Modelling daily precipitation features in the Volta Basin of West Africa



The combination of a conventional Markov chain model (zero and first order) and a gamma distribution model are found to be applicable to derive meaningful agricultural features from precipitation in the Volta Basin (West Africa). Since the analysis of the monthly or annual precipitation amount does not provide any adequate information on rainfall timing and sufficiency of crop water requirement, rainfall modelling was performed on a daily time scale for 29 rainfall stations. The modelled rainfall features follow distinct spatial patterns, which will be presented as maps of(1) rainfall occurrence probabilities and (2) recommendations of optimal planting dates. In addition, the effective drought index (EDI) working on daily time scales is calculated in order to assess drought properties of five different rainfall regions within the Volta Basin. Apart from the common way of separately modelling the duration and intensity due to their different distributions, a copula approach is chosen in this study to construct a bivariate drought distribution. Application of the measures derived to agricultural decision support will be discussed briefly. Copyright © 2009 Royal Meteorological Society