Stewart: Defining migration rates of Porophrys vetulus 



479 



A Spring (May or June) 



North! (model 8) 



Soutfi (model 8) 



North (models 6. 7, 9) 



Soutfi (models 6, 7, 9) 



Proportion of population moving 



Figure 3 



Prior (horizontal lines) and posterior distributions for the movement para- 

 meters (proportion of the population moving) in the two-season models (6-9) 

 allowing movement in either the spring (May or June) or fall (October or 

 November). Solid lines indicate movement to the north and dashed lines indi- 

 cated movement to the south for model 8. Movement to the north is indicated 

 by the dashed and dotted lines, movement to the south indicated by the dotted 

 lines for models 6-9 (not separated in the legend). 



published analyses of these data sets: English sole are 

 not highly mixed across the entire coast but are also 

 not sedentary at the scale of hundreds of kilometers. 

 The approach of modeling the open coast separately is 

 supported by the observation that only rarely are tagged 

 English sole observed to move between the open coast 

 and Puget Sound or the Strait of Georgia. This result 

 may be specific to the biological and ecological habits of 

 the species but it is commonly assumed to be the result 

 for other species as well. 



Also in concordance with historical observations about 

 the seasonal and latitudinal movements of flatfish, the 

 current analysis supports models that include more 

 than just one movement per year. Movement appears 

 to be of greater magnitude in the fall, just before the 

 spawning season, but this pattern is not supported 

 when monthly hypotheses are explored. This lack of a 

 consistent pattern could be due to interannual variabil- 

 ity in spawning activity. Research shows that spawning. 



in the case of English sole, is related to temperature 

 (Kruse and Tyler, 1983; Peterman et al., 1987), as well 

 as latitude (Castillo, 1995). Because this analysis lacked 

 temperature as a covariate, there may not be adequate 

 data support for specific spring and fall movement rates 

 with the potential variation in timing of spawning ac- 

 tivity (when compared to uniform monthly movement). 

 If specific environmental information were available for 

 each year of the analysis, covariates could be developed 

 to improve assignment of the correct month for pre- 

 spawning migration to the south and post-spawning 

 migration to the north. 



Of potential importance to stock assessment is the 

 net movement to the south predicted by nearly all of 

 the models. Although this pattern does not fit the data 

 substantially better than simple diffusion (equal move- 

 ment north and south) throughout the year, increased 

 movement to the south estimated in simpler models 

 may be worth further investigation. The effect of net 



