Combining weather radar and raingauge data for hydrologic applications

In 2001 the paper titled “A Bayesian technique for conditioning RADAR precipitation estimates to raingauge measurements” (Todini, 2001) introduced a new technique based upon the use of block-Kriging and of Kalman filtering to combine, optimally in a Bayesian sense, areal precipitation fields estimated from meteorological radar to point measurements of precipitation, such as are provided by a network of raingauges. Block Kriging was used to estimate the average field over the radar pixels and its variance from the point rain gauge measurements, while a Kalman filter was taken to find the a posteriori estimates by combining the a priori estimates provided by the RADAR with the block Kriged measurements provided by the gauges, in a Bayesian framework. An original Block Kriging approach to the problem of spatial interpolation was introduced, which also included a new formulation of Kriging with uncertain point precipitation measurements. A Maximum Likelihood estimator was used at each step in time to estimate the semi-variogram parameters, while a new non negativity constraints were added to the Kriging system to prevent negative values in the Kriging weights. The paper summarizes the new potentialities of the new system and shows the results obtained in real world applications.