9404 - Laserfiche WebLink

Home Browse Search

each of these results are discussed, however, it is important to review population models and their assumptions. Open vs closed population models.---Modelling of capture history is defined by the idea of population closure. An open population is one in which study organisms enter and leave (via birth, death, immigration, emigration, or ontogeny). A closed population does not change composition during the course of the study (Nichols, 1992). While open populations are the norm in wildlife invesAgations, closed models approximate the short-duration realities of nature (Skalski and Robson, 1992). In fact, Pollock (1982) recommended as an ideal survey design a sequence of intense trapping sessions each followed by a longer period of cessation of trapping. Data from each session would be analyzed separately using closed models (as done herein). Survival rates derived from the time-duration between trapping sessions could then serve as input for open-population models (M. E. Douglas and P. C. Marsh, in prep.). Three assumptions are crucial to closed-population studies: Closure is substantiated; organisms do not lose marks during the course of the experiment; and all marks are correctly recorded at each trapping occasion. The most critical is the first. Closure for the duration of a trapping session allows the resulting estimate to represent a "snapshot" of the population at a given point in space and time. In the present study, sampling each month was brief, and movements between reaches were negligible during sampling. Thus, closure both by reach/ month and by month for the entire LCR is indeed supported, and the resulting population estimates are robust. Past and present population estimates in the LCR.---Population estimates for G. cypha in the LCR are presented in Table 1--5. Our May, 1992 estimate at the confluence was 1,320 adult G. cypha. This is a reduction of 27% and 54%, respectively, from estimates of 1,800 and 2,900 individuals in May of 1987 and 1988. An estimate for the entire 14.9 km length of the LCR during May of 1992 was 4,346 (summed estimate for the three reaches = 4,602). This contrasts with the estimate of 25,000 chub in 1989. The best-fitting population estimate for our entire 19-month study (4,508 individuals) was obtained using Pollock and Otto's estimator (Mbb). This model is one of the most realistic and useful for a mark-recapture experiment, in that it allows for individual variance in behavioral response to capture (Otis et al., 1978). Its estimate is larger than two average estimates for the 19-month study (i.e., 2,992 (monthly summed over reaches) and 2,434 (monthly for LCR)]. Although results from a model utilizing 19-months of data should be superior to an average of those data, any such long-term estimate must be viewed skeptically, given the violations of demographic and temporal closure mentioned earlier. 9