Figure 1. Map of Study area showing location of Red Creek drainage basin.
GIS in Water Resources
Fall
1999
Term Paper
Paul Grams
Department of Geography and Earth Resources
Utah State University
grams@cc.usu.edu
Table of Contents
The sediment budget is among the most fundamental and valuable tools that can be applied to fluvial geomorphic studies. A sediment budget may be quantified at many different scales and levels of detail. At their most basic, sediment budgets are a quantification of the change in sediment storage for a given reach:
change in storage = sediment influx - sediment outflux.
More complex sediment budgets can separate between components of sediment storage and sources of sediment input. However, even simple sediment budgets can be extremely difficult to quantify because of the limited availability of accurate long-term sediment transport data. Yet, the need remains to interpret past and present patterns of fluvial erosion and deposition, and more importantly, to predict future sedimentation trends. These data can be especially important to scientists and resource managers charged with formulating management options for heavily impacted rivers. Such goals cannot be achieved without quantification of sediment budgets.
Completion of large hydroelectric dams on many rivers of the western United States has significantly altered the sediment transport regimes, and consequently the downstream patterns of erosion and deposition (Collier et al., 1996). These issues have been studied extensively on the Colorado River in Grand Canyon and more recently on the Green River below Flaming Gorge Dam. Patterns of sediment erosion and deposition are of interest because they affect the formation and maintenance of characteristic channel and bank features that are the foundation of the aquatic and riparian ecosystems. On the Colorado River system, to which the Green River is a principal tributary, sand bars create important nursery habitat and gravel bars create spawning habitat for endangered species of native fish (Stanford, 1994). The regeneration and recruitment of cottonwood trees and the life history of a threatened orchid that inhabits the flood plain are also dependent on sedimentation patterns. Sedimentation and channel-conveyence capacity in populated reaches of the river could also be concerns in the future if flooding of human settlements becomes a problem.
Because large reservoirs effectively trap nearly all upstream sediment (Williams and Wolman, 1984), management of downstream sediment resources depends on tributary contributions. Sediment inputs from large tributaries with hydrologic and sediment monitoring stations as well as the hundreds of ungaged tributaries is a principle research focus in Grand Canyon (Webb et al., 1999). Research on the Green River has documented channel narrowing in Dinosaur National Monument, 60 km downstream from Flaming Gorge Dam (Grams, 1997). These findings indicate that, although Flaming Gorge Dam blocks all upstream sediment, sediment transport in downstream reaches is limited by transport capacity rather than sediment supply. Possible sediment sources in the reach between the dam and the reach of documented channel narrowing include two large tributaries, many smaller tributaries, and the sand bed and banks of the Green River in the 30-km long alluvial reach in Browns Park. In the first 30 km downstream from Flaming Gorge Dam, the Green River flows in a confined bedrock canyon and the bed is primarily gravel.
The purpose of this project is to test one method that may be used to estimate the sediment contributions of ungaged tributaries to the Green River downstream from Flaming Gorge Dam. The approach applies the Universal Soil Loss Equation (USLE) in a semi-distributed manor using GIS for analysis and presentation. Sediment production data predicted by the model will be compared with measured sediment transport data, allowing estimation of a sediment delivery ratio. Red Creek was chosen for the study because it is the only tributary between the dam and the Canyon of Lodore in Dinosaur National Monument for which hydrologic and sediment transport data are available.
Red Creek drains approximately 397 km2 (153 mi2) and flows into the Green River 18 km (11 mi) downstream from Flaming Gorge Dam (Figure 1). The relief in the basin is 1,242 m between the confluence with the Green River and the top of Pine Mountain at the far eastern edge of the basin. The geology of the basin is dominated by highly erodible Cretaceous and Tertiary shales and sandstones.
Figure 1. Map of Study area showing
location of Red Creek drainage basin.
Only 8 km before joining the Green River, does Red Creek encounter the highly resistant Precambrian quartzites of the Uinta Mountain Group and the Red Creek Formation. For these last 8 km the stream flows through a canyon deeply incised into these rocks. Deposits at the mouth of Red Creek indicate that sediment is transported through this canyon by debris flow processes during intense summer storms (Figure 2).
Figure 2. Aerial photograph of confluence of Red Creek
and Green River. The Green River is flowing from top to bottom of
photo.
Streamflow data are available for the US Geological Survey station Red Creek near Dutch John, UT (station no. 9234700) and only for a six year period from 1971 to 1976. Peak flow and suspended sediment data are available for this same period. Average flow is 0.2 m3/s and the average flood is 17 m3/s, with the maximum flood recorded during the brief period of record being 41 m3/s. These data demonstrate the episodic pattern of runoff on Red Creek and the dominance of summer and occasional winter large events (Figure 3). Nearly all sediment is transported during these large runoff events.
Figure 3. Daily sediment load and mean daily discharge of Red Creek near
Dutch John, UT for the period of record.
The universal soil loss equation (USLE) was developed in the 1950's by the US
Department of Agriculture for the purpose of estimating soil loss due to erosion
on agricultural fields. The equation is empirical and based on over 10,000
plot-years of data gathered from runoff-erosion studies on small agricultural
plots, under both natural and simulated rainfall conditions (Wischmeier and
Smith, 1978, Renard, 1997). The standard plot on which the data for the
USLE is based is 22.1 m in length with a uniform slope of 9% and the soil
surface in a "continuous clean-tilled fallow" condition (Wischmeier and Smith,
1978). The factors of the equation are designed to adjust for deviation
from that standard condition. The equation is, therefore, well grounded in
field data for agricultural settings but not designed for application to
undisturbed soils and landscapes at the watershed scale. The equation has,
however, been used for these types of studies because it is easy to use and
there are very few alternatives (Wilson, 1996, De Roo, 1998). The
components of the USLE are described below:
The Universal Soil Loss EquationRainfall-Runoff Erosivity (R)A = RKLSCP,
where,
A = soil loss per unit area (t/ha)
R = Rainfall-Runoff erositivity (Mj*mm/ha*h*yr)
K = Soil erodibility (t*ha*h/ha*Mj*mm)
L = Slope length factor
S = Slope steepness factor
C = Crop (land cover) factor
P = Practice (erosion controls) factor
Figure 4. Isoerodent map of the western United States. The
value of the contour line near Red Creek is 10 hundred
(ft*tonf*in/ac*h*yr). This value is multiplied by 17.02 to obtain the
metric value in (Mj*mm/ha*h*yr). From Renard et al. (1997).
Soil Erodibility (K)
Soil erodibility is a lumped parameter
that represents an integrated average annual value of the total soil and soil
profile reaction to a large number of erosion and hydrologic processes (Renard
et al., 1997). Like the rainfall erosivity factor, the soil erodibility
factor is determined from long-term measurements at standard soil plots.
From these plots, scientists have derived relationships between the K factor and
soil properties. Values for the K factor determined by soil scientists are
included in the STATSGO GIS soil coverages. The K factor is estimated for
each soil layer of every soil component and is therefore included in the
soillayers file in the STATSGO database. I averaged the value for
the upper two soil layers, which were usually nearly the same, and used it
for that soil component. I then applied a weighted average of these K
factors for each soil component based on the comppct field in the soil
components file in the STATSGO database to determine an average value for a
soil map unit. These values were entered in a new field for the K factor
in the soil map unit coverage for the Red Creek basin. This coverage was
clipped from the STATSGO coverages for Utah and Wyoming that had been projected
from geographic coordinates to UTM Zone 12 (Figure
5). The shape file was converted to a grid for use in the map
calculator. The K factors provided in the STATSGO database are in english
units of (t*ha*h/ha*Mj*mm), which are multiplied by 0.1317 to obtain SI units
of, (t*ha*h/ha*Mj*mm).
Figure 5. Soil units of the Red Creek Basin. The soil units
are shown by the red outline. The color shades are elevation intervals
from the 30m DEM for the Red Creek drainage basin. The location of the
gaging station Red Creek near Dutch John, UT is indicated by the green and black
bullseye. The drainage network is from the EPA river-reach file 3.
Slope Length and Steepness (LS)
The slope length and
steepness factors are typically represented as a combined topographic
factor. The topographic factor in the original formulation of the USLE is
based on LS equal to 1 for the conditions of the standard soil plot. The
USLE reference manuals contain tables that contain the ratio of soil erodibility
under field conditions to erodibility of the standard plot, all other factors
being equal (Wischmeier and Smith, 1978). These tables are based on the
gradient and length of the field plot and are not suited to values determined
from DEM data. Moore and Burch (1986) described a formulation of the LS
factor based on unit stream power that can be calculated from a DEM:
Slope and specific catchment area were calculated from the 30 m DEM of the Red Creek basin using the SinMap extension for ArcView and the TARDEM program, respectively (Pack et al., 1998; Tarboton, 1999). The LS factor is most strongly weighted by slope as shown by comparison between maps of slope (Figure 6) and the LS factor (Figure 7) for the Red CReek basin.LS = (a/22.13)0.4 (s/0.0896)1.3
where,
a = specific catchment area, and
s = slope
Figure 6. Map of slope in the Red Creek basin. Steepest
slopes are shaded blue.
Figure 7. Map of the LS factor for the Red Creek basin.
Regions of greatest LS factor are shaded red.
Land Cover Factor (C)
The land cover factor is the ratio of
soil loss under specified field conditions to the corresponding loss from the
standard soil plot. The USLE reference manual contains tables of values
for for the C factor for undisturbed forest and range land (Wischmeier and
Smith, 1978). Using these tables, I determined the C factor for the
different land cover classifications contained in the land cover GIS coverage
for the Red Creek basin. This coverage was obtained from the EPA BASINS
database. I added the C factors value to a new field that I added to the
land coverage shape file (Figure
8). The shape file was then converted to a grid to be used in the map
calculator (Figure
9).
Figure 8. Table of C factor (Ave_C_Fact column) values
determined for the land cover categories (Level2 column) of the Red Creek
basin.
Figure 9. Land cover categories of the Red Creek basin from the
Land Use/Land Cover coverage in the EPA BASINS database. The C factor
values for each land cover type are shown in Figure 8.
Calculation of Sediment Production
Sediment production per
hectare (the SI units of the parameters) was calculated by multiplying the
USLE factors using the map calculator in ArcView. The P factor for soil
conservation practices was assumed to be 1, which is common for non-agricultural
applications of the USLE (Molnar and Julien, 1997). The result was
multiplied by 0.09 hectares per grid cell to obtain the result in sediment
production per grid cell.
Ignoring sediment deposition
Rainfall is represented uniformly over
entire catchment
Hillslopes are represented as “uniform slope facets” in
original model
Soil erodibility and land cover are adequately represented in
the available soil and land cover GIS coverages
Sediment Production
The pattern of sediment production
strongly reflects the LS factor and slope steepness maps (Figure
10).
Figure 10. Sediment production per grid cell in the Red Creek
basin. The values shown are the logarithm of soil loss to better
illustrate the areas of greater sediment production, which correspond with the
LS factor. Areas of greatest sediment production are shaded
red.
Sediment Delivery Ratio
The total sediment production for
the basin was determined by the mean soil loss for the basin by the number of
cells in the basin. This was calculated separately for the entire Red
Creek basin and for the area above the Red Creek near Dutch John gage
station. The entire basin had a mean soil loss of 0.19 Mg per cell for a
basin of 441,486 cells resulting in a total sediment production of about 82,000
Mg. The basin above the gage had the same mean soil loss and a basin of
405,283 cells resulting in a total sediment production of about 76,000 Mg.
The sediment delivery ratio was calculated as the ratio of the mean annual
sediment load at the Red Creek gage to the total sediment production for the
basin above the gage. The mean annual sediment load for the Red Creek near
Dutch John gage was estimated by dividing the sum of the daily loads for the
period of record by the length of the record, in years. The mean daily
load is 209 Mg per day and the mean annual load is 75,000 Mg per year.
This results in a sediment delivery ratio of 0.99, which means the average mass
of suspended sediment transported at the gage is essentially equal to the mass
of sediment production on the hillslopes. Thus, the entire Red Creek
basin would be estimated to supply about 82,000 Mg of sediment annually to the
Green River.
The agreement between the estimated sediment production and the measured mean annual sediment load at the gage station is encouraging and indicates order of magnitude accuracy in the estimate. This agreement cannot, however, be taken to mean that there truly is a direct coupling between hillslope sediment production and sediment yield on an annual basis. Moreover, it is likely that, due to episodic nature of sediment transport processes that occur in the Red Creek drainage, there is significant storage of sediment throughout the basin and that this stored sediment is most likely transported to the Green River in pulses and not uniformly in time. No other data are available for the Red Creek basin with which these results can be compared.
The above results predict that approximately 200 Mg of sediment per km2 are eroded in the Red Creek basin and supplied to the Green River annually. This value is consistent with values for other similar basins on the Colorado Plateau and for the Colorado River basin in total (Webb et al., 1999; David J. Topping, US Geological Survey, personal communication). Thus, the estimated sediment yield based on the USLE is of a reasonable magnitude. Measurements of sediment load in other Colorado Basin tributaries should be analysed for additional comparison.
Sediment transport measurements are available for the Green River at Gates of Lodore, about 40 km downstream from Red Creek. These data are also generally consistent with the estimated contribution of Red Creek. Martin et al. (1998) calculated a suspended sediment transport relation, based on field measurements made in 1995-97. If the sediment-transport relation is assumed to be stationary, it estimates a pre-dam annual load of 186,000 Mg and a post-dam annual load of 60,000 Mg. In post-dam years with large floods, the load may be as high as 212,000 Mg. These values indicate that the estimated 82,000 Mg input from Red Creek is a significant part of the Green River sediment budget. With only minor inputs from other sources, the annual supply of sediment to the Green River downstream from Red Creek could easily exceed the capacity of the river to transport that supply during years of normal dam operations, which do not include floods. A transport limited system is consistent with the process of channel narrowing that has been documented for downstream reaches. Much additional research is needed to validate the results of the sediment yield estimated for Red Creek; and that research must include methods that can independently verify the result. Most importantly, these results demonstrate the utility of estimates of sediment yield and the potential importance of small tributaries to the sediment budget of large rivers, particularly those downstream from large dams.
Collier, M., Webb, R.H., and Schmidt, J.C., 1996, Dams and rivers: Primer on the downstream effects of dams, US Geological Survey Circular 1126, 94 p.
De Roo, A.P.J., 1998, Modeling runoff and sediment transport in catchments using GIS, Hydrological Processes, 12:905-922.
Grams, P. E., 1997, Geomorphology of the Green River in Dinosaur National Monument: Unpublished Masters Thesis, Utah State University, Logan, Utah, 140 p.
Martin, J.A., Grams, P.E., Kammerer, M.T., and Schmidt, J.C., 1998, Sediment transport and channel response of the Green River in the Canyon of Lodore between 1995-1997, including measurements during high flows, Dinosaur National Monument, Colorado: Final Report, US Bureau of Reclamation, Salt Lake City, 190 p.
Molnar, D.K. and Julien, P.Y., 1998, Estimation of upland erosion using GIS, Computers and Geosciences, 24:183-192.
Moore, I.D. and Burch, G.J., 1986, Physical basis of the length-slope factor in the universal soil loss equation, Soil Sci. Am. J., 50:1294-1298.
Pack, R. T., D. G. Tarboton and C. N. Goodwin, (1998), "Terrain Stability Mapping with SINMAP, technical description and users guide for version 1.00," Report Number 4114-0, Terratech Consulting Ltd., Salmon Arm, B.C., Canada (report and software available from http://www.engineering.usu.edu/dtarb/).
Renard, K.G., Foster, G.R., Weesies, G.A., McCool, D.K., and Yoder, D.C., 1997, Predicting soil erosion by water: A guide to conservation planning with the revised universal soil loss equation (RUSLE), Agriculture Handbook No. 703, U.S. Department of Agriculture, Washington D.C., 404 p.
Stanford, J.A., 1994, Instream flows to assist the recovery of endangered fishes of the upper Colorado River basin: Review and synthesis of ecological information, issues, methods, and rational, US Fish and Wildlife Service, FLBS Open-File Report 130-93, Denver, 89 p.
Tarboton, D.G., 1999, TARDEM, A suite of programs for the Analysis of Digital Elevation Data, http://www.engineering.usu.edu/dtarb/
Webb, R.H., Griffiths, P.G., Melis, T.S., and Hartley, D.R., 1999, Sediment delivery by ungaged tributaries of the Colorado River in Grand Canyon, US Geological Survey Water-Resources Investigations Report, in review.
Williams, G.P., and Wolman, M.G., 1984, Downstream effects of dams on alluvial rivers, US Geological Survey Professional Paper 1286.
Wilson, J.P., 1996, GIS-based land surface/subsurface modeling: New potential for new models, in Third International Conference/Workshop on Integrating GIS and Environmental Modeling, Sante Fe, New Mexico, USA, January 21-25, 1996. http://www.sbg.ac.at/geo/idrisi/GIS_Environmental_modeling/program.html
Wischmeier, W.H. and Smith, D.D., 1978, Predicting rainfall erosion losses: A guide to conservation planning, Agriculture Handbook No. 537, U.S. Department of Agriculture, Washington D.C., 58 p.