Gap Filling Daily Rainfall Data using Quantile Method

Abstract

Filling the gap of daily rainfall data is an important step to obtain a complete data set before further study that related to water resources. Generally, methods that use to fill the gap are arithmetic mean and inverse distant weighting (IDW). Some limitations of these traditional methods are underestimation of daily rainfall and overestimation of the number of rainy days, for example. Thus, this research attempts to study a gap filling daily rainfall data base upon statistical approach, quantile method (QT). By this method, a mixed probability distribution of Bernoulli-Gamma function is used to derive the target and source stations daily rainfall distributions of six selected rain stations located in upper Ping River basin. This study has tested such method to fill the missing daily rainfall at 20% and 40% percent missing. The study result reveals that QT generally yields a statistical value such as mean, maximum, variance, and 99% percentile better than IDW, but QT yields greater RMSE and MAE because QT produce higher in variance.