Supercomputing and climate research: high resolution long-time simulations to improve our understanding

Have you ever watched a newscast and the on-air meteorologist mentions “according to our weather models, our best chances for rain will occur between the hours of 6 and 9 pm”? Have you wondered what exactly are those models they’re talking about?

A weather model is a series of equations that take a look at an initial state of the atmosphere (such as the temperature, pressure, and humidity) and look ahead to see how the conditions will change over time. As early as the 1920’s, weather forecasts have been produced using this method. It wasn’t until the 1950’s that the era of numerical weather prediction began using computers. Since then, the use and improvement of these models has been imperative to successful weather forecasts, as well as climate simulations.

Different from weather forecast models, climate models look at the interactions between the atmosphere, oceans, land surface and ice. By looking at the interactions between these spheres through the past, scientists are able to project temperature change that results from concentrations of various atmospheric gases, like carbon dioxide. Recently, the National Center for Atmospheric Research (NCAR), an organization under the parent University Corporation for Atmospheric Research (UCAR), opened the Wyoming Supercomputing System.

The building housing the NCAR-Wyoming Supercomputing Center in Cheyenne, Wyoming, officially opened on October 15, 2012. (©UCAR. Photo by Carlye Calvin.)

This center is important to not only NCAR, but the atmospheric science community as it attempts to provide the computing capabilities for improvement of tornadoes, hurricanes, earthquakes, droughts and climate simulations.  Models are limited by the computers they run on.  The launching of this new system allows scientists to do what they couldn’t before.

The computers housed at NCAR-Supercomputing Wyoming Center (NSWC) allow for high resolution modeling.   High resolution means that the difference between grid points (the physical points where the equations are solved) is smaller.  High resolutions allow for small scale weather phenomenon, such as tornadoes, to be resolved as well as run long simulations in time.  Resolving smaller scale weather phenomenon means that the features of the weather system are able to been in output.

Earth’s climate system. This image depicts a single month from a simulation of the 20th century by the NCAR-based Community Climate System Model (version 4). This image captures wind directions, ocean surface temperatures, and sea ice concentrations. The arrows show just how close together the model grid points reside (©UCAR, image courtesy Gary Strand, NCAR).

Think of it this way: if you’re looking at a glass of water sitting on a table, you’ll have a really hard time seeing that small piece of dirt floating on top if you’re looking at it while you’re standing up.  If you look at it very closely, you’ll be able to see it better.  This is what these new high resolution climate models hope to do: simulate more realistically the climate system, including its variability and response to different forcings.

-Jessica Mackaro