Calambokidis et al.: Distribution and abundance of marine mammals off the northern Washington coast 



567 



Data analysis 



For ship surveys between 1995 and 2000, position and 

 oceanographic data (including depth, sea surface tem- 

 perature) logged by the ship's computer were later rec- 

 onciled with the sighting and effort data recorded by 

 the observers. Sighting positions were analyzed for each 

 species for water depth, distance from shore, distance 

 from shelf edge (200-m depth contour) and sea surface 

 temperature. Data analysis and mapping were conducted 

 by using a geographic information system (GIS) with 

 Arclnfo software (ESRI, Redlands, CA). Data from the 

 shorter 2002 ship survey were included in the summary 

 of sightings but were not available for the analyses of 

 sightings related to oceanographic features. 



Line-transect analysis to determine density and abun- 

 dance was conducted for the two species with more than 

 30 sightings (humpback whales and Dall's porpoise). 

 We used the program (Distance, version 3.5, Research 

 Unit for Wildlife Population Assessment, University of 

 St. Andrews, St. Andrews. UK) to conduct analyses. 

 For these analyses, we used only effort and sightings 

 from the regular east-west transect lines and did not 

 include on-effort data from opportunistic lines or cross- 

 tracks. We included sightings made by secondary as 

 well as the primary observer. Although whales were 

 reportedly seen out to 6 nmi, we truncated the sight- 

 ings at 3 nmi for humpback whales and 2.5 nmi for 

 Dall's porpoise. For humpback whales we included 16 

 sightings of unidentified whales (unidentified mainly 

 because of distance). These were probably humpback 

 whales because the only other large whales that were 

 seen in the surveys were a few gray whales seen close 

 to shore. Distance position data were incomplete for 13 

 of the 188 whale sightings and 14 of 82 Dall's porpoise 

 sightings; for these the missing value was randomly 

 selected from the observed measurements. 



The Distance program was used to select the best 

 model for sighting probability in relation to distance off 

 the transect. We allowed the program to select among 

 models (half-normal, uniform, hazard-rate, and nega- 

 tive exponential) and varying numbers of adjustment 

 terms (cosine and simple polynomials) based on lowest 

 Akaike's information criterion (AIC) score. All years 

 were pooled for the model of sighting probability, but 

 encounter rate and group size were calculated by year. 

 An adjustment to group size was calculated if there 

 was a significant group size bias with distance from 

 the track line, which was not the case for humpback 

 whales but was present in some years (1996 and 1997) 

 for Dall's porpoise. 



Area was calculated for abundance estimation based 

 on the zone covered by the regularly scheduled transect 

 lines covered in most years (study area was considered 

 to encompass waters 2.5 nmi north of the northernmost 

 line and 2.5 nmi south of the southernmost line). The 

 only annual adjustment for area was for humpback 

 whales in 2002. Surveys in that year did not cover 

 the southern end of the study area (because of limited 

 ship time), an area with a typically lower abundance 



of whales. To avoid extrapolating the higher density 

 of whales from the northern portion of the study area 

 to this region, we excluded this missed area from the 

 abundance estimates. 



Estimates of abundance for humpback whales were 

 also calculated by using capture-recapture models (Se- 

 ber, 1982; Hammond, 1986). We used identifications 

 obtained in pairs of adjacent years taken from 1994 to 

 2002 to generate Petersen capture-recapture estimates. 

 The Chapman modification of the Petersen estimate 

 (Seber, 1982) was used because it was appropriate for 

 sampling without replacement (Hammond, 1986). 



Results 



In total, there were 706 sightings of 2467 animals over 

 the six ship surveys combined (Table 3). Fifteen differ- 

 ent marine mammal species were seen: nine cetacean 

 species, five pinniped species, and the sea otter were 

 identified. Each year, 9 to 12 different species were seen, 

 except in 2002 when only six species were observed. 

 This 2002 survey, although shorter than those of the 

 other years, showed a dramatic change in the species 

 diversity and numbers of animals. We saw many more 

 humpback and Dall's porpoise than in previous years. 

 We also noted the absence of six regularly observed spe- 

 cies: harbor porpoise, gray whales, Pacific white-sided 

 dolphins (Lagenorhynchus obliquidens), Risso's dolphin 

 (Grampus griseus), harbor seals, and California sea lions 

 iZalophus californianus). 



Humpback whales 



Of the large cetaceans, humpback whales were the most 

 common species seen; there were 232 sightings of 402 

 animals during ship surveys (Table 3). Largest numbers 

 of humpback whales were seen in 2002, when there 

 were 79 sightings of 139 individuals during the one- 

 week survey. Group sizes ranged from 1 to 8 animals 

 (mean=1.7, SD=1.1). Only six calves were recorded from 

 the ship surveys — probably because it was difficult to 

 identify calves at the distance at which most sightings 

 were made. Of these six sightings of mothers with calves, 

 four sightings were outside the primary areas where 

 other humpback whale groups were seen. 



Sightings were concentrated in the northern part of 

 the study area between Juan de Fuca Canyon and the 

 outer edge of the continental shelf, an area known as 

 "the Prairie" (Fig. 2). A small area east of the mouth of 

 Barkley Canyon and north of the Nitnat Canyon where 

 the water depth was 125-145 m had a high density of 

 sightings in all years. A smaller number of humpback 

 whales were also seen on Swiftsure Bank. Sightings in 

 2002 were not only more numerous but more broadly 

 distributed; sightings were recorded in the areas de- 

 scribed above and also farther south and closer to shore 

 than those seen in previous years. 



Line-transect estimates for humpback whales were 

 very consistent in the first five surveys (1995 to 



