Channel-Pattern Adjustments and Geomorphic Characteristics of Elkhead Creek, Colorado, 1937–97

By John G. Elliott and Stevan Gyetvai

U.S. GEOLOGICAL SURVEY

Water-Resources Investigations Report 99 – 4098

Prepared in cooperation with the COLORADO RIVER WATER CONSERVATION DISTRICT

Denver, Colorado 1999 U.S. DEPARTMENT OF THE INTERIOR BRUCE BABBITT, Secretary

U.S. GEOLOGICAL SURVEY Charles G. Groat, Director

The use of firm, trade, and brand names in this report is for identification purposes only and does not constitute endorsement by the U.S. Geological Survey.

For additional information write to: Copies of this report can be purchased from:

District Chief U.S. Geological Survey U.S. Geological Survey Information Services Box 25046, Mail Stop 415 Box 25286 Denver Federal Center Federal Center Denver, CO 80225-0046 Denver, CO 80225 CONTENTS

Abstract...... 1 Introduction...... 2 Purpose and Scope...... 2 Acknowledgments ...... 2 Study Area ...... 3 Geographic Setting ...... 3 Hydrology ...... 4 Channel-Pattern Adjustments and Geomorphic Characteristics...... 4 Streamflow...... 4 Aerial Photography...... 7 Planimetric Geomorphology...... 11 Cross-Sectional Geomorphology...... 15 Sediment Size ...... 26 Shear Stress and Sediment Entrainment ...... 26 Sedimentation in Elkhead Reservoir...... 31 Factors Affecting Channel-Pattern Adjustments ...... 33 Summary and Conclusions ...... 37 References Cited...... 38

FIGURES 1. Map of Elkhead Creek showing study area, streamflow-gaging stations, and Elkhead Reservoir...... 3 2. Plot showing annual peak discharges and estimated bankfull discharge from streamflow-gaging station 09245000 Elkhead Creek near Elkhead, 1953–96...... 5 3. Digital orthophoto quadrangle of Elkhead Creek showing channel positions in 1937 and 1993, meander study sites, and cross-section locations, (A) reach upstream from confluence with the Yampa River, (B) reach immediately downstream from Elkhead Reservoir, (C) reach upstream from Elkhead Reservoir ...... 8 4. Channel centerline of Elkhead Creek meander 20.30 in 1937, 1953, 1970, 1977, 1981, and 1993 showing amplitude increase during the period 1938 through 1970 and translational shift during the period 1971 through 1993 ...... 14 5. Elkhead Creek cross sections showing geomorphic features, alluvial stratigraphy, and the bankfull stage, (A) cross section PT-2, (B) cross section PT-1, (C) cross section LG-2, (D) cross section LG-1, (E) cross section NR-Q, (F) cross section LW-1, (G) cross section VT-2, and (H) cross section VT-1 ...... 18 6. Plot showing cumulative size distribution curves for streambed sediment at Elkhead Creek study sites ...... 27 7. Plot showing cross-channel distribution of boundary shear stress for three discharges and the critical shear stress for entrainment of various sediment sizes at a typical Elkhead Creek site, cross section LG-1...... 29 8. Aerial photograph showing deltaic sedimentation in Elkhead Reservoir and the location of the delta distal margin in 1977, 1981, and 1993...... 32 9. Graph showing Elkhead Creek meander migration rates for time intervals of equal duration: (A) reach downstream from Elkhead Reservoir, and (B) reach upstream from Elkhead Reservoir ...... 34

TABLES 1. Streamflow characteristics at streamflow-gaging station 09245000 Elkhead Creek near Elkhead, Colorado, 1954 through 1996...... 6 2. Statistics from the digital processing and rectification of aerial photographs to the 1997 digital orthophoto quadrangles ...... 11 3. Elkhead Creek meander migration rates for time intervals determined from aerial photographs...... 16 4. Streambed sediment-size characteristics from Elkhead Creek study sites...... 28 5. Elkhead Creek delta progradation rates, estimated sediment volume, and extrapolated bedload-transport rates, 1975 through 1993...... 33

CONTENTS III CONVERSION FACTORS, VERTICAL DATUM, AND ABBREVIATIONS

Multiply By To obtain millimeter (mm) 0.03937 inch meter (m) 3.281 foot (ft) kilograms per square meter (kg/m2) 0.2048 pounds per square foot (lb/ft2) newtons per square meter (N/m2) 0.02088 pounds per square foot (lb/ft2) foot (ft) 0.3048 meter (m) mile 1.609 kilometers (km) acre 4,047 square meter (m2) square mile (mi2) 2.590 square kilometers (km2) cubic foot per second (ft3/s) 0.02832 cubic meter per second (m3/s) acre-foot (acre-ft) 1,233 cubic meter (m3) ton, short 0.9072 megagram (Mg)

Sea level: In this report, “sea level” refers to the National Geodetic Vertical Datum of 1929 (NGVD of 1929)—a geodetic datum derived from a general adjustment of the first-order level nets of both the United States and Canada, formerly called Sea Level Datum of 1929.

IV CONTENTS Channel-Pattern Adjustments and Geomorphic Characteristics of Elkhead Creek, Colorado, 1937–97

By John G. Elliott and Stevan Gyetvai

Abstract by impoundment and was used as the control reach. Mean meander migration rates in the Onsite channel surveys and sediment downstream study reach were 1.2 ft/yr from 1938 measurements made in 1997, aerial photographs to 1953, 2.5 ft/yr from 1954 to 1970, and 4.8 ft/yr taken from 1937 through 1993, and streamflow- from 1978 to 1993, compared to rates of 0.5 ft/yr, gaging-station record from 1954 to 1996 were 1.6 ft/yr, and 6.6 ft/yr for the same periods in the used to determine the probable cause of acceler- upstream study reach. ated streambed and streambank erosion in the Sediment and channel-geometry measure- lower reaches of Elkhead Creek, a perennial, ments and estimated hydraulic conditions at eight meandering tributary of the Yampa River. cross sections indicate that most of the sediment Concern about the possible effects of Elkhead sizes represented in the streambed are mobile at Reservoir, constructed in 1974, has been frequently occurring streamflows; those stream- expressed by landowners living downstream. flows are less than or equal to the bankfull Evidence cited as an indication of reservoir- discharge of approximately 1,800 to 2,200 cubic related effects include the trapping of bedload- feet per second. Discharge data from 1954 transported sediment in the reservoir, vertical through 1996 recorded at a site upstream from incision of the streambed, and lateral erosion the reservoir were examined to determine the causing loss of agricultural land. A large deltaic effect of hydrology on meander migration rates. deposit composed of approximately 163 acre-ft of The discharge data were assumed to be represent- bedload-transported sediment formed in Elkhead ative of the total streamflow and flood hydrology Reservoir between 1974 and 1993, the contempo- of both the upstream and downstream reaches rary bankfull stage of Elkhead Creek is several because Elkhead Reservoir normally has a feet below the elevation of a broad terrace that full pool. Mean annual streamflow increased previously was the flood plain, and lateral erosion 122 percent, and the mean annual flood increased at meander bends occurs at a higher rate than in 130 percent from the pre-regulation period (1954 previous periods at some locations. to 1970) to the post-regulation period (1978 to Elkhead Creek meander migration rates 1993), a possible explanation for much of the were used as a measure of lateral instability in the increase observed in meander migration rate in study reaches. Meander migration rates based on both the upstream and downstream reaches in the changes in channel centerline position were calcu- period after reservoir construction. lated for three periods from five sets of rectified Channel instability, quantified by meander aerial photographs for reaches upstream and migration rates, has increased throughout Elkhead downstream from the reservoir. The creek Creek since 1977. The most probable cause is a upstream from Elkhead Reservoir was unaffected combination of external factors affecting the

Abstract 1 entire watershed, such as changes in annual runoff tant as they relate to the effectiveness of potentially and flood magnitude and sedimentation in costly erosion-mitigation activities and irrigation Elkhead Reservoir. Local land-use practices, such diversion repairs. Results of the evaluation also will as intentional meander cutoff and riparian vegeta- further the understanding of channel-pattern adjust- tion removal, also can decrease channel stability, ments and bank-erosion processes. but these factors were not addressed in this study. Purpose and Scope

INTRODUCTION This report will identify some likely causes of recent streambank erosion in lower Elkhead Creek and The lower portion of Elkhead Creek is a mean- will provide information useful in guiding decisions dering tributary of the Yampa River in northwestern about mitigation activities where appropriate. The Colorado and flows through a relatively broad, alluvial following are specific objectives of the investigation: valley that has been utilized for ranching since the 1. determine the history of channel-pattern adjust- 19th century. Recently, several landowners along ments and the rate of meander-bend migration Elkhead Creek have expressed concern over the extent and cutoff before and since reservoir construc- and rate of streambank erosion and the loss of production; tive pasture acreage. Questions have arisen over the potential effects of Elkhead Reservoir, a 13,000-acre-ft 2. determine the contemporary depth of incision impoundment located a short distance upstream from below the flood plain; the eroding reaches, on sediment delivery and stream- 3. quantify the sediment characteristics of the flow fluctuations and, consequently, on channel streambed and the adjacent flood plain; stability and streambank erosion. A proposal to 4. examine the relation between peak runoff and enlarge the storage capacity of Elkhead Reservoir has channel-pattern adjustments resulting in stream- been challenged on the basis of concern that down- bank erosion; stream erosion rates could increase. Channel insta- 5. determine whether these channel adjustments are bility recently rendered the Smith Ditch irrigation unique to the reach downstream from Elkhead diversion from Elkhead Creek nonfunctional and has Reservoir or are common to all meandering compounded efforts to repair it. reaches of lower Elkhead Creek; and Many natural and human factors affect the 6. estimate the accumulation of bedload-transported behavior of a meandering alluvial stream. The sediment in the Elkhead Reservoir deltaic hydrology of Elkhead Creek is dominated by snow- deposit. melt and is complicated by the effects of reservoir impoundment and numerous irrigation diversions and returns. U.S. Geological Survey (USGS) topographic Acknowledgments maps and aerial photographs suggest that channel- pattern shifts and meander-bend avulsions were The study was made possible with the permis- common in the sinuous, lower 12 miles of Elkhead sion and cooperation of the many landowners and resi- Creek before Elkhead Reservoir was constructed in dents along Elkhead Creek. The authors are especially 1974. Cattle grazing in the watershed may have grateful to Jonathan Evans (USGS) for invaluable affected sediment yield, riparian vegetation growth, logistical support and assistance in the field and to and streambank morphology to an unknown extent. Jennifer Sieverling (USGS) for assistance in prepara- Consequently, the degree to which channel instability tion of aerial photographs for GIS analysis. R.E. and bank erosion rates are influenced by Elkhead Florida and Ed Neilson of the Natural Resources Reservoir, land-use practices, geology, sediment char- Conservation Service (NRCS) supplied some of the acteristics, climate variability, and hydrology has not cross-section survey data. Additional field assistance been established. The USGS, in cooperation with the was provided by Ray D. Tenney (CRWCD). Technical Colorado River Water Conservation District reviews of the report were provided by Janet S. Heiny (CRWCD), evaluated these issues, which are impor- and Paul J. Kinzel (USGS).

2 Channel-Pattern Adjustments and Geomorphic Characteristics of Elkhead Creek, Colorado, 1937–97 STUDY AREA Geographic Setting

Geomorphic data were collected from two prin- Elkhead Creek is a large, perennial tributary of cipal areas in the lower part of the Elkhead Creek the Yampa River in northwestern Colorado (fig. 1). watershed. A reach approximately 5 river miles long The Elkhead Creek watershed is located southwest of upstream from Elkhead Reservoir was selected to the Elkhead Mountains and has a drainage area of 2 represent relatively unregulated conditions in the about 250 mi . Elevations range between 10,500 and 6,220 ft. The geology of the headwater areas includes watershed and to identify trends in fluvial geomor- the sedimentary Iles Formation, the Williams Fork phology unrelated to reservoir effects. A reach approx- Formation, and the Lewis Shale of Cretaceous age; the imately 9 river miles long downstream from the sedimentary Fort Union and Wasatch Formations of reservoir was studied to identify any detectable reser- Tertiary age; and isolated Tertiary basaltic intrusives. voir effects. (A river mile is the in-channel distance The Elkhead Creek main stem and most of the prin- measured along a river or stream course.) cipal tributaries flow over shales and sandstones of the

107°30' 107°15'

k STUDY AREA r o F ER IV h YAMPA R R t PLATTE r Craig o

N RIVER 09245000 O D RA COLO DENVER

SOUTH Grand Junction Colorado Springs DOLORES G U ISON NN R AR KANSAS RIVER Pueblo RIO G R RANDE ek Cre

L d a o 40°37'30" e n h g k l

G E Ralph White u

Lake h Dr Figure 3C y F o rk Figure 8 09246200 voir ser Re 13

789 d a e h k l

E k e e r Figure 3B C

n M io t a EXPLANATION ca y ifi 09246400 n rt a Fo Figure 3A r 09246400 d Streamflow-gaging station G

u and number l c h

40 0 1 2 3 MILES Craig River pa 0 1 2 3 KILOMETERS Yam

40°30'

Figure 1. Map of Elkhead Creek showing study area, streamflow-gaging stations, and Elkhead Reservoir.

STUDY AREA 3 Upper Cretaceous Lance Formation (Tweto, 1976) and station 09246200. The second new gage, station Holocene alluvium. The valley gradient in the lower 09246400 Elkhead Creek below Maynard Gulch, is 16 miles of the watershed ranges from 0.0045 to located approximately 3.3 river miles downstream 0.0022 ft/ft. from the dam and 5.9 river miles upstream from the Elkhead Creek is a meandering stream in the confluence with the Yampa River. The downstream lower part of the watershed with sinuosity ranging gage at station 09246400 records runoff from approxi- from 1.1 to 2.2 and averaging 1.6. Oxbow ponds, mately 85 percent of the watershed, which includes meander scrolls, and arcuate groves of cottonwood outflow from Elkhead Reservoir. trees in the valley of Elkhead Creek are evidence of Annual peak discharges vary widely in magni- previous meander development and channel-pattern tude and reflect the water content of the preceding adjustment typical of meandering stream systems. The winter snowpack and the spring weather patterns. streambed is composed predominantly of gravel and Annual peak discharges recorded at station 09245000 cobbles, and sand deposits cover some streambank and have ranged from 224 ft3/s in 1992 to 2,850 ft3/s in point-bar surfaces. Valley-fill material exposed in 1984 (fig. 2); both extremes occurred subsequent to cutbanks is typical of fining-upward flood-plain sedi- construction of Elkhead Reservoir in 1974. The mean ment; gravel and sand strata are overlain by a thick, annual peak discharge at this station was 1,028 ft3/s. loamy deposit. Annual streamflow, the total runoff from the water- The valley floor and adjacent hillslopes in the shed in a year, ranged from 12,010 acre-ft in 1977 to lower watershed have been used for grazing and hay 82,380 acre-ft in 1984. The mean annual streamflow production since the late 19th century (Wes Signs, for 1954–96 was 40,200 acre-ft (Crowfoot and others, Colorado River Water Conservation District, oral 1997). commun., 1998). The existing riparian vegetation is dominated by cottonwood trees, willows, and other woody shrubs. Elkhead Reservoir was constructed in CHANNEL-PATTERN ADJUSTMENTS AND 1974 and has a normal pool capacity of 13,000 acre-ft. GEOMORPHIC CHARACTERISTICS The reservoir is located 9 river miles upstream from the mouth of Elkhead Creek. A combination of data from the streamflow- gaging stations, aerial-photographic interpretation, and onsite measurements was used to quantify channel- Hydrology pattern adjustments and describe geomorphic characteristics in two reaches of Elkhead Creek. Elkhead Creek is a snowmelt-dominated stream with annual discharge peaks usually occurring from late April to late May. The USGS has operated six Streamflow streamflow-gaging stations at various times on the main stem and on the North Fork, a large tributary of Peak flow frequency and the duration of daily Elkhead Creek. Station 09245000 Elkhead Creek near mean discharges were calculated by standard USGS Elkhead, in the middle part of the watershed, has the procedures (U.S. Interagency Advisory Committee on longest period of record (1953–96) and is used for Water Data, 1982) with the record from station hydrologic reference in this study. This gage, which 09245000 Elkhead Creek near Elkhead (fig. 1). The was discontinued after water year 1996, recorded 100-year flood at this location of the watershed was runoff from approximately 27 percent of the water- 2,580 ft3/s and the 10-year flood was about 1,700 ft3/s shed. (table 1). Based on discharge record analysis, observa- Two new gaging stations were established in late tions made during high streamflow in 1998, and 1995 to replace the discontinued station 09245000. commonly accepted statistical properties for the bank- Station 09246200 Elkhead Creek above Long Gulch is full discharge, 950 ft3/s (the 2-year flood) was identi- located approximately 10.6 river miles upstream from fied as a reasonable approximation of the bankfull Elkhead Dam and records most of the inflow to the discharge in Elkhead Creek at station 09245000. This reservoir (fig. 1). Runoff from approximately 68 discharge, as an instantaneous annual flood peak, was percent of the watershed passes the upstream gage at equaled or exceeded in 25 of the 44 years of record

4 Channel-Pattern Adjustments and Geomorphic Characteristics of Elkhead Creek, Colorado, 1937–97

53–96. 1993

1981 1977

/s 3 constructed Reservoir /s

WATER YEAR 1970

Annual instantaneous discharge peak, ft Estimated bankfull discharge, 950 ft Aerial photography years 1953 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 0 Annual peak discharges and estimated bankfull discharge from streamflow-gaging station 09245000 Elkhead Creek near Elkhead, 19 Elkhead, Creek near Elkhead 09245000 station from streamflow-gaging discharge bankfull and estimated peak discharges Annual 500

3,000 2,500 2,000 1,500 1,000

DISCHARGE, IN CUBIC FEET PER SECOND (ft /s) (ft SECOND PER FEET CUBIC IN DISCHARGE, 3 Figure 2. Figure

CHANNEL-PATTERN ADJUSTMENTS AND GEOMORPHIC CHARACTERISTICS 5 Table 1. Streamflow characteristics at streamflow-gaging from the long-term gage several miles upstream station 09245000 Elkhead Creek near Elkhead, Colorado, (09245000) and the new gage above Long Gulch 1954 through 1996 (09246200) resulted in a coefficient of determination 2 [Recurrence interval in years equals reciprocal of (R ) of 0.92. exceedance probability; streamflow duration, in percentage The time difference between diurnal peaks at the 3 of time specific discharge is equaled or exceeded; ft /s, long-term gage and the new gage below Maynard cubic foot per second] Gulch (09246400) was approximately 9–12 hours. The Peak flow frequency Streamflow duration regression between data from the long-term gage and Percentage the new gage below Maynard Gulch resulted in a coef- Recur- of time Discharge 2 rence Discharge ficient of determination (R ) of 0.87. A weaker corre- equaled or (ft3/s) interval lation between these station records resulted when exceeded 1.05 403 95 1.20 data from the entire year were used in the regression 1.11 493 90 2.30 analysis. Small, low-elevation tributaries introducing 1.25 623 85 2.30 runoff to Elkhead Creek between the long-term gage 2.00 950 80 3.31 and the new gages before the annual spring peak may 5.00 1,402 75 3.71 have contributed to the weaker correlation at low 10.0 1,697 70 4.20 discharges. 25.0 2,060 65 4.59 Some of the effects of Elkhead Reservoir on the 50.0 2,324 60 4.98 hydrology of Elkhead Creek can be seen in the brief 100 2,580 55 6.00 streamflow record at the two new gaging stations. 200 2,832 50 6.94 Elkhead Reservoir is approximately 3.8 miles long and 500 3,160 45 7.92 has a normal pool capacity of 13,000 acre-ft. The 40 9.06 reservoir is operated in such a way that it is normally 35 11.4 full or nearly full. Spring snowmelt runoff generally 30 14.2 flows into and through the reservoir with a short resi- 25 21.1 dence time. Station-to-station linear regression 20 37.8 between daily mean discharges recorded in April 15 83.3 through September 1996 at the gage on Elkhead Creek 10 175 above Long Gulch (09246200) and the gage below 2 5 342 Maynard Gulch (09246400) resulted in an R of 0.98. Although there is a very strong correlation between the stations in daily mean discharge (the streamflow aver- (fig. 2). As a daily mean discharge, 950 ft3/s was aged over a 24-hour period), there are notable differ- equaled or exceeded approximately 0.73 percent of the ences in the diurnal range and volume of streamflow. time on average, or about 3 days per year. Instantaneous diurnal discharge peaks recorded Peak flow frequency and flow duration statistics at the Maynard Gulch gage downstream from the have not been calculated for the two new gaging reservoir lag the peaks at the Long Gulch gage stations on Elkhead Creek—09246200 upstream from upstream from the reservoir by about 7 hours. The the reservoir and 09246400 downstream from the diurnal range in discharge and the instantaneous daily reservoir—because of inadequate record lengths. discharge peaks tend to be less at the downstream However, an approximate correlation was made Maynard Gulch gage than at the upstream Long Gulch between streamflow at the long-term gaging station gage; however the daily mean discharge, adjusted for (09245000) and streamflow at the two new gaging traveltime, tends to be greater at the Maynard Gulch stations based on 1 year of concurrent record. Daily gage downstream from the reservoir. The attenuation mean discharges during April through September were of diurnal peaks recorded at the Maynard Gulch gage regressed for the one concurrent year (1996). No probably is due to streamflow passing through traveltime lag was used in the regression, but the time Elkhead Reservoir. The greater daily mean discharges difference between diurnal peaks at the long-term gage recorded at the downstream Maynard Gulch gage and the new gage above Long Gulch was approxi- probably are due to additional runoff contributed by mately 2–4 hours. The linear regression between data tributaries entering between the two gages. The

6 Channel-Pattern Adjustments and Geomorphic Characteristics of Elkhead Creek, Colorado, 1937–97 drainage area of Elkhead Creek increases by graphs. A DOQ map is a photographic image in which 24 percent from the Long Gulch gage to the Maynard all points have been made planimetric and have been Gulch gage, an increase representing 17 percent of the rectified to a standard map-projection coordinate total Elkhead Creek watershed measured at the mouth. system (latitude/longitude, Universal Transverse Mercator, and so forth). The 1937, 1953, 1970, 1977, and 1981 aerial photographs were registered and recti- Aerial Photography fied to the 1997 DOQ images before geomorphic measurements and analysis were conducted. Digital Aerial photographs of the Elkhead Creek water- image processing and geometric correction are shed spanning several years reveal the channel condi- described in greater detail by Jensen (1996). tion and position at discrete times. Year-to-year comparisons of the channel position in photographs Digital image processing, rectification, and rectified to remove optical distortion allows quantifi- registration to the 1997 DOQ maps were done with cation of channel change during the periods between ArcView Spatial Analyst software (Environmental the photograph dates. Meander scrolls, oxbow lakes, Systems Research Institute, 1996). Several semiper- and other landforms visible in these photographs also manent features such as buildings, road intersections, provide a means to interpret geomorphic conditions and trees, visible in all photographs and in the DOQ, from a period before the first photographs were taken. were used as rectification control points. These points Gurnell (1997) described a method for quantifying were at elevations similar to the stream elevation, temporal and spatial river channel changes using aerial thereby reducing the potential for rectification error photography, onsite cross-section surveys, and a caused by relief displacement (Gurnell, 1997). The geographical information system (GIS). original aerial photographs were rectified to the scale Aerial photographs (fig. 3) were used to deter- and coordinates of the DOQ image with second-order mine the channel position in 6 specific years and were polynomial affine transformation functions fitted to the basis for calculating meander migration rates and the control points using a least-squares criterion deltaic sedimentation rates. Photographs selected for (Jensen, 1996, p.127). Initially, the fit between the the Elkhead Creek geomorphic analysis were taken in rectified image and the DOQ standard was evaluated 1937, 1953, 1970, 1977, 1981, and 1993. Most of the qualitatively by observing the degree of registration of photographs were taken in August or September when the control points and other semipermanent geo- seasonal streamflow had receded to low levels. Photo- graphic features, such as roads and fence lines, on an graphs taken in 1953 and 1970 bracket a period of overlay of the rectified image and the DOQ. In addi- relatively small flood peaks when only about one-half tion, the fit was evaluated quantitatively with root- χ2 of the annual flood peaks exceeded the estimated mean-square error (RMSE) and chi-squared ( ) bankfull discharge (fig. 2). Photographs taken in 1977 statistics calculated by the ArcView Spatial Analyst and 1993 bracket a comparable length of time in which extension. The RMSE is an estimate of the standard annual flood peaks exceeded the estimated bankfull deviation of the dependent variable, in this case the discharge about the same number of times as in the plotting position of the control points in the rectified 1953 through 1970 time period, but with much greater image. The RMSE describes the deviation between the magnitude. The photographs taken in 1977 and there- control-point locations on the rectified photograph and after also document the period since construction of the control-point coordinates calculated by the trans- χ2 Elkhead Reservoir in 1974. formation function. The statistic is a measure of the The Elkhead Creek aerial photographs were association between two distributions and, in this anal- enlarged to 18 × 18-inch prints having scales of ysis, is strongly correlated with RMSE. approximately 1:6,000. The enlarged photographs The initial photograph-rectification step were digitally scanned, and these images were elec- produced a reasonably good fit to the DOQ, but for tronically processed to remove optical distortion and some photographs the fit could be improved by elimi- photographic parallax common to many aerial photo- nating one or more specific control points with large graphs. Digital orthophoto quadrangle (DOQ) maps of individual RMSE and χ2. Selected control points were areas in the Elkhead Creek watershed were created in removed iteratively, and the unrectified image was 1997 by the USGS from 1993 National Aerial Photog- rectified again until the stepwise decrease in total raphy Program (NAPP) high-altitude aerial photo- RMSE and χ2 variables became negligible. Table 2

CHANNEL-PATTERN ADJUSTMENTS AND GEOMORPHIC CHARACTERISTICS 7

, d s a n e e m e o i n n a h t i i n l l k e a l r r o r c i E t r t

e e o l t t e s

e c t m i n n n v p e o i o s r e e i u f s

R c c c y

s s m l l e O d a s a s e I e e - l u e p i o s T r t n n t r s s A s n n m m N c o

p r

r a a a l N r c u

e Y h h e e

A d h c c d v n n c L i

m n a a r n

3 7

P e , o

a a r

s 9 3 r

n k X i f e

) h e

9 9 t e i

E C

( 1 1 M C s e

y r r i d 1 o - u v

t C r T s 0

e r V 4 s . e e 0 d 0 R 7 n

a 9 d e a 1 e

h , d k n l 3 n 9 E o

a i 9

s 1 m l

3 o d r u 5 f n

v 9 a

m a 1

7 a

r 3 e T n r 9 e t E 1 e s d

E n

e 1 n n - i F

w w a s o

t W n 0 e

L d

e o 0 i y t l M b 0 i , e s t 2 o a i p 0

d l 0 e e 5 n , m n 1 m a i

0 h h c 0

c 0 g , a n e 1 i r

w ) o 0 B h ( 0

s , 5

r k e e v i e r R

0 C

d p

a a m

a e h

Y k h l

e k E l

h f

E o

h t e i l g w

n e a c r r n d

e e a u

v u l

i f q

R o o t c o

n h m p o i o o r s f h

l t r u m o

a v 7 l e a a 3 r t

t i

9 r a s g i

p 1 e p

D u .

r . e i h

n r m 3 o c

A a o a e a r f r e e e

r e Y u

s ) g M b e i A F ( R

8 Channel-Pattern Adjustments and Geomorphic Characteristics of Elkhead Creek, Colorado, 1937–97 - r e s e R

, d s a n e o i h t

h k a T lc l

u c E G E

o l E

m n F o o r i f t

0 c 0 m e

d 0 s a

, r - e

2 s r a t s

s 0 n o

p r 0

y c u 5 a

d h 1 M 3 n c a a 0 4

. e , 0 r s

8 0 ) , e t i

1 C ( s

, y r 0 i d o 0 u r v t 5 r i s d

e r o s a e e v e d r 0 R n h e a d e k s a l e m e

h h , E k R l 3 c 0 t 9 E i

n 9 0 . 1 o D m

7 o d s r h f n r t

i a e

m 7 v a m i 3 e r S d 9 0 t 3 1 s 6

5 . n n i 9

7 w s 1 o

n d

o d i y t n l i n e s o 7 t a i o

a 1 i s p 5 . l

7 d l 8 6 u e e 3 k . n v m 9 e 6 n a e 1 m

r i

r C h n h c e

d c

e a g d

e a n e h e n lk i E r

w a w ) t o

e B h e (

s , M b

r k n e e v i o e i e e m r t R 3

n n C a a i i

a 2 t l l e d . p r r r s a t 6

r 0 m e e e t t e s 0 a g e h t i 4 n n Y v p k n

l i

i 6 s e e

u e

E 4

N g

R c c h

f 2 y t

s a

r l l o 9 O d

a h e I g 0 e e e t l e i u - l p i T t b n n g w

w s A n n n m m N 3

e m

o a r a a a c l N r r 5 u f n e d Y h h e 9

A n e a c c d v m 1 L u i

u l

m d n f r a q 3 7

d n o n a e o n r 9 3 n o t r n X i f a e c t o 9 9 o

a i

E h

1 1 M S s m p l 7 o o r u 3 f h

t v 9 r m a o 1 3

. a 0 l

r d e 2 0 a r n . t e e t i 4 e s u 6 g 6 d i p n 4 e i n D t u 2

. n w a 9 h t 3 o 0 e c

e C e B a r M b e r — u

r i ) g i A o F ( v

CHANNEL-PATTERN ADJUSTMENTS AND GEOMORPHIC CHARACTERISTICS 9

m - h o a r y f c c

l r o l m

u a n D e o i r G t t c s T p e E s u -

E s h F s c o a r 0 e c r

) d 0 , n C ( 2 a

, , r 0 i s 0 o e t v 5 i r , s

e 1 y s 0 d e u 0 R t

0 s , d

r 1 a e e d h 0 n k l 0 a E e 5

m m

, o r 3 f

9 0 9 m 1 a

e d r n t s a

k o 9

e d 1

o e

i r y

e e n m l t n i

e n n C a a t o s i i n i t l l a e n i r r o s s r i o l d

t i r t e e t e i u t t e s g e c t s

v 7 i m n n v p n

e o i i s d a 3 e e m

p u s

a l R 9 c c r

h e s a s r l l e 1 e O c d n a

s e I g

e e e a l h d n - u p i e

o b T e t

a n n N n r k r r

h s A n n l m m ) a o c

m c

o f

l a a a B l N r e u E ( f g e

e Y h h e e , n n

A i M b r

c c d v m n L i e

w d m n a r v n o i 3 7

P n o a e h a R r 9 3 n r s

X a i f e

t h

9 9 a k

E p S 1 1 M C e e m r a C 1 Y

- d 0 e a . G h t e 0 0

L h 0 h 2 k t 2 l i 6 E w

4 f e 2 o

c 9 e n 0 l e g u n l f a n r o d c . a

d u m e q

o u 0 r o n f t 0 i

t o 2 n m h ch 6 l o a u p G 4 e

g o C 2

n r h o t 9 L t — s r 0 r i p o

o u l

v a r h t i c e g s i a e e D r

. R )

A d ( e a

r ,

C e s u h n k g l i o i F t E

10 Channel-Pattern Adjustments and Geomorphic Characteristics of Elkhead Creek, Colorado, 1937–97 presents summary statistics from the photograph- ArcView Spatial Analyst extension (Environmental rectification procedure. Systems Research Institute, Inc., 1996). Use of from 7 to 14 control points near to, and at elevations similar to, the stream, and use of second- Planimetric Geomorphology order polynomial rectification functions produced reasonably good fits of the rectified photographs to Planimetric geomorphic characteristics of the DOQ (table 2). The range of RMSE (0.360 to Elkhead Creek were determined using 1971 topo- 2.613 m) is within the error range noted by Gurnell graphic maps and six sets of aerial photographs span- (1997). Consequently, there was a high degree of ning several decades. Stream-channel and valley gradients over distances of several miles were calcu- confidence that the apparent channel-position change lated from USGS 1:24,000-scale quadrangle maps observed during the five photograph-defined periods with 20- and 40-ft topographic contours. Valley represented a true change on the land surface. After gradient, calculated with this resolution, was relatively rectification, channel locations from all 6 years were uniform and ranged from 0.0022 to 0.0045 ft/ft. digitized and plotted on a single map. Changes in the Stream-channel gradient ranged from 0.0010 to stream channel position were measured with the 0.0028 ft/ft.

Table 2. Statistics from the digital processing and rectification of aerial photographs to the 1997 digital orthophoto quadrangles

[Photographic images rectified with second-order polynomial affine transformation function; Upper reach, upstream from Elkhead Reservoir; Dam reach, immediately downstream from Elkhead Dam; Lower reach, upstream from confluence with Yampa River; Root mean square error, deviation between control point location on the rectified photograph and the control-point coordinates calculated by the transformation function, calculated for the Easting (x) and Northing (y) directions]

Root mean Root mean Number of square square Chi Chi Photograph date control error, error, square, square, and reach points Easting Northing Easting Northing (meters) (meters) 1937 Upper reach 9 1.149 0.519 11.877 2.424 1937 Dam reach 10 0.745 1.579 5.546 34.923 1937 Lower reach 7 1.161 2.613 9.443 47.777

1953 Upper reach 10 2.346 1.759 55.015 30.955 1953 Dam reach 9 1.313 1.148 15.527 11.835 1953 Lower reach 9 0.722 0.811 4.693 5.923

1970 Upper reach 8 0.525 0.559 2.208 2.501 1970 Dam reach 9 0.746 1.833 5.005 30.240 1970 Lower reach 6 0.745 1.578 5.546 24.923

1977 Upper reach 10 1.440 1.809 20.748 32.708 1977 Dam reach 10 0.523 0.808 2.735 6.530 1977 Lower reach 14 1.854 1.616 48.126 56.571

1981 Upper reach 10 0.360 1.097 1.274 12.833 1981 Dam reach 9 1.131 1.109 11.521 11.074 1981 Lower reach 12 0.909 2.025 9.016 49.215

CHANNEL-PATTERN ADJUSTMENTS AND GEOMORPHIC CHARACTERISTICS 11 Sinuosity, a descriptor of channel pattern easiest to detect. Ranchers, landowners, and resource defined as the downstream in-channel distance divided managers are sensitive to losses of productive land and by the down-valley linear distance, is spatially and observe the greatest change at these meander loca- temporally variable and is influenced by valley tions. gradient, streamflow magnitude, sediment load, and Meander bends included in this study were the mechanics of meander development (Schumm, selected to be representative of other meanders in the 1977; Schumm and Khan, 1972). Elkhead Creek sinu- valley. Seventeen meanders were selected in the reach osity, calculated from the 1971 topographic maps, downstream from Elkhead Dam (figs. 3A and 3B), and ranged from 1.08 to 2.57 and averaged 1.57 between 13 meanders were selected in the reach upstream from Dry Gulch and the confluence with the Yampa River. the backwater effects of Elkhead Reservoir (fig. 3C). A stream reach may have a constant average sinuosity, The upstream reach was considered the control reach, yet within the reach the channel position may change unaffected by Elkhead Reservoir (completed in 1974). substantially (for example, a meander cutoff in one Meanders selected for study generally were free to bend may be offset by meander amplitude increase in migrate through the Elkhead Creek flood plain; another nearby bend). however, some meander bends impinged against adja- Channel migration rates are strongly controlled cent valley side slopes from time to time. Valley side- by the unit stream power (essentially a function of slope impingement usually resulted in a directional discharge and channel gradient), bank erosion resis- change in meander migration, which is typical of tance (a function of sediment size, cohesiveness, strati- meanders throughout the upper reach and in the lower graphic relations, and vegetation characteristics), bank reach immediately downstream from Elkhead Dam height, meander-bend radius of curvature, channel where the valley side slopes are closely spaced. The width, and sediment transport (Hickin and Nanson, study excluded meanders that had been significantly 1984). Hickin and Nanson found that the maximum affected by human activity, such as those fixed in posi- channel migration rates occur where the ratio of tion by bridge works or those meanders intentionally meander radius of curvature to channel width, r/w, is cut off by mechanical means. between about 2.0 and 3.0 in gravel-bed rivers in Elkhead Creek meander migration rates were western Canada. In straighter reaches (r/w greater than calculated using a variation of the method developed about 3.0), channel migration usually is very slow. In by Hickin (1974) and Hickin and Nanson (1984). reaches with tight bends (r/w less than about 2.0), Hickin (1974) calculated meander migration rates channel migration rates decrease because flow breaks along erosion pathlines, or orthogonals, connecting the down into large eddies that markedly increase flow point of minimum radius of curvature of time-sequen- resistance (Hickin, 1974). Flow resistance in a tial meander bend locations. The erosional axis of the meander bend can be minimized as the channel gradu- meander was the orthogonal reflecting the greatest ally changes its cross-section form or bend radius of lateral migration rate of the meander on the flood curvature through bank erosion, or by abruptly plain. Hickin’s (1974) technique was developed using changing its position through a meander cutoff, or aerial photographs of western Canadian meander avulsion. Either process involves erosion and remobili- bends that had pronounced ridge-and-swale topog- zation of sediment that may be a valuable resource, raphy accentuated by riparian vegetation. The ridge- such as a pasture, hay meadow, road, or building site. and-swale features were formed by the stream as the Channel adjustments from 1938 through 1993 meander bend migrated laterally, forming a point-bar were studied in individual meander bends to assess complex. Meander bend location and morphology trends in erosion rates along Elkhead Creek. Meander from several periods before the photograph date were migration was measured along the erosional axis of the well preserved in the ridge-and-swale topography meander, generally at the apex of the meander, or visible in Hickin’s photographs. As a result, the where the meander bend radius of curvature, r, was a orthogonals Hickin plotted generally reflected a minimum. This is the location where the channel continuous progression of meander development migration rate usually is greatest (Hickin and Nanson, through time. 1984) and, although greater than the overall channel The Elkhead Creek meanders occur in a region migration rate through the entire valley, it is where where riparian vegetation and land-use practices do increases or decreases in channel migration rates are not accentuate the meander ridge-and-swale topog-

12 Channel-Pattern Adjustments and Geomorphic Characteristics of Elkhead Creek, Colorado, 1937–97 raphy preserved in point-bar complexes. Conse- Meander migration rates were determined for quently, only the active channel position visible in the five time periods defined by aerial photographs taken photograph was used in the Elkhead Creek analysis. in 1937, 1953, 1970, 1977, 1981 and 1993. The aerial Because some of the Elkhead Creek photographs are photographs were rectified to the 1997 DOQ, which separated by as much as 16 or 17 years of time, the was created from the 1993 photograph (fig. 3). The channel position changes quantified for this analysis rectification procedure is described in the preceding reflect average rates for the time interval. The channel section. The periods 1938 through 1953 and 1954 center line, rather than the eroding concave bank, was through 1970 are representative of the period before digitized from each rectified aerial photograph. The regulation by Elkhead Reservoir. The periods 1978 center line was more easily visible and more objec- through 1981 and 1982 through 1993 are representa- tively determined from the Elkhead Creek photo- tive of the regulated period (downstream from the graphs than was the concave bank. Also, reversals in dam). The period 1971 through 1977 encompasses the channel position (negative migration rates) were easier construction (1974) and filling of Elkhead Reservoir. to detect in the channel center line than along the Meander migration rates were calculated for concave bank. The error in locating the channel center 17 meanders in the reach downstream from Elkhead line partly is a function of the aerial photograph-recti- Dam and for 13 meanders upstream from Elkhead fication precision (table 2) and the resolution of fluvial Reservoir (table 3). The migration mode for most of features in the photograph. The latter error probably is these meanders in all time periods was amplitude less than one-half the width of the low-discharge water increase, although translational shift occurred in a few surface, which ranged from 31 to 50 ft when channel meanders in one or more periods. None of the mean- surveys were made in 1997. ders included in the migration-rate calculations under- Meander bends develop and migrate in different went avulsion. Negative meander migration rates were directions and modes. Hickin (1974) noted that cross- recorded in some time periods when the channel valley erosion dominated in early meander develop- reversed its lateral movement. An assumption was ment when bends had a low amplitude and large radius made that, by assigning negative values to position of curvature, whereas down-valley erosion dominated reversals, the net migration rate was a better reflection in older, higher amplitude meanders. Elkhead Creek of the overall, long-term disruption of the alluvial meanders included in this study were fully developed valley floor. For example, a meander that progres- meander bends in the 1937 and subsequent photo- sively increased in amplitude by eroding the concave graphs. Two predominant meander migration modes bank would have a significantly greater detrimental were observed in the study areas—amplitude increase effect on nearby land use than would a meander that and translational shift. Amplitude increase involved initially increased in amplitude but subsequently lateral migration, usually normal to the down-valley decreased in amplitude (through position reversal) by direction, near the point of minimum radius of curva- eroding recently deposited point-bar sediment on the ture, and was the most common migration mode (inner) convex bank rather than undisturbed valley-fill observed in Elkhead Creek. The meander amplitude material on the (outer) concave bank. increased, more or less, along an orthogonal that Meander-bend avulsions, or cutoffs, are another bisected the point bar and was typical of the migration natural mechanism of channel adjustment. An avul- of the Canadian meanders studied by Hickin (1974) sion shortens the distance between two points on a (fig. 4). When amplitude increase dominated, the meandering stream, and the immediate effect is to meander-bend limbs upstream and downstream from increase the local stream gradient, stream power, and the point of minimum radius of curvature, or the apex, boundary shear stresses that in turn tend to promote usually did not change position significantly. Transla- additional bed and bank scour. Over time, some of tional shift involved an abrupt change in migration these local effects are mitigated by additional adjust- direction at the meander-bend apex, generally ments upstream and downstream from the avulsion in changing from cross valley to down valley. Upstream gradient, sinuosity, channel cross section, and rough- and/or downstream limbs migrated laterally along with ness. Human-induced, mechanical cutoffs produce the location of minimum radius of curvature when the similar channel and pattern adjustments over time, meander bend underwent a translational shift (fig. 4). although the specific effects may be unpredictable.

CHANNEL-PATTERN ADJUSTMENTS AND GEOMORPHIC CHARACTERISTICS 13

e y e l l a v s s o r C h t

g n i r u d

e N y s e a l l e r a c n v i

e n d w u t i o l p 0 D 7 m a 9

g 3 1 n i 5 7 w 9 3 o 1 9 h f s 1

o 5 3

9 1 b . 9 1 0

2 m 7 i d 2 l

3 3 n r a 9 9

m , e 1 9 1 a d

d 8 1 1 e n n

n 9 i r 1 a o t e

t , T s e 7

b 7

E x p 3 7 7 r

M . E e 9 9 3 U e 3 1 F p 1 9

d 9 ,

a 1 9 0 n 0 n

1 7 i , 0

a 3 9 h 2 0 e 1 g 7

3 , 0 u . 7 M 3 o 5 0 5 r 9 1 9 h 2 1 t

0 r , 1 0 7 e 7 1 3 9 d 9 1

0 n 1

d 5 1 a n o i i

8 e r 0 e 9 0 3 M 5 p . 1

f 0 e 2 o h

t r

5 e b 3 g d e e 1 n 9 i t t t . n m s s r f f f i e i i i 9 a l a a u 0 d

e h h h 7 1 d e e 2

o r r s s s

t 3 m

f m

l l l c c r i M 9 k

a a a a n n h e i i e s 1

n n n n e

e l e e r o o o o r i i i i t a n n t t t t d d i C n s

a a a a a u u l l l o t t r d i n i i e t s s s l l a g i a n n n p p e w l M a a a h s M o m m r r r k n l A A T T T a D

E r

t f

o d

n 3 0 7 1 3 e a 5 7 7 8 9 n s

i r l – – – – – 0 r a 7 8 4 1 8 2 e t e 9 3 5 7 7 8 n Y 1 9 9 9 9 9

e 1 1 1 1 1 0 h c

g l 3

. e u

0 n o

n r

2 n

a r t

h e 1 2 3 4 5 o 8 C d 3

h . t n 9 r 4 a 1

O e e d r o M u i r g i e F p

14 Channel-Pattern Adjustments and Geomorphic Characteristics of Elkhead Creek, Colorado, 1937–97 Several meander avulsions occurred in Elkhead characteristics at cross sections PT-1 (fig. 5B) and Creek during the periods defined by the aerial photo- LG-1 (fig. 5D), the 1997 high water marks (peak graphs (figs. 3A and 3B), and two avulsions were discharge 2,760 ft3/s), and water-surface observations observed that occurred since 1993. The remnants of in the spring of 1998. This bankfull discharge estimate several avulsions that occurred before 1937 also could is applicable to all of the surveyed cross sections in be seen on terraces and the flood plain (figs. 3A and the study reach upstream from Elkhead Reservoir 3C). Some avulsions were human induced (Donald (figs. 5A–5E). Based on the limited streamflow record Van Tassel, local resident, oral commun., 1997); others at the new gage below Maynard Gulch (09246400), probably were the result of progressive channel adjust- the correlation between streamflow at the Long Gulch ments. The avulsion rate in the reach downstream from and Maynard Gulch gages (described in a previous Elkhead Reservoir was 4.6 avulsions per valley mile section) and channel surveys, the estimated bankfull per century for the 1938 through 1970 period and 2.1 discharge is between 1,800 and 2,200 ft3/s in the avulsions per valley mile per century for the 1971 downstream reach. Bankfull width of the eight through 1997 period. By comparison, the avulsion surveyed cross sections ranged between 67 and 122 ft rate in the reach upstream from Elkhead Reservoir was and averaged 88 ft. The surveys were conducted at low 2.0 avulsions per valley mile per century for the 1938 flow, approximately 17 ft3/s, and the average water- through 1970 period and 1.4 avulsions per valley mile surface width was 41 ft. The bankfull width to depth per century for the 1971 through 1997 period. (W/D) ratio ranged from 12 to 24 and averaged 16.5. Meander avulsion rate is not an accurate indicator of Bankfull water-surface slope ranged from 0.0012 to channel-pattern instability for Elkhead Creek because 0.0039 and averaged 0.0023. of the intentional cutoff of some meanders. However, Anecdotal evidence from local landowners and because adjustments to either natural or human- residents suggests the channel of Elkhead Creek may induced avulsions tend to propagate to nearby channel have incised in recent decades both upstream and segments, a higher incidence of avulsion may correlate downstream from Elkhead Reservoir (Donald Van with a higher rate of channel migration, bank erosion, Tassel and Leroy Lawton, local residents, oral or sedimentation in nearby reaches. commun., 1997). A broad terrace surface forms the valley floor along Elkhead Creek in the study reaches. This terrace surface is used for grazing and hay Cross-Sectional Geomorphology production throughout the valley. If this terrace surface is used as a very crude vertical reference to a former Channel-geometry characteristics and the water- flood-plain surface, an estimation of subsequent surface slope of Elkhead Creek were surveyed at eight vertical incision can be made. The contemporary cross sections using a total-station laser theodolite in bankfull surface ranges from 2.5 to 3.8 ft below the 1997 (fig. 5). Characteristics of sediment in the terrace surface at cross sections upstream from channel and the stratigraphy of the valley-fill material Elkhead Reservoir (figs. 5A–5E) and 2.9 to 6.2 ft were determined from onsite inspection during the below the terrace at cross sections downstream from surveys. Cross sections selected were representative of the reservoir (figs. 5F–5H). Buried, intact tree trunks the channel morphology upstream and downstream and dense root masses were observed in the valley-fill from Elkhead Reservoir. All sections except VT-1 material exposed by the cutbank at cross section LW-1 were located in relatively straight reaches; cross (fig. 5F), indicating vertical aggradation of a former section VT-1 was located in a meander bend. The flood-plain surface in this reach and subsequent inci- surveys allowed determination of bankfull width and sion to a lower elevation, possibly within the last depth, the depth of incision below the valley floor or century. No ages were obtained for the contemporary terrace surfaces, and the elevation of alternate bars and valley-floor terrace surface or the buried surface on point bars. Water-surface slope and cross-sectional which these trees were rooted; however, one of the characteristics were used to estimate boundary shear partially buried cottonwood trees (Populus deltoides) stress for selected streamflows. was still living in 1997. Bankfull discharge at the new gaging station Stratigraphy of the uppermost valley-fill mate- above Long Gulch (09246200) was estimated to be rial was recorded from cutbank exposures onsite and is approximately 1,800 ft3/s, based on surveyed channel summarized in the cross sections in figure 5. The

CHANNEL-PATTERN ADJUSTMENTS AND GEOMORPHIC CHARACTERISTICS 15 Table 3. Elkhead Creek meander migration rates for time intervals determined from aerial photographs

[River mile, channel distance upstream from Yampa River; Orthogonal distance, meander displacement between points of minimum meander bend radius of curvature; T, meander migration mode by translational shift, otherwise by amplitude increase; feet/year, feet per year; MA, mean annual; acre-ft/yr, acre-foot per year; ft3/s, cubic foot per second; nr, no record]

1938–53 1954–70 1971–77 1938–53 1954–70 1971–77 Meander Meander Meander Orthogonal Orthogonal Orthogonal River mile migration migration migration distance distance distance rate rate rate (feet) (feet) (feet) (feet/year) (feet/year) (feet/year) Meanders Downstream from Elkhead Reservoir 0.40 15.8 1.0 90.3 5.3 4.3 0.6 0.70 16.7 1.0 69.4 4.1 36.5 5.2 0.90 39.1 2.4 66.6 3.9 13.0 1.9 0.95 99.3 6.2 17.4 1.0 6.1 0.9 1.15 6.0 0.4 47.8 2.8 7.1 1.0 1.35 57.8 3.6 43.0 2.5 8.0 1.1 2.82 37.3 2.3 19.6 1.2 8.6 1.2 3.00 27.5 1.7 27.1 1.6 20.5 2.9 3.84 –29.7 –1.9 54.9 3.2 7.8 1.1 4.32 51.8 3.2 24.8 1.5 57.5 8.2 4.93 69.8 4.4 36.0 2.1 12.9 1.8 6.17 –18.4 –1.1 11.3 0.7 –23.2 –3.3 6.23 16.2 1.0 43.5 2.6 27.6 3.9 6.85 9.1 0.6 22.5 1.3 84.2 12.0 7.00 7.7 0.5 36.8 2.2 –26.6 –3.8 7.60 –44.3 –2.8 33.7 2.0 27.4 3.9 8.43 –26.2 –1.6 69.3 4.1 62.0 8.9 Mean migration rate 1.2 2.5 2.8 Standard deviation 2.4 1.3 4.0 MA flow, acre-ft/yr nr 36,062 39,956 MA flood, ft3/s nr 923 945

Meanders Upstream from Elkhead Reservoir 20.00 54.7 3.4 19.0 1.1 14.4 2.1 20.15 T 72.6 4.5 –29.2 –1.7 T 85.4 12.2 20.30 11.2 0.7 62.8 3.7 T 211.6 30.2 20.44 5.6 0.3 67.3 4.0 –24.4 –3.5 21.00 49.1 3.1 –65.1 –3.8 36.4 5.2 21.20 –40.6 –2.5 T 77.3 4.5 59.1 8.4 21.62 T –101.3 –6.3 120.9 7.1 85.6 12.2 22.22 31.5 2.0 –89.3 –5.3 41.5 5.9 22.48 30.4 1.9 –73.0 –4.3 89.7 12.8 23.06 16.9 1.1 0.0 0.0 92.7 13.2 23.35 –31.7 –2.0 54.3 3.2 –52.4 –7.5 23.60 0.5 0.0 95.9 5.6 –45.4 –6.5 23.85 7.1 0.4 118.8 7.0 –55.1 –7.9 Mean migration rate 0.5 1.6 5.9 Standard deviation 2.9 4.3 10.8 MA flow, acre-ft/yr nr 36,062 39,956 MA flood, ft3/s nr 923 945

16 Channel-Pattern Adjustments and Geomorphic Characteristics of Elkhead Creek, Colorado, 1937–97 Table 3. Elkhead Creek meander migration rates for time intervals determined from aerial photographs—Continued

1978–81 1982–93 1938–70 1978–93 1978–81 1982–93 Meander Meander Pre-regulation Post-regulation Orthogonal Orthogonal River mile migration migration meander meander distance distance rate rate migration rate migration rate (feet) (feet) (feet/year) (feet/year) (feet/year) (feet/year) Meanders Downstream from Elkhead Reservoir 0.40 25.4 6.4 47.6 4.0 3.2 4.6 0.70 15.8 3.9 62.0 5.2 2.6 4.9 0.90 13.4 3.3 26.3 2.2 3.2 2.5 0.95 19.6 4.9 31.6 2.6 3.5 3.2 1.15 0.9 0.2 33.1 2.8 1.6 2.1 1.35 15.8 3.9 66.7 5.6 3.1 5.2 2.82 –8.6 –2.1 48.6 4.0 1.7 2.5 3.00 64.0 16.0 21.1 1.8 1.7 5.3 3.84 34.0 8.5 19.2 1.6 0.8 3.3 4.32 14.5 3.6 T 127.8 10.6 2.3 8.9 4.93 4.2 1.1 58.5 4.9 3.2 3.9 6.17 66.3 16.6 –2.6 –0.2 –0.2 4.0 6.23 T 77.3 19.3 48.1 4.0 1.8 7.8 6.85 T 63.6 15.9 49.6 4.1 1.0 7.1 7.00 T 142.0 35.5 –193.6 –16.1 1.3 –3.2 7.60 39.5 9.9 T 154.6 12.9 –0.3 12.1 8.43 78.1 19.5 34.9 2.9 1.3 7.1 Mean migration rate 9.8 3.1 1.9 4.8 Standard deviation 9.6 5.9 1.2 3.3 MA flow, acre-ft/yr 48,460 42,630 nr 44,088 MA flood, ft3/s 1,383 1,142 nr 1,202

Meanders Upstream from Elkhead Reservoir 20.00 20.4 5.1 57.2 4.8 2.2 4.8 20.15 T 123.3 30.8 T 79.4 6.6 1.3 12.7 20.30 T 81.1 20.3 T 136.3 11.4 2.2 13.6 20.44 26.9 6.7 T 199.6 16.6 2.2 14.2 21.00 4.1 1.0 57.1 4.8 –0.5 3.8 21.20 44.7 11.2 58.8 4.9 1.1 6.5 21.62 44.4 11.1 5.4 0.4 0.6 3.1 22.22 61.5 15.4 T 74.6 6.2 –1.8 8.5 22.48 41.2 10.3 T 170.8 14.2 –1.3 13.2 23.06 48.8 12.2 25.0 2.1 0.5 4.6 23.35 98.0 24.5 –80.1 –6.7 0.7 1.1 23.60 30.6 7.6 14.4 1.2 2.9 2.8 23.85 –2.8 –0.7 –43.8 –3.6 3.8 –2.9 Mean migration rate 12.0 4.8 1.1 6.6 Standard deviation 9.0 6.6 1.6 5.4 MA flow, acre-ft/yr 48,460 42,630 nr 44,088 MA flood, ft3/s 1,383 1,142 nr 1,202

CHANNEL-PATTERN ADJUSTMENTS AND GEOMORPHIC CHARACTERISTICS 17 ) cross section PT-1, PT-1, section cross )

B ) cross section VT-1. Location of of Location VT-1. section cross ) Cross section Water surface at 17 cubic feet per second Bankfull stage

H Berm ) cross section PT-2, ( PT-2, section ) cross

A terrace Grassy ) cross section VT-2, and ( and VT-2, section cross )

G ) cross section LW-1, ( LW-1, section ) cross

F bar Cobble ) cross section NR-Q, ( NR-Q, section ) cross

E HORIZONTAL DISTANCE FROM LEFT TERRACE, IN FEET Bedrock ) cross section LG-1, ( LG-1, section ) cross

A Loam Gravel Sand Gravel Gleyed clay Loamy sand 280 300 320 340 360 380 400 420 440 460 480 500 520 540 560 580 Elkhead Creek cross sections showing geomorphic features, alluvial stratigraphy, and the bankfull stage, ( stage, bankfull the and stratigraphy, alluvial features, geomorphic showing sections Creek cross Elkhead 98 96 94 92 90 88 86

106 104 102 100 ELEVATION, IN FEET ABOVE ARBITRARY DATUM ARBITRARY ABOVE FEET IN ELEVATION, ) cross section LG-2, (

C Figure 5. Figure sections shown in figure 3. figure in shown sections (

18 Channel-Pattern Adjustments and Geomorphic Characteristics of Elkhead Creek, Colorado, 1937–97 Willows ) cross section PT-1, PT-1, section cross )

B ) cross section VT-1. Location of of Location VT-1. section cross )

H Cobbles Cross section Water surface at 17 cubic feet per second Bankfull stage ) cross section PT-2, ( PT-2, section ) cross

) cross section VT-2, and ( and VT-2, section cross ) High water 1997 water High

G Sand Sand ) cross section LW-1, ( LW-1, section ) cross

F Oxbow ) cross section NR-Q, ( NR-Q, section ) cross

E HORIZONTAL DISTANCE FROM LEFT TERRACE, IN FEET ) cross section LG-1, ( LG-1, section ) cross

D Hay meadow

B 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 Elkhead Creek cross sections showing geomorphic features, alluvial stratigraphy, and the bankfull stage, ( stage, bankfull the and stratigraphy, alluvial features, geomorphic showing sections Creek cross Elkhead 98 96 94 92 90 88 86

106 104 102 100 ELEVATION, IN FEET ABOVE ARBITRARY DATUM ARBITRARY ABOVE FEET IN ELEVATION, ) cross section LG-2, (

C Figure 5. Figure sections shown in figure 3—Continued. 3—Continued. figure in shown sections (

CHANNEL-PATTERN ADJUSTMENTS AND GEOMORPHIC CHARACTERISTICS 19 ) cross section PT-1, PT-1, section cross )

B Willow terrace ) cross section VT-1. Location of of Location VT-1. section cross ) Cross section Water surface at 18 cubic feet per second Bankfull stage H ) cross section PT-2, ( PT-2, section ) cross

A Sand bar (over gravel?) ) cross section VT-2, and ( and VT-2, section cross )

G ) cross section LW-1, ( LW-1, section ) cross

F Cobbles ) cross section NR-Q, ( NR-Q, section ) cross

E HORIZONTAL DISTANCE FROM LEFT TERRACE, IN FEET Loam and clay Fine sand Grassy terrace ) cross section LG-1, ( LG-1, section ) cross

C –20 0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 Elkhead Creek cross sections showing geomorphic features, alluvial stratigraphy, and the bankfull stage, ( stage, bankfull the and stratigraphy, alluvial features, geomorphic showing sections Creek cross Elkhead 98 96 94 92 90

110 108 106 104 102 100 ELEVATION, IN FEET ABOVE ARBITRARY DATUM ARBITRARY ABOVE FEET IN ELEVATION, ) cross section LG-2, (

C Figure 5. Figure sections shown in figure 3—Continued. 3—Continued. figure in shown sections (

20 Channel-Pattern Adjustments and Geomorphic Characteristics of Elkhead Creek, Colorado, 1937–97 ) cross section PT-1, PT-1, section cross )

B Willow terrace ) cross section VT-1. Location of of Location VT-1. section cross )

Cross section Water surface at 18 cubic feet per second Bankfull stage

H Driftwood ) cross section PT-2, ( PT-2, section ) cross

A ) cross section VT-2, and ( and VT-2, section cross )

G Sand bar (over gravel ?) ) cross section LW-1, ( LW-1, section ) cross

Cobbles F ) cross section NR-Q, ( NR-Q, section ) cross

E HORIZONTAL DISTANCE FROM LEFT TERRACE, IN FEET Grassy terrace Sand and cobbles Fine sand Loam Clay loam Cobbles ) cross section LG-1, ( LG-1, section ) cross

D Elkhead Creek cross sections showing geomorphic features, alluvial stratigraphy, and the bankfull stage, ( stage, bankfull the and stratigraphy, alluvial features, geomorphic showing sections Creek cross Elkhead 160 180 200 220 240 260 280 300 320 340 360 380 400 420 440 460 98 96 94 92 90

110 108 106 104 102 100 ELEVATION, IN FEET ABOVE ARBITRARY DATUM ARBITRARY ABOVE FEET IN ELEVATION, ) cross section LG-2, (

C Figure 5. 5. Figure sections shown in figure 3—Continued. 3—Continued. figure in shown sections (

CHANNEL-PATTERN ADJUSTMENTS AND GEOMORPHIC CHARACTERISTICS 21 ) cross section PT-1, PT-1, section cross )

B ) cross section VT-1. Location of of Location VT-1. section cross ) Cross section Water surface at 18 cubic feet per second Bankfull stage

H ) cross section PT-2, ( PT-2, section ) cross

A Grassy terrace ) cross section VT-2, and ( and VT-2, section cross )

bar

Sand High water 1998 water High ) cross section LW-1, ( LW-1, section ) cross

F Gravel ) cross section NR-Q, ( NR-Q, section ) cross

E HORIZONTAL DISTANCE FROM LEFT TERRACE, IN FEET brush Heavy ) cross section LG-1, ( LG-1, section ) cross

D Loam Cobbles

E Elkhead Creek cross sections showing geomorphic features, alluvial stratigraphy, and the bankfull stage, ( stage, bankfull the and stratigraphy, alluvial features, geomorphic showing sections Creek cross Elkhead –20 0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 98 96 94 92 90 88 86

106 104 102 100 ELEVATION, IN FEET ABOVE ARBITRARY DATUM ARBITRARY ABOVE FEET IN ELEVATION, ) cross section LG-2, (

C Figure 5. Figure sections shown in figure 3—Continued. 3—Continued. figure in shown sections (

22 Channel-Pattern Adjustments and Geomorphic Characteristics of Elkhead Creek, Colorado, 1937–97 ) cross section PT-1, PT-1, section cross )

B Grass and sage Cross section Water surface at 16 cubic feet per second Bankfull stage ) cross section VT-1. Location of of Location VT-1. section cross )

) cross section PT-2, ( PT-2, section ) cross Willows A Sand and gravel ) cross section VT-2, and ( and VT-2, section cross )

G Cobbles

) cross section LW-1, ( LW-1, section ) cross Buried cottonwood Buried F ) cross section NR-Q, ( NR-Q, section ) cross

E Loam Clay Fine sand Gleyed clay Cobbles/gravel and coarse sand HORIZONTAL DISTANCE FROM LEFT TERRACE, IN FEET and grass Cottonwoods ) cross section LG-1, ( LG-1, section ) cross

F Elkhead Creek cross sections showing geomorphic features, alluvial stratigraphy, and the bankfull stage, ( stage, bankfull the and stratigraphy, alluvial features, geomorphic showing sections Creek cross Elkhead –20 0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 98 96 94 92 90 88

108 106 104 102 100 ELEVATION, IN FEET ABOVE ARBITRARY DATUM ARBITRARY ABOVE FEET IN ELEVATION, ) cross section LG-2, (

C Figure 5. 5. Figure sections shown in figure 3—Continued. 3—Continued. figure in shown sections (

CHANNEL-PATTERN ADJUSTMENTS AND GEOMORPHIC CHARACTERISTICS 23 ) cross section PT-1, PT-1, section cross )

B terrace Cross section Water surface at 18 cubic feet per second Bankfull stage Cottonwood ) cross section VT-1. Location of of Location VT-1. section cross )

H ) cross section PT-2, ( PT-2, section ) cross

A ) cross section VT-2, and ( and VT-2, section cross )

G Sand ) cross section LW-1, ( LW-1, section ) cross

F Cobbles ) cross section NR-Q, ( NR-Q, section ) cross

E HORIZONTAL DISTANCE FROM LEFT TERRACE, IN FEET and sage Hay meadows ) cross section LG-1, ( LG-1, section ) cross

D Loam Gritty clay Gritty clay Calcareous clay Fine sand Cobbles

G Elkhead Creek cross sections showing geomorphic features, alluvial stratigraphy, and the bankfull stage, ( stage, bankfull the and stratigraphy, alluvial features, geomorphic showing sections Creek cross Elkhead –20 0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 98 96 94 92 90 88 86

106 104 102 100 ELEVATION, IN FEET ABOVE ARBITRARY DATUM ARBITRARY ABOVE FEET IN ELEVATION, ) cross section LG-2, (

C Figure 5. Figure sections shown in figure 3—Continued. 3—Continued. figure in shown sections (

24 Channel-Pattern Adjustments and Geomorphic Characteristics of Elkhead Creek, Colorado, 1937–97 Cottonwoods ) cross section PT-1, PT-1, section cross )

B Cross section Water surface at 18 cubic feet per second Bankfull stage ) cross section VT-1. Location of of Location VT-1. section cross )

H ) cross section PT-2, ( PT-2, section ) cross

A Willow terrace ) cross section VT-2, and ( and VT-2, section cross )

G Driftwood Sand ) cross section LW-1, ( LW-1, section ) cross

F Cobbles ) cross section NR-Q, ( NR-Q, section ) cross

E HORIZONTAL DISTANCE FROM LEFT TERRACE, IN FEET ) cross section LG-1, ( LG-1, section ) cross

D Sage

H Loam Gritty clay Cross- bedded cobbles (cemented) Elkhead Creek cross sections showing geomorphic features, alluvial stratigraphy, and the bankfull stage, ( stage, bankfull the and stratigraphy, alluvial features, geomorphic showing sections Creek cross Elkhead 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 98 96 94 92 90 88 86

106 104 102 100 ELEVATION, IN FEET ABOVE ARBITRARY DATUM ARBITRARY ABOVE FEET IN ELEVATION, ) cross section LG-2, (

C Figure 5. 5. Figure sections shown in figure 3—Continued. 3—Continued. figure in shown sections (

CHANNEL-PATTERN ADJUSTMENTS AND GEOMORPHIC CHARACTERISTICS 25 valley-fill material exposed at most Elkhead Creek channel boundary below the approximate bankfull cross sections consisted of a layer of coarse-grained elevation. The measurements excluded flood-plain fluvial sediments overlain by a thick deposit of fine- sediment exposed in cutbanks and sediment that had grained overbank sediments. The uppermost fine- slumped into the stream. grained overbank sediments consisted of distinctive Sediment sizes in Elkhead Creek ranged from horizons of fine sand or gritty clay, clay loam, and fine sand to large cobbles (fig. 6). The abrupt inflec- loam. The upper surface of the lower coarse-grained tions of the cumulative size curves at 0.25 mm result layer, where present, consisted of sandy gravel and or from an onsite assessment of all fine sand as having a cobbles and was exposed at depths of about 3.3 to diameter of 0.25 mm. A greater variation of particle 9.8 ft below the scarp formed by the valley floor. diameters exists within the sand-size range, but the These sediments tended to be noncohesive and easily Wolman method is not sensitive to measurements of erodible. The coarse-grained alluvial layer was not material finer than about 1 or 2 mm, and no attempt observed in the stratigraphy at cross sections PT-1 was made to measure the intermediate diameter of this (fig. 5B) and LG-2 (fig. 5C). Crossbedded and imbri- fine sediment. The Elkhead Creek cumulative size cated gravel and cobbles approximately 4.9 ft below distribution curves cluster together in the coarse-gravel the terrace scarp were well cemented with calcium and cobble-size regions, indicating that stream-bed carbonate at cross section VT-1 (fig. 5H). Bedrock was sediments from all sampled sites have generally exposed in the streambed at cross section PT-2 similar characteristics in the coarsest range. However, (fig. 5A), approximately 8 ft below the valley-floor differences in size distributions are visually apparent scarp. Both conditions may have had an effect on the in the sand- and gravel-size ranges. vertical incision or channel geometry of these sections. Riparian vegetation is an important variable in The median grain size, or d50, of the streambed streambank stability that affects sediment resistance to sediment in Elkhead Creek ranged from 2 to 41 mm erosion and bank-sediment mass cohesiveness and (table 4), comparable to size categories of very fine induces deposition of sediment carried in the stream- gravel to coarse gravel. The largest sediment particles flow (Thorne, 1990). Scattered stands of trees and sampled were classified as medium to large cobbles shrubs grew along the banks and terraces of Elkhead and generally were found in the thalweg, the deepest Creek when onsite cross-section surveys were made in part of the cross sections (fig. 5). The percentage of 1997. However, riparian vegetation type and density sand-size or finer sediment in the streambed ranged were difficult to evaluate in the older aerial photo- from 30 to 50 percent at the cross sections upstream graphs; consequently, changes in these variables were from Elkhead Reservoir (PT-2, PT-1, LG-2, LG-1, and not examined in this study. NR-Q). At the two sampled cross sections downstream from the reservoir, the percentage of sand-size or finer sediment was 8 percent (LW-1) and 16 percent (VT-2). Sediment Size

Sediment from the streambed surface was sampled in 1997 at the Elkhead Creek cross sections Shear Stress and Sediment Entrainment using the Wolman method for measuring the sediment- particle intermediate diameter (Wolman, 1954). Size Sediment entrainment in stream channels is distributions were computed for all cross sections partly a function of the boundary shear stress acting on except VT-1, where sediment measurements were not sediment particles in the streambed or other inundated made. Much of the left bank at VT-1 consisted of alluvial surfaces. Shear stress is proportional to the cemented, crossbedded gravel- and cobble-size sedi- square of streamflow velocity and is most accurately ment. In other respects, the sediment at VT-1 was determined by measurements of velocity vectors in similar to sediment a short distance upstream at VT-2. downstream, lateral, and vertical directions. When The Wolman measurements were made at regularly velocity data are unavailable, mean shear stress in a spaced intervals across the channel in a linear traverse channel cross section commonly is approximated by at or very near the surveyed cross section. These the relation between boundary shear stress, flow depth, measurements were made from bank to bank and and energy gradient given by the duBoys equation included all fluvially deposited material on the (Chow, 1959, p. 168):

26 Channel-Pattern Adjustments and Geomorphic Characteristics of Elkhead Creek, Colorado, 1937–97 Boulders PT-2 PT-1 LG-2 LG-1 NR-Q LW-1 VT-2 Cobbles Gravel SEDIMENT PARTICLE SIZE, IN MILLIMETERS 1 10 100 1,000 Sand Cumulative size distribution curves for streambed sediment at Elkhead Creek study sites. Creek study at Elkhead sediment streambed for curves distribution size Cumulative 0.1 0

90 80 70 60 50 40 30 20 10

100 CUMULATIVE PERCENT FINER THAN PARTICLE SIZE SIZE PARTICLE THAN FINER PERCENT CUMULATIVE Figure 6. 6. Figure

CHANNEL-PATTERN ADJUSTMENTS AND GEOMORPHIC CHARACTERISTICS 27 Table 4. Streambed sediment-size characteristics from Elkhead Creek study sites

[Percentile, percentage of the sediment sample finer than the indicated particle size; Sorting index = [(d84/d50)+(d50/d16)/2)]; Critical shear stress for d50 calculated with equation 2; mm, millimeter; lb/ft2, pounds per square foot. Size category lower limit: "boulder" = 256 mm, "cobble" = 64 mm, "gravel" = 2 mm, "very coarse sand" = 1.0 mm, "coarse sand" = 0.50 mm, "medium sand" = 0.25 mm, fine sand = 0.125 mm, "very fine sand" = 0.062 mm]

16th 25th 50th 75th 84th 95th per- per- per- per- per- per- Max- Critical Percent d / d / d / Cross cent- cent- cent- cent- cent- cent- imum 84 84 50 Sorting shear sand d d d section ile ile ile ile ile ile size 50 16 16 index stress for size ratio ratio ratio 2 (d16) (d25) (d50) (d75) (d84) (d95) (mm) d50 (lb/ft ) (mm) (mm) (mm) (mm) (mm) (mm) PT-2 30 0.25 0.38 23 52 62 85 98 2.7 248.0 92.0 47.3 0.23 PT-1 40 0.25 0.25 6 49 63 89 122 10.5 252.0 24.0 17.3 0.06 LG-2 50 0.25 0.25 2 34 67 105 140 33.5 268.0 8.0 20.8 0.02 LG-1 40 0.25 0.25 40 80 101 150 180 2.5 404.0 160.0 81.3 0.41 NR-Q 30 0.25 0.50 38 66 86 114 250 2.3 344.0 152.0 77.1 0.39 LW-1 8 12 20 41 55 65 83 112 1.6 5.4 3.4 2.5 0.42 VT-2 16 2 17 40 64 82 110 225 2.1 41.0 20.0 11.0 0.41

τ γ o = D S (1) section were substituted for the cross-section mean flow depth in equation 1 to illustrate the spatially vari- where able nature of shear stress in the Elkhead Creek cross τ sections. Shear stresses associated with three reference o is the mean boundary shear stress (in pounds per square foot), discharges were calculated for the surveyed Elkhead γ is the specific weight of water (62.4 lb/ft3), Creek cross sections to illustrate the effect of varying D is the mean flow depth (in feet), and flow depths and water-surface slopes on sediment entrainment potential. The reference discharges are S is the energy gradient (foot per foot) for a representative of low-flow conditions (an average of given discharge. approximately 17 ft3/s), bankfull conditions (1,800 to Assumptions for using equation 1 are: (1) the channel 2,200 ft3/s), and a flood discharge that fills the existing cross section has a regular, or trapezoidal, shape and channel to the approximate elevation of the valley width is at least 10 times greater than depth, (2) floor and former flood-plain surface (discharge streamflow is steady (there is a continuity of discharge unknown). within the reach), and (3) streamflow is uniform (velocity is constant in both magnitude and direction Figure 7 illustrates the cross-channel shear- through the reach). Application of equation 1 is inap- stress variation at the three reference discharges for propriate in channel sections where there is a strong cross section LG-1. Cross section LG-1 has geomor- lateral variation in acceleration or where abrupt, local phic features typical of other Elkhead Creek channel changes in streambed gradient occur. Cross sections sections (figs. 5A through 5H) including a well- in this study were not trapezoidal, although all had defined, single-thread, low-discharge channel, an allu- single-thread channels and all had width at least vial bar in or adjacent to the low-discharge channel 10 times mean flow depth. that is inundated by the bankfull discharge, a low- Most natural streams do not completely satisfy elevation terrace inundated by moderate flood the assumptions for equation 1, and the boundary discharges slightly greater than the bankfull discharge, shear stress associated with any specific discharge is and a high-elevation terrace inundated by infrequent nonuniformly distributed across the channel. Lateral floods. Cobble- and gravel-size material is exposed in and downstream variations in cross-section the deepest part of the channel at most cross sections morphology and variations in energy gradient with surveyed. The alluvial bars at most cross sections are discharge result in a wide range of boundary shear composed of sand and gravel. Low-elevation terraces stresses and, consequently, produce variable condi- have a similar sediment composition as the alluvial tions for sediment entrainment, sorting, and deposi- bars but tend to be more densely vegetated. The tion. Streamflow depths at points along each cross boundary shear stress calculated with equation 1 is

28 Channel-Pattern Adjustments and Geomorphic Characteristics of Elkhead Creek, Colorado, 1937–97

ABOVE ARBITRARY DATUM ARBITRARY ABOVE STREAMBED ELEVATION, IN FEET IN ELEVATION, STREAMBED 98 94 90 106 102 = 40 mm 50 ous sediment sizes at a typical typical a at sizes sediment ous Shear stress at valley floor stage Shear stress at bankfull stage Shear stress at 18 cubic feet per second Critical shear stress for cobbles = 64 mm Critical shear stress for d Critical shear stress for medium gravel = 23 mm Critical shear stress for sand less than 2 mm Bankfull water surface Streambed and terraces 0.65 0.41 0.23 0.02 HORIZONTAL DISTANCE, IN FEET 160 180 200 220 240 260 280 300 320 340 360 380 400 420 440 460 Cross-channel distribution of boundary shear stress for three discharges and the critical shear stress for entrainment vari of entrainment for stress shear the critical and discharges three for stress shear boundary of distribution Cross-channel 0

1.0 0.8 0.6 0.4 0.2 IN POUNDS PER SQUARE FOOT SQUARE PER POUNDS IN SHEAR STRESS, STRESS, SHEAR Figure 7. Figure Elkhead Creek site, cross section LG-1. section cross site, Creek Elkhead

CHANNEL-PATTERN ADJUSTMENTS AND GEOMORPHIC CHARACTERISTICS 29 τ proportional to flow depth. Consequently, the shear to entrain sediment d50) was estimated by equating o τ stress generated over the higher surfaces is less than with c and using the relation between shear stress and that generated over more deeply inundated surfaces, the reference discharges at a specific cross-section for a given discharge (fig. 7). Shear stress generally location. Wilcock (1992, p. 297) states that equation 2 increased with discharge because flow depths also and d50 can be used to provide a minimum estimate of increased. However, water-surface slope decreased the shear stress necessary to initiate general movement with increasing discharge at some cross sections. At of a mixed-size sediment based on experiments using those cross sections, the rate of shear-stress increase sediments with d50 up to about 20 mm. Lane’s (1955) was less than at other cross sections where water- data indicate that the Shields equation is applicable surface slope remained relatively constant or increased for sediments with a 75th percentile (d75) up to about with discharge. 100 mm. Elliott and Parker (1997) and Elliott and τ The critical shear stress ( c), the shear stress at Hammack (2000) used the Shields (1936) equation to which general movement of sediment begins, has been estimate the critical shear stress for entrainment of related to sediment size characteristics (Shields, 1936; sediment d50 in the Gunnison River of Colorado. The Lane, 1955; Komar, 1987; Wilcock, 1992). Critical critical shear stress associated with sediment entrain- shear stress for entrainment of the sediment median ment (equation 2) is, at best, a minimum estimate of particle size can be estimated using the Shields (1936) the critical discharge because only a small area of the equation: entire surface or a few particles of the d50 size may be entrained by the critical discharge (Lisle and others, τ τ γ γ c = *c ( s – ) d50 (2) 1993; Milhous, 1982). Wilcock and McArdell (1993) observed that complete mobilization of a size fraction, where such as d50, occurred at about twice the shear stress τ c is the critical shear stress (in newtons per necessary for incipient motion of individual particles square meter); in that size fraction. τ * is the dimensionless critical shear stress or c It is possible to evaluate the sediment entrain- the Shields parameter; γ ment potential of each reference discharge with s is the specific weight of sediment (assumed 3 respect to a particular geomorphic surface or location to be 25,990 N/m ); τ γ on a cross section when the boundary shear stress, o, is the specific weight of water τ (9,807 N/m3); and is compared with the critical shear stress, c, for the sediment at various locations on the cross sections d50 is the median sediment size (in meters). Critical shear stress in equation 2 (newtons per square (fig. 7 and table 4). At cross section LG-1 and at most meter) is converted to English units (pounds per Elkhead Creek cross sections, sand-size (up to 2-mm square foot) by multiplying by 0.02088. diameter) and medium gravel-size (up to 23-mm diam- Neill (1968) recommended a Shields parameter, eter) sediment theoretically is entrainable from the τ * , of 0.030 for streambeds composed of coarse deepest part of the channel by discharges as low as c 3 materials; however, other investigators have demon- 16–18 ft /s. This is supported by observation of the τ streambed at low flow, which reveals a dominant strated a variable *c in channels with mixed-size sediments (Komar, 1987) and in channels where the streambed particle size (coarse gravel and cobbles) bed-surface particle size is significantly larger than the that is greater than the theoretically entrainable subsurface particle size (Parker and others, 1982; particle sizes (sand- and medium-size gravel). At the Andrews, 1983). The Shields parameter could not be bankfull reference discharge, the alluvial bar adjacent precisely determined in this study as in some flume to the low-discharge channel at LG-1 is inundated. The and instrumented field studies; therefore, based on shear stress generated over this inundated surface published research and the particle-size ranges theoretically is sufficient to entrain sand and fine measured at the Elkhead Creek study sites, critical gravel, and the shear stress generated over the deepest shear stresses for each sediment measurement were part of the streambed should be able to entrain all τ calculated with Shields *c values of 0.030. gravel sizes and some sediment as coarse as cobble τ Once c was identified for a specific deposit, the size (64-mm diameter) or larger. Because this entrain- critical discharge (the minimum streamflow required able size range includes the d50 at cross section LG-1

30 Channel-Pattern Adjustments and Geomorphic Characteristics of Elkhead Creek, Colorado, 1937–97 (40 mm), it can be stated that the streambed at LG-1 is that passed through the reservoir. By 1993, the delta mobile at the bankfull discharge. had prograded approximately another 1,560 ft beyond the 1981 margin, and the delta area increased by 27.1 acres, a rate of approximately 2.3 acres per year, Sedimentation in Elkhead Reservoir in the 12-year period from 1982 through 1993. The Elkhead Creek delta volume was estimated Elkhead Dam impounded a 3.8-mile-long reser- using a contour map created by Ayres Associates voir in 1974, inundating the former Elkhead Creek (1995) from a topographic survey made when Elkhead flood plain and trapping all of the bedload and a large Reservoir was drawn down to a low level. The delta amount of suspended sediment previously transported surface has relatively low relief and was assumed to through this reach of the Elkhead Creek valley. Since be horizontal with an average surface elevation of impoundment, much of the bedload transported into 6,365 ft, approximately equal to the normal reservoir the reservoir has been deposited in a delta at the north- pool elevation. Surveyed points on the former flood eastern (upstream) end of the reservoir. The mass of plain immediately downstream from the delta distal sediment deposited in the delta provides some indica- margin and approximately 3,700 ft downstream from tion of the natural bedload transport rate in this reach the proximal (upstream) margin had an average eleva- of Elkhead Creek and can be considered an approxi- tion of 6,355 ft. It was assumed that the former valley mate surrogate for long-term bedload measurements. floor sloped uniformly downstream beneath the new Delta area was digitized for two time periods delta; therefore, the delta thickness increased bracketed by the aerial photographs from 1977, 1981, uniformly from zero to 10 ft over a 3,700-ft length. and 1993. These photographs were taken in September Delta thickness at the distal margins in 1981 and 1993 of each year and, as such, reflect the condition of the was calculated as proportional to the distance between delta at the end of the runoff season. The distal the proximal margin and a point 3,700 ft downstream margins of the deltaic deposits in 1977, 1981, and where the depth to the former flood plain was approxi- 1993 are shown in figure 8. After 3 years of impound- mately 10 ft. Because of the assumed wedge-shaped ment, the deltaic deposits were barely visible and the longitudinal cross section (horizontal upper surface vestigial channel of the inundated Elkhead Creek still and sloping lower surface), the average delta thickness was identifiable in 1977 (fig. 8). Sedimentation during was one-half the thickness at the distal margin. Based this period probably was relatively low because two of on valley-side topography and subaqueous delta-front the three annual flood peaks were relatively insignifi- conditions observed in 1997, the lateral and distal cant. The 1975 flood peak was 1,460 ft3/s (recurrence boundaries of the deltaic deposit are known to slope interval of 5.5 years), but the 1976 and 1977 flood steeply but, for the purpose of these volumetric calcu- peaks were considerably smaller than the approximate lations, were assumed to be vertical. By 1981, the bankfull discharge. delta had an estimated volume of approximately By 1981, the distal margin of the Elkhead Creek 27.3 acre-ft, which had accumulated at a rate of delta had prograded approximately 1,180 ft down- approximately 3.9 acre-ft per year. By 1993, the delta valley from the 1977 limit of flood-plain inundation, had a volume of approximately 163.4 acre-ft and, and the total surface area of the delta was 17.1 acres since 1981, accumulated at a rate of approximately (fig. 8). Delta area, in these calculations, includes the 11.3 acre-ft per year (table 5). distributary channel area as well as the subaerially The delta mass was estimated from the previ- exposed sediment, but not subaqueous sediment ously calculated delta volume and an assumed sedi- deposited in the reservoir beyond the distal margin, ment density for the deltaic deposits. The density of which could not be seen in some of the aerial photo- sediment deposits in a reservoir is dependent on graphs. During 7 years of sediment deposition from several variables such as the sediment-size distribution 1975 (the first runoff season following completion of (the percentage of sand-, silt-, and clay-size particles), Elkhead Reservoir) through 1981, the delta advanced sediment compaction, and the relative amount of time at a rate of approximately 168 ft per year and added the sediment is submerged or exposed to the air approximately 2.4 acres per year of exposed sediment (Lara and Pemberton, 1965). Prolonged submersion in the upper end of Elkhead Reservoir (table 5). The has little effect on the density or unit weight of sand- amount of subaqueous deposition in the reservoir is size sediment, whereas prolonged submersion tends to not known nor is the amount of suspended sediment significantly decrease the unit weight of clay-size sedi-

CHANNEL-PATTERN ADJUSTMENTS AND GEOMORPHIC CHARACTERISTICS 31 n i

. g 3 r N 9 a 9 O I 1 m k

T r 1 d e l 8 n a a A 9

e a t

r 1 , s

N i , y N C 7 d A

d 7 L a

d 9 n t P 1 l a a

e X n e

D h n

i k

g l r

a E m

l a t s i d

a t l T e E d

E e F h t

0 f 0 o

0 , n o 2 i t a c 0 o 0 l

5 e , h 1 t

d 0 n a 0

0 r i , o 1 v r e s e 0 R 0

5 d a e h k l E 0

n i

n o i t a t n e m i d e s

r c i i

h d c a t

l o l u a v e

G e d r ll h g e e n

d k s i

a l e W w E o R h s

h p a r g o t o h p

l a i r e A

. 8

e r u g i F

32 Channel-Pattern Adjustments and Geomorphic Characteristics of Elkhead Creek, Colorado, 1937–97 Table 5. Elkhead Creek delta progradation rates, estimated sediment volume, and extrapolated bedload-transport rates, 1975 through 1993

[ft/yr, feet per year; acre-ft, acre-foot; tons/yr, tons per year; progradation distance and delta surface area digitized from aerial photographs; delta thickness estimated from predeposition topography and an assumed uniform postdepositional surface topography; delta mass assumed a density of 97 pounds per cubic foot; ranked floods from gage 09245000 Elkhead Creek near Elkhead, 1954–96, with number 1 the largest flood discharge, number 2 the second largest, and so forth]

Sediment Prograda- Delta total Estimated Estimated Estimated Delta total deposition Ranked tion rate surface mean delta delta total delta total Period length rate in floods in in period area thickness volume mass (feet) period period (ft/yr) (acres) (feet) (acre-ft) (tons) (tons/yr) 1975–81 1,177 168 17.1 1.6 27.3 57,698 8,243 8, 3, 4, 11 1982–93 2,740 130 44.2 3.7 163.4 345,192 23,958 6, 2, 1, 14, 7, 10 1975–93 2,740 144 44.2 3.7 163.4 345,192 18,168 ment and slightly increase the unit weight of silt-size FACTORS AFFECTING CHANNEL- sediment. Elkhead Reservoir has a relatively constant PATTERN ADJUSTMENTS pool elevation and, consequently, deltaic sediments up to the elevation of the pool surface probably are satu- Several variables affect channel instability in rated most of the time. The sediment-size distribution Elkhead Creek, and an emphasis of this study has been of Elkhead Creek deltaic sediments was not deter- to distinguish reservoir effects from external variables mined; however, onsite examination of sediment and natural meandering processes. The presence of deposited on the delta in 1997 indicated a predomi- Elkhead Reservoir probably has both mitigating and nance of sand-size material. Based on an assumption exacerbating effects on channel erosion downstream that silt- and clay-size particles constituted a very (Williams and Wolman, 1984). The brief streamflow small part of the Elkhead Creek delta and based on the record at gages 09246200 (upstream) and 09246400 calculations of Lara and Pemberton (1965), a density (downstream) illustrates that Elkhead Reservoir of 97 pounds per cubic foot was adopted for the delta reduces the range of diurnal discharge fluctuations and mass estimation. The delta, in 1981, contained approx- attenuates the daily discharge peak downstream (R2 = imately 57,700 tons of sediment and, in 1993, 0.98 for daily mean discharges April–September contained approximately 345,000 tons of sediment 1996). A decrease in stage during the annual flood (table 5). period would generally be accompanied by a decrease in flow depth, boundary shear stress, and erosion The mass of deltaic sediment was used to assess potential. Elkhead Reservoir also is very efficient at the long-term bedload transport rate in Elkhead Creek. trapping bedload-transported sediment, as illustrated Although the streambed at the seven study cross by the growth of the delta (fig. 8). It is reasonable to sections was composed of varying amounts of mobile presume that sediment stored in the delta is sediment gravel-size material (figs. 5 and 6, table 4), the large that eventually would have been transported through percentage of sand in bed-material samples upstream the lower reaches had the reservoir not been from the Elkhead Reservoir (table 4) and observed in constructed. Streambed sediment upstream from the the delta indicates that a significant quantity of the reservoir generally has a higher sand content than transported bedload was in the sand-size category. The streambed sediment downstream from the reservoir, bedload transport of sand, estimated from the delta although only two reaches downstream were sampled mass estimates, was approximately 8,240 tons per (table 4). Clear-water, bedload-deficient releases from year from 1975 through 1981 and approximately Elkhead Reservoir may partly explain the increase in 24,000 tons per year from 1982 through 1993 meander migration rates at several meanders immedi- (table 5). Because of the sediment-trapping efficiency ately downstream from Elkhead Dam in the 1978–93 of Elkhead Reservoir, the bedload transport rate imme- period (fig. 9A). Some downstream effects of a reser- diately downstream from Elkhead Dam presumably voir may continue to develop and propagate for years, was much lower than these estimates. as Williams and Wolman (1984) demonstrated.

FACTORS AFFECTING CHANNEL-PATTERN ADJUSTMENTS 33

8.43

7.60 7.00

) reach upstream from from upstream ) reach

6.85

6.23

6.17

4.93

4.32 3.84

) reach downstream from Elkhead Reservoir, and ( and Reservoir, Elkhead from downstream reach )

3.00

2.82 1.35

UPSTREAM DISTANCE, IN RIVER MILES

1.15

0.95 0.90

1938-53 1954-70 1978-93

0.70 0.40

A Elkhead Creek meander migration rates for time intervals of equal duration: ( duration: of equal intervals time for rates migration Creek meander Elkhead

9 6 3 0

–3 –6 –9 15 12 MIGRATION RATE, IN FEET PER YEAR PER FEET IN RATE, MIGRATION Elkhead Reservoir. Elkhead Figure 9. Figure

34 Channel-Pattern Adjustments and Geomorphic Characteristics of Elkhead Creek, Colorado, 1937–97

23.85 23.60

1938-53 1954-70 1978-93

23.35

23.06 22.48

) reach downstream from Elkhead Elkhead from downstream reach )

22.22

21.62

21.20

21.00 20.44

UPSTREAM DISTANCE, IN RIVER MILES

20.30 20.15

) reach upstream from Elkhead Reservoir—Continued. Elkhead from upstream reach ) 20.00 B

B Elkhead Creek meander migration rates for time intervals of equal duration: ( duration: equal of intervals time for rates migration meander Creek Elkhead

9 6 3 0

–3 –6 –9 15 12 MIGRATION RATE, IN FEET PER YEAR PER FEET IN RATE, MIGRATION Figure 9. Figure Reservoir, and ( and Reservoir,

FACTORS AFFECTING CHANNEL-PATTERN ADJUSTMENTS 35 The preexisting characteristics of the flood-plain failure adds sediment having a variety of sizes to the and channel sediments and the natural variability of bedload, much of it transportable, but the larger parti- climate and streamflow also affect channel behavior. cles may form a lag-type pavement or armor that The influence of these variables can be seen both resists further degradation except by very large floods upstream and downstream from Elkhead Reservoir. A (Milhous, 1982). The presence of a pavement or large proportion of the streambed sediment in Elkhead armor, or the presence of nonentrainable surfaces such Creek is sand and fine gravel that can be transported as the bedrock exposed in the streambed at cross by relatively low discharges. The cross-channel shear- section PT-2 (fig. 5A) and the cemented cobbles at stress distributions and critical shear-stress values for cross section VT-1 (fig. 5H), could make lateral various sediments in the channel at cross section LG-1 erosion a more common channel adjustment than stre- (fig. 7) and the other surveyed cross sections indicate ambed scour. that sand is entrainable at almost any discharge. Meander migration rates were calculated for Medium gravel (up to 23 mm) can be entrainable from three periods of generally comparable length 1938–53, the deeper parts of the channel at low discharges and 1954–70, and 1978–93, inclusive (fig. 9). The 1978– from nearly all channel boundary locations when 93 period represents the post-reservoir period; stream- discharge is approximately bankfull (approximately flow in the earlier two periods was not affected by 1,800–2,200 ft3/s). Larger or more frequent floods Elkhead Reservoir. Only the second and third periods would be more effective at transporting sediment in span the time for which continuous streamflow record Elkhead Creek. is available and, because bank erosion and meander The valley-fill stratigraphy exposed at several migration are partly dependent on streamflow, only the cross sections provides evidence of the past behavior latter two periods are discussed here. of Elkhead Creek and the possible mechanisms of Meander migration rates in the reach down- future channel adjustments. The buried and subse- stream from Elkhead Reservoir (fig. 9A and table 3) quently exhumed Populus deltoides tree trunks near had a mean of 2.5 ft/yr and a standard deviation of 1.3 cross section LW-1 (fig. 5F) became established on a in the 1954–70 period, compared to a mean of 4.8 ft/yr sandy surface that may have been an earlier stream- and a standard deviation of 3.3 in the 1978–93 period, bank or flood-plain surface. The root crown of the tree following reservoir construction (F-test p=0.0004; at cross section LW-1 is buried by clay and loam, t-test p=0.0150; table 3). Based on these statistics typical of overbank sediments, and is approximately 2 alone, one might conclude that Elkhead Reservoir had to 4 ft below the surface of the adjacent valley floor, a substantial effect on hydrology and sediment trans- indicating a period of vertical aggradation after the port leading to the doubled mean meander-migration tree became established. The age of the tree is rate. However, a similar trend in meander migration unknown but may be on the order of about a century rates was observed in the study reach upstream from (Wes Signs, Colorado River Water Conservation Elkhead Reservoir (fig. 9B and table 3). The mean District, oral commun., 1998). The elevation differ- migration rate in the upstream study reach was ence between the contemporary valley-floor terrace 1.6 ft/yr and the standard deviation was 4.3 in the surface and the contemporary bankfull surface (2.5 to 1954–70 period, whereas the mean was 6.6 ft/yr and 6.2 ft) is an indication of vertical incision during the standard deviation was 5.4 in the 1978–93 period time since the valley floor functioned as the flood (F-test p=0.4352; t-test p=0.0156). The fourfold plain. Anecdotal evidence indicates that this incision increase in the mean meander-migration rate in the has occurred in recent decades (Donald Van Tassel and reach unaffected by Elkhead Reservoir requires Leroy Lawton, local residents, oral commun., 1997). consideration of other variables that influenced the The physical properties of the valley-fill mate- entire Elkhead Creek watershed. rial exposed in cutbanks along the study reaches affect Channel instability throughout the Elkhead the processes of channel adjustment in Elkhead Creek. Creek valley may have been influenced by changes in The sandy gravel and cobble layer of the valley-fill land use, climate, or runoff. Mean annual streamflow sediment exposed near stream level in many reaches of at the 09245000 gage increased from 36,062 acre-ft/yr Elkhead Creek (fig. 5) is easily erodible when inun- to 44,088 acre-ft/yr between the 1954–70 and 1978–93 dated and, when removed, leaves the overlying loamy periods, and the mean annual flood increased from 923 sediment unsupported and prone to collapse. Bank ft3/s to 1,202 ft3/s between the same periods. Sedi-

36 Channel-Pattern Adjustments and Geomorphic Characteristics of Elkhead Creek, Colorado, 1937–97 ment-transport rates estimated from the delta volume regulated streamflow and sediment entrapment. These and mass calculations have potential inaccuracies issues were addressed by evaluating the channel because the actual thickness of the delta is not known patterns preserved in aerial photographs in the pre- and nor is the sediment-size composition and the density post-reservoir periods and by comparing the reaches of deltaic deposits. However, assuming that these esti- upstream from and downstream from the reservoir. In mates are reasonable, one explanation for the 190- addition, 44 years of streamflow record, channel cross- percent increase in sediment transport rate from the section characteristics, and sediment entrainment 1975–81 period to the 1982–93 period is based on the potential were examined. magnitude of seasonal floods in the two periods. A meandering channel pattern is part of a Floods greater than the approximate bankfull dis- continuum of planimetric channel configurations that charge, about 950 ft3/s, occurred, on average, about reflect discharge, valley gradient, sediment load, bank every other year in both periods; however, the 1982–93 strength, and other factors. The process of meandering period included the flood of record (1984) and five involves a constant adjustment of the channel cross consecutive years (1982–86) when flood peaks section and pattern around average conditions. These exceeded the bankfull discharge (fig. 2). It is possible adjustments can involve bank erosion, point bar devel- that the large floods of the early and middle 1980’s opment, pattern adjustment (meander-bend amplitude initiated a period of increased bank erosion and sedi- increase, avulsion), streambed scour, and aggradation. ment remobilization from the flood plain and adjacent All of these adjustments can be considered normal in a terraces, both upstream and downstream from Elkhead “stable channel” that maintains an average channel Reservoir. geometry (width, depth, gradient) and pattern (sinu- The larger volume of runoff and the larger flood osity) over time. Meander migration rates, sediment peaks after the late 1970’s may have resulted in more transport rates, and delta progradation rates probably erosive conditions in Elkhead Creek, all else assumed are highly variable from year to year but, averaged constant. A cause-and-effect relationship between over time, show some correlation with discharge, with these hydrologic changes and an increase in upstream- and without the effects of Elkhead Reservoir. reach meander migration rates seems probable. A case The principal conclusions of the study are: for a similar, though more subtle, causal relationship 1. Bankfull discharge is approximately 1,800 ft3/s in with downstream meander migration rates also can be the study reach upstream from Elkhead Reservoir made because of the operational characteristics of and approximately 1,800–2,200 ft3/s in the study Elkhead Reservoir. Maintenance of a relatively stable, reach downstream from Elkhead Reservoir. The full pool and spillway passage of excess inflow attenu- bankfull width of eight surveyed cross sections ates but does not eliminate flood events downstream, ranged from 67 to 122 ft and averaged 88 ft. The as indicated by streamflow analysis of discharge data bankfull width-to-depth ratio ranged from 12 to from the concurrently operated streamflow-gaging 24 and averaged 16.5. Bankfull water-surface stations upstream (09246200) and downstream slope ranged from 0.0012 to 0.0039 and averaged (09246400) from the reservoir. Periods of large floods 0.0023. and increased meander migration in the upstream 2. Buried tree trunks at cross section LW-1 indicate reach probably were concurrent with periods of large that vertical flood-plain aggradation has occurred floods and increased meander migration in the reach in some reaches of Elkhead Creek, possibly downstream from Elkhead Reservoir. within the last century. The contemporary bankfull elevation at the eight cross sections is several feet below the broad, valley-floor terrace that SUMMARY AND CONCLUSIONS may have been a former flood plain. The elevation difference between the valley-floor terrace The effect of Elkhead Reservoir on downstream and the contemporary bankfull elevation is an channel instability has become an issue in the Elkhead indication of more recent incision, possibly Creek watershed. A study was conducted to identify within the last few decades. The elevation differ- the likely causes of streambank erosion in lower ence between the valley-floor terrace and the Elkhead Creek and to distinguish the relative effects of contemporary bankfull stage ranges from 2.5 to natural or independent variables from the effects of 3.8 ft at cross sections in the reach upstream from

SUMMARY AND CONCLUSIONS 37 Elkhead Reservoir and 2.9 to 6.2 ft at cross 09245000 gage increased from 36,062 acre-ft/yr sections in the reach downstream from Elkhead to 44,088 acre-ft/yr between the 1954–70 and Reservoir. 1978–93 periods, and the mean annual flood 3. The streambed of Elkhead Creek upstream and increased from 923 ft3/s to 1,202 ft3/s over the downstream from Elkhead Reservoir is same periods. composed largely of sediment particle sizes 7. Channel instability, represented by meander migra- entrainable by commonly occurring discharges tion rates, has increased since 1977 due to a (up to the bankfull stage). The median grain size combination of external factors affecting the ranged from 2 to 41 mm (very fine gravel to entire watershed, such as annual runoff and flood coarse gravel), and the largest particle sizes magnitude, as well as local factors, such as reser- ranged from 98 to 250 mm (medium to large voir sedimentation. Land-use practices that cobbles). The percentage of sand-size or finer potentially affect channel instability, such as sediment in the streambed ranged from 30 to intentional meander cutoff and vegetation 50 percent at the cross sections upstream from removal, were not addressed in this study. Elkhead Reservoir, whereas at the two sampled cross sections downstream from the reservoir, the percentage of sand-size or finer sediment was REFERENCES CITED 8 percent and 16 percent. 4. Elkhead Reservoir decreases the range of diurnal Andrews, E.D., 1983, Entrainment of gravel from naturally discharge fluctuations and attenuates but does not sorted riverbed material: Geological Society of eliminate flood peaks downstream. The reservoir America Bulletin, v. 94, p. 1225–1231. is very efficient at trapping bedload-transported Ayres Associates, 1995, Elkhead Reservoir Contour sediment. A delta containing approximately 163 Map 34–0246.00: Fort Collins, Colorado. acre-ft of sediment in 1993 has developed since Chow, V.T., 1959, Open-channel hydraulics: New York, Elkhead Reservoir was constructed in 1974. McGraw-Hill, Inc., 680 p. Some lateral erosion of Elkhead Creek immedi- Crowfoot, R.M., Paillet, A.V., Ritz, G.F., Smith, M.E., ately downstream from Elkhead Reservoir may Jenkins, R.A., and O’Neill, G.B., 1997, Water be related to bedload-deficient releases from the resources data, Colorado, water year 1996, vol. 2, reservoir. Colorado River Basin: U.S. Geological Survey Water- 5. Mean meander migration rates increased in the Data Report CO–96–2, 551 p. 1978–93 period relative to earlier periods of Elliott, J.G., and Hammack, L.A., 2000, Entrainment of comparable length in reaches both upstream and riparian gravel and cobbles in an alluvial reach of a downstream from Elkhead Reservoir. Meander regulated canyon river: Regulated Rivers—Research migration rates in the reach downstream from and Management, v. 16, no. 1, p. 37–50. Elkhead Reservoir had a mean of 2.5 ft/yr in the Elliott, J.G., and Parker, R.S., 1997, Altered streamflow and 1954–70 period compared to a mean of 4.8 ft/yr sediment entrainment in the Gunnison Gorge: Journal of the American Water Resources Association, v. 33, in the 1978–93 period, following reservoir no. 5, October 1997, p. 1041–1054. construction. A similar trend in meander migra- Environmental Systems Research Institute, Inc., 1996, tion rates was observed in the study reach ArcView spatial analyst user’s manual: 147 p. upstream from Elkhead Reservoir. The mean Gurnell, A.M., 1997, Channel change on the River Dee migration rate in the upstream study reach was meanders, 1946–1992, from the analysis of air photo- 1.6 ft/yr in the 1954–70 period, whereas the mean graphs: Regulated Rivers—Research and Manage- was 6.6 ft/yr in the 1978–93 period. The fourfold ment, v. 13, p. 13–26. increase in the mean meander migration rate in Hickin, E.J., 1974, The development of meanders in natural the reach unaffected by Elkhead Reservoir river channels: American Journal of Science, v. 274, requires consideration of other variables that April 1974, p. 414–442. influenced the entire Elkhead Creek watershed. Hickin, E.J., and Nanson, G.C., 1984, Lateral migration 6. Mean annual streamflow and annual flood magni- rates of river bends: American Society of Civil Engi- tudes affecting both reaches also increased in the neers, Journal of the Hydraulics Division, v. 110, 1978–93 period. Mean annual streamflow at the no. 11, p. 1557–1567.

38 Channel-Pattern Adjustments and Geomorphic Characteristics of Elkhead Creek, Colorado, 1937–97 Jensen, J.R., 1996, Introductory digital image processing— Shields, A., 1936, Application of similarity principles and A remote sensing perspective—Upper Saddle River, turbulence research to bedload movement, translated New Jersey: Englewood Cliffs, N.J., Prentice Hall, from Anwendung der Achnlichkaitsmechanik und der 316 p. Turbulenzforschung auf die Geschiebewegung: Mittei- Komar, P.D., 1987, Selective gravel entrainment and the lung Preussischen Versuchanstalt fur Wasserbau und empirical evaluation of flow competence: Sedimen- Schiffbau, Berlin, No. 26, by W.P. Ott and J.C. von tology, v. 34, p. 1165–1176. Uchelen, California Institute of Technology Hydro- Lane, E.W., 1955, Design of stable channels: Transactions, dynamics, Pasadena, Calif., report no. 167, 43 p. American Society of Civil Engineers, v. 120, no. 2776, p. 1234–1279. Thorne, C.R., 1990, Effects of vegetation on riverbank Lara, J.M., and Pemberton, E.L., 1965, Initial weight of erosion and stability, in Thornes, J.B. ed., Vegetation deposited sediments: Proceedings of the Federal and erosion: New York, John Wiley, p. 125–144. Inter-Agency Sedimentation Conference, 1963, Tweto, Ogden, 1976, Geologic map of the Craig 1° × 2° U.S. Department of Agriculture, Agricultural Research quadrangle, Northwestern Colorado: U.S. Geological Service Miscellaneous Publication 970, p. 818–845. Survey Miscellaneous Investigations Series Lisle, T.E., Iseya, F., and Ikeda, H., 1993, Response of a Map I–972. channel with alternate bars to a decrease in supply of mixed-size bed load—A flume experiment: Water U.S. Interagency Advisory Committee on Water Data, 1982, Resources Research v. 29, no. 11, p. 3623–3629. Guidelines for determining flood-flow frequency: Milhous, R.T., 1982, Effect of sediment transport and flow Reston, Va., U.S. Geological Survey, Office of Water regulation on the ecology of gravel-bed rivers, in Hey, Data Coordination, Bulletin 17B of the Hydrology R.D., Bathhurst, J.C., and Thorne, C.R., eds., Gravel- Subcommittee, 183 p. bed rivers: Chichester, England, John Wiley and Sons, Wilcock, P.R., 1992, Flow competence—A criticism of a Limited, p. 819–842. classic concept: Earth Surface Processes and Land- Neill, C.R., 1968, A Re-examination of the beginning of forms, v. 17, p. 289–298. movement for coarse granular bed materials: Walling- ford, United Kingdom, Hydraulics Research Station, Wilcock, P.R., and McArdell, B.W., 1993, Surface-based Report No. INT 68, 37 p. fractional transport rates—Mobilization thresholds and Parker, Gary, Klingman, P.C., and McLean, D.G., 1982, partial transport of a sand-gravel sediment: Water Bedload and size distribution in paved gravel-bed Resources Research v. 29, no. 4, p. 1297–1312. streams: American Society of Civil Engineers, Journal Williams, G.P., and Wolman, M.G., 1984, Downstream of the Hydraulics Division 108(HY4): 544–571. effects of dams on alluvial rivers: U.S. Geological Schumm, S.A., 1977, The fluvial system: New York, John Survey Professional Paper 1286, 83 p. Wiley and Sons, 338 p. Schumm, S.A. and Khan, H.R., 1972, Experimental study Wolman, M.G., 1954, A method of sampling coarse river- of channel patterns: Geological Society of America bed material: American Geophysical Union Trans- Bulletin v. 83, p. 1755–1770. actions, v. 35, p. 951–956.

REFERENCES CITED 39