By Dan Reed, Scalable and Multicore Computing Strategist, Microsoft Corporation

Climate Change, Computing and Public Policy

We know the Earth’s climate has changed during the planet’s history, due to the complex interplay of the oceans, land masses and atmosphere, the solar flux and the biosphere. Recently, the U.S. Climate Change Science Program and the Intergovernmental Panel on Climate Change (IPCC)¹ concluded that climate change will accelerate rapidly during the 21st century unless there are dramatic reductions in greenhouse emissions.

As a species, we now face true life and death questions — the potential effects of human activities and natural processes on our planet’s ecosystem. High-performance computing (HPC) tools and technologies provide one of the most promising options for gaining that understanding by allowing us to develop and test complex models, validating them against observational data and testing hypotheses related to practices and policies.

The issue is not simply climate change but determining the local and regional effects, such as weather changes (e.g., precipitation levels), storm surges and flooding in coastal areas and hurricane frequency and severity. Changes in precipitation patterns affect agriculture and commerce, rises in sea levels create coastal flooding due to storm surge, and long-range trends shift the frequency and severity of local weather. Local and state agencies need information to formulate public policies, such as flood plain insurance rates, building codes and evacuation plans.

As a specific example, consider the potential effects of rising sea levels on coastal areas due to melting ice. Not only do sea levels rise, inundating low-lying regions, they can exacerbate the effects of storm surge — the coastal waves experienced during storms. Policy makers and regional planners need to understand what construction zoning and flood plain insurance rates are appropriate.

Answering such questions is possible by combining terrain elevation data, typically obtained by satellite imaging and LIDAR (light detection and ranging), with models of ocean currents, atmospheric patterns and coastal geography. These models are complex and require weeks to months of calculation on high-performance computing systems to produce the data needed for regional and coastal planning. However, the benefits are enormous — protecting environmentally sensitive coastal regions, reducing property loss and saving lives during severe weather.

Computing: The Third Pillar of Scientific Discovery

In 2005, I was privileged to chair the computational science subcommittee of the U.S. President’s Information Technology Advisory Committee (PITAC), which examined the relationship between computing and scientific discovery. In our report, Computational Science: Ensuring America’s Competitiveness², we noted that

Computational science is now indispensable to the solution of complex problems in every sector, from traditional science and engineering domains to such key areas as national security, homeland security, and public health. Advances in computing and connectivity make it possible to develop computational models and capture and analyze unprecedented amounts of experimental and observational data to address problems previously deemed intractable or beyond imagination.

Computational science now constitutes the third pillar of the scientific enterprise, a peer alongside theory and physical experimentation. This is especially important in a field such as climate change studies, where the models are complex — multidisciplinary and multivariate — and one cannot conduct parametric experiments at planetary scale. Fortunately, the same technology used to build desktop and laptop computers now enables us to evaluate these complex models.

In the late 1970s and the 1980s, HPC was defined by exotic supercomputers — high-speed, expensive systems designed for a select class of experts and used for national defense and national security tasks. By analogy, a supercomputer is a massive earth moving machine, with expensive support systems and highly trained operators.

With the birth of the PC, a new approach to HPC emerged that dramatically lowered the cost of HPC and opened access to researchers across the country. This new model interconnects thousands or even tens of thousands of desktop-class computers to perform complex calculations in parallel. Much as one partitions a complex construction task across a group of cooperating construction workers with shovels, commodity supercomputing divides complex tasks across large numbers of cooperating PCs. Microsoft’s Windows HPC Server ³ exemplifies this capability, allowing researchers to create commodity supercomputers based on industry-standard PCs that can tackle complex business, scientific and societal problems.

High-Performance Computing: Enabling Climate Change Insight

Why is this commodity approach to high-performance computing especially critical to climate change studies? First, one must simulate hundreds to thousands of Earth years to validate models and to assess long-term consequences. This is practical only if one can simulate a year of climate in at most a few hours of elapsed time at reasonable cost. Each of these simulations must be of sufficient fidelity (i.e., temporal and spatial resolution) to capture salient features. Today, for example, most climate models that are run for several hundred to several thousand simulated years do not explicitly resolve important regional features like hurricanes. These are large-scale, capability computing problems (i.e., ones requiring the most powerful computing systems).

Second, to understand the effects of environmental changes and to validate climate models, one must conduct parameter studies (e.g., to assess sensitivity to different conditions such as the rate of CO2 emissions or changes in the planet’s albedo — its reflectivity and solar energy absorption). Each of these studies involves hundreds to thousands of individual simulations. This is only practical if each simulation in the ensemble takes a modest amount of time. These are large-scale, capacity computing problems (i.e., ones requiring ongoing access to multiple, large-scale computing systems).

Third, understanding the sensitivity of physical and biogeochemical processes to social, behavioral and economic policies requires evaluation of statistical ensembles and many model variants. These are hypothesis-driven computational scenarios that are only possible after the physical and biogeochemical processes are understood, requiring additional capacity and capability computing.

This is a daunting problem — developing, validating and evaluating multidisciplinary climate models in time to provide the necessary answers to critical questions:

How many simulation scenarios are necessary (minimally and optimally)

What model elements are needed for each scenario?

What temporal and spatial resolution, along with physical models, is affordable?

What are the errors and uncertainties in model predictions?

When must research end and production simulation begin to produce policy guidance?

Underlying these questions is the need for powerful computers to model climate change at regional and local scales and to support the sophisticated and computationally expensive algorithms needed to represent the complexities of both natural and human effects. These same computers and storage systems manage the tsunami of observational data now being captured via a new generation of environmental sensors. This allows climate modelers to couple high-resolution Earth system models with assimilated satellite and other data, supported by large data archives and intelligent data mining and management. This fusion of sensor data with complex models is large-scale computational science in its clearest and most compelling form, enabled by standard hardware and software.

Today, we face both great opportunities and great challenges. Computational science, enabled by inexpensive but powerful computing clusters, truly is the "third pillar" of the scientific process that can help ensure the health of our planet.

About the Author

Daniel A. Reed is Director of Scalable and Multicore Computing Strategy at Microsoft. Previously, he was the Chancellor’s Eminent Professor at the University of North Carolina at Chapel Hill, as well as the Director of the Renaissance Computing Institute (RENCI), which explored the interactions of computing technology with the sciences, arts and humanities. He is a former Director of the National Center for Supercomputing Applications (NCSA) at the University of Illinois at Urbana-Champaign, where he also led National Computational Science Alliance, a consortium of roughly fifty academic institutions and national laboratories to develop next-generation software infrastructure of scientific computing. He was also one of the principal investigators and chief architect for the NSF TeraGrid.

Dr. Reed is a member of President Bush’s Council of Advisors on Science and Technology (PCAST) and a former member of the President’s Information Technology Advisory Committee (PITAC). He recently chaired a review of the federal networking and IT research (NITRD) portfolio, and he is chair of the board of directors of the Computing Research Association (CRA), which represents the research interests of universities, government laboratories and industry. He received his PhD in computer science in 1983 from Purdue University.