A B S T R A C T

The aim of this study is to construct a mathematical model for the temperature of Umudike which could be used to predict reliable and dependable future temperature values. In particular, it was intended to estimate and isolate the components of time series: Trend, Seasonal effect, cyclical effect and irregular effect present in the data using the descriptive Lime series approach. This estimation and isolation was done bsed on the models of the original and logerithmthally transformed data. The study, which was carried out in Umudike, covered the period 1987 to 1.96. Data were collected from the National Root Crops Research Institute Umudike (NJRCRI). The results of the analysis showed amongst others that (a) All the conponnts of time series were present in the data; (b) On the averae the temperature of Umudike within the period under consideration is 26.7°c with a standard deviation of 1.2°C. Arising from the results, it was suggested that farmers lanted those crops that strive well under high temperature during the period February to April and those that strive well under moderate temperature be planted during the period July to September. It was concluded that, within the limits of analytical error, there would be reduced crop failures if farmers and others can adhere strictly to the advice provided based on the forecast values of 1997."