Earth

1
Clouds and climate the range of scales.
Mixing in laboratory cloud chamber
Small cumulus clouds
Earth in visible light
1,000 km
10 cm
2
Super-parameterization Climate context for
cloud-scale processes

• Wojciech Grabowski
• Mesoscale and Microscale Meteorology Division
• National Center for Atmospheric Research
• Boulder, Colorado, USA

3
(No Transcript)
4
cloud-scale processes

• cloud-scale scale at which coupling between
  cloud dynamics and cloud microphysics takes place
• from 100s of meters (e.g., moist convection) to
  tens of kilometers (e.g., mesoscale convective
  systems, cirrus)
• processes rising motion cools the air cloud
  droplets form when water saturation is reached
  latent heating affects buoyancy droplets become
  drops and they fall out drops evaporate below
  the cloud base droplets freeze to form ice
  crystals ice crystals grow. etc

5
Representation of clouds in contemporary climate
models is one of the most uncertain aspects of
modeling of climate and climate change. Such
models have horizontal grid spacing of the order
of 100 km and cloud processes are typically
subgrid-scale. However, the impact of clouds on
radiative transfer and energy transformations
(e.g., latent heating) is critical for the
climate system. This aspect is especially
significant in the tropics, where clouds are
associated with moist convection
6
Cloud-resolving models (CRMs)

• Models appropriate for small-scale atmospheric
  dynamics (i.e., nonhydrostatic either
  compressible or anelastic)
• Include representation of moist precipitating
  thermodynamics (cloud processes)
• Usually include representation of exchange
  between atmosphere and surface (surface fluxes)
• Often include radiative transfer
• Grid spacing horizontal around 1 km, vertical a
  few hundred meters (often stretched)
• Can be 2D (x-z) or 3D (x-y-z)
• With higher spatial resolution Large Eddy
  Simulation (LES) models

7
Coupling of physical processes in a
cloud-resolving model
8
Cloud-resolving modeling of GATE cloud
systems(Grabowski et al. JAS 1996)
2 Sept, 1800 Z
400 x 400 km horizontal domain, doubly-periodic,
2 km horizontal grid length Driven by observed
large-scale conditions
4 Sept, 1800 Z
7 Sept, 1800 Z
9

• Grabowski et al. JAS 1998
• low resolution two-dimensional simulations can
  be used as realizations of tropical cloud systems
  in the climate problem and for improving and/or
  testing cloud parameterizations for large-scale
  models
• - Can we use 2D CRM in all columns of a climate
  model to represent deep convection?
• - Can we move other parameterizations (radiative
  transfer, land surface model, etc) into 2D CRM?

10
Cloud-Resolving Convection Parameterization
(CRCP) (super-parameterization)Grabowski and
Smolarkiewicz, Physica D 1999Grabowski, JAS 2001

• The idea is to represent subgrid scales of
  the 3D large-scale model (horizontal resolution
  of 100s km) by embedding periodic-domain 2D CRM
  (horizontal resolution around 1 km) in each
  column of the large-scale model
• Another (better?) way to think about CRCP
  CRCP involves hundreds or thousands of 2D CRMs
  interacting in a manner consistent with the
  large-scale dynamics

11
Original CRCP proposal
12

• CRCP is a parameterization because scale
  separation between large-scale dynamics and
  cloud-scale processes is assumed cloud models
  have periodic horizontal domains and they
  communicate only through large scales
• CRCP is embarrassingly parallel a climate
  model with CRCP can run efficiently in 1000s of
  processors
• Most (if not all) of physical (and
  chemical,biological, etc.) processes that are
  parameterized in the climate model can be
  included into CRCP framework (super-parameterizat
  ion)

13
(No Transcript)
14
(No Transcript)
15
A day, a year, a millennium paradigm

• With the same amount of computer time, one can
  perform
• about a day-long simulation using cloud-resolving
  AGCM
• about a year-long climate simulation using AGCM
  with super-parameterization
• about a millennium-long climate simulation using
  a traditional AGCM

16
Examples of application

• Madden-Julian Oscillation (MJO) as a holy
  grail of atmospheric tropical dynamics (Raymond
  2001) this will be covered in my next talk
  (Tuesday)
• Application to a real climate model Marat
  Khairoutdinov and Dave Randall (CSU)

17
Results from a traditional climate model versus
SP climate model
Khairoutdinov et al. JAS 2005
Traditional SP Observations
18
Results from a traditional climate model versus
SP climate model
Khairoutdinov et al. JAS 2005
Traditional SP Observations
19
Results from a traditional climate model versus
SP climate model
Khairoutdinov et al. JAS 2005
Meridional distributions of various
climatologically important quantities
20
Results from a traditional climate model versus
SP climate model
Khairoutdinov et al. JAS 2005
21
Super-parameterization domains are periodic,
which implies ltwgt0
22
Super-parameterization domains are periodic,
which implies ltwgt0
23
New Super-Parameterization (NSP)
Super-parameterization domains cannot be periodic
for full fields (this implies ltwgt0), but can be
periodic for small-scale perturbations. As a
results, they can include large-scale gradients
of thermodynamic variables, topography, surface
characteristics, etc.
24
(No Transcript)
25
The original CRCP (or SP) proposal cannot include
large-scale gradients (of SST, topography,
land-surface processes, etc.)
26
SP system with cloud-scale models that include
large-scale gradients and nonzero mean vertical
velocity within cloud-scale models
27
Cloud-resolving modeling of tropical
circulations driven by large-scale SST
gradients Grabowski et al. (JAS 2000)
dx 1 km dz 0.5 km
prescribed radiative cooling (1.5 K/day across
troposphere)
28
Surface precipitation rate in a cloud-resolving
simulation (benchmark or reference simulation)
29
Quasi-two-day oscillations result from the
interaction between convection and propagating
gravity waves period comes from gravity wave
speed and horizontal extent of the computational
domain.
30
Grabowski et al. JAS 2000
31
The same problem with the original
super-parameterization
32
and with the new super-parameterization approach.
33
original SP
new SP
34
ORIGINAL SP (SP2004)
8 large-scale model columns, each 500-km
horizontal extent
80 large-scale model columns, each 50-km
horizontal extent
35
8 large-scale model columns, each 500-km
horizontal extent
80 large-scale model columns, each 50-km
horizontal extent
colder!
36
8 x 500 km 20 x 200 km 80 x 50 km
200 x 20 km
original SP
new SP
fully-resolved
37
8 large-scale model columns, each 500-km
horizontal extent
80 large-scale model columns, each 50-km
horizontal extent
Comparison between the original SP and the new
approach suggests that the improvement of the
coupling between large-scale and cloud-scale
models (i.e., coupling of the vertical velocity)
improves interactions between convection and
convectively-generated gravity waves.
38
Conclusions

• CRCP (super-parameterization, SP) is a
  stepping-stone in a quest for a global
  cloud-system resolving model (a day, a year, a
  millennium paradigm). SP allows elements of
  cloud-scale and mesoscale dynamics (cloud
  processes) to be included in contemporary
  climate models. Other processes (physical,
  chemical, etc) can be included in the SP model
  (e.g., land-surface SP as a physics coupler).
• Work is underway to further extend the original
  SP concept. The new development is particularly
  relevant to the SP application over land
  (topography, land-surface processes) and in
  midlatitudes (baroclinic processes).

39
Results from a traditional climate model versus
SP climate model
Khairoutdinov et al. JAS 2005
Traditional SP Observations
40
Grabowski et al. JAS 2000