9711 - Laserfiche WebLink

Home Browse Search

14 Abundance Trends and the Status of the Little Colorado River Population of Humpback Chub 198!-2006 mark-recapture estimates in the LCR. However, given the low precision in the LCR estimates and the possible influence of migration magnitude and timing on the results of this program, it is, again, not too surprising that these assessments are not strongly correlated with the ASMR 3 results. The low preci- sion of these estimates may not permit detection of a 25% increase in adult abundance. Additionally, preliminary analy- ses of data collected during 2007 suggests that the abundance estimate for 2007 may be twice as large as the 2006 estimate (R. Van Haverbeke, U.S. Fish and Wildlife Service, oral com- mun., 2007). Though this result would provide support for the ASMR 3 results, it would also call into question the ability of the LCR program to provide a consistent meu'ic of overall population size. One would have to reconcile whether that level of change was related to a very large age class entering the sanlpled population, a larger than normal traction of the population entering the LCR during the sampling period, or some other factor. Though the GCDAMP is fortunate to have such a large mark-recapture database for these high-profile endangered animals, significant changes in sampling protocol over time continue to cause anlbiguity. As identified by Melis and others (2006), reu"Ospective analyses of the data suggest a continual updating of the adult mortality rate estimates as additional infonnation has been collected since 2000. Following addition of the 2006 data, this updating is again apparent (fig. 29). It appears that adult mortality rate estimates may be stabilizing as more data are collected, but it is difficult to be certain. The likely cause of this updating is the sampling program essen- tially having to "catch-up" following the low sampling effort during period 3. When focused analyses of this dataset began with open population models in 2000 (U.S. Geological Survey, unpub. data, 2000), there had been so little sampling in the mid- to late-1990s that the models interpreted the lack of old fish captures as a relatively high adult mortality rate. As addi- tional data were collected through a more rigorous sampling program during 2000-06, mortality rate and capture probabil- ity estimates were adjusted. The hope is that if the GCMRC continues with a fairly uniform sampling program over time, adult mortality rate will stabilize and only abundance esti- mates in the last few years of the dataset will be subject to significant updating. An additional finding, identified by Martell (2006) and in this assessment, is the major change in gear selectivity between sampling period 2 and sampling period 4. Though it is presently unknown what may have caused this change, several possibilities have been suggested. First, it is possible that elimination of the extensive trammel-netting effort during sampling period 4 may have reduced the capture probability of middle-aged fish. If this is U'ue, fitting the ASMR 3 model only to data collected in the LCR should indicate similar patterns in capture probability during both sampling periods. However, when this fit was conducted, the patterns were essentially unchanged ti-om those predicted from the entire dataset (U.S. Geological Survey, unpuh. data, 2(){)7). This is not too surprising, since fish captured in the LCR intlow reach of the Colorado River represent only about 11 % of the entire HBC mark-recapture database. Second, it is possible that reducing the use of large hoop nets in the LCR during sampling period 4 reduced the catch rate of larger fish. Though this is possible, the ASMR 3 results indicate a reduction in capture probability of mid-sized tish, rather than in the largest individuals. Finally, it is possible that sampling in the LCR only 4 months of the year during period 4, as opposed to 10-12 months of the year during period 2, may be the cause. This is certainly possible, particularly if there is some differ- ential migration timing for the middle-aged fish relative the oldest individuals. One obvious result of all this confusion is that large changes in sampling protocol should be carefully considered in light of how those changes may atTect the ability to infer population change. This is particularly true for populations that are in low abundance and exhibit low capture probabil- ity. A careful simulation of considered changes may help to expose potential problems or, at the very least, help to clarify thinking related to proposed changes in sampling protocol. Finally, those considering implementing a mark-recapture- based monitoring program should plan to expend considerable sampling effort using similar protocols for the duration of the monitoring program. The results herein support the recom- mendations of Williams and others (2001) that the objectives of the monitoring program, with regard to issues such as precision of measured quantities, should not only be clearly identified, but that the measured quantities should be directly linked to the management objectives. Estimating the Humpback Chub Growth Function Using Mark-Recapture Data Understanding the relationship between fish age and fish length is necessary to address a host of fundamental issues across a broad spectrum of fisheries management. Though this relationship is typically estimated using paired observations of individual fish age and length (Quinn and Delisa, ] 999), this often requires sacrificing the animal to obtain the age informa- tion. The TIGM and TDGM seek to obtain this information through non-lethal sampling and using information that is frequently collected in routine mark-recapture studies. Par- ticularly for endangered species such as the HBC, a non-lethal method to obtain information on growth rate is mandatory. In the postdam Colorado River, temperature is thought to be a limiting factor affecting native fish spawning, rearing, growth, and survival (Kaeding and Zimmemlan, 1983: Valdez and Ryel, 1995; Gonnan and Stone, 1999). As a result, much effort has been expended attempting to better understand how temperature affects basic functions such as growth (Clarkson and Childs, 2000; Robinson and Childs, 200 I; Petersen and Paukert, 2005), swimming ability (Ward and others, 2002), and predation risk (Ward and Bonar, 20(3). However, because of the sensitivity of HBC at larval and juvenile stages, much of the emphasis on understanding the effects of temperature