Priority Project "EX-CELO"
EXtension of Cosmo-EuLag Operationalization

Last updated: Aug 2018

Project leader: Zbigniew Piotrowski (IMGW)

Project resources

IMGW

0.65/CY 2017

1.65/CY 2018

1.25/CY 2019

Total:

3.55 FTE

Introduction

The goal of EX-CELO is to augment the implementation of the consistent anelastic/compressible COSMO-EULAG dynamical core (DC) within the operational version of the COSMO framework. Complementary to the CELO Priority Project, this effort embraces the data assimilation capability of COSMO-EULAG, as well as aims at extending the COSMO formulation to fully exploit the lowest model level at the surface specific to COSMO-EULAG A-grid formulation.

Motivation

The future very-high resolution weather forecasting of COSMO relies on both: robust dynamical cores allowing for accurate representation of convection interacting with steep mountain slopes, and computationally effective frameworks for integration of COSMO software. The former aspect is already addressed with the development of ICON-LAM and COSMO-EULAG dynamical cores, whereas the latter is realized with the GridTools/OpenACC port of COSMO/ICON software. In the broader context, substantial scientific progress is made within the PantaRhei (IFS-FVM) and ESCAPE H2020 projects led by ECMWF, benefiting the efforts of COSMO Consortium within the mentioned areas.

The newly funded “Numerical weather prediction for sustainable Europe” project, financed by the European Regional Development Fund within the Smart Growth Operational Programme 2014 - 2020, Measure 4.4. “Boosting human potential in R&D sector” through the First Team 1/2016 award of Foundation for Polish Science contributes to the aforementioned efforts. Thus it is justified to establish new COSMO priority project, complementary to the CELO priority project of COSMO, aiming at covering aspects that are not a part of the existing efforts but are of key importance to exploit COSMO-EULAG in operational production at COSMO weather services.

The goal of this priority project is to synthesize efforts at the leading edge of research and development in high-resolution numerical weather modelling. In particular, the project aims at further development of COSMO-EULAG numerics, especially concerning enlarging the permitted timestep and improving the convergence of the elliptic solver. Also, it aims at formulating alternative diffusion operator conformant with the A-grid formulation of EULAG as well as the ILES property of MPDATA schemes. Last but not least, from the forecasting perspective, the project addresses the data assimilation capabilities of COSMO-EULAG based on the outcomes of KENDA-O project.

Actions proposed

• Full implementation of the EULAG surface level in COSMO framework with new strategies of the implicit (ILES) diffusion realization (Task 1)
• Enhance the efficiency of COSMO-EULAG numerical algorithms (Task 2)
• Interoperability of COSMO-EULAG with KENDA-O (Task 3)

The tasks will be carried out in parallel starting from the beginning of the project.

Tasks, deliverables and participants

Task 1: Adaptation of COSMO framework to embrace 0-m level of EULAG DC

EULAG dynamical core employs A-grid formulation of prognostic variables, which results in the lowest mass point at the surface (0 m), where all scalar and vector variables are defined and prognosed. This brings both scientific and technical challenges. While standard COSMO formulation also defines model variables at the surface for the turbulent diffusion package, they are not subject to the fully prognostic treatment (consistently with the model domain where actual integration in time is conducted). Moreover, standard COSMO formulation assumes that wind is vanishing there, which is only an option for EULAG dynamical core (non-zero wind is also physically meaningful when finite-volume perspective is taken into the account). However, admitting non-zero wind at the surface may require adaptation of the ICON turbulence package to waive zero-surface wind approximation therein.

It is possible to raise the lowest mass point of EULAG above the surface (albeit as a price of introducing approximation in the form of vanishing vertical velocity – resulting from the co-located velocity components on A-grid), This configuration has already been successfully (although preliminarily) tested during COSMO-EULAG development with CELO project. This approach, however, should be connected with reconsideration of the formulation of the surface boundary conditions for advection, implicit solver and diffusion, which is proposed within this Task.

The appropriate subtasks includes:

• systematic research on the realization of the turbulent and scalar fluxes from the surface within COSMO NWP setup. This involves conducting both: series of idealized experiments evaluating the effective transfer with several strategies of diffusion realization (1-D, 3-D, simple wall model + implicit MPDATA diffusion).. Define suitable strategy for operational use, based on objective forecasting performance in Alpine weather scenario
• adaptation, if feasible, of ICON physics ported to COSMO to accommodate the extra level at the surface and to provide meaningful tendencies needed for integrating atmospheric dynamics at 0 m level in time. This involves e.g. evaluation of water phase changes together with appropriate correction for the moist bouyancy term of EULAG DC and turbulent mixing coefficients.
• implementation of the correct boundary condition for diffusion at the surface, depending on the choice of the lowest model level height in EULAG
• adaptation of the COSMO I/O and communication layer, if feasible
• adaptation of COSMO interpolators to accommodate 0 m level, if feasible.

The aim of the EX-CELO effort is to provide formulation of recent COSMO/ICON physics that is fully compatible with the COSMO-EULAG capability to solve atmospheric dynamics at the 0 m level and to assess its significance for weather prediction. It will be realized within the frame of the PhD thesis of a new IMGW employee.

Deliverables:

(09.2017) Custom diffusion operator at the surface for COSMO-EULAG

(09.2018) Intercomparison of several numerical strategies to realize turbulent transport from the surface

(03.2019) Adaptation of ICON physics to fully exploit 0 m level

(09.2019) Technical adaptation of COSMO environment to accommodate surface level

Task 2: Enhance the efficiency of COSMO-EULAG algorithms

For competitive computational performance of COSMO-EULAG, it is needed to investigate new strategies concerning the existing explicit three-dimensional advection and efficiency of the iterative solver. To address this aspect, the following actions are planned in concert with IFS-FVM development at ECMWF:

• implementation of the horizontal multigrid preconditioning of iterative solver
• implementation of the time-adaptive timestep capability
• assessment of the possibility of advection substepping, implicit vertical advection or other numerical technique (e.g. semi-lagrangian approach, lagrangian vertical coordinate) allowing for long timesteps

Deliverables:

(09.2018) Multigrid preconditioner for COSMO-EULAG

(09.2018) Adaptive timestep capability for COSMO-EULAG

(03.2019) Alternative advection strategies for COSMO-EULAG

Task 3: Data assimilation capability of COSMO-EULAG

COSMO-EULAG needs to be coupled to the existing LETKF data assimilation provided by KENDA-O. As of January, 2018, COSMO-EULAG is being successfully tested with the standard COSMO nudging scheme without any modifications.

The following subtasks are proposed:

• investigation of the existing KENDA implementation towards the different set of prognostic variables of COSMO-EULAG, evaluation of the consistency of hydrostatic approximation in COSMO-EULAG and COSMO-RK in the context of DA
• implementation of necessary modifications and testing of LETKF on a sample meteorological scenario

Deliverables:

(09.2018) Report on compatibility of KENDA implementation with COSMO-EULAG formulation

(09.2018) COSMO-EULAG with LETKF data assimiliation

WG1 will give support by providing advice (without FTE).

Risks

Timely progress of this priority project relies on synchronization with and the timely progress of CELO priority project. Task 1 will be realized by the new IMGW employee who is not fully recruited yet and his/hers capabilities are not completely clear. Task 3 is in a large part a “terra incognita” and is connected with a necessity to gain novel (in the context IMGW) expertise.

References

Smolarkiewicz P.K., W. Deconinck, M. Hamrud, C. Kuhnlein, G. Mozdzynski, J. Szmelter and N. P. Wedi (2016) A finite-volume module for simulating global all-scale atmospheric flows, J. Comput. Phys., 314, pp.287-304, doi: 10.1016/j.jcp.2016.03.015

Smolarkiewicz, P.K., C. Kuehnlein, N.P. Wedi (2014), A consistent framework for discrete integrations of soundproof and compressible PDEs of atmospheric dynamics. Journal of Computational Physics 263, 185 – 205