Prepared in cooperation with the Grand Canyon Monitoring and Research Center Monitoring of Coarse Sediment Inputs to the Colorado River in Grand Canyon

Introduction Debris flows can have an immediate rapids, potentially altering the eddy pat- and dramatic effect on the river corri- tern and increasing the length of the rapid. Coarse sediment (particles with an dor. Even a single small debris flow As a result, debris fans and rapids may be intermediate diameter > 64 mm) affects may significantly alter the topography aggrading over the long term. the primary components of the Colorado and hydraulics of a debris fan and rapid River ecosystem. The deposition of in a matter of minutes. However, the Monitoring Debris Fans coarse sediment at tributary junctures Colorado River redistributes the coarse builds large debris fans that constrict the sediment introduced by debris flows The effective monitoring of coarse river and form rapids (fig. 1). Debris fans, almost immediately after deposition sediment requires both the short-term and the debris bars that develop below and during subsequent high flows. documentation of inputs by debris flow rapids, provide stable substrate for aquatic Before closure of Glen Canyon Dam, and the long-term evaluation of the redis- organisms, notably the alga Cladophora large floods on the river routinely tribution of that sediment by the Colorado glomerata. The pool above and recirculat- removed all fine sediment and some River. Both efforts involve measuring the ing eddy below the debris fan effectively coarse sediment from aggraded debris volume and particle-size distribution of trap fine sediment for storage on the bed fans (a process called reworking), trans- sediment delivered, as well as the effects or in sand bars. Debris fans and debris porting coarse sediment through the of its redistribution on the morphology bars form the fan-eddy complex that pool below the rapid and depositing it and hydraulics of the river channel. Moni- attracts humpback chub (Gila cypha), an as debris bars (fig. 1). In the regulated toring debris flows at regular intervals endangered species. Monitoring the input river, floods of reduced magnitude do will not only alert managers and research- of coarse sediment to the Colorado River not have sufficient stream power to ers to sudden, potentially important ecosystem and its long-term redistribution rework aggraded debris fans as thor- changes to channel resources but also will by the river is critical to the understanding oughly (Webb and others, 1999a, add to an existing database designed to and management of these valued 1999b). Coarse particles that are enable modeling of the interaction of resources. This fact sheet presents an entrained by these lower discharges coarse sediment and the Colorado River. overview of methods for monitoring may be deposited in the pools below The effective and efficient monitoring of coarse sediment input and redistribution in Grand Canyon. These methods are dis- cussed more thoroughly in Melis (1997), Melis and others (1994, 1997), and Webb and others (1999a, 1999b, 2000).

Debris Flows and the River

In small tributaries of the Colorado River between Powell and Mead reser- voirs, coarse sediment is transported to the river almost exclusively by debris flow. While tributary streamflow deposits are well-sorted and typically have less than 3% coarse sediment by weight, debris-flow deposits are poorly sorted and contain 5 to 76% coarse sediment (Webb and others, 2000). Figure 1. Schematic diagram of the fan-eddy complex on the Colorado River.

U.S. Department of the Interior USGS Fact Sheet 019Ð01 U.S. Geological Survey February 2001

information content, all monitoring should be done at river discharges that are as equivalent and as low as possible. Flow from the dam typically is low in the fall and early spring when heating and cool- ing demands for electricity are low. Several options exist for measuring debris-fan volume and area (Table 1). The most accurate measurement is obtained by combining direct survey of subaerial fan topography with multi-beam bathy- metric measurements of the subaqueous debris fan. This is also the most expensive method, both in terms of field-work and data processing, and data for the topogra- phy of the debris fan before the debris flow are seldom available. Remote-sensing techniques can over- come these limitations at the expense of lower resolution and accuracy. The most promising technique uses image analysis of digital aerial photography. If high-gain digital aerial photography is taken over clear, non-turbulent water, the images can be analyzed to reveal subaqueous topog- raphy in shallow water. Digital topogra- Figure 2. Aerial photographs of the debris fan at Lava Falls Rapid. A, (March 24, 1996) phy combining subaerial and subaqueous Lava Falls Rapid was constricted by a 1995 debris flow from Prospect Canyon. B, (April 3 features can be developed from stereo 9, 1996) Reworking during the rising limb of the 1996 controlled flood removed 5,900 m images. The technique is new (started in of the edge of the fan, increasing the width of the rapid by an average of 5 m. 2000) and cannot be used retrospectively channel change is highly dependent on debris flows is best done annually for pre-debris flow conditions. current efforts by the Grand Canyon between fall and early spring. Despite the A slightly less accurate but time-sav- Monitoring and Research Center to large number of tributaries in Grand Can- ing alternative is the use of LIDAR (LIght develop a baseline topographic map of the yon, debris flows are relatively infre- Detection And Ranging), an airborne laser entire river channel. quent; no more than 8 debris flows have device which is expected to be flown Most debris flows in Grand Canyon been documented in any given year dur- annually for the entire river corridor occur during the summer monsoon. Given ing the past decade (Melis and others, beginning in 2001. The accuracy of LIDAR data as collected in Grand Can- that reworking can be substantial during 1994; Webb and others, 2000). For pur- yon is considerably less than the diameter high flows in the river, documenting new poses of comparison and maximizing of most of the coarse particles being mon- Table 1. Types and accuracies of techniques for measuring debris-fan geometry itored. LIDAR topography must be com- bined with either field assessments or Expected Horizontal Vertical aerial photography to accurately map Spacing Technique Frequency of Accuracy Accuracy debris-fan boundaries and with multi- (m) Measurement (m) (m) beam bathymetric data to calculate a com- plete fan volume. Survey On demand ~0.01* ~0.01* Variable Aerial mapping photography, which (annual) typically is flown annually at a discharge Bathymetry On demand 0.05 0.05-0.06 Variable of 8,000 ft3/s, is the least accurate but

*** cheapest and most widely available tech- Digital aerial Annual 1-5 n.a. 0.18 nique. Aerial photography suitable for photos debris-fan monitoring has been flown at LIDAR Annual 0.30 0.15 2 least annually since the mid-1980s, pro- viding an excellent baseline of pre-flow ** *** Aerial photos Annual ~1-10 n.a. Variable fan conditions. Comparison of pre- and * Depends upon instrument setup and rodman accuracy. post-event photography greatly aids in ** Depends upon the quality of control points and the camera and flight characteristics. delineating the boundaries of debris flows *** Topography can not be extracted without stereo photography and control panels. as well as reworking by the Colorado

River (fig. 2). Standard aerial photographs ing, or the interlocking of particles. Sutur- Percent constriction of the river chan- must be digitized, georeferenced, and rec- ing is caused by the rearrangement and nel, a ratio of the average channel width tified before use, requiring the establish- wearing together of particles, and its through the rapid to the average channel ment of control points in the field. When occurrence makes debris fans difficult to width above and below the rapid (Webb pre-event topography is not available, fan rework. and others, 1999a), is a useful measure of volume can be calculated by multiplying the impact of a debris flow on river chan- Other measurements include survey- the fan area measured from aerial photo- nel morphology (fig. 5). For ease of mea- ing the water-surface fall through the graphs by an average thickness. surement and consistency, measures of rapid, which can be used to calculate Particle-size distributions are best channel width are best obtained from geo- stream power for a given discharge (fig. measured by combining point counts in referenced remote-sensing data, particu- 4). Surface velocity through the rapid can the field with standard sieve analysis in larly aerial photographs. Aerial be the laboratory to capture the full range of measured by timing the passage of photography may also reveal qualitative particles found in debris flow deposits. floats through the rapid. Other on-site changes in the hydraulics and navigability Larger particles may be measured on site. measurements include documentation of of rapids, reflecting underlying changes in Both unreworked deposits on fan surfaces changes in hydraulic features such as the the positions of boulders, and can be com- and reworked deposits along distal fan shifting of waves and holes, and the bined with field observations as a qualita- edges are evaluated (fig. 3). Particle-size movement, appearance, or disappearance tive measure of channel change. measurements should document sutur- of rocks. Because debris bars typically are unstable, much of the monitoring of these bars is best performed using remote sens- ing. This monitoring is most accurate when digital aerial photographs georefer- enced by geographical positioning sys- tems (GPS) are used because stable, long- term control points may be difficult to locate. Alternatively, LIDAR data of sub- aerial topography and multi-beam bathy- metric data can be combined to give an important three-dimensional portrait of bar evolution over time. Particle-size dis- tributions, and particularly lithologic counts, may be useful for assessing the stability and formation of debris bars.

Long-Term Monitoring

Figure 3. Particle-size distributions reflecting reworking on distal margins of a recently Repeat monitoring of older fans could aggraded debris fan at mile 127.6 in Grand Canyon. The curve for partially river be performed at the same time as monitor- reworked deposit is the result of releases from Glen Canyon Dam. The curve of fully ing at recently aggraded debris fans, river reworked deposit is for reworking under pre-dam conditions. though additional monitoring should be done following the occurrence of atypi- cally large reworking floods at other times of the year. Monitoring could start with an examination of new aerial photography and recalculation of debris-fan area, width of the reworked zone, and channel con- striction. Because regulated floods do not overtop most fans but instead are eroded laterally, lack of change in any of these parameters indicates a stable fan where armoring is likely. Field work on these fans could be limited to annual point counts along the fan edge to evaluate the degree of armoring. Multi-year stability in particle-size distribution as well as fan area and channel constriction indicates a Figure 4. Water-surface profiles through Lava Falls Rapid showing the effects of stable fan and monitoring could be sus- reworking by the March 1996 flood. pended. However, any river discharge

exceeding the maximum flood that has that is important to the aquatic food base. passed that fan to date would be cause to Aquatic plants use coarse particles as again monitor the debris fan for change. anchor points, and aquatic invertebrates Conversely, change in any of these use the spaces between coarse sediment parameters, particularly after large floods, as habitat. Identification of stable versus is suggestive of debris-fan reworking and unstable deposits of coarse particles will additional field-work is necessary. New aide in long-term assessments of aquatic measures of fan volume and particle-size productivity. distribution are necessary to determine Debris-flow monitoring should be part the fraction of deposit and classes of par- of an adaptive management program, par- ticles reworked by a given flood discharge ticularly with respect to changes in both at the site. Effects on the rapid are again aquatic and terrestrial habitat. Addressing measured by water-surface fall through the processes of debris-fan reworking and the rapid, surface velocity, and calcula- armoring makes possible the determina- tions of stream power. New measures also tion of how much and which size classes reset the baseline for future measures of of coarse particles actually enter the river. reworking. Changes in fan volume cou- This empirical information should be pled with particle-size data can be used to combined with sediment-transport models estimate amount and sizes of coarse parti- to estimate the zone of influence debris cles redistributed downstream to the pool flows and debris-fan reworking have on or cobble bars. downstream aquatic habitat. The long-term effects of debris flows Figure 5. Percent constrictions of debris on the river corridor are also an important Fine Sediment fans. A. Error bar indicates change in consideration, particularly in a river sys- constriction as a result of the 1996 flood with Changes in debris fans can be used to tem with decreasing amounts of fine-sedi- date of debris flow (Webb and others, 1999). assess the overall contributions of sand ment mantling the channel margins. B. All debris fans in Grand Canyon. from small tributaries to the upper reaches Debris-flow inputs may form a significant part of fine-grained sediment inputs, as National Park: peak discharges, flow transfor- of Grand Canyon, where sand bars are mations, and hydrographs, in Chen, C. (ed.), diminished. Sand bars may be covered by well as modifying channel geometry to enhance storage of fine-grained sedi- Debris-flow Hazards and Mitigation: Mechan- debris flows, eliminating their usefulness ics, prediction, and assessment, Proceedings of ment. To date, these quantities have been as camping beaches. Monitoring of sand- first International Conference: New York, bar size downstream from aggraded estimated using a probability model of American Society of Civil Engineers, p. 727- debris fans should be a consideration of debris-flow occurrence coupled with gen- 736. any long-term monitoring plan for coarse eralized ranges of debris-flow volume and Webb, R.H., Griffiths, P.G., Melis, T.S., Elliott, sediment inputs. The volume of fine sedi- particle-size distribution (Webb and oth- J.G., and Pizzuto, J.E., 1999a, Reworking of ment in the upper pool could be moni- ers, 2000). Building a database of debris debris fans by the 1996 controlled flood in tored in concert with debris fans by multi- flow input and reworking will permit the Grand Canyon, in Webb, R.H., Schmidt, J.C., Valdez, R.A., and Marzolf, G.R., The 1996 beam bathymetric measurements. As testing of these estimates as well as the flood in Grand Canyon: Scientific experiment reworking floods change the debris fan, development of a more realistic model of debris-flow inputs. and management demonstration: Washington, fine sediment is liberated from the upper D.C. American Geophysical Union, Geophysi- pool and is made available for storage cal Monograph, p. 37-51. ÐRobert H. Webb and Peter G. Griffiths downstream in sand bars. Webb, R.H., Melis, T.S., Griffiths, P.G., Elliott, J.G., Cerling, T.E., Poreda, R.J, Wise, T.W, and Benefits of Debris-Fan Selected References Pizzuto, J.E., 1999b, Lava Falls Rapid in Grand Canyon, Effects of late Holocene debris Monitoring Melis, T.S., 1997, Geomorphology of debris flows on the Colorado River: U.S. Geological flows and alluvial fans in Grand Canyon Survey Professional Paper 1591, 90 p. Data collected from the monitoring of National Park and their influences on the Webb, R.H., Griffiths, P.G., Melis, T.S., and Hart- debris flows will have a number of impor- Colorado River below Glen Canyon Dam, ley, D.R., 2000, Sediment delivery by ungaged tant uses in the research and management Arizona [Ph.D. Dissertation]: Tucson, Uni- tributaries of the Colorado River in Grand Can- of the Colorado River ecosystem in Grand versity of Arizona, 495 p. yon, U.S. Geological Survey Water-Resources Canyon. Annual monitoring provides Melis, T.S., Webb, R.H., Griffiths, P.G, and Investigations Report 00-4055, 67 p. detailed, up-to-date information on Wise, T.J., 1994, Magnitude and frequency changes in Grand Canyon rapids and data for historic debris flows in Grand For further information, contact: Canyon National Park and vicinity, eddies. Monitoring also provides an Robert H. Webb Arizona: U.S. Geological Survey Water empirical measure of the quantity and Resources Investigations Report 94-4214, U.S. Geological Survey particle-size distribution of sediment 285 p. 1675 W. Anklam Road input to the Colorado River annually by Melis, T.S., Webb, R.H, and Griffiths, P.G., Tucson, AZ 85745 debris flow as well as changes in substrate 1997, Debris flow in Grand Canyon E-mail: [email protected]