STATISTICAL MODELING OF THE VARIATION OF SOME CLIMATIC VARIABLES FOR WEST AFRICA: TOWARDS FUTURE PREDICTION

Abstract

Globally, the obvious impact of climate change on human and ecosystem is of great concern. For agricultural productivity and sustainability, Africa especially the West Africa sub-region depend on apposite climate. However, the recent change in climate poses a challenge to the sub-region. Limited technical, institutional and financial supports in dealing with the impact of climate change also worsen the situation in this region. This work models the variation of some climatic variables in West Africa using Kalman Filter. Eight variables; maximum temperature, minimum temperature, rainfall, land use, population, solar radiation, relative humidity and wind speed were considered. Climatic data for West Africa spanning 19752015 were used. To model the impact/relationship, Kalman Filter, a recursive estimator that produces the optimal estimate of a dynamic system, was used. Control variables (population and land use) and transfer functions were introduced to control the dynamism of the model in MATLAB environment. Results show that there is a drastic variation in climate in the region within the period of the data. It also establishes the fact that population and land use pattern, especially loss of vegetation, had significant impact on the climatic variability in the West African climate system. This can be mitigated when institutional frameworks are effectuated in order to minimize the rapid change in population and land use pattern in West Africa.