HOLT ET AL.: MONITORING DOLPHIN ABUNDANCE 



Tc = days searched aboard the Cromwell , 

 and 



C = ratio of schools detected per day by 

 observers aboard the Jordan in the 

 calibration area during 1979 to 

 schools detected per day by observ- 

 ers aboard the Cromwell in the cali- 

 bration area during 1979. 



The percent of searching days when one to 

 three ships were used was allocated to each stra- 

 tum (Table 3) by the square root of school density. 

 The number of schools which would be expected to 

 be detected based on the standardized detection 

 rates then was calculated (Table 4). 



Table 3. — Percent of searching days 

 allocated by square root of density to 

 each stratum in the north and total 

 areas. 



Stratum 



North area 



Total area 



Inside 

 Middle 

 West 

 South 



45.6 

 36.5 

 17.9 



35.8 

 28.7 

 14.0 

 21.5 



Table 4. — Number of days searched and number of schools de- 

 tected per year of effort with use of 1 , 2, or 3 ships allocated to the 

 various strata by square root of density. 



North 



Total 



Stratum 



Number Number 

 days schools 



Number 

 days 



Number 

 schools 



1 ship = 120 days 

 Inside 

 Middle 

 West 

 South 



Total 



2 ships = 

 Inside 

 Middle 

 West 

 South 



240 days 



Total 



3 ships = 

 Inside 

 Middle 

 West 

 South 



Total 



360 days 



55 

 44 

 21 



120 



110 

 88 

 42 



240 



165 



132 



63 



360 



149 

 65 

 14 



228 



298 



130 



28 



456 



447 



195 



42 



684 



43 

 34 

 17 

 26 



120 



86 

 68 

 34 

 52 



240 



129 



102 



51 



78 



360 



116 

 50 

 12 

 40 



218 



232 



100 



24 



80 



436 



348 



150 



36 



120 



654 



which uses pooled density and school size esti- 

 mates, than when using method A, which uses 

 estimates calculated for each stratum (Table 5). 

 This is because large variances associated with 

 the method A population size estimates occur due 

 to small sample sizes in some strata. Therefore, 

 method B was used in subsequent calculations. 



The same number of years is required to detect 

 a specific trend if the north or total areas are 

 surveyed (Table 5). This result is true only if the 

 1977-83 data, which contain small sample sizes 

 in the south stratum, are representative of future 

 data. However, the northern offshore spotted dol- 

 phin stock occurs only in the north area and elim- 

 ination of the south stratum will ensure better 

 coverage of this north area, especially in the west 

 stratum where sample sizes are minimal for ap- 

 plying the Fourier series model (Table 4). There- 

 fore, subsequent calculations were made only for 

 the north area. Annual population estimates for 

 the northern stock would be biased only if sub- 

 stantial variation in the amount of dolphin mi- 

 gration between the north area and south stra- 

 tum occurred during survey years. 



Table 5. — Number of years required to detect an annual 5% de- 

 crease in spotted dolphin population size using 1 , 2, or 3 ships and 

 2 different methods of pooling data. Method A utilized Equation (2) 

 in text while method B utilized Equation (3). Alpha and beta levels 

 equal 0.05, and effort was allocated to the various strata by square 

 root of density. Number of schools expected to be detected each 

 year determined using detection rates from Equation (9). K deter- 

 mined using Equation (8). CV (N) denotes coefficient of variation 

 of population abundance estimate. 



Stratum 



Number 

 ships 



Number 

 schools 



K CV (A/) 



Years 

 required 



North area 

 Method A 



Method B 



Total area 

 Method A 



Method B 



1 

 2 



3 



1 



2 

 3 



1 

 2 

 3 



1 

 2 

 3 



228 



456 

 684 



228 

 456 

 684 



218 

 436 

 654 



218 



436 

 654 



4.05 



3.06 



3.93 



2.94 



RESULTS 



For either the north or total area, the same 

 decrease in spotted dolphin populations can be 

 detected 2 to 4 years earlier using method B, 



At the 5% error level, only rates of change of 

 11% per year or greater can be detected in a 5-yr 

 survey period, even using three ships per year 

 (Table 6). This is a rather high rate of decrease. 



443 



