WSP11169 - Laserfiche WebLink

Home Browse Search

i.!r. 't' Ii I It, ,:'... "'.;l{ j ;,' r CURRENT STOCK ASSESSMENT METHODS IN SUPPORT OF ADAPTIVE MANAGEMENT FOR GRAND CANYON HUMPBACK CHUB Recent analyses of historical data on humpback chub in Grand Canyon have caused considerable consternation because of uncertainties about the current size of the population and because of thc strong probability that the population has been declining steadily for at least a decade. Our most recent assessment models indicate that the current spawning population is probably somewhere between 2000 and 4000 age 4 and older fish, Le. suggesting this population might be considered as contributing to ESA deli sting based on the current population abundance, but has likely declined by at least 50% since 1990, Le. does not meet the stable population criterion for deli sting (Figures 1-3). We remain quite uncertain about the absolute population size because of uncertainties about whether field procedures have met some assumptions of the main method used to estimate absolute abundance (mark-recapture sampling) and because of limited sample sizes, but all the assessment methods clearly agree that the population is in decline. This includes not only the mark-recapture population estimates, but also several population trend indices based on catch-per-effort (CPUE) in sampling gear that has been fished consistently over the years (Figure 4). Only one ttend indexing method, trammel netting in the Colorado River mainstem, fails to indicate a downward trend in abundance. This is likely due to high variability in trammel net CPUE and previous sampling targeting known aggregations of humpback chub. One assessment model (called "Supertag") resulted in a considerably lower estimate for recent adult abundance (1100-1200 fish in 2001), but we now believe that estimate was biased downward because of using two inappropriate assumptions in the calculations: the population was assumed to have a stable age structure in the early 1990s, and older chubs were assumed to be equally vulnerable to sampling programs. Grand Canyon assessments and data analyses are greatly complicated by the migratory life history of the chubs that spawn in the Little Colorado River (LCR) and by inconsistency over the years in sampling relative to thc timing of the spawning migration. Older fish are over-represented in samples taken in the LCR during spawning runs, but are underrepresented in samples taken there outside the spawning season. The opposite effect occurs in mainstem sampling. Further, there are indications that older fish do not spawn every year, making them less vulnerable to sampling when sampling effort was/is concentrated in the LCR where it is easy to catch fish for marking. The Supertag method did not account for these complexities in interpretation of historical data. There are two strategic options for monitoring arid population assessment in Grand Canyon: (I) make independent population (and/or trend index) estimates each year using multiple-trip mark- recapture experiments (mark fish on successive trips and measure the proportion of the population made up by these "known" marked numbers) along with index CPUE sampling; and/or (2) use more elaborate stock assessment models to integrate current and past information into more complex estimators of current abundance. It should be noted that virtually all fisheries management programs for important harvested fish stocks are based on integrated assessment approaches, particularly considering that annual point estimates can "bounce around" a lot due to chance sampling factors so that if each estimate were takcn too seriously there would be inappropriate (unnecessary or even dangerous) management responses to those chance factors. It might make scientists more comfortable to pretend that only the most recent, independently Draft - April2l, 2003 2