Methods for the statistical evaluation of African precipitation

African livelihoods are extremely closely tied to the spatial and temporal distribution of rainfall for a variety of reasons. Due to a lack of spatial or temporally complete records, the understanding of African climatologies, the implications of observed rainfall amounts, and the significance of forecasts is largely unknown. This research attempts to improve the understanding of historical, current, and future rainfall based on fitting probability distributions to modeled historical data.; Using modeled historical rainfall data as the climatic record it is possible to estimate gamma distribution parameters, along with a mixture coefficient representing the probability of no rain, to define a probability distribution function for a particular location and time interval. Goodness-of-fit tests show that the gamma distribution is adequate for the modeled data at a monthly time scale. Gamma parameters for all 12 months are classified resulting in a map of rainfall regimes.; The Standardized Precipitation Index (SPI) is a spatial and temporally invariant measure of observed rainfall. There are complications involved with implementing the SPI at a continental level, and methods for overcoming these obstacles are discussed. While the SPI does a good job of reporting the intensity of rainfall during an interval, it does not calculate the persistence of intensity. The Intense SPI Event Accumulation (ASEA) is proposed as a potential method to capture this persistence. Characteristics of SPI and ASEA data for Africa are presented and evaluated.; The Forecast Interpretation Tool (FIT) represents and attempt to blend climatological distribution parameters with probability forecasts. Through Monte Carlo resampling, new distribution parameters can be estimated, which reflect the forecasts given the historical data. This allows for powerful analysis of rainfall scenarios for an upcoming growing season, which can prove valuable in forecasting food insecurity. Analysis of the FIT output evaluate this technique as a method for implementing the forecasts in a quantitative manner.; This dissertation presents a coherent method for evaluating past, present and forecasted rainfall amounts. It develops new techniques that show promise in the analysis of precipitation at a multi-country or continental scale. The methods presented could be used to benefit millions of lives in Africa by reducing the impacts of food insecurity.