Shrimp and redfish studies; Bryan Mound brine disposal sited off Freeport, Texas, 1979-1981. Volume I(B): Texas coast shrimp catch and effort data.

The analyses reported herein address the potential impacts to the Texas shrimp fishery from offshore disposal of brine associated with the U.S. Department of Energy Strategic Petroleum Reserve Program at the Bryan Mound storage site, through the analysis of the historical data base (Gulf Coast Shrimp Data). Spatial and temporal patterns in indicators of shrimping activity have been determined through the use of analysis of variance, multiple regression analysis and orthogonal polynomial analysis. In addition, predictive multiple regression models have been developed to test the null hypothesis that no significant differences in catch (or catch/area) and catch/effort (or catch/effort/area) are attributable to brine discharge. Multiple regression analysis, relating catch/non-directed nominal effort (C/F') to vessel characteristics and utilizing dummy variables to account for spatial and temporal variations in abundance was conducted for the data for 1965 to 1977. Results indicated that vessel characteristics accounted for only about four percent of the variation in C/F' (log-log model), with spatial and temporal patterns in abundance accounting for another thirty percent of the variance. A large amount of the variance in C/F' remained unexplained and is probably due to within-cell heterogeneity in abundance and errors in the reporting process. Based on the small amount of variance in C/F' explained by vessel characteristics, effort was not standardized, and non-standardized effort was used in subsequent ANOVA and multiple regression analyses. Five way, factorial design fixed effects ANOVA's, using the five way interaction mean square as the error term, were conducted for catch (C), catch/area (C/A), catch/directed nominal effort (C/F), and catch/directed nominal effort/area (C/F/A) for brown shrimp, white shrimp and both species combined. The ANOVA model included eighteen years (1960 to 1977), twelve months (January to December), four statistical areas (18 to 21), three depth intervals (0-10, 11-20, and 21-50 fathoms), and three size classes (20 count or less, 21-30 count and greater than 30 count). The results of these analyses are displayed in means tables, histograms of main effects and print grid density displays of important two and three factor interactions. For each taxa group, the results of the analyses for C and C/A were very similar, as were the results for C/F and C/F/A. Brown shrimp results were quite different (as compared to the results for catch) when effort was incorporated into the dependent variable. Major trends in the results are discussed in light of our knowledge of the penaeid shrimp life cycle. Orthogonal polynomial models were developed for main effects and two and three factor interactions from the ANOVA's. In general, linear, quadratic and cubic components of the orthogonal polynomial models were inadequate to explain all significant trends in the data, with especially poor results for year and month. The results for C for both species showed many more significant orthogonal polynomial components than did the results for C/A, C/F,and C/F/A. Stepwise multiple regression analysis was used to provide mathematical predictors of C and C/F' for brown shrimp and for white shrimp. Equations were developed for each species for both interview and total data. For both interview and total data, three models were developed for each species relating C in statistical area 19, 11 to 15 fathoms depths to catch variables, effort variables, and catch and effort variables. For C/F', one equation was developed for each species for both interview and total data relating C/F' in statistical area 19, 11 to 15 fathom depths to C/F' variables in adjacent spatial strata and statistical areas. In general, the models for interview data explained more variance with fewer independent variables than did the models based on total data, and more variance was explained for brown shrimp variables than for white shrimp variables. White shrimp C appeared to be more closely related to bay catch and trips than did brown shrimp C. In all cases good predictors are developed, and their use in assessing post-discharge impacts is discussed.