Urban sprawl constantly reconfigures the land use pattern, and such transformations may significantly modify regional carbon stocks. Utilizing Tai’an City as the study site, this research established a comprehensive integrated Patch-generating Land Use Simulation (PLUS), Integrated Valuation of Ecosystem Services and Trade-offs (InVEST), and Optimal Parameters-based Geographical Detector (OPGD) system to reconstruct carbon storage shifts from 2000 to 2020, project its reaction to four diverse development trajectories in 2030, and investigate the drivers underlying spatial disparities. The results indicate a persistent decline in carbon storage throughout the past two decades, with peak concentrations primarily gathered in mountain regions dominated by forest and grassland, whereas lesser amounts were grouped in urban and suburban areas defined by built-up land. Compared to 2020, the projected carbon stock in 2030 drops by 1,803,966 t under the natural growth trajectory and by 2,417,778 t under the high-quality economic growth pathway, whereas it rises by 47,326 t under cultivated land conservation and by 7679 t under ecological conservation. Elevation represents the most crucial driver among the selected variables in clarifying the spatial fluctuation of carbon storage (q = 0.3985), followed by slope (0.3323), mean annual temperature (0.2382), and the Normalized Difference Vegetation Index (NDVI) (0.1219). The synergy between elevation and NDVI produces the highest integrated explanatory power (q = 0.4906). These outcomes imply that constraining construction land growth while protecting agricultural and ecological land is vital for preserving and enhancing regional carbon sink potential.

Under the Yellow River Basin’s strategic imperative of ecological protection and high-quality development, the critical question is whether rapid urbanization’s population agglomeration and land expansion are genuinely aligned with-or increasingly decoupled from-underlying economic vitality. To address this, we construct a unified analytical framework that synchronously captures coupling coordination, decoupling trajectories, and land-use efficiency, operationalized through multi-source county-level panel data for the Jinan Metropolitan Area in lower Yellow River Basin across 2000, 2010, and 2020. Unlike conventional two-dimensional assessments, the framework distinguishes system coordination, evolutionary direction, and input–output performance while using harmonized nighttime light data to capture the spatial intensity of economic vitality. The results show that: (1) the mean county-level coordination grade increased from 4.76 to 5.32, although strong core–periphery disparities persisted; (2) the decoupling structure shifted from relative homogeneity to pronounced differentiation, and strongly decoupled counties accounted for 48% of all units by 2020; and (3) average efficiency rose from 0.25 to 0.39, but most counties remained below the efficiency frontier, producing a pattern of central agglomeration and peripheral dispersion. Cross-classification further identifies coordinated but inefficient, efficient but dysfunctional, and low-coordination–low-efficiency trajectories that cannot be detected using a single model. The framework therefore provides a transferable diagnostic tool for differentiated territorial planning in metropolitan regions undergoing rapid structural transformation.