ABSTRACT

The data used in this project was collected from the meterological unit of the National Root Crop Research Institute (N.R.C.R.I) Umudike. It covers rainfall data (mm) in Umudike from (1989-1998). The ultimate objective of this study is to construct a mathematicL model which could be used to predict reliable and dependable future rainfall values. Time series analysis was used to estimate and Isolate the time series components existing in the original rainfall data and the transformed data. The time series model employed in this analysis is the multiplicative model. Forecast for 1999 rainfall was made and the error term was calculated for the original and the transformed data. The model based on original data appeared to have per- formed better than that based on the transformed data in the forecasting. Farmers are advised not to start planting in the months of January and February based on the fact that the analysis in the past ten years show insufficient amount of rainfall in these two months mentioned. It would be wise if they start planting in the month of March which is not distant from rainy season (April - September) to enhance the production.