[Previous] | [Session 15] | [Next]
H. Houben, R. W. Bergstrom (Bay Area Environmental Research Institute), J. Hollingsworth (San Jose State University Foundation)
Mars Global Surveyor (MGS) has produced an overwhelming atmospheric dataset, principally from the Thermal Emission Spectrometer (TES). Four-dimensional variational data assimilation of TES radiances, using a specially constructed general circulation model, reproduces the observations to within 1 erg/cm2/s/sr/cm-1 in the center of the 15 micrometer CO2 band and--with an improved ground temperature formulation--almost as well in the wings. One-sol forecasts have an rms error of order 2 erg/cm2/s/sr/cm-1. The product of the assimilation is a daily global meteorology of the planet. The calibration and validation of these results is quite challenging given the lack of ground truth measurements. However, we have begun systematic comparisons of our results with other MGS instrument meteorological observations: Radio Science occultation profiles; Mars Horizon Sensor Assembly measures of broadband 15 micrometer radiation; Mars Orbiter Camera and Mars Orbiter Laser Altimeter water, dust, and carbon dioxide cloud detections. Because our assimilated fields are global, these supplementary observations need not be co-located or simultaneous with TES measurements. In addition, since the daily datastream is much bigger than the number of variables required to constrain the model, our results are insensitive to model assumptions or parameterizations (indeed, they are used to improve these model features). As a result, our compact model state can be efficiently used to reproduce the data (at the model resolution). This suggests a standard that should be incorporated on future missions devoted to the study of the Martian or other planetary atmospheres: high level data assimilation products should be produced in near real time. These will allow all researchers in the field to make the best, timely use of the data and may contribute to operational considerations as well. The latter will certainly be enhanced when predictions are optimized, probably through the use of sequential assimilation.
Supported by the Mars Data Analysis Program under RTOPS 344-34-21-04 and 390-90-21-08.
If the author provided an email address or URL for general inquiries,
it is as follows:
Bulletin of the American Astronomical Society, 34, #3< br> © 2002. The American Astronomical Soceity.