Early warning systems for rainfall-induced debris flows often rely on rainfall thresholds, intended as combinations of rain intensity and duration, which, when exceeded, are expected to trigger debris flows. Rainfall thresholds are generally based on rain gauge observations. However, rain-gauge-based estimates of debris flow triggering rainfall have been shown to be affected by large uncertainties, mainly due to the distance between debris flows and rain gauges, which may lead to a step degradation of the performances of the warning systems. A potential solution to the observational limitations of rain gauges lies on the use of weather radar observations. Radar rainfall estimates in mountainous areas are affected by a number of errors, which require specific correction schemes. This chapter assesses the use of radar rainfall estimates for the estimation of debris flows triggering rainfall and analyzes the performance of different radar and gauge estimates for the identification of debris flows’ triggering thresholds. The study is based on seven storm events that triggered 117 debris flows in the Eastern Italian Alps. Data from a C-band Doppler weather radar are reanalyzed by combining physical corrections and quantitative adjustment. Four radar rainfall scenarios characterized by increasing accuracy and requirements in terms of data and elaborations are used to derive rainfall thresholds. Results show that raw radar estimates lead to underestimation, and bias adjustment does not provide significant improvement. Implementing physically based corrections provides thresholds with less than 6% error. Thresholds derived from gauges as well as from gauge-colocated radar estimates are examined and found substantially underestimated. This suggests that great benefit can be provided by radar rainfall estimates when appropriate correction schemes are implemented.
Radar rainfall estimates for debris-flow early warning systems: Effect of different correction procedures on the identification of intensity-duration thresholds
Marra F.;Nikolopoulos E. I.;Borga M.
2017
Abstract
Early warning systems for rainfall-induced debris flows often rely on rainfall thresholds, intended as combinations of rain intensity and duration, which, when exceeded, are expected to trigger debris flows. Rainfall thresholds are generally based on rain gauge observations. However, rain-gauge-based estimates of debris flow triggering rainfall have been shown to be affected by large uncertainties, mainly due to the distance between debris flows and rain gauges, which may lead to a step degradation of the performances of the warning systems. A potential solution to the observational limitations of rain gauges lies on the use of weather radar observations. Radar rainfall estimates in mountainous areas are affected by a number of errors, which require specific correction schemes. This chapter assesses the use of radar rainfall estimates for the estimation of debris flows triggering rainfall and analyzes the performance of different radar and gauge estimates for the identification of debris flows’ triggering thresholds. The study is based on seven storm events that triggered 117 debris flows in the Eastern Italian Alps. Data from a C-band Doppler weather radar are reanalyzed by combining physical corrections and quantitative adjustment. Four radar rainfall scenarios characterized by increasing accuracy and requirements in terms of data and elaborations are used to derive rainfall thresholds. Results show that raw radar estimates lead to underestimation, and bias adjustment does not provide significant improvement. Implementing physically based corrections provides thresholds with less than 6% error. Thresholds derived from gauges as well as from gauge-colocated radar estimates are examined and found substantially underestimated. This suggests that great benefit can be provided by radar rainfall estimates when appropriate correction schemes are implemented.
Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
