Scientists at the Fish and Wildlife Research Institute (FWRI) have been developing habitat suitability models that predict spatial distributions and relative abundance of fish and invertebrate species in Florida estuaries.
Habitat Suitability Index (HSI) modeling was developed by the U.S. Fish and Wildlife Service (USFWS) in the early 1980s as part of the Habitat Evaluation Program. The models were used to support rapid decision-making in data-poor situations. Expert-opinion and literature sources were used to develop suitability indices (SI) indicative of habitat preferences across gradients. These indices were then combined to produce the index. The geometric-mean algorithm is most commonly used to determine the index.
In 1997, the Strategic Environmental Assessment Division of the National Oceanographic and Atmospheric Administration (NOAA) worked with USFWS to develop HSI models with geographic information systems (GIS) using qualitative methods to predict spatial distributions of estuarine species in Maine's Casco and Sheepscot Bays. Similar methods were used to predict the spatial distributions of American oyster, white shrimp, and spotted seatrout in Pensacola Bay, Florida.
In 1998, FWRI initiated studies to develop quantitative HSI models in collaboration with NOAA and the University of Miami (Rubec et al. 1998). FWRI's Fisheries-Independent Monitoring (FIM) data were analyzed by species life stages and season in Tampa Bay and in Charlotte Harbor (Rubec et al. 1999, 2001). Modeling was conducted initially with spotted seatrout (Cynoscion nebulosus), pinfish (Lagodon rhomboides), and bay anchovy (Anchoa mitchilli).
Methods were developed to average environmental data from FIM sampling and other agencies within sampling grids across each estuary on a monthly basis (Figure 1). Aerial photography is used to determine the distribution of submerged aquatic vegetation (SAV) for the creation of maps of bottom type. Soundings data obtained from NOAA were interpolated to create bathymetry maps. Data points for temperature, salinity, and dissolved oxygen were interpolated to produce monthly surface and bottom habitat layers. Then, monthly habitat layers were averaged using the ArcView GIS Spatial Analyst extension to create seasonal maps for each habitat type (Figure 2).
Habitat preferences for each species life stage were determined by fitting polynomial regressions to mean CPUEs across environmental gradients (Figure 3). Higher mean CPUEs indicate that the species is more abundant in parts of the gradient. For example, a peak in the suitability curve indicates that early juvenile spotted seatrout are most abundant at 28-32°C (Figure 3). Hence, the life stage has a habitat affinity for the peak in the suitability curve. Similar suitability curves for Tampa Bay and Charlotte Harbor indicate similar habitat preferences for the life stages of each species.
Mean suitability values from the curves (Figure 4) are used as input to habitat suitability models. The HSI model was used to calculate the geometric mean of the SIs associated with each habitat layer across grid cells to create a predicted HSI map in each estuary (Figure 5). The HSI model was used to create seasonal maps (spring, summer, fall, winter) depicting the spatial distribution of each species life stage.
Raw CPUE data were overlaid within four zones (Low, Moderate, High, and Optimum) of the predicted map (Figure 6) and mean CPUEs calculated to create a histogram (Figure 7). The model was verified when mean CPUEs were found to increase across the zones. Hence, the Optimal zone should have the highest mean CPUE.
Suitability indices from a nearby estuary were also used with the habitat layers from the first estuary to test transferability of the model. For example, Charlotte Harbor SIs were transferred to Tampa Bay to create a second map. Statistics were used to compare the similarity of each pair of seasonal maps within each estuary.
Figures 6 and 7
Suitability indices transferred from another estuary can be used to infer species distributions and relative abundance of species in estuaries lacking fisheries monitoring. For example, data from Tampa Bay has been applied to predict the distribution of juvenile pinfish in Charlotte Harbor (Figure 8).
The spatial modeling can be used to define which habitats are most important for each species. The Optimum zones have the potential of being designated Habitat Areas of Particular Concern (HAPC) associated with Essential Fish Habitat (Rubec et al. 1998). The approach can assist decision-making associated with habitat protection and fisheries management. This research was funded by USFWS through the Sport Fish Restoration Program.
Rubec, P. J., J.C.W. Bexley, H. Norris, M.S. Coyne, M.E. Monaco, S.G. Smith, and J.S. Ault. 1999. Suitability modeling to delineate habitat essential to sustainable fisheries. Pages 108-133, In: L.R. Benaka (ed.). Fish Habitat: Essential Fish Habitat and Restoration, American Fisheries Society Symposium 22.
Rubec, P.J., M.S. Coyne, R.H. McMichael, Jr., and M.E. Monaco. 1998. Spatial methods being developed in Florida to determine essential fish habitat. Fisheries 23(7):21-25.
Rubec, P.J., S.G. Smith, M.S. Coyne, M. White, A. Sullivan, D. Wilder, T. MacDonald, R.H. McMichael, Jr., M.E. Monaco, and J.S. Ault. 2001. Spatial modeling of fish habitat suitability in Florida estuaries. Pages 1-18, In: GH. Kruse, N. Bez, A. Booth, M.W. Dorn, S. Hills, R.N. Lipcus, D. Pelltier, C. Roy, S.J. Smith, and D. Witherell (eds.) Spatial Processes and Management of Marine Populations, Alaska Sea Grant College Program, Fairbanks Alaska, AG-SG-01-02.