A Nonparametric Renewal Model for Modeling Daily Precipitation
Stochastic and Statistical Methods in Hydrology and Environmental Engineering: An International Conference in Honour of Professor T. E. Unny
University of Waterloo, Waterloo, Ontario, Canada
A nonparametric wet/dry spell model is developed for describing daily precipitation at a site. The model considers alternating sequences of wet and dry days in a given season of the year. All the probability densities of interest are estimated nonparametrically using kernel probability density estimators. The model is data adaptive, and yields stochastic realizations of daily precipitation sequences for different seasons at a site. Applications of the model to data from rain gauges in Utah indicate good performance of the model.
Lall, Upmanu L.; Rajagopalan, B.; and Tarboton, David G., "A Nonparametric Renewal Model for Modeling Daily Precipitation" (1993). Civil and Environmental Engineering Faculty Publications. Paper 2597.