Low (web) Quality - BALTEX

183
The North American Regional Climate Change Assessment Program
(NARCCAP): Status and results
Raymond W. Arritt for the NARCCAP Team
Department of Agronomy, Iowa State University, Ames, Iowa 50011 USA (rwarritt@bruce.agron.iastate.edu)
1. Overview
The North American Regional Climate Change Assessment
Program (NARCCAP) is an international program that is
generating projections of climate change for the U.S.,
Canada, and northern Mexico at decision-relevant regional
scales. NARCCAP uses multiple limited-area regional
climate models (RCMs) with a nominal grid spacing of 50
km nested within results from multiple atmosphere-ocean
general circulation models (AOGCMs). The use of multiple
regional and global models allows us to investigate the
uncertainty in model responses to future emissions (here, the
A2 SRES scenario).
Six regional climate models are used in NARCCAP:
• Regional Spectral Model, being run by project
participants at the Scripps Institution of
Oceanography / University of California at San
Diego
• Canadian Regional Climate Model, by participants at
the consortium Ouranos
• RegCM3, by participants at the University of
California at Santa Cruz
• MM5, by participants at Iowa State University
• WRF (Weather Research and Forecasting Model), by
participants at Pacific Northwest National
Laboratories
• HadRM3, by participants at the UK Meteorological
Office Hadley Centre
The four global climate models providing initial and
boundary conditions for the regional models are:
• HadCM3, from the UK Meteorological Office
Hadley Centre
• CGCM3, from the Canadian Centre for Climate
Modelling and Analysis
• CCSM3.0, from the National Center for Atmospheric
Research
• CM2.1 from the NOAA Geophysical Fluid
Dynamics Laboratory
The project also includes global time-slice experiments at
the same discretization used for the regional models (50 km)
by the GFDL atmospheric model (AM2.1) and the NCAR
atmospheric model (CAM3).
Phase I has been completed and the results to be shown
include pronounced variations in skill across the model
domain as well as findings that spectral nudging is
beneficial in some regions but not in others. Phase II is
nearing completion and some preliminary results will be
shown.
3. Users and Data
NARCCAP seeks to accommodate the needs of three
classes of users of the model results:
• Users who intend to perform analyses of the
NARCCAP output, such as process-oriented studies
or climate change analyses for specific regions.
• Users who will apply NARCCAP results to climate
change impacts studies.
• Users who will perform further downscaling of
NARCAP results, whether using dynamical
downscaling (e.g., using the NARCCAP results as
initial and boundary conditions for a fine-resolution
numerical model) or statistical downscaling.
NARCCAP data are formatted into a netCDF-based
standard largely following that used for the AOGCM data
archive developed for the IPCC Fourth Assessment
Report. Data and metadata are checked for adherence to
project standards and then are made available through the
Earth System Grid (http://www.earthsystemgrid.org/).
The uniform data format greatly simplifies the ability of
end-users to compare results from different NARCCAP
simulations.
4. Further Information
For further information please consult the project web site
at http://www.narccap.ucar.edu
2. NARCCAP Phases and Progress
Phase I of the experiment uses the regional models nested
within the NCEP/DOE reanalysis for the period 1979-2004
in order to establish uncertainty attributable to the RCMs
themselves. Phase II of the project then nests the RCMs
within results from the current and future runs of the
AOGCMs to explore the cascade of uncertainty from the
global to the regional models.
In NARCCAP Phase II each RCM-AOGCM pair performs
two simulations: one for the period 1971-2000, and another
for 2041-2070. The difference between these two can then
be used to infer regional climate change. Performance of the
coupled RCM-AOGCM simulation for 1971-2000 can be
compared to the reanalysis-driven results from Phase I in
order to deduce errors and uncertainty arising from the use
of the AOGCM to represent current climate.