Improving the spatial monitoring of rain gauges for flood assessment

Abstract

Flood assessment for early warning systems rely on the confidence on data. Rainfalls are usually estimated by radar with a low accuracy but a fine spatial mesh and rain gauges with a high accuracy but a low density. Adding rain gauges could be of added value for assessing the flood.

First, a method for analysing the role of added rain gauges on the reduction of uncertainties of spatial average rainfall is proposed. Kriging method with external drift was applied for interpolation estimation. The uncertainty of the interpolation is also quantified. Also the distribution of the weights of the rain gauges was analysed simultaneously using Thiessen polygon. When comparing rainfall from rain gauges with different amounts of additions, the impact of change is evaluated using the indicator of mean area rainfall. In this study, we analyse the distribution of rain gauges in the Landro basin and the corresponding changes in precipitation. The spatial distribution of rain gauges within the basin was optimised and improved.

Finally the rainfall data was applied to the Iber software and the changes in river flow were obtained by means of the corresponding hydrological model. The evolution of the probability of false alarm and the probability of good assessment are also estimated.

This work is carried aout in the AAFLOOD Atlantic Area Interreg project.