Preparing the Pacific

As the world prepares for the effects of climate change, many smaller nations are bracing for the impact. With extreme weather patterns set to increase, these regions must prepare now.

Dr Michelle McCrystall’s career has been dedicated to uncovering the science of our changing climate, from polar climate research to tropical cyclones. Her latest project is an online accessible resource called EMPIRIC_AI, designed to help regions most affected by the increasing extreme weather events expected as a result.

Previously based in Canada, Michelle’s primary area of expertise is large-scale climate dynamics and atmospheric processes, seeking to understand the impacts on the climate of the Arctic and, more recently, forecasting the region’s future climate. 

Her work has taken her around the world. “I have been to Svalbard, Norway, and I spent a lot of time in Canada, including Hudson Bay, which is said to be the polar bear capital of Canada – although I have never seen one.”

Since she moved to New Zealand, the focus of her work has shifted to the South Pacific, specifically assessing the impact of tropical cyclones on health infrastructure and, therefore, the health of the populations of these communities. 

She has been concentrating on the impact of climate change on the region and developing the EMPIRIC_AI – the AI-enabled ensemble projection of cyclone risk for Pacific Island Countries and Territories health infrastructure – a tool for affected areas to inform their future planning using artificial intelligence (AI).

“We know climate change is happening, and we know it will have an impact. So, moving forward as a scientist, I believe we need to bring the science to the people who can use this information to plan ahead."

The goal is to create a free network and online platform to inform the future planning of health infrastructure for these countries as an increase in the frequency and severity of tropical cyclones in the region is anticipated.

“It’s very different to what I’ve done before. Until now, my research has concentrated on understanding science as a phenomenon. However, this project uses scientific processes to identify the local impact for future generations."

The project was initiated when Chris Hovart was approached by a group of surgeons and public health officials to assist them in making climate predictions for hospitals.

At the time, Chris was a senior lecturer at the University of Auckland and the original Principal Investigator (PI) on the project. Now at Brown University, Chris is a Co-Investigator on the EMPIRIC_AI project.

“Climate models do not represent the Pacific islands, and running long simulations at the resolution required to ‘see’ them is very expensive. So, we need to devise clever ways to understand the realities of how large-scale climate patterns affect local extreme weather and the statistics of tropical cyclones. We do this by using a complex model we have developed for creating tropical cyclone trajectories and running this for long periods to train a model for cyclone impacts,” Chris says.

The cross-disciplinary nature of the project presented a new opportunity for Michelle that piqued her interest. This factor was a significant element of the grant application, which involved collaborating with interested people and vested stakeholders in climate change.

“We are working with people in fields that do not immediately come to mind in the context of climate change. But by bringing the information to different industries, organisations, and groups, we are bridging a gap in communication and collaboration.”

The key stakeholders and collaborators Michelle identifies include the Pacific Community – a multi-diverse organisation, various health facilities across the islands, doctors at Harvard, Melbourne, and Cairns Universities who have all worked as doctors and clinicians in New Caledonia, Fiji, and Vanuatu. More recently, the Fijian Ministry of Health has indicated interest in the project.

She also says the Fijian Ministry of Health has said it is willing to provide data on climate events that significantly impact health and health infrastructure so that projections are tailored to include what is most impactful and important for those countries.

Significant cleanup will always occur after a cyclone along the coastline. However, Michelle explains there are connected events that occur off the back of the cyclones that can be overlooked.

“Extreme precipitation is one example. In 2019, a storm hit Vila Central in Vanuatu that caused extreme precipitation and flooded the entire hospital, and they are still rebuilding.”

Contamination of drinking water is another big problem. While groundwater was not an initial focus for the team, it is an example of how variable the research can be, depending on the needs of that community. This is where EMPIRIC_AI will become a valuable tool for these regions. The AI programme is tasked with speeding up data processing and determining various metrics to predict future climate events.

Pre-AI, Michelle explains a model might take anywhere from one to two days to produce feedback, but now, results are available in hours, and the plan is to make that even faster again.

“It’s not a Magic 8 ball or a looking glass that tells you an exact answer. More accurately, it provides a range of possibilities based on the collected data,” says Michelle.

While this project is very much centred around informing policy and healthcare, the potential applications of this resource are far-reaching.

“Once the model is developed, it should easily adapt to anywhere else impacted by tropical cyclones. For example, parts of the Caribbean that are as heavily affected by tropical cyclones could use this model.”

On an even larger scale, the project has changed the trajectory of Michelle’s research plans. The broad applicability of the model makes it possible to look at other climate phenomena, like extratropical cyclones, or projects further removed, like building a wind farm. With a deadline looming at the end of the year, the team has made impressive progress since the project’s inception in January.

The EMPIRIC_AI will be up and running within the next couple of years and promises much hope to regions most impacted by climate change, with potential applications that extend beyond the Pacific.

“Often, the research I work on is science for scientists, with the hope that it gets disseminated. But this project has real, tangible benefits that will positively impact people and communities, which is really fulfilling.”