Detection of spurious precipitation signals from automatic weather stations in irrigated areas

Tipping-bucket rain gauges are convenient and reliable sensors of rainfall measurements; however, like all other field sensors, they are subject to different kinds of errors. Due to their location, rain gauges in this research can record accidental pulses produced by vibrations from works of farm machineries near the station, or may receive water from sprinkler irrigation systems. These spurious inputs are recorded as precipitation data, although they do not correspond to rain, so it is necessary to detect them in order to avoid their inclusion in the future soil-water balance. The main objective of this work is to design a simple quality control procedure to validate precipitation data generated in several stations of the Agroclimatic Information Network of Andalusia (southern Spain), and valid for similar agro-meteorological station networks. The relationship between the degree of cloudiness through attenuation of solar radiation (atmospheric transmittance coefficient), relative humidity and rainfall measurements has been studied in order to separate true and false precipitation records. Meteorological data from 2002 to 2011 were used for the study.