Remote Sensing Inversion Of Forest Margin Grassland Living Combustibles Moisture Content Based On RADARSAT-2 Dual-polarization SAR Data

With the rapid changes in the global environment,grassland fires frequently occur,and the harm caused by grassland fires is becoming more and more serious,especially in the border areas of China and Mongolia.Among them,live fuel moisture content,defined as the moisture content per unit mass of fuel,is a key vegetation variable affecting wildfire behavior because it is highly correlated with fire ignition and propagation.The Dural Forest Farm is located in the northwest of Arshan,Inner Mongolia Autonomous Region,and is at the downwind.It is vulnerable to the influence of the northwest wind and the Mongolian fire spreads to this area,making the Dural Forest Farm in a high fire hazard area.In this paper,the live fuel moisture content in the forest margin grassland of Arxan Dural Forest Farm in the Sino-Mongolian border area is used as the research object.Using the Radarsat-2 dual-polarization SAR data and the ground measured data in September 2021,a semi-empirical model was constructed for retrieving the water content of grassland live fuel.According to Sentinel-2 satellite data,the grassland spatial distribution in the study area was extracted,and finally,the spatial distribution map of forest margin grassland living fuel moisture content was generated and its characteristics were analyzed.The results obtained in this paper are as follows:（1）Based on the water cloud model and the bare soil scattering ray model,a semi-empirical model for inversion of grassland Fuel Moisture Content is constructed.There is a good correlation between the HH polarimetric radar backscatter coefficient obtained by model simulation and the HH polarimetric radar backscatter coefficient extracted from SAR images.Among them,the coefficient of determination（R~2）and root mean square error（RMSE）of model training were 0.57 and 1.137d B,respectively,R~2=0.42 and RMSE=2.002d B for model validation,while R~2=0.44 and RMSE=1.404d B for all samples.（2）Using the look-up-table method to reverse the established semi-empirical model can realize the high-precision inversion of the moisture content of the grassland living combustibles.There is a high correlation between the moisture content of grassland live fuel inverted by the look-up table algorithm and the measured grassland live fuel moisture content.The R~2 and RMSE of model training are 0.45 and 5.69%,respectively,the R~2and RMSE of model validation are 0.44 and 5.47%,respectively,and the R~2 and RMSE of all samples are 0.46 and 5.64%,respectively.Therefore,it is shown that this model is feasible for inversion of grassland live fuel moisture content in this area.（3）The established look-up table was used to invert the moisture content of grassland live fuels in the entire study area.The results showed that the moisture content of live combustibles in forest margin grassland in the border between China and Mongolia was mainly between 35%and55%.Among them,the areas with low moisture content（<35%）are mainly distributed in the border areas between China and Mongolia in the southwest of Dural Forest Farm.The areas with higher water content（>55%）are mainly concentrated in the northern and central regions.