Low-Latitude Cloud Feedbacks Climate Process Team (CPT)

Overall Goal:  Increase our understanding of tropical and subtropical cloud feedbacks on climate sensitivity, and reduce the large uncertainty in GCM simulations of these feedbacks


Cloud feedbacks on natural climate variations and anthropogenic climate change continue to be a leading source of uncertainty in climate predictions. In low latitudes, most clouds are generated by physical processes such as cumulus convection or boundary-layer eddies which involve turbulent circulations of very small length scales (meters to a few km) not resolvable by the grid of a typical numerical model used for global climate prediction, which typically has cells exceeding 100 km in each horizontal dimension and 20-60 vertical levels. These processes must be represented in the model as parameterizations which represent the effects of the small scale motions on heating, moistening, condensed water, etc. based only on the large-scale variables prognosed by the model, such as temperature, moisture, and winds.  Formulation of parameterizations for 'moist' processes involving clouds has been a long-standing challenge since the development of global climate models in the 1960s. However, modern satellite and in-situ observations, coupled with cloud-resolving models that can explicitly simulate the turbulent circulations that generate most low-latitude clouds over a domain of limited size, and computers that allow us to globally simulate with finer and finer numerical mesh sizes, provide hope for making more rapid progress on moist-process parameterization and on reducing uncertainties in cloud feedbacks on climate projections.

Specific Tasks

  1. Diagnose in detail the reasons for different trends of NCAR vs. GFDL low-latitude cloud distribution with 2xCO2, in which we see low-lying clouds tending to increase in the NCAR (CCSM) climate model but decrease in the GFDL model.
  2. Constrain relevant cloud feedbacks using current and historical data, esp. the low cloud feedbacks relevant to (1).
  3. Improve relevant GCM parameterizations using best available physics, focussing especially on cloud microphysics, shallow and deep cumulus convection, and cloud-topped PBLs.

For a fuller description of the CPT concept, the three current CPTs and their goals, see:

Peer-reviewed publications


Lead PI: Christopher S. Bretherton, University of Washington

Participating Modeling Centers (and public CPT web sites):

Center Lead PI CPT Liaison Scientist Other participants
GFDL Isaac Held Ming Zhao Leo Donner, Brian Soden
NCAR Jeff Kiehl Cecile Hannay Bill Collins, Phil Rasch, Jim Hack
GMAO (NASA) Julio Bacmeister   Max Suarez

Core PIs (and public CPT web sites) outside modeling centers and their CPT group members.

CPT PI Institution Group members Task area
Chris Bretherton U. Washington Matt Wyant PBL, cumulus parameterization; GCM cloud diagnostics
Marat Khairoutdinov Colo. St. U.   'Superparameterization' and CRM simulations
Cara-lyn Lappen Colo. St. U.   Unified turbulence/cumulus param.; SCAM in GCSS
Brian Mapes U. Miami   Deep convective cloud diagnosis/parameterization.
Joel Norris Scripps   Use of historical cloud observations
Robert Pincus NOAA/CDC Crispian Batstone Subgrid microphysical/radiative modeling
Bjorn Stevens UCLA Brian Meideiros Idealized planets and bulk modeling of cloud-topped PBL
Kuan-Man Xu NASA Takmeng Wong New CERES satellite products
Minghua Zhang Stony Brook U.   GCM cloud/convection diagnosis and validation

Advisory panel

CPT workshops:
Introductory all-hands meeting. 20-21 November 2003, NCAR
7 July 2004 mini-meeting, Santa Fe CCSM:
All-hands Meeting II. 21-22 October 2004, Seattle
  • Agenda (including CPT password-protected links to presentations).
All-hands Meeting III. 29-30 November 2005, GFDL
  • Agenda (including CPT password-protected links to presentations).
Selected background readings on cloud feedbacks
Column lat/lons for special study.
Internal CPT web page (password required)

Chris Bretherton <breth@atmos.washington.edu>