UCAR > Communications > Quarterly > Winter 1999 Search

Winter 1999

It's about TIME: A new upper-atmosphere model

Byron Boville, Rolando Garcia, and Ray Roble. (Photo by Carlye Calvin.)

Onward and upward

For years, some scientists at NCAR have averred that the atmosphere can be thought of as extending from the solar corona to the earth. In a few years, it will be possible to study the implications of that idea. With funds from NSF's Knowledge and Distributed Intelligence program, principal investigator Bob Clauer (University of Michigan), Ray Roble, and colleagues at Michigan and Rice Universities have begun work on a high-performance supercomputer model that extends from the corona to the thermosphere.

"The component models exist and have been quite successful independently," Roble notes. Michigan has a magnetohydrodynamics model that goes from the corona to the outer magnetosphere (the region of space near earth, in which the earth's magnetic fields are the dominant force); Rice has a model of the inner magnetosphere. Roble's contribution is a version of his model that includes only the thermosphere and ionosphere, named TIE-GCM. "It's a very complex coupling problem," says Roble.

To cope with the enormous distances it must cover, the model will probably use an adaptive grid. "In the vicinity of the sun, you have a lot of grid points," Roble explains. "In the vast region through space where there isn't very much variation, there will be a smaller number of grid points. Then when you get in the vicinity of the earth there's a lot of grid points again. The logic is built in [to the model] to build up the grid where the action is."

One of the things Roble is enjoying most as the effort gets under way is a chance to collaborate with specialists from other fields. "The scientists are working with programmers and computer engineers, and working with massively parallel machines. It's great for us."

Ray Roble, a senior scientist in the High Altitude Observatory, has spent the last 20 years or so creating and refining a general circulation model of the upper atmosphere--the transition zone between the bulk of the atmosphere below 30 kilometers and the void above 500 km. Now, the knowledge Roble has amassed in developing his thermosphere-ionosphere-mesosphere electrodynamics general circulation model (TIME-GCM) is being incorporated into the middle atmosphere community climate model, an offspring of NCAR's CCM family that was created by Byron Boville (Climate and Global Dynamics Division). A third co-principal investigator, Rolando Garcia, and colleagues in NCAR's Atmospheric Chemistry Division (ACD) will be adding chemistry components to the model.

"A year from now, we expect to have a model that extends from the surface to 120 kilometers and includes ozone chemistry," says Boville. The yet-unnamed composite model will be one of only two or three general circulation models to reach that altitude. Some GCMs only extend to 30 km, including the troposphere and part of the lower stratosphere; a few have been extended into the mesosphere, with upper boundaries in the range of 60-80 km. The new model will include the entire mesosphere as well as the lower thermosphere. But the plan doesn't stop there. Further into the future, the scientists will add thermospheric and ionospheric dynamics and extend the model a few hundred kilometers higher.

A simple way to conceptualize the atmosphere is by its thermal structure, a sort of four-layer cake governed by radiative and dynamic processes. From the bottom up, each layer--troposphere, stratosphere, mesosphere, thermosphere--is alternately cooling or warming with altitude. The top layer, however, has a more complicated structure, because the thermosphere coexists in space with the ionosphere, a region ruled by electric and magnetic fields. The electrically charged atoms and molecules of the ionosphere are ionized by solar radiation and aurora particle precipitation. This happens throughout the atmosphere, but in the denser air lower down, charged particles quickly bump into oppositely charged particles and recombine. In the sparsely populated ionosphere, ionized particles zip around for long distances without losing their charge.

"Most of the variability in the ionosphere and thermosphere has been attributed to variation in the aurora, heating due to dissipation of electric currents, and variability in the sun's ultraviolet and extreme ultraviolet radiation," says Roble--in other words, to processes within the upper atmosphere or coming from the sun. "But we have never been able to account for the observed variability in the lower part of the thermosphere, so we have not been able to model it." A few years ago, Roble attempted to overcome this problem by coupling the TIME-GCM with the CCM2, which extends to 30 km. "We took off [the CCM's] rigid lid and put my model on the top. Then we could understand that the troposphere was an important source of variability.

"As a demonstration project to show that variability was coming across the boundary between the models, it worked reasonably well," Roble continues. However, the differences between the two models were daunting. The TIME-GCM is a gridpoint model with a 5-minute time step; the CCM2 is a spectral model with a 20-minute time step. "To do the problem right, we have to have one model that does it all."

The scientists hope that the completed model will help answer a longstanding question. "Walt Roberts's dream was to couple the sun to the weather," Roble comments. "There has never been a model that was able to address those issues. Hopefully the extended version of this model with chemical coupling will be a start." In the coupling with the CCM2, Roble found that "effects from above get smaller and smaller as you go down, and they get buried in weather noise. It was much clearer to see the lower atmosphere affecting the upper atmosphere." Others will use the extended model to look at, for example, the chemistry and dynamics of the ozone hole and the importance of solar activity to processes throughout the entire atmosphere.

In this issue... Other issues of UCAR Quarterly

Edited by Carol Rasmussen, carolr@ucar.edu
Prepared for the Web by Jacque Marshall
Last revised: Tue Apr 4 15:11:41 MDT 2000