Synergistic use of multi-satellite remote sensing to detect forest fires: A case study in South Korea

Forest fire frequency is increasing owing to climate change. Therefore, better forest fire monitoring strategies are required, as they can start unexpectedly and spread rapidly. Earth observation satellites can efficiently prompt rapid responses to forest fires. In this study, Burned Area Index (BAI) and difference Normalized Burn Ratio (dNBR) were analysed to detect and monitor a forest fire in Korea using data from four sun-synchronous satellites and one geostationary satellite, and the results were compared in terms of their spatial, temporal, and spectral resolutions. KOMPSAT-3A efficiently estimated detailed information of the fire on a local-scale for its spatial resolution but was limited to only observing the local-scale fire due to its narrow swath. Sentinel-2 and Landsat-8 were adequate for observing the forest fire on both local- and large-scales and provided more spectral bands and temporal information, which increased the accuracy of detecting the fire damage. Visible Infrared Imaging Radiometer Suite (VIIRS) and GK-2A showed the highest temporal resolution and enabled early detection of the wildfire and its duration, but their low spatial resolutions limited damage estimates to the local-scale. Thus, satellites worldwide may be used synergistically to ensure efficient responses to frequent and massive forest fires.

受气候变化影响，森林火灾发生频率正持续攀升。由于森林火灾往往突发且蔓延迅猛，因此亟需更为完善的森林火灾监测策略。对地观测卫星能够高效助力森林火灾的快速响应工作。本研究借助4颗太阳同步卫星与1颗地球静止卫星获取的数据，通过分析过火面积指数（Burned Area Index, BAI）与差值归一化燃烧比（difference Normalized Burn Ratio, dNBR），对韩国境内的森林火灾开展检测与监测，并从空间分辨率、时间分辨率与光谱分辨率三个维度对结果展开对比。KOMPSAT-3A凭借优异的空间分辨率，能够精准估算局部尺度下火灾的详细信息，但由于其幅宽较窄，仅可用于局部尺度火灾的观测。Sentinel-2与Landsat-8则可适用于局部与大尺度的森林火灾观测，且提供了更为丰富的光谱波段与时间序列信息，从而提升了火灾损毁检测的精度。可见红外成像辐射计套件（Visible Infrared Imaging Radiometer Suite, VIIRS）与GK-2A拥有最高的时间分辨率，能够实现野火的早期检测及其蔓延时长的监测，但较低的空间分辨率使其仅可开展局部尺度的损毁估算。综上，全球范围内的卫星可协同应用，以保障对频发且大规模森林火灾的高效应对。