Supercomputers are helping scientists predict the future of climate change better than ever before
Supercomputers, those mega-computers located at national laboratory sites that can process calculations in a matter of nanoseconds, have long been on the front lines of combatting climate change. Researchers have used them to run models and simulations to better predict future weather as well as crunch data to better reveal the impacts of climate change.
Now, the latest crop of supercomputers -- the fastest yet -- are enabling scientists to make the best models to date. Researchers have been able to create high resolution simulations of global climate models that are far more accurate and revealing greater details.
To test the accuracy of the models, researchers at the Lawrence Berkeley National Laboratory chose the years 1979 to 2005 to run analyses at 25 km, 100 km, and 200 km resolutions. The researchers found that not only were the simulations much closer to the actual observed weather events, but the models were far better at reproducing intense storms, such as hurricanes and cyclones.
The simulations, which produced 100 terabytes of data, or 100,000 gigabytes, were notably superior to previous models in mountainous regions because earlier computers would just average the altitude in a grid. The high resolution version had a more accurate representation of the terrain, which led to more precise simulations of snow and rain.
The newer model also produced more storms and hurricanes, which was truer to the actual events. This is especially important for helping to predict changing storm patterns associated with climate change as scientists believe extreme storms will become more frequent.
The lead authors of the paper say the new models will help them to better illustrate how climate change will bring greater rainfall, snow and storms to some places and drier conditions to others. They plan on running a future case scenario next to do just that.
Watch the video below to see a high resolution simulation and how it compares to an earlier version.