Locating spatial variation in the association between road network and forest biomass carbon accumulation

•Kernel density estimation was applied to evaluate density of road network.•Used GWR to observe the spatial association between road network and biomass C.•The association varied significantly across the study area.•The tradeoff and synergistic relationships existed simultaneously.•Differentiated measures can be implemented according to the GWR outcome. The effects of road network on disturbance regimes of forests are less investigated in biomass carbon accumulation than in landscape pattern. To fills in this knowledge gap, we sought to explore the spatial variations in the relationships between road network and biomass carbon accumulation, taking the upstream area of the Minjiang River, Fujian Province in China, as a case. Our results showed that the biomass carbon density increased gradually as the increase of distances from road in both of the study years 2007 and 2012, with a concentration of the biomass carbon loss within the 1000–1500m distance zone during the study period. The regression outcomes indicated that the geographically weighted regression models fit better than the ordinary least squares, with all the road network measures being statistically significant at the 1% level for the biomass carbon density in both of the study years. Basing on the sign and size of the coefficients estimated by the geographically weighted regressions for each grid, we found the tradeoff and synergistic relationships between the distribution of road network and biomass carbon density existed simultaneously in the study area. Geographical clusters (i.e., hot spots), where the marginal effects of the road network indictors on biomass carbon sequestration ability varied significantly across locations, were also identified to present spatially explicit and quantitative assessments of the geographic variations in these multiple relationships. Our analysis fills the research gap, which assumes that the road’s impact holds the same everywhere within a given geographical range. Moreover, identified hot spots could facilitate the implement of more elaborate forest management policies (i.e., grid-based) to dialectically deal with the effect of road network construction.