Daily rainfall data are readily available, but in many locations there are few rain gauges that provide subdaily rainfall data. Hydrologic models often deal with this problem by generating synthetic hourly rainfall series by assuming a temporal scaling behavior. However, Beuchat et al. found that the particular scaling laws may apply only to specific climates, and it is difficult to produce regional scaling laws. Therefore, they propose a new approach to estimating subdaily rainfall statistics using a mathematical technique called multivariate adaptive regression splines (MARS).