Rainfall intermittency and the sampling error of tipping-bucket rain gauges

Abstract

This work amplifies upon the influence of rainfall intermittency on pluviometric time series. Intermittency is here defined as the percentage of no-rain periods within a rainfall event, and can be interpreted as a sort of ‘stochastic intermittency’. When measuring an intermittent or, in general, ‘erratic’ signal, a relevant source of error is associated with the sampling procedure. By using numerical simulation of intermittent rainfall events, we focus on the deriving pattern of sampling errors within rainfall measurements obtained from any classic tipping-bucket rain gauge. The analysis, performed in both the time and frequency domain, reveals a strong inverse linear dependence between rainfall intermittency and the sampling errors, while a weak dependence on the autocorrelation of synthetic rainfall events is shown. The average absolute error is around 30% for the events analysed while the error at each sampling period peaks values higher than 100%. A direct influence of intermittency on the statistical characteristics of measured rainfall events is observed.