Downstream Effects of Flaming Gorge Reservoir on the Green River, Colorado and Utah

Downstream effects of Flaming Gorge Reservoir on the Green River, Colorado and Utah

E. D. ANDREWS U.S. Geological Survey, Water Resources Division, Box 25046 M.S. 413, Denver Federal Center, Denver, Colorado 80225

ABSTRACT of sediment in the bed, banks, or flood plain, the mouth. Sayre and Kennedy (1978) attributed average hydraulic characteristics width, depth, this channel disequilibrium, in part, to res- The Green River is one of the principal velocity, roughness, slope, and channel pattern, ervoirs. tributaries in the Colorado River basin and through a reach of channel at a given discharge, The opportunity to study the downstream ef- drains 44,700 rii2 in Wyoming, Colorado, will be nearly constant. Such river channels are fects of reservoirs has been limited by two pri- and Utah. Since October 1962, flows of the in quasi-equilibrium. Although this condition mary difficulties. In most alluvial rivers, the Green River have been regulated by Flaming may not be exact, the uniform elevation of flood mean annual sediment discharge is small com- Gorge ReservoLr, which is located 412 river plains along many rivers indicates that quasi- pared to the quantity of sediment stored within a miles upstream from its confluence with the equilibrium is approached for periods of several reach of a few hundred channel widths in length. Colorado River. Mean annual runoff has not decades to centuries. An appreciable and per- Consequently, the annual sediment deficit been affected by the reservoir. The duration sistent change in the water discharge, sediment caused by reservoir storage also is small com- of the relatively large discharges that trans- load, or sediment size will cause a disequilib- pared to the volume of sediment available for port most of the annual sediment load, how- rium between the quantity of sediment supplied transport. Major channel adjustments may occur ever, has decreased significantly. As a result, to the reach and the quantity transported out of only after an appreciable change in quantity of the mean annual sediment discharge has de- the reach. Thus, the hydraulic characteristics sediment within the channel. This condition creased by 54% Ho 3.21 x 106 tons from 6.92 x will readjust, so as to attain a new quasi-equilib- may require several decades to develop, depend- 106 tons at the Jlensen gage located 105 river rium. ing upon the distance downstream from the res- miles downstream from the reservoir and by The quasi-equilibrium adjustment of a river ervoir. Reservoirs with storage greater than 6 6 48% to 8.83 x 10 tons from 17.0 x 10 tons at channel located downstream from a reservoir 1,000,000 acre-feet have been built only within the Green River, Utah, gage located 290 river typically will be altered to a substantial degree the past 50 yr or so. The vast majority have been miles downstresun from the reservoir. Sedi- by the storage of sediment in the reservoir and constructed since 1950. The far-downstream ef- ment supply to i:he channel equals the annual the decrease of river discharge, especially the fects of these reservoirs upon channel equilib- transport within a relatively short distance, peak flows. The nature of the disequilibrium will rium are only now becoming evident. 68 river miles, downstream from the reser- vary longitudinally downstream in magnitude, The second difficulty is closely related. Com- voir. Downstream from river mile 166, the direction (sediment surplus or deficit), and dura- prehensive, long-term records of river flows and supply of sediment from upstream plus tribution. Williams and Wolman (1984) described sediment transport at several locations down- tary inflow exceeds the transport of sediment the complex channel changes that have occurred stream from the reservoirs, as well as on major 6 by ~5.4 x 10 ions per year on an average. downstream of 21 dams. Degradation of the riv- tributaries, usually do not exist (Petts and The quasi-equilibrium that appears to have erbed immediately downstream of a reservoir Lewin, 1979). Thus, the information req uired to existed prior to the reservoir no longer occurs has been the most commonly studied channel describe the characteristics of river flows and along a majority of the Green River. impact (for example, see Hathaway, 1948; sediment transport prior to a reservoir and the In response to the reduced peak dis- Komura and Simons, 1967; and Petts, 1979). change since regulation is usually unavailable. charges, the fcinkfull channel width of the These investigations usually have been limited to Several of the longest records of daily sedi- Green River has decreased by -10%. Ad- a reach extending only a few hundred channel ment transport that exist for North American justment of the channel to decreased peak widths downstream from the dam. Investiga- rivers have been collected at gaging stations in flows and altered sediment loads is nowhere tions into the downstream effects of reservoirs the Colorado River basin. As of IS'83, the complete. At piresent, it appears that a cen- have rarely considered channel adjustments that length of most of these records of sediment tury or more will be required for the Green might occur downstream from the confluence of transport was almost 40 yr. Large reservoirs River to adjust to the effects of Flaming the first major tributary, although there is evi- were constructed on each of the three major Gorge Reservoir. dence that the changes are quite significant. headwater tributaries during the early 1960s. Lawson (1925) described extensive channel ag- The pre- and post-reservoir periods of record INTRODUCTION gradation in a reach of the Rio Grande begin- therefore are now ~20 yr. This investigation ning > 100 mi downstream from Elephant Butte considered the downstream effects of Flaming Alluvial channels adjust over a period of Reservoir. Degradation and aggradation of the Gorge Reservoir on the Green River. years, so that the sediment supplied to the chan- Missouri River channel have become serious Long-term channel change is of particular in- nel is transported with the available discharge. problems in the nearly 800-mi reach between terest for the Green River because of the impacts When there is no net accumulation or depletion the last downstream reservoir and the river it may have on the survival of several species of

Geological Society of America Bulletin, v. 97, p. 1012-1023, 10 figs., 3 tables, August 1986.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/97/8/1012/3444989/i0016-7606-97-8-1012.pdf by guest on 30 September 2021 DOWNSTREAM EFFECTS OF RESERVOIR ON RIVER, COLORADO AND UTAH 1013

tober 1962. Fontenelle Reservoir, also located on the main-stem Green River, is the second largest impoundment in the Green River basin and was completed in April 1964. Tributaries to the Green River have numerous small im- poundments, especially in their headwaters. Ex- cept for the Duchesne River, however, the tributaries are generally free flowing and unregulated at present (1985). Several reservoirs have been proposed and are being considered for the principal tributaries. The location of the long-term gaging stations in the Green River basin are shown in Figure 1. Water discharge has been recorded daily at most gages for several decades. Extensive records of suspended-sediment concentration also have been collected at the gages shown in Figure 1. At those gages where the suspended-sediment concentration has been measured daily for five years or more, the mean annual sediment discharge was determined by averaging the series of annual values. At the other gages, the relation between sediment discharge and water discharge (sediment-discharge rating curve) was computed from a least-squares regression of the log- transformed values. The mean annual sediment discharge was computed by the sediment- discharge rating curve and flow-duration method. The water- and sediment-discharge records were compiled to determine mean annual water and sediment budgets prior to construction of Flaming Gorge Reservoir, for three reaches of the Green River. Reach 1 extends downstream from Flaming Gorge Reservoir to the Jensen gage. Reach 2 extends downstream from the Jensen gage to the Ouray gage. Reach 3 extends downstream from the Ouray gage to the Green River, Utah, gage. The mean annual inflow and outflow of water and sediment to each reach are summarized in Table 1. The periods of record used to Figure 1. Locations of principal water and sediment gaging stations. compute the mean annual values are indicated. The mean annual water and sediment discharges at the Jensen and Green River, Utah, gages since fish (Miller and others, 1982). These fish have HYDROLOGY OF THE GREEN regulation by Flaming Gorge Dam began were adapted to a riverbed with areas of both sand RIVER BASIN computed from the measured daily values. and gravel. Gravel-bed parts of the channel, es- Sampling of suspended-sediment concentration pecially riffles, are essential for spawning. Sand- The Green River drains -44,700 mi2 along in the Green River at the Ouray gage was dis- bed parts of the channel, especially backwater the west flank of the Rocky Mountains in continued in September 1966, 4 yr after the be- along banks or in the lee of bars, are utilized by Wyoming, Colorado, and Utah (Fig. 1). It is one ginning of reservoir regulation. This record was the juvenile fish. Either aggradation that results of the principal tributaries in the Colorado River extended by correlating the 4 yr of measured in an all sand-bed channel or degradation that basin. The main-stem Green River originates in sediment loads with the sediment loads deter- results in an all gravel-bed channel would elimi- the Wind River Mountains of Wyoming and mined for the Green River, Utah, gage. The nate essential habitat. Although the habitat of flows southerly to its confluence with the Colorado coefficient of determination for the relation is native fish is sensitive to channel disequilibrium, River near Moab, Utah. The principal tributar- 0.95. appreciable channel aggradation or degradation ies of the Green River are the Blacks Fork, For some tributaries, the mean annual contri- also will have serious consequences for engineer- Yampa, Duchesne, White, Price, and San Raf- butions of water and sediment to the Green ing structures and the use of land adjacent to the ael Rivers. Flow in the Green River has been River were computed from daily measurements river (Lane, 1955). regulated by Flaming Gorge Reservoir since Oc- collected after regulation of flow by Flaming

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/97/8/1012/3444989/i0016-7606-97-8-1012.pdf by guest on 30 September 2021 1014 E. D. ANDREWS

TABLE 1. SJMMARY OF MEAN ANNUAL INFLOW AND OUTFLOW OF WATER AND SEDIMENT TO SELECTED REACHES OF THE GREEN RIVER BEFORE AND AFTER THE CONSTRUCTION OF FLAMING GORGE RESERVOIR

Period of Drainage Pre-reservoir Post-reservoir sediment record area Mean annual Suspended Mean annual Suspended (water years) (mi2) discharge sediment discharge sediment (ft3/!) (tons x lo'/yr) (ft3/s) (tons * lO6/^ )

Reach 1

Inflow Green River at Greendale, Utah 1956-19S9 15,090 2,130 3.60 2,130 0 Little Snake at Lily, Colorado 1958-1964 3,730 571 1.29 571 1.29 Yampa near Maybell, Colorado 1950-1958 3,410 1,470 .388 1,470 .388 Red Creek near Dutch John, Utah 1971-1977 140 7.76 .0844 7.76 .0844 Ungaged area 3,030 141 1.55 20 1.55 Total 25,400 4,320 6.92 4,199 3.31 Outflow Green River at Jensen, Uuh 1947-1979 25,400 4,320 6.92 4,199 3.21

Reach 2

Inflow Green River at Jensen, Utili 1947-1979 25,400 4,320 6.92 4,199 3.21 White River at mouth 1974-1982 5,120 632 1.67 654 1.67 Ungaged area 4,980 498 4.18 741 4.18 Total 35,500 5,450 12.8 5,500 9.06 Outflow Green River at Ouray, Uuh 1951-1966 35,500 5,450 12.8 5,500 6.61

Reach 3

Inflow Green River at Ouray, UtiJi 1951-1966 35,500 5,450 12.8 5,500 6.61 Price River at Woodsidi, Utah 1975-1982 1,540 105 2.18 105 2.18 Ungaged area 3,560 275 2.04 -200 2.04 Total 40,600 5,830 17.0 5,805 10.8 Outflow Green River at Green Rivi:r, Utah 1944-1982 40,600 5,830 17.0 5,580* 8.83

*-200 fl^/s are diverted upstream from the gage.

Gorge Reservoir had begun. These tributaries, shown in Figure 2. Immediately upstream from CHANGES IN CHANNEL however, have ne t been affected significantly by Flaming Gorge Reservoir, the unregulated mean EQUILIBRIUM DOWNSTREAM OF any regulation or impoundment. Mean annual annual flow was 1,575 ft3/s or 27% of the total FLAMING GORGE RESERVOIR values shown in Table 1, therefore, are a reason- basin outflow. The mean annual sediment dis- able estimate of inflow and outflow of water and charge at this gage, however, was only 0.37 x The mean annual inflow and outflow of sediment to reaches of the Green River prior to 106 tons or 2.2% of the basin outflow. Similar water and sediment to three reaches of the appreciable water-resources development. The contrasts are shown by the comparison between Green River after the construction of Haming estimated contributions from ungaged areas the water and sediment discharges of the Green Gorge Reservoir are summarized in Table 1. were computed by assuming that inflow and River basin and the quantities measured by the Comparing the pre- and post-reservoir ]periods, outflow to the «aches were equal. In terms of farthest upstream gages on the Yampa and mean annual water discharge at the Green River the sediment budgets, this assumes that the White Rivers. The combined mean annual dis- near Jensen, Utah, gage has not been affected. reaches were in long-term quasi-equilibrium charge at the 3 farthest upstream gages is 2,680 Mean annual discharge of the Green River near prior to flow regulation. The presence of exten- ft3/s or 46% of the basin discharge, whereas the Jensen, Utah, gage was 4,200 ft3/s from 1963 to sive flood plains along most alluvial reaches sediment contribution is only 0.44 x 106 tons or 1981 compared to 4,320 ft3/s from 1.947 to supports this assu mption. 2.6% of the basin-sediment discharge. The con- 1962. The range of daily flows throughout the tributing drainage area is -27% of the total year, however, has been altered considerably SOURCE AREAS OF RUNOFF basin. and will be discussed below. The mean annual AND SEDIMENT The large sediment-contributing areas in the sediment discharge at the Jensen gage has de- 6 Green River basin are indicated, similarly, in creased by 54% to 3.21 x 10 tons. Mean annual Water and sediment are not contributed to Figure 2. All tributaries joining the Green River inflow of sediment to the river reach down- the channel network uniformly across the Green downstream from the gage near Green River, stream from Flaming Gorge Reservoir to the 6 River basin (Iorns and others, 1965). Further- Wyoming, supply relatively large percentages of Jensen gage is 3.31 x 10 tons. Thus, there exists more, the principal source areas of water and the basin-sediment outflow. In all instances, the a near balance of sediment supply and transport sediment differ greatly. A majority of the annual sediment is contributed primarily by the down- in this reach of Green River after regulation. water discharge from the basin is supplied by the stream part of the tributary drainage basins. Downstream from Flaming Gorge Dam, the headwater areas. Conversely, the semiarid parts Thus, the large water-contributing parts of the Green River flows through a narrow, bedrock of the basin at lower elevations contribute most Green River basin lie around the rim, especially canyon for 12 river miles. Degradation of the of the sediment. The mean annual water dis- along the northeast divide. Conversely, the large riverbed in this reach has been limited, i'n 1982, charge and sediment discharge prior to 1962 at sediment-contributing areas are located in the bed material in this reach consisted primarily of gaging stations in the Green River basin are central and southern parts of the basin. coarse gravel, cobbles, and boulders. Photo-

6 112° 111° 110° 109° 108° 107° 106° utes -1.9 x 10 tons/yr. An equilibrium be- 44° tween sediment supply and transport occurs downstream from the mouth of the Yampa River and may exist for some distance upstream. Thus, the reach of active channel degradation is relatively short, no more than 68 river miles. This result is a consequence of the location of the reservoir just upstream from the high sediment-yielding portion of the drainage basin. Mean annual flow of the Green River at the Ouray gage has not changed appreciably since the construction of Flaming Gorge Dam. Mean annual water discharge of the Green River at the Ouray, Utah, gage was 5,500 ft3/s from 1951 to 1962 compared to 5,450 ft3/s from 1963 to 1981. The estimated mean annual sediment discharge at the Ouray gage, however, decreased 48% from 12.8 x 106 tons prior to 1962, to 6.61 x 106 tons during the post-reservoir period. The decrease in the mean annual sediment discharge at the Ouray gage, 6.2 x 106 tons, is significantly greater than the estimated annual quantity of material deposited in Flaming Gorge Reservoir, 3.6 x 106 tons. A comparison of mean annual sediment inflow and outflow to the Green River between the Jensen and Ouray gages (reach 2) shows that, on an average, 2.4 x 106 tons/yr have been deposited in this reach (Table 1). Substantial aggradation of reach 2, therefore, has occurred since the construction of Flaming Gorge Dam. This aggradation probably is con- centrated near the downstream end of the reach. The two principal tributaries to reach 2, the White and Duchesne Rivers, join the Green River within a short distance upstream from the Ouray gage. These tributaries deliver to the Green River an estimated mean annual sediment discharge of 4.8 x 106 tons or 85% of the total quantity of material supplied to reach 2. Conse- quently, aggradation in reach 2 is most likely greatest in the reach immediately upstream from the Ouray gage. Figure 2. Mean annual runoff and sediment load prior to 1962 at selected gaging stations. Upstream from the mouth of the Duchesne River, there is an approximate balance between graphs taken of the channel prior to 1962 indi- tion of the riverbed is inhibited by the coarse the supply and transport of sediment within cate that the size of bed material has not bed-material armor that has developed. reach 2 over a period of years. Consequently, changed significantly. Entrainment of such Downstream from the canyon, the Green the channel of the Green River appears to be in coarse bed material probably was only a few River flows through the wide Browns Park Val- equilibrium from the mouth of the Yampa River days a year on an average before flows were ley for 40 river miles before entering Ladore downstream to the mouth of the Duchesne regulated (Graf, 1980). As a result of flow regu- Canyon. The size of bed material decreases River under the hydraulic conditions that have lation, discharges larger than 5,000 ft3/s no through the reach, until the channel bed be- existed since flow regulation by Flaming Gorge longer occur. Prior to 1962, a daily mean dis- comes entirely sand with a median diameter of Reservoir began in October 1982. charge of 5,000 ft3/s was equaled or exceeded 0.40 mm in the last 15 river miles. Owing to Mean annual water discharge of the Green 10% of the time, and the maximum value for the channel degradation and tributary inflow, the River at the Green River, Utah, gage was 5,580 water years from 1951 to 1962 was 19,000 quantity of sediment transported increases rap- ft3/s from 1963 to 1981 compared to 5,830 ft3/s. Since 1962, discharges large enough to en- idly through the reach from between river miles ft3/s from 1944 to 1962. As shown previously train bed material and thus degrade the riverbed 34 to 50 downstream from Flaming Gorge Res- for the Green River gages at Jensen and Ouray, have occurred only very rarely as a result of ervoir. The Yampa River joins the Green River Flaming Gorge Reservoir has not had an appre- extreme floods on tributaries. Further degrada- 68 mi downstream of the reservoir and contrib- ciable effect on the mean annual flow of the

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/97/8/1012/3444989/i0016-7606-97-8-1012.pdf by guest on 30 September 2021 1016 E. D. ANDREWS

Green River. The estimated mean annual began in 1962 are especially noteworthy. First, mm, as well as for all sand-sized material and all sediment discharge at the Green River, Utah, the reach of channel degradation has a much material for the pre- and post-reservoir jwriods gage, however, has decreased 48% from 17.0 * more limited extent than do either the reaches of at the Green River near the Jensen, Utah, gage. 106 to 8.83 x I D6 tons. The decrease in the equilibrium or of aggradation. Second, the abso- The transport rate of suspended sediment in 4 mean annual sediment discharge at the Green lute magnitude of the disequilibrium between size fractions, 0.004-0.016 mm, 0.0625-0.125 River, Utah, gage, 8.1 x 106 tons/yr, is much sediment inflow and outflow throughout the ag- mm, sand-sized, and all material measured dur- greater than the estimated annual quantity of grading reach is as large as or larger than within ing the pre- and post-reservoir periods, are plot- material deposited in Flaming Gorge Reservoir, the degrading reach. The aggrading reach of the ted in Figure 3 versus the associated water 3.6 x 106 tons. The comparison of mean annual Green River is much longer than the degrading discharge. Regardless of particle size, no appre- sediment inflow and outflow in Table 1 shows reach, and the volume of accumulated sediment ciable difference in the suspended-sediment that 2 x 106 tons/yr, on an average, have ac- is much larger. In terms of channel equilibrium, transport rate at a given discharge between the cumulated in reach 3 since 1962. Tributaries the greatest impact of Flaming Gorge Dam is pre- and post-reservoir periods is apparen t at the deliver ~ 4.2 x 106 tons/yr to reach 3. The not immediately downstream, but instead sev- Jensen gage. principal tributary, Price River, joins the eral hundred miles downstream. This effect is a For each sediment-sized fraction, E. least- Green River 18 mi upstream from the Green direct result of the location of Flaming Gorge squares linear regression was fit to the log- River, Utah, gage and supplies an estimated Reservoir within the drainage basin downstream transformed values of water discharge ar d daily 6 mean annual load of 2.18 x 10 tons. The other from those parts of the basin with large water sediment transport rate measured during i:he pre- tributaries delivei sediment throughout reach 3. yield, but upstream of those parts of the basin and post-reservoir periods. The regressio n equa- Therefore, aggradation of the Green River with large sediment yields. tions are summarized in Table 2. The regression channel probably occurs along the entire length The sediment-transporting characteristics of a equations for the pre- and post-reservoir periods of reach 3. river are altered in complex and manifold as- were compared, using the F-test, to detect The discussion has emphasized the down- pects by a storage reservoir. Principally, a stor- whether a statistically significant changi: in the stream impact on channel equilibrium (sediment age reservoir may change the magnitude and relation between sediment transport rate and budget) of the Green River due to flow regula- frequency of river flows, as well as the quantity water discharge has occurred. The results of this tion at Flaming Gorge Reservoir. Immediately of sediment transported by a given discharge due analysis are summarized in Table 2. The level of downstream from the dam, potential transport to alteration of the channel morphology and/or confidence at which there is no significant dif- of sand-sized material greatly exceeds availabil- the availability of sediment within the channel. ference between the pre- and post-reservoir pe- ity. Tributaries draining areas of relatively large Each of these factors varies downstream from riods varies somewhat, but without a.ny ap- sediment yields, however, significantly increase the reservoir. The following discussions describe parent trend, among the several size fractions. the sediment load of the Green River within the adjustments of (1) the relation between sediment For sand-sized sediment, there is no significant first 68 river mil» downstream from the reser- transport rate and water discharge for various difference in the relation between transport rate voir. As a result, the supply and transport of size fractions; (2) the magnitude of effective and water discharge during the pre- ar.d post- sediment probably are in equilibrium down- water discharge; and (3) bankfull channel di- reservoir periods at the 90th percentile level. For stream from the mouth of the Yampa River. The mensions in the vicinity of the Jensen and Green all sediment sizes, there is no significant differ- zone of equilibrium under present (1985) condi- River, Utah, gaging stations. ence in the pre- and post-reservoir transport relations probably extends downstream to the tions at the 75th percentile level of confidence. These tests are quite strict. It is thus concluded mouth of the Duchesne River, which joins the Sediment Transport Rate Green River at river mile 166. Within this reach that the sediment-transport rate at a given discharge has changed very little, if at all, at the of the Green River, there has been no net ac- It has been shown that the mean annual sed- Jensen gage as a result of flow regulation and cumulation or depletion of sediment. Down- iment discharge at the Jensen gage has decreased sediment storage by Flaming Gorge Reservoir. stream from the mouth of the Duchesne River, by 54% from 6.92 x 106 to 3.21 x 106 tons since Any changes in the sediment-transport lelations the mean annual, supply of sediment from up- 1962. This change was determined from mea- probably are limited to those particles <0.062 stream and tributaries has exceeded the trans- sured daily values of suspended-sediment con- mm in diameter. port. As a result, there is a long-term net centration and water discharge. During the accumulation of sediment. The zone of aggrada- period of record at the Jensen gage, water years This analysis is in good agreement \vith the tion probably extends downstream on the Green 1947-1979, the size distribution of suspended pre- and post-reservoir sediment budgets sum- River to its confluence with the Colorado, al- sediment was determined for 218 of the daily marized in Table 1. Although the meari annual though there are no gage records of annual concentration samples, 161 before October sediment load decreased by 54% after 1J62, the water and sediment discharge downstream from 1962 and 57 after October 1962. The daily sed- inflow and outflow of sediment to the reach Green River, Utah, to confirm this conclusion. iment transport (Ik) for a given size fraction, k, upstream from the Jensen gage have remained The principal tributary to the Green River was computed from the measured percentage of in approximate equilibrium. Consequently, the downstream from Green River, Utah, is the San sediment in a size fraction, Pk, the total concen- quantity of sediment stored in the reach up- Rafael River. Ioms and others (1965) estimated tration, C, and the daily mean discharge, Q: stream has remained nearly constant. Assuming that the San Rafael River supplied on an aver- the size distribution of sediment has not changed age in excess of 1 x 10® tons/yr to the Green Ik = 0.0027 (Pk)(C)(Q). (1) appreciably, the transport rate of sediment at a River, but relatively small mean annual water given discharge should be unaffected oy pres- discharge of ~ 140 ft3/s. Daily sediment transport rates were deter- ence and operation of Flaming Gorge Rservoir. Two aspects concerning the downstream se- mined for 6 size fractions—<0.004 mm, At the Green River, Utah, gage, the size dis- quence of degradation, equilibrium, and aggra- 0.004-0.016 mm, 0.016-0.0625 mm, 0.0625- tribution of suspended sediment was determined dation that have existed since flow regulation 0.125 mm, 0.125-0.250 mm, and 0.250-0.500 for 286 of the daily concentration samples, 220

1,000,000 Figure 3. Measured suspended-sediment transport rates for various-sized material as a function of water discharge at the Green 100,000 River near Jensen, Utah, gage during the pre- reservoir (open triangles) and post-reservoir (solid dots) periods. A. 0.004-0.016 mm fraction. B. 0.0625-0.125 mm fraction. C. Sand- 10,000 size fraction. D. All sizes.

1000 ment transport rate of any particle size at a given discharge is apparent between the pre- and post- reservoir periods. 100 For each sediment-sized fraction, a least- squares linear regression was fit to log-trans- > < formed values of water discharge and daily Q sediment transport rate measured at the Green w 10 River, Utah, gage during the pre- and post- o reservoir periods. The regression equations determined for the pre- and post-reservoir periods (D oc were compared, using the F-test statistic. The < regression equations and results of this analysis I a 1,000,000 are summarized in Table 3. As was the case for w o values measured at the Jensen gage, the level of confidence at which no significant difference between the pre- and post-reservoir periods is de- 100,000 tected varies considerably among the several size o LÜ fractions. In general, the confidence level im- CO proves with increasing particle size. For the sand-sized sediment, there is no significant dif- 10,000 - ference in the relation between transport rate and water discharge during the pre- and post- reservoir periods at the 90th percentile level. For

1000 all sediment sizes, there is no significant difference in the pre- and post-reservoir transport relations at the 95th percentile level of confidence. The confidence levels are slightly less than those 100 determined for the Jensen gage. Nevertheless, the analysis indicates that there has been no significant change in the sediment transport rate at the Green River, Utah, gage as a result of Flam- 10 ing Gorge Reservoir. As shown earlier, the inflow of sediment to the reach upstream from the Green River, Utah, gage, exceeds the outflow by a considerable quantity, and the reach is accu- 1 mulating nearly 2.0 x 106 tons/yr on an average. 100 1000 10,000 100,000 100 1000 10,000 100,000 In spite of the large quantity of material that has accumulated in the reach upstream from this WATER DISCHARGE (FF/S) gage, the sediment transport relations have remained remarkably constant.

before October 1962 and 66 after October River at the Green River, Utah, gage. The trans- The theory of alluvial river channels (see 1962. Using these measurements, daily sediment port rate of suspended sediment in 4 size Mackin, 1948; Leopold and Maddock, 1953) transport rates were determined for 6 size fractions—0.004-0.016 mm, 0.0625-0.125 mm, holds that hydraulic characteristics of a channel fractions—<0.004 mm, 0.004-0.016 mm, sand-sized, and all material measured during the will adjust over a period of years to transport the 0.016-0.0625 mm, 0.0625-0.125 mm, 0.125- pre- and post-reservoir periods—are plotted in quantity of sediment supplied with the available 0.250 mm, and 0.250-0.500 mm—as well as Figure 4 versus the associated water discharge. discharge. With regard to the current condition for all sand-sized material and all material for As shown previously for the Jensen gage, no of the Green River in the vicinity of the Green the pre- and post-reservoir periods at the Green appreciable difference in the suspended-sedi- River, Utah, gage, this principle indicates that

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/97/8/1012/3444989/i0016-7606-97-8-1012.pdf by guest on 30 September 2021 1018 E. D. ANDREWS

the sediment transport rate at a given discharge TABLE 2. COMPARISON OF REGRESSION EQUATIONS RELATING SUSPENDED-SEDIMENT TRANSPORT RATES TO WATER DISCHARGE DURING THE PRE- AND POST-RESERVOIR PERIODS FOR THE GREEN RIVER NEAR JENSEN, UTAH should be increasing with time. The comparison

of sediment transport rates as a function of water Sediment Period Relation between sediment Correlation Level of size of discharge (Lj) and water coefficient significance discharge at the < jreen River, Utah, gage for the fraction record discharge (Q) pre- and post-reservoir periods, however, found no evidence of an appreciable increase in sedi- <0.004 <1962 L,= 1.08Q1 00 0.31 s 1963 L = 0.336Q''" .28 1 2 ment transport rate for a given flow after the 0.004-0.016 <1962 LS = 0.0821Q ' ® .42 91963 L = 0.0108Q1-36 .42 completion of Flaming Gorge Dam in 1962. 0.016-0.0625 <1962 L= 2.16 » lO^Q1 88 .77 >1963 L, = 5.03 X IO"6«226 .79 0.0625-0.125 <1962 L* = 5.04 X 10-V-49 Change in Effec tive Discharge >1963 L, = 2.26 X IO-'Q2 30 0.125-0.250 <1962 L, = 5.59 X IO-'Q2 73 >1963 1^=1.67X10-5Q2'» .72 2 53 0.250-0.500 <1962 L, = 1.24 X lO-'Q .82 One of the principal downstream effects of 7 2 34 >1963 Lj = 5.04 X 10" Q - .69 2 Flaming Gorge Reservoir is to decrease the Sand <1962 Ls = 3.38 « lO-'Q « .87 >1963 L = 2.04 X IO-5!}216 .79 3 1 70 range of daily mean flows. As noted previously, All sizes <1962 L! = 6.16X10- Q - .62 2 154 the mean annual discharges of the Green River >1963 L,= 1.72 X 10- Q .62 during the pre- and post-reservoir periods are virtually identical at both the Jensen and the

Green River, Utah, gages. The percentage of TABLE 3. COMPARISON OF REGRESSION EQUATIONS RELATING SUSPENDED-SEDIMENT TRANSPORT RATES TO WATER time that various daily mean discharges are DISCHARGE DURING THE PRE- AND POST-RESERVOIR PERIODS FOR THE GREEN RIVER NEAR GREEN RIVER, UTAH equaled or exceeded, however, is substantially Sediment Period Relation between sediment Correlation Level of different for most flows. The durations for daily size of discharge (Lj) and water coefficient significane; mean discharge for the pre- and post-reservoir fraction record discharge (Q)

105 periods are compared in Figure 5 for the Green- <0.004 <1962 i M.49Q 0.35 1 34 dale gage, Figure 6 for the Jensen gage, and >1963 L = 0.0693Q .40 0.004-0.016 <1962 [j-0.2310™ .40 2 160 Figure 7 for the Green River, Utah, gage. The >1963 Lj = 0.271 X 10" Q .44 4 95 0.016-0.0625 <1962 L, = 1.99 X l1963 I, = 4.54 X IO-SQ .67 2 68 0.0625-0.125 <1962 L, - 1.02 « io-'q - .86 the time has been significantly decreased since 3 04 >1963 I, = 3.07 X IO-'Q .84 7 2 60 regulation of the Green River by Flaming Gorge 0.125-0.250 <1962 L, = 1.24 X IO" !} - .89 2 47 >1963 I, = 3.74 X lO^Q - .75 Reservoir began. For a given duration within 2 23 0.250-0.500 <1962 I, = 7.52 X lO-'Q - .87 5 87 >1963 L = 2.04 X 10" Q' .41 this range, the decrease in discharge is about the 2 71 Sand <1962 L, = 1.24 X lO^Q - same at all three: gages. Maximum daily dis- >1963 I, = 2.06 X IO-SQ290 1 53 All sizes <1962 L = 0.0503Q ' .58 charge, for example, has decreased by 14,000 -2 78 >1963 Lj = 0.363 X 10 Q' - .61 ft3/s, whereas the 2% exceedence discharge has decreased by 7,OCO ft3/s. Discharges equaled or exceeded <5% of the time at the Jensen and computed for three reaches. These reaches are able to consider only those sediment sis«s that Green River, Utah, gages have decreased by representative of conditions in alluvial sections are present in the channel bed and banks in -25%. Thus, although the drainage area is 2.7 of channel in each of the downstream zones— quantities of more than a few percent. The times greater at the Green River, Utah, gage degradation, equilibrium, and aggradation— transport rate of those particle sizes that are than it is at Flaming Gorge Reservoir, the effect defined by the sediment-budget analysis. present in the channel perimeter tends to be of the reservoir is still substantial 290 river miles Selected reaches are (1) through Browns Park more closely correlated with flow, because the downstream. (degrading); (2) immediately downstream from quantity of material available does not vary Discharges equaled or exceeded < 10% of the the Jensen gage (equilibrium); and (3) in the greatly over a period of years. Analysis of the time usually transport a majority of the annual vicinity of the Green River, Utah, gage (aggrad- relations between sediment transport rate and sediment load (Wolman and Miller, 1960). ing). The duration of discharges and the relation discharge for various sediment-size fractions 2 Furthermore, the increment of discharge that of sediment transport rate to discharge need to found that the coefficient of determination (r ) transports the largest quantity of sediment over a be known to compute the effective discharge. increased significantly for particles larger than period of years, called the effective discharge, This information is provided at the latter two 0.016 mm in diameter (see Tables 2 £.nd 3). typically is equaled or exceeded <5% of the reaches by the records of the nearby gages. Consequently, particles with a fall diameter of time, but it determines the bankfull-channel di- There are no comparable gage records of dis- 0.016 mm were assumed to be the smalhst par- mensions and pattern. Andrews (1980) com- charge and sediment transport in the Browns ticles present to an appreciable degree in the puted the effective discharge for 15 gages in the Park reach. The duration of discharges recorded channel bed and banks in the vicinity of the Yampa River basin and showed that the effec- at the Greendale gage therefore were assumed to Jensen and Green River, Utah, gages. tive discharge wai; nearly identical to the bank- be representative, because the contribution by Sediment-load duration relations for each size full discharge at all sites. The bankfull-channel intervening tributaries is rarely significant. The fraction >0.016 mm were computed from the dimensions thus appeared to be adjusted to the quantity of bed-material transport by various transport rate versus discharge (Figs. 3 ind 4) effective discharge. discharges was computed, using the Engelund- and the flow duration relations. The ranges of In order to dete rmine the effect of flow regu- Hansen relation (Engelund and Hansen, 1967), discharge were divided into -30 equal incre- lation on the effective discharge in the Green with measured values for slope, width, and bed- ments. The sediment-discharge duration rela- River, the mean annual quantity of sediment material size distribution. tions then were integrated between the limits of transported during the pre- and post-reservoir The object of the analysis is to assess the po- each increment and the result multiplied by 365 periods by various increments of discharge was tential for channel changes, and so it is prefer- days. The mean annual quantity of sediment

1,000,000 effective discharge (modal value) for the Green River in the Browns Park reach was computed to be 7,450 ft3/s during the pre-reservoir period, water years 1951-1962. This discharge was 100,000 equaled or exceeded 5.5% of the time, or 20.0 days/yr on an average. During the post- reservoir period, water years 1966-1981, the effective discharge decreased by 63% to 2,750 10,000 ft3/s. This discharge has a duration of 27% of the time, or 99 days/yr on an average. The effective discharge (modal value) for the Green River near Jensen gage was 20,500 ft3/s during the 1000 4Ï A A pre-reservoir period, water years 1947-1962. > A This discharge was equaled or exceeded 3% of the time, or 11 days/yr on an average. During ¿rfi A AÛA the post-reservoir period, water years 1966- 100 1981, the effective discharge decreased by 44% 3 <> to 11,500 ft /s. This discharge has a duration of o 7.5% of the time or 27.4 days/yr. The effective CO 10 discharge of the Green River in the vicinity of -Z. o the Green River, Utah, gage was 26,500 ft3/s during the pre-reservoir period, water years a 1944-1962. The effective discharge was equaled £ or exceeded 2.8% of the time or 10.2 days/yr. 5 1,000,000 During the post-reservoir period, water years C/D 1966-1981, the effective discharge decreased by 23% to 20,500 ft3/s. The duration of effective discharge has decreased slightly to 2.4% of the 100,000 time or 8.8 days/yr. Q LU Magnitude of the effective discharge has de- CO creased significantly in all three reaches as a consequence of flow regulation. The percentage 10,000 change is greatest at the farthest upstream reach and decreases downstream as the influence of the reservoir diminishes. Nevertheless, the effect

1000 is still significant 290 river miles downstream at the Green River, Utah, gage, where the contributing area is 2.7 times that of the reservoir. No directly comparable studies showing the down- 100 stream influence of a reservoir are known. Gregory and Park (1974), however, described an example in which the decrease in channel area extended downstream to a point where the 10 contributing area was four times that at the reservoir.

ADJUSTMENT OF BANKFULL- CHANNEL DIMENSIONS 100 1000 10,000 100,000 100 1000 10,000 100,000

Computed effective discharges were shown to WATER DISCHARGE (FT3/S) be in excellent agreement with measured bank- Figure 4. Measured suspended-sediment transport rates for various-sized material as a full discharges at unaltered watersheds with self- function of water discharge at the Green River at Green River, Utah, gage during the pre- formed channels in the Yampa River basin reservoir (open triangles) and post-reservoir (solid dots) periods. A. 0.004-0.016 mm frac- (Andrews, 1980). The bankfull characteristics of tion. B. 0.062S-0.12S mm fraction. C. Sand-size fraction. D. All sizes. these self-formed channels therefore appeared to be determined by the effective discharge over a transported by an increment of discharge was and post-reservoir periods are compared in period of years. As shown above, the effective determined by summing over the several size Figure 8 for the Browns Park reach, in Figure 9 discharge computed for selected reaches of the fractions. for alluvial sections in the vicinity of the Jensen Green River downstream from Flaming Gorge The quantities of sediment transported by dif- gage, and in Figure 10 for alluvial sections in the Reservoir has decreased significantly. Conse- ferent increments of discharge during the pre- vicinity of the Green River, Utah, gage. The quently, it is to be expected that the bankfull

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/97/8/1012/3444989/i0016-7606-97-8-1012.pdf by guest on 30 September 2021 1020 E. D. ANDREWS

"1—I—I—I 1—I—m—I—r —I—I 1—r 40,000

M 5000 10,000 - — uj t

l>RE-RESERVOIR

l'OST-RESERVOlR

500

J I ! I ! I L I I I I I I I I I L I I 0.05 0.1 0.5 1 2 5 10 20 30 40 50 60 70 80 90 95 98 99 99.8 99.9 0.05 0.1 0.5 1 2 5 10 20 30 40 50 60 70 B0 90 95 98 99 99.8 99.9 DURATION OF DISCHARGE (% TIMEI DURATION OF DISCHARGE (% TIME)

Figure 5. Comparison of the duration for daily mean discharges Figure 6. Comparison of the duration for daily mean discharges during the pre- and post-reservoir periods measured at the Green during the pre- and post-reservoir periods measured at the Green River near Greendale, Utah, gage. River near Jensen, Utah, gage.

channel of the Green River will adjust over a cause of their greater stability. At all locations, some reaches of the Green River downstream period of years to the locally prevailing effective one or both of the banks are bedrock, and, ex- from Flaming Gorge Dam were taken prior to discharge. cept for the Ouray gage, the bed material is 1962 and after 1980. Thus, it is possible: to in- Detailed information describing the bankfull- coarse gravel. The gaged cross sections thus can- vestigate some of the channel changes resulting channel dimensions of the Green River is lim- not be considered typical of the channel in allu- from flow regulation over ~20 yr. Although aer- ited, especially prior to construction of naming vial, self-formed reaches. ial photographs do not provide a complete Gorge Dam. All of the gaging stations used in Aerial and terrestrial photographs are the description of the bankfull-channel characteris- this investigation have rated cross sections where primary sources of information from which the tics, several attributes can be measured and de- the discharge is measured several times a year. hydraulic adjustment of Green River can be reduced, especially when combined wit'i field These cross sections, however, were selected be- constructed. Large-scale aerial photographs of inspection. (1) Bankfull-channel width a.nd the

III 11 1 1 1 1 1 1 1 1 1 1 I 1 30,000

\ \ _ \ \ \ \ \\ t \\ \\ S 10,000 I - A - A - A \\ - - \\ - \ ^ FRE-RESERVOIR FOST-RESERVOIR \

1000 1 1 800 "li 1 1 1 1 1 1 0.05 0.1 0.5 1 2 5 10 20 30 40 50 60 70 80 90 95 98 99 99.8 7500 10,000 12,500 15,000 17.50C 20,000 DURATION OF DISCHARGE (% TIMEI DISCHARGE (FP/S) Figure 7. Comparison of the duration for daily mean discharges Figure 8. Comparison of the quantity of bed material transported by during the pre- and post-reservoir periods measured at the Green increments of discharge during the pre- and post-reservoir periods River at Green I liver, Utah, gage. computed for the Green River through the Browns Park reach.

300,000 1 1 T I IPRE-RESERVOIR ^POST-RESERVOIR

250,000

200,000

Ö 300,000 —

150,000

1 o

100,000

50,000 20,000 30,000 DISCHARGE (FP/S)

Figure 10. Comparison of the quantity of bed material transported by increments of discharge during the pre- and post-reservoir periods computed for the Green River in the vicinity of the Green River, 20,000 30,000 40,000 Utah, gage. DISCHARGE (FT'/S)

Figure 9. Comparison of the quantity of bed material transported by increments of discharge during the pre- and post-reservoir periods charge (Qi), and estimated future bankfull dis- computed for the Green River in the vicinity of the Jensen, Utah, charge (Q2), the future bankfull width (W2) can gage. be estimated by equation 3, assuming that ai = a.2. This is not an unreasonable assumption as long as such factors as channel slope, composi- extent of channel migration were determined di- These relations are called the "downstream hy- tion and density of bank vegetation, bed- rectly from the photographs. (2) Changes in the draulic geometry" of a river. Langbein (1964), material-size distribution, and suspended-sedi- channel pattern were described. (3) The flood Engelund and Hansen (1967), and Parker ment concentration versus discharge relations plain constructed during the pre- and post- (1978, 1979) showed that the width-versus- remain unchanged. If the measured bankfull reservoir periods was identified on photographs. discharge equation is the most consistent rela- width after a large change in the effective dis- Subsequently, the change in flood-plain elevation. The coefficient of determination (r2) for charge differs significantly from the estimated tion was surveyed during a field inspection. the width equation typically is 0.95 or larger, value, then one may conclude that the adjust- Leopold and Maddock (1953) found that the and the exponent value, b, rarely differs much ment of channel width is incomplete. variation of the hydraulic variables—mean ve- from 0.5. Underlying the hydraulic geometry Although the Green River has been affected locity (u), width (w), and mean depth (d), with analysis there is the assumption or determination by the spread of tamarisk during the past cen- increasing bankfull discharge (QB) in the down- that the hydraulic variables have attained a tury, riparian vegetation has not been a major stream direction—could be described broadly quasi-equilibrium adjustment to the magnitude cause of any channel changes since 1951. Graf by a set of simple power equations: and frequency of flows that have occurred over (1978) examined channel changes at 18 cross a period of years. sections first photographed by river explorers m u = k QB , (2) A significant deviation in the value of a hy- prior to 1915 and found that thick stands of draulic variable from the expected value is evi- tamarisk had grown along the banks in many b w = aQB , (3) dence that the particular variable is not in reaches. Concomitantly, channel width has de- and adjustment with the prevailing range and dura- creased by an average of 27%. These data indi- tion of sediment and water discharges. Given the cate, however, that all of this change occurred f before 1951. d = cQB . (4) prior bankfull width (wj), the prior bankfull dis-

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/97/8/1012/3444989/i0016-7606-97-8-1012.pdf by guest on 30 September 2021 1022 E. D. ANDREWS

Adjustment of the Green River Channel Adjustment of the Green River Channel measured at 14 cross sections located in a reach in Browns Park Downstream from Jensen, Utah of nearly 15 river miles. The channel was narrower at all cross sections in 1981 compared to Bankfull-channel width of the Green River A few miles downstream from the Jensen 1952. On an average, the bankfull-channel through Browns Park was measured at 24 cross gage, the Green River enters an alluvial valley width decreased by 10% from 515 to 465 ft. sections spaced evenly throughout a reach of 22 within which it flows for nearly 60 river miles. Adjustment of channel width in the Green river miles, using large-scale aerial photographs Bankfull-channel width in this reach was mea- River downstream from the Green River, Utah, taken in 1951 and 1980. The channel was nar- sured at 15 cross sections, using large-scale aerial gage to the flow regulation by Flaming Gorge rower at all locations except one. On an average, photographs taken in 1964 and 1978. Although Reservoir appears to be nearly complete. The channel width decreased by 13% from 560 to the earlier series of photographs were taken 2 yr expected channel width under quasi-equilibrium 485 ft during the p;riod. As described above, the after flow regulation had begun, they probably conditions, given the decreased effective dis- effective discharge of the Green River in Browns show a good representation of the pre-reservoir charge, is 450 ft. This value is only slightly less Park decreased by 63% to 2,750 ft3/s from river channel. The channel was narrower at all than the measured value of 465 ft. 7,450 ft3/s as a result of flow regulation by locations examined except at one in 1978 com- Flaming Gorge Reservoir. pared to 1964. On an average, bankfull-channel SUMMARY AND CONCLUSIONS At some time since the 1951 aerial photo- width decreased by 13% from 700 ft to 610 ft. graphs were taken, the Green River in Browns The computed effective discharge of the Green Park began building a new flood plain ~4.0 ft River at the Jensen, Utah, gage decreased by The contribution of runoff and sediment per 3 lower than the previous one. The lower flood- 44% from 20,500 to 11,500 ft /s during the unit area to the channel network varies greatly plain elevation is due both to a decrease in bank- post-reservoir period, water years 1966-1981. within the Green River basin. Furthermore, the full depth associated with the decreased effective Adjustment of Green River channel width principal source areas of runoff and sediment are discharge and to degradation of the riverbed. downstream from the Jensen, Utah, gage was different. A majority of the basin-wide runoff is Given the measured river slope, bankfull width, incomplete in 1978. The estimated quasi- supplied by the relatively high-elevation areas and bed-material-size distribution, stage-disequilibrium channel width associated with the (>10,000 ft) near the rim of the basin. These charge relations were computed for the pre- and decreased effective discharge as computed by areas, however, have very small sediment yields. post-reservoir periods, using the Engelund- equation 3 is 524 ft. This estimate is substan- The most important sediment-contributing areas Hansen roughness equation (Engelund and tially less than the measured value of 610 ft. If a are far downstream, within the middle- and Hansen, 1967). This analysis indicated that the constant rate of channel narrowing is assumed, lower-elevation parts of the basin. bankfull depth has: decreased from 4.7 ft to 3.1 -30 yr will be required for the channel width to The downstream effects of Flaming Gorge ft. The mean channel degradation therefore has attain the expected value associated with the Reservoir are profoundly affected by its location 6 been -2.4 ft. An estimated 9.5 x 10 tons of post-reservoir effective discharge. At many cross in the drainage basin relative to the principal sand-sized bed material was eroded from the 22 sections, the bankfull channel has become nar- runoff and sediment-contributing areas. Com- mi of alluvial channel within Browns Park be- rower as a result of accretion of material along pared to the farthest downstream gage, the tween 1951 and 1980, when the aerial photo- one or both banks. Although the accretion of Green River at Green River, Utah, the area up- graphs were taken. All of the degradation bank material was common, the most significant stream from Flaming Gorge Reservoir, prior to probably occurred since 1962, when the up- process narrowing the channel occurred where a its construction, contributed 37% of the annual stream supply of s jdiment was substantially de- distributary channel has filled with bed material runoff but only 21% of the sediment load from creased by Flaming Gorge Reservoir. and the mid-channel bar has become attached to 37% of the drainage area. Consequently, tie res- Adjustment of channel width to the regulated the bank. Concomitantly, a thick vegetation ervoir controls a proportional share of the basin flow and decreased sediment supply to the cover has become established on these areas. runoff but traps only a moderate proport ion of Green River through Browns Park is far from Both series of aerial photographs show the the basin-sediment yield. The downstream ef- Green River at approximately the same dis- fects of Flaming Gorge Reservoir on the Green complete. The estimated quasi-equilibrium 3 channel width associated with the decreased ef- charge, 6,000 to 7,000 ft /s. The number and River channel would be markedly different if fective discharge as computed by equation 3 is areal extent of mid-channel bars was apprecia- either a larger or smaller proportion of the basin- 340 ft. This estimate is significantly less than the bly greater in 1978 than in 1964, in spite of the sediment yield were trapped within the measured value of 485 ft. On the basis of the fact that some bars have become attached to reservoir. comparison of aerial photographs, as well as banks. Pre- and post-reservoir sediment budgets cross sections surveyed in 1983, the present were computed for three reaches of the Green channel primarily is the result of riverbed degra- Adjustment of the Green River Channel River, using measured daily water and sediment dation. Deposition of new bank material does Downstream from Green River, Utah discharges. Prior to the construction of a dam in not appear to have been a significant factor con- Flaming Gorge, a quasi-equilibrium condition tributing to the decrease in channel width. An alluvial reach located downstream from appears to have existed downstream \n the Rather, the preseni;, smaller channel was formed the Green River, Utah, gage was selected to in- Green River channel; that is, over a period of by entrenchment within the pre-1962 channel. vestigate changes in bankfull-channel width in years, the transport of sediment out of a given Given the deficit in sediment supply compared reach 3, the aggrading part of the river. Large- river reach equaled the supply of sediment into to transport, net deposition of material along the scale aerial photographs of this reach were taken the reach. Since reservoir regulation began in banks would be inconsistent. in 1952 and 1981. Bankfull-channel width was 1962, the mean annual sediment discharge at

downstream gages has decreased substantially. nual hydrograph rather than from a decrease in ACKNOWLEDGMENTS The mean annual sediment discharge has de- the annual runoff. creased by 54% from 6.92 x 106 to 3.21 x 106 The quantity of sediment in transport at a Robert Hirsch and Brent Troutman contrib- tons at the Jensen gage, by 48% from 12.8 x 106 given discharge does not appear to have been uted to this investigation in many ways. Their to 6.62 x io6 tons at the Ouray gage, and by affected by naming Gorge Reservoir. For both thoughtful discussions and critical comments 48% from 17.0 x 106 to 8.83 x io6 tons at the the Jensen and Green River, Utah, gages, the concerning the analysis of long-term trends in Green River, Utah, gage. The decrease in mean variation of daily suspended-sediment transport sediment transport rates have been greatly ap- annual sediment discharge at the Ouray and rate as a function of water discharge was com- preciated. Robert Meade, John Lewin, and Green River, Utah, gages far exceeds the quan- puted for each size fraction from <0.004 mm to Donald Hillier reviewed drafts of the manu- tity of sediment trapped in the reservoir. the 0.250-0.500 mm fraction by a least-squares script. Their suggestions have been especially Post-reservoir annual sediment budgets show regression of the log-transform data during the helpful. that a majority of the Green River channel is no pre- and post-reservoir periods. No statistically longer in quasi-equilibrium. Three distinct longi- significant change in the relations was detected REFERENCES CITED by the F-statistic at the 95th-percentile level of tudinal zones involving channel degradation, Andrews, E. D., 1980, Effective and bankfull discharges of streams in the quasi-equilibrium, and aggradation were identi- confidence. Yampa River basin, Colorado and Wyoming: Journal of Hydrology, v. 46, p. 311-330. fied. Beginning immediately downstream from Over a period of years, the bankfull dimen- Engelund, F., and Hansen, E., 1967, A monograph on sediment transport in alluvial streams, in Teknisk Verlag: Copenhagen, Denmark, Technical the dam and extending downstream 68 river sions of an alluvial channel adjust to the magni- University of Denmark, 63 p. miles, sediment transport out of the reach ex- tude and duration of discharges. The effective Graf, W. L„ 197S, Fluvial adjustments to the spread of tamarisk in the Colo- rado Plateau region: Geological Society of America Bulletin, v. 89, ceeds the tributary contribution, and the channel discharge is the increment of discharge that p. 1491-1501. 1980, The effect of dam closure on downstream rapids: Water Re- is degrading. The length of the degrading reach, transports the largest quantity of sediment over a sources Research, v. 16(1), p. 129-136. however, is relatively limited, due to the large period of years. Computed effective discharges Gregory, K. J., and Park, C., 1974, Adjustment of river channel capacity downstream from a reservoir: Water Resources Research, v. 10(4), quantity of sediment supplied by tributaries. for 15 alluvial reaches in the Yampa River basin p. 870-873. Hathaway, G. A., 1948, Observations on channel changes, degradation and In the reach between river miles 68 and 166 were found to be nearly identical to the bankful scour below dams: International Association for Hydraulic Research, discharge (Andrews, 1980). Therefore, the Second Meeting, Report, p. 267-307. downstream from Flaming Gorge Reservoir, the Iorns, W. V , Hembree, C. H . and Oakland, G. L„ 196S, Water Resources of quantity of sediment supplied to the reach from bankfull-channel dimensions appear to be ad- the Upper Colorado River Basin—Technical Report: U.S. Geological Survey Professional Paper 441,370 p. upstream plus tributaries approximately equals justed to the effective discharge. Komura, S., and Simons, D. B., 1967, River bed degradation below dams: American Society of Civil Engineers, Journal of the Hydraulics Divi- the transport of sediment out of the reach over a The effective discharges that occurred during sion, HY4, p. 1-14. period of years. This reach appears to be in the pre- and post-reservoir periods were com- Lane, E. W., 1955, The importance of fluvial morphology in hydraulic engineering: American Society of Civil Engineers, Proceedings, v. 81, quasi-equilibrium, as there is no net accumula- puted for three reaches downstream from Flam- p. 1-17. Langbein, W. B., 1964, Geometry of river channels: American Society of Civil tion or depletion of bed material. Downstream ing Gorge Reservoir. Since 1962, the effective Engineers, Proceedings, Journal of the Hydraulics Division, v. 90, HY2, from river mile 166 to the mouth (river mile discharge of the Green River has decreased from p. 301-312. Lawson, L. M., 1925, Effects of Rio Grande storage on river erosion and 3 412), the supply of sediment from upstream and 7,450 to 2,750 ft /s, in Browns Park, 20,500 to deposition: Engineering News-Record, v. 95(10), p. 372-374. 3 Leopold, L. B., and Maddock, T., Jr., 1953, The hydraulic geometry of stream tributary inflow exceeds the transport of sedi- 11,500 ft /s in the vicinity of the Jensen gage, channels and some physiographic implications: U.S. Geological Survey 6 ment out of the reach by 5.4 x 10 tons/yr on an and 26,500 to 20,500 ft3/s in the vicinity of the Professional Paper 252,57 p. Mackin, J. H., 1948, Concept of the graded river: Geological Society of Amer- average. Green River, Utah, gage. ica Bulletin, v. 59, p. 463-512. Miller, W. H„ Archer, Donald, Tyrus, H. M„ and McNatt, R. M„ 1982, The decrease in mean annual sediment trans- An analysis of aerial photographs and inspec- Yampa River fishes study—Final report: Salt Lake City, Utah, U.S. Fish and Wildlife Service, 78 p. port at the Jensen and Green River, Utah, gages tion of channel morphology indicate that the Parker, Gary, 1978, Self-formed straight riven with equilibrium banks and since 1962 is due entirely to a decrease in the bankfull-channel width of the Green River has mobile bed—Part 1, The sand-silt river: Journal of Fluid Mechanics, v. 89, pt. 1, p. 109-125. magnitude of river flows that are equaled or decreased in response to the decreased effective 1979, Hydraulic geometry of active gravel rivers: American Society of Civil Engineers, Journal of the Hydraulics Division, v. 105, HY9, exceeded <30% of the time. Daily mean water discharge caused by flow regulation. On an av- p. 1185-1201. discharges with a duration of 5% or less have erage, bankfull-channel width has decreased by Petts, G. E„ 1979, Complex response of river channel morphology subsequent to reservoir construction: Progress in Physical Geography, v. 3(3), decreased in magnitude by 25% during the post- 13% to 485 ft from 560 ft through Browns Park, p. 329-362. Petts, G. E., and Lewin, J., 1979, Physical effects of reservoirs on river systems, reservoir period at both the Jensen and Green 13% to 610 ft from 700 ft downstream from the in Hollis, G. E., ed., The impacts of man on the hydrological cycle: River, Utah, gages. The magnitude of daily Jensen gage, and 10% to 465 ft from 515 ft Norwich, England, Geoboolcs, p. 79-91. Sayre, W. W., and Kennedy, J. F., 1978, Degradation and aggradation of the mean discharges with a duration >30%, how- downstream from the Green River, Utah, gage. Missouri River: University of Iowa, Iowa Institute of Hydraulic Re- search Report 215,61 p. ever, has increased to the extent that the mean Although the decrease in bankfull-channel Williams, G. P., and Wolman, M. G., 1984, Downstream effects of dams on annual runoff measured during the pre- and width has been significant in all reaches, a chan- alluvial rivers: U.S. Geological Survey Professional Paper 1286, 83 p. Wolman, M. G., and Miller, J. P., 1960, Magnitude and frequency of forces in post-reservoir periods is virtually unchanged at nel width consistent with the prevailing effective geomorphic processes: Journal of Geology, v. 68, p. 54-74.

both gages. The decrease in annual sediment discharge exists only in the reach downstream MANUSCRIPT RECEIVED BY THE SOCIETY MAY 20,1985 REVISED MANUSCRIPT RECEIVED FEBRUARY 12,1986 transport thus results from a more uniform an- from the Green River, Utah, gage. MANUSCRIPT ACCEPTED FEBRUARY 14, 1986

Printed in U.S.A.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/97/8/1012/3444989/i0016-7606-97-8-1012.pdf by guest on 30 September 2021