Geological Survey Professional Paper 1547 Sedimentology, Behavior, and Hazards of Debris Flows at Mount Rainier Washington

METHODS OF STUDY

The distribution of past flows in time and space is an excellent guide to the probability and extent of future flows. This study amplifies the landmark work of Crandell (1971) by likewise focusing on lahars formed during postglacial time at Mount Rainier. Older lahars, however extensive (lahar deposit along lower Cowlitz River, Bethel, 1981), are not included because of their possible disturbance by, and potential confusion with, extensive glacial deposits. (See Dethier and Bethel, 1981.) Also, not only is the record of older lahars probably incomplete, but the conditions of their formation probably differed from present conditions.

Certain flow deposits that are distinctive and correlative over significant distances are named informally in accordance with the North American Stratigraphic Code for key or marker beds (North American Commission on Stratigraphic Nomenclature, 1983). Crandell's (1971) nomenclature is retained for the units he recognized. Given the textural differences in lahar and lahar-runout facies (Scott, 1988b, fig. 10), the older flow units are dated and correlated based on tephra units, radiocarbon dating, soil formation, and clast characteristics. Table 1 summarizes the prominent tephra units and other major postglacial volcanic events at Mount Rainier.

¹Years before 1950 in radiocarbon years, except as otherwise indicated.

The dynamics of flows are integral to any discussion of flow behavior and hazards. Original flow-wave volumes are estimated from deposit volumes as discussed for specific flows. Ancient flow discharges are determined from flow velocities, as calculated from paleohydrologic techniques using vertical runup or superelevation around bends (see Costa, 1984, p. 304—305; Johnson, 1984, p. 305—309), and from cross-sectional areas derived from levels of flow deposits on valley-side slopes. Recently, discharges calculated using superelevation around bends were called into question by Webb and others (1989, p. 22, table 10), who noted large differences in cross-sectional areas calculated for flows in straight and curved reaches. We believe such discrepancies result because flow surfaces, particularly in very sharp bends, become markedly concave, as is well shown in photographs of 1990 debris flows at Jiangila Ravine in Yunnan Province, China, by R.J. Janda and K.M. Scott (U.S. Geological Survey). Most cross sections at Mount Rainier (pl. 1) are not from comparably sharp bends, but the observation of Webb and others still stands as a cautionary note for any discharge calculated from a section in a bend (such as section T—1, pl. 1).

Post-flow valley erosion is generally small and, in dealing with flow depths of many tens of meters, is generally not more than several percent of the flow cross-section. The dated tephra layers (table 1) define the levels of valley bottoms at successive postglacial intervals, as well as the erosion of valley-side slopes through time. Cycles of aggradation and degradation related to Neoglacial advances and retreats are critical only in defining the cross-sectional areas of the relatively small glacial-outburst flows. In general, the accuracy of discharges determined with these techniques, proportional to the size of the flows, compares with that of other indirect-discharge determinations.

Some velocity measurements determined from runup and superelevation are suspect, however, because the techniques are unverified for debris flows and because of some locally high values (Costa, 1984), especially those above 30 m/s. Some high values measured for 1980 flows at Mount St. Helens (Fairchild, 1985; Pierson, 1985; Scott, 1988b) resulted from the lateral momentum provided by catastrophic ejection of wet debris that settled and flowed as a lahar. The most relevant velocities in this study are those determined near the points where flows left the confined valleys of the Cascade Range and inundated the Puget Sound lowland.

Some aspects of flow rheology can be inferred from texture and fabric comparisons with modern flows (Scott, 1988b). The texture of the deposits is determined from a combination of field measurements of the b axes of gravel-size (>2 mm) particles at a level or in a grid-defined area (Wolman, 1954), and laboratory measurements of the sand, silt, and clay fractions by sieve and pipet. This combination of techniques is statistically valid (Kellerhals and Bray, 1971), and a technique for combining the two is described by Scott (1988b). The approach thus incorporates the complete spectrum of sizes in the deposit, whereas the size analyses of lahars reported in Crandell (1971) do not include sediment in the coarse cobble and boulder-size ranges. Crandell's approach could result in a reported clay content, for example, of twice the actual clay content for a lahar that contains 50 percent coarse cobbles and boulders. Differences in results were assessed by comparing his analyses with field counts of the coarser fractions at many localities. The analyses reported by Crandell (1971) are useful in showing the relative but not the absolute differences in clay content. The grain-size measures reported here are graphically determined, following the method of Folk (1980).