Julian and Beeson: Estimates of marine mammal, turtle, and seabird mortality 
277 
Table 5 
Observed and estimated cetacean, pinniped, turtle, and seabird entanglement, stratified by year, in the California halibut and 
angel shark set gillnet fishery during the NMFS Observer Program, July 1990-December 1995. Estimates of entanglement are 
reported to the nearest individual. Estimated coefficients of variation (CV) are included in parentheses; ( — ) indicates CV was 
undefined. Effort and estimates for 1990 pertain to the third and fourth quarters only. 
Year 
1990 
1991 
1992 
1993 
1994 
1995 
1 
Unidentified sea lion 
67 
867 
(0.22) 
143 1,850 
(0.16) 
341 
3,438 
(0.28) 
239 
1,977 
(0.13) 
109 
905 
(0.15) — 
729 
(0.08) 
Harbor seal 
30 
411 
(0.23) 
43 
615 
(0.23) 
90 
1,204 
(0.47) 
71 
475 
(0.13) 
23 
227 
(0.33) — 
228 
(0.13) 
Unidentified turtle 
0 
0 
(— ) 
0 
0 
(— ) 
0 
0 
(— ) 
2 
13 
(0.64) 
0 
0 
(— ) — 
5 
(0.59) 
Common loon 
0 
0 
(— ) 
4 
48 
(0.60) 
1 
7 
(0.92) 
0 
0 
(— ) 
0 
0 
(— ) — 
— 
(— ) 
Western grebe 
0 
0 
(— ) 
2 
22 
(0.68) 
3 
23 
(0.70) 
1 
6 
(0.90) 
0 
0 
(— ) — 
2 
(0.61) 
Brandt’s cormorant 
2 
41 
(0.78) 
41 
494 
(0.37) 
20 
321 
(0.58) 
5 
25 
(0.45) 
2 
16 
(0.66) — 
9 
(0.40) 
Unidentified seabird 
0 
0 
(— ) 
2 
22 
(0.68) 
5 
37 
(0.41) 
1 
6 
(0.90) 
0 
0 
(— ) — 
3 
(0.83) 
1 Estimates for 1995 were based on stratified rates from 1993 results. 
m = Dr , 
(3) 
~ 2 
n 2 ^ 2 
= U o r 
(4) 
Variables k f and d i represent the observed kill and 
number of days for the i th trip; d avg is the sampled 
mean number of days per trip; <rj, 6 and d d k are 
the sample variances and covariance of d i and k ( ; d 
and n are the observed number of days and trips, 
and D represents the total number of days of effort. 
The finite population correction factor, (1 - n/N), 
where N is the total number of trips, was approxi- 
mated by using (T - d ID) because the total number 
of driftnet trips was not determined during the esti- 
mation of effort. 
Mortality estimation in the set gillnet fishery In 
the setnet fishery, a trip was selected and all net pulls 
during the trip were observed. Because greater than 
99% of all observed set gillnet trips were one day in 
length, a trip was considered equivalent to an effort- 
day. Net pulls were not randomly sampled and total 
number of net pulls per trip was not recorded for 
unobserved trips, therefore a trip was considered the 
sampling unit. For observed trips, the overall aver- 
age number of net pulls per trip (or effort-day) was 
3.08 (CV=0.43), and the overall average number of 
marine mammal entanglements per trip was 0.48 
(CV=3.10) during the NMFS Observer Program. 
Researchers had previously shown ( Perkins et al. 5 ) 
that quarter of year and set location were signifi- 
cant predictors of sea lion and harbor seal entangle- 
ment. Consequently, estimation of incidental kill in 
the setnet fishery was stratified by area for all mam- 
mal species. The geographic area fished was divided 
into four strata: southern California, Channel Is- 
lands, Ventura, and central California. Central Cali- 
fornia included all effort south of Bodega Bay and 
north of Point Conception, 34°30’N (CDFG blocks 
437-650). There was no legal coastal setnet fishing 
north of Bodega Bay. The Channel Islands stratum 
included all effort within CDFG blocks containing 
part of any Channel Islands (CDFG blocks 684-690, 
707-713, 760-762, 765, 806-807, 813-814, 829, 849- 
850, and 867). The southern California stratum in- 
cluded all effort south of 33°50' N (blocks 718-918), 
excluding CDFG block 776 and the Channel Islands 
stratum. The Ventura stratum contained the remain- 
ing area. Incidental kill estimation for sea lions and 
harbor seals included additional stratification by 
quarter of the year for the southern California and 
Ventura areas because for these species, overstrati- 
fication was not a problem. 
Yearly estimates of incidental kill in the setnet fish- 
eries were calculated for all species of marine mam- 
mals observed entangled (Tables 4 and 5). Estimates 
for turtles and seabirds were calculated by the same 
method. For each species, kill rate, r , and total inci- 
dental kill, m s , were estimated for each area by us- 
ing a mean-per-unit (MPU) estimator with effort- 
days as the sampling unit (Diamond and Hanan 1 ). 
Formulae for MPU estimators and their estimated 
variances for each area are (Cochran, 1977) 
= 
£A, S 
d a 
l--=2- 
D 
d. Y 1 A 
s J 
\ d sJ 
d ls > 
m s = D s r s , 
(5) 
( 6 ) 
(7) 
