Study of the Geographically Weighted Regression Application on Climate Data

Abstract

This study used Geographical Weighted Regression (GWR) technique to find spatial relationship between Elevation and climate (Rainfall, Temperature) in Northern Nigeria using climate (Rainfall, Temperature) data from weather stations from 1980 – 2010 obtained from Nigerian Meteorological Agency (Nimet). From the results of the analysis it was shown that there is significant relationship between the elevation and climate variables (Rainfall, Tmax and Tmin). The study also shows that GWR has smaller residual sum of square than OLS in analysing the relationship between Elevation and Climate data. This may be due to the consideration of the spatial variation of the relationship over the study region. When mapping the results of GWR model it was observed that the effect of Elevation on climate variables appears to vary geographically

Keyword: Geographical Weighted Regression (GWR), Ordinary Least square (OLS),