Search

In recent years we increased the typical maximum peak field on axis gradients obtained in L-band superconducting RF (SRF) photoinjector cavities at vertical tests to around 55 MV/m. This was achieved with niobium cathode plugs directly screwed to the cavity back wall omitting an RF choke filter and a load lock system for cathodes. Copper demonstrated being a suitable cathode material in normal conducting injector cavities used at X-Ray Free Electron Lasers (XFELs) operating with pulsed RF. In this article we present the first experimental confirmation that peak field on axis gradients around 55 MV/m and beyond can be achieved in L-band SRF photoinjector cavities with copper cathode plugs screwed to the cavity back wall. We view this as a major milestone for the development of a high gradient photoinjector operating continuous wave (CW)., Comment: 4 pages, 7 figures, to be submitted to PRAB

Aeolian sediment transport is observed to occur on Mars as well as other extraterrestrial environments, generating ripples and dunes as on Earth. The search for terrestrial analogues of planetary bedforms, as well as environmental simulation experiments able to reproduce their formation in planetary conditions, are powerful ways to question our understanding of geomorphological processes towards unusual environmental conditions. Here, we perform sediment transport laboratory experiments in a closed-circuit wind tunnel placed in a vacuum chamber and operated at extremely low pressures to show that Martian conditions belong to a previously unexplored saltation regime. The threshold wind speed required to initiate saltation is only quantitatively predicted by state-of-the art models up to a density ratio between grain and air of $4 \times 10^5$, but unexpectedly falls to much lower values for higher density ratios. In contrast, impact ripples, whose emergence is continuously observed on the granular bed over the whole pressure range investigated, display a characteristic wavelength and propagation velocity essentially independent of pressure. A comparison of these findings with existing models suggests that sediment transport at low Reynolds number but high grain-to-fluid density ratio may be dominated by collective effects associated with grain inertia in the granular collisional layer., Comment: 38 pages, 21 figures

The objective is to establish an empirical relationship between measurements of radar, aeolian, and surface roughness on a variety of natural surfaces and to understand the underlying physical causes. This relationship will form the basis for developing a predictive equation to derive aeolian roughness from radar backscatter. Results are given from investigations carried out in 1989 on the principal elements of the project, with separate sections on field studies, radar data analysis, laboratory simulations, and development of theory for planetary applications.

The interaction between winds and desert surfaces has important implications for sediment transport on Earth, Mars, and Venus, and for understanding the relationships between radar backscatter and aerodynamic roughness as part of the NASA Shuttle Imaging radar (SIR-C) Mission. Here, researchers report results from measurements of boundary layer wind profiles and surface roughness at sites in Death Valley and discuss their implications. The sites included a flat to undulating gravel and sand reg, alluvial fans, and a playa. Estimates of average particle size composition of Death Valley sites and arithmetic mean values of aerodynamic roughness are given in tabular form.

The interaction between wind and desert surfaces has important implications for sediment transport on Earth, Mars, and Venus, and for understanding the relationship between radar backscatter and aerodynamic roughness. Here, researchers report results from measurements of atmospheric boundary layer profiles, assessment of radar backscatter at P, L, and C wavelengths, and surface roughness in Death Valley, the Mojave Desert, and Lunar Lake, NV, and discuss the implications for aeolian process. The sites include playas, gravel and sand regs, alluvial fans, and lava flows. Boundary layer wind profiles were measured using anemometers at heights of 0.75, 1.25, 2.07, 3.44, 5.72, and 9.5 m; temperature sensors at heights of 1.3 and 9.6 m; and wind vanes at 9.7 and 1.5 m. Microtopographic measurements were made using a template and a laser-photo device to obtain RMS height. This study demonstrates that radar backscatter coefficients obtained from airborne and perhaps orbiting instruments could permit the derivation of aerodynamic roughness values for large areas. Such values, when combined with wind frequency data, could enable assessment of aeolian processes on a regional scale.

The aeolian wind tunnel is a special case of a larger subset of the wind tunnel family which is designed to simulate the atmospheric surface layer winds to small scale (a member of this larger subset is usually called an atmospheric boundary layer wind tunnel or environmental wind tunnel). The atmospheric boundary layer wind tunnel is designed to simulate, as closely as possible, the mean velocity and turbulence that occur naturally in the atmospheric boundary layer (defined as the lowest portion of the atmosphere, of the order of 500 m, in which the winds are most greatly affected by surface roughness and topography). The aeolian wind tunnel is used for two purposes: to simulate the physics of the saltation process and to model at small scale the erosional and depositional processes associated with topographic surface features. For purposes of studying aeolian effects on the surface of Mars and Venus as well as on Earth, the aeolian wind tunnel continues to prove to be a useful tool for estimating wind speeds necessary to move small particles on the three planets as well as to determine the effects of topography on the evolution of aeolian features such as wind streaks and dune patterns.

In the study of planetary aeolian processes the effect of gravity is not readily modeled. Gravity appears in the equations of particle motion along with interparticle forces but the two terms are not separable. A wind tunnel that would permit variable gravity would allow separation of the forces and aid greatly in understanding planetary aeolian processes. The design Carousel Wind Tunnel (CWT) allows for a long flow distance in a small sized tunnel since the test section is a continuo us circuit and allows for a variable pseudo gravity. A prototype design was built and calibrated to gain some understanding of the characteristics of the design and the results presented.

The Carousel wind tunnel (CWT) proposed to study aeolian processes aboard a space station consists of two concentric rotating drums. The space between the two drums comprises the wind tunnel test section. Differential rates of rotation of the two drums would provide a wind velocity with respect to either drum surface. Preliminary results of measured velocity profiles made in a CWT prototype indicate that the wall bounded boundary layer profiles are suitable to simulate flat plate turbulent boundary layer flow. The two dimensional flat plate Cartesian coordinate equations of motion of a particle moving through the air are explained. In order to assess the suitability of CWT in the analysis of the trajectories of windblown particles, a series of calculations were conducted comparing cases for gravity with those of zero gravity. Results from the calculations demonstrate that a wind tunnel of the carousel design could be fabricted to operate in a space station environment and that experiments could be conducted which would yield significant results contributing to the understanding of the physics of particle dynamics.

One of the important parameters in the analysis of sediment entrainment and transport is gravitational attraction. The availability of a laboratory in Earth orbit would afford an opportunity to conduct experiments in zero and variable gravity environments. Elimination of gravitational attraction as a factor in such experiments would enable other critical parameters (such as particle cohesion and aerodynamic forces) to be evaluated much more accurately. A Carousel Wind Tunnel (CWT) is proposed for use in conducting experiments concerning sediment particle entrainment and transport in a space station. In order to test the concept of this wind tunnel design a one third scale model CWT was constructed and calibrated. Experiments were conducted in the prototype to determine the feasibility of studying various aeolian processes and the results were compared with various numerical analysis. Several types of experiments appear to be feasible utilizing the proposed apparatus.

A method is presented for predicting the geometry of a vortex sheet trailing a lift generating wing in its vicinity. It differs from others in that it uses a continuous vorticity distribution rather than discrete vortex filaments. It was found that the results were in good agreement with available experimental data, after the initial cycle of what was to be a relaxation scheme, so that iterations for this end were unnecessary. While the technique is computing intensive, it becomes attractive when used in conjunction with panel methods applications for complete aircraft configurations. The panel method solution needs to be found only once for each streamwise station downstream of the wing. Corrections applied are geometric in nature and are independent of other computational aspects.

The thickness growth characteristics of flowing lava are investigated using a heat balance model and a two-dimensional model for flow of a Bingham plastic fluid down an inclined plane. It is found that yield strength plays a crucial role in the thickening of a lava flow of given flow rate. To illustrate this point, downstream thickness profiles and yield strength distributions were calculated for flows with mass flow rates of 10,000 and 100,000 kg/m-sec. Higher flow rates led to slow cooling rates which resulted in slow rate of increase of yield strength and thus greater flow lengths.

The mechanism determining the length of dark erosional (type-Id) streaks near craters and hills on the surface of Mars is investigated using Mariner-9 images and the results of wind-tunnel studies (Greeley et al., 1974; Iversen et al., 1975 and 1976). An empirical model of time-dependent growth of crater-wake streaks in the downwind direction involving deflation by sand-grain saltation is developed and shown to give relatively good agreement with streak growth rates observed over a 38-day period of Mariner observations and with the streak-length/crater-diameter and streak-length/crater-height ratios analyzed statistically by Lee (1984). This time-dependent mechanism is thus considered a reasonable alternative (for dark streaks) to the blocking model proposed by Lee for both bright and dark streaks. Preliminary analysis of further wind-tunnel experiments (Iversen et al., 1982) supports the blocking model of bright-streak formation.

Small particles and winds of sufficient strength to move them have been detected from Venera and Pioneer-Venus data and suggest the existence of aeolian processes on Venus. The Venus wind tunnel (VWT) was fabricated in order to investigate the behavior of windblown particles in a simulated Venusian environment. Preliminary results show that sand-size material is readily entrained at the wind speeds detected on Venus and that saltating grains achieve velocities closely matching those of the wind. Measurements of saltation threshold and particle flux for various particle sizes have been compared with theoretical models which were developed by extrapolation of findings from Martian and terrestrial simulations. Results are in general agreement with theory, although certain discrepancies are apparent which may be attributed to experimental and/or theoretical-modeling procedures. Present findings enable a better understanding of Venusian surface processes and suggest that aeolian processes are important in the geological evolution of Venus.

A feasibility study was undertaken to determine if a suitable apparatus could be designed to analyze aeolian processes for operation in space and to assess the feasibility of conducting meaningful experiments to address key aspects of aeolian processes. To meet this objective a prototype apparatus was fabricated and some limited experiments were run to determine its suitability for this application. At least three general types of experiments were devised that could be carried out aboard a space station: threshold studies, swirl (dust devil) experiments, and analyses of windblown particle trajectories. How experiments in a zero-g environment could advance knowledge of aeolian processes was studied.

Forty-six different fins, which were members of twelve plan-form families, were tested. A two dimensional Boeing single element airfoil at an angle of attack of eight degrees and a sweepback angle of thirty-two was used to simulate a portion of the wing of a generator aircraft. Various free stream velocities were used to test any individual fin at its particular angle of attack. While the fin itself was mounted on the upper surface of the generator model, the angle of attack of each fin was varied until stall was reached and/or passed. The relative fin vortex strengths were measured in two ways. First, the maximum angular velocity of a four blade rotor placed in the fin vortex center was measured with the use of a stroboscope. Second, the maximum rolling moment on a following wing model placed in the fin vortex center was measured by a force balance.

A similitude parameter is derived which is based on theoretical considerations of erosion due to sand in saltation. This parameter has been used to correlate wind tunnel experiments of particle flow over model craters. The characteristics of the flow field in the vicinity and downstream of a crater are discussed and it is shown that erosion is initiated in areas lying under a pair of trailing vortices. The erosion rate parameter is used to calculate erosion rates on Mars, reported in Part 2, to be published later.

The Mariner 9 photographs of the Martian surface include many which show crater-associated streaks caused by atmospheric winds. Due to the low density atmosphere of Mars, the wind speeds necessary to move surface particles and thus cause the streaks are apparently very high. The atmospheric-boundary-layer wind tunnel has been used to determine threshold speeds for particles of varying density and diameter and the results for small particles have been extrapolated to estimate somewhat lower threshold speeds on Mars than previous estimates. A series of streak modeling tests is used to derive an erosion-rate correlation function, which is in turn used to estimate erosion rates to the lee of some of the craters on Mars which exhibited time-dependent streak behavior during the mission.

The aeolian mechanics of saltating grains on Mars is studied using both an experimental investigation in an atmospheric wind tunnel and numerical solutions to the equations of motion of a single grain under Martian surface conditions. Numerical solutions for terrestrial conditions are empirically correlated with experimental data for Mars, and the modified equations are solved to estimate Martian grain motion. The calculations demonstrate the importance of a lifting force in initiating grain motion on both planets. It is found that Martian grain-trajectory length scales are longer than those on earth and fall with a higher terminal grain velocity. The wind-tunnel experiments show that saltating grains have an enhanced erosive capability in crater wakes. It is suggested that this enhanced capability may well explain the streaks associated with craters in certain regions on the Martian surface.

To study the eolian mechanics of saltating particles, both an experimental investigation of the flow field around a model crater in an atmospheric boundary layer wind tunnel and numerical solutions of the two- and three-dimensional equations of motion of a single particle under the influence of a turbulent boundary layer were conducted. Two-dimensional particle motion was calculated for flow near the surfaces of both Earth and Mars. For the case of Earth both a turbulent boundary layer with a viscous sublayer and one without were calculated. For the case of Mars it was only necessary to calculate turbulent boundary layer flow with a laminar sublayer because of the low values of friction Reynolds number; however, it was necessary to include the effects of slip flow on a particle caused by the rarefied Martian atmosphere. In the equations of motion the lift force functions were developed to act on a single particle only in the laminar sublayer or a corresponding small region of high shear near the surface for a fully turbulent boundary layer. The lift force functions were developed from the analytical work by Saffman concerning the lift force acting on a particle in simple shear flow.

The plateau region in which trailing vortices initially exhibit slow decay of maximum tangential speed is demonstrated with the aid of constant and variable eddy viscosity models. The existence of the plateau is explained by the presence of the viscous core and by the necessity for transition of the initial, nearly inviscid profile to similarity far downstream. Span loading is shown to affect the initial maximum tangential speed as well as the duration of the plateau region. Triangular span loading is shown to present less hazard to trailing aircraft, at least in the near field, than elliptic loading because the inviscid profile for triangular loading exhibits a finite tangential speed at the vortex centerline and a large inviscid core.

In order to determine the nature of Martian aeolian processes, an investigation is in progress which involves wind tunnel simulations, geologic field studies, theoretical model studies, and analyses of Mariner 9 imagery; this report presents the preliminary results. Threshold speed experiments were conducted for particles ranging in relative density from 1.3 to 11.35 and diameter from 10.2 to 1290 microns to verify and better define Bagnold's (1941) expression for grain movement, particularly for low particle Reynolds numbers and to study the effects of aerodynamic lift and surface roughness. Wind tunnel simulations were conducted to determine the flow field over raised rim craters and associated zones of deposition and erosion.

The characteristics of the plateau region in the vortex system which trails from a lifting wing are discussed. The decay of the vortex due to viscous or turbulent shear is very slow in the plateau so that the maximum tangential speed in the vortices remains nearly constant for some distance downstream of roll-up and then begins to decrease, becoming inversely proportional to the square root of the distance downstream. Mathematical models are developed to analyze the structure of the plateau area. Solutions are obtained for both constant and variable eddy viscosity models.

Wind tunnel experiments have revealed a characteristic flow field pattern over raised-rim craters which causes distinctive zones of aeolian erosion and deposition. Comparisons of the results with Mariner 9 images of Mars show that some crater-associated dark zones result from wind erosion and that some crater-associated light streaks are depositional.

Wind friction threshold speeds for particle movement were determined in a wind tunnel operating at martian surface pressure with a 95 percent CO2 and 5 percent air atmosphere. The relationship between friction speed and free-stream velocity is extended to the critical case for Mars of momentum thickness Reynolds numbers between 425 and 2000. It is determined that the dynamic pressure required to initiate saltation is nearly constant for pressures between 1 bar and 4 mb for atmospheres of both air and CO2.

The interaction and eventual merger of corotational vortices and the decay of a single vortex have been studied by employing zero-, one- and two-equation turbulent-flow models in order to gain a better understanding of the role of turbulence. An implicit finite-difference procedure is used to integrate the unsteady, two-dimensional equations in a cross-plane. The zero- and one-equation formulations utilize a mixing-length model, which incorporates the streamline curvature effect by prescribing a spatially-varying mixing-length. In the two-equation model, the turbulence kinetic energy equation and a modified rate of dissipation equation which includes a streamline curvature correction are solved. Computational results of different models applied to various flow-configurations are presented and compared with available experimental data whenever possible.

A computational study has been conducted to gain a better understanding of the role of turbulence in the merger of corotating vortices and in the decay of a single vortex. The flow field is simulated by using an inviscid flow model and zero-, one-, and two-equation turbulent-flow models. An implicit finite-difference procedure is employed to integrate the unsteady, two-dimensional, incompressible flow equations in a cross-plane. In the zero- and one-equation models, a spatially varying mixing-length model that simulates the streamline curvature effect is used to define an eddy viscosity. The computational merging distance criterion is in good agreement with the experimental one. In the case of the two-equation model, the turbulent kinetic energy and a modified rate of dissipation equation incorporating a curvature correction term are solved. A comparative study of the computational results corresponding to the various flow models and the available experimental data is presented.

The merger of two corotating vortices was studied with a laser velocimeter designed to measure the two cross-stream components of velocity. Measurements were made at several downstream distances in the vortex wake shed by two semispan wings mounted on the wind-tunnel walls. The velocity data provided wall-defined contours of crossflow velocity, stream function, and vorticity for a variety of test conditions. Downstream of the merger point, the vorticity was found to be independent of the downstream distance for radii smaller than r/b = 0.05. For larger radii, the vorticity depended on the distance from the wing. Upstream of the merger, a multicell vorticity pattern was found.

Single and multiple trailing vortices shed from semi-span wings and a transport model in a wind tunnel were studied by means of a laser-velocimeter, hot-wire anemometer, and a trailing model incorporating a 6-component force balance. Velocity profile and turbulence data from the laser-velocimeter and hot-wire anemometer are presented and shown to compare well with the Betz inviscid circulation model. Lift and rolling moment measurements on the following model are compared with those predicted from the flow field measurements.

The merger of two vortices was studied with a laser velocimeter designed to measure the two cross-stream components of velocity. Measurements were made at several downstream distances in the vortex wake shed by two semispan wings mounted on the wind tunnel walls. The velocity data provided well difined contours of cross-flow velocity, stream function and vorticity. Downstream of the merger point the vorticity was shown to be independent of the downstream distance for small radii, and at larger radii was dependent on the distance from the wing rather than from the merger point. Upstream of the merger point a multicell vorticity pattern was shown.

A computational study has been conducted to gain a better understanding of the criteria governing the merger of corotating vortices. The zero-equation and one-equation turbulent-flow models are used to model the interaction of two rolled-up, corotational vortices. A model appropriate for the swirling flow under consideration has been proposed to prescribe the algebraic variation of mixing-length in the flow field. The unsteady, two-dimensional equations in a cross-plane are solved for primitive variables using an implicit finite-difference code. The computed downstream merging distances are in good agreement with available wind tunnel data.

A merging-distance criterion for equal-strength corotational vortices is derived from low-turbulence wind-tunnel flow-visualization data. The vortex separation distance is normalized by defining a vortex core diameter based on circulation defect and angular-momentum defect. Merging may take place for larger separation distances than predicted from earlier two-dimensional inviscid calculations, which indicates that viscosity and possibly three-dimensional effects are important factors in the merging phenomenon. Hot-wire velocity distributions and rolling-moment measurements show that attenuation of the vortex hazard is associated with vortex merging.

Attempts to predict threshold speeds for windblown dust and sand on Mars and Venus have raised new questions about the mechanism of soil and sand transport by wind. Estimates of threshold speeds on Mars and Venus are made including the effects of both Reynolds number and interparticle forces of cohesion. The current estimates show lower threshold speeds for small particles than previous estimates by Hess, who assumed that Reynolds number is the only important parameter. These estimates, on the other hand, show somewhat higher threshold speeds than those of Sagan and Bagnold, who assumed that the particles are cohesionless. An optimum particle size results, for which threshold speed is a minimum, contrary to Sagan and Bagnold's prediction.

Results are reported for wind-tunnel experiments performed to determine the threshold friction speed of particles with different densities. Experimentally determined threshold speeds are plotted as a function of particle diameter and in terms of threshold parameter vs particle friction Reynolds number. The curves are compared with those of previous experiments, and an A-B curve is plotted to show differences in threshold speed due to differences in size distributions and particle shapes. Effects of particle diameter are investigated, an expression for threshold speed is derived by considering the equilibrium forces acting on a single particle, and other approximately valid expressions are evaluated. It is shown that the assumption of universality of the A-B curve is in error at very low pressures for small particles and that only predictions which take account of both Reynolds number and effects of interparticle forces yield reasonable agreement with experimental data. Effects of nonerodible surface roughness are examined, and threshold speeds computed with allowance for this factor are compared with experimental values. Threshold friction speeds on Mars are then estimated for a surface pressure of 5 mbar, taking into account all the factors considered.

Mariner 9 spacecraft images showing evidence of variable surface features and surface erosion resulting from atmospheric wind on Mars have caused a renewed interest in the eolian mechanics of saltating grains. To study this phenomenon, both experimental investigation in an atmospheric wind tunnel and numerical solutions of the equations of motion of a single grain under Martian surface conditions were conducted. The numerical solutions for earth were used and empirically adjusted to correlate with existing experimental data for Mars. These modified equations were then solved to estimate grain motion for Mars. These calculations show the importance of a lifting force on the grain to initiate motion in both earth and Mars calculations. Major findings include a comparison of earth and Mars grain trajectories that show Mars length scales to be longer and to fall with a higher terminal grain velocity. The grains in the Mars calculation also made a smaller collision angle with the surface on rebound.

Threshold wind speeds for setting particles into motion on Mars are estimated by evaluating experimentally observed threshold friction velocities and determining the ratio of this velocity to the threshold wind speed at the top of earth's atmospheric boundary layer (ABL). Turning angles between the direction of the wind at the top of the ABL and the wind stress at the surface are also estimated. Detailed consideration is given to the dependence of the threshold wind speed at the top of the ABL on particle diameter, surface pressure, air temperature, atmospheric stability and composition, surface roughness, and interparticle cohesion. The results are applied to interpret a number of phenomena that have been observed on Mars and are attributable to aeolian processes. It is shown that: (1) minimum threshold wind speeds of about 50 to 100 m/sec are required to cause particle motion on Mars under 'favorable' conditions; (2) particle motion should be infrequent and strongly correlated with proximity to small topographical features; (3) in general, particle motion occurs more readily at night than during the day, in winter polar areas than equatorial areas around noon, and for H2O or CO2 ice particles than for silicate particles; and (4) the boundary between saltating and suspendible particles is located at a particle diameter of about 100 microns.

Theoretical mass flow rate and the particle trajectory equations of motion of granular material in saltation are used to correlate experimental data for the rate of erosion in the wake regions of wind-tunnel-model Martian craters. Vertical geometric distortion is inherent in the simulation because of the equivalent roughness height characteristic of a turbulent boundary layer, which is affected by material in saltation. It is thus necessary to distort topographic model geometry in the vertical direction. A systematic similitude which is based on erosion rate and equivalent roughness in saltation is shown to correlate time-dependent model data as long as the model Reynolds number is higher than a critical value.

Wind friction threshold speeds for particle movement were determined in a low pressure boundary layer wind tunnel at an atmospheric pressure of 5.3 mb. The results imply that for comparable pressures on Mars, the minimum wind friction threshold speed is about 2.5 m/sec, which would require free-stream winds of 50 to 135 m/sec, depending on the character of the surface and the atmospheric conditions. The corresponding wind speeds at the height of the Viking lander meteorology instrument would be about a factor of two less than the free-stream wind speed. The particle size most easily moved by winds on Mars is about 160 microns; particles both larger and smaller than this (at least down to about 5 microns) require stronger winds to initiate movement.

A correlation function, derived on the basis of self-similar variable eddy-viscosity decay, is introduced and utilized to correlate aircraft trailing vortex velocity data from ground and flight experiments. The correlation function collapses maximum tangential velocity data from scale-model and flight tests to a single curve. The resulting curve clearly shows both the inviscid plateau and the downstream decay regions. A comparison between experimental data and numerical solution shows closer agreement with the variable eddy viscosity solution than the constant viscosity analytical solution.

A correlation function, derived on the basis of self similar variable eddy viscosity decay, is introduced and utilized to correlate aircraft trailing vortex velocity data from ground and flight experiments. The correlation function collapses maximum tangential velocity data from scale model and flight tests to a single curve. The resulting curve clearly shows both the inviscid plateau and the downstream decay regions. A comparison between experimental data and numerical solution shows closer agreement with the variable eddy viscosity solution than the constant viscosity analytical solution.

Preliminary results are reported of an investigation which involves wind tunnel simulations, geologic field studies, theoretical model studies, and analyses of Mariner 9 imagery. Threshold speed experiments were conducted for particles ranging in specific gravity from 1.3 to 11.35 and diameter from 10.2 micron to 1290 micron to verify and better define Bagnold's (1941) expressions for grain movement, particularly for low particle Reynolds numbers and to study the effects of aerodynamic lift and surface roughness. Wind tunnel simulations were conducted to determine the flow field over raised rim craters and associated zones of deposition and erosion. A horseshoe vortex forms around the crater, resulting in two axial velocity maxima in the lee of the crater which cause a zone of preferential erosion in the wake of the crater. Reverse flow direction occurs on the floor of the crater. The result is a distinct pattern of erosion and deposition which is similar to some martian craters and which indicates that some dark zones around Martian craters are erosional and some light zones are depositional.