.93 0.29 3.50 336 2020 1.85 0.04 39,924 0.2690 3.47 1,357 0.93 0.29 3.47 339 2021 1.85 0.04 40,633 0.2809 3.44 1,357 0.93 0.29 3.44 342 2022 1.85 0.04 41,343 0.2934 3.40 1,357 0.93 0.29 3.40 345 2023 1.85 0.04 42,052 0.3065 3.37 1,357 0.93 0.30 3.37 348 2024 1.85 0.04 42,762 0.3204 3.33 1,357 0.93 0.30 3.33 351 2025 1.85 0.04 43,471 0.3351 3.29 1,357 0.93 0.30 3.29 355 2026 1.85 0.04 44,141 0.3504 3.26 1,357 0.93 0.31 3.26 359 2027 1.85 0.04 44,812 0.3666 3.21 1,357 0.93 0.31 3.21 364 2028 1.85 0.04 45,482 0.3838 3.17 1,357 0.93 0.32 3.17 369 2029 1.85 0.04 46,152 0.4022 3.12 1,357 0.93 0.32 3.12 374 2030 1.85 0.04 46,822 0.4218 3.07 1,357 0.93 0.33 3.07 380

PAGE 215

215 LIST OF REFERENCES Adalberth, K. 1997. Energy use during the life cycle of buildings: A method. Building and Environment 32 (4) (7): 317 20. Alcamo, J.; Doll, P.; Heinrichs, T.; Kaspar, F.; Lehner, B.; Rosch, T.; Siebert, S. 2003a. f water use and 337. Alcamo, J.; Doll, P.; Heinrichs, T.; Kaspar, F.; Lehner, B.; Rosch, T.; Siebert, S. 2003b. Global estimates of water withdrawals and availability under current and future busines s as usual conditions. Hydrological Sciences Journal 48 (3), 339 348. political Management, Lo ndon, UK: Overseas Development Administration: 13 26. (eds) Water in the Arab World: perspectives and prognoses. Cambridge, MA: Harvard University Press: 65 100. Allan, J.A. 2003. Virtual Water: the Water, Food, and Trade Nexus: Useful Concept or Misleading Metaphor? Water International 28(1), 4 11 Allen, R.G., L.S. Pereira, D. Raes, and M. Smith. 1998. Crop evapotranspiration Guidelines for computing crop water requirements FAO Irrigation and drainage paper 56 . Rome: Food and Agriculture Organization of the United Nations. Accessed on July 21, 2014: http://www.fao.org/docrep/X0490E/x049 0e00.htm#Contents Alley, W.M.; T.E. Reilly and O.L. Franke. 1999. Sustainability of ground water resources. U.S. Geological Survey circular: 1186. 79 p. Anderson, J.R., E.E. Hardy, J.T. Roach and R.E. Witmer. 1976. A land use and land cover classification system for use with remote sensor data. Geological Survey Professional Paper 964. Retrieved from: http://landcover.usgs.gov/pdf/anderson.pdf Arnold Jr., C. L. and C.J. Gibbons. 1996. Impervious surf ace coverage. Journal of th e American Planning Association. 62(2) : 243 258. Arora, V. K. 2002. The use of the aridity index to assess climate change effect on annual runoff. Journal of Hydrology 265 (1 4) (8/30): 164 77. Arpke, A. and N. Hutzler. 2006. Dom estic water use in the United States: A life cycle approach. Journal of Indust rial Ecology, 10(1 2), 169 184.

PAGE 216

216 Bagrov, N. A. 1953. Mean long term evaporation from land surface [in Russian], Meteorol. Gidrol., 10: 20 25. Baumann, H., and A.M. Tillman. 2004. life cycle assessment methodology and application. Lund, Sweden: Studentlitteratur. Barnosky, A.D., E.A. Hadly, J. Bascompte, E.L. Berlow, J.H. Brown, M. Fortelius, and A.B. Smith. 2012. Approaching a state shift in earth's biosphere. Nature, 486(7401): 52 58. Retrieved from http://dx.doi.org/10.1038/nature11018 Bauer C, Zapp P. Generic characterisation factors for land use and water consumption. In: Dubr euil A, editor. Life cycle assessment of metals issues and research directions. Pensacola, Fl, USA: SETAC Press; 2005. Bayart, J., C. Bulle, L. Deschênes, M. Margni, S. Pfister, F. Vince, and A. Koehler. 2010. A framework for assessing off stream freshwate r use in LCA. The International Journal of Life Cycle Assessment, 15, 439 453. Berger, M., and M. Finkbeiner. 2013. Methodological challenges in volumetric and impact oriented water footprints. Journal of Industrial Ecology 17 (1): 79 89. Booth D . B . and C . R . Jackson . 1997 . Urbanization of aquatic systems: Degradation thresholds, stormwater detection, and the limits of mitigation. Journal of the American Water Resources Association . 33(5):1077 1090. Borghi, A, C. Strazza, M. Gallo, S. Messineo, and M. Naso. 2013. Water supply and sustainability: Life cycle assessment of water collection, treatment and distribution service. International Journal of Life Cycle Assessment 18 (5) (06/01): 1158 68, http:/ /dx.doi.org/10.1007/s11367 013 0549 5 . Bredehoeft, J.D. 2002. The water budget myth revisited: Why hydrogeologists model. Ground Water 40(4): 340 345. Budyko, M. I. 1958. The heat balance of the earth's surface. Washington: U.S. Dept. of Commerce, Weather Bureau. Bureau of Economic and Business Research (BEBR). 2010. Projections of Florida population by county, 2009 2035. 43 (156). Accessed on July 22, 2014: http://www.bebr.ufl.edu Chhabra, N. 2011. Developing a life cycle assessment impact indicator for water resources in the built environment [electronic resource]. Gainesville, Fla: University of Florida, http://purl.fcla.edu/fcla/etd/UFE0043484 . Choudhury, B.J . 1999. Evaluation of an empirical equation for annual evaporation using field observations and results from a biophysical model. Journal of Hydrology 216 (1 2) (3/8): 99 110.

PAGE 217

217 Chow, V.T. 1964. Handbook of applied hydrology. McGraw Hill Book Company. Cianfr ani C.M., W.C. Hession and D . M . Rizzo. 2006. Watershed imperviousness impacts on stream channel condition in southeastern Pennsylvania. Journal of the American Water Resources Association . 42(4):941 956. Cole R.J., and Kernan, P.C. 1996. Life Cycle Energy Use in Office Buildings. Building and Environment 31(4): 307 317. Cole R.J., and Rousseau D. 1992. Environmental Auditing for Building Construction: Energy and Air Pollution Indices for Building Materials. Building and Environment 27(1) : 23 30. Costa, J.E., 1987. Hydraulics and basin morphometry of the largest flash floods in the conterminous United States. J. Hydrol., 93: 313 338. Journal of Hydraulics Division. 91: 123 137. Dietrich, T. Stanton. 1978. The urbanization of Florida's population: an historical perspective of county growth, 1830 1970 . Gainesville, FL: Bureau of Economic and Business Research, University of Florida. Döll, P., F. Kaspar, and B. Lehner . 2003. A Global Hydrological Model for Deriving Water Availability Indicators: Model Tuning and Validation. Journal of Hydrology 270 (1): 105. Donohue, R. J., M. L. Roderick and T.R. McVicar. 2007. On the importance of including vegetation dynamics in Budy ko's hydrological model. Hydrol . Earth Syst. Sci. 11: 983 995. Dresen, Boris, and Michael Jandewerth. 2012. Integration of spatial analyses into LCA calculating GHG emissions with geoinformation systems. International Journal of Life Cycle Assessm ent 17 (9 ) (11/01): 1094 103. http://dx.doi.org/10.1007/s11367 011 0378 3 . Drummond, M.A. and T.R. Loveland. 2010. Land use pressure and a transition to forest cover loss in the eastern United States. Bios cience 60(4): 286 298. El Sayed M., M. Mahgoub, N. P. van der Steen, K. Abu Zeid and K. Vairavamoorthy. 2010. Towards sustainability in urban water: A life cycle analysis of the urban water system of Alexandria city, Egypt. Journal of Cleaner P roduction 18 (10 11): 1100 1106. Falcone, J. 2011. GAGES II: Geospatial Attributes of Gages for Evaluating Streamflow. U.S. Geological Survey. Accessed online: http://water.usgs.gov/lookup/getspat ial?gagesII_Sept2011

PAGE 218

218 FDEP (Florida Department of Environmental Protection). 2006a. Pasco County master reuse system wet weather performance assessment. Prepared by King Engineering Associates for Pasco County Utilities Department in 2004. Accessed on July 23, 2014: http://depedms.dep.state.fl.us/Oculus/servlet/login FDEP (Florida Department of Environmental Protection). 2006b. State of Florida Industrial wastewater facility permit FL0435490 . Accessed on July 23, 2014: http://depedms.dep.state.fl.us/Oculus/servlet/login FDEP (Florida Department of Environmental Protection). 2011. Disposal rating evaluation Hudson rapid rate i nfiltration basin site Pasco County, Florida, for Pasco County utilities division. FDEP Permit No. FLA12735 001 DWP . Prepared by Qore Property Sciences for Pasco County Utilities Division in 2007. Accessed on July 23, 2014: http://depedms.dep.state.fl.us/Oculus/servlet/login FDOT (Florida Department of Transportation). 1999. Florida land use, cover and forms classification system. Department of Transportation Surveying and Mapping, Geograp hic Mapping Section. Retrieved from: http://www.dot.state.fl.us/surveyingandmapping/documentsandpubs/fluccmanual 1999.pdf Finkenbine J . K . , J . W . Atwater and D.S. Mavinic . 2000 . Stream health after urbanization. Journal of the American Water Resources Association . 36(5):1149 1160. Finnveden, G., M.Z. Hauschild, T. Ekvall, J. Guinée, R. Heijungs, S. Hellweg, A. Koehler, D. Pennington, and S. Suh. 2009. Recent de velopments in life cycle assessment. Journal of Environmental Management 91 (1) (10): 1 21. Foley, J.A., R. DeFries, G. P. Asner, C. Barford, G. Bonan, S. R. Carpenter and P.K. Snyder. 2005. Global consequences of land Use . Science. 309 (5734): 570 574. Fr y, J., G. Xian, S. Jin, J. Dewitz, C. Homer, L. Yang, C. Barnes, N. Herold and J. Wickham. 2011 . Completion of the 2006 National Land Cover Database for the Conterminous United States, PE&RS, 77(9):858 864. Fry, J.A., Coan, M.J., Homer, C.G., Meyer, D.K., and Wickham, J.D. 2009. Completion of the National Land Cover Database (NLCD) 1992 2001 Land Cover Change Retrofit product: U.S. Geological Survey Open File Report 2008 1379, 18. Fu, B.P. 1981. On the calculation of evaporation from land surface (in Chines e). Sci. Atmos. Sin. 5, 23 31. Geolytics. 2013. Time Series Research Package. Haase, D. and H. Nuissl. 2007. Does urban sprawl drive changes in the water balance and policy?: The case of Leipzig (Germany) 1870 2003. Landscape and Urban Planning, 80(1 2) : 1 13.

PAGE 219

219 Hargreaves, G. H. and Z. A. Samani. 1982. Estimating potential evapotranspiratio n. J. Irrig. Drain. Eng. 108 (3 ): 223 230. Hendrickson, C. T., Horvath, A. (2000). Resource Use and Environmental Emissions of U.S. Construction Sectors. Journal of Constr uction Engineering and Management, 126 (1) : 38 44. Heuvelmans, G., B. Muys, and J. Feyen. 2005 a . Extending the life cycle methodology to cover impacts of land use systems on the water balance. International Journal of Life Cy cle Assessment, 10(2): 113 119. Heuvelmans, G., J. F. Garcia Qujano, B. Muys, J. Feyen, and P. Coppin. 2005 b . Modelling the water balance with SWAT as part of the land use impact evaluation in a life cycle study of CO2 emission reduction scenarios. Hydrologi cal Processes, 19(3): 729 748 . Hoekstra, A. Y., A.K. Chapagain, M.M. Al daya, and M.M. Mekonnen. 2011 . The Water Footprint Assessment Manual. Earthscan. 224 p. Hoekstra, A. Y., M. M. Mekonnen, A. K. Chapagain, R. E. Mathews and B. D. Richter. 2012. Global monthly water scarcity: Blue water footprints versus blue water availability. PLoS ONE, 7(2) : e32688. Retrieved from http://dx.doi.org/10.1371%2Fjournal.pone.0032688 Homer, C., J. Dewitz, J. Fry, M. Coan, N. Hossain, C. Larson, N. Herold, A. McKerrow, J. N. VanDriel, and J. Wickham. 2007. Completion of the 2001 National Land Cover Database for the Conterminous United States. Photogrammetric Engineering and Remote Sensing, 73(4): 337 341. Horvath, A., and C. Hendrickson. 1 998. Steel versus steel reinforced concrete bridges: Environmental assessment. Journal of Infrastructure Systems 4 (3) (09/01; 2014/06): 111 7 . Humbert, S. 2013. ISO 14046: water footprint. Summary of the project. Presentation (Accessed on July 2, 2014: http://wulca waterlca.org/pdf/ISO14046_Sebastien_humbert_SETAC2013.pdf Huston, S., N.L. Barber, J.F. Kenny, K.S. Linsey, D.L. Lumia, M.A. Maupin. 2005. Estimated Use of Wat er in the United States in 2000. United States Geologica l Survey. http://pubs.usgs.gov/circ/2004/circ1268/ International Organization for Standardization (ISO). 1997. ISO: 14040 Environmental management Life cycle assessment Principles and framework. Istanbulluoglu, E . , T . Wang, O . M. Wright and J . D. Lenters. 2012. Interpretation of hydrologic trends from a water balance perspective: The role of groundwater storage in the B udyko hypothesis. Water Resources Research 48 (3) : W00H16.

PAGE 220

220 Jacobs, J., J. Mecikalski, and S. Paech. 2008. Satellite based solar radiation, net radiation, and potential and reference evapotranspiration estimates over Florida. Technical report submitted to the USGS. Retrieved from http://hdwp.er.usgs.gov/ET/GOES_FinalReport.pdf Jensen, M. E. , R. D. Burman and R. G. Allen. 1990. Evapotranspiration and irrigation water requirements. ASCE manuals and reports on engin eering practice no. 70, ASCE, New York. Jin, S., L. Yang, P. Danielson, C. Homer, J. Fry, and G. Xian. 2013. A comprehensive change detection method for updating the National Land Cover Database to circa 2011. Remote Sensing of Environment 132: 159 175. Kennedy, C., J. Cuddihy and J. Engel Yan. 2007 . The changing metabolism of cities . Journal of Industrial Ecology. 11(2) : 43 59. Kenny, J.F., N.L. Barber, S. Hutson, K.S. Linsey, K.K. Lovelace, M.A. Maupin. 2009. Estimated Use of Water in the United States in 2005. United St ates Geological Survey. http://pubs.er.usgs.gov/publication/cir1344 Kenway, S., A. Gregory and J. McMahon. 2011. Urban water mass balance analysis. Journal of Industrial Ecology . 15(5) : 693 706. (1 3) (8): 1 23. Kounina, A., Margni, M., Bayart, J., Boulay, A., Berger, M., Bulle, C . and S. Humbert . 2013 . Review of methods addressing freshwater use in life cycle inventory and impact assessment. International Journal of Life Cycle Assessment 18(3), 707 721 . K undzewicz , Z . W. and P. D öll . 2009. Will groundwater ease freshwater stress under climate change? Hydrological Sciences Journal . 54 (4) : 665 75. http://dx.doi.org/10.1623/hysj.54.4.665 . Levin, R.B., P.R. Epstein, T.E. Ford, H. Winston, E. Olson, and E.G. Reichard. 2002. U.S. Drink ing Water Challenges in the Twenty First Century. Environmental He alth Perspectives 110 : 43 52. Lindfors, L. G., Christiansen, K., Hoffman, L., Virtanen, Y., Jun tilla, V., Hanssen, O. J., Rønning, A ., Ekvall, T. and Finnveden, G. 1995. Nordic Guidelines for Life Cycle Assessment. Nordic Council of Ministers, Copenhagen, Denmark, Nord 1995: 20 . Livingston, E. H. and E. McCarron . 1992. Stormwater management: A guide for floridians. Tallahassee, Fla: Florida Dept. of Environmental Regulation. Loveland, T.R., T.L. Sohl, S.V. Stehman, A.L. Gallant, K.L. Sayler and D.E. Napton . 2002 . A Strategy for Estimating the Rates of Recent United States Land Cover

PAGE 221

221 Changes. Photogrammetric Engineering and Remote Sensing, 68 (1 0 ) : 1091 1099. Lu, J., G. Sun, S.G. McNulty, D.M. Amatya. 2005. A comparison of six potential evapotranspiration me thods for regional use in the southeastern United States. JAWRA Journal of the American Water Resour ces Association 41(3) : 621 633. Lundie, S., G.M. Peters, and P.C. Beavis. 2004. Life cycle assessment for sustainable metropolitan water systems planning. E nvironmental Science & Technology 38(13) : 3465 3473. Lundin, M. and G. M. Morrison. 2002. A life cycle assessment based procedure for development of environmental sustainability indicators for urban water syst ems. Urban Water 4(2) : 145 152. Magnusson, W.E. 2001. Catchments as basic units of management in conservation biology courses. Conservation biology 15(5): 1464 1465. Mahdavi, A., and R. Ries. 1998. Towards computational eco analysis of building designs. Computers & Structures 67 (5) (6/1): 375 87. Mare lla, R.L. 2008. Water Use in Florida, 2005 and Trends 1950 2005. USGS Fact sheet 2008 3080. Accessed on June 28, 2014: http://pubs.usgs.gov/fs/2008/3080/#Water_Wit hdrawal_Trends,%201950 2005 Marella, R.L. 2014. Water withdrawals, use, and trends in Florida, 2010: USGS Scientific Investigations Report 2014 5088, 59. http://dx.doi.org/10.3133/sir20145088 . Martinez, C., and M. Thepadia. 2010. Estimating reference evapotranspiration with minimum data in florida. Journal of Irrigation and Drainage Engineering 136 (7) (07/01; 2014/06): 494 501, http://dx.doi.org/10.1061/(ASCE)IR.1943 4774.0000214 Maurer, E.P., A.W. Wood, J.C. Adam, D.P. Lettenmaier, and B. Nijssen. 2002. A Long Term Hydrologically Based Data Set of Land Surface Fluxes and States for the Conterminous United States, J. Climate 15 : 3237 3251. Mecikalski, J.R., D.M. Sumner, J.M. Jacobs, C.S. Pathak, S.J. Paech and E.M. Douglas. 2011. Use of Visible Geostationary Operational Meteorological Satellite Imagery in Mapping Reference and Potential Evapotranspiration over Florida. In Evapot ranspiration, edited by Prof. L. Labedzki. InTech, DOI: 10.5772/14478. Available from: h ttp://www.intechopen.com/books/evapotranspiration/use of visible geostationary operational meteorological satellite imagery in mapping reference and p Mezentsev, V. S. 1955. More on the calculation of average total evaporation [in Russian], Meteorol. Gidro l., 5 : 24 26.

PAGE 222

222 Milà i Canals, L., A. Dubreuil, G. Gaillard and R. Müller Wenk. 2007. Key elements in a framework for land use impact assessment within LCA. International Journal of Life Cycle Assessment 12(1): 5 15. Milà i Canals, L., J. Chenoweth, A. Chapa gain, S. O rr, A. Antón, and R. Clift. 2009. Assessing freshwater use impacts in LCA: Part I inventory modelling and characterisation factors for the main impact pathways. The International Journal of Life C ycle Assessment, 14 (1) : 28 42. Milly, P. C. D. 19 94. Climate, soil water storage, and the average annual water balance. Water Resources Research 30 (7): 2143 56. Milly, P.C.D., J. Betancourt, M. Falkenmark, R.M. Hirsch, Z.W. Kundzewicz, D.P. Lettenmaier, R.J. Stouffer. 2008. Stationarity is dead: whither water management? Science, 319: 573 574 . Milly, P. C. D., and K. A. Dunne. 2002. Macroscale water fluxes: 1.Quantifying errors in the estimation of basin mean precipitation, Water Resour. Res. 38(10): 1205. Moglen, G. E. and S. Kim . 2007. Limiting impervio usness. Journal of the American Planni ng Association. 73(2): 161 171. Motoshita, M., N. Itsubo and A. Inaba. 2011. Development of impact factors on damage to health by infectious diseases caused by domestic water scarcity. International Journal of Life Cyc le Assessment 16 (1): 65 73. Napton, D.E., R.F. Auch, R. Headley, and J.L. Taylor. 2010. Land changes and their driving forces in the Southeastern United States. Region. Environ. Change 10(1):37 53. OECD. 2008. OECD environmental outlook to 2030. Organisat ion for Economi c Co operation and Development. OECD. 2012. OECD environmental outlook to 2050 Organisation for Economi c Co operation and Development. Observ. University of Tartu, 4 : 200. Omernik, J.M., 1987. Ecoregions of the conterminous United States, Annals of the Association of American Geographers, 77:118 125. Omernik, J.M. 2003. The Misuse of Hydrologic Unit Maps for Extrapolation, Reporting, and Ecosystem Manageme nt. J. of the American Water Resources Association. 39(3): 563 573. Omernik, J.M., R.M. Hughes, G.E. Griffith and G.M. Hellyer. 2011. Common geographic frameworks. In U.S. EPA (Environmental Protection Agency) 2011. Landscape

PAGE 223

223 and predictive tools: a guide to spatial analysis for environmental assessment. Risk Assessment Forum. Washington D.C.: EPA/100/R 11/002. Ortiz O., F. Castells and G. Sonnemann. 2007. Important issues in LCA and Ecodesign within the Building Sector for Developing Countries. Paper prese nted at international conference on life cycle assessment CILCA 2007 Sao Paulo, Brazil. Oudin, L., V. Andréassian, J. Lerat, and Claude Michel. 2008. Has land cover a significant impact on mean annual streamflow? An international assessment using 1508 catchments. Journal of Hydrology 357 (3 4) (8/15): 303 16. Owens, J. W. 2001. Water resources in life cycle impact assessment: Considerations in choosing category indicators. Journal of Industrial Ecology 5(2): 37 54. Patton, P.C. and V.R. Baker. 1976. Mor phometry and floods in small drainage basins subject to diverse hydrogeomorphic controls. Water Resour. Res. 12(5): 941 952. Pennington, D. W., J. Potting, G. Finnveden, E. Lindeijer, O. Jolliet, T. Rydberg, and G. Rebitzer. 2004. Life cycle assessment par t 2: Current impact assessment practice. Environment International 30 (5) (7): 721 39. Pfister, S., A. Koehler and S. Hellweg. 2009. Assessing the environmental impacts of freshwater consumption in LCA. Environmental Science & Technology 43(11): 40 98 4104. Pfister, S., P. Bayer, A. Koehler and S. Hellweg. 2011. Environmental impacts of water use in global crop production: hotspots and trade offs with land use Environmental Science & Technology 45 (13): 5761 5768. Pike, J.G. 1964. The estimation of annual ru n off from meteorological data in a tropical climate. J. Hydrol. 2: 116 123. Ponce, V., R. Pandey, and S. Ercan. 2000. Characterization of drought across climatic spectrum. Journal of Hydrologic Engineering 5 (2) (04/01; 2013/12): 222 4, http://dx.doi.org/10.1061/(ASCE)1084 0699(2000)5:2(222) . Potere, D., A. Schneider, S. Angel and D.L. Civco. 2009. Mapping urban areas on a global scale: Which of the eight maps now available is more accurate? Internat ional Journal of Remote Sensing. 30(24) : 6531 6558. Price, C.V., N. Nakagaki, K.J. Hitt, and R.M. Clawges. 2006. Enhanced Historical Land Use and Land Cover Data Sets of the U.S. Geological Survey, U.S. Geological Survey Digital Data Seri es 240. Retrieved from http://pubs.usgs.gov/ds/2006/240

PAGE 224

224 Priestley, C.H.B. and R.J. Taylor. 1972. On the assessment of surface heat flux and evaporation using large scale parameters. Monthl y Weather Review 10 0(2): 81 92. Randolph, J. 2004. Environmental land use planning and management. Washington: Island Press. Richter, B.D. 2009. Re thinking environmental flows: from allocations and reserves to sustainability boundaries. River. Res. Applic. 26(8): 1052 1063. Richter, B.D., M.M. Davis, C. Apse and C. Konrad. 2012. Short communication: A presumptive standard for environmental flow protection. River. Res. Applic. 28: 1312 1321. Ridoutt, B.and S.A. Pfister. 2012. A new water footprint calculation method integrati ng consumptive and degradative water use into a single stand alone weighted indicator. International Journal of Life Cycle Assessment 18(1 ): 204 207. Ries, R., and A. Mahdavi. 2001. Integrated computational life cycle assessment of buildings. Journal of Co mputing in Civil Engineering 15(1) (01/01; 2014/06): 59 66. Roderick, M.L. and G.D. Farquhar. 2011. A simple framework for relating variations in runoff to variations in climatic conditions and catchment properties. W ater Resources Research 47(12). Saad, R ., M. Margni, T. Koellner, B. Wittstock and L. Deschênes. 2011. Assessment of land use impacts on soil ecological functions: Development of spatially differentiated characterization factors within a canadian context . The International Journal of Life Cycle Assessment 1 6 : 198 211 . Sayrs, L.W. 1989. Pooled time series analysis. Iowa: Sage Publications. Scalenghe, R., and F.A. Marsan . 2009. The anthropogenic sealing of soils in urban area s. Landscape and Urban Planning. 90(1) : 1 10. Retrieved from http://www.sciencedirect.com/science/article/pii/S0169204608001710 . Schneider, A., M . A . Friedl and D.Potere. 2009. A new map of global urban extent from MODIS satellite data. Environmental Research Letters . 4(4) : 044003. Retrieved from http://stacks.iop.org/1748 9326/4/i=4/a=044003 SETAC (Society of Environmental Toxicology and Chemistry) 1992. Life Cycle Assessment inventory, Classification, Valuation, Data Bases. Workshop report, Leiden 1992 . Schreiber, P. 1904. Über die Beziehungen zwischen dem Niederschlag und der Wasserfu¨hrung der Flu¨ße in Mitteleuropa. Z. Meteorol. 21: 441 452.

PAGE 225

225 Sleeter, B.M., T.S. Wilson, C .E. Soulard, and J. Liu. 2011. Estimation of late twentieth century land cover change in California. Environ. Monit. Assess. 173(1 4): 251 266. Sophocleous, M. 2000. From safe yield to sustainable development of water resources: The Kansas experience. Jour nal of Hydrology 235 : 27 43. Splinter, D. K., D.C. Dauwalter, R.A. Marston and W.L. Fisher. 2011. Watershed Morphology of Highland and Mountain Ecoregions in Eastern Oklahoma. The Professional Geographer. 63(1): 131 143. Stein, R.G. 1977. Energy Cost of Bu ilding Construction. Energy and Buildings 1: 27 29 Stepenuck K . F . , R.L. Crunkilton and L . Wang . 2002 . Impacts of urban landuse on macroinvertebrate communities in southeastern Wisconsin streams. Journal of the American Water Resources Association . 38(4): 1041 1051. St ewart, M. and B.P. Weidema. 2005 . A Consistent Framework for Assessing the Impacts from Resource Use A Focus on Resource Functionality. The International Journal of Life Cycle Assessment 10 (4): 240 247. Theis, C.V. 1940 . The source of water derived from wells. Civil Engineering 10(5): 277 280. Thepadia, M., and C. Martinez. 2012. Regional calibration of solar radiation and reference evapotranspiration estimates with minimal data in Florida. Journal of Irrigation and Drain age Engineering 138 (2) (02/01; 2014/06): 111 9, http://dx.doi.org/10.1061/(ASCE)IR.1943 4774.0000394 Turc, L. ion et : 491 569. UN (United Nations). 1997. Comprehensive Assessment of Freshwater Resources of the World. The United Nations Report on Sustainable Development, accessed October 17, 2011, http://daccess dds ny.un.org/doc/UNDOC/GEN/N97/003/65/IMG/N9700365.pdf?OpenElement UN (United Nations). 2012. World Urbanization Prospects: The 2011 Revision. Population Division of the Depart¬men t of Economic and Social Affairs of the United Nations Secre tariat. http://esa.un.org/unpd/wup/unup/p2k0data.asp US Census. 2011. Current estimates data. Retrieved on July 22 nd , 2014: http://www.census.gov/popest/data/index.html USEIA (United States Energy Information Administration). 2012. Annual Energy Review 2011. Energy Information Administration. DOE/EIA 0384.

PAGE 226

226 USEPA ( U.S. E nvironmental Protection Agency ). 199 3. Guidance specifying management measures for sources of nonpoint pollution in coastal waters. 840 B 92 002. USEPA ( U.S. Environmental Protection Agency ). 1999 . Level III ecoregions of the continental United States (revision of Omernik 1987). U.S. Environmental Protection Agency, National Health and Environmental Effects Research Laboratory, Western Ecology Division, Corvallis, Oregon. USGS (United States Geologi cal Survey). 2014 a . National Water Use Information Program. Site Inventory for the Nation. http://waterdata.usgs.gov/nwis/inventory?search_c riteria=search_site_no&submit ted_form=introduction (accessed June 30 th , 2014). USGS (United States Geological Survey). 2014 b . Historical Water Use in Florida. Individual Counties 1965 2000. http://fl.water.usgs.gov/infodata/wateruse/historical.html (accessed June 30th, 2014). Van Ek, R.; Lindeijer, E.; van Oers, L.; van der Voet, E.; Witte, J.P. (2002) TNO report 42/02.002943 Towards including desiccation in LCA. TNO Industrial Techn ology. Accessed on July 5th, 2012: www.repository.tudelft.nl/.../dc_lindeijer_2002.pdf Van Zelm, R., A.M. Schipper, M. Rombouts, J. Snepvangers and M.A.J. Huijbregts. 2011. Implemen ting groundwater extraction in life cycle impact assessment: characterization factors based on plant species richness for the Netherlands. Environ. Sci. Technol. 45 (2): 629 635. Vörösmarty, C.J., P .Green, J. Salisbury and R.B. Lammers. 2000. Global water resources: vul nerability from climate change and population growth. Science 289: 284 288. Walsh C . J . , A . H . Roy, J . W . Feminella, P.D. Cottingham , P.M. Groffman and R.P. Morgan II. 2005 . The urban stream syndrome: Current knowledge and the search for a cur e. Journal of the North American Benthological Society . 24(3): 706 723. Wiken, E. , F . J. Nava and G . Griffith. 2011. North American Terrestrial Ecoregions Level III. Commission for Environmental Cooperation, Montreal, Canada. http://www.epa.gov/wed/pages/ecoregions/na_eco.htm#Downloads (accessed July 12 th , 2014). WWAP. 2012. The United Nations World Water Development Report 4: Managing Water under U ncertainty and Risk. Paris, UNESCO. Yang, G., L.C. Bowling, K.A. Cherkauer and B.C. Pijanowski. 2011. The impact of urban development on hydrologic regime from catchment to basin scales. Landscape and U rban Planning. 103(2): 237 247.

PAGE 227

227 Yang, H., D. Yang, Z. Lei and F. Sun. 2008. New analytical derivation of the mean annual water energy balance equation. Water Resour. Res. 44 : W03410. Zhang, L., W.R. Dawes and G.R. Walker. 2001. Response of mean annual evapotranspiration to vegetation changes at catchment scal e. Water Resources Research 37(3): 701 708. Zhang, L., K. Hickel, W. R. Dawes, F. H. S. Chiew, A. W. Western, and P. R. Briggs. 2004. A rational function approach for estimating mean annual evapotranspiration. Water Resources R esearch 40 (2) (02/05): W0250 2. Zhang, L., N. Potter, K. Hickel, Y. Zhang and Q. Shao. 2008. Water balance modeling over variable time scales based on the budyko framework model development and testing. Journal of Hydrology 360(1 4): 117 131.

PAGE 228

228 B IOGRAPHICAL SKETCH Originally from Venezuela, Rodrigo started in 1998 his career as an architect in a construction management firm house design department . In 2000 his desire to better coordinate the building process led him to pursue a MSc. in Building Construction at the M.E. Rinker Sr., School of Building Construction at the University of Florida through the Fulbright program . 2003 gave Rodrigo the oppor tunity to study at Birkbeck College (London) a post graduate certificate in Economics and Finance. From 2003 until 2007, Rodrigo served as an assistant professor at the Universidad Simón Bolívar in Venezuela in the professional Architecture program while a t the same time sharing partnership in a local design and construction firm. In 2010, Rodrigo returned to academia to pursue his interest in evaluating building performance through life cycle assessment , and seeks to understand the impact of the built envi ronment on freshwater resources. His next steps involve moving to Denmark and marry ing his fiancée.