CORRELATION OF UPWELLING INDEX AND DUNGENESS CRAB CATCH 



Louis W. Botsford and Daniel E. Wickham' 



ABSTRACT 



The statistical relationship between upwelling intensity and the annual Cancer magister catch is 

 examined using techniques of time series analysis. The statistical relationships between upwelling in 

 different years and crab catch in different years are examined. The statistic used is the correlation 

 coefficient for varying time lag. The results correlating upwelling and crab catch indicate that time lag 

 is short enough to demonstrate that adults are affected, and that the time lag is shorter in California 

 than in Oregon or Washington. The crab catches appear to be cyclic with a period of 9 yr in California 

 and 12 yr in Washington and Oregon. Possible causal mechanisms for the statistical relationships are 

 presented. Results suggest that variations in crab catches may be due primarily to biological rather 

 than abiotic influences. 



Large fluctuations in yearly catch in the Pacific 

 Coast Dungeness crab fishery have been observed 

 over the years for which records were kept. The 

 catch in some years is as low as 20% of the catch in 

 good years. Furthermore, the fluctuations seem to 

 be of a cyclic nature: 3 or 4 yr of low catch are often 

 followed by 3 or 4 yr of high catch after which the 

 same pattern occurs. Based on the fact that the 

 variation in catch was similar along the entire 

 west coast, the Pacific Marine Fisheries Commis- 

 sion (1965:38) concluded that a large-scale abiotic 

 influence was responsible. 



Many other explanations of the fluctuation in 

 crab catch have rested on Cleaver's (1949) conten- 

 tion that abundance of adult crabs fluctuates 

 widely in accordance with conditions affecting the 

 early stages, e.g., the Pacific Marine Fisheries 

 Commission (1965:38) suggested that seasonal 

 currents possibly sweep the larvae to unfavorable 

 settling areas. However, although the larvae may 

 be more sensitive than juveniles to environmental 

 changes and the percentage of them surviving is 

 certainly less, there has been no experimental 

 proof that fluctuation in abundance depends 

 primarily on larval survival. 



Peterson (1973) was the first to examine the 

 statistical relationship between the Dungeness 

 crab fishery and upwelling. (His paper serves as a 

 good introduction to our work; much valuable in- 

 formation is not repeated here.) He compared the 

 upwelling index developed by Bakun (1973) to the 



'Bodega Marine Laboratory, P.O. Box 247, Bodega Bay, CA 

 94923. 



Manuscript accepted January 197.5. 

 FISHERY BULLETIN: VOL. 73, NO. 4, 1975. 



crab catch records for the years 1949 to 1972. The 

 comparison was made for three different loca- 

 tions: Washington (lat. 48°N, long. 125°W), 

 Oregon (lat. 45°N, long. 125°W), and northern 

 California (lat. 42°N, long. 125°W). 



Peterson used graphical comparison and two 

 statistical techniques— the corner test and the 

 contingency table— to evaluate the relationship 

 between upwelling index (summed over the 

 summer months) and the annual crab catch for 

 several different lag times at each location. The 

 conclusion was that good crab catch followed a 

 good upwelling summer by IV2 yr in California and 

 Oregon and V2 yr in Washington. Two aspects of 

 these results are unexpected: 1) the short lag times 

 and 2) shorter lag time in Washington than in 

 California and Oregon. The short lag times seem 

 to preclude the effect of upwelling on larval or 

 early stages being the dominant cause of the large 

 catch. The shorter lag time in Washington is 

 unexpected because of the shorter generation 

 times of most species in the marine food chain at 

 the higher temperatures of more southerly waters. 



Peterson explained the lag differences by 

 proposing that the final molt from sublegal to legal 

 size was the determining factor. Dungeness crabs 

 molt later in more northerly waters (late fall in 

 Washington, late summer in Oregon, and early 

 summer in California). It was hypothesized that 

 the Washington animals were molting after the 

 summer upwelling and benefited from the 

 increased food supply, whereas animals in more 

 southerly waters molted prior to the increased 

 food supply and did not benefit until the following 

 summer molt. 



901 



