Bayesian belief networks to analyse trade-offs among ecosystem services at the regional scale

Knowledge on trade-offs and synergies among ecosystem services is crucial for the design of land use strategies that optimize ecosystem service delivery. Correlation coefficients, obtained through pairwise comparison of ecosystem service provision maps, have been put forward as suitable indicators to quantify these interactions. However, for more in depth analyses of trade-offs and synergies where driving forces of interactions need to be determined, more sophisticated methods are needed. Although Bayesian belief networks have been frequently mentioned as promising tools to investigate interactions among ecosystem services, up till now, no structured approaches to do so have been suggested. This paper presents a way to analyse trade-offs and synergies among ecosystem services together with their driving forces. Joint probability distributions of ecosystem service pairs, which can be calculated by using Bayesian belief network models, are used to quantify interactions. The paper demonstrates the approach by quantifying trade-offs and synergies among several ecosystem services in Flanders, Belgium. Our analysis identifies two bundles of ecosystem services which react synergistically. Wood production and several regulating services on the one hand and food production and soil formation on the other hand. Trade-offs are identified among food production and most of the other services that were included into the analysis. In addition to these general findings, the analysis shows that the identified interactions may change depending on the considered environmental conditions, specified through soil type, land cover and land use.