As the staff at the Cascades Volcano Observatory (CVO) became more familiar with the behavior of streams near Mount St. Helens, data collection and analysis methods were modified and improved. Improvements were based on experience gained from frequent data collection during storm flows. Examination of sediment-discharge records made data needs apparent to field personnel.
Standard gage houses and cableways were constructed on streams near Mount St. Helens during 1980-82. However, standard techniques for monitoring those streams did not easily accommodate the extraordinary flow conditions. To obtain hydrologic data under the extremes of high stream velocities, debris- and sediment-laden flows, and unpredictable channel fluctuations, modifications to field equipment were developed and tested. Equipment operation and reliability were evaluated under arduous field conditions.
Discharge and sediment-concentration measurements were made frequently, with several visits per month to gaging stations. Between 1980 and 1982, the volume of data available for sediment-discharge records overwhelmed the existing system of manual computation. The extra accuracy afforded by the high frequency of sediment sampling required additional collation and computation of data. In the chronology, principal observations or developments in methods and analyses are listed with brief descriptions under each water year.
A small pond that had formed on the debris-avalanche deposit breached on August 27, 1980. The sediment-laden flow was sampled at several sites as it traveled through the North Fork and mainstem the Toutle Rivers. The sediment data provided insight into the magnitude of sediment discharges that could be expected in future storm flows.
Extensive cross-section networks were established in the Toutle and the Lewis River basins. Surveys of channel geometry were made periodically, and profiles were published in several open-file reports (Martinson and others, 1984, 1986; Meyer and others, 1986; Meyer and Dodge, 1988).
Crews sampled and measured storm flow at several gaging stations simultaneously in November and December 1980. Episodes of continuous data collection often lasted more than 36 hours. Synoptic data were critical in evaluating sediment transport through the Toutle River system.
Standard manometers, float switches, and stage sensors were linked to satellite telemetry to provide timely warning of hazardous changes in river stage. Some warning stations were established in rugged terrain near Mount St. Helens and were maintained by regular helicopter visits. Rises in river stage provided logistical data for synoptic sampling during storm flows (Childers and Carpenter, 1985).
Automatic pumping sediment samplers were equipped with air compressors to ensure that sampling would continue if the diaphragm pump were submerged under flood water. The pumping samplers, modified from the U.S. PS-69 (designed in 1969), collected critical water samples when technicians could not.
Flood debris and sediment deposition interfered with manometer orifices and automatic-sampler intakes, causing critical interruptions in storm-flow records. Comments on discharge measurements frequently read, "orifice gone," or "PS-69 not operating." When manometer sensors or sampler intakes were set at stationary reference points, the streambed often filled above their level and stopped operations. Sensors with multiple orifices in a vertical arrangement were designed and became a standard installation. The multiple orifices were only effective, however, when personnel were at the site to change tubing hookups.
Field measurements of vertical profiles of stream velocity and sediment concentration were made beginning in 1981. Analysis of the profiles showed possible sources of error in velocity measurements due to non-logarithmic profiles. Velocity measurements near the surface were often significantly greater than law-of-the-wall predictions. Vertical profiles of sediment concentration deviated significantly from the Rouse distribution (Dinehart, 1987).
On March 19, 1982, an opportunity for observation of a lahar provided rare synoptic measurements at three gaging stations on the Toutle River separated by 23 mi. High concentrations of fine sediment apparently enhanced the ability of the flow to transport high concentrations of sand. The observations prompted sedimentologists to interpret some ancient flow deposits in the Cascades Range as products of a similar process (Pierson and Scott, 1985).
The need for current information on stream and sediment discharge led to the streamlining of office and lab procedures. Sediment-discharge data were provided to the U.S. Army Corps of Engineers and other interested agencies within 10 working days of a storm flow.
Sediment data collected at gaging stations were used to calculate the expected volume of sediment that the SRS on the North Fork Toutle River would need to retain. Several factors, including the retention of a hypothetical flow from a breakout of Spirit Lake, were included in the volume estimates.
Channel cross-section data were entered into a detailed database of channels in the Mount St. Helens area. A database/graphics program (MAPLE) was designed to calculate elevations from field surveys and to plot channel cross sections in several formats.
The downstream drag of sounding weights and sediment samplers in high stream velocities was counteracted with staylines installed upstream of several gaging-station cableways. A radio remote-control feature, tested at CVO, allowed precise positioning of the tether across the river section.
At Mount St. Helens, stream-discharge measurements were made about five times as often during the year as at other gaging stations. A data entry, computation, and retrieval program (CHEK) was designed to reduce the manual computations involved in checking discharge measurements. Plotting routines and programs for detailed hydraulic calculations allowed versatile use of the discharge data.
Sediment data were made available in an organized format to data users as soon as laboratory analyses were completed. The SSDS included data entry and accounting procedures for the sediment laboratory, retrieval and analysis programs for hydrologic technicians, and remote access to the data for users in other U.S. Geological Survey offices. Laboratory data were directly accessible for computation of sediment-discharge records.
Chart traces of river stage and sediment concentration were digitized into computer format with a program (WASH) designed to compute daily water and sediment discharge. The program retrieved sample data from the SSDS, which were plotted adjacent to continuous discharge records. Manual computations were minimized, and tabular outputs were designed to meet the needs of reviewers within the U.S. Geological Survey.
Helley-Smith bedload samplers with 3 x 3 in. openings were too small for the large bedload transported in the Toutle River. Several modifications to the existing design in 1986 (6- x 12-in. opening, larger sample bag) improved bedload sampling during storm flows (Childers, 1992).
To evaluate sediment discharge from the North Fork Toutle River debris-avalanche deposit, a 1,000-ft cableway was installed over a typical braided reach, upstream from the existing and proposed sediment dams. Hydrographers encountered extremes of velocity and bedload transport at this site. The sediment data were used in estimates of sediment delivery to the planned sediment-retention structure.
When an overflow tunnel to control the elevation of Spirit Lake was completed, the runoff was sent into South Coldwater Creek where sustained high discharges had not occurred since its inundation by the debris avalanche in 1980. The erosion and deposition of sediment along the channel and in Coldwater Lake were measured by several crews during April and May of 1985 (D.F. Meyer, U.S. Geological Survey, written. commun., 1992).
An ultrasonic sediment-concentration meter was installed at the North Fork Toutle River at Kid Valley for evaluation. The commercial unit (Markland Co., Canada) showed near-instantaneous fluctuations in concentration. Long-term trends on the instrument chart corresponded well with changes in stream discharge and were comparable to sediment-concentration curves from samples (S.A. Gustafson, U.S. Geological Survey, written commun., 1987). Although the meter was operable in concentrations from 1,000 to 70,000 mg/L, precision increased with concentrations greater than 10,000 mg/L.
Depth-sounding sonar equipment was evaluated on the Toutle River. Sonar detected the passage of dune bedforms composed of fine gravel. An installation for sonar measurements at high flows was constructed at the gaging station on the North Fork Toutle River and was used throughout several high water periods. Several small dunes were grouped into longer bedforms (30 to 60 ft) that evolved during storm flows (Dinehart, 1989).
Manually operated cable cars with battery-powered reels were replaced by gas-powered hydraulic cable cars at two gaging stations on the Toutle River. Hydraulic motors on the cable-car wheels and on the sampling reels eased the strain caused by heavy bedload samplers used in the Toutle River.
Vibrating U-tube technology, used in industry for mass-density measurements of fluids (Dynatrol, Automation Products), was adapted by the U.S. Geological Survey for measurement of sediment concentration in streams. Two prototype units were sent to CVO and put into use on the Toutle River. Results obtained during 1987-90 were comparable to sampled sediment concentrations (S.A. Gustafson, U.S. Geological Survey, written commun., 1989).
A series of cross sections were surveyed over a 1,000-ft reach of channel at the North Fork Toutle River gaging station at Kid Valley. A three-dimensional computer model of flow and sediment transport (Nelson and Smith, 1989) was used to simulate the scour and fill observed at the station (Shimizu and others, 1989). A modeled prediction of fill at the cableway cross section occurred in December 1989.
Although bedload discharge is usually less than suspended-sediment discharge, bedload movement measurably affects channel geometry and roughness. Sediment data collection was sufficient for indirect measurements of total sediment discharge, which includes bedload. Bedload-transport formulas were applied to the sediment data to judge the applicability of various formulas to the Toutle River (Hammond, 1989).
Direct measurements of bedform wavelengths were needed to describe gravel bedforms. An observation platform was built to measure a 4- x 25-ft swath of river bed with three sonar transducers on a moving carriage. The platform was installed in October 1988 at the North Fork Toutle River gaging station. Wavelengths of gravel bedforms were at least as long as the platform. Transport rates of gravel bedload, sampled adjacent to the platform during stationary measurements of bed elevation, corresponded with bedform migration (Dinehart, 1992a).
Electromagnetic velocity meters give continuous velocity readings that can be acquired by portable computers. Velocity profiles were measured with three vertically mounted meters throughout a range of discharge conditions for up to 8 hours an episode. Records of this type showed velocity pulsations of several minutes duration (Dinehart, 1992a).
Gaging stations were installed above and below the SRS on the North Fork Toutle River. The gage on the lake behind the SRS was equipped with a new design pressure transducer having a 100-ft range (Paroscientific Digiquartz).
Monumented channel cross sections that had been established in the early 1980s (Meyer and others, 1986) were resurveyed to evaluate the effects on the North Fork and the mainstem Toutle Rivers downstream from the SRS. Degradation of the Toutle River channel was monitored with periodic cross-section surveys. Streambeds were scoured locally by infrequent high flows, as observed at the North Fork Toutle River at Kid Valley. Overall, minimal scour was measured at most cross sections along the Toutle River.
Bedload was sampled simultaneously with sonar measurements of coarse-gravel bedforms at the North Fork Toutle River at Kid Valley. Growth of bedforms in gravel beds during storm-flow recession was identified as a recurrent process. The changes in mean bed elevation and the increased form drag induced by gravel bedforms affected the stage-discharge relation measurably (Dinehart, 1992a).