|
<br />
<br />20
<br />
<br />6000
<br />
<br />COPEIA, 1996, NO. 1
<br />
<br />DOUGLAS;
<br />
<br />5000
<br />
<br />~\
<br />;.: .
<br />./
<br />. , ~-
<br />. ,
<br />,I ,
<br />I I
<br />/. I I
<br />./ 1/
<br />/ / II
<br />, .
<br />
<br />. .,.,:
<br />'/, \ ' 10
<br />. , ,
<br />. , ,
<br />, .,
<br />, I
<br />.
<br />
<br />4000
<br />
<br />
<br />c:
<br />.2
<br />;;;3000
<br />~
<br />w
<br />
<br />2000
<br />
<br />.
<br />
<br />
<br />\}>:
<br />
<br />'.
<br />~mm~e~0~~~~~~~~~~~~
<br />~ ~ ~ m ~ ~ ~ - ~ ~ ~ ~ ~ ~ ~ m m ~ ~
<br />
<br />1000
<br />
<br />o
<br />
<br />MonthlYear
<br />
<br />Fig. 4. Bivariate plot of population estimates for
<br />1991-1992 by month (solid line) and by month
<br />summed over reach (dashed line).
<br />
<br />50). The model with the second-highest crite-
<br />rion (i.e., the Null model; Mo) is also the simplest
<br />in that it presumes that all members of the pop-
<br />ulation are equally at risk of capture on every
<br />trapping occasion. Burhnam 's estimator (Mb) as-
<br />sumes capture probabilities vary with time and
<br />with behavioral effects (such as trap-happiness,
<br />trap-shyness). The last two models Gackknife
<br />estimator Mh, and Chao's Mh) accept that cap-
<br />ture probabilities vary by individual animal.
<br />
<br />Movement by season within and between reaches.-
<br />To determine extent of movement by G. 0Pha
<br />within the LCR, capture and subsequent recap-
<br />ture(s) for 1992 were compiled by reach and
<br />season (Table 3). Because these data reflect
<br />numbers of individuals tagged within each reach
<br />for a given season then subsequently recap-
<br />tured, percentages for each reach and season
<br />total 100%. Direct measurements of upstream
<br />movement by some tagged fish are provided in
<br />Table 3. For G. c)pha tagged at confluence dur-
<br />ing winter and subsequently recaptured, 49%
<br />(n = 47) were taken upstream in Powell or Salt
<br />
<br />reaches during winter/spring. Similarly, of those
<br />tagged at confluence during spring and subse-
<br />quently recaptured, 51 % (n = 96) were taken
<br />upstream during spring/summer. For Powell
<br />reach, 18% (n = 7) of recaptures initially tagged
<br />there during winter were taken in Salt reach
<br />during winter/spring, whereas 31 % (n = 59) of
<br />recaptures tagged there in spring were taken at
<br />Salt during spring/summer. Overall, 21 % of
<br />total movements in 1992 (ascertained by mark/
<br />recapture) was upstream.
<br />Elevated population estimates at confluence
<br />in Jan./Feb. of 1992 (Fig. 3B), followed by up-
<br />stream movement, argue strongly for staging.
<br />Estimates at the confluence peaked in early
<br />March then gradually decreased through June.
<br />A similar peak occurred within Powell Canyon
<br />reach in late March, extended into April, then
<br />decreased into June. Population size did not
<br />peak in Salt Canyon reach until April; estimates
<br />remained elevated through June. The last six
<br />months of 1992 were similar to that of 1991
<br />(Fig. 3), with estimated population sizes dwin-
<br />dling through late summer. However, estimates
<br />rose again in autumn 1991 but remained low
<br />during a similar period in 1992.
<br />Evidence for downstream movement is less
<br />convincing (Table 3). OfG. cypha tagged in Pow-
<br />ell reach during winter and subsequently re-
<br />captured, 21 % (n = 8) were taken at confluence
<br />in the remainder of the year. Similarly, 16% (n
<br />= 30), and 15% (n = 15) of recaptures tagged
<br />at Powell in spring and summer, respectively,
<br />were taken at confluence in the remainder of
<br />the year. At Salt, 16% (n = 33) and 7% (n = 12)
<br />of recaptures tagged in spring and summer, re-
<br />spectively, were taken in the two lower reaches
<br />over the remaining seasons. Overall, 9% of re-
<br />captures in 1992 indicated downstream move-
<br />ment.
<br />Table 3 primarily reflects population stasis by
<br />reach, particularly summer through winter. At
<br />confluence, 17% (n = 33) of individuals tagged
<br />in spring were subsequently retaken there sum-
<br />mer through winter, whereas 76% (n = 54) of
<br />
<br />TABLE 3 . ADULT Gila cyp,
<br />RIVER DURING A GIVEN SE
<br />SEASON (= RCP.SEASON)
<br />
<br />TAG REACH TAG SEASO
<br />Confluence Winter
<br />
<br />Confluence Spring
<br />
<br />Confluence Summer
<br />
<br />Confluence Autumn
<br />
<br />Powell Winter
<br />
<br />Powell Spring
<br />
<br />Powell Summer
<br />
<br />Powell Autumn
<br />
<br />Salt Winter
<br />
<br />Salt Spring
<br />
<br />Salt Summer
<br />
<br />Salt Autumn
<br />
<br />chub tagged in summel
<br />same reach summer th
<br />77% (n = 10) of chub
<br />autumn were retaken ti
<br />Powell, 23% (n = 45)
<br />spring were again re(
<br />through winter; 78% (
<br />during summer wer,
<br />through winter. In ad,
<br />those tagged in autumr
<br />same reach autumn/wi
<br />occurred at Salt, wher
<br />dividuals tagged durinf
<br />
<br />TABLE 2. POPULATION ESTIMATES GENERATED UNDER FIVE DIFFERENT MODELS (= MODEL) FOR ADULT Gila
<br />cyplza WITHIN THE LiTTLE COLORADO RIVER, FROM JULY 1991 THROUGH DECEMBER 1992. Also provided are
<br />goodness-of-fit (= CRITERION), with standard deviation of the estimate (= SO), and 95% lower and upper
<br />confidence intervals (= LOWER CI; UPPER CI). Models are defined in text.
<br />
<br />MODEL CRITERION ESTIMATE SD LOWER CI UPPER CI
<br />Pollock and Otto (Mbh) 0.61 4508 120 4330 48 II
<br />Null Model (Mo) 0.49 6793 110 6585 7017
<br />Burnham's (M'b) 0.48 8724 920 7242 10,901
<br />Jackknife (Mh) 0.42 10,444 329 9833 11,121
<br />Chao's (Mh) 0.42 8039 210 7648 8472
<br />
|