Using Big Data to understand our oceans
The ocean is changing at an alarming rate and this may affect marine life in ways that scientists don’t fully understand yet. That’s why Willi Rath and the Ocean Dynamics team at GEOMAR are trying to harness the increasing amounts of data on marine organisms that are being generated.
Oceans cover 71 percent of the Earth’s surface and play a critical role in regulating the global climate. As the planet has warmed due to human-caused climate change, the oceans have been feeling the heat. They absorb one-third of the carbon dioxide produced from human activities and more than 90 percent of the heat trapped by increased greenhouse gases in the atmosphere. As the ocean changes, there may be direct implications for marine life in these vast waters that are sensitive to even slight changes in the environment.
For instance, in the future, when fish larvae hatch in the Bay of Biscay on the western coast of France, they might grow more vigorously and be swept by ocean currents into a direction that’s opposite from those of previous generations. This may also have a rippling effect on animals higher up in the food chain that depend on them for sustenance. These are only some of the many effects that changes in the ocean could have on the life cycle of marine organisms and their ecosystems.
To improve our understanding of the intricate forces that drive the ocean circulation, the Ocean Dynamics team at the GEOMAR Helmholtz Centre for Ocean Research in Kiel develops and runs complex numerical models to simulate ocean currents. With this, the team can potentially predict the effects of climate change on the ocean system and how marine organisms will adapt over time. “We’re currently working with data that comes from a collaboration with marine biologists who are interested in how fish populations sustain themselves in the Mediterranean Sea,” says Dr. Willi Rath, a data scientist who has been working on the project for several years.
The role of data analysis in better understanding our oceans
Fish eggs and larvae are difficult to follow because of their minuscule size and the many species that lay eggs in large numbers. “Of course, there are people who track individual fish larva and try to follow them during the first couple of months of their lives to see how they grow, where they go and how they swim, etc.,” he explains over the phone. “But in a model or a simulation, it’s much easier to generate a lot of data and statistics on it. I don’t think it would be feasible to track 15 million fish larvae for several months in a real ocean.”
"It is difficult to track fish eggs and larvae directly because they are extremely small and the many species lay countless eggs. With a model or simulation, it's much easier to generate corresponding amounts of data."
Willi Rath, Data Scientist
For scientists, performing a computer simulation of the journey of a fish larva is undoubtedly helpful in understanding where it could drift to over time. However, when there are 15 million journeys to track, things can get tricky. “Usually, it’s difficult to foresee what kind of details – or variables – we’re interested in upfront so we can’t simply reduce the amount of data being generated, because we potentially need all the information we can get from the simulation,” says Rath, who helped to develop an ocean analysis simulation tool called OceanParcels, which is available to other scientists around the world.
The gap between the amount of data generated and effective methods of data analysis
While the “supercomputers” that Rath and his team are using can perform numerical experiments and generate high volumes of simulated data on the changing ocean circulation and climate system, the problem the researchers face is that the methods of data analysis and visualization haven’t caught up yet. “We’re basically overwhelmed by the amount of data we can produce at the moment,” he says. “There’s an increasing gap between the amount of data that we can produce and our ability to visualize it. And this is something that’s probably inhibiting scientific progress a lot more than we think.”
Ultimately, developing a technique to visualize large sets of data could help scientists not only improve the method of predicting how changing oceans will impact different marine populations, but also, more broadly, understand the complexity of other large data sets related to predicting the future effects of climate change. “Only through the process of analyzing the data and trying to understanding what’s happening to a single larva, or animal, will we be able to understand what needs to changed in the bigger picture if the ocean circulation changes,” explains Rath. “Potentially, all the details matter.”
Text by Charmaine Li
