EFFICIENT USE OF AVAILABLE DATA FOR REGIONAL PREDICTION OF DEBRIS FLOW HAZARDS

The widespread availability of digital elevation data coupled with the computational power of modern desktop computers provides an unprecedented opportunity for evaluation of regional debris flow hazards. Guided by process-based conceptual models, we can build simple, spatially distributed numerical models to predict debris flow behavior in terms of readily derived topographic attributes.

In devising such models, we have examined debris flow occurrence in terms of three factors: initiation, transport, and deposition. We propose a set of topographic attributes that act as primary controls on each of these and then use field observations to test and calibrate the models. We expect that debris-flow-triggering landslide locations are primarily correlated with surface gradient; that the propensity for debris flow transport and channel scour are correlated with channel gradient and valley width; and that depositional zones are correlated with cumulative upslope source volume and valley or fan geometry.

All of these topographic factors can be estimated using digital elevation data, allowing us to make specific, testable predictions – at regional scales – of where debris flows occur and of the spatial extent of areas impacted by debris flow transport and deposition. We’ll discuss the conceptual models, tests of their validity and applicability in light of available data, and show results for regional mapping of debris flow hazards in western Oregon.