Analysis of the relationship between NDVI and climate variables in minnesota using geographically weighted regression and spatial interpolation

In order to better understand the effects of climate change on ecosystems, the relationship between Normalized Difference Vegetation Index (NDVI) and atmospheric constituents have been explored widely by scientists using the global technique of Ordinary Least Squared (OLS) regression analysis. However, recent studies exploring such relationships at different spatial scales have revealed that local statistical approaches are more appropriate when the assumption of spatial stationarity is invalid. This study aims to explore the relationships between NDVI and the local level atmospheric constituents consisting of precipitation and temperature in the state of Minnesota from 1990 to 1997 using Geographically Weighted Regression (GWR) and spatial interpolation techniques. The analysis focuses on the summer months, when such relationships are more apparent in northern mid-latitude regions. In comparison to traditional OLS, there is a substantial improvement in the analysis using GWR with the average r2 value improved from 0.24 to 0.67. The overall relationship between the different atmospheric constituents and NDVI were broadly consistent with the different types of land uses across the state with the highest correlation located in forested areas. The spatial patterns of the association between different climatic variables and NDVI in the form of regression coefficients were not very consistent over the years as result of inter-annual variations in the local climate.