Hypothesis

Data

Data available to address those questions include Corps of Engineers stream surveys in the fall of 1959 and water level and terrace surveys in 1989 by TRT. These surveys (Figure 3) document an evolution which differs from the simple model above in some ways, but follows it in others. Only the earliest (9/10/59) profile is dominantly linear; the modern water surface profile is kinked, reflecting the effect of lag boulders in stabilizing steeper slopes in the early, coarse sediments. The terraces are best preserved (e.g., T2, T5) as the locally highest surface, but are also preserved as fragments inset into higher, older deposits.

Interpretation

Terrace Formation

Although terrace formation (incision/downcutting) reflects an increase in the ratio of power to load, that change in turn can be caused by increasing discharge, decreasing load, or complex responses within the system which cause local instabilities (S. A. Schumm, 1973, in Fluvial Geomorphology, Fourth Binghamton Symposium, M. Morisawa, ed., 299-310). In the case of the Madison River terraces, comparison of sequential aerial photography brackets the times of terrace formation. Each of those times also includes a period of high discharge (note that the 10-year flood is about 113 cms [m³s^-1] and the 50-year flood is about 191 cms). It seems clear that in the case of the Madison River downstream from the Madison 'Slide, incision to form terraces requires a relatively high stream discharge.

Table 1 - Terrace Times and Discharges

Terrace Profiles

Longitudinal profiles of the Madison River downstream from the slide (Figure 4) show a concave-upward shape for both the modern water surface and the highest terraces. The two profiles converge at the point where modern aggradation is taking place (see linked photo above). Aggradation is now occuring 2-3 kilometers downstream from the outlet of 'Quake Lake, as predicted by the conceptual model of Figure 2. It seems evident from the convergence of pre-slide, present, and terrace elevations not far from the margin of the slide debris (about 1000 m downstream from the lake outlet) that the disequilibrium imposed on the river by the point addition of sediment load is being assimilated across a relatively short distance.

Conclusions

Terrace Evolution - Space

The amount of disequilibrium is extremely well fit (r² exceeds 0.99) by a power function (Figure 5). Here "disequilibrium" is defined as the deviation in profile elevation from the pre-slide profile (assumed to have been at dynamic equilibrium). Extrapolation of this model to twice the length of available data (possibly justified by the extremely strong fit) predicts that the Madison River will never aggrade significantly beyond about 5-6 km downstream from the slide. Here "significantly" is defined by a 1-m envelope of natural variation around the pre-slide profile. The equation for the curve of maximum aggradation, thus limiting the geomorphic effect of the slide, is clearly empirical. It is affected by the amount of load available and the caliber of that load, by the shape of the valley confining the slide (thus superelevating the river) and below the slide (thus storing fluvial sediment), and by the discharge of the river. The propagation of the "sediment wave" from placer mining studied by Gilbert and James for hundreds of km from its source certainly results from efficient transmission through the canyons of the Sierra Nevada to the Central Valley of California.

Terrace Evolution - Time

Although the distance of terrace progradation is apparently largely deterministic, the time required for ultimate progradation is logically probablistic. In the extreme case, slide dam failure would have eroded the slide mass back to near the pre-slide profile, geologically instantaneously. In the process, terraces such as described here would have formed over minutes to hours. Similarly, an extreme event (like failure of the Hebgen dam, immediately upstream from 'Quake Lake) could at any time cause nearly complete equilibration of the system. In the absence of such an event, however, it will require progressively larger discharges to remobilize the coarse sediment on the progressively gentler terrace profiles, thus formation of terraces should become less frequent. A crude regression on the ages of the terraces formed to date (Figure 6) suggests that it will require at least hundreds of years and may require tens of thousands of years for the locus of incision following aggradation to migrate to a point 5-6 km downstream from the slide.

Implications

The implications of this study are multiple and manifold. The coarse sediment generated as point loads (glacier termini, placer mines...) will cause local aggradation downstream, followed by incision, thus terrace formation. "Local" will vary depending on characteristics of the load, discharge, and valley, but will generally be less than a few tens of kilometers in unconfined valleys, even in the case of large glacier termini (see Paradise Valley in these pages). This progradation will occur rapidly at first, but more slowly as disequilibrium decreases. It is possible that pulses of aggradation relating to the last deglaciation are still affecting river valleys tens of kilometers downstream from the former glacier termini! Finally, although net aggradation ("fill terraces") will be confined to the immediate vicinity of the load source, other effects will certainly be felt farther downstream. The long-term increase of bedload at a caliber transportable by the river will result in decreased downcutting and channel widening downstream, thus formation of a broad floodplain. When incision recommences, a cut terrace transitional to the fill terraces upstream will be formed, as documented by J. H. Moss (1982, in Glacial Geomorphology , D. R. Coates, ed., Fifth Binghamton Symposium, 293-314).