MiKlip first phase: TORUS

TOwards Regionally focUsed modelling of decadal climate predictionS

The proposed joint research of TORUS focus on the following technical and scientific key question: How will decadal prediction benefit from an improvement in simulating the North Atlantic- Arctic key region? Therefore we will implement a global setup, the TORUS model system, with the finite-element ocean-sea ice model FESOM and the atmospheric model ECHAM. The use of FESOM with its unstructured grid allows a strong regional refinement of Arctic key regions. The project aims on studying the influence of an improved regional modelling of the North Atlantic-Arctic on the simulation of decadal variability and predictability.
For the validation, improvement and initialization of the TORUS and of the MiKlip system new field oceanographic data in key areas of the Arctic Ocean will be provided. The global TORUS model system will contribute to the MIKLIP system particularly by performing sensitivity studies with respect to the influence of a different ocean model formulation and to the influence of regionally refined resolution of the Arctic aiming at better understanding of key mechanisms of decadal variability and predictability. Additionally, ensemble simulations for the next decade will be performed with the TORUS model system in order to assess the influence of an improved representation of the North Atlantic-Arctic key region on decadal predictions.
Accordingly, the project TORUS consists in different work packages. The global coupled TORUS model system, consisting in the ocean-sea ice model FESOM with an unstructured grid and the atmospheric model ECHAM will be implemented and validated. Control runs will be carried out and sensitivity studies concerning the influence of regionally focused modelling on decadal variability will be performed with two different refinement grids. New oceanographic data sets will be collected and provided for the validation and initialization of the TORUS and MiKlip system. During the last phase of MiKlip, TORUS will contribute to the compilation of decadal climate predictions by providing ensemble simulations with the TORUS system.

The meshes of the atmospheric model ECHAM (right) and of the ocean-sea ice model FESOM (left), which are coupled via an exchange grid and the coupler OASIS4.

The TORUS project will contribute to the research challenges of MiKlip

  1. by providing the global TORUS model system which allows regional refinement of resolution in high latitude key regions,
  2. by contributing a model with a different dynamical core for the coupled sea ice-ocean system to the prototypical system for decadal prediction in order to assess the uncertainty introduced by climate model deficiencies,
  3. by an improved understanding of physical mechanisms especially those in high latitudes, governing decadal climate variability and predictability in the Arctic-North Atlantic region.

Goals

TORUS aims at achieving the following specific objectives:

  • Providing a model system with an alternative formulation of the ocean dynamics which allows regionally focused modelling of the Arctic key region. To avoid the numerical, physical and technical problems of model nesting or open boundaries, we will implement a global setup with the unstructured grid, finite-element ocean-sea ice model FESOM and the atmospheric model ECHAM. 
  • Assessing the effects of enhanced resolution. Within TORUS sensitivity studies will be performed on the influence of regionally enhanced resolution over the Arctic ocean on the model climate and its simulation of decadal variability and predictability.
  • Investigating mechanisms of decadal variability.  TORUS will study in particular dynamical processes in the Arctic key region which are fundamental for generating large-scale atmospheric and oceanic variability patterns and hence decadal variability and predictability.
    Coupling of additional climate subsystems. The proposed model set-up will enable us to study whether the regionally enhanced modelling of especially sea-ice has the potential to improve the ability of current climate models in simulating the observed climate and decadal variability in particular over the Arctic.
  • Providing new Arctic observations for model validation and initialization. The areas of specific interest of Arctic observations include northern Kara Sea, the Laptev Sea shelf and continental margin, and the junction of the Lomonosov Ridge with Siberian shelf in order to exploring regional impact from coastal polynias, dynamics of the Arctic halocline water and intermediate Atlantic water.

 

TORUS - Results from DS1

The focus of the project TORUS concentrates on the improvement oft the simulation oft the Arctic climate system by regionally focused modelling within a global coupled model system. In order to achieve this goal, a new coupled atmosphere-ocean-sea ice model has been developed. The ocean-sea ice model component FESOM applies unstructured grids und enables us to regionally refine the spatial resolution in particular regions in accordance with the scientific question of interest. First, the activities on the model development concentrated on the technical development of the coupled model system ECHAM5-FESOM. After the successful implementation of ECHAM5-FESOM and the official release of the atmospheric model ECHAM6 the technical development of the new coupled model system ECHAM6-FESOM had been started and successfully finished by the end of 2012 within the DS1.

Since then is the model ECHAM6-FESOM being tested with validation simulations and a long control simulation has been started. The results of the first 300 years of this control simulation have been described in detail in Sidorenko et al. (2013). The results show that ECHAM6-FESOM performs at least as well as some of the most sophisticated climate models participating in CMIP5. There are still model deficits like, e.g., a too weak meridional overturning ocean circulation, deviations in the position and strength oft he North-Atlantic current and the unrealistic periodic freezing of the Labrador sea. Such model deficits are typical for ocean models of this complexity. In the future, grid configurations with increased regional refinement, e.g. in the North Atlantic ocean, will be tested to improve the estimated model deficits.

Since the beginning of the DS2 of MiKlip additional preparatory work for the planned predictability studies has been started. This work includes sensitivity studies and ensemble simulations to investigate the influence of a stochastic sea ice parameterisation. The sensitivity studies prove the acceleration of the Arctic and Antarctic sea-ice drift within the coupled stochastic system in comparison to a deterministic multi-decadal reference simulation and an increase of the multi-year sea-ice at the Arctic Canadian Archipelago. In the framework of the cooperation with the MiKlip-project SPARCS sensitivity studies to investigate the influence of the parameterization of surface fluxes taking onto account a form-drag have been carried out. The preliminary results of these studies revealed no general improvement for the Arctic by applying the new parameterisation, but regional changes did occur. In contrast, improved sea-ice concentrations have been simulated in winter and spring for the Antarctic.

More on TORUSNews-Icon

Contact

Alfred-Wegener-Institut für Polar- und Meeresforschung (AWI) AWI Potsdam
Dr. Dörthe Handorf
Prof. Klaus Dethloff
Dr. Wolfgang Dorn

AWI Bremerhaven
Prof. Thomas Jung
Dr. Jens Schröter
Dr. Sergey Danilov
Dr. Dmitry Sidorenko
Prof. Wolfgang Hiller
Dr. Dirk Barbi

Towards multi-resolution global climate modeling with ECHAM6-FESOM. Part II: climate variability

2016 - Climate Dynamics, pp. 1–26

Rackow, T. | H. F. Goessling, T. Jung, D. Sidorenko, T. Semmler, D. Barbi, and D. Handorf

Potential sea ice predictability and the role of stochastic sea ice strength perturbations

2014 - GRL, Vol. 41 (23), pp. 8396–8403

Juricke, S. | H. F. Goessling, and T. Jung

Towards multi-resolution global climate modeling with ECHAM6–FESOM. Part I: model formulation and mean climate

2014 - Clim. Dyn, Vol. 44 (3), pp. 757-780

Sidorenko, D. | T. Rackow, T. Jung, T. Semmler, D. Barbi, S. Danilov, K. Dethloff, W. Dorn, K. Fieg, H.F. Goessling, D. Handorf, S. Harig, W. Hiller, S. Juricke, M. Losch, J. Schröter,D. V. Sein, and Q. Wang