Sediment-transport monitoring through 1990 by the Cascades Volcano Observatory yielded information about lahar behavior and sediment-transport processes that is summarized here. Examples of existing and potential research in sedimentation, both regional and basic, are presented as benefits of monitoring. Suggestions for additional data collection also are described. The U.S. Geological Survey has prepared and released numerous publications that present detailed aspects of sediment transport at Mount St. Helens (see sources in "Availability of Data").
Eruption-induced flows of volcanic debris and mud leave characteristic deposits that have been identified in the river valleys of several Cascades Range mountains (Crandell and Waldron, 1956; Mullineaux and Crandell, 1962; Schmincke,1967; Crandell, 1971; Scott, 1988; Scott and others, 1992). The flows have been described with the inclusive Indonesian word "lahar" to denote rapid, transient flows of sediment and water from a volcano. The sediment in lahars is mobilized by melted ice and snow and by stream water in channels that the lahar invades. As lahars travel and diminish in flow quantity, they deposit sediment that is later recognized by the angular, floating clasts supported in a sandy or clayey matrix.
The quantity of sediment transported from Mount St. Helens as lahars on May 18, 1980, was estimated from relatively few samples of the flowages. The samples were collected without knowledge of the extended range of sediment concentration that lahars undergo. The lahar in the South Fork Toutle River was sampled on recession, at a stream discharge of about 500 ft3/s, 3 hours after an estimated peak discharge of 45,000 ft3/s. Seven hours later, the lahar from the North Fork Toutle River was sampled on the rising stage at a discharge of about 60,000 ft3/s. The sample of the North Fork lahar and a sample taken 9 hours after peak discharge were collected at the Toutle River at Highway 99 (Dinehart and others, 1981). Curves of sediment concentration and stream discharge were drawn from these data ( fig. 82 ; J.M. Knott, U.S. Geological Survey, written commun., 1981), and were used to estimate the sediment discharge of the two lahars at 153 million tons on May 18-19, 1980.
On May 19, 1980, the waning lahar from the Toutle River was sampled from the bridge over the Cowlitz River at Castle Rock throughout the day. These were cross-section samples, collected every hour and later composited in the sediment laboratory for a single analysis at each hour (Dinehart and others, 1981). Stream discharge of the lahars past the Cowlitz River at Castle Rock gaging station was calculated by comparison with the Toutle River at Highway 99.
Lahars can be debris flows that become diluted to a hyperconcentrated phase where sediment concentrations range from 40 to 80 percent sediment by weight (Beverage and Culbertson, 1964), and sediment settles differentially by hindered fall velocity. A hyperconcentrated phase of a lahar was observed at three gaging stations on the Toutle River on March 20, 1982, as it flowed to the Cowlitz River. This flow was the result of a minor eruption in the crater of Mount St. Helens that began explosively on March 19, 1982 (Waitt and others, 1983). The deposits in the Toutle River ( fig. 83 ) were well-sorted sand and fine gravel, unstratified or massive and crudely stratified, with thicknesses greater than 1 m (Scott, 1988).
The experience of under-sampling the May 18-19, 1980, lahars prompted the collection of abundant samples during the lahar of March 19-20, 1982, in the Toutle River. Sediment samples collected at the North Fork Toutle River at Kid Valley, the Toutle River at Tower Road, and the Toutle River at Highway 99 included concentrations around 1 million mg/L. Over 100 samples of the main flow were collected in total at the three gaging stations (Dinehart, 1986) in spite of floating woody debris and the highly-viscous material that would not flow easily into sample nozzles and bottles.
Records of river stage and discharge measurements were combined with sediment samples to compute sediment discharge for the March 19-20, 1982, lahar. Sediment discharge for the hyperconcentrated flow at the North Fork Toutle River at Kid Valley was 5,430,000 tons, which decreased to 3,480,000 tons at the Toutle River at Highway 99. Although the flow was depositional, the fine sediment measured at high concentrations by sampling was found in only small quantities in the lahar deposits. Suspended-sediment samples collected at the Cowlitz River at Castle Rock during the flow were mostly fine sediment. These observations showed that the distribution of sediment sizes found in the lahar deposits did not fully represent the sediment that flowed as the lahar. The fine sediment, which is fundamental in maintaining high sediment concentrations in the flow, was transported past the depositional area, and its presence was not recorded in the deposits.
Particle-size analyses of the March 19, 1982, lahar samples were used to draw curves of fine and sand concentration ( fig. 84 ). Fine concentration attained a range of 300,000 mg/L, and gradually decreased. Sand concentration, however, reached much higher concentrations and receded more rapidly. Rapid decreases in concentration may be diagnostic of hyperconcentrated-phase lahars, which are differentiated from lahars with higher, but more constant, sediment concentrations.
A small lahar reached the North Fork Toutle River on May 14, 1984, which was sampled at Kid Valley and at the Toutle River at Tower Road (see "Peaks and Lags of Sediment Concentration"). Peak measured concentrations were about 79,000 and 46,000 mg/L at Kid Valley and Tower Road, respectively. The sediment concentrations were an order of magnitude lower than those of the March 19-20, 1982, lahar. No other sediment-laden flows directly attributed to volcanic action were sampled in the Toutle River between the May 14, 1984, flow and the end of the study period.
Monitoring and Sediment-Transport Processes
Monitoring sediment transport in streams at Mount St. Helens for the purpose of sediment discharge measurements also resulted in extensive collection of research data. High sediment concentrations and stream velocities are seldom measured as frequently as was done in the study area. Therefore, analysis of the field data provided examples of sediment transport at extreme flow conditions. Some field data illustrate sediment-transport variability, such as velocity fluctuations, rapid changes in cross-sectional area and bed elevation, and random changes in sediment concentration. This section notes benefits of monitoring sediment-transport processes for research and engineering. Monitoring of this kind provides data to compute percentage deviations of sediment discharge values and to document the hydraulics of sediment transport.
When collecting samples and measuring velocity in storm flows, lower flow depths are difficult to reach with cable-suspended equipment. For samples and velocity measurements in the upper region of the flow, the measurements can be extended to represent the entire flow by application of velocity and concentration-distribution laws. Empirical corrections for concentration and velocity distributions are developed from simplified conditions in nearly steady flows, and their applicability to high velocities and sediment concentrations is not widely documented. Therefore, vertical distributions of velocity and sediment were measured at several gaging stations, in association with routine sediment-discharge measurements (Dinehart, 1987). Velocity profiles, as measured by Price AA velocity meters, corroborated other observations (Marchand and others, 1984) of surface velocities that were higher than predicted. Sand suspension near the bed was greater than predicted by concentration distribution laws. The field data showed that, in storm flows, partial-depth measurements could not be adjusted reliably to derive vertically-averaged measurements. This finding justified the use of larger sounding weights and sediment samplers.
Velocity and concentration distributions varied rapidly with time. Repetitive stream discharge measurements during single storm flows showed that mean bed elevation varied significantly. Bedform migration was investigated as a source of the variability (Dinehart, 1989). Sonar was first deployed in 1986 to detect bedform movement. Dune-like bedforms were found migrating at 1 to 3 in./s, with dune heights from 5 to 20 in. Concurrent samples of bed material and bedload collected during sonar observations demonstrated that the dunes were composed of fine-to-coarse gravel (Dinehart, 1992b). Continuous measurements of velocity at two or three points above the bed showed that the velocity profile was directly affected by bedform migration (Dinehart, 1992b).
Rapid changes in bed regime during storm flows were apparent from water-surface features. Upstream-moving surface waves indicated the formation of sandy antidunes in some streams during high-concentration storm flows in 1980-82. Periodic alternation in the stream surface from dark, sandy boils to smooth, shooting flow indicated transition from dune to upper-regime plane bed. Periodic fluctuations of water surface seen in stage records of storm flow indicated bedform growth during flow recession. Continuous sonar observations of streambeds during storm flow confirmed that stage fluctuations corresponded to bedform migration.
Measured bedload discharges were compared with discharges from formulas that use hydraulic and sediment data to estimate bedload discharge (S.E. Hammond, U.S. Geological Survey, written. commun., 1992). Several formulas (for example, Bagnold, 1966; Shen and Hung, 1972; Yang, 1973) gave transport rates that were comparable to field measurements. Other transport formulas were less applicable for the range of conditions in the North Fork and the mainstem Toutle Rivers.
Additional sediment-transport data that are available for research are listed here:
Following the closure of the SRS on the North Fork Toutle River, streambeds downstream coarsened measurably. Bedload samples collected during that period contained sand and fine gravel, whereas subpavement samples became deficient in the same range of sediment sizes. River channels often degrade downstream from dams, and the bedload samples can reveal details of the armoring process.
Further analysis of the total sediment-discharge data collected in streams near Mount St. Helens should prove valuable. The coarse sediment, high stream velocities, and high sediment-discharge rates are distinctly different from sand-dominated streams where total sediment-discharge studies have been performed in the past. Data for geomorphic evaluations of the instantaneous discharge and sediment transport measurements are available (Childers and others, 1988; Hammond, 1989).
A byproduct of data collection for calculations of total sediment discharge is a database of particle-size distributions of suspended sediment. The size distributions vary unpredictably with flow conditions and basin characteristics. Simple relations between size statistics, stream discharge, and sediment concentration were not found in sediment data spanning several years. Because suspended-sediment samples were collected in the study area under a wide range of conditions, a large population of size distributions is available for further analysis.
Runoff response of the Toutle River was altered by the 1980 eruption (Datta and others, 1983). Interest in the altered rainfall-runoff relation decreased after erosion of ashfall had progressed for several years. However, precipitation data collected at several sites near Mount St. Helens through 1990 (Uhrich, 1990) are available for analysis.
Eleven years of sediment-transport observations at Mount St. Helens gave a detailed sequence of sediment discharge from the Toutle and the Lewis River basins after the 1980 eruption. Although steady efforts to monitor stream discharge and sediment concentration defined the overall trends, additional data collection would enhance sediment-transport studies. The retrospective look at sediment transport at gaging stations suggested needs for additional data collection and strategic changes in stream monitoring.
Frequent sediment samples were not obtained at some gaging stations during 1980-81, which precluded complete records of daily sediment discharges at those sites. The first 12 months of sediment discharge from the North Fork and the South Fork Toutle Rivers were observed by sediment measurements on 35 separate days. No daily values of sediment discharge in water years 1980-81 were computed for the Green River, the Muddy River, or Pine Creek, due to the scarcity of sediment data. The limited amount of time and personnel to install automatic sampling equipment hindered the acquisition of useful sediment data throughout the winter of 1980-81. Future developments in sediment data collection might include automatic sampling equipment that can be deployed rapidly (in a few days) at remote sites.
Uncertainty in estimates of total sand discharge persisted throughout the sediment-sampling program. Dunne and Leopold (1981) highlighted the lack of sand divisions for samples collected during 1980 in streams near Mount St. Helens. Because sand concentration is useful for estimates of sediment deposition, sand-division analyses were made of all samples beginning in 1981. Dependence on automatic sediment sampling eventually reduced the number of samples with known sand concentrations. To monitor sand transport during storm flow, additional depth-integrated samples would be essential. Automatic sampling equipment that collects representative sand concentrations is desirable.
There are few data available to describe streambed coarsening and pavement formation. The limited bed-material data suggest that methods of sampling bed material could have been modified and expanded as soon as 1981. The standard bed-material sampler (BM-54) was inadequate for subaqueous sampling of gravel bed material, and the samples were not fully representative of gravel beds. A sampler for gravel bed material that can be used at gaging stations with gravel streambeds may be required. Established methods of sampling exposed gravel bars also can be used periodically (Church and others, 1987).
A variety of methods to measure bed elevation by sonar can be used at gaging stations (Dinehart, 1992a). Bed-elevation data can improve the application of stage/discharge relations, detect processes of fill and scour, and identify modes of sediment transport not detectable by sampling alone.
Bedforms in heterogenous gravel beds may cause stage-discharge relations to shift by changing mean bed elevation and mean velocity. When the physical processes of migrating gravel dunes are better understood, flow resistance in erodible gravel beds can be estimated more reliably (Dinehart, 1992a).
At high sediment concentrations, small-scale turbulence in river flow is visibly dampened, and surface velocities are occasionally greater than expected from distribution laws. Field observations of vertical velocity profiles did not provide reliable contrasts between clear-water and sediment-laden flows, because bed roughnesses were not known. Field investigations of sediment-laden flow, accompanied by measurements of bedforms and bedload transport, would define significant effects of high sediment concentrations on flow velocity.
Sediment discharges are greatest during brief periods of unsteady storm flow, when channel geometry is altered by scour and fill and by migration of bedforms of several scales. Erodible banks and mobile streambeds produce changes in channel geometry during high flows, but understanding the sequence of erosion and deposition in relation to the passage of flood waves requires more frequent measurements of channel geometry.
River flow and data-collection activities can benefit from recording to videotape, especially during storm flows and unique sediment-laden flows. When a flow is short-lived and field data will be analyzed in depth, the videotapes will provide information not recorded in field notes. Visual images of water-surface and channel changes can be used to assess flow behavior, especially from steady views with long duration.
