Application of Vector Autoregression (Var) on Modelling and Forecasting Average Monthly Rainfall and Temperature

Rainfall and temperature have become the two most natural factor that determines the standard of agricultural production. Sensitivity in climate variability over a long period of time need to be recorded, looking at difference in temporal and spatial scale. The need to understand the nature of the differences in the climate system and their impact on the society and environment is of great interest. This paper tends to apply Vector auto-regressive on modelling and forecasting average monthly rainfall and temperature in Nigeria. A monthly data sourced from World Bank climate portal, from January 1986 to December 2021. Augmented Dickey-Fuller (ADF) a test used to test for stationarity of the trends. Also, the criterion, Alkaike information criterion (AIC) is considered in the model lag selection and the VAR model favored VAR at lag 8. Ordinary least square has been used to estimate the VAR model parameter. Granger causality shows a bi-lateral causation from the temperature during rainfall and from rainfall during different temperature. “Impulse Response Functions” (IRF) and “Forecast Error Variance Decomposition” (FEVD) were further carried out as a structural analysis between the two variables, it revealed that, rainfall and temperature are interrelated.