Rapid adaptation to climatic and environmental changes is crucial for critically-stressed regions worldwide. Large communities in arid areas, such as the Chihuahua Desert on the USA-Mexico border, are severely water-stressed. Nevertheless, these areas are becoming increasingly populated, as well as experiencing land use and land cover changes (LULCC), including urban expansion and agricultural intensification. Acute and chronic events attributed to hydroclimatic changes (e.g., prolonged droughts and extreme heatwaves) can push these communities beyond their resilience, making recovery difficult. To understand the sustainability of these LULCCs, comprehensive data with adequate spatiotemporal coverage and appropriate modeling tools are needed, connecting social and natural systems. Another confounding factor for the international transboundary region is the drastic difference in the availability and reliability of data along with governance frameworks. Remotely sensed data and convergent models can be crucial to mitigating this transboundary data inequity. Modeling methods that combine the stakeholder institutional knowledge of LULCC with observations can further enhance the use and impact of remote sensing to aid decision-making and design a sustainable policy. This study proposes methods and models that integrate and interpret both social and natural systems data to aid decision-making and design sustainable policies. The datasets created by matching temporal and spatial locations with LULCC events over the last decade are used to learn a quantitative Bayesian Network. Stakeholders’ opinions can be incorporated to improve the model. The Bayesian Network is a graphical model with probabilistic relations, which allows analysis of ‘what-if’ scenarios. It is also explainable and allows for incorporating stakeholders’ knowledge. The developed model can be used with future climate and market predictions to project future LULCCs, creating an operational decision-support tool.