Homework Assignment 1 - Due Tuesday, April 8, 2008
Introduction
Explore Earth's energy balance and model errors using output from the
CCSM3 climate model compared to various observational data sets.
The CCSM Atmosphere Model Working Group (AMWG) diagnostic
package was applied to climatologies from a series of model runs. See
http://www.cgd.ucar.edu/cms/mstevens/ccsm3/ccsm3.T85.html
Runs b30.009 (a long 1990s control) and b30.030a, b30.030b etc (an
ensemble of 20th century runs). The "ES01" suffix means the run was done
on the Earth Simulator, which is mostly irrelevant. Also Ignore the others
runs, b30.010, b30.011, b30.023, etc.
These runs were conducted for the IPCC AR4 at T85 resolution in the
atmosphere and land and roughly 1 deg resolution in the ocean and sea
ice. It is a straight climate model, with no biogeochemistry or reactive
atmospheric chemistry, so anthropogenic greenhouse gases, aerosols,
and ozone are all specified (constant in the control and variable in
the 20th century ensemble).
The diagnostic comparison is always conducted for a "Test Case",
which is a CCSM run, and a "Control Case", which can be observations
(OBS) or any another CCSM run. Later you will be running the diagnostics
package on your own runs. This assignment is designed for you to get to
know it.
Explore
Begin with b30.030c compared to b30.009
http://www.cgd.ucar.edu/cms/mstevens/ccsm3/ccsm3/b30.030c-b30.009_1979-1999/sets.htm
Click on Set 1 Tables and then click on the link to bring up the table
with global averaged annual means. The url is
(link corrected after hardcopy print)
http://www.cgd.ucar.edu/cms/mstevens/ccsm3/ccsm3/b30.030c-b30.009_1979-1999/set1/table_GLBL_ANN.asc
Most of the variable names and units can be found by opening another
browser window and pointing it to Set 5 Horizontal Contour Plots,
http://www.cgd.ucar.edu/cms/mstevens/ccsm3/ccsm3/b30.030c-b30.009_1979-1999/set5_6/set5_6.htm
However, the first two variables are more cryptic:
RESTROM = Top of Model Residual Energy Flux
and RESSURF = Surface Residual Energy Flux.
Also FLNT = FLUT = Outgoing longwave radiation at the top of model.
Note that RESTROM = FSNT - FLNT
and RESSURF = FSNS - FLNS - LHFLX - SHFLX (LW corrected to LH after hardcopy)
Scan down the column of b30.030c averages and think about the variables
that interest you. Don't worry about understanding all of them.
Written part
1) Go to the same table of diagnostics but for a comparison between
b30.030c and OBS. (Some of the numbers are slightly different than
before and their variable names have been altered slightly too.)
a) What are the global average annual means for RESTROM and RESURF in
this run and what should they be in nature?
Observations of RESTROM and RESURF in the diagnostics table are
missing because uncertainty in the satellite measurement is too large to
reliably estimate the top of atmosphere energy balance.
Notice that FSNTOA (similar to FSNT) and FLUT are in poor agreement
between the CERES and ERBE satellite instruments.
b) A few of the model's parameters were tuned to make RESTROM near
zero in the 1990s control. Does this make sense? Can this model be used
reliably for greenhouse warming experiments given this method of
tuning. Discuss. (This question is open-ended on purpose. Talk to each
other about it. )
2) How does the model compare to the OBS for the basic top of
atmosphere/surface radiative fluxes on the global annual mean and also
for the annual and DJF zonal means in set 3 (corrected 4/2) and maps in set 5. Write
about a one page summary or list of the major features. Keep in mind
that the satellite error is about +/- 10 W/m2. Do not write about the
longwave and shortwave cloud forcing variables at this time. Try to look
at all three observational sets CIRES, ERBE and ISCCP for at least one
of the variables so you can see that the observations vary. (last sentence added 4/2)
3) Check out the zonal mean temperature and wind in b30.030c compared to
OBS (set 4). Discuss the errors in these fields.