Spatially and temporally varying dust lifting thresholds in a Martian GCM

Introduction: Atmospheric dust is one of the key constituents of the Martian atmosphere, and variations in dust loading can have a profound effect on the development of the mean circulation [1], and affect ice cloud formation [2] and other aspects of the climate system. While parameterisations for predicting rates of dust emission from the surface are now an important part of Martian general circulation models (MGCMs) [3, 4, 5], the high degree of interannual variability and apparent unpredictability in the occurrence of large Martian dust storms [6] has proved difficult to capture in models thus far. One important reason for this is a limited knowledge of Martian soil properties, and of the spatial and temporal variability of the surface state (particle surface dust density, particle size distribution). Here we use what information there is to produce a physically-based dust lifting scheme for the UKMGCM (a Mars GCM which couples the UK spectral dynamical core to physics from the LMD model [7]) that rivals, as much as possible, the complexity of current terrestrial dust emission parameterisations. Attention is paid to the ability of the model to reproduce observed timings of dust storm initiation and decay, the spatial pattern of surface dust removal, and interannual variability in global dust opacity. MGCM dust lifting parameterisations: Dust lifting in the UKMGCM was originally implemented by [3], and features two parameterised processes, in common with other Mars GCMs [4, 5]. Convective lifting, by ‘dust devils’, occurs at a rate proportional to an activity parameter [8], dependent on the thermodynamic efficiency of the convective heat engine and the sensible heat flux. A thresholddependency may additionally be imposed, using a cyclostrophic estimate of the tangential wind speed of the vortex. The proportionality constant, defining the efficiency of the lifting process, is tuned to provide model dust opacities in broad agreement with observations over the period Ls=0-180°, when lifting by other mechanisms is thought to be infrequent. Greater focus has been applied to the other model dust lifting mechanism, namely near-surface wind stress, reflecting its importance in the development of regional and global dust storms. We calculate the threshold stress for the initiation of dust lifting using equation (24) of [9], applied to sand-sized particles of diameter ~180μm. Above threshold, micron-sized vertical dust flux is set proportional to the horizontal saltation flux, estimated using a function based on the numerical model of [10]. This functional dependence is similar to the more commonly-used formula of [11], but saltation flux levels off somewhat with increasing frictional velocities, due to the influence of electrostatic effects on saltation. It thus implies a slightly more explosive lifting process than the previously-used flux function. GCM mean surface winds do not include subgridscale windspeed variation and transient gusts, which are of crucial importance to a strongly threshold-dependent process such as wind stress dust lifting. In addition, a two-threshold situation has recently been theorised to exist on Mars [12, 13], implying that very different frictional velocities are required respectively to initiate and subsequently to sustain saltation. At present, both these considerations have been subsumed into a reductive scaling factor applied to the calculated threshold friction velocity. The value of this factor, determined by the model’s ability to produce significant dust lifting at the correct locations and time of year, and found to be ~0.7 at low model resolution (resulting in a threshold stress of around 0.019Nm), is physically plausible given the order-of-magnitude difference in impact and fluid thresholds found by [13]. This is, however, likely to be a key source of error in the dust lifting scheme, and would benefit from greater insight into smallscale surface wind variability.