Scientists want to use artificial intelligence to save Maine's coast

Oct. 13—Forecasting isn't just for weather. A new center at Bigelow Laboratory is using cutting-edge artificial intelligence algorithms to forecast ocean activity, from toxic algal blooms to right whale migration, with the hopes of benefitting both coastal industries and the environment.

"There's a big demand for forecasting. People are expecting forecasts of all different kinds now, from COVID forecasts to political forecasts," said Nick Record, a senior research scientist at Bigelow Laboratory for Ocean Sciences in East Boothbay. "We're trying to tap into this societal need and demand for forecasts and apply it to ocean systems that we live in and rely on."

The ability to accurately forecast complex ocean dynamics alone, such as temperature and salinity, is useful for the industries that use the coastline and the scientists that study it. With artificial intelligence, though, these forecasts will be constantly improving in accuracy even as the climate changes — and, with it, Maine's ability to adapt to the changing coastline will improve as well.

Record runs the new Tandy Center for Ocean Forecasting, which launched in the summer of 2021. He has been leading forecasting projects for years, like Ecocaster, which uses predictive modeling for a range of phenomena, from jellyfish populations to vehicular moose crashes. The new forecasting center combines a number of projects throughout Bigelow Laboratory with the overarching goal of more accurately predicting Maine's coastline.

Take, for example, those smelly, toxic algal blooms that crop up along the coast and have only been exacerbated in recent years by the Gulf of Maine's warming waters. Not only can a toxic algal bloom ruin a day at the beach, but the Maine Department of Marine Resources also forces shellfish harvesters and growers to shut down until the toxic threat subsides in order to protect consumer health.

"It's hard for the industry to plan and prepare for those closures or the reopening," Record said.

Previously, scientists at the Maine DMR were only able to make rough predictions about forthcoming blooms after years of sampling to see which toxin levels would spike before a toxic algal bloom.

At his Center for Ocean Forecasting, though, Record is using the data collected by places like the Maine DMR along with other data that are already being collected — from atmospheric data to ocean temperatures — to more accurately predict when a toxic algal bloom might be on the horizon.

"Turning it into a model which is empirical in the sense that it's using data and not somebody's feelings, or maybe their experience was limited to a window without blooms," said Kohl Kanwit, director of the Bureau of Public Health at the Maine DMR. "The model can take a lot more information and be unbiased in its assessment."

Record said that his algorithm — the way that the computer crunches the numbers to determine whether or not a toxic algal bloom is coming — uses elements of "deep learning," which means that it learns over time from its mistakes and improves.

"On this time scale, you know right away if you're right or you're wrong," Record said. "Each time you do that the algorithm gets a little bit better."

During the summer of 2021, Record worked with growers and harvesters in the industry to develop this tool for predicting toxic algal blooms, which he sees as an essential element to creating a good forecasting algorithm.

"What's really useful to people is something different from what a scientist thinks is useful," Record said. "One of the things that I'm really trying hard to figure out is what are the forecasts that people need."

The results have been promising. Record said he can now predict about a week ahead of time what the odds are that toxin levels at a specific site will hit the level that requires closure.

"We predicted all closures correctly," Record said. "We feel like we have the confidence now to open it up more broadly. The plan is for next year to involve the Maine coastal aquaculture industry and get feedback."

While algorithms will never replace the need for sampling and testing entirely, forecasting does help the DMR allocate its resources to the sites that are predicted to need them the most, Kanwit said. That allows industry to better adapt to the changing conditions.

"It's really difficult for people to adapt to something when a big closure happens overnight," Kanwit said. "This provides a lot more warning and the ability to keep things consistent."

The forecasting center doesn't just look at algae. Forecasting the movement of endangered right whales, which are so often in conflict with coastal industry because of mortality related to lobstering and shipping, is also a goal for the Center, Record said.

As the Tandy Center for Ocean Forecasting continues its work, Record aims to create algorithms that will help paint a more complete picture of the ocean as the climate changes and for future decisions, like how offshore wind power might affect species as they adjust to warming waters.

"We know where species are now but we've done some forecasting about what the ocean will look like in 30 years," Record said. "There's that kind of longer term perspective as well. When you think about forecasting, the sky is the limit."