The Greenness–Moisture and Structure Gap as a Diagnostic of Post-Fire Recovery Quality: Toward Sustainable, Evidence-Based Restoration After the 2025 Gyeongbuk Wildfire

Reliable assessment of post-fire recovery is essential for sustainable land management and for tracking progress toward Land-Degradation Neutrality (LDN; Sustainable Development Goal (SDG) target 15.3), yet greenness-only metrics risk overstating recovery and misdirecting scarce restoration resources. After the March 2025 Gyeongbuk megafires in the Republic of Korea, we developed a multi-indicator framework in Google Earth Engine to assess first-year recovery and mapped 105,550 ha of burned land. Using Sentinel-2 imagery, burn severity was mapped with the differenced Normalized Burn Ratio (dNBR) from a same-season image pair. Recovery was quantified with a greenness index, the Normalized Difference Vegetation Index (NDVI), and a moisture- and structure-sensitive index, the Normalized Difference Moisture Index (NDMI), each expressed as a recovery fraction on an identical pixel mask. Both indices declined with increasing severity, and their divergence widened: the normalized greenness–structure gap, defined as 1 − RF(NDMI)/RF(NDVI), rose from 0.05 in low-severity stands to 0.32 in moderate-low and 0.47 in moderate-high stands and did not increase further in high-severity stands (0.45). Block-bootstrap tests showed that the increases up to moderate–high severity were significant. These increases persisted after correcting for phenology with unburned control pixels, with severity derived from Landsat, and with severity and recovery computed from disjoint images. In the most severely burned stands, NDVI recovery exceeded NDMI recovery by roughly 80%. A random-forest analysis under spatially blocked validation identified burn severity as the dominant control (63–92% of importance, depending on the measure), with topography secondary. These results are consistent with early greenness gains in a severely burned forest being driven by herbaceous and shrub regrowth rather than canopy-structure recovery, so that NDVI alone overestimates ecological recovery. We propose the greenness–structure gap as an interpretable, map-based diagnostic that delineates degradation hotspots and supports cost-effective restoration prioritization through a scalable, low-cost, and openly transferable workflow relevant to sustainable post-fire land governance and LDN reporting.