Legault and Ehrhardt: Correcting annual catches from seasonal fisheries for use in virtual population analysis 
287 
Figure 6 
Monthly observed and corrected catches from an initial population of 10,000 fish 
when the fishery operates late in the year (top panel). The resulting population 
decline by month under the two conditions is shown in the lower panel. 
derestimated the true population size. The June and 
July observed catches produced low bias, demonstrat- 
ing the validity of Pope’s cohort analysis approxima- 
tion (Pope, 1972). The increasing bias at age for all 
three scenarios with corrected catch data was due to 
a choice of a small constant N K+1 value for all ages 
and years. This method of choosing N K+1 is not rec- 
ommended during normal application of the algo- 
rithm but was done to demonstrate the increased or 
decreased bias relative to observed catches, depend- 
ing upon the timing of the catch. Much better guesses 
of N k+1 for the corrected catch would be the results 
of VTPA with the observed catch matrix. In this case 
the recorrected catches will usually be so similar to 
the corrected catches that no further iterations are 
required. 
