Image above: PlanetScope image of rice fields in Griffith, New South Wales, Australia captured September 3, 2022. © 2022, Planet Labs PBC. All Rights Reserved.
Our world is faced with increasing pressures to feed a growing global population, and we are simultaneously challenged with climate change impacts and water scarcity. Staple food crops like rice offer a way to provide nutrition to millions of individuals around the world each day, however, this crop can be intensive on water resources, especially in countries faced with drought like Australia. Rice grows best when farmers irrigate the land until they have flooded the rice paddies, a state that is also called “ponded.” So, often growers are confronted with the trade-off of working to obtain higher rice yields at the expense of water resources. For this reason, it’s important to gain scientifically-informed insights on these trade-offs to help enable efficient and sustainable rice production. Currently, scientists from the University of New England in Australia, New South Wales Department of Primary Industries in Australia, and Planet are researching just that. These teams are leveraging Sentinel-2 and Planet Fusion data to better understand this agricultural system and make predictions based on a variety of water management strategies. Using Sentinel-2’s shortwave infrared bands and the gap-free, daily fused satellite insights from Planet, they collected water data to train models covering harvest seasons from 2018 through 2021. They analyzed regions before and after (ponded and not ponded) to evaluate crop trends. This research demonstrated the use of machine learning models, powered by satellite data, to monitor management plans and estimate effective ponding dates for the future. This type of information can have widespread applications for farmers and growers as well as the broader industry and government sectors of Australia.
The full study can be found in Agricultural Water Management.