Some Questions to Ponder on Nested Climate Modeling
These questions are NOT in the form of a test based on
the last class that addressed this issue. Rather, they are a subset of
continuing things to ponder based on the material presented by Dr Jenkins,
the papers read in preparation and other works. Please think about this more
as it is one of the likely growth regions in climate modeling....
(i) What are the situations most/least suited to nested
climate modeling?
(ii) One of the fundamental assumptions (at least in Giorgi
and Mearns 1991) is that physical processes on the ``large scales'' and the
``mesoscales'' are essentially separable. Is this a valid assumption?
(iii) How important is parameterization type
matching across the limited area model (LAM) and the GCM? For example,
is use of BATS in the LAM warranted if only bucket hydrology is used in
the GCM? Why/why not?
(iv) In some situations (such as convective clouds
potentially), the assumptions underlying parameterizations that are valid
for the typical GCM gridscale are no longer acceptable at finer resolution.
How can this be accounted for?
(v) In one of the modeling papers the radiation scheme in
the LAM was diurnal, but the GCM was not. What sort of effects would you
expect from this?
(vi) A successful GCM simulation over the region of
interest (ROI) would seem to be a minimal precondition to nested modeling.
Without access to high resolution diagnostics, what are the appropriate
measures for ``success'' here?
(vii) After the first 24-48 hours of simulation, is there
any time at which the GCM boundary conditions cannot be said to affect the
entire LAM domain?
(viii) Could artificial neural networks (ANN) or some
other nonlinear relational technique provide more potential for useful
information on the regional scale than direct model nesting? Should direct
process studies, perhaps coupled with synoptic climatological sorting be
pursued more vigorously?
Last Updated: April 15, 1997