Spatial heterogeneity and driving mechanisms of carbon storage in the urban agglomeration within complex terrain: Multi-scale analyses under localized SSP-RCP narratives.

• Localized socioeconomic and natural narratives are applied on the regional scale. • Propose ecological zoning management in mountainous agglomerations. • POP, GDP, DEM, and Slope have an inverted U-shaped relationship with carbon storage. • Urbanization expands into the western mountainous under the localized SSP5–8.5 narrative. Ecological carbon storage is critical to climate change mitigation by absorbing CO 2 from the atmosphere. Mountainous ecosystems, crucial carbon reservoirs, are experiencing persistent carbon storage declines, yet management mechanisms targeting such complex terrains remain challenging. Focusing on the Sichuan-Chongqing urban agglomeration, interwoven as mountains and basins, we employed the MPI-coupled (Improved Markov-PLUS-InVEST) model and localized "socio-economic-climatic" narratives to simulate the dynamic evolution of carbon storage over extended periods. Diverse impacts and gradient effects of social, economic, and natural (incl. terrain, climate) drivers on carbon storage are further examined. The findings revealed that: (1) The most carbon storage growth projected from 2030 to 2060 reaches 1.53 %, but of which localized SSP5–8.5 narrative the urbanization encroached upon the undeveloped mountainous, severely impairing carbon sequestration by 81.5Tg. (2) In addition to landscape types, carbon storage exhibits complex spatial responses to economic development, population size, and terrain with an inverted U-shaped relationship. The proposed methodology offers more accurate carbon storage estimates and assists policy-makers in similar terrains to effectively manage the spatial heterogeneity of carbon storage. This study facilitates the implementation of tailored ecological conservation through urban zoning, thus contributing to enhanced subregional climate change mitigation efforts globally. [ABSTRACT FROM AUTHOR]