P31B-0979 0800h
Numerical Simulation of the Atmospheric Circulations and Emission, Transport and Dispersion of Dust Particles in the Tarim Basin
It is widely accepted that the Tarim Basin within the Taklimakan desert is one of the major sources of aeolian dust in East Asia. However, the mechanisms responsible for dust emission, transport, and dispersion from surface to the free atmosphere are not clear and evaluations of dust supply to the free atmosphere have large uncertainties. We have performed a numerical simulation of the dust event in spring there to examine the mechanisms and to evaluate the dust supply with more certainty. The Regional Spectral Model (RSM; the equivalent horizontal grid size is 20km) and the Non-Hydrostatic Model (NHM; 10km) of the Japan Meteorological Agency are employed to represent atmospheric circulation. The Lagrangian transport model with a simple spectral dust emission scheme by Shao (2003) is coupled with the NHM to simulate dust emission, transport, and dispersion. Effects of soil texture, vegetation, and snow cover fraction on dust emission are formulated in Shao's scheme. The model reasonably simulates space and time variations of surface wind, and transport and dispersion of dust. Three characteristic types of mesoscale wind system can be distinguished in the simulation. The dust dispersion is dependent on not only the surface wind intensification, but also the vertical motions arising in or between these wind systems.
P31B-0980 0800h
Did Dust From the 1930s US Dust Bowl Make it to Greenland?
The "Dust Bowl" phenomenon during the 1930s in the southwestern United States generated huge amounts of airborne dust with eastward transport of mineral aerosol. Amidst the normal flux of dust from China and Mongolia to the Greenland ice cap, were there any "Dust Bowl" years during which North American dust made significant deposits, or were even detectable? Donarummo et al. (2003) reported a possible instance of Dust-Bowl dust during 1933 from the GISP2 ice core drilled at Summit in central Greenland. Did this occur elsewhere in other years? At the NorthGRIP (75 N; 042 W) ice camp in 2001, we drilled four shallow firn cores (~ 1 m apart and 25.0 m deep) to the level of 1930 firn. We divided the cores into the four periods 1930-1945, 1945-1960, 1960-1980, and 1980-1990, and combined the appropriate intervals of all four cores in order to recover sufficient dust. If the southwestern U.S. dust were to be detected during the primary Dust-Bowl years (1930-1945 interval), we expected to see the signal return to pure East Asian characteristics over the course of the subsequent three periods. The four composite firn samples were melted, the dust extracted in the field by super-centrifugation, and returned to LDEO. Non-destructive XRD analyses for clay mineralogy preceded dissolution, chemical purification and analysis of radiogenic isotope composition (ONd(0) and 87Sr/86Sr) by Thermal Ionization Mass Spectrometry. The results of these analyses were compared to our samples of Chinese and Mongolian desert material, to samples of known Dust-Bowl dust, and to all the samples of ice-core and snow-pit dust from Greenland we had previously analyzed. These comparisons reveal no hint of compositional change from Asian toward Dust-Bowl characteristics. We conclude that the single possible occurrence of Dust-Bowl dust in 1933 ice reported by Donarummo et al. (2003) did not at all characterize the entire Dust-Bowl era in Greenland. We have also previously shown that the mineralogical and radiogenic isotope composition of dust in Greenland ice is the same at multiple high-altitude, interior, ice-cap locations (Bory et al., 2003), so the 300 km that separate our NorthGRIP results and the single possible result at GISP2 (Summit) is probably not significant. The answer to the question posed in the title is: "Perhaps once, but not often, and not volumetrically significantly." References: -Bory et al., Geochem., Geophys., Geosys., v.4, 1107,doi:10.1029/2003GC000627, 2003 -Donarummo et al., Geophys. Res. Let. 30(6):1269, , doi:10.1029/2002GL016641, 2003
P31B-0981 0800h
Influence on the atmospheric general circulation caused by the direct effect which dust exerts on radiation process
The total amount of the soil particle (Aeolian-dust) danced by the wind from the dryness area is called annual 1000-3000Tg on the whole earth. The thing from Sahara Desert in North Africa occupies more than the half of these. The small particle of particle diameter piles up into the atmosphere among the danced soil particles for a long period of time (being weight quantity around about 1%). The particle which piles up into the atmosphere has the work which is scattered about or absorbs solar radiation. On the other hand, aeolian dust has the work which performs absorption and discharge as black-body to infrared radiation, and serves as a substance which has greenhouse effect to an earth air system. We developed an general circulation model (MASINGAR) incorporating each model about the radiation process in consideration of the influence on the solar radiation by generating of aeolian dust, transportation, each self-possessed process, and dust, and infrared radiation. The numerical experiment about influence done to the atmospheric general circulation of aeolian dust using this was conducted.
P31B-0982 0800h
Dust Devils: Experimental Results for Vortex Sediment Flux
Laboratory experiments using the {\it Arizona State University Vortex Generator} ("dust devil machine") yield new results in simulating the amount of sand and dust (flux) raised by dust devils. Flux experiments involved measuring mass loss from a test bed as a function of time for known vortex parameters yielding a calculated value for the flux. Sediment fluxes of dust (2 $\mu$m in diameter), silica sands (5 sizes ranging in diameter from 90 to 500 $\mu$m), and walnut shells (2 sizes ranging in diameter from 590 to 1700 $\mu$m) are compared to the Reynold's number, {\it u{$\theta$}r{$\theta$}/$\nu$} (4000-18,000), and a dimensionless lifting parameter, {\it $\Delta$P/($\rho$$_{p}$gD$_{p}$)}, (0.001-1.0), in which {\it u{$\theta$}} and {\it r{$\theta$}} are the vortex tangential velocity and core radius, {\it $\nu$} is the kinematic viscosity of the air, {\it $\Delta$P} is the pressure drop across the vortex, {\it $\rho$$_{p}$} and {\it D$_{p}$} are the particle density and diameter, and {\it g} is gravitational acceleration. Results show that in general, flux increases with Reynold's number and with lifting parameter. Lower-density walnut shells show a higher flux than silica sands and dust suggesting that on Mars (where {\it g} is lower) the flux would be greater in comparison to Earth. Lower-density walnut shell particles are used to simulate the lower {\it g} of Mars. Future work involves further expansion of the experiment matrix with other materials of varying sizes and densities. Use of the Mars Surface Wind Tunnel facility at NASA Ames Research Center will also allow this study to be simulated at Mars atmospheric pressures.
P31B-0983 0800h
Searching for Active Dust Devils in Gusev Crater from Orbit by {\it Mars Express} and the Ground from {\it Spirit}
Prior to the landing of the {\it Mars Exploration Rover, Spirit}, orbital data revealed numerous low albedo features in the Gusev crater landing ellipse attributed to the passage of dust devils (DD), and it was anticipated that active DD would be seen during {\it Spirit} operations. {\it Spirit} landed on a linear dark feature considered to be a DD track. {\it Mars Express} images taken after the landing were compared to images taken just before the landing; results show that many of the dark DD tracks in the area had faded over one martian year (inferred to be the result of dust settling on the surface) and that new tracks had formed in areas north and west of the final {\it Spirit} landing site. The three camera systems on {\it Spirit} (Pancam, Hazcam, Navcam) were used to search for active DD in a campaign similar to that used successfully during the {\it Mars Pathfinder} mission. However, thus far no active DD have been seen from the surface. We attribute this result to one or more of the following: 1) no DD formed within sight of {\it Spirit}, 2) observations were not taken at the time or place of DD activity, 3) camera filters/compression values were insufficient to image active DD, or 4) current image processing techniques are inadequate for detecting active DD. As Spirit continues operations, additional searches will be made when resources allow.
P31B-0984 0800h
Mars Digital Dune Database
Currently, there is no comprehensive, global, digital database for dune deposits on Mars. The advent of a series of successful Mars missions, coupled with advances in technology enabling a significant increase in instrument resolution, have provided a large compilation of data covering a wide range of wavelengths for the Martian surface. Given the recent availability of high-resolution data and detailed surficial information returned from orbital and rover missions, it is critical that we update the Mars global information base by creating a digital database of dune deposits that includes this new influx of data. As of spring 2004, the Thermal Emission Imaging System (THEMIS) infrared (IR) coverage of the surface of Mars was 98% for nighttime and 75% for daytime acquired images, forming a data set of global coverage at a resolution not previously possible. The combination of high-resolution and global coverage makes the THEMIS IR data set the logical choice for a planet wide inventory of dune deposits. Data sets of a global scale like those of Mars Orbiter Laser Altimeter (MOLA) and Thermal Emission Spectrometer (TES) will enable rapid and contiguous comparisons with the dune database. Other imagery like that of Mars Orbiter Camera (MOC) will provide very high-resolution, localized visual data for accurate interpretations of morphological characterizations. The dune database will provide researchers with an extensive, comprehensive and stable database for use in a wide-array of global studies. The database will also offer researchers a centralized depository for updating physical parameters with newly validated findings. The initial construction of the database is based upon dune forms or deposits identified, classified and digitized using only THEMIS IR images. These digitized polygons are converted from THEMIS image coordinates to ARCMAP aerographical coordinates, allowing delineation of areal extent of the deposits and preserving relevant THEMIS image information such as L$_{s}$, local time, and sun azimuth/angle. The ARCMAP polygons will also retain reference to all THEMIS IR images used in their construction. Where available, THEMIS VIS and/or MOC images will be used to confirm, modify or refine original classifications. In addition to providing an improved resolution for features below the IR image threshold, this secondary examination will also provide a list of cross-referenced THEMIS VIS and MOC images for future investigations. Physical parameters such as wind direction based on slip-face geometry, dune wavelength, elevation, and volume of the deposits will be incorporated into the database on a priority-based schedule. In addition to THEMIS VIS and MOC images, supplemental data sets, such as TES and others, will be used where available to further refine and/or validate existing data on global wind patterns, sediment transport, sources and sinks, and stratigraphic units.
P31B-0985 0800h
The Search for Dune Movement on Mars
Although aeolian processes are active on Mars, questions remain about whether the current wind regime is adequate to cause dune movement. Early studies using Viking images might have detected movement in the northern erg, but that study was limited by image resolution. Since 1997, the Mars Orbiter Camera (MOC) on Mars Global Surveyor has been returning narrow angle (NA) images at resolutions of 1.5-12 m/pixel, and overlapping pairs of these images acquired at sufficiently spaced times can be used to search for dune movement. Whereas most images acquired during the Mapping Phase of the mission were near-nadir looking, images from the Extended and Relay portions of the mission often had look directions off nadir (ROTO images). This geometry causes parallax and complicates the search for dune movement but does not prevent it. Dune locations in MOC (and now THEMIS) images can also be compared to locations in Viking images, but movement would need to be on a larger scale because of the larger pixel sizes of Viking and THEMIS images. Results showing evidence for dune movement or for dune immobility would both be important for understanding the current martian aeolian setting and how it might have changed over time. The MSSS website and a USGS database of potentially overlapping MOC narrow angle images have been used to identify image pairs that could be used to determine whether dunes moved during the time spanned by the images. Of the over 150,000 MOC NA images released, only a small subset are overlapping pairs, and only a subset of those contain dunes. At this time, no definite dune movement has yet been detected, but the study is being expanded to cover more areas on Mars and to include more images as they become available.
P31B-0986 0800h
Wind Eroded Rocks in Gusev Crater, Mars, seen from the {\it Mars Exploration Rover}, Spirit
Ventifacts (wind abraded rocks) are common along the 3.2 km traverse of the rover, {\it Spirit}, from the landing at Columbia Memorial Station to the Columbia Hills (CH), and include flat faces on rocks ("facets"), flutes, and grooves. Facets range in size from 2 to 50 cm across and commonly occur on the north side of rocks from about 2 cm above the soil horizon up to the topmost peak of the rock. The upper third of these rock facet faces tend to be highly polished and relatively dust free and typically also is the zone of groove etching. Although facets occur on about 20-30% of rocks seen along the traverse, they are less than 10% on the southwest side of the West Spur in the CH and up to 75% occurrence on top of West Spur near Clovis. Flutes are U-shaped features on a rock that are closed on the upwind side and open on the downwind side in map view creating an "arrowhead" shape and grooves are longer and more linear than flutes, both resulting from aeolian scouring. Grooves range in size from 2 cm long and 0.25 cm wide to 9 cm long and 1 cm wide, while flutes range in size from 1 cm long and 0.5 cm wide to 8 cm long and 5 cm wide. Flutes and grooves are commonly on the northwest sides of rocks, typically on a facet face, and are oriented northwest-southeast along the long axes. Flat, light surfaces in the CH, such as Clovis, are thought to be outcrops of "bedrock," many of which have grooves as long as 10 cm and oriented 20 degrees east of north. Together, the ventifacts observed via {\it Spirit} suggest formative winds from the northwest, consistent with indicators, such as wind streaks, indicative of current wind directions in the region. The relative lack of ventifacts in the area immediately south of West Spur suggest this area is shielded from winds from the north-northwest as the wind is topographically guided more southward along the west flank of the CH complex. The grooves on Clovis are oriented in response to the wind flow over local topography.
P31B-0987 0800h
Aeolian Processes at Meridiani Planum
The Opportunity Rover has observed a variety of aeolian features at Meridiani Planum. These features imply wind-related processes that operate on a range of time-scales, so that at least a {\it relative} time-scale of formation and modification can be assigned to many. (a) Features forming and changing in the shortest time include probable impact ripples (cm-scale) that occur in dark basalt sand on the floors of local depressions throughout the plains. Also in this category are deposits of bright airfall dust (in the form of streaks) that are not removed downwind of topographic features such as crater rims. Analysis of MOC imagery indicates that streaks change orientation after intense dust storms. The similar orientations of impact ripples and bright streaks are thought to indicate the prevailing direction of the most-recent vigorous wind regime. (b) Forming and changing on a longer, intermediate time-scale (and hence older) are deflationary ripples armored with well-sorted mm-sized hematitic grains that likely propagate by creep (i.e., pushed by the impacts of smaller saltating grains). The removal of dust from the plains during storms implies saltation of sand ($\sim 100 \mu$ m) which indicates shear velocities approaching what is required to roll the $\sim 1$ mm hematitic grains. The ``plains ripples'' rarely occupy local depressions and cover the plains at Meridiani Planum ($\sim 1$ cm tall, $\sim 10$ cm wide, up to $\sim 1$ m long). Plains ripples indicate two prominent orientations: the orientation of individual bedforms as well as that of en-echelon ripple trains. As deflationary structures that form by the winnowing of small grains, these features require relatively long periods to form. The presence of multiple orientations indicates that, once formed, multiple episodes of vigorous winds may be required to change markedly the orientations of these features. (c) Requiring probably the longest formation times are tails of protected rock downwind of hematite spherules embedded in outcrop. The spherule-tails are thought to develop during periods of intense sand-blasting of the rock surface, and may indicate the consequence of wind abrasion over very long periods (i.e., many intense storms). These tails possibly indicate mean wind directions over long periods, and therefore the directions to sources of sediments deposited at Meridiani, or deposition-zones for sediments removed from Meridiani. In addition to these features, Opportunity has observed a collection of dunes at the bottom of Endurance crater. These basalt-sand dunes occur in a cellular pattern with star-shaped junctions and solitary arms that extend several meters from the main sand body. The cellular arrangement implies a complex pattern of winds in multiple directions at the crater floor. High cohesion of the sands is implied by lineations in dune faces that suggest cross-bedding. Red-filter observations indicate the presence of bright dust along crest-lines, implying that these are not currently active. Slip faces are steep, indicating minimal degradation. Well-defined high-frequency ripples that occur on or adjacent to most dune faces are interpreted as potentially active impact ripples.
P31B-0988 0800h
Trace and Rare Earth Element Geochemical Signatures Used to Assess the Sources of Silicate Materials in Atacama Desert Soils
The Atacama Desert of northern Chile is one of the driest deserts on Earth with precipitation levels as low as 2 mm y$^{-1}$. Soils in the driest parts of the Atacama record the effects of long-term hyperaridity, retaining atmospherically-derived elements in quantities rarely seen on Earth. Recent work has demonstrated that the hyper-arid environment allows Atacama soils to accumulate large amounts of sulfate (e.g. anhydrite, gypsum), nitrate (NaNO$_{3}$), and chloride (halite) that result in radical volumetric expansion of the soil. Possible salt sources include eolian redistribution from Atacama playas and/or marine aerosols. The objective of this work is to examine the trace and rare earth element (REE) chemistry of Atacama soils to assess the sources of silicate materials. Three soils developed on fluvial landforms were examined along a south to north transect (Copiapo $ > $ Altamira $ > $ Yungay) that coincides with decreasing moisture levels (~15mm to ~2 mm yr$^{-1}$, south to north) and with differing eolian sources. Trace and REE chemistry from material collected in soil parent materials, dust traps, and local stream beds will be compared to soils to assess the magnitude of local and distal eolian source contributions to Atacama soils.
P31B-0989 0800h
A Dune Simulation Wind Tunnel for Studies of Lee Face Processes
Sand is deposited on the lee slope of dunes by grainfall, avalanching (grainflow), and wind ripple migration. These processes play major roles in the formation of aeolian cross strata. Grainfall is produced by saltating grains that are blown over the dune crest and fall on the lee slope. Avalanching occurs when sand on the lee slope fails and the resulting grainflow will deposit tongues of sand downslope. Lee slope deposits are often preserved in the rock record and an understanding of the conditions that produce them as well as avalanche frequency and magnitude could provide important information about dune morphology, sediment flux and availability, air flow, and the environment of deposition. Despite their importance there have been very few studies of grainfall and avalanching because the lee slope of dunes is a very fragile and easily disturbed environment. Designing research strategies to study these processes presents significant technical challenges. To overcome these obstacles construction and testing of a dune simulation wind tunnel was recently completed as a joint project between the Wind Erosion Laboratory, Department of Geography, University of Guelph and the Desert Research Institute. This wind tunnel contains a small, but true-scale sand dune that is 8.5 m long, 1.2 m high, and 1 m wide and is capable of producing wind speeds of 16 m/s at 30cm above the dune crest. The wind tunnel provides an opportunity to study lee slope processes in a well-constrained environment by controlling wind speeds and direction, dune geometry and composition, and allows for extensive instrumentation and close observation of depositional processes. Preliminary experiments in the dune simulation wind tunnel indicate that avalanche frequency is predictable: increased wind speeds resulted in more frequent avalanches. Avalanches commonly originate in the mid-lee slope region at or near the point of reattachment of the return cell caused by flow separation at the crest. Downslope saltation and/or reptation on the lee slope surface appear to slow at the point of reattachment and appears to be an important factor in the location of avalanche origination. Preliminary measurements of the distribution of sediment deposited by grainfall at three different wind speeds suggest a possible depositional bulge in the mid-lee slope region. Research supported by NSF EAR-0207833
P31B-0990 0800h
Representation of Vegetation and other Non-erodible Elements in Aeolian Sediment Transport Models
The presence of non-erodible elements is well understood to be a reducing factor for soil erosion by wind, but the limits of its protection of the surface and erosion threshold prediction is complicated by the varying geometry, spatial organization and density of the elements. The predictive capabilities of some of the most recent models for estimating wind driven particle fluxes are reduced due to the poor representation of the effectiveness of vegetation to reduce wind erosion and dust emissions. Two approaches have been taken to account for roughness effects on sediment transport thresholds. Marticorena and Bergametti (1995) in their dust emission model parameterize the effect of roughness on threshold with the assumption that there is a relationship between roughness density and the aerodynamic roughness length of a surface and that the partitioning of shear stress can be estimated based on this relationship. Raupach et al. (1993) offer a different approach based on physical modeling of wake developments behind individual roughness elements and the partition of the surface stress and the total stress over a roughened surface. This geometric and drag coefficient based-predictive model allows the user to specify available inputs to describe most environments and generate the ratio that characterizes the erosion threshold of an initially bare erodible surface with the threshold once non-erodible elements are present. The objective of this presentation is to quantify the precision and accuracy of these two threshold wind erosion models by comparing model predictions with field and wind tunnel measurements with airflow over a range of roughness element scales, shapes, and distributions. The comparison between the models shows the partitioning approach to be a good framework to explain the effect of roughness on entrainment (and potentially transport) of sediment by wind. Both models provided very good agreement for the wind tunnel experiments using solid objects on a non-erodible surface, however their performance when tested in a field environment is less favourable. The Marticorena and Bergametti (1995) approach displays a scaling dependency when the difference between the roughness length of the surface and the overall roughness length is too great, creating a negative partition. While the Raupach et al. (1993) model requires more inputs than that of Marticorena and Bergametti (1995) its predictions are better due to the incorporation of more variations in the roughness geometry and the alterations to the flow they can cause.
P31B-0991 0800h
Shear Stress Partitioning in Airflow over Rough Surfaces: Roughness Form Effects and Influence on the Distribution of Shear Stress
Roughness elements distributed across a surface can significantly decrease the entrainment and transport of underlying fine-grained sediments by wind. The parameterization of roughness effects on wind erosion thresholds and sediment transport is critical to the development of models that can provide realistic predictions of sediment thresholds and fluxes due to wind erosion. Raupach et al. (1993) present a model for predicting the protective role of roughness elements in terms of a threshold friction velocity ratio as a function of the roughness geometry and the aerodynamic properties of the surface and roughness elements. The predictive capacity of this model remains uncertain and the work presented here represents part of an on-going effort of our group to improve the parameterization of the Raupach et al. (1993) model. To gain additional understanding of how roughness elements influence the magnitude and nature of the shear stress acting on the surface among the elements and evaluate strength and weaknesses of the roughness density parameter to characterize these effects, a wind tunnel study using model roughness arrays of similar roughness density composed of cube-shaped elements of different length dimensions was undertaken. Roughness density is defined as the total frontal area of all the elements to the total surface area that they occupy. Shear stress in the above element air flow was determined from vertical wind speed profile measurements. Point measurements of near surface shear stresses within the roughness array were made with simple omni-directional skin friction meters in order to investigate the partitioning of shear stress to the intervening surface. The results suggest that the roughness density parameter has severe limitations in describing the shear stress partitioning for these regularly arrayed rough surfaces. For surfaces with identical roughness densities, the surface composed of more and smaller elements was observed to have average and maximum shear stresses up to 20 percent lower than the surface composed of larger and less numerous elements, for similar freestream wind speeds. The results from the wind tunnel testing also suggest an explanation for the physical basis of the effect roughness has on particle transport thresholds due to alterations in the distribution of shear stress. Raupach et al. (1993) assumed that this effect was a result of spatial inhomogeneities in the surface shear however data from this study indicate the roughness modifies the frequency distribution of the range of shear stresses reaching the surface.
P31B-0992 0800h
Wind Flow Characteristics over Rough Surfaces
The presence of roughness elements such as isolated boulders or sparse vegetation, common to many arid and semi-arid environments, protect the underlying finer-grained sediments from wind erosion, principally by absorbing momentum from the wind. As a result, wind shear stress is partitioned between the isolated roughness elements and the intervening surface. As part of a larger wind erosion study, a series of wind profile measurement experiments were conducted over a relatively flat un-vegetated, crusted field with an area of approximately 10,000 meters square, located at the Jornada Experimental Range, USDA ARS, Las Cruces, NM. Near the upwind margin of the field a triangular shaped area 2000 meters square (with the triangle base located downwind and perpendicular to the dominant wind direction) was covered with up to 1,9250, commercially available, 5-gallon buckets in regularly spaced staggered arrays. The buckets were used to produce four different roughness densities to investigate changes in wind flow patterns and momentum over the roughness configurations. Wind profiles were measured on two 9 m towers located at the upwind and downwind margins of the triangular arrays along the central axis. Wind speed was measured with 8 anemometers spaced logarithmically from 0.5 to 9 m. Wind direction was measured at the top of each tower with a wind vane. Wind speed and direction data were recorded with a data logger at 1s intervals with an averaging time of 10 minutes. Wind profiles at the leading edge of the arrays approaching at angles less than or equal to 10 degrees from normal to the front of the roughness array are associated with the wind flow over the flat, crusted surfaces and are well described by the Prandtl-von Karman log-linear relationship. For the same free stream wind speeds, the profiles at the downwind edge of the arrays are also characteristically log-linear when a displacement height (d) is included in the Prandtl-von Karman equation. A strong linear relationship (R$^{2}$= 0.97) was found between the ratio of d to inter-element spacing (l) and the roughness density. As well, for similar free stream wind speeds, shear velocity increased markedly between the upwind and down wind towers, the difference of which was dependent on roughness density. Average momentum losses across the arrays (determined from the momentum deficit law) increased linearly with roughness density, indicating the increasing absorption of momentum by the roughness elements as the number of elements per unit area increased.
P31B-0993 0800h
Interannual Variability of Mars Global Dust Storms: an Example of Self-Organized Criticality?
Previous simulations of Martian global dust storms with a simple low-order model showed the desired interannual variability of storms if one of the model parameters - the threshold wind speed for starting saltation and lifting dust from the surface - was finely tuned. In this paper we show that the fine-tuning of this parameter could be the result of negative feedback in which processes associated with global dust storms raise the threshold and small-scale processes like dust devils, which are active in years between the storms, lower the threshold. The mechanisms responsible for the increase and decrease in the threshold wind speed could be the removal of sand grains and exposure of non-erodible elements, and gradual re-introduction of sand by small-scale processes and mantling of the non-erodible elements, respectively. We show how the atmosphere/dust system could organize itself to stay close to the critical state, so that global dust storms occur some years but not every year. According to our model, if the climate changed and the peak winds increased, the saltation threshold wind speed would increase so that global dust storms would continue to occur irregularly. The same thing would happen if the peak winds decreased. In both cases the system would continue to exhibit interannual variability. This takes away the mystery of why the system is so finely tuned: It tunes itself. Regardless of the initial conditions, the winds redistribute the sand grains in surface reservoirs until the threshold wind speed is close to the critical value. This explanation of the interannual variability of the global dust storms is not unique and we have only demonstrated this mechanism in the context of a low order model of the circulation. Our aim is not to resolve all the issues related to the interannual variability of global dust storms on Mars entirely, but to stimulate further thinking and follow-on GCM studies.
P31B-0994 0800h
Simulating the Martian Dust Cycle with Prognostic surface dust deposits
Changes in the surface distribution of dust as a result of dust storms have been observed from orbit [Christensen, 1988; Smith et al., 2004] and has been suggested as an important part of interannual variability of global dust storms [Haberle, 1986]. The observations of air and surface temperatures following the 2001 global dust storm provide the first direct evidence for climatic coupling between changes in surface dust and the atmosphere [Smith et al., 2004]. In this work, we examine the impact of finite dust sources on: the seasonal cycle of background dust; on the evolution of dust storms; on the pattern of global dust storm interannual variability; and to examine the response of circulation changes in the surface dust distribution (and hence albedo and surface temperature) following the 2001 global dust storm. The surface dust deposits are introduced in the model as 2 dimensional surface budgets(one for each dust size used). Dust is added to the surface in response to fluxes from the lowest atmospheric level predicted by the sedimentation scheme. Dust is removed from the surface so as to provide the injection amounts predicted by the lifting schemes. If a surface element is exhausted, dust is not injected into the atmosphere regardless of the injection rates required by the lifting schemes. In some cases, we couple the dust amount to the surface albedo using the MGS maps, and results of surface scattering models.