Rain event properties in nature and in rainfall simulation experiments: a comparative review with recommendations for increasingly systematic study and reporting

In hydrology and geomorphology, less attention has been paid to rain event properties such as duration, mean and peak rain rate than to rain properties such as drop size or kinetic energy. A literature review shows a lack of correspondence between natural and simulated rain events. For example, 26 studies that report event statistics from substantial records of natural rain reveal a mean rain rate of just 3·47 mm h −1 (s.d. 2·38 mm h −1 ). In 17 comparable studies dealing with extreme rain rates including events in cyclonic, tropical convective, and typhoon conditions, a mean maximum rain rate (either hourly or mean event rain rate) of 86·3 mm h −1 (s.d. 57·7 mm h −1 ) is demonstrated. However, 49 studies using rainfall simulation involve a mean maximum rain rate of 103·1 mm h −1 (s.d. 81·3 mm h −1 ), often sustained for > 1 h, exceeding even than of extreme rain events, and nearly 30 times the mean rain rate in ordinary, non‐exceptional, rain events. Thus rainfall simulation is often biased toward high rain rates, and many of the rates employed (in several instances exceeding 150 mm h −1 ) appear to have limited relevance to ordinary field conditions. Generally, simulations should resemble natural rain events in each study region. Attention is also drawn to the raindrop arrival rate at the surface. In natural rain, this is known to vary from < 100 m −2 s −1 to > 5000 m −2 s −1 . Arrival rate may need to be added to the list of parameters that must be reproduced realistically in rainfall simulation studies. Copyright © 2008 John Wiley & Sons, Ltd.