The monitoring and forecasting of severe weather has the potential to enable civil defence organisations to alleviate the natural hazards severe weather can cause, such as floods. This project has focussed on developing the understanding of weather radar data, and its application to flood forecasting. Research has also applied radar and satellite data in developing the initialisation of a numerical weather prediction model to improve the forecasting of severe weather. Standard weather radars, commonly used throughout Europe, are unable to differentiate between rain, hail, and melting particles. Neither can they distinguish between precipitation, and echoes caused by non precipitation, such as buildings or hills, or anaprop which can occur during occasions when a temperature inversion exists. In this project we have had access to data from both polarisation and Doppler radars, and ideas and experience from one part of Europe have been successfully applied elsewhere. For example, a method of anaprop detection devised for the Baltic has been applied operationally in the Po Valley in Italy. Moreover, modelling has helped explain the occurrence of a Baltic anaprop event. Theoretical understanding of polarisation signatures has been practically applied, also in the Po valley, to the interpretation of echoes from a severe storm event, in which two supercell storms occurred producing a large hailswath over 200 km long. This has significantly helped the understanding of the storm dynamics. The analysis of data from a research radar has provided insight into the measurement problems that can occur with the use of advanced radars, and on the relative merits of different polarisation schemes. An investigation has been made into the effects for hydrological modelling of errors in radar data on rainfall. Advances have also been made in the three-dimensional visualisation of storms, and in the retrieval of wind information from Doppler radars.