The sensitivity of the Martian atmospheric circulation to a number of poorly known or strongly varying parameters (surface roughness length, atmospheric optical depth, CO2 ice albedo, and thermal emissivity) is investigated through experiments performed with the Martian version of the atmospheric general circulation model of Laboratoire de M¿teorologie Dynamique, with a rather coarse horizontal resolution (a grid with 32 points in longitude and 24 points in latitude). The results are evaluated primarily on the basis of comparisons with the surface pressure records of the Viking mission. To the end, the records are decomposed into long-period seasonal variations due to mass exchange with the polar caps and latitudinal redistribution of mass, and short-period variations due to transient longitudinally propagating waves. The sensitivity experiments include a 5-year control simulation and shorter simulations (a little longer than 1 year) performed with ''perturbed'' parameters values.

The main conclusions are that (1) a change of horizontal resolution (twice as many points in each direction) mostly affects the transient waves, (2) surface roughness lengths have a significant impact on the near-surface wind and, as a matter of consequence, on the latitudinal redistribution of mass, (3) atmospheric dust optical depth has a significant impact on radiative balance and dynamics, and (4) CO2 ice albedo and thermal emissivity strongly influence mass exchange between the atmosphere and the polar caps. In view of this last conclusion, an automatic procedure is implemented through which the albedo and emissivity of each of the two polar caps are determined, together with the total (i.e., including the caps) atmospheric CO2 content, in such a way as to get the closets fit of the model to the Viking pressure measurements. ¿American Geophysical Union 1995