Estimating Paddy Rice Production and Cropping Intensification in the Mekong Delta, Vietnam

Abstract

Rice is Asia’s most important food crop and contributes significantly to the GDP through exports. Knowledge about its production is significant in the contexts of, not only food security, but also water management, trade policy, greenhouse gas emission and hydrological modelling. Paddy fields in Asia are often located in coastal lowlands and river deltas, which are vulnerable to sea level rise, floods and droughts, salt water intrusion and other hazards. At the same time these regions are battling issues of urban encroachment, stagnating yields, economic pressure and increased demand due to population growth. We performed two studies in the Mekong Delta to detect changes in seasonality, referring to multiple rice crops cultivated on a field within a year, as well as estimating rice yield and production. The Mekong Delta was chosen as the study area since it exhibits a large amount of rice agriculture and is facing issues that can be related to most other Deltas with major rice production, chief among which is an increase in salinity intrusion, intensification of rice agriculture in high yield areas and switch to pond aquaculture in low yield areas. Changes in seasonality have been studied using a 15-year spanning MODIS time-series. We calculated phenological metrics, such as length of season, start and end dates, from the time-series to show the number of rice crops planted for each year as well as changes in this seasonality in the last 15 years. We can show spatial patterns of intensification, floods and droughts and water regulatory measures. Copernicus Sentinel-1 time series have been used to estimate rice yield and production. We used a segmentation approach to create object-wise time series. Using in-situ yield data we trained gradient boosted regression trees to estimate rice yield and production for each of the three major rice growing seasons.