Technology  August 27, 2010

Writing code to peer into gobal climate future

BOULDER – For decades the National Center for Atmospheric Research has been creating models to help simulate what Earth will look like in the future.

While scientists and researchers are testing various hypotheses, software developers are entering the equations and ensuring the parameters to test those hypotheses work correctly.

SPONSORED CONTENT

Select your Republic Services residential cart now!

In preparation for Republic Services becoming the primary provider of residential recycling, yard trimmings, and trash, residents should now select the best cart size and service schedule for their household needs.

Sam Levis, an NCAR project scientist who develops and analyzes the modeling software, said there are hundreds of people helping write code for various models, including the recently announced Community Earth System Model.

The software and technology isn’t new, but having numerous people working on a project at the same time while keeping it accurate despite constant scientific discoveries is what makes NCAR’s modeling impressive.

The new Community Earth System Model, which was started decades ago, is continually evolving as scientists find out more about how the Earth works and is changing. With that data, Levis and a group of other scientists input equations into the software to achieve accurate modeling.

At the most basic level, the group is inputting equations like F=MA (for force equals mass times acceleration), which helps determine how fluids move, Levis said.

“We’ll use what we understand about science that may be in the form of equations about how the atmosphere operates and how the ocean and things on the land behave,” he said. “We use equations to represent physical processes and chemical and biological processes.”

With each process, such as how wind changes over time, how ocean currents move and how soil erodes, transformed into an equation and input into the modeling software, scientists can, with a fair amount of accuracy, predict what the Earth will look like in the future.

Constant adjustments

But scientists and researchers are constantly making adjustments, because, as Levis pointed out, “The simulation is far from perfect.”

After starting with a base – which NCAR calls preindustrial conditions in the year 1850 – scientists compare that model with what information they know and what they think is reasonable for that time, given factors like the atmospheric composition and how close the Earth is to the sun.

With the base set as accurately as possible, scientists can determine what snow pack will be in 10 years, how the ocean currents will change, and what areas of the Earth will erode faster than others.

To allow researchers to tweak various conditions as needed for various hypotheses, Levis and other developers enter the information in segments that can be accessed individually and changed as needed.

“We write code in newer versions of Fortran, mainly just because of tradition, and it works well,” Levis said.

The programs are very modularized, he said. “We try to isolate each process that we simulate to facilitate further development so when new science comes around we can easily incorporate it into the model.”

Levis said hundreds of people are involved with the community model at NCAR. With the module concept, he said someone from a university could work on a problem within a segregated group without impacting the others. It also allows scientists to turn various processes on and off or play around with individual parameters to see what would happen.

The Community Earth System Model has new features that haven’t been done on supercomputers studying climate. Its projections will be used for government reports on climate change, providing data for future environmental policy.

BOULDER – For decades the National Center for Atmospheric Research has been creating models to help simulate what Earth will look like in the future.

While scientists and researchers are testing various hypotheses, software developers are entering the equations and ensuring the parameters to test those hypotheses work correctly.

Sam Levis, an NCAR project scientist who develops and analyzes the modeling software, said there are hundreds of people helping write code for various models, including the recently announced Community Earth System Model.

The software and technology isn’t new, but having numerous people working on a project at the same time…

Sign up for BizWest Daily Alerts