FISHERY BULLETIN: VOL. 86, NO. 3 



tery reflection of sunlight, and causes whitecaps 

 which obscure subsurface observation. Cloud cover 

 decreases penetration of sunlight into the sea and 

 causes its surface to appear dark and glazed. It is 

 therefore not surprising that calm seas and clear 

 skies result in higher apparent densities. 



It is more surprising that apparent density did not 

 vary with observer's subjective appraisal of surface 

 penetration. This may have been because surface 

 penetration was only tested in paired cases for which 

 sea state and cloud cover were identical. In these 

 cases, differences in surface penetration may have 

 been due primarily to subjective differences in the 

 way individual observers were coding it. If all cases 

 are considered, surface penetration is very highly 

 correlated with sea state and cloud cover and prob- 

 ably could be used as an alternative measure of 

 sighting conditions. We prefer, however, to use sea 

 state and cloud cover because their measure is less 

 subjective than surface penetration and could more 

 easily be used by other researchers. 



Missed Animals 



A principal assumption of strip transect methods 

 is that all individuals within the strip are counted. 

 We cannot necessarily meet this assumption just by 

 eliminating the categories of sighting conditions 

 with significantly lower density. In fact, it is possi- 

 ble that this method of selectively eliminating data 

 could overestimate density by eliminating a category 

 of sighting conditions which (by random chance) had 

 a significantly lower density. We do not believe that 

 this is likely in the cases of sea state or cloud cover 

 because the trends were the same for both surveys 

 and because the categories that were eliminated 

 were judged a priori as being poorer sighting con- 

 ditions. We believe that porpoise density is more 

 likely underestimated due to missed animals. Kraus 

 et al. (1983) found that observers in aircraft saw only 

 14% of the harbor porpoise groups known to be pres- 

 ent based on shore-based observations. Missed 

 animals may include some individuals that were near 

 the surface and visible but were not seen, as well 

 as others that were diving and were too deep to be 

 seen. 



We infer that some near-surface animals were 

 missed based on three reasons. First, apparent den- 

 sity decreased with increasing sea state and cloud 

 cover, hence near-surface animals must be missed 

 in (at least) the poorer conditions. (An alternative 

 explanation is that porpoise spend less time near the 

 surface when sighting conditions are poor.) Second, 

 of the two principal observers in 1985, the less ex- 



perienced observer may have missed more porpoise. 

 Third, in five instances in 1984 and one instance in 

 1985, the data recorder saw harbor porpoise in the 

 inside swath that were missed by the observer. In 

 three of these cases, conditions were Beaufort 1 with 

 <25% cloud cover. Recorders searched only occa- 

 sionally as conditions permitted, so it is not possi- 

 ble to use these data to quantify how many near- 

 surface animals were missed under good conditions. 



Based on behavioral observations, we can also be 

 certain that some harbor porpoise are missed be- 

 cause they are too deep to be seen. Water visibility 

 was typically only 2-5 m during the surveys. Aerial 

 observers have seen harbor porpoise dive out of view 

 during the passing of the plane or while circling. 

 Harbor porpoise were not visible from the helicopter 

 during dives, even in very calm water. Some frac- 

 tion of porpoise must be missed because they are 

 too deep to be seen. 



We have tried to account for the fraction of diving 

 animals by dividing density estimates by the frac- 

 tion of time harbor porpoise are known to be near 

 the surface (i.e., within surfacing series). Uncor- 

 rected estimates of harbor porpoise density are 

 based on the assumption that porpoise are never too 

 deep to be seen; this undoubtedly results in an 

 underestimate of porpoise density. Our method of 

 adjusted estimates of porpoise density assumes that, 

 when diving, porpoise are always too deep to be 

 seen; this has been corroborated by helicopter ob- 

 servations. The latter estimate should therefore be 

 closer to the true value of porpoise density (if biases 

 due to other factors have been eliminated). 



Offshore Distribution of 

 Harbor Porpoise 



In 1985, aerial transects were flown at 0.61 and 

 1.85 km from the shore. The latter value was chosen 

 to correspond approximately to the ship transects 

 along the 10-fathom (18.3 m) isobath. The intent was 

 to directly compare aerial estimates of density to 

 estimates made from ships at the same distance 

 from shore (considered below) and to provide a 

 means to extrapolate ship estimates to regions that 

 were too shallow to survey by ship. In regard to the 

 latter, we wish to know whether harbor porpoise 

 density at 0.61 km from shore is different from that 

 at 1.85 km. 



In 1985, surveys at 0.61 km were never flown 

 under good sighting conditions (Beaufort & 1 and 

 <25% cloud cover), so direct comparisons between 

 0.61 and 1.85 km are not possible. Considering only 

 the best category of sighting conditions, it is possi- 



442 



