Early Adopter
Commonwealth Scientific and Industrial Research Organisation (CSIRO)
Applied Research Topic
Potential Applications ► Water quality; HABs; machine learning& AI; tools to support decision-making
Description
CSIRO's AquaWatch mission is to enhance AI capabilities for water quality monitoring and forecasting. A central aspect of this project is the development of deep-learning-based algorithms capable of translating spectral reflectance into precise water quality parameters.
Training these networks using remote sensing imagery is challenging due to insufficient data density and the scarcity of cloud-free images. To address these difficulties, this project will leverage the eReefs model to simulate the environment of the Great Barrier Reef region and generate comprehensive synthetic datasets of satellite imagery. Ultimately, this project would simulate a decade’s worth of PACE data over this sensitive region, offering a solution to the data scarcity issue and providing a resource for pre-training of advanced deep learning networks.
The project will then employ various neural network architectures to refine these models using real PACE imagery and in-situ water quality sensors to ensure continued model applicability to PACE imagery as it becomes available. Additionally, we plan to explore the capabilities of the Python Top Of Atmosphere Simulation Tool (PyTOAST) for generating "virtual swaths" that match the spatial resolution and format of real PACE data, streamlining the data preparation process and creating a valuable resource for researchers in the domain.
Significance
The objective of AquaWatch Australia is to pioneer a comprehensive 'weather service' for water quality through the establishment a seamless ground-to-space national water quality monitoring system. This initiative is designed to bolster water management practices through the provision of precise data and predictive insights, facilitating informed decision-making and effective policy development. Some key objectives include:
1. Ensuring Drinking Water Safety through HAB detection and forecasting; analyzing dissolved and particulate matter, observing water quality trends, and identifying pathogens.
2. Recreational Water Quality Management: Monitoring of recreational waters, aiming to maintain and enhance their aesthetic and ecological value.
3. Coastal and Aquaculture Ecosystem Monitoring: The initiative places a significant emphasis on coastal waters and aquaculture farms, vital for both the economy and biodiversity.
4. Addressing the Impact of Agriculture and Mining: AquaWatch recognizes the critical need to monitor and manage the environmental footprint of agriculture and mining, industries integral to Australia's economy but potential sources of ecological stress.
Strategic Approach to Extreme Events
In addition to routine monitoring, a pivotal aspect of AquaWatch is the capacity to forecast and mitigate the impacts of extreme events such as bushfires, floods, and droughts. By providing advanced warnings and actionable data, AquaWatch empowers stakeholders to implement effective planning and response strategies, minimizing environmental and public health impacts.
Why PACE
The novel hyperspectral capability and broad spectral range PACE's Ocean Color Instrument (OCI) allow for the detection and quantification of various water quality parameters with unprecedented detail and accuracy and provides an incredibly rich set of training data for deep learning networks. This also provides an opportunity to further fine tune models using in-situ hyperspectral and water quality data from CSIRO's HydraSpectra array.End User(s)
Key stakeholders include:- Industries involved with water quality management.
- Regional state governments and State-based water quality management agencies.
- Australian Commonwealth agencies with national and international outreach in water quality policy.
- International aid agencies involved in improving water quality outcomes in the Pacific region.
- Companies and agencies involved in the design, development and construction of sensor and satellite systems that have an interest in implementing EO systems to improve water quality monitoring.
SAT Partner(s)
Brian BarnesPublications
Baird, M. E., Cherukuru, N., Jones, et al. (2016). Remote-sensing reflectance and true colour produced by a coupled hydrodynamic, optical, sediment, biogeochemical model of the Great Barrier Reef, Australia: Comparison with satellite data. Environmental Modelling & Software, 78, 79–96. doi:10.1016/j.envsoft.2015.11.025
Steven, A. D. L., Baird, M. E., Brinkman, R., et al. (2019). eReefs: An operational information system for managing the Great Barrier Reef. Journal of Operational Oceanography, 12(sup2), S12–S28. doi:10.1080/1755876x.2019.1650589