Workshop Intercomparison Information
Mars Atmosphere Modelling and Observations Workshop: Intercomparison Information
January 13 - 15, 2003
Palacio de Congresos y Exposiciones, Granada, Spain
The aim of this page is to provide up-to-date information on the coordinators, data requirements and progress of the model intercomparisons.
Coordinators and contact details:
General coordination : Claire Newman (newmanc@atm.ox.ac.uk)
Data requirements and instructions for delivery:
Hannu Savijarvi, Dave Crisp and Ari-Matti Harri are trying to organize an intercomparison of Mars model radiation schemes for the lower atmosphere (0-50 km, LTE conditions). This project was already underway before these intercomparison studies were suggested, but they will hopefully be presenting some results at the workshop.
If you want to join in this effort, just feed the input from below into your scheme(s), code preferably lifted directly from the host model, and email the broadband flux output of the 6 cases to them. They will do the rest.
Dave provides a line-by-line reference calculation, but other LBL code results are also most welcome. Please circulate this information freely as somebody may have been forgotten unintentionally. They might enlargen the comparison into other cases and spectral ranges; if you have ideas, suggestions or comments, please let them know. Contact details are: hannu.savijarvi at helsinki.fi, david.crisp at jpl.nasa.gov and ari-matti.harri at fmi.fi.
INPUT
Profiles of temperature, water vapour mass mixing ratio and dust visible optical depth for tau=1 at the surface: (other tau profiles are obtained simply by scaling)
Well-mixed gases; use constant mass mixing ratios of: CO2: 9.53E-01 O3: 2.00E-08 CO: 3.80E-04 O2: 9.50E-04
Ground temperature: 210.0 K, surface broadband albedo: 0.24, surface broadband IR emissivity: 0.96, solar zenith top-of-the-atmosphere flux: 610.270 W/m2
Calculate at each altitude fluxes (thermal up, down, solar up, down) for the following 6 cases: CO2 only; All gases + tau = 0, 0.3, 0.6, 1.0, 5.0, using solar zenith angles: 0, 15, 30, 45, 60, 75, 80, 85 deg (8 angles) (or 0 and 60 deg at least)
OUTPUT
Tables of fluxes (in W/m2) at each altitude, e.g. (sample table).
Email your output tables to Ari-Matti.Harri at fmi.fi, indicating clearly, which case it is. State also whether you prefer staying anonymous (e.g. scheme "A", "B", "C" etc.) in the publications that may follow.
There are two different studies proposed for the upper atmosphere (above about 50 km). A first intercomparison is for models in the UV (below about 400 nm), and a second one for radiative models in the IR (1 to 20 um), this last one is particularly focussed to non-LTE problems.
UV INTERCOMPARISON
The main objective is the comparison of UV+EUV atmospheric heating rates in the Martian mesosphere and thermosphere. Also monochromatic or integrated photolysis rate profiles could be compared, at least for CO2 and O2. And in addition I would also propose to compute the downward UV flux at lest at the lowest boundary of our interest, around 50 km, since this is a pure transmission calculation and useful for low-atmosphere models.
The idea is to run our different models as they are currently used, without changes or modifications for the game. Just feed it with the input data we
supply below, run it and extract (some of) the results mentioned above. Use your default/prefered values for any other data required by your model
and not specified here (like absorption cross sections, etc).
Input atmospheres and fluxes:
- AtmRef_1a.dat: Profiles of temperature, pressure, number density and VMR of CO2, N2, O2, CO, atomic oxygen, and H2 for solar minimum conditions in an regular altitude grid from 0 to 250 km. From 0 to 50 km the values of Temp and Press coincide with the reference profile for the low-atmosphere intercomparison.
- AtmRef_1b.dat: Additional VMR for other minor atmospheric constituents.
- SfMin.dat: Flux at Mars's Top-Of-Atmosphere for solar minimum conditions (fluxes integrated to 1 nm bins)
- AtmRef_2a.dat: Same as AtmRef_1 but for solar maximum conditions.
- AtmRef_2b.dat: " AtmRef_1b " " "
- SfMax.dat: Same as SfMin but for solar maximum conditions
Other input parameters:
- Solar Zenith Angles: 0, 60, 85 degree
- Heating rate efficiency (constant for all altitudes, wavelength and solar cycle conditions): 0.15
Outputs to be compared:
- (1) Profiles of atmospheric heating rates (erg s-1 cm-3) or (K/day) for solar minimum and maximum conditions, in (some of) the following spectral regions:
- EUV Integrated : 0 - 120 nm
- FUV Integrated : 120 - 200 nm (includes Ly alpha)
- UV Integrated : 200 - 400 nm
- UV+FUV+EUV Integ: 0 - 400 nm
If your model allows for very fine spectral resolution, please send your "monochromatic" heating rates between 0 and 400 nm at altitudes 50, 80, 100, 120, 160 and 200 km above the surface. Note that the input solar flux is integrated in 1 nm intervals. This could be our maximum spectral resolution.
- (2) Monochromatic and/or integrated photolysis rate of CO2
- (2b) Same as (2) but for O2
- (2c) Same as (2) but for H2
- (3) Downward fluxes, monochromatic (from 0 to 400 nm) or integrated in some spectral regions (specify the ranges clearly so we can compare) at 100 and 50 km above surface, for solar min and max conditions.
IR Non-LTE INTERCOMPARISON
Same idea as in the UV: run your model as it is normally used; just feed the following input data in, run it and supply (some of) the results specified below. Use your default values for any information not contained in the following inputs files and parameteres.
Input Files:
For daytime calculations, we use 2 input ref. atmospheres, as in the UV comparison, for min and max solar conditions, but only one file for the solar flux since this does not change in the IR during the solar cycle:
- SfIR.dat: Solar flux (at TOA in Mars) in the IR and near-IR ( fluxes integrated to 1 nm bins )
For nighttime calculations we propose the following files:
- AtmRef_3.dat: Profile of temperature, pressure, number density and vmr of CO2, O2, atomic oxygen, N2, CO and H2 in an regular altitude grid from 0 to 250 km.
Other parameters :
- Collisional quenching rate CO2(0,v,0)-O3P : Kvt=3e12 cm3 s-1
- For daytime calculations: SZA = 0, 60, 85 degree
Outputs :
- (1) Profiles of near-IR solar heating rates ([erg s-1 cm-3] or [K day-1]) by the CO2 bands in the 4.3 um, 2.7 um, and 1.0-2.0 um spectral regions, for minimum and maximum solar conditions
- (2) Profiles of photoabsorption rates ( [erg s-1] or [photons s-1] ) of the major CO2 bands in those spectral ranges and solar conditions.
- (3) Profile of IR thermal cooling rate by CO2 in the 15 um region (in [erg s-1 cm-3] or [K day-1] ) for the nighttime atmosphere.
- (4) Vibrational temperatures or level populations of the major CO2 vibrational levels in the IR, for nighttime and daytime conditions
Please send your results during November to Miguel A. Lopez-Valverde at
valverde@iaa.es, indicating clearly which case it is, format, etc.
Basic requirements are:
Use a horizontal model resolution of ~ 5 degrees.
Produce zonally-averaged (on pressure surfaces) temperatures and zonal winds for a low dust run. E.g., visible opacity=0.3.
Provide time-averaged data for the 4 seasons centered on Ls 0, 90, 180, and 270, i.e.: Ls 345-15, 75-105, 165-195, and 255-285.
E.g. at 25 latitudes x 15 pressure levels (TO BE DECIDED ON - check with Alison Bridger to find out which pressure levels to interpolate to, prior to taking zonal averages).
Ascii or binary files preferred, together with a guide to reading the data.
Further comparisons may include a high dust run (tauvis > 1), a clear atmosphere run (tau=0) or an 'MGS-like' run (using a setup which produces the closest match to MGS observations during the first year).
Some results!
WILL BE BASED ON SAME RUNS AS THOSE FOR ZONAL MEAN
Basic requirements are:
Use horizontal model resolution of ~ 5 degrees.
Produce zonal wind, meridional wind, temperature and (please, if possible!) geopotential height output at least 2 times per sol for the seasons centered on Ls=0 and 270 (e.g., Ls 345-15, and 255-285) using tau=0, then the same dust scenarios as used for the zonal mean runs.
Provide data on 11-15 pressure levels (TO BE DECIDED ON - check with Jef Hollingsworth).
Data format negotiable.
WILL BE BASED ON SAME RUNS AS THOSE FOR ZONAL MEAN
Basic requirements are:
Use horizontal model resolution of ~ 5 degrees (actually this is flexible).
Primary requirement: produce surface pressure output 12 times per sol for one year using tau=0, then the same dust scenarios as used for the zonal mean runs.
Secondary requirement: negotiable!: produce temperatures at either one or a selection of pressures (T at 50Pa would be useful, but contact John Wilson for details) output 12 times per sol for one year for the same simulations.
Binary files preferred, together with a guide to reading the data.
Some results!
Two experiments are suggested, both using radiatively INACTIVE dust transport, with a LOW background dust loading (e.g., tau611=0.3 with ~ the Conrath 1975 vertical distribution in the Ames model, or tau700=0.2 using the 'low dust scenario' vertical distribution in the Oxford-LMD model):
-
1. To compare transport in the models:
The experiment
- Starting from a spun-up 'low dustiness' atmosphere at Ls=240 degrees, lift dust into the lowest model level at a constant rate (1e-7 kg/m/s) at all longitudes between latitudes 20 S and 45 S for 10 sols, then run for a further 30 sols. [E.g., in the Oxford GCM, this would involve lifting dust at all 72 longitude grid points for grid latitudes 22.5, 27.5, 32.5, 37.5, 42.5 S).]
- Use 2 micron diameter particles, with gravitational sedimentation and atmospheric mixing turned on.
- Use MOLA topography and preferably a horizontal grid resolution of about 5 degrees. Regarding vertical resolution, in an ideal world we would all have our lowest level close to the surface, e.g. about 5 metres, and the model top above 70km, but everyone's standard setup will do.
Data output: all of the following, or as much as possible, would be appreciated!
- Output surface pressure, dust mixing ratios for all grid points at a minimum of 15 levels, and 611Pa visible dust opacities, 4 times per sol.
- Also output, at a rate of 4 times per sol, the dust mass lifted (1st. 10 sols only; used as a check) and the dust mass deposited for each surface grid point since the last output. If this time integration is too awkward, just note that you will be sending instantaneous values 4 times per sol instead.
- It would also be useful to output temperatures and winds at the same times and locations, though these are not immediately requested.
Data transfer
- I would ideally like surface pressures, dust lifting and deposition data, mixing ratios and opacities in netcdf form or as unformatted binary, together with details of the grid spacing and levels used (e.g. sigma values).
- A basic IDL or grads program to read in and plot the data would also be useful if possible.
- This should come to about 50Mbytes, and the best option would be for me to get it from your ftp site. Please e-mail me to arrange this or to discuss alternatives!
Some results!
-
2. To compare dust cycles produced with parameterized lifting:
The experiment
- Starting from a spun-up 'low dustiness' atmosphere at Ls=180 degrees, run the model for up to one year with parameterized dust lifting of 2 micron diameter particles.
- At each time step, lift dust if the surface is free of CO2 ice, and if the near-surface wind stress, tau, exceeds a threshold value, tau_t, of 0.02N/m^2, i.e., if tau > 0.02N/m^2.
- Lift dust according to this formula.
- Caution: depending on the model, this choice of threshold may of course produce very different results, with almost no lifting in some models and huge amounts in others. If possible, it would be more useful to perform several experiments (simultaneously if possible, e.g. by specifying a number of different tracers) using a range of threshold values. For example, tau_t=0.01,[0.02], 0.03 and 0.04N/m^2.
Data output: again, all of the following, or as much as possible, would be appreciated!
- Output surface pressure, near-surface wind stress and visible dust opacity at 611Pa, 4 times per sol.
- Also output, at a rate of 4 times per sol, the dust mass lifted and the dust mass deposited at each surface grid point since the last output. If this time integration is too difficult, just note that you will be sending instantaneous values 4 times per sol instead.
Data transfer
- I would ideally all of the above in netcdf form or as unformatted binary, together with details of the grid spacing used.
- A basic IDL or grads program to read in and plot the data would also be useful if possible.
- This should come to about 140Mbytes, and the best option would be for me to get it from your ftp site. Again, please e-mail me to arrange this or to discuss alternatives!
Some results!
The intercomparison will involve:
- Total water vapor and ice inventories (northern and southern hemispheres & global) as a function of Ls
- Zonally averaged cloud opacities in the visible and IR as a function of latitude and Ls
- "Snapshots" of the cloud opacity in the visible and IR as a function of Lat and Lon at Ls 90 and 270
The intercomparison will consist of a test experiment to simulate the diurnal cycles of winds, temperatures and pressures at the VL-2 site in summer (where observed cycles do not change much from day to day). Some details are as follows, but complete information (and many results too!) are given here:
Use Ls=145 degrees
Use a fixed, horizontally uniform visible dust opacity of 0.4 (approximately as observed by the VL-2 imager) in the GCM
Use MOLA topography, TES thermal inertia and albedo
The horizontal resolution and domain should be as similar as possible to the following: a mother domain resolution (grid spacing) of 243km (with a roughly hemispheric or larger extent), with the resolution of each nest increasing by a factor of 3, i.e., 81km, 27km, 9km and 3km
The vertical resolutions should also be as similar as possible to the following: for direct comparison to the VL-2 data, a lowest level at 1-2 metres would make sense with the greatest resolution near the surface, e.g. with the model top at 0.03 mbar and 32 levels
It would be highly desirable for everyone to use the same map projection for the simulations - polar stereographic looks like the best choice
The most essential figures for comparison would be the average diurnal cycles of T (at ~ 1.6m), Ps and winds (at ~ 1.6m)
Some results!
The experiments and data required:
For solar medium (F10.7=130), Equinox (Ls = 0) and Solstice (Ls = 270 and 90):
- zonal mean fields (u,v,w,rho,T): [ALT vs LAT slices] [z=100 to 200 km]
- global distribution @ 12 UT, z=125, 200 (u,v,w,rho,T): [LAT vs. SLT at constant ALT]
- longitude vs. altitude @ 12 UT (T) at equator
- daily mean heating rates (EUV and conduction, separate terms): (ALT vs. SLT) that will largely approximate SZA variations:
- (a) Equinox (Ls = 0) : Lat = 0, z=100 to 200 km, 12UT
- (b) Solstice (Ls = 90) : Lat = 25.0, z=100 to 200 km, 12UT
- (c) Solstice (Ls =270) : Lat = -25.0, z=100 to 200 km, 12UT
- species concentrations: O and CO2 since these are the major ones over 100 to 200 km; O-fractional abundance (fo) is important for CO2 cooling and ionospheric chemistry.
- for global distribution @ 12 UT, pressure = 1.2-nanobar for comparison with Figure 13 of Stewart et al. (1992).
- latitude slices of fo removing solar declination effects, leaving scaled Mars solar flux effects:
- (a) Equinox (Ls = 180) : Lat = 0, z=100 to 200 km, 12UT
- (b) Solstice (Ls = 90) : Lat = 25.0, z=100 to 200 km, 12UT
- (c) Solstice (Ls =270) : Lat = -22.5, z=100 to 200 km, 12UT
References:
Stewart, A. I. F., M. J. Alexander, R. R. Meier, L. J. Paxton,
S. W. Bougher, and C. G. Fesen, Atomic oxygen in the Martian thermosphere, J. Geophys. Res., 97, 91-102, (1992).
To main workshop page
To Claire Newman's Home Page