BIOSTATISTICAL ANALYSIS 19 



and pattern of movement can be forecast. The situation is analogous 

 to the famiUar one of weather prediction. Ahhough it is not possible 

 to forecast the exact day-to-day variations in the weather in a 

 particular region, we can forecast the general climate in an area. 

 Throughout this report we will be more concerned with the "climate" 

 than with the "daily" fluctuations. 



Typical time-of-entry curves for each of the four species were 

 constructed from two main sources of data. The first of these is the 

 troll catches off the west coast of Vancouver Island and the second 

 the seasonal distribution of net catches in the San Juan Island area. 

 The catch per unit of effort in the offshore troll fishery on a given 

 day is a measure of the population abundance in the area. The 

 relative index of population abundance can be converted into num- 

 bers at the end of the season when the total catch and escapement 

 have been tallied. The San Juan catches were lagged from three- 

 and-one-half days to several weeks, depending on the species and 

 the time of year, to give a similar index of relative abundance. 



The "standardized" entry patterns shown are those that were 

 used in the simulation program. The entry patterns represent the 

 daily entry of fish into Area 1 or Strait area (see Fig. 9). The timing 

 of the fish's appearance in the inner fishing areas (Areas 2, 3, 4, and 

 5) is determined by the rate of movement along the various parts 

 of their migratory route. As fish travel through the fishery, there 

 is also a tendency for the temporal abundance pattern to become 

 more spread out (less peaked). 



The rate of travel is usually sharply reduced in the estuarine 

 areas just off the natal spawning streams. The exact length of the 

 delay time may vary considerably from species to species and between 

 the early and late races or parts of a run for one species. Delay 

 time is also affected by the stream flow conditions. The migratory 

 times used in the simulation model are shown in Figs. 10, 11, 12, 

 and 13 for sockeye, pink, silver, and chum salmon, respectively. 

 The data were obtained, in part, from published results of tagging 

 experiments conducted by the International Pacific Salmon Fisheries 

 Commission, the Washington State Department of Fisheries, and 

 Fisheries Research Board of Canada. 



The tagging data were supplemented with an analysis of the 

 timing of the catch curves in the various areas. From each catch 

 curve it is possible to estimate the time of peak abundance of the 

 fish in an area. The time lag between the peaks of abundance in 



