Accounting for rainfall systematic spatial variability in flash flood forecasting

Abstract

Just as with the storms that cause them, flash floods are highly variable and non-linear phenomena in both time and space; hence understanding and anticipating the genesis of flash floods is far from straightforward. There is therefore a huge requirement for tools with the potential to provide advance warning of situations likely to lead to flash floods, and thus provide additional time for the flood forecasting services. The Flash Flood Guidance (FFG) method is used on US catchments to estimate the average number of inches of rainfall for gJust as with the storms that cause them, flash floods are highly variable and non-linear phenomena in both time and space; hence understanding and anticipating the genesis of flash floods is far from straightforward. There is therefore a huge requirement for tools with the potential to provide advance warning of situations likely to lead to flash floods, and thus provide additional time for the flood forecasting services. The Flash Flood Guidance (FFG) method is used on US catchments to estimate the average number of inches of rainfall for given durations required to produce flash flooding. This rainfall amount is used afterwards as a flood warning threshold. In Europe, flash floods often occur on small catchments (approximately 100 km2) and it has already been shown that the spatial variability of rainfall has a great impact on the catchment response (Le Lay and Saulnier, 2007). ThereforeJust as with the storms that cause them, flash floods are highly variable and non-linear phenomena in both time and space; hence understanding and anticipating the genesis of flash floods is far from straightforward. There is therefore a huge requirement for tools with the potential to provide advance warning of situations likely to lead to flash floods, and thus provide additional time for the flood forecasting services. The Flash Flood Guidance (FFG) method is used on US catchments to estimate the average number of inches of rainfall for given durations required to produce flash flooding. This rainfall amount is used afterwards as a flood warning threshold. In Europe, flash floods often occur on small catchments (approximately 100 km2) and it has already been shown that the spatial variability of rainfall has a great impact on the catchment response (Le Lay and Saulnier, 2007). Therefore, in this study, an improved FFG method which accounts for rainfall spatial variability is proposed. The objectives of this paper are (i) to assess the FFG method applicability on French Mediterranean catchments with a distributed process-oriented hydrological model and (ii) to assess the effect of the rainfall spatial variability on this method. The results confirm the influence of the spatial variability of rainfall events in relation with its interaction with soil properties., in this study, an improved FFG method which accounts for rainfall spatial variability is proposed. The objectives of this paper are (i) to assess the FFG method applicability on French Mediterranean catchments with a distributed process-oriented hydrological model and (ii) to assess the effect of the rainfall spatial variability on this method. The results confirm the influence of the spatial variability of rainfall events in relation with its interaction with soil properties.iven durations required to produce flash flooding. This rainfall amount is used afterwards as a flood warning threshold. In Europe, flash floods often occur on small catchments (approximately 100 km2) and it has already been shown that the spatial variability of rainfall has a great impact on the catchment response (Le Lay and Saulnier, 2007). Therefore, in this study, an improved FFG method which accounts for rainfall spatial variability is proposed. The objectives of this paper are (i) to assess the FFG method applicability on French Mediterranean catchments with a distributed process-oriented hydrological model and (ii) to assess the effect of the rainfall spatial variability on this method. The results confirm the influence of the spatial variability of rainfall events in relation with its interaction with soil properties.