Climate Mission Outcome

1
Overview of Climate Predictions and Projections
Program

• Climate Mission Outcome
• A predictive understanding of the global climate
  system on time scales of weeks to decades with
  quantified uncertainties sufficient for making
  informed and reasoned decisions (Criteria for
  success progress measured by indicators of
  predictive understanding and skill scores)

2
Future New and Improved Products (preliminary
modeling work going on for most of these already)

• Climate
• Improved intraseasonal to seasonal to decadal
  forecasts
• Scenarios for future climate mitigation and
  adaptation studies
• Assessments of potential for abrupt changes -
  surprises
• Utilization of Earth System models in expanding
  product suite
• Water resource drought forecasts including
  nutrient runoff
• Climate related health and disease forecasts
• Projections of sea level changes
• Ecosystems
• Ecological assessments and predictions from
  climate change
• Fisheries productivity forecasts that incorporate
  the effects of climate
• Improved assessments of sea level change on
  coastal resources and ecosystems
• Forecasts and mitigation strategies related to
  air/water quality and quantity in coastal zone
• Weather and Water
• Improved 10-14 day forecasts
• Regional and continental scale air-quality and
  atmospheric chemistry predictions
• Improved forecasts for water resources (droughts,
  floods) including interactions with estuaries and
  coasts

3
Functional Structure of Predictions and
Projections Program (Seasonal to Interannual
Component Shown)
Operational Forecasts
New and Improved Products
Information Products
Test Bed - transition to operations

• Systematic Research forecasts and applications
  (Research PMs)
• establish systematic research multi-model SI
  prediction activity
• establish multi-model Hydrological prediction
  system
• Test application models drought, fire, water
• Improve consolidation tools

• Routine Attribution reports

• Multi-model-based predictability studies
• Predictability studies
• Experimental predictions
• Studies supporting process research
• Data Distribution capability

Model Data Assimilation System Development
in Environmental Modeling Program

• Process research, hypothesis testing and
  diagnostic studies
• Targeted efforts for improving climate models
  (CPTs, parameterizations,)
• Field experiments in support of model
  improvements CPTs
• global tropical interactions with new focus on
  Indo-Pacific and Atlantic regions
• Monsoon related studies
• Emerging applications (coastal ecosystems air
  quality fisheries,)

Observations, reanalyses, forcings research
4
What can lead to improvements in S/I forecasts -
our strategy

• Develop a (community) research strategy (FY06/Q2)
• Improved dynamical prediction models
• Enhanced use of ensemble information from a
  single model
• Multi-model ensembles
• Improved empirical prediction tools
• Improvements in consolidation procedures
• Improved SST predictions
• Climate Nowcasts (Dynamical OCN)
• Predictability beyond ENSO SSTs

5
A Number of Approaches can Improve Skill Scores
Example - Objective Consolidation Tool
Pink Operational Forecasts (avg. score
17) Blue Objective consolidation forecast tool
(avg. score23)
6
Proposed Structure for Improving Skill of SI
Forecasts Metric for incorporation into
operations improves skill over period of
operational forecasts
Operational SI Forecasts/Skill
Objective Consolidation Tool
OCN
CFS
Empirical Methods
Assessment
Research Foci
Dynamical OCN
Multimodel CDC/IRI/
Research SI Forecasts/Skill
Objective Consolidation Tool
7
Priorities Next 1-5 years resulting from our
CLIVAR planning in 5 year research plans -
NCEP needs?

• Seasonal to Interannual (working towards
  regional capabilities)
• Improve skill of SI predictions
• Establish systematic community based multi-model
  forecasting capability/infrastructure
• Incorporate impacts of Indo-Pacific and Atlantic
  SST anomalies
• Develop dynamical understanding of trends
  incorporate in forecasts
• Implement routine attribution capability
• Develop seasonal hydrological forecasting
  capability (a national drought prediction
  experiment)
• Predictive understanding of influence of climate
  on environment (a new focus)
• Coastal ecosystems and fisheries regimes
• Decadal to Centennial- working towards regional
  capabilities where possible
• Develop experimental decadal trends forecasts
  resulting from predictive understanding of
  anthropogenic and natural variations (Atlantic
  focus) also links directly to SI predictions
• Attribution of climate of 20th C to natural
  versus anthropogenic influences
• Understanding past decadal variability abrupt
  changes
• Reduce uncertainty in future projections
• Implement earth system modeling capability
• Intraseasonal Forecasting
• Improve week2 skill scores
• Develop capability to predict extremes for weeks
  2,3,4.
• Predictive understanding of climate on statistics
  of extremes (hurricanes others)

8
Uses of Multi-model Ensembles
Research - forecasts and AMIP runs - A
distributed activity
Climate Testbed - centralized activity
Application models hydrology, etc.
Attribution and predictability studies
Research forecasts
Operations

• The above need to be more systematic and be
  linked to other national/international activities
• COPES
• CliPAS (APEC-Korea)
• C20C runs
• others

9
MM Ensemble for Attribution and Predictability
Assessments

• What NOAA supported activities currently exist
• Seasonal Diagnostics Consortium
• Continuously updated AMIP runs forced with
  global SSTs. Participating models are from NCEP,
  GFDL, CDC (running CCM3), IRI, GMAO, and ECPC
• Although updating the AMIP runs is a distributed
  activity, centralized collection of data and
  display of basic results is done at CPC
• C20C simulations with different natural and
  anthropogenic forcings
• Need to formalize predictability studies and link
  to NCPO research programs

10
MM Ensemble for Predictions

• What NOAA supported activities currently exist
• MM ensemble predictions at IRI (based on tier-2
  approach with skill assessments for participating
  models obtained from AMIP simulations)
• Empirical-Dynamical SI prediction System at CDC
  (based on a set of tier-2 AMIP model runs)
• Both are distributed approaches. Both are
  pragmatic in the sense that there is no
  consistent set of hindcasts. There are strong
  ties with the multi-model attribution and
  predictability assessment activities.
• Need to have a formal comparison of these
  forecasts with the operational approaches

11
MM Ensembles for Predictions Future

• What more is desirable
• A multi-model tier-1 prediction capability that
  would include several national coupled models
• A consensus among participating entities as to
  what is required to achieve a 1-tier multi-model
  ensemble goal, e.g.,
• What should be the length for the hindcasts?
• Need for a consistent ODA?
• Minimum size of ensemble for each coupled model?
• Distributed or centralized activity?
• What gets implemented on Test Bed?
• What can be achieved under the available
  resource? And if enough resources are not
  available, does meeting requirements halfway
  still beneficial (e.g., reduced length of
  hindcasts)? OR it HAS to be an all or nothing
  approach.
• MM for regional downscaling (S-I, CC scenarios)
• Linking to application models, e.g., hydrological
  predictions