A logistic-based method for rice monitoring from multi-temporal MODIS-Landsat fusion data

Information on rice cropping activities and growing areas is critical for crop management. This study developed a logistic-based method to monitor rice sowing and harvesting activities and, accordingly, to map rice growing areas from the MODIS-Landsat fusion data in An Giang Province, Vietnam. The EVI2 data derived from the fusion data compared with that of Landsat data indicated a close correlation (R2 = 0.93). The comparisons between the estimated sowing and harvesting dates and the field survey data revealed the RMSE values of around 8 and 5 days for the winter-spring crop and 9 and 12 days for the summer-autumn crop, respectively. The rice mapping results compared with the ground reference data indicated an overall accuracy and Kappa coefficient of 93.2% and 0.86 for the winter-spring crop, and 91.7% and 0.83 for the summer-autumn crop, respectively. These results were reaffirmed by the government's rice areas statistics, with the relative error in area values smaller than 3.3%.