For remote sensing of precipitation from radar and satellites to fulfill its ultimate potential it will be necessary to improve the precision of rain retrievals and to optimize the assimilation of the data into numerical models. It is well known from developments in data assimilation that for optimal performance models require information on the error covariance of the parameter being measured. In general, this means information is required on the measurement error of instruments as well as the representativeness of the measurements themselves. In the case of precipitation measurement it is necessary to develop more information on the variability of precipitation fields within precipitating cloud systems. This presentation considers approaches to the determination of space time variability of precipitating cloud systems using profilers and scanning radars.