Motivation

Some recent examples of regional scale climate variability and change such as the prolonged drought in the southwestern U.S., increased Atlantic hurricane activity, changes in commercial fish production in the North Pacific and northern North Atlantic and the recent melting of the outlet glaciers on Greenland have raised awareness about climate change and its impact on society on decadal time scales, including amongst decision and policy makers. On these 10-30 year scales, regional variations in climate and their impacts largely represent natural, i.e., internal, variability of the climate system primarily driven by the slowly varying oceans, significantly affecting the anthropogenic (forced) climate change signals.

Decadal prediction is a new, but rapidly growing field. Some recent review articles (e.g., Meehl et al. 2009) as well as white papers (Hurrell et al. 2009; Latif et al. 2009; Balmaseda et al. 2009) detail the need for and usefulness of such decadal predictions, and outstanding issues and challenges with both observations and models. A unique aspect of the decadal prediction problem is that it represents a joint initial and boundary value problem. This implies that best possible initial conditions of the climate system need to be provided. The global ocean is the primary source of the longer-term temporal 'memory' of the climate system. Therefore, robust decadal predictability and prediction assessment require that the ocean be initialized using observational information, synthesized into appropriate initial conditions. A suite of coordinated decadal hindcast and prediction experiments for the period 1960-2035 are being carried out as part of the Coupled Model Intercomparison Project Phase 5 (CMIP5) to improve our understanding of decadal climate variability and predictability. Results from these experiments, which partly are also initialized using ocean data or ocean syntheses, will be evaluated for the Intergovernmental Panel on Climate Change Fifth Assessment Report (IPCC AR5).

Observational examples of decadal variability include the Pacific Decadal Oscillation or Inter-decadal Pacific Oscillation (PDO/IPO) and the multidecadal variability in the Sea Surface Temperatures (SSTs) in the Atlantic Basin, usually referred to as the Atlantic Multidecadal Oscillation (AMO). Coupled general circulation models used in climate studies usually exhibit significant decadal variability / oscillation in their Atlantic Meridional Overturning Circulations (AMOCs). Furthermore, some studies show a broad resemblance between the observed and model simulated SST variability patterns in the North Atlantic that is usually associated with the AMOC variability. The link between the AMO and AMOC cannot be verified by the observations due to the lack of long-term AMOC records. Nevertheless, presence of such long-lived variability as depicted by either the PDO or AMOC forms the basis for decadal prediction studies. Because of its prominent role in the Earths climate system, the AMOC variability and its potential predictability have received much recent attention. Decadal variability and prediction exists in other oceans including the Southern Ocean, Indian and South Pacific Oceans, although lack of long-term observational data severely limits detection of the decadal signal in these regions.

Despite increasing number of studies, many important aspects of decadal variability and prediction remain controversial. For example, the amplitude and period of the AMOC variability differ considerably across ocean models and the associated SST variability patterns, magnitudes, and periods do not match the observational AMO properties in most models. This holds also for ocean models constrained by ocean observations and a detailed study is required as to determine why those differences remain and what observational data base is required to better constrain and understand them. Furthermore, understanding of decadal variability mechanisms is severely lacking. Some recent studies indicate that proper initialization of oceanic indices, e.g., the AMOC and the global upper-ocean heat content, is important for predictability of the climate system. However, robustness of such initialization approaches - attempting to accurately represent present-day low frequency variability in the ocean - across different models remains unclear.

The WGOMD has recently finalized an experimental protocol for the Coordinated Ocean-ice Reference Experiments (CORE-II) forced with interannually varying surface data sets for the period 1948-2007 from Large and Yeager (2008). The CORE-II hindcast simulations provide a highly anticipated framework to evaluate ocean model performance, to study mechanisms of ocean phenomena and their variability from seasonal to decadal timescales, to identify forced variability changes, and to develop mechanistic descriptions of observed climate variability and change. Furthermore, CORE-II experiments will serve as a useful direct comparison to observational studies, and they can be used to optimize the ocean observing systems for the present climate. The natural variability differences amongst models participating in CORE-II experiments will be interesting for understanding and evaluating the robustness of modeled ocean variability and the formation, propagation and decay of anomaly signals that has been observed during the 20th Century. Given the reliability of data assimilation prior to the Argo period is under question, initializing decadal predictions from CORE-II hindcast simulations without data assimilation can be an option. Similarly, anomalies can be extracted from the CORE-II climatology and then applied to the decadal prediction simulations.

An alternative approach is being followed by GSOP: The synthesis of all available ocean data sets by merging them over many years with ocean circulation models. Results of mathematically consistent approaches are dynamically self-consistent and can serve as a complementary basis to study the dynamics of ocean variability, to improve ocean and coupled models, to improve the observing system and to initialize coupled forecast models. GSOP performed now for several years an evaluation of existing ocean syntheses (Lee et al., 2009; Heimbach et al., 2009; Stammer et al., 2009). A next step includes the use of those syntheses for initialization, but especially also for studying ocean decadal variability. Part of those efforts needs to be to assess the adequacy of ocean syntheses for such studies as well as identifying and improving existing shortcomings.

Despite its prominent role in decadal variability and predictability, understanding of the underlying physical mechanisms of oceanic natural variability is clearly missing. As indicated above, so far many efforts have focused on the Atlantic Ocean, with the AMOC as a key player, although the Pacific Ocean also contains intriguing variability on the decadal time scales. There are of course significant difficulties with the decadal variability and prediction problem associated with a paucity of observational data, the long time scales involved, and ocean (and climate) model limitations. The CLIVAR WGOMD and GSOP are motivated to hold a joint workshop on decadal variability, predictability, and prediction, specifically focusing on the oceans role in understanding and modeling the decadal variability. In particular, we believe that the availability of the CORE-II hindcast simulations and the recent advances in ocean syntheses will be of significant help in addressing these issues.

The main goals of the workshop are:

To assess how well the ocean models and ocean syntheses reproduce observed decadal variability,

To understand and evaluate the robustness of simulated ocean internal variability,

To identify the underlying physical mechanisms in the ocean in decadal climate variability,

To evaluate the outcomes of the CMIP5 decadal prediction experiments.

References

Balmaseda, M.A. & Co-authors (2010). “Initialization for seasonal and decadal forecasts” in Proceedings of OceanObs'09: Sustained Ocean Observations and Information for Society (Vol. 2), Venice, Italy, 21-25 September 2009, Hall, J., Harrison D.E. & Stammer, D., Eds., ESA Publication WPP-306.

Heimbach, P. & Co-Authors (2010). "Observational Requirements for Global-Scale Ocean Climate Analysis: Lessons from Ocean State Estimation" in Proceedings of OceanObs'09: Sustained Ocean Observations and Information for Society (Vol. 2), Venice, Italy, 21-25 September 2009, Hall, J., Harrison D.E. & Stammer, D., Eds., ESA Publication WPP-306.

Hurrell, J. & Co-Authors (2010). "Decadal Climate Prediction: Opportunities and Challenges" in Proceedings of OceanObs'09: Sustained Ocean Observations and Information for Society (Vol. 2), Venice, Italy, 21-25 September 2009, Hall, J., Harrison D.E. & Stammer, D., Eds., ESA Publication WPP-306

Large, W.G., and S.G. Yeager, 2008: The global climatology of an interannually varying air sea flux data set. Climate Dynamics, doi:10.1007/s00382-008-0441-3, 24pp.

Latif, M. & Co-Authors (2010). "Dynamics of Decadal Climate Variability and Implications for its Prediction" in Proceedings of OceanObs'09: Sustained Ocean Observations and Information for Society (Vol. 2), Venice, Italy, 21-25 September 2009, Hall, J., Harrison D.E. & Stammer, D., Eds., ESA Publication WPP-306.

Lee, T. & Co-Authors (2010). "Ocean State Estimation for Climate Research" in Proceedings of OceanObs'09: Sustained Ocean Observations and Information for Society (Vol. 2), Venice, Italy, 21-25 September 2009, Hall, J., Harrison D.E. & Stammer, D., Eds., ESA Publication WPP-306.

Meehl, G.A., and Co-authors, 2009: Decadal prediction. Can it be skillful? BAMS, 90, 1467-1485.

Stammer, D. & Co-Authors (2010). "Ocean Information Provided through Ensemble Ocean Syntheses" in Proceedings of OceanObs'09: Sustained Ocean Observations and Information for Society (Vol. 2), Venice, Italy, 21-25 September 2009, Hall, J., Harrison D.E. & Stammer, D., Eds., ESA Publication WPP-306.