In the days following the May 18, 1980, eruption, earth scientists recognized that the unstable landscape around Mount St. Helens presented an opportunity to describe rarely observed processes of erosion and deposition. The colossal scale of sedimentation attracted scientific interest partly because the devastation occurred near a populated area and was reasonably accessible. As the data needs for mitigation structures and sediment-transport evaluation became clear, the Federal Government made considerable funding available for hazard assessment. Technical and research personnel throughout the U.S. Geological Survey specified the ideal components of a successful data-collection program.
This section describes the scientific goals for study of the Mount St. Helens area and the objectives of the monitoring program. An overview of gaging stations is followed by a summary of data-collection methods.
Scientific Goals and Objectives
The immediate scientific goal was to obtain hydrologic and geomorphic data about the unstable landscape. Decreased infiltration of rainfall, widespread destruction of forests, and a vast supply of erodible sediment were expected to alter drastically the hydrologic responses of affected drainage basins. Consequences of the 1980 eruption to local hydrology were to be assessed in detail. Stream-channel changes and erosion of the surrounding landscape were likely to occur rapidly. Long-term research in the study area would thus benefit from the "compression" of the geomorphic time scale.
Effects of the eruption on water quality in streams and lakes were to be evaluated by numerous sampling programs. Laboratory and field studies were proposed to investigate particle size, mass density, critical velocity, tractive force, and roughness of the debris and ash deposits. Documentation of extreme sediment-transport conditions was essential for projecting annual sediment discharges to be contained by sediment-control works. Mathematical modeling of sediment transport was to be supported with extensive collection of samples of suspended sediment and bed material, and with repeated surveying of stream-channel cross sections. Data for calculation of total sediment discharge, collected regularly at several gaging stations, were to be suitable for modeling of sediment transport.
Records of sediment yield from lands affected by the 1980 eruption were required for resource planning and for geomorphic studies of the rapidly adjusting river systems. Channel surveys of streams near Mount St. Helens were used to measure erosion and deposition of sediment. Research into the initiation, rheology, and sedimentary deposits of volcanic debris flows was planned.
Many of the highest mountains in the Cascades Range are considered dormant volcanoes, and the 1980 eruption of Mount St. Helens demonstrated the potential hydrologic hazards from those mountains. The 1980 eruption was only a recent example of volcanic sedimentation. River valleys in the vicinity of Mount St. Helens contain fills of ancient volcanic alluvium that extend many miles from the mountain, according to previous studies in the area (Mullineaux and Crandell, 1962). The sedimentologic information acquired at Mount St. Helens was intended to aid in hazard assessment at other volcanoes.
To define the changing stream conditions and provide a basis for long-term interpretive studies, a network of gaging stations for continuous monitoring of water discharge and sediment transport was installed by the Washington District of the U.S. Geological Survey (fig. 3). Hazard-warning systems were operated with the network of gaging stations to provide advance notice of stream flooding (Childers and Carpenter, 1985).
Objectives of the monitoring program were:
To meet each objective, specific procedures were developed, as discussed below.
The first significant seismic activity and sporadic eruptions at Mount St. Helens in this century occurred in March 1980. At that time, long-term gaging stations were operational in the Cowlitz and the Lewis River basins. Stream monitoring in the vicinity was immediately increased to watch for hydrologic hazards posed by the threat of eruption. In April 1980, new sampling sites at Pine Creek and the North Fork Toutle River were equipped with water-quality monitors and telemetry relays to GOES (Geostationary Operational Environmental Satellite). Both of the new sampling sites were destroyed by lahars from the May 18, 1980, eruption. When the devastation was assessed, and an urgent need for flood warning was declared, a gaging-station network was planned that would provide standard river monitoring and real-time alerts of flood hazards (Childers and Carpenter, 1985).
In June 1980, continuous monitoring sites were established on streams in the eruption-affected drainage basins of Mount St. Helens (fig. 3, table 1 ). Most sediment data were collected at gaging stations on seven streams: the Green River , the North and South Fork Toutle Rivers, the Toutle River, the Cowlitz River, Clearwater Creek, and the Muddy River. (figs. 4-7). Some gaging stations were located near the stream mouth to estimate sediment yield from the entire drainage basin. Most gaging stations were easily accessible by road, except for the South Fork Toutle River above Herrington Creek and Clearwater Creek above mouth, which were often visited by helicopter. By 1990, six gaging stations were still operated for sediment discharge records in the Mount St. Helens area. The number and location of gaging stations varied from year to year as better measuring sites were established and unneeded ones were discontinued. In this report, abbreviated forms of gaging-station names (for example, Toutle River at Tower Road) are used in text and figures. Complete names and station numbers are given in table 1 .
Periodic sampling sites, for which daily sediment-discharge records were not computed ("none" in table 1), also were established in most drainage basins. These sites were visited less often, and efforts to obtain continuous gage-height records were limited.
Additional sediment data were collected from streams in other drainage basins, including the Cispus River (tributary to the Cowlitz River above the Toutle River) and the Kalama River (tributary to the Cowlitz River below the Toutle River) ( table 2 ). These streams were affected primarily by airborne volcanic ash; increased sediment transport from erosion of the ash was temporary and negligible after 1981. Sediment-transport data were collected at those gaging stations for periods of several months.
Basic Methods of Data Collection
The data used to compute suspended-sediment discharge were collected at gaging stations with standard methods, as described in the following sections. Suspended-sediment discharge in a stream usually is computed from the product of measured water discharge and measured suspended-sediment concentration. The equation
defines the computation, where QS is suspended-sediment discharge, in tons per day; QW is water discharge, in cubic feet per second; and C is suspended-sediment concentration, in milligrams per liter. The coefficient 0.0027 converts the mixed units to inch-pound units of tons per day.
Inch-pound units are used by the U.S. Geological Survey for length and weight measurements in routine hydrologic work. Stream depths and widths are measured in feet (ft), velocities in feet per second (ft/s), and flows in cubic feet per second (ft3/s), whereas sediment size is expressed in millimeters (mm) and sediment concentration is expressed in milligrams per liter (mg/L). Units in this report are identical to those used to record, calculate, and archive the sediment data collected at Mount St. Helens. Other measurements, such as drainage area and river mile, also are expressed in inch-pound units to maintain consistency.
Because sediment is transported by turbulent flows, because alluvial channels are unstable, and because sediment sizes can range from clays to boulders, sediment sampling and discharge measurements are subject to error from temporal variability. The error is compounded by spatial variability in the stream cross section. Rapid fluctuations in water discharge and sediment concentration were resolved by frequent measurement and sampling to improve the accuracy of sediment discharge records.
Discharge and sediment concentration tended to change rapidly after peak stage. Measurement methods that were designed to define sediment discharge during steady flow were not reliable during unsteady storm flows. Therefore, methods were adopted that improved time resolution during rapidly changing flow. The frequency of cross-section samples was reduced, the frequency of single-vertical samples was increased, and discharge measurements were completed in about 30 to 40 minutes using flood-measurement techniques (Buchanan and Somers, 1969).
If stream discharge changes rapidly during storm flow, sediment concentration also will change in response to erosion or deposition. Concentration curves often do not coincide with discharge hydrographs, and unexpected changes can be detected only by frequent sampling. In streams near Mount St. Helens, sediment concentration was sampled during storm flows from once an hour to as often as every 5 minutes. Consequently, concentration curves for storm flows lasting several days were defined by dozens of sediment-sample concentrations. Several discharge measurements were made over the same period to detect rapid changes in stream discharge not shown by river stage. (Specific techniques are described in the following sections on "Water Discharge" and "Sediment Concentration".)
The spatial resolution of sediment-transport characteristics along a stream channel is limited because data are usually collected only at gaging-station cross sections. However, greater spatial resolution was attained with the basin-wide surveying of channel cross-section geometry at many points along disrupted or developing channels (Martinson and others, 1984, 1986; Meyer and others, 1986; Meyer and Dodge, 1988). Changes in channel volume were determined from channel geometry for comparison with sediment discharge records of nearby gaging stations (Meyer and Janda, 1986).