Transcript LES Intercomparison Case of Precipitating Shallow Cumulus

LES Intercomparison Case of
Precipitating Shallow Cumulus
Margreet van Zanten, Bjorn Stevens, Pier Siebesma, Louise Nuijens
A.S. Ackerman, H. Jiang, D. Mechem, D.C. Lewellen, S. Wang,
B. Shipway, F. Couvreux, M. Khairoutdinov, A. Noda,
P. Bogenschutz, J. Slawinska, A. Cheng, F. Burnet
A short reminder
The RICO LES Intercomparison Case focuses on precipitating shallow
cumulus convection and is based on observations taken during a threeweek period in a relatively undisturbed trade-wind regime in the Caribbean.
Fourteen participants provided output for 24-hour simulations with and
without microphysics:
explicit bin microphysics (3), one-moment bulk schemes (6), two-moment
bulk schemes (5)
Models agree quite well!
When looking into detail:
The character and amount of precipitation differs substantially among the
simulations
Status
First version of the paper has been written and distributed among
participants (Margreet van Zanten)
Work is in progress to put the paper into its final shape, that is:

to highlight the main results

shaping up the figures (combining model and observational data)
In final form: 2 - 4 weeks?
The focus of this talk is to inform you about the main points we came up
with after asking ourselves:
what is the main message that we wish to bring across with this paper?
Our main points

Why did we have an intercomparison?

What did we learn from the intercomparison which we could not
have learned from running our models individually?
1.
How sensitive is precipitation formation to the microphysical versus
numerical and dynamical representation of the flow?
2.
Despite precipitation differences among models, does consistent
behavior emerge when simulations start to precipitate?
3.
Can observations from RICO constrain aspects of the simulations that
are relatively robust?
4.
What makes RICO an interesting case?
(1) Sensitivity to microphysics
If dynamics would constrain precipitation, a positive relationship between incloud liquid water (for simulations without microphysics) and precipitation (for
simulations with microphysics) may be expected
(2) Consistent behavior once it rains?
Yes
No
(3) Comparison with RICO data
In the data as well as in LES rain seems to scale with echo fraction, which
suggests that to a first approximation once a cloud begins to rain the microphysical details do not matter
(3) Comparison with RICO data
'wet'
'dry'
What threshold defines cloud?
(3) Comparison with RICO data
When looking at intermittency of
precipitation, one needs to keep in
mind that the precipitation fraction
or 'chances of finding rain
somewhere in a domain', are very
scale-dependent
(4) RICO versus BOMEX
Mass flux decreasing less sharply with height : importance entrainment and
properties of air entrained (Pier)
Summary
What did we learn from the LES Intercomparison of precipitating shallow
cumulus?
1.
Precipitation formation appears very sensitive to microphysics
2.
Notwithstanding precipitation differences among models, some
consistent behavior does emerge once simulations begin to
precipitate (different lapse rates, shallower clouds)
3.
Given that past intercomparisons were cases without in-situ data to
compare to, RICO data can help us constrain some robust aspects of
the simulations
'To a first approximation rain is quite simple. Clouds rain or they don’t, and
when they rain the character of the rain is universal. More rain then means
more clouds are raining, or clouds raining for longer, but not that clouds are
raining harder.'
by: Bjorn Stevens