Integrating local perceptions and Bayesian belief networks into model ecosystem services multifunctionality in the Colombian Andes

Access to water in rural watersheds is affected by land-use changes and limited integration of governance factors into ecosystem service (ES) analysis. This study implemented a participatory spatial Bayesian belief network (BBN) for the practical application of ES multifunctionality in Rural water supply systems (RWSS) to enhance their role in watershed planning and governance. The framework applies indicators of ES supply and demand (derived from the InVEST and ARIES models), socioeconomic data (demographic and multidimensional poverty), and local perceptions collected through interviews. Two BBN models were developed: Model 1 used quantitative biophysical and socioeconomic indicators. Model 2: Incorporating governance and perception variables from a local community perspective. Probabilistic spatial inference and uncertainty analysis were implemented to map multifunctionality and to identify priority zones for governance. The results showed that incorporating local perceptions increased the probability of highly multifunctional areas by approximately 10%, particularly in the upper sub-watersheds characterized by strong local participation. The uncertainty analysis identified transitional agricultural–urban zones as areas that require additional data collection and participatory validation. The participatory BBN framework successfully integrated socio-ecological and governance dimensions, facilitating the identification of ES multifunctionality zones that align ecological integrity with community priorities. A spatial BBN is a flexible methodology for data-scarce Andean regions that supports watershed management plans. This approach contributes to a method that links stakeholder perceptions to spatial decision-making and benefits achievement to Sustainable Development Goals (SDG), SDG 6 (Clean Water and Sanitation), and SDG 15 (Life on Land) by advancing adaptive, inclusive, and evidence-based watershed planning.