9484 - Laserfiche WebLink

Home Browse Search

December 2002 FLOW-SEDIMENT EFFECTS ON RIVERINE FISH 1500 1200 `. E ? 900 CU 600 cu 300 0stratum9 tL to stratum S 01 1995 1996 1997 1998 1999 2000 2001 Year FIG. 7. Peak discharge (highest mean daily discharge of the year) in strata 8 and 9 during the DFS monitoring years of 1996-2001. The antecedent peak flow of 1995 is also shown. Horizontal lines indicate thresholds for widespread bed mobilization in stratum 8 (upper line) and stratum 9 (low- er line). Because total DFS can vary by rock size, relative DFS (scaled to median rock diameter) was examined as a means to isolate changes in DFS attributable to changes in fine sediment accumulation alone. By fall 2001, 4 yr following the relatively high flows of 1997, mean relative DFS had declined from 1.4 to 0.5 rock diameters (rd) in runs of stratum 8, and from 1.9 to 0.5 rd in stratum-9 runs. Relative DFS also declined in riffles, though values were generally higher than in runs. In 2001, relative DFS in stratum-8 riffles averaged 0.7 rd; in stratum 9, 1.1 rd (Fig. 6). Relative DFS was also significantly (P < 0.05) correlated with peak dis- charge in both runs and riffles of stratum 9 and runs of stratum 8. No correlations were found between total. or relative DFS and the mean August to October base flow discharge. Frequency of substrate-mobilizing flows.-Because deposits of fine sediment were found to accurrttlate in the bed over time, we were most interested in potential changes in the mean interval between flushing events. Pitlick and Van Steeter (1998) found that the minimum flow necessary to produce widespread bed mobilization in strata 7, 8, and 9 corresponded with the bankfull flow, and results from our study indicate this to be the case for strata 10 and 11 as well (see Pitlick and Cress 2000). Because of geomorphic similarities, we assumed this is also true for stratum 6. However, downstream of the Dolores River confluence (strata 1-5), discharge thresholds for widespread bed movement have not yet been determined and these strata were therefore not included in our analysis of flushing-flow frequencies. Relatively large blocks of years are required for cal- culating mean recurrence intervals, and we were for- tunate that gauge records for annual peak flow for strata 6-11 were available back to 1908. The early period 1733 was characterized by a series of wet years (1908-1930) followed by dry years (1931-1940); the latter period, also had wet (1982-1987) and dry periods (1988-1992) but were of shorter duration (N. Doeskin, personal communication). For our comparison of mean recurrence intervals, we assumed that thresholds for bed-mobilizing discharges were similar between historic and recent periods. How- ever, because these calculated thresholds were based on recent channel and substrate characteristics, we preface our results with the caveat that recent threshold discharges are only approximations of historic thresh- olds. If the bankfull channel was historically wider or the substrate coarser than today, historic thresholds for bed mobilization would have been higher, likely length- ening average recurrence intervals. Although we know that wetted area of the main channel at base flow has narrowed by a mean of 15% in strata 8-10 since 1937 (Van Steeter and Pitlick 1998), historic bankfull chan- nel width and historic substrate size distributions are unknown. The frequencies of flows necessary for initial motion and for widespread mobilization of the bed in strata 6- 11 were evidently much lower during recent years than during the early half of the century prior to most river regulation. During the preregulation period of record (1908-1942), mean recurrence intervals of widespread bed mobilization (assuming current discharge thresh- olds) were 1.2-1.4 yr in strata 9-11 (upstream of the Gunnison River confluence); this increased to 2.7-3.6 yr in the "post-development" period (1966-2000). Be- tween the inflows of the Dolores and Gunnison rivers (strata 6-8), the mean interval during 1908-1942 was 1.4-2.7 yr and during 1966-2000, 4.6-13.5 yr (Fig. 8). Most of the difference between periods can be attri- buted to water development, although the numbers may also have been influenced by climate to some unknown degree (snowpack records for the Colorado basin did not begin until around 1950). DISC U S S10N Longitudinal trends Abundance of adult pikeminnow, and other main- channel native fish, generally declined in a downstream direction. Differences in food resource availability can explain much of this observed pattern. Our study iden- tified three important patterns related to longitudinal changes in lower trophic level structure: (1) benthic biomass in both riffles and runs decreased downstream, (2) riffle substrates contained more biota and detritus than run substrates of a similar water depth, and (3) the total surface area of riffles (the most productive habitat type) decreased downstream. Principal component analysis showed a strong pos- itive relationship between biomass of both primary pro- ducers and invertebrate consumers and the degree to which the substrate was free of fine sediment. The