Climatological probability forecasting of rainfall in Kenya

Abstract

Climate variability has significant impact on performance of the economy of a country especially in a developing country like Kenya. Certain sectors of the economy like agriculture, manufacturing, transport and communication are vulnerable to unanticipated weather changes. Thus a good climate and weather forecasting models with high degree of certainty would playa pivotal role in future planning and policy formulation of any economy of a developing country. Climatological weather forecast in Kenya is used to classify amount of rainfall expected in each of the homogenous climatic zones into three categories: Below Normal, Near Normal and Above Normal using terciles. These terciles are practical (actual) values obtained by arranging the rainfall data from a given base period to a specific year into ascending order anddividing the data into three equal parts. This mode of classifying rainfall data has been proved to have inconsistencies and in- homogeneities. This study intends to improve the classifications of rainfall into the above stated categories by fitting a theoretical probability distribution to the rainfall data and . determining theoretical values of the terciles using distribution function fitted. Probability forecasts offer several benefits. They contain more information than weather forecasts and the uncertainty in the forecast is specifically expressed, thus the user is madeaware of that uncertainty and can use this information in decision-making. Probability forecasts can be used with thresholds to make decisions, where the thresholds canvary from user to user and purpose to purpose. Availability of probability forecasts would allow users to make decisions based on a quantitative uncertainties and his /her threshold for making the decision. Themainobjectives of the study is to determine the distribution ofterciles which are usedtoclassify the amount of rainfall received in aparticular region into the following thethree categories: Below normal, Near normal and Above normal in order to improve Climatological forecasting in Kenya. Specificobjectives of the study are: 1. To fit the theoretical probability distribution to the cumulative rainfall data for a season for a given weather station in the country. 2. To find the estimates of the probability distribution. 3. To obtain the theoretical (actual) distribution of the tercilesTI andT2,that is F (TI, T2) arid find out how this information can be used to improve on the forecast classification-rainfall data.