Understanding ocean chemistry through the power of Cloud Computing

Ocean temperatures and chemistry are changing dramatically and posing a risk to certain life forms, including shellfish such as oysters grown and harvested in Washington state. Microsoft Research teamed up with University of Washington scientists to take data from a complex modeling system run on supercomputers and bring it to the cloud. Soon, it will be widely available to help predict growing conditions for the shellfish industry and may help other industries adapt to ocean changes.

Climate change driven fluctuations in ocean chemistry have been linked to die-offs of baby oysters along the Northeast Pacific Ocean coast. The gradual change known as ocean acidification is making it difficult for shellfish, corals, sea urchin, and other creatures to form their calcium-based shells or other structures. The Northwest region’s thriving oyster hatcheries were struck by high mortality rates. While climate is changing worldwide, the Northwest is particularly vulnerable to ocean acidification because of the upwelling of colder more acidic water into the bays and estuaries of Washington, Oregon and British Columbia.

Modeling complex currents and chemistry

The Washington state legislature asked the University of Washington to study ocean acidification through the Washington Ocean Acidification Center (WOAC). Microsoft Research joined with Parker MacCready, professor of physical oceanography, to bring complex information from a variety of data sources into a system called LiveOcean, which would provide a model of currents and chemistry and predict a few days into the future.

Just like a numerical weather forecast model, LiveOcean will soon provide a forecast that predicts the acidity of water in a specific bay or other coast region three to seven days in advance. Bill Dewey, Director of Public Affairs for Taylor Shellfish, needs that prediction to know when and where it is safe to plant oyster larvae and raise juvenile oysters.

Estimates are that the West Coast oyster industry generates 3,000 jobs and makes an annual economic impact of about $207 million.

More than 30 percent of Puget Sound’s marine species are vulnerable to ocean acidification by virtue of their dependence on the mineral calcium carbonate to make shells, skeletons and other body parts.

A baby oyster uses carbonate ions in the water to make their first shell. If the water is too acidic, the baby oyster uses too much energy and dies. Taylor Shellfish has hatcheries for the baby oysters and separate “planting” beds where young oysters are moved to continue growing. Forecasts would help Taylor Shellfish know the safest planting locations.

Dewey remembers hearing the phrase “ocean acidification” for the first time in 2007, with other industry leaders at a meeting, and knows the extreme challenges from ocean acidification. Besides working for Taylor Shellfish, Dewey also farms his own shellfish beds and has a personal stake in the challenge.

Making a model open to everyone

MacCready is the scientist leading the LiveOcean team who stepped up to help the shellfish industry. But he also looks forward to other scientists and industries drawing their own insights from this same model. The LiveOcean design incorporates the open use by others.
MacCready was a visiting scientist who spent four months at Microsoft Research in Redmond, Washington collaborating with scientists there, including Rob Fatland. Fatland is now the Director of Cloud and Data Solutions at the University of Washington.

MacCready’s team used Azure tools to draw data from a large model run on a high-performance computing cluster. The model is known as the Regional Ocean Modeling System (ROMS). They push data out of Azure with Python then write scripts for websites. Using the cloud is “the way of the future” he said, for complex systems like this one. “It gives the ability to create and use different resources without having to go out and buy hardware yourself.”

MacCready and Microsoft researchers built a forecast system open to everyone. Using the cloud, a non-scientist will be able to reach into ROMS forecast data and pull out information through a smart phone, laptop, tablet, or other devices. The Azure component uses Python and the Django web framework to provide these forecasts in an easy-to-consume format. To produce the forecasts, the Live Ocean system relies on other sources: US Geological Survey data for river flow, atmospheric forecasts, and another ocean model called HYCOM.

One crucial element of LiveOcean is the careful validation of its results. MacCready and others have spent years validating the modeling system by direct observation from physical instruments paired to predictions.

Policy and public understanding

Predictions will be vital for both policy-makers and scientists, according to Dewey. The impact of discoveries from the model could be vast. It will show a bigger and better picture to every part of society as decisions loom due to challenges from climate change. Beyond the industries and legislators, MacCready also sees LiveOcean as providing a new and important window on the coastal ocean globally as scientists and others adopt the model and begin to use it.
Dewey sees the model waking people up to changes. “The ocean acidification issue has really come to light here on the Pacific, where we have these upwelling events. But we are the tip of the spear for this. It has woken up the industry across the country,” he said.