How explainable AI can optimize supply-demand networks for sustainable cities

The use of explainable AI allows for transparent identification of critical drivers, offering policymakers actionable evidence free from the subjectivity that often complicates sustainability planning. By integrating artificial intelligence with ecological modeling, the approach demonstrates how emerging technologies can move beyond prediction to provide clear guidance on thresholds, trade-offs, and interventions.

A team of Chinese researchers has unveiled a new artificial intelligence framework designed to tackle the mounting resource imbalances that threaten the sustainability of one of the country’s fastest-growing urban clusters. The study, published in Land, focuses on aligning water, energy, and food systems within supply–demand networks to secure ecological stability in the Bohai Rim urban agglomeration.

The research, titled Explainable AI-Driven Integration of Water–Energy–Food Nexus into Supply–Demand Networks, introduces an explainable artificial intelligence (XAI) approach that integrates ecological modeling and supply–demand analysis, offering policymakers a practical roadmap for addressing sustainability challenges in densely populated and industrialized regions.

Identifying sources, corridors and critical zones

The study applies an innovative XGBoost-SHAP model combined with the Zonation 5 spatial prioritization tool and circuit theory to build interconnected networks of water, energy, and food supply and demand. By running this hybrid framework across the Bohai Rim region, the researchers identified 114 stable supply sources and 128 deficit sources. These findings were linked by 472 efficient ecological corridors and 296 general corridors, forming the backbone of a region-wide supply–demand network.

The team also defined six major supply potential zones, each demonstrating specialization in water provision, food production, energy security, or composite roles that combine multiple functions. These zones form critical anchors for resource management, acting as both supply hubs and resilience buffers for neighboring urban centers. The networked approach ensures that surplus resources in one area can be redirected to address shortages in another, reducing systemic vulnerabilities that often surface during droughts, energy shortages, or disruptions in food supply chains.

Understanding the drivers of resource imbalances

Beyond mapping the physical flow of resources, the study sought to uncover the underlying drivers of imbalance. Using the explainable AI approach, the researchers identified five dominant factors shaping the water–energy–food nexus in the Bohai Rim. These included annual precipitation, vegetation coverage, human activity intensity, cropland distribution, and temperature.

Precipitation emerged as the most influential determinant of supply stability, highlighting the region’s acute vulnerability to water scarcity. Vegetation coverage also played a pivotal role, directly influencing ecological resilience and the capacity to support food and water systems. In contrast, human activity intensity proved to be a destabilizing factor, exerting negative interactions with precipitation and vegetation. Cropland distribution further amplified pressures when expansion exceeded ecological thresholds, while temperature trends, particularly linked to urban heat island effects, added to the complexity of the nexus.

The model revealed hidden thresholds within these factors. For instance, vegetation coverage above 92 percent was identified as a tipping point for ecological recovery, while cropland expansion beyond 68 percent began to strain system balance. Similarly, average temperatures around 13 degrees Celsius were flagged as critical for avoiding the escalation of heat-related resource stress. These insights underline the non-linear and interdependent nature of urban sustainability challenges.

Proposing strategies for sustainable urban development

Armed with these insights, the researchers outlined a suite of optimization strategies tailored to the Bohai Rim’s unique conditions. Among them was a call for large-scale ecological restoration, including targeted vegetation recovery programs to surpass the 92 percent resilience threshold. The study also recommended tighter regulation of cropland expansion, ensuring farmland use remains within ecologically safe limits.

Urban planning featured prominently, with proposals to mitigate heat island effects through climate-sensitive design and vegetation buffers. The creation of buffer forests along ecological corridors, 100 meters for efficient corridors and 50 meters for general corridors, was proposed as a means of reducing ecological resistance while reinforcing connectivity across supply and demand zones.

At the zone-specific level, the authors call for tailored interventions such as sponge city construction to improve water retention, smart irrigation technologies to minimize waste, and the promotion of renewable energy sources to relieve pressure on fossil-fuel-intensive grids. Collectively, these strategies aim to balance urban growth with ecological resilience, ensuring the region’s long-term sustainability.

Broader implications and global relevance

While focused on the Bohai Rim, the framework developed in this study carries broader significance. The use of explainable AI allows for transparent identification of critical drivers, offering policymakers actionable evidence free from the subjectivity that often complicates sustainability planning. By integrating artificial intelligence with ecological modeling, the approach demonstrates how emerging technologies can move beyond prediction to provide clear guidance on thresholds, trade-offs, and interventions.

Rapid urbanization, while economically advantageous, places immense pressure on natural resources unless managed within a scientifically informed framework. The Bohai Rim, characterized by its industrial intensity, food production role, and water scarcity, mirrors challenges faced by other global urban clusters where resources are stretched thin by population growth and economic expansion.