Analysis of weather radar and rain gauges for flood forecasting

Real time rainfall for flood forecasting is predominately measured by rain gauges and weather radars. Analysis of rainfall data measured simultaneously by a dense rain gauge network and weather radar over the Brue catchment, during a period of six years, was undertaken using Principal Component Analysis and Cluster Analysis. Wind, temperature, relative humidity for the major storms events over the Brue were modelled by the numerical weather model MM5. PCA provided information regarding the state of the existing rain gauge network in terms of redundancy and adequacy of the rain gauge network. Cluster analysis was used to identify the best rain gauge locations. The research using the eight greatest storm events assisted the integration of measured rainfall and minimisation of the weaknesses of the three main information sources relevant to real time rainfall estimations (rain gauge, weather radar and numerical weather model rainfall)