FISHERY BULLETIN: VOL. 81, NO. 2 

 150° 160° 170° 180° 170° 160° 150° 140° 130° 120° 110° 



50° - 



FIGURE 1.— The approximate migration 

 route of the California stock of gray 20 



CANADA 



GRAY WHALE MIGRATION ROUTES - 



20° 



whales. 



150° 160° 170° 180° 170° 160° 150° 140° 130° 120° 110° 



To estimate total abundance by extrapolating from 

 recorded counts of passing whales one must deter- 

 mine the following: 



1) What proportion of the population, if any, 

 passes beyond sight of the observers? Does this 

 change with time or experience? How does the 

 observer's accuracy in estimating the distance to 

 passing whales vary with distance? 



2) Are there diel variations in migration rate? How 

 can daylight counts be used to estimate the num- 

 ber of whales passing at night? 



3) How do weather (visibility) conditions affect 

 census results? 



4) Does the observer's ability to count the number 

 of individuals within a passing group vary with 

 group size? 



5) Are the initiation and termination of the migra- 

 tion fully represented in the data? 



During the 1978-79 southward migration we con- 

 ducted two types of verification experiments aimed 

 at addressing the questions of points 1 and 4 above. 

 These were reported in detail in Reilly et al. (1980). 

 In one experiment we tested 12 observers simul- 

 taneously for accuracy in estimating distances to and 

 numbers within 50 events in which whales passed the 

 Granite Canyon station. The observers estimated the 

 distance offshore to within one of seven predefined 

 distance intervals, as during the actual annual cen- 

 suses (see Methods). We found significant hetero- 

 geneity between observers for both distance and 

 count estimates. Given this heterogeneity, there were 



also consistent biases recorded: In placing whales to 

 within correct intervals out to 1 mi (1.6 km) and 

 beyond 1.5 mi (2.4 km), and in estimating the true 

 number of individuals present in groups of one whale, 

 and four or more. Further analysis of this data (Reilly 

 1981) indicated that "experienced" observers were 

 on average no more accurate than inexperienced ob- 

 servers, but somewhat more precise. 



A second experiment was conducted during 1978- 

 79 to characterize the width of the migration corridor 

 offshore from the Monterey counting stations (Reilly 

 et al. 1980). A small aircraft flew a series of transects 

 perpendicular to the coast in the vicinity of the 

 stations, recording locations of sighted whales (Fig. 

 2). The results indicated that, contrary to previous 

 assumptions and characterizations of 95% of the 

 population passing within 1.6 km (Rice and Wolman 

 1971; Sund and O'Connor 1974), we found only 

 about 40% within 1.6 km, with significant numbers 

 passing offshore between 1.6 and 4.5 km. This ex- 

 periment was repeated during 1979-80, with results 

 reported here. 



Regarding night migration rate (point 2 above), af- 

 ter a review of all available information, we accepted 

 an assumption of a constant 24-h rate. Contrary to 

 the earlier report of Ramsey (1968), we found no 

 evidence of a diurnal fluctuation from the shore cen- 

 sus data. During the 1979-80 migration a new (pro- 

 totype) infrared image sensor, supplied by the U.S. 

 Department of Defense, was tested at Granite Can- 

 yon. As with previously tested night-vision devices 

 (Reilly et al. 1980), it proved unsatisfactory. 



The possible effect of visibility conditions on cen- 



268 



