8161 - Laserfiche WebLink

Home Browse Search

690 OSMUNDSON ET AL. dom-number sequences. Mean and variance of 7-year-old fish as determined by scales was used to generate an initial size distribution. For a sim- ulated population of 103 fish in each age, lengths at each age were calculated based upon simulated distributions of sizes in the previous age-class. Range of ages expected for given lengths was also calculated from these simulations. To aid in constructing growth simulations, sta- tistical tests were conducted to determine if growth in one year was related to growth the following year (i.e., whether individual fish consistently grew more than the average) and whether mean growth in some years was greater than in other years. Contingency table analysis was conducted with fish captured in multiple years to test the re- lationship of individual growth between years. We used analysis of covariance (initial length as co- variate) to test for differences among years for 1991-1995. Survival Estimates.-Survival rates were estimated for Colorado squawfish 550 mm and longer. We as- sumed that survival rates of juveniles and suba- dults were probably different from adults but had no way of assessing them. Estimates were limited to upper-reach adults (Figure 1) because longitu- dinal variation in size distributions, coupled with unequal sampling effort between reaches, could have biased whole-river length distributions. The estimates were limited to fish 550 mm and longer because Seethaler (1978) reported immature fish as large as 503 mm TL. Estimates were made with a modified Chapman-Robson approach (Seber 1982), in which survival is based on declining numbers of individuals by age in the population. Because captured fish were not reliably aged, survival estimates were made by comparing lengths of captured fish with theoretical length dis- tributions under the following assumptions (con- sistent with Ricker 1975): (1) survival rate is uni- form with age over the range of ages examined; (2) survival rate is uniform over time and does not vary among years; (3) recruitment to the first size examined is equal among years; and (4) the sample is uniformly drawn from all ages-lengths consid- ered (i.e., there is no effect of gear selectivity). Theoretical length distributions (termed stable length distributions) were calculated from age dis- tributions, assuming constant survival rates and constant recruitment into the youngest age-class (termed stable age distributions), and the gener- ated age-length distributions (Figure 2). The sta- - *' - A >- o C 0> :J C- O> L. U. Age (years) - E E - B ~ +oJ 0) C 0> CCS +oJ t2 1ttH t Age (years) - *' - >- o C 0> :J c- O> L. U. Length (mm) FIGURE 2.-Calculation of stable length distribution (C) from stable age distribution (A) and age-length dis- tribution (B). The stable age distribution assumes a con- stant survival rate and rate of recruitment into the youn- gest age-class. Age-length distributions are calculated from distributions of growth increments based on total length. The stable length distribution was calculated by using relative number of fish for each length as deter- mined by age-length frequency distribution and relative number of fish by age.