Feature: The Earth System Grid: Climate Data for the Global Community
For the past two years, the Earth System Grid has helped scientists around the world study, analyze and predict all facets of the Earth's climate. This U.S. data grid provides almost 160 terabytes of climate-related simulation data to scientists worldwide through two portals.
Through the first portal, more than 3,000 scientists can find and download 130 terabytes of data hosted at four U.S. supercomputing centers: The National Center for Atmospheric Research; and Oak Ridge, Lawrence Berkeley, and Los Alamos National Laboratories. The data represents six years of U.S. earth system and climate modeling simulations.
"This is Community Climate System Model data, which has atmospheric, sea ice and ocean components," says Don Middleton from the National Center for Atmospheric Research. "Scientists run experiments using the data, predicting future climate based on changes in conditions, or simulating past climate and comparing to observed records."
The second portal, from the Program for Climate Model Diagnosis and Intercomparison at Lawrence Livermore National Laboratory, is the gateway to the data used for the fourth Intergovernmental Panel on Climate Change.
"We gathered simulations from 23 different climate models used by the international community for the 4th IPCC in 2003," says Dean Williams from Lawrence Livermore National Laboratory. "The Earth System Grid was a natural fit to disseminate the data."
The analysis of the 4th IPCC data is still ongoing; to date more than 200 papers have been published that describe climate change research performed using data downloaded from the Earth System Grid. This research ranges from the evolution of polar sea ice to the variability of soybean yields.
ESG-I, a prototype begun in the year 2000, showed how burgeoning grid computing technologies could be used to move and replicate large data sets. Today's Earth System Grid is the end stage of the ESG-II project, which was financed by the DOE's Scientific Discovery for Advanced Computing program. The collaboration has recently received funding from the second phase of that program to develop an Earth System Grid Center for Enabling Technologies.
"We're at the transition between classic climate models and earth system models that include more than just atmospheric and ocean components," explains Middleton. "It's a much more comprehensive simulation of our planet, and it's a global effort."
The Earth System Grid Center for Enabling Technologies, or ESG-CET, will move toward a gateway approach, with scientists using the two portals to access earth system data stored at centers around the world. ESG-CET architecture will support visualization or analysis tools that those same centers host, so that scientists don't have to download enormous data sets tying up their own resources.
"Right now we're hosting about 30 terabytes of data for the IPCC," says Williams. "The next IPCC, which will start in 3 years, will require at least 300 terabytes. It doesn't make sense anymore for everyone to send us their data."
The ESG-CET will need to be fully functional by 2011, when more than 1,000 scientists, politicians and others will participate in the 5th IPCC.
-Katie Yurkewicz, iSGTW