MEPS 265:197-212 (2003)  -  doi:10.3354/meps265197

Development and evaluation of statistical habitat suitability models: an example based on juvenile spotted seatrout Cynoscion nebulosus

Sven Kupschus*

Fish and Wildlife Conservation Commission, Florida Marine Research Institute, 100 8th Avenue SE, St. Petersburg, Florida 33701, USA *Present address: The Centre for Environment, Fisheries and Aquaculture Science, Lowestoft Laboratory, Pakefield Road, Lowestoft, Suffolk NR33 0HT, UK

ABSTRACT: Conservation and fisheries managers require models that describe the abundance and distribution of fishes in order to protect and manage exploited fish stocks in the face of anthropogenically induced habitat loss and exploitation. In this study, models describing environmental preferences were developed for juvenile spotted seatrout Cynoscion nebulosus within 3 Florida estuaries: the Indian River Lagoon, Tampa Bay, and Charlotte Harbor. A generalized additive model (GAM) was developed to describe the environmental preferences, spatial distribution, and temporal fluctuations of juvenile seatrout in each estuary based on a 4 yr time-series of fisheries-independent catches. All 3 models indicated similar environmental preferences for all populations examined and were also qualitatively consistent with the findings from other studies. Consequently, habitat-preference models based on GAMs are useful tools to predict fish abundances in estuaries lacking fisheries-independent data, given knowledge of the local environmental conditions. These initial findings were further supported by quantitative analyses of the models¹ abilities to predict abundance in independent datasets, despite complications with multicollinearity of independent variables and temporal differences in the recruitment periodicity between the 3 Florida populations. Estimates of environmentally based habitat value on a relative scale will aid conservation managers in protecting vital nursery habitats in estuaries currently lacking fisheries-independent information and in predicting the effects of future environmental change. Finally, assuming similar environmental conditions between years, the year effect from these models can serve as an index of relative abundance for fisheries managers to predict future recruitment strength and to tune catch-at-age stock-assessment models.


KEY WORDS: Cynoscion nebulosus · Fish distributions · Habitat suitability modelling · Generalized additive models (GAM)


Full text in pdf format