EXPERIMENTAL CLIMATE PREDICTION CENTER

John Roads (SIO)

Link to NOAA Strategic Plan: NOAA’s Mission Goal 2: Understand Climate Variability and Change to Enhance Society’s Ability to Plan and Respond; Goal 3: Serve Society’s Needs for Weather and Water Information

RESEARCH OBJECTIVES AND SPECIFIC PLANS TO ACHIEVE THEM

The goal of the Experimental Climate Prediction Center (ECPC) is to develop an integrated, global to regional, climate prediction capability. ECPC undertakes research to:

Identify coupled modes of interannual variability

Persistent surface anomalies in the ocean and land have a strong influence upon atmospheric features that would otherwise be unpredictable beyond a few weeks.

Develop models capable of predicting these modes

Experimental state of the art numerical models are being acquired from national centers (e.g. National Centers for Environmental Prediction) as well as being developed within the ECPC.

Evaluate the predictive capability of these models

Experimental predictions are made, routinely, to evaluate how well these numerical models can predict at various time scales ranging from days to years. As these predictions are evaluated, experimental predictions are posted to the world-wide-web for download by interested researchers.

Transfer ECPC methodologies to other application centers

ECPC collaborates with researchers at a number of regional application centers to transition experimental prediction methodologies to operation.

ECPC emphasizes research leading to routine experimental predictions, applications, and products with an integrated modeling system that fully couples the atmosphere, land, and ocean at global to regional scales.

RESEARCH ACCOMPLISHMENTS

During the past year, we have worked to develop and apply an updated version of NCEP's seasonal forecast model (SFM), which is based on updated physics from the NCEP/DOE reanalysis II (RII). These forecasts are initialized from the NCEP analysis and compared to previous forecasts initialized from the RII analysis.

We have continued to compute a parallel seasonal forecast every month for IRI, which is also used by other institutions. In contrast to the initial condition forecast above, we compute a 12-member 7-month forecast using predicted SSTs provided by IRI, as well as a 10-member 4-month forecast using persisted SSTs. These runs are initialized from previous runs forced by observed SSTs. In addition we run a 12-member 3-month downscaling over the western U.S. We have assisted the IRI in running an extended series of hindcasts using the ECPC SFM. We are also nearing completion of a 12-member 50-year AMIP-style run using a new cloud scheme.

The ECPC SFM and RSM atmospheric-land modeling systems have now been merged and are hereafter referred to collectively as G-RSM. As requested by the USFS, our short-range climate forecast system is now being focused on 4 month forecasts made at the beginning of each month along with corresponding historical runs made for the corresponding months from 1982-present. Global and regional historical simulations are being run to better understand the model capabilities. We are contributing these simulations to regional model comparison projects in order to better understand the capabilities of our G-RSM. In addition, we have contributed an augmented global analysis from the RII and an updated version of the RII—using the SFM in place of the original RII model—to the Coordinated Enhanced Observing Experiment.

There are only a few physical parameterization differences between the ECPC G-RSM and the NCEP Coupled Forecast System (CFS) global forecast model. In particular, the ECPC G-RSM has an updated land surface model (Noah), similar to the medium range Global Forecast System (GFS) in use at NCEP. Other model improvements currently being evaluated are concerned with improved land surface and cloud parameterization schemes. We fully intend, that our ECPC G-RSM will continue to remain similar and relevant to the NCEP operational large-scale seasonal forecast system.

We are looking at even longer prediction horizons. Our new G-RSM is now coupled to the MIT ocean model, uses the Jet Propulsion Laboratory (JPL) monthly ocean analysis as initial conditions and our independent experimental ECPC Coupled Prediction Model (ECPM) could eventually contribute to a community coupled model forecast ensemble. In addition, we have also developed a coupling of the RSM to the Regional Ocean Modeling System (ROMS), which we refer to as the Scripps Coupled Ocean-Atmosphere Regional (SCOAR) model. This regional model has been used to simulate a number of regional ocean phenomena and in conjunctions with the ECPM may eventually be used to make experimental regional ocean predictions.

Further details about ECPC progress are provided on the ECPC site: http://ecpc.ucsd.edu/

Fig. 1 Near surface temperature (degrees C) in June 2001. SMALL: downscaling run over Europe, NH: downscaling run over the NH, CRU: gridded observation from CRU, and NCEP-NCAR Reanalysis.

Fig. 2 Percentage variance from the CMORPH precipitation product (2003-2005) described by the (a) annual, (b) intraseasonal, (c) slow synoptic, (d) fast synoptic, (e) high-frequency, and (f) diurnal variance categories. The percentages listed show the theoretical values of each variance category from a white-noise spectrum .