Spatiotemporal Evolution of CO2 Emissions from Fossil Fuel Combustion, Cement Production, and Gas Flaring at the County Scale in the Beijing–Tianjin–Hebei Region Based on Landscape Pattern Metrics, 2000–2020

How landscape configuration relates to carbon emissions at the county scale, including spatial spillovers and local heterogeneity, remains insufficiently resolved. Here, we analyze 199 county-level units in the Beijing–Tianjin–Hebei (BTH) region using data from 2000, 2005, 2010, 2015, and 2020. We combine landscape pattern metrics derived from 30 m land-use data with ODIAC carbon emissions and socioeconomic indicators. Spatial autocorrelation analysis, two-way fixed-effects ordinary least squares, the Spatial Durbin Model, and Geographically Weighted Regression are used to quantify average associations, spatial spillovers, and local variation. County-scale carbon emissions increased continuously from 2000 to 2020, although growth slowed after 2010, and high-emission areas remained concentrated in urban cores. Emissions showed significant positive spatial autocorrelation, with stable High–High clusters after 2010. Landscape configuration was associated with carbon emission density. Patch Density generally showed negative associations, whereas the Contagion Index showed positive associations in the global and spatial models. The Largest Patch Index and Patch Cohesion Index also exhibited spillover effects, while GWR revealed substantial local variation in the direction and magnitude of these relationships. These findings demonstrate that landscape patterns are linked to county-scale carbon emissions through both local and cross-county pathways, highlighting the need for spatially differentiated low-carbon governance and territorial spatial planning.