400 likes | 494 Views
Climate Historical Forecast Project (CHFP)-WGSIP. Ben Kirtman Rosenstiel School of Marine and Atmospheric Science University of Miami. Multi-model and multi-institutional experimental framework for sub-seasonal to decadal complete physical climate system prediction.
E N D
Climate Historical Forecast Project (CHFP)-WGSIP Ben Kirtman Rosenstiel School of Marine and Atmospheric Science University of Miami
Multi-model and multi-institutional experimental framework for sub-seasonal to decadal complete physical climate system prediction • Baseline assessment of seasonal prediction capabilities using the best available models and data for initialization • Diagnostic Studies, Decision Support Science, Application Modeling • Experimental framework for focused research on how various components of the climate system interact and affect one another • GLACE, SHFP, Sea-Ice HFP, Initialization Strategies … • Test bed for evaluating IPCC class models in ISI prediction • Test bed for evaluating NWP models in ISI prediction
CHFP • Largely a Multi-Model Hindcast or Re-Forecast Effort • Large Array of Saved Fields in Support of Diverse Diagnostic Studies • Application Models, Decision Support Science • Framework for Coordinated Multi-Model Numerical Hypothesis Testing Experiments • Not Serving the Needs of Making Real-Time Multi-Model Forecasts
Experimental Protocol • Coupled models and resolution are left to the individual participants • Complexity of Components is left to individual participants • Atmospheric initial states: NCEP (or ECMWF) reanalysis each February, May, August and November of each year from 1979-present. • 10 member ensembles, 7-months lead • Addition retrospective forecasts using each month of each year from 1979-present are encouraged. • Additional retrospective forecast using initial conditions from each February, May, August, and November 1960-1978 are encouraged • Oceanic initial states: (if appropriate) to be taken from most appropriate ocean data assimilation system. • Sea Ice initial states: (if appropriate) to be taken from best available observational data. • Land initial states: (if appropriate) to be taken from most appropriate land data assimilation system or consistent offline analyses driven by observed meteorology (i.e., GSWP; GLACE2). • No Future Information as the Prediction Evolves – Mimic Operational Setting
Climate Historical Forecast Project • ISI Predictability and Prediction Assessments • Current Forecast Capability (ENSO, Global T2m, P) • Maximum Predictability Not Achieved • Improving Prediction Quality • Untapped Sources of Predictability • Improving the building blocks of forecast systems • Lessons Learned Outstanding Issues
1st WCRP Seasonal Prediction Workshop Kirtman and Pirani (2009) Assessment of Intraseasonal to Interannual Climate Prediction and Predictability US National Academies Maximum Predictability has Not been Achieved http://www.nap.edu/catalog.php?record_id=12878
Predictability - “The extent to which a process contributes to prediction quality” Many sources of predictability remain to be fully exploited by ISI forecast systems • Land Interactions (e.g., Soil Moisture, Snow Cover; Vegetation changes) • Sea-Ice Interactions (i.e., atmosphere-ice; ocean-ice) • Troposphere-Stratosphere Interactions • Sub-Seasonal Variability (e.g., MJO)
Multi-Model vs. Single Model Large Ensemble vs. Multi-Model
Brier Skill Score for Lower/Upper tercile (1980-2001) Temperature and Precipitation
Climate Historical Forecast Project (CHFP) Many sources of predictability remain to be fully exploited by ISI forecast systems • Land Interactions (e.g., Soil Moisture, Snow Cover; Vegetation changes) • Sea-Ice Interactions (i.e., atmosphere-ice; ocean-ice) • Troposphere-Stratosphere Interactions • Sub-Seasonal Variability (e.g., MJO)
Forecast skill: r2 with land ICs minus that obtained w/o land ICs -2 • Initialization is a challenge due to spatial and temporal heterogeneity in soil moisture • Procedures for measuring of land-atmosphere coupling strength are still being developed • Land Data Assimilation Systems (LDAS) coupled with satellite observations could contribute to initialization • Further evaluation and intercomparison of models are necessary
Additional Predictability Likely Associated with Stratospheric Dynamics Stratosphere resolving HFP Goal: Quantifying Skill Gained Initializing and Resolving Stratosphere in Seasonal Forecast Systems • Parallel hindcasts from stratosphere resolving and • non-resolving models • Action from WGSIP-12: Endorse as subproject of CHFP • SPARC to recommend diagnostics
Links across WCRP Explore Seasonal Predictability Associated with Sea-Ice • Sea-Ice Initialization Experiment: • Follow CHFP Protocols for Other Components, Data • Initializing with observed Sea-Ice vs. Climatology • 1 May, 1 November 1996 and 2007 • 8 Member Ensembles • Spring snow melt into soil moisture and influence on spring • temperature anomalies
Links across WMO Several areas of potential collaboration on intraseaonal time-scales: • Investigate how much ocean-atmosphere coupling impacts skill • Role of resolution on skill • Multi-Scale interactions • Ensemble techniques • Intraseasonal Variability (e.g., MJO) • Testing NWP Models in Coupled Mode
Forecasting of MJO is relatively new; many dynamical models still represent MJO poorly 23
Niño 3.4 SST anomaly histogram • mu = mean • std = standard deviation • skew = skewness coef. • kurt = kurtosis coef. • Normal distribution fitted
Improving Forecast System Building Blocks • Sustaining and Enhancing Observing Systems • Improving Data Assimilation Systems (component wise and the coupled system) • Initialization vs. Assimilation • Reducing Model Errors – Resolution Issues – Modeling the Model Uncertainty
The Initial Condition Problem • Best State Estimate • Data Assimilation in the Separate Component Models • Maybe Not Ready to Do the Coupled Assimilation Problem • Coupled Model Climate ≠ Observed Climate • Anomaly Initialization • Coupled “Modes” of Coupled Model ≠ Observed Coupled “Modes” • Initializing the Coupled Modes • Identifying the Coupled Modes: EOFs, SVDs, … • Context of the Forecast Environment
Initializing the Coupled Modes of the Coupled Model Coupled Data Assimilation Nino34 SSTA Evolution CFS Control Initialized Coupled Modes GODAS
Bias Removed Bias Included
CCSM3.0 Jan 1982 IC CFS Jan 1982 IC
CCSM3.0 Jan 1982 IC CCSM3.5 Jan 1982 IC
HRC, LRC Observational Estimate Surface Current Speeds
Rainfall: HRC, and LRC Rainfall: Observational Estimate
HRC Equatorial Thermal Structure LRC
Local Air-Sea Feedbacks: Point Correlation SST and Latent Heat Flux “Best” Observational Estimate Coupled Model Simulation Model of the Model Uncertainty
Contemporaneous Latent Heat Flux - SST Correlation Observational Estimates Increased “Randomness” Coupled Model Control Coupled Model Random Interactive Ensemble: Increased the Whiteness of the Atmosphere forcing the Ocean
Random IE Control 4 4 3 3 2 2 1 1 0 0 -1 -1 -2 -2 -3 -3 -4 -4 Nino34 Regression on Equatorial Pacific SSTA
Outstanding Issues • Quantifying Forecast Uncertainty Due to Uncertainty in Model Formulation • Multi-Model Helps, but Ad-Hoc; Need Models of Model Error (e.g., Stochastic physics) • Quantifying Forecast Uncertainty Due to Uncertainty in Observational Estimates • Initial Condition Problem • Model Error • Need for International Coordinated Effort at Improving Models • Multi-Model is Not an Excuse for Neglecting Model Improvement; Resolution • Data Assimilation (Coupled Assimilation) and Forecast Initialization • Sustained and Enhanced Observing Systems • Climate System Component Interactions • Coupled Ocean-Land-Ice-Atmosphere; External Forcing vs. Natural Variability • Quantifying the Limit of Predictability • Identifying Sources and Mechanisms for Predictability • Data Dissemination Strategy Needs to be Resolved at the Outset