30 SALMON GEAR LIMITATION 



the basis of the sockeye run and catch. On August 30. the regulation 

 was switched to pink sahiion and continued until October 18. After 

 October 18, it was assumed that the fishery would operate on a four- 

 day week for the remainder of the season. In addition to this rule 

 which permitted the desired escapement, a second decision was made 

 which could close either the American or Canadian fisheries inde- 

 pendently. If the Canadian catch exceeded the American catch, the 

 Canadian fisheries were closed. However, if the American catch ex- 

 ceeded the Canadian catch by over 5 per cent of the total catch, 

 the American fishery was closed until the Canadian fishery caught up. 

 The extra 5 per cent for the American fishery was allowed because 

 the fish are available to the Canadians in the Fraser River after the 

 bulk of the run has passed through the American fishery. The Ca- 

 nadian fishery has an opportunity to catch up if it is slightly behind. 

 These regulations were imposed for both sockeye and pink salmon. 

 The number of days actually fished during the season was recorded 

 by the computer for each week. The actual fishing patterns produced 

 by this regulatory scheme were fairly close to those observed in the 

 past. 



Validation of the Model 



One way of testing the validity of a complex and detailed simu- 

 lation model is to put into the model an actual entry pattern and gear 

 distribution and compare the output generated by the model with the 

 actual catch data. The sockeye salmon are an especially good species 

 for use in such hindcasts because of the wealth of good statistical data 

 available for the fishery. One such validation of the simulation model 

 was carried out using a more detailed description of migratory routes 

 than shown in Fig. 10. A schematic outline of the biological and 

 gear sectors in this simulation model (which will be referred to as 

 simulation model No. 1 to avoid confusion) is shown in Fig. 15. 

 Results using model No. 1 were tabulated by month for each of the 

 subareas. A comparison between the output generated by the com- 

 puter and the actual catch data is shown in Fig. 16. The greatest 

 discrepancy between the computer results and the actual observed 

 results is in the September catches by the Canadian fleet off the mouth 

 of the Fraser River and in the South Vancouver Island area. (This 

 area included statistical area 17 which contains the high September 

 catches of the Canadian gill net fleet. ) In later runs and in simulation 

 model No. 2 this deficiency in the model was corrected by decreasing 



