8161 - Laserfiche WebLink

Home Browse Search

696 OSMUNDSON ET AL. be reached by most individuals in the present pop- ulation. Natural mortality probably occurs before these fish reach that point. There are anecdotal accounts of Colorado squawfish historically reaching 1,200-1,800 mm TL in the upper Colorado River basin (upstream of Lees Ferry, Arizona), a length considerably larger than that of any Colorado squawfish caught in the past 20 years (e.g., Jordan 189l; Quarterone 1993). Two explanations proposed for the absence of such individuals are that the growth rate has declined or that the survival rate is reduced. Behn- ke and Benson (1983) suggested that the extir- pation of bony tail Gila elegans has impacted the food supply of Colorado squawfish, thereby re- ducing growth. Gilpin (1993) showed with simu- lations that average and maximum sizes would sig- nificantly increase if adult survival rate was in- creased to 0.95, and he suggested that rates have declined through angling mortality. Neither hy- pothesis, however, adequately explains the dis- appearance of very large individuals. Kaeding and Osmundson (1988) concluded that slow growth in the upper Colorado River basin was an historic norm because temperature regimes have, with few local exceptions, remained unchanged and poten- tial foods, such as native Gila species and suckers, remain plentiful. Forage for young Colorado squawfish may even be greater than in the past because of the addition of non-native minnows. Also, size structure of Gilpin's simulated popu- lation with a survival rate of 0.95 indicated the largest individuals would not exceed 1,000 mm TL. We offer a third hypothesis. Large fish may have attained their size in the lower Colorado River ba- sin (downstream of Lees Ferry) and later moved upstream where they were eventually captured. Kaeding and Osmundson (1988) demonstrated that longer growing seasons and warmer temperatures in the lower basin historically provided 1.5-2.3 times the annual thermal units for growth than in upper basin reaches. Additionally, Colorado squawfish are capable of long-distance move- ments: radio-tagged adults have traversed the en- tire length of their current range in the upper Col- orado River (313 km) in less than 3 months (McAda and Kaeding 1991), and similarly long spawning migrations in the Green River have been reported (Tyus 1990). If the lower basin was once a source of upper basin large fish, blockage of upstream movement by main-stem dams and even- tual extirpation of downstream populations may explain the disappearance of very large individ- uals. Survival Our range of suitable (P < 0.05) survival esti- mates of 0.83-0.87 (1991-l994 data) is higher than the estimate of Gilpin (1993) for the Green River population (0.81), which may reflect differ- ences in environment between the two rivers. However, had the methodology used by Gilpin (1993) provided a range rather than a point esti- mate, it may well have overlapped ours, suggesting no difference in survival rates. Deviations from a stable population size in either river would affect comparability of estimates because each was based on that assumption. Our estimates of survival were also higher than that (0.65 for females; 0.80 for males) found for the piscivorous walleye Stizoste- dion vitreum in a lightly harvested population (Schneider et aI. 1977). Our method of determining survival rate was not designed to replace capture-recapture models for open populations, such as Jolly-Seber (Jolly 1965; Seber 1965) and variations thereof. It can be applied, however, when the consistent capture effort and long capture histories necessary for Jol- ly-Seber modeling (see Pollock et aI. 1990) are unavailable. Chief limitations of our method are a need for population structure data over several years and capture methods that sample various- sized fish at rates representing the actual popula- tion. Our tests indicated that population size struc- ture of Colorado squawfish 550 mm and longer did not change significantly during 1990-1995 and was similar to that in 1982. We were unable to test for capture bias in trammelnetting, but dispropor- tionate numbers of large fish were caught by elec- trofishing, consistent with known biases of that gear type (Reynolds 1983). When captures by elec- trofishing were included in the analysis, estimated survival increased, as expected (not shown). Though historic survival rates are unknown, there are new sources of mortality that may have recently lowered rates, although other sources of mortality are probably unchanged. Accidents dur- ing high water (abrasions, strandings, etc.), stress during spawning, and predation by great blue her- ons (Ardea herodias) and bald eagles (Haliaeetus leucocephalus) must have been primary sources of adult mortality. Significant change in these factors is unlikely, except perhaps the frequency of strand- ings, which may have increased because of ex- cavation of floodplain gravel pits. Colorado squaw fish are attracted to such habitats during "