DisTrad热像元分解模型运用于高植被覆盖区的问题及改进

DisTrad Model for Thermal Sub-pixel Mapping in High Vegetation Area

验证了DisTrad热像元分解模型在华南地区的可行性,发现在高植被覆盖区,由于生物量高,NDVI(Normalized Difference Vegetation Index)指数已接近饱和,在热像元分解时敏感性差。提出用EVI(Enhanced Vegetation Index)指数来代替DisTrad模型中的NDVI参数进行热像元分解。通过华南地区MODIS高分辨率EVI及NDVI资料(250 m)与地表温度(LST)的关系,获得了高分辨率(250 m)LST图像,并利用同步ASTER高分辨率LST图像(90 m)进行了验证。结果表明:即使在NDVI指数已接近饱和的高生物量地区,EVI指数仍然保持较高的敏感度,在高植被覆盖区(如华南地区)用EVI代替DisTrad模型中的NDVI指数能够减少运算量,并能获得更好的热像元分解结果。

This paper validates the applicability of DisTrad model for thermal sub-pixel mapping over south China.We found the index of NDVI would reach to saturate state in high biomass regions,and its sensitivity would also become decreased.In our study,we proposed to replace NDVI by EVI in DisTrad model for thermal sub-pixel mapping in South China area.The high-resolution LST image（250 m） was got by the relationships between high-resolution EVI/NDVI and LST images,and the results were verified by co-temporal ASTER LST data with 90 m resolution.The results show that this method can reduce the computation time and remain sensitivity in high-biomass areas.