As the USA shivers under yet another blast of cold air from the Arctic, a study presented at the American Geophysical Union meeting in December has shown that Arctic storms were 40 per cent more frequent between 2000 and 2010 than previously thought.
Critical to the study were calculations run on the Ohio Supercomputer Centre (OSC) which was used to generate complex visualisations from the Polar Weather Research and Forecasting model (Polar WRF). To generate the visualisations, the Arctic System Reanalysis (ASR) group used thousands of cores on OSC’s HP-Intel Xeon ‘Oakley Cluster’ and IBM 1350 Opteron ‘Glenn Cluster’.
The Polar WRF is a modification of the Weather Research and Forecasting model widely used by researchers and most US federal agencies created by the Polar Meteorology Group of the Byrd Polar Research Center at Ohio State.
The findings are important to researchers who want to get a clear picture of current weather patterns, and a better understanding of potential climate change in the future, explained Dr David Bromwich, professor of geography at The Ohio State University. The cyclone study, was presented at the American Geophysical Union, in a poster co-authored by Bromwich's colleagues Natalia Tilinina and Sergey Gulev of the Russian Academy of Sciences and Moscow State University.
Extreme Arctic cyclones are of special concern to climate scientists because they melt sea ice, 'When a cyclone goes over water, it mixes the water up. In the tropical latitudes, surface water is warm, and hurricanes churn cold water from the deep up to the surface. In the Arctic, it’s the exact opposite: there’s warmer water below, and the cyclone churns that warm water up to the surface, so the ice melts,’ Bromwich said.
Bromwich leads the Arctic System Reanalysis (ASR) collaboration, which uses statistics and algorithms to combine and re-examine diverse sources of historical weather information, such as satellite imagery, weather balloons, buoys and weather stations on the ground. ASR provides researchers with high-resolution information against which they can validate climate prediction tools.
The ASR group analysed 17 surface variables, 71 forecast surface variables, 13 forecast upper air variables, and three soil variables. The data accumulated for and generated by the model filled hundreds of terabytes of disk space on the center’s IBM Mass Storage System.
One global data set used for comparison was ERA-Interim, which is generated by the European Centre for Medium-Range Weather Forecasts. Focusing on ERA-Interim data for latitudes north of 55 degrees, Tilinina and Gulev identified more than 1,200 cyclones per year between 2000 and 2010. For the same time period, ASR data yielded more than 1,900 cyclones per year.
The Arctic-centered ASR appeared to catch smaller, shorter-lived cyclones that escaped detection in the larger, global data sets. The ASR data also provided more detail on the biggest cyclones, capturing the very beginning of the storms earlier and tracking their decay longer.
ASR is a collaborative effort involving Ohio State, the National Center for Atmospheric Research, the University of Illinois at Urbana-Champaign and the University of Colorado-Boulder. It is funded by the National Science Foundation as an International Polar Year project.
