Blog icon

By Jane Nicholls 27 April 2022 7 min read

This satellite image of the mouth of the Clarence River at Yamba shows a large sediment plume. The plume extended some 30km into the Tasman Sea. Image: European Union. Modified data from Copernicus Sentinel-2, processed with the Sentinel Hub EO Browser.

At CSIRO’s Black Mountain campus in Canberra, three passionate CSIRO scientists have turned decades of satellite data into powerful knowledge. Their work includes assembling a rich data cube and analysing the remote sensing data within it to understand how water quality has changed over the past 20 years.

A project with the NSW Department of Planning and Environment (DPE) is filling in vast knowledge gaps about our waterways. This will help inform better decisions and mitigation strategies to preserve water quality and biodiversity.

First, build your data cube

“We started talking about a single problem: how do we help monitor the water quality of coastal waters over thousands of kilometres,” said Dr Nagur Cherukuru, Senior Research Scientist from CSIRO’s Oceans and Atmosphere.

CSIRO worked with DPE, the management agency for NSW coasts and estuarine water quality. They provided their field vessels, equipment and expertise on NSW ocean and estuarine aquatic ecosystems to conduct this research. Dr Cherukuru also worked with CSIRO colleagues Mr Matt Paget, technical manager for CSIRO’s Earth Analytics Science Innovation (EASI) platform, and Dr Eric Lehmann from the Remote Sensing and Image Integration team. The trio built on work by Geoscience Australia and CSIRO to build an Open Data Cube, which accesses petabytes of Earth Observation data from multiple satellites.

A wealth of expertise across CSIRO and NSW DPE is driving innovation in water quality monitoring

The masses of satellite data released by the EU and the US, along with open-source tools to better wrangle that data, and the rise of cloud computing which enables engineers to spin up powerful computer resources on demand, “has been a game changer”, said Mr Paget.

“It allows us to collate these satellite data series over our regions of interest," said Mr Paget. "We build analytical tools to drill down into the stack of satellite data over those two decades, organising it into a metaphorical cube of data. Then we do our deep-stack space and time analyses.”

Extensive Earth observation satellite data

For the NSW DPE project and using the cloud computing resources, it took them a little over a month to create the cube structure of 20 years of Earth Observation (EO) data covering an area over the NSW shelf and the marine estate.

“That includes rainfall data for 20 years, land use and land cover data, and for the water, how much phytoplankton, sediment and dissolved substances are there,” said Dr Cherukuru.

“It’s not images from one satellite – we brought multiple sensors into this data cube. Pick a day and there will be 10 parameters for that date, including wind speed and wind direction.”

“In addition, we are building a mosaic of collective understanding of how these systems work,” adds Mr Paget. “This research is part of CSIRO’s AquaWatch Mission in development and will help us further refine our understanding of the complexity of the processes.”

Furthermore, scientists can interrogate the data cube from many angles.

“We see we have 10mg of sediment and we ask, where did it come from?” said Dr Cherukuru. “Was it pushed by wind, or did it come from the land next to the river? The analysis enables us to understand what happened, when, and why.”

Earth Observation images show the mouth of the Clarence River before the flood on February 9 (left) and after, on March 3, with a sediment plume calculated to be 10km extending well off the coast into the Tasman Sea. Image: European Union. Modified data from Copernicus Sentinel-2, processed with the Sentinel Hub EO Browser.

Putting the data lessons into action

The historical satellite data – which can be kept up to date in the data cube in near real-time – provides a graphic illustration of how climate, land and river use impact coastal waters. The analysis for the NSW DPE shows a marked decline in light availability for photosynthesis off the NSW coast between 2002 and 2022.

“Coastal water quality impacts the benthic habitats, such as seagrass, kelp, sponges and coral,” said Dr Cherukuru.

“A sediment plume cuts out the light that is available for photosynthesis, the plant growth is less and there is less food available for fish. We want to know how and where the water quality degrades.”

“For NSW, the government is interested in understanding variability in ocean water quality across their marine jurisdiction and adjacent shelf. They also want to know how large the river plumes are when they occur, and how they might impact on biodiversity assets.”

Using remote sensors and data cube technology increases the efficiency of water quality monitoring over larger areas

The 2022 NSW floods caused significant sediment plumes. Dr Cherukuru compared two images side-by-side from the Copernicus Sentinel-2 satellite constellation. They showed the mouth of Clarence River at Yamba on February 9, and on March 3. In the first, he calculated the river mouth as 2km wide. In the second, it was 10km wide.

“You can see a large plume jetting out from the land and going about 30km out into the Tasman Sea,” he said.

Shining the light on water quality

Dr Lehmann has visualised the data to show how euphotic depth – the light availability in the water column in the coastal waters off NSW – has changed over two decades. The analysis visualises a clear degradation trend in coastal water quality.

“The amount of light available in the water column has reduced by 30cm every year,” said Dr Lehmann. “That means that either sediment or dissolved substances coming from the land is blocking the light, and there is no light available for photosynthesis. Close to the coast, there’s a much higher loss of water column available for primary production.”

The EO data cube delivers graphic evidence of the long-term degradation in water quality to environmental managers and industry. The scientists plan on doing further work to drill into the detail. They are also applying this research methodology for a project in Malaysia, where they analysed more than a decade of satellite data to link upstream land use to coastal water degradation.

“In Malaysia, we looked at the interactions between land indicators. This enabled us to link the water quality and land cover in each catchment,” added Dr Lehmann.

The resulting knowledge and tools are helping local managers develop water quality monitoring programs and better management and mitigation strategies.

Bringing in AI to dig deeper

To date, the team has used statistical data analysis to understand historical changes in the water quality.

“We’ve used physics, biology, chemistry and oceanography models. Now we are proposing adding Artificial Intelligence and Machine Learning modelling, to enable us to bring predictive capability,” said Dr Cherukuru.

The magic happens when the team connects the data dots. For example, connecting data around climate processes (rainfall, cyclones) with terrestrial conditions (bushfire or drought), and with what’s flowing down the river and into the ocean. Analysing these layers of connected data over time, with AI boosting the predictive power, will help them to give advance warning of extreme events. It will cover a range of scenarios to help them plan mitigation strategies.

“By understanding the connections between climate, land and ocean, we are able to help ecosystem managers,” said Dr Cherukuru.

“It can feed into protecting seagrass and rocky reefs, better management of coastal ecosystems, and reducing the risk for the blue economy and related industries. This is very useful for the aquaculture industry. For example, we can say, ‘The last time all these things happened, the water quality degraded for 10 days’, so we can give industry an indication of what they’re dealing with.”

“We have discovered a lot using oceanography. Coupled with AI, our knowledge will accelerate even further. It will pick through the data cube and tell us even more about the interactions between land, ocean and climate.”

Preliminary results of the analysis of the data cube – which accesses data from 2003 to 2021 – show a considerable loss of euphotic depth over time. This means there is less light available in the water column for photosynthesis, leading to less food available for fish. The scientists estimate that the amount of available light in the water column has reduced by 30cm each year. Image: CSIRO, Nagur Cherukuru, Matt Paget, Eric Lehmann/NSW Coastal outflows

With so much satellite data in their hands, said Dr Lehmann, “the sky’s the limit.”

This research was co-funded by CSIRO and the NSW Government under the Marine Estate Management Strategy. The ten-year Strategy was developed by the NSW Marine Estate Management Authority to coordinate the management of the marine estate.

Contact us

Find out how we can help you and your business. Get in touch using the form below and our experts will get in contact soon!

CSIRO will handle your personal information in accordance with the Privacy Act 1988 (Cth) and our Privacy Policy.

First name must be filled in

Surname must be filled in

I am representing *

Please choose an option

Please provide a subject for the enquriy

0 / 100

We'll need to know what you want to contact us about so we can give you an answer

0 / 1900

You shouldn't be able to see this field. Please try again and leave the field blank.