An Artificial Intelligence–Based Index for Climate Resilience Assessment of Community Aqueducts in Tropical Andean Watersheds: Application in Colombia

Abstract Community-managed water supply systems, or community aqueducts, are complex socioecological systems that are highly vulnerable to climate variability and change. Assessing climate resilience is therefore essential to inform decision-making processes aimed at strengthening socioenvironmental sustainability. This study developed and applied an artificial intelligence (AI)-based index, termed the climate resilience index for community aqueducts in tropical Andean watersheds (CRI-CATAW), to assess the climate resilience (CR) of these systems. The CRI-CATAW is grounded in three climate resilience capacities (absorption, adaptation, and transformation) and integrates four cross-cutting dimensions: environmental, sociocultural, administrative management, and structural–operational. The methodological framework comprised three stages: (1) indicator selection using the Delphi method, (2) indicator operationalization through fuzzy logic, and (3) application of the index to 13 community aqueducts within the Bolo River watershed, Valle del Cauca, Colombia. The Delphi process resulted in the definition of 28 qualitative and quantitative indicators that were integrated using an AI-based fuzzy logic model. The index incorporated 486 decision rules, modeled with trapezoidal, triangular, and singleton fuzzy membership functions. Climate resilience was estimated by processing inference rules in MATLAB, yielding a classification into three levels: low, medium, and high. Application of the CRI-CATAW to 13 aqueducts in the Bolo River basin indicated that 69.2% exhibited low levels of climate resilience, with significant implications for their socioenvironmental sustainability, and produced results consistent with observed local socioenvironmental conditions. CRI-CATAW constitutes a novel, comprehensive, generic, and replicable AI-based tool for assessing the climate resilience of community-managed water supply systems. Further applications across diverse socioenvironmental and cultural contexts are required to strengthen the validation of indicators, membership functions, and decision rules.