Winter wheat yield estimation model with MODIS normalized near-infrared spectral index

Terra/MODIS has spectral and spatial resolution advantage over NOAA/AVHRR. To probe into using MODIS near-infrared spectrum further, winter wheat yield estimation was taken as example in Hebei Province, China. Firstly, according to winter wheat biological characteristic, three MODIS near-infrared spectrum data were retrieved in heading stage, which central wavelength is 860 nm, 1240 nm and 1640 nm. Secondly, the normalized near-infrared spectral index (NNSI) is defined by every two near-infrared spectrum, such as (860 nm, 1240 nm), (860 nm, 1640 nm) and (1240 nm, 1640 nm). Thirdly, the statistical correlation analysis with yield were carried on and set up models for yield forecasting with NNSI. The result shows their coefficient correlations are greater than 0.77 and better than with NDVI. Especially the NNSI defined by (860 nm, 1640 nm), its coefficient correlation is 0.815. So NNSI can do well to forecast winter wheat yield. So we can conclude that normalized index in near-infrared spectrum can do better and more reliable than normalized index in visual and near-infrared spectrums for yield forecasting. And given play to the hysperspectral advantage of MODIS, it can service to crop condition monitoring and crop yield estimation of Ministry of Agriculture.