9724 - Laserfiche WebLink

Home Browse Search

LITTLE COLORADO RIVER HUMPBACK CHUB 235 TABLE I.-Sample sizes of humpback chub by year, gear type, and location. Values for nets are the numbers of net sets; values for electrofishing are the numbers of shoreline sections (typically 100-400 m in length) sampled. The Colorado River samples were collected in the Little Colorado River inflow reach (approximately 9 Ian upstream and II Ian downstream of the confluence with the Little Colorado River). The little Colorado River samples were collected within 15 Ian of the confluence with the Colorado River. Colorada River Little Colorado River Year Electroflshing Trammel netting Hoop netting Trammel netting Hoop netting 1987 85 142 1988 179 399 1989 177 454 1990 20 142 73 356 1991 231 1,076 4 319 2,826 1992 407 847 198 3,712 1993 558 1.104 15 146 4,602 1994 37 19 2 103 3,885 1995 137 71 2,371 1996 89 122 4 1.114 1997 89 163 20 936 1998 70 42 372 27 1,435 1999 106 54 550 12 1.046 2000 225 344 284 7 2,700 2001 47 1,098 605 64 2,651 2002 168 120 199 I 2,998 2003 119 100 108 8 2.786 Total 2,303 5,231 2.139 1,494 34,413 integrated transponder (PIT) tags to provide a unique, long-term identification mark. Capture-recapture data were used in a simple, two-sample closed-population model to estimate population size (N) and two Jolly- Seber-type open-population models to generate esti- mates of capture probability (fJ), rate of population change (),.), and population size (Seber 1982; Williams et al. 2(02). We used an extension of the Jolly-Seber population model to incorporate information on age at first capture (pollock 1981) to estimate mortality, capture probability, and LCR population size with the program MARK (White and Burnham 1999). We also used methods presented by Pradel (1996) to estimate the rate of population change directly without having to estimate population size. In addition, we developed a new age-structured, open-population capture-recapture model (ASMR) similar to the Jolly-Seber models but developed specifically to address analysis needs related to this robust capture-recapture data set. The ASMR model predicts age at first capture from length, the numbers of marked and unmarked fish at risk of capture from the age-specific survival rate, and age- and time-specific capture probabilities. Details of the ASMR model are provided in the companion paper in this issue (Coggins et al. 2006, this issue). Model background.--Capture-recapture methods have rigorous assumptions and are broadly defined as "closed" or "open" depending on whether the population is allowed to change over the sample interval (open models) or not (closed models). Short- term capture-recapture studies are usually closed- population estimates and are based on the assumption that all members of the population are vulnerable to sampling and that no animals leave or enter the population between sampling occasions. Open-popula- tion models allow for changes in the population due to births, deaths, and movement into and out of the area and are generally more suited for longer-tenn studies (Pine et al. 2003). Most multiyear capture~ecapture studies use some variation of the classic open Jolly--Seber model to estimate population size, recruitment, and survival (Jolly 1965; Seber 1965; Pollock et al. 1990). A key assumption of open-population models is that every animal (both marked and unmarked) present in the population during a sampling period has the same probability of being captured. Parameter estimates from open-population models can be biased if capture probabilities are heterogeneous (pollock et al. 1990). In many fisheries applications, much of this heteroge- neity in capture probability is related to animal age. Incorporating age information into the Jolly-Seber model reduces estimation bias by accounting for this heterogeneity in capture probability (pollock 1981; Pollock et al. 1990). Closed-population mode/s.-Between 2001 and 2003, we made a series of two-pass, closed-population abundance estimates of humpback chub (TL > 150 mm) in the lower 14.2 km of the LCR during two 12- 14-d annual spring sampling trips. Population size and variance were estimated using a Chapman-modified,