64 



Fishery Bulletin 97(1), 1999 



ber of index fishermen within each statistical dis- 

 trict, total effort for each district was calculated by 

 dividing its total catch by the average catch per trap 

 haul (CPTH) for index fishermen in that district. 



Because seasonal trends were evident in tempera- 

 ture, CPTH, and effort data, any analysis involving 

 daily observations of these parameters were 

 detrended by using first differencing i.e. /-(;'- 1), 

 where / is the individual observation of temperature, 

 CPTH, or catch. This avoided spurious negative cor- 

 relations between the decreasing catches through- 

 out the season that were due to depletion of the fish- 

 able biomas, and increasing spring-summer tempera- 

 tures. Because most fishermen on the eastern shore 

 attempt to haul their traps daily and because longer 

 soaks are usually associated with unfavorable weather 

 (which tends to affect most fishermen in the same area ), 

 there was no attempt to adjust for differences in soak 

 times within or between each fisherman's time series. 

 Inspection of the data showed that gaps in the indi- 

 vidual time series tended to occur on the same days, 

 which coincided with storm days and Sundays. 



Temperature information was also available from 

 DFO moored recorders moored along the eastern 

 shore during the period when logbooks were kept. 

 Years with a strong association between tempera- 

 ture and catch were identified for further study at 

 smaller spatial and temporal scales as follows: daily, 

 temperature, catch, and effort data from the 1986- 

 94 fishing seasons at the location with the most com- 

 plete longer-term records, i.e. Port Bickerton, were 

 detrended as described above and the resulting indi- 

 vidual series correlated. This analysis identified 1991 

 and 1994 as years with significant correlations. The 

 latter year was chosen for further investigation into 

 the nature of this relationship because of the avail- 

 ability of data at small spatial s.cales throughout the 

 area from FSRS records. Three matrices of port ver- 

 sus date, one each for the detrended temperature, 

 CPTH and effort data, were then developed for fur- 

 ther correlation analysis. Each matrix resulted in a 



port versus port matrix of correlation coefficients, i.e 

 indices of similarity between port pairs in terms of 

 the daily patterns of change for the respective pa- 

 rameters. Correlations between port pairs could then 

 be plotted against distance between them. 



Results and discussion 



Effort information is not available at the largest spa- 

 tial and longest temporal scale examined (Atlantic 

 coast of Nova Scotia, 50 years). At this scale, how- 

 ever, there is a significant correlation between an- 

 nual mean sea surface temperatures in the approxi- 

 mate center of the area (i.e. Halifax harbor) and sum- 

 marized lobster catches ( Fig. 2; Table 1 ) at short lags 

 prior to 1974. This correlation was also noted by 

 Campbell et al (1991). Catches are significantly cor- 

 related (P<0.01) with temperatures for three cycles 

 prior to 1974 at lags of 1-3 years on the eastern shore 

 and 0-2 years on the south shore. After 1974, corre- 

 lations are highest (although marginally insignifi- 

 cant) at lags of 8 years on the eastern shore and 6 

 years on the south shore, approximately the time 

 lobsters take to recruit to the fishery in these areas. 

 This finding suggests that during this period, tem- 

 perature or an unknown covariate also influenced 

 the survival of larvae and subsequently increased 

 recruitment to the fishery. The increase in catches 

 in both areas after 1974 is consistent with the north- 

 west Atlantic coast-wide "recruitment pulse" 

 ( Pezzack 1992 ), which is generally attributed to a com- 

 mon environmental influence, although Drinkwater et 

 al ( 1996) concluded that this environmental influence 

 was not temperature. Prior to 1974 the near in-phase 

 changes in temperature and catch along the eastern 

 and south shore could be due to temperature-induced 

 changes in catchability during warmer years i.e. to 

 greater activity of lobsters on the grounds or faster 

 growth to legal size, or to both. However, without long- 

 term effort information, changes in effort cannot be 



