CSIRO sets itself water quality mission

Water quality is getting harder to predict. At the same time, it’s becoming dramatically more important to more people in more ways.

That’s why it is the latest addition to the CSIRO’s list of eight core missions.

The CSIRO has set itself the goal of providing an accessible near-real-time monitoring and forecasting service – in much the same way the Bureau of Meteorology covers the weather – for key parts of Australia by 2026.

“The importance of water quality has been quite obvious in recent months with the mass fish kills,” says CSIRO AquaWatch Mission Lead Dr Alex Held. “Maybe this time that was an effect of too much water running off the landscape.”

Such a disaster doesn’t have to be a surprise. An early warning may help reduce the impact.

“We also have fish kills in dry years, with the likes of blue-green algae taking the oxygen out of the water,” he says. “Clearly, there’s real value in us developing a program to identify and warn of the arrival of such dead water across Australia.”

Advanced warning is also of real value when securing community drinking water supplies, livestock access and aquaculture industries.

“These people often say they really need to know what the water quality is going to be tomorrow or the day after so they can do something about it beforehand,” says Held.

Water induction pipes can be closed before the unwanted algae floods in, livestock can be herded to safer fields, and tuna and shellfish farms can act before pollutants arrive.

Moreton Bay. Spencer Gulf. Cockburn Sound. Lake Tuggeranong. Lake Hume. Fitzroy River. All contain key industries and facilities that will be the first to benefit from the new service.

“We’ve set ourselves some lofty goals,” Held says. “By 2026, we want our partners at five to 10 hotspots around Australia to have an app on their phone that can give them up-to-date water quality reports and some prediction of any change over the next few days. We want to offer the same to the general public across the whole country a few years after that.”

The need for clarity

Water is increasingly precious. Rainfall is getting more erratic and the stresses imposed upon environments by climate change are ratcheting-up stresses on water systems.

“It’s getting worse,” says Held. “About three billion people worldwide really don’t have good information about the water they need.”

And that’s not restricted to inland freshwater sources.

Urban sprawl and intensified runoff are creating issues with coastal water bodies.

“Imagine a fish farm somewhere on the coast,” Held says. “They can have red tides – toxic algal blooms – flowing by. It’d be fantastic if we can warn them two or three days in advance so they can harvest their fish or move the pens out of harm’s way.”

The CSIRO and AquaWatch foundation partner SmartSat CRC are establishing pilot projects to push the limits of modern technology to provide accurate, timely surveillance.

“The initial suite of cooperative projects brings together research partners, government and industry and covers integrated ground-to-space water quality monitoring systems,” says SmartSat CRC Chief Executive Officer Professor Andy Koronios. “This collaborative effort to harness the data gathered from satellites will play an essential role in safeguarding our future water supply and improving our natural environments.”

AquaWatch needs a host of satellite and ground-based sensors to provide the quality and scope of data needed for a forecasting service.

“We are bringing things together that have already been developed, but maybe also developing new ways of integrating them and modelling the data,” says Held. “There are quite a few interesting engineering and scientific challenges in bringing all these datasets together in a standardised way, and how to present them to farmers and policymakers in a meaningful format.”

Eyes in the sky

“We can do chemical analysis now, basically, from space,” Held explains. “It’s not just a bunch of pretty pictures anymore.”

The ever-increasing supply of orbital multispectral cameras opens up a host of new opportunities. These capture the full spectrum of light reflected from the landscape into each sensor pixel. The resulting distribution of colours can reveal what materials are there, and in what quantities, such as the concentration of chemicals in water.

“The level of greenness could show how much chlorophyll is there, for instance, and that could be an indicator of perhaps too much fertiliser leaching into the water,” Held says. “Or we can fingerprint the presence or absence of cyanobacteria – the toxic algae that can affect animals and humans.”

But water is a tough subject for hyperspectral analysis.

“In the visible spectrum of sunlight, there tends not to be many photons reflected back from water to a satellite,” Held says. “So you need sensors that are very, very sensitive.”

Part of AquaWatch’s role is identifying the minimum technical requirements to capture useful data. This will then guide the development of future cameras better suited to water monitoring.

“A lot of those satellites we have access to at the moment are from other countries and are not completely customised for water quality issues, especially inland water,” he says. “So, we need to design satellite systems that can do high spatial resolution – or show a lot more detail – than we get from existing ocean colour satellites.”

Water is not the sole subject of AquaWatch monitoring.

“These satellites will see everything going on around it, the land use, the industrial use, the cultural use,” he adds. “We can use this information to try and understand what’s causing water quality problems. Is it erosion? Is it something else?”

Eyes on the ground

Satellites can’t see – or sense – everything. So, an enormous array of land-based monitoring systems must be developed and deployed across Australia.

These will verify and train hyperspectral imaging systems, but also fill vital gaps in the data.

“Because we’re only planning to use optical satellites, they won’t be able to measure salinity,” says Held.

“That’s not an optically active chemical. So that’s why we need to combine land and space-based data to produce the most robust data.”

Such sensors already exist. But they must be made resilient enough to withstand Australia’s harsh environments while being easy to produce and deploy.

“The continental-scale of AquaWatch will require a mind-boggling number of sensors in Australia alone,” says Held.

One sensor, the CSIRO’s HydraSpectra, has already been designed in-house after researchers found existing sources too expensive for such a large-scale project.

“We can’t build them quickly enough at the moment,” says Held. “There’s a real opportunity here for an Australian industry partner who can take our design, improve on it, and manufacture it on a more economical scale.”

Once deployed, they’ll need to be maintained.

“If we were going to have thousands of these sensors around Australia, one model would be to provide basic training to regional communities and indigenous groups and get them to look after this instrumentation.”

Connectivity is also an issue.

A nationwide network of land-based sensors must deliver their data within 15 or 20 minutes to be helpful. And that needs low-power, affordable and reliable satellite uplink systems.

“It’s a massive challenge to bring these different data sets from these sensors into a single place where you can bring it to analyse it together,” says Held. “So we’ll be testing different technologies for every step of every process.”

More from Cosmos Weekly