Brad Stith, Eric Swain, James Reid, Daniel Slone
U.S. Department of Agriculture - Natural Resources Conservation Service (NRCS) Department of the Interior - U.S. Geological Survey Department of Commerce - National Oceanic and Atmospheric Administration (NOAA) Environmental Protection Agency (EPA) Smithsonian Institution - National Museum of Natural History (NMNH)
2201 NW 40th Terrace
1. Linking an individual-based manatee model with TIME and a new TTI hydrology model to assess CERP restoration effects on the Everglades and Ten Thousand Islands estuaries
USGS hydrologists and manatee researchers will collaborate on various tasks to provide an integrated modeling effort. USGS hydrologists have developed a coupled surface- and ground-water hydrology application (TIME), using the FTLOADDS model code (Swain and others, 2004), which can generate salinity and water temperature data for the major river systems, bays, and near-shore gulf region within ENP. The primary hydrology task for FY06 is to develop an application of FTLOADDS for the TTI region, which in combination with the existing TIME application will enable salinity and water temperature data to be generated across most of southwest coast below Naples. The TTI region is west and mostly north of the existing TIME area. The northern boundary of the TTI model will be defined with measurable surface-water control locations such as canals and waterways, and the off-shore boundary with tidal information. The aquifer characteristics will be estimated for the ground-water model, and bathymetry for the surface-water model will be developed from existing data. The smaller-scale bathymetric features, such as channels, will be represented by several modifications to the code. Current modifications allow for model grid-face frictional resistance to be modified to represent smaller features. Further modifications will be made to represent the storage and dimensions of such features. The FTLOADDS code incorporates salinity transport in both the ground-water and surface-water, and is designed to simulate heat transport, thus enabling the modeling of salinity and water temperature. Boundary conditions must be defined for all parameters modeled. Water monitoring stations that log salinity and water temperature are distributed across much of the region and will be used for model calibration, as well as available point measurements of these parameters. Data for areal variations in temperature are available from various sources, and the necessary solar radiation data exists at several locations. Once properly calibrated, the applications will be used to generate a time series of salinity and water temperature values at a specific number of nodes across a habitat network used by manatees throughout the TTI/ENP region. This time series will be averaged from the short timestep used in the hydrologic model to a 6-hour time step during a period for which manatee telemetry data are available. These hydrologic data will be provided to the manatee model, which will produce manatee movement patterns and distributions that can be compared to actual manatee data for the same period.
USGS manatee researchers will process newly acquired telemetry data and complete various analyses needed to parameterize the individual-based manatee model. This includes data from manatees recently tagged in Whitewater Bay, as well as from 32 manatees tagged in TTI between June 2000 and June 2005, and miscellaneous tagged manatees that used the area (e.g. rescued animals). We will analyze the telemetry data using several approaches reflecting the hierarchical structure of the individual-based model. To capture the variability of individual manatee behavior, we will analyze manatee movement in a 3-level hierarchy. At the broadest scale, we will analyze migratory behavior of manatees in the study area in response to major cold fronts. Our approach will follow Deutsch et al. (2003), who identified several discrete categories of migratory behavior of manatees tagged on the east coast of Florida, ranging from long distance migrants to year-round residents. At the intermediate scale, we will develop seasonal home ranges for each tagged animal for the dry, wet, and cold seasons using fixed kernel analysis with least-square cross validation. Resulting home ranges will provide measures of variability among individuals in home range characteristics, such as seasonal home range size and distance from critical resources. Accounting for this variability will be important, since substantial heterogeneity in home ranges has been observed in tagged manatees on the east coast of Florida (Deutsch et al., 2003). The observed distribution of individual home ranges will be used to parameterize the home range allocation module of the individual-based model. At the finest scale, we will analyze movements between different habitat zones during the three primary seasons (dry, wet, cold) using multi-state modeling. This new statistical approach implemented in programs MARK and SURGE (Williams et al. 2002) will generate transition probabilities for movement between several broad habitat zones (e.g. offshore, bay, river) for each individual at a 6-hour time interval (reflecting the sampling interval of the Argos telemetry tags). The resulting distribution of Markovian probabilities will be used to parameterize the individual-based model, providing a useful technique for quantifying individual heterogeneity in movement behavior observed in tagged manatees as they make regular movements between offshore foraging zones and inshore zones with freshwater or thermal refugia (See Stith et al. 2004 and Reid et al. 2003 for further details on the observed movement patterns).
The manatee model was written in C++, allowing for the development of a flexible interface to read hydrologic output from the FTLOADDS model. The binary output from FTLOADDS will be processed to extract the data for the specific nodes representing destination sites for feeding, drinking, and sheltering from cold, with connections representing travel corridors. This network data structure will allow directed movements to be simulated in an efficient manner using well-known graph theory algorithms. The manatee model will increment through the seasonally changing hydrologic data, and manatees will respond to the availability of freshwater or warm water as they move across the network, using the parameters developed from the telemetry data. Large-scale migratory movements to winter home ranges are triggered by offshore water temperatures falling below 20 degrees Celsius on the network. Within a seasonal home range, transition probabilities obtained from the multi-state analysis are used to simulate the movement of manatees among ecological zones as a Markov chain process. To calibrate the model, simulations are run where each manatee is initially assigned a randomly located home range. As they move around the network seeking different resources, manatees are exposed to the simulated hydrologic conditions and experience positive or negative reinforcement (e.g. while transitioning up a river to find freshwater). Initially, manatees have no preferences for different parts of the network within their home range, but as they explore the network these preferences change based on a simple reinforcement model (see Sutton and Barto, 1998). During this learning phase, individuals shift their home ranges in response to positive and negative reinforcement. Once all individuals converge on stable home ranges, snapshots of the aggregate distribution of individuals will be generated and compared to aerial survey data collected for the same time period under similar hydrologic conditions. The aerial survey data will be subdivided, with one subset used to calibrate the model, and a holdout set to validate the model. Model calibration will primarily involve modifying the 2 or 3 parameters of the reinforcement model, which control the tradeoff between exploring less rewarding sites and maintaining site fidelity.
2. Hydrologic modeling and manatee winter use patterns at passive thermal refuges
The focus will be on field work to collect data at Port of the Islands in support of parameterization of the three-dimensional model, including field measurements of salinity and temperature, along with high-resolution bathymetry of the basin to be simulated. Measurements of these relevant parameters at some locations have already been made by the USGS in cooperation with the South Florida Water Management District. These measurements are not at the specific manatee habitat locations, but are necessary for model boundaries and calibration. In order to get data at the habitat locations, boat-based measurements and data-logging probes will be used to obtain vertical profiles of salinity and temperature at strategic locations. Bathymetric measurements will be made manually with GPS reference from a boat, or by acoustic Doppler meter, providing crucial discharge information as well. A continuously recording salinity and temperature gage will be installed to supplement existing continuous measurements in the area. A series of data logging temperature probes will be deployed at Port of the Islands basin, Wooten's basin, and Big Cypress Preserve Oasis Ranger Station canals. These hourly records of temperatures at different depths, combined with similar data collected over the previous two winters, will complement other hydrologic measurements and provide a better understanding of annual variations.
We plan to further characterize the manatee use of these passive thermal refuges and correlate with hydrologic findings at each site. Manatees have been tracked in the TTI as part of a study on manatee use patterns and freshwater flows within the region, which primarily focused on movements outside the winter season. Also, a smaller number of manatees initially tagged in the southern Everglades continue to be tracked in the TTI. Additional information collected during ground and aerial surveys at specific aggregation sites provides data on numbers of manatees present before, during, and after major cold fronts. Collectively, tracking and survey data will help identify manatee movement patterns associated with winter cold fronts. These findings, integrated in manatee movement models, will be coupled with hydrologic models on basin temperature and salinity to better understand the dynamics of these sites as passive thermal refuges.
2201 NW 40th Terrace
U.S. Department of the Interior, U.S. Geological Survey, Center for
Comments and suggestions? Contact: Heather Henkel - Webmaster
Generated by mp version 2.8.18 on Wed Jan 31 16:24:56 2007