Topographic variation in south Florida, though generally small, is very important for the flora and fauna of the region. Recently the U.S. Geological Survey (USGS) started the High Accuracy Elevation Data (HAED) collection project with the goal of obtaining an estimate of topography in south Florida at a high spatial resolution (nominally 400x400 meters) with high vertical accuracy (vertical estimates are accurate to within about 3 cm). The project has now collected, processed, and made available topographic data for a large portion of the natural areas in south Florida, including most of the Everglades National Park (ENP), portions of Big Cypress National Preserve (BCNP) and Water Conservation Area 3.

Prior to the existence of the HAED project there were no topographic maps of south Florida with both the spatial resolution and spatial extent needed to model the ecology of fauna and flora in south Florida. To fill this void, the Across Trophic Level System Simulation (ATLSS) project developed the High Resolution Topography (HRT) model. This model estimates elevation for 30x30 meter grid cells across the natural regions of south Florida. The estimation is based on the types of vegetation and the patterns of hydrology at each location. The founding assumption of the HRT model is that plants tend to be in places where local topography and hydrology combine to create a suitable habitat. Knowledge of the vegetation and hydrology at a location can be used to estimate elevation by choosing an elevation that results in hydrologic conditions suitable for local vegetation.

Three data sets form the basic inputs into the HRT model: a map of the distribution of vegetation in south Florida, a map of the history of hydroperiod distributions, and a table of hydroperiod preferences for each of the vegetation types in the vegetation map. The hydroperiod preferences are derived from a review of available literature. The vegetation distribution map used is the Florida GAP (FGAP) map (version 2.1) created by Leonard Pearlstine at the University of Florida. The FGAP map assigns one of 43 vegetation types to each 30x30 meter cell over most of South Florida.

The hydroperiod data for the HRT model is created from the Calibration/Verification run of the South Florida Water Management Model (SFWMM). The SFWMM is managed by the South Florida Water Management District (SFWMD) and is the standard hydrologic model for South Florida. The Calibration/Verification run of this model is considered to be the one that most accurately reflects the historical pattern of hydrology in South Florida from 1979 to 1995. The SFWMM partitions south Florida into a grid, where each grid cell is 2x2 miles. The model estimates water depth in each 2x2 mile cell on a daily time step from January 1, 1979 to December 31, 1995.

Over the past two years there have been a number of advances in the ATLSS High Resolution Topography (HRT) modeling project. An extensive literature review, Plant Community Parameter Estimates and Documentation for the Across Trophic Level System Simulation (ATLSS) by Paul Wetzel has been completed and peer-reviewed. This document describes the hydroperiod preferences of the natural vegetation types used in the 6.1 version of the Florida GAP (FGAP) map.

The second major advance is the development of a new HRT map. This new version is based on FGAP 6.1, hydrology data from the South Florida Water Management Model (SFWMM) version 3.7, and the hydroperiod estimates that appear in Wetzel's report. We are currently waiting for a more current version of the SFWMM Calibration/Verification output before completing a final version of the new HRT map.

Finally, we have begun an analysis that looks at the relation between HAED elevations, distribution of vegetation as provided in the FGAP map, and hydrologic parameters as predicted by the SFWMM. This analysis uses a multiple regression model with vegetation type and a number of yearly average hydrologic variables as predictors, and elevation estimates from the HAED collection project as the response variable. There are two major purposes in performing this analysis. The first is to test an assumption made by the HRT model that there is a relation between the elevation of a location and the vegetation associated with that location by FGAP. The second is to determine if a multiple regression model can be used to predict HAED elevation based on vegetation and hydrology. Both of these results provide a basis for evaluating the output of the HRT model.

This analysis has been completed for the Big Boy Lake HAED sampling unit. The regression coefficients for several of the vegetation types in this region are significantly non-zero. This indicates that vegetation provides some information about variation in topography. However, the r² and PRESS (Prediction Sum of Squares Statistics) values for the regression model are not sensitive to the presence of vegetation as a descriptor variable. This indicates that in the Big Boy Lake region the overall contribution of vegetation to explaining variation in elevation is small. The conclusion we draw from this result is that for the Big Boy Lake HAED region we do not expect the HRT to perform well when compared to the HAED data because the inputs to the HRT model have very little relation to the HAED elevation data. The limited scope of this analysis leaves a number of unanswered questions. The relation between vegetation and HAED elevation in other regions and at other spatial scales has not yet been explored. While working with the HAED data we have noticed that elevations were not sampled in some tree islands. According to Greg Desmond, HAED project leader at USGS, the methods used to collect elevation data for HAED result in the systematic omission of tree islands. The effect of the sampling bias that arises from this omission is to compress the variance in measured elevations relative to the actual variance.

The HAED data will become the primary basis for topography used by ATLSS models. However, the availability of the HAED topography does not completely eliminate the need for the HRT model. In the short term, there is still a need to use HRT estimates of topography for regions not yet covered by the HAED project. In the long term, the methodologies used in the HRT model may be needed to augment the topography generated by the HAED. We propose using the HRT map and the HRT methodologies to augment the HAED topography in areas where HAED has omitted vegetation structures, or has compressed variation. Work has begun on identifying regions where the HRT model can be useful for augmenting the HAED topography and the methods needed to incorporate HRT output into a HAED-based topography.

Contact: Duke-Sylvester, Scott M., University of Tennessee, 569 Dabney Hall, Knoxville, TN, 37996-1610, Phone: 865-974-0223, Fax: 865-974-3067, sylv@tiem.utk.edu

(This abstract was taken from the Greater Everglades Ecosystem Restoration (GEER) Open File Report 03-54)

Back to Project Homepage