Real-time remote-sensing based monitoring for the rice industry
Project start: 30 November 2021
Project close: 30 November 2024
Objectives
This project aims to provide timely information to rice growers, which will support decision making on the timing of crop management to increase water productivity (yield per water use, tonnes/hectare). The seven objectives are:
- Predict start-of-season and permanent water dates per field.
- Predict end-of-season grain moisture vs time, supporting drainage and harvest timing decisions.
- Identify growth stages of each field, including panicle initiation, flowering, and maturity.
- Produce real-time information on current rice growth versus bench marked information.
- Map rice fields early in the growing season.
- Forecast potential yield per field.
- Combine the insights from the above objectives to perform industry-level analysis, generating information on productivity improvements through adoption of various water management strategies.
Background
The Australian rice industry has achieved great progress in the last 20 years, with substantial gains in yield and water productivity; as a result of ongoing agronomic research, varietal improvements, and industry-research-extension integration. However, there is a need to further improve water productivity to ensure resilience with increasing pressure on water availability.
Real-time monitoring of rice crops using remote sensing has been adopted in other international growing regions. Information derived from satellite imagery and spatial weather datasets can be delivered to the whole industry, throughout the growing season, in near real time. As well as measuring crop vigour and variability, remote sensing data can be used to determine agronomic parameters that are important in optimising crop management and decision-making.
Notifications alerting rice growers to field-specific key growth stages, grain moisture content and harvest readiness have not previously been delivered. This project will bring together remote sensing, weather, and field observations, applying machine learning techniques to develop models to enable field-scale predictions. These will be delivered to growers regularly, supporting improved water productivity through the optimal timing of inputs (such as water and nutrients) and harvest management decisions.
Information from remote sensing will augment industry databases, enabling analysis leading to insights on optimised growing practices, adoption of new water management strategies and prediction of future productivity. These insights will facilitate grower decision making leading to adoption of best-practice crop management strategies.
Research
The project will develop predictive models, which will be validated against in-field measurements of crop parameters. The models will be trained and deployed using data from historical field trials (collected by NSW DPI, RRAPL and collaborating growers), frequently updated satellite imagery (from the Sentinel and Planet satellites) and gridded weather observations and forecasts. Ongoing commercial-scale field trials will be established, encompassing the main growing regions, varieties and crop management strategies. Models will be updated and tested as new data becomes available, and feedback from growers will be used to improve them. Outputs will be automated and calculated on a daily and per-field basis, including water application dates, key phenological stages, grain moisture and plant growth monitoring. Delivery of outputs to growers will be through established channels, including from SunRice Grower Services.
Expected outcomes and implications:
This project is focused on the ‘when’ questions of rice crop management, providing tools to support timely grower decision making on critical crop management steps. The provision of timing information to growers, such as bench marked rice growth profiles and key management dates (water application, PI nitrogen, flowering, maturity, harvest) will inform understanding of crop progression on a per-field basis. This will augment information available to growers guiding the timing of water and fertiliser application, with potential increase in yield and water productivity.
Grain moisture and maturity predictions have not been previously available. These will enable better decisions on drainage and harvest, improve quality and overall farm water productivity and aid post-harvest planning.
Automation of rice field detection, water application and harvest dates will alleviate the onus on growers to accurately enter this data, and also improve the accuracy of industry databases to empower enhanced bench marking analysis. Water and phenology date predictions, together with yield data, will underpin industry-wide analysis, generating insights on management factors that lead to water productivity increase.
This project builds on existing collaboration between industry (SunRice, Grower Services, RRAPL and individual growers), research organisations (UNE, DPI), extension (Rice Extension) and commercial imagery and analytics providers (Planet, Google). This cross-sectoral engagement ensures the industry participants can provide directiozn to ensure the project outputs are relevant to growers; the research participants ensure scientifically rigorous validation considering variability across sites, seasons and rice varieties; whilst the inclusion of commercially available technologies ensures widespread adoption by growers can be fast tracked.
Publications
- Darbyshire, R., Crean, E., Dunn, T., & Dunn, B. (2019). Predicting panicle initiation timing in rice grown using water efficient systems. Field Crops Research, 239, 159-164. https://doi.org/10.1016/j.fcr.2019.05.018
- Dunn, B. W., & Gaydon, D. S. (2011). Rice growth, yield and water productivity responses to irrigation scheduling prior to the delayed application of continuous flooding in south-east Australia. Agricultural Water Management, 98(12), 1799-1807. https://doi.org/10.1016/j.agwat.2011.07.004
- Van Niel, T. G., & McVicar, T. R. (2004). Current and potential uses of optical remote sensing in rice-based irrigation systems: a review. Australian Journal of Agricultural Research, 55(2), 155-185. https://doi.org/10.1071/AR03149
Acknowledgements
NSW Department of Primary Industries
SunRice (including RRAPL and Grower Services)
University of New England's Applied Agricultural Remote Sensing Centre
PhD Opportunities
Decision Making Tools for the Rice Industry, Using Remote Sensing, Meteorology and Machine Learning
