Agricultural drought threatens food security and agricultural sustainable development. There have been numerous spectral indices from remote sensing images developed for monitoring crop drought. However, most present spectral indices are focusing on crop growth and Land Surface Temperature (LST), and the crop canopy water content are in less consideration simultaneously. Additionally, the Normalized Difference Vegetation Index (NDVI) is used for characterizing crop growth in almost all spectral drought indices, with the spectral saturation problem of NDVI for closed crop canopy. When vegetation cover is high, NDVI values tend to saturate, which makes them insensitive to further changes in crop health. Therefore, the NDVI saturation phenomenon may lead to an underestimation of the extent of crop drought, as it is not effective in identifying subtle changes in crops under high-density vegetation conditions. Hence, we propose three novel triangular spectral indices for characterizing winter wheat drought using three features including LAI, Land Surface Water Index (LSWI) and LST. For validating the proposed spectral indices, we compared the agreement between these indices with measured Relative Soil Moisture (RSM) and Volumetric Water Content (VWC) of soil in agricultural meteorological station and present popular indices including Crop Water Stress Index (CWSI), Temperature-Vegetation Drought Index (TVDI), and Vegetation Health Index (VHI). The results revealed that our proposed indices including Euclidean distance Crop Health Index (ECHI), Difference Crop Health Index (DCHI) and Perpendicular Water Stress Index (PWSI) outperformed the popular CWSI, TVDI and VHI, with stronger correlations with measured RSM and VWC in agricultural meteorological station. Secondly, there are spatial consistencies for characterizing winter wheat drought between proposed ECHI, DCHI and PWSI with popular CWSI, TVDI and VHI. In addition, our proposed ECHI, DCHI and PWSI have achieved good performance of drought monitoring both in irrigated and rainfed croplands. All these results suggest that our proposed indices have great potential in crop drought monitoring.

中文翻译：

---

表征冬小麦干旱的新型三角光谱指数


农业干旱威胁粮食安全和农业可持续发展。已经开发出许多用于监测作物干旱的遥感图像光谱指数。然而，目前的光谱指数大多关注作物生长和地表温度（LST），而很少同时考虑作物冠层含水量。此外，归一化植被指数（NDVI）用于表征几乎所有光谱干旱指数中的作物生长，以及封闭作物冠层的NDVI光谱饱和问题。当植被覆盖率较高时，NDVI 值趋于饱和，这使得它们对作物健康状况的进一步变化不敏感。因此，NDVI饱和现象可能导致低估作物干旱程度，因为它不能有效识别高植被条件下作物的细微变化。因此，我们提出了三个新颖的三角光谱指数，利用 LAI、地表水指数 (LSWI) 和 LST 等三个特征来表征冬小麦干旱。为了验证所提出的光谱指数，我们将这些指数与农业气象站测量的土壤相对湿度（RSM）和土壤体积含水量（VWC）之间的一致性进行了比较，并提出了流行的指数，包括作物水分胁迫指数（CWSI）、温度-植被干旱指数 (TVDI) 和植被健康指数 (VHI)。结果显示，我们提出的指数，包括欧氏距离作物健康指数（ECHI）、差异作物健康指数（DCHI）和垂直水分胁迫指数（PWSI）优于流行的CWSI、TVDI和VHI，与测量的RSM和VWC具有更强的相关性。农业气象站。 其次，所提出的 ECHI、DCHI 和 PWSI 与流行的 CWSI、TVDI 和 VHI 之间在表征冬小麦干旱方面存在空间一致性。此外，我们提出的 ECHI、DCHI 和 PWSI 在灌溉农田和雨养农田的干旱监测中都取得了良好的效果。所有这些结果表明我们提出的指数在作物干旱监测方面具有巨大潜力。