Bayesian Belief Network models as trade-off tools of ecosystem services in the Guayas River Basin in Ecuador

Abstract

Food production often leads to environmental degradation. Consequently, insights into ecosystem functioning in relation to exploitation are needed as a basis for socioeconomically acceptable mitigation of these impacts. A Bayesian Belief Network (BBN) model is developed to link three major ecosystem services (ES), i.e. food production, water provision and ecotourism, and determine the effect of local agricultural practices and management on the ES in the Guayas Basin (Ecuador). Several data sources were integrated into the BBN model, including processed spatial data from primary and secondary sources, sampling and survey data, and expert knowledge. The model suggests that banana and sugar cane generate the highest yield but provide low ecotourism benefits. In contrast, cacao produces the lowest yields but contributes to better water quality. Scenario analyses suggest that environmental gains are possible by optimising the land use (LU) based on the edaphoclimatic requirements of crops. Moreover, the integration of LU optimisation with upscaling and farming intensification can allow for additional advantages in water provision and ecotourism while mitigating productivity losses. The BBN development approach can serve as a reference for other case studies, where data scarcity plays a limiting factor in the assessment of interactions between key ES.