Climate change has been at the top of the news agenda this week, in part because of the attention around a massive public protest in New York City. The People’s Climate March, which drew an estimated 400,000 people, is calling attention to a number of climate-related initiatives in advance of a climate summit among members of the United Nations.
Leaving aside the political arguments around climate change, it is one thing to understand the potential causes, but another entirely to specifically target more isolated geographical and event-related issues.At the end of the day, the most powerful weapon to combat climate change (and the speculation and arguments around whether it’s happening) is certainly solid data. The problem, of course, is that large-scale modeling of complex atmospheric and earth systems data takes a lot of compute power–and increasingly sophisticated modeling and simulation tools.
Supercomputers are essential to providing actionable information to both the public and groups that have the capability to enact change through policy and process. Most systems advanced enough to handle these applications are occupied with a range of scientific applications. This means that having a dedicated national resource devoted to advancing our understanding of climate changes is critical. Among the most notable supercomputers in this category is the Yellowstone system at the National Center for Atmospheric Research (NCAR), which is the world’s foremost machine dedicated exclusively to understanding the causes and impacts of climate change worldwide.
As noted previously, the scale and complexity of weather and climate models is rivaled only by a select few other massive simulations. To prepare for this, the 1.5 petaflop system, which ranked at #29 on the most recent June 2014 Top 500 list, has been outfitted with the necessary memory, compute, and storage to handle advanced climate studies. Yellowstone is an IBM iDataPlex cluster comprised of over 72,576 cores (Xeon E5-2670) and a total of 144.6 TB of memory. This is complemented by the rest of the Yellowstone environment, which includes the data analysis Geyser machine and Caldera, the center’s visualization cluster, as well as the GPFS-based 16 PB GLADE central disk resource and the center’s high performance storage system.
The system is housed in a 153,000 square foot facility at the NCAR-Wyoming Supercomputing Center in Cheyenne and hit production readiness in summer of 2012. While climate modeling and related applications are at the heart of the approximately $30 million machine’s initial purpose, there are a number of other applications running, including those devoted to space weather, aviation safety, drought, wildfires, and other specific phenomena.
According to the NSF’s datasheet on the climate change areas of study, the following areas in atmospheric science alone highlight the focus of applications for the system:
Other target applications include helping researchers understand tornadoes and other dangerous weather events, analyze how geomagnetic storms affect communications systems and the power grid, target potential water issues as related to drought and flood conditions, analyze earthquakes and tsunamis and more.
The following are the broader societal impacts of this research according to the NSF: