Quality control of precipitation data

Abstract

Valuable information about natural processes is acquired from measurements and the data they produce. However, such sources of information are subject to disturbances of different kinds: device specific errors, uncertainties over data storage, transmission problems and data transformation effects to name but a few. Therefore, data acquired from measurements do not always represent the real behavior of the observed processes. To read the data and identify their reliable and unreliable components, normally, requires experience and a lot of effort. Nevertheless, for many applications, from pipe design to flood warning, a good quality of data is required, usually reflecting an economic value.