FISHERY BULLETIN: VOL. 84, NO. 2 



Table 6.— Escapement estimates obtained by correcting for differential 

 catchability of fresh and decayed carcasses for three methods of estimating 

 escapement. For each correction, the ratio of the average fresh to decayed 

 catchabilities that was assumed to obtain the corrected estimate is given. 



Assumed fresh/decayed 

 Catchabilities 



Corrected escapement 



Jolly-Seber Manly and Parr Method 3 



Table 7.— Escapement estimates obtained by simulation of reduced sampling effort 

 for three methods of estimating escapement. For each simulation the fraction of 

 decayed top carcass captures and the fraction of decayed bottom carcass captures 

 ignored is given. 



Fraction of decayed 

 Carcass captures ignored 



Escapement estimate 



method are less biased than those obtained by the 

 other two methods. 



DISCUSSION 



The estimates of total immigration are all remark- 

 ably close to the weir count. This accuracy is even 

 more remarkable in light of the fact that CDF&G 

 has traditionally used a correction factor of 0.95 to 

 account for an estimated 5% of the spawning 

 grounds that is not sampled on Bogus Creek. Inclu- 

 sion of this factor brings all of the estimates to 

 within 1.4% of the weir count. Since the third 

 method provides a high degree of precision (Table 

 2) at much less sampling cost, it is preferable over 

 the other two methods. We can compare the preci- 

 sion of the third method with the Jolly-Seber and 

 Manly and Parr methods by comparing the standard 

 error estimates that are available for those two 

 methods (Table 1). The Jolly-Seber method is more 

 precise in estimates of AT, B, and <t>. This is expected, 

 since both the Manly and Parr method and the third 

 method use fewer individuals in estimates than the 

 Jolly-Seber method does. However, the precision of 

 the third method is more than adequate: 95% con- 

 fidence intervals are +5.3% and -5.5% of the 

 escapement estimate. 



The detected violations of assumptions, age- 

 dependent catchability and heterogeneity of capture 

 probabilities and survival, are those that would be 

 expected on the basis of physical considerations. 



Survival of carcasses is a function of two processes: 

 fresh carcasses being removed by carnivores, and 

 old carcasses decaying and becoming buried in the 

 stream bed. Rates of disappearance could thus be 

 affected by condition, and therefore age and size, 

 of carcasses. Older carcasses and smaller carcasses, 

 which decay more quickly and are buried more easily 

 than larger carcasses, would be expected to have 

 lower survival rates. 



Catchability is a function of both visibility and loca- 

 tion, both of which would be expected to vary with 

 condition and size of carcasses. This causes two dif- 

 ferent types of problems: age-dependent catchability 

 and size-dependent catchability. Shiny, fresh car- 

 casses were much more visible on the bottom of the 

 stream than the brown, decayed carcasses. Car- 

 casses on the stream surface were in general visi- 

 ble regardless of their condition. Since carcasses lost 

 their high visibility in about a week, no marked car- 

 casses will be in this high visibility category, and un- 

 marked carcasses will on the average be more catch- 

 able than marked carcasses. This can be thought of 

 as age-dependent catchability. Size-dependent catch- 

 ability stems from the fact that decayed individuals 

 that were large were more visible than those that 

 were small. This can be viewed as capture heter- 

 ogeneity. Since fresh fish were high visible regard- 

 less of their size, this heterogeneity existed only 

 among decayed individuals. Based on these con- 

 siderations we would expect catchability to vary 

 with age and size according to Figure 4. 



268 



