Monitoring rice farming activities in the Mekong Delta region

Abstract:

Half of the world's population depends on rice for survival. Rice agriculture thus plays an important role in the developing world’s economy. Vietnam is one of the largest rice producers and suppliers on earth and more than 80% of the exported rice was produced from the Mekong Delta region, which is situated in the southwestern Vietnam and encompasses approximately 40,000 km². Changes in climate conditions could likely trigger the increase of insect populations and rice diseases, causing the potential loss of rice yields. Monitoring rice-farming activities through crop phenology detection can provide policymakers with timely strategies to mitigate possible impacts on the potential yield as well as rice grain exports to ensure food security for the region. The main objective of this study is to develop a logistic-based algorithm to investigate rice sowing and harvesting activities from the multi-temporal Moderate Resolution Imaging Spectroradiometer (MODIS)–Landsat fusion data. We processed the data for two main cropping seasons (i.e., winter–spring and summer–autumn seasons) through a three-step procedure: (1) MODIS–Landsat data fusion, (2) construction of the time-series enhanced vegetation index 2 (EVI2) data, (3) rice crop phenology detection. The EVI2 data derived from the fusion results between MODIS and Landsat data were compared with that of Landsat data indicated close correlation between the two datasets (R² = 0.93). The time-series EVI2 data were processed using the double logistic method to detect the progress of sowing and harvesting activities in the region. The comparisons between the estimated sowing and harvesting dates and the field survey data revealed the root mean squared error (RMSE) values of 8.4 and 5.5 days for the winter–spring crop and 9.4 and 12.8 days for the summer–autumn crop, respectively. This study demonstrates the effectiveness of the double logistic-based algorithm for rice crop monitoring from temporal MODIS−Landsat fusion data. The results in forms of spatialtemporal and quantitative information of rice sowing and harvesting activities were vital for crop management, and the methods are thus suggested for rice crop monitoring in the study region and could be transferable to other regions for crop monitoring.