March/April 2000 Table of Contents
fun to watch play across a field, but little did we know the role
dust devils play in dust trasport on our world. Now we can see them
regularly on arid Mars, too.
Earth, wind is an agent of erosion, the importance of which is often
overlooked in favor of fluvial processes (running water). Waves
on oceans that break on shores giving rise to longshore transport
are the direct result of the frictional drag of wind on water. In
arid environments in which vegetation is sparse, water often originating
from catastrophic thunder showers is the most potent agent of erosion.
But, due to the aridity, wind is an important contributor to the
erosion and sedimentation processes, especially contributing to
airborne particulate. The global scale of wind transport is only
recently recognized, as dust originating in the Sahara Desert of
northern Africa is now determined to contribute to airborne particulate
in Miami, Florida, and Brazil.
more localized example of dust entrainment and transport on Earth
is that of the "dust devil." These relatively small, tornado-like
phenomena are the result of daytime warming of surfaces when humidity
is low. In arid to semiarid environments, these miniature cyclones
reveal themselves when they entrain dust and debris, pumping material
several tens, or occasionally hundreds, of meters skyward. If originating
over vegetated surfaces, dust devils may not be visible because
dust is not entrained, but they can still be felt, often sustaining
winds in excess of 100 km/hr. Because of their high winds and transport
of dust and debris, dust devils are often considered more of an
irritant than an important geologic process.
notion is starting to change. In arid to semiarid regions of the
United States, in particular the Great Basin, the Desert Southwest,
and western Texas, dust devils are now considered to be a substantive
reason that some communities are found to be noncompliant with U.S.
Environmental Protection Agency mandated limits on PM10 and PM2.5
single dust devil may be responsible for entraining and transporting
several hundred kilograms of sand, dust, and debris. A particular
local region may have several tens, or hundreds, of dust devils
develop in a particular day. One hour's observation southeast of
Ely, Nevada, in summer 1998, recorded ten dust devils. This is an
example to suggest that up to 100 dust devils a day in any given
environment conducive to dust devil formation is a reasonable estimate,
given that dust devils in late spring and summer begin to develop
between 9:00 a.m. and 10:00 a.m. and may continue to form until
6:00 p.m. This number of dust devils, and the amount of material
entrained and transported by any one, illustrates the significant
amount of erosion and transport caused by dust devils.
certain desert surfaces, desert pavement in particular, may be the
direct result of dust devil activity. The formation of desert pavement-relatively
flat surfaces covered by rocks with few fines-is a controversial
subject. Most explanations involve the action of wind, but disagreement
exists on how wind acts to remove only fines resulting in a surface
covered by larger particles. Dust devils act like natural vacuum
cleaners, easily sucking finer dust particles up into their vortices.
These dust devils lack sufficient energy to entrain particles much
larger than sand (4 mm in diameter or so), so they may be more efficient
sorters of sediment size than prevailing winds. Superior sorting
of sediment size by dust devils may be a superior model for the
development of desert pavement.
of 1985, dust devils are known to exist on Mars, providing yet another
example of the geologic similarity of that planet with Earth. As
with the study of Earth, the importance of dust devils to present
Martian geomorphological processes is probably understated. In defense
of this claim, let's discuss the physics of dust devil formation.
receives practically no direct heat (thermal electromagnetic energy)
from the Sun. Most electromagnetic energy originating from the Sun
and reaching Earth is in the form of intense, visible light. Earth
materials (rock, soil, water) absorb visible light. This absorption
brings about molecular excitation in these materials, and this excitation
causes a release of the absorbed, electromagnetic energy. The released
energy is in the thermal (heat) frequency range. Thus, Earth is
warmed indirectly by the Sun, as light is absorbed, converted in
frequency to heat, then re-emitted into the atmosphere effecting
barefoot across pavement, rock, or soil exposed to intense sunlight
is one way to appreciate this process. As such surfaces release
heat, warming air directly above, the warm air rises because air
higher above the surface is cooler. Differences, in this case of
temperature, cause motion-air circulation in this example. Although
heat is not visible to the human eye, thermal energy rising from
a surface does distort light. This is the cause of mirages, especially
evident during summer months.
Dust devil imaged southeast of Ely, NV, in late
June, 1998. Image courtesy of the author.
air rises through cooler surrounding air much like gas bubbles rise
through an opened bottle of soda water. As the warmer air rises,
a localized cell of low pressure is created. Air under higher pressure
surrounding the local cell rushes in, forcing the warmer air farther
upward (upwelling). Coincidentally, the rising air rotates, counterclockwise
if in the northern Hemisphere, forming a vortex. If the rise of
air is rapid, this rotation will likewise be rapid. In this event,
a dust devil is born.
weather conditions for dust devil formation are a relatively clear
day (letting sufficient sunlight fall on ground surfaces) with associated
low humidity. Lower humidity results in cooler nighttime air temperatures.
After dawn, as surfaces are warmed, a higher thermal gradient develops
between warm surface air and cooler air above. The larger this gradient,
the faster the heat from the surface will rise. The summer season
is most ideal, because sunlight will fall on ground surfaces more
A large Terran dust devil formed in Australia.
Note the distinctive vortex shpae and the sheer quantity of dust
entrained by this single dust devil. Image (c)2000 by John Roenfeldt
the ideal geographic environment is one that is arid to semiarid,
wherein ground surfaces are sparsely vegetated. Soil and rock absorb
sunlight, releasing it as heat more efficiently than does vegetation;
in fact, vegetation may act to cool a surface. Valleys in the North
American Great Basin are an example of such an ideal environment.
From late spring through early autumn, nighttime temperature in
many of these valleys drops below 12 degrees Celsius. As day breaks,
the ground surface is quickly warmed by intense sunlight. Dust devils
commonly develop between 9:00 a.m. and 10:00 a.m., local time, after
sufficient warming occurs and thermal plumes of rising air start
course, the ideal climatic and geographic conditions mentioned for
dust devil formation on Earth are closely matched on Mars. There,
atmospheric pressure approximately 1/100 that of Earth, little cloud
cover, very low humidity (especially away from the polar ice caps),
and ground surfaces perfectly barren of vegetation allow sunlight
to efficiently heat ground surfaces.
of the very low atmospheric pressure, the thermal gradient in the
air above these surfaces is large. During one experiment, the Mars
Pathfinder lander recorded a temperature of 16 degrees Celsius,
30 cm above the ground surface, whereas a temperature of -7 degrees
Celsius was recorded 150 cm above the ground surface. Given the
lack of humidity, light from the Sun having an intensity less than
that reaching Earth but with sufficient energy to effect warming,
barren ground surfaces, and very cold air above these surfaces,
dust devils should theoretically be as common, or more so, on Mars
as they are on Earth.
was first imaged by the Mariner spacecraft-Mariner 4
in 1964, Mariners 6 and 7 in 1969, and a detailed
mapping from orbit by Mariner 9, launched in 1971. None of
the Mariner missions, though, provided the resolution necessary
to observe a relatively small dust devil from space. Such resolution
was available with the next Mars mission, Viking, consisting
of two lander/orbiter combinations, both of which were launched
and reached Mars in 1976.
Viking I and II orbiters acquired images of the planet
over two full Martian years, enough time and seasonal variation
to image dust devils if they occur more than rarely on that planet.
Although the Viking spacecraft acquired images in the Earth
period, 1976-1980, it was five years before dust devils were identified
in their images. This marked the first substantiation of dust devil
features on Mars.
Five dust devils imaged by the
Viking orbiter in 1976. Image courtesy of NASA.
represented a two-component system. One component remained in orbit,
imaging Mars over the period, 1976-1980. The other component soft
landed, via parachute, on the surface. The Viking landers
conducted several experiments, most notably soil analysis to detect
evidence for life, results from which were inconclusive. Additionally,
each lander was equipped with a camera that acquired digital images
from the Martian surface. The manner in which these cameras operated,
especially their long exposure times, favored the imaging of stationary
objects on the Martian surface. Dynamic surface processes, such
as dust devils, could not be imaged with sufficient resolution by
the Viking landers.
Pathfinder, July 1998
twenty years subsequent to Viking lander operation, a more
sophisticated lander was successfully deployed. Known as Pathfinder,
this ground-based system was equipped with an ASI/MET weather monitoring
system that recorded temperature (results from which are quoted
earlier), barometric pressure, wind speed and direction. Additionally,
a digital camera developed by Peter Smith and colleagues at the
Lunar and Planetary Laboratory, University of Arizona, provided
stereo imaging. Each "eye" of the stereo camera was associated with
a filter wheel. Twelve geology filters covering 443 nanometers to
1003 nanometers, including three repeated for stereo coverage, were
loaded into those wheels (in addition to eight solar filters and
a close-up diopter); their bandwidths range from 19 nanometers to
41 nanometers. Most "news-worthy" of Pathfinder's systems was a
mobile robot, Sojourner, itself capable of imaging and mineral identification.
to Pathfinder's launch, I, along with Stephen Metzger, then
a doctoral student at the University of Nevada, Reno, proposed to
NASA what we felt would be a necessary image processing technique
to "see" dust devils on Mars against a heavily dust-laden background
sky. Although the Viking landers did not provide sufficient
resolution to clearly image dust devils, their images revealed a
sky so dust laden that its color was red to reddish-white. Given
that Martian dust is also red to reddish-brown, resulting from oxidation
of iron minerals, consequently Martian dust devils were likely to
have a relatively high visible red signature, we felt that image
differencing would be necessary to reveal dust devils against the
Composite (blue, green, and red
filter) image of Big Crater from the Mars Pathfinder camera.
The hill in the distance has been alternatively referred to as
"Far Knob" and "Misty Mountain." Notice the hazy atmosphere. The
haze is entirely attributable to dust. Some researchers have speculated
that if not for the dust, the Martian sky would be almost black
due to the low atmospheric density. Image courtesy of NASA.
of the filters on each eye of the Pathfinder camera were
visible blue and visible red. We hypothesized that Martian dust
devils would have a higher red reflectance, due to the color of
the dust, and a lower blue reflectance. Martian dust is red to bright
red, not blue. Consequently, we believed that a Martian dust devil
should appear as a brighter feature in the red image and a darker
feature in the blue image. On this basis, we designed a process
involving the subtraction of a blue-filter image from a red-filter
image to enhance and reveal dust devils against the background Martian
sky. Experiments applied to digital images of Earth dust devils
suggested that this technique would be successful. Nevertheless,
our proposal to be involved in the Pathfinder mission to look for
dust devils in this manner was assigned relatively low priority
by NASA and we were not funded.
After applying the sky-image subtraction, then
red-minus-blue differencing, two large dust devils are revealed
in the Big Crater image. Image courtesy of the author and NASA.
the Pathfinder mission, the ASI/MET instrument recorded several
"events" that involved a brief, sudden drop in temperature with
associated changes in barometric pressure, wind speed, and wind
direction. These "events" lasted only for a short duration of time,
and one of only a few plausible explanations for these events was
the passage of a dust devil directly over the instrument. This could
not, however, be confirmed by the Pathfinder camera. Although dust
devils were the most likely cause of the ASI/MET instrument fluctuations,
images from the camera displayed as true color images of the Martian
horizon, created by combining blue, green, and red filter images,
did not reveal visible dust devils. This was true, even though one
imaging sequence was designed specifically to search for dust devils
to help confirm the ASI/MET fluctuations.
to the Pathfinder mission, Metzger downloaded several hundred
of its digital images that were acquired at a local Martian time
deemed ideal for dust devil formation (11 a.m. to 4:00 p.m., using
the analogy of Earth). His initial experiments of subtracting blue-filter
images from red-filter images failed to identify dust devils. One
aspect of image processing omitted from our original proposal to
NASA was the correction of camera artifacts.
types of artifact interfered with the ability to resolve dust devils.
One was the CCD of the camera that converted analog sunlight from
the optical system to the digital signal transmitted to Earth. This
CCD was associated with a stationary interference pattern that imparted
a noise to each digital image. The second artifact involved Martian
dust that contaminated the outer surface of the camera lens. Metzger
selected "sky" images, images of mostly Martian sky with minimal
horizon, acquired closely in time to images in which dust devils
were suspected. These "sky" images were first subtracted from images
to be inspected to remove camera artifacts. The thinking was that
CCD interference and lens dust contaminated all images in precisely
the same way. Subtracting "sky" images would then remove these noises.
Having minimal horizon minimizes the amount of horizon artifact
introduced into images by this subtraction method. Once the sky
image subtraction was completed, the red-minus-blue scheme was attempted
again. This time, Metzger's experiments were successful in identifying
several dust devil features in Pathfinder images. His discovery
marked the first time dust devils were sighted from the Martian
surface. A total of five different dust devils were ultimately identified
in Pathfinder images.
of camera data by Smith's team at the University of Arizona showed
that Mars's atmosphere is more dusty by late afternoon compared
to early morning. They cited Metzger's discovery and suggested that
dust devils are the best explanation for this daily fluctuation
in atmospheric dust opacity.
Global Surveyor, 1999
devils must be commonplace on Mars. The Viking orbiters imaged
many dust devils during 1976-1980. Mars Pathfinder, over
a fairly short interval of time, imaged five separate dust devils.
Given that two different missions, twenty years apart and acquiring
images that were tantamount to random glimpses, imaged dust devils
suggests that they are not a rare event on Mars. Instead, they must
be considered to be so common that they play a major role in dust
entrainment and transport.
Mars Global Surveyor imaged
seven dust devils in June 1999. The arrow indicates one of them
in this image, courtesy of NASA.
test of this hypothesis is provided by Mars Global Surveyor,
a program to image Mars from orbit at significantly better resolution
than what was provided by the Viking orbiters. If dust devils
are indeed common on Mars, then this current imaging program should
capture them in at least several images. In June 1999, Mars Global
Surveyor indeed captured dust devils in two images. No fewer
than seven dust devils are visible in one of these images. This
represents the third mission in a row that has imaged dust devils,
strengthening the hypothesis that dust devils are the dominant dust-transport
mechanism on Mars.
the Martian atmosphere presently have sufficient density to enable
the formation and subsequent mobility of sand dunes? Certainly,
Mars has numerous sand dunes. All satellite missions to date have
imaged them. What is not clear is whether dunes are presently forming,
or if existing dunes are mobile.
is more difficult to move by wind than is dust. Its size is usually
1 to 4 mm. On Earth, wind speeds of 15 m/s or greater are required
to move sand. On Mars, with an atmospheric density 1/100 that of
Earth, much greater wind velocities are needed to move sand-sized
particles. Are such wind velocities possible on Mars given its current
(Global Surveyor) and future mapping missions will enable
us to analyze the mobility of existing dunes by comparing their
images to earlier images from Viking. If dunes are considered
to be mobile, either they are composed of particles much smaller
than sand ("dust dunes"), or significant evidence for very high
wind velocities will be realized. If dunes are found to be immobile,
then at least a couple of explanations are possible. One, dunes
formed when Mars's atmosphere was more dense. Two, winds were calm
in the interim between satellite mapping missions. If the atmospheric
density has changed, then dust devils may offer the only current
mechanism for sediment transport on Mars.
R. CARR is a professor in the Department of Geological
Sciences at the University of Nevada in Reno. He is the author of
one book and over fifty journal articles on various aspects of data
analysis, image processing, and geological engineering. He is married
and has two children, five poodles, three cats, one guinea pig,
and various fish. His favorite song? "At the Zoo," by Simon and
Garfunkle. His email address is email@example.com.