- Email: [email protected]
Dust activity over the Hellas basin of Mars during the period of southern spring equinox
Icarus 311 (2018) 306–316
Contents lists available at ScienceDirect
Icarus journal homepage: www.elsevier.com/locate/icarus
Dust activity over the Hellas basin of Mars during the period of southern spring equinox Kim-Chiu Chow∗, Kwing-lam Chan, Jing Xiao Space Science Institute/ Lunar and Planetary Science Laboratory, Macau University of Science and Technology, Macau
a r t i c l e
i n f o
Article history: Received 1 September 2017 Revised 9 April 2018 Accepted 13 April 2018 Available online 22 April 2018 Keywords: Mars Dust storms Hellas basin Cap-edge flow Sublimation flow Slope wind
a b s t r a c t Dust storms that last for more than one Martian day frequently occur in Hellas basin of Mars around southern spring equinox. The dynamics behind the formation of these “Hellas storms”, however, is still unclear. In this study, a ten-year climatology of Martian atmospheric dust is simulated with the MarsWRF global climate model. Results suggest that occurrence of Hellas storms during southern spring equinox is related to the abrupt increase in the surface temperature contrast between the southern edge of the Hellas basin and the carbon dioxide ice covered south polar cap. Significant dust lifting over the southern edge of the Hellas basin occurs mainly during night time, due to the particular strength of the prevailing downslope flows at the site. © 2018 Elsevier Inc. All rights reserved.
1. Introduction It has been reported (e.g., Martin and Zurek, 1993; Wang et al., 2015) that Hellas basin is a region on Mars where dust storms can be commonly observed. The basin is a huge crater region about 30 0 0 km in diameter and 8 km in depth. The steep topography of the basin boundary likely generates a slope-wind circulation associated with the diurnal solar cycle. Observational studies (e.g., Strausberg et al., 20 05; Cantor, 20 07) have shown that the wellknown planet-encircling dust event that occurred in 2001 was initiated in Hellas Basin around the time of the southern spring equinox (LS = 180°). In particular, Strausberg et al. (2005) pointed out that in the initiation phase of this severe dust event, a sequence of dust storm events was observed between the southern edge of Hellas basin and the south polar ice cap edge. The dust was transported out of the basin and propagated to the east in later stage, and facilitated the development of a second dust lifting center over the Syria/Solis/Daedalia region. The occurrence of dust storms over Hellas Basin around the southern spring equinox (LS = 180°) period has been noted in some general circulation model simulations (e.g., Newman et al., 2002; Basu et al., 2006; Newman and Richardson, 2015). In those studies, the small-scale to regional-scale dust storms (“Hellas storms”) typically lasted for a few sols (a Martian solar day) and usually initiated at the southern or southwestern edge of the basin. Results ∗
Corresponding author. E-mail address: [email protected] (K.-C. Chow).
https://doi.org/10.1016/j.icarus.2018.04.011 0019-1035/© 2018 Elsevier Inc. All rights reserved.
of a high resolution mesoscale model by Toigo et al. (2002) were shown in Strausberg et al. (2005); in which a high-stress region in the southern edge of Hellas Basin was displayed around LS = 180°. Furthermore, the numerical experiments of Ogohara and Satomura (2008) showed that dust injected at the southwestern edge of the Hellas basin during southern spring equinox may form a regional dust storm over the basin, similar to the Hellas storms simulated by Newman et al. (2002), Newman and Richardson (2015), and Basu et al. (2006). In the modeling studies mentioned above, the occurrence of Hellas storms are believed to be due to the temperature difference between the basin and the southern region. However, there has been only limited analysis (e.g., Strausberg et al., 2005) to explain why the Hellas storms are initiated only during this particular short period, and in this particular region. Nevertheless, there were still some modeling and dynamical studies related to this topic. In Siili et al. (1997, 1999), a two-dimensional mesoscale model was used to study the circulation over a flat ice covered cap edge region. From the idealized numerical simulations, they found that if only the thermally driven circulation due to ice-land temperature contrast is present, the resulting surface stress may be insufficient for dust lifting. However, the stress may be sufficient if a sublimation flow due to the ice cap is present as well. They also found that the presence of an ice cover on the top edge of the valley may significantly enhance the downslope wind (referred as ice-edge forcing) and surface stress during the night time. In the day time, the enhanced temperature contrast between the ice-covered valley slope at the top and the ice free ground surface in the bottom may
K.-C. Chow et al. / Icarus 311 (2018) 306–316
significantly weaken the upslope wind. The addition of sublimation flow over the edge of the valley may also substantially increase the downslope wind. In Toigo et al. (2002), a three-dimensional mesoscale model with realistic physics and nested with a Mars general circulation model was used to examine dust lifting over the ice cap edge during southern summer. In particular, the three major forcings related to dust lifting over cap edge: thermal contrast of cap edge, slope flow, and sublimation flow due to CO2 ice were examined. With a more realistic simulation of the ice cap, they found that surface stress for dust lifting over the cap edge region is generally large during local afternoon. They also found that the sublimation flow is not important to dust lifting during the period considered in their study, and slope flow due to topography could be more important to dust lifting compared with the effect of thermal contrast over the cap edge region. In particular, their results presented in Strausberg et al. (2005) suggest that a high surface stress region can be found over the southern edge of the Hellas Basin during the period of the southern spring equinox, and the stress is at maximum during night time. In the afternoon, upslope wind is dominant over the western and northern rims of the basin. These previous studies provide some insight on the dynamics of dust lifting over the cap edge region. However, the effect of the above three forcing over the Hellas basin, particularly during the period of the southern spring equinox was basically not investigated in detail. In the present study, the formation of Hellas storms has been investigated based on the results from a general circulation model for Mars. This study is attempting to address the question why Hellas storms are commonly initiated and active in the basin for a short period around the southern spring equinox, which could be related to the retreat of ice cap as will be discussed in this paper. This issue has not been addressed in the studies mentioned above (Silli et al., 1997, 1999; Toigo et al., 2002; Strausberg et al., 2005). We believe that a better understanding on the initiation of Hellas storms may help us understand more about the regular dust cycle, and hopefully may provide more insight into the initiation mechanism of the 2001 planet-encircling dust event occurred around the time of the southern spring equinox. 2. The numerical model The general circulation model (GCM) MarsWRF (Richardson et al., 2007; Toigo et al., 2012) is used in this study to simulate the climate of Mars. MarsWRF is the Mars version of the PlanetWRF model (Richardson et al., 2007), which was developed by the National Center for Atmospheric Research (NCAR) Weather Research and Forecasting (WRF) model for Earth (Skamarock and Klemp, 2008). It has been illustrated in Richardson et al. (2007) and Toigo et al. (2012) that MarsWRF is capable of realistically reproducing climate features such as the large-scale general circulation and temperature field on Mars. In this study, the global domain of the MarsWRF model has been set to 36 latitude × 72 longitude grid points (horizontal spatial resolution of about 5° or 300 km in the equatorial region). There are 52 vertical levels in terrain-following hydrostatic-pressure vertical coordinate (Skamarock and Klemp, 2008), and the model top is set at 0.0057 Pa (about 80 km in altitude). Half of these 52 levels are located between the model top and the pressure level of 100 Pa. Hydrostatic dynamics is used for the model runs, and traditional Rayleigh damping is applied to the dynamical variables at the top three vertical layers. The radiation scheme for short- and longwave radiation employed in these simulations is the "wide band model" scheme as described in Richardson et al. (2007), Toigo et al. (2012) and Newman and Richardson (2015). This scheme considers the heating/cooling effects of dust and carbon dioxide (CO2 ). The planetary boundary layer scheme and land surface scheme were
307
Fig. 1. Seasonal evolution of annually average zonal-mean column dust optical depth at 9.3 μm (normalized to a 700 Pa surface) from the ten-year average climatology of model simulation results (upper panel) and the re-constructed observational data (Montabone et al., 2015) averaged for the six Martian years of observations that do not contain a global-scale dust storm (lower panel).
largely adapted from existing schemes in WRF forf Earth, as described in Richardson et al. (2007) and Toigo et al. (2012). The model also includes other physical process parameterizations specific to Mars, such as the carbon dioxide cycle (Guo et al., 2009) and the dust cycle (Newman and Richardson, 2015). However, no parameterization of the water cycle is included in these simulations. The parameterization of dust processes in the model includes two interactive dust schemes similar to those used in Newman and Richardson (2015), and dust is assumed to be available everywhere and at all times over the whole planet surface except those surfaces with ice cover. The first scheme is similar to usual dust models on Earth, in which the lifting of dust is proportional to the surface wind stress. Dust lifting occurs over the surface when the local near-surface stress exceeds a particular threshold value (constant value 0.042 N m−2 in this study). The second scheme provides most of the background dust, which is parameterized as dust lifting by dust devils. The amount of dust lifting is dependent on the temperature difference between the surface and the air above, as well as the sensible heat flux. The radiation scheme in the model is interactive with dust so that the suspended dust may change the atmospheric radiation and thus the circulation. The model was run for eleven Martian years, initially starting from the time of northern spring equinox (LS = 0°). The first year is considered as a spin-up period. Therefore, only ten years of simulation results are used in this study and their ensemble mean is defined as the climatology of the model results. Some important model parameters such as the threshold stress of dust lifting and lifting rate of dust devils were tuned to simulate the regular climate of Mars. When comparing the annual variation of the zonal-mean column dust optical depth (CDOD) from the tenyear averaged model climatology with the 6-year averaged observational data (described below), the model simulation is able to reproduce the essential features of the climatological dust distribution of Mars (Fig. 1), particularly the significant increase in global dust opacity after LS = 135°, and the two episodes of increased dust opacity around LS = 240° and LS = 320° (although there is a certain time shift of the episodes). However, the simulation overestimates the dust opacity over the northern winter polar region, and during the observed winter “solsticial pause” period (e.g.,
308
K.-C. Chow et al. / Icarus 311 (2018) 306–316
Fig. 2. Annual variation of zonal-mean carbon dioxide ice (kg m−2 ) of the four-year averaged data (MY24–MY27) from the Mars Analysis Correction Data Assimilation (MACDA v1.0, contours). The corresponding result from the ten-year model climatology is shown in shading.
Lewis et al., 2016; Kass et al., 2016). Although the dust climate simulated in these simulations may not produce all of the details when compared with observations, the results of the model climatology suggest that these simulations are reproducing the basic general circulation of Mars (see figures in Toigo et al., 2012). For the carbon dioxide cycle, the model climatology suggests that the simulation is capable of capturing the evolution of CO2 ice cap (Fig. 2) when compared with the Mars Analysis Correction Data Assimilation (MACDA v1.0), publicly available from the Centre for Environmental Data Analysis (Montabone et al., 2014). Although a small time delay can be observed in the southward retreat of the ice cap edge during the southern spring equinox period, this comparison might be biased by the fact that the CO2 ice cap in the
Fig. 3. The Hellas Basin region in the model with topography shown in contours. Shading is the surface carbon dioxide ice (kg m−2 ) from the ten-year model climatology averaged for ten sols from LS = 196.3° to 202.2°. The rectangles R1, R2, and R3 are some regions defined in this study. R1: Hellas Basin (30° S to 60° S, 45° E to 90° E). R2: southern edge of Hellas Basin (50° S to 60° S, 45° E to 90° E). R3: neighboring region south of the southern edge (60° S to 70° S, 45° E to 90° E).
reanalysis itself has not yet been fully validated by observations. The CO2 ice cap over the region of Hellas basin in the model climatology can also be seen from the distribution of surface CO2 ice averaged for ten sols during the southern spring equinox (Fig. 3). 3. The occurrence of hellas storms To compare the model results with observations, the CDOD data re-constructed by Montabone et al. (2015, M2015 hereafter) has
Fig. 4. Time series of column dust optical depth at 9.3 μm (normalized to the 700 Pa constant pressure level) over the Hellas basin (region R1). Blue crosses are from the re-constructed observational data (Montabone et al., 2015) averaged for six years without global-scale dust event. Black triangles are from the corresponding daily-mean ten-year climatology of model results. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
K.-C. Chow et al. / Icarus 311 (2018) 306–316
309
Fig. 5. Column dust optical depth at 9.3 μm (normalized to surface level 700 Pa) around the southern spring equinox. Contours are topography with intervals 20 0 0 m. (a) Observational data (Montabone et al., 2015) averaged for the six Mars years excluding the 2001 and 2007 global dust storms, and averaged for 6 sols from LS = 168.9° to 171.7° (b) From the ten-year averaged model climatology, further averaged for 10 sols from LS = 196.3° to 202.2°. Vectors are near surface (2 m) wind averaged in this period. .
310
K.-C. Chow et al. / Icarus 311 (2018) 306–316
Fig. 6. Vertical profiles of dust mixing ratio (10−5 kg/ kg) and velocity vectors (m s−1 , vertical component multiplied by 100) over the Hellas basin region from the model results (ten-year climatology ) averaged over 10 sols from LS = 196.3° to 202.2°. (a) Height-latitude profile averaged over the longitudes from 45° E to 90° E. (b) The corresponding height-longitude profile averaged over latitudes from 30° S to 60° S.
K.-C. Chow et al. / Icarus 311 (2018) 306–316
311
Fig. 7. Time series of daily mean results of the ten-year model climatology over the southern edge region of the Hellas basin (region R2). Blue crosses denote surface air density (kg m−3 , right axis), and black curve denotes surface carbon dioxide ice (kg m−2 , left axis). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
been used. The data are based on the Thermal Emission Spectrometer (TES) and the Thermal Emission Imaging System (THEMIS) observations, as well as the estimates of CDOD from the Mars Reconnaissance Orbiter (MRO)/Mars Climate Sounder in MY 28–33. These gridded, daily maps cover Martian years (MY) 24 to 32. As in M2015, we excluded the years of the 2001 and 2007 global dust storms and used the six years without them (MY24, 26, 27, 29, 30, and 31) to obtain the climatology of dust optical depth. 3.1. Climatology of Hellas storms from observational data The region of the Hellas basin focused in this study has the topography shown in Fig. 3. In this study, the analyzes of the results over the region of Hellas Basin are referred to three rectangular regions R1, R2, and R3 (Fig. 3). The Hellas Basin is referred to the region R1, and the southern edge of the Hellas Basin is referred to the region R2, while R3 is a neighboring region south of the southern edge and has the same area as R2. From the time series of the M2015 climatological data averaged over the Hellas basin (blue crosses in Fig. 4), we can see that there is a prominent episode of dust storms (also described in Wang and Richardson, 2015; Kass et al., 2016) leading to a climatological increase in column dust optical depth around Ls = 170°, which is close to the southern spring equinox (LS = 180°). The corresponding global distribution of CDOD averaged for six days around this period is shown in Fig. 5a. A belt of relatively high CDOD in the southern latitudes encompassing the Hellas basin can be observed (Fig. 5a) during this period, but these values (mainly present in years observed by TES) are the subject of current CDOD retrieval revision (L. Montabone, personal communication). Nonetheless, higher CDOD values are generally present at the Hellas basin and the northeastern region of the Argyre basin. It is worth noting that there are two other stronger episodes of observed CDOD at the Hellas basin. One episode is the period between LS = 220° and LS = 240°, and another episode is around
LS = 320° with a significant dip around the solstice period (blue crosses in Fig. 4). In fact, near the southern summer solstice large dust optical depth is not limited to the Hellas Basin but is a general feature in the southern hemisphere. This double-episode pattern is consistent with the observed “solsticial pause” pattern of regular dust climate (e.g., Wang and Richardson, 2015; Kass et al., 2016; Lewis et al., 2016). The high dust amount during the doublepeak episode periods could be contributed by the cross-equatorial “flushing storms” from the northern hemisphere (e.g., Wang and Richardson, 2015). In this study, the episodes of CDOD over the Hellas basin during the above-mentioned double peak period are not examined here because they likely do not originate in Hellas (Wang and Richardson, 2015) and we solely focus on the Hellas storms which occur around the southern spring equinox. Unless otherwise specified, Hellas storms mentioned hereafter in this study are reffered to dust storms occurring in the Hellas basin around the southern spring equinox. 3.2. Climatology of Hellas storms from model results In the ten years of model simulations, the occurrence of Hellas storms can be found in seven years in different periods between LS = 180° and LS = 200°. The storms usually last for a period between 1 to 5 sols. In this study, we consider that there is an occurrence of Hellas storms in the model results if the CDOD averaged over the Hellas basin is above 0.35 and lasts for at least one day during the period between LS = 180° and LS = 220° The occurrence and evolution of a typical simulated Hellas storm can be seen in the movie supplementary to this paper (Supporting information). The occurrence of Hellas storms from the model results can be seen from the corresponding time series of the ten-year model climatology in Fig. 4 (black triangles, the episode around LS = 200°). The simulation reproduces a maximum in dust optical depth near the southern summer solstice, although it is not
312
K.-C. Chow et al. / Icarus 311 (2018) 306–316
in the period between LS = 220° and LS = 240° in the observational data. However, the simulation does not reproduce the third episode around LS = 320° as in the observational data (Fig. 4). Although the focus of this paper is solely on the equinoctial Hellas storms, future work will be devoted to improve the overall dust cycle in our version of MarsWRF. The corresponding global distribution of the modeled CDOD in the period near the southern spring equinox (Fig. 5a and b) shows a prominent increase in dust amounts in the Hellas Basin (Fig. 5b) similar to that in observations (Fig. 5a). However, the significant dust distribution in the northeast of the Argyre basin from the M2015 data is not reproduced in these model simulations. The vertical profiles of the simulated equinoctial Hellas storm climatology (Fig. 6) show that the dust in the storms is mainly concentrated in the southwestern edge of the basin, which is consistent with previous modeling studies discussed in the introduction. The mean surface wind in the basin is southerly (Fig. 6a), and so the mean flow in the southern edge (region R2) is downslope, which is consistent with the idealized modeling in Siili et al. (1999). It is worth noting that the mean upslope flow in the northern edge of the basin (Fig. 6a) appears to join the upward flow of the Hadley circulation in the equatorial region. As a result, some of the dust in the basin could be brought to higher altitudes in the equatorial region, something that could have also happened during the onset of the 2001 planet-encircling dust storm (Wilson et al.,2008). The corresponding height-longitude profile (Fig. 6b) in the basin shows that the downslope flow in the western edge and the upslope flow in the eastern edge are relatively weak compared with the north-south direction in Fig. 6a, and the prevailing westerly flow does not show an upward transport of dust from the basin. 3.3. Possible cause of Hellas storms occurrence As suggested by the idealized numerical study of Siili et al. (1999), it is possible that the episode of increased dust in the Hellas basin near the period of southern spring equinox is related to three kinds of flow or forcing. They are the buoyancydriven slope wind (strongest at night), the thermally driven flow due to the temperature contrast between the ice-covered seasonal cap and the ice-free surface, and the flow associated with the sublimation of CO2 ice in the ice cap edge region. While the slope wind at night time may be present for the whole year, the latter two forcings should be particularly significant during the southern spring equinox period. During this period, the CO2 ice starts to sublimate, and this can be seen from the model climatology as a sharp decrease in CO2 ice over the southern edge (region R2) of the Hellas basin around the southern spring equinox (Fig. 7). Due to the seasonal increase in near-surface air temperature, the surface air density in the southern edge is decreasing during this period (Fig. 7), although the mass of the atmosphere is increasing due to the increase in CO2 gas released to the atmosphere. This decrease in surface air density does not favor the sharp increase in surface stress over the southern edge of the basin during this period (Fig. 8a). Therefore, the increase in surface stress should be mainly due to the increase in surface winds during this time (Fig. 8b and c). In fact, the time series of model results over the southern edge region (region R2) of the Hellas basin (Fig. 8) suggests that the increase in downslope flow and thus the increase in surface stress over the southern edge of the Hellas basin near the southern spring equinox may be associated with the strong surface temperature difference between the southern edge of the basin and the ice-covered surface in the south (Fig. 8a). Well before the southern spring equinox (
Fig. 8. Time series of some results of the ten-year model climatology over the southern edge region of the Hellas basin (region R2). (a) Daily-mean surface stress (N m−2 , black curve, left axis) and surface temperature difference (K, blue crosses, right axis) between the southern edge region and a region of the same area in the south (region R3). (b) Surface (2 m) meridional wind (m s−1 ) at local night time at 22:00 and 04:00 (blue and black curves respectively), and local day time at 10:00 and 16:00 (red and green curves respectively). The corresponding daily mean result is shown in black crosses. Positive and negative values denote downslope and upslope winds respectively. (c) is similar to (b) but for a close-up view in the period between LS = 120° and LS = 220°. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
K.-C. Chow et al. / Icarus 311 (2018) 306–316
313
Fig. 9. Dust mixing ratio (shading, 10−5 kg/kg) and wind velocity vectors (m s−1 ) at 2 m altitude averaged at 4 local times from each of 10 sols from Ls = 196.3 to 202.2 taken from the ten-year model ensemble average. (a) 04:00. (b) 10:00. (c) 16:00. (d) 22:00. Contours show the topography as in Fig. 3.
though surface stress can be strong during this period due to the intense katabatic wind, dust lifting does not occur over the icecovered regolith. When solar radiation is further increasing in the south, the ice over the southern edge starts to sublimate and so the temperature difference increases significantly. This is accompanied by the sharp increase in surface stress over the southern edge, despite the decrease in surface air density during this period as mentioned above. The increase in temperature difference and stress stops around LS = 200°. Beyond this time the ice in the neighboring region to the south of the southern edge also sublimates as the ice cap edge is retreating to the south, and so the temperature difference and surface stress decrease significantly over the southern edge. As mentioned before, the mean surface wind over the southern edge of the basin is downslope around the southern spring equinox. That the downslope wind prevails during most of the year can be seen from the time series of the daily-mean surface wind (Fig. 8b). When the southern edge is still covered by CO2 ice between LS = 40° and LS = 120° (see Fig. 7), there is no diurnal variation in the surface meridional wind direction and downslope wind prevails over the entire diurnal cycle (Fig. 8b). As the southern
spring equinox approaches, diurnal variation of the surface meridional wind starts to appear. However, downslope wind still prevails over the edge region, not only in the night time but also in most of the day time period, and the overall daily-mean wind is downslope (Fig. 8b). Upslope winds appear in only one of the four daily local times examined (04:0 0, 10:0 0, 16:0 0, and 22:0 0). This reflects that the ice-edge forcing is strong in this period. The magnitude of the downslope wind is generally much stronger than that of the upslope wind. In fact, the model results indicate that during the period of southern spring equinox, the lifting of dust in the southern edge (region R2) of the Hellas basin occurs mainly during the night time around 04:00 (Fig. 9), which can be seen from the largest mixing ratios at this local time (Fig. 9a). The daily-mean wind increases significantly near the southern spring equinox but the increase is mainly for the downslope wind in the night time between LS = 140° and LS = 180° (see Fig. 8b). The downslope wind attains its maximum magnitude of nearly 7 m s−1 around LS = 200°, while at the same period the corresponding maximum value of the upslope wind is about 4 m s−1 , and decreases in magnitude when approaching LS = 200°. The maximum downslope wind around LS = 200° is consistent with the maximum
314
K.-C. Chow et al. / Icarus 311 (2018) 306–316
Fig. 10. Time series (between LS = 150° and LS = 290°) of some variables over the southern edge region of the Hellas basin (region R2) from the control experiment (EXP_CTRL) and the sensitivity experiment with earlier sublimation of CO2 ice (EXP_ESUB). (a) and (b) show respectively the surface temperature difference (K) and the daily-mean surface stress (N m−2 ) similar to Fig. 8a. Black and red curves are from the results of EXP_CTRL and EXP_ESUB respectively. (c) The corresponding time series of column dust optical depth at 700 Pa averaged over the Hellas basin (region R2). The black and red curves are from EXP_CTRL and EXP_ESUB respectively. (d) Surface meridional winds (m s−1 ) at local night time and day time. Black and red curves are from EXP_CTRL and EXP_ESUB respectively, and are averages of two local night times (22:0 0 and 04:0 0) in Fig. 9b. Blue and purple curves are respectively from the two experiments averaged for the two local day times (10:00 and 16:00) in Fig. 8b. Positive and negative values denote downslope and upslope winds respectively. Notice that the time series of (b) and (d) are up to LS = 230°. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
in surface stress during this period (Fig. 8a). After LS = 220°, when the surface temperature difference decreases significantly (Fig. 8a) the magnitudes of the downslope winds in the night time decrease significantly, while the upslope winds in the day time increase significantly. Since the strength of the day time upslope wind is dependent on the solar radiation over the southern slope of the basin, it attains its maximum near the southern summer solstice (LS = 270°). 3.4. Results of sensitivity experiment To further investigate the importance of thermally driven capedge flow and sublimation flow on the occurrence of Hellas storms, a sensitivity experiment has been performed based on the first
year of the ten-year model results. The simulation in this first year is considered as the control experiment (labeled as EXP_CTRL). Typical Hellas storms as described in Section 3.2 do occur around Ls = 200° in this EXP_CTRL simulation. In the sensitivity experiment EXP_ESUB, the sublimation rate of the CO2 ice was increased after LS = 172.9° (Sol 361) by reducing the value of latent heat of the CO2 ice by half. The purpose of this experiment is to speed up the southward retreat of the ice cap edge. The results of this experiment show that the faster retreat of the ice cap edge may lead to an earlier occurrence (around LS = 180° ) of maximum surface temperature difference between the southern edge of the basin (region R2) and the ice covered surface in the south (Fig. 10a). The maximum in surface stress (Fig. 10b) is also coincident with the maximum in surface tem-
K.-C. Chow et al. / Icarus 311 (2018) 306–316
perature difference as in the model climatology (Fig. 8a), and an earlier occurrence of Hellas storms can be seen (Fig. 10c) which is coincident with the maximum stress during this period. The maximum in downslope wind during night time also occurs earlier in EXP_ESUB (Fig. 10d) and is also coincident with the maximum stress. It is worth-noting that the maximum values of downslope wind and surface stress in EXP_ESUB are slightly greater than that in EXP_CTRL, although the maximum surface temperature difference is smaller than that in EXP_CTRL. This may be due to the stronger sublimation flow in EXP_ESUB. 4. Conclusion and discussions The results of this study indicate that the occurrence of Hellas storms during a short period (usually a few sols) around the southern spring equinox may be related to the abrupt increase in surface temperature difference between the southern edge (region R2) of the Hellas basin and the ice-covered region to its south. This may induce a strong thermally-direct "sea-breeze"-type flow during this period. Together with the usual downslope katabatic winds at night-time, the overall abrupt increase in surface stress over the southern edge of the basin may lead to the lifting of dust. Before the equinox period, the southern edge of the basin is still covered by ice, and so the temperature difference is very small. Downslope wind is prevalent over the southern edge region but dust lifting is limited because either the wind stress is not sufficient or the land surface is still covered by CO2 ice. Around the equinox period, when the downward solar radiation is increasing in the southern hemisphere, the ice over the southern edge of the Hellas Basin starts to sublimate and the surface temperature increases. However, at the same time the neighboring region to the south of the southern basin edge (region R3) is still covered by ice. This may lead to a rapid increase in temperature difference, which is not limited to the day time. The ice-free surface temperature over the edge in the night time is still greater than the ice-covered temperature (day or night) in the neighboring region to its south, and thus the regional temperature contrast exists for a whole day. The corresponding southerly cap-edge wind, together with the downslope katabatic wind over the basin may lead to the sufficient wind stress for dust lifting at night. In the local day time, the slope forcing is upward in the southern edge of the basin, which is opposite to the other two forcings, and so is not favorable for the occurrence of dust lifting. The large temperature difference only lasts for a relatively short period and decreases quickly when the ice cap in the south is retreating southward. Without the thermally driven capedge wind, the slope wind over the basin may be not sufficient to initiate dust lifting over the southern edge. The above results are generally consistent with the results of the numerical simulations by Toigo et al. (2002) discussed in Strausberg et al. (2005) , in which the surface stress at the southern edge of the Hellas Basin is maximum around midnight. In the ten-year model climatology simulated in this study, there are seven out of the ten years that Hellas storms do occur. This issue of interannual variability in the simulations results should be further investigated. In fact, results of some preliminarily investigation (results not shown) suggest that the absence of Hellas storms in some years of simulations may be related to the relatively dustier atmosphere during the southern spring equinox period. It is speculated that an atmosphere relatively rich in background dust may reduce the surface temperature difference between the ice cap and the adjacent land surface, and so may reduce the strength of the cap-edge wind. Finally, it should be mentioned that there are some deficiencies in the model simulations. First, our simulation does not reproduce perfectly the observed dust distribution. However, we believe that this should not significantly invalidate the findings discussed in this study, since the occurrence of the modeled Hellas storms
315
considered is mainly associated with surface wind stress, which in turn is mainly dependent on the topography, the annual evolution of solar radiation and the seasonal ice cap. Second, there is a slight discrepancy in the occurrence of the storms between the model climatology (LS = 200°) and observations (LS = 180°). Despite this discrepancy, we believe that the model results have revealed some basic features of the dynamics in the Hellas Basin during the spring equinox period. The discrepancy may be related to the coarse resolution (5°) of the model domain. For this resolution, the boundary of the ice cape edge may not be in the true latitudinal position during that period. In fact, the results of the present study suggest that the occurrence of Hellas storms is rather sensitive to the latitudinal retreat of the ice cap edge. For a further study, mesoscale modeling coupled to global circulation modeling (possibly with the inclusion of data assimilation) can help to get more insights into the phenomenon of Hellas storms, as well as their connection with equinoctial global-scale dust events.
Acknowledgments This research is funded by the grants from the FDCT of Macau (grant no. 039/2013/A2 and 080/2015/A3). The data of the numerical simulations, as well as the model codes are freely available from the first author on request. The observational data of column dust optical depth based on Montabone et al. (2015) was obtained from the Laboratoire de Météorologie Dynamique du CNRS (LMD) and is available at http://www-mars.lmd.jussieu.fr/ mars/dust_climatology/. The two present reviewer and Claire Newman who previously reviewed this paper are highly appreciated for their efforts in correcting numerous errors in the manuscript and their suggestion in improving the manuscript.
Supplementary materials Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.icarus.2018.04.011.
References Basu, S., Wilson, J., Richardson, M.I., et al., 2006. Simulation of spontaneous and variable global dust storms with the GFDL Mars GCM. J. Geophys. Res. 111 (E10), 9004. doi:10.1029/2005JE002660. Cantor, B.A., 2007. MOC observations of the 2001 Mars planet-encircling dust storm. Icarus 186, 60–96. Guo, X., Lawson, W.G., Richardson, M.I., et al., 2009. Fitting the Viking Lander surface pressure cycle with a Mars general circulation model. J. Geophys. Res 114, E07006. doi:10.1029/2008JE003302. Kass, D.M., Kleinbohl, A., McCleese, D.J., et al., 2016. Interannual similarity in the Martian atmosphere during the dust storm season. Geophys. Res. Lett 43, 6111–6118. Lewis, S.R., Mulholland, D.P., Read, P.L., et al., 2016. The solsticial pause on Mars: 1. A planetary wave reanalysis. Icarus 264, 456–464. doi:10.1016/j.icarus.2015.08. 039. Martin, L.J., Zurek, R.W., 1993. An analysis of the history of dust activity on Mars. J. Geophys. Res. 98 (E2), 3221–3246. Montabone, L., Forget, F., Millour, E., et al., 2015. Eight-year climatology of dust optical depth on Mars. Icarus 252, 65–95. Montabone, L., Marsh, K., Lewis, S.R., et al., 2014. The Mars analysis correction data assimilation (MACDA) dataset V1.0. Geosci. Data J 1 129–139. doi:10.1002/gdj3. 13. Newman, C.E., Lewis, S.R., Read, P.L., et al., 2002. Modeling the Martian dust cycle: 2. Multiannual radiatively active dust transport simulations. J. Geophys. Res 107 (E12), 5124. doi:10.1029/20 02JE0 01920. Newman, C.E., Richardson, M.I., 2015. The impact of surface dust source exhaustion on the martian dust cycle, dust storms and interannual variability, as simulated by the MarsWRF general circulation model. Icarus 257, 47–87. Ogohara, K., Satomura, T., 2008. Northward movement of Martian dust localized in the region of the Hellas basin. Geophys. Res. Lett. 35, L13201. doi:10.1029/ 2008GL034546.
316
K.-C. Chow et al. / Icarus 311 (2018) 306–316
Richardson, M.I., Toigo, A.D., Newman, C.E., 2007. PlanetWRF: a general purpose, local to global numerical model for planetary atmospheric and climate dynamics. J. Geophys. Res. 112. doi:10.1029/20 06JE0 02825. Siili, T., Haberle, R.M., Murphy, J.R., 1997. Sensitivity of Martian southern polar cap edge winds and surface stresses to dust optical thickness and to the large-scale sublimation flow. Planet. Space Sci. 47, 951–970. Siili, T., Haberle, R.M., Murphy, J.R., Savijarvi, H., 1999. Modelling of the combined late winter ice cap edge and slope winds in Mars’ Hellas and Argyre regions. Planet. Space Sci. 47, 951–970. Skamarock, W.C., Klemp, J.B., 2008. A time-split non-hydrostatic atmospheric model for weather research and forecasting applications. J. Comput. Phys. 227 (7), 3465–3485. Strausberg, M.J., Wang, H., Richardson, M.I., et al., 2005. Observations of the initiation and evolution of the 2001 Mars global dust storm. J. Geophys. Res. 110, E02006. doi:10.1029/2004JE002361.
Toigo, A.D., Richardson, M.I., Wilson, R.J., et al., 2002. A first look at dust lifting and dust storms near the south pole of Mars with a mesoscale model. J. Geophys. Res. 107, 5050–5062. doi:10.1029/20 01JE0 01592. Toigo, A.D., Lee, C., Newman, C.E., et al., 2012. The impact of resolution on the dynamics of the Martian global atmosphere: varying resolution studies with the MarsWRF GCM. Icarus 221 (1), 276–288. Wang, H., Richardson, M.I., 2015. The origin, evolution, and trajectory of large dust storms on Mars during Mars years 24–30 (1999–2011). Icarus 251, 112–127. Wilson, R.J., Haberle, R.M., Noble, J., et al., 2008. Simulation of the 2001 planet-encircling dust storm with the NASA/NOAA Mars general circulation model. Third International Workshop on the Mars Atmosphere: Modeling and Observations Abstract no. 9023.
Contents lists available at ScienceDirect
Icarus journal homepage: www.elsevier.com/locate/icarus
Dust activity over the Hellas basin of Mars during the period of southern spring equinox Kim-Chiu Chow∗, Kwing-lam Chan, Jing Xiao Space Science Institute/ Lunar and Planetary Science Laboratory, Macau University of Science and Technology, Macau
a r t i c l e
i n f o
Article history: Received 1 September 2017 Revised 9 April 2018 Accepted 13 April 2018 Available online 22 April 2018 Keywords: Mars Dust storms Hellas basin Cap-edge flow Sublimation flow Slope wind
a b s t r a c t Dust storms that last for more than one Martian day frequently occur in Hellas basin of Mars around southern spring equinox. The dynamics behind the formation of these “Hellas storms”, however, is still unclear. In this study, a ten-year climatology of Martian atmospheric dust is simulated with the MarsWRF global climate model. Results suggest that occurrence of Hellas storms during southern spring equinox is related to the abrupt increase in the surface temperature contrast between the southern edge of the Hellas basin and the carbon dioxide ice covered south polar cap. Significant dust lifting over the southern edge of the Hellas basin occurs mainly during night time, due to the particular strength of the prevailing downslope flows at the site. © 2018 Elsevier Inc. All rights reserved.
1. Introduction It has been reported (e.g., Martin and Zurek, 1993; Wang et al., 2015) that Hellas basin is a region on Mars where dust storms can be commonly observed. The basin is a huge crater region about 30 0 0 km in diameter and 8 km in depth. The steep topography of the basin boundary likely generates a slope-wind circulation associated with the diurnal solar cycle. Observational studies (e.g., Strausberg et al., 20 05; Cantor, 20 07) have shown that the wellknown planet-encircling dust event that occurred in 2001 was initiated in Hellas Basin around the time of the southern spring equinox (LS = 180°). In particular, Strausberg et al. (2005) pointed out that in the initiation phase of this severe dust event, a sequence of dust storm events was observed between the southern edge of Hellas basin and the south polar ice cap edge. The dust was transported out of the basin and propagated to the east in later stage, and facilitated the development of a second dust lifting center over the Syria/Solis/Daedalia region. The occurrence of dust storms over Hellas Basin around the southern spring equinox (LS = 180°) period has been noted in some general circulation model simulations (e.g., Newman et al., 2002; Basu et al., 2006; Newman and Richardson, 2015). In those studies, the small-scale to regional-scale dust storms (“Hellas storms”) typically lasted for a few sols (a Martian solar day) and usually initiated at the southern or southwestern edge of the basin. Results ∗
Corresponding author. E-mail address: [email protected] (K.-C. Chow).
https://doi.org/10.1016/j.icarus.2018.04.011 0019-1035/© 2018 Elsevier Inc. All rights reserved.
of a high resolution mesoscale model by Toigo et al. (2002) were shown in Strausberg et al. (2005); in which a high-stress region in the southern edge of Hellas Basin was displayed around LS = 180°. Furthermore, the numerical experiments of Ogohara and Satomura (2008) showed that dust injected at the southwestern edge of the Hellas basin during southern spring equinox may form a regional dust storm over the basin, similar to the Hellas storms simulated by Newman et al. (2002), Newman and Richardson (2015), and Basu et al. (2006). In the modeling studies mentioned above, the occurrence of Hellas storms are believed to be due to the temperature difference between the basin and the southern region. However, there has been only limited analysis (e.g., Strausberg et al., 2005) to explain why the Hellas storms are initiated only during this particular short period, and in this particular region. Nevertheless, there were still some modeling and dynamical studies related to this topic. In Siili et al. (1997, 1999), a two-dimensional mesoscale model was used to study the circulation over a flat ice covered cap edge region. From the idealized numerical simulations, they found that if only the thermally driven circulation due to ice-land temperature contrast is present, the resulting surface stress may be insufficient for dust lifting. However, the stress may be sufficient if a sublimation flow due to the ice cap is present as well. They also found that the presence of an ice cover on the top edge of the valley may significantly enhance the downslope wind (referred as ice-edge forcing) and surface stress during the night time. In the day time, the enhanced temperature contrast between the ice-covered valley slope at the top and the ice free ground surface in the bottom may
K.-C. Chow et al. / Icarus 311 (2018) 306–316
significantly weaken the upslope wind. The addition of sublimation flow over the edge of the valley may also substantially increase the downslope wind. In Toigo et al. (2002), a three-dimensional mesoscale model with realistic physics and nested with a Mars general circulation model was used to examine dust lifting over the ice cap edge during southern summer. In particular, the three major forcings related to dust lifting over cap edge: thermal contrast of cap edge, slope flow, and sublimation flow due to CO2 ice were examined. With a more realistic simulation of the ice cap, they found that surface stress for dust lifting over the cap edge region is generally large during local afternoon. They also found that the sublimation flow is not important to dust lifting during the period considered in their study, and slope flow due to topography could be more important to dust lifting compared with the effect of thermal contrast over the cap edge region. In particular, their results presented in Strausberg et al. (2005) suggest that a high surface stress region can be found over the southern edge of the Hellas Basin during the period of the southern spring equinox, and the stress is at maximum during night time. In the afternoon, upslope wind is dominant over the western and northern rims of the basin. These previous studies provide some insight on the dynamics of dust lifting over the cap edge region. However, the effect of the above three forcing over the Hellas basin, particularly during the period of the southern spring equinox was basically not investigated in detail. In the present study, the formation of Hellas storms has been investigated based on the results from a general circulation model for Mars. This study is attempting to address the question why Hellas storms are commonly initiated and active in the basin for a short period around the southern spring equinox, which could be related to the retreat of ice cap as will be discussed in this paper. This issue has not been addressed in the studies mentioned above (Silli et al., 1997, 1999; Toigo et al., 2002; Strausberg et al., 2005). We believe that a better understanding on the initiation of Hellas storms may help us understand more about the regular dust cycle, and hopefully may provide more insight into the initiation mechanism of the 2001 planet-encircling dust event occurred around the time of the southern spring equinox. 2. The numerical model The general circulation model (GCM) MarsWRF (Richardson et al., 2007; Toigo et al., 2012) is used in this study to simulate the climate of Mars. MarsWRF is the Mars version of the PlanetWRF model (Richardson et al., 2007), which was developed by the National Center for Atmospheric Research (NCAR) Weather Research and Forecasting (WRF) model for Earth (Skamarock and Klemp, 2008). It has been illustrated in Richardson et al. (2007) and Toigo et al. (2012) that MarsWRF is capable of realistically reproducing climate features such as the large-scale general circulation and temperature field on Mars. In this study, the global domain of the MarsWRF model has been set to 36 latitude × 72 longitude grid points (horizontal spatial resolution of about 5° or 300 km in the equatorial region). There are 52 vertical levels in terrain-following hydrostatic-pressure vertical coordinate (Skamarock and Klemp, 2008), and the model top is set at 0.0057 Pa (about 80 km in altitude). Half of these 52 levels are located between the model top and the pressure level of 100 Pa. Hydrostatic dynamics is used for the model runs, and traditional Rayleigh damping is applied to the dynamical variables at the top three vertical layers. The radiation scheme for short- and longwave radiation employed in these simulations is the "wide band model" scheme as described in Richardson et al. (2007), Toigo et al. (2012) and Newman and Richardson (2015). This scheme considers the heating/cooling effects of dust and carbon dioxide (CO2 ). The planetary boundary layer scheme and land surface scheme were
307
Fig. 1. Seasonal evolution of annually average zonal-mean column dust optical depth at 9.3 μm (normalized to a 700 Pa surface) from the ten-year average climatology of model simulation results (upper panel) and the re-constructed observational data (Montabone et al., 2015) averaged for the six Martian years of observations that do not contain a global-scale dust storm (lower panel).
largely adapted from existing schemes in WRF forf Earth, as described in Richardson et al. (2007) and Toigo et al. (2012). The model also includes other physical process parameterizations specific to Mars, such as the carbon dioxide cycle (Guo et al., 2009) and the dust cycle (Newman and Richardson, 2015). However, no parameterization of the water cycle is included in these simulations. The parameterization of dust processes in the model includes two interactive dust schemes similar to those used in Newman and Richardson (2015), and dust is assumed to be available everywhere and at all times over the whole planet surface except those surfaces with ice cover. The first scheme is similar to usual dust models on Earth, in which the lifting of dust is proportional to the surface wind stress. Dust lifting occurs over the surface when the local near-surface stress exceeds a particular threshold value (constant value 0.042 N m−2 in this study). The second scheme provides most of the background dust, which is parameterized as dust lifting by dust devils. The amount of dust lifting is dependent on the temperature difference between the surface and the air above, as well as the sensible heat flux. The radiation scheme in the model is interactive with dust so that the suspended dust may change the atmospheric radiation and thus the circulation. The model was run for eleven Martian years, initially starting from the time of northern spring equinox (LS = 0°). The first year is considered as a spin-up period. Therefore, only ten years of simulation results are used in this study and their ensemble mean is defined as the climatology of the model results. Some important model parameters such as the threshold stress of dust lifting and lifting rate of dust devils were tuned to simulate the regular climate of Mars. When comparing the annual variation of the zonal-mean column dust optical depth (CDOD) from the tenyear averaged model climatology with the 6-year averaged observational data (described below), the model simulation is able to reproduce the essential features of the climatological dust distribution of Mars (Fig. 1), particularly the significant increase in global dust opacity after LS = 135°, and the two episodes of increased dust opacity around LS = 240° and LS = 320° (although there is a certain time shift of the episodes). However, the simulation overestimates the dust opacity over the northern winter polar region, and during the observed winter “solsticial pause” period (e.g.,
308
K.-C. Chow et al. / Icarus 311 (2018) 306–316
Fig. 2. Annual variation of zonal-mean carbon dioxide ice (kg m−2 ) of the four-year averaged data (MY24–MY27) from the Mars Analysis Correction Data Assimilation (MACDA v1.0, contours). The corresponding result from the ten-year model climatology is shown in shading.
Lewis et al., 2016; Kass et al., 2016). Although the dust climate simulated in these simulations may not produce all of the details when compared with observations, the results of the model climatology suggest that these simulations are reproducing the basic general circulation of Mars (see figures in Toigo et al., 2012). For the carbon dioxide cycle, the model climatology suggests that the simulation is capable of capturing the evolution of CO2 ice cap (Fig. 2) when compared with the Mars Analysis Correction Data Assimilation (MACDA v1.0), publicly available from the Centre for Environmental Data Analysis (Montabone et al., 2014). Although a small time delay can be observed in the southward retreat of the ice cap edge during the southern spring equinox period, this comparison might be biased by the fact that the CO2 ice cap in the
Fig. 3. The Hellas Basin region in the model with topography shown in contours. Shading is the surface carbon dioxide ice (kg m−2 ) from the ten-year model climatology averaged for ten sols from LS = 196.3° to 202.2°. The rectangles R1, R2, and R3 are some regions defined in this study. R1: Hellas Basin (30° S to 60° S, 45° E to 90° E). R2: southern edge of Hellas Basin (50° S to 60° S, 45° E to 90° E). R3: neighboring region south of the southern edge (60° S to 70° S, 45° E to 90° E).
reanalysis itself has not yet been fully validated by observations. The CO2 ice cap over the region of Hellas basin in the model climatology can also be seen from the distribution of surface CO2 ice averaged for ten sols during the southern spring equinox (Fig. 3). 3. The occurrence of hellas storms To compare the model results with observations, the CDOD data re-constructed by Montabone et al. (2015, M2015 hereafter) has
Fig. 4. Time series of column dust optical depth at 9.3 μm (normalized to the 700 Pa constant pressure level) over the Hellas basin (region R1). Blue crosses are from the re-constructed observational data (Montabone et al., 2015) averaged for six years without global-scale dust event. Black triangles are from the corresponding daily-mean ten-year climatology of model results. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
K.-C. Chow et al. / Icarus 311 (2018) 306–316
309
Fig. 5. Column dust optical depth at 9.3 μm (normalized to surface level 700 Pa) around the southern spring equinox. Contours are topography with intervals 20 0 0 m. (a) Observational data (Montabone et al., 2015) averaged for the six Mars years excluding the 2001 and 2007 global dust storms, and averaged for 6 sols from LS = 168.9° to 171.7° (b) From the ten-year averaged model climatology, further averaged for 10 sols from LS = 196.3° to 202.2°. Vectors are near surface (2 m) wind averaged in this period. .
310
K.-C. Chow et al. / Icarus 311 (2018) 306–316
Fig. 6. Vertical profiles of dust mixing ratio (10−5 kg/ kg) and velocity vectors (m s−1 , vertical component multiplied by 100) over the Hellas basin region from the model results (ten-year climatology ) averaged over 10 sols from LS = 196.3° to 202.2°. (a) Height-latitude profile averaged over the longitudes from 45° E to 90° E. (b) The corresponding height-longitude profile averaged over latitudes from 30° S to 60° S.
K.-C. Chow et al. / Icarus 311 (2018) 306–316
311
Fig. 7. Time series of daily mean results of the ten-year model climatology over the southern edge region of the Hellas basin (region R2). Blue crosses denote surface air density (kg m−3 , right axis), and black curve denotes surface carbon dioxide ice (kg m−2 , left axis). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
been used. The data are based on the Thermal Emission Spectrometer (TES) and the Thermal Emission Imaging System (THEMIS) observations, as well as the estimates of CDOD from the Mars Reconnaissance Orbiter (MRO)/Mars Climate Sounder in MY 28–33. These gridded, daily maps cover Martian years (MY) 24 to 32. As in M2015, we excluded the years of the 2001 and 2007 global dust storms and used the six years without them (MY24, 26, 27, 29, 30, and 31) to obtain the climatology of dust optical depth. 3.1. Climatology of Hellas storms from observational data The region of the Hellas basin focused in this study has the topography shown in Fig. 3. In this study, the analyzes of the results over the region of Hellas Basin are referred to three rectangular regions R1, R2, and R3 (Fig. 3). The Hellas Basin is referred to the region R1, and the southern edge of the Hellas Basin is referred to the region R2, while R3 is a neighboring region south of the southern edge and has the same area as R2. From the time series of the M2015 climatological data averaged over the Hellas basin (blue crosses in Fig. 4), we can see that there is a prominent episode of dust storms (also described in Wang and Richardson, 2015; Kass et al., 2016) leading to a climatological increase in column dust optical depth around Ls = 170°, which is close to the southern spring equinox (LS = 180°). The corresponding global distribution of CDOD averaged for six days around this period is shown in Fig. 5a. A belt of relatively high CDOD in the southern latitudes encompassing the Hellas basin can be observed (Fig. 5a) during this period, but these values (mainly present in years observed by TES) are the subject of current CDOD retrieval revision (L. Montabone, personal communication). Nonetheless, higher CDOD values are generally present at the Hellas basin and the northeastern region of the Argyre basin. It is worth noting that there are two other stronger episodes of observed CDOD at the Hellas basin. One episode is the period between LS = 220° and LS = 240°, and another episode is around
LS = 320° with a significant dip around the solstice period (blue crosses in Fig. 4). In fact, near the southern summer solstice large dust optical depth is not limited to the Hellas Basin but is a general feature in the southern hemisphere. This double-episode pattern is consistent with the observed “solsticial pause” pattern of regular dust climate (e.g., Wang and Richardson, 2015; Kass et al., 2016; Lewis et al., 2016). The high dust amount during the doublepeak episode periods could be contributed by the cross-equatorial “flushing storms” from the northern hemisphere (e.g., Wang and Richardson, 2015). In this study, the episodes of CDOD over the Hellas basin during the above-mentioned double peak period are not examined here because they likely do not originate in Hellas (Wang and Richardson, 2015) and we solely focus on the Hellas storms which occur around the southern spring equinox. Unless otherwise specified, Hellas storms mentioned hereafter in this study are reffered to dust storms occurring in the Hellas basin around the southern spring equinox. 3.2. Climatology of Hellas storms from model results In the ten years of model simulations, the occurrence of Hellas storms can be found in seven years in different periods between LS = 180° and LS = 200°. The storms usually last for a period between 1 to 5 sols. In this study, we consider that there is an occurrence of Hellas storms in the model results if the CDOD averaged over the Hellas basin is above 0.35 and lasts for at least one day during the period between LS = 180° and LS = 220° The occurrence and evolution of a typical simulated Hellas storm can be seen in the movie supplementary to this paper (Supporting information). The occurrence of Hellas storms from the model results can be seen from the corresponding time series of the ten-year model climatology in Fig. 4 (black triangles, the episode around LS = 200°). The simulation reproduces a maximum in dust optical depth near the southern summer solstice, although it is not
312
K.-C. Chow et al. / Icarus 311 (2018) 306–316
in the period between LS = 220° and LS = 240° in the observational data. However, the simulation does not reproduce the third episode around LS = 320° as in the observational data (Fig. 4). Although the focus of this paper is solely on the equinoctial Hellas storms, future work will be devoted to improve the overall dust cycle in our version of MarsWRF. The corresponding global distribution of the modeled CDOD in the period near the southern spring equinox (Fig. 5a and b) shows a prominent increase in dust amounts in the Hellas Basin (Fig. 5b) similar to that in observations (Fig. 5a). However, the significant dust distribution in the northeast of the Argyre basin from the M2015 data is not reproduced in these model simulations. The vertical profiles of the simulated equinoctial Hellas storm climatology (Fig. 6) show that the dust in the storms is mainly concentrated in the southwestern edge of the basin, which is consistent with previous modeling studies discussed in the introduction. The mean surface wind in the basin is southerly (Fig. 6a), and so the mean flow in the southern edge (region R2) is downslope, which is consistent with the idealized modeling in Siili et al. (1999). It is worth noting that the mean upslope flow in the northern edge of the basin (Fig. 6a) appears to join the upward flow of the Hadley circulation in the equatorial region. As a result, some of the dust in the basin could be brought to higher altitudes in the equatorial region, something that could have also happened during the onset of the 2001 planet-encircling dust storm (Wilson et al.,2008). The corresponding height-longitude profile (Fig. 6b) in the basin shows that the downslope flow in the western edge and the upslope flow in the eastern edge are relatively weak compared with the north-south direction in Fig. 6a, and the prevailing westerly flow does not show an upward transport of dust from the basin. 3.3. Possible cause of Hellas storms occurrence As suggested by the idealized numerical study of Siili et al. (1999), it is possible that the episode of increased dust in the Hellas basin near the period of southern spring equinox is related to three kinds of flow or forcing. They are the buoyancydriven slope wind (strongest at night), the thermally driven flow due to the temperature contrast between the ice-covered seasonal cap and the ice-free surface, and the flow associated with the sublimation of CO2 ice in the ice cap edge region. While the slope wind at night time may be present for the whole year, the latter two forcings should be particularly significant during the southern spring equinox period. During this period, the CO2 ice starts to sublimate, and this can be seen from the model climatology as a sharp decrease in CO2 ice over the southern edge (region R2) of the Hellas basin around the southern spring equinox (Fig. 7). Due to the seasonal increase in near-surface air temperature, the surface air density in the southern edge is decreasing during this period (Fig. 7), although the mass of the atmosphere is increasing due to the increase in CO2 gas released to the atmosphere. This decrease in surface air density does not favor the sharp increase in surface stress over the southern edge of the basin during this period (Fig. 8a). Therefore, the increase in surface stress should be mainly due to the increase in surface winds during this time (Fig. 8b and c). In fact, the time series of model results over the southern edge region (region R2) of the Hellas basin (Fig. 8) suggests that the increase in downslope flow and thus the increase in surface stress over the southern edge of the Hellas basin near the southern spring equinox may be associated with the strong surface temperature difference between the southern edge of the basin and the ice-covered surface in the south (Fig. 8a). Well before the southern spring equinox (
Fig. 8. Time series of some results of the ten-year model climatology over the southern edge region of the Hellas basin (region R2). (a) Daily-mean surface stress (N m−2 , black curve, left axis) and surface temperature difference (K, blue crosses, right axis) between the southern edge region and a region of the same area in the south (region R3). (b) Surface (2 m) meridional wind (m s−1 ) at local night time at 22:00 and 04:00 (blue and black curves respectively), and local day time at 10:00 and 16:00 (red and green curves respectively). The corresponding daily mean result is shown in black crosses. Positive and negative values denote downslope and upslope winds respectively. (c) is similar to (b) but for a close-up view in the period between LS = 120° and LS = 220°. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
K.-C. Chow et al. / Icarus 311 (2018) 306–316
313
Fig. 9. Dust mixing ratio (shading, 10−5 kg/kg) and wind velocity vectors (m s−1 ) at 2 m altitude averaged at 4 local times from each of 10 sols from Ls = 196.3 to 202.2 taken from the ten-year model ensemble average. (a) 04:00. (b) 10:00. (c) 16:00. (d) 22:00. Contours show the topography as in Fig. 3.
though surface stress can be strong during this period due to the intense katabatic wind, dust lifting does not occur over the icecovered regolith. When solar radiation is further increasing in the south, the ice over the southern edge starts to sublimate and so the temperature difference increases significantly. This is accompanied by the sharp increase in surface stress over the southern edge, despite the decrease in surface air density during this period as mentioned above. The increase in temperature difference and stress stops around LS = 200°. Beyond this time the ice in the neighboring region to the south of the southern edge also sublimates as the ice cap edge is retreating to the south, and so the temperature difference and surface stress decrease significantly over the southern edge. As mentioned before, the mean surface wind over the southern edge of the basin is downslope around the southern spring equinox. That the downslope wind prevails during most of the year can be seen from the time series of the daily-mean surface wind (Fig. 8b). When the southern edge is still covered by CO2 ice between LS = 40° and LS = 120° (see Fig. 7), there is no diurnal variation in the surface meridional wind direction and downslope wind prevails over the entire diurnal cycle (Fig. 8b). As the southern
spring equinox approaches, diurnal variation of the surface meridional wind starts to appear. However, downslope wind still prevails over the edge region, not only in the night time but also in most of the day time period, and the overall daily-mean wind is downslope (Fig. 8b). Upslope winds appear in only one of the four daily local times examined (04:0 0, 10:0 0, 16:0 0, and 22:0 0). This reflects that the ice-edge forcing is strong in this period. The magnitude of the downslope wind is generally much stronger than that of the upslope wind. In fact, the model results indicate that during the period of southern spring equinox, the lifting of dust in the southern edge (region R2) of the Hellas basin occurs mainly during the night time around 04:00 (Fig. 9), which can be seen from the largest mixing ratios at this local time (Fig. 9a). The daily-mean wind increases significantly near the southern spring equinox but the increase is mainly for the downslope wind in the night time between LS = 140° and LS = 180° (see Fig. 8b). The downslope wind attains its maximum magnitude of nearly 7 m s−1 around LS = 200°, while at the same period the corresponding maximum value of the upslope wind is about 4 m s−1 , and decreases in magnitude when approaching LS = 200°. The maximum downslope wind around LS = 200° is consistent with the maximum
314
K.-C. Chow et al. / Icarus 311 (2018) 306–316
Fig. 10. Time series (between LS = 150° and LS = 290°) of some variables over the southern edge region of the Hellas basin (region R2) from the control experiment (EXP_CTRL) and the sensitivity experiment with earlier sublimation of CO2 ice (EXP_ESUB). (a) and (b) show respectively the surface temperature difference (K) and the daily-mean surface stress (N m−2 ) similar to Fig. 8a. Black and red curves are from the results of EXP_CTRL and EXP_ESUB respectively. (c) The corresponding time series of column dust optical depth at 700 Pa averaged over the Hellas basin (region R2). The black and red curves are from EXP_CTRL and EXP_ESUB respectively. (d) Surface meridional winds (m s−1 ) at local night time and day time. Black and red curves are from EXP_CTRL and EXP_ESUB respectively, and are averages of two local night times (22:0 0 and 04:0 0) in Fig. 9b. Blue and purple curves are respectively from the two experiments averaged for the two local day times (10:00 and 16:00) in Fig. 8b. Positive and negative values denote downslope and upslope winds respectively. Notice that the time series of (b) and (d) are up to LS = 230°. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
in surface stress during this period (Fig. 8a). After LS = 220°, when the surface temperature difference decreases significantly (Fig. 8a) the magnitudes of the downslope winds in the night time decrease significantly, while the upslope winds in the day time increase significantly. Since the strength of the day time upslope wind is dependent on the solar radiation over the southern slope of the basin, it attains its maximum near the southern summer solstice (LS = 270°). 3.4. Results of sensitivity experiment To further investigate the importance of thermally driven capedge flow and sublimation flow on the occurrence of Hellas storms, a sensitivity experiment has been performed based on the first
year of the ten-year model results. The simulation in this first year is considered as the control experiment (labeled as EXP_CTRL). Typical Hellas storms as described in Section 3.2 do occur around Ls = 200° in this EXP_CTRL simulation. In the sensitivity experiment EXP_ESUB, the sublimation rate of the CO2 ice was increased after LS = 172.9° (Sol 361) by reducing the value of latent heat of the CO2 ice by half. The purpose of this experiment is to speed up the southward retreat of the ice cap edge. The results of this experiment show that the faster retreat of the ice cap edge may lead to an earlier occurrence (around LS = 180° ) of maximum surface temperature difference between the southern edge of the basin (region R2) and the ice covered surface in the south (Fig. 10a). The maximum in surface stress (Fig. 10b) is also coincident with the maximum in surface tem-
K.-C. Chow et al. / Icarus 311 (2018) 306–316
perature difference as in the model climatology (Fig. 8a), and an earlier occurrence of Hellas storms can be seen (Fig. 10c) which is coincident with the maximum stress during this period. The maximum in downslope wind during night time also occurs earlier in EXP_ESUB (Fig. 10d) and is also coincident with the maximum stress. It is worth-noting that the maximum values of downslope wind and surface stress in EXP_ESUB are slightly greater than that in EXP_CTRL, although the maximum surface temperature difference is smaller than that in EXP_CTRL. This may be due to the stronger sublimation flow in EXP_ESUB. 4. Conclusion and discussions The results of this study indicate that the occurrence of Hellas storms during a short period (usually a few sols) around the southern spring equinox may be related to the abrupt increase in surface temperature difference between the southern edge (region R2) of the Hellas basin and the ice-covered region to its south. This may induce a strong thermally-direct "sea-breeze"-type flow during this period. Together with the usual downslope katabatic winds at night-time, the overall abrupt increase in surface stress over the southern edge of the basin may lead to the lifting of dust. Before the equinox period, the southern edge of the basin is still covered by ice, and so the temperature difference is very small. Downslope wind is prevalent over the southern edge region but dust lifting is limited because either the wind stress is not sufficient or the land surface is still covered by CO2 ice. Around the equinox period, when the downward solar radiation is increasing in the southern hemisphere, the ice over the southern edge of the Hellas Basin starts to sublimate and the surface temperature increases. However, at the same time the neighboring region to the south of the southern basin edge (region R3) is still covered by ice. This may lead to a rapid increase in temperature difference, which is not limited to the day time. The ice-free surface temperature over the edge in the night time is still greater than the ice-covered temperature (day or night) in the neighboring region to its south, and thus the regional temperature contrast exists for a whole day. The corresponding southerly cap-edge wind, together with the downslope katabatic wind over the basin may lead to the sufficient wind stress for dust lifting at night. In the local day time, the slope forcing is upward in the southern edge of the basin, which is opposite to the other two forcings, and so is not favorable for the occurrence of dust lifting. The large temperature difference only lasts for a relatively short period and decreases quickly when the ice cap in the south is retreating southward. Without the thermally driven capedge wind, the slope wind over the basin may be not sufficient to initiate dust lifting over the southern edge. The above results are generally consistent with the results of the numerical simulations by Toigo et al. (2002) discussed in Strausberg et al. (2005) , in which the surface stress at the southern edge of the Hellas Basin is maximum around midnight. In the ten-year model climatology simulated in this study, there are seven out of the ten years that Hellas storms do occur. This issue of interannual variability in the simulations results should be further investigated. In fact, results of some preliminarily investigation (results not shown) suggest that the absence of Hellas storms in some years of simulations may be related to the relatively dustier atmosphere during the southern spring equinox period. It is speculated that an atmosphere relatively rich in background dust may reduce the surface temperature difference between the ice cap and the adjacent land surface, and so may reduce the strength of the cap-edge wind. Finally, it should be mentioned that there are some deficiencies in the model simulations. First, our simulation does not reproduce perfectly the observed dust distribution. However, we believe that this should not significantly invalidate the findings discussed in this study, since the occurrence of the modeled Hellas storms
315
considered is mainly associated with surface wind stress, which in turn is mainly dependent on the topography, the annual evolution of solar radiation and the seasonal ice cap. Second, there is a slight discrepancy in the occurrence of the storms between the model climatology (LS = 200°) and observations (LS = 180°). Despite this discrepancy, we believe that the model results have revealed some basic features of the dynamics in the Hellas Basin during the spring equinox period. The discrepancy may be related to the coarse resolution (5°) of the model domain. For this resolution, the boundary of the ice cape edge may not be in the true latitudinal position during that period. In fact, the results of the present study suggest that the occurrence of Hellas storms is rather sensitive to the latitudinal retreat of the ice cap edge. For a further study, mesoscale modeling coupled to global circulation modeling (possibly with the inclusion of data assimilation) can help to get more insights into the phenomenon of Hellas storms, as well as their connection with equinoctial global-scale dust events.
Acknowledgments This research is funded by the grants from the FDCT of Macau (grant no. 039/2013/A2 and 080/2015/A3). The data of the numerical simulations, as well as the model codes are freely available from the first author on request. The observational data of column dust optical depth based on Montabone et al. (2015) was obtained from the Laboratoire de Météorologie Dynamique du CNRS (LMD) and is available at http://www-mars.lmd.jussieu.fr/ mars/dust_climatology/. The two present reviewer and Claire Newman who previously reviewed this paper are highly appreciated for their efforts in correcting numerous errors in the manuscript and their suggestion in improving the manuscript.
Supplementary materials Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.icarus.2018.04.011.
References Basu, S., Wilson, J., Richardson, M.I., et al., 2006. Simulation of spontaneous and variable global dust storms with the GFDL Mars GCM. J. Geophys. Res. 111 (E10), 9004. doi:10.1029/2005JE002660. Cantor, B.A., 2007. MOC observations of the 2001 Mars planet-encircling dust storm. Icarus 186, 60–96. Guo, X., Lawson, W.G., Richardson, M.I., et al., 2009. Fitting the Viking Lander surface pressure cycle with a Mars general circulation model. J. Geophys. Res 114, E07006. doi:10.1029/2008JE003302. Kass, D.M., Kleinbohl, A., McCleese, D.J., et al., 2016. Interannual similarity in the Martian atmosphere during the dust storm season. Geophys. Res. Lett 43, 6111–6118. Lewis, S.R., Mulholland, D.P., Read, P.L., et al., 2016. The solsticial pause on Mars: 1. A planetary wave reanalysis. Icarus 264, 456–464. doi:10.1016/j.icarus.2015.08. 039. Martin, L.J., Zurek, R.W., 1993. An analysis of the history of dust activity on Mars. J. Geophys. Res. 98 (E2), 3221–3246. Montabone, L., Forget, F., Millour, E., et al., 2015. Eight-year climatology of dust optical depth on Mars. Icarus 252, 65–95. Montabone, L., Marsh, K., Lewis, S.R., et al., 2014. The Mars analysis correction data assimilation (MACDA) dataset V1.0. Geosci. Data J 1 129–139. doi:10.1002/gdj3. 13. Newman, C.E., Lewis, S.R., Read, P.L., et al., 2002. Modeling the Martian dust cycle: 2. Multiannual radiatively active dust transport simulations. J. Geophys. Res 107 (E12), 5124. doi:10.1029/20 02JE0 01920. Newman, C.E., Richardson, M.I., 2015. The impact of surface dust source exhaustion on the martian dust cycle, dust storms and interannual variability, as simulated by the MarsWRF general circulation model. Icarus 257, 47–87. Ogohara, K., Satomura, T., 2008. Northward movement of Martian dust localized in the region of the Hellas basin. Geophys. Res. Lett. 35, L13201. doi:10.1029/ 2008GL034546.
316
K.-C. Chow et al. / Icarus 311 (2018) 306–316
Richardson, M.I., Toigo, A.D., Newman, C.E., 2007. PlanetWRF: a general purpose, local to global numerical model for planetary atmospheric and climate dynamics. J. Geophys. Res. 112. doi:10.1029/20 06JE0 02825. Siili, T., Haberle, R.M., Murphy, J.R., 1997. Sensitivity of Martian southern polar cap edge winds and surface stresses to dust optical thickness and to the large-scale sublimation flow. Planet. Space Sci. 47, 951–970. Siili, T., Haberle, R.M., Murphy, J.R., Savijarvi, H., 1999. Modelling of the combined late winter ice cap edge and slope winds in Mars’ Hellas and Argyre regions. Planet. Space Sci. 47, 951–970. Skamarock, W.C., Klemp, J.B., 2008. A time-split non-hydrostatic atmospheric model for weather research and forecasting applications. J. Comput. Phys. 227 (7), 3465–3485. Strausberg, M.J., Wang, H., Richardson, M.I., et al., 2005. Observations of the initiation and evolution of the 2001 Mars global dust storm. J. Geophys. Res. 110, E02006. doi:10.1029/2004JE002361.
Toigo, A.D., Richardson, M.I., Wilson, R.J., et al., 2002. A first look at dust lifting and dust storms near the south pole of Mars with a mesoscale model. J. Geophys. Res. 107, 5050–5062. doi:10.1029/20 01JE0 01592. Toigo, A.D., Lee, C., Newman, C.E., et al., 2012. The impact of resolution on the dynamics of the Martian global atmosphere: varying resolution studies with the MarsWRF GCM. Icarus 221 (1), 276–288. Wang, H., Richardson, M.I., 2015. The origin, evolution, and trajectory of large dust storms on Mars during Mars years 24–30 (1999–2011). Icarus 251, 112–127. Wilson, R.J., Haberle, R.M., Noble, J., et al., 2008. Simulation of the 2001 planet-encircling dust storm with the NASA/NOAA Mars general circulation model. Third International Workshop on the Mars Atmosphere: Modeling and Observations Abstract no. 9023.