9588 - Laserfiche WebLink

Home Browse Search

survival relationship was also different as the best fit model included a linear term for the covariate fish TL but not a quadratic term. An exploratory analysis revealed that an added quadratic term was positive, which indicated the slope continued to rise, rather than decline, unlike survival relationship plotted for 2000-2003 data. Mean probabilities of capture for fish used in the above survival analyses were higher in the 2000 to 2003 period compared to the 1991 to 1999 period. Probabilities of capture.--Capture probabilities for Colorado pikeminnow generated from abundance estimation data also demonstrated a strong quadratic effect of size (parameter values and their SE's for TL and TL2 terms are TL = 0.2392, SE = 0.0845; TLZ = -0.1904, SE = 0.0526; intercepts for the individual reaches and their SE's are in Table 8). This is because small and large fish had relatively low capture probabilities and fish from about 500 to 600-mm TL had the highest ones (e.g., Fig. 10, Yampa River used as representative example of variation). It should be noted that estimated probabilities of capture are potentially a function of fish abundance as well as other factors (e.g., behavior or habitat use) that may make fish in a certain size class more or less available for capture. Thus, the most common size class in a population does not necessarily have the highest probability of capture. An intuitive explanation to understand how capture probabilities are used to generate abundance estimates is to simply divide the number of animals in the capture sample by the probability of capture, given the unlikely but simplifying assumption that all animals at risk of capture in the population have identical probabilities of capture. For example, if 100 animals are captured in a sample and the animals have a known and constant probability of capture of 0. 10, there must be 1,000 animals in the population (100/0.10 = 1,000). In this example, each fish in the capture sample contributes equally to the abundance estimate. This assumption of non-heterogeneity among individual capture probabilities is 32