Legault and Ehrhardf: Correcting annual catches from seasonal fisheries for use in virtual population analysis 
285 
6,000 — 
+ 
+ 
5,500 — 
+ 
xT 
+ 
u-< — 
o 
+ 
6 
a 
+ 
o 5,000 — 
+ 
s 
<D 
+ 
O 
<D — 
£ 
+ 
o 
O 
+ 
4,500 — 
+ 
+ 
|| 
Jan Mar May Jul Sep Nov 
Month of catch 
Figure 4 
Monthly corrected catch values based on the assumption that 5,000 fish were 
caught during only one given month, N K+1 is 10,000, and the annual natural 
mortality rate is 0.3. 
population analysis when catches from seasonal fish- 
eries are used directly without correction. 
Examples 
Three scenarios were examined by means of simula- 
tion to demonstrate the algorithm. An initial popu- 
lation structure and recruitment pattern were set, 
and a given catch was removed at a constant fishing 
mortality rate under three scenarios: 1) the catch 
occurred only in January, 2) the catch was split evenly 
between June and July, and 3) the catch occurred 
only in December. The natural mortality rate was 
constant for all ages and years at 0.5 per year. The 
selectivity curve was sigmoid to follow a trawl-type 
pattern. A tuning index was collected from the popu- 
lation without error for use in calibrated virtual popu- 
lation analysis (VPA). The VPAused in the examples 
was FADAPT3 (Restrepo 1 ). Any sequential popula- 
1 Restrepo, V. 1996. Cooperative Unit for Fisheries Education 
and Research, Rosenstiel School of Marine and Atmospheric 
Science, Univ. Miami, 4600 Rickenbacker Causeway, Miami, FL 
33149. Personal commun. 
tion analysis program could be used for these simu- 
lated examples because of a lack of error distribu- 
tions for the data. Seven years were simulated for 
each scenario by applying the catch during the ap- 
propriate month(s) and the observed catch at age 
recorded for each year. For each scenario, 3 different 
sets of catch data were used as input for the VPA: 1 ) 
the observed catch-at-age data, 2) the corrected catch- 
at-age data with guesses for the N K+1 values, and 3) 
the recorrected catch-at-age data with the popula- 
tion numbers at age taken from the results of the 
corrected catch-at-age VPA. The resulting population 
numbers at age from the three VPA’s were compared 
with the true values, and the percent bias in the es- 
timates was computed. The true population numbers, 
3 catch matrices, and 3 bias matrices are given for 
the January, June and July, and December catch sce- 
narios in Tables 1, 2, and 3, respectively. 
Although each scenario had the same initial popu- 
lation, recruitment pattern, and annual catches, the 
populations became quite different, depending upon 
when the catch was removed. The earlier the catch 
was taken in the year, the lower the resulting fish- 
ing mortality and the larger the remaining popula- 
