<%BANNER%>

Potential for remote sensing to locate the ordinary high water line in Florida: a case study of Lakes Hatchineha and Kis...

University of Florida Institutional Repository

PAGE 1

POTENTIAL FOR REMOTE SENSING TO LOCATE THE ORDINARY HIGH WATER LINE IN FLORIDA: A CASE STUDY OF LAKES HATCHINEHA AND KISSIMMEE By JANET C. NUNLEY A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2002

PAGE 2

Copyright 2002 by Janet C. Nunley

PAGE 3

ACKNOWLEDGMENTS I give special thanks to Scot Smith, Bon Dewitt, and Levent Genc who spent endless nights getting the early images and thoughts in order. Special thanks also go to Mark Clark for his patience and endurance. Many short meetings turned into marathon events and I learned a lot. I would like to thank Ben Heilman, John Clyatt, Robby Bickel and Hannibal Snipes for their assistance with the fieldwork. Without them, this work could never have been completed. Thanks go to the Florida Department of Environmental Protection for funding this research. Finally, many thanks to my husband, Bryant, for his patience and support. Without him, I never could have done it! iii

PAGE 4

TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iii LIST OF TABLES.............................................................................................................vi LIST OF FIGURES..........................................................................................................vii ABSTRACT.......................................................................................................................ix CHAPTER 1 INTRODUCTION...........................................................................................................1 Early Cases......................................................................................................................2 Cornerstone Cases...........................................................................................................3 A Recent Case.................................................................................................................5 Debate Rages On.............................................................................................................5 Opposing Sides...............................................................................................................7 Legislative Action...........................................................................................................9 How It All Fits Together...............................................................................................11 Site Description.............................................................................................................13 2 LITERATURE REVIEW..............................................................................................15 3 MATERIALS AND METHODS..................................................................................31 Imagery.........................................................................................................................31 Landsat...................................................................................................................31 Indian Remote Sensing Satellite............................................................................33 The SPOT Satellite.................................................................................................33 The IKONOS Satellite...........................................................................................34 Compact Airborne Spectrographic Imager............................................................34 Image Processing..........................................................................................................35 Visual Interpretation.....................................................................................................37 Vegetation Sampling.....................................................................................................38 Vegetation Analysis......................................................................................................40 iv

PAGE 5

4 RESULTS......................................................................................................................42 Imagery Results............................................................................................................42 Landsat Images......................................................................................................42 Indian Remote Sensing Images..............................................................................42 The SPOT Images..................................................................................................47 The IKONOS Images.............................................................................................47 The CASI Images...................................................................................................47 Vegetation Results........................................................................................................58 5 DISCUSSION................................................................................................................64 Remote Sensing Aspects...............................................................................................64 Vegetation Aspects.......................................................................................................71 Conclusions...................................................................................................................73 APPENDIX A VEGETATION CHARTS.............................................................................................75 B VEGETATION TABLES..............................................................................................93 LIST OF REFERENCES.................................................................................................108 BIOGRAPHICAL SKETCH...........................................................................................112 v

PAGE 6

LIST OF TABLES Table page 4-1 Summary table of imagery results.........................................................................58 4-2 Twenty-three most frequently occurring species on the banks of Lake Hatchineha and Lake Kissimmee, combined.........................................................59 4-3 Average minimum distance from the OHWL that the closest species were found for Lake Hatchineha and Lake Kissimmee, combined................................60 4-4 Ten species with the highest usability index for combined lakes and land use types.......................................................................................................................61 4-5 Twelve species with highest usability index for all land use types of Lake Hatchineha only.....................................................................................................61 4-6 Ten species with highest usability index for all land use types of Lake Kissimmee only.....................................................................................................62 4-7 Ten species with highest usability index for ranch areas of combined lakes........62 4-8 Ten species with highest usability index for natural areas of combined lakes......62 4-9 Ten species with highest usability index for urban areas of combined lakes........63 B-1 List of species ordered by frequency of occurrence for Lake Hatchineha and Lake Kissimmee combined....................................................................................94 B-2 List of species ordered by frequency of occurrence for Lake Hatchineha............97 B-3 List of species ordered by frequency of occurrence for Lake Kissimmee.............99 B-4 Average minimum distance from the OHWL for Lake Hatchineha and Lake Kissimmee, combined..........................................................................................101 B-5 Average minimum distance from the OHWL for Lake Hatchineha....................104 B-6 Average minimum distance from the OHWL for Lake Kissimmee....................106 vi

PAGE 7

LIST OF FIGURES Figure page 1-2 Average monthly stage data for Lake Kissimmee covering a 7-year period.........14 3-1 Different ground cell sizes of imagery used in this study......................................35 3-3 Highest elevation along a transect.........................................................................40 3-4 Idealized vegetation transect chart and terminology used in vegetation analysis...................................................................................................................41 4-1 Transect 31, classified Landsat image of Lake Hatchineha...................................43 4-2 Transect 9, classified Landsat image of Lake Kissimmee.....................................44 4-3 Transect 33, classified IRS image of Lake Hatchineha.........................................45 4-4 Transect 9, classified IRS image of Lake Kissimmee...........................................46 4-5 Transect 31, classified SPOT image of Lake Hatchineha......................................48 4-6 Transect 9, classified SPOT image of Lake Kissimmee........................................49 4-7 Transect 31, classified IKONOS image of Lake Hatchineha................................50 4-8 Transect 4, classified IKONOS image of Lake Kissimmee..................................51 4-9 Transect 7, classified IKONOS image of Lake Kissimmee..................................52 4-10 Histogram of FACU, FAC, FACW or OBL species classified in each CASI hyperspectral class.................................................................................................53 4-11 Histogram of species classified in each CASI hyperspectral class........................54 4-12 Transect 31, classified CASI image of Lake Hatchineha......................................55 4-13 Transect 9, classified CASI image of a portion of Lake Kissimmee.....................56 4-14 Vegetation in the area of Transect 33. Note watermarks on cypress trees...........57 5-1 Catfish Creek entrance to Lake Hatchineha. Note water within the cypress........70 vii

PAGE 8

A-1 Vegetation Transect 1, ranch land use, Lake Kissimmee......................................76 A-2 Vegetation Transect 2, ranch land use, Lake Kissimmee......................................77 A-3 Vegetation Transect 3A, natural area land use, Lake Kissimmee.........................78 A-4 Vegetation Transect 4, ranch land use, Lake Kissimmee......................................79 A-5 Vegetation Transect 5A, ranch land use, Lake Kissimmee...................................80 A-6 Vegetation Transect 6, natural area land use, Lake Kissimmee............................81 A-7 Vegetation Transect 7, urban land use, Lake Kissimmee......................................82 A-8 Vegetation Transect 8, urban land use, Lake Kissimmee......................................83 A-9 Vegetation Transect 9, ranch land use, Lake Kissimmee......................................84 A-10 Vegetation Transect 10A, natural area land use, Lake Kissimmee.......................85 A-11 Vegetation Transect 31, natural area land use, Lake Hatchineha..........................86 A-12 Vegetation Transect 32, natural area land use, Lake Hatchineha..........................87 A-13 Vegetation Transect 33, natural area land use, Lake Hatchineha..........................88 A-14 Vegetation Transect 34, natural area land use, Lake Hatchineha..........................89 A-15 Vegetation Transect 36, ranch land use, Lake Hatchineha....................................90 A-16 Vegetation Transect 37, natural area land use, Lake Hatchineha..........................91 A-17 Vegetation Transect 38, urban land use, Lake Hatchineha....................................92 viii

PAGE 9

Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science POTENTIAL FOR REMOTE SENSING TO LOCATE THE ORDINARY HIGH WATER LINE IN FLORIDA: A CASE STUDY OF LAKES HATCHINEHA AND KISSIMMEE By Janet C. Nunley December 2002 Chair: Scot E. Smith Major Department: College of Natural Resources and Environment Locating the ordinary high water line (OHWL) of Floridas freshwater bodies has been a contentious issue for many decades. At the time of statehood in 1845, all property below the OHWL was designated as State sovereignty land, to be held in trust for public use. Legal battles have raged over the boundary line dividing sovereignty land and private ownership. Florida has approximately 7800 lakes and numerous rivers. Ordinary high water determinations are made only on a case-by-case basis, as it would be prohibitively expensive to locate the OHWL on each of Floridas freshwater bodies. Case law requires that the best methods available' be used to locate the OHWL. Historically, OHWL determinations have been made by interpreting aerial photography and by locating other indicators such as watermarks, debris lines, lichen lines, and vegetation indicators. In an effort to calculate the area of Floridas sovereignty ix

PAGE 10

lands, the State legislature appropriated funds to determine the usefulness of remote sensing for such a task. This study used Landsat, SPOT, Indian Remote Sensing (IRS), and IKONOS satellite imagery as well as CASI airborne hyperspectral images to assess the ability of remote sensing to detect water and to detect a vegetation change or shift corresponding to the OHWL. Open water was detectable in all the images, but was not detectable in the presence of dense emergent vegetation. The relatively large ground cell size of Landsat, SPOT, and IRS severely limited their ability to detect vegetation shifts corresponding to the OHWL within a desired 1 m margin of error. IKONOS was better than the other sensors at detecting vegetation shifts near the OHWL; however, it was only successful 14% of the time. When shifts were detected, the 4 m ground cell size of IKONOS did not meet the desired 1 m margin of error when determining a boundary line. With its 1 m ground cell size, CASI had the best potential as a method for locating the OHWL. Vegetation shifts corresponding to the OHWL were detected in three of the four transects located on the CASI images. However, two were coincidentally located at the edge of cypress. Further study is needed to determine if CASI can consistently detect changes at the OHWL. Vegetation analysis of the study site showed no consistent useful vegetation indicators of the OHWL. This may be the result of an extended antecedent drought at the time of this study. x

PAGE 11

CHAPTER 1 INTRODUCTION The ordinary high water line (OHWL) has been a contentious issue in Florida for more than a century. When Florida became a state in 1845, all lands beneath navigable waters were conveyed to the state for public use as state sovereignty lands. Navigability does not require the presence of water at all times, but does require the water body to be navigable during its ordinary state (Reimer 2001). According to the Department of Environmental Protection (DEP), Division of Surveying and Mapping (Wilkinson and Warner 2002 Section 5 p. 2), OHWL means The boundary between sovereignty lands owned by the State of Florida by virtue of its sovereignty and adjacent riparian lands. The OHWL approximates the point to which the water normally rises during the high water season. The OHWL excludes exceptionally high water events caused by unusual floods or freshets. The OHWL encompasses all parts of the waterbody however shallow the water may be at the outside lines or elsewhere if the water is in fact a part of the waterbody that is navigable for useful purposes. It is an ambulatory line that shifts in response to long term, gradual, natural changes in water levels or in the shoreline. In general, accretion, erosion, reliction and submergence shift the OHWL while avulsion, artificial accretion, artificial erosion, artificial reliction and artificial submergence do not shift the OHWL. Designation as sovereignty land means that these lands are the property of the public and are to be used and maintained for the benefit of the public (Guest 1991). This sounds straightforward, but it is a misconception to believe that the OHWL is clear. As Warner (1999) notes, when Florida received title to all the navigable water bodies within the state in 1845, there were no lists or maps of all these water bodies and, consequently, the extent of the states ownership was unclear. This is still true today. Because of limited financial resources, OHWL determinations are still made only on a case-by-case basis. 1

PAGE 12

2 The OHWL and state sovereignty lands are inextricably linked. The concept of protecting lands for public use originates in Greek philosophy and Roman law. During the reign of the Roman Emperor Justinian (529-534 A.D.), the law specified the public nature of rivers, ports, and seashores: By the law of nature these things are common to mankindthe air, running water, the sea, and consequently the shores of the sea. . All rivers and ports are public. . The public use of the banks of a river is part of the law of nations, just as is that of the river itself. . The public use of the sea-shore, too, is part of the law of nations, as is that of the sea itself; . the shores may be said to be the property of no man. . (Sanders 1876 p. 158-159) This concept was later entwined in English common law and ultimately in American law (more specifically, in Florida law). It is because ordinary high water covers the shore on a fairly regular basis that the OHWL includes the shore. If an average or low watermark were used instead, the legal boundary between private and sovereignty, or public, lands would not include the shore (Warner 1999). The boundary between public and private land, however, has been at issue almost since Florida entered the Union. Since the late 1800s, court battles have raged over navigability issues and the line demarcating ordinary high water. Early Cases The earliest cases dealt with navigability. In an 1889 case, Bucki v. Cone, the Florida Supreme Court determined that a water body is navigable if it is a stream of sufficient capacity and volume of water to float to market the products of the country. . [and it is] not essential . that the stream should be continuously, at all seasons of the year, in a state suited for such flotage (Bucki v. Cone 1889 p. 160). In 1909, the Florida Supreme Court, in Broward v. Mabry, reaffirmed that the state holds in trust for the people the navigable waters and the lands thereunder, including the shore or space

PAGE 13

3 between ordinary high and low water marks and that capacity for navigation, not usage for that purpose, determines the navigable character of waters (Broward v. Mabry 1909 p. 827). Another problem that Broward v. Mabry dealt with resulted from the Swamp and Overflowed Lands Act that Congress passed in 1850. By the 1850 Act, states were granted all unsold swamp and overflowed lands within their borders. Florida received approximately 22 million acres that were above the ordinary high water mark (Warner 1999). According to Assistant Attorney General Monica Reimer (2001): The expectation [of Congress] was that the states would convey these lands to large companies which would then be responsible for drainage and reclamation. In an attempt to orderly manage these federal land grants, the 1855 Florida Legislature created the Internal Improvement Fund, which consisted of title to internal improvement lands and swamp land and all proceeds from the sale of such lands. The legislature then vested title to the internal improvement lands and swamp lands in the Trustees of the Internal Improvement Fund and gave them authority to sell and transfer the lands as prescribed by statute (p. 13). Broward v. Mabry (1909) clearly stated that the trustees of the internal improvement fund, who have the disposal of the swamp and overflowed lands of the state, have no authority to convey the title to the lands under navigable waters that properly belong to the sovereignty of the state (p. 827). Cornerstone Cases Two cases decided in 1927 became cornerstones of ordinary high water line law in Florida: Martin v. Busch and Tilden v. Smith. Both cases helped define the OHWL and introduced some of the problems involved in delineating the OHWL on Floridas freshwater bodies (Warner 1999). In Martin v. Busch (1927), the complainant (Busch) was claiming ownership of land that was once part of the lakebed of Lake Okeechobee before drainage operations lowered lake levels and exposed land. He claimed that his

PAGE 14

4 title to the land originated from the Swamp and Overflowed Lands Act of 1850. In hearing the case, the Supreme Court of Florida reiterated that swamp and overflow deeds did not convey title to lands below the OHWL and that all lands below the OHWL were designated as state sovereignty lands (Cooner n.d.). The Court also stated that it was the States responsibility to establish the boundary between the sovereignty lands and uplands (Warner 1999). This case also addressed the problem of locating the OHWL on low, flat-banked water bodies. . In flat territory or because of peculiar conditions, there may be little if any shore to navigable waters, or the elevation may be slight and the water at the outer edges may be shallow and affected by vegetable growth or other conditions, the line of ordinary high-water mark may be difficult of accurate ascertainment; but, when the duty of determining the line of high-water mark is imposed or assumed, the best evidence attainable and the best methods available should be utilized in determining and establishing the line of true ordinary high-water markMarks upon the ground or upon local objects that are more or less permanent may be considered in connection with competent testimony and other evidence in determining the true line of ordinary high-water mark . .(Martin v. Busch 1927 p. 274). Martin v. Busch is relevant to this study because the Court ruled that the best evidence attainable and the best methods available should be utilized in determining and establishing the line of true ordinary high-water mark . . (Martin v. Busch, 1927, p. 274). The Tilden v. Smith (1927) case clarified the concepts of ordinary and level (as in ordinary high water level), and found that the OHWL is . . a line at which the water usually stands when free from disturbing causes, or as the usual or ordinary height (p. 710). It also found that there can be considerable variation in water level resulting from unusually high rainfall or drought, but . . nevertheless the character of the vegetation and trees around the lake gave some evidence of an average or ordinary high-water mark . . (p. 709). Furthermore, the Court went on to say that since water in low, flat water

PAGE 15

5 bodies does not impress any well-defined line of demarcation, vegetation must serve as the principal criterion. Tilden v. Smith (1927) concluded that in ascertaining the normal level and average high-water mark of a lake, evidence may be introduced as to the character of the vegetation upon its banks (p. 708). A Recent Case A 1994 Appellate case validated the methodology (i.e., locating water marks and documenting vegetation patterns) used to locate the OHWL. In McNamara v. Kissimmee River Valley Sportsmans Association (and Board of Trustees of the Internal Improvement Trust Fund), action was brought by the Sportsmans Association over the right of Roger McNamara to fence off land, including a spoil island, on Lake Hatchineha. The Court found that the area in controversy was vegetated lake bottom and that spoil island was retained in public ownership (p. 155). The Court also upheld the cornerstone cases and reiterated that the OHWL was the boundary of normal reach during the high water season, that the OHWL is the place where water prevents the cultivation of ordinary crops, and that flat shorelines require using the best evidence, including water marks and historical aerial photos, for determining the OHWL (Wilkinson and Warner 2002). Debate Rages On An ongoing issue surrounding ordinary high water level began in 1976. The David Smith case is unique to the ordinary high water issue because it involves ditches and dikes along Lake Poinsett in Brevard County. Several dikes had been built below the OHWL of 13 ft (3.96 m) above mean sea level (AMSL), one about 1969, the other about 1973. These dikes came to the attention of the State in 1976 when Smith (and two neighbors) wanted to construct a new dike and ditch along the 13.00 ft contour line

PAGE 16

6 (OHWL in this case). A 1978 agreement was reached with the State contingent upon a restoration plan for the 1969 and 1973 dikes. This required the breaching of the 1969 and 1973 dikes to restore the Lake Poinsett marsh (Division of State Lands Response to David Smiths Fact Sheet. 1997. Florida Department of Environmental Protection, Tallahassee, FL, USA). The U.S. Army Corps of Engineers finally issued a permit in 1981 that required breaching the lower dike/ditch system and having restoration inspected and approved before beginning construction of the upper [13.00 ft contour] dike/ditch system (Division of State Lands 1997 p. 2). The 1997 DEP report goes on to say that in 1982, St. Johns River Water Management District (SJRWMD) notified Smith of their concern that he was attempting to proceed with development of the property prior to the completion of the mandated restoration. A 1983 letter from Smiths agent contended that Smith had been unable to complete the restoration due to inundation of the dikes by high water in Lake Poinsett. SJRWMD reported that same year that the restoration was 90% complete, and that 22 gaps had been created. Smith later asserted that the dikes remained functional (even when the gaps were cut) though aerial photos clearly showing inundation refute this assertion. Some time in 1986 Smith refilled the gaps in the two illegal dikes. In 1995, Trustees for the State filed suit against Smith charging that he had illegally diked, drained, and occupied several hundred acres of sovereignty land in Lake Poinsetts lakebed. Smith went to the Governors Cabinet and asked for a disclaimer on the property, asserting that the disputed property was swamp and overflowed land, which encouraged improvement by draining and ditching as set forth in the Swamp and

PAGE 17

7 Overflowed Lands Act of 1850 (Warner, S. 2002. Unreal estate, political waters and the authority of marshmallows: controversies surrounding Floridas ordinary high water line. Unpublished manuscript). As of 2002, the case remains open, but has not yet reached the courts. Opposing Sides On one side of the issue are those who believe the state is attempting a land grab from landowners like Mr. David Smith. The State and Trustees of the Internal Improvement Trust Fund, who are charged with protecting Floridas sovereignty lands, sit on the opposing side. Mr. Smith has become the poster boy of Florida Cattlemens Association (FCA) and the Florida Agricultural Coalition (FAC). A FAC position paper on ownership of private lands (1999) gives a brief description of Floridas acquisition of swamp and overflowed lands and goes on to suggest that the Smith case (referred to indirectly) may change Florida Supreme Court case law that has existed since the early part of the century (p. 1). FACs (1999) position paper goes on to say, BTIITF [Trustees] over the years has sold almost all of the land to private owners. Many of these lands were actually sold with the intent to drain and develop them, or the land would revert back to the State (p. 1). They do not mention that draining and development could not occur on sovereignty lands. FAC (1999) claims the gist of the issue is that the Attorney General is attempting to change the law to have OHWL be The highest reach of water during an ordinary year. This means the highest reach of water during the wet season, which everyone knows, is the rainy season or flood stage. This wouldtake up over 600,000 acres of lands that are now owned, farmed, or occupied by private landowners, ON WHICH TAXES ARE BEING PAID. (p. 1).

PAGE 18

8 There is no mention of previous Supreme Court rulings that held that sovereignty lands were not conveyed with swamp and overflowed land deeds, nor is there any mention of the 1994 MacNamara case in which the Appellate Court held that the OHWL on fresh waters is the ordinary or normal reach of water during the high water season (MacNamara v. Kissimmee River Valley 1994 p. 156). The MacNamara case reaffirmed the cornerstone cases, it did not change the definition of OHWL. Furthermore, there is no mention of case law, which addresses the issue of taxes. MacNamara v. Kissimmee River Valley (1994) states Even if taxes had been paid, such payment cannot form the basis of equitable estoppel because it is the Trustees of the Internal Improvement Fund rather than the tax assessor who are authorized to speak for the state on the subject of boundaries on navigable lake bottoms. Sec. 253.12(1), Fla. Stat. If a taxing error has taken place, the remedy is a tax refund rather than conversion of lake bottoms to private ownership (p. 163). Finally, FAC (1999) states The members of the FAC believe landowners should be able to go to a public library, get a copy of the law, and go to their property and identify where the property ends and the publics begins. By [the] same token, a member of the public should be able to look at the bank of a water body and identify where public lands end and private property begins. These owners should be able to fence and post their land against trespassers and poachers. (p. 1) It sounds simple enough, go to the lake, walk the shore, and locate the water line. If the OHWL were so easy to find, Floridas courts would not have ruled on numerous cases regarding navigability and ordinary high water lines. Furthermore, case law would not mention low, flat banked water bodies, nor would there be any directions as to the ascertainment of the OHWL such as water marks on objects, competent testimony, and vegetation indicators. This study confirmed the difficulty of unequivocally determining the OHWL on such water bodies.

PAGE 19

9 The FAC (1999) position paper also includes an editorial opinion by FCA President Larry Barthle which goes into further detail on the Smith case. In that editorial, Barthle expresses his view that the courts are changing the rules and that the OHWL is no longer the mark on the bank that any fool can see which has been caused by the continuous action of water for eons (p. 2). This editorial also fails to mention that sovereignty lands were not conveyed with swamp and overflowed land deeds. Mr. Barthle also suggests that any fool can see the high water mark. However, Lake Poinsett is a low bank lake where there is no mark on the bank caused by continuous wave action. Interestingly, FAC and FCA both contend that Congress intended swamp and overflowed lands for development after draining and ditching. This is accurate providing any such efforts did not include sovereignty lands. Neither organization mentions that Mr. Smiths own privately hired engineers estimated the OHWL of Lake Poinsett at 13 ft (3.96 m) AMSL in 1977. Staff for the Trustees then concluded that the OHWL was not below the 13 ft (3.96 m) contour. However, Mr. Smiths dikes and ditches built in 1969 and 1973 extend below the 8.3 ft (2.53 m) elevation (Warner 2002). The State, on the other hand points to previous case law in which the Courts have repeatedly upheld the States ownership of land below the OHWL. Legislative Action Periodically, lobbying activities by FAC, FCA, and others have attempted to get the Legislature to unequivocally define the OHWL. The most recent attempt was a bill introduced during the 2000 Legislative session. Senate Bill 1824 and House Bill 1807, the Florida Land Title Protection Act, proposed to do just that. According to Larry Barthle of FCA, HB1807 and SB 1824, will end the repugnant and arrogant land grab perpetuated by Attorney General Butterworth. . [It] will put into law the definition of

PAGE 20

10 ordinary high water mark that was in use for more than 140 years. (FAC 1999 p. 2) In the analysis of the Florida Land Title Protection Act, HB1807 (2000) defines the OHWL as . the point upon which the presence and action of the water is so continuous as to destroy the value of the land for agricultural purposes, preventing the growth of vegetation, constituting what may be an ordinary agricultural crop. This mark is ambulatory and shifts in response to long-term changes in the water level. The high water is not the highest point to where the water rises in times of floods or rains, but instead it is the line in which the water impresses upon the soil and covers it for such periods so as to deprive it of vegetation and destroy its agricultural value. In addition, this section clarifies that lands that are subject to such periodical overflow so as to require drainage, levees or embankments to keep the water out in order to render the land suitable for cultivation are not sovereignty lands, below the ordinary high water mark. Furthermore, this water-mark does not include: nonnavigable creeks, sloughs, swamps, canals, and low and overflowed lands that are adjacent to the navigable water body (p. 4). This definition of OHWL diverges substantially from case law. Had this legislation passed, the OHWL would have moved waterward and would no longer include any land that could be ditched or drained to prevent the presence of water during the high water season. Excluding the seasonality of rainfall from ordinary high water line determinations makes no sense because Florida has both a distinct wet and dry season. The wet season plays an important revitalizing role in Floridas ecosystems and, therefore, must be included as part of an ordinary year as stated in case law. If one excludes the seasonal rains then, in fact, the water line would be measured only during the dry season when water levels are low. In that event, case law would be addressing ordinary low water line not an ordinary high water line. The 2000 bill eventually died on the floor. It was not the first time that legislation was proposed to try to define, with certainty, the OHWL and it will undoubtedly not be the last. Sam Ard, lobbyist for FCA, lobbied for this bill. Since this bill died, Mr. Ard and other lobbyists for FAC and FCA will undoubtedly continue lobbying for legislative action to define the OHWL.

PAGE 21

11 How It All Fits Together As noted above, OHWL issues continually recur in Florida. OHWL determinations continue to be complicated and contentious. The Mr. Smith case has revitalized the fight in recent years and inspired introduction of the Land Title Protection Act in the 2000 legislative session. After the bill died, lobbying activities by Sam Ard of FCA sought to have funds appropriated for research to determine the amount of acreage held as sovereignty land. To that end, the 2001 Legislature granted the following: From the funds in Specific Appropriation 1661, up to $300,000 is provided to the Department of Environmental Protection for the purpose of a pilot project to map and determine acreage of sovereignty lands using remote sensing satellite data obtained through NASA or its affiliates (Interim Land Management p. 229). The intent of this grant was to determine the viability of using remote sensing technologies to accurately ascertain the area of Floridas sovereignty lands. Lakes Hatchineha and Kissimmee were ultimately chosen for this pilot study because the OHWL for Lake Hatchineha had been adjudicated at 52.50 ft (16.00 m) AMSL in MacNamara v. Kissimmee River Valley and long term stage data exists for both lakes. Furthermore, it has been commonly accepted that the OHWL for Lake Kissimmee is 52.00 ft (15.85 m) AMSL. Knowing the OHWL is essential to the calculation of the total acreage of sovereignty lands. To do this, one would need to delineate the OHWL by detecting the presence of water within the emergent vegetation or be able to detect vegetation shifts corresponding to the OHWL via changes in pixel classes. In a phone conversation in June 2002 with Mr. Terry Wilkinson, of DEP, Mr. Wilkinson indicated that the desired margin of error when determining a boundary line is m.

PAGE 22

12 Historical or traditional methods of determining the OHWL include manual interpretation of aerial photography in conjunction with field surveys. In field surveys, the presence of such things as water lines on stationary objects, debris lines, lichen lines, vegetation types and soil types are all useful for delineation purposes. According to Davis (1973), many vegetation types are distinctly related to the amount, position, and duration of water in the substratum. The vegetation can then be classified as obligate wetland, facultative wetland, facultative, facultative upland, or upland or terrestrial based on its frequency of occurrence along the hydrologic gradient (Reed 1988). By determining the relative positions of the terrestrial plants and wetland plants, one can determine, with some reliability, the present position and former positions of water levels (Davis 1973). The goal of this study was to assess the current state-of-the-art satellite and aircraft based remote sensing (Landsat, SPOT, IRS, IKONOS, and CASI) technologies for the purpose of corroborating the OHWL of 52.50 ft (16.00 m) and 52.00 ft (15.85 m) AMSL for Lakes Hatchineha and Kissimmee, respectively. The use of this new technology will be in line with the intent of case law requiring that the best methods available should be used to locate the OHWL. If a line of demarcation can be found using remote sensing, the State would have a less expensive, yet reliable method for locating the OHWL of Florida lakes. The hypotheses for this study were as follows: Vegetation can be used as an indicator to locate the OHWL. Landsat will not be effective for locating the OHWL. SPOT will not be effective for locating the OHWL. IRS will not be effective for locating the OHWL. IKONOS will not be effective for locating the OHWL. CASI will be useful for locating the OHWL.

PAGE 23

13 Site Description Lakes Hatchineha and Kissimmee are located in Osceola and Polk Counties in south central Florida (Figure 1-1). Kissimmee is approximately 140 km 2 (Brenner et al. 1990). Lake Hatchineha is smaller, covering approximately 27 km 2 (Bishop 1967). The OHWL for Lake Hatchineha has been adjudicated to be 52.50 ft (16.00 m) AMSL and the commonly accepted OHWL for Lake Kissimmee is 52.00 ft (15.85 m) AMSL. This area of Florida had undergone an extended drought at the time of this study and water levels in Lake Kissimmee had not reached 52.00 ft (15.85 m) AMSL since January 2000 (Figure 1-2). Figure 1-1 Study area

PAGE 24

14 OHWL 52.00 ft Time of study Figure 1-2 Average monthly stage data for Lake Kissimmee covering a 7-year period

PAGE 25

CHAPTER 2 LITERATURE REVIEW The origin of remote sensing began with aerial photography in 1858 over the outskirts of Paris, France. Gaspard Felix Tournachon, who suspected that aerial photography would one day become very important, took the first-known successful aerial photograph from a balloon. Aerial photographys first military use came during the American Civil War. By the end of World War I and during World War II, aerial photography had become key in planning military strategy and was also important during the cold war (Jensen 2000). In 1960, the first satellite image was successfully recorded onboard the Corona satellite, and the era of satellite imagery was born (Jensen 2000). Early remote sensing technology seems to have been primarily used by the military. However, in 1972, Landsat-1 began recording data to test the feasibility of collecting land-surface data about the Earth from an unmanned satellite. Since that time, a number of other satellites from the United States and elsewhere have been launched and images are used for both government and non-government programs (Jensen 2000). Remote sensing has become an important technology in environmental and ecological studies. Large and remote ecosystems can now be classified into biologically significant categories in a timely and cost-effective way (Hall-Atkinson and Smith 2000). For example, the National Biological Services Gap Analysis Program is an important program that utilizes remote sensing technologies to determine land-use/land-cover classes, which can help identify gaps in the protection of biological diversity (Edwards 15

PAGE 26

16 et al. 1998). Remote sensing plays an important role in analyzing changes in land use and cover because of its ability to record changes efficiently, which then provides information important to the management of natural resources. In heterogeneous environments, remote sensors can provide large amounts of data both efficiently and economically (Ruiz-Luna and Berlanga-Robles 1999). The ability of remote sensing to provide a cost-effective means of classifying the environment has been an important factor in its increasing popularity for large ecosystem analyses. One of the first ecological programs developed using remote sensing technology was US Fish and Wildlife Services National Wetlands Inventory (NWI) project. Wetlands mapping began in 1979. NWIs goal was to generate and disseminate scientific information about wetlands that could be used for decision-making about resource use and allocation. There needed to be a resource that would provide information about the characteristics and extent of the Nations wetlands (Wilen and Bates 1995, Loveland 2001). NWI researchers developed a classification system that divides ecological systems, with certain common characteristics, into particular groups. The resulting classification divided habitats into five major systems. Each system is divided into subsystems that reflect hydrologic conditions, followed by classes that describe the appearance of wetland vegetation (e.g. emergent wetland) or substrate. Classes are further divided into subclasses that are named based on the predominant vegetation life form. Descriptors are also included that provide information about hydrology, water chemistry and human influence (Cowardin et al. 1979).

PAGE 27

17 Once a classification system was developed, a remote sensing technology was chosen. Researchers decided that satellite imagery from Landsat would not be able to provide the necessary detail for wetland classification. Therefore, midand high-altitude color infrared aerial photography, at a scale of 1:24,000, was chosen. Once images are obtained, a long process is involved in creating wetlands maps that includes photo-interpretation, field studies, review of soil maps, and review of existing information about an areas wetlands to assure accurate identification (Wilen and Bates 1995). Turnaround time from photo acquisition to map production can be as long as 10 years (Wilen and Frayer 1990). As of 1995, the NWI project had continued to investigate the use of satellite imagery for use in wetlands mapping. Landsat technology was reevaluated and the French satellite SPOT was also investigated. After a yearlong test, it was determined that satellite imagery was still not able to provide the needed level of detail for wetland classification. Therefore, NWI continues to use aerial photography for wetland mapping (Wilen and Bates 1995). Other researchers have found satellite imagery to be quite useful for their purposes. The US Geological Surveys (USGS) Gap Analysis Program (GAP) that began about 10 years ago found Landsat, with its 30 m resolution, to be sufficient for mapping vegetation and habitat for conservation status assessment of vertebrate species in order to allow individuals, planners, managers, and policy makers to make informed decisions (Scott et al. 1993, Loveland 2001). One problem of GAP is that consistent methods of classification are not used by each state. To overcome this problem, GAP has developed a classification template to assist states in providing consistent maps. Additionally, GAP

PAGE 28

18 promotes the use of a common data source and the use of consistent vegetation definitions. It is recognized, however, that there are still limitations of satellite remotely sensed data for producing these maps, mainly due to spatial resolution (Loveland 2001). Another major program was developed by USGS in 1995. The purpose of the USGS Land Cover Characterization Program is to provide consistent national land use and land cover data that is nationally consistent and compatible with past and future USGS land cover classifications. This program also uses Landsat imagery and uses an unsupervised classification approach augmented by post-classification modeling that include key variables such as elevation and population density. The mapping process takes about three to four years. Therefore, maps are already dated by the time they are released. As of 2001, accuracy assessments had not been completed and all results were preliminary (Loveland 2001). National vegetation maps are not the only maps of interest. Many researchers are interested in producing global vegetation maps. Historically, scientists have attempted to classify global vegetation using Earth Systems models. Early global vegetation databases were derived from published maps, atlases and national databases, and they attempted to represent existing global vegetation. Problems with these early databases include inconsistency in vegetation classification, variable measurement techniques and variable spatial sampling resolutions (Running et al. 1995). Prentice et al. (1992) and Neilson et al. (1992) developed models of geographic distributions of vegetation based on specific physiological responses to cold tolerance, growing season heat sums, and drought stress. In these models, equivalent vegetation types, such as forests, can be defined separately into classes like boreal, temperate, or tropical because their

PAGE 29

19 classification scheme includes climatalogical parameters as integral components. Still, some researchers argue that these models, though improved, still only provide maps of potential vegetation types (Running et al. 1995). Townshend et al. (1991) suggested that global vegetation classifications must be realistic measures of existing land cover. A consistent, remote sensing-based classification scheme would remove many of the ambiguities in global vegetation maps because they would no longer be derived from varying methodologies. However, remote sensing technology cannot produce maps with the large number of classes seen in the earlier maps because they remove climate classes from the definition (Loveland et al. 1991). Therefore, Running et al. (1995) suggested that a new logic be introduced for global vegetation classification that would be based on simple, observable, unambiguous characteristics of vegetation structure related to ecosystem biogeochemistry. Such characteristics would need to be repeatedly observable with remote sensing, such that they could be classified the same way each time and they would be translatable into biophysical parameters of interest by global climate and biogeochemical models. Running et al. (1995) suggests that this classification logic should be the basis for all classifications, and that those requiring more exacting classifications build hierarchically from this one. This is a logical suggestion because a single basis for all classification schemes would then start out on a level playing field. As remote sensing technology improves, vegetation classification should become more accurate, especially if all classifications start with the same foundation. This would improve global vegetation maps as well as smaller scale maps that are used to analyze change over time in specific ecosystems.

PAGE 30

20 Landscape ecologists using satellite imagery to characterize vegetation, species distributions, and communities are also seeking ways to make classifications more accurate. Unfortunately, classification maps often simplistically characterize heterogeneous communities as single, discrete units (Townsend 2000). Townsend (2000) reports that among potential options for improved classification schemes are fuzzy classifications and mixture/subpixel models. However, these models merely convey information about variations in spectral signature between pixels, not necessarily differences in on-the-ground content of mapped vegetation. Despite these problems, some method of delineating classes on maps is essential in order to communicate information about spatial patterns and the distribution and abundance of species. According to Townsend (2000), the resulting maps show transition zones as discrete boundaries. Furthermore, some recognizable on-the-ground vegetation associations may not actually be discernable with satellite or other remotely sensed images. Townsend and Walsh (2001) report that complex community relationships and the nature of their spatial variability are rarely captured with satellite imagery. Broad ecotones are especially difficult to portray accurately because they are made up of transition zones that grade into each other across complex soil, hydrologic, and topographic gradients. The key to making vegetation maps more appropriate for ecological interpretation is to find a classification scheme that more accurately portrays complex vegetation communities. Some of the most complex communities are coastal wetlands. High spatial resolution photography is useful for classifying and identifying changes in coastal wetlands. NWI is a primary producer of these high quality, detailed, coastal wetland maps from high spatial resolution photography (Peters, D.D. 1994. Use of aerial

PAGE 31

21 photography for mapping wetlands in the United States: National Wetlands Inventory. Proceedings of the First International Airborne Remote Sensing Conference and Exhibition. Strasbourg, France). Production of these maps is enormously expensive and turnaround time from photo acquisition to map production can be as long as ten years (Wilen and Frayer 1990). Obviously, maps with such long turnaround times are not useful for monitoring more immediate changes in coastal wetlands. In an effort to find a less expensive and faster way to monitor wetlands, Ramsey and Laine (1997) compared the use of Landsat and high spatial resolution photography for identifying changes in complex coastal wetlands. Landsat images are less expensive, but researchers needed to determine their viability for identifying changes in coastal wetlands. In their study, Ramsey and Laine (1997) used Landsat to monitor preand post-hurricane changes in Louisianas coastal wetlands. They encountered problems with using Landsat, including misclassification of emergent vegetation and misclassification in areas that included a mix of emergent vegetation and water. Landsat images did, however, provide good separation of classes when one class dominated areas >30 m. This study illustrates both the problems and the potential for using Landsat images to identify complex wetland ecosystems. In 1998, researchers in Sango Bay, Uganda also attempted to assess complex ecosystems. A biodiversity assessment was conducted in an attempt to determine which of the areas ecosystems have the greatest biodiversity value. The area includes swamps, grasslands, cultivated land, and forests that border Lake Victoria in Uganda. It was important to obtain a detailed map of the area, and since accessibility was a problem in many of the areas, remote sensing was the logical method of choice. Landsat was chosen

PAGE 32

22 over SPOT because even though SPOT has a finer resolution (20 m as opposed to Landsats 30 m), Landsat records an additional infrared channel that is important for discriminating different vegetation types. A supervised classification scheme (Maximum Likelihood Classification) was used. This type of classification scheme requires that the people involved have a good working knowledge of actual ground vegetation and how these vegetation types appear on the image (Fuller et al. 1998). Final maps identified fourteen land-cover classes with the broadest categories being: water, swamp, dry grassland, degraded woody vegetation, semi-natural woody vegetation, and intensive land usage. Broad categories were subdivided (Fuller et al. 1998). Of interest here is the subcategory tall grass swamp. Five tall grass swamps were identified and mapped, but there was tremendous variability within the class. For example, four sites included five to ten grass species while the fifth site included 33 species (Fuller et al. 1998). The Sango Bay research brings two significant factors to light. First, one must have a thorough knowledge of the vegetation communities being studied, and second, that Landsat was unable to distinguish between a swamp with five species and a swamp with 33 species. This implies that Landsat would not be very useful in delineating or locating a line such as the OHWL on Lakes Kissimmee and Hatchineha. Satellite imagery is continuing to grow in popularity for use in mapping wetlands. While NWI continues to use aerial photography in order to produce detailed wetland maps, other researchers are testing some of the newer satellite images available. In Punjab, India, researchers prepared a land use/land cover map of the Harike wetland ecosystem from Indian Remote Sensing Satellite (IRS) imagery with a 23 m spatial

PAGE 33

23 resolution or ground cell size. The Harike wetland covers 285.1 km 2 and is rich in aquatic flora and fauna. This ecosystem assumed international importance because it serves as a breeding ground and habitat for a large variety of migratory and domiciled birds. Anthropogenic pressure in recent years has created severe ecological problems and resulted in the need to conserve this wetland. Since no reliable data were available for Harike, researchers wanted to develop a baseline map that would identify spatial distribution and variation in the waterspread, turbidity and aquatic vegetation in different seasons. IRS images were obtained from different years for Harike, both pre-monsoon and post-monsoon (Chopra et al. 2001). A map was generated using visual interpretation of IRS images and a classification system recommended by the Space Applications Centre of Ahmedabad, India. The Harike wetland ecosystem was ultimately divided into five major classes: built-up land, agricultural land, forest, wasteland, and wetland. When the pre-monsoon and post-monsoon images were compared, researchers were able to record differences in the waterspread and turbidity. Furthermore, based on visual interpretation, three types of aquatic vegetation were identified: water hyacinth, grassland 1 and grassland 2. The difference between the two grassland categories resulted from the portion of the spectrum with the highest reflectance (grassland 1 had high reflectance in the blue, green, and red regions while grassland 2 had high reflectance in the infrared region). Though these were recorded as two distinct classes on the map, no verification of differences in vegetation species between the classes on the ground was possible due to inaccessibility of the area (Chopra et al. 2001). The Harike study illustrates the potential for using IRS images to

PAGE 34

24 broadly classify wetlands, but challenges still exist in terms of accurately assessing heterogeneous communities whether they are wetland or upland. Satellite imagery still poses challenges for wetland mapping due, in part, to problems with spatial resolution. Vegetation gradients may not be detected with satellite imagery because one pixel is likely to be wider than the associated vegetation shift on the ground. Furthermore, the spectral resolution of satellite imagery may restrict its ability to detect differences in vegetation types, which are often spectrally similar in wetlands (Harvey and Hill 2001). When comparing various remote sensing technologies, researchers found large-scale aerial photography to be superior to both Landsat and SPOT satellite imagery for detailed wetland mapping. More importantly, the use of context and texture information obtained with aerial photography allowed researchers to enhance vegetation class delineation and identification capabilities. The use of contextual and textural characteristics is especially useful and desirable for delineation of vegetation communities in heterogeneous wetland ecosystems. However, no significant difference existed in the ability to map key structural components with any of the remotely sensed images (Harvey and Hill 2001). The key, then, is to be aware of the different capabilities of the imagery and to select the ones that would be most likely to produce the required information. This again illustrates the likely failure of Landsat, SPOT, IRS, and IKONOS satellite imagery to prove useful in locating the OHWL on Lakes Hatchineha and Kissimmee. Remote sensing images used in the Lake Kissimmee and Lake Hatchineha study mainly provide vegetation information. Consequently, it was important to look at

PAGE 35

25 vegetation communities and how they fall out along shorelines. Lakeshores often provide large areas of wetland vegetation including swamps and marshes (Keddy and Fraser 2000). Environmental conditions such as water levels, soil fertility, grazing and other disturbances are known to affect plant species composition in wetlands (Hall et al. 2001, Keddy and Fraser 2000). The shoreline merely stretches wetland vegetation out along several gradients (Keddy and Fraser 2000). One would expect to find narrow bands of wetland vegetation along steep topographic gradients and much broader bands in flat topographic areas such as those found along Lakes Hatchineha and Kissimmee. Associations between hydrophytic vegetation communities and hydric soils are usually found along the shoreline gradient as well. Some researchers have looked at the relationship between hydrophytic vegetation and hydric soils. Wetland delineation standards developed by the U.S. Army Corps of Engineers require evidence of hydrophytic vegetation, hydric soils, and wetland hydrology. Logically, one would expect to find a direct relationship between hydrophytic vegetation and hydric soils. However, the flat topography of Florida has been problematic in linking hydric soils to hydrophytic vegetation in some areas (Davis et al. 1996). According to Reed (1988), the 1987 Corps Manual defines a hydrophytic plant community as one that has greater than 50% of dominant species classified as obligate (OBL), facultative wetland (FACW) or facultative (FAC) on regional lists of plants that occur in wetlands. Some hydric soil indicators found in Floridas sandy soils include stratified layers within the upper 15 cm of the surface with value 3 or less and chroma 1 or less, presence of muck, stripped matrix beginning within the upper 15 cm, and presence of 2% or more organic bodies starting within the upper 15 cm of the surface (USDA 1998).

PAGE 36

26 Davis et al. (1996) studied the relationship between hydrophytic vegetation, hydric soils, and hydrology in north-central Florida. Researchers had no difficulty identifying wetland or upland communities that occurred at extremes of the hydrologic gradients. However, flatwoods communities were not consistently identified as wetlands according to the 1987 Corps Manual, but were frequently classified as wetlands according to the 1989 Corps Manual. Davis et al. (1996) found that Floridas flat topography obscured the upland extent of hydrophytic vegetation and the expected corresponding hydric soils. In flat topography, small variations in water level can extend over a greater area, and for a longer period, than in steep terrain. Consequently, hydrophytic vegetation can be found upslope from the wetland hydrology or hydric soil boundary. Under natural conditions, shorelines of freshwater lakes are subject to both wave action and frequent changes in water level. Such natural disturbances can lead to shorelines with high biological diversity. Vegetation types are directly related to the duration of flooding. Thus, the degree of flooding results in distinct vegetation zones on the shoreline (Keddy and Fraser 2000). Since the duration of flooding varies over time, vegetation zonation is a dynamic property. In years with extraordinarily high water levels, upland shrubs and trees may die. Alternatively, during periods of drought, existing seed banks may germinate, increasing the diversity of species close to the lakeshore (Keddy and Fraser 2000). Another factor controlling species composition of shorelines is a species competitive ability. Competitive ability is linked to environmental conditions and in some cases soil fertility. In fertile soils, the pioneer and generalist species tend to

PAGE 37

27 dominate, leaving specialist species to occupy the remaining spaces. There also appear to be hierarchies of competitive ability that control the distribution of species along shorelines. Soil fertility and dominance hierarchies are only some of the factors controlling vegetation species in wetlands. Disturbance also plays a role in species distribution. Finally, various combinations of fertility, disturbance, and hydrology will have different impacts on wetland vegetation gradients (Keddy and Fraser 2000). Lakes Kissimmee and Hatchineha have both been impacted by disturbance, shifts in hydrology, and probably soil fertility. The Kissimmee River was once a long, meandering river, originating in Orlando and winding its way to its terminus in Lake Okeechobee (Kushlan 1990). However, between 1962 and 1971, the U.S. Army Corps of Engineers channelized the river to provide drainage and flood control for agricultural and residential development. The first lock on canal C-38 is located at the southern end of Lake Kissimmee. The channelization of the river had significant negative impacts on area flora and fauna (Wetzel et al. 2001). Both lakes have changed dramatically since the channelization of the river. Water levels are now controlled, and are lower than historic levels. Most importantly, water no longer flows naturally. Stagnation of the lakes has contributed to numerous other problems, primarily the invasion of hydrilla, water hyacinth and other aquatic vegetation that have taken a firm hold. Consequently, there are now periodic drawdowns and muck removals, leaving behind spoil islands. Undoubtedly, changes in the Kissimmee River influenced the nature of the lakes and probably changed the vegetation communities along the shorelines as well. Under normal circumstances, one would expect to find distinct vegetation gradients on a lakeshore. Many Florida lakes experience dramatic water level

PAGE 38

28 fluctuations during the year due to a distinct wet and dry season. These fluctuations, in addition to Floridas relatively flat topography, combine to create wide littoral zones. In some areas, shorelines have distinct zones marking low-water to high-water lines, called the upper, middle, and lower zones (Davis 1973). The upper zone is the zone of interest for this study because it is where water reaches the highest level assuming an ordinary year. The plant community in this zone includes plants that can tolerate water for only a short period of time. According to Hall (1987), some plant species can be found in more than one habitat while others are restricted to a single habitat. Thus, the make-up of the plant community in the area of interest can provide valuable information. For example, the presence of predominantly upland species indicates the area never floods, or floods only in unusually high water events. In such communities, a flood event would cause nearly all species to die (Hall 1987). If an area can be identified as upland then, clearly, the OHWL would be located at a lower elevation. According to Davis (1973), the upper zone, where water reaches the highest level, may contain such species as sweet gum, magnolia, saw palmetto, the sand cypress bushes Hypericum spp., carpet grass Axonopus sp., Bermuda grass, and crab grass. Hall (1987) recommends locating both the upland line, where flooding never occurs, and the lower line, where flooding always occurs, before attempting to locate the OHWL. Some species indicative of upland communities are turkey oak (Quercus geminata), wire grass (Aristida stricta), and longleaf pine (Pinus palustris) (Davis 1973, Hall 1987). Conversely, species indicative of the lower zone, where water is always or nearly always present, include cattail (Typha sp.), pickerelweed (Pontederia sp.), and maidencane

PAGE 39

29 (Panicum hemitomon) (Davis 1973). Under natural circumstances, cypress trees (Taxodium sp.) are good indicators of flooded conditions (Hall 1987). Once these upper and lower zones are located, other indicators of OHWL can be used. These include lake stage data, drift lines, lichen lines, algal mats, cypress knees, and scarps, among others. Following identification of these indicators, vegetation in the area should be identified (Hall 1987). Some species, however, are problematic because of their regional differences in distribution (Davis 1973, Hall 1987). Floridas climate extends from temperate to subtropical and the physical and chemical properties of the soil vary widely throughout the state. Classification problems arise when plant species occur only in wet habitats in one region, but occur in other habitats, even uplands, in other regions (Hall 1987). Davis (1973) studied numerous lakes, including both Lakes Kissimmee and Hatchineha, to develop a method for determining former high water levels by identifying plants living in or near Floridas lakes. On Lakes Kissimmee and Hatchineha, Davis (1973) was able to determine historic water levels by comparison of oak tree (UPL) and cypress tree (OBL) ages. For both lakes, oaks had occurred predominantly on shoreline berms where water never reached. At the time of his study of Lake Hatchineha in 1971, the water level was approximately 50.20 ft (15.30 m) AMSL. Davis found that the historic high water level was 56.70 ft (17.29 m) AMSL since about 1870. A vertical distance of greater than 45.10 m exists between the water level on the day of Davis study and the historic high water level. At the time of statehood, the high water position was probably about 57.00 ft (17.37 m) AMSL. On Lake Kissimmee, Davis determined that the high water level was probably never as high as 57.00-60.00 ft (17.37-18.29 m) AMSL

PAGE 40

30 during the 20 th century. At the time of his study, Davis (1973) noted that water levels in Lake Kissimmee had gradually been lowered by about 3.00 ft (0.91 m) and were being regulated by the new Kissimmee River controls. There is good evidence that both remote sensing and vegetation identification and analysis may be useful tools in locating the OHWL on Lakes Hatchineha and Kissimmee. Remote sensing technologies have been used successfully to broadly delineate wetlands. Research by both Davis and Hall show that vegetation, though problematic at times, can be used to help locate the ordinary high water line.

PAGE 41

CHAPTER 3 MATERIALS AND METHODS Imagery Four commercial earth-looking satellites provided suitable imagery for mapping vegetation and shoreline for use in this study: Landsat 7 ETM (Landsat) Indian Remote Sensing Satellite (IRS) SPOT IKONOS In addition to satellite imagery, airborne hyperspectral images from Canadas Compact Airborne Spectrographic Imager (CASI) were obtained for portions of the western shores of both Lakes Hatchineha and Kissimmee. Landsat, IRS, SPOT, and IKONOS images were purchased when lake elevations were known to be near ordinary high (16.00 m AMSL for Lake Hatchineha and 15.85 m AMSL for Lake Kissimmee based on NGVD 29). Landsat The Landsat sensing instrument is a multispectral scanner that measures radiation reflected or emitted by the Earths surface. There are eight bands sensitive to different wavelengths of visible and infrared radiation. Ground cell size is nominally 30 m x 30 m. The spectral parameters are as follows: Band 1: 0.45 to 0.52 m (blue). Provides increased penetration of water bodies, as well as supporting analyses of land use, soil, and vegetation characteristics. Band 2: 0.52 to 0.60 m (green). This band spans the region between the blue and red chlorophyll absorption bands and therefore corresponds to the green reflectance of healthy vegetation. 31

PAGE 42

32 Band 3: 0.63 to 0.69 m (red). This is the red chlorophyll absorption band of healthy green vegetation and represents one of the most important bands for vegetation discrimination. It is also useful for soil boundary and geological boundary delineations. This band may exhibit more contrast than bands 1 and 2 because of the reduced effect of atmospheric attenuation. The 0.69 m cutoff is significant because it represents the beginning of a spectral region from 0.68 to 0.75 m, where vegetation reflectance crossovers take place that can reduce the accuracy of vegetation investigations. Band 4: 0.76 to 0.90 m (reflective infrared). This band is especially responsive to the amount of vegetation biomass present in a scene. It is useful for crop identification and emphasizes soil/crop and land/water contrasts. Band 5: 1.55 to 1.75 m (mid-infrared). This band is sensitive to the turgidity or amount of water in plants. Such information is useful in crop drought studies and in plant vigor investigations. Band 6: 10.4 to 12.5 m (thermal infrared). This band measures the amount of radiant flux emitted from surfaces. The apparent temperature is a function of the emissivities and the true or kinetic temperature of the surface. It is useful for locating geothermal activity, vegetation classification, vegetation stress analysis, and soil moisture studies. Band 7: 2.08 to 2.35 m (mid-infrared). This is an important band for the discrimination of geologic rock formations. It has been shown to be particularly effective in identifying zones of hydrothermal alteration in rocks (Jensen 2000 p. 194). Finally, the panchromatic band is sensitive to wavelengths between 0.52 m and 0.90 m. Bands 1 to 5 and 7 have a 30 m x 30 m spatial resolution or ground cell size. The thermal infrared band 6 has 60 m x 60 m ground cell size, and the panchromatic band has a 15 m x 15 m ground cell size (Jensen 2000). Landsat satellites pass over the same place on earth every 16 days. The image used for this project was taken on October 23, 1999 (scene identification #23000310450487412). On that date, the average water level at the United States Geological Survey (USGS) gauging station located in Lake Kissimmee was 15.85 m AMSL.

PAGE 43

33 Indian Remote Sensing Satellite The IRS sensor has many of the same characteristics as Landsat. It was designed to fill in for times that Landsat was unavailable. The spectral parameters of IRS LISS-III multispectral sensor are as follows (Jensen 2000): Band 2: 0.52 to 0.59 m with a 23 m x 23 m spatial resolution. Band 3: 0.62 to 0.68 m with a 23 m x 23 m spatial resolution. Band 4: 0.77 to 0.86 m with a 23 m x 23 m spatial resolution. Band 5: 1.55 to 1.70 m with a 23 m x 23 m spatial resolution. Panchromatic: 0.50 to 0.75 m with a 5.8 m x 5.8 m spatial resolution. A color (23 m pixel size) IRS scene of the study area taken on December 4, 1999 was acquired for this project. The scene identification number is 1C28705200. The average water level at the USGS gauging station located in Lake Kissimmee was 15.96 m AMSL on that date. The SPOT Satellite Two sensors are onboard the SPOT satellite. One has a 20 m x 20 m ground cell size and records data in 3 spectral bands (Band 1 has a spectral resolution of 0.50 to 0.59 m; Band 2 has a spectral resolution of 0.61 to 0.68m, and Band 3 has a spectral resolution of 0.79 to 0.89 m). The second sensor uses one panchromatic band (0.51 to 0.73 m) with a ground cell size of 10 m x 10 m (Jensen 2000). The SPOT scene acquired for this project was taken on March 2, 1993 with the 3-band multi-spectral sensor. The scene identification number was 2-260-293-930302-163230-1-X. The average water level at the USGS gauging station located in Lake Kissimmee was 15.47 m AMSL on that date.

PAGE 44

34 The IKONOS Satellite One of the latest commercial earth-looking remote sensing satellites is IKONOS. The IKONOS satellite sensor has a panchromatic band with a spectral resolution of 0.45 to 0.90 m and a ground cell size of 1 m x 1 m. IKONOS also has 4 multispectral visible and near-infrared bands (Band 1: 0.45 to 0.52 m, Band 2: 0.52 to 0.60 m, Band 3: 0.63 to 0.69 m, and Band 4: 0.76 to 0.90 m) with a 4 m x 4 m ground cell size (Jensen 2000). IKONOS imagery from both panchromatic and multispectral scanners was obtained for this project. The IKONOS Space Imaging Corporations product acquired for this project was a Geo 1 m + 4 m Bundle taken on March 29, 2000. Scene identification numbers for the 5 scenes required to cover the study area were: (1) 20000239154547300000011603012, (2) 2000023915454730000001 1603013, (3) 20000239154547300000011603014, (4) 20000239154547300000011603 015, and (5) 20000239154547300000011603016. The average water level at the USGS gauging station located in Lake Kissimmee was 15.41 m AMSL on that date. Compact Airborne Spectrographic Imager The CASI hyperspectral sensor detects an array of narrow spectral bands in the visible and infrared wavelengths. The spectral range covered by the 288 channels is between 0.4 and 0.9 m. Each band covers a wavelength range of 0.018 m. While ground cell size depends on the altitude of the aircraft, the spectral bands measured and the bandwidths used are all programmable. For this study, data from 12 spectral bands were collected for portions of the western shore of Lakes Kissimmee and Hatchineha (Band 1 midpoint: 0.45 m, Band 2 midpoint: 0.54 m, Band 3 midpoint: 0.55 m, Band 4 midpoint: 0.56 m, Band 5 midpoint: 0.63 m, Band 6 midpoint: 0.65 m, Band 7 midpoint: 0.0.69 m, Band 8 midpoint: 0.70 m Band 9 midpoint: 0.77 m, Band 10

PAGE 45

35 midpoint: 0.78 m, Band 11 midpoint: 0.79 m, Band 12 midpoint: 0.83 m). The ground cell size for CASI imagery was 1 m x 1 m. The CASI images were acquired on March 20, 2002 for this project. The average water levels at the USGS gauging stations located in Lake Kissimmee and Lake Hatchineha were 15.38 m AMSL and 15.37 m AMSL, respectively. Images from each of the above sources have different ground cell sizes (Figure 3-1). Each type of image was used to test its ability to detect and display an indication of the OHWL. IRS 23 m x 23 m Landsat 30 m x 30 m 1 m Desired Margin of Error for Boundary Lines CASI 1 m x 1 m IKONOS 4 m x 4 m SPOT 20 m x 20 m Figure 3-1 Different ground cell sizes of imagery used in this study. The red line represents a desired margin of error ( m) when locating a boundary line. Image Processing The Landsat image was geometrically rectified to a Universal Transverse Mercator (UTM), Zone 17 projection (datum WGS 84) using ground control points taken from USGS digital orthophoto quarter quadrangles (DOQQ). Radiometric correction was

PAGE 46

36 not performed on the imagery because the objective of this study was not to compare land use/land cover change over time. All remaining images were geometrically rectified to the DOQQs. Images were classified using an unsupervised classification approach. In a multispectral image, each pixel has a spectral signature determined by the reflectance of that pixel in each of the spectral bands. Multispectral classification is a process whereby spectral signatures are analyzed and then pixels are assigned to classes based on similar signatures (Jensen 1996). This classification system is based on the principle that all pixels representing a homogeneous land cover type should have roughly the same spectral signature, with only relatively small differences due to variables such as shadows. Classification procedures attempt to group together such similar pixels. The detail of the classes depends on the spectral and spatial resolution (Jensen 2000). Unsupervised classification is a method in which a computer searches for natural groupings of similar pixels called clusters. The fewer clusters there are, the more the pixels within each cluster will vary in terms of spectral signature, and vice versa. ERDAS IMAGINE TM image processing software was used and unsupervised classification was performed using an algorithm called the Iterative Self-Organizing Data Analysis Technique (ISODATA). In this algorithm, the analyst inputs the number of clusters desired (six for this study) and a confidence threshold. The computer program will then build clusters iteratively, meaning with each new iteration, the clusters become more and more refined. The iterations stop when the confidence level is reached. The confidence threshold for this study was 95%. Once the clusters were built, they were assigned to a land cover class. The image clusters were then recoded into a GIS layer,

PAGE 47

37 which displayed each land cover class with a different color. Images were prepared with three classes, five classes, six classes, nine classes, twelve classes, and 25 classes. Final interpretation was done using seven class and six class designations for Lakes Hatchineha and Kissimmee, respectively. Using more classes made the image too noisy, and fewer classes did not provide enough contrast. The seven classes for Lake Hatchineha were as follows: Open water Hydric 1 Hydric 2 Hydric 3 Hydric 4 Cypress Mesic There is no cypress surrounding Lake Kissimmee, so the cypress class was eliminated for those images. The SPOT imagery for the study area was processed and classified in the same manner as the Landsat imagery. The IRS, IKONOS and CASI images were preprocessed radiometrically and geometrically. Images were then classified in the same manner as the Landsat and SPOT images. Visual Interpretation Upon completion of image processing, ground truth field surveys were conducted. When vegetation collection was complete, transect lines were drawn on Landsat, IRS, SPOT, IKONOS, and CASI images by first identifying the GPS coordinate corresponding to the OHWL, and then drawing a line along the azimuth recorded in the field. Images were visually interpreted in a qualitative assessment to identify changes in

PAGE 48

38 pixel class designations corresponding to vegetation shifts that may correlate with the OHWL. Vegetation Sampling The areas surrounding the lakes were divided into land use types and included ranching, natural, and urban. Distance along the lake margin was calculated for each land use type and transect sites (Figure 3-2) were randomly selected. In some cases, however, line transects were not taken at the exact point due to inaccessibility of the selected location. In such cases, line transects were run at the nearest accessible point. Transects 1, 2, 3A, 4, 5A, 6, 7, 8, 9, 10A were recorded on Lake Kissimmee. Transects 31, 32, 33, 34, 36, 37, 38 were recorded on Lake Hatchineha. Transects 9, 31, 32, and 33 correspond to the CASI images. Six transects were on privately owned ranch land, 3 at urban sites, and 8 in natural areas. Of the eight transects designated as natural, four were impacted by cattle ranching. Though these areas were designated either as state parks or water management district lands, the current drought has reduced water levels well below fence lines that would normally keep cattle from grazing on these properties. In all four cases, either cattle or cattle dung was a present. Real time kinematics GPS and other surveying techniques were used to find the vertical elevation corresponding to the known ordinary high water elevation for each transect. Without knowing the horizontal position of the line, the point was found by intersecting the known elevation with the shoreline. Transects were then set up starting from the OHWL and running perpendicular to the lake margin. Transects extended laterally 50 m below the OHWL and up to 50 m landward of the OHWL. In cases where the highest elevation occurred less than 50 m landward of the OHWL, with a lower

PAGE 49

39 Figure 3-2 Index map showing approximate transect locations elevation on the landward side, transects extended only to the point of highest elevation (Figure 3-3). Vegetation was recorded along the entire length of the transect and identified to species level (when feasible) with the expertise of botanist, David Hall, and some assistance from S. Barry Davis, N. Williams, M. Whitten, and K. Perkins of the University of Florida Herbarium. Vegetation was classified according to hydrologic indicators. Indicators in parentheses after species names are based on the National List of Plant Species That Occur in Wetlands (Reed 1988). If a plant could not be identified to species level, no designation was given due to variations of designations within the genera.

PAGE 50

40 Figure 3-3 Highest elevation along a transect. When highest elevation occurred less than 50 m (e.g., 23 m) landward of the OHWL, transect ended at that elevation. Vegetation Analysis Vegetation lists were composed for each transect and charted according to their location on the transect so that vegetation shifts could be easily identified (Figure 3-4). This information was used in the visual interpretation of the remote sensing imagery. Descriptive statistics were used to identify associations between vegetation and the OHWL. The purpose was to establish the accuracy of vegetation as an indicator of the OHWL. Frequency of occurrence was also determined for each vegetation species by Equation 3-1. Freq. of Occur. = (# of transects in which a species occurred)/(total # of transects surveyed) (3-1) Finally, a usability index was developed using Equation 3-2. Usability Index = [(Freq. of Occur)/(Minimum Avg. Distance)] 100 (3-2) The range of possible values for the Usability Index is from 0.1 to 1000. Higher values equate to better use of a species as an indicator of the OHWL. For example, if a species had a value of 1000 it would indicate that the species was present in every transect surveyed and that its distribution always began or ended at the OHWL

PAGE 51

41 5 4 3 2 1 -1 -2 -3 -4 -5 Species 1 Species 2 Species 3 of 4. Red line represents OHWL. Green line represents water edge for Species 2 with a value Green line represents land edge fo r Species 2 with a value of 1. Figure 3-4 Idealized vegetation transect chart and terminology used in vegetation analysis. In this figure, Species 2 has a landward edge value of 1 and a waterward edge value of 4. These values indicate that on the landward side of OHWL, Species 2 had a minimum distance of 1 m from OHWL, and on the waterward side, Species 2 had a minimum distance of 4 m from OHWL.

PAGE 52

CHAPTER 4 RESULTS Imagery Results Open water could be visually identified with Landsat, SPOT, IRS, IKONOS, and CASI imagery. Water could not be identified in areas with dense emergent wetland vegetation. All images had observable changes in pixel class designations, but no discernable change in class corresponded to the OHWL for Landsat, SPOT, and IRS images. The following figures are excerpts from the classified images. The contour associated with the elevation of 52.50 ft AMSL is shown in black on the Lake Hatchineha figures. Landsat Images There were changes in pixel class designations observable in the Landsat images, but none were indicative of OHWL. Figures 4-1 and 4-2 are sections of Landsat images classified into land cover classes for Lake Hatchineha and Lake Kissimmee, respectively. The changes in pixel class designations in the figures likely indicate a change in the plant community. However, these plant communities were not directly correlated with the OHWL. Indian Remote Sensing Images Pixel class designation differences are again evident in the IRS imagery, but none were indicative of the OHWL. As with Landsat imagery, pixel designations likely indicate a change in plant community, but there is no direct correlation between this change and the OHWL. Figures 4-3 and 4-4 are classified IRS images of Lakes 42

PAGE 53

43 Hatchineha and Kissimmee, respectively. The shift detectable in Figure 4-3 does not fall within 2 m of the OHWL. Figure 4-1 Transect 31, classified Landsat image of Lake Hatchineha. The small-scale image includes a square that is represented by the larger scale image to the right. The black line represents the contour of the known OHWL.

PAGE 54

44 Figure 4-2 Transect 9, classified Landsat image of Lake Kissimmee

PAGE 55

45 Figure 4-3 Transect 33, classified IRS image of Lake Hatchineha *Radiometric dropouts occurred in the image as indicated.

PAGE 56

46 Figure 4-4 Transect 9, classified IRS image of Lake Kissimmee

PAGE 57

47 The SPOT Images The multi-spectral SPOT imagery was classified with respect to land cover and the results are shown in Figures 4-5 and 4-6 for Lakes Hatchineha and Kissimmee, respectively. The changes in pixel class designations were not indicative of the OHWL. The IKONOS Images The IKONOS results are shown in Figures 4-7, 4-8, and 4-9, which correspond to Transects 31, 4, and 7, respectively. There appears to be a correlation between a change in the pixel class designation and the OHWL in Figure 4-9. The estimated accuracy of this edge is 4 m. The most prominent vegetation was Bahia grass (Paspalum notatum). This was an urban site with a high elevation of 16.65 m AMSL. Transect 7 extended only 34 m above the OHWL because a house blocked the way. Vegetation above the OHWL included FACU, FAC, FACW, and OBL species, but their combination resulted in classification as mesic. The CASI Images The CASI imagery exhibited the best results in terms of discerning plant community changes. A change in pixel class designation is visible approximately 3 m above the OHWL. Vegetation transect chart (Appendix A, Figure A-9) illustrates a similar shift in vegetation closer to 4 m above the OHWL. Five species (Euthamia caroliniana, Paspalum notatum, Sesbania herbacea, Axonopus furcatus, and Centella asiatica) occurred on both sides of the OHWL, while two occurred only above the OHWL and thirteen occurred only below the OHWL. The vegetation occurring above the OHWL included FACU, FAC, FACW, and OBL species, but still provided a distinct classification as mesic in this study. The following figures (4-10 and 4-11) illustrate the inability of the CASI hyperspectral scanner to distinguish between wetlands designations

PAGE 58

48 within the plant community. This results from the high degree of variance in the reflectance received by the scanner. Figure 4-5 Transect 31, classified SPOT image of Lake Hatchineha

PAGE 59

49 Figure 4-6 Transect 9, classified SPOT image of Lake Kissimmee

PAGE 60

50 Figure 4-7 Transect 31, classified IKONOS image of Lake Hatchineha

PAGE 61

51 Figure 4-8 Transect 4, classified IKONOS image of Lake Kissimmee

PAGE 62

52 Figure 4-9 Transect 7, classified IKONOS image of Lake Kissimmee CASI images were classified with respect to land cover type and the results are shown in Figures 4-12 and 4-13. Classified CASI imagery for Transects 31 and 32 on Lake Hatchineha was useful in detecting plant community changes corresponding to the OHWL. The change detected resulted from a spectral reflectance pattern sensed by CASI, and was located at the edge of the cypress canopy. No such change was visible for Transect 33 where the OHWL fell within the cypress. Inspection of the original

PAGE 63

53 classified images revealed the cypress pattern clearly, however, the excerpt shown in Figure 4-12 does not capture the details as well as the originals. There is a clear change in pixel classes at the OHWL for Transect 9 on Lake Kissimmee. A photograph taken at Transect 33 shows watermarks on the cypress trunks (Figure 4-14). 0.010.020.030.040.050.060.070.0open waterhydric 1hydric 2hydric 3hydric 4mesicHyperspectral Classes% Frequency of Occurance FAC FACU FACW OBL Figure 4-10 Histogram of FACU, FAC, FACW or OBL species classified in each CASI hyperspectral class. The variation within each class illustrates the inability of the imagery to discriminate wetland designations within this littoral zone.

PAGE 64

54 open waterhydric 1hydric 2hydric 3hydric 4mesic Taxodium distichumUrochloa muticaPanicum hemitomonPanicum repensPontederia cordataSagittaria lancifoliaPolygonum punctatumPouzolzia zeylanicaJuncus sp. Commelina diffusaPhyla nodifloraPaspalum notatumSesbania herbaceaEuthamia carolinianaAxonopus furcatusRhexia cubensisJuncus effususCentella asiaticaRhynchospora sp.Andropogon spAlternanthera philoxeroidesPtilimnium capillaceumLudwigia arcuataHydrocotyle umbellataBacopa carolinianaCynodon dactylonEleocharis 0.010.020.030.040.050.060.070.080.090.0100.0% FrequencyOccurranceSpecies Figure 4-11 Histogram of species classified in each CASI hyperspectral class. This further illustrates the variation of the species and the respective classes to which they were assigned.

PAGE 65

55 Figure 4-12 Transect 31, classified CASI image of Lake Hatchineha

PAGE 66

56 igure 4-13 Transect 9, classified CASI image of a portion of Lake Kissimmee F

PAGE 67

57 Figure 4-14 Vegetation in the area of Transect 33. Note watermarks on cypress trees. Table 4-1 summarizes the ability of both satellite and aircraft based images used in this study to show a change in pixel classes associated with the ordinary high water line. Column headings represent the transects and row headings represent the source of the images. Any transect with a corresponding image showing a change in pixel classes within 2 m of the OHWL is represented by a letter or a letter and a number within the cell as follows: H2 = Hydric 2 class, H3 = Hydric 3, H4 = Hydric 4 and M = Mesic. The value of 2 m was chosen based on a statistical analogy. If one assumes the 1 meter tolerance (mentioned earlier) to be a standard deviation, 2 meter would give a 95% confidence level. The Total column is the number of transects showing a change at the OHWL divided by the total number of transects represented in the imagery.

PAGE 68

58 Table 4-1 Summary table of imagery results. Dashes (-) indicate no data and nc indicates no change in pixel class designations within 2 m of the OHWL. The top entry within a cell indicates the pixel class on the landward side of the OHWL while the bottom entry indicates the pixel class on the waterward side. Transects Image 1 2 3A 4 5A 6 7 8 9 10A 31 32 33 34 36 37 38 % Total Landsat nc nc nc nc nc nc nc nc nc nc H3 H4 nc nc nc nc nc nc 5.88 IRS nc nc nc nc nc nc nc nc H3 H2 nc nc nc nc 7.69 SPOT nc nc nc nc nc nc nc nc nc nc nc H3 H2 nc nc nc nc nc 5.88 IKONOS nc nc nc nc M H4 nc nc nc nc nc nc nc H4 H3 nc 14.29 CASI M H2 H3 H2 H4 H2 nc 75.00 Vegetation Results Vegetation transect charts illustrating how vegetation fell out along each line transect are located in Appendix A, Figures A-1 through A-17 for both Lake Kissimmee and Lake Hatchineha. Table 4-2 and 4-3 summarize species data for combined lakes. Both tables include 23 species, which represent 25% of the total species identified to that level. The tables include the most frequently occurring species and the species found closest to the OHWL. There are only two species contained in both tables, Cyperus lecontei and Sesbania herbacea. Only three species dropped out at the OHWL: Bidens alba, Carex longii and Euthamia minor. A complete list of species sorted by frequency of occurrence and by average minimum distance from the OHWL for both combined lakes, and for each lake individually can be found in the Appendix B (Tables B-1 through B-6).

PAGE 69

59 Table 4-2 Twenty-three most frequently occurring species on the banks of Lake Hatchineha and Lake Kissimmee, combined. The number of occurrences represents the number of transects that the species occurred in. Frequency of occurrence is the number of occurrences divided by the total number of transects. Top 23 occurring species Number of occurrences Frequency of occurrence Pontederia cordata (OBL) 13 0.76 Polygonum punctatum (FACW+) 13 0.76 Hydrocotyle umbellata (OBL) 12 0.71 Paspalum notatum (FACU+) 10 0.59 Bacopa caroliniana (OBL) 9 0.53 Phyla nodiflora (FACW) 9 0.53 Centella asiatica (FACW) 9 0.53 Sagittaria lancifolia (OBL) 8 0.47 Alternanthera philoxeroides (OBL) 8 0.47 Sesbania herbacea (FACW-) 8 0.47 Panicum repens (FACW-) 7 0.41 Eupatorium capillifolium (FACU) 7 0.41 Taxodium distichum (OBL) 6 0.35 Axonopus furcatus (OBL) 6 0.35 Ptilimnium capillaceum (OBL) 6 0.35 Panicum hemitomon (OBL) 5 0.29 Hydrochloa caroliniensis (OBL) 5 0.29 Euthamia caroliniana (FAC) 5 0.29 Diodia virginiana (FACW) 4 0.24 Quercus geminata oak 4 0.24 Myrica cerifera (FAC+) 4 0.24 Cyperus polystachyos (FACW) 4 0.24 Cyperus lecontei (FACW) 4 0.24

PAGE 70

60 Table 4-3 Average minimum distance from the OHWL that the closest species were found for Lake Hatchineha and Lake Kissimmee, combined. Species with a value of 0.1 occurred at the edge of the OHWL, but for calculation purposes were given a value other than 0. Species name Average minimum distance from OHWL (m) Bidens alba (FACW-) 0.10 Caxex longii (OBL) 0.10 Euthamia minor (FAC) 0.10 Coreopsis leavenworthii (FACW) 1.00 Desmodium incanum 1.00 Digitaria serotina (FAC) 1.00 Cuphea carthagenensis (FACW) 1.05 Erechtites hieraciifolia (FAC-) 2.00 Hypericum mutilum (FACW) 2.00 Lythrum alatum (FACW+) 2.00 Rhynchospora microcarpa (FACW+) 2.37 Baccharis halimifolia (FAC) 2.50 Eryngium prostratum (FACW) 2.67 Ambrosia artemisiifolia (FACU) 3.00 Diospyros virginiana (FAC) 3.00 Juncus dichotomus (FACW) 3.00 Murdannia nudiflora (FAC) 3.00 Schinus terebinthifolius (FAC) 3.00 Cyperus lecontei (FACW) 3.75 Cyperus rotundus (FAC-) 4.00 Juncus effusus (FACW+) 4.00 Rhexia cubensis (FACW+) 4.00 Sesbania herbacea (FACW-) 4.40 The following tables give the top ten species sorted by the usability index (UI) for: Combined lakes and land use types All land use types on Lake Hatchineha only All land use types on Lake Kissimmee only All ranch land of combined lakes All natural areas of combined lakes All urban areas of combined lakes Only Sesbania herbacea occurred in five of these tables with a high UI value of 18.44 (out of a possible 1,000) and a low UI value of 10.7. Urena lobata, Bidens alba and

PAGE 71

61 Carex longii had the highest UI values of 333.33, with Bidens alba and Carex longii showing up in three of the six tables. Three species had consistently high UI values in each of the tables where they show up: Bidens alba, Carex longii, and Euthamia minor. Table 4-4 Ten species with the highest usability index for combined lakes and land use types Species name Usability index Bidens alba (FACW-) 58.82 Caxex longii (OBL) 58.82 Euthamia minor (FAC) 58.82 Centella asiatica (FACW) 11.32 Cuphea carthagenensis (FACW) 11.20 Sesbania herbacea (FACW-) 10.70 Alternanthera philoxeroides (OBL) 9.88 Ptilimnium capillaceum (OBL) 7.84 Phyla nodiflora (FACW) 7.80 Paspalum notatum (FACU+) 7.70 Table 4-5 Twelve species with highest usability index for all land use types of Lake Hatchineha only. Twelve species are included because species 10, 11, and 12 all have the same ranking on the UI. Species name Usability index Bidens alba (FACW-) 142.86 Quercus geminata oak 14.29 Baccharis halimifolia (FAC) 11.43 Sesbania herbacea (FACW-) 11.43 Phyla nodiflora (FACW) 10.83 Ptilimnium capillaceum (OBL) 9.89 Hydrocotyle umbellata (OBL) 9.14 Eupatorium capillifolium (FACU) 8.40 Taxodium distichum (OBL) 7.65 Thelypetris interrupta (FAC+) 7.14 Erechtites hieraciifolia (FAC-) 7.14 Cuphea carthagenensis (FACW) 7.14

PAGE 72

62 Table 4-6 Ten species with highest usability index for all land use types of Lake Kissimmee only Species name Usability index Caxex longii (OBL) 100.00 Euthamia minor (FAC) 100.00 Cuphea carthagenensis (FACW) 100.00 Centella asiatica (FACW) 16.80 Alternanthera philoxeroides (OBL) 13.81 Rhynchospora microcarpa (FACW+) 12.68 Sesbania herbacea (FACW-) 11.92 Eryngium prostratum (FACW) 11.25 Cyperus lecontei (FACW) 10.67 Paspalum notatum (FACU+) 10.38 Table 4-7 Ten species with highest usability index for ranch areas of combined lakes Species name Usability index Euthamia minor (FAC) 166.67 Cuphea carthagenensis (FACW) 166.67 Centella asiatica (FACW) 21.35 Myrica cerifera (FAC+) 16.67 Rhynchospora microcarpa (FACW+) 13.07 Sesbania herbacea (FACW-) 11.54 Euthamia caroliniana (FAC) 11.11 Eryngium prostratum (FACW) 11.11 Axonopus furcatus (OBL) 9.38 Paspalum notatum (FACU+) 8.45 Table 4-8 Ten species with highest usability index for natural areas of combined lakes Species name Usability index Urena lobata (FACU) 125.00 Axonopus furcatus (OBL) 125.00 Sesbania herbacea (FACW-) 18.44 Paspalum notatum (FACU+) 16.67 Scoparia dulcis (FAC) 12.50 Cyperus polystachyos (FACW) 12.50 Digitaria serotina (FAC) 12.50 Cyperus lecontei (FACW) 12.50 Ptilimnium capillaceum (OBL) 12.50 Phyla nodiflora (FACW) 11.70

PAGE 73

63 Table 4-9 Ten species with highest usability index for urban areas of combined lakes Species name Usability index Urena lobata (FACU) 333.33 Bidens alba (FACW-) 333.33 Caxex longii (OBL) 333.33 Alternanthera philoxeroides (OBL) 60.00 Sida acuata 33.33 Baccharis halimifolia (FAC) 33.33 Coreopsis leavenworthii (FACW) 33.33 Desmodium incanum 33.33 Eryngium prostratum (FACW) 33.33 Polygonum punctatum (FACW+) 27.03

PAGE 74

CHAPTER 5 DISCUSSION Remote Sensing Aspects Case law requires that the best methods available should be used for making ordinary high water determinations. There is little potential for using satellite remote sensing to locate the OHWL. The degree of usefulness is related to the type of imagery used and to the nature of the shoreline. The areas studied in this project were broad, low bank shorelines. Landsat, SPOT, and IRS were not usefulness because of their relatively large ground cell size. Townsend and Walsh (2001) reported that complex community relationships and the nature of their spatial variability are rarely captured with satellite imagery. Broad ecotones are especially difficult to portray accurately because they are made up of transition zones that grade into each other across complex soil, hydrologic, and topographic gradients. Evidence from this study supports that contention. The variety of vegetation types found in different areas in this study impacts the spectral signatures, leading to a range of classes, none of which led to a clear demarcation of the OHWL at or near the upper littoral zone. As illustrated in the summary table of pixel class designation changes corresponding to the known 52.50 ft (Hatchineha) or 52.00 ft (Kissimmee) AMSL ordinary high water line for each of the images used in this study (Table 4-1), the only one that emerged as having potential to delineate a boundary line such as the OHWL was CASI. In the case of Landsat and SPOT, only one of 17 (5.88%) of the transects had a change in pixel class designations within 2 m of the OHWL. The pixel class designation 64

PAGE 75

65 on the landward side of the OHWL for Landsat Transect 31 was Hydric 3 and the waterward side was Hydric 4. For SPOT Transect 32, the landward side was Hydric 3 and the waterward side was Hydric 2. The IRS image was able to detect one change in pixel class designations within 2 m of the OHWL for a total of one of thirteen (7.69%). The IRS image corresponded to a change at Transect 33. The landward side of the OHWL was classified as Hydric 3 while the waterward side was classified as Hydric 2. This transect corresponds to an area of Lake Hatchineha where the OHWL falls within the cypress. Even CASI, with its smaller ground cell size, was unable to detect a shift at the OHWL for this transect. IKONOS detected a change in pixel class designations corresponding to the OHWL for two of fourteen transects (14.29%). These results are better, but the change in pixel class designation may not be within the desired 1 m margin of error due to the 4 m ground cell size. Even though each of the satellites was able to detect a shift corresponding to the OHWL for at least one transect, results were not consistent. Consequently, none were useful for locating a boundary line such as the OHWL. Landsat, SPOT and IRS have the largest ground cell size of the sensors tested and showed virtually no potential for delineating a boundary line or detecting subtle vegetation shifts on the ground. Fuller et al. (1998) found that Landsat was even unable to distinguish between a swamp with five species and a swamp with 33 species. Landsat has been shown to be useful for classifying broad categories of land cover, and it is therefore likely that SPOT and IRS, which have slightly smaller ground cell sizes, would also be useful for broad classification schemes. However, parameters for this study were much narrower, requiring the delineation of a boundary line.

PAGE 76

66 As Harvey and Hill (2001) reported, the characteristic vegetation gradients found in wetlands are narrower than the pixel size of current sensors, which often results in the inability to distinguish between different vegetation types. It is conceivable that use of a different classification scheme would have improved the discrimination capabilities of this study. Townsend (2000) assessed the quantitative fuzzy approach to assess classifications of mapped vegetation. In this approach, fuzzy classifications and mixture/subpixel models provide information about the degree of similarity between ground locations of natural vegetation and multiple classes on a map. It facilitates a more detailed interpretation of map classification and, therefore, reduces the arbitrariness of evaluations. It also provides a greater understanding of the subtleties in species composition and how they relate to the fuzziness of the natural vegetation classes. In order to perform such a classification, it is necessary to know the ecological relationships within plant communities. The vegetation data must be characterized as a percent of an area (e.g., 100% cypress or 86% soil). Vegetation information obtained from Lakes Hatchineha and Kissimmee for this study was only determined along line transects, so there was no way to assess percent of land cover. However, even if a fuzzy classification could have been done, it is doubtful that a narrow boundary line could have been detected from Landsat, SPOT or IRS images because of their resolution limitations. Another constraint of this study was the lack of specific ground cover information to correlate to the satellite images. Conditions on the ground had almost certainly changed between the time that Landsat, SPOT and IRS images were taken (October 1999, March 1993 and December 1999, respectively) and the time vegetation was sampled in April and May 2002. However, this constraint was known at the outset and

PAGE 77

67 was not considered significant to the conclusions of this study. The generally accepted OHWL was known for both lakes, and vertical elevations measured in the field placed the starting point for each transect on the OHWL. GPS coordinates allowed placement of each transects point associated with the OHWL on the satellite images. Even though vegetation had probably changed in the years since the satellite images were recorded, the vertical elevation had probably not. Therefore, once the point corresponding to the OHWL was located on the image, a change in pixel class designations associated with that location would be discernible if one existed. A change would indicate a vegetation shift corresponding to the OHWL. Since this study did not attempt to assess change in land use/land cover, it was unnecessary to know what vegetation was present at the time the images were taken. The key constraint was ground cell size, which determines the ability to discern very small changes in vegetation or to discern a boundary line such as OHWL within the preferred 1 m margin of error. Because of their relatively large ground cell sizes, Landsat, SPOT and IRS were not well suited for this task. The IKONOS satellite, with a smaller ground cell size of 4 m, theoretically has more potential to provide information that would be useful in locating the OHWL. However, the images used in this study demonstrated only one change each at the OHWL for Lake Hatchineha and Lake Kissimmee using this sensor. The area where the OHWL coincided with a distinct change in pixel class designation on Lake Kissimmee (Transect 7) corresponded to an area with a predominance of Bahia grass (Paspalum notatum), although six other species also occurred in this area. Bahia grass is common in both ranching and urban sites around Lake Kissimmee. Transect 7 is an urban site, and it is likely that the Bahia recorded in this study was also present at the time these images were

PAGE 78

68 taken in March 2000. Interestingly, Bahia does not appear on the list of species that occur closest to the OHWL, and had a usability index (UI) rating of only 16.67 (out of a possible 1000). Therefore, Bahia could not be relied upon as a good indicator of the OHWL. It is likely that the pixel class change seen in the IKONOS images related to a natural berm on the shore. Perhaps the sensors were detecting more sand and dirt because vegetation was not as dense as it was in other areas. Furthermore, because the ground cell size of these images was 4 m, the ability to locate a boundary line within a 1 m margin of error is limited. CASI hyperspectral imagery, with a ground cell size of 1 m, had the best results in terms of discerning plant community changes. However, this result is based on a small sample size. In the summary table (Table 4-1), CASI images showed the best results with three of four transects (75%) detecting a vegetation shift corresponding to the OHWL. CASIs 1 m ground cell size is within the desired 1 m margin of error needed for defining a boundary line. The OHWL coincided with a break in pixel class designations on the images for two of three transects (Transects 31 and 32) on Lake Hatchineha. However, in both of these transects the break in pixel class designations at the OHWL coincided with the edge of the cypress canopy. In the third Lake Hatchineha transect (Transect 33), the OHWL fell within the cypress canopy and could not be detected in the CASI imagery. On Lake Kissimmee, only Transect 9 was included in the CASI imagery data set. There was a clear change in pixel class designations at the OHWL, but it was not clear what caused this change. There were no distinct vegetation shifts on the ground at the OHWL, although a natural berm did occur there. It is possible that the sensors picked up

PAGE 79

69 reflections at different angles because of this berm, resulting in the classification change in the CASI imagery. Additional studies should be conducted with CASI imagery to more accurately assess its ability to demonstrate changes in pixel classes corresponding to the OHWL. However, because CASI imagery is expensive compared to satellite imagery (CASI $18,215/km 2 vs. Landsat $0.60/km 2 ), its use as a cost effective means of locating the OHWL may be limited. One of the largest hurdles in attempting an automated approach to the determination of the OHWL using remote sensing was the variation in vegetation indicators. For example, oak trees (Quercus geminata) around Lake Kissimmee are good indicators of the uppermost reach of water. The OHWL sometimes fell at the edge of oaks, but never beyond them. This upland species of oak can be detected with remote sensing and, therefore, can sometimes be used to delineate a line beyond which water will not be present. Oaks around Lake Kissimmee sit on a natural berm and the OHWL usually falls somewhere on the lakeside of the berm. However, many oaks are far removed from the lake, growing beyond small wetlands with a wide variety of vegetation types in between, and therefore, are not good indicators of the uppermost reach of water in those areas. Lake Hatchineha and some other Florida lakes are surrounded by cypress (Taxodium) stands. When the ordinary high water stage occurs within a cypress stand, it would be nearly impossible to detect this water using the imagery available for this study. Figure 5-1 illustrates this circumstance by showing an area of Lake Hatchineha with water present within the cypress that was not clearly detectable with the satellite or CASI imagery.

PAGE 80

70 Figure 5-1 Catfish Creek entrance to Lake Hatchineha. Note water within the cypress. Another drawback related to the use of remote sensing for locating the OHWL relates to the nature of low bank lakes such as Lakes Kissimmee and Hatchineha. This characteristic presents a special problem in finding evidence of the OHWL using remote sensing techniques. Slight water elevation changes cause large shifts in the lateral position of the shore. This results in large littoral zones with diverse vegetation types and becomes problematic for determining distinct vegetation indicators of the OHWL, particularly by remote sensing. On the other hand, remote sensing may detect a line at ordinary high water for steep bank lakes because vegetation may not be as prevalent along a steep bank, and open water would provide a distinct spectral signature.

PAGE 81

71 There are broad ecological uses for satellite remote sensing in delineating land use/land cover, but its application for accurately locating a distinct line is limited, especially on low bank lakes. This limitation is largely due to the fact that vegetation shifts on the ground are not clearly evident in imagery of these resolutions. Even under normal hydrologic regimes when vegetation would likely fall out in clearer patterns, satellite imagery would probably not be able to accurately ascertain the location of the OHWL because of the difference between the narrow vegetation bands and the large ground cell size of the imagery. CASI has the greatest potential because of its 1 m ground cell size, but may be limited due to its expense. Further study of CASI imagery should be conducted to determine whether it can consistently detect spectral shifts corresponding to the OHWL. Vegetation Aspects There is historic evidence that water levels were significantly higher and less aquatic vegetation existed in the lakes prior to the channelization of the Kissimmee River. Hydrologic modifications have resulted in stabilized water levels and have altered vegetation associations within the marshes. Lake Kissimmee and its associated marsh now have floating tussocks, mixed grass marshes, and nearly floating pickerelweed marshes (Kushlan 1990). At the time of this study, south Florida had experienced an extended drought that very likely caused further alteration of the vegetation community. According to local ranchers and fishermen, since lake levels are now controlled, water flow in and out of the lakes is reduced and hydrilla and water hyacinth have become problems in the lakes as well. The increased vegetation has resulted in a need for periodic lake draw-downs and muck removal. The resulting waste is piled up along various locations of the lakeshore and creates spoil islands. These islands further alter

PAGE 82

72 plant communities as they move seed banks around and place muck and debris in areas that were once either open water are flat shoreline. Vegetation analysis elicited no good vegetation indicators of the OHWL. Shorelines are dynamic environments and during periods of drought, as existed in the study area, existing seed banks may germinate, increasing the diversity of species close to the lakeshore (Keddy and Fraser 2000). It is likely that the FACU vegetation seen in the lower littoral zone occurred as the result of prolonged low water levels in the lakes. Furthermore, cattle ranching in many areas has impacted Lake Kissimmee, and vegetation types seen in these areas include improved pasture consisting of non-native species. The result is a conglomeration of FACU, FAC, FACW, and even OBL species in many areas of the upper and lower littoral zones. Normally, the lower littoral zone would have standing water or be adjacent to it (Davis 1973). Obligate wetland vegetation would be expected in this zone, including species such as Pontedaria cordata, Sagittaria lancifolia, Typha latifolia, and Panicum hemitomon. These species did occur in the lower littoral zones of Lakes Kissimmee and Hatchineha, but they failed to fall into a consistent pixel class designation on the classified images. In vegetation analysis, frequency of occurrence data showed that only seven species (out of 93 total) showed up in more than half the transects thus indicating that there were no consistent vegetation types that could be used as potential indicators of the OHWL. When lakes were divided by land use type, the highest usability index value (333.33) occurred within the urban class. However, because the three species with a 333 UI occurred in only one of 17 transects (Bidens alba and Carex longii) or three of 17 transects (Urena lobata), they are of limited importance. It appears that when these

PAGE 83

73 species occur, they occur at the edge of the OHWL, but because these species were not abundant, they could not be relied upon as good indicators of the OHWL. The fact that the frequency of occurrence table and the average minimum distance table shared only two common species serves as further evidence that vegetation was not very useful for locating the OHWL. The most frequently occurring species (Polygonum punctatum, Pontedaria cordata, Hydrocotyle umbellata and Paspalum notatum) occurred in 13, 13, 12, or 10 of the 17 transects, respectively, but none of them consistently occurred near the OHWL, again confirming that there were no good vegetation indicators of the OHWL. Conclusions If one wanted to obtain only a crude approximation of the acreage of sovereignty land in Florida, IKONOS could be used. However, in order to obtain an accurate determination of acreage, the OHWL must be pinpointed because it serves as the boundary line between sovereignty and private land. To that end, the only potentially useful remote sensing technology assessed in this study was CASI. Landsat, SPOT and IRS would not provide a reasonable estimate because their ground cell size limits their ability to distinguish vegetation shifts corresponding to the OHWL. Even IKONOS is limited because the 4 m ground cell size does not meet the desired 1 m margin of error. Consequently, this study will not resolve the continuing legal battles surrounding the OHWL. Case law has dealt with this issue for nearly a century. The hope that future legal battles would be easily settled because the OHWL could quickly be pinpointed using remote sensing technologies has been dampened. The results of this study found no good and inexpensive technology for the job.

PAGE 84

74 In terms of vegetation analysis, there were no consistent indicators of the OHWL found in this study. Those vegetation types that occurred at or near the OHWL did not occur frequently enough to be useful. However, the drought conditions experienced in the lakes prior to and during the sampling period have likely altered the vegetation significantly and it would be useful to reassess vegetation indicators under more normal hydrologic regimes. Under more normal conditions, vegetation would be expected to fall within distinct hydrologic zones. It would be expected that areas closest to the water would be dominated by OBL and FACW species, while FAC and FACU species would be found further landward. At the time of this study, FAC and FACU species were often found well below the OHWL. Use of quadrat field investigation method rather than the line transect method to collect and analyze vegetation may also have been more informative. Quadrats would have provided information about percent coverage of vegetation and could have been used in combination with fuzzy classification schemes to assess the ability of various remote sensing technologies to detect a shift at the OWHL. Elevation information would also have been useful to collect in combination with vegetation. Hydrology and vegetation are inextricably linked, and knowing vegetation elevation would have provided useful information that linked hydroperiods and frequency of inundation to vegetation by comparing elevations to lake stage data. It is possible that under normal hydrologic conditions, vegetation would have been a useful indicator of the OHWL as Davis found in his 1973 research.

PAGE 85

APPENDIX A VEGETATION CHARTS In each of the following charts, the red line represents the OHWL. Each column represents a 1 meter distance on the ground. Columns with positive numbers represent land above the OHWL while columns with negative numbers represent land below the OHWL. Any vegetation that occurred along the transect was recorded and is represented by a colored box in the appropriate column. Wetland designations are noted in parenthesis next to the species name. Vegetation listed without a wetland designation are those not listed in the National List of Plant Species that Occur in Wetlands. 75

PAGE 86

76 igure A-1 Vegetation Transect 1, ranch land use, Lake Kissimmee F

PAGE 87

Figure A-2 Vegetation Transect 2, ranch land use, Lake Kissimmee 77

PAGE 88

Figure A-3 Vegetation Transect 3A, natural area land use, Lake Kissimmee 78

PAGE 89

Figure A-4 Vegetation Transect 4, ranch land use, Lake Kissimmee 79

PAGE 90

Figure A-5 Vegetation Transect 5A, ranch land use, Lake Kissimmee 80

PAGE 91

Figure A-6 Vegetation Transect 6, natural area land use, Lake Kissimmee 81

PAGE 92

82 Figure A-7 Vegetation Transect 7, urban land use, Lake Kissimmee

PAGE 93

Figure A-8 Vegetation Transect 8, urban land use, Lake Kissimmee 83

PAGE 94

84 Figure A-9 Vegetation Transect 9, ranch land use, Lake Kissimmee

PAGE 95

Figure A-10 Vegetation Transect 10A, natural area land use, Lake Kissimmee 85

PAGE 96

Figure A-11 Vegetation Transect 31, natural area land use, Lake Hatchineha 86

PAGE 97

Figure A-12 Vegetation Transect 32, natural area land use, Lake Hatchineha 87

PAGE 98

Figure A-13 Vegetation Transect 33, natural area land use, Lake Hatchineha 88

PAGE 99

89 Figure A-14 Vegetation Transect 34, natural area land use, Lake Hatchineha

PAGE 100

90 Figure A-15 Vegetation Transect 36, ranch land use, Lake Hatchineha

PAGE 101

91 Figure A-16 Vegetation Transect 37, natural area land use, Lake Hatchineha

PAGE 102

92 Figure A-17 Vegetation Transect 38, urban land use, Lake Hatchineha

PAGE 103

APPENDIX B VEGETATION TABLES

PAGE 104

Table B-1 List of species ordered by frequency of occurrence for Lake Hatchineha and Lake Kissimmee combined. The number of occurrences represents the number of transects that the species occurred in. Frequency of occurrence is the number of occurrences divided by the total number of transects. Species name Number of occurrences Frequency of occurrence Pontederia cordata (OBL) 13 0.76 Polygonum punctatum (FACW+) 13 0.76 Hydrocotyle umbellata (OBL) 12 0.71 Paspalum notatum (FACU+) 10 0.59 Bacopa caroliniana (OBL) 9 0.53 Phyla nodiflora (FACW) 9 0.53 Centella asiatica (FACW) 9 0.53 Sagittaria lancifolia (OBL) 8 0.47 Alternanthera philoxeroides (OBL) 8 0.47 Sesbania herbacea (FACW-) 8 0.47 Panicum repens (FACW-) 7 0.41 Eupatorium capillifolium (FACU) 7 0.41 Eleocharis sp. 6 0.35 Taxodium distichum (OBL) 6 0.35 Axonopus furcatus (OBL) 6 0.35 Ptilimnium capillaceum (OBL) 6 0.35 Panicum hemitomon (OBL) 5 0.29 Hydrochloa caroliniensis (OBL) 5 0.29 Euthamia caroliniana (FAC) 5 0.29 Diodia virginiana (FACW) 4 0.24 Quercus geminata oak 4 0.24 Myrica cerifera (FAC+) 4 0.24 Cyperus polystachyos (FACW) 4 0.24 Cyperus lecontei (FACW) 4 0.24 Typha latifolia (OBL) 3 0.18 Sabal palmetto (FAC) 3 0.18 Cynodon dactylon (FACU) 3 0.18 Richardia brasiliensis 3 0.18 Juncus sp. 3 0.18 Sida acuata 3 0.18 Vitis rotundifolia (FAC) 3 0.18 Urena lobata (FACU) 3 0.18 Urochloa mutica (FACW) 3 0.18 Eryngium prostratum (FACW) 3 0.18 Rhynchospora microcarpa (FACW+) 3 0.18 Conyza canadensis (FACU) 2 0.12 Scoparia dulcis (FAC) 2 0.12 Paspalum conjugatum (FAC) 2 0.12 94

PAGE 105

95 Table B-1 Continued Species name Number of occurrences Frequency of occurrence Thelypetris interrupta (FAC+) 2 0.12 Bacopa monnieri (OBL) 2 0.12 Ludwigia arcuata (OBL) 2 0.12 Cyperus croceus (FAC) 2 0.12 Eclipta prostrata 2 0.12 Lindernia grandiflora (OBL) 2 0.12 Pouzolzia zeylanica 2 0.12 Baccharis halimifolia (FAC) 2 0.12 Cuphea carthagenensis (FACW) 2 0.12 Nymphia odorata (OBL) 1 0.06 Phytolacca americana (FACU+) 1 0.06 Solanum diphyllum 1 0.06 Nelumbo lutea (OBL) 1 0.06 Chenopodium ambrosioides (FACU) 1 0.06 Momordica charantia 1 0.06 Plantago virginica (FACU-) 1 0.06 Smilax bona-nox (FAC) 1 0.06 Ludwigia peruviana (OBL) 1 0.06 Kyllinga brevifolia (FACW) 1 0.06 Bidens laevis (OBL) 1 0.06 Cyperus odoratus (FACW) 1 0.06 Cyperus articulatus (OBL) 1 0.06 Scleria ciliata (FAC) 1 0.06 Cyperus surinamensis (FACW) 1 0.06 Cirsium nuttallii (FAC) 1 0.06 Commelina diffusa (FACW) 1 0.06 Lycopus rubellus (OBL) 1 0.06 Salix caroliniana (OBL) 1 0.06 Cyperus haspan (OBL) 1 0.06 Hibiscus grandiflorus (OBL) 1 0.06 Sacciolepis striata (OBL) 1 0.06 Toxicodendron radicans (FAC) 1 0.06 Lachnocaulon engleri (FACW+) 1 0.06 Ludwigia palustris (OBL) 1 0.06 Cyperus strigosus (FACW) 1 0.06 Teucrium canadense (FACW-) 1 0.06 Cyperus rotundus (FAC-) 1 0.06 Juncus effusus (FACW+) 1 0.06 Rhexia cubensis (FACW+) 1 0.06 Ambrosia artemisiifolia (FACU) 1 0.06 Diospyros virginiana (FAC) 1 0.06

PAGE 106

96 Table B-1 Continued Species name Number of occurrences Frequency of occurrence Juncus dichotomus (FACW) 1 0.06 Murdannia nudiflora (FAC) 1 0.06 Schinus terebinthifolius (FAC) 1 0.06 Erechtites hieraciifolia (FAC-) 1 0.06 Hypericum mutilum (FACW) 1 0.06 Lythrum alatum (FACW+) 1 0.06 Coreopsis leavenworthii (FACW) 1 0.06 Desmodium incanum 1 0.06 Digitaria serotina (FAC) 1 0.06 Bidens alba (FACW-) 1 0.06 Caxex longii (OBL) 1 0.06 Euthamia minor (FAC) 1 0.06

PAGE 107

97 Table B-2 List of species ordered by frequency of occurrence for Lake Hatchineha Species name Number of occurrences Frequency of occurrence Polygonum punctatum (FACW+) 7 1.00 Taxodium distichum (OBL) 6 0.86 Pontederia cordata (OBL) 5 0.71 Phyla nodiflora (FACW) 4 0.57 Hydrocotyle umbellata (OBL) 4 0.57 Eupatorium capillifolium (FACU) 4 0.57 Ptilimnium capillaceum (OBL) 3 0.43 Alternanthera philoxeroides (OBL) 3 0.43 Urena lobata (FACU) 3 0.43 Vitis rotundifolia (FAC) 3 0.43 Panicum hemitomon (OBL) 3 0.43 Sida acuata 3 0.43 Sagittaria lancifolia (OBL) 3 0.43 Baccharis halimifolia (FAC) 2 0.29 Sesbania herbacea (FACW-) 2 0.29 Pouzolzia zeylanica 2 0.29 Eclipta prostrata 2 0.29 Urochloa mutica (FACW) 2 0.29 Bacopa caroliniana (OBL) 2 0.29 Cynodon dactylon (FACU) 2 0.29 Panicum repens (FACW-) 2 0.29 Richardia brasiliensis 2 0.29 Paspalum conjugatum (FAC) 2 0.29 Sabal palmetto (FAC) 2 0.29 Myrica cerifera (FAC+) 2 0.29 Juncus sp. 2 0.29 Bidens alba (FACW-) 1 0.14 Quercus geminata oak 1 0.14 Thelypetris interrupta (FAC+) 1 0.14 Erechtites hieraciifolia (FAC-) 1 0.14 Cuphea carthagenensis (FACW) 1 0.14 Paspalum notatum (FACU+) 1 0.14 Cyperus rotundus (FAC-) 1 0.14 Centella asiatica (FACW) 1 0.14 Cyperus strigosus (FACW) 1 0.14 Teucrium canadense (FACW-) 1 0.14 Sacciolepis striata (OBL) 1 0.14 Toxicodendron radicans (FAC) 1 0.14 Lycopus rubellus (OBL) 1 0.14 Commelina diffusa (FACW) 1 0.14

PAGE 108

98 Table B-2 Continued Species name Number of occurrences Frequency of occurrence Cirsium nuttallii (FAC) 1 0.14 Hydrochloa caroliniensis (OBL) 1 0.14 Cyperus articulatus (OBL) 1 0.14 Eleocharis sp. 1 0.14 Momordica charantia 1 0.14 Smilax bona-nox (FAC) 1 0.14 Chenopodium ambrosioides (FACU) 1 0.14 Solanum diphyllum 1 0.14 Diodia virginiana (FACW) 1 0.14 Typha latifolia (OBL) 1 0.14

PAGE 109

99 Table B-3 List of species ordered by frequency of occurrence for Lake Kissimmee Species name Number of occurrences Frequency of occurrence Paspalum notatum (FACU+) 9 0.90 Centella asiatica (FACW) 8 0.80 Hydrocotyle umbellata (OBL) 8 0.80 Pontederia cordata (OBL) 8 0.80 Bacopa caroliniana (OBL) 7 0.70 Sesbania herbacea (FACW-) 6 0.60 Axonopus furcatus (OBL) 6 0.60 Polygonum punctatum (FACW+) 6 0.60 Alternanthera philoxeroides (OBL) 5 0.50 Euthamia caroliniana (FAC) 5 0.50 Phyla nodiflora (FACW) 5 0.50 Panicum repens (FACW-) 5 0.50 Eleocharis sp. 5 0.50 Sagittaria lancifolia (OBL) 5 0.50 Cyperus lecontei (FACW) 4 0.40 Cyperus polystachyos (FACW) 4 0.40 Hydrochloa caroliniensis (OBL) 4 0.40 Rhynchospora microcarpa (FACW+) 3 0.30 Eryngium prostratum (FACW) 3 0.30 Ptilimnium capillaceum (OBL) 3 0.30 Eupatorium capillifolium (FACU) 3 0.30 Diodia virginiana (FACW) 3 0.30 Quercus geminata oak 3 0.30 Myrica cerifera (FAC+) 2 0.20 Lindernia grandiflora (OBL) 2 0.20 Cyperus croceus (FAC) 2 0.20 Ludwigia arcuata (OBL) 2 0.20 Bacopa monnieri (OBL) 2 0.20 Scoparia dulcis (FAC) 2 0.20 Conyza canadensis (FACU) 2 0.20 Panicum hemitomon (OBL) 2 0.20 Typha latifolia (OBL) 2 0.20 Caxex longii (OBL) 1 0.10 Euthamia minor (FAC) 1 0.10 Cuphea carthagenensis (FACW) 1 0.10 Coreopsis leavenworthii (FACW) 1 0.10 Desmodium incanum 1 0.10 Digitaria serotina (FAC) 1 0.10 Hypericum mutilum (FACW) 1 0.10 Lythrum alatum (FACW+) 1 0.10

PAGE 110

100 Table B-3 Continued Species name Number of occurrences Frequency of occurrence Urochloa mutica (FACW) 1 0.10 Juncus sp. 1 0.10 Ambrosia artemisiifolia (FACU) 1 0.10 Diospyros virginiana (FAC) 1 0.10 Juncus dichotomus (FACW) 1 0.10 Murdannia nudiflora (FAC) 1 0.10 Schinus terebinthifolius (FAC) 1 0.10 Juncus effusus (FACW+) 1 0.10 Rhexia cubensis (FACW+) 1 0.10 Lachnocaulon engleri (FACW+) 1 0.10 Ludwigia palustris (OBL) 1 0.10 Cyperus haspan (OBL) 1 0.10 Hibiscus grandiflorus (OBL) 1 0.10 Salix caroliniana (OBL) 1 0.10 Cyperus surinamensis (FACW) 1 0.10 Richardia brasiliensis 1 0.10 Scleria ciliata (FAC) 1 0.10 Cyperus odoratus (FACW) 1 0.10 Bidens laevis (OBL) 1 0.10 Kyllinga brevifolia (FACW) 1 0.10 Ludwigia peruviana (OBL) 1 0.10 Plantago virginica (FACU-) 1 0.10 Thelypetris interrupta (FAC+) 1 0.10 Nelumbo lutea (OBL) 1 0.10 Cynodon dactylon (FACU) 1 0.10 Sabal palmetto (FAC) 1 0.10 Nymphia odorata (OBL) 1 0.10 Phytolacca americana (FACU+) 1 0.10

PAGE 111

101 Table B-4 Average minimum distance from the OHWL for Lake Hatchineha and Lake Kissimmee, combined Species name Average minimum distance from OHWL (m) Coreopsis leavenworthii (FACW) 1.00 Desmodium incanum 1.00 Digitaria serotina (FAC) 1.00 Bidens alba (FACW-) 1.00 Caxex longii (OBL) 1.00 Euthamia minor (FAC) 1.00 Cuphea carthagenensis (FACW) 1.05 Erechtites hieraciifolia (FAC-) 2.00 Hypericum mutilum (FACW) 2.00 Lythrum alatum (FACW+) 2.00 Rhynchospora microcarpa (FACW+) 2.37 Baccharis halimifolia (FAC) 2.50 Eryngium prostratum (FACW) 2.67 Ambrosia artemisiifolia (FACU) 3.00 Diospyros virginiana (FAC) 3.00 Juncus dichotomus (FACW) 3.00 Murdannia nudiflora (FAC) 3.00 Schinus terebinthifolius (FAC) 3.00 Cyperus lecontei (FACW) 3.75 Cyperus rotundus (FAC-) 4.00 Juncus effusus (FACW+) 4.00 Rhexia cubensis (FACW+) 4.00 Sesbania herbacea (FACW-) 4.40 Pouzolzia zeylanica 4.50 Ptilimnium capillaceum (OBL) 4.50 Urochloa mutica (FACW) 4.67 Centella asiatica (FACW) 4.68 Alternanthera philoxeroides (OBL) 4.76 Cyperus strigosus (FACW) 5.00 Teucrium canadense (FACW-) 5.00 Lindernia grandiflora (OBL) 5.00 Eclipta prostrata 5.50 Lachnocaulon engleri (FACW+) 6.00 Ludwigia palustris (OBL) 6.00 Euthamia caroliniana (FAC) 6.00 Cyperus polystachyos (FACW) 6.50 Urena lobata (FACU) 6.67 Phyla nodiflora (FACW) 6.79 Cyperus haspan (OBL) 7.00 Hibiscus grandiflorus (OBL) 7.00

PAGE 112

102 Table B-4 Continued Species name Average minimum distance from OHWL (m) Sacciolepis striata (OBL) 7.00 Toxicodendron radicans (FAC) 7.00 Paspalum notatum (FACU+) 7.64 Axonopus furcatus (OBL) 7.67 Lycopus rubellus (OBL) 8.00 Salix caroliniana (OBL) 8.00 Cyperus croceus (FAC) 8.00 Vitis rotundifolia (FAC) 8.33 Cirsium nuttallii (FAC) 9.00 Commelina diffusa (FACW) 9.00 Eupatorium capillifolium (FACU) 9.38 Ludwigia arcuata (OBL) 9.50 Cyperus surinamensis (FACW) 10.00 Myrica cerifera (FAC+) 10.25 Panicum repens (FACW-) 10.71 Bacopa caroliniana (OBL) 10.79 Taxodium distichum (OBL) 11.20 Bacopa monnieri (OBL) 11.50 Cyperus articulatus (OBL) 12.00 Scleria ciliata (FAC) 12.00 Sida acuata 12.67 Juncus sp. 13.00 Richardia brasiliensis 13.67 Hydrocotyle umbellata (OBL) 13.83 Thelypetris interrupta (FAC+) 14.00 Quercus geminata oak 14.75 Cyperus odoratus (FACW) 15.00 Polygonum punctatum (FACW+) Paspalum conjugatum (FAC) 15.05 Bidens laevis (OBL) 16.00 Hydrochloa caroliniensis (OBL) 16.40 Kyllinga brevifolia (FACW) 17.00 Panicum hemitomon (OBL) 17.00 Scoparia dulcis (FAC) 17.50 Cynodon dactylon (FACU) 18.33 Ludwigia peruviana (OBL) 19.00 Momordica charantia 21.00 Plantago virginica (FACU-) 21.00 Smilax bona-nox (FAC) 21.00 Diodia virginiana (FACW) 21.00 Eleocharis sp. 21.35 15.01

PAGE 113

103 Table B-4 Continued Species name Average minimum distance from OHWL (m) Sabal palmetto (FAC) 22.00 Conyza canadensis (FACU) 23.00 Pontederia cordata (OBL) 23.62 Chenopodium ambrosioides (FACU) 26.00 Nelumbo lutea (OBL) 29.00 Sagittaria lancifolia (OBL) 30.88 Solanum diphyllum 32.00 Typha latifolia (OBL) 39.00 Nymphia odorata (OBL) 46.00 Phytolacca americana (FACU+) 46.00

PAGE 114

104 Table B-5 Average minimum distance from the OHWL for Lake Hatchineha. Species with a value of 0.1 occurred at the edge of the OHWL, but for calculation purposes were given a value other than 0. Species name Average minimum distance from OHWL (m) Bidens alba (FACW-) 0.10 Quercus geminata oak 1.00 Thelypetris interrupta (FAC+) 2.00 Erechtites hieraciifolia (FAC-) 2.00 Cuphea carthagenensis (FACW) 2.00 Baccharis halimifolia (FAC) 2.50 Sesbania herbacea (FACW-) 2.50 Paspalum notatum (FACU+) 3.00 Cyperus rotundus (FAC-) 4.00 Centella asiatica (FACW) 4.00 Ptilimnium capillaceum (OBL) 4.33 Pouzolzia zeylanica 4.50 Cyperus strigosus (FACW) 5.00 Teucrium canadense (FACW-) 5.00 Phyla nodiflora (FACW) 5.28 Eclipta prostrata 5.50 Urochloa mutica (FACW) 6.00 Hydrocotyle umbellata (OBL) 6.25 Alternanthera philoxeroides (OBL) 6.67 Urena lobata (FACU) 6.67 Eupatorium capillifolium (FACU) 6.80 Sacciolepis striata (OBL) 7.00 Toxicodendron radicans (FAC) 7.00 Lycopus rubellus (OBL) 8.00 Vitis rotundifolia (FAC) 8.33 Commelina diffusa (FACW) 9.00 Cirsium nuttallii (FAC) 9.00 Bacopa caroliniana (OBL) 10.00 Panicum hemitomon (OBL) 10.33 Cynodon dactylon (FACU) 10.50 Hydrochloa caroliniensis (OBL) 11.00 Taxodium distichum (OBL) 11.20 Panicum repens (FACW-) 12.00 Cyperus articulatus (OBL) 12.00 Sida acuata 12.67 Richardia brasiliensis 15.00 Paspalum conjugatum (FAC) 15.05 Sabal palmetto (FAC) 15.50

PAGE 115

105 Table B-5 Continued Species name Average minimum distance from OHWL (m) Eleocharis sp. 16.00 Polygonum punctatum (FACW+) 17.29 Myrica cerifera (FAC+) 17.50 Juncus sp. 18.00 Momordica charantia 21.00 Smilax bona-nox (FAC) 21.00 Chenopodium ambrosioides (FACU) 26.00 Pontederia cordata (OBL) 27.40 Solanum diphyllum 32.00 Diodia virginiana (FACW) 36.00 Sagittaria lancifolia (OBL) 37.33 Typha latifolia (OBL) 41.00

PAGE 116

106 Table B-6 Average minimum distance from the OHWL for Lake Kissimmee. Species with a value of 0.1 occurred at the edge of the OHWL, but for calculation purposes were given a value other than 0. Species name Average minimum distance from OHWL (m) Caxex longii (OBL) 0.10 Euthamia minor (FAC) 0.10 Cuphea carthagenensis (FACW) 0.10 Coreopsis leavenworthii (FACW) 1.00 Desmodium incanum 1.00 Digitaria serotina (FAC) 1.00 Hypericum mutilum (FACW) 2.00 Lythrum alatum (FACW+) 2.00 Urochloa mutica (FACW) 2.00 Rhynchospora microcarpa (FACW+) 2.37 Eryngium prostratum (FACW) 2.67 Myrica cerifera (FAC+) 3.00 Juncus sp. 3.00 Ambrosia artemisiifolia (FACU) 3.00 Diospyros virginiana (FAC) 3.00 Juncus dichotomus (FACW) 3.00 Murdannia nudiflora (FAC) 3.00 Schinus terebinthifolius (FAC) 3.00 Alternanthera philoxeroides (OBL) 3.62 Cyperus lecontei (FACW) 3.75 Juncus effusus (FACW+) 4.00 Rhexia cubensis (FACW+) 4.00 Ptilimnium capillaceum (OBL) 4.67 Centella asiatica (FACW) 4.76 Lindernia grandiflora (OBL) 5.00 Sesbania herbacea (FACW-) 5.03 Euthamia caroliniana (FAC) 6.00 Lachnocaulon engleri (FACW+) 6.00 Ludwigia palustris (OBL) 6.00 Cyperus polystachyos (FACW) 6.50 Cyperus haspan (OBL) 7.00 Hibiscus grandiflorus (OBL) 7.00 Axonopus furcatus (OBL) 7.67 Phyla nodiflora (FACW) 8.00 Cyperus croceus (FAC) 8.00 Salix caroliniana (OBL) 8.00 Paspalum notatum (FACU+) 8.67 Ludwigia arcuata (OBL) 9.50 Cyperus surinamensis (FACW) 10.00

PAGE 117

107 Table B-6 Continued Species name Average minimum distance from OHWL (m) Richardia brasiliensis 10.00 Panicum repens (FACW-) 10.20 Bacopa caroliniana (OBL) 11.01 Bacopa monnieri (OBL) 11.50 Scleria ciliata (FAC) 12.00 Polygonum punctatum (FACW+) 12.35 Eupatorium capillifolium (FACU) 13.67 Cyperus odoratus (FACW) 15.00 Diodia virginiana (FACW) 16.00 Bidens laevis (OBL) 16.00 Kyllinga brevifolia (FACW) 17.00 Scoparia dulcis (FAC) 17.50 Hydrocotyle umbellata (OBL) 17.63 Hydrochloa caroliniensis (OBL) 17.75 Ludwigia peruviana (OBL) 19.00 Quercus geminata oak 19.33 Plantago virginica (FACU-) 21.00 Pontederia cordata (OBL) 21.25 Eleocharis sp. 22.42 Conyza canadensis (FACU) 23.00 Thelypetris interrupta (FAC+) 26.00 Sagittaria lancifolia (OBL) 27.00 Panicum hemitomon (OBL) 27.00 Nelumbo lutea (OBL) 29.00 Cynodon dactylon (FACU) 34.00 Sabal palmetto (FAC) 35.00 Typha latifolia (OBL) 38.00 Nymphia odorata (OBL) 46.00 Phytolacca americana (FACU+) 46.00

PAGE 118

LIST OF REFERENCES Bishop, E.W. 1967. Florida Lakes, Part I: A Study of High Water Lines of Some Florida Lakes. Florida Board of Conservation, Division of Water Resources, Tallahassee, FL, USA. Brenner, M., M.W. Binford, and E.S. Deevey. 1990. Lakes. p. 364-391. In R.L. Myers and J.J. Ewel (eds.) Ecosystems of Florida. University of Florida Press, Orlando, FL, USA. Broward v. Mabry, 58 Fla. 393, 50 So. 826 (1909). Bucki v. Cone, 25 Fla. 1, 6 So. 160 (1889). Chopra, R., V.K. Verma, and P.K. Sharma. 2001. Mapping, monitoring and conservation of Harike wetland ecosystem, Punjab, India, through remote sensing. International Journal of Remote Sensing 22:89-98. Cooner, J.C. No date. Navigability and Ordinary High Water. DBPR Course Number: CEP 003-128. Florida Surveying and Mapping Society, Tallahassee, FL, USA. Cowardin, L.M., V. Carter, F.C. Golet, and E.T. LaRoe. 1979. Classification of wetlands and deepwater habitats of the United States. U.S. Fish and Wildlife Service, Washington, DC, USA. FWS/OBS-79/31. Davis, J.H., Jr. 1973. Establishment of Mean High Water Lines in Florida Lakes. Publication No. 24. Florida Water Resources Research Center, Gainesville, FL, USA. OWRR Project Number A-015-FLA. Davis, M.M., S.W. Sprecher, J.S. Wakeley, and G.R. Best. 1996. Environmental gradients and identification of wetlands in north-central Florida. Wetlands 16:512-523. Edwards, T.C., Jr., G.G. Moisen, and D.R. Cutler. 1998. Assessing map accuracy in a remotely sensed, ecoregion-scale cover map. Remote Sensing of Environment 63:73-83. Florida Agricultural Coalition. 1999. Position statement on ownership of private lands. Online www.flagcoalition.com/fl_ag_coalition/RenderPage.asp ?pageid=118 [Accessed 18 July 2002]. 108

PAGE 119

109 The Florida Land Title Protection Act of 2000, H.R. 1807. Online www.leg.state.fl.us/data/session/2000/House/bills/analysis/pdf/HB1807S1.GG.pd f [Accessed 26 July 2002] Fuller, R.M., G.B. Groom, S. Mugisha, P. Ipulet, D. Pomeroy, A. Katende, R. Bailey, and R. Ogutu-Ohwayo. 1998. The integration of field survey and remote sensing for biodiversity assessment: a case study in the tropical forests and wetlands of Sango Bay, Uganda. Biological Conservation 86:379-391. Guest, D. 1991. The ordinary high water boundary on freshwater lakes and streams: origin, theory, and constitutional restrictions. Journal of Land Use & Environmental Law 6:205 Hall, B.R., J.R. Dudley, and D.J. Leopold. 2001. Environmental influences on plant species composition in ground-water seeps in the Catskill Mountains of New York. Wetlands 21:125-134. Hall, D.W. 1987. The use of plants as an aid to determine the ordinary high water line in Florida. In D.W. Gibson (Principal Investigator) Development of Procedural Methodology to Enable Florida Land Surveyors to Determine the Ordinary High Water Line Boundary of Sovereign Lands, Phase I Area Research. Department of Civil Engineering, University of Florida, Gainesville, FL, USA. Star Project 86-043. Hall-Atkinson, C. and L.C. Smith. 2001. Delineation of delta ecozones using interferometric SAR phase coherence MacKenzie River Delta, N.W.T., Canada. Remote Sensing of Environment 78:229-238. Harvey, K.R. and G.J.E. Hill. 2001. Vegetation mapping of a tropicalfreshwater swamp in th eNorthern Territory, Australia: a comparison of aeirial photography, Landsat TM and SPOT satellite imagery. International Journal of Remote Sensing 22:2911-2925. Interim Land Management of Conservation and Recreation Lands Program from Conservation and Recreation Lands Trust Fund. 2001. S. 1661. Senate Bill 2000, Third Engrossed, p. 229. Online www.leg.state.fl.us/data/session/2001/senate/appbills/pdf/enroll.pdf [Accessed 26 July 2002]. Jensen, J.R. 2000. Remote Sensing of the Environment: An Earth Resource Perspective. p. 181-240. Prentice Hall, Upper Saddle River, NJ, USA. Keddy, P. and L.H. Fraser. 2000. Four general principles for the management and conservation of wetlands in large lakes: the role of water levels, nutrients, competitive hierarchies and centrifugal organization. Lakes & Reservoirs: Research and Management 5:177-185.

PAGE 120

110 Kelly, N.M. 2001. Changes in the landscape pattern of coastal North Carolina wetlands under the Clean Water Act, 1984-1992. Landscape Ecology 16:3-16. Kushlan, J.A. 1990. Freshwater Marshes. p. 324-363. In R.L. Myers and J.J. Ewel (eds.) Ecosystems of Florida. University of Florida Press, Orlando, FL, USA. Loveland, T.R. 2001. Toward a national fuels mapping strategy: lessons from selected mapping programs. International Journal of Wildland Fire 10:289-299. Loveland, T., J. Merchant, D. Ohlen, and J. Brown. 1991. Development of a land-cover characteristics database for the conterminous U.S. Photogrammetric Engineering and Remote Sensing 57:1453-1463. MacNamara v. Kissimmee River Valley Sportsmans Association and Board of Trustees of the Internal Improvement Trust Fund, 648 So.2d 155 (Fla. App. 2 Dist. 1994). Martin v. Busch, 93 Fla. 535, 562, 112 So. 274 (1927). Neilson, R., G. King, and G. Koerper. 1992. Toward a rule based biome model. Landscape Ecology 7:27-49. Prentice, C., W. Cramer, S. Harrison, R. Leemans, R. Monserud, and R. Solomon. 1992. A global biome model based on plant physiology and dominance, soil properties and climate. Journal of Biogeography 19:117-134. Ramsey, E.W., III and S.C. Laine. 1997. Comparison of Landsat Thematic Mapper and high resolution photography to identify change in complex coastal wetlands. Journal of Coastal Research 13:281-292. Reed, P.B. 1988. National list of plant species that occur in wetlands: 1988 Florida. U.S. Fish and Wildlife Service, Washington, DC, USA. NERC-88/18.09, 1-140. Reimer, M.K. 2001. The Public Trust Doctrine: historic protection for Floridas navigable rivers and lakes. The Florida Bar Journal April:10-20. Ruiz-Luna, A. and C.A. Berlanga-Robles. 1999. Modifications in coverage patterns and land use around the Huizache-Caimanero lagoon system, Sinaloa, Mexico: a multi-temporal analysis using Landsat images. Estuarine, Coastal and Shelf Science 49:37-44. Running, S.W., T.R. Loveland, L.L. Pierce, R.R. Namani, and E.R. Hunt, Jr. 1995. A remote sensing based vegetation classification logic for global land cover analysis. Remote Sensing of Environment 51:39-48. Sanders, T.C. 1876. The Institutes of Justinian; with English Introduction, Translation, and Notes. p. 158-159. Callaghan & Company, Chicago, IL, USA.

PAGE 121

111 Scott, J.M., F. Davis, B. Csuti, R. Noss, B. Butterfield, C. Groves, H. Anderson, S. Caicco, F. DErchia, T.C. Edwards, Jr., J. Ulliman, R.G. Wright. 1993. Gap analysis: a geographic approach to protection of biological diversity. Wildlife Monographs No. 123. Tilden v. Smith, 94 Fla. 502, 113 So. 708 (1927). Townsend, P.A. 2000. A quantitative fuzzy approach to assess mapped vegetation classifications for ecological applications. Remote Sensing of Environment 72:253-267. Townsend, P.A. and S.J. Walsh. 2001. Remote sensing of forested wetlands: application of multitemporal and multispectral satellite imagery to determine plant community composition and structure in southeastern USA. Plant Ecology 157:129-149. Townshend, J., C. Justice, W. Li, C. Gurney, and J. McManus. 1991. Global land cover classification remote sensing: present capabilities and future possibilities. Remote Sensing of Environment 35:243-255. US Department of Agriculture, Natural Resources Conservation Service. 1998. Field Indicators of Hydric Soils in the United States, Version 4.0. G.W. Hurt, P.M. Whited, and R.F. Pringle (eds). USDA, NRCS, Ft. Worth, TX, USA. Warner, S. 1999. Unreal Estate, Political Waters and the Authority of Marshmallows: Controversies Surrounding Floridas Ordinary High Water Line (Property Rights, Public Trust, Ecosystem Management, Water Protection) (Doctoral dissertation, Emory University, 1999). Dissertation Abstracts International 60, (06B):2590. Wetzel, P.R., A.G. van der Valk, and L.A. Toth. 2001. Restoration of wetland vegetation on the Kissimmee River floodplains: potential role of seed banks. Wetlands 21:189-198. Wilen, B.O. and M.K. Bates. 1995. The US Fish and Wildlife Services National Wetlands Inventory project. Vegetatio 118:153-169. Wilen, B.O. and W.E. Frayer. 1990. Status and trends of U.S. wetlands and deepwater habitats. Forest Ecology and Management 33/34:181-192. Wilkinson T. and Warner, S. 2002. Floridas Non-Tidal Water Boundaries. Florida Department of Environmental Protection, Bureau of Surveying and Mapping. Tallahassee, FL, USA.

PAGE 122

BIOGRAPHICAL SKETCH Janet Nunley was born and raised outside Buffalo, New York. After graduating from high school in 1982, Janet was a single mom who worked full-time to support her daughter. Janet moved to Vero Beach, Florida in 1990 and graduated from Indian River Community College with Highest Honors in 1997. She transferred to the University of Florida and was awarded a Bachelor of Science degree in Zoology, with Honors, in 2000. Her passion for nature, wildlife, and the environment prompted her to continue her education. She will receive a Master of Science degree in Interdisciplinary Ecology in December 2002. Janet plans to join AmeriCorps for 2 years before either pursing a doctoral degree or working in the private sector. She is married to husband, Bryant, and has one daughter, Heather. 112


Permanent Link: http://ufdc.ufl.edu/UFE0000582/00001

Material Information

Title: Potential for remote sensing to locate the ordinary high water line in Florida: a case study of Lakes Hatchineha and Kissimmee
Physical Description: Mixed Material
Creator: Nunley, Janet C. ( Author, Primary )
Publication Date: 2002
Copyright Date: 2002

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0000582:00001

Permanent Link: http://ufdc.ufl.edu/UFE0000582/00001

Material Information

Title: Potential for remote sensing to locate the ordinary high water line in Florida: a case study of Lakes Hatchineha and Kissimmee
Physical Description: Mixed Material
Creator: Nunley, Janet C. ( Author, Primary )
Publication Date: 2002
Copyright Date: 2002

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0000582:00001


This item has the following downloads:


Full Text











POTENTIAL FOR REMOTE SENSING TO LOCATE THE ORDINARY HIGH
WATER LINE IN FLORIDA: A CASE STUDY OF LAKES HATCHINEHA AND
KISSIMMEE















By

JANET C. NUNLEY


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA


2002



































Copyright 2002

by

Janet C. Nunley















ACKNOWLEDGMENTS

I give special thanks to Scot Smith, Bon Dewitt, and Levent Genc who spent

endless nights getting the early images and thoughts in order. Special thanks also go to

Mark Clark for his patience and endurance. Many short meetings turned into marathon

events and I learned a lot. I would like to thank Ben Heilman, John Clyatt, Robby Bickel

and Hannibal Snipes for their assistance with the fieldwork. Without them, this work

could never have been completed. Thanks go to the Florida Department of

Environmental Protection for funding this research. Finally, many thanks to my husband,

Bryant, for his patience and support. Without him, I never could have done it!
















TABLE OF CONTENTS
page

A C K N O W L E D G M E N T S ......... ...................................................................................... iii

LIST OF TABLES ........... ............................... ............... ............. vi

L IST O F F IG U R E S .... ...... ................................................ .. .. ..... .............. vii

ABSTRACT ........ .............. ............. ...... .......... .......... ix

CHAPTER

1 IN TR OD U CTION .................. ............................ .. .......... .............. ..

E early C ases..................................................................... 2
C o rn ersto n e C ases........................................................................................................... 3
A R ecen t C a se ....................................................................... 5
D debate R ages O n .................................................... .............. 5
O p p o sin g S id es .......................................................... ........ ...... 7
L legislative A action ..................................................................... 9
H ow It A ll Fits Together............................. .............. 11
Site D description ............................. .............. ...... 13

2 LITERATURE REVIEW ................................. ................................... 15

3 MATERIALS AND METHODS .................................................31

Imagery ....................................... .............. 31
L an d sat ................... ........ .................................... 3 1
Indian Remote Sensing Satellite ...................................... ............ 33
The SPO T Satellite............................. .............. 33
The IK ON O S Satellite .................................................. ........ 34
Compact Airborne Spectrographic Imager ........................................................... 34
Im ag e P ro c e ssin g ................................................................. ................................. 3 5
Visual Interpretation ......................................... 37
Vegetation Sampling....................................................... 38
V vegetation A analysis ................................................................................... .. 40










4 R E SU L T S ....................................................... 42

Im agery R results ........................................ 42
L an d sat Im ag es ....................................................................... 4 2
Indian R em ote Sensing Im ages.................................................... ... ................. 42
The SPOT Images .. ................. ........................ .... .......... .. 47
The IK O N O S Im ages ......... ................................. ...................... .............. 47
The CA SI Im ages ........... ...... ....... .... ......... .............. .. 47
Vegetation Results ............ .............. .... ... ...... .............. 58

5 D IS C U S S IO N .............. ...... ........ .................................................................6 4

R em ote Sensing A spects................... ........................................... ........... .............. 64
Vegetation Aspects ........................................................... 71
Conclusions...................... ............... ..... .............. 73

APPENDIX

A V E G E TA TIO N CH A R T S ..................................................................... ..................75

B V E G E TA TIO N TA B LE S..................................................................... ...................93

LIST O F R EFEREN CE S ... ... ................. ........................................... ............... 108

BIOGRAPHICAL SKETCH .................................. ............................. 12




























v
















LIST OF TABLES


Table page

4-1 Summary table of imagery results .............. ................................ ...............58

4-2 Twenty-three most frequently occurring species on the banks of Lake
Hatchineha and Lake Kissimmee, combined.....................................................59

4-3 Average minimum distance from the OHWL that the closest species were
found for Lake Hatchineha and Lake Kissimmee, combined..............................60

4-4 Ten species with the highest usability index for combined lakes and land use
ty p es. .............................................................................. 6 1

4-5 Twelve species with highest usability index for all land use types of Lake
H watch in eh a on ly ................................................... ................ 6 1

4-6 Ten species with highest usability index for all land use types of Lake
K issim m ee on ly ................................................... ................ 62

4-7 Ten species with highest usability index for ranch areas of combined lakes ........62

4-8 Ten species with highest usability index for natural areas of combined lakes ......62

4-9 Ten species with highest usability index for urban areas of combined lakes ........63

B-l List of species ordered by frequency of occurrence for Lake Hatchineha and
Lake K issim m ee com bined...................... ................................. ............... 94

B-2 List of species ordered by frequency of occurrence for Lake Hatchineha ............97

B-3 List of species ordered by frequency of occurrence for Lake Kissimmee.............99

B-4 Average minimum distance from the OHWL for Lake Hatchineha and Lake
K issim m ee, com bined ................................................ .............................. 101

B-5 Average minimum distance from the OHWL for Lake Hatchineha..................104

B-6 Average minimum distance from the OHWL for Lake Kissimmee....................106
















LIST OF FIGURES


Figure p

1-2 Average monthly stage data for Lake Kissimmee covering a 7-year period.........14

3-1 Different ground cell sizes of imagery used in this study............................... 35

3-3 H highest elevation along a transect .............................................. ............... 40

3-4 Idealized vegetation transect chart and terminology used in vegetation
a n a ly sis ...................................................................... 4 1

4-1 Transect 31, classified Landsat image of Lake Hatchineha...............................43

4-2 Transect 9, classified Landsat image of Lake Kissimmee................................44

4-3 Transect 33, classified IRS image of Lake Hatchineha................ ................45

4-4 Transect 9, classified IRS image of Lake Kissimmee ...................................46

4-5 Transect 31, classified SPOT image of Lake Hatchineha................... ............... 48

4-6 Transect 9, classified SPOT image of Lake Kissimmee .................................49

4-7 Transect 31, classified IKONOS image of Lake Hatchineha............................50

4-8 Transect 4, classified IKONOS image of Lake Kissimmee .............................. 51

4-9 Transect 7, classified IKONOS image of Lake Kissimmee ...............................52

4-10 Histogram ofFACU, FAC, FACW or OBL species classified in each CASI
hyperspectral class ...................... .. ........................... .. ...... .... ........... 53

4-11 Histogram of species classified in each CASI hyperspectral class........................54

4-12 Transect 31, classified CASI image of Lake Hatchineha................................55

4-13 Transect 9, classified CASI image of a portion of Lake Kissimmee...................56

4-14 Vegetation in the area of Transect 33. Note watermarks on cypress trees. .........57

5-1 Catfish Creek entrance to Lake Hatchineha. Note water within the cypress........70









A-1 Vegetation Transect 1, ranch land use, Lake Kissimmee ...................................76

A-2 Vegetation Transect 2, ranch land use, Lake Kissimmee ....................................77

A-3 Vegetation Transect 3A, natural area land use, Lake Kissimmee .......................78

A-4 Vegetation Transect 4, ranch land use, Lake Kissimmee ...................................79

A-5 Vegetation Transect 5A, ranch land use, Lake Kissimmee ................................80

A-6 Vegetation Transect 6, natural area land use, Lake Kissimmee ............................81

A-7 Vegetation Transect 7, urban land use, Lake Kissimmee ..................................82

A-8 Vegetation Transect 8, urban land use, Lake Kissimmee ..................................83

A-9 Vegetation Transect 9, ranch land use, Lake Kissimmee .....................................84

A-10 Vegetation Transect 10A, natural area land use, Lake Kissimmee .....................85

A-11 Vegetation Transect 31, natural area land use, Lake Hatchineha ..........................86

A-12 Vegetation Transect 32, natural area land use, Lake Hatchineha ..........................87

A-13 Vegetation Transect 33, natural area land use, Lake Hatchineha ..........................88

A-14 Vegetation Transect 34, natural area land use, Lake Hatchineha ..........................89

A-15 Vegetation Transect 36, ranch land use, Lake Hatchineha...............................90

A-16 Vegetation Transect 37, natural area land use, Lake Hatchineha........................91

A-17 Vegetation Transect 38, urban land use, Lake Hatchineha.............................. 92















Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

POTENTIAL FOR REMOTE SENSING TO LOCATE THE ORDINARY HIGH
WATER LINE IN FLORIDA: A CASE STUDY OF LAKES HATCHINEHA AND
KISSIMMEE

By

Janet C. Nunley

December 2002

Chair: Scot E. Smith
Major Department: College of Natural Resources and Environment

Locating the ordinary high water line (OHWL) of Florida's freshwater bodies has

been a contentious issue for many decades. At the time of statehood in 1845, all property

below the OHWL was designated as 'State sovereignty' land, to be held in trust for

public use. Legal battles have raged over the boundary line dividing sovereignty land

and private ownership. Florida has approximately 7800 lakes and numerous rivers.

Ordinary high water determinations are made only on a case-by-case basis, as it would be

prohibitively expensive to locate the OHWL on each of Florida's freshwater bodies.

Case law requires that 'the best methods available' be used to locate the OHWL.

Historically, OHWL determinations have been made by interpreting aerial

photography and by locating other indicators such as watermarks, debris lines, lichen

lines, and vegetation indicators. In an effort to calculate the area of Florida's sovereignty









lands, the State legislature appropriated funds to determine the usefulness of remote

sensing for such a task.

This study used Landsat, SPOT, Indian Remote Sensing (IRS), and IKONOS satellite

imagery as well as CASI airborne hyperspectral images to assess the ability of remote

sensing to detect water and to detect a vegetation change or shift corresponding to the

OHWL. Open water was detectable in all the images, but was not detectable in the

presence of dense emergent vegetation. The relatively large ground cell size of Landsat,

SPOT, and IRS severely limited their ability to detect vegetation shifts corresponding to

the OHWL within a desired 1 m margin of error. IKONOS was better than the other

sensors at detecting vegetation shifts near the OHWL; however, it was only successful

14% of the time. When shifts were detected, the 4 m ground cell size of IKONOS did not

meet the desired 1 m margin of error when determining a boundary line. With its 1 m

ground cell size, CASI had the best potential as a method for locating the OHWL.

Vegetation shifts corresponding to the OHWL were detected in three of the four transects

located on the CASI images. However, two were coincidentally located at the edge of

cypress. Further study is needed to determine if CASI can consistently detect changes at

the OHWL. Vegetation analysis of the study site showed no consistent useful vegetation

indicators of the OHWL. This may be the result of an extended antecedent drought at the

time of this study.














CHAPTER 1
INTRODUCTION

The ordinary high water line (OHWL) has been a contentious issue in Florida for

more than a century. When Florida became a state in 1845, all lands beneath navigable

waters were conveyed to the state for public use as state sovereignty lands. Navigability

does not require the presence of water at all times, but does require the water body to be

navigable during its ordinary state (Reimer 2001). According to the Department of

Environmental Protection (DEP), Division of Surveying and Mapping (Wilkinson and

Warner 2002 Section 5 p. 2), OHWL means

The boundary between sovereignty lands owned by the State of Florida by virtue
of its sovereignty and adjacent riparian lands. The OHWL approximates the point
to which the water normally rises during the high water season. The OHWL
excludes exceptionally high water events caused by unusual floods or freshets.
The OHWL encompasses all parts of the waterbody however shallow the water
may be at the outside lines or elsewhere if the water is in fact a part of the
waterbody that is navigable for useful purposes. It is an ambulatory line that
shifts in response to long term, gradual, natural changes in water levels or in the
shoreline. In general, accretion, erosion, reliction and submergence shift the
OHWL while avulsion, artificial accretion, artificial erosion, artificial reliction
and artificial submergence do not shift the OHWL.

Designation as sovereignty land means that these lands are the property of the public and

are to be used and maintained for the benefit of the public (Guest 1991). This sounds

straightforward, but it is a misconception to believe that the OHWL is clear. As Warner

(1999) notes, when Florida received title to all the navigable water bodies within the state

in 1845, there were no lists or maps of all these water bodies and, consequently, the

extent of the state's ownership was unclear. This is still true today. Because of limited

financial resources, OHWL determinations are still made only on a case-by-case basis.









The OHWL and state sovereignty lands are inextricably linked. The concept of

protecting lands for public use originates in Greek philosophy and Roman law. During

the reign of the Roman Emperor Justinian (529-534 A.D.), the law specified the public

nature of rivers, ports, and seashores:

By the law of nature these things are common to mankind-the air, running
water, the sea, and consequently the shores of the sea .... All rivers and ports are
public .... The public use of the banks of a river is part of the law of nations, just
as is that of the river itself. The public use of the sea-shore, too, is part of the
law of nations, as is that of the sea itself; the shores may be said to be the
property of no man.... (Sanders 1876 p. 158-159)

This concept was later entwined in English common law and ultimately in American law

(more specifically, in Florida law). It is because ordinary high water covers the shore on

a fairly regular basis that the OHWL includes the shore. If an average or low watermark

were used instead, the legal boundary between private and sovereignty, or public, lands

would not include the shore (Warner 1999). The boundary between public and private

land, however, has been at issue almost since Florida entered the Union. Since the late

1800s, court battles have raged over navigability issues and the line demarcating ordinary

high water.

Early Cases

The earliest cases dealt with navigability. In an 1889 case, Bucki v. Cone, the

Florida Supreme Court determined that a water body is navigable if it is a "stream of

sufficient capacity and volume of water to float to market the products of the country. ..,

[and it is] not essential that the stream should be continuously, at all seasons of the

year, in a state suited for such flotage" (Bucki v. Cone 1889 p. 160). In 1909, the Florida

Supreme Court, in Broward v. Mabry, reaffirmed that the state holds in trust for the

people "the navigable waters and the lands thereunder, including the shore or space









between ordinary high and low water marks" and that "capacity for navigation, not usage

for that purpose, determines the navigable character of waters" (Broward v. Mabry 1909

p. 827).

Another problem that Broward v. Mabry dealt with resulted from the Swamp and

Overflowed Lands Act that Congress passed in 1850. By the 1850 Act, states were

granted all unsold swamp and overflowed lands within their borders. Florida received

approximately 22 million acres that were above the ordinary high water mark (Warner

1999). According to Assistant Attorney General Monica Reimer (2001):

The expectation [of Congress] was that the states would convey these lands to
large companies which would then be responsible for drainage and reclamation.
In an attempt to orderly manage these federal land grants, the 1855 Florida
Legislature created the Internal Improvement Fund, which consisted of title to
internal improvement lands and swamp land and all proceeds from the sale of
such lands. The legislature then vested title to the internal improvement lands and
swamp lands in the "Trustees of the Internal Improvement Fund" and gave them
authority to sell and transfer the lands as prescribed by statute (p. 13).

Broward v. Mabry (1909) clearly stated that "the trustees of the internal improvement

fund, who have the disposal of the swamp and overflowed lands of the state, have no

authority to convey the title to the lands under navigable waters that properly belong to

the sovereignty of the state" (p. 827).

Cornerstone Cases

Two cases decided in 1927 became cornerstones of ordinary high water line law

in Florida: Martin v. Busch and Tilden v. Smith Both cases helped define the OHWL

and introduced some of the problems involved in delineating the OHWL on Florida's

freshwater bodies (Warner 1999). In Martin v. Busch (1927), the complainant (Busch)

was claiming ownership of land that was once part of the lakebed of Lake Okeechobee

before drainage operations lowered lake levels and exposed land. He claimed that his









title to the land originated from the Swamp and Overflowed Lands Act of 1850. In

hearing the case, the Supreme Court of Florida reiterated that swamp and overflow deeds

did not convey title to lands below the OHWL and that all lands below the OHWL were

designated as state sovereignty lands (Cooner n.d.). The Court also stated that it was the

State's responsibility to establish the boundary between the sovereignty lands and

uplands (Warner 1999). This case also addressed the problem of locating the OHWL on

low, flat-banked water bodies.

... In flat territory or because of peculiar conditions, there may be little if any
shore to navigable waters, or the elevation may be slight and the water at the outer
edges may be shallow and affected by vegetable growth or other conditions, the
line of ordinary high-water mark may be difficult of accurate ascertainment; but,
when the duty of determining the line of high-water mark is imposed or assumed,
the best evidence attainable and the best methods available should be utilized in
determining and establishing the line of true ordinary high-water mark... Marks
upon the ground or upon local objects that are more or less permanent may be
considered in connection with competent testimony and other evidence in
determining the true line of ordinary high-water mark .. .(Martin v. Busch 1927 p.
274).

Martin v. Busch is relevant to this study because the Court ruled that "the best evidence

attainable and the best methods available should be utilized in determining and

establishing the line of true ordinary high-water mark ." (Martin v. Busch, 1927, p.

274).

The Tilden v. S.ithl (1927) case clarified the concepts of 'ordinary' and 'level' (as

in ordinary high water level), and found that the OHWL is ". .. a line at which the water

usually stands when free from disturbing causes, or as the usual or ordinary height" (p.

710). It also found that there can be considerable variation in water level resulting from

unusually high rainfall or drought, but ". .. nevertheless the character of the vegetation

and trees around the lake gave some evidence of an average or ordinary high-water mark

.." (p. 709). Furthermore, the Court went on to say that since water in low, flat water









bodies does not impress any well-defined line of demarcation, vegetation must serve as

the principal criterion. Tilden v. S.mitl (1927) concluded that in "ascertaining the normal

level and average high-water mark of a lake, evidence may be introduced as to the

character of the vegetation upon its banks" (p. 708).

A Recent Case

A 1994 Appellate case validated the methodology (i.e., locating water marks and

documenting vegetation patterns) used to locate the OHWL. In McNamara v. Kissimmee

River Valley Sportsmans' Association (and Board of Trustees of the Internal

Improvement Trust Fund), action was brought by the Sportsmans' Association over the

right of Roger McNamara to fence off land, including a spoil island, on Lake Hatchineha.

The Court found that the "area in controversy was vegetated lake bottom and that spoil

island was retained in public ownership" (p. 155). The Court also upheld the

'cornerstone' cases and reiterated that the OHWL was the boundary of normal reach

during the high water season, that the OHWL is the place where water prevents the

cultivation of ordinary crops, and that flat shorelines require using the best evidence,

including water marks and historical aerial photos, for determining the OHWL

(Wilkinson and Warner 2002).

Debate Rages On

An ongoing issue surrounding ordinary high water level began in 1976. The

David Smith case is unique to the ordinary high water issue because it involves ditches

and dikes along Lake Poinsett in Brevard County. Several dikes had been built below the

OHWL of 13 ft (3.96 m) above mean sea level (AMSL), one about 1969, the other about

1973. These dikes came to the attention of the State in 1976 when Smith (and two

neighbors) wanted to construct a new dike and ditch along the 13.00 ft contour line









(OHWL in this case). A 1978 agreement was reached with the State contingent upon a

restoration plan for the 1969 and 1973 dikes. This required the breaching of the 1969 and

1973 dikes to restore the Lake Poinsett marsh (Division of State Lands Response to

David Smith's "Fact Sheet". 1997. Florida Department of Environmental Protection,

Tallahassee, FL, USA).

The U.S. Army Corps of Engineers finally issued a permit in 1981 that "required

breaching the lower dike/ditch system and having restoration inspected and approved

before beginning construction of the upper [13.00 ft contour] dike/ditch system"

(Division of State Lands 1997 p. 2). The 1997 DEP report goes on to say that in 1982,

St. John's River Water Management District (SJRWMD) notified Smith of their concern

that he was attempting to proceed with development of the property prior to the

completion of the mandated restoration. A 1983 letter from Smith's agent contended that

Smith had been unable to complete the restoration due to inundation of the dikes by high

water in Lake Poinsett. SJRWMD reported that same year that the restoration was 90%

complete, and that 22 gaps had been created. Smith later asserted that the dikes remained

functional (even when the gaps were cut) though aerial photos clearly showing

inundation refute this assertion. Some time in 1986 Smith refilled the gaps in the two

illegal dikes.

In 1995, Trustees for the State filed suit against Smith charging that he had

illegally diked, drained, and occupied several hundred acres of sovereignty land in Lake

Poinsett's lakebed. Smith went to the Governor's Cabinet and asked for a disclaimer on

the property, asserting that the disputed property was swamp and overflowed land, which

encouraged improvement by draining and ditching as set forth in the Swamp and









Overflowed Lands Act of 1850 (Warner, S. 2002. Unreal estate, political waters and the

authority of marshmallows: controversies surrounding Florida's ordinary high water line.

Unpublished manuscript). As of 2002, the case remains open, but has not yet reached the

courts.

Opposing Sides

On one side of the issue are those who believe the state is attempting a 'land grab'

from landowners like Mr. David Smith. The State and Trustees of the Internal

Improvement Trust Fund, who are charged with protecting Florida's sovereignty lands,

sit on the opposing side. Mr. Smith has become the 'poster boy' of Florida Cattlemen's

Association (FCA) and the Florida Agricultural Coalition (FAC). A FAC position paper

on ownership of private lands (1999) gives a brief description of Florida's acquisition of

swamp and overflowed lands and goes on to suggest that the Smith case (referred to

indirectly) may change Florida Supreme Court case law "that has existed since the early

part of the century" (p. 1).

FAC's (1999) position paper goes on to say, "BTIITF [Trustees] over the years

has sold almost all of the land to private owners. Many of these lands were actually sold

with the intent to drain and develop them, or the land would revert back to the State" (p.

1). They do not mention that draining and development could not occur on sovereignty

lands. FAC (1999) claims the gist of the issue is that the Attorney General is attempting

to change the law to have OHWL be

The highest reach of water during an ordinary year. This means the highest reach
of water during the wet season, which everyone knows, is the 'rainy season' or
'flood stage'. This would...take up over 600,000 acres of lands that are now
owned, farmed, or occupied by private landowners, ON WHICH TAXES ARE
BEING PAID. (p. 1).









There is no mention of previous Supreme Court rulings that held that sovereignty lands

were not conveyed with swamp and overflowed land deeds, nor is there any mention of

the 1994 MacNamara case in which the Appellate Court held that the OHWL "on fresh

waters is the ordinary or normal reach of water during the high water season"

(MacNamara v. Kissimmee River Valley 1994 p. 156). The MacNamara case reaffirmed

the 'cornerstone' cases, it did not change the definition of OHWL. Furthermore, there is

no mention of case law, which addresses the issue of taxes. MacNamara v. Kissimmee

River Valley (1994) states

Even if taxes had been paid, such payment cannot form the basis of equitable
estoppel because it is the Trustees of the Internal Improvement Fund rather than
the tax assessor who are authorized to speak for the state on the subject of
boundaries on navigable lake bottoms. Sec. 253.12(1), Fla. Stat. If a taxing error
has taken place, the remedy is a tax refund rather than conversion of lake bottoms
to private ownership (p. 163).

Finally, FAC (1999) states

The members of the FAC believe landowners should be able to go to a public
library, get a copy of the law, and go to their property and identify where the
property ends and the public's begins. By [the] same token, a member of the
public should be able to look at the bank of a water body and identify where
public lands end and private property begins. These owners should be able to
fence and post their land against trespassers and poachers. (p. 1)

It sounds simple enough, go to the lake, walk the shore, and locate the water line. If the

OHWL were so easy to find, Florida's courts would not have ruled on numerous cases

regarding navigability and ordinary high water lines. Furthermore, case law would not

mention low, flat banked water bodies, nor would there be any directions as to the

ascertainment of the OHWL such as water marks on objects, competent testimony, and

vegetation indicators. This study confirmed the difficulty of unequivocally determining

the OHWL on such water bodies.









The FAC (1999) position paper also includes an editorial opinion by FCA

President Larry Barthle which goes into further detail on the Smith case. In that editorial,

Barthle expresses his view that the courts are changing the rules and that the OHWL is

"no longer the mark on the bank that any fool can see which has been caused by the

continuous action of water for eons" (p. 2). This editorial also fails to mention that

sovereignty lands were not conveyed with swamp and overflowed land deeds. Mr.

Barthle also suggests that "any fool" can see the high water mark. However, Lake

Poinsett is a low bank lake where there is no mark on the bank caused by continuous

wave action. Interestingly, FAC and FCA both contend that Congress intended swamp

and overflowed lands for development after draining and ditching. This is accurate

providing any such efforts did not include sovereignty lands. Neither organization

mentions that Mr. Smith's own privately hired engineers estimated the OHWL of Lake

Poinsett at 13 ft (3.96 m) AMSL in 1977. Staff for the Trustees then concluded that the

OHWL was not below the 13 ft (3.96 m) contour. However, Mr. Smith's dikes and

ditches built in 1969 and 1973 extend below the 8.3 ft (2.53 m) elevation (Warner 2002).

The State, on the other hand points to previous case law in which the Courts have

repeatedly upheld the State's ownership of land below the OHWL.

Legislative Action

Periodically, lobbying activities by FAC, FCA, and others have attempted to get

the Legislature to unequivocally define the OHWL. The most recent attempt was a bill

introduced during the 2000 Legislative session. Senate Bill 1824 and House Bill 1807,

the Florida Land Title Protection Act, proposed to do just that. According to Larry

Barthle of FCA, HB 1807 and SB 1824, "will end the repugnant and arrogant land grab

perpetuated by Attorney General Butterworth .... [It] will put into law the definition of









ordinary high water mark that was in use for more than 140 years." (FAC 1999 p. 2) In

the analysis of the Florida Land Title Protection Act, HB 1807 (2000) defines the OHWL

as

... the point upon which the presence and action of the water is so continuous as
to destroy the value of the land for agricultural purposes, preventing the growth of
vegetation, constituting what may be an ordinary agricultural crop. This mark is
ambulatory and shifts in response to long-term changes in the water level. The
high water is not the highest point to where the water rises in times of floods or
rains, but instead it is the line in which the water impresses upon the soil and
covers it for such periods so as to deprive it of vegetation and destroy its
agricultural value. In addition, this section clarifies that lands that are subject to
such periodical overflow so as to require drainage, levees or embankments to
keep the water out in order to render the land suitable for cultivation are not
sovereignty lands, below the ordinary high water mark. Furthermore, this water-
mark does not include: nonnavigable creeks, sloughs, swamps, canals, and low
and overflowed lands that are adjacent to the navigable water body (p. 4).

This definition of OHWL diverges substantially from case law. Had this legislation

passed, the OHWL would have moved waterward and would no longer include any land

that could be ditched or drained to prevent the presence of water during the high water

season. Excluding the seasonality of rainfall from ordinary high water line

determinations makes no sense because Florida has both a distinct wet and dry season.

The wet season plays an important revitalizing role in Florida's ecosystems and,

therefore, must be included as part of an "ordinary" year as stated in case law. If one

excludes the seasonal rains then, in fact, the water line would be measured only during

the dry season when water levels are low. In that event, case law would be addressing

ordinary low water line not an ordinary high water line. The 2000 bill eventually died on

the floor. It was not the first time that legislation was proposed to try to define, with

certainty, the OHWL and it will undoubtedly not be the last. Sam Ard, lobbyist for FCA,

lobbied for this bill. Since this bill died, Mr. Ard and other lobbyists for FAC and FCA

will undoubtedly continue lobbying for legislative action to define the OHWL.









How It All Fits Together

As noted above, OHWL issues continually recur in Florida. OHWL

determinations continue to be complicated and contentious. The Mr. Smith case has

revitalized the fight in recent years and inspired introduction of the "Land Title

Protection Act" in the 2000 legislative session. After the bill died, lobbying activities by

Sam Ard of FCA sought to have funds appropriated for research to determine the amount

of acreage held as sovereignty land. To that end, the 2001 Legislature granted the

following:

From the funds in Specific Appropriation 1661, up to $300,000 is provided to the
Department of Environmental Protection for the purpose of a pilot project to map
and determine acreage of sovereignty lands using remote sensing satellite data
obtained through NASA or its affiliates (Interim Land Management p. 229).

The intent of this grant was to determine the viability of using remote sensing

technologies to accurately ascertain the area of Florida's sovereignty lands. Lakes

Hatchineha and Kissimmee were ultimately chosen for this pilot study because the

OHWL for Lake Hatchineha had been adjudicated at 52.50 ft (16.00 m) AMSL in

MacNamara v. Kissimmee River Valley and long term stage data exists for both lakes.

Furthermore, it has been commonly accepted that the OHWL for Lake Kissimmee is

52.00 ft (15.85 m) AMSL. Knowing the OHWL is essential to the calculation of the total

acreage of sovereignty lands. To do this, one would need to delineate the OHWL by

detecting the presence of water within the emergent vegetation or be able to detect

vegetation shifts corresponding to the OHWL via changes in pixel classes. In a phone

conversation in June 2002 with Mr. Terry Wilkinson, of DEP, Mr. Wilkinson indicated

that the desired margin of error when determining a boundary line is +1 m.









Historical or traditional methods of determining the OHWL include manual

interpretation of aerial photography in conjunction with field surveys. In field surveys,

the presence of such things as water lines on stationary objects, debris lines, lichen lines,

vegetation types and soil types are all useful for delineation purposes. According to

Davis (1973), many vegetation types are distinctly related to the amount, position, and

duration of water in the substratum. The vegetation can then be classified as obligate

wetland, facultative wetland, facultative, facultative upland, or upland or terrestrial based

on its frequency of occurrence along the hydrologic gradient (Reed 1988). By

determining the relative positions of the terrestrial plants and wetland plants, one can

determine, with some reliability, the present position and former positions of water levels

(Davis 1973).

The goal of this study was to assess the current state-of-the-art satellite and

aircraft based remote sensing (Landsat, SPOT, IRS, IKONOS, and CASI) technologies

for the purpose of corroborating the OHWL of 52.50 ft (16.00 m) and 52.00 ft (15.85 m)

AMSL for Lakes Hatchineha and Kissimmee, respectively. The use of this new

technology will be in line with the intent of case law requiring that the best methods

available should be used to locate the OHWL. If a line of demarcation can be found

using remote sensing, the State would have a less expensive, yet reliable method for

locating the OHWL of Florida lakes. The hypotheses for this study were as follows:

* Vegetation can be used as an indicator to locate the OHWL.
* Landsat will not be effective for locating the OHWL.
* SPOT will not be effective for locating the OHWL.
* IRS will not be effective for locating the OHWL.
* IKONOS will not be effective for locating the OHWL.
* CASI will be useful for locating the OHWL.









Site Description

Lakes Hatchineha and Kissimmee are located in Osceola and Polk Counties in

south central Florida (Figure 1-1). Kissimmee is approximately 140 km2 (Brenner et al.

1990). Lake Hatchineha is smaller, covering approximately 27 km2 (Bishop 1967). The

OHWL for Lake Hatchineha has been adjudicated to be 52.50 ft (16.00 m) AMSL and

the commonly accepted OHWL for Lake Kissimmee is 52.00 ft (15.85 m) AMSL. This

area of Florida had undergone an extended drought at the time of this study and water

levels in Lake Kissimmee had not reached 52.00 ft (15.85 m) AMSL since January 2000

(Figure 1-2).


4


Figure 1-1 Study area


0 2,550 5,10 10,200 15,300 20,400
MeWs














54

53


51

50

49
48


46
45 4 i -.,!!- h-II.I

44


MonthlYear



Figure 1-2 Average monthly stage data for Lake Kissimmee covering a 7-year period














CHAPTER 2
LITERATURE REVIEW

The origin of remote sensing began with aerial photography in 1858 over the

outskirts of Paris, France. Gaspard Felix Tournachon, who suspected that aerial

photography would one day become very important, took the first-known successful

aerial photograph from a balloon. Aerial photography's first military use came during the

American Civil War. By the end of World War I and during World War II, aerial

photography had become key in planning military strategy and was also important during

the cold war (Jensen 2000).

In 1960, the first satellite image was successfully recorded onboard the Corona

satellite, and the era of satellite imagery was born (Jensen 2000). Early remote sensing

technology seems to have been primarily used by the military. However, in 1972,

Landsat-1 began recording data to test the feasibility of collecting land-surface data about

the Earth from an unmanned satellite. Since that time, a number of other satellites from

the United States and elsewhere have been launched and images are used for both

government and non-government programs (Jensen 2000).

Remote sensing has become an important technology in environmental and

ecological studies. Large and remote ecosystems can now be classified into biologically

significant categories in a timely and cost-effective way (Hall-Atkinson and Smith 2000).

For example, the National Biological Service's Gap Analysis Program is an important

program that utilizes remote sensing technologies to determine land-use/land-cover

classes, which can help identify "gaps" in the protection of biological diversity (Edwards









et al. 1998). Remote sensing plays an important role in analyzing changes in land use

and cover because of its ability to record changes efficiently, which then provides

information important to the management of natural resources. In heterogeneous

environments, remote sensors can provide large amounts of data both efficiently and

economically (Ruiz-Luna and Berlanga-Robles 1999). The ability of remote sensing to

provide a cost-effective means of classifying the environment has been an important

factor in its increasing popularity for large ecosystem analyses.

One of the first ecological programs developed using remote sensing technology

was US Fish and Wildlife Service's National Wetlands Inventory (NWI) project.

Wetlands mapping began in 1979. NWI's goal was to generate and disseminate scientific

information about wetlands that could be used for decision-making about resource use

and allocation. There needed to be a resource that would provide information about the

characteristics and extent of the Nation's wetlands (Wilen and Bates 1995, Loveland

2001).

NWI researchers developed a classification system that divides ecological

systems, with certain common characteristics, into particular groups. The resulting

classification divided habitats into five major systems. Each system is divided into

subsystems that reflect hydrologic conditions, followed by classes that describe the

appearance of wetland vegetation (e.g. emergent wetland) or substrate. Classes are

further divided into subclasses that are named based on the predominant vegetation life

form. Descriptors are also included that provide information about hydrology, water

chemistry and human influence (Cowardin et al. 1979).









Once a classification system was developed, a remote sensing technology was

chosen. Researchers decided that satellite imagery from Landsat would not be able to

provide the necessary detail for wetland classification. Therefore, mid- and high-altitude

color infrared aerial photography, at a scale of 1:24,000, was chosen. Once images are

obtained, a long process is involved in creating wetlands maps that includes photo-

interpretation, field studies, review of soil maps, and review of existing information about

an area's wetlands to assure accurate identification (Wilen and Bates 1995). Turnaround

time from photo acquisition to map production can be as long as 10 years (Wilen and

Frayer 1990).

As of 1995, the NWI project had continued to investigate the use of satellite

imagery for use in wetlands mapping. Landsat technology was reevaluated and the

French satellite SPOT was also investigated. After a yearlong test, it was determined that

satellite imagery was still not able to provide the needed level of detail for wetland

classification. Therefore, NWI continues to use aerial photography for wetland mapping

(Wilen and Bates 1995).

Other researchers have found satellite imagery to be quite useful for their

purposes. The US Geological Survey's (USGS) Gap Analysis Program (GAP) that began

about 10 years ago found Landsat, with its 30 m resolution, to be sufficient for mapping

vegetation and habitat for conservation status assessment of vertebrate species in order to

allow individuals, planners, managers, and policy makers to make informed decisions

(Scott et al. 1993, Loveland 2001). One problem of GAP is that consistent methods of

classification are not used by each state. To overcome this problem, GAP has developed

a classification template to assist states in providing consistent maps. Additionally, GAP









promotes the use of a common data source and the use of consistent vegetation

definitions. It is recognized, however, that there are still limitations of satellite remotely

sensed data for producing these maps, mainly due to spatial resolution (Loveland 2001).

Another major program was developed by USGS in 1995. The purpose of the

USGS Land Cover Characterization Program is to provide consistent national land use

and land cover data that is nationally consistent and compatible with past and future

USGS land cover classifications. This program also uses Landsat imagery and uses an

unsupervised classification approach augmented by post-classification modeling that

include key variables such as elevation and population density. The mapping process

takes about three to four years. Therefore, maps are already dated by the time they are

released. As of 2001, accuracy assessments had not been completed and all results were

preliminary (Loveland 2001).

National vegetation maps are not the only maps of interest. Many researchers are

interested in producing global vegetation maps. Historically, scientists have attempted to

classify global vegetation using "Earth Systems" models. Early global vegetation

databases were derived from published maps, atlases and national databases, and they

attempted to represent existing global vegetation. Problems with these early databases

include inconsistency in vegetation classification, variable measurement techniques and

variable spatial sampling resolutions (Running et al. 1995). Prentice et al. (1992) and

Neilson et al. (1992) developed models of geographic distributions of vegetation based

on specific physiological responses to cold tolerance, growing season heat sums, and

drought stress. In these models, equivalent vegetation types, such as forests, can be

defined separately into classes like boreal, temperate, or tropical because their









classification scheme includes climatalogical parameters as integral components. Still,

some researchers argue that these models, though improved, still only provide maps of

potential vegetation types (Running et al. 1995).

Townshend et al. (1991) suggested that global vegetation classifications must be

realistic measures of existing land cover. A consistent, remote sensing-based

classification scheme would remove many of the ambiguities in global vegetation maps

because they would no longer be derived from varying methodologies. However, remote

sensing technology cannot produce maps with the large number of classes seen in the

earlier maps because they remove climate classes from the definition (Loveland et al.

1991). Therefore, Running et al. (1995) suggested that a new logic be introduced for

global vegetation classification that would be based on simple, observable, unambiguous

characteristics of vegetation structure related to ecosystem biogeochemistry. Such

characteristics would need to be repeatedly observable with remote sensing, such that

they could be classified the same way each time and they would be translatable into

biophysical parameters of interest by global climate and biogeochemical models.

Running et al. (1995) suggests that this classification logic should be the basis for all

classifications, and that those requiring more exacting classifications build hierarchically

from this one. This is a logical suggestion because a single basis for all classification

schemes would then start out on a level playing field. As remote sensing technology

improves, vegetation classification should become more accurate, especially if all

classifications start with the same foundation. This would improve global vegetation

maps as well as smaller scale maps that are used to analyze change over time in specific

ecosystems.









Landscape ecologists using satellite imagery to characterize vegetation, species

distributions, and communities are also seeking ways to make classifications more

accurate. Unfortunately, classification maps often simplistically characterize

heterogeneous communities as single, discrete units (Townsend 2000). Townsend (2000)

reports that among potential options for improved classification schemes are fuzzy

classifications and mixture/subpixel models. However, these models merely convey

information about variations in spectral signature between pixels, not necessarily

differences in on-the-ground content of mapped vegetation. Despite these problems,

some method of delineating classes on maps is essential in order to communicate

information about spatial patterns and the distribution and abundance of species.

According to Townsend (2000), the resulting maps show transition zones as discrete

boundaries. Furthermore, some recognizable on-the-ground vegetation associations may

not actually be discernable with satellite or other remotely sensed images. Townsend and

Walsh (2001) report that complex community relationships and the nature of their spatial

variability are rarely captured with satellite imagery. Broad ecotones are especially

difficult to portray accurately because they are made up of transition zones that grade into

each other across complex soil, hydrologic, and topographic gradients. The key to

making vegetation maps more appropriate for ecological interpretation is to find a

classification scheme that more accurately portrays complex vegetation communities.

Some of the most complex communities are coastal wetlands. High spatial

resolution photography is useful for classifying and identifying changes in coastal

wetlands. NWI is a primary producer of these high quality, detailed, coastal wetland

maps from high spatial resolution photography (Peters, D.D. 1994. Use of aerial









photography for mapping wetlands in the United States: National Wetlands Inventory.

Proceedings of the First International Airborne Remote Sensing Conference and

Exhibition. Strasbourg, France). Production of these maps is enormously expensive and

turnaround time from photo acquisition to map production can be as long as ten years

(Wilen and Frayer 1990). Obviously, maps with such long turnaround times are not

useful for monitoring more immediate changes in coastal wetlands. In an effort to find a

less expensive and faster way to monitor wetlands, Ramsey and Laine (1997) compared

the use of Landsat and high spatial resolution photography for identifying changes in

complex coastal wetlands. Landsat images are less expensive, but researchers needed to

determine their viability for identifying changes in coastal wetlands. In their study,

Ramsey and Laine (1997) used Landsat to monitor pre- and post-hurricane changes in

Louisiana's coastal wetlands. They encountered problems with using Landsat, including

misclassification of emergent vegetation and misclassification in areas that included a

mix of emergent vegetation and water. Landsat images did, however, provide good

separation of classes when one class dominated areas >30 m. This study illustrates both

the problems and the potential for using Landsat images to identify complex wetland

ecosystems.

In 1998, researchers in Sango Bay, Uganda also attempted to assess complex

ecosystems. A biodiversity assessment was conducted in an attempt to determine which

of the area's ecosystems have the greatest biodiversity value. The area includes swamps,

grasslands, cultivated land, and forests that border Lake Victoria in Uganda. It was

important to obtain a detailed map of the area, and since accessibility was a problem in

many of the areas, remote sensing was the logical method of choice. Landsat was chosen









over SPOT because even though SPOT has a finer resolution (20 m as opposed to

Landsat's 30 m), Landsat records an additional infrared channel that is important for

discriminating different vegetation types. A supervised classification scheme (Maximum

Likelihood Classification) was used. This type of classification scheme requires that the

people involved have a good working knowledge of actual ground vegetation and how

these vegetation types appear on the image (Fuller et al. 1998).

Final maps identified fourteen land-cover classes with the broadest categories

being: water, swamp, dry grassland, degraded woody vegetation, semi-natural woody

vegetation, and intensive land usage. Broad categories were subdivided (Fuller et al.

1998). Of interest here is the subcategory "tall grass swamp". Five tall grass swamps

were identified and mapped, but there was tremendous variability within the class. For

example, four sites included five to ten grass species while the fifth site included 33

species (Fuller et al. 1998).

The Sango Bay research brings two significant factors to light. First, one must

have a thorough knowledge of the vegetation communities being studied, and second,

that Landsat was unable to distinguish between a swamp with five species and a swamp

with 33 species. This implies that Landsat would not be very useful in delineating or

locating a line such as the OHWL on Lakes Kissimmee and Hatchineha.

Satellite imagery is continuing to grow in popularity for use in mapping wetlands.

While NWI continues to use aerial photography in order to produce detailed wetland

maps, other researchers are testing some of the newer satellite images available. In

Punjab, India, researchers prepared a land use/land cover map of the Harike wetland

ecosystem from Indian Remote Sensing Satellite (IRS) imagery with a 23 m spatial









resolution or ground cell size. The Harike wetland covers 285.1 km2 and is rich in

aquatic flora and fauna. This ecosystem assumed international importance because it

serves as a breeding ground and habitat for a large variety of migratory and domiciled

birds. Anthropogenic pressure in recent years has created severe ecological problems and

resulted in the need to conserve this wetland. Since no reliable data were available for

Harike, researchers wanted to develop a baseline map that would identify spatial

distribution and variation in the waterspread, turbidity and aquatic vegetation in different

seasons. IRS images were obtained from different years for Harike, both pre-monsoon

and post-monsoon (Chopra et al. 2001).

A map was generated using visual interpretation of IRS images and a

classification system recommended by the Space Applications Centre of Ahmedabad,

India. The Harike wetland ecosystem was ultimately divided into five major classes:

built-up land, agricultural land, forest, wasteland, and wetland. When the pre-monsoon

and post-monsoon images were compared, researchers were able to record differences in

the waterspread and turbidity. Furthermore, based on visual interpretation, three types of

aquatic vegetation were identified: water hyacinth, grassland 1 and grassland 2. The

difference between the two grassland categories resulted from the portion of the spectrum

with the highest reflectance (grassland 1 had high reflectance in the blue, green, and red

regions while grassland 2 had high reflectance in the infrared region). Though these were

recorded as two distinct classes on the map, no verification of differences in vegetation

species between the classes on the ground was possible due to inaccessibility of the area

(Chopra et al. 2001). The Harike study illustrates the potential for using IRS images to









broadly classify wetlands, but challenges still exist in terms of accurately assessing

heterogeneous communities whether they are wetland or upland.

Satellite imagery still poses challenges for wetland mapping due, in part, to

problems with spatial resolution. Vegetation gradients may not be detected with satellite

imagery because one pixel is likely to be wider than the associated vegetation shift on the

ground. Furthermore, the spectral resolution of satellite imagery may restrict its ability to

detect differences in vegetation types, which are often spectrally similar in wetlands

(Harvey and Hill 2001).

When comparing various remote sensing technologies, researchers found large-

scale aerial photography to be superior to both Landsat and SPOT satellite imagery for

detailed wetland mapping. More importantly, the use of context and texture information

obtained with aerial photography allowed researchers to enhance vegetation class

delineation and identification capabilities. The use of contextual and textural

characteristics is especially useful and desirable for delineation of vegetation

communities in heterogeneous wetland ecosystems. However, no significant difference

existed in the ability to map key structural components with any of the remotely sensed

images (Harvey and Hill 2001). The key, then, is to be aware of the different capabilities

of the imagery and to select the ones that would be most likely to produce the required

information. This again illustrates the likely failure of Landsat, SPOT, IRS, and

IKONOS satellite imagery to prove useful in locating the OHWL on Lakes Hatchineha

and Kissimmee.

Remote sensing images used in the Lake Kissimmee and Lake Hatchineha study

mainly provide vegetation information. Consequently, it was important to look at









vegetation communities and how they 'fall out' along shorelines. Lakeshores often

provide large areas of wetland vegetation including swamps and marshes (Keddy and

Fraser 2000). Environmental conditions such as water levels, soil fertility, grazing and

other disturbances are known to affect plant species composition in wetlands (Hall et al.

2001, Keddy and Fraser 2000). The shoreline merely stretches wetland vegetation out

along several gradients (Keddy and Fraser 2000). One would expect to find narrow

bands of wetland vegetation along steep topographic gradients and much broader bands

in flat topographic areas such as those found along Lakes Hatchineha and Kissimmee.

Associations between hydrophytic vegetation communities and hydric soils are

usually found along the shoreline gradient as well. Some researchers have looked at the

relationship between hydrophytic vegetation and hydric soils. Wetland delineation

standards developed by the U.S. Army Corps of Engineers require evidence of

hydrophytic vegetation, hydric soils, and wetland hydrology. Logically, one would

expect to find a direct relationship between hydrophytic vegetation and hydric soils.

However, the flat topography of Florida has been problematic in linking hydric soils to

hydrophytic vegetation in some areas (Davis et al. 1996). According to Reed (1988), the

1987 Corps Manual defines a hydrophytic plant community as one that has greater than

50% of dominant species classified as obligate (OBL), facultative wetland (FACW) or

facultative (FAC) on regional lists of plants that occur in wetlands. Some hydric soil

indicators found in Florida's sandy soils include stratified layers within the upper 15 cm

of the surface with value 3 or less and chroma 1 or less, presence of muck, stripped

matrix beginning within the upper 15 cm, and presence of 2% or more organic bodies

starting within the upper 15 cm of the surface (USDA 1998).









Davis et al. (1996) studied the relationship between hydrophytic vegetation,

hydric soils, and hydrology in north-central Florida. Researchers had no difficulty

identifying wetland or upland communities that occurred at extremes of the hydrologic

gradients. However, flatwoods communities were not consistently identified as wetlands

according to the 1987 Corps Manual, but were frequently classified as wetlands

according to the 1989 Corps Manual. Davis et al. (1996) found that Florida's flat

topography obscured the upland extent of hydrophytic vegetation and the expected

corresponding hydric soils. In flat topography, small variations in water level can extend

over a greater area, and for a longer period, than in steep terrain. Consequently,

hydrophytic vegetation can be found upslope from the wetland hydrology or hydric soil

boundary.

Under natural conditions, shorelines of freshwater lakes are subject to both wave

action and frequent changes in water level. Such natural disturbances can lead to

shorelines with high biological diversity. Vegetation types are directly related to the

duration of flooding. Thus, the degree of flooding results in distinct vegetation zones on

the shoreline (Keddy and Fraser 2000). Since the duration of flooding varies over time,

vegetation zonation is a dynamic property. In years with extraordinarily high water

levels, upland shrubs and trees may die. Alternatively, during periods of drought,

existing seed banks may germinate, increasing the diversity of species close to the

lakeshore (Keddy and Fraser 2000).

Another factor controlling species composition of shorelines is a species

competitive ability. Competitive ability is linked to environmental conditions and in

some cases soil fertility. In fertile soils, the pioneer and generalist species tend to









dominate, leaving specialist species to occupy the remaining spaces. There also appear to

be hierarchies of competitive ability that control the distribution of species along

shorelines. Soil fertility and dominance hierarchies are only some of the factors

controlling vegetation species in wetlands. Disturbance also plays a role in species

distribution. Finally, various combinations of fertility, disturbance, and hydrology will

have different impacts on wetland vegetation gradients (Keddy and Fraser 2000).

Lakes Kissimmee and Hatchineha have both been impacted by disturbance, shifts

in hydrology, and probably soil fertility. The Kissimmee River was once a long,

meandering river, originating in Orlando and winding its way to its terminus in Lake

Okeechobee (Kushlan 1990). However, between 1962 and 1971, the U.S. Army Corps of

Engineers channelized the river to provide drainage and flood control for agricultural and

residential development. The first lock on canal C-38 is located at the southern end of

Lake Kissimmee. The channelization of the river had significant negative impacts on

area flora and fauna (Wetzel et al. 2001). Both lakes have changed dramatically since the

channelization of the river. Water levels are now controlled, and are lower than historic

levels. Most importantly, water no longer flows naturally. Stagnation of the lakes has

contributed to numerous other problems, primarily the invasion of hydrilla, water

hyacinth and other aquatic vegetation that have taken a firm hold. Consequently, there

are now periodic drawdowns and muck removals, leaving behind spoil islands.

Undoubtedly, changes in the Kissimmee River influenced the nature of the lakes and

probably changed the vegetation communities along the shorelines as well.

Under normal circumstances, one would expect to find distinct vegetation

gradients on a lakeshore. Many Florida lakes experience dramatic water level









fluctuations during the year due to a distinct wet and dry season. These fluctuations, in

addition to Florida's relatively flat topography, combine to create wide littoral zones. In

some areas, shorelines have distinct zones marking low-water to high-water lines, called

the upper, middle, and lower zones (Davis 1973). The upper zone is the zone of interest

for this study because it is where water reaches the highest level assuming an ordinary

year. The plant community in this zone includes plants that can tolerate water for only a

short period of time. According to Hall (1987), some plant species can be found in more

than one habitat while others are restricted to a single habitat. Thus, the make-up of the

plant community in the area of interest can provide valuable information. For example,

the presence of predominantly upland species indicates the area never floods, or floods

only in unusually high water events. In such communities, a flood event would cause

nearly all species to die (Hall 1987). If an area can be identified as upland then, clearly,

the OHWL would be located at a lower elevation.

According to Davis (1973), the upper zone, where water reaches the highest level,

may contain such species as sweet gum, magnolia, saw palmetto, the sand cypress bushes

Hypericum spp., carpet grass Axonopus sp., Bermuda grass, and crab grass. Hall (1987)

recommends locating both the upland line, where flooding never occurs, and the lower

line, where flooding always occurs, before attempting to locate the OHWL. Some

species indicative of upland communities are turkey oak (Quercus geminata), wire grass

(Aristida stricta, and longleaf pine (Pinuspalustris) (Davis 1973, Hall 1987).

Conversely, species indicative of the lower zone, where water is always or nearly always

present, include cattail (Typha sp.), pickerelweed (Pontederia sp.), and maidencane









(Panicum hemitomon) (Davis 1973). Under natural circumstances, cypress trees

(Taxodium sp.) are good indicators of flooded conditions (Hall 1987).

Once these upper and lower zones are located, other indicators of OHWL can be

used. These include lake stage data, drift lines, lichen lines, algal mats, cypress knees,

and scarps, among others. Following identification of these indicators, vegetation in the

area should be identified (Hall 1987). Some species, however, are problematic because

of their regional differences in distribution (Davis 1973, Hall 1987). Florida's climate

extends from temperate to subtropical and the physical and chemical properties of the soil

vary widely throughout the state. Classification problems arise when plant species occur

only in wet habitats in one region, but occur in other habitats, even uplands, in other

regions (Hall 1987).

Davis (1973) studied numerous lakes, including both Lakes Kissimmee and

Hatchineha, to develop a method for determining former high water levels by identifying

plants living in or near Florida's lakes. On Lakes Kissimmee and Hatchineha, Davis

(1973) was able to determine historic water levels by comparison of oak tree (UPL) and

cypress tree (OBL) ages. For both lakes, oaks had occurred predominantly on shoreline

berms where water never reached. At the time of his study of Lake Hatchineha in 1971,

the water level was approximately 50.20 ft (15.30 m) AMSL. Davis found that the

historic high water level was 56.70 ft (17.29 m) AMSL since about 1870. A vertical

distance of greater than 45.10 m exists between the water level on the day of Davis' study

and the historic high water level. At the time of statehood, the high water position was

probably about 57.00 ft (17.37 m) AMSL. On Lake Kissimmee, Davis determined that

the high water level was probably never as high as 57.00-60.00 ft (17.37-18.29 m) AMSL









during the 20th century. At the time of his study, Davis (1973) noted that water levels in

Lake Kissimmee had gradually been lowered by about 3.00 ft (0.91 m) and were being

regulated by the new Kissimmee River controls.

There is good evidence that both remote sensing and vegetation identification and

analysis may be useful tools in locating the OHWL on Lakes Hatchineha and Kissimmee.

Remote sensing technologies have been used successfully to broadly delineate wetlands.

Research by both Davis and Hall show that vegetation, though problematic at times, can

be used to help locate the ordinary high water line.














CHAPTER 3
MATERIALS AND METHODS

Imagery

Four commercial earth-looking satellites provided suitable imagery for mapping

vegetation and shoreline for use in this study:

* Landsat 7 ETM (Landsat)
* Indian Remote Sensing Satellite (IRS)
* SPOT
* IKONOS

In addition to satellite imagery, airborne hyperspectral images from Canada's Compact

Airborne Spectrographic Imager (CASI) were obtained for portions of the western shores

of both Lakes Hatchineha and Kissimmee. Landsat, IRS, SPOT, and IKONOS images

were purchased when lake elevations were known to be near ordinary high (16.00 m

AMSL for Lake Hatchineha and 15.85 m AMSL for Lake Kissimmee based on NGVD

29).

Landsat

The Landsat sensing instrument is a multispectral scanner that measures radiation

reflected or emitted by the Earth's surface. There are eight bands sensitive to different

wavelengths of visible and infrared radiation. Ground cell size is nominally 30 m x 30 m.

The spectral parameters are as follows:

Band 1: 0.45 to 0.52 gm (blue). Provides increased penetration of water bodies,
as well as supporting analyses of land use, soil, and vegetation characteristics.

Band 2: 0.52 to 0.60 gm (green). This band spans the region between the blue
and red chlorophyll absorption bands and therefore corresponds to the green
reflectance of healthy vegetation.









Band 3: 0.63 to 0.69 [m (red). This is the red chlorophyll absorption band of
healthy green vegetation and represents one of the most important bands for
vegetation discrimination. It is also useful for soil boundary and geological
boundary delineations. This band may exhibit more contrast than bands 1 and 2
because of the reduced effect of atmospheric attenuation. The 0.69 .im cutoff is
significant because it represents the beginning of a spectral region from 0.68 to
0.75 jim, where vegetation reflectance crossovers take place that can reduce the
accuracy of vegetation investigations.

Band 4: 0.76 to 0.90 rm (reflective infrared). This band is especially responsive
to the amount of vegetation biomass present in a scene. It is useful for crop
identification and emphasizes soil/crop and land/water contrasts.

Band 5: 1.55 to 1.75 rm (mid-infrared). This band is sensitive to the turgidity or
amount of water in plants. Such information is useful in crop drought studies and
in plant vigor investigations.

Band 6: 10.4 to 12.5 [m (thermal infrared). This band measures the amount of
radiant flux emitted from surfaces. The apparent temperature is a function of the
emissivities and the true or kinetic temperature of the surface. It is useful for
locating geothermal activity, ..., vegetation classification, vegetation stress
analysis, and soil moisture studies.

Band 7: 2.08 to 2.35 rm (mid-infrared). This is an important band for the
discrimination of geologic rock formations. It has been shown to be particularly
effective in identifying zones of hydrothermal alteration in rocks (Jensen 2000 p.
194).

Finally, the panchromatic band is sensitive to wavelengths between 0.52 itm and

0.90 m. Bands 1 to 5 and 7 have a 30 m x 30 m spatial resolution or ground cell size.

The thermal infrared band 6 has 60 m x 60 m ground cell size, and the panchromatic band

has a 15 m x 15 m ground cell size (Jensen 2000).

Landsat satellites pass over the same place on earth every 16 days. The image

used for this project was taken on October 23, 1999 (scene identification

#23000310450487412). On that date, the average water level at the United States

Geological Survey (USGS) gauging station located in Lake Kissimmee was 15.85 m


AMSL.









Indian Remote Sensing Satellite

The IRS sensor has many of the same characteristics as Landsat. It was designed

to fill in for times that Landsat was unavailable. The spectral parameters of IRS' LISS-

III multispectral sensor are as follows (Jensen 2000):

Band 2: 0.52 to 0.59 [m with a 23 m x 23 m spatial resolution.

Band 3: 0.62 to 0.68 [m with a 23 m x 23 m spatial resolution.

Band 4: 0.77 to 0.86 [m with a 23 m x 23 m spatial resolution.

Band 5: 1.55 to 1.70 [m with a 23 m x 23 m spatial resolution.

Panchromatic: 0.50 to 0.75 [im with a 5.8 m x 5.8 m spatial resolution.

A color (23 m pixel size) IRS scene of the study area taken on December 4, 1999

was acquired for this project. The scene identification number is 1C28705200. The

average water level at the USGS gauging station located in Lake Kissimmee was 15.96 m

AMSL on that date.

The SPOT Satellite

Two sensors are onboard the SPOT satellite. One has a 20 m x 20 m ground cell

size and records data in 3 spectral bands (Band 1 has a spectral resolution of 0.50 to 0.59

[lm; Band 2 has a spectral resolution of 0.61 to 0.68[m, and Band 3 has a spectral

resolution of 0.79 to 0.89 pm). The second sensor uses one panchromatic band (0.51 to

0.73 [m) with a ground cell size of 10 m x 10 m (Jensen 2000). The SPOT scene

acquired for this project was taken on March 2, 1993 with the 3-band multi-spectral

sensor. The scene identification number was 2-260-293-930302-163230-1-X. The

average water level at the USGS gauging station located in Lake Kissimmee was 15.47 m

AMSL on that date.









The IKONOS Satellite

One of the latest commercial earth-looking remote sensing satellites is IKONOS.

The IKONOS satellite sensor has a panchromatic band with a spectral resolution of 0.45

to 0.90 tm and a ground cell size of 1 m x 1 m. IKONOS also has 4 multispectral visible

and near-infrared bands (Band 1: 0.45 to 0.52 inm, Band 2: 0.52 to 0.60 inm, Band 3:

0.63 to 0.69 rim, and Band 4: 0.76 to 0.90 [im) with a 4 m x 4 m ground cell size

(Jensen 2000). IKONOS imagery from both panchromatic and multispectral scanners

was obtained for this project. The IKONOS Space Imaging Corporation's product

acquired for this project was a "Geo 1 m + 4 m Bundle" taken on March 29, 2000. Scene

identification numbers for the 5 scenes required to cover the study area were: (1)

20000239154547300000011603012, (2) 2000023915454730000001 1603013, (3)

20000239154547300000011603014, (4) 20000239154547300000011603 015, and (5)

20000239154547300000011603016. The average water level at the USGS gauging

station located in Lake Kissimmee was 15.41 m AMSL on that date.

Compact Airborne Spectrographic Imager

The CASI hyperspectral sensor detects an array of narrow spectral bands in the

visible and infrared wavelengths. The spectral range covered by the 288 channels is

between 0.4 and 0.9 [im. Each band covers a wavelength range of 0.018 [im. While

ground cell size depends on the altitude of the aircraft, the spectral bands measured and

the bandwidths used are all programmable. For this study, data from 12 spectral bands

were collected for portions of the western shore of Lakes Kissimmee and Hatchineha

(Band 1 midpoint: 0.45 itm, Band 2 midpoint: 0.54 itm, Band 3 midpoint: 0.55 tim, Band

4 midpoint: 0.56 [im, Band 5 midpoint: 0.63 [im, Band 6 midpoint: 0.65 [im, Band 7

midpoint: 0.0.69 [im, Band 8 midpoint: 0.70 [im Band 9 midpoint: 0.77 [im, Band 10










midpoint: 0.78 rim, Band 11 midpoint: 0.79 rim, Band 12 midpoint: 0.83 [pm). The

ground cell size for CASI imagery was 1 m x 1 m.

The CASI images were acquired on March 20, 2002 for this project. The average

water levels at the USGS gauging stations located in Lake Kissimmee and Lake

Hatchineha were 15.38 m AMSL and 15.37 m AMSL, respectively. Images from each of

the above sources have different ground cell sizes (Figure 3-1). Each type of image was

used to test its ability to detect and display an indication of the OHWL.




1 m Desired
Margin of 30
Error for
Boundary
Lines Landsat 30 m x 30 m
25


20 IRS 23 m x 23 m


15 SPOT 20 m x 20 m


10 -
CASI 1 mx 1 m

5 IKONOS4mx4m







Figure 3-1 Different ground cell sizes of imagery used in this study. The red line
represents a desired margin of error (+1 m) when locating a boundary line.

Image Processing

The Landsat image was geometrically rectified to a Universal Transverse

Mercator (UTM), Zone 17 projection (datum WGS 84) using ground control points taken

from USGS digital orthophoto quarter quadrangles (DOQQ). Radiometric correction was









not performed on the imagery because the objective of this study was not to compare land

use/land cover change over time. All remaining images were geometrically rectified to

the DOQQs.

Images were classified using an 'unsupervised' classification approach. In a

multispectral image, each pixel has a spectral signature determined by the reflectance of

that pixel in each of the spectral bands. Multispectral classification is a process whereby

spectral signatures are analyzed and then pixels are assigned to classes based on similar

signatures (Jensen 1996). This classification system is based on the principle that all

pixels representing a homogeneous land cover type should have roughly the same

spectral signature, with only relatively small differences due to variables such as

shadows. Classification procedures attempt to group together such similar pixels. The

detail of the classes depends on the spectral and spatial resolution (Jensen 2000).

Unsupervised classification is a method in which a computer searches for natural

groupings of similar pixels called clusters. The fewer clusters there are, the more the

pixels within each cluster will vary in terms of spectral signature, and vice versa.

ERDAS IMAGINETM image processing software was used and unsupervised

classification was performed using an algorithm called the Iterative Self-Organizing Data

Analysis Technique (ISODATA). In this algorithm, the analyst inputs the number of

clusters desired (six for this study) and a 'confidence threshold.' The computer program

will then build clusters iteratively, meaning with each new iteration, the clusters become

more and more 'refined.' The iterations stop when the confidence level is reached. The

confidence threshold for this study was 95%. Once the clusters were built, they were

assigned to a land cover class. The image clusters were then recorded into a GIS layer,









which displayed each land cover class with a different color. Images were prepared with

three classes, five classes, six classes, nine classes, twelve classes, and 25 classes. Final

interpretation was done using seven class and six class designations for Lakes Hatchineha

and Kissimmee, respectively. Using more classes made the image too 'noisy', and fewer

classes did not provide enough contrast. The seven classes for Lake Hatchineha were as

follows:

* Open water
* Hydric 1
* Hydric 2
* Hydric 3
* Hydric 4
* Cypress
* Mesic

There is no cypress surrounding Lake Kissimmee, so the cypress class was eliminated for

those images.

The SPOT imagery for the study area was processed and classified in the same

manner as the Landsat imagery. The IRS, IKONOS and CASI images were preprocessed

radiometrically and geometrically. Images were then classified in the same manner as the

Landsat and SPOT images.

Visual Interpretation

Upon completion of image processing, ground truth field surveys were conducted.

When vegetation collection was complete, transect lines were drawn on Landsat, IRS,

SPOT, IKONOS, and CASI images by first identifying the GPS coordinate

corresponding to the OHWL, and then drawing a line along the azimuth recorded in the

field. Images were visually interpreted in a qualitative assessment to identify changes in









pixel class designations corresponding to vegetation shifts that may correlate with the

OHWL.

Vegetation Sampling

The areas surrounding the lakes were divided into land use types and included

ranching, natural, and urban. Distance along the lake margin was calculated for each

land use type and transect sites (Figure 3-2) were randomly selected. In some cases,

however, line transects were not taken at the exact point due to inaccessibility of the

selected location. In such cases, line transects were run at the nearest accessible point.

Transects 1, 2, 3A, 4, 5A, 6, 7, 8, 9, 10A were recorded on Lake Kissimmee. Transects

31, 32, 33, 34, 36, 37, 38 were recorded on Lake Hatchineha. Transects 9, 31, 32, and 33

correspond to the CASI images.

Six transects were on privately owned ranch land, 3 at urban sites, and 8 in natural

areas. Of the eight transects designated as natural, four were impacted by cattle

ranching. Though these areas were designated either as state parks or water management

district lands, the current drought has reduced water levels well below fence lines that

would normally keep cattle from grazing on these properties. In all four cases, either

cattle or cattle dung was a present.

Real time kinematics GPS and other surveying techniques were used to find the

vertical elevation corresponding to the known ordinary high water elevation for each

transect. Without knowing the horizontal position of the line, the point was found by

intersecting the known elevation with the shoreline. Transects were then set up starting

from the OHWL and running perpendicular to the lake margin. Transects extended

laterally 50 m below the OHWL and up to 50 m landward of the OHWL. In cases where

the highest elevation occurred less than 50 m landward of the OHWL, with a lower



































0 2.550 5,100 10,200 15,300 20,400 -V

Figure 3-2 Index map showing approximate transect locations

elevation on the landward side, transects extended only to the point of highest elevation

(Figure 3-3). Vegetation was recorded along the entire length of the transect and

identified to species level (when feasible) with the expertise of botanist, David Hall, and

some assistance from S. Barry Davis, N. Williams, M. Whitten, and K. Perkins of the

University of Florida Herbarium. Vegetation was classified according to hydrologic

indicators. Indicators in parentheses after species names are based on the National List of

Plant Species That Occur in Wetlands (Reed 1988). If a plant could not be identified to

species level, no designation was given due to variations of designations within the

genera.










23 m
Lower elevation, OWHL
landward of the ---------
OHWL, perhaps
moist conditions Highest elevation,
with obligate landward transect stops at
wetland vegetation this point (e.g., at 23 m) Lake


Figure 3-3 Highest elevation along a transect. When highest elevation occurred less
than 50 m (e.g., 23 m) landward of the OHWL, transect ended at that
elevation.

Vegetation Analysis

Vegetation lists were composed for each transect and charted according to their

location on the transect so that vegetation shifts could be easily identified (Figure 3-4).

This information was used in the visual interpretation of the remote sensing imagery.

Descriptive statistics were used to identify associations between vegetation and the

OHWL. The purpose was to establish the accuracy of vegetation as an indicator of the

OHWL. Frequency of occurrence was also determined for each vegetation species by

Equation 3-1.

Freq. of Occur. = (# of transects in which a species occurred)/(total # of transects
surveyed) (3-1)

Finally, a usability index was developed using Equation 3-2.

Usability Index = [(Freq. of Occur)/(Minimum Avg. Distance)] 100 (3-2)

The range of possible values for the Usability Index is from 0.1 to 1000. Higher values

equate to better use of a species as an indicator of the OHWL. For example, if a species

had a value of 1000 it would indicate that the species was present in every transect

surveyed and that its distribution always began or ended at the OHWL























Figure 3-4 Idealized vegetation transect chart and terminology used in vegetation
analysis. In this figure, Species 2 has a landward edge value of 1 and a
waterward edge value of 4. These values indicate that on the landward
side of OHWL, Species 2 had a minimum distance of 1 m from OHWL,
and on the waterward side, Species 2 had a minimum distance of 4 m from
OHWL.














CHAPTER 4
RESULTS

Imagery Results

Open water could be visually identified with Landsat, SPOT, IRS, IKONOS, and

CASI imagery. Water could not be identified in areas with dense emergent wetland

vegetation. All images had observable changes in pixel class designations, but no

discernable change in class corresponded to the OHWL for Landsat, SPOT, and IRS

images. The following figures are excerpts from the classified images. The contour

associated with the elevation of 52.50 ft AMSL is shown in black on the Lake Hatchineha

figures.

Landsat Images

There were changes in pixel class designations observable in the Landsat images,

but none were indicative of OHWL. Figures 4-1 and 4-2 are sections of Landsat images

classified into land cover classes for Lake Hatchineha and Lake Kissimmee, respectively.

The changes in pixel class designations in the figures likely indicate a change in the plant

community. However, these plant communities were not directly correlated with the

OHWL.

Indian Remote Sensing Images

Pixel class designation differences are again evident in the IRS imagery, but none

were indicative of the OHWL. As with Landsat imagery, pixel designations likely

indicate a change in plant community, but there is no direct correlation between this

change and the OHWL. Figures 4-3 and 4-4 are classified IRS images of Lakes









Hatchineha and Kissimmee, respectively. The shift detectable in Figure 4-3 does not fall

within 2 m of the OHWL.


4.


Legend
OHWL

Class sNBmam
- dyi trlt 1
-lHydric 2
Sflydrc 3
___ -linEl '4
-d~
-~~ ss


S


) 1 3.700 7.4M WO-3
I- I- 1 I L L-r I I


Figure 4-1 Transect 31, classified Landsat image of Lake Hatchineha. The small-
scale image includes a square that is represented by the larger scale image
to the right. The black line represents the contour of the known OHWL.


p'i~








44









































Legend
Cla Namer 2,0 "0 8

SHJ.IL :

Hydri 3
I l Hydric 4

0 2,200 4,400 ,8
OpelWI I i I I


Transect 9, classified Landsat image of Lake Kissimmee


Figure 4-2









































Legend
OHWL


ClassaNamis
I Hydcc 1
SHydric 2
HyWic 3
W rHyefc 4i


I OpenrWater


S

7.4ID Meter
i


3.700


Figure 4-3 Transect 33, classified IRS image of Lake Hatchineha *Radiometric
dropouts occurred in the image as indicated.


i


1,o





46






it^


/


Legend
Class_Names
M HVtC 1
Hy H c 2


I OpenWater


0 2,200 '400 8,800 Meters
I I I I I I I I


Figure 4-4 Transect 9, classified IRS image of Lake Kissimmee


i

U


N

S
R-
$









The SPOT Images

The multi-spectral SPOT imagery was classified with respect to land cover and

the results are shown in Figures 4-5 and 4-6 for Lakes Hatchineha and Kissimmee,

respectively. The changes in pixel class designations were not indicative of the OHWL.

The IKONOS Images

The IKONOS results are shown in Figures 4-7, 4-8, and 4-9, which correspond to

Transects 31, 4, and 7, respectively. There appears to be a correlation between a change

in the pixel class designation and the OHWL in Figure 4-9. The estimated accuracy of

this edge is 4 m. The most prominent vegetation was Bahia grass (Paspalum notatum).

This was an urban site with a high elevation of 16.65 m AMSL. Transect 7 extended

only 34 m above the OHWL because a house blocked the way. Vegetation above the

OHWL included FACU, FAC, FACW, and OBL species, but their combination resulted

in classification as mesicc'.

The CASI Images

The CASI imagery exhibited the best results in terms of discerning plant

community changes. A change in pixel class designation is visible approximately 3 m

above the OHWL. Vegetation transect chart (Appendix A, Figure A-9) illustrates a

similar shift in vegetation closer to 4 m above the OHWL. Five species (Euthamia

caroliniana, Paspalum notatum, Sesbania herbacea, Axonopusfurcatus, and Centella

asiatica) occurred on both sides of the OHWL, while two occurred only above the

OHWL and thirteen occurred only below the OHWL. The vegetation occurring above

the OHWL included FACU, FAC, FACW, and OBL species, but still provided a distinct

classification as mesicc' in this study. The following figures (4-10 and 4-11) illustrate the

inability of the CASI hyperspectral scanner to distinguish between wetlands designations







48



within the plant community. This results from the high degree of variance in the


reflectance received by the scanner.


..
: i.

^ 8-:: ^-


:1


19%.."


* .. ... .. S


Legend




ClassNames

SHydrlc 1
1-lydrlc 2


Hydric 4


openw afterr


N
'I'd"^


5a


T5O


3,T700
II


7.400 Meter
I


Figure 4-5 Transect 31, classified SPOT image of Lake Hatchineha












































Legend
Class_Name
- Hrr.r I
1 1 I nH irl: 2
F-- Hydr l 3
=' Hydrir 4


M Open Water


P4


S


0 2,200 4A.e 008
I I I I I ,I I


Transect 9, classified SPOT image of Lake Kissimmee


Figure 4-6
















F-.-


,, I
id r-.


-4.


Legend

I62.5

Class_Names
- Hydric 1
1Hydric 2
H1 Mydrlc 3
S1Mylrc 4


M Open Water


1.960


3,700


7,40 Meteis


Transect 31, classified IKONOS image of Lake Hatchineha


I I ~ I I I r I


Figure 4-7













,.

h* *


Legend



- Hytdf 1

|I aName2


:ic L,4c)


Figure 4-8 Transect 4, classified IKONOS image of Lake Kissimmee


N8

S


'10


I I








































-__ Mr'K 1
2 N



hrt lI: a ipss spaa Tfaa Muirli S





Figure 4-9 Transect 7, classified IKONOS image of Lake Kissimmee

CASI images were classified with respect to land cover type and the results are

shown in Figures 4-12 and 4-13. Classified CASI imagery for Transects 31 and 32 on

Lake Hatchineha was useful in detecting plant community changes corresponding to the

OHWL. The change detected resulted from a spectral reflectance pattern sensed by

CASI, and was located at the edge of the cypress canopy. No such change was visible for

Transect 33 where the OHWL fell within the cypress. Inspection of the original


~~~
~' r
Lt~









classified images revealed the cypress pattern clearly, however, the excerpt shown in

Figure 4-12 does not capture the details as well as the originals. There is a clear change

in pixel classes at the OHWL for Transect 9 on Lake Kissimmee. A photograph taken at

Transect 33 shows watermarks on the cypress trunks (Figure 4-14).

70.0 .I


60.0



50.0
-

40.0
0
-.

, 30.0-
-


20.0



10.0



0.0
-


hydric 1


I


hydric 2 hydric 3
Hyperspectral Classes


hR


hydric 4


Figure 4-10 Histogram ofFACU, FAC, FACW or OBL species classified in each
CASI hyperspectral class. The variation within each class illustrates the
inability of the imagery to discriminate wetland designations within this
littoral zone.


open water
open water


SFAC
*FACU
O FACW
OOBL


mes
mesic








































open


'a c ,5 to -

f m -iB Q) B ) y
-s .y ~
E ~ 2 g


r-


c-
0- E MM -


I~~~ I~) !B I I ^ |
| ~3 -8 3 i
c Q) -t x


8 Ca
Q) M 0 > 0
Q, 0 (O o G 0
r "
0 E Q

0I O a I
0m


Species


Figure 4-11 Histogram of species classified in each CASI hyperspectral class. This further illustrates the variation of the
species and the respective classes to which they were assigned.


100.0

90.0

. 0.0

70.0

o0.O0

50.0

-10 0

20 0

100

00


0/
O
c
F
c
r
u
e
q r
u
a
e
n
n
c
c
e
y































7%

S iI



m

'









Legend
OHWL
- 52.5
C lssNames
I ydric 1
W 1Opnyda 2
j tydnc 3
yI I Hync 4
Mesic
Open Water


N



S


0 750 1,500
i i i I I


3,JD Meters
i I I I


Figure 4-12 Transect 31, classified CASI image of Lake Hatchineha



































l~F LL' %
-c
'?. :


Legend
Class_NamneB


__ ] hi:: 2

=-- Hy:rr. S
r-1 Hydri4


openwater


N



S


1.6o


3.200


MaDO !Mters


Figure 4-13 Transect 9, classified CASI image of a portion of Lake Kissimmee


I _I I /t /I I I I 1




































Figure 4-14 Vegetation in the area of Transect 33. Note watermarks on cypress trees.

Table 4-1 summarizes the ability of both satellite and aircraft based images used

in this study to show a change in pixel classes associated with the ordinary high water

line. Column headings represent the transects and row headings represent the source of

the images. Any transect with a corresponding image showing a change in pixel classes

within 2 m of the OHWL is represented by a letter or a letter and a number within the cell

as follows: H2 = Hydric 2 class, H3 = Hydric 3, H4 = Hydric 4 and M = Mesic. The

value of 2 m was chosen based on a statistical analogy. If one assumes the 1 meter

tolerance (mentioned earlier) to be a standard deviation, 2 meter would give a 95%

confidence level. The 'Total' column is the number of transects showing a change at the

OHWL divided by the total number of transects represented in the imagery.






58


Table 4-1 Summary table of imagery results. Dashes (-) indicate no data and "nc"
indicates no change in pixel class designations within 2 m of the OHWL.
The top entry within a cell indicates the pixel class on the landward side of
the OHWL while the bottom entry indicates the pixel class on the waterward
side.

Transects
Image 1 2 3A 4 5A 6 7 8 9 10A 31 32 33 34 36 37 38 % Total
H3
Landsat nc nc nc nc nc nc nc nc nc nc H4 nc nc nc nc nc nc 5.88
H3
IRS nc nc nc nc nc nc nc nc H2 nc nc nc nc 7.69
H3
SPOT nc nc nc nc nc nc nc nc nc nc nc H2 nc nc nc nc nc 5.88
M H4
IKONOS nc nc nc nc H4 nc nc nc nc nc nc nc H3 nc 14.29
M H3 H4
CASI -- H2 H2 H2 nc 75.00


Vegetation Results

Vegetation transect charts illustrating how vegetation fell out along each line

transect are located in Appendix A, Figures A-i through A-17 for both Lake Kissimmee

and Lake Hatchineha. Table 4-2 and 4-3 summarize species data for combined lakes.

Both tables include 23 species, which represent 25% of the total species identified to that

level. The tables include the most frequently occurring species and the species found

closest to the OHWL. There are only two species contained in both tables, Cyperus

lecontei and Sesbania herbacea. Only three species dropped out at the OHWL: Bidens

alba, Carex longii and Euthamia minor. A complete list of species sorted by frequency

of occurrence and by average minimum distance from the OHWL for both combined

lakes, and for each lake individually can be found in the Appendix B (Tables B-1 through

B-6).









Table 4-2 Twenty-three most frequently occurring species on the banks of Lake
Hatchineha and Lake Kissimmee, combined. The number of occurrences
represents the number of transects that the species occurred in. Frequency
of occurrence is the number of occurrences divided by the total number of
transects.


Top 23 occurring species
Pontederia cordata (OBL)
Polygonum punctatum (FACW+)
Hydrocotyle umbellata (OBL)
Paspalum notatum (FACU+)
Bacopa caroliniana (OBL)
Phyla nodiflora (FACW)
Centella asiatica (FACW)
Sagittaria lancifolia (OBL)
Alternanthera philoxeroides (OBL)
Sesbania herbacea (FACW-)
Panicum repens (FACW-)
Eupatorium capillifolium (FACU)
Taxodium distichum (OBL)
Axonopusfurcatus (OBL)
Ptilimnium capillaceum (OBL)
Panicum hemitomon (OBL)
Hydrochloa caroliniensis (OBL)
Euthamia caroliniana (FAC)
Diodia virginiana (FACW)
Quercus geminata oak
Myrica cerifera (FAC+)
Cyperus polystachyos (FACW)
Cyperus lecontei (FACW)


Number of
occurrences
13
13
12
10
9
9
9
8
8
8
7
7
6
6
6
5
5
5
4
4
4
4
4


Frequency of
occurrence
0.76
0.76
0.71
0.59
0.53
0.53
0.53
0.47
0.47
0.47
0.41
0.41
0.35
0.35
0.35
0.29
0.29
0.29
0.24
0.24
0.24
0.24
0.24









Table 4-3 Average minimum distance from the OHWL that the closest species were
found for Lake Hatchineha and Lake Kissimmee, combined. Species with a
value of 0.1 occurred at the edge of the OHWL, but for calculation purposes
were given a value other than 0.


Species name
Bidens alba (FACW-)
Caxex longii (OBL)
Euthamia minor (FAC)
Coreopsis le ciu'ni /hii (FACW)
Desmodium incanum
Digitaria serotina (FAC)
Cuphea carthagenensis (FACW)
Erechtites hieraciifolia (FAC-)
Hypericum mutilum (FACW)
Lythrum alatum (FACW+)
Rhynchospora microcarpa (FACW+)
Baccharis halimifolia (FAC)
Eryngium prostratum (FACW)
Ambrosia artemisiifolia (FACU)
Diospyros virginiana (FAC)
Juncus dichotomus (FACW)
Murdannia nudiflora (FAC)
Schinus terebinthifolius (FAC)
Cyperus lecontei (FACW)
Cyperus rotundus (FAC-)
Juncus effusus (FACW+)
Rhexia cubensis (FACW+)
Sesbania herbacea (FACW-)


Average minimum
distance from OHWL (m)
0.10
0.10
0.10
1.00
1.00
1.00
1.05
2.00
2.00
2.00
2.37
2.50
2.67
3.00
3.00
3.00
3.00
3.00
3.75
4.00
4.00
4.00
4.40


The following tables give the top ten species sorted by the usability index (UI)


Combined lakes and land use types
All land use types on Lake Hatchineha only
All land use types on Lake Kissimmee only
All ranch land of combined lakes
All natural areas of combined lakes
All urban areas of combined lakes


Only Sesbania herbacea occurred in five of these tables with a high UI value of 18.44

(out of a possible 1,000) and a low UI value of 10.7. Urena lobata, Bidens alba and









Carex longii had the highest UI values of 333.33, with Bidens alba and Carex longii

showing up in three of the six tables. Three species had consistently high UI values in

each of the tables where they show up: Bidens alba, Carex longii, and Euthamia minor.

Table 4-4 Ten species with the highest usability index for combined lakes and land
use types


Species name
Bidens alba (FACW-)
Caxex longii (OBL)
Euthamia minor (FAC)
Centella asiatica (FACW)
Cuphea carthagenensis (FACW)
Sesbania herbacea (FACW-)
Alternanthera philoxeroides (OBL)
Ptilimnium capillaceum (OBL)
Phyla nodiflora (FACW)
Paspalum notatum (FACU+)


Usability index
58.82
58.82
58.82
11.32
11.20
10.70
9.88
7.84
7.80
7.70


Table 4-5 Twelve species with highest usability index for all land use types of Lake
Hatchineha only. Twelve species are included because species 10, 11, and
12 all have the same ranking on the UI.


Species name
Bidens alba (FACW-)
Quercus geminata oak
Baccharis halimifolia (FAC)
Sesbania herbacea (FACW-)
Phyla nodiflora (FACW)
Ptilimnium capillaceum (OBL)
Hydrocotyle umbellata (OBL)
Eupatorium capillifolium (FACU)
Taxodium distichum (OBL)
Thelypetris interrupt (FAC+)
Erechtites hieraciifolia (FAC-)
Cuphea c iaii ti/'//e'/ii (FACW)


Usability index
142.86
14.29
11.43
11.43
10.83
9.89
9.14
8.40
7.65
7.14
7.14
7.14









Table 4-6 Ten species with highest usability index for all land use types of Lake
Kissimmee only


Species name
Caxex longii (OBL)
Euthamia minor (FAC)
Cuphea c,1 iith,/genill',i (FACW)
Centella asiatica (FACW)
Alternanthera philoxeroides (OBL)
Rhynchospora microcarpa (FACW+)
Sesbania herbacea (FACW-)
Eryngium prostratum (FACW)
Cyperus lecontei (FACW)
Paspalum notatum (FACU+)


Usability index
100.00
100.00
100.00
16.80
13.81
12.68
11.92
11.25
10.67
10.38


Table 4-7 Ten species with highest usability index for ranch areas of combined lakes
Species name Usability index
Euthamia minor (FAC) 166.67
Cuphea c, ii thigencll\i (FACW) 166.67
Centella asiatica (FACW) 21.35
Myrica cerifera (FAC+) 16.67
Rhynchospora microcarpa (FACW+) 13.07
Sesbania herbacea (FACW-) 11.54
Euthamia caroliniana (FAC) 11.11
Eryngium prostratum (FACW) 11.11
Axonopusfurcatus (OBL) 9.38
Paspalum notatum (FACU+) 8.45


Table 4-8 Ten species with highest usability index for natural areas of combined lakes
Species name Usability index
Urena lobata (FACU) 125.00
Axonopusfurcatus (OBL) 125.00
Sesbania herbacea (FACW-) 18.44
Paspalum notatum (FACU+) 16.67
Scoparia dulcis (FAC) 12.50
Cyperuspolystachyos (FACW) 12.50
Digitaria serotina (FAC) 12.50
Cyperus lecontei (FACW) 12.50
Ptilimnium capillaceum (OBL) 12.50
Phyla nodiflora (FACW) 11.70









Table 4-9 Ten species with highest usability index for urban areas of combined lakes
Species name Usability index
Urena lobata (FACU) 333.33
Bidens alba (FACW-) 333.33
Caxex longii (OBL) 333.33
Alternanthera philoxeroides (OBL) 60.00
Sida acuata 33.33
Baccharis halimifolia (FAC) 33.33
Coreopsis leavenworthii (FACW) 33.33
Desmodium incanum 33.33
Eryngium prostratum (FACW) 33.33
Polygonumpunctatum (FACW+) 27.03














CHAPTER 5
DISCUSSION

Remote Sensing Aspects

Case law requires that the best methods available should be used for making

ordinary high water determinations. There is little potential for using satellite remote

sensing to locate the OHWL. The degree of usefulness is related to the type of imagery

used and to the nature of the shoreline. The areas studied in this project were broad, low

bank shorelines. Landsat, SPOT, and IRS were not usefulness because of their relatively

large ground cell size. Townsend and Walsh (2001) reported that complex community

relationships and the nature of their spatial variability are rarely captured with satellite

imagery. Broad ecotones are especially difficult to portray accurately because they are

made up of transition zones that grade into each other across complex soil, hydrologic,

and topographic gradients. Evidence from this study supports that contention. The

variety of vegetation types found in different areas in this study impacts the spectral

signatures, leading to a range of classes, none of which led to a clear demarcation of the

OHWL at or near the upper littoral zone.

As illustrated in the summary table of pixel class designation changes

corresponding to the known 52.50 ft (Hatchineha) or 52.00 ft (Kissimmee) AMSL

ordinary high water line for each of the images used in this study (Table 4-1), the only

one that emerged as having potential to delineate a boundary line such as the OHWL was

CASI. In the case of Landsat and SPOT, only one of 17 (5.88%) of the transects had a

change in pixel class designations within 2 m of the OHWL. The pixel class designation









on the landward side of the OHWL for Landsat Transect 31 was Hydric 3 and the

waterward side was Hydric 4. For SPOT Transect 32, the landward side was Hydric 3

and the waterward side was Hydric 2.

The IRS image was able to detect one change in pixel class designations within 2

m of the OHWL for a total of one of thirteen (7.69%). The IRS image corresponded to a

change at Transect 33. The landward side of the OHWL was classified as Hydric 3 while

the waterward side was classified as Hydric 2. This transect corresponds to an area of

Lake Hatchineha where the OHWL falls within the cypress. Even CASI, with its smaller

ground cell size, was unable to detect a shift at the OHWL for this transect.

IKONOS detected a change in pixel class designations corresponding to the

OHWL for two of fourteen transects (14.29%). These results are better, but the change in

pixel class designation may not be within the desired 1 m margin of error due to the 4 m

ground cell size. Even though each of the satellites was able to detect a shift

corresponding to the OHWL for at least one transect, results were not consistent.

Consequently, none were useful for locating a boundary line such as the OHWL.

Landsat, SPOT and IRS have the largest ground cell size of the sensors tested and

showed virtually no potential for delineating a boundary line or detecting subtle

vegetation shifts on the ground. Fuller et al. (1998) found that Landsat was even unable

to distinguish between a swamp with five species and a swamp with 33 species. Landsat

has been shown to be useful for classifying broad categories of land cover, and it is

therefore likely that SPOT and IRS, which have slightly smaller ground cell sizes, would

also be useful for broad classification schemes. However, parameters for this study were

much narrower, requiring the delineation of a boundary line.









As Harvey and Hill (2001) reported, the characteristic vegetation gradients found

in wetlands are narrower than the pixel size of current sensors, which often results in the

inability to distinguish between different vegetation types. It is conceivable that use of a

different classification scheme would have improved the discrimination capabilities of

this study. Townsend (2000) assessed the 'quantitative fuzzy approach' to assess

classifications of mapped vegetation. In this approach, fuzzy classifications and

mixture/subpixel models provide information about the degree of similarity between

ground locations of natural vegetation and multiple classes on a map. It facilitates a more

detailed interpretation of map classification and, therefore, reduces the arbitrariness of

evaluations. It also provides a greater understanding of the subtleties in species

composition and how they relate to the fuzziness of the natural vegetation classes. In

order to perform such a classification, it is necessary to know the ecological relationships

within plant communities. The vegetation data must be characterized as a percent of an

area (e.g., 100% cypress or 86% soil). Vegetation information obtained from Lakes

Hatchineha and Kissimmee for this study was only determined along line transects, so

there was no way to assess percent of land cover. However, even if a fuzzy classification

could have been done, it is doubtful that a narrow boundary line could have been detected

from Landsat, SPOT or IRS images because of their resolution limitations.

Another constraint of this study was the lack of specific ground cover information

to correlate to the satellite images. Conditions on the ground had almost certainly

changed between the time that Landsat, SPOT and IRS images were taken (October

1999, March 1993 and December 1999, respectively) and the time vegetation was

sampled in April and May 2002. However, this constraint was known at the outset and









was not considered significant to the conclusions of this study. The generally accepted

OHWL was known for both lakes, and vertical elevations measured in the field placed

the starting point for each transect on the OHWL. GPS coordinates allowed placement of

each transect's point associated with the OHWL on the satellite images. Even though

vegetation had probably changed in the years since the satellite images were recorded, the

vertical elevation had probably not. Therefore, once the point corresponding to the

OHWL was located on the image, a change in pixel class designations associated with

that location would be discernible if one existed. A change would indicate a vegetation

shift corresponding to the OHWL. Since this study did not attempt to assess change in

land use/land cover, it was unnecessary to know what vegetation was present at the time

the images were taken. The key constraint was ground cell size, which determines the

ability to discern very small changes in vegetation or to discern a boundary line such as

OHWL within the preferred 1 m margin of error. Because of their relatively large ground

cell sizes, Landsat, SPOT and IRS were not well suited for this task.

The IKONOS satellite, with a smaller ground cell size of 4 m, theoretically has

more potential to provide information that would be useful in locating the OHWL.

However, the images used in this study demonstrated only one change each at the OHWL

for Lake Hatchineha and Lake Kissimmee using this sensor. The area where the OHWL

coincided with a distinct change in pixel class designation on Lake Kissimmee (Transect

7) corresponded to an area with a predominance of Bahia grass (Paspalum notatum),

although six other species also occurred in this area. Bahia grass is common in both

ranching and urban sites around Lake Kissimmee. Transect 7 is an urban site, and it is

likely that the Bahia recorded in this study was also present at the time these images were









taken in March 2000. Interestingly, Bahia does not appear on the list of species that

occur closest to the OHWL, and had a usability index (UI) rating of only 16.67 (out of a

possible 1000). Therefore, Bahia could not be relied upon as a good indicator of the

OHWL. It is likely that the pixel class change seen in the IKONOS images related to a

natural berm on the shore. Perhaps the sensors were detecting more sand and dirt

because vegetation was not as dense as it was in other areas. Furthermore, because the

ground cell size of these images was 4 m, the ability to locate a boundary line within a 1

m margin of error is limited.

CASI hyperspectral imagery, with a ground cell size of 1 m, had the best results

in terms of discerning plant community changes. However, this result is based on a small

sample size. In the summary table (Table 4-1), CASI images showed the best results

with three of four transects (75%) detecting a vegetation shift corresponding to the

OHWL. CASI's 1 m ground cell size is within the desired 1 m margin of error needed

for defining a boundary line. The OHWL coincided with a break in pixel class

designations on the images for two of three transects (Transects 31 and 32) on Lake

Hatchineha. However, in both of these transects the break in pixel class designations at

the OHWL coincided with the edge of the cypress canopy. In the third Lake Hatchineha

transect (Transect 33), the OHWL fell within the cypress canopy and could not be

detected in the CASI imagery.

On Lake Kissimmee, only Transect 9 was included in the CASI imagery data set.

There was a clear change in pixel class designations at the OHWL, but it was not clear

what caused this change. There were no distinct vegetation shifts on the ground at the

OHWL, although a natural berm did occur there. It is possible that the sensors picked up









reflections at different angles because of this berm, resulting in the classification change

in the CASI imagery. Additional studies should be conducted with CASI imagery to

more accurately assess its ability to demonstrate changes in pixel classes corresponding

to the OHWL. However, because CASI imagery is expensive compared to satellite

imagery (CASI $18,215/km2 vs. Landsat $0.60/km2), its use as a cost effective means

of locating the OHWL may be limited.

One of the largest hurdles in attempting an automated approach to the

determination of the OHWL using remote sensing was the variation in vegetation

indicators. For example, oak trees (Quercus geminata) around Lake Kissimmee are good

indicators of the uppermost reach of water. The OHWL sometimes fell at the edge of

oaks, but never beyond them. This upland species of oak can be detected with remote

sensing and, therefore, can sometimes be used to delineate a line beyond which water will

not be present. Oaks around Lake Kissimmee sit on a natural berm and the OHWL

usually falls somewhere on the lakeside of the berm. However, many oaks are far

removed from the lake, growing beyond small wetlands with a wide variety of vegetation

types in between, and therefore, are not good indicators of the uppermost reach of water

in those areas.

Lake Hatchineha and some other Florida lakes are surrounded by cypress

(Taxodium) stands. When the ordinary high water stage occurs within a cypress stand, it

would be nearly impossible to detect this water using the imagery available for this study.

Figure 5-1 illustrates this circumstance by showing an area of Lake Hatchineha with

water present within the cypress that was not clearly detectable with the satellite or CASI

imagery.




































Figure 5-1 Catfish Creek entrance to Lake Hatchineha. Note water within the
cypress.

Another drawback related to the use of remote sensing for locating the OHWL

relates to the nature of 'low bank lakes' such as Lakes Kissimmee and Hatchineha. This

characteristic presents a special problem in finding evidence of the OHWL using remote

sensing techniques. Slight water elevation changes cause large shifts in the lateral

position of the shore. This results in large littoral zones with diverse vegetation types and

becomes problematic for determining distinct vegetation indicators of the OHWL,

particularly by remote sensing. On the other hand, remote sensing may detect a line at

ordinary high water for steep bank lakes because vegetation may not be as prevalent

along a steep bank, and open water would provide a distinct spectral signature.









There are broad ecological uses for satellite remote sensing in delineating land

use/land cover, but its application for accurately locating a distinct line is limited,

especially on low bank lakes. This limitation is largely due to the fact that vegetation

shifts on the ground are not clearly evident in imagery of these resolutions. Even under

normal hydrologic regimes when vegetation would likely fall out in clearer patterns,

satellite imagery would probably not be able to accurately ascertain the location of the

OHWL because of the difference between the narrow vegetation bands and the large

ground cell size of the imagery. CASI has the greatest potential because of its 1 m

ground cell size, but may be limited due to its expense. Further study of CASI imagery

should be conducted to determine whether it can consistently detect spectral shifts

corresponding to the OHWL.

Vegetation Aspects

There is historic evidence that water levels were significantly higher and less

aquatic vegetation existed in the lakes prior to the channelization of the Kissimmee River.

Hydrologic modifications have resulted in stabilized water levels and have altered

vegetation associations within the marshes. Lake Kissimmee and its associated marsh

now have floating tussocks, mixed grass marshes, and nearly floating pickerelweed

marshes (Kushlan 1990). At the time of this study, south Florida had experienced an

extended drought that very likely caused further alteration of the vegetation community.

According to local ranchers and fishermen, since lake levels are now controlled, water

flow in and out of the lakes is reduced and hydrilla and water hyacinth have become

problems in the lakes as well. The increased vegetation has resulted in a need for

periodic lake draw-downs and muck removal. The resulting waste is piled up along

various locations of the lakeshore and creates 'spoil' islands. These islands further alter









plant communities as they move seed banks around and place muck and debris in areas

that were once either open water are flat shoreline.

Vegetation analysis elicited no good vegetation indicators of the OHWL.

Shorelines are dynamic environments and during periods of drought, as existed in the

study area, existing seed banks may germinate, increasing the diversity of species close to

the lakeshore (Keddy and Fraser 2000). It is likely that the FACU vegetation seen in the

lower littoral zone occurred as the result of prolonged low water levels in the lakes.

Furthermore, cattle ranching in many areas has impacted Lake Kissimmee, and

vegetation types seen in these areas include 'improved' pasture consisting of non-native

species. The result is a conglomeration of FACU, FAC, FACW, and even OBL species

in many areas of the upper and lower littoral zones. Normally, the lower littoral zone

would have standing water or be adjacent to it (Davis 1973). Obligate wetland vegetation

would be expected in this zone, including species such as Pontedaria cordata, Sagittaria

lancifolia, Typha latifolia, and Panicum hemitomon. These species did occur in the lower

littoral zones of Lakes Kissimmee and Hatchineha, but they failed to fall into a consistent

pixel class designation on the classified images.

In vegetation analysis, frequency of occurrence data showed that only seven

species (out of 93 total) showed up in more than half the transects thus indicating that

there were no consistent vegetation types that could be used as potential indicators of the

OHWL. When lakes were divided by land use type, the highest usability index value

(333.33) occurred within the urban class. However, because the three species with a 333

UI occurred in only one of 17 transects (Bidens alba and Carex longii) or three of 17

transects (Urena lobata), they are of limited importance. It appears that when these









species occur, they occur at the edge of the OHWL, but because these species were not

abundant, they could not be relied upon as good indicators of the OHWL.

The fact that the frequency of occurrence table and the average minimum distance

table shared only two common species serves as further evidence that vegetation was not

very useful for locating the OHWL. The most frequently occurring species (Polygonum

punctatum, Pontedaria cordata, Hydrocotyle umbellata and Paspalum notatum) occurred

in 13, 13, 12, or 10 of the 17 transects, respectively, but none of them consistently

occurred near the OHWL, again confirming that there were no good vegetation indicators

of the OHWL.

Conclusions

If one wanted to obtain only a crude approximation of the acreage of sovereignty

land in Florida, IKONOS could be used. However, in order to obtain an accurate

determination of acreage, the OHWL must be pinpointed because it serves as the

boundary line between sovereignty and private land. To that end, the only potentially

useful remote sensing technology assessed in this study was CASI. Landsat, SPOT and

IRS would not provide a reasonable estimate because their ground cell size limits their

ability to distinguish vegetation shifts corresponding to the OHWL. Even IKONOS is

limited because the 4 m ground cell size does not meet the desired 1 m margin of error.

Consequently, this study will not resolve the continuing legal battles surrounding the

OHWL. Case law has dealt with this issue for nearly a century. The hope that future

legal battles would be easily settled because the OHWL could quickly be pinpointed

using remote sensing technologies has been dampened. The results of this study found

no good and inexpensive technology for the job.









In terms of vegetation analysis, there were no consistent indicators of the OHWL

found in this study. Those vegetation types that occurred at or near the OHWL did not

occur frequently enough to be useful. However, the drought conditions experienced in

the lakes prior to and during the sampling period have likely altered the vegetation

significantly and it would be useful to reassess vegetation indicators under more normal

hydrologic regimes. Under more normal conditions, vegetation would be expected to fall

within distinct hydrologic zones. It would be expected that areas closest to the water

would be dominated by OBL and FACW species, while FAC and FACU species would

be found further landward. At the time of this study, FAC and FACU species were often

found well below the OHWL.

Use of quadrat field investigation method rather than the line transect method to

collect and analyze vegetation may also have been more informative. Quadrats would

have provided information about percent coverage of vegetation and could have been

used in combination with fuzzy classification schemes to assess the ability of various

remote sensing technologies to detect a shift at the OWHL. Elevation information would

also have been useful to collect in combination with vegetation. Hydrology and

vegetation are inextricably linked, and knowing vegetation elevation would have

provided useful information that linked hydroperiods and frequency of inundation to

vegetation by comparing elevations to lake stage data. It is possible that under normal

hydrologic conditions, vegetation would have been a useful indicator of the OHWL as

Davis found in his 1973 research.















APPENDIX A
VEGETATION CHARTS

In each of the following charts, the red line represents the OHWL. Each column

represents a 1 meter distance on the ground. Columns with positive numbers represent

land above the OHWL while columns with negative numbers represent land below the

OHWL. Any vegetation that occurred along the transect was recorded and is represented

by a colored box in the appropriate column. Wetland designations are noted in

parenthesis next to the species name. Vegetation listed without a wetland designation are

those not listed in the National List ofPlant Species that Occur in Wetlands.



















Transect 1 Ranching
Lake Kissimmee
Name/Classification


Paspaium notatum (FACU)
Sesbama herbacea (FAC)
Wynca cenfere (FAC)


'enielia asrat.ca (FACW)
'yperus haspan (OBL)
Bacopa caroinana (OBL)
Ludwgsa arcuata (OBL)
4temranthera phi.oxero.des (OBL)
Hydrocotyle umbeblta (FACW)J
Dodia vwrgmana (FACW)J
Hydrochioa caroiinrensos (OBL)


PoThgonumpunciatum (OBL)ii! i
Ponieder.a cordata (OBL)
Sag9ltaira lancfola (OBL)


: 'I '1, I' 1 I 'I' I 11 0 1 011
S: 7 6 4 3 2 19 65


1 23 4 5 678 9 0 1


- *j I I :


2 2 2
2 3 4


111111111111111111111111111


T__

R__


ST,,, i ii I -.


II I I I I I I I I I I I I I I I I I I I I I I


S II I I II I


Vegetation Transect 1, ranch land use, Lake Kissimmee


Figure A-


c


SI



































Figure A-2 Vegetation Transect 2, ranch land use, Lake Kissimmee


SAN

SAND


5 5 61
s1d1'1i




















rransect 3A
Natural Area (was ranching)
-ake Kissimmee
\lame/Classification


S1 1 1 1 2 2 2 2 2 2 2 3333 3 3 3 3 33 44444444445
111 22 -4 7 -9 1 1 2-3 -5 -7 -0 -3 -5 -7 -9 -2 -4 -6 a9 -2 -4 -6 -8 -0


.rcs g..em.a.a oak I i i
"aspfutm notatum (FACU)
-uthamia carodnana (FAC)
Ecoprna dlcis (FAC)
tesbamn ferbacea (FAC)
yp-rus lecontea (FACW)
;y&Vocofe umbeJlafa (FACW)
loadoa v -grana (FACW)
Digitana serotrna (FAC)
4temantheraphlioxerodes (OBL)
acYpa carohn-anar (OBL)
Eleochans sp (OBL)
4xonopus furcaus (FAC)
yperus poaystachyos (FACW)
lonfedena comrata (OBL)


A Highet eviaran is 5 16 and oc ur 329 m ovethe OIlWL Beyndothis point, land dop do seanwih wetland etan te oaK canpy tr endat 38 m, Tie under ote K cp i sad


Vegetation Transect 3A, natural area land use, Lake Kissimmee


Figure A-3


. i i-i 1. 1 1-1 -


".














Sransect 4 Kanching
Lake Kissimmee
Name/Classification
Paspalum notatum (FACU)
Euthamia carohnIana (FAC)
Eryngwum prostratum (FACW)
Sesbana herbacea (FAC)
Rhynchospora macrocarpa (OBL)
Centelia asiatca (FACW)
Pontedena cordata (OBL)
Bacopa carolinMan (OBL)
Ludwigja palustns (OBL)
Pan cum repens (FACW)
EBeochanssp (OBL)
Hydrocotyle umbeIata (FACW)
Sagfttana lancfola (OBL)


Figure A-4


4- 1 1 1 1
7 6 5 4 3 2 1112 3 --61718 9 0 11213 5


I.ThTFTh Th.


Vegetation Transect 4, ranch land use, Lake Kissimmee


A ,11- "-- -I- [.,. _1 ,, I I [ I I 1 .. .- ,- -.,H I,-, I:I I I ,II ,T js I1 11 :,-i C--,-1 h, '-1 -[h'I" II,-[ ir1[-- 1, l,[ l .I I ,,-III. I- II ,[11

















TransectA Ranching
Lake Kissim mee
Name/Classification
QuecusgemFndaa oak
Sab! paitdto (FA)
J,- us (OBL


: 1: 11 1 1 1Id 3.1: 11 10 2111.:-1 1 :: 1 1. 1I I
44 I 1 M I I I + I2


IIIII-


2 2 21 2 2 2 212 3 3
34567890 1



Al RBOAT
TRAIL


I II I I I I I


111111111111111 II II II111111


Vegetation Transect 5A, ranch land use, Lake Kissimmee


Figure A-5
























L5 4 4 4 4 4 4 4 4 4 4 3 3 3 3 3 3 3 3 3 3 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 3 3 3 I 1 4 4 4 5
ame las 9s 1 5 t 3 1 1 9 0 1 65 3 2 1 0 9 6 5 3 2 1 0 9 6 5 3 2 1 8 7 6 54 2 1 1 2 3 5 6 7 9 0 1 2 3 56 7 9 01 0 9 12 1 1 2 4 5 0


ETqo, llI l l J, AC)l
h v o (FA084 C)
-. .emna a t 'ACU)
)/p-r Ie-tea (FAC'V)
igure A 6.ettn s (FAC6eL

W-- 0(FAC)
-Jy d nO rarJ (FACW) -AOp



Dyp iu*ohacIhyl (FTC EI l


oIt l(FEACW)
s', ,', ~a/tz':of- -O L -- -- ---- -




y retf- ta (FAC)
ypha ..t'a (opL)





Figure A-6 Vegetation Transect 6, natural area land use, Lake Kissimmee
I Ill~l

















rransect 7 Urban
Lake Kissimmee
Mlaie/Classification
:oryza canacsns.s (FACU)
?yperus croceus (FAC)
easpaSa noataum (FACU)
'hyla nodtora (FAC)
Richabtra btasrens.s
iyperus poystachyos (FACW)
4xonopus fulcatus (FAC)
'utham-a carofniana (FAC)
nyperus suinameins (FACW)
WVuroama nuafMora (FAC)
iuncus dclhotoQUS (OBL)
4terantheray phloxeriodes (OBL)
Zetrel.a as.at&ca (FACW)
Dyperus ecoe, (FACW)
'ontedera coreata (OBL)


3acopa carolniana (OBL)
'agttara fancltol.a (OBL)
vefuro ltea (OBL)


-1 1 -1 -1 l I I, I I
1 0 5.Z. -.1j j : 16 c -I I' I: 1 19 1.17 1 j III j 111 2 3 4 5 6 7 8 0 2 0121
F II1111


'-AhAuse sits at342 rabethe OHL. [ highele'ain H4&),tthereforetransctbegaat pin


Figure A-7


4 4 4 4 4 5
546474B49 0


L -II


00
Vegetation Transect 7, urban land use, Lake Kissimmee


1.1^111111111111111111-^1


- - - - - -







































Figure A-8 Vegetation Transect 8, urban land use, Lake Kissimmee



00
w



















Transect 9 Ranching
Lake Kissimmee 11 111 222 222 333333
Name/Classification I 9 7654321 9876543211 234567890 123456789012345678901234567 89 123456789
Raspasum noteturn (FACU)
Sesbana herbacea (FAC)
Eulhama carohnmana (FAC)
Axonopas turcaOus (FAC)
Rhexa cubensls (FACW)
Junus effusus (OBL)
Certeda asrabca (FACWV)
Rhynchospore sp .
Andropogon sp ....
Phy5i nodrflora (FAC.
A/ernaninera pn.loxeroide3 (OBL)
PtIrmn.um capa ceum (OBL)
Ludwmga ercuata (OBL)
Pancum repens (FACVW) ...
Poy/onum punclalum (OBL)
Hydrocoiyie umbelala (FACW)
Bacopa caromnlana (OBL)
Cynodon dectyon (FACU) ..
Eleochans sp (OBL)
Ponledenra cordala (OBL)
00
I I I i Lh I L -- J ,. I. J


Vegetation Transect 9, ranch land use, Lake Kissimmee


Figure A-9















Transect 10A Natural Area
Lake Kissimmee .. .. .. 1 1 1 1 11 1 11 1 1 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 4 4 4 4 4 4 4 4 4 4 5
Name/Classification ai5 4 312 11 2 314151617 8 9101 234567890111234 567 90 112 3141516 7 8 9 01 2 3 4 5 6 7 8 9 0
Mynca cenfera (FAC)
Schnus terebnthifoi~us (FAC)
Diospyros virginiana (FAC)
Urochioa mut.ca (FACW)
Hibiscus grandiflorus (OBL)
Polygonumpunctatum (OBL) III II
Panicumr hemtomon (OBL)I II
Sagitana lancifo.,a (OBL) I I
Djoda virgniana (FACW) I
Nympha odorata (OBL)I I I I

1 1 .I I I I I I I I I I I I I III I III II I I II 1, I II


Figure A-10


Vegetation Transect 10A, natural area land use, Lake Kissimmee



















Transect 31 Natural Area
LakeHatchineha 1111111 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 4 4 4 4 4 4 4 4 4 4 5
Name/Classification 5 4 3 2 1 0 7 4 54 32112345 678 2 1 2 3 4 7 9 0 1 2 3 4 5 7 1 2 3 4 7 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0
Taxodaum dsticham (OBL) A
Urochloa maf ca (FACW) T
Pancuhem somon (OBL) R R
Panum repens (OBL) B A
Pontedela cordaia (OBL) O
Saglaraliactolfa (OBL) A L
Polygonum punctatum (OBL) T


Figure A-11 Vegetation Transect 31, natural area land use, Lake Hatchineha


















Lake Hatchineha 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 23 3 3 3 3 3 3 3 3 3 4 4 4 44 4 4 4 4 4 4
Name/Classification -_ 1 1 9876543 2112345 678901234567890123456789 12 4 567890 123 789 0
Taxodrum &Ochum (OBL)
Sabai ,ameiio 0(FAC) 0 0
rum nutti (FAC) S A N D A P P
Paspum conjuaItum (FAC) ..I T E E
Pouzoiza zeylaca ___ R R N N
Pmyconumpundalum (OBL) B A
pJuncus sp OBL L) 0 o W W
Comment dffusa (FACW) .A L A A
Phyla nodiora (FAC) .T T T
UroNo mutbca (FACVJV) .. E E
Pamcum hemromon (OBL) R .R
Sapana Ianc.foiua (OBL)

i I :, I I i .. .. ii ~ hi i~ i i i i I I h~ ,, I 11 II ,I


Figure A-12 Vegetation Transect 32, natural area land use, Lake Hatchineha
































Figure A-13 Vegetation Transect 33, natural area land use, Lake Hatchineha








00
00



















Transect 34
Natural Area (was ranching)
Lake Hatchineha 1 1 1 ....1 11111111 12 2 2 2 3 3 3 3 3 3 3 4 4 4 4
Name/Classification 5 o31098765432112345678901234567890123456789012345678901234567890
Taxodium dostchum (OBL)
Vhs rtounfoha (FAC)SAND
Swda acuata A
Rjchardia bras.ilenss
Eupalonum capr 'wftum (FAC)
Phyia nOdcflora (FAC B1
Pblhmnmm capWcaceum (FACW)0
Eclpte prosir aa
Hydrocoiyle umbellaa (FACVV) T
Hydrochloa carohmenss (OBL)
Cyperus arbculaus (OBL0 T
Bacopa carolrnana (OBL) I
AMternarnhera pnloxerodes (OBL) A
Eeochanssp (OBL)
Polygonumpandatum (OL) L
Cynodon dactyon (FACU)
Ponptederja cordata (OBL)

11- [ -[I I I I I B I- [ I I I,
F, I II IO J ,
h I ,, ,,, I-, lh LL .. ,, : A


Figure A-14 Vegetation Transect 34, natural area land use, Lake Hatchineha

















Iransect 36 Kancning
Lake Hatchineha
Name/Classification
Taxodm dstichum (OBL)
Soiarnim dtphyium
S'da acurta
Uiena lobata (FACU)
Richardia brasifensfs


Cyodon dactyon (FACU)
Aitermanthera phsoxefordes (OBL)


,J :.:. : ;: [:-';,, ,: .lU


;AND


I.=HIHl111


' 1 112


I-
J tIItI :-





I rI


w
A
P
T
E
E
N
R


Cyperw sfgoeas (FAC_
Pti!mnm ? eum p. (FAC. )
Ba3pa caro-ri-ana (OBL)1 1 1 1 1 1 1 1 1 1 1 1
Phyia nodioa (FAC)


EcUptaprostrata


Figure A-15 Vegetation Transect 36, ranch land use, Lake Hatchineha o


_ __


' '


I


II