Geophysical Research Letters Supporting Information for Extreme Detached Dust Layers near Martian Volcanoes: Evidence for Dust Transport by Mesoscale Circulations Forced by High Topography N.G. Heavens1,*, B.A. Cantor2, P.O Hayne3, D.M. Kass3, A. Kleinböhl3, D.J. McCleese3, S. Piqueux3, J.T. Schofield3, and J.H. Shirley3 1 Department of Atmospheric and Planetary Sciences, Hampton University, 23 E. Tyler St., Hampton, Virginia, 23669 2 Malin Space Science Systems, San Diego, CA, 92191 3 NASA Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA, 91109 * Corresponding Author’s E-mail: Nicholas.Heavens@hamptonu.edu Contents of this file Text S1 to S7 Figures S1 and S2 Additional Supporting Information (Files uploaded separately) Captions for Tables S1 and S2 Introduction This file contains a description of the version of the Mars Climate Sounder retrievals used in the paper, a short primer on converting density-scaled opacity to mass mixing ratio, a justification for the definition of “extreme detached dust layers”, descriptions of the automated retrieval survey and manual survey of the radiance observations, the reasoning behind our estimate of the formation frequency of layers, references for the supplemental text sections, two supplemental figures, and two supplementary tables providing additional details about the surveys of Mars Climate Sounder data referred to in the paper. 1 Text S1. Mars Climate Sounder Retrieval Database Version 4.3 Mars Climate Sounder retrievals at present consist of vertical profiles of dust opacity at 463 cm-1 (km-1), water ice opacity (km-1) at 842 cm-1, and temperature (K) on pressure (Pa) coordinates retrieved from observations of radiance from the Mars Climate Sounder (MCS) on board Mars Reconnaissance Orbiter (MRO) [McCleese et al., 2007, 2010; Kleinböhl et al., 2009, 2011]. The retrievals also contain height information based on the geometric pointing of the MCS instrument. Version 4.3 of the retrieval dataset differs from previous versions in several respects (A. Kleinböhl et al., No widespread dust in the middle atmosphere of Mars from Mars Climate Sounder observations, submitted to Icarus): (1) where the limb is too opaque to retrieve surface pressure, a climatological surface pressure is assumed; (2) a hard altitude cut-off for aerosol retrieval was removed and replaced with a requirement that the radiance at the maximum retrieved altitude exceed 1% of the maximum radiance in the relevant channel and exceeds 5 times the noise level (this threshold is raised to up to 3% over cold surfaces); and (3) when no on-planet measurements are available to retrieve surface temperature, a climatological surface temperature is assumed but corrected on the basis of the extrapolated column dust optical depth. These changes increase the number of retrieved profiles during the times of year when column dust opacity is high and allows measurement of significant aerosol opacity along a vertical range comparable to the temperature retrievals. Text S2. Density-Scaled Opacity Using the ideal gas law, the ratio between opacity and atmospheric density can be calculated from the retrieved profiles, forming the basis for an estimate of aerosol mass 2 mixing ratio from assumptions about the aerosol properties [Heavens et al., 2011, 2014]. The formula for conversion is: q 4 A reff dz 3 Qext (S1) where density-scaled opacity is the term in brackets, A is the assumed density of the aerosol (3000 kg m-3 for dust; 910 kg m-3 for water ice) and reff and Qext are the effective radius and extinction coefficient for the particle size distribution assumed by the retrieval algorithm. The conversion factors (the unbracketed term on the r.h.s. of Eq. S1) between density-scaled opacity of dust and water ice in m2 kg-1 and mass mixing ratio in ppm are 12000 and 2200 respectively [Heavens et al., 2010, 2011]. The uncertainties in converting density-scaled opacity to mass mixing ratio for water ice are discussed in Heavens et al. [2010]. The uncertainties of the conversion process for dust are discussed in Heavens et al. [2014]. Text S3. EDDL Definition The definition for an “extreme detached dust layer” is based on comparing the vertical dust distribution of Mars’s dayside tropical atmosphere inside and outside of global and regional dust storm activity. The focus on the tropical dust distribution is due to the need to utilize the dust storm catalog of Wang and Richardson [2015], which focuses on lowlatitude dust storm activity. Extreme detached dust layers may occur at higher latitudes. The dust distribution is characterized by two parameters in all individual retrieved profiles within 30º S–30º N from a particular period: (1) the largest dust mass mixing ratio in the profile and (2) the altitude at which it occurs above the areoid. These 3 parameters differ from the magnitude and altitude of a detached dust layer, because there is no guarantee that the largest mass mixing ratio corresponds to a resolved layer. However, it is a good approximation. The vertical dust distribution during four periods was then examined with bivariate histograms of the two parameters (Figure S1). The first period (the 2007 Global Dust Storm) was Ls=261.6º–305.5º of MY 28, during which a global dust storm formed and began to decay [Wang and Richardson, 2015]. The second period was Ls=261.6º– 305.5º of MY 30, a period during which no low-latitude dust storm was observed in visible imagery [Wang and Richardson, 2015]. The third period was Ls=231.4º–237.3º of MY 29 (MY 29A Regional Storm), a period during which a regional dust storm in Acidalia expanded to a maximum area of 1.61 107 km2 [Wang and Richardson, 2015]. Smaller regional dust storms do generate DDLs with properties similar to the highest magnitude and altitude DDLs in global dust storms. Nevertheless, the Acidalia storm was the only non-global storm in the Wang and Richardson [2015] catalog that was both observed by MCS and met the size distribution-based size criterion for regional dust storm proposed by Cantor et al. [2001]. The fourth period was Ls=231.4º–237.3º of MY 30, a period during which no low-latitude dust storm was observed in visible imagery [Wang and Richardson, 2015]. Note that the third period is truncated at the beginning from the period of activity listed in the dust storm catalog in order to avoid overlap with a period in MY 30 with a low-latitude regional dust storm. The vertical dust distributions in the non-dust storm periods are broadly similar. There is a broad peak in the bivariate histograms at a dust mass mixing ratio value of ~10 ppm and altitudes from 15–30 km above the areoid, a broad peak at very low dust values 4 at 40–55 km altitude, and a diffuse bell-shaped part of the distribution centered on ~40 km altitude and peaking at ~50 ppm (Figures S1c–d). The first feature is the High Altitude Tropical Dust Maximum (HATDM) described by Heavens et al. [2011a, 2011b, 2014]. The second feature is due to the cutoff of retrievals at high altitudes due to the limb becoming optically thick because of high dust or water ice opacity. The third feature is not readily identifiable. It may be indicate a population of DDLs intermediate in properties between the extreme population in global dust storms and the HATDM, perhaps the result of local dust storms. However, its origin is irrelevant here. It is not due to regional or global dust storm activity. The bivariate histogram of the two dust profile parameters during the regional dust storm period is similar to the non-dust storm periods (Figure S1b). However, the diffuse bell-shaped part of the distribution is not bell-shaped but expands higher in mass mixing ratio and altitude space. In the global dust storm case, the HATDM part of the distribution is much smaller, either because it is hard to see or no longer exists. Just as in the regional dust storm case, the diffuse bell-shaped part of the distribution is also not bell-shaped and expands even higher in mass mixing ratio and altitude space (Figure S1a). Whatever the exact origin of these distributions, these bivariate histograms show that global and regional dust periods are distinguished from non-dust storm periods in the same seasons of the year by the presence of significant maxima in dust mass mixing ratio at altitudes above 50 km. This criterion provides a threshold for EDDL altitude. The threshold for EDDL magnitude is based on taking the peak column dust opacity observed by the Mars Exploration Rovers during the 2007 global dust event 5 [Montabone et al., 2015], converting it to MCS dust opacity by dividing it by a factor of 7.0 [Montabone et al., 2015], and calculating the uniform mass mixing ratio that would correspond to this column opacity in an isothermal atmosphere with surface pressure of 610 Pa and temperature of 220 K. It is therefore a first guess for the average dust mass mixing ratio in the tropics during a global dust event. Applying both of these thresholds results in 0% of the retrievals meeting the EDDL definition during the non-dust storm periods, 0.5% of the retrievals meeting it during the MY 29A Regional Dust Storm, and 13% of the retrievals meeting it during the 2007 Global Dust Storm (Figure S1), demonstrating that it is a good filter for DDLs associated with regional and global dust storm activity. Text S4. The Automated Survey of the Retrievals The 23,578 retrievals that met the EDDL definition were filtered by automated methods according to four criteria. Criterion 1: to minimize any confusion of dust with carbon dioxide ice, the survey excluded retrieved detached dust layers in air colder than the sum of the CO2 frost point temperature, twice the statistical error in the temperature retrieval, and 20 K (to account for gravity wave effects (see Spiga et al. [2012]). This criterion eliminated 210 retrievals. Criterion 2: the survey excluded retrievals in which water ice opacity exceeds dust opacity in the layer. When water ice opacity is high, the retrieval algorithm may introduce dust to fit errors in water ice spectroscopy [Heavens et al., 2011a, 2014]. This criterion eliminated 134 retrievals. 6 Criterion 3: the survey excluded nightside retrievals (solar zenith angles > 90º). A high mass mixing ratio detached dust layer would be expected to rapidly heat the atmosphere during the day and cool it rapidly at night [Heavens et al., 2011a,b]. Retrievals without a distinct warming feature coincident with the dust layer are reported, as long as temperatures meet Criterion 2. But at night, it is possible that temperatures near the CO2 frost point coincident with an high mass mixing ratio aerosol layer are due to rapid infrared cooling of dust and do not necessarily demonstrate that carbon dioxide ice is condensing. To avoid this ambiguity, nightside retrievals were excluded. This criterion eliminated 766 retrievals. Criterion 4: the survey excluded all retrievals from observations made during regional and global dust storms catalogued by Wang and Richardson [2015] during Mars Years 28–30. This catalog focuses on low latitude dust storm activity, and no comparable peer-reviewed survey yet exists for the succeeding years. This criterion excluded 21,755 retrievals, 21,529 of which were observed during the global dust storm of 2007 as catalogued by Wang and Richardson [2015]. Text S5. Manual Survey of the Radiance Observations The radiance observations were surveyed in two ways. In the first survey, the radiance data in each MCS channel was converted to brightness temperature (or counts for A6) and manually inspected. Note that each MCS Level 1B file contains four hours of calibrated radiance observations. The castellation features were identified and their time of occurrence and location were noted. Table S2 also contains additional information such as whether they are associated with EDDL-containing retrievals in the region of 7 interest and whether there was regional dust storm activity in the area at the same time, as inferred from Wang and Richardson [2015] and the MARCI Weather Reports (http://www.msss.com/msss_images/subject/weather_reports.html). In the second survey, the radiance data associated with all of the EDDLcontaining retrievals not associated with regional or global dust storm activity was manually inspected in the same way as the first survey. It was determined whether the EDDL-containing retrievals overlapped with a castellation feature or a loop feature. In the process, it was discovered that a small number of castellation features had been missed, so the results of the first survey were updated. (The underlying issue in most cases was castellation features spanning two files.) In both surveys, castellation and loop features were required to be apparent in MCS channels A1–A6 in order to identify features with stronger contributions from dust than from condensate clouds. Dust would be expected to warm the atmosphere, resulting in emission features in MCS channels A1–A3 as well as direct emission in A4 and A5 and reflection in A6, which is not necessarily true of water ice or CO2 ice. The castellation feature survey (Table S2) contains three instances in which a limb castellation is in the region of interest, is not contemporaneous with any known regional or global dust storm activity, but is not associated with EDDL-containing retrievals. These instances occur in three different bins of 10º of Ls and suggest that the systematic bias against detecting EDDL-containing retrievals due to problems with retrieving from castellation features is on the order of one orbit per bin. Text S6. Estimating the Frequency of EDDL Formation 8 We estimate their frequency of formation of EDDLs in a given region (e.g., 30 S–60 N, 100–180 W) by estimating the expected fraction of orbits MCS would observe such a layer if one were always present, fE, and then comparing fE to the fraction of orbits MCS actually observed such a layer, fO. If fO < fE, fO/ fE is the fraction of time that a layer is present. If layers do not last multiple sols, fO/ fE is the frequency of layer formation in units of sol-1. In most cases, layers do not last multiple sols. (A possible exception is shown in Figure 4b in the main text.) A dust layer of a few ppm mass-mixing ratio at 60 km above the datum (~0.3 Pa) would be quickly removed by sedimentation. Following the equations of Murphy et al. [1990], a dust particle of 1.06 m radius (the effective radius of dust assumed by the retrieval algorithm) would fall initially at 3.5 ms-1 and be below 40 km in less than a sol. Mesoscale modeling by Spiga et al. [2013] suggests that detached dust layers of 10s of ppm mass mixing ratio may rise during the day due to buoyancy obtained by radiative self-heating, counteracting some of the effects of sedimentation. In addition, smaller dust particles will sediment more slowly. However, horizontal advection should gradually dilute layers. Yet even if layers do last multiple sols, assuming they do not do so results in an upper bound for frequency. We estimate fE under the assumption that the average longitudinal dimension of a layer is on the order of 1000 km. In 17 cases (making up 120 of the 211 EDDLcontaining retrievals that are outside of dust storm activity and in the ROI), layers were observed in successive orbits, arguing that their longitudinal dimension sometimes exceeds 1500 km. However, there are 33 cases in which layers in the ROI were not observed on successive orbits. Mesoscale modeling suggests a typical longitudinal 9 dimension of 1000 km is plausible [Rafkin et al., 2002; Michaels et al., 2006]. Assuming the layer is randomly distributed in longitude and using the fact that the average longitudinal extent of the region is 4000 km, it is found that MCS has a 25% probability of observing a layer during each MRO orbit. The parameter fO, is then the fraction of orbits on which EDDL-containing retrievals were observed (Figure 2f), such that the formation rate of EDDLs is simply 4 fO sol-1, where fO is expressed as a fraction rather than as a percentage. Text S7. References Cantor, B.A. (2007), MOC observations of the 2001 Mars planet-encircling dust storm, Icarus, 186, 60–96. Heavens, N. G., J. L. Benson, D. M. Kass, A. Kleinböhl, W. A. Abdou, D. J. McCleese, M. I. Richardson, J. T. Schofield, J. H. Shirley, and P. M. Wolkenberg (2010), Water ice clouds over the Martian tropics during northern summer, Geophys. Res. Lett., 37, L18202, doi:10.1029/2010GL044610. Heavens, N. G., M. I. Richardson, A. Kleinböhl, D. M. Kass, D. J. McCleese, W. Abdou, J. L. Benson, J. T. Schofield, J. H. Shirley, and P. M. Wolkenberg (2011a), The vertical distribution of dust in the Martian atmosphere during northern spring and summer: Observations by the Mars Climate Sounder and analysis of zonal average vertical dust profiles, J. Geophys. Res., 116, E04003, doi:10.1029/2010JE003691. Heavens, N. G., M. I. Richardson, A. Kleinböhl, D. M. Kass, D. J. McCleese, W. Abdou, J. L. Benson, J. T. Schofield, J. H. Shirley, and P. M. Wolkenberg (2011b), Vertical distribution of dust in the Martian atmosphere during northern spring and summer: Highaltitude tropical dust maximum at northern summer solstice, J. Geophys. Res., 116, E01007, doi:10.1029/2010JE003692. Heavens, N. G., M. S. Johnson, W. A. Abdou, D. M. Kass, A. Kleinböhl, D. J. McCleese, J. H. Shirley, and R. J. Wilson (2014), Seasonal and diurnal variability of detached dust layers in the tropical Martian atmosphere, J. Geophys. Res. Planets, 119, 1748–1774, doi:10.1002/2014JE004619. Kleinböhl , A., J. T. Schofield, D. M. Kass, W. A. Abdou, C. R. Backus, B. Sen, J. H. Shirley, W. G. Lawson, M. I. Richardson, F. W. Taylor, N. A. Teanby, and D. J. McCleese (2009), Mars Climate Sounder limb profile retrieval of atmospheric 10 temperature, pressure, dust, and water ice opacity, J. Geophys. Res., 114, E10006, doi: 10.1029/2009JE003358. Kleinböhl, A., J. T. Schofield, W. A. Abdou, P. G. J. Irwin, R. J. de Kok (2011), A single-scattering approximation for infrared radiative transfer in limb geometry in the Martian atmosphere, J. Quant. Spectrosc. Radiat. Transfer, 112, 1568–1580. Kleinböhl, A., R.J. Wilson, D. Kass, J. T. Schofield, and D. J. McCleese (2013), The semidiurnal tide in the middle atmosphere of Mars, Geophys. Res. Lett., 40, 1952–1959, doi:10.1002/grl.50497. McCleese, D. J., J. T. Schofield, F. W. Taylor, S. B. Calcutt, M. C. Foote, D. M. Kass, C. B. Leovy, D. A. Paige, P. L. Read, and R. W. Zurek (2007), Mars Climate Sounder: An investigation of thermal and water vapor structure, dust and condensate distributions in the atmosphere, and energy balance of the polar regions, J. Geophys. Res., 112, E05S06, doi:10.1029/2006JE002790. McCleese, D. J., N.G. Heavens, J.T. Schofield, W.A. Abdou, J.L. Bandfield, S.B. Calcutt, P.G.J. Irwin, D.M. Kass, A. Kleinböhl, S.R. Lewis, D.A. Paige, P.L. Read, M.I. Richardson, J.H. Shirley, F.W. Taylor, N. Teanby, R.W. Zurek (2010), Structure and dynamics of the Martian lower and middle atmosphere as observed by the Mars Climate Sounder: Seasonal variations in zonal mean temperature, dust, and water ice aerosols, J. Geophys. Res. Planets, 115, E12016, doi:10.1029/2010JE003677. Michaels, T. I., A. Colaprete, and S. C. R. Rafkin (2006), Significant vertical water transport by mountain-induced circulations on Mars, Geophys. Res. Lett., 33, L16201, doi:10.1029/2006GL026562. Montabone, L., F. Forget, E. Millour, R.J. Wilson, S.R. Lewis, B. Cantor, D. Kass, A. Kleinböhl, M.T. Lemmon, M.D. Smith, M.J. Wolff, (2015), Eight-year Climatology of Dust Optical Depth on Mars, Icarus, in press, doi: 10.16/j.icarus.2014.12.034. Rafkin, S. C. R., M. R. V. Sta. Maria, and T. I. Michaels (2002), Simulation of the atmospheric thermal circulation of a martian volcano using a mesoscale numerical model. Nature, 419, 697-699. Spiga, A., F. González-Galindo, M.-Á. López-Valverde, and F. Forget (2012), Gravity waves, cold pockets and CO2 clouds in the Martian mesosphere, Geophys. Res. Lett., 39, L02201, doi:10.1029/2011GL050343. Spiga, A., J. Faure, J.-B. Madeleine, A. Määttänen, and F. Forget (2013), Rocket dust storms and detached dust layers in the Martian atmosphere, J. Geophys. Res. Planets, 118, 746–767, doi:10.1002/jgre.20046. 11 Wang, H. and M.I. Richardson (2015), The origin, evolution, and trajectory of large dust storms on Mars during Mars years 24–30 (1999–2011), Icarus, 251, 112–127, doi: 101.1016/j.icarus.2013.10.033. 12 Figure S1. Bivariate histogram (% of total retrievals) of the highest value of mass mixing ratio in a retrieval and the altitude above the areoid at which it occurs in all dayside retrievals from 30° S–30° N during: (a) Ls=261.6º–305.5º of MY 28; (b) Ls=231.4º–237.3º of MY 29; (c) Ls=261.6º–305.5º of MY 30; (d) Ls=231.4º–237.3º of MY 30. Binning is in intervals of 1 ppm and 1 km. The red lines mark the thresholds for the EDDL definition. The percentage of retrievals that meet the EDDL definition is given in the titles. 13 Figure S2. A dust cloud over Arsia Mons observed by MRO-MARCI at Ls=177.05° of MY 29 (B04_011255_1771_MA_00N121W, observed around 12:00 UTC on 20 December 2008). The data has been calibrated and photometrically corrected as in Cantor [2007] to emphasize the dust cloud and projected to 1 km resolution in simple cylindrical coordinates. Table S1. Results of the Manual Survey of Extreme Detached Dust Layers (attached as a separate file). Table lists individual orbits with EDDL-containing retrievals. Latitudes and longitudes are of the approximate center of the EDDL-containing retrievals. The magnitude of the layer is defined as the highest magnitude EDDL-containing retrieval in the orbit, while the altitude is defined as the altitude at which the maximum dust mass mixing ratio (the magnitude) occurs in the highest magnitude retrieval in the orbit. Orbits with retrievals identified as being associated with regional dust storms are marked in red. Some of these instances span multiple orbits and so ranges of parameters are provided. Orbits with EDDL-containing retrievals observed synchronously with dust storm activity 14 over the volcanoes are marked in yellow. As in the main text, the Region of Interest is defined as 30º S–60º N, 100º –180º W. When volcanoes are referred to, only Olympus, Ascraeus, Pavonis, and Arsia Montes are meant. Table S2. Results of the Castellation Feature Survey of MRO-MCS Calibrated Radiance Observations (attached as separate file). As in the main text, the Region of Interest is defined as 30º S–60º N, 100º –180º W. Latitudes and longitudes are of the approximate centers of the features at the tangent point of the observation. Castellation features that are in the Region of Interest, but are neither associated with regional or global dust storm activity nor with EDDL-containing retrievals included in Table S1 are in bold. Further details concerning the compilation of the table are in Text S6. 15