Uncertainty in Estimation of Debris-Flow Triggering Rainfall: Evaluation and Impact on Identification of Threshold Relationships

Operational debris-flow warning systems are often based on the use of empirical rainfall thresholds derived from rain gauge observations. However, rain gauges are usually located away from the actual debris-flow locations thus estimation of triggering rainfall properties from rain gauges can be associated with considerable uncertainty. This work examines the uncertainty in gauge-based estimation of debris-flow triggering rainfall and evaluates its impact on the identification of rainfall thresholds used for debris-flows prediction. These issues are assessed by using high-resolution radar data to represent "actual" space-time patterns of precipitation at and around the debris-flow initiation points. Rain-gauge network sampling is simulated by randomly sampling radar-rainfall fields. Rainfall is estimated by using three rainfall interpolation methods: Nearest neighbor (NN), inverse distance weighting (IDW), and ordinary kriging (OK). Comparison of results from these three methods shows that no particular benefit in intensity-duration threshold estimation is obtained by using approaches that are more complex than the NN method. NN provides estimates with smaller bias than IDW and OK but larger estimation variance. On average, decrease in gauge density leads to increased underestimation of debris-flow rainfall and subsequently this results in large underestimation of the intensity-duration thresholds.