A Methodology for Estimating Spatial Non-stationarity in Ecosystem Service Valuation
Austin Troy
A hedonic analysis of housing transactions in West Baltimore was conducted to determine whether there is spatial variability in willingess to pay for proximity to trees and parks, using geographically weighted regression (GWR). GWR is a regression method that allows parameter estimates and test statistics to vary continuously over space. By enabling statistical models to vary over space, we can visualize spatial patterns in complex socio-economic relationships and use this information to help delineate functional socio-economic “patches.” We find that valuations of tree cover and proximity to parks is spatially non-stationary within the study area.
Keywords: patch analysis, social, economic, project, Baltimore

[ Back ]