(abstract for AGU Fall 2008 meeting)

Atmospheric tides driven by solar heating are readily detectable at Earth’s surface as variations in air pressure. Above the lower stratosphere the tides attain large amplitudes and can be a significant part of atmospheric motion. Output from the general circulation model WACCM, the Whole-Atmosphere Community Climate Model, contains tidal oscillations in its middle and upper atmosphere, but it has not previously been examined for the surface signature of the tides. We have done so both for WACCM Version 1 and for the climate models contributing to the latest assessment report of the Intergovernmental Panel on Climate Change (from the Coupled Model Intercomparison Project phase 3 [CMIP3] archive, http://www-pcmdi.llnl.gov/ipcc/about_ipcc.php).

Surface pressure tides in WACCM-1's output occur mainly in the tropics and are dominated by the semidiurnal component, as observed. Quantitatively, however, WACCM-1 overestimates the diurnal tide’s amplitude by up to a factor of two over tropical land while underestimating it by about one-third globally. Similar though less egregious errors occur in WACCM-1’s semidiurnal tide. Although the observed pressure field is contaminated by sampling errors, it is clear that WACCM-1 misses robust real-world features such as the large diurnal tides over mid-latitude high terrain and the non-migrating component of both diurnal and semidiurnal tides. We have not yet examined output from the most recent version of the model (WACCM-3), however.

Climate modelers generally do not have atmospheric tides or the middle atmosphere in mind during the model development process, and conventional wisdom holds that the dominant semidiurnal surface-pressure tide is mainly forced by ozone heating in the middle atmosphere. One might therefore expect that climate models with tops below the ozone heating peak (~50 km altitude) or with poor resolution in the middle atmosphere would produce surface pressure tides that are weaker than observed and perhaps dominated by the diurnal (once a day) component. In fact, our preliminary analysis reveals no such correlation between model features and the quality of tide simulation. Even GCMs that include only half the stratosphere produce reasonable-looking semidiurnal surface pressure tides! We will discuss possible resolutions of this paradox.

Last Updated: 2008-09-03