Analyzing the Relationship Between Remotely-Sensed Environmental Conditions and Arsenic Absorption in Rice During the Growing Season in Cambodia

Arsenic (As) contamination, a known carcinogen, impacts the health of nearly 200 million people around the world, especially those who live along the floodplains of Southeast Asia. As naturally mobilizes into pore water via the reduction of iron oxides that bind it. Public exposure to As increases as it accumulates in staple agricultural products, specifically rice, when it is irrigated with high As water (finding over 80% of samples exceed the 100 ppb safety guideline). Concentrations of As within a harvested rice grain vary significantly both annually and geographically, making it difficult to identify the driving environmental and geochemical variables. Hence, we currently lack the ability to predict As levels in rice (and water) and effectively manage them, critical steps in solving this public health crisis. We posit that rice As levels vary in response to variable soil redox conditions (flooding), nutrient status (fertilization) and rice variety (affecting how those conditions impact the plant). In this study, we measure the chemical composition of over 200 Cambodian rice samples grown by subsistence farmers in known paddy fields across three years with widely variable flooding conditions. Mean As concentrations and variability were highest in 2018 at 251.1 ppb (decreasing to 163.4 and 141.7 ppb in 2019 and 2020, respectively). Flooding conditions and rice growth were estimated from remotely-sensed measurements of flooding frequency and normalized difference vegetation index (NDVI) at each site during each growing season. NDVI data trends suggest that it may also be used as a more accurate proxy for water presence and indication of individual field growing conditions, including the use of fertilizer. Predictive models of grain As would require more environmental factors, such as precipitation and temperature, to be incorporated alongside rice physiological factors affecting when As uptake occurs. A more precise sample extraction technique may improve rice composition data; and further analysis of site-specific data should be filtered and viewed through lens of regional grouping, rice varieties, and individual variation. With these future steps, we anticipate that key drivers controlling rice As levels will become evident and our ability to predict these concentrations will subsequently improve.