Crop phenology estimation by time series MODIS medium-resolution data

Crop phenology is a critical component of implementing agricultural activities and providing important indicators for climate related research. The needs for crop phenology estimation not only exist in large scale but also in medium or small scale. In this work, Jianhu County was chosen to test the ability of crop phenology estimation by Moderate Resolution Imaging Spectroradiometer (MODIS) data. 794 cloudy-free MODIS 250-m resolution images of Jianhu County from 2000 to 2013 were finally chosen. Two vegetation indices were compared to evaluate their effectiveness in crop phenology estimation. These were the Normalized Difference Vegetation Index (NDVI) and the Enhanced Vegetation Index (EVI). EVI was less sensitive than NDVI to aerosol effects. Therefore, EVI was chosen to estimate crop phenology of Jianhu County. Then, the crop spatial distribution data were generated from RapidEye 5-m resolution images by visual interpretation to exclude those mixed pixels. In the long-term monitoring, ground survey and correlation analyses, we found that the MODIS-EVI time series data were suitable to estimate the turning green, the heading and the maturation of winter wheat and were suitable to estimate the transplanting, the heading and the maturation of rice. Because of the free availability and frequent coverage of MODIS data, the cost-effective assessment here by using MODIS 250-m resolution images may serve as a template applicable to other counties of Jiangsu province. Moreover, crop phenology estimation in county scale will be particularly helpful in providing some baseline information to yield estimation and food security, and will be useful to promote crop phenology estimation methods.