INSIDE AN ACTIVE MARTIAN STORM IN JEZERO CRATER
M.T. Lemmon, Space Science Institute, Boulder, CO, USA (mlemmon@spacescience.org), M.D. Smith, Goddard Space Flight Center, Greenbelt, MD, USA, R. Hueso, A. Munguira, A. Sanchez-Lavega, UPV/EHU, Bilbao, Spain, D. Viudez-Moreiras, A. Vicente-Retortillo, J.A. Rodriguez-Manfredi, Centro de Astrobiologia
(INTA-CSIC), Madrid, Spain, G. Martinez, Lunar and Planetary Institute, Houston, TX, USA, C. Newman, Aeolis Research, Chandler, AZ, USA, R. Sullivan, D. Banfield, Cornell Univ., Ithaca, NY, USA, M. Baker, Smithsonian National Air and Space Museum, Washington, DC, USA, J.F. Bell, Arizona State Univ., Tempe, AZ, USA,
J.N. Maki, M. de la Torre-Juarez, L. Tamppari, JPL-Caltech, Pasadena, CA, USA, V. Apestigue, D. Toledo,
INTA, Madrid, Spain.
Introduction:
Martian dust storms are a key part of the current
climate that remain poorly predictable [1]. They range
in scale from local to planet-encircling dust events
(PEDE). Orbital data have been used to track and classify storms, showing patterns in where and when
storms occur and move [2-3]. Yet there is no predictor
of why one local storm grows to regional and another
dies, or one regional storm triggers or joins a PEDE
and another dissipates.
In situ data on storm processes is a necessary element to develop a complete picture of storm evolution. While landers and rovers have experienced dust
storms or been impacted by the lifted dust, there has
been no prior comprehensive collection of meteorological data within the active lifting area of a storm.
The Viking landers documented the effects of planetary scale events, including winds associated with
storm onset at VL1, without documenting local lifting
or activity [4]. The Spirit and Opportunity rovers,
without meteorological capabilities, experienced local, regional, and planetary events [5]. Local events
passed before additional imaging could be commanded, while the regional events were associated
with a reduction in local dust lifting [5-6]. While the
2008 PEDE involved a dynamic environment at each
site, the solar-powered rovers had reduced operations
[5]. The Curiosity rover experienced optical depths of
9 yet saw only a reduction in local activity [7].
During January 2022, a regional dust storm affected the Perseverance, InSight, and Curiosity sites,
causing the solar-powered InSight to cease operations
for several days. We report on the events at the Perseverance site in Jezero crater: six sols of dynamic and
dusty weather that erased rover tracks, cleaned some
surfaces, dirtied others, and partly damaged the
rover’s wind sensors.
Data and methods:
The Perseverance rover landed on 19 February
2021 in Jezero crater, at 18.4°E longitude, 77.5°N latitude. Among its instruments were several cameras
and a meteorology package. Mastcam-Z is a multispectral, stereo camera pair with zoom lenses [8].
Navcam is a color, stereo camera pair with a wide
field of view (FOV) and the front and rear Hazcams
are each color, stereo camera pairs with fisheye lenses
[9]. MEDA, the Mars Environmental Dynamics Analyzer, is a meteorological package that measures

pressure, temperature, winds, ultraviolet, visible, and
infrared fluxes and radiances, optical depth, and humidity; it includes a zenith-looking fisheye camera,
Skycam, as part of its Radiations and Dust Sensor
(RDS) [10].
Meteorology. MEDA functions independently of
rover wake-sleep cycles at a cadence that is specified
in observation tables. Typical behavior was to measure with all sensors (excluding the camera) at 1 Hz,
for a minimum of 5 minutes of each Mars-hour (1.03
hours) and to run continuously for even hours on even
sols and odd hours on odd sols. Additional data collection was commanded during the storm.
Optical depth. Mastcam-Z measured dust optical
depth via Sun imaging using solar filters (neutral density filters combined with a narrow 880-nm filter for
the right camera and a short-pass filter to get red,
green, and blue optical depths for the left camera). Solar images were calibrated to radiance on sensor [11]
and then solar flux was calculated and used to determine atmospheric extinction (see also [5]).
Skycam’s 126° diameter view of the sky is partly
obscured by a neutral density annulus, allowing direct
solar imaging in mid-morning and mid-afternoon.
Due to the limited range of solar elevations, Skycam
optical depths were calibrated to match Mastcam-Z
near-infrared optical depths using measurements
close in time. (Skycam has been zenith-looking and
uncovered and has accumulated dust on the optics.)
MEDA sensor data were also used to determine
optical depth. TIRS (Thermal Infrared Radiometer)
measures broadband and 15-µm radiance, which have
been used to derive 9- µm dust optical depth [12]. The
RDS-Top-7 photodiode measures broadband solar
downward flux with an approximately Lambertian response. The rover was stationary over sols 289-327,
and the TOP7 flux was calibrated to optical depth using a simple 2-stream model to tune the results to
match the Mastcam-Z 880-nm channel.
Site imaging. Upon arrival at the site, the rover
acquired a site and rover deck panorama with
Navcam, as well as images of the area immediately in
front of and behind the rover with Hazcam. During the
next sols, a site panorama was acquired with Mastcam-Z. All the scene was acquired with the 34-mm

(a)

(b)

(c)

(d)

(e)

Figure 1. Meteorology summary from the start of sol 309 to the end of sol 321. (a) Pressure; (b) Temperature for
surface (magenta), local air (black), and lower 200 m (orange); (c) wind direction, with speed shown using colorbar with a linear scale from 0-20+ m/s, and vertical lines showing the two boom failures; (d) fluxes in broadband
visible (black, saturated near noon outside of storm) and thermal infrared (orange), shown on an arbitrary scale;
(e) visible optical depth from Mastcam-Z (black circles), Skycam (blue diamonds), RDS (dark gray line fragments); and TIRS 9-µm optical depth (orange, scale 2x different).
zoom setting; many areas were imaged at zooms as
high as 110-mm (resulting in 3x better spatial resolution). Hazcam images were acquired episodically to
support contact and proximity science operations.
Once the storm was recognized, additional Hazcams
were acquired to assess the storm’s impact. In addition, the Navcam and the 34-mm Mastcam-Z site panoramas were repeated after the storm. Albedo was
monitored by MEDA’s RDS and TIRS [13].
While at the site, Navcam acquired quasi-periodic
environmental survey images. These included dust
devil surveys (DDS), which used 3 monochrome images at each of 5 aims to survey the complete horizon
for motion; dust devil movies (DDM) which used 2145 monochrome exposures at a single aim to find and
track motion; cloud surveys , which used 5 color images to mosaic nearly the full sky hemisphere; and
cloud movies, which were 8 color images at a single
aim [14,15]. The ‘dust devil’ activities were subframed for the horizon at the full width of the FOV,
while the ‘cloud’ activities used full 96°x72° FOV.
Results:
The team was aware that a storm existed, but not
that it would come to Jezero, when the sol 312-313
plan was made, but sol 313 included an unusually
large amount of MEDA data due to available power
and projected downlink volume. By the time the sol
314 plan was made, the team was aware of unusual
activity. Extra environmental monitoring was included in that and subsequent plans, but the rover also
continued with rock-coring operations.

Progression of the storm. Figure 1 shows key meteorological measurements through the storm. Generally normal behavior was seen pre-storm (sols 309312). The pressure profile showed tides [16] and a
vortex near noon on sol 311 (see also [17]). Pressure
tides increased in magnitude over sols 311-313, due
to the storm’s atmospheric heating even while distant.
The pressure profile returned to normal by sol 321.
The storm moved into Jezero crater late in the
morning of sol 313. About 10 AM, the winds increased in strength, with gusts >15 m/s, and turned
clockwise around the site over several hours. During
11:30-12:00, the winds were ~6 m/s from the north;
imaging showed a mottled sky with rolling dust
clouds, variable shadowing, and extreme dust lifting
activity. The DDS and DDM combined (see Fig. 2)
over 11:35-11:59 to show 14 dust devils, more than
any previous sol; 3 indeterminate dust lifting events;
and 4 dust-carrying gust events, more than previously
identified (cf., [14,18]). In late morning, RDS showed
sky darkening in the north, followed by erratic brightness. The RDS showed 10% brightness variations
over several minute periods from shadowing and local
dust lifting. While dust devils moved S near noon,
high-altitude dust clouds moved north (west of the
rover) and south (east of the rover), suggesting CW
motion or complex dynamics at 10s km scale.
The afternoon of sol 313 had the deepest pressure
minimum, and through the day the optical depth increased and became variable. After noon, the winds
rotated CW until they were roughly easterly. At 13:42
LTST, winds peaked at 22 m/s from 110° as a west-

optical depth: while visible optical depth >2 implied
near-surface extinction >0.2 km-1 for a well-mixed atmosphere), a hill 40 km to the east remained barely
visible even on sol 315, implying opacity <0.1 km -1.
Thus, the boundary layer was dust-depleted by a factor of ~2 compared to the air above during sols 313316. Local dust lifting, while significant, remained
small compared to the storm’s total dust load.
Over sols 316-319, storm activity receded, followed by reductions in dust amount. Visibility was
improved on sols 316-318; the boundary layer was
dust-depleted by a factor of 3-4 by sol 318 as the
crater cleared faster than the air above. Despite that,
daytime optical depths remained elevated on sols 316317, but decreased on sol 318 to become normal on
319. Dust lifting events were seen in 4 of 7
DDS+DDM sequences over sols 314-318. The wind
remained variable, as dust devils were observed moving south at 13:00 on sol 318. At the dawn pressure
maximum on sols 319 and 320, pressure oscillations
with 9- and 15-minute periods were observed, possibly gravity waves caused by the storm to the east.
The daytime maximum surface temperature was
reduced by ~10 K on sols 316-7, while the overnight
low was highest around sol 315 [19]. 1.5-m and lowest-200-m peak air temperature increased several K
on sols 312-5 and were depressed on 317.
Prior to the storm, high-altitude ice contributed to
excess visible optical depth in the mornings; however,
during the storm the 9-µm dust optical depth was generally ½ of the visible optical depth [12]. Residual differences may relate to changes in the vertical distribution of the dust, with relatively enhanced IR opacity
when dust filled the crater.

Figure 2. Dust activity seen in Navcam frames: (a) sol
313, ncam00535, azimuth 305°; (b) same frame after
mean frame removal; (c, d) additional frames after
mean frame removal; (e) sol 318, ncam00536, azimuth
310°; (f) same frame after mean frame removal.
moving, CW-rotating vortex passed to the north,
when the wind sensor on boom 2 stopped reporting
due to damage from blown debris. During the afternoon, visibility across the crater dropped as dust increased in the boundary layer.
Sols 314-315 had a high and variable optical depth
through the day, with elevated but more moderate IR
optical depth at night. On sol 315, the other WS boom
stopped reporting at 14:24 LTST. That time was coincident with a 200-s optical depth spike that may
have been due to local gusting. Visibility remained
below normal, but not as much as implied by column

Storm-induced changes. There were multiple instances in which winds from different directions, primarily east, were strong enough to mobilize sand and
dust. Rover-disturbed areas, no longer in equilibrium
with the environment, were affected the most. Smallscale ripples were mobilized, and local deflation and
deposition occurred on natural surfaces and the rover.
Sol 316 rear Hazcam images showed substantial
modification and erasure of tracks to the NE (Fig. 3),

Figure 3. Changes immediately around the rover as
seen in Rear (a, sol 286; b, sol 316) and Front Hazcam
(c, sol 311; d, sol 315; e, sol 322).

and the optics became dust coated. Front Hazcam images on sol 314 showed removal of material from one
abrasion site and two drill holes, as well as changes in
other rover-modified terrain, by winds from the east.
Additional removal occurred between ~13:00 on sols
314 and 315, and the cuttings were later scoured by
winds from ENE before sol 320.
Site imaging after the storm (Fig. 4) showed a series of changes. (1) The rover itself was affected. The
deck had a substantial increase in its sand load (at
least some of which was last modified by winds from
the west, based on piling against obstacles). Wheels
were scoured clean. Clods were removed from the
deck near the Mastcam-Z calibration target on sol
311-312 (possibly during a pre-storm vortex) and
314-315. Over 315-316 there was a debris dump on
the targets, but material was removed from magnets
by winds from the SE (Fig. 4g). (2) Rover-modified
terrain was substantially reworked. Tracks to the east,
south, and west were impacted. Some tracks disappeared; others were smoothed by motion of sand. (3)
Small ripples all around the rover moved westward,
and sand moved on rocks. Notably, undisturbed features other than small ripples were unmodified. While
the storm was significant for disturbed material and
loose sand, equilibrium materials were unimpressed.
Conclusions:
A significant dust storm impacted the Perseverance site for 6 sols. The leading edge of the storm
brought a dynamic local environment on the first sol,
unlike prior in situ storm encounters [4,5,7]. The next
two sols brought high dust loads within the crater.
Over the last 3 sols, the crater cleared faster than the
column dust optical depth, suggesting the rover was
experiencing outflow from increasingly distant lifting. During the storm’s approach, winds rotated

around the area clockwise, increased in speed, and
drove dust devils and dusty gusts. As the storm’s leading edge went over the site, there were high and variable daytime optical depths, warmer overnight and
cooler daytime temperatures, and variable sky brightness due to clouds of dust. Vortices continued to form,
and in multiple events sand was mobilized in ripples,
sand and/or finer material was transported to the rover
deck and likely damaged the MEDA wind sensors,
and rover modifications to the terrain were erased or
reworked by winds. No winds >22 m/s were measured, but larger winds may have occurred after the
loss of the wind sensor.
References:
1. C.E. Newman and M.I. Richardson, 2015, Icarus
257, 47-87. 2. B.A. Cantor, et al., 2001, J. Geophys.
Res. 106, E10, 23653-87. 3. H. Wang and M.I. Richardson, 2015, Icarus, 251, 112-127. 4. J.A. Ryan and
R.M. Henry, 1979, J. Geophys. Res. 84, B6, 2821-9.
5. M.T. Lemmon, et al., 2015, Icarus 251, 96-111. 6.
R. Greeley, et al., 2010, J. Geophys. Res. 115,
E00F02. 7. S.D. Guzewich, et al., 2019, Geophys.
Res. Lett. 46, 71-79. 8. J.F. Bell III, et al., 2021, Space
Sci. Rev. 217, 24. 9. J.N. Maki, et al., 2020, Space Sci.
Rev. 216, 137. 10. J.A. Rodriguez-Manfredi, et al.,
2021, Space Sci. Rev. 217, 48. 11. A.G. Hayes, et al.,
2021, Space Sci. Rev. 217, 29. 12. M.D. Smith, et al.,
2022, MAMO-7. 13. A. Vicente-Retortillo et al.,
2022, MAMO-7. 14. C.E. Newman et al., 2022,
MAMO-7. 15. P. Patel et al., 2022, MAMO-7. 16. A.
Sanchez-Lavega et al., in preparation. 17. R. Hueso et
al., 2022, MAMO-7. 18. C.E. Newman, et al., 2022,
Sci. Adv., in press. 19. A. Munguira, et al., 2022,
MAMO-7. 20. S. Maurice, et al., 2021, Space Sci.
Rev. 217, 47.

Figure 4. Changes in the rover site are shown in (a) sol 286-310 (top) and 321 (bottom) Mastcam-Z panoramas; (b) Navcam deck images on sols 286 (top) and 321 (bottom); (c) sol 318 Mastcam-Z 110-mm image of
drill cuttings removal; (d) Mastcam-Z 110-mm image of ripple migration from sol 290 (left) to 320 (right);
(e) Navcam images showing tracks, ripples, and sand-blasting of wheels, sol 286 (left) and 321 (right); (f)
Mastcam-Z calibration target images shown debris deposition and wind effects; and (g) a calibration target
detail with Supercam [20] image showing removal of dust from the magnet by winds from SE (right).