BOTSFORD ET AL.: CYCLIC COVARIATION IN CALIFORNIA FISHERIES 



San Francisco and Monterey. The unit of salmon 

 catch is kilograms of dressed fish with heads on. 



Crab catch (kilograms) was summarized by 

 season rather than calendar year so that a sea- 

 son's catch includes catch from November and 

 December of the previous year. The geographic 

 breakdown of crab catch was the same as the 

 salmon catch. Seasonal distribution of crab catch 

 and effort was available only from 1952 to 1976 4 

 for a northern California region which included 

 an average of 93% of the total northern California 

 catch. In addition to crab catch data, we have 

 also used recent estimates of preseason legal 

 abundance (Methot and Botsford 1982). These 

 estimates were computed from the decline in 

 catch-per-unit-effort within each season accord- 

 ing to the Leslie method. Gotshall( 1978) also esti- 

 mated preseason legal abundance for the years 

 1967-72 using the Leslie method. His results for 

 those years are essentially the same as those used 

 here. McKelvey et al. (1980) also estimated pre- 

 season legal abundance, but used a method that 

 depended on the estimated total number of pots 

 in the fishery and annual catch. Since the rela- 

 tionship between these variables can change 

 from year to year in this fishery, we did not use 

 their estimates. 



We present first the statistical characteristics 

 of cyclic fluctuations in the northern California 

 salmon and crab catch records as reflected in 

 estimates of their auto- and cross-correlation 

 functions. We then examine characteristics of 

 each of the two salmon species in the fishery. The 

 northern California salmon fishery is composed 

 of king (or chinook) salmon, Oneorhynchus tsha- 

 wytscha, which originate primarily in coastal 

 rivers of northern California and Oregon, and 

 silver (or coho) salmon, 0. kisutch, which origi- 

 nated primarily in coastal rivers of northern 

 California and Oregon in the 1950's, but depend 

 increasingly on hatchery production in Oregon 

 in the 1960's and 1970's (Pacific Fisheries Man- 

 agement Council 1978). 



We then compare the characteristics of the 

 northern California fishery to the central Cali- 

 fornia fishery which differs in two respects: 1) It 

 includes a period of protracted decline in the 

 crab fishery and 2) it involves salmon stocks that 

 originated in the Sacramento and San Joaquin 

 River systems (Pacific Fisheries Management 

 Council 1978). The central California crab fish- 



ery declined near 1960 and has remained at a low 

 level since then. Putative causes of this decline 

 and continued low level include an increase in 

 sea temperature (Wild 1980), a predatory worm 

 (Wickham 1979), and an increase in individual 

 growth rate (Botsford 1981). We compare char- 

 acteristics of the northern California fisheries 

 with the central California fishery both before 

 and after the decline. 



We then examine a specific potential cause of 

 the observed covariation. The most obvious and 

 perhaps the most parsimonious explanation of 

 the observed covariation in catch records is 

 switching behavior of fishermen. The proposed 

 hypothesis is simply that, during years of high 

 crab abundance, more fishermen continue to fish 

 for crab rather than beginning to fish for salmon 

 when the salmon season opens. The legal crab 

 season opened in December and continued at 

 least through June in the years of interest. The 

 salmon season opened in April or May, depend- 

 ing on year and species. Although most of the 

 crab catch is landed early in the season, crab and 

 salmon seasons do overlap, thus providing an 

 opportunity to switch. The possibility that this 

 mechanism is responsible for the observed be- 

 havior is examined here from three points of 

 view: 1) A comparison of catch during overlap- 

 ping and nonoverlapping segments of the salmon 

 season, 2) analysis of changes in mean date of the 

 salmon catch, and 3) calculation of the relation- 

 ship between salmon catch and crab catch per 

 delivery during May and June. 



RESULTS 



Northern California Total Catch 



Estimates of the auto-correlation function for 

 both total northern California Dungeness crab 

 landings and total northern California salmon 

 landings for the years 1940 to 1976 are of the 

 form that would arise from cyclic processes of 

 period 9 or 10 yr (Fig. 2). They both fall to a sig- 

 nificant negative value of correlation then rise to 

 a significant positive value of correlation. 5 The 

 auto-correlation of crab abundance is not shown 



"Annual Market Crab-Statewide Reports, California Depart- 

 ment of Fish and Game. 1952-77. 



5 Estimates of the same functions for the time period 1952 to 

 1976 imply that crab is more cyclic (dropping to 0.7 then in- 

 creasing to 0.8) while salmon is less cyclic (dropping to -0.3 

 then increasing to 0.4). In this analysis, as in the raw data (Fig. 

 1), salmon catch appears to be more cyclic in earlier rather 

 than later years, while the crab catch appears to be more cyclic 

 in later years. 



793 



