[RegCNET] Bug fixing release 4.3.5.3

[ggiulian at ictp.it]

Hi all!

The 4.3.5.3 bugfix release has been posted in the Files Section.

Fixed in this relase incrementally from the 4.3.5.2:

- Typo in the OXBC filename in the model different from the one produced by chem_icbc preventing the AERO simulation run. The file produced by the chem_icbc is now OXBC, while the model was searching for OXCL. Reported by James Ciarlo'.

- Rounding of longitudes prevented grads helper programs from catching terrain output in BAND case. This annoyance was reported by Hernandez Garces Anel which has also tested all input files in the Testing directory (some of them were outdated). Thanks Anel!

- Fix for the restart of the lake model in BATS. This receives the fix proposed by Yafang Zhong on the mailing list: https://lists.ictp.it/pipermail/regcnet/2013/006412.html. Thanks Yafang!*

*- Fix an off by one error in the lake initialization. The error was in mod_param in setting the values for dhlake1 with indexes going for dot point external, whereas the matrix is now on cross points internal.

- Fix unit of measure of lake surface temperatures in the SUB file output. The SUB output now contains surface lake temperature. The netCDF file had units in K, the variable was in Celsius. Now it has been put in K.

- Fix iseaice model induced instabilities. The instabilities have been reported by Michal Belda on an European Domain extending to the North. The error was changing abruptly the sea ice depth to zero from one meter depth in the reading of boundary conditions. Now the model itself melts the last ice.

- Fix stability control numbers printout to have same numbers with different number of processors. The print of model control stability parameters in model run (first and second order mean derivatives of ps over the internal domain) was different with different number of processors for floating divide operation rounding. Now the total value is collected from all processors before dividing for the total number of grid points. Before each processor was dividing for the number of assigned points and the master node was summing up all numbers.

New Minor Features:

- Added an ismthlev to terrain parameters to allow user controlled level of smoothing of topography.

-Added user configurable Critical Richardson Number for land/ocean points in the Holtslag PBL option in the input namelist with the new holtslagparam stanza. Default is the previous value.

-Removed dependency on the order of stanzas in the input namelist file, some of the stanzas have been made optional (the model printout reports using default values for the parameters).

Ciao!

Graziano.