9711 - Laserfiche WebLink

Home Browse Search

10 Abundance Trends and the Status of the little Colorado River Population of Humpback Chub 1989-2006 number of categories equal to the number of ages. Assigned age for a single fish having length in bin l is therefore a m~ltinomial random vmiable with probabilities Pea 17) with a smgle draw. The multinomial random number generator within program R (R Development Core Team, 2006) was used to randomly assign age to tagged fish. Age-at-recapture was calculated as the sum of age-at-tagging and time-at-large. For each resulting dataset of captures- and recaptures-at-age, adult (age-4+) abundance and recruitment was estimated using ASMR 3. Additionally, ADMB was used to compute the 95% profile confidence interval for adult ablmdance and recruit- ment. This procedure was repeated to generate and analyze 1,000 datasets (i.e., Monte Carlo trials). Resu Its 2006 Humpback Chub Assessment Update with Refinements Index-Based Assessments Between 1987-99 and 2002-present, the Arizona Ganle and Fish Department sampled HBC using hoop nets in the lower 1 ,200-m section of the LCR. Examination of this catch-rate index suggests that abundance of both subadult (l50-199-mm TL) and adult (2:200-mm TL) HBC declined between 1987 and 1992 and remained relatively constant through much of the 1990s (fig. la). Since 2003, this index suggests a slight upward trend in the abundance of subadult fish. Note that several data points in the index are shifted slightly relative to those reported by Coggins and others (2006a). This adjustment resulted from additional standm'd- ization of the data used to construct this index (David Ward, AIizona Game and Fish Department, oral commun., 2007). The index of a!lult abundance in the LCR inflow reach of the Colorado River indicates a similar trend in adult fish (2:200 mm, fig. Ib). In general, this index shows a stable to declin- ing trend in relative abundance through the 1990s with a slight indication of increased abundance in most recent years. All monthly trammel-net samples from the LCR inflow reach for 1990-2006 are presented in figure 1 b. However, only samples from 1990-93,2001, and 2005--06 (i.e., dark circles in figure 1 b) represent robust sampling coverage throughout the el;tire reach. Annual sample sizes in 1994-2000 and 2002--03 (i.e., hollow circles in figure I b) were between 2% and 50% of the 1990-93 average sample size, and in some years effort was focused near the LCR confluence where HBC density may have been highest. Thus, the 1990-93, 200 I , and 2005--06 data are likely to best depict the overall trend of relative abundance within this reach. Simple linear regression analyses provide estimated slopes that are not significantly different from z.ero (p = 0.16 for all data and p = 0.26 for the preferred data; fig. Ib). Tagging-Based Assessments As desclibed above, the data required for the ASMR models are numbers of fish marked and recaptured each year and for each age. For the results contained in this section, all ages were assigned based on the standard von Bertalanffy growth curve as described in Coggins and others (2006b). With that in mind, examination of the age distribution of fish marked and recaptured since 1989 provides insight into the trends in sampling effon and also provides important infonna- tion related to HBC survival (fig. 2). The top panel of figure 2 shows the numbers of newly tagged fish by age and year. The total number of fish marked annually is influenced by both trends in sampling effort and numbers of unmarked fish alive. The most consistent period of sampling has been 200 I-pres- ent; approximately 1,100 fish have been tagged annually. Because a large fraction of the population was tagged in sampling periods 1 and 2, the majority of fish tagged in recent years are young fish and the number of new fish tagged each year declined with fish age. The bottom panel of figure 2 represents the numbers of fish of each age recaplllred each year. The same patterns related to sampling eflort are evident, but there are some very interesting patterns that result from the high sampling eflort in the early to mid-1990s (fig. 2). For example, in 1995 a total of 1,244 HBC were collected and 902 of these fish had been marked in previous years. This pattern is evident for several years of data, indicating that the high sampling efTort in the early 1990s resulted in marking upwards of 70% of the HBC population. The sampling program since 2000 (period 4) has been plagued by the low sampling effort in the mid- to late-1990s. For example, figure 2 shows that because few age-3 to age-5 fish were tagged during period 3, there were few age-8 to age-I 0 fish recaptured in the early 2000s. This contributes to the "spoon" shapes in the lower panel of figure 2, where there were relatively large numbers of fish younger than age-lO recaptured in period 4, lower catches of age-l 0 to age-IS fish, and relatively stable numbers of older than age-15 fish. Another finding is the extreme longevity of HBC. This is most evident by examining the number of HBC of each age marked in each year and recaptured in subsequent years (fig. 3). Figure 3a shows the number of HBC of each age marked in 1989-92 and recaptured in subsequent years. This figure shows that a remarkable number of old fish (>age-15), first tagged in the eady 1990s, continued to be recaptured into 2006. This slow decay pattem of tagged fi.sh demonstrates the low mortality rate suffered by older HBC. Closed Population Models The time series of abundance estimates for HBC greater than or equal to 150 mm TL in the LCR during spring indicate