基于改进温度植被干旱指数的农田土壤水分反演方法

基于植被指数 地表温度（VI Ts）特征空间的温度植被干旱指数（TVDI）被广泛应用于土壤水分监测，但TVDI为土壤水分相对值，而且利用散点图确定干湿边会造成很大的不确定性。基于能量平衡方程和TVDI，该文提出一种定量干湿边选取方法和改进的TVDI模型——定量温度植被指数（Temperature Vegetation Quantitative Index，TVQI），以MODIS遥感数据为基础，实现了定量干湿边真实土壤水分的遥感估算。结果表明：TVQI估算结果与所观测土壤水分呈0.01水平显著相关，总体上的平均绝对误差小于0.02cm3/cm3，均方根误差RMSE小于0.035cm3/cm3；相对TVDI，TVQI克服了传统干边计算中对植被覆盖类型的限制，更能够准确反应土壤深度在0～10cm、10cm～20cm的土壤水分值，尤其与10cm～20cm土壤水分值更为贴近。

Soil Moisture Estimation Based on Improved Temperature Vegetation Dryness Index

Temperature vegetation dryness index （TVDI） has been widely used to monitor soil moisture.However，TVDI represents the relative value of soil moisture and the dry edge is fitted by the observations，which causes a great uncertainty.This research proposes a method for quantitative selection of dry edge and wet edge based on the energy balance equation and an improved TVDI model：quantitative temperature vegetation index （TVQI）.With this method，the real value of soil moisture was estimated using moderate resolution imaging spectroradiometer （MODIS） data in this research and validated with in situ measurements at depth of 0～10cm and 10cm～20cm.The bias and RMSE are less than 0.02cm3/cm3 and 0.035cm3/cm3，respectively，with a significant correlation at the 0.01 level （Pearson correlation coefficient>0.67）.Compared with TVDI，TVQI can overcome the limitations of the vegetation index land surface temperature （VI Ts） space on vegetation cover types and correctly retrieve the soil moisture at the depth of 0～10cm and 10cm～20cm，especially at 10cm～20cm.