BARLOW: SHIP SURVEYS OF HARBOR PORPOISE 



ficient of variation in the pooled ship estimates is 

 nearly 25%, these estimates are in very close agree- 

 ment. However, because the aerial estimates are 

 based only on the small fraction of the coastline that 

 was surveyed under optimal conditions, the ship 

 estimates are probably a better representation of 

 porpoise density for the entire coast. 



Of the areas surveyed, harbor porpoise density is 

 highest in northern California and Oregon. The 

 highest density was seen in northern Oregon (region 

 7) during survey 1. The second highest density was 

 observed in northern California between Bodega 

 Head and Cape Mendocino (region 4) on survey 3. 



Two areas in central California (regions 1 and 3) 

 were found to have very low densities. Region 1 in- 

 cludes the Big Sur coastline from Point Conception 

 to Point Sur. This area is characterized by steep 

 depth gradients and hence has little habitat that is 

 suitable for harbor porpoise. Region 1 was relatively 

 well covered, with 378 km of trackline surveyed at 

 Beaufort sea states 0-2. In contrast, region 3 in- 

 cludes the Gulf of the Farallons with its broad 

 coastal shelf within the 100 m isobath. Based on 

 surveys of 764 km, Szczepaniak (fn. 9) estimated 

 1,033 harbor porpoise are found in the Gulf of the 

 Farallones alone. This is much greater than my 

 estimate of 112 animals in region 3 based on only 

 175 km of survey effort. Because of his greater 

 amount of search effort in this area, I believe that 

 Szczepaniak' s estimates for region 3 are more ac- 

 curate than mine. Although regions 1 and 3 were 

 both identified as low density areas, more confidence 

 can be placed on this conclusion for region 1 than 

 for region 3. 



Abundance 



The size and behavioral characteristics of harbor 

 porpoise make estimating their abundance difficult. 

 Harbor porpoise are small, occur in groups of only 

 a few individuals, and surface without conspicuous 

 splashes; their distribution is extremely patchy. 

 Even v^th 5 observers, the effective path width that 

 can be searched from a ship is <1 km, and that path 

 width decreases very rapidly in rougher sea states. 

 All of these factors contribute to high variability in 

 the abundance estimates presented here. Seasonal 

 and year-to-year changes in the distribution of har- 

 bor porpoise may also contribute to the variability 

 seen within geographic strata. These are, however, 

 the best (and, for some regions, the only) estimates 

 of harbor porpoise abundance for the study area. 



Although there are no prior estimates for Oregon 

 or Washington coasts, Dohl et al. (fn. 4) estimated 



harbor porpoise abundance in central and northern 

 California. Their estimates range from 3,000 har- 

 bor porpoise in autumn to 1,600 in summer, which 

 correspond (approximately) to the pooled estimate 

 of 11,457 for regions 1-4 based on the present study. 

 There are, however, several problems with the ap- 

 plication of their methods to the estimation of har- 

 bor porpoise abundance. In a direct comparison with 

 shore counts, Kraus et al. (1983) showed that 

 observers on aircraft saw only 10-20% of harbor 

 porpoise groups. Dohl et al. (fn. 4) did not apply a 

 correction to account for harbor porpoise groups 

 that are submerged at the time the aircraft passed. 

 Also, Dohl et al. did not stratify estimates by 

 distance from shore or depth. Although most of their 

 harbor porpoise sightings were within 0.5 km (0.25 

 nmi) of shore, their density estimates were extrap- 

 olated to an area extending 166 km from the coast. 

 Estimates from the current study are based on 

 better methodology than previous estimates. 



In addition to exposed coastal habitats, harbor 

 porpoise are also found in bays along the coasts of 

 California, Oregon, and Washington. Goetz (1983) 

 reported that harbor porpoise are found throughout 

 the year in Humboldt Bay, CA. Harbor porpoise 

 have been seen in San Francisco Bay, but are 

 described as rarely present^^. Abundance of harbor 

 porpoise in inland waters may, however, vary 

 seasonally (Taylor and Dawson 1984). No estimates 

 exist for the total number of harbor porpoise in- 

 habiting bays. Survey effort in the present study 

 was limited to exposed coastal areas (including 

 Monterey Bay, but excluding San Francisco Bay, 

 Humboldt Bay, Coos Bay, Yaquina Bay, the mouth 

 of the Columbia River, Willapa Bay, and Grays Har- 

 bor). If harbor porpoise density in bays were the 

 same as that which was observed along the 18 m 

 isobath, population sizes presented here could be in- 

 creased by approximately 3.1% to account for por- 

 poise inhabiting 900 km"^ (the approximate com- 

 bined area of Humboldt Bay, Coos Bay, Yaquina 

 Bay, the mouth of the Columbia River, Willapa Bay, 

 and Grays Harbor). 



Line Transect Assumptions 



Biases in abundance estimates can be an even 

 greater problem than high variability. In the case 

 of estimates presented here, biases could be intro- 

 duced if the assumptions of line transect sampling 



'^Szczepaniak, I. D., and M. A. VV^ebber. 1985. Status of the 

 harbor porpoise {Phocoena phocoena) in the eastern North Pacific, 

 with an emphasis on California. Contract report to the Center 

 for Environmental Education, Washington, D.C., 52 p. 



429 



