산사태재난신속대응을위한HLS와Sentinel-2 단독의위성관측가용성비교평가* = Comparative Evaluation of Satellite Observation Availability Between HLS and Sentinel-2 for Rapid Landslide Disaster Response*

다국어 초록 (Multilingual Abstract)

This study evaluated and compared satellite observation availability between Harmonized Landsat-Sentinel-2 (HLS) and Sentinel-2 standalone imagery for rapid response immediately following landslide disasters. An operational window from one to fourteen days after the landslide event (D+1 to D+14) was defined, and the Opportunity-Usable-Bridge (OUB) framework was proposed to separately quantify image acquisition opportunity (Opportunity) and quality-masked usable observations (Usable). An analysis of 2,021 landslide events that occurred in South Korea from 2019 to 2023 revealed that HLS provided approximately 3.6 times more observation days (oppDays) than Sentinel-2 standalone (12.24 vs. 3.36 days), with HLS outperforming Sentinel-2 in 98.6% of events. The maximum consecutive observation gap (oppMaxGap) was also shorter for HLS by an average of 2.39 days (1.13 vs. 3.52 days). However, after applying the quality mask (τ = 0.3), the number of usable observation days (usableDays) was higher for Sentinel-2 (0.54 vs. 0.34 days), and the median value for both sensors was 0.0 days, indicating that in more than half of the events, not a single usable observation was secured within two weeks after the landslide. This result is attributed to the combined effects of HLS's 30 m spatial resolution, its Fmask-based quality masking policy, and the persistence of cloud cover during the summer monsoon season. This study quantitatively demonstrates that improved temporal resolution of satellite imagery does not directly translate into increased usable observations, clearly revealing that the current optical-satellite-only operational system has structural limitations for rapid landslide response during the monsoon season. At the same time, the high acquisition opportunity (oppDays) of HLS holds potential value that could be converted into actual usable observations as cloud-removal techniques, such as SAR fusion, advance. The OUB framework can thus serve as a practical tool for evaluating satellite observation availability in establishing a multi-sensor-based rapid disaster response system.