744 



Fishery Bulletin 93(4), 1995 



The important difference is that the five-year time 

 series did not reveal a trend in abundance, whereas 

 the analysis including the new (1991 and 1993) data 

 indicated a decline in harbor porpoise abundance in 

 central California during the eight-year period 1986- 

 93. To investigate the possibility that animals may 

 have moved northward into northern California, the 

 analysis was repeated with northern California as a 

 third area stiatum. The results were essentially the 

 same: the year effect was slightly less pronounced 

 but still significant at a = 0.10 (Table 3). Figure 3 

 shows a plot of relative abundance (porpoise observed 

 per kilometer) in each of the three defined areas for 

 the period 1986-93, adjusted for the effects of sea 

 state and cloud cover (based on the parameters of 

 the best-fitting model). The combined relative abun- 



dance for all of central California, 

 calculated as an average of the val- 

 ues of porpoise per kilometer for ar- 

 eas 1 and 2, weighted by the propor- 

 tion of the total study area encom- 

 passed by each (33.6% for area 1; 

 66.4% for area 2), is indicated by a 

 dashed line in Figure 3. 



Discussion 



The indication of a declining trend in 

 abundance is somewhat surprising, 

 given that the central California 

 population of harbor porpoise was 

 expected to be recovering from im- 

 pacts of heavy fishery-related mortal- 

 ity prior to about 1987 (Barlow and 

 Forney, 1994). The point estimate for 

 the decline corresponds to a 9.3% per 

 year decrease (coefficient of variation, CV=0.56) in 

 harbor porpoise abundance for central California (or 

 8.3%, CV=0.56, if northern California surveys are in- 

 cluded in the analysis); however, the confidence in 

 this value is low because of the low power of the test. 

 There are a number of possible explanations for the 

 observed decline, including 1) statistical error, 2) 

 movement of animals out of the study area, 3) ef- 

 fects of fishery-related mortality, and 4) change in 

 natural mortality and net reproduction. Each of these 

 possibilities will be discussed separately below. 



Statistical error 



It is possible that the results of the test are incor- 

 rect, that is, an a-error ( detecting a trend although 



