To determine whether our shipboard and aerial transect coverage 

 adequately sampled the SCB and the full range of environmental gradients 

 found within the study area, we determined the frequency distribution of 

 these variable characteristics for all 1,000 quadrats and compared these 

 statistically with the same parameters represented by, for example, 

 transect lines. Chi-square analysis showed no statistical difference 

 between the range of environmental parameters sampled in our transects 

 and that of all the quadrats in the SCB. We therefore conclude that our 

 shipboard and aerial transects adquately sampled the environmental vari- 

 ables with no significant bias. 



Since seasonal and geographical variations and sea surface temper- 

 atures affect the potential availability of cetacean prey, and therefore 

 cetacean distribution, we mapped water temperatures along our survey 

 tracks for each period and each trip. Surface temperatures were measured 

 hourly from the ship with a through-hull thermometer calibrated with a 

 bucket thermometer. For offshore or coastal areas or quadrats not visited 

 on a particular trip, we referred to maps generated by the U.S. Coast 

 Guard aerial surveys and the biweekly temperature projections published 

 by the National Marine Fisheries Service. Temperatures to .5°C were 

 assigned to each quadrat for a given survey period and entered into the 

 permanent computer files. The resultant computer-generated maps were 

 used to obtain mean sea surface temperature profiles throughout the Bight 

 and in each zone for comparison of cetacean densities with the above 

 stated environmental factors. 



For each cruise or flight series a battery of computer data output 

 was produced: 1) a chronological listing of all cetacean sightings 

 including species, number of individuals, and locations, 2) a listing by 

 specific location (quadrat) of all cetacean sightings, density and num- 

 bers, 3) a generated cetacean density map of the entire Bight, 4) a 

 computer-generated graphic display Bight-wide of cetacean density in 

 relation to environmental factors, 5) a computer-generated level of 

 effort map indicating quadrats visited and number of visitations. 



Stepwise multiple linear regression programs were designed to detect 

 and rank those environmental features which related significantly to 

 variation in cetacean density. The output of the regression program, 

 carried out on an IBM 360 computer, determined which environmental 

 features best predicted geographic variation in density. Independent 

 variables entered into the regression analysis included water depth, sea 

 surface temperature, bottom slope, distance to nearest land, distance to 

 mainland, as well as the inverse of these variables. If density along 

 the transect line was functionally related to some variables, for instance, 

 water depth, its measured value would be expected to increase or decrease 

 with an increase or decrease in the value of the independent variable. 

 This analysis calculated a sequence of equations beginning with regres- 

 sion of density against that single variable that made the greatest 

 reduction in the error sums of squares and continuing to rank and include 

 variables until no additional variation in density could be explained. 



B-67 



