Interactive pathways of agriculture, urbanization, and ecosystem vulnerability in U.S. carbon emissions: insights for carbon neutrality and climate governance

The complex interdependencies among agricultural systems, urbanization, and ecosystem vulnerability represent critical yet underexplored determinants of carbon emissions. To move beyond single-factor analyses, this study constructs a multidimensional framework to quantify the synergistic effects of these factors on U.S. carbon emissions from 1995 to 2022. By integrating data from The Notre Dame Global Adaptation Initiative (ND-GAIN) with national emission inventories, we adopt a multi-method analytical approach: Gaussian fitting validates the Environmental Kuznets Curve (EKC) hypothesis, through which an inflection point is identified at a per capita GDP of about $40,000, while the results also reveal the model’s limitations in capturing policy-induced nonlinearities. Partial least squares regression (PLSR) exhibits superior explanatory power with an R² value of 0.869 compared to ridge regression, and its latent component analysis further uncovers significant nonlinear interactions. Key findings show that food vulnerability and urban concentration act as significant suppressors of carbon emissions, whereas agricultural production capacity and water vulnerability function as major driving factors. Scenario simulations based on the Stochastic Impacts by Regression on Population, Affluence and Technology model further confirm the emission-reducing effect of ecological restoration, which leads to an approximate 1.29% decrease in carbon emissions, while also indicating that adaptive agriculture development may result in a short-term increase of about 3.44%. This study underscores the necessity of multi-system governance, proposes a “natural restoration-technological innovation-institutional synergy” policy framework, and provides a data-driven pathway for U.S. carbon neutrality as well as a novel paradigm for multi-scale climate governance analysis.