Fragmented Networks, Detoured Access: A Nationwide Graph-Theoretic Analysis of Forest Road Connectivity in the Republic of Korea

Forest road networks are almost universally characterised by road density alone, which says nothing about how the roads are arranged. We assembled the 2025 national forest road network of the Republic of Korea (14,007 routes; 25,958 km) with the 1:5000 national road centreline layer into a weighted graph and analysed it with graph-theoretic connectivity indices, centrality and two measures of circuity. Considered alone, the registered forest road system is a disconnected network of 4742 components in which circuitry is close to absent. Public roads bring 96.0% of forest road length into one connected component, almost entirely through minor and unclassified roads, so the system is functionally connected; the quality of that connection is what is measured here. Route-level circuity averaged 1.574 (95% CI 1.568–1.579), whereas the median of the individual circuity ratios between adjacent components was 2.92 (2.82–3.06). The median straight-line gap was 708 m and the median network distance 3002 m, separate medians that need not describe the same pair. Ridge height accounts for part of that variation (Spearman +0.37) but not all of it. Assuming the statutory 10% grade ceiling, the geometrically estimated length of a connecting link was shorter than the existing network path for most adjacent pairs; the estimate is geometric and is not an alignment design. Two-thirds of pairs already lie closer together than half the spacing implied by the national density, so connectivity is a second planning criterion alongside density rather than a substitute for it.