Multi-Scale Drivers of Food-Water-Energy Related Ecosystem Services in the Context of Land Degradation: Insights from the Beijing-Tianjin-Hebei Agglomeration

Land degradation poses significant challenges to maintaining food–water–energy (FWE) security, particularly in rapidly urbanizing regions. The ecosystem services (ES) related to FWE are critical for sustainable regional development. Yet, rapid urbanization in the Beijing–Tianjin–Hebei region has altered land use and ecological patterns, intensifying degradation risks and complicating the maintenance of FWE-related ES (FWE-ES) across spatial scales. However, few studies have systematically examined how degradation-sensitive FWE-ES respond to land use change across multiple spatial scales, or how scale-dependent drivers shape trade-offs and synergies. To address this gap, this study investigates the key drivers of FWE-ES changes from 2000 to 2020 at watershed, county, and built-up area scales, while analyzing spatial trade-offs and synergies among multiple ES to guide scale-specific land degradation mitigation and restoration strategies. Focusing on the Beijing–Tianjin–Hebei urban agglomeration, we assessed five ES indicators—habitat quality (HQ), carbon sequestration (CS), soil retention (SDR), water yield (WY), and food production (FP)—using the MGWR model and geographic detectors to examine spatial relationships and driving mechanisms. Results indicate that WY and FP increased over the study period, but HQ and CS declined, reflecting biodiversity loss and carbon storage depletion. SDR improved at broader scales (indicating partial restoration potential) but deteriorated in built-up areas, signaling localized erosion and soil degradation risks. Key drivers included patch density (PD) and Contagion for HQ, DEM for CS, slope for SDR, Precipitation for WY, and NDVI and slope for FP. Synergies were observed between HQ and CS, whereas WY and CS exhibited trade-offs. The findings highlight the scale-dependent nature of FWE-ES dynamics and their links to land degradation processes, with watershed-scale patterns explaining most variations in HQ and SDR. At the same time, county-scale factors dominated FP and CS. These insights provide an empirical basis for integrating multi-scale ES management into targeted land degradation prevention and restoration policies.