![[menu bar image map]](/menubar.gif)
Funds which support the development and operation of the GSMM come largely from the Stratospheric General Circulation with Chemistry Project (SGCCP). SGCCP is a broad-based effort within the Atmospheric Chemistry and Dynamics Branch whose objective is to numerically model the dynamics and chemistry of the stratosphere and mesosphere, and to compare the modeled results to observations of the real atmosphere (UARS, SBUV, TOMS, SAGE). The latter step, in essence, measures the extent of our understanding of the physical processes which govern the evolution of the winds, temperatures, and trace constituents in this region of the atmosphere, which lies between approximately 15 km and 90 km altitude.
In order to accomplish the above objective, we must obtain or create datasets of dynamically balanced winds and temperatures covering the entire globe for long time intervals. Since continuous observed data is not available above about 55 km, we must rely on numerical models to "generate" a simulated atmosphere for us. Here, we integrate the equations of motion in time and periodically dump the basic state variables to disk for later analysis. The GSMM is one of the three-- dimensional (3D), global coverage models which we use to fulfill this task.
A particular feature and, hence, an advantage of the GSMM is that its upper boundary lies near the mesopause, which is located at approximately 90 km altitude, or 0.01 hPa. This is in contrast to the other important 3D dynamics model which we use, the assimilated data model, which terminates at about 55 km, or 0.40 hPa. In other words, the GSMM fills the need for simulations above the domain of data assimilation. Even though the density of the atmosphere in this region is four orders of magnitude below that at the earth's surface, UARS observations and results from recent numerical investigations make it apparent that inclusion of the mesosphere is necessary to completely describe the long-term transport circulation even in the stratosphere. For information on the assimilated data model, see the Data Assimilation Office's (Code 913.0) Home Page.
Mechanistic models in general have been implemented by several research groups in order to study analytical dynamics issues and to address some problems which appear to be generic to low and moderate resolution numerical simulations of the stratosphere. For our concerns, the most crucial is that of weak wintertime planetary waves, which in the real atmosphere periodically amplify. Without a robust planetary wave spectrum, the simulated wintertime polar vortices become anomalously strong, and the observed interannual variability of the circulation of the stratosphere, especially in the northern winter, is not reproduced. Planetary waves play the dominant role in reducing the equator-to-pole temperature gradient and in the mixing and distribution of trace constituents, such as ozone, nitrous oxide, and methane.
These planetary waves originate largely in the troposphere and propagate upward. Thus, to faithfully reproduce stratospheric dynamics, we need accurate simulations of the lower atmosphere and, at the same time, need to assure the upward transmission of Rossby waves through the tropopause. In fact, this has proven a difficult task, for it involves such diverse considerations as the domain in which the models are formulated, the incorporation of topography and clouds, and the time-stepping scheme which is used to integrate the model forward. Indeed the solution of these, and other fundamental modeling problems, is not mandated for Code 916. Rather, we are here to study transport issues.
The mechanistic approach offers a simple, yet effective method by which we may continue our studies. We do not attempt to model the troposphere at all. Instead, the lower boundary of the model is placed on a constant pressure surface which typically lies in the lower stratosphere and/or the upper troposphere. We generally choose 100 hPa. That surface, which is often called the "reference surface," is then continuously updated with observed geopotentials as the integration proceeds. That is, borrowing from the traditional terminology, we "mechanically force" the GSMM's lower stratospheric planetary wave spectrum to conform to that which is observed. Above the reference surface, the GSMM remains a prediction model since the planetary waves freely interact as they propagate upward. This obviates difficulties with transmission through the tropopause and imposes a degree of interannual variability in the generated winds.
The source of the dynamical core of the GSMM is version 0 (old, but tractable!) of the Community Climate Model from the National Center for Atmospheric Research. Modifications include relaxing the requirement for zero vertical velocity at the model's reference surface, addition of Rayleigh friction as a crude parameterization of momentum dissipation in the stratosphere and mesosphere, and incorporation of the accurate SGCCP radiative transfer algorithm. The mixing ratios of ozone and water vapor, which are used by the radiative transfer algorithm to determine local heating and cooling rates, are supplied by climatology rather than by prediction. A detailed description of the GSMM can be found in Nielsen et al. (1994).
Click here for full image (18 Kb)
Figure 1 illustrates the performance of the mechanistic model at 80 N during two Northern hemisphere winters, Nov. 1986 through March 1987 and Nov. 1989 through March 1990. The top two panels display the zonally averaged wind and the bottom two panels display the zonally averaged temperature, each as a function of time and pressure. Both winters are punctuated by large planetary wave-number one events, indicated by the deceleration in the mean zonal winds and in the rapid increase in temperatures near 1 hPa. The primary interannual difference is that after the January 1987 event, winds in the stratosphere do not substantially recover, and temperatures at this latitude in the lower stratosphere are generally above 220 K. In contrast, after the event in early February 1989, the winds recover to at least 30 m/sec. This indicates that the polar vortex survives the attack of planetary wave energy relatively intact. In fact, lower stratospheric temperatures in March 1990 are quite cold, hovering near 210 K in the zonal mean at this latitude. The TOMS retrievals indicate anomalously low total ozone near the North Pole in March 1990.
Click here for full image (27 Kb)
Figure 2 illustrates the evolution of a conservative tracer in a five-month CTM experiment which used GSMM winds. Each panel shows the zonally averaged tracer mixing ratio in arbitrary units, say parts per billion by volume (ppbv), at the indicated date. The range is 0 ppbv to approximately 10 ppbv with the darkest shading indicating highest concentrations. The result is instructive with respect to the long-term transport circulation in the stratosphere and mesosphere, the so-called "Brewer-Dobson circulation."
On 1 Nov 1989, we initialize the tracer in two "solid columns," one roughly located over each respective pole. In the northern hemisphere, winter is approaching and the polar atmosphere is rapidly cooling. Thus the tracer, which is largely confined to the polar vortex, rapidly descends. By 1 Dec 1989, the tracer material which was originally in the mesosphere has descended into the stratosphere. By 1 Apr 1990, much of the material has descended to the lower stratosphere, if not below the reference surface. The magnitude of decent is important since its attendant adiabatic heating may to some degree determine whether the lower stratosphere remains cold enough to support the formation of polar stratospheric clouds in winter.
Over the South Pole the integration begins as summer is approaching. Heating rates at this time are large and positive in the upper stratosphere over the pole, and at pressures less than 10 hPa, the tracer material ascends. This ascent, along with descent over the opposite pole, establishes a pole-to-pole circulation near solstice, which transports the tracer material northward across the equator and into the northern hemisphere. By 1 Dec 1989, material which was originally located over the South Pole can be seen crossing the equator, and one month later, it can be seen beginning to descend in the mesosphere over the North Pole. At the end of the integration, this material is descending well into the stratosphere.
The mean meridional transport circulation which is revealed by the evolution of the inert tracer is qualitatively what we expect. That is, we see the solstitial pole-to-pole Brewer-Dobson cell. As the equinox approaches, and the sun migrates toward small declinations, the strength of this circulation weakens. At equinox, the one-cell structure is "replaced" by a two-cell structure in the zonal mean, with gentle ascent over the equator, flow from the tropics to both poles in the mesosphere, and descent over both poles. At the following (June) solstice, the one-cell structure reappears, but with the mean circulation in the opposite direction.
Click here for full image (12 Kb)
To examine the transport properties of GSMM winds in three-dimensions and to address a more practical issue, we present Figure 3, which shows some aspects of the evolution of nitrous oxide during the February 1990 planetary wave event (see Figure 1 for the winds and temperatures). In general, the N2O mixing ratio falls steadily with height, from approximately 250 ppbv in the lower stratosphere to less than 0.5 ppbv in the mesosphere. It is fairly inert in the darkness of polar night, and thus can be regarded there as a conserved tracer. That is, we should see depleted N2O in winter polar stratosphere if it is descending in the wintertime polar vortex like the tracer does in Figure 2.
On the left in Figure 3 is a north polar orthographic projection of the CTM's N2O on the 1000 K potential temperature surface on 11 Feb 1990. The location of the polar vortex is indicated by the 1, 2, 5, and 10 ppbv contours. Indeed, these closed contours represent mesospheric concentrations of N2O. On the equatorward side of the vortex, tropical N2O is being transported rapidly northeastward across Siberia and is indicated by mixing ratios above 40 ppbv. This illustrates how effectively the planetary waves "mix" low-latitude and high-latitude air.
The center panel of Figure 3 is a longitude-height contour plot of N2O along latitude 69 degrees north. This cross-section cuts through the middle of the westward tilting polar vortex between 30 and 120 degrees E longitude. Dramatically illustrated is the descent of mesospheric N2O in the core of the vortex. The extent of this descent is confirmed by UARS observations of methane which itself behaves like N2O in polar night.
This three-dimensional picture of descent thus augments the zonally averaged description from Figure 2. In fact, wintertime polar descent cannot be considered simply as constant downward motion over the pole. Rather, the model demonstrates that descent occurs within the polar vortex which is occasionally displaced from the pole by the action of planetary waves. On the right in Figure 3 we show the zonally averaged N2O mixing ratio as the mid-February 1990 planetary wave starts to decay. At this time, wave-one amplitudes have nearly faded in the middle stratosphere and the polar vortex has migrated back onto the pole at pressures greater than 1 hPa, carrying it's depleted N2O with it. In the upper stratosphere and mesosphere amplitudes are still large. Thus the zonal average shows the 5 ppbv contour "cut-off." This "cut-off" feature is seen in HALOE methane observations during the decaying phase of a similar wave event in October 1992 over Antarctica. In both cases, reference only to the mean circulation is insufficient to explain the development of the "cut-off" feature.
Last Updated: 4 Oct 2000
Author: Eric Nielsen (Emergent Information Technologies, Inc.) (eric.nielsen@gsfc.nasa.gov)
Web Curator: Leslie R. Lait (Hughes STX) (lrlait@ertel.gsfc.nasa.gov)
Responsible NASA organization/official: Dr. P. K. Bhartia, Atmospheric Chemistry and Dynamics Branch/Head
![[]](/icons/back.gif){kind=icon}