Accessibility-Associated Forest-Cover Loss and Fragmentation Around Mining Roads and Rural Settlements in the Miombo Woodlands of Southeastern DR Congo

The miombo woodlands of the southeastern Democratic Republic of the Congo (DR Congo) are increasingly exposed to mining expansion and associated land-use changes. However, the spatial association between mining infrastructure, rural settlements, and forest-cover dynamics remains poorly quantified. This study assessed long-term forest-cover change and fragmentation in relation to accessibility gradients around a mining exploration road and four rural settlements in Mutshatsha Territory, a rapidly transforming mining frontier within the Katangese Copperbelt. Land-cover dynamics from 1998 to 2023 were reconstructed from six Landsat observations using Random Forest classification implemented in Google Earth Engine. Landscape metrics were used to quantify changes in forest composition and configuration, while distance-based analyses assessed spatial patterns along accessibility gradients. Forest cover declined from 39.9% in 1998 to 12.5% in 2023, corresponding to an annual forest-cover loss rate of 2.75%, while agricultural and built-up/bare-soil classes expanded. Forest fragmentation increased, as reflected by greater patch subdivision and declining dominance and cohesion of large forest patches. Forest cover generally increased with distance from the mining road after accounting for year, indicating a significant accessibility-associated spatial gradient. Linear regressions showed strong statistical associations between forest cover and road-distance gradients (R2 = 0.71–0.94), while village-distance relationships reached R2 values of up to 0.80. These R2 values represent the proportion of variance accounted for by the fitted spatial relationships and should not be interpreted as causal effects of accessibility. Forest-cover patterns also varied among villages and years, indicating that accessibility-associated gradients occurred within broader temporal and landscape dynamics. Overall, the results demonstrate substantial forest-cover loss and landscape reorganization associated spatially with accessibility gradients around mining infrastructure and rural settlements. These findings support conservation and land-use planning that considers remaining forest patches, landscape connectivity, and accessibility patterns beyond the immediate infrastructure footprint.