| The identification of primary climate controls and their utilization in developing statistical models for precipitation forecast in northern Nigeria is the main goal of this research. The input consists of time series of seven ocean-atmospheric variables with 34 years of data, namely, north and south Atlantic sea surface temperature, sea level pressure, the north Atlantic oscillation, the southern oscillation index, Nigerian temperature and historic rainfall. Nineteen synoptic stations in the study area are selected and grouped into three sub-regions using cluster analysis. Fluctuation in rainfall within each sub-region is determined and the resultant trend modeled by least square multiple regression. Using Pearson's product moment correlation, the identified climate forcing agents are analyzed for association with rainfall in each of the three sub-regions. The correlation coefficients at different rainfall lags are tested for significance. Results indicate that different combinations of climate agents force rainfall in each sub-region, with some agents, such as sea level pressure and sea surface temperature, common to all sub-regions. In particular, the research identified south Atlantic sea surface temperature domains common to all sub-regions for the period 1961 to 1990 as 21s/15w, 21s/1w and 21s/3w.; By employing identified climate agents in rainfall forecast for the sub-regions, seasonal and monthly models are derived, comparison of which, in time and space, indicate that the standardized seasonal model yields the least residual or error component. The standardized seasonal model successfully improved forecast in two out of three regions, and in a test forecast, the version that utilized both NDJ values of the explanatory variables and last year's rainfall. |
