Science Briefs

What's Wrong with Models of the Stratosphere?

Figures 1
Figure 1: Vertical profiles of observed and simulated mean age in the tropics. The blue shaded region represents values simulated by 20 different models. Individual lines are results from GCMs at NCAR (red) and GISS (violet). The symbols are averages of mean age inferred from SF6 and CO2 from several balloon flights in 1997. The uncertainty is represented by the red shaded region, which shows one standard deviation on either side of the observations.
Figure 2
Figure 2: Simulated values of Cly mixing ratios versus mean age. Each number represents a different model result. Blue symbols represent latitudinal zone average values at 35°N, 18km and red symbols at 55°N, 22km. Observational estimates of the mean at these two locations are about 2 years and 5 years, respectively.

Complex interactions between chemistry, radiation, and transport determine the environmental effects of industrial pollutants, such as chlorine, on the stratosphere. Models used to predict the impact of pollutants must include all these processes simultaneously.

To determine the causes of unrealistic features in a model, however, individual components need to be evaluated separately. Recently, researchers from several institutions, including the NASA Goddard Institute for Space Studies, examined transport in a range of stratospheric models and concluded that transport inaccuracies may limit the ability to assess the impacts of pollutants on stratospheric ozone (Hall et al. 1998). These inaccuracies are a major source of uncertainty for simulations of important chemical species in the stratosphere.

Atmospheric transport is best quantified from observations of trace gases, or "tracers". Sulfur hexafluoride (SF6) is an example of a good tracer since it is both long-lived (only photochemically destroyed after thousands of years in the atmosphere) and steadily increasing from well-known industrial surface sources. From SF6, we can derive a useful measure of transport, called the "mean age", which is the average time required for air to reach the stratosphere from the troposphere.

Fig. 1 compares observations of the vertical distribution of the mean age in the stratosphere to various models. Although mean ages deduced from the models vary considerably, most are significantly lower than observed values. Comparisons made at middle and high latitudes show similar discrepancies. Results from general circulation models at the National Center for Atmospheric Research and at GISS are indicated separately in Fig. 1. These models display the "oldest", and therefore most "realistic" air (i.e., are in closest agreement with the observations), although they too display unrealistic features outside the tropics.

Who cares? After all, it is the distribution of pollutants with ozone-depleting potential that we want to simulate accurately and not the mean age of the stratosphere. One such pollutant is inorganic chlorine, Cly, which has reactive forms implicated in rapid ozone destruction. As illustrated in Figure 2 for two regions of the stratosphere, there is a clear correlation between Cly and the mean age across models. This implies that most of the "uncertainty" in simulation of Cly (as defined by the range across model results) is due solely to uncertainties in model transport, as measured by the mean age. Moreover, the uncertainty in both Cly and mean age is more than a factor of 2 at these locations. Such large Cly variations will produce significant uncertainties in simulations of ozone chemistry.

These results indicate that there is room for improvement in models that are used to predict the impact of industrial gases on the ozone layer, and inaccuracies in transport are a major source of the error. Further research is critical to better understand and simulate atmospheric transport.


Hall, T. M., D.W. Waugh, K.A. Boering, and R.A. Plumb 1999. Evaluation of transport in stratospheric models. J. Geophys. Res. 104, 18815-18839.


Please address all inquiries about this research to Dr. Timothy Hall.