ABSTRACT

Rainfall temperature and Relative humidity data for 6 years (2004-2009) For Umudike in Umuahia Abia State were used in the Analysis in order to detect the recent trend in the components (Rainfall, temperature and relative humidity). The main objective of this study is to construct time series models which could be to predict reliable dependable estimate of future occurrence of this three component and also to dictate and compare the relationship that exist among these three components. The data for this study is a secondary from agro metrology department of National Root Crops Research Institute Umudike. The data is presented in monthly basics and covers this period of January 2004 to December 2009. The box-jenkins Iterative procedure is used to outline steps for model identification, model fitting and the diagnostic checking. The original data showed seasonal variation from the time plot ACF and this time plot of ACF of and PACF of the values of the three components. There is a downward trend for rainfall which is insignificant,, There is also an upward trend for temperature which is also insignificant and a straight horizontal trend for relative humidity which is insignificant. This lead to differencing of the series to make it stationary. The ACF and PACF of the first non seasonal and seasonal differencing suggested ARIMA model (l1 l)(l 10)xl2xl for temperature data, ARIMA (OO1)(111)x12x1. For rainfall and ARIMA (101)(1 l l)x12xl For Relative humidity data which were subsequently fitted after the model was found adequate from residual analysis. At the end relationship's were dictated and compared for the three components and the data for the three components for the months of the year 2009 were predicted using the identified model.