Uncertainty of Tipping-Bucket Data May Hamper Detection and Analysis of Secular Changes in Short-Term Rainfall Rates

Exploring the secular tendency to intensification of short-interval rainfall intensities, such as those associated with convective storms, requires rainfall data having sufficient accuracy and temporal resolution. Light rainfalls also exhibit secular change, and documenting these imposes considerable demands on data quality. Tipping-bucket rain gauges are the most widely deployed globally for data collection, but they cannot record rainfall amount or rainfall rate instantaneously. Both require data to be collected over some finite time interval, the accumulation time (AT), during which one or more buckets must fill and tip. Relatively short ATs, such as when analysing 15 min rainfall amounts and rates, are associated with increased uncertainty in TBRG data. Quantifying the resulting uncertainty forms the subject of the present work. Worst-case rainfall depth and rainfall rate errors that would arise in TBRG data for constant rainfall rates of 5, 10, 20, 30, 40, and 50 mm h−1 are determined for ATs from 5 min to 50 min. Errors frequently considerably exceed the 1–2% accuracy levels claimed by many manufacturers of TBRGs. The errors found pose challenges for the detection of secular change in rainfall rates. The present results point to the need for fuller analysis of errors in TBRG data for short-duration rainfalls and for gauge specifications to specify uncertainty separately for rainfall depth and rainfall rate.