Assimilation of Mars Global Surveyor meteorological data

The Mars Global Surveyor (MGS) Mission is more than just a return to Mars. It represents a qualitatively new type of planetary mission. This is true not only because of the capable instrumentation aboard the spacecraft and the choice of orbit and data rate, but also because of a major beneficial change in our understanding of the interplay between observation and modeling. The Thermal Emission Spectrometer (TES), for example, is capable of taking thousands of infrared spectra per day. The 15 micrometer carbon dioxide band in these spectra can be inverted to obtain atmospheric temperature profiles, amounting to some 15,000–25,000 separate measurements each day. In addition, radio occultations produce fewer, but much higher resolution, temperature and pressure profiles. How is such a quantity of data to be handled? Clearly not in the old-fashioned manner in which a single profile is studied at great length and detail. The quantity of data does not allow this; and the quality of the data does not support it. (A single atmospheric profile is bound to contain small-scale spatial and temporal variation—e.g., gravity waves—that cannot be removed unambiguously, as well as possible errors due to noise in the observations or non-uniqueness of the inversions). Furthermore, much of the atmospheric science of interest (winds, for example) depends on quantities like horizontal temperature gradients that are not directly observed. Instead, the data should be examined collectively with the aid of a model that incorporates our knowledge of the governing physics. Described here are the first results of such a data assimilation exercise with TES observations during the aerobraking hiatus period at L_s ≈ 200.