Building the resilient food waste supply chain for the megacity: Based on the Multi-scale Progressive Fusion framework

Abstract

Food waste (FW), a significant component of municipal solid waste (MSW), exhibits notable spatiotemporal fluctuations and environmental impacts, especially in megacities. To enable efficient FW recycling, we developed a resilient supply chain prediction model by leveraging the Multi-scale Progressive Fusion (MPF) framework. The framework integrates models for annual MSW prediction, monthly fluctuations, FW separation rates, and spatial downscaling, enabling it to accurately capture spatiotemporal patterns at a monthly resolution and 1 km² scale (R² = 0.8130)., significantly outperforming the Baseline framework (R² = 0.1383). Our analysis predicted that FW in Beijing would exhibit seasonal variations by 2035, with daily FW during the peak season (July and August) being 36 % higher than in the off-season (February). These findings supported seasonal resilient strategies for addressing spatiotemporal fluctuations in FW. Scenario analysis demonstrated that the MPF framework, when combined with resilient measures, significantly improved supply chain resilience. Compared to non-resilient supply chains reliant on baseline FW predictions, it enhanced traffic adaptation by 5.6 %, reduced costs by 16.9 %, and lowered greenhouse gas (GHG) emissions by 40.1 %. This study demonstrated the critical role of accurate FW spatiotemporal forecasting in enhancing resilience within megacities and provided a practical pathway for building resilient FW supply chains globally.