Comparison Of Five Drought Indices For Agricultural Drought Monitoring And Impacts On Winter Wheat Yields Analysis
Wen Zhuo,Jianxi Huang,Xiaodong Zhang,Hao Sun,Dehai Zhu,Wei Su,Chao Zhang,Zhe Liu
DOI: https://doi.org/10.1109/Agro-Geoinformatics.2016.7577702
2016-01-01
Abstract:Various remote sensing-based drought indices have been developed during the past decades. These indices have their own applicability. Examination of adaption associated with limitations in drought indices would be useful for index selection and improvement. In this study, five drought indices (e.g. the temperature condition index (TCI), vegetation condition index (VCI), vegetation health index (VHI), temperature vegetation dryness index (TVDI), drought severity index (DSI)) generated by time series of MODIS data, which include 8-day MOD11 LST products, MOD13 NDVI products and MOD16 ET/PET products at 1km resolution from 2000 to 2012, were performed in Henan, and Shaanxi province in the northern China, aiming to better understand the differences and potential monitoring ability among these drought indices. 8-day drought indices were compared against winter wheat yield at the prefecture-city scale from 2000 to 2012. Results showed that TVDI outperformed the five drought indices in correlation with winter wheat yield, and followed by DSI, and next is the VHI, whereas TCI and VCI have the lowest correlation. Furthermore, during the winter wheat growth period, drought shows varying agricultural impacts on winter wheat yield at different phenological stages, jointing-heading stage and flowering-filling stage have higher correlations than others, and the highest correlation coefficient can reach 0.8, which indicated that winter wheat requires more water during these stages. As a result, the sensitive phenological stages can be used as the key period for effective agricultural drought monitoring. And during the two highest correlation growth period-jointing-heading stage and flowering-filling stage, higher correlations were found between TVDI and the winter wheat yield, which the correlation coefficient can reach 0.5. Furthermore, the correlation between drought indices and yield in Shaanxi province is generally stronger than that of Henan province, which showed that remote sensing drought indices have a better indication in rain fed area on drought.