Low (web) Quality - BALTEX

184
MRED: Multi-RCM Ensemble Downscaling of global seasonal forecasts
Raymond W. Arritt for the MRED Team
Department of Agronomy, Iowa State University, Ames, Iowa 50011 USA (rwarritt@bruce.agron.iastate.edu)
1. Motivation
The Multi-Regional Climate Model Ensemble Downscaling
(MRED) project was recently initiated to address the
question, Do regional climate models provide additional
useful information for seasonal forecasts? Regional climate
models (RCMs) have long been used to downscale global
climate simulations. In contrast the ability of RCMs to
downscale seasonal climate forecasts has received little
attention. MRED will systematically test the RCM
downscaling methodology by using a suite of RCMs to
downscale seasonal forecasts produced by the National
Centers for Environmental Prediction (NCEP) Climate
Forecast System (CFS) and the NASA GEOS5 system. The
initial focus will be on wintertime forecasts in order to
evaluate topographic forcing, snowmelt, and the potential
usefulness of higher resolution, especially for near-surface
fields influenced by highly irregular orography.
Each RCM will cover the conterminous United States at
approximately 32 km node spacing, comparable to the grid
spacing of the North American Regional Reanalysis which
will be used to evaluate the models (Figure 1). The forecast
ensemble for each RCM will be comprised of 15 members
over a period of 22+ years (from 1982 to 2003+) for the
forecast period 1 December – 30 April. The RCMs will be
continually updated at their lateral boundaries using 6-
hourly output from CFS or GEOS5. Each RCM will provide
hydrometeorological output in a standard netCDF-based
format for a common analysis grid. MRED will compare
individual RCM and global forecasts as well as ensemble
precipitation and temperature forecasts which are currently
being used to drive macroscale land surface models (LSMs).
The project also will evaluate wind, humidity, radiation, and
turbulent heat fluxes, which are important for more
advanced coupled macro-scale hydrologic models. Metrics
of ensemble spread will also be evaluated. Extensive
analysis will be performed to link improvements in
downscaled forecast skill to regional forcings and physical
mechanisms. Our overarching goal is to determine what
additional skill can be provided by a community ensemble
of high resolution regional models, which we believe will
eventually define a strategy for more skillful and useful
regional seasonal climate forecasts.
2. Participating Models
The regional climate models participating in MRED are:
• MM5, being executed at Iowa State University.
• WRF-NMM, at Iowa State University and NOAA
Global Systems Division. This is the version of WRF
used for U.S. operational short-range weather
forecasting by NCEP.
• WRF-ARW, the Advanced Research version of
WRF, at Pacific Northwest National Laboratories.
• CWRF, a version of the WRF model that has been
specially adapted for climate simulation, at the
Illinois State Water Survey.
• Eta/SSiB, at the University of California Los
Angeles.
• RAMS (Regional Atmospheric Modeling System) at
Colorado State University and the University of
Colorado.
• RSM (Regional Spectral Model), at the Scripps
Institution of Oceanography / University of
California at Santa Barbara.
The initial global seasonal forecast model to be
downscaled is the T62L64 version of the NOAA CFS. A
seasonal forecast model based on the NASA Goddard
Space Flight Center GEOS5 GCM, currently in
development, also will be downscaled when results
become available.
3. Further Information
For further information consult the project web site at
http://rcmlab.agron.iastate.edu/mred/
Figure 1. MRED analysis domain and vegetation