基于不同遥感数据的林火污染物排放估算对比研究

Evaluating the influence of spatial resolution on forest fire pollutant emission estimation

遥感技术是进行林火污染物排放估算的重要手段,而空间分辨率的差异会对林火污染物排放估算产生重要影响。文中采用“自下而上”的遥感估算方法,针对三种不同空间分辨率遥感数据(哨兵2号、Landsat OLI、MODIS)分别进行了林火污染物排放量估算,结果表明:1)由空间分辨率造成的火烧迹地整体差异范围介于1.81%~9.82%之间。2)从可燃物载量来看,林地可燃物载量最大相差20.69%,灌丛最大相差13.78%,草地最大相差31.13%。3)从各燃烧强度面积来看,重度燃烧的差异范围为12.93%~45.08%,中度燃烧的差异范围为18.22%~18.84%,轻度燃烧的差异范围为6.20%~10.96%。4)基于MODIS、Landsat OLI以及哨兵2号遥感数据计算的污染物总排放量分别为4.13×10^(5)t、4.44×10^(5)t以及4.92×10^(5)t,哨兵2号与Landsat OLI计算结果的差异为9.77%,与MODIS计算结果的差异为16.01%。

Remote sensing is an important tool to estimate pollutant emissions derived from forest fires,and the difference in spatial resolution can impact significantly the estimated values.In this study,a"down to up"method was used to estimate forest fire pollutant emissions based on remote sensing data with three different spatial resolutions(Sentinel-2,Landsat OLI,and MODIS).The results show that:1)Remote sensing data with a high spatial resolution describes the detailed characteristics of burned areas more clearly,while those with low spatial resolution present a certain omission error in the extraction of fire traces and cannot monitor fire patches with a small area.The overall variation of the burned area associated with spatial resolution ranges from 1.81%to 9.82%.2)Spatial resolution can affect land cover classification which potentially leads to different amounts of fuel load.The maximum differences of fuel loads for forest,shrubland,and grassland are 20.69%,13.78%,and 31.13%,respectively.3)In terms of the area of each combustion intensity level,the difference in severely burned area is between 12.93%and 45.08%,the difference in moderately burned area is between 18.22%and 18.84%,and the difference in mildly burned area is between 6.20%and 10.96%.4)The total amounts of pollutant emissions estimated based on MODIS,Landsat OLI and Sentinel-2 are 4.13×10^(5)t,4.44×10^(5)t,and 4.92×10^(5)t,respectively.The difference between the results of Sentinel-2 and Landsat OLI is 9.77%,and the difference between the results of Sentinel-2 and MODIS is 16.01%.