Clark et al.: A habitat-use model for juvenile Farfantepenaeus aztecus in Galveston Bay 



273 



1.2 

 1.0 



0.8 

 0.6 

 0.4 

 02 

 0.0 

 -0.2 



1.0 

 0.8 

 0.6 

 0.4 

 0.2 

 0.0 

 -0.2 



0.2 0.4 0.6 0.8 1.0 1.2 



0.2 0.4 0.6 0.8 1.0 1.2 1.4 



Observed log density 



ME + SAV x SNB 



Figure 9 



Relationships between observed densities of brown shrimp iF. aztecus) in 

 Aransas, Matagorda, and San Antonio Bays and predicted densities from 

 the Galveston Bay model. Relationships for all bays combined are shown 

 in the upper left graph. For each relationship, the r 2 is shown for the least 

 squares regression, and the number of observations (n ) and the total number 

 of samples in parentheses. ME = marsh edge; SAV = submerged aquatic 

 vegetation; SNB = shallow nonvegetated bottom. 



(USFWS, 1981). Recently, Christensen et al. 1 and Brown 

 et al., 2000, developed suitability indices, based on lit- 

 erature reviews and expert opinion, and raster-based GIS 

 models that produce a spatial view of relative suitability. 

 The Florida Fish and Wildlife Conservation Commis- 

 sion-Marine Research Institute (FMRI) and the National 

 Ocean Service's Center for Coastal Monitoring and As- 

 sessment (NOS/CCMA) collaborated to develop a suite of 

 quantitative HSI modeling approaches, using fisheries- 

 independent monitoring catch-per-unit-of-effort (CPUE) 

 data (Rubec et al., 1998, 1999, 2001). These studies used 

 an unweighted geometric mean formula as part of the HSI 

 models to assess overall suitability. This approach assigns 

 equal weight to all factors by using scaled suitability indi- 

 ces as inputs to the model. The regression approach used 

 in this study more appropriately weights density according 

 to the factors in the model and allows a more robust tech- 

 nique to elucidate spatial patterns of habitat use by using 

 actual CPUE data. In addition, the method described 

 in our study can support more complex analyses, such 

 as interaction effects or trophic relationships (or both). 



Our ANOVA (Table 1) revealed that season, bottom type, 

 salinity, and the interaction between salinity and bottom 

 type are significant factors that influence the distribution 

 of juvenile brown shrimp in Galveston Bay. The addition of 

 the interaction effect to the model increases the coefficient 

 of determination from 0.63 to 0.73. Without this term in 

 the model, predicted values for brown shrimp density are 

 overestimated compared to the observed density data. 

 Seagrass beds in salinities greater than 15 ppt supported 

 significantly greater densities of brown shrimp than did 

 marsh edge. However, in locations with salinities less 

 than 15 ppt, brown shrimp densities were not significantly 

 different between the two bottom types. These results in- 

 dicate significantly lower use among all the bottom types 

 analyzed in the fresher portion of the estuary. It is likely 

 that salinity and a combination of other environmental 

 factors directly or indirectly (or directly and indirectly) 

 affect abundance on bottom types and habitat quality in 

 this region. The results indicate that SAV supports greater 

 brown shrimp density than do ME and SNB; however, SAV 

 accounts for less than 1% of the total bottom type within 



