A Record of Water-Ice Cloud at the Phoenix Landing Site Derived from Modelling MET Temperature Data
G. Bischof1, J.E. Moores1, H.M. Sapers1, Centre for Research in Earth and Space Science, York University,
Toronto, Canada (gbischof@yorku.ca)1, B.A. Cooper2, NOIRlab, Gemini North Observatory, Hilo, USA2
heat contamination given off from the lander.
Introduction:
The Phoenix mission played a key role in our
Surface Energy Balance and Cloud Emission. This
current understanding of polar clouds. Previous
work seeks to build a record of clouds at the Phoenix
studies have investigated the opacity (Dickinson et
site by analyzing the radiative contribution of the
al., 2010), morphology (Moores et al., 2010), and
water-ice clouds to the surface energy balance. The
transport (Whiteway et al., 2009) of the water-ice
energy balance used in this work is adapted from
clouds throughout the 151-sol mission in the Martian
Martínez et al. (2014) and is given by
northern polar region (68.2°N).
The lidar and Surface Stereo Imager (SSI)
onboard the Phoenix lander were the two main instruments used for characterization of the clouds.
Where G is the net flux into the surface, S is the solar
However, both instruments leave large temporal gaps
flux minus a portion reflected back to space given by
in the coverage of clouds. The lidar was used only a
the surface albedo, α. LW↓ is the longwave flux
few times per sol, with no overnight coverage at the
emitted from atmospheric gases and dust. LW↑ is the
beginning and end of the mission. Similarly, atmoslongwave flux emitted from the surface. H is the
pheric movies captured with the SSI encompass only
latent heat flux, describing the exchange in energy
0.76% of the length of the entire mission (Moores et
between the surface and near-surface atmosphere. LE
al., 2010).
is the latent heat, representing the phase change of
The Meteorological (MET) station onboard the
water. The last term, R, is the longwave flux emitted
Phoenix lander, however, ran near-continuously
from the clouds, referred to as “cloud emission”
throughout the mission, taking measurements of the
throughout. R is an independent parameter in the
air temperature at 2 m, 1.5 m, and 1 m above the
model, calculated in 2-hour intervals.
surface every 2 seconds. Investigations into waterBecause the Phoenix lander was not equipped
ice clouds in other regions of Mars show that the
with
a ground temperature sensor, the thermal enviabsorption and emission of longwave flux by clouds
ronment
of the site is modelled numerically. This
can influence the surface and near-surface atmoswas
achieved
using the subsurface conduction model
pheric temperature, particularly in the nighttime
of
Schörghofer
(2020), with physical parameters
(Wilson et al., 2007; Cooper et al., 2021).
tuned to the Phoenix site. The model included a 2We can exploit the near-continuous coverage of
layer scheme, where subsurface ice began at 10.4 cm
the MET instrument to model a record of clouds
deep, within the range of depths uncovered by the
throughout the Phoenix mission by isolating cloud
Robotic Arm (Mellon et al., 2009).
emission from the surface energy balance. We comThe ground and 2 m air temperature are related to
pare the cloud emission with water-ice optical depths
the
sensible heat flux by
retrieved from the lidar (Dickson et al. 2010) to
provide a unique examination of the radiative properties of Martian water-ice clouds.
[2]

Methods:
MET Suite on Phoenix. The MET instrument suite
onboard the Phoenix lander carried three fine-wire
thermocouple temperature sensors, located 1 m, 1.5
m, and 2 m from the surface, with a measurement
frequency of 0.5 Hz. The temperature sensors had an
uncertainty of ± 1 K (Taylor et al., 2008). A pressure
sensor took measurements of the atmospheric pressure at 2 m at the same cadence as the temperature
measurement. A tell-tale wind-indicator was included to monitor the direction and speed of the wind.
Reduced Data Records (RDRs) form each sol
were acquired for this work from the Planetary Data
System: Atmospheres Node. We use the data
from the 2 m air temperature sensor to avoid any

where k is the Von Karman constant, cp is the specific heat capacity of CO2, u is the wind-speed, ρa is the
density of the air at 2 m, f(Rb) is a thermal stability
term, z0 is the height at which the air temperature is
measured, and za is the surface roughness.
The air temperature is plotted in time steps of 120
s and compared to the data acquired from the MET
instrument. If the modelled temperature and MET
temperature are equal, R is set to zero over the entire
run of the model. If the temperatures differ, R is
varied on 2-hour intervals to match the modelled
temperature to the MET data. By doing this for each
sol of the mission, the amount of cloud emission in
the energy balance necessary to match the MET

temperature data is found, creating a record of clouds
throughout the mission.
It is important to note that, while the temperature
is modelled over a full diurnal cycle, cloud emission
is only considered between 22:00 to 10:00 local true
solar time (LTST). During the day, the near-surface
layer is super-adiabatic (Smith et al., 2006), meaning
convection is influential when determining the temperature. This can be seen in the turbulent daytime
temperatures in Figure 1. However, at nighttime, the
atmosphere becomes much more stable. Because the
daytime convection is not included in the model,
isolating the cloud emission from the energy balance
is not possible. As such, we limit the scope of this
work to the time of day when radiation is the dominant heat transporting mechanism.
Results:
The modelled temperature, energy balance, and
cloud emission for sols 9-10, 64-65, and 137-138 are
shown in Figure 1. These sols were chosen as they
represent a range in the values of R that were found
throughout the mission. The rightmost panels show
the buildup and decay of the cloud emission through
the night, representing the formation and dissipation
of clouds. On sol 9-10, the modelled temperature fit
the MET temperature well, so no cloud emission was
needed in the surface energy balance, representing a
day with no clouds. On sol 64-65, up to 5 W m-2
emitted from clouds is needed to ensure the modelled temperature fits the MET data. On sol 137-138,
the highest value of R is 10 W m-2.
Sol 9: a)

Sol 64: d)

Sol 137: g)

b)

e)

h)

c)

f)

i)

Figure 1: 2 m air temperature on (a) sols 9-10, (d)
sols 64-65, and (c) sols 137-138. The red line represents the temperature when the model is run with R
= 0 over the entire run. The blue line represents the
temperature with R values from the corresponding
righthand panel (c, f, and i). The dashed gray line is

the MET temperature. The center panels (b, e, and h)
show the remaining terms in Equation 1.
We complete this analysis for each sol to investigate cloud emission as a function of time over the
entire mission, as shown in Figure 2. At the beginning of the mission, cloud emission is sparse, with
low values around 5 W m-2. Toward the middle of

the mission, from sols 55 to 80, the cloud emission
reaches a
Figure 2: A near-continuous record of nighttime
clouds is built by analyzing the cloud emission as a
function of time throughout the entire Phoenix mission.
minimum. There are several sols with no emission
whatsoever, indicating no clouds formed throughout
the night. During this time, on the sols where clouds
are present, the cloud emission is shown to
begin around 02:00 LTST, corresponding
with the coldest point of the night. This minimum in cloud emission coincides with the
time after summer solstice when the
nighttime atmospheric temperature is highest.
From sol 90 until the end of the mission, the
cloud emission reaches its highest values.
Most sols show the presence of clouds, with
values up to and surpassing 20 W m-2.
Throughout the entire mission, the cloud
emission gets larger throughout the night,
with the highest flux in the energy balance
from the clouds occurring between 02:00 and
06:00 LTST.
As clouds emit longwave flux toward the
surface during the nighttime, there is an increase in the surface and near-surface temperature. We investigate this temperature
change by taking the difference between the
temperature model when it is run with and
without R introduced, shown in Figure 3. The clouds
typically induced a small warming effect, between 1
– 4 K. Toward the end of the mission, when the
cloud emission was highest, the temperature was up
to 8 K warmer than it would have been without the
presence of clouds.

Figure 3: The maximum change in the 2 m air temperature from 22:00 to 10:00 LTST due to the longwave flux
emitted from clouds above the Phoenix lander. Errors bars represent an uncertainty of  1K, consistent with
uncertainty
from
the
MET
temperature
sensor.
sion is due to the annular cloud reaching the landing
site. Additionally, Sánchez-Lavega et al. 2018 note
Discussion:
This work suggests clouds were present at the
that years in which there was high coverage of the
Phoenix landing site before they were detected by
annular cloud show a rapid sol-tol-sol evolution of
the SSI or lidar. The lidar was precluded from
the cloud. This agrees with Figure 3, where the temnighttime operation until sol 38, where water ice-fog
perature change varies from each sol.
was observed. Water-ice retrievals from the Thermal
The lidar ran frequently overnight after sol 60,
Emission Spectrometer on Mars Global Surveyor in
showing the ice-water content and water-ice optical
the late northern spring show the presence of clouds
depth of the clouds throughout the latter half the
in northern polar region of Mars (Tamparri et al.
mission (Dickinson et al. 2010). We investigated the
2008). Sporadic water-ice clouds were detected in
properties of the water-ice cloud by analyzing the
the period between 12:30 and 14:30 LTST. The
cloud emission as a function of water-ice optical
presence of afternoon clouds suggests clouds could
depth retrieved from the lidar (Figure 4). We see that
have formed over the Phoenix site during the coolest
optically
part of the night. Additionally, Moores et al. (2010)
note an increase in the optical depth at the beginning
of the mission, where they noted this increase could
be due to dust and not necessarily indicate water-ice
clouds. However, the increase in optical depth in the
beginning of the mission and drop in optical depth
near the middle of the mission shown in Moores et
al. (2010) is similar to the trend in cloud emission
seen in Figure 2.
Each year, around Ls = 120 at 60 N, an annular
cloud forms, seen in images taken from the Mars
Reconnaissance Orbiter and Mars Express (SánchezLavega et al. 2018). During the year the Phoenix
mission occurred, a double cyclone produced cloud
phenomena, beginning at Ls = 118 (Sánchez-Lavega
et al. 2018). The increase in cloud emission around
sol 90 in Figure 2 coincides with the formation of the
Figure 4: Cloud emission as a function of water-ice
annular cloud, indicating the increase in cloud emisoptical depth values measured with the lidar. Con-

tour lines represent the temperature of the cloud.
Horizontal error bars represent the uncertainty in
the lidar measurements of ±10%, cited in Dickinson
et al. (2010), The vertical error bars are given as ± 5
W m-2, consistent with the uncertainty in the MET
instrument.
thicker clouds radiate more longwave flux toward
the
surface, with an increase in optical depth throughout
the night. On some occasions, clouds with different
optical depths have similar values of emission. Terrestrial studies have shown that a cloud composed of
aggregates emits more thermal radiation than that
made of spherical particles (Wendish et al. 2007). A
more thorough understanding of cloud microphysics
is needed to explain the thermal emission from water-ice clouds.
Conclusions
A record of water-ice cloud throughout the Phoenix mission was built by modelling the 2 m air temperature collected by the MET temperature sensor.
The amount of cloud emission was found over the
mission, filling in temporal gaps left by the lidar and
SSI.
A diurnal cycle was seen, where the cloud emission increased throughout the night, typically reaching a peak between 02:00 and 06:00 LTST, and
dissipating into the morning. In addition, as a seasonal cycle was observed, where sporadic clouds
were seen in the beginning of the mission and
reached a minimum by the middle of the mission.
From sol 90 onward, the highest amount of cloud
emission occurred, reaching values upwards of 20 W
m-2. This induced a warming effect at the surface,
raising the 2 m air temperature, usually by 1 – 3 K.
This analysis can be expanded to other regions of
Mars to build cloud records by analyzing the energy
balance. This would build a cloud record at other
locations where temporal gaps may exist. In addition, analyzing the cloud emission as a function of
water-ice optical depth may provide a valuable comparison between equatorial and polar clouds.
Acknowledgements
This work was funded in part by the Natural Sciences and Engineering Research Council of Canada
(NSERC) Technologies for Exo-planetary Science
(TEPS) Collaborative Research and Training Experience (CREATE) program. The authors wish to thank
Jim Whiteway for providing the optical depth data
obtained from the lidar used in the discussion of this
analysis.
References
Cooper, B.A., de la Torre Juárez, M., Mischna, M. et
al. 2021., Journal of Geophysical Research: Planets,
126, e2020JE006737, doi:
https://doi.org/10.1029/2020JE006737

Dickinson, C., Whiteway, J. A., Komguem, L.,
Moores, J. E., & Lemmon, M. T. 2010, Geophysical
Research
Letters,
37,
L18203,
doi:
https://doi.org/10.1029/2010GL044317
art ne ,
, enn o, N., Fischer, E., et al. 2014,
Journal of Geophysical Research: Planets, 119, 1822,
doi: https://doi.org/10.1002/2014JE004618
Mellon, M. T., Arvidson, R. E., Sizemore, H. G., et
al. 2009, Journal of Geophysical Research: Planets,
114,
E00E07,
doi:
https://doi.org/10.1029/2009JE003417
Moores, J. E., Lemmon, M. T., Smith, P. H.,
Komguem, L., & Whiteway, J. A. 2010, Journal of
Geophysical Research: Planets, 115, E00E08, doi:
https://doi.org/10.1029/2009JE003409
Sánchez-Lavega, A., Garro, A., del RíoGaztelurrutia, T., et al. 2018, Journal of Geophysical
Research: Planets, 123, 3020– 3034, doi:
https://doi.org/10.1029/2018JE005740
Schorghofer, N. 2020, Planetary-Code-Collection:
Thermal and Ice Evolution Models for Planetary
Surfaces
v1.1.7,
GitHub.
doi:10.5281/zenodo.594268
https://github.com/nschorgh/
Planetary-CodeCollection/
Smith, M. D., Wolff, M. J., Spanovich, N., et al.
2006, Journal of Geophysical Research: Planets, 111,
E12S13, https://doi.org/doi:10.1029/2006JE002770
Tamppari, L.K., Smith, M.D., Bass, D.S., et al. 2008,
Planetary and Space Science, 56, 227-245, doi:
https://doi.org/10.1016/j.pss.2007.08.011
Taylor, P. A., Catling, D. C., Daly, M., et al. 2008,
Journal of Geophysical Research: Planets, 113,
E00A10, doi: https://doi.org/10.1029/2007JE003015

Whiteway, J. A., Komguem, L., Dickinson, C., et
al.
2009,
Science,
325,
68,
doi:
https://doi.org/10.1126/science.1172344
Wilson, R. J., Neumann, G. A., & Smith, M. D.
2007, Geophysical Research Letters, 34, L02710,
doi: https://doi.org/10.1029/2006GL027976
Wendisch, M., Yang, P., Pilewskie, P. 2007, Journal
of Geophysical Research, 112, D08201, doi:
https://doi.org/10.1029/2006JD007899