Modeling Titans atmosphere with a threedimensional general circulation model, TitanWRF

1
Modeling Titans atmosphere with a
three-dimensional general circulation model,
TitanWRF

• Yuk lunch seminar
• Caltech
• August 22nd 2006

2
Outline of talk

• Introduction (key features of interest on Titan)
• Method used (TitanWRF)
• Results to date (from TitanWRF)
• Other work (related to Titan and TitanWRF)
• Future plans

3
Introduction
What do we know about Titan and its atmosphere?

• Orbital obliquity of 26.7 has seasons
• Slowly rotating (15.95 times slower than Earth)
• Long year (29.5 times Earths) slowly
  changing seasons
• Thick (1.5Bar) N2-CH4 atmosphere (CH4 0.05,
  0-7km)
• Gravity is only 1.35 ms-2
• Weak solar forcing (1 of solar radiation at
  Earth)
• Stratospheric haze layer absorbs solar radiation
• Long timescales overall slow to adjust to
  changes

4
Introduction
What do we know about Titans circulation?

• Voyager observed from Ls 8-16, Cassini from
  292-??
• Cassini will soon overlap Voyager season
  (2010)
• all spacecraft observations late winter /
  early spring
• Mapping observations of temperature
  composition only
• Zonal winds mainly derived (using gradient wind
  balance)
• Few direct observations of zonal winds (Huygens
  Earth-based Doppler cloud tracking from
  Cassini VIMS, etc.)
• Meridional circulation (v,w) largely inferred
  from seasonal changes in composition (e.g. polar
  accummulation)
• Tropospheric circulation largely unknown

5
Introduction
Key features of interest - super-rotating winds
Huygens Doppler wind experiment confirmed strong
prograde (westerly) winds aloft at the equator,
and showed large vertical variations in zonal
wind speed
from Bird et al. 2005
6
Introduction

• Hides theorem no extrema of absolute angular
  momentum except at the boundaries in a steady
  state axisymmetric atmosphere
• Extremum contour of constant angular momentum
  surrounding it
• Steady state no mass or angular momentum flux
  across contour
• But diffusion unbalanced downgradient flux
  across contour
• ? such an extremum is impossible away from
  boundaries
• maximum angular momentum must occur at the
  surface in a region of easterlies, so diffusive
  losses can be balanced by surface gains
• M a cos ? (? a cos ? u), so the maximum
  cannot exceed ? a2 (the max value at any latitude
  for u 0, occuring at ? 0)
• u ? a sin2 ? / cos ? everywhere, and u 0 at
  the equator, in such an atmosphere if angular
  momentum is conserved
• Hence equatorial westerlies require upgradient
  eddy fluxes of angular momentum

7
Introduction
Temperature observations and derived zonal winds

• If temperatures were purely radiative-convective,
  we would expect far larger latitudinal
  temperature gradients
• Weak gradients demonstrate importance of
  dynamics, transporting heat polewards
• NB - CIRS data sheds light on only a portion of
  Titans murky atmosphere

Temperature
S. pole
CIRS data from Flasar et al. Science 2005
Mesosphere
Derived zonal wind
300
Haze - strongly stabilizes thick stratospheric
layers
200
Stratosphere
Main haze
100
CH4 clouds
Troposphere
0
S. pole
8
Introduction
Past modeling work - 3D models

• LMD model - produced right amount of
  super-rotation in stratosphere
• Attributed to Gierasch mechanism, but no full
  wave / eddy-momentum analysis
• No gravitational tides due to Saturn assumes
  zero surface thermal inertia
• Tokano model - very weak super-rotation in
  stratosphere
• Puzzling because uses same radiative transfer as
  LMD model
• Includes gravitational tide effect surface
  temperatures vary more realistically
• Other models coming, e.g. NASA Ames GCM with
  full microphysics

Past modeling work - 2D (latitude-height) models

• LMD model - uses parameterized up-gradient
  momentum transport
• Used for many studies, e.g. to look at coupled
  dynamics-haze-photochemistry
• Other models underway using similar
  parameterizations

9
Method used
Overview of the planetWRF model

• Can be run as a three-dimensional global or
  mesoscale model
• Highly accurate (mass and momentum conserving)
• Fast and flexible (can be run on single or
  muliple processors)
• Consists of dynamical core parameterized
  sub-grid scale physical processes (e.g. radiative
  transfer, boundary layer mixing)

Titan-specific features used for TitanWRF

• Titan constants used (e.g. rotation rate, gas
  constant)
• Titan radiative transfer used (McKay et al.
  1989) - includes
• Collision-induced absorption of IR in
  troposphere / lower stratosphere
• Absorption and scattering of visible by CH4 and
  stratospheric haze
• Emission in IR by stratospheric gases C2H2 and
  C2H6 and haze

10
Method used
Initial experiments

• Model started from rest with top sponge layer
  damping winds to 0
• As (1), but with damping to the zonal mean
• As (1), but without a diurnal cycle (diurnally
  averaged radiation)
• As (2), but without a diurnal cycle (diurnally
  averaged radiation)

Timing issues

• A Titan year takes 2 1/2 days on 16
  processors on CITerra, the GPS
  cluster
• But we expect atmospheric spin-up (gaining
  angular momentum from the surface) to take 30
  Titan years to near equilibrium
• We are exploring methods to jump to the end of
  this process, but have currently run for 15 Titan
  years taking about 5 1/2 weeks

11
Results to date
Zonal mean surface temperatures through year 9
12
Results to date
Streamfunctions at solstice and equinox
Ls 90º (northern summer)
Ls 180º (northern fall)
13
Results to date
Zonal mean temperatures and winds southern
summer solstice
Year 1
Year 14
14
Results to date
Super-rotation index - from TitanWRF
Super-rotation index - from the LMD 3D Titan model
Nudged it on its way
(Hourdin et al. Icarus 1995)
15
Results to date
Super-rotation index - seasonal cycle long-term
increase
Year 3
Net upward transport of angular momentum in
stratosphere and troposphere
Strong stratospheric winter jet
Equinox
Net downward transport of angular momentum from
stratosphere
Solstice
Summer
Winter
Total super-rotation of atmosphere is due to net
transfer of angular momentum from surface into
atmosphere. It stops increasing when downward
transport at solstice balances upward transport
at equinox
16
Results to date
Recall that equatorial westerlies require
upgradient eddy fluxes of angular momentum For a
slowly rotating planet, expect waves due to
barotropic instability A necessary condition is
that d2u/dy2 - ? change sign in the domain
Modeled d2u/dy2 - ? (shaded) and zonal wind, u
(thin contours) for
Ls 90º
Ls 180º
Ls 270º
Thick contour shows d2u/dy2 - ? 0 occurs on
equatorward side of jets
17
Task - identify eddy transports
Results to date

• Lack of published work on identifying the wave
  types in 3D Titan models that are responsible for
  transporting angular momentum equatorwards
• We need to make sure we can establish a physical
  mechanism for this transport of angular momentum
  in TitanWRF, and to rule out e.g. model noise as
  the cause

18
Results to date
19
Other work
Titan balloon trajectories in the lower atmosphere

• The challenge to predict possible passive
  balloon trajectories given the poor current state
  of knowledge about Titans lower atmosphere
• Need a model with tidal forcing - future TitanWRF
  application!
• Or one can cheat by using zonal mean background
  winds with tidal winds superimposed - e.g.

Background zonal winds from Tidal winds
from Tokano Neubauer (Icarus 02) Tokano
Neubauer (GRL 05)
Height (km)
Latitude (degrees E)
Can realistic meridional and zonal winds in the
troposphere enable an unsteered balloon (with
perhaps altitude control) to a) travel to a
range of latitudes, and b) hover over / return
to points of interest?
20
Other work
Titan balloon trajectories in the lower atmosphere

• Current 3D Titan models predict quite different
  tropospheric winds

LMD zonal winds at equinox
Tokano GCM zonal winds at Ls180
Weaker westerlies and more easterlies
Height (km)
Strong westerlies almost everywhere
Latitude (degrees E)

• This impacts the relative speed of propagation
  through the tidal flow

In the other case its more likely to produce
loops, perhaps almost hovering at one
location, or to travel more westwards
With strong zonal winds the balloon travels
mostly eastwards and roughly around latitude
circles
Latitude
x
Start
Longitude
21
Other work
Titan balloon trajectories in the lower atmosphere
z
Strong westerly wind
Strong westerly wind
Balloon moved to lower altitude
Weak easterly wind
Weak easterly wind

• We will use TitanWRF to ask - is the above
  scenario for using altitude control to change
  balloon direction feasible (e.g,. are there
  enough wind reversals with height)?
• We will assess the impact of model uncertainty
  (currently huge due to lack of data for
  comparison!) on predicted trajectories
• We will look at other problems for predicting
  trajectories - e.g., background wind strength and
  local time of day (hence phase with respect to
  the gravitational tide) can have a very large
  impact on the zonal and meridional extent of the
  balloons flight path

22
Other work
Titan mesoscale (limited area) modeling

• Modest (
• Interesting to consider the flows due to such
  obstacles (e.g. orographic methane cloud
  production gravity wave production)
• Simplest to look at mesoscale model with enforced
  background flow
• Enables high resolution experiments to examine
  e.g. new microphysics / cloud scheme at
  relatively low cost
• In future could also use nested domains within
  global model

Some simple experiments to demonstrate this
capability - not meant to be realistic of a
particular region on Titan!
23
(No Transcript)
24
(No Transcript)
25
(No Transcript)
26
Future plans
Continue to spin up global simulations described,
others using a) different radiative parameters
e.g. haze production rate, coagulation b)
different surface properties e.g. TI
Analyze eddy momentum transports, EP flux
divergence - goal is to attribute equatorial
super-rotation to certain waves
Perform simulations with no diurnal cycle to
gauge its importance and effect on results
Compare the simulated spun-up atmospheres with
Voyager and Cassini observations in all cases
Using the spun-up model atmospheres, use basic
cloud schemes to assess when and where methane
clouds would form, and compare with observations
27
Future plans
Extend to include the gravitational tide effect
Allow haze advection - this enables
radiative-dynamical feedbacks, as haze is
radiatively active, and also enables comparison
between observed modeled albedo changes
Cassini ISS
Clouds seen from Cassini
Include a more detailed cloud scheme, with
methane microphysics and with ethane ice
particles initially acting as condensation nuclei
Jiafang Xiaos research
Keck
and Earth