PurposeRoad traffic has long been recognized as a considerable source of ecological disturbance, compromising the integrity of natural ecosystems from their pristine baseline state. Pollutant deposition from vehicular emissions significantly contributes to environmental contamination, while associated human activities exacerbate ecosystem disruptions. In designated conservation zones, including nature reserves, land-use practices such as pastoralism and agriculture are subject to regulation. Nevertheless, the influence of road traffic on ecosystem structure and processes, particularly concerning carbon (C) sequestration and its spatial heterogeneity, warrants further research. This consideration is critically relevant in the understanding the intricate C fractions and their dynamic interplay with environmental factors. Materials and methodsIn this study, we meticulously selected four distinct locations along a principal roadway within a national reserve. For each site, experimental measuring 2 m * 2 m were established at intervals of 2 m, 10 m, 20 m, 30 m, 40 m, and 50 m from the road's edge. Observational data were accrued during the peak of the vegetative growth seasons in August of both 2020 and 2021. Within each plot, soil and cut-ring samples were procured for comprehensive analysis of soil carbon fractions (Total Carbon, TC, g/kg; Carbon Density, g/m2; Readily Oxidizable Carbon, ROC, g/kg; Dissolved Organic Carbon, DOC, mg/kg; Microbial Biomass Carbon, MBC, mg/kg; Cumulative Mineralization Carbon, CMC, mg(CO2-C)*(kg*soil)−1; Soil Inorganic Carbon, SIC, g/kg), chemical properties (pH; Electrical Conductivity, EC, us/cm; Soil Organic Matter, SOM), and physical properties (Soil Water, SW, %; Bulk Density, BD, g/cm3; area conditions). Vegetation metrics, including above-ground biomass (AGB) and below-ground biomass (BGB), species composition, height, and coverage, were meticulously documented. The four sites were replicated, culminating in 24 uniformly oriented plots to mitigate wind influence. ResultsOur analysis revealed that C stock, as determined by TC and C density increased with distance from the road edge, reaching a peak at approximately 30 m (TC: 19.43 g/kg, C density: 39.26 g/m2, SOM: 27.57 g/kg), followed by a subsequent decline. At the 30 m distance, there was a 60.6 %, 41.2 %, and 38.7 % enhancement in TC, C density, and SOM, respectively, compared to the 2 m distance, and a 23.9 %, 25.9 %, and 19.8 % increase relative to the 50 m distance. Levels of DOC, MBC, and CMC exhibited an upward trend from the road edge to 50 m, suggesting more favorable microbiological conditions in less disturbed locales. SEM demonstrated that SIC is directly influenced by BGB (β = −0.31), pH (β = 0.46), and labile C components (β = 0.56), culminating in a reduction of total C storage (total effect, β = −0.28). The dynamic changes of soil C stocks were primarily explained directly by soil labile C (β = 0.475, p < 0.001), and indirectly by vegetation biomass (β = 0.460) and pH (β = −0.570), and were closely correlated with ROC (0.54, the highest eigenvalue among Labile_C). ConclusionThe findings suggest that the buffering effect of the roadside in these nature reserves up to a distance of 30 m has an signficant impact on C stock, and that C losses attributable to biochemical processes and laible soil C fractions may contribute to diminished C levels in less disturbed zones 50 m away from the road. Our findings also substantiate the moderate disturbance hypothesis, particularly in the milieu of road traffic within the Hulunbuir Nature Reserve, Mongolia. A moderate distance from roadways enhances soil C storage compared to areas near the road or undisturbed natural landscapes.