Bogdan Rosa 1, Marcin Kurowski 1, Damian Wójcik 1, and Michał Ziemiański 1 Acknowledgements: Oliver Fuhrer 2, Zbigniew Piotrowski 1,3 1. Institute of Meteorology

Bogdan Rosa1, Marcin Kurowski1, Damian Wójcik1, and Michał Ziemiański1

Acknowledgements: Oliver Fuhrer2, Zbigniew Piotrowski1,3

1. Institute of Meteorology and Water Management

2. Meteo Swiss

3. National Center for Atmospheric Research

Recent progress in the anelastic branch

COSMO General Meeting, Rome, 5-9 September 2011

Outline


Testing of the EULAG dynamical core for realistic flows over the Alpine topography with simplified physics parameterization

Coupling of EULAG dynamical core with COSMO via an interface

Technical testing of the coupled model by idealized cases

Testing the EULAG core for Alpine flows - a convection case study


Task 1.3: Tests of the EULAG dynamical core for realistic flows over the Alpine topography with simplified physics parameterization

To simulate realistic flows over the Alpine topography, with resolution ranging from 2.2, 1.1 km to 0.55 km, applying simplified parameterizations of basic subgrid processes. Compare results between EULAG and COSMO simulations.

Moist processes are only simulated on explicit grid (i.e. no shallow convection parameterization, no moist turbulence) with a simple microphysics (Kessler-scheme) in both models. Turbulent diffusion with a one-equation (TKE)-model (not necessarily the same in both models) without interactive parameterization, but using the surface and radiation fluxes from independent COSMO 2.2 km model runs (or simplified surface fluxes)


Description of the experiments

Experiments involve case study of summer Alpine convection on 12 July 2006.

Simplified parameterizations:• Boundary layer processes are represented by TKE (turbulent kinetic energy) model.• Surface fluxes and drag taken from the operation run of COSMO2 model for Switzerland.• Simple representation of moist processes (warm rain Kessler-scheme)

Experiment setup:• Horizontal resolution 1.1 km, vertical resolution as in COSMO2• The computational domain is restricted to 234x198 km and covers the Southern Alps• Initial, boundary conditions from COSMO2 operational run

General meteorological situation in the Alpine region - 12 July 2006

MSG (Meteosat Second Genertion) 12:00 UTC

Synoptic situation in the area: slow-moving cold front in a shallow surface trough of low pressure

Synoptic map – 2:00 UTC, 12 July 2006

This is representative case study for summer (convective) situations.



Diurnal cycle of potential temperature at the ground. Results from COSMO and EULAG experiments.


Time evolution of cloud water and stream lines from EULAG simulation.

12:00 UTC14:00 UTC

16:00 UTC 18:00 UTC

Comparison of the EULAG simulation with satellite images

12:00 UTC

15:00 UTC

12:00 UTC

15:00 UTC

Temporal and spatial structure of the simulated convection in the EULAG experiment closely resembles the actual development


Realization of the CDC plan – coupling


CDC Project Plan

Task 1.6: Coupling of EULAG dynamical core with COSMO via an interface.

To better compare the behavior of the new dynamical core in more realistic model applications with full physics parameterizations, the dynamical core of EULAG will be coupled with the COSMO-model. As an intermediate step and to keep the amount of work in a reasonable range this will be done via an interface.

This means that the EULAG dyn. core keeps his own variables, data structure, etc. It is not the aim to have a very efficient code version at this stage but to have a useable model version.

EULAG in Fortran 90


Migration of F77 EULAG code into F90

Dynamic memory allocation (ALLOCATE/DEALLOCATE), no COMMON blocks, no DATA and BLOCKDATA statements

Modularization (data and source code separation, logical code decomposition into specialized modules)

Fortran 90 language syntax (free format syntax, upper/lower cases, names and comments in English, COSMO-like code indentations, new operators (>, <, etc.), KIND argument for real and integer types, no GOTO instructions,...)

C&E coupling – plugging the new dynamical core in


A new configuration of EULAG dyn. core is done via namelists.

EULAG dynamical core is temporarily plugged instead of default R-K dynamical core in src_runge_kutta module. It uses a set of variables defined in data_eufields module.

In src_runge_kutta:

Sections 1-3a: unchanged(initial time step,

boundary tendencies, slow tendencies fromradiation, convection,

Rayleigh damping, etc.)

Sections 4a-6: unchanged (advection and diffusion

of scalars, saturation adjustment)

Sections 3b-3cR-K time

integration

EULAGtime

integration

COSMO fields to EULAG

beginof a time step

endof a time step

EULAG fields to COSMO

C&E coupling – parallelization


MPI communication is basically organized in COSMO (init_procgrid). A new subroutine (emulate_geomset) is used to initialize EULAG's MPI variables following COSMO structure. This ensures identical domain decomposition in both models, i.e. the same grid points lie on the same processor domain.

Domain decomposition in EULAG Single processor neighborhood

C&E coupling – grid adaptation


COSMO staggered C-grid and EULAG unstaggered A-grid

There are at least two options for distribution of vertical levels: 1) interpolation from COSMO to EULAG levels2) 1:1 transformation for mass levels with interpolation of the velocities only (operational?)

COSMO (1) (2)EULAG

dz

dz

dz

dz/2

dz/2

n levels n+1 levels n+2 levels

dom

ain

heig

ht

C&E coupling – data flow


Mutual data transfer between COSMO and EULAG

When data flow is required?1. at the initialization stage (to initialize EULAG dynamical core)2. at each time step

R-K integr.

EULAGintegr.

COSMO fields to EULAG

EULAG fields to COSMO

Sections 1-3a.

Sections 4a-6.

C-grid(u

C,v

C,w

C,T, p, forcings)

A-grid(u

E,v

E,w

E,Th, forcings)

(uC,v

C,w

C,T)

Realization of the CDC plan – testing C&E


Task 1.7: Technical testing with COSMO by idealised cases

The correct coupling of the EULAG dynamical core into COSMO can be at first tested with the implemented idealized test cases. This testing can be performed ‘by a press of a button’ in COSMO. The staff is now well trained with the idealized tests (see task 1.1), therefore it is not necessary to perform an extended analysis of such idealized tests, but simply to check if any technical coupling problems occur.

Performed tests:1. inertia-gravity wave (Skamarock and Klemp,

1994)2. cold density current (Straka et al., 1993)

Two dimensional time dependent simulation of inertia-gravity waves


Skamarock W. C. and Klemp J. B. Efficiency and accuracy of Klemp-Wilhelmson time-splitting technique. Mon. Wea. Rev. 122: 2623-2630, 1994

Initial potential temperature perturbation

Setup overview:

domain size 300x10 km resolution 1x1km, 0.5x0.5 km, 0.25x0.25 km rigid free-slip b.c. periodic lateral boundaries constant horizontal flow 20m/s at inlet no subgrid mixing hydrostatic balance stable stratification N=0.01 s-1

max. temperature perturbation 0.01K Coriolis force included

Constant ambient flow within channel 300 km and 6000 km long

km

km

outletinlet 220

/1

)/sin()0,,(

axx

Hztzx

c

Initial potential temperature perturbation

smtw

smtv

smtu

/0)0(

/0)0(

/20)0(

Initial velocity


Results - gravity waves in a short channel

EulagC&E

EulagC&E

EulagC&E

Comparison with analytical solution


Eulag

C&E

Analytical

Profiles of potential temperature along 5000m height


'

C&E

Analytical

Gravity waves in a long channel


Eulag

C&E

Eulag

C&E

Eulag

C&E

Gravity waves in a long channel


C&E

C&E

C&E

Analytical

Analytical

Analytical


Profiles of potential temperature along 5000m height

C&E

Analytical


Two dimensional time dependent simulation of cold blob descending to the ground

)300( Kconst

Straka, J. M., Wilhelmson, Robert B., Wicker, Louis J., Anderson, John R., Droegemeier, Kelvin K., Numerical solutions of a non-linear density current:A benchmark solution and comparison International Journal for Numerical Methods in Fluids, (17), 1993

free-slip b.c.

open b.c.

periodic b.c.periodic b.c.

r)(rTT

Experiment configuration:

• isentropic atmosphere, θ(z)=const (300K)• periodic lateral boundaries• free-slip bottom b.c.• constant subgrid mixing, K=75m2/s• domain size 51.2km x 6.4km• bubble min. temperature -15K• bubble size 8km x 4km• no initial flow• integration time 15min


Distribution of potential temperature after 900 sec

C&E inviscid 100m

C&E viscous – diffusive forcing from COSMO parameterizations100m

Comparison of the potential temperature distribution


C&E

Eulag

Cosmo100m


C&E

Eulag

Cosmo

Comparison of the horizontal and vertical velocities obtained from three different models

100m

Eulag

C&E

Cosmo100m


C&E

Eulag

Cosmo

Comparison of potential temperature distribution at resolution 25 m

25m


C&E

Eulag

Cosmo25m

C&E

Eulag

Cosmo25m

Comparison of the horizontal and vertical velocities obtained from three different models

1. MOTIVATION CONCLUSIONS


Temporal and spatial structure of the simulated convection in the EULAG experiment closely resembles the actual development.

EULAG code has been successfully implemented in COSMO. Results of the idealized tests obtained using the hybrid E&C

model are in good qualitative and quantitative agreement both with reference and analytical solutions.

Small differences indicate the need for further testing and verification of the E&C code.

Dynamical core of the developed prototype, cooperates correctly with the diffusive forcing from COSMO parameterizations.

Realization of the CDC plan – documentation of the already performed tasks.


B. Rosa, M. J. Kurowski, and M. Z. Ziemiański, 2011, Testing the anelastic nonhydrostatic model EULAG as a prospective dynamical core of a numerical weather prediction model. Part I: Dry Benchmarks, Acta Geophysica, 59 (6), DOI: 10.2478/s11600-011-0041-1. M. J. Kurowski, B. Rosa and M. Z. Ziemiański, 2011, Testing the anelastic nonhydrostatic model EULAG as a prospective dynamical core of numerical weather prediction model. Part II: Simulations of a supercell, Acta Geophysica, 59 (6), DOI: 10.2478/s11600-011-0051-z. M. Z. Ziemiański, M. J. Kurowski, Z. P. Piotrowski, B. Rosa and O. Fuhrer, 2011, Toward very high resolution NWP over Alps: Influence of the increasing model resolution on the flow pattern, Acta Geophysica, 59 (6), DOI: 10.2478/s11600-011-0054-9

Documents

Bogdan Rosa 1, Marcin Kurowski 1, Damian Wójcik 1, and Michał Ziemiański 1 Acknowledgements: Oliver Fuhrer 2, Zbigniew Piotrowski 1,3 1. Institute of Meteorology