Debris flows in Grand Canyon are one of the main controls on the geomorphic framework of the Colorado River, depositing large boulders that define the river’s longitudinal profile. These mass movements are initiated when weathered bedrock or colluvial wedges fail during intense rainfall. These failures can be classified according to four failure mechanisms: 1) direct failure of weathered bedrock (12 percent); 2) failure of colluvium from the impact of runoff cascading from cliff pourovers, known as the “firehose effect” (36 percent); 3) failure of saturated colluvium (21 percent); and 4) combinations of these three mechanisms (30 percent).
A variety of geomorphic factors relating to climate, exposed lithologic strata, geologic structure, and drainage-basin morphology play various roles in the process of debris-flow initiation. Among these factors, the exposure of shale source areas at heights greater than 100 m above the mainstem is particularly critical in each of the four initiation mechanisms. Shales fail readily as weathered bedrock, produce abundant colluvial source material, and form slopes where colluvium collects. Shales also provide the fine particles and clay minerals (for example, kaolinite and illite) that appear to be essential to the stability of slurries, enabling debris flows to transport poorly-sorted sediments up to several kilometers from tributary sources to the Colorado River.
Earlier work has investigated the statistical significance of various morphometric variables in relation to flood magnitude, usually emphasizing drainage-basin area, mean-basin elevation, and amount or intensity of precipitation (Roeske, 1978; Thomas and others, 1994). Patton and Baker (1976) suggest that stream magnitude may also be a good predictor of flash-flood potential for small drainage basins. They argue that transient controls, such as climatic variability, also play a significant role. Shown (1970) includes all of these variables in modeling sediment transport in the southwest, as well as factors relating to surface geology and soils such as rock type, hardness, weathering, and texture.
Those drainage-basin variables that are most significant in influencing the binomial frequency of debris flows are well illustrated by a plot of the modeled probability of debris flows in all 600 geomorphically significant tributaries of Grand Canyon between Lees Ferry and Surprise Canyon (river miles 0-248; fig. 18). The results of the logistic regression model clearly demonstrate that debris flows in Grand Canyon do not occur randomly in space. The binomial frequency of debris flows is substantially higher in eastern Grand Canyon, where 51 percent of the tributaries have had a probability of debris flows greater than 75 percent during the last 100 years (fig. 19). In contrast, 17 percent of tributaries in western Grand Canyon have had a probability of debris flows greater than 75 percent. For all of Grand Canyon, 60 percent of the tributaries have had a debris-flow probability greater than 50 percent, although that probability was 74 percent in eastern Grand Canyon and 47 percent in western Grand Canyon.
|
| Figure 18. Map showing the probabilities of debris-flow occurrence during the last century in 600 tributaries of the Colorado River in Grand Canyon. |
 |
| Figure 19. Histograms of the probability of debris-flow occurrence in eastern and western Grand Canyon. |
Distinct patterns are evident in the debris-flow probability map, such as an overall higher probability of debris flows in eastern Grand Canyon (fig. 18). This is due to a combination of an increase in precipitation and a general trend in river aspect to the southwest that takes maximum advantage of increased precipitation. Also significant in eastern Grand Canyon is the presence of the Hermit Shale at an optimum height above the river. All three elements combine to make the stretch from Lees Ferry to 75-Mile Canyon the most prolific zone of debris-flow production in the Grand Canyon, as evidenced by the high probability of debris-flow occurrence (fig. 18).
These same trends also occur in western Grand Canyon. Debris flows fall off notably where the canyon turns to the north-west, perpendicular to the movement of severe storms. Although climatic variables were not selected as statistically significant in the models, they still seem to have an underlying influence. Debris-flow probability is lowest downstream from Diamond Creek, where precipitation is lowest and the river trends northwest. The results of the logistic regression model clearly demonstrate that variation in lithologic, morphometric, and climatic variables have a strong effect on the probability of debris flows in Grand Canyon.
Back: Logistic-Regression Analyses
| AccessibilityFOIAPrivacyPolicies and Notices | |
  |
|