County-level rice area estimation in southern China using remote sensing data

Rice acreage estimation is a key aspect of assessing rice production. A method of estimating rice acreage at the county level is explored, using data from the HJ-1A/B Chinese Environmental Satellite for Hunan Province, which has complex rice cropping patterns. The method combines supervised and unsupervised classification using a mixed-pixel decomposition model. The rice acreage estimation results were validated by interpretation of RapidEye images for early-season rice and ground survey data for medium-season and late-season rice. The results showed a good correlation between the estimates derived from RapidEye and from HJ CCD data for pure rice pixels (R 2 =0.99 , 0.99, and 0.97 for early-, medium-, and late-season rice). The discrepancy was <10% at the plot level (6.5×6.5 km ) for early-season rice, while it was 12.20 and 12.36% at the plot level (1×1 km ) for medium-season and late-season rice. These results suggested that the method proposed in this study is capable of rice acreage estimation at the county level. The method can also be used in mountainous regions and areas of fragmented planting.