Data assimilation of three mars years of thermal emission spectrometer observations: Large-scale transient and stationary waves

Introduction: Large-scale traveling and stationary planetary waves are diagnosed from an analysis of profiles retrieved from the Thermal Emission Spectrometer (TES) [1] aboard the Mars Global Surveyor (MGS) spacecraft during its scientific mapping phase. The analysis was conducted by assimilating the TES temperature profile and total dust opacity retrievals [2] into a pseudo-spectral Mars general circulation model to produce a full, physically self consistent record of all atmospheric variables stored at an interval of two hours over the entire MGS mapping phase. The data cover a period of about three Mars years, corresponding to the interval 1999–2004 on Earth. These include the year which contained the 2001 global dust storm [3] and two years of more moderate dust activity, although large regional storms occurred during southern hemisphere summer in both years and there was considerable atmospheric variability between all three years [4]. We focus on the planetary wave activity, both traveling and stationary large-scale waves, in the assimilated record. Data assimilation is a particularly useful technique for the analysis of transient wave behaviour since it is capable of producing global, time-dependent atmospheric fields, which the assimilation scheme endeavours to make as consistent as possible with whatever observations are available. These atmospheric variables may be sampled from the model as often as desired, on a regular grid of points. If particular variables, or regions of the atmosphere, are not observed directly, the model will at least ensure that they are consistent with the laws of physics incorporated within its framework. A complex climatology of transient waves is revealed, modulated by the large-scale topography and surface thermal properties, the time of year and, crucially, the amount of dust suspended in the atmosphere. Some individual case studies show the temporal and spatial structures of the waves in the assimilation record, although the large data set has by no means been fully explored. Companion papers discuss the thermal atmospheric tides [5] and the processes associated with the initiation of dust storms [6] from the same assimilated analysis. Output from the same assimilation has also been used to identify potential deficiencies in the model, such as the lack of water ice clouds [7]