Evaluation of Typical Spectral Vegetation Indices for Drought Monitoring in Cropland of the North China Plain

Abstract

Spectral vegetation index (SVI) is indispensable for comprehensively monitoring of agricultural drought based on remote sensing. Greenness sensitive (e.g., normalized difference vegetation index) and water sensitive SVIs (e.g., normalized difference infrared index) were mostly concerned in drought monitoring. Their sensitivity to drought conditions has been explored yet distinct results were found. This contradiction has not been fully understood in previous studies. In this study, the greenness and water sensitive SVIs were further evaluated in a cropland area of the North China Plain by using MODIS reflectance products from 2000 to 2015. Results indicate that fractional vegetation coverage (FVC) plays an important role in the sensitivity of SVIs to drought condition. During early months of the crop growth, when FVC is relatively low, greenness sensitive indices had closer correlation with in situ drought indices. On the contrary, water sensitive indices presented better correlation during the later months of the crop growth when FVC is high. This study further suggests that variations in vegetation water content are more dynamic than changes in vegetation greenness properties under the influence of drought. However, soil background effects will make greenness sensitive indices outperform water sensitive indices for mid and low density vegetation coverage.