Assessing forest fire properties in Northeastern Asia and Southern China with satellite microwave Emissivity Difference Vegetation Index (EDVI)
Rui Li,Yuyun Fu,Yves Bergeron,Osvaldo Valeria,Raphaël D. Chavardès,Jiheng Hu,Yipu Wang,Jiawei Duan,Dong Li,Yuanxi Cheng
DOI: https://doi.org/10.1016/j.isprsjprs.2021.10.019
IF: 12.7
2022-01-01
ISPRS Journal of Photogrammetry and Remote Sensing
Abstract:In the context of global warming, forest fires are expected to occur more frequently and intensively, and impose more significant impacts on human society, terrestrial ecosystems, and atmosphere. Most of the existing methods in monitoring large-scale forest fire are based either on satellite visible and infrared observations or weather-based indices. This work explored the advantages of a new satellite microwave-based vegetation index in monitoring forest fire occurrence and fire intensity in Northeastern Asia and Southern China. Specifically, we used satellite observations during 2002–2011 to investigate the correlation at different temporal scales between forest fire properties (fire count, FC; fire radiative power, FRP) and the vegetation water content proxy of the Microwave Emissivity Difference Vegetation Index (EDVI) derived from the Moderate Resolution Imaging Spectroradiometer and the Advanced Microwave Scanning Radiometer-EOS on Aqua satellite. The correlations were compared to that with three weather-based indices including the Fine Fuel Moisture Code, Initial Spread Index and Fire Weather Index (FWI) to determine whether EDVI provides new independent information of forest fires. Finally, EDVI and the weather-based indices FWI were combined to establish multivariate linear regression models to estimate FC and FRP. Results show that: 1) the temporal variations of FC and total FRP are negatively correlated with EDVI using the daily and monthly observations at 1° grid and regional scales; and overall opposite annual cycles and interannual variations between FC (and total FRP) and EDVI are observed in Northeastern Asia and Southern China; 2) compared to the weather-based indices, EDVI shows higher correlation with the temperate forest fire properties in Southern China while shows weaker correlation with the forest fire properties in Northeastern Asia; and a combination of the two kind of indices is found to improve the explained variance for fire properties in both regions; 3) multivariate linear regression models based on EDVI and FWI provide better estimation of FC and FRP compared to the linear regression models based on FWI alone. To our knowledge, this is the first work that comprehensively investigates the potential application of the microwave-based vegetation water content index in forest fire count and fire intensity.
imaging science & photographic technology,remote sensing,geography, physical,geosciences, multidisciplinary