Regional Assessment of Sub-Hourly Annual Rainfall Maxima

Abstract

The assessment of rainfall extremes at sub-hourly scales is generally hindered by a lack of rainfall data at small timescale resolutions. This study proposes a methodology for assessing mean annual maximum rainfall at the sub-hourly scale by blending historical time series of annual maxima recorded by mechanical stations (operating at hourly scales) up to the end of the past century with newer time series of annual maxima at higher time resolutions recorded by automatic stations installed over the past twenty years. A linear correlation was found at the regional scale between the shape parameter controlling the dependency of rainfall maxima with a duration longer than one hour and the shape parameter of the dependency of rainfall maxima with the durations shorter than one hour. Thanks to this correlation, data recorded at the mechanical stations could be exploited to assess sub-hourly mean annual maxima. The proposed hybrid procedure was verified and was found to provide estimates with an accuracy close to those obtained with the high-resolution data, i.e., our best estimates. Moreover, the proposed procedure outperforms what could be achieved by spatially interpolating the best estimates at those locations where only hourly data are available.