[RegCNET] Bug fixing release 4.3.5.2

Jian-Hua (Joshua) Qian jqian at iri.columbia.edu
Fri Jan 18 03:02:54 CET 2013


Hi,

Where can I download the new version of the RegCM4.3 code and data?

thanks!
Joshua Qian

----- Original Message -----
From: "Graziano Giuliani" <ggiulian at ictp.it>
To: "regcnet‎" <RegCNET at lists.ictp.it>, regcm-news at gforge.ictp.it
Sent: Wednesday, January 16, 2013 8:31:56 AM
Subject: [RegCNET] Bug fixing release 4.3.5.2

The 4.3.5.2 bugfix release has been posted in the Files Section.

Fixed in this relase incrementally from the 4.3.5.1:

- Fix chem_icbc for PGI compiler.

An allocatable type was offending the PGI compiler in writing aerosol boundary 
conditions. Ahmed Shalaby first spotted this, the allocation has been changed in 
a fixed maximum dimension.

- Fixed MP_XX , ECENS and FNEST input subroutines for ECHAM6 MPI CMIP5 GCM, 
ECMWF seasonal ensemble, and regcm nesting.

The change of vertical interpolation subroutines used in the sigma2p and some 
icbc input layer got screwed up in 4.3.5 because of the different vertical 
ordering used in the two programs. This is a regression on 4.3.4 and has been 
fixed to allow both top to bottom and bottom to top ordering in all vertical 
interpolation subroutines.

- Fixed FVGCM binary input layer. Fixed GISST input layer.

The 4.3.5 change in the icbc program from real(4) to real(8) broke from 4.3.4 
some input layer where data where not stored in a packed dataset. In particular 
Yafang Zhong provided us a patch for the GISST input dataset spotting the 
problem. This has lead to a check of all input layer, and, upon discovery of an 
old US FVGCM dataset, to a rewrite of the FVGCM input layer. More, on some 
machines the gfortran built code refuses to open data not writable without 
explicitly setting the action='read' argument in the open. This has been 
implemented. Again, the usage of the ioparam namelist stanza and ibytes 
parameter can be avoided using the implicit inquire fortran function for iolenght.

- Fixed FNEST input layer for new CORDEX output naming convention

Another needed change after the variable naming convention to CORDEX archive 
specification. The FNEST was searching for u,v,t,qv and not ua,va,ta,qas. The 
new variables are now the default, but the compatibility with old names has also 
been kept.

- Small Fix in chem diagnostic output.

The sign in the diagnostic tendency from cumulus removal and boundary conditions 
effect was inverted, and the emission diagnostic tendency output was wrongly 
decoupled from the surface pressure. Dry deposition also in the pollen case was 
calculated regardless of the value of the ichdrdepo flag.

Ciao!

Graziano.

-- 
Graziano Giuliani - Earth System Physics Section
The Abdus Salam International Centre for Theoretical Physics
Strada Costiera, 11 - I - 34151 Trieste Italy
_______________________________________________
RegCNET mailing list
RegCNET at lists.ictp.it
https://lists.ictp.it/mailman/listinfo.cgi/regcnet


More information about the RegCNET mailing list