Skip to main content
NCBI home page
NASA Author Manuscripts logoLink to NASA Author Manuscripts
. Author manuscript; available in PMC: 2019 Jul 29.
Published in final edited form as: J Geophys Res Planets. 2018 Jun 19;123(7):1732–1745. doi: 10.1029/2018JE005623

Cross-Equatorial Flushing Dust Storms and Northern Hemisphere Transient Eddies: An Analysis for Mars Year 24

H Wang 1
PMCID: PMC6662229  NIHMSID: NIHMS1029892  PMID: 31360622

Abstract

The relationship between frontal/flushing dust storms and northern hemisphere synoptic period transient eddies in Mars year 24 is examined in this paper. Frontal dust storms are observed roughly continuously during the presolstice (early/middle fall) and postsolstice (middle/late winter) time periods, but flushing dust storms that cross the equator are confined to shorter seasonal windows on both sides of the solsticial pause. In the lower atmosphere, the timing of cross-equatorial flushing dust storms correlates better with eddy temperature than with eddy meridional wind; in the middle atmosphere, it correlates better with eddy meridional wind than with eddy temperature. This is because both the lower atmosphere eddy temperature and the middle atmosphere eddy meridional wind are dominated by zonal wave number m = 3 eastward traveling waves during the cross-equatorial flushing dust storm periods. Frontal dust storms do not seem to be limited to any particular wave mode, but cross-equatorial flushing dust storms appear to be closely related to m = 3 eastward traveling waves, at least in Mars year 24. The effectiveness of m = 3 waves in this regard is partially due to their amplitudes but more importantly due to their seasonal distributions and latitudinal positions. During the time periods when m = 3 waves are strong, the m = 3 waves are also located at lower latitudes, closer in distance to the fairly strong southward mean meridional wind in the low latitudes. Dust in frontal dust storms at high latitudes can be easily entrained into the low-latitude circulation and be efficiently transported southward.

1. Introduction

A dust storm on Mars can attain planetary scale, making it an impressive meteorological phenomenon. Planet-encircling dust storms occur during the dust storm season of northern fall and winter (Martin & Zurek, 1993; Montabone et al., 2015). They are associated with dramatic changes in the thermal structure and circulation of the Martian atmosphere (Kass et al., 2016; Liu et al., 2003). Moreover, these major dust storms pose significant concern for the Entry-Descent-Landings and surface operation of Mars missions. Therefore, it is important to understand how they initiate and interact with the circulation.

During Mars Years (MY) 24–30, northern hemisphere-originated major dust storms occurred more often than southern hemisphere-originated ones, though both types showed rapid zonal expansion in the southern hemisphere (Wang & Richardson, 2015). The initiation phases of northern hemisphere-originated major dust storms usually involve flushing dust storms. Flushing dust storms transport dust from northern middle/high latitudes to low latitudes through confined longitudinal corridors that connect with the northern storm zones in Acidalia, Utopia, and Arcadia (Hollingsworth et al., 1996). They often originate as frontal dust storms that have curvilinear shapes and significant temperature perturbations indicative of their association with midlatitude/high-latitude baroclinic eddies (Hinson & Wang, 2009). These eddies are characterized by synoptic variations whose wave periods (P) are between 2 and 30 sols (Banfield et al., 2004; Barnes, 1980). Their far-reaching influence was detected at Gale crater in the southern hemisphere (Haberle et al., 2018). The link between major dust storms and northern hemisphere transient eddies is the motivation for this study.

Previous studies established that the temporal distribution of northern hemisphere frontal/flushing dust storms generally agrees with the temporal variation of near-surface transient eddies in northern fall and winter (Hinson et al., 2012). Both exhibit a solsticial pause—a minimum around the northern winter solstice (Lee et al., 2018; Lewis et al., 2016; Mulholland et al., 2016). In particular, the occurrence of large flushing dust storms appears to correlate well with the power of zonal wave number m = 3 eastward traveling waves in the near-surface temperature field (Hinson & Wang, 2009; Wang et al., 2013).

This paper takes a closer look at the dust storm versus transient eddy relationship and finds it important to use different meteorological variables at different heights to understand the variability. This is essentially due to the vertical structures of different wave modes represented in different variables. It further implies that wave mode transition (i.e., the change of the dominant wave mode with time; Collins et al., 1996) is expressed differently at different heights in different variables. This paper also examines the relationship between the southward advancement of cross-equatorial dust storms and the meridional winds in the lower atmosphere. The proper combination of synoptic eddy and zonal mean meridional wind is found to be important for the seasonality of cross-equatorial flushing dust storms.

2. Data

The Version 1.0 Mars Analysis Correction Data Assimilation (MACDA) data set (Montabone et al., 2014) is used to analyze the variability of meteorological variables in this paper. MACDA assimilates Mars Global Surveyor (MGS) Thermal Emission Spectrometer (TES) column dust opacities and nadir temperature profiles into the UK/LMD Mars General Circulation Model (GCM), providing a more faithful representation of atmospheric state than a free running model (Montabone et al., 2006). The data set contains globally gridded meteorological fields at 5° latitude × 5° longitude × 25 level resolution every 2 Mars hours from MY 24 Ls = 141° to MY 27 Ls = 86°. It has been used to study the seasonality, structure, and energetics of transient eddies in the Martian atmosphere (e.g., Battalio et al., 2018; Lewis et al., 2016; Mooring & Wilson, 2015). This paper uses MACDA to analyze northern hemisphere transient eddies from a different perspective. Specifically, it examines how the variabilities of various meteorological variables correlate with the dust storms concurrently observed in MGS Mars Orbiter Camera (MOC) Mars Daily Global Maps (MDGMs) (Wang, 2016; Wang & Ingersoll, 2002).

Each MDGM is made from13 consecutive sets of MGS MOC global map swaths taken during approximately a day. It is composed of a north polar (45°N-90°N, polar stereographic), a south polar (45°S–90°S, polar stereographic), and an equatorial (60°S–60°N, simple cylindrical) map at 0.1° × 0.1° resolution. Adjacent MDGMs have a common image swath. Dust storms with clear boundaries can be identified in MDGMs through visual inspection. The locations and Ls of frontal/flushing dust storms were previously collected using the equatorial MDGMs in Wang (2007). These data are used in this study.

In Wang (2007), frontal dust storms were identified as curvilinear-shaped dust storms analogous to terrestrial baroclinic fronts; flushing dust storms included (a) frontal dust storms that exhibited significant southward motion, (b) frontal dust storms that extended into the topographic corridors through Acidalia, Arcadia, or Utopia, (c) irregularly shaped dust storms probably evolved from or triggered by frontal dust storms, and (d) other dust storms along the trajectories of dust storm sequences that involve flushing dust storms in their histories. Among the storms in (a)–(d), some extend from middle/high latitudes to low latitudes, others remain at low latitudes. Since most area in a MDGM is imaged only once per day (except for the polar region), it is difficult to tell for sure whether/how a storm in (c) or (d) is related to other dust storms. Dust storms in (c) and (d) are included because they are involved in dust storm sequences. Ambiguous cases are included for completeness, but they represent a small fraction. A local or regional dust storm that is isolated in space and time from other dust storms does not lead to a major dust storm (Cantor, 2007; Toigo et al., 2018); however, when local and regional dust storms become members of a dust storm sequence that keeps a large area dusty over an extended period of time, the dust storm sequence has the potential to grow into a major dust storm (Wang & Richardson, 2015).

To better account for high-latitude dust storms (some of which are not covered by equatorial MDGMs), this study counts the total number of frontal dust storms in each north polar MDGM (supporting information Table S1). These dust storms comprise both flushing and nonflushing dust storms. They include (1) frontal storms that are entirely north of 60°N and (2) frontal storms that are totally or partially between 45°N and 60°N.The former are observable only in north polar MDGMs; the latter are observable in both north polar and equatorial MDGMs. Although circumpolar streaks are common from midfall to midwinter (Wang & Ingersoll, 2002), they are excluded in this study as they are confined within the polar hood and do not reach lower latitudes.

The dust within frontal/flushing dust storms is either lifted locally or transported from elsewhere. This paper examines the spatial and temporal distributions of the dust storms regardless of where/how dust is lifted. Among the years covered by MACDA, MY 24 provides the best example for this study. Moreover, the error in TES spectra grew with time (Pankine, 2016). Therefore, this paper primarily focuses on MY 24.

3. Results

3.1. Correlations

The solsticial pause of near-surface transient eddies in MACDA was well documented by Lewis et al. (2016). This behavior is reproduced in Figure 1 for the 1.2 < P ≤ 20 sols synoptic period transient eddies in the temperature (T) and meridional wind (V) field at σ= 0.90 (z ~ 1 km) for MY 24. The color-filled contours show the zonal means of the standard deviations as a function of Ls and latitude. The standard deviation at each grid point is calculated every 1° of Ls using linearly detrended and band-pass filtered (1.2 < P≤ 20 sols) time series for each 20-sol time window.

Figure 1.

Figure 1.

Open in a new tab

Standard deviations of 1.2 < P ≤ 20.0 sols transient eddies derived from (top) temperature T in K and (bottom) meridional wind V in m/s at σ = 0.90 as a function of Ls and latitude for MY 24. Symbols indicate the southern tips of frontal/flushing dust storms observed in equatorial MDGMs (60°S–60°N). Multiday events are connected with lines. Short vertical lines denote the Ls of frontal dust storms observed in north polar MDGMs (45°N–90°N). The lengths of the lines are proportional to the number of events. The shortest line corresponds to 1. MY = Mars year; MDGMs = Mars Daily Global Maps.

To examine the correlations of eddy fields with dust storms, Figure 1 indicates the Ls and latitudes of the southern tips of the frontal/flushing dust storms observed in equatorial MDGMs (60°S–60°N) using symbols. Potential multiday dust storms are connected with lines (Wang, 2007). The timing of the frontal dust storms observed in north polar MDGMs (45°N–90°N) is indicated by short vertical lines in each panel. In this paper, no effort is made to match the storms in north polar maps with those in equatorial maps. As a consequence, even if a line and a symbol in Figure 1 are at the same Ls, they may or may not refer to the same dust storm.

Figure 1 shows that both the eddy T and the eddy V have their strongest variation in northern midlatitude and high latitude before and after the northern winter solstice. The solsticial pause is also exhibited by the seasonal distribution of frontal/flushing dust storms. Therefore, there is a general positive correlation between these dust storms and synoptic eddies.

However, a closer examination of Figure 1 reveals that the cross-equatorial flushing dust storms correlate better with the eddy T than with the eddy V in the lower atmosphere. Although frontal dust storms are observed more or less continuously during Ls = 180°–230° and Ls = 310°–360°, the largest flushing dust storm sequences are observed during shorter seasonal windows (Ls = 210°–230° and Ls = 320°–335°), right next to the solsticial pause. While eddy T is roughly stronger during these flushing dust storm periods than during the rest of the presolstice and postsolstice periods (except for Ls = 350°−360°), eddy V tends to maximize earlier in the fall and later in the winter. Consequently, despite southward dust transport being more directly related to V than T, the standard deviation of eddy T near the surface seems to be a better indicator of the timing of cross-equatorial flushing dust storms.

Figure 2 is the same as Figure 1 but for σ = 0.072—a higher vertical level (z ~ 28 km). The situation is reversed. There is a strong variation in eddy T during the winter solstice period when no frontal/flushing dust storms are observed. In comparison, the standard deviation of eddy V maximizes during the presolstice and postsolstice periods when cross-equatorial flushing dust storms are observed. Thus, eddy V in the middle atmosphere (instead of eddy V near the surface) appears to correlate better with cross-equatorial flushing dust storms.

Figure 2.

Figure 2.

Open in a new tab

Same as Figure 1 but for σ = 0.072.

3.2. Wave Modes

To understand the results in section 3.1, this study examines the dominant wave mode of 1.2 < P ≤ 20 sols eastward traveling waves on both σ levels. Although westward traveling waves are present during major dust storms, they are mostly filtered out in this study due to their long wave periods (Banfield et al., 2004; Wang, 2017). Figures 3 and 4 show the results derived from the T and V fields, respectively. The amplitude, wave frequency, and zonal wave number of the dominant wave mode are shown as a function of Ls and latitude in the top, middle, and bottom rows of each figure. The dominant wave mode is the one with the largest amplitude among all the 1.2 < P ≤ 20 sols eastward traveling waves. These waves are derived from the longitude versus time array of the corresponding variable at each latitude and σ level using fast Fourier transform. The calculation is performed every 1° of Ls for each 20-sol time window. The window length is based on the wave periods and coherence time of traveling waves (Banfield et al., 2004; Barnes, 1980; Hinson, 2006). The mean and trend of the 20-sol time series at each grid point are removed before the space-time spectral analysis.

Figure 3.

Figure 3.

Open in a new tab

Ls versus latitude distribution of the amplitude (top row, K), frequency (middle row, cycle/sol), and zonal wave number (bottom row) of the dominant wave mode of the P > 1.2 sol eastward traveling waves derived from the temperature field at (left column) σ = 0.90 and (right column) σ = 0.072 for Mars year 24. For simplicity, only the pixels corresponding to wave amplitudes >3.0 K in the top row are plotted in the bottom two rows. The same symbols and lines as those in Figure 1 are superimposed in the bottom row.

Figure 4.

Figure 4.

Open in a new tab

Same as Figure 3 but for meridional wind (m/s). For simplicity, only the pixels corresponding to amplitudes greater than 40% of the maxima represented by the color bars of the top row are plotted in the bottom two rows.

Figure 3 shows that the dominant wave mode in T differs between the lower and middle atmosphere. At σ = 0.90, the eddy amplitudes show a solsticial pause; the wave mode transitions progress through zonal wave numbers m = 2,1, 3, 2 during the presolstice period and m = 3, 2,1, 2 during the post-solstice period. In comparison, at σ = 0.072, the eddy amplitudes are large during Ls = 210°–340°; the wave mode transitions involve only m = 1 of different wave periods (Wilson et al., 2002) and m = 2. These results agree with those in Lewis et al. (2016) and are consistent with the vertically confined character of m = 3 traveling waves derived from temperature data (Banfield et al., 2004).

Although frontal dust storms show no apparent preference for a particular dominant zonal wave number, cross-equatorial flushing dust storms are preferably observed during the time periods when m = 3 eastward traveling waves are strong in the near-surface temperature field (Figure 3, bottom left panel). During Ls = 355°–360°, although the near-surface temperature variation is strong, the variation is dominated by m = 2 and there are no cross-equatorial flushing dust storms. The flushing dust storm sequence during Ls = 210°–214° in MY 24 coincides with a dominant m = 1 wave in T at σ = 0.90; however, an m = 3 wave is still significant at the time and is only slightly weaker than the m = 1 wave. This suggests that large flushing dust storms correlate positively with the amplitudes of m = 3 waves in eddy T in the lower atmosphere. The eddy T in the middle atmosphere is dominated by m = 1 or m = 2 waves and shows no apparent resemblance to the seasonal distribution of frontal/flushing events.

A comparison between Figures 3 and 4 shows that the wave modes and their transitions in V are completely different than those in T at σ = 0.072. Strong variations in eddy V in the middle atmosphere are attributed to m = 3 eastward traveling waves whose timing corresponds nicely to that of cross-equatorial flushing dust storms (Figure 4, right column). At σ = 0.90, the m = 3 waves are also dominant during Ls = 210°–230° and Ls = 320°–335°, but the amplitudes of these waves are smaller than the amplitudes of the m = 2 waves earlier in the fall, later in the winter, or at higher latitudes (Figure 4, left column).

The results above indicate that the cross-equatorial flushing dust storms in MY 24 are closely related to the m = 3 traveling waves. These waves are manifest in T in the lower atmosphere and in V in the middle atmosphere. Figure 5 shows the dominant eastward traveling wave in T and V as a function of latitude and height at Ls = 219.3° in MY 24. In the T field, the largest wave amplitude is at midlatitudes below z ~ 10 km, and is attributed to the P = 2.2 sols m = 3 wave; the second largest amplitude is at higher latitudes in the middle atmosphere and is attributed to the P = 20 sols m = 1 wave. In the V field, the P = 2.2 sols m = 3 wave dominates the midlatitude and high latitude throughout much of the lower and middle atmosphere and attains its maximum at high latitudes in the middle atmosphere. Wang et al. (2013) pointed out that m = 3 traveling waves extend higher from the surface in V than in T, which is consistent with the expectation from thermal wind balance. The additional information in Figure 5 is that the rank of this wave mode with respect to others also follows the same pattern, that is, the dominant status of the m = 3 wave mode extends higher from the surface in V than in T. Thus, the results in section 3.1 can be understood in terms of the differences in the expressions of various wave modes in different meteorological variables.

Figure 5.

Figure 5.

Open in a new tab

Latitude versus height distribution of the amplitude (top row), frequency (middle row, cycle/sol), and zonal wave number (bottom row) of the dominant wave mode of the P > 1.2 sol eastward traveling waves derived from the (left column) temperature (K) and (right column) meridional wind (m/s) field at Ls = 219.3° of Mars year 24. For simplicity, only the pixels with amplitudes greater than 30% of the maxima represented by the color bars of the top row are plotted in each panel.

To further examine the relationship between dust storms and eastward traveling waves of different zonal wave numbers, Figure 6 shows the square root of the total power of 1.2 < P ≤ 20 sols m = 1,2, and 3 eastward traveling waves derived from T at σ = 0.90 (left) and V at σ = 0.072 (right). All the panels show a solsticial pause and a degree of positive correlation with the frontal/flushing dust storm distribution. However, cross-equatorial flushing dust storms appear to correlate better with m = 3 waves in both the T and V fields. Although there are flushing dust storms when m = 2 waves dominate (Ls = 180°–210° and Ls = 340°–360°), those storms do not reach as far south; therefore, they are not cross-equatorial flushing dust storms. On a side note, although m = 3 waves are dominant when cross-equatorial flushing storms occur, m = 2 and/or m = 1 waves are usually present during the same time periods. The combination of these waves can lead to sharpened gradients across fronts and/or enhanced variability in storm zones (Banfield et al., 2004).

Figure 6.

Figure 6.

Open in a new tab

Square root of the total power of 1.2 < P ≤ 20.0 sols (top row) m = 1, (middle row) m = 2 and (bottom row) m = 3 eastward traveling waves as a function of Ls and latitude for Mars year 24. The left column is derived from the temperature (K) at σ = 0.90; the right column is derived from the meridional wind (m/s) at σ = 0.072. The same symbols and lines as those in Figure 1 are superimposed in each panel.

To get an idea of the overall influence of various zonal wave numbers throughout the vertical column, this study calculates the mass weighted average of the total power of 1.2 < P ≤ 20 sols m = 1,2. and 3 eastward traveling waves. The square roots of the results derived from T and V fields are plotted in Figure 7. In the vertically averaged sense, the m = 1 waves in T do not exhibit a solsticial pause due to the contribution of the middle atmosphere, though the m = 1 waves in V still do. For the other two zonal wave numbers, results for both T and V lead to the same conclusion—cross-equatorial flushing dust storms are observed when m = 3 eastward traveling waves are strong.

Figure 7.

Figure 7.

Open in a new tab

Same as Figure 6, but the total power is averaged vertically according to mass before the square root is taken.

Surface pressure is another variable representing the behavior of the vertical column (of mass). Figure 8 is the same as Figure 6 but for surface pressure. It shows that all three zonal wave numbers exhibit a solsticial pause and some degree of positive correlation with frontal/flushing dust storms. The amplitudes of m = 1 and m = 2 waves are much larger than those of m = 3 waves in the surface pressure field, which is in contrast to the results for T and V. In other words, even though m = 3 waves are weaker than other zonal wave numbers in terms of surface pressure, they are sometimes stronger in terms of temperature and meridional wind. Therefore, even a weak signal of m = 3 in surface pressure can be significant for studies related to frontal/flushing dust storms (Haberle et al., 2018). Since a wave mode may dominate in some fields but not others, wave mode transition depends on meteorological variable. Collins et al. (1996) first discussed wave mode transition between m = 1 and m = 2. Hinson et al. (2012) discussed wave mode transitions among m = 1,2, and 3. It is worth noting that Collins et al. (1996) analyzed surface pressure, while Hinson et al. (2012) mainly analyzed near-surface temperature.

Figure 8.

Figure 8.

Open in a new tab

Same as Figure 6 but for surface pressure (Pa) in Mars year 24.

3.3. Meridional Winds

The previous section shows that m = 3 eastward traveling waves are good indicators of cross-equatorial flushing dust storms; however, it is puzzling why the strongest synoptic variability in the near surface V field (Figure 1, bottom panel) did not flush any dust storms across the equator in MY 24.

Figure 9 shows the zonal means of the 20-sol means of V at σ = 0.90 as a function of Ls and latitude for MY 24. There are weak northward winds (0 ≤ V ≤ 2 m/s) north of ~40°N and strong southward winds (—12 < V< —4 m/s) in the low latitudes (30°S–30°N). During the period around the northern winter solstice, the mean V generally shifts toward the negative direction (i.e., northward V becomes weaker and southward becomes V stronger), making it more favorable for southward tracer transport in the low latitudes. However,as shown by the standard deviations of 1.2 < P≤ 20 sols eddy Vat σ = 0.90 (Figure 9, contours), the transient eddies in the midlatitudes are the weakest during the winter solstice period, making it less favorable for southward dust transport from the high latitudes.

Figure 9.

Figure 9.

Open in a new tab

Ls versus latitude distributions of (color) the zonal mean of the 20-sol mean V (m/s) at σ = 0.90 and (contour) the zonal mean of the standard deviation of the 1.2 < P ≤ 20.0 sols eddy V (m/s) at σ = 0.90 for (top row) Mars year (MY) 24, (middle row) MY 25, and (bottom row) MY 26. The southern tips of frontal/flushing dust storms observed in the equatorial MDGMs are superimposed in each panel. Major data gaps in MGS TES (\) and MOC (/) data are indicated with gray vertical bars. MDGMs = Mars Daily Global Maps; MGS = Mars Global Surveyor; TES = Thermal Emission Spectrometer; MOC = Mars Orbiter Camera.

During the transitional time periods of midfall (Ls = 210°–240°) and midwinter (Ls = 300°–330°), although neither the transient eddy V nor the mean V is at its peak (across Ls = 180°–360°), the configuration is probably more favorable for tracer transport from high to low latitudes and across the equator. During Ls = 210°– 220° and Ls = 320°–330°, the latitudes of synoptic eddies are at their southernmost positions. For example, the contours for standard deviation = 3.0 and 6.0 m/s descend to ~10°N and 30°N which are well within the domain controlled by strong zonal mean V (Figure 9 top panel). This probably makes it easier for dust in northern frontal systems to be transported continuously southward once it is entrained into the low-latitude circulation. It is consistent with the occurrence of the cross-equatorial dust storms at Ls ~ 210° and Ls ~ 320° in MY 24. At Ls ~ 220°, synoptic variability begins to weaken, but southward zonal mean V continues to strengthen, resulting in another opportunity for cross-equatorial flushing dust storms. After Ls ~ 230°, synoptic eddies enter the solsticial pause period when they are probably too weak for frontal dust storms to develop. This situation lasts until Ls ~ 320° when the postsolstice flushing dust storm season of MY 24 begins.

Similarly, in MY 26 (bottom panel of Figure 9), the cross-equatorial dust storms around Ls ~ 210° and Ls ~ 310° occur roughly when northern hemisphere synoptic eddies descend to low latitudes. At the time, both the synoptic eddy V and the mean V are moderately strong, but neither is as strong as their maximum values across the Ls = 180°–360° period. The smaller cross-equatorial dust storm at Ls ~ 230° is under weaker eddy V and stronger mean V condition. The Ls ~ 340° flushing dust storm sequence (which comes close to the equator) is under stronger eddy V and weaker mean V condition. In the fall of MY 25 (Figure 9, middle panel), synoptic eddies are suppressed during a global dust storm (Cantor, 2007; Smith et al., 2002); in the meantime, the latitudes of synoptic eddies remain high. Thus, a favorable condition for cross-equatorial flushing dust storm is not attained. In the postsolstice season of MY 25, cross-equatorial flushing dust storms return roughly at a time when moderately strong eddy V descends to low latitudes where moderately strong southward mean V exists.

In Figure 10, the 1.2 < P ≤ 20 sols eddy Vat σ = 0.90 in MY 24 is decomposed into different zonal wave numbers, and the square root of the wave power is plotted in color. In general, m = 3 waves are located at lower latitudes than m = 2 waves, which are in turn located at lower latitudes than m = 1 waves (Lee et al., 2018; Lewis et al., 2016). The synoptic variability during the transitional time periods when cross-equatorial dust storms are active is mostly provided by m = 3 traveling waves. Although the maximum amplitudes of the m = 1 and m = 2 waves during Ls = 180°–360° are larger than the maximum amplitude of the m = 3 waves during the same time period, the m = 3 waves are closer in distance to the low latitudes where relatively strong southward mean V exists. In addition, the m = 3 waves are closer in time to the peak of the low-latitude southward mean V. As a result, the latitudinal and seasonal distributions of the m = 3 eastward traveling waves are suitable for cross-equatorial flushing dust storms. In fact, m = 3 are present during all cross- equatorial flushing dust storm periods of MY 24–26, though they are occasionally not dominant. For instance, in the fall of MY 26, m = 2 traveling waves are especially strong and clearly outweigh other wave modes. The Ls ~ 210° cross-equatorial dust storm sequence in MY 26 actually coincides with a dominant m = 2 wave (Wang, 2007), though the sequence occurs after the peak of the m = 2 waves in that fall. It should be noted that there is a nearby TES data gap (indicated in Figure 9) which probably has some influence on MACDA (Rogberg et al., 2010).

Figure 10.

Figure 10.

Open in a new tab

Colors show the Ls versus latitude distribution of the square root of the total power of the 1.2 < P ≤ 20.0 sols eddy V (m/s) at σ = 0.90 for (top row) m = 1, (middle row) m = 2, and (bottom row) m = 3 eastward traveling waves in Mars year 24. Contours show the zonal mean of the 20-sol mean V (m/s). Symbols indicate the southern tips of frontal/flushing dust storms observed in equatorial Mars Daily Global Maps.

4. Summary and Discussion

The relationship between frontal/flushing dust storms and northern hemisphere synoptic transient eddies in MY 24 is re-examined using MDGM and MACDA. Although frontal dust storms are observed more or less continuously during the presolstice and postsolstice time periods, cross-equatorial flushing dust storms are confined to shorter seasonal windows on both sides of the solsticial pause.

In the temperature field near the surface, there is strong synoptic variability during the time periods of cross-equatorial flushing dust storms; however, in the meridional wind field near the surface, the variability peaks earlier in the fall and later in the winter. Thus, large flushing dust storms appear to be better correlated with near-surface eddy T than with near-surface eddy V, despite V being more directly related to southward tracer transport. In the temperature field in the middle atmosphere, there is strong synoptic variability during the winter solstice period when essentially no frontal/flushing dust storms are observed; however, in the meridional wind field in the middle atmosphere, the time periods of strong synoptic variability coincide with the time periods of large flushing dust storms. Thus, large flushing dust storms appear to be better correlated with the eddy V in the middle atmosphere than with the eddy V near the surface.

The dominant wave modes in different variables are examined to understand the correlations. The strong eddy Tin the middle atmosphere is dominated by m = 1 and m = 2 traveling waves, and exhibits no solsticial pause. During the time periods of large flushing dust storms (in midfall and midwinter), both the lower atmosphere eddy T and the middle atmosphere eddy V are prominent and are dominated by m = 3 eastward traveling waves. In the eddy V field in the lower atmosphere, the m = 3 waves are also dominant during large flushing dust storm periods, but they are weaker than the m = 2 waves earlier in the fall and later in the winter. Toward the end of northern winter, there are no cross-equatorial dust storms, the strong near-surface eddy T is dominated by m = 2 traveling waves whose eddy V in the middle atmosphere is not particularly strong. Thus, frontal dust storms do not seem to be limited to any particular wave mode, but large flushing dust storms appear to be closely related to m = 3 traveling waves, at least in MY 24. Wave mode transition to/from m = 3 is a useful indicator for potential large flushing dust storms.

The relative importance of waves in different meteorological variables is useful fora comprehensive understanding of the variability of the Martian atmosphere. This paper primarily analyzes eddy T and eddy V. Selected results for eddy U (similar to Figures 4 and 7) can be found in the supporting information (Figures S1 and S2). Eddy U also suggests that cross-equatorial flushing dust storms are best correlated with m = 3 traveling waves in MY 24, but the details of its spatial and temporal distributions are different from those of eddy V. In particular, eddy U in the middle atmosphere is more susceptible to m = 1 than eddy V. Since a wave mode is expressed differently in different meteorological fields due to its horizontal and vertical structures, the rank of a wave mode can be different for different variables. This implies that wave mode transition depends on meteorological variable. It is important to choose the appropriate variables for different applications.

The m = 3 traveling waves are well correlated with the large flushing dust storms in MY 24. This is in part due to their wave amplitudes, but more importantly due to their seasonal distributions and latitudinal positions. During the time periods when m = 3 waves are strong, the m = 3 waves are also located at lower latitudes and are connected with strong southward mean V in the low latitudes. Dust in frontal storms at high latitudes can be easily entrained into the low-latitude circulation and be efficiently transported southward. Other zonal wave numbers (m = 1 and 2) tend to peak under less favorable conditions for cross-equatorial flushing dust storms—In late fall and early winter, transient eddy winds are probably too weak for frontal dust storms to develop at high latitudes; in early fall and late winter, the latitudinal separation between the strongest synoptic eddy V and the strongest southward mean V is large, making it difficult for high-latitude dust to be transported to low latitudes. In order for m = 1 and m = 2 waves to initiate cross-equatorial flushing dust storms, they probably need to be extremely strong and/or shifted to lower latitudes (e.g., fall of MY 26). In short, the seasonal and latitudinal distributions of traveling waves are important for the development of cross-equatorial flushing dust storms.

In addition to synoptic eddies and mean winds, other circulation components also play important roles in the development of flushing dust storms. For instance, stationary waves can facilitate flushing in Acidalia and oppose flushing in Arcadia (Hinson & Wang, 2009); thermal tides can facilitate flushing when synchronized with traveling waves and oppose flushing when not (Wang et al., 2003). It is possible that the seasonal distribution of the flushing dust storms in a particular geographic location or within a certain stretch of time can be influenced by stationary waves and thermal tides, but it represents the next level of complexity. This paper focuses on the general seasonality of frontal/flushing dust storms and does not discuss other circulation components further. However, to fully understand the Martian dust cycle, it is important to consider the interactions of synoptic eddies with other circulation components.

Supplementary Material

Supp
TableS1

Key Points:

  • Cross-equatorial flushing storms are in shorter seasonal windows than frontal storms

  • Cross-equatorial flushing storms are closely related to m = 3 traveling waves in MY24

  • Seasonality and latitudes of m = 3 waves favor cross-equatorial flushing storms

Acknowledgments

This paper is partially supported by NASA MDAP grant 80NSSC17K0475 and partially supported by the Smithsonian Institution. The MACDA data set is available from the Centre for Environmental Data Analysis (http://catalogue.ceda.ac.uk/uuid/c69013e492b4412380630ed77bee9543). The MGS MOC MDGM is available at the Harvard Dataverse (https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/WWRT1V). H. Wang would like to thank Don Banfield, Claire Newman, and an anonymous reviewer for their review

References

  1. Banfield D, Conrath BJ, Gierasch PJ, Wilson RJ, & Smith MD (2004). Traveling waves in the Martian atmosphere from MGS TES nadir data. Icarus, 170(2), 365–403. 10.1016/j.icarus.2004.03.015 [DOI] [Google Scholar]
  2. Barnes JR (1980). Time spectral analysis of mid-latitude disturbances in the Martian atmosphere. Journal of Geophysical Research, 37(9), 2002–2015. [Google Scholar]
  3. Battalio M, Szunyogh I, & Lemmon M (2018). Wave energetics of the southern hemisphere of Mars. Icarus, 309, 220–240. 10.1016/j.icarus.2018.03.015 [DOI] [Google Scholar]
  4. Cantor BA (2007). MOC observations of the 2001 Mars planet-encircling dust storm. Icarus, 186(1), 60–96. 10.1016/j.icarus.2006.08.019 [DOI] [Google Scholar]
  5. Collins M, Lewis SR, Read PL, & Hourdin F (1996). Baroclinic wave transitions in the Martian atmosphere. Icarus, 120(2), 344–357. 10.1006/icar.1996.0055 [DOI] [Google Scholar]
  6. Haberle RM, de la TooreJuarez M, Kahre MA, Kass DM, Barnes JR, Hollingsworth JL, et al. (2018). Detection ofnorthern hemisphere transient eddies at Gale crater Mars. Icarus, 307, 150–160. 10.1016/j.icarus.2018.02.013 [DOI] [Google Scholar]
  7. Hinson DP (2006). Radio occultation measurements of transient eddies in the northern hemisphere of Mars. Journal of Geophysical Research, 111, E05002. 10.1029/2005JE002612 [DOI] [Google Scholar]
  8. Hinson DP, & Wang H (2009). Further observations of regional dust storms and baroclinic eddies in the northern hemisphere of Mars. Icarus, 206, 290–305. 10.1016/j.icarus.2009.08.019 [DOI] [Google Scholar]
  9. Hinson DP, Wang H, & Smith MD (2012). A multi-year survey of dynamics near the surface in the northern hemisphere of Mars: Short-period baroclinic waves and dust storms. Icarus, 219(1), 307–320. 10.1016/j.icarus.2012.03.001 [DOI] [Google Scholar]
  10. Hollingsworth JL, Haberle RM, Barnes JR, Bridger AFC, Pollack JB, Lee H,& Schaeffer J (1996). Orographic control ofstorm zones on Mars. Nature, 380(6573), 413–416. 10.1038/380413a0 [DOI] [Google Scholar]
  11. Kass DM, Kleinbohl A, McCleese DJ, Schofield JT, & Smith MD (2016). Interannual similarity in the Martian atmosphere during the dust storm season. Geophys/ca/ Research Letters, 43, 6111–6118. 10.1002/2016GL068978 [DOI] [Google Scholar]
  12. Lee C, Richardson MI, Newman CE, & Mischna MA (2018). The sensitivity of solsticial pauses to atmospheric ice and dust in the MarsWRF general circulation model. Icarus, 311, 23–34. 10.1016/2018.03.019 [DOI] [Google Scholar]
  13. Lewis SR, Mulholland DP, Read PL, Montabone L, Wilson RJ, & Smith MD (2016). The solsticial pause on Mars: 1. A planetary wave reanalysis. Icarus, 264, 456–464. 10.1016/j.icarus.2015.08.039 [DOI] [Google Scholar]
  14. Liu JJ, Richardson MI, & Wilson RJ (2003). An assessment of the global, seasonal, and interannual spacecraft record of Martian climate in the thermal infrared. Journal of Geophysical Research, 108(E8), 5089 10.1029/2002JE001921 [DOI] [Google Scholar]
  15. Martin LJ, & Zurek RW (1993). An analysis of the history of dust activity on Mars. Journal of Geophysical Research, 98(E2), 3221–3246. 10.1029/92JE02937 [DOI] [Google Scholar]
  16. Montabone L, Forget F, Millour E, Wilson RJ, Lewis SR, Cantor B, et al. (2015). Eight-year climatology of dust optical depth on Mars. Icarus, 251, 65–95. 10.1016/j.icarus.2014.12.034 [DOI] [Google Scholar]
  17. Montabone L, Lewis SR, Read PL, & Hinson DP (2006).Validation of Martian meteorological data assimilation for MGS/TES using radio occultation measurements. Icarus, 185(1), 113–132. 10.1016/j.icarus.2006.07.012 [DOI] [Google Scholar]
  18. Montabone L, Marsh K, Lewis SR, Read PL, Smith MD, Holmes J, et al. (2014).The Mars analysis correction data assimilation (MACDA) dataset V1.0. Geoscience Data Journal, 1(2), 129–139. 10.1002/gdj3.13 [DOI] [Google Scholar]
  19. Mooring TA, & Wilson RJ (2015). Transient eddies in the MACDA Mars reanalysis. Journal of Geophysical Research: Planets, 120,1671–1696. 10.1002/2015JE004824 [DOI] [Google Scholar]
  20. Mulholland DP, Lewis SR, Read PL, Madeleine J, & Forget F (2016). The solsticial pause on Mars: 2 Modeling and investigation of causes. Icarus, 264, 465–477. 10.1016/jjcarus.2015.08.038 [DOI] [Google Scholar]
  21. Pankine AA (2016). Radiometric error and re-calibration ofthe MGSTES spectra. Planetary and Space Science, 134,112–121. 10.1016/j.pss.2016.10.015 [DOI] [Google Scholar]
  22. Rogberg P, Read PL, Lewis SR,& Montabone L. (2010). Assessing atmospheric predictabilityon Mars using numerical weather prediction and data assimilation. Quarterly Journal of the Royal Meteorological Society, 136(651), 1614–1635. 10.1002/qj.677 [DOI] [Google Scholar]
  23. Smith MD, Conrath RJ, Pearl JC, & Christensen PR (2002). Thermal emission spectrometer observations of Martian planet-encircling dust storm 2001A. Icarus, 157(1), 259–263. 10.1006/icar.2001.6797 [DOI] [Google Scholar]
  24. Toigo AD, Richardson MI, Wang H, Guzewich SD, & Newman CE (2018). The cascade from local to global dust storms on Mars: Temporal and spatial thresholds on thermal and dynamical feedback. Icarus, 302, 514–536. 10.1016/j.icarus.2017.11.032 [DOI] [Google Scholar]
  25. Wang H (2007). Dust storms originating in the northern hemisphere during the third mapping year ofMars Global Surveyor. Icarus, 189(2), 325–343. 10.1016/j.icarus.2007.01.014 [DOI] [Google Scholar]
  26. Wang, H. (2016). “MGS MOC Mars daily global maps”, doi: 10.7910/DVN/WWRT1V, Harvard Dataverse, V1.
  27. Wang H (2017). Major dust storms and westward traveling waves on Mars. Geophysical Research Letters, 44, 3493–3501. 10.1002/2017GL072894 [DOI] [Google Scholar]
  28. Wang H, & Ingersoll AP (2002). Martian clouds observed by Mars Global Surveyor Mars orbiter camera. Journal of Geophysical Research, 107(E10), 5078 10.1029/2001JE001815 [DOI] [Google Scholar]
  29. Wang H, & Richardson MI (2015). The origin, evolution, and trajectory of large dust storms on Mars during Mars years 24–30 (1999–2011). Icarus, 251, 112–127. https://doi.org/10.1016Zj.icarus.2013.10.033 [Google Scholar]
  30. Wang H, Richardson MI, Toigo AD, & Newman CE (2013). Zonal wavenumber three traveling waves in the northern hemisphere of Mars simulated with a general circulation model. Icarus, 223(2), 654–676. 10.1016/j.icarus.2013.01.004 [DOI] [Google Scholar]
  31. Wang H, Richardson MI,Wilson RJ, Ingersoll AP,& Zurek RW. (2003).Cyclones,tides,and theorigin ofmajorduststorms on Mars. GRL, 30(9), 1488 10.1029/2002GL016828 [DOI] [Google Scholar]
  32. Wilson RJ, Banfield D, Conrath BJ, & Smith MD (2002). Traveling waves in the northern hemisphere of Mars. Geophysical Research Letters, 29(14), 1684 10.1029/2002GL014866 [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supp
NIHMS1029892-supplement-Supp.pdf (535.3KB, pdf)
TableS1
NIHMS1029892-supplement-TableS1.xlsx (51.7KB, xlsx)

RESOURCES