Application of Generalized Additive Models to Assess Spatial and Temporal Shifts in Suitable Habitat of Juvenile Southern Flounder

Suitable habitats of juvenile southern flounder (Paralichthys lethostigma) within the Galveston Bay Complex (GBC), Texas, were assessed using Generalized Additive Models (GAM). Fishery independent data collected with bag seines throughout the GBC from 2000-2009 were used for GAM construction. In addition to flounder presence or absence, models also included numerous environmental variables (e.g. temperature, salinity, substrate, and depth). These variables, along with spatially dependent variables such as distance to habitat or hydrologic features generated within a Geographic Information System (GIS), were used to predict the probability of southern flounder presence or absence. Separate models were generated for newly settled and older young of year (YOY) flounder based on size. In addition, 2 seasonal models (summer and fall) were generated for YOY flounder. After model completion, the availability and spatial distribution of suitable habitat within the GBC was generated within a GIS using available environmental and spatial data.  Factors affecting flounder presence changed seasonally, as did the corresponding shifts in the availability of suitable habitat. Newly settled and YOY founder distributions were influenced by temperature and salinity, and habitat associations were stronger for YOY flounder.  Spatial distributions of suitable habitat stress the relative importance of East Bay within the GBC throughout the first year of life. These models demonstrate the potential dynamics of suitable habitats for juvenile southern flounder and provide insight into ontogenetic shifts in habitat preference during the first year of life.