The Role of Coupled Versus Uncoupled Reanalyses for Initialization of Decadal Climate Predictions
Final Report Abstract
Ocean has been described in previous studies as the memory of the climate system. Ocean’s internal variability has been shown to be predictable decades ahead within climate models, providing the hope that synchronizing ocean circulation with observations can improve decadal climate predictions (DCPs). Predictions require a starting point derived with data assimilation (DA) methods that blend available observations and climate model states together. A variety of methods used for the state-of-the-art DCPs within the Climate Model Intercomparison Project Phase 6 and Decadal Climate Prediction Project ranges from coupled to uncoupled approaches, online to offline, initializing different media of the climate system and using different data streams. Multi-model predictions are synchronized with similar observational sources but are carried out with different climate models and from different DA products, thus leading to somewhat different initial ocean circulations. If the ocean is the memory of the climate system, why is this variety not more pronounced in the prediction skill of DCPs? Instead, the multi-model DCPs show similar prediction skill which, recently, hardly improves over that from the un-initialized simulations. In this project, we aimed to identify promising developments in terms of data assimilation and initialization as this topic is of essence for the future development of DCPs. Originally, our study aimed at comparing the idealized experiments based on the coupled adjoint and coupled nudging carried out with the Earth System Model of intermediate complexity CESAM (Centrum für Erdsystemforschung und Nachhaltigkeit Erdsystem Assimilations-Model). Most of the planned experiments are completed, apart from the DCPs starting from the coupled adjoint DA, which we decided not to run after the quality check. More specifically, from the test experiments follows that different DA setting choices strongly affect the resulting Atlantic Meridional Overturning Circulation (AMOC). For instance, both the fullfield coupled nudging and adjoint DA, resulted in a split of the AMOC cell, which is a guarantee for initialization shocks in DCPs. How critical the consequences of such issues at the assimilation step are for the prediction skill, e.g., at the surface of the ocean, has not been fully understood. To follow up on this, in addition to the CESAM runs, we analysed the AMOC from the 12 operational DCP systems from the WMO Lead Centre for Annual-to-Decadal Climate Prediction. We identified that the AMOC initialization influences the quality of temperature predictions in the North Atlantic subpolar gyre (SPG) region. When predictions show a large initial AMOC error, they usually have low skill for predicting internal variability of the SPG 5 years after the initialization. While the anomaly-initialized predictions do not experience large drifts, they show low SPG skill when skill (due to external forcing) also present in historical runs is removed using a residual correlation metric. Thus, our study demonstrates that initial shock (in full-field DA) and model biases (in anomaly DA) in the ocean circulation in prediction systems reduce prediction quality for the SPG beyond 5 years. However, we do not find any clear benefit for skill from the level of advancement of DA methods (nudging versus other methods). The results of our study are of direct relevance for the producing centres and those interested in skilful predictions in the North Atlantic sector.
Publications
-
A design of the optimal setup for a coupled data assimilation for decadal climate predictions, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-3616, 2022 (Oral)
Polkova I., Lyu, G., Stammer, D., Schubert, S., Lunkeit, F. & Köhl, A.
-
A design of the optimal setup for a coupled data assimilation for decadal climate predictions, OS56B-03, AGU 12- 16 December 2022, hybrid (Oral)
Polkova I., Lyu G., Stammer G., Schubert S., Lunkeit F. & Köhl A.
-
Optimal coupled data assimilation for decadal climate predictions, Workshop Seamless seasonalto-decadal predictions state of the art and opportunities JRC Ispra, 8-10 October 2022 (Oral)
Polkova I., Swingedouw D., Lyu G., Schubert S., Stammer D., Lunkeit F. & Köhl. A.
-
How good should the initial conditions for decadal climate predictions be in terms of the Atlantic meridional overturning circulation? ECMWF’s 2023 Annual Seminar “Earth System Reanalysis”, 4-8 September 2023, ECMWF’s Headquarters in Reading, UK (Pitch and Poster)
Polkova I., Swingedouw D., Hermanson L., Smith D., Köhl A., Lyu G., Stammer D., Schubert S. & Lunkeit F.
-
How good should the initial conditions for decadal climate predictions be in terms of the Atlantic meridional overturning circulation? WCRP Open Science Conference 2023, Kigali, Rwanda (Pitch and Poster)
Polkova I., Swingedouw D., Hermanson L., Köhl A., Stammer D., Smith D., Kröger J., Bethke I., Yang X., Zhang L., Nicolì D., Athanasiadis P.J., Karami M.P., Pankatz K., Pohlmann H., Wu B., Bilbao R., Ortega P., Yang S. & Matear R.J.
-
How good should the initial conditions for decadal climate predictions be in terms of the Atlantic meridional overturning circulation? WCRP Workshop on Improving climate models and projections using observation MIT, Cambridge MA, USA, June 12 – 14, 2023 (Oral)
Polkova I., Swingedouw D., Hermanson L., Smith D., Köhl A., Lyu G., Stammer D., Schubert S. & Lunkeit F.
-
How good should the initial conditions for decadal climate predictions be in terms of the Atlantic meridional overturning circulation? Workshop "Meeting AMOC Observation Needs in a Changing Climate", 18-20 July 2023, Hamburg (Pitch and Poster)
Polkova I., Swingedouw D., Hermanson L., Smith D., Köhl A., Lyu G., Stammer D., Schubert S. & Lunkeit F.
-
Initialization shock in the ocean circulation reduces skill in decadal predictions of the North Atlantic subpolar gyre. Frontiers in Climate, 5.
Polkova, Iuliia; Swingedouw, Didier; Hermanson, Leon; Köhl, Armin; Stammer, Detlef; Smith, Doug; Kröger, Jürgen; Bethke, Ingo; Yang, Xiaosong; Zhang, Liping; Nicolì, Dario; Athanasiadis, Panos J.; Karami, Mehdi Pasha; Pankatz, Klaus; Pohlmann, Holger; Wu, Bo; Bilbao, Roberto; Ortega, Pablo; Yang, Shuting ... & Matear, Richard J.
-
Toward model-consistent initialization of decadal climate predictions, CMCC Webinar, 21 February 2023 (Oral)
Polkova I.
