Test Case 4: Toga-coare deep convection

This is only 2D test case and the only one to require a forcing file, force...dat. And, like test case 3, it requires the radiation code. This page provides the background information on test case 4 and examples of run of test case 4 on various machines.

Background information (from Jon Petch)

The TOGA-COARE observational campaign took place over the tropical west Pacific warm pool between November 1992 and February 1993. Deep convection is prevalent in this area and the experiment was able to provide comprehensive datasets of convective systems over that period. These datasets have been used to to run CRMs and hence provide diagnostics which can be used to evaluate convective parametrization schemes in SCMs.

This case is based on the CRM sensitivity tests of Petch and Gray (submitted to QJ, 2000+) who studied the effect of different microphysics parametrizations, dimensionality, domain size, resolution and radiation on the 20-26 Dec 1992 period selected by the GCSS Working Group 4. The standard case provided here is run in 2-D with interactive radiation and three-phase microphysics. Single moment representations are used for rain, snow and graupel, and double moment for ice. The effects of large-scale vertical motion are represented by temperature and humidity forcings derived form the TOGA-COARE observations, and the sea surface temperature is also varied in line with the observations. This particular case is run with a 1km horizontal grid spacing in a 256km wide domain. The example of typical mean profiles from the 6 day integration shown below can be compared with the results presented in Petch and Gray.

Typical 6 day mean profiles from the TOGA case. Typical 6 day time series from the TOGA case.

This simulation is a well studied example of tropical deep convection, and tests the model's use of interative radiation with multi-phase microphysics, along with its use in a 2-D configuration.

Files required to run test case 4

The files needed to run this test case can be found in /home/lem/VER2.3/TEST_CASES/TEST4_2.3 and the table below indicates which files you'll need for which machines

	Turing (T3E - no longer available)	Department machines	Origins (i.e. wren, fermat or green)	Newton
run...f	runTEST4_2.3.f	runTEST4_2.3-m.f	runTEST4_2.3-w.f	runTEST4_2.3-w.f
end...f	endTEST4_2.3.f	endTEST4_2.3.f	endTEST4_2.3.f	endTEST4_2.3.f
nmlsetup...	nmlsetupTEST4_2.3-t	nmlsetupTEST4_2.3-m	nmlsetupTEST4_2.3-w	nmlsetupTEST4_2.3-w
nmlchain...	nmlchainTEST4_2.3	nmlchainTEST4_2.3	nmlchainTEST4_2.3	nmlchainTEST4_2.3
model...	modelTEST4_2.3	modelTEST4_2.3	modelTEST4_2.3-n	modelTEST4_2.3-n
force...dat	forceTEST4_2.3	forceTEST4_2.3	forceTEST4_2.3	forceTEST4_2.3

Runs for Test Case 4

Computer	Run No	Version	PEs	IIP, JJP, KKP	Time Taken	Results
Newton	136	2.3	1	1, 256, 60	8.85 hrs	Mean Profiles Time series
Origins	32	2.3	1	1, 256, 60	17.6 hrs	Mean Profiles Time series
GRID	53	2.3	1	1, 256, 60	~211 hrs (8.8 days)	Mean Profiles Time series
T3E at Met Office	N/A	2.2	1	1, 256, 60	unknown	Mean profiles Time series
Turing	114	2.2	1	1, 256, 60	unknown	Mean profiles Time series