SEASONAL VERTICAL WATER VAPOR DISTRIBUTION AT THE
PHOENIX LANDING SITE
C. W.S. Leung, Jet Propulsion Laboratory/Caltech, Pasadena, USA (cecilia.leung@jpl.nasa.gov), L.
Tamppari, Jet Propulsion Laboratory/Caltech, Pasadena, USA, D. Kass, Jet Propulsion Laboratory/Caltech,
Pasadena, USA, M. Smith, NASA Goddard Space Flight Center, Greenbelt, USA, G. Martínez, Lunar and
Planetary Institute, Universities Space Research Association, Houston, USA, E. Fischer, University of Michigan, Ann Arbor, USA.
Introduction:
The objective of this study is to use a combination of orbital and surface observations to constrain
the seasonal vertical water vapor distribution at the
Phoenix Mars Lander location (68°N, 234°E) in the
Martian north polar region. The water cycle is key to
understanding Mars’ current and past climate, and
the vertical distribution of water reflects the complex
interactions involving temperature variations, winds,
cloud microphysics, convective and turbulent mixing, regolith-atmosphere exchange, and other processes that influence the inventory of water. While
remote sensing observations from orbit have provided a global, multi-year interannual climatology of
water vapor column abundances in the Martian atmosphere [1,2,3,4], no such extensive climatology
exists for the vertical distribution of water vapor.
Furthermore, orbital assets typically have difficulties
observing the bottom-most scale height due to limited path transmission resulting from increased dust
and cloud optical thickness near the surface [5,6].
Thus, the vertical water vapor distribution particularly in the planetary boundary layer remains poorly
constrained due to the limited direct observations.
A key emphasis of this study is the usage of a recently recalibrated surface vapor pressure dataset,
provided by the of the Phoenix (PHX) thermal and
electrical conductivity probe (TECP) [7]. These newly recalibrated measurements provide a new constraint for studying the vertical profiles of water vapor along with using coordinated measurements of
water vapor column abundances from TES and
CRISM, near-surface pressure and temperature from
the Phoenix Meteorology Package (MET), and atmospheric profiles of temperature and water ice
clouds from the Mars Climate Sounder (MCS). We
will test whether modeled water vapor column abundances are consistent with the common hypothesis
that water vapor distribution is well mixed from the
surface to the cloud saturation level. Furthermore,
this study will evaluate whether water may be enhanced or depleted in the near surface layer, and determine the maximum height of a well-mixed layer as
a function of season for the duration of the Phoenix
mission.
Observational Constraints:
For this study, we take a seasonal approach to
understanding the vertical distribution of water vapor

between Ls = 75° to Ls =150°, partitioning this seasonal period into 5-degree solar longitude (°Ls) bins.
Relative Humidity. The aggregate TECP dataset
taken over the full duration of the Phoenix mission is
shown in Figure 1. A diurnal cycle can be seen
where peak water vapor pressure values occur during
the mid-afternoon and drop to a minimum value ~
0200 LTST. While relative humidity measurements
were taken near continuously over the full diurnal
cycle for a few individual sols, relative humidity
measurements were typically observed for only a few
hours each sol resulting in large gaps when looking
at the full diurnal cycle for most individual sols. The
mid-afternoon 1500 LTST TECP vapor pressure
value extracted from each of the resulting diurnal
curves is then used to represent the surface water
vapor constraint in each seasonal bin.

Figure 1. All TECP vapor pressure measurements
collected over the 151 sols of the Phoenix mission
are shown in grey. Overlaying the individual measurements are diurnal curves fitted to the TECP data
in each 5°Ls interval.
Water Vapor Column Abundance. We consider
mid-afternoon measurements from TES and CRISM
which fall within ± 5° latitude and ± 15° longitude of
the Phoenix landing site. In our climatological approach, we consider the average water vapor column
abundances taken across all Mars Years where measurements are available, even though interannual variability is present in the Martian water cycle. we partition the combined TES and CRISM data sets into 5°
Ls bins and find the average value of each bin as the
representative water vapor column abundance at that

season. The seasonal trend in water vapor column
abundances between Ls ~30° to 180° is shown in
Figure 2. The overall historic seasonal trend indicates a gradual rise from ~22.8 pr-μm near the beginning of the Phoenix mission (Ls ~ 76°), to a peak
of ~37.8 pr-μm at Ls =125°, and then follows a relatively sharp drop in column water vapor until the end
of the Phoenix mission.

Figure 2. Seasonal column water vapor abundances
for TES (MY 24-27) & CRISM (MY 28-32) over the
Phoenix landing site. Black circles represent the 5°
Ls binned average values. The seasonal trend of the
binned average shows a gradual increase in water
vapor starting in mid-northern spring, peaking at
around Ls=120°, then a sharper decrease towards
northern autumn equinox.
Temperature Profile. The temperature profile is used
to calculate saturation vapor pressures and determine
saturation conditions at each atmospheric level.
Above the boundary layer, the seasonal average temperature profile is determined from MCS temperature
retrievals. For our seasonal study of water vapor variability at the Phoenix landing site, we generated a
set of MCS profiles for each 5° Ls bin that included
retrievals from both limb and nadir sounding based
on a query matching local time between 1350 to
1650 LTST, 63° to 73°N latitude, and -135° to 115°E longitude. Between 216 to 301 individual
MCS profiles were included for each seasonal bin,
from which a mean MCS temperature profile was
derived (Fig. 3). The average temperature profile in
the lowest 1 to 1.5 scale height, depending on the
season, is estimated using a combination of MCS
surface temperatures, MET near-surface measurements, and from potential temperature profiles determined from column models and large eddy simulations at the Phoenix landing site. We incorporated a
surface layer with the superadiabatic lapse rate, then
a layer with an adiabatic lapse rate of 4.5 Kelvin per
km, up to a height of 5.25 km, based on PHX EDL
modeling work [8].

Figure 3. Temperature profile for the 5° Ls bin representing Ls = 85-90°. The average MCS temperature profile (thick black line) was determined from
263 profiles between MY28 to 35 that satisfy our
latitude-longitude-local time requirements.

Testing the Well-mixed Assumption for Vertical Water Vapor Distribution:
We test the validity of the uniformly well-mixed
assumption of water vapor distribution by constraining the constant water vapor mixing ratio in the nearsurface to that of the observed TECP water vapor
pressure, then calculating the resulting total water
vapor column abundance and comparing this value
with the column vapor abundance as observed from
orbit. A number of retrieval algorithms use the vertically uniform assumption to determine the column
vapor abundances. We consider the water vapor profiles for two cases where water is well-mixed in the
lowest atmospheric levels: one where the nearsurface vapor mixing ratio is constant up to the saturation or cloud condensation level (CCL), and the
other where the near-surface vapor mixing ratio is
constant up to the top of the planetary boundary layer.
Well-mixed to the Cloud Condensation Level. To
determine the water vapor profile for this first scenario, the observed TECP water vapor mixing ratio is
assumed to be constant and well-mixed from the surface up to the cloud condensation level, and then
follows the saturation vapor pressure curve to the top
of the atmosphere. The saturation vapor pressure as a
function of height is determined from the seasonally
averaged temperature profile, as described above.
The altitudes of the cloud condensation level varies
with season, reflecting the observed vertical temperature seasonal variability. We found that the cloud

condensation height near the Phoenix landing site is
~15 km at Ls~90° and decreases to ~8 km near
Ls=125°.
Well-Mixed to the Top of the Boundary Layer. In
the second scenario, vapor is well-mixed only to the
top of the planetary boundary layer, falls off, and
then follows a saturation vapor pressure curve. We
assume the constant water vapor mixing ratio, constrained by the near-surface TECP value, extends
from the ground up to the top of the boundary layer
at 4 km altitude. Above the planetary boundary layer
height, the water vapor vertical profile is represented
by a diagonal transect, the slope of which is determined by the altitude it intersects with the inflection
point on the saturation vapor pressure curve. Above
this point, we assume the vapor follows the saturation vapor pressure curve to the top of the atmosphere.
Excess & Depleted Near Surface Layer. The total
water vapor column abundance is determined by
integrating the water vapor mixing ratio from the
surface up to the top of the atmosphere. Figure 4
shows the resulting seasonal variability in modeled
water vapor column abundances for case 1a (uniformly well-mixed to the cloud condensation level)
and case 1b (uniformly well-mixed to top of the
planetary boundary layer) compared to the water
vapor column abundances observed from TES &
CRISM with uncertainties.
The overall results show that, generally speaking,
the assumption of a uniform water vapor mixing ratio
from the surface up to the cloud condensation level
and then following a saturation vapor pressure curve
to the top of the atmosphere yields column abundances that are too high from the beginning of the
Phoenix mission Ls~76° to Ls ~120°and too low
from Ls ~120° to the end of the Phoenix mission at
Ls~150°, when compared to orbital measurements. A
similar conclusion is found even when the uniformly
well mixed water vapor extends only up to the top of
the boundary layer (case 1b). This suggests that,
prior to Ls =120°, there is an overabundance of water
near the surface, concentrated at a height below that
of the planetary boundary layer. After Ls~120°, this
trend is reversed, suggesting that there is a layer
above the surface layer that TECP measures that has
a higher vapor mixing ratio, or stated another way,
that the surface layer is depleted in vapor.
The timing of this transition is coincident with the
peak water column abundances observed from orbit
(Ls=110°-120°), when water ceases to increase in the
polar atmosphere and begins to decrease again.
Therefore, it may be that our determination of this
seasonal overabundance of vapor followed by a depletion of vapor in the lower layer is caused by net
outgassing of the surface from the beginning of the
Phoenix mission through Ls=120°, which then transi-

tions to a net uptake of vapor into the surface until
the end of the mission. The physical mechanism
causing the net outgassing maybe due to either sublimation of shallow-subsurface ice escaping through
the regolith to the atmosphere, and/or net desorption
(regolith becoming drier over the pass of the sols).
We also do not rule out the possibility of low-level
advection of vapor-laden airmasses (presumably northerly)
followed by advection of dry airmasses, or some combination of these mechanisms.

Figure 4. Seasonal variability in modeled water vapor column abundances for case 1a (uniformly wellmixed to the cloud condensation level), and case 1b
(uniformly well-mixed to top of the planetary boundary layer) versus the vapor column abundances from
TES & CRISM with uncertainties. The error bars on
the TES-CRISM data represent +/- 1-sigma variability.
Maximum Well-Mixed Layer Height:
Using both the TECP water vapor pressure as a constraint at the surface, as well as TES and CRISM
vapor abundances to constrain the total water vapor
in the integrated column, we can determine the maximum height of the near surface well-mixed water
vapor layer. We consider two scenarios to account
for cases when the well-mixed assumption results in
water vapor column abundances greater than those
measured by TES & CRISM (which applies between
Ls =75°-120°, and Ls=130°-135°), versus those cases where the uniformly well-mixed assumption results in water vapor column abundances less than
those measured by TES & CRISM (which applies

between Ls =120°-130°, and Ls=135°-150°).
Our results show that the height of the maximum
mixed layer tends to rise between Ls=75° until it
reaches the highest maximum height of 13.3 km at
Ls=95°-100° (Fig. 5). The maximum mixed layer
height then rapidly falls until it reaches the lowest
value of 3.45 km around Ls=115°-120°. Between
Ls=120° to 150°, the maximum height of the near
surface well-mixed water vapor layer stays fairly
consistent with a value hovering around a mean value
of 6.86 km.

Figure 5. Seasonal variability in the maximum
height of a well-mixed layer constrained by TES and
CRISM total water vapor column abundances as well
as by TECP representing the vapor quantity in the
well-mixed region. The height of the cloud saturation
level is plotted as the dash line in gray.

We compare our results with the values found in
Pankine and Tamppari (2015), whose retrieval algorithm matched the combined daytime and nighttime
TES water vapor spectral signatures for the same
location and time period to determine the vertical
extent of vapor in their ‘wet’ atmospheric layer and
the value of the mixing ratio [9]. Their results found
slight seasonal variability in the the height of the
‘wet’ layer in MY25, but remained mostly between
7–10 km from Ls = 75° and 140°, except during the
period between Ls =90–95° when it increased to 17
km (15-18km Ls= 80°-100°). Considering the uncertainty of 2–3 km in Pankine and Tamppari (2015)’s
retrieved ‘wet’ layer heights, our maximum wellmixed layer of 13.3 km at Ls=95° falls only slightly
below their lower limit, meanwhile, our maximum
mixed-layer height of 7 km at Ls=140° is well within
their stated range. Interestingly, both our results and
Pankine and Tamppari (2015) showed that the highest well-mixed layer heights during the period of the
Phoenix mission could be expected around Ls~95°100°.
Conclusion:
Using a combination of orbital and surface observa-

tions, we constrain the seasonal vertical distribution
of water vapor in the planetary boundary layer and
below the cloud condensation height the Phoenix
Mars Lander location (68°N, 234°E) in the Martian
north polar region. Previous observations have
shown significant discrepancies regarding the vapor
distribution in the boundary layer, drawing conflicting conclusions as to whether water is uniformly
mixed below the cloud condensation height, or
whether water is mostly confined in the near surface
layer. We conclude that the uniformly well-mixed
assumption leads to an over-estimation of the total
water vapor column abundances from Ls=75-120°,
and an under-estimation of the total water vapor column abundances at Ls=120-150°. The overestimation of vapor in the column is particularly evident
during peak surface water vapor pressures
(~Ls=110°-120°). We interpret this trend to mean
there is a lower layer in which water is generally enhanced prior to Ls=120° and depleted after Ls=120°.
This suggests the possibility that this layer is owing
to net vapor release from the surface prior to
Ls=120°, such that daytime convective processes
cannot mix it fast enough, and a concentration of
water vapor at the surface that actively participates in
subsurface exchange in a layer that is mostly decoupled from the daytime convective mixed layer. Further, it suggests the possibility that after Ls=120°,
there is net adsorption of water into the surface, leaving behind a low-level deficit of water. Using both
TECP and column vapor abundance as constraints
for the vertical vapor distribution, we evaluated the
maximum height of the well-mixed vapor layer to
ranges between 3.45–13.3 km, with the lowest values
during the time period of peak water vapor column
abundances (Ls=110-120°). Between Ls=120° to
150°, the maximum height of the well-mixed layer
remains fairly consistent with a mean altitude of 6.86
km. These results are particularly important for
providing insight into the seasonal transport of water
and the role of regolith-atmospheric exchange.
References:
[1] Smith, (2002) JGR: Planets, 107(11), 1–19; [2]
Smith, (2004) Icarus, 167(1), 148–165; [3] Smith et
al., (2009) JGR, 114, 1–13; [4] Pankine et al., (2010)
Icarus, 210, 58-71; [5] Kleinböhl et al., (2009) JGR,
114, E10006; [6] Guzewich et al., (2017)
JGR:Planets, 122, 2779-2792; [7] Fischer et al.,
(2019) JGR:Planets, 124, 11, 2780-2792; [8] Tyler et
al. (2008) JGR, 113, E00A12; [9] Pankine and
Tamppari (2015) Icarus, 252, 107-120.