Scientists' plan for real-time virtual Earth would simulate...everything
Living Earth Simulator would draw on databases, supercomputing to create a global model of human activity
- By Henry Kenyon
- Jan 11, 2011
Modeling and simulation has become a vital scientific tool capable of predicting weather patterns or charting chemical reactions. But a new effort launched by an international team of researchers plans to simulate everything that happens on the planet.
The Living Earth Simulator is part of a European program designed to collect, aggregate and fuse data from a variety of sources, such as NASA’s Planetary Skin project, into one all-encompassing model. The simulator is part of the Future ICT Knowledge Accelerator program at the Swiss Federal Institute of Technology (ETH) in Zurich. The Future ICT effort seeks to team hundreds of the top scientists in Europe in a 10-year, 1 billion euro project to explore social life on Earth and everything it relates to. The effort is funded by the European Commission with the goal of becoming operational by 2022.
According to the BBC, the goal of the Living Earth Simulator is to further scientific understanding about events on the planet by looking at the human actions that affect societies and the environmental forces at work. Dr. Dirk Helbing, chairman of the Future ICT Project, told the BBC that many of today’s global problems, such as wars, social and economic instability and the spread of disease, are related to human behavior. But he noted that there is a fundamental lack of understanding about how societies and the economy work, adding that we know more about the early universe than our own planet.
Borrowing a term from physics, Helbing described the modeling projects as a “knowledge accelerator” that will bring different branches of science together to produce data much in the same way (metaphorically speaking) that a particle accelerator smashes atoms together to unlock scientific secrets. One of the key goals of the simulator is to model the behavior of entire economies or ecosystems in real time to detect and, hopefully, head off any crises.
MIT Technology Review compared the effort to a kind of Google Earth for society. But instead of using the map function to zoom into a home, a similar function could be applied to monetary transactions, health trends, global tourism patterns and carbon dioxide emissions. Helbing described this process as “reality mining.”
Besides modeling current events, the simulator will be used to predict potential future events such as financial crises and pandemics. With access to vast amounts of data, the model is also intended as a tool for finding solutions to such problems. According to Technology Review, Helbing also wants to establish situation rooms that will allow world political and business leaders to view and manage crises as they occur.
Future ICT’s Living Earth Simulator will pull data from some of the world’s largest supercomputers. Some of the machines that will supply computing and modeling muscle include ETH Zurich’s Brutus supercomputing cluster and supercomputers at the Los Alamos National Laboratory and the Brookings Institution in the United States. Science Daily reported that European researchers involved with the simulator are already running separate simulation programs, studying the travel activities of all of Switzerland’s population or the origins of international conflicts. All of these massive databases and powerful machines could be tapped to support the Living Earth Simulator.
Making sense of all of the data will be a challenge. Helbing told Daily Tech that his team will feed massive amounts of data to various global supercomputer clusters to cover all of the daily activity on the planet. He said that much of this data already exists and that his team is currently using more than 70 online information sources, such as Google Maps and Wikipedia.
After the data is collected and integrated, the research team’s computer and social scientists and engineers will build an architecture to convert the data into a real-time model of planetary activity. Although supercomputers will help make sense of some of the information, Helbing told Daily Tech that researchers will use Semantic Web technology to encode data, which allows the simulator to better understand what it is reading. Some of the technology necessary for the project, especially the analysis computing elements, will not become available in the coming decade, he said.
Although the simulator will follow human behavior, all personal data will be scrubbed out of the model. An ethics committee and targeted research will be put in place to ensure that personal data is not misused. The goal of the process is to identify statistical interconnections when many people interact, but not to track or predict individual behavior, Science Daily reported.