Enhancing ecosystem services maps combining field and environmental data

Abstract

Ecosystem service maps are increasingly being used to prioritize management and conservation decisions. Most of these maps rely on estimates of ecosystem services estimated for individual land cover classes rather than incorporating field data. We developed combined field models (CFM) using regression analysis to estimate ecosystem services based on the observed relationship between environmental and land cover data and field measurements of ecosystem services. Local ecosystem service supply was estimated from vegetation data measured at fifty sites covering the widest range of environmental conditions across a watershed in Mexico. We compared the accuracy of the CFM approach for forage, timber, firewood and carbon storage over a more commonly “look up table” method relying on a uniform estimate of ecosystem service supply by land cover type. The CFM revealed higher accuracy when compared to the “look up table” approach. The resulting CFM models explained a large fraction of the variance (42–89%) using a combination of land cover, remote sensing data, hydrology and distance from developed areas. In addition, mapping residuals from Geographically Weighted Regressions provided an estimate of uncertainty across the CFM model results. This approach provides better estimates of ecosystem service delivery and uncertainty for land managers and decision-makers.