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



275 



the Galveston Bay model (with interaction term) to brown 

 shrimp density data from Aransas, Matagorda, and San 

 Antonio bays. The results indicated similar habitat-use 

 patterns in Aransas and San Antonio bays; there were 

 higher densities in high-salinity seagrass beds and a de- 

 clining density as salinity decreased in these bay systems. 

 No SAV samples were taken in Matagorda Bay; however, 

 the model performed well in predicting greater brown 

 shrimp density in higher-salinity marsh-edge habitats. 

 Our analysis suggests that although the empirical model 

 is complex, it is general enough to be applicable across a 

 broader range of habitat types. The model results may, 

 however, have some geographic limitations. For instance, 

 the model may not perform well within the Laguna Madre 

 in south Texas, where freshwater inflow is diminished and 

 hypersaline conditions exist. This conclusion is consistent 

 with Rubec et al. (1999), who used similar methods to 

 demonstrate that HSI models are applicable across estuar- 

 ies in central Florida. Our results are promising in view 

 of previous efforts where predictions of nekton abundance 

 with empirical models have proven difficult. 



Currently, estuarine EFH for most federally managed 

 species in the Gulf of Mexico exists as mapped estimates 

 of relative abundance from NOS's estuarine living marine 

 resources (ELMR) database (GMFMC, 1998; Nelson and 

 Monaco, 2000). The entire Galveston Bay complex was 

 considered EFH for brown shrimp based on ELMR relative 

 abundance data. Our model, generated by using brown 

 shrimp density data, provides a more spatially resolved 

 delineation of EFH (in waters <1 m depth) for brown 

 shrimp <100 mm. 



The analyses described in the present study focused 

 on bottom types in waters less than 1 m which comprise 

 about 25% of the available habitat in Galveston Bay. 

 Trawl CPUE data from Texas Parks and Wildlife De- 

 partment (TPWD) were analyzed to compare abundance 

 and distribution patterns in waters >1 m. These trawls 

 (3.8-cm stretched mesh) do not capture small size classes 

 (<50 mm TL) of brown shrimp efficiently; thus the trawl 

 analysis provides information only on larger size classes 

 (mean=89 mm). However, few individuals in smaller size 

 classes of shrimp (<50 mm TL) are likely to inhabit deeper 

 bay waters; density estimates of small nekton, including 

 shrimp, decline rapidly with depth (Mock, 1966; Baltz et 

 al., 1993; Rozas, 1993; Rozas and Zimmerman, 2000). In 

 addition, these CPUE values are likely underestimates of 

 brown shrimp density; catch efficiency for shrimp in trawls 

 can be roughly estimated at 20 f /f (Zimmerman et al., 1984; 

 Rozas and Minello, 1997). Despite these problems, shrimp 

 abundance estimates in water >1 m appear low; abun- 

 dance estimates from TPWD trawl data in deep open-bay 

 waters were almost two orders of magnitude lower than 

 densities in shallow water habitats. 



Brown shrimp population estimates from the present 

 study (Table 3) were highest in the lower bay (224,568 per 

 ha.). Seagrass beds accounted for more than 607i of the es- 

 timate ( 145,142 per ha.) and marsh edge and nonvegetated 

 bottom types combined were estimated at approximately 

 79,000 per ha. As noted earlier, the NWI regularly flooded 

 emergent vegetation classification is not all marsh edge but 



is a complex of SNB, marsh edge, and inner marsh with 

 different shrimp densities associated with each of these 

 microhabitat types. Minello and Rozas (in press) modeled 

 small-scale density patterns on the marsh surface in a 

 437-ha. salt marsh of lower Galveston Bay and applied 

 these data to a GIS analysis of marsh landscape patterns. 

 In this highly fragmented marsh complex that was 37% 

 SNB and 63% marsh vegetation, they estimated brown 

 shrimp populations at 37,000 per ha. We could not estimate 

 brown shrimp populations in irregularly flooded emergent 

 vegetation, although the areal coverage of this habitat type 

 was large. Compared with the regularly flooded wetlands, 

 overall densities of brown shrimp in these irregularly 

 flooded systems should be relatively low because of higher 

 marsh surface elevations (Rozas and Reed, 1993; Minello 

 et al., 1994; Minello and Webb, 1997) and restricted tidal 

 access (Rozas and Minello, 1999). We also were unable to 

 assess the contribution of oyster reef as habitat for brown 

 shrimp. Coen et al. ( 1999), however, reported brown shrimp 

 on oyster reefs, and Powell ( 1993 ) estimated that there was 

 108 km 2 of this habitat in Galveston Bay. 



Our modeling results provide evidence that estuarine 

 habitat types are discriminately used by brown shrimp. 

 The success of transferring our empirical model from 

 Galveston Bay to adjacent bay systems in Texas suggests 

 that the model has a broad application and can possibly 

 be used to simulate patterns of habitat use in systems 

 that lack sufficient density data. Continuing collections 

 of density data in Gulf estuaries are necessary to make 

 additional interestuary comparisons and to determine 

 whether these habitat-use patterns differ throughout the 

 distributional range of brown shrimp. The use of other 

 habitat types also needs to be examined. For example, 

 other available habitat types from Galveston Bay, such as 

 oyster reef and inner marsh, and from other Gulf estuar- 

 ies, such as mangrove, calcium carbonate rock formations, 

 and sponge communities, may be important habitats for 

 this federally managed species. 



Acknowledgments 



Funding and support for this work was provided by the 

 Southeast Region of NOAA's National Marine Fisheries 

 Service, The Southeast Fisheries Science Center, and the 

 Biogeography Program of the National Ocean Service. We 

 would like to thank Pete Sheridan, Lawrence Rozas, Ken 

 Heck, and Roger Zimmerman for providing access to pub- 

 lished and unpublished data sets. John Boyd helped with 

 construction of the nekton density database. 



Literature cited 



Baltz, D. M., J. W. Fleeger, C.F. Rakocinski, and J. N. McCall. 

 1998. Food, density, and microhabitat: factors affecting 

 growth and recruitment potential of juvenile saltmarsh 

 fishes. Environ. Biol. Fish. 53:89-103. 

 Baltz, D. M., C. Rakocinski, and J. W. Fleeger. 



1993. Microhabitat use by marsh-edge fishes in a Louisiana 

 estuary. Environ. Biolog. Fish. 36:109-126. 



