Estimation of Southeast Asian rice paddy areas with different ecosystems from moderate-resolution satellite imagery

Multi-temporal satellite imagery from the Moderate Resolution Imaging Spectrometer (MODIS) was used to map the different ecosystems of Southeast Asian (SEA) rice paddies. The algorithm was based on temporal profiles of vegetation strength and/or water content, using MODIS surface reflectance in visible to near-IR range. The results obtained from the analysis were compared to national statistics. Estimated SEA regional rice area was 42 × 106 ha, which agrees with published values. The model performance was dependent on rice ecosystems. Good linear relationships between the model results and the national statistics were observed for rainfed rice. High linear coefficients of determination, R2, were also found for irrigated rice and upland rice, but the model tended to underestimate irrigated rice and overestimate upland rice. However, these high R2 values indicated that the model effectively simulated spatial distribution of these rice areas. These R2 values were either of similar magnitude or larger than those reported in literature, regardless of the rice ecosystem. Poor correlation was observed for deepwater rice.

► We mapped SE Asian rice paddies with different ecosystems using the satellite data.
► Mapping model is derived from observed vegetation–water phenology.
► The model well simulated area distribution of rainfed, irrigated, and upland rices.
► This model is more appropriate than that without rice ecosystem extraction.