Development of rainfall disaggregation model in the Awash River Basin, Ethiopia

Abstract

This study aims at developing a model that can generate synthetic hourly rainfall data from the existing daily rainfall data of Awash river basin. Fifteen minutes rainfall data collected from national meteorological agency for 13 active stations and daily data collected from 54 stations were considered. Stochastic rainfall disaggregation and Hyetos temporal precipitation model was tested using the available fifteen minutes data. Three regions with close climate condition and rainfall pattern were identified and tested to be homogeneous in the stochastic method. Both methods are tested by using statistical comparison of variance, skew-ness, probability of dry period, and Lag-1 ACF. The result of the stochastic method showed very good performance in preserving the probability of zero rainfall and the daily rainfall total. But it has limitation in dis-aggregating rainfall magnitudes with high return period. Statistical comparison of Hyetos model indicated very good agreement with the original data. Especially the daily total statistical properties were well preserved. The comparison of the two methods showed that Hyetos is better in preserving the statistical property. Generally the methods are capable in preserving statistical properties and the daily total rainfall depth. Therefore, Hyetos model is pertinent for only temporal disaggregation, whereas the stochastic method is applicable for both spatial and temporal disaggregation in the basin.