Correlation of winter temperature 

 and landings of pink shrimp 

 Penaeus duorarum in IMorth CavoWna 



William F. Hettler 



Beaufort Laboratory, Southeast Fisheries Science Center 



National Marine Fisheries Service, NOAA, Beaufort, North Carolina 28516-9722 



In habitats where low water temper- 

 ature is not a Hmiting factor, pink 

 shrimp Penaeus duorarum produc- 

 tion has been related to rainfall 

 and surface-water inflow (Browder 

 1985, Sheridan 1991). In contrast, 

 North Carolina landings of pink 

 shrimp were correlated with water 

 temperature during the previous 

 winter, but not to rainfall (Hettler 

 and Chester 1982). In that study, 

 the average water temperature of 

 the two coldest consecutive weeks 

 of each year recorded at a single 

 temperature station located at the 

 Beaufort Laboratory was a predic- 

 tor of spring landings (through July) 

 for the entire North Carolina 

 fishery. Fifteen years of tempera- 

 ture records and landings were used 

 to determine this relationship. Since 

 the last year reported (1981), 10 ad- 

 ditional years of temperature and 



landings data have become avail- 

 able. This note presents these new 

 data and uses the resulting 25-year 

 time-series to report that average 

 minimum winter water tempera- 

 ture remains a reliable basis for 

 forecasting landings of this species. 

 The temperature/landings relation- 

 ship previously published (Hettler 

 and Chester 1982) was recalculated 

 after adding the 1982-91 tempera- 

 ture and landings data (Table 1, Fig. 

 1). No evidence of curvilinearity in 

 the relationship could be found by 

 fitting higher-order polynomial 

 models. A time-series model was 

 not appropriate because pink 

 shrimp are 'annuals' and their an- 

 nual population levels generally 

 show low autocorrelation as sug- 

 gested by the 1962-91 North Caro- 

 lina pink shrimp heads-off landings 

 data (Fig. 1). Thus the simple linear 



Table I 



Actual and predicted landings (heads-off) of pink shrimp Penaeus duorarum in the 

 North Carolina spring fishery, February-July, based on average water temperature 

 of the two coldest consecutive weeks of the preceding winter. 



Year 



1982 

 1983 

 1984 

 1985 

 1986 

 1987 

 1988 

 1989 

 1990 

 1991 



Temp. 

 °C 



Landings (kg) 



Actual 



Predicted 



Percent over ( + ) 

 or under ( - ) 



5.0 

 8.8 

 5.9 

 5.4 

 6,9 

 8.3 

 6.1 

 8.1 

 3.7 

 10.0 



197,630 

 451,163 

 184,380 

 126,797 

 307,514 

 551,521 

 433,125 

 639,166 

 66,853 

 592,.381 



173,527 

 491,765 

 248,899 

 207,025 

 332,646 

 449,892 

 265,648 

 433,142 

 64,656 

 592,262 



+ 13.9 



-8.3 



-25.9 



-38.7 



-7.6 



-H22.6 



-1-63.0 



-I- 47.5 



+ 3.4 



< + 0.1 



model was retained. The new re- 

 gression to determine spring pink 

 shrimp landings in North Carolina 

 was 



Landings (kg) = 83747(T)- 245208, 



where T was the average tempera- 

 ture of the two coldest consecutive 

 weeks (°C). The relationship was 

 significant (P<0.001, r- 0.803). 

 The more general relationships of 

 average winter water temperature 

 (Dec-Mar) or average midwinter 

 water temperature (Jan-Feb) did 

 not correlate with landings over the 

 25-year time-series. 



Predicted landings of pink shrimp 

 were calculated and averaged with- 

 in 25% of the actual landings for the 

 recent 10-year period. Landings in 

 1991 were within >0.1% of the pre- 

 diction. Possible causes of the rela- 

 tively large deviations in some 

 years' landings from the predicted 

 are discussed in Hettler and Ches- 

 ter (1982) and include errors in the 

 process of estimating landings, 

 year-to-year changes in fishing ef- 

 fort, and, in addition, possible local 

 thermal anomalies. 



These new data continue to sup- 

 port the hypothesis that reduced 

 pink shrimp landings in North Caro- 

 lina are probably a result of cold kill 

 of overwintering shrimp caused by 

 cold water temperatures. In the 

 coldest years (1963, 1977, 1978, and 

 1990) when spring landings were 

 less than 100,000kg, lethal cold 

 water probably penetrated all but 

 the most highly protected over- 

 wintering estuarine habitat. North 

 Carolina is the northern limit in the 

 range of pink shrimp, thus this 

 species is more likely to encounter 

 low temperature stress in this loca- 

 tion than in more southerly loca- 

 tions. The linearity of the model is 

 perhaps a consequence of these 

 shrimp's inherent vulnerability to 

 cold water temperatures interact- 



Manuscript accepted 20 March 1992. 

 Fishery Bulletin, U.S. 90:405-406 (1992). 



405 



