Project Summary Sheet

Study Title: Development of Selected Model Components of an Across-Trophic-Level System Simulation (ATLSS) for the Wetland Systems of South Florida
Study Start Date: 1997 Study End Date: 12/31/2008
Web Sites: ATLSS.ORG
Location: The Greater Everglades ecosystem
Funding Source: ENP Critical Ecosystems Studies Initiative (CESI) and USGS Priority Ecosystems Science (PES) Initiative
Principal Investigators: Louis J. Gross, The Institute for Environmental Modeling (TIEM), University of Tennessee, Knoxville, TN 37996-1610 Phone: 865-974-4295 e-mail: gross@tiem.utk.edu
Study Personnel: Donald L. DeAngelis, Phone: 305-284-1690 e-mail: ddeangelis@umiami.ir.miami.edu
Other Supporting Organizations: USGS Greater Everglades Priority Ecosystems Science (GE PES) Funding, NPS, ACE, EPA
Associated Projects: Component of ATLSS Program, SFWMD Hydrology and Environmental Modeling; USGS National Wetlands Research Center

Overview & Objectives: The ongoing goals in this project have been the following: 1) As part of the ATLSS Program combine biological and physical models using multiple approaches, including trophic system components at differing spatial and temporal scales, to estimate landscape-level responses of biotic systems to environmental changes. 2) Produce models capable of using detailed behavioral and physiological data and dynamically linking these with spatially-explicit abiotic information. 3) Produce models capable of generating testable hypotheses about trophic component responses to alternative possible anthropogenic influences. ATLSS provides a mechanism to evaluate the relative impacts of alternative hydrologic scenarios on various trophic components. The methodology to integrate components involves: i) a landscape structure for dynamic communication between models; ii) a high resolution topography to estimate high resolution water depth across the landscape; and, iii) a variety of visualization tools to aid model development, validation, and comparison to field data.

Status: Current CESI funding has been extended to 12/31/2006. This includes additional FY04 CESI funding and additional PES funding added in FY05

Recent Progress: Progress is being made on testing the reliability of ATLSS's Spatially Explicit Species Index (SESI) models by comparing year-to-year trends in SESI outputs with year-to-year trends in empirical data. An example of such comparisons is the Cape Sable sparrow population for a given sub-region, the 'western subpopulation,' of its range. Data are available for 1981 and the period 1992-present. Because the SESI model produces values that are on the interval 0 to 1, in order to compare SESI model trends with population data, the SESI values must first be normalized to the population data. When this is done the SESI index does seem to track population values over most of the period from 1992 to 2002. An ATLSS vegetation succession model, VSMod, encompassing 24 vegetation types, has been developed. VSMod simulates the pattern of spatial and temporal changes in the distribution of vegetation in the Greater Everglades landscape as a function of the hydrologic regime, patterns of fire disturbance and nutrients. A primary goal is to quantify the relative differences between various hydrologic scenarios as reflected in their impacts on vegetation succession. VSMod incorporates a spatially explicit, stochastic cellular automata model to simulate vegetation succession. At any given time, each 500x500-meter plot is in one of a finite number of states. The transition between states occurs with a probability that varies in both space and time, dependent on local hydrologic and fire history as well as on the current vegetation. The model runs on a yearly time step, synchronized with the fire model, and produces annual maps of vegetation over the model area. VSMod rules are based on reports by Wetzel (2001a, 2001b). Three modeled factors influence the succession of one plant association to another: fire, nutrient change, and prolonged change in hydrologic conditions. The model is now available for use, though parameterization of the fire component is still ongoing. The fire distribution data available to us is restricted to Everglades National Park, and this limits the ability to compare model estimates of fire size distribution across the entire ATLSS modeling region. To address this, a series of extensive simulations are underway, using parallel processors, to provide best estimates of model fire parameters of the mean fire size distribution restricted to the region and time period for which data are available. A set of sample VSMod simulations have been posted on the ATLSS web pages to allow input from interested agencies on the process. These simulations have been done using a range of fire scenarios (high, medium and low) to allow comparisons of the effect of different fire regimes.

To make possible direct access to ATLSS model results for natural resource managers, a new ecosystem modeling service suite has been developed. Program Model Interface, developed through the assistance of the National Science Foundation, provides a convenient means for natural resource managers to use ATLSS's high performance ecological modeling package through their own PCs. The interface is developed to provide an intuitive interface for model modification, while removing the burden of model software and hardware management and acquisition. The ecosystem modeling service suite allows natural resource managers in South Florida to remotely launch ecosystem modeling tasks on, and harvest model results from, a high performance computing grid at the University of Tennessee with the use of minimal local resources.

A new SESI model has been developed for the Florida panther. The Florida Panther Spatially-Explicit Species Index Model (FPSESI) provides annual indicators of potential suitability for panther movement, dispersal, and prey resources across the Everglades and Big Cypress landscapes. FPSESI was developed to assess the relative suitability of alternative hydrologic scenarios by evaluating beneficial and adverse effects on the panther and its primary prey, white-tailed deer. The model considers the panther's typical behaviors and life-history traits that influence responses to a range of spatial and temporal hydrologic conditions. The user will be provided with options for weighting positive and negative factors to produce a combined suitability index. Water depth is modeled as a factor in daily movements of panthers within home ranges and in the dispersal of subadults from the natal range. Metrics of connectivity across Shark River Slough include depth, spatial extent, and periodicity of innundation. Impacts to panthers via prey are evaluated as effects of water depths on the movement behaviors and breeding and foraging habitat of white-tailed deer. Restoring and maintaining 4-to-7 month hydroperiods in the habitats that produce high quality deer forage would have a beneficial effect, while prolonged periods of innundation would have adverse impacts on deer movements and breeding success.

Planned Products: ATLSS has been requested to provide model output to DOI and other agencies for CSOP. This will be done through the ATLSS User Interface. Continuing work will be done on testing and upgrading the SESI and other ATLSS models. The vegetation model (VSMod) will be tested and integrated into the SESI models.

A new ATLSS landscape class structure is being developed that will allow ATLSS runs to be effectively carried out on any hydrology model output, including hydrologic models that do not use uniform grid elements (e.g., finite element methods). As part of the CESI funding, some effort will be made to upgrade the SESI models to use this more general hydrology. However, there are numerous scaling issues that will require careful thought in all the models. Some FY05 PES funding has been used to adapt the SESI models to the more general hydrologic output. As of this date, the new landscape classes have been developed which allow ATLSS to utilize general mesh hydrology model output, including RSM from SFWMD. The ATLSS SESI Deer model has been modified to allow the use of general mesh grids, with several other SESI models planned to be modified in a similar manner.

There will be continued Testing and Validation of ATLSS Models. At the time the ATLSS SESI models were developed, much of the parameterization and determination of model rules were decided based upon the expertise of those with long field experience in the system. This was done in part due to the very limited and spatially-sparse data available at that time. Since then, a wide variety of data sets for the above mentioned species have been elaborated, and some of these have been done in a spatially-explicit manner (particularly the SRF data sets). We propose to collect these new data in a format amenable to comparison to particular outputs of the SESI models, in order to evaluate and modify the models as needed. This involves placing the data in an appropriate geo-referenced format, and applying needed spatial averaging in order to utilize it in comparisons to model output. In particular, we wish to have the data readable in the ATLSS Dataviewer so that independent assessments of the model performance may be made by various researchers external to UT. These comparisons will be carried out for all SESI models for which data are available.

Improvements will be made in the delivery of models to client agencies. All ATLSS SESI models have been made available to ENP staff members, but use of these has been limited due to computer constraints (ATLSS models are developed to run on a Sun Microsystems platform rather than a Windows platform). A major task will be to develop methods to allow either agency representatives to readily and rapidly make model runs they desire on a remote system of Sun machines, or to move the codes to be platform independent. A NSF-funded project at the University of Tennessee is currently underway that has allowed dispersed resource managers to access remotely, through the Web, the capabilities of the SInRG (Scalable Intracampus Research Grid) at the University of Tennessee. This allows users at resource agencies in South Florida, with relatively little computer expertise, to initiate ATLSS simulations on the computers at the University of Tennessee. The Web-based use of the ATLSS models has been available throughout 2005, using software, called NetSolve, developed at the University of Tennessee. Representatives from U. S. Fish and Wildlife Service and the National Park Service have access to ATLSS model runs through the interface. Discussions are ongoing with various agency representatives to provide a more efficient mechanism for ATLSS model runs, particularly for revised hydrology scenarios which are not able to be developed using the current Model Interface. Two papers have been published in 2005 which describe the approach taken for the ATLSS Model interface - this effort has been supported by NSF.

Recent Products:

Comiskey, E. J., A. C. Eller, Jr., and D. W. Perkins. 2004. Evaluating impacts to Florida panther habitat: how porous is the umbrella. Southeastern Naturalist. 3(1):51-74.

Immanuel, A., M. W. Berry, L. J. Gross, M. Palmer, and D. Wang. 2004. A parallel implementation of ALFISH: simulating hydrological compartmentalization effects on fish dynamics in the Florida Everglades. Simulation Modelling Practice and Theory 13:55-76.

Duke-Sylvester, S. M. 2004. LygoMod : A model for the spread and optimal treatment of Lygodium microphyllum in the Arthur R. Marshall Loxahatchee National Wildlife Refuge; version 1.0.. Delivered to U.S. Fish and Wildlife Service staff at the A.R.M. Loxahatchee NWR, August 2004.

Duke-Sylvester, S. M. 2004. The ATLSS High Resolution Multi-Data Source Topography (HMDT).
http://www.atlss.org/~sylv/HTML/Everglades/HMDT-ShortReport/index.html

Duke-Sylvester, S. M. and E. J. Comiskey. 2004. Comparison of Calibration/Verification v5.0 to Calibration/Validation v3.4.
http://www.atlss.org/~sylv/HTML/Everglades/HydroComparison/index.html

Immanuel, A., M. W. Berry, L. J. Gross, M. Palmer, and D. Wang. 2005. A parallel implementation of ALFISH: simulating hydrological compartmentalization effects on fish dynamics in the Florida Everglades. Simulation Modelling Practice and Theory 13:55-76.

Wang, D., E. Carr, M. Palmer, M. W. Berry, and L. J. Gross. 2005. A Grid Service Module for Natural Resource Managers. IEEE Internet Computing 9:35-41.

Wang, D., E. Carr, L. J. Gross, and M. W. Berry. 2005. Toward ecosystem modeling on computing grids. Computing in Science and Engineering 7:44-52.

Specific Task Products:

Integration of the vegetation dynamics model into ATLSS SESI models.
New landscape class structure for models
Tests of one or more SESI models
Development of small fish SESI model with Joel Trexler and Bill Loftus
Perform simulation runs for CERP

Specific Relevance to Information Needs Identified in DOI's Science Plan in Support of Ecosystem Restoration, Preservation, and Protection in South Florida (DOI's Everglades Science Plan) [Page numbers listed below are from the DOI's Everglades Science Plan. See Plan on SOFIA's Web site: http://sofia.usgs.gov/publications/reports/doi-science-plan/]:

The ATLSS project links detailed biotic models with spatially-explicit abiotic data at regional extents, in order to provide a scientifically-defensible basis for regional planning that accounts for the complexity of biotic responses from individual-organism levels to that of communities (p. 79). ATLSS models were applied extensively in analyzing Restudy and Mod Water plans. ATLSS products are requested regularly by various agencies in South Florida including the South Florida Water Management District, U.S. Fish and Wildlife Service and Everglades National Park.

Key Findings:

1. Version 1.0 of Everglades vegetation succession model (VSMod 1.0)
2. Creation of ATLSS High-Resolution Multi-data source Topography (HMDT)
3. Publication of paper on Florida panther habitat: Comiskey, E. J., A. C. Eller, Jr., and D. W. Perkins. 2004. Evaluating impacts to Florida panther habitat: How porous is the umbrella? Southeastern Naturalist 3(1):51-74.
4. Publication of two papers on the development of ATLSS for computational grids.