Modeling and Data Assimilation Task Team

The 2014 TPOS 2020 Project Workshop identified inadequacies in models and in data assimilation as the major limiting factors for effective use of TPOS observations in seasonal-to-interannual climate predictions and the accuracy of related products, including both the analysis of the ocean state and the predictive skill of coupled model forecasts. In order to address this, inadequacy, the Modeling and Data Assimilation Task Team was created.

The Modeling and Data Assimilation Task Team will specifically address the following:

  • Develop strategies for coordinated modeling and assimilation activities for designing and planning the future TPOS observing systems, such as those proposed by the other task teams.
  • Identify pathways that will contribute to improved understanding of systematic errors and subsequent model improvements, especially through promotion of joint activities with other bodies that have mandates to improve models.
  • Contribute modeling and data assimilation insights into the identification of observational requirements.
  • Provide guidance on the assessment of the impact of modeling and assimilation, including through systematic continuous evaluation (metrics and process-oriented diagnostics), OSEs, and OSSEs, of the TPOS and its design, especially using the multi-model approach.
  • Recommend strategies for model initialization that will promote the efficient use of TPOS information.
  • Provide recommendations on improving coordination among centers currently engaged in ocean analysis and prediction towards assessment of TPOS and its influence on predictions.

Project Function

A significant fraction of the observational effort will need to be directed towards improved understanding of processes and mechanisms, which in turn should be coordinated with a program of improved model parameterizations and reduced systematic error; an additional benefit is that such data/model studies also contribute to improved design and reduced inefficiencies of the observing system.

Task Team Publications

  • Dec. 2014: Terms of Reference
  • Task Team Details

    Task Team Members and Affiliations:

  • Arun Kumar, NOAA Center for Weather and Climate Prediction, USA*
  • Eric Guilyardi, Laboratoire A1:D64 et du Climat : Expérimentations et Approches Numériques, France*
  • Oscar Alves, Bureau of Meteorology, Australia
  • Yosuke Fujii, Japan Meteorological Agency, Meteorological Research Institute, Japan
  • Alicia Karspeck, National Center for Atmospheric Research, USA
  • Yukio Masumoto, Observational Research System for Global Change, Japan
  • Tim Stockdale, European Centre for Medium-Range Weather Forecasts, United Kingdom
  • Jennifer Waters, UK Met Office, United Kingdom
  • Andrew Wittenberg, NOAA Geophysical Fluid Dynamics Laboratory, USA
  • *Denotes Co-chairs

    Terms of Reference

  • To develop strategies for coordinated modelling and assimilation activities for designing and planning the future TPOS observing systems, such as those proposed by the other task teams.
  • To identify pathways that will contribute to improved understanding of systematic errors and subsequent model improvements, especially through promotion of joint activities with other bodies that have mandates to improve models.
  • To contribute modelling and data assimilation insights into the identification of observational requirements.
  • To provide guidance on the assessment of the impact of modelling and assimilation, including through systematic continuous evaluation (metrics and process-oriented diagnostics), OSEs, and OSSEs, of the TPOS and its design, especially using the multi-model approach.
  • As appropriate, recommend strategies for model initialization that will promote the efficient use of TPOS information.
  • To provide recommendations on improving coordination among centers currently engaged in ocean analysis and prediction towards assessment of TPOS and its influence on predictions.
  • Background

    From the La Jolla Workshop and associated White Paper(s)

    The Workshop identified inadequacies in models and in data assimilation as the major limiting factors for effective use of TPOS observations in seasonal-to-interannual climate predictions and the accuracy of related products, including both the analysis of the ocean state and the predictive skill of coupled model forecasts. Inadequacies could be model errors associated with either the atmospheric or oceanic component of the coupled models, or could be related to data assimilation methodologies.

    Experience with atmospheric reanalyses and weather forecast systems clearly indicates that more useful information can be extracted from observations as the forward models and assimilation systems improve. Further, just as multi-model forecasts for sub-seasonal and seasonal time scales have led to greater forecast reliability and increased accuracy, so too multi-model analyses could lead to greater reliability in the estimation of the ocean state and for the quantification of analysis errors. In summary, the “route to impact” for the TPOS 2020 observations is inextricably linked to improvements in the modelling and assimilation systems, and to enhanced coordination across various centers engaged in ocean analysis and prediction.

    Given the remarks above about the criticality of such efforts for achieving scientific and societal impact of the TPOS 2020 observations, the TPOS 2020 needs to embrace modelling and assimilation activities as part of its overall strategy. The results from these efforts will also assist in the identification of model errors, areas of large uncertainty where model/reanalyses diverge (and additional observational constraints may be required), and process studies needed for improving relevant model components. In addition, further coordinated observing system and process study experiments may be needed to assist the design of a future observing system beyond TPOS 2020. Modelling and assimilation activities, therefore, should involve (ocean and coupled) model and forecast system developers and TPOS observationalists.

    TPOS observational requirements will also vary with the evolution of the forecasting systems both in the context of providing adequate observations and to further challenge analysis and forecasting systems. As model resolution increases, the observational needs for forecast initialization, process evaluation, parameterization and verification may change. For TPOS 2020, the typical ocean resolution would range from about 1° x ½° for climate applications (likely finer by 2020) to 1/12 of degree (or higher) for global medium-range ocean forecasting. The vertical resolution of the upper ocean is already about 1 m in some models but is coarser around the depth of the thermocline. The increased model resolution and complexity of the forecasting system in the future (coupled ocean-atmosphere-wave models for the medium range) also has implications on the requirements for initialization.

    Guidance from the 1st TPOS 2020 Steering Committee meeting

    Improvement in the forward model for ocean and intraseasonal to interannual climate prediction systems was regarded as a priority; no amount of improvement in the observing system or assimilation methodology will be able to adequately mitigate the issues arising from systematic model errors. Similarly, the accurate depiction of key processes across a wide range of time and space scales (diurnal, intraseasonal, interannual, decadal) is germane to Pacific climate and its global impacts. Analysis/assimilation issues have been identified and will need to be addressed in parallel. The diversity among ENSO events, and the changes in variability and predictability and model skill between different epochs, also underlines a need for improved understanding of the two-way interaction of ENSO (and other variability such as the Madden-Julian Oscillation) with the background climate based on existing and enhanced observations and models.

    The TPOS 2020 Project has a number of potential strategies available to enhance the value derived from the observing system: improvements in assimilation systems to optimise the value derived from existing observational assets; and sensitivity studies to contribute to the design of the observing system, leading to a more efficient and effective use of existing and planned resources.

    The immediate implication for TPOS 2020 is that a significant fraction of the observational effort will need to be directed towards improved understanding of processes and mechanisms, which in turn should be coordinated with a program of improved model parameterisations and reduced systematic error; an additional benefit is that such data/model studies also contribute to improved design and reduced inefficiencies of the observing system.

    A number of processes/mechanisms received attention:

    • Processes in the fresh pool region of the western, including moist convection in the vicinity of the ITCZ on timescales from diurnal to intraseasonal to interannual and decadal
    • The role of salinity variations in variability, predictability and prediction on intraseasonal to interannual timescales;
    • Improved understanding of coupling between the surface and the thermocline
    • mpact of the slowly varying large scale environment (e.g., decadal oscillations) on ENSO predictability diversity and forecast skill
    • Interactions and teleconnections with other tropical regions and the extra tropics/mid-latitudes (e.g., the NW Pacific).

    A number of options were considered in order to address these and related issues:

    • Investigate opportunities to engage with the WCRP/GEWEX Grand Challenge on Clouds Convection and Climate Sensitivity to address Convection/ITCZ issues.
    • Engage with the Year of the Maritime Continent initiative, perhaps in conjunction with the Chinese Western Pacific Big Cross effort, SPICE-related activities etc., which are likely to deliver model improvement and improved knowledge of the East Asian monsoon, especially at intraseasonal time scales (i.e. the MJO).
    • Improve understanding of coupling between the surface and the thermocline; a model/data process experiment in the cold tongue mixing and upwelling region was discussed.
    • Promote intercomparison studies on the assimilation of salinity in ocean and climate models.
    • Develop an improved framework for observing sensitivity studies, one that draws more widely on the accrued scientific knowledge in addition to model guidance.
    • Examine options for a tropical ocean systematic errors workshop, involving observational, modelling and assimilation experts, perhaps in conjunction with WGCM and WGNE.

    The TPOS 2020 SC accepts that it may not have the mandate to lead in all of these areas so engagement and cooperation with existing mechanisms (e.g., WGCM, WGNE, GEWEX-GASS, WCRP Cloud, Circulation and Climate System Grand Challenge) will be important.