Sheridan Forecasting the fishery for Penaeus duorarum 



749 



10 100 1000 10000 



September L-67 discharge (m 3 x 100,000) 



Figure 4 



Scatterplots of surface water variables against subsequent 

 fishing year landings, 1966-94. 



•-• •' 





T 



~r~ 



~r 



45 50 55 60 65 



August well P38 level (cm) 



-. - - 1 '■ 



~t 



40 







45 



T " 

 SO 



~~1 



60 



September well P37 level (cm) 



Figure 5 



Scatterplots of well water levels against subsequent fish- 

 ing year landings, 1966-94. 



year's model had become "the" forecast model (Table 

 3). The results indicate that a model that success- 

 fully forecasts for the year it was developed does not 

 always function well in succeeding years (e.g. mod- 

 els 1, 3, and 9) and that initially poor models can 

 later become successful (e.g. models 5 and 6). Gener- 

 ating new forecast models each year from updated 

 databases appears to be more accurate than using a 

 fixed forecast model. 



Discussion 



These models were quite accurate at forecasting 

 whether landings would increase or decrease and 

 only moderately accurate at forecasting amounts 

 landed, but they did not address cause-effect rela- 

 tionships between environmental variables and pink 

 shrimp recruitment. The regular selection of only a 

 few variables in the regression models, however, 



implies that determining the mechanisms behind 

 selected variables describing Florida Bay and adja- 

 cent waters could provide the requisite data for more 

 accurate forecasting of pink shrimp landings. Two 

 classes of data regularly occurred as forecast vari- 

 ables: 1) measurements of upland freshwater sup- 

 ply during the rainy season, and 2) fishing activity 

 on the Tortugas grounds during the waning months 

 of a given fishing year. 



Correlation between penaeid shrimp production 

 and freshwater inputs have been recorded world- 

 wide, even though the exact relationship between 

 landings and freshwater may be site- or species-spe- 

 cific. In a previous study of the Tortugas pink shrimp 

 fishery, quarterly landings were found to be either 

 positively or negatively correlated with ENP ground 

 water levels (Browder, 1985). Elsewhere, correlations 

 between penaeid shrimp landings and rainfall, river 

 discharge, or well-water levels have been positive 

 (Hildebrand and Gunter, 1953; Gunter and Edwards, 



