Influence of systematic errors from tipping bucket rain gauges on recorded rainfall data

Abstract

Tipping bucket rain gauges (TBR) are widely used in urban hydrology. The present study investigated the uncertainties in recorded rainfall intensity induced by the following properties of the TBR: depth resolution i.e. the bucket volume, calibration parameters, wetting and evaporation losses and the method of data recording (time between tips or tips per minute). The errors were analysed by means of a TBR simulator i.e. a simulation program that models the behaviour of a TBR. Rainfall data disaggregated to 6 seconds from measured 1-min data and randomly varied were taken as input to the simulator. Different TBR data series were produced by changing the properties of the simulated rain gauge. These data series together with the original rainfall events were used as input to a rainfall-runoff model. Computed overflow volume and peak discharge from a combined sewer overflow (CSO) weir were compared. Errors due to depth resolution (i.e. the bucket size) proved to be small. Therefore TBRs with a depth resolution up to 0.254 mm can be used in urban hydrology without inducing significant errors. Wetting and evaporation losses caused small errors. The method of data recording had also little influence. For larger bucket volumes variable time step recording induced smaller errors than tips per minute recording.