Get documents about "The NOAA Climate Test Bed
Document Sample
The NOAA Climate Test Bed
R. W. Higgins1, H. Pan2, M. Ji3 and Stephen Lord2
1
Climate Prediction Center, NOAA/NWS/NCEP, Washington, D.C., USA
2
Environmental Modeling Center, NOAA/NWS/NCEP, Washington, D.C., USA
3
NOAA Climate Office, NOAA/OAR, Washington, D.C., USA
1.0 Introduction extending beyond 10-14 days also show
increased skill. This advancement provides
The operational Seasonal-to-Interannual a major motivation to extend current “Week
(S/I) forecast is a challenging scientific 2” forecasts into monthly, and longer, time
problem that has primarily been the domains. Considerable scientific
province of sophisticated statistical forecast investigation and development will be
algorithms, with a lesser role given to fully needed to accelerate improvements in these
dynamical forecast systems. The relative forecasts and provide useful products and
lack of skill in dynamical forecast systems applications on the 3 week-to-S/I time
has dissuaded efforts at extending scales.
operational forecasts past the traditional
predictability limit of two weeks. Despite
many important applications, not the least of
which is the prediction of tropical
intraseasonal variability (e.g. the Madden-
Julian Oscillation), and despite considerable
research aimed at improving prediction on 2
week to 12 month time scales, progress has
been slow.
Recent developments in coupled
atmosphere-ocean-land surface forecast
systems for S/I climate prediction, however,
have reinvigorated attention on this critical
NOAA product area. For example, in the
past year NOAA has implemented a coupled Figure 1. Forecast anomaly correlation for
Climate Forecast System (CFS), composed NINO3.4 SST for all months (25 seasons)
of the NCEP Global atmospheric Forecast over the period DJF 1997-1998 to DJF
System (GFS) model and the GFDL 2003-2004. CFS: new NCEP coupled
Modular Ocean Model (MOM3), that is a Climate Forecast System; CMP14: old
significant step forward in forecasting NCEP coupled system; CCA: Canonical
ENSO related SST variability in the Correlation Analysis (statistical); CA:
Tropical Pacific using dynamical forecast Constructed Analog (statistical); CONS:
systems and, indeed, shows skill at least official CPC consolidation forecast;
equal to that of current statistical algorithms MARKOV: Markov (statistical) technique
(Fig. 1). Moreover, recent tests with fully (Courtesy Huug VandenDool, CPC).
coupled climate forecast systems have Building on recent momentum, it
demonstrated that atmospheric forecasts behooves NOAA to encourage and facilitate
1
faster improvement of NOAA operational The mission of the NOAA Climate Test
climate forecast systems, and at the same Bed (CTB) is to accelerate the transition of
time increase the exposure of NOAA research and development into improved
climate forecast products in an attempt to NOAA operational climate forecasts,
link them to application models, and products, and applications.
thereby, extend the utility and scope of
NOAA climate services. 3.0 Operations Concept
It is now clear that, having achieved at
least parity with statistical algorithms, The CTB mission is intended to provide
coupled numerical forecast systems have the more traction and visibility to intraseasonal-
greatest potential for improving climate to-decadal research and development efforts
forecasts. Moreover, they now appear by accelerating the transition of this research
mature enough to be used as realistic into NOAA climate forecast operations.
vehicles for basic research on predictability The CTB will provide an operational testing
of the “Earth” (atmosphere-ocean-land environment to support projects that result
surface-cryosphere) system. Supporting the in a direct influence on NOAA climate
NOAA climate forecast systems for use by forecast operations and provide
climate researchers is an optimal way to infrastructure and resources for (long term,
leverage the expertise of the external competitive, community-wide) projects on
scientific community as well as that of broader research issues affecting NOAA
NOAA researchers in order to accelerate the climate forecast operations.
necessary improvements.
The “Predictions and Projections” In order to accomplish its mission, the CTB
component of the FY05 NOAA Climate has the following objectives:
Program Plan has identified support for a
systematic research to operations transition • to assess scientific advances
program as a key deficiency. The plan (models, tools, data sets, observing
emphasizes that there is a lack of support to systems) that have potential for a
transition research and operational climate direct influence on NOAA
models to a community based infrastructure operations;
to accelerate model advances from the • to accelerate the synthesis and
external and internal research communities implementation of these advances
to operations. In order to achieve this for NOAA operational climate
synergistic blend of internal and external forecasts;
research and development efforts required to • to develop new tools and
accelerate improvements in NOAA applications in a quasi-operational
operational climate forecast systems, a environment subject to metrics that
Climate Test Bed (CTB) facility has been mandate good scientific
organized. performance while meeting ease-of
use criteria and time constraints;
• to utilize advanced statistical and
numerical model output, and
stimulate model improvements in
2.0 Mission climate analysis and forecasting
applications;
• to facilitate the transfer of tested
2
software into the NCEP Announcements of Opportunity. Base
development cluster while funded activities will generally be near-term
incorporating adjustments necessary activities, funded by NOAA organizations,
to generate climate forecast that are likely to influence NOAA climate
products that are forecaster-friendly forecast operations within 1 year. These
and time-efficient; projects have definite outcomes and
• to prepare documentation, training lifecycles. They will be selected by CTB
materials, and evaluations of management in consultation with the
performance characteristics of Oversight Board (see below).
successful products to facilitate Announcements of Opportunity (AO’s) will
their use by NOAA climate forecast support long-term, competitive, reviewed
operations staff. community wide activities that are likely to
influence NOAA climate forecast operations
The CTB framework includes both in 2-3 years. These activities will be funded
“infrastructure” and “transition projects”. by NOAA and potentially other U.S.
CTB infrastructure includes computing agencies, proposed and selected through a
support (e.g. the NCEP Research formal review process. Proposal selection
Supercomputer), software support (e.g., and peer review will be managed by the
models, data, diagnostics), contractors, NOAA Climate Program Office. If other
technical assistants, system administrators, organizations contribute, their input is
management and administrative staff. CTB considered in the selection process.
infrastructure will support the exchange of Science priorities for base funded
climate forecast system software (e.g. activities include climate forecast system
models, data assimilation, codes, testing assessments (e.g. model diagnostics and
data sets, etc.). This software will be testing), and climate products and
donated by NCEP and others who wish to applications (e.g. consolidation of forecast
participate in the CTB. Augmentation tools, skill masks, objective verification,
funding for CTB infrastructure will be product development for climate monitoring
needed to support this software. The CTB and assessment). AO’s will support
infrastructure will also support a capability community-wide participation in the CTB
for generating and producing multi-model on broader science issues such as
ensembles (subject to CTB implementation
guidelines). The CTB will maintain Earth • Multi-model ensemble system
System Modeling Framework (ESMF) development; MM climate forecasts
compatible software when it becomes and applications; Attribution;
capable of supporting a full S/I forecast • Model development through Climate
system and the software has been converted Process and modeling Teams;
to ESMF-compatible structures. While the • Climate product and applications
bulk of the computing resources will be development (e.g. predictability
supplied by the NOAA “Research studies, drought monitoring and
computer” resource, augmentation funding prediction, extreme events);
for supporting infrastructure, such as system • Climate reanalysis and data impact;
administration, will be required. • Advanced forecast capabilities (e.g.
Scientific work in the CTB will be ecosystems; air chemistry; carbon
accomplished via transition projects, that cycle; fisheries)
include both base funded activities and
3
Collaboration, via AO-driven proposals, may produce additional breakthroughs in
with organizations maintaining climate- forecasting.
relevant software will be encouraged. In the
more distant future, expansion to 2) ESMF
international climate forecasting by adding
ECMWF, Canadian, Met Office, and Meteo The recent emergence of the Earth System
France products is a major goal. Modeling Framework (ESMF) provides an
The CTB will maintain strong linkages excellent opportunity for the CTB to
to CLIVAR, GEWEX, and joint CLIVAR- investigate new model components and to
GEWEX programs such as the Climate improve operational climate forecast
Prediction Program for the Americas systems using ESMF compatible software.
(CPPA), the GEWEX Americas Prediction GFDL’s climate forecast codes are
Project (GAPP), Pan American Climate becoming ESMF compatible. NASA is
Studies (PACS), CLIVAR Atlantic and building an ESMF compatible model and
Pacific, Variability of the American NCEP is constructing ESMF-compatible
Monsoons (VAMOS), the Climate versions of its next-generation global model.
Dynamics and Experimental Prediction
(CDEP) program and others. Some specific 4) JCSDA
examples include:
The NASA-NOAA-DOD Joint Center for
1) CLIVAR/Climate Process and Modeling Satellite Data Assimilation provides a
Teams (CPT’s) leveraged resource for addressing climate
data assimilation and modeling issues. The
The CTB is closely related in concept to the JCSDA supports observations and data
CLIVAR Climate Process and modeling assimilation (ocean, land-surface,
Teams (CPTs). The CTB priority of atmosphere) software and proposals in this
improving physical parameterizations in arena can be executed more efficiently due
NOAA climate forecast models can be to the supported infrastructure.
considered as the NOAA implementation of
the CPT concept by providing computing 5) NOMADS
resources and an operational testing
environment to address high priority science Dissemination and archiving of CTB data
issues that are common among operational sets will be done through the NOAA
climate prediction models. The NOAA Operational Model Archive and Distribution
Climate Test Bed emphasizes S/I time scales System (NOMADS). GFDL and NCEP
and is uniquely oriented towards scientists have been part of a larger
applications that can be transitioned to consortium that has been developing
operational implementation at NOAA. The NOMADS capabilities for broad data set
NOAA CTB will support competitive, peer availability.
reviewed proposals oriented toward
contributions to NOAA operational It is important to emphasize that the
missions. With this strategy, the CTB will CTB will not take the lead on the following
maintain a focus on areas of investigation activities under the purview of NOAA
most likely to directly improve NOAA’s climate forecast operations or the
operational numerical forecast systems yet participating NFCs:
remain open to advanced technologies that
4
• Operational forecasting; • Science Advisory Board (SAB): to
• Real-time monitoring or integrated coordinate with the broader science
climate assessments; community and with other programs,
• NCEP operational climate forecast and to provide “independent” advice
system development, integration, and on high-priority scientific
calibration; challenges.
• Model development for participating
NFC’s. Complete ToR and responsibilities for the
OB, CST and SAB are also found on the
CTB web page.
4.0 Management Structure
5.0 Supported Software
The CTB management structure includes
the following areas: oversight and scientific NCEP’s next generation model will be
direction, programmatic and scientific on the development path for upgrades to
management and computing. The CTB staff NCEP’s operational climate forecast system.
will include a Director, a Deputy Director, a The large amount of testing necessary for
Program Manager, Administrative Staff, putting any new system upgrade into
Technical Assistants, Focal Points, and other operations, which includes lengthy
associated personnel (e.g. Visiting Scientists calibration runs, ensures that these upgrades
and Students). Personnel will consist of will occur every 3-4 years, rather than
civil servants, contractors, visiting scientists several times annually as in weather
and students. Complete Terms of Reference applications. Thus, the next upgrade to
(ToR) and responsibilities for CTB staff are NCEP’s system will probably occur in 2007.
found on the CTB web page at: Therefore, there is now an excellent
opportunity for the community to work with
http://www.cpc.ncep.noaa.gov/products/ctb NCEP on its next operational upgrade. All
components of NCEP’s system are now in a
The CTB also includes an Oversight position to be upgraded, including the
Board (OB), a Climate Science Team (CST) atmospheric, ocean, land-surface and sea ice
and an external Science Advisory Board forecast models, standardized (formatted)
(SAB). These groups have unique observations, data assimilation systems
responsibilities and charges that are briefly (ocean, atmosphere, land-surface), post-
summarized as follows: processing and archive capability.
• Oversight Board (OB): to make In order to improve NCEP forecasts
recommendations to the Directors of from Week 2 to S/I, a multi-model ensemble
the NOAA Climate Office and strategy will be pursued in the CTB. The
NCEP concerning the CTB on an CTB will host an ESMF-compatible
annual basis; software framework that will promote the
• Climate Science Team (CST): to experimentation and development of multi-
guide CTB activities at the working model ensemble products.
level, including evaluation of CTB
activities and recommendations to The CTB will collaborate with the
CTB management on computing JCSDA to develop improved data
resource usage and access; assimilation techniques for the CTB
5
applications. Common software will be operational climate forecasts, products and
maintained jointly between the JCSDA and applications. These activities are expected
CTB, with the JCSDA focusing on the data to directly influence NOAA climate forecast
assimilation modules (primarily ocean and operations within 1 year, with definite
land surface) and the CTB focusing on outcomes and lifecycles. Visiting scientists
specialized data sets and output products. will be supported by participating NOAA
Any Reanalysis capability will draw upon organizations (includes reassigned personnel
JCSDA software but be executed within from other centers). Annual
CTB resources. Announcements of Opportunity (AO’s) will
be funded by NOAA and other agencies to
6.0 Resources support community wide participation on
broader research issues. The NOAA Office
A three-year ramp up of CTB resources of Global Programs will undertake
is planned. CPC and EMC FTEs have been administrative tasks.
reallocated to the CTB. As mentioned NOAA Climate Office funding will
above, the bulk of the computing resources provide a venue for scientists in the external
will be supplied by the NOAA “Research research community to work on NCEP
Supercomputer”. Augmentation funding models and products through the CTB.
will be needed to support contractors in NCEP has limited manpower to work with
order to transfer and enhance CTB software external PI’s on AO-driven CTB projects
and datasets, and to maintain computing and complete its own base funded projects.
infrastructure. The CTB management will Thus, the CTB will require a few “bridge”
establish an overall policy on allocation of scientists, to help infuse AO-driven research
computing resources (with approval by the into operations.
OB). NCEP/NCO will provide the
management of the computing resources
(user access, accounting, etc) according to
established policies.
Base funded activities will be funded by
NOAA organizations to accelerate the
transition of R&D to improved NOAA
6
