Assessing post-disaster recovery through high-resolution land use mapping: evidence from Fukushima, Japan

Monitoring land-use and land-cover (LULC) dynamics is essential for understanding long-term recovery processes in post-disaster environments. Following the 2011 Great East Japan Earthquake and the Fukushima Daiichi nuclear accident, many areas of Fukushima experienced prolonged evacuation, abandoned buildings, vegetation expansion, and uneven redevelopment. Assessing these recovery processes remains challenging because conventional land-use surveys are difficult to conduct in formerly evacuated areas, and existing LULC mapping studies often focus on land-cover classification rather than extracting recovery indicators needed to evaluate localized recovery dynamics. This study develops a recovery-oriented assessment framework that integrates high-resolution aerial imagery, Attention U-Net semantic segmentation, and recovery indicator extraction to assess post-disaster recovery. Multi-temporal LULC maps are used to identify patterns of building demolition, vegetation succession, and redevelopment associated with post-disaster recovery in Namie and Odaka, Fukushima. High-resolution 5 m LULC maps were generated from 0.5 m aerial imagery using an Attention U-Net segmentation model and subsequently used to assess recovery trajectories from 2012 to 2022. Building demolition estimates were consistent with independent survey observations and official demolition statistics, supporting the reliability of the recovery indicators derived from the proposed framework. The results reveal substantial spatial heterogeneity in recovery processes, characterized by vegetation expansion in abandoned areas, extensive demolition activities in urban centers, and localized redevelopment following the lifting of evacuation orders. The proposed framework demonstrates how high-resolution LULC information can be systematically transformed into spatially explicit recovery indicators, providing a practical approach for monitoring post-disaster recovery processes, supporting land management, and informing evidence-based redevelopment planning.