A Single Column Model version of HIRLAM

1
A Single Column Model version of HIRLAM

• Javier Calvo

2
Motivation

• The have a reference SCM version of HIRLAM for
  research and validation of model physics.
• The idea is to maintain a code very close to the
  reference complete HIRLAM or even included within
  it.
• Including different case studies to be used as
  reference cases.
• Trying to gather as many users as possible within
  HIRLAM community.

3
Background

• Currently several SCMs are used by Hirlam
  researchers.
• Its difficult to maintain them updated with the
  reference model.
• Comparisons are always uncertain due to the
  different set ups.
• Version most widely used is the one developed by
  B. Sass and previously used for the turbulence
  intercomparisons.
• Complete model change significantly due to the
  inclusion of ISBA surface scheme (tile approach).
• New cases coded for some versions from EUROCS
  cloud project.

4
The new Single Column Model

• Case set up and philosophy from B. Sasss 1D
  model lifted to the new SCM.
• Dinamics follow quite closely the complete HIRLAM
• SCM versions of some routines included.
• 3D structure of the arrays kept.
• This facilitates updates from the complete model.
• The SCM can be better used for debugging and
  optimization.
• Maybe the model could be extent to 2 point model
  or 2D model.

5
How to run the model

• Simple steering unix script
• Makefile
• Control through namelist
• Output in ASCII format
• In principle plotting bu gnuplot or grads
• Input data in ASCII format
• Source versions based on RCS

H1d/EXP/bin/Run1d,Plot Data/iniARM_Cu.dat Grdy1d p
hys
6
Calling tree dynamics
reads namelists
reads departure data
vertical interpolation to model levels
dynamical tendencies
prescribed tendecies if needed
Postprocesing and printinds
7
Calling tree
8
Calling tree
9
Select physics routines with NOPTION()
Probably more options than in the complete HIRLAM
CBR turbulence NOPTION(2)2
10
Send diagnostics from physics to printing routine

• Deviation from the complete model physics trough
  directives H1D
• SCM physics shoud work without modications in the
  complete HIRLAM

CBR turbulence
i fdef H1D endif
11
Clouds and Convection STRACO
STRACO NOPTION(2)1
12
Clouds and Convection KF/RK
KF/RK NOPTION(2)4
CBR turbulence NOPTION(2)2
13
SCM cases studies
Dry boundary layer 1.- Convective BL with some
shear (case WC05 in Ayotte et al., BLM, 1996) 4.-
Pure convective case with no wind (Ayotte,
1995) 5.- Lepzig case thermally neutral and
shear driven with strong winds 6.- Intermittent
turbulence in a thermally stable BL layer (from
SABLES) BL clouds 2.- Sc case from GCSS case 3
(Duynkerke, 1998) 8.- Diurnal cycle of Sc over
sea (EUROCS) 3.- Shallow Cu over sea from BOMEX
(Siebesma and Holtslag, 1996) 7.- Diurnal cycle
of Shallow Cu over land (ARMEUROCS) Deep
convection 10.- Diurnal cycle of Deep Cu over
land (ARMGCSS 3a case) 9.- Diurnal cycle of
Deep Cu over land (ARMidealized case from EUROCS)

• 0.- Initialization and dynamical tendencies
  from the complete HIRLAM model.
• xx.- Sensitivity of Deep Cu to environmental
  humidity (idealized case from EUROCS)
• xx.- Long integration (at least 30 days) from
  TOGA COARE

14
To be done

• Coding of all case studies
• Inclusion of new physics schemes under
  development
• moist turbulence and statistical clouds (Sander
  volunteers)
• Tiedtke convection with statistical clouds?
• Documentation of the SCM and the cases.
• Basic plotting tool?
• Based on gnuplot/grads?
• Inclusion of observations and LES results for
  comparison.

15
Feedback

• Feedback and help is welcome to improve the SCM
  and to achieve a wide use of it.

16
The end