Ann Tihansky; Don DeAngelis; Wiley Kitchens; Franklin Percival; Susan Walls; Dan Slone; Ann Foster; Eric Swain; David Sumner; Hal Davis; Nathaniel Plant
We propose to down-scale predictions from a suite of coupled Atmosphere-Ocean General Circulation Models (AOGCMs) to make regional scale predictions for the southeastern United States. For the time being the hydrologic and biologic models are confined to Florida. Climate outputs will then be used as inputs to a suite of species / habitat / ecosystem models that are currently being used in two key areas: the Greater Everglades and Suwannee River-Big Bend as a proof of concept that down-scaled climate results can work in ecological forecast models. We will run three scenarios of Land Use/Land Cover (LULC): past (circa 1900), present, and future (2041-2070). Additional climate model runs will address the contribution of green house gasses to climate variability and change over the Florida peninsula. Model perturbation experiments will be performed to address sources of variability and their contribution to the output regional climate change scenarios. We will develop scenarios that specifically address potential changes in temperature (land and near sea surface) and rainfall fields over the peninsula. We will then provide these scenarios and modeling results to resource management groups (NGOs, state and federal) via workshops in which the scenarios will be used to predict responses of additional selected species, habitats and ecosystems.
Our approach is to develop regional climate predictions and subsequent ecological predictions for two 30-year long time periods as well as for the present. The first 30-year period is the recent past, spanning the period from 1971-2000. This will be used as a control, with copious observations of both climate variables (e.g. rainfall, ET) and species (e.g. densities, ranges) to verify both climate and ecology model outputs and to serve as a baseline to systematically judge the impacts of an altered climate. The second 30-year time period will begin 30 years in the future and extend for the thirty years from 2041-2070. This is a time horizon that is immediately relevant to habitat management.
The Florida peninsula is long and narrow (700 x 150 km) and separates the warm waters of the Gulf of Mexico from those of the southeast Atlantic Ocean. This juxtaposition results in an incredibly variable, heterogeneous climate, with mean annual rainfall exceeding 150 cm in most parts of the state, but also includes frequent regional droughts. Peninsular Florida is unique in that it spans the temperate/sub-tropical boundary, and will be a key region to monitor as the freeze line shifts due to climate change. A comprehensive evaluation of the uncertainties in the regional climate projections for peninsular Florida is necessary to plan for future risk mitigation strategies and adapt to any changes from anthropogenic influence of climate. Such a comprehensive regional climate change study has not been conducted, and could be used for initiating policies and planning activities for conserving Florida’s biodiversity and resource management (Scott, 2008).
Florida has been highly impacted by land use and land cover change (LULC) since the turn of the century. Given such evidence from observations and previous modeling studies (Marshall, C. H., Jr, et al, 2004; Pielke, R.A. Et al, 1999; and Zhao, M. et al, 2001), there is motivation to understand the impact of LULC on regional climate projections in addition to changes that result of increasing greenhouse gas concentrations.
This project is designed to examine two areas of Florida that are particularly vulnerable to climate change: the Greater Everglades of south Florida and the Suwannee River / Big Bend region on the west coast. These two ecosystems provide an opportunity to compare the effects of climate change on a highly impacted/tropical system and a relatively unimpacted/subtropical system. The Florida Everglades have been heavily impacted by urban development, a large network of drainage canals that resulted in drastic changes to water flow, altered wetland plant communities, and spread of non-native plants and animals. The Suwannee River is the largest river in the eastern US with no dams or water control structures. It is a relatively pristine aquatic ecosystem with comparatively natural hydrology, aquatic plant composition, and relatively few non-native species. The system is also located near the freeze line and thus is likely to be impacted by small changes in temperature.
Vasu Misra, Co-PI, Ass’t Professor, Lead-Climate Sub-team; Eric Chassignet: Co-PI, Director/Professor; and Lydia Stefanova, Co-PI, Senior Research Associate, Center for Ocean-Atmospheric Prediction Studies (COAPS), Dept. Meteorology, Florida State University
Brad Stith, Jacobs Technology, Inc. Co-PI, individual-based modeling;
Christa Zweig, Florida Fish & Wildlife Co-op Unit, Co-PI, vegetation modeling;
Michael Allen, UF, Forest Resources & Conservation, Co-PI, fisheries ecology & modeling
Pielke, Roger A., Sr.; Steyaert, Louis T.; Willard, Debra A.
accessed as of 7/18/2011
Wolfert, Melinda A.; Bales, Jerad D.; Goodwin, Carl R.
Walko, R. L.; Steyaert, L. T.; Vidale, P. L.; Liston, G. E.; Lyons, W. A. Chase, T. N.
Pitman, A. J.; Chase, T. N.
This report is from the Climate Change Summit held October 1-3, 2008 in Orlando, FL.
Kanamatsu, M
Knamitsu, M.
Tomlinson, S. A.; Tillis, G.
Prepared in cooperation with the Suwannee River Water Management District
Collingham, Y. C.; Willis, S. G.; Gregory, R. D.; Smith, K. W.; Huntley, B.
Crumpacker, D. W.; Hardin, D. E.
The full article is available via journal subscription or single article purchase. The abstract may be viewed on the SpringerLink website
Decker, J. D.
The full article is available via journal subscription or single article purchase. The abstract may be viewed on the website below.
An ongoing international project, the North American Regional Climate Change Assessment Program (NARCCAP) is dynamically downscaling the Atmosphere-Ocean General Circulation Model (AOGCM) for the North American region. Our project will conduct a similar regional climate change study for peninsular Florida that will parameterize the hydrology, plant, and animal models.
We will use the Regional Spectral Model (RSM) as our regional climate model. It was developed at the National Centers for Environmental Prediction (NCEP; Juang and Kanamatsu, 1994) and is maintained at the Experimental Climate Prediction Center (ECPC) at UCSD (Kanamaru and Kanmitsu, 2007).
The proposed regional domain for downscaling is at 10 km horizontal resolution over the entire Florida peninsula and adjacent oceans. Using Scale Selective Bias Correction (SSBC), the sponge zone along the 4 sides (lateral boundaries) of the regional domain where the solution of the regional climate model is gradually nudged towards the coarser model solution are much narrower in the RSM compared to other regional climate models. Conventionally, the grid points within this sponge zone are ignored in any analysis of the regional model results. However in the case of the RSM, far fewer points near the lateral boundaries have to be ignored allowing for maximal usage of the of the grid points in the regional domain
We will downscale three of the Intergovernmental Panel on Climate Change 4th Assessemnt Report (IPCC-AR4) AOGCM’s: Community Climate System Model version 3.0 (CCSM3.0), Geophysical Fluid Dynamics Laboratory version 2.1 (GFDL2.1), Hadley Centre Coupled Model version 3 (HadCM3). These AOGCM’s are part of the NARCCAP and their outputs have been stored at 6 hourly intervals, which is essential for nesting in the regional climate model. AOGCM outputs beginning from 2001 are not yet available. As useable outputs become available we will attempt to incorporate them. For the present, we will use data available through 2000. The RSM will be nested into these three AOGCM’s for the current period (1971-2000) and for the future period (2041-2070) forced with the IPCC Special Report on Emissions Scenarios A2 scenario for the 21st century (IPCC). For each downscaled model integration we will conduct pre-industrial (Goldewijk, 2001), present (ISLSCP-II) and projected land cover (UCAR) scenarios. This combination of 2 30yr periods, by 3 LULCs, by 3 models provides a total of 540 years of modeled climate output. For all these model integrations we propose to use the Sea Surface Temperature from the corresponding AOGCM linearly interpolated to the RSM grid.
Task 2: Hydrological & Ecological modeling and forecasting
Several approaches will be used in applying the down-scaled climate predictions to forecasts of potential ecological effects for species, habitats and ecosystems of concern to resource managers. These will include all of the models described above, climate envelope models and Bayesian statistical methods for handling uncertainty. The Across Trophic Level System Simulation (ATLSS), Spatially-Explicit Species Index (SESI), and Individually Based Models (IBMs) for American alligator, snail kite, American crocodile, and the Florida manatee and functional groups in southern Florida as well as the new Suwannee models all use vegetation succession model and hydrological inputs. Hydrologic inputs are provided by landscape hydrology models that work on a relevant scale of spatial resolution (e.g., 500 x 500 m or similar scale pixels based on species life history). These inputs include at least daily water depths, and sometimes, additionally, salinity and temperature, in a pixel. These hydrologic data for the ecological models will be provided by the Tides and Inflows to the Mangrove Ecotone (TIME) model (Swain, 2005; Swain and Decker, 2008; Swain et al, 2003) and the Lower Suwannee River hydrologic model (Bales et al, 2006). TIME will itself be driven by 10 x 10 km precipitation and temperature data over a 30-year time frame. The Suwannee River model will be updated to reflect current and future conditions.
Vegetation maps will be provided by the ATLSS vegetation succession model, VSMod, which uses the Florida Gap Analysis Program (GAP) model (FGAP v6.6) as the vegetation map input. This map uses a 30x30 meter grid. It contains 78 habitat types. Of these, approximately 58 represent natural vegetation communities. From these 58 types the ATLSS Vegetation Succession Model (VSMod) uses 24 that are relevant to southern Florida. Vegetation is likely to change both as a result of CERP and of projected climate changes. We will attempt to project those changes using VSMod. However, the expected uncertainty in such changes is high, so one option will be to use the most recent data on vegetation types available. In the Suwannee, monitoring data from vegetation transects along north/south and east/west gradients will be used to develop a vegetation succession model. Remote sensing will also be used to extend the period of record for the model and to link hydrologic and vegetation relationships at multiple scales. Output from both TIME and VSMod (either current vegetation data or best possible projections of future vegetation) and the Suwannee models will then be used as input for three types of models:
Habitat suitability models: These include the ATLSS SESI models for Cape Sable seaside sparrow, short and long legged wading birds, the white tailed deer, the American alligator, and Florida panther. Also included will be the "sea grass distribution" model. Similar habitat/species index models will be developed for the Suwannee including amphibians (based on an existing model developed for the SW Florida Feasibility Study in the Everglades), rodents, and raptors. These data will be collected at present conditions to develop relationships for future habitat changes. To the extent possible, the Everglades models will be reparameterized to the Suwannee basin for comparison to the southern modeling domain.
Stage-based & individual-based population models: We will use the climate change scenarios to explore changes to the American alligator population in the Everglades using the ATLSS Alligator Population Model. This model will be updated and used for the Suwannee basin as well. We will also develop a new fish model with data from current monitoring and fish community data from Fish and Wildlife Commission (FWC) from the last 12 years. We will analyze the fish and amphibian data and link them with the hydrology model to the habitat metrics (plant types) and fish and amphibian community metrics. Outputs from the model will include indicator species that support important fisheries (e.g., red drum, largemouth bass), vulnerable species (e.g., Gulf sturgeon), and nongame species (river frog) of potential value for assessing and monitoring long-term change associated with diminished freshwater flows and increased salinity and/or sea-level rise. Individual-based models will be similarly developed using climate inputs. Species to be included are: snail kite, American crocodile, and the manatee.
Climate envelope models (CEMs) for vegetation: Subsets of important vegetation types will be analyzed using climate envelope models (Green et al, 2008). These Include cypress, several pines, oaks, tropical hardwoods and mangroves. CEMs have been developed for some of these groups in the past, but at a scale 100x larger than our proposed work (Box et al, 1999). Our models should be a significant improvement for use by resource managers. Additionally we will examine the climate response of invasive plant species such as Brazilian pepper, Melaleuca, Australian pine, and climbing fern. Output from the GCM down-scaling will serve as input to the envelope models.
Lastly, amphibian and reptile population data from the USGS Amphibian Research and Monitoring Initiative (ARMI) network sites in Florida will also be analyzed via climate envelope modeling techniques.
Ecological and hydrological models also have uncertainties due to methods of parameter estimations. We can estimate uncertainties by several methods. Within a climate-species scenario uncertainty can be estimated using a brute-force sensitivity analyses, where each input parameter is varied and the contribution to total output variance due to each parameter is calculated. Likewise uncertainty between different climate scenarios for each species can be estimated by comparing grids of habitat distributions derived from each scenario. The grid for each scenario will have a mean or best estimate of habitat suitability values, and the difference between each scenario and a reference scenario will provide an index of the uncertainty. Finally, we can add an estimate of model fidelity through data analysis and model-data comparison. The data analysis will capture the internal system variability that is not resolved by the models and the model-data comparison will identify the magnitude of model errors.
Task 3: Information Transfer and Data Dissemination Workshop
An initial training workshop will be held during the first year of the project.
To inform resource managers and cooperators about progress for the year, we will conduct yearly technology transfer workshops. These will be informal workshops designed to encourage questions and answers between the PI’s and resource managers. This annual 1-day workshop will be held toward the end of each fiscal year in a central location.
We propose to hold a technical workshop as part of an international scientific meeting, specifically the Coastal and Estuarine Research Federation International Biennial Conference in 2011 in Daytona Beach, FL.
U.S. Department of the Interior, U.S. Geological Survey
Comments and suggestions? Contact: Heather
Henkel - Webmaster
Generated by mp version 2.8.18 on Tue Jul 19 12:54:13 2011