Hi,

Ok, I made a small mistake earlier, having the rigth domain in the bacground but the wrong cords on the display.
Anyway, from playing with rotated maps a bit I think that the rotation of -120 is not really necessary, -90 or -100 looks ok for me.
Sample:

Thanks! that image actually looks like the area that I'm trying to cover. Taking account all the suggestions, I change a little bit the domain making it a little bit width in Longitude in order to make that the Andes hill will be center and don't be forced by boundary conditions.

NSChile)
NLON=198 (=2^6*3+2*(NPBPTS+1) )
NLAT=260
NLEV=65
SOUTH=-41.00
WEST=-82.80
NORTH=-15.00
EAST=-63.00
POLAT=-90.00
POLON=0.0
NPBPTS=2
NBNDRY=8
GPHALO=8
NDTIME=600 # What this number means? Time step.
HDFSET=regular
HIRES=0.025
;;

Any suggestion or Comments.

I have other questions, but related with some test and variables names of the scripts in config-sh folder. The thing is that we wrote 4 config-sh files considering paths and that kind of things related with our cluster (The name is levque), but I have some questions about the name variables in our config.levque-gnu:

What is $SCRATCH folder? I understand that is like a root path to the climate files.

What does $HL_ARC folder contains? Is it the hdf climate files?

I'll try starting climate file generation with your latest domain definition.

$SCRATCH folder is the home of $HL_DATA, a runtime temporary directory. O.K., it should not be very temporary but allowing to files to stay e.g. for a couple of weeks, to allow studying log files and stuff. Also the compiled libraries are here, $HL_LIB should point to $HL_DATA/lib . Your experiment definitions and modified code should stay forever, thus they are all in $HL_WD under your home directory.

$HL_ARC is where the resulting grib files and other archived files are archived, so a long-term storage directory.

Perhaps Martynas could explain more, based on the Lithuanian cluster definitions.

Hi,

Well, Laura explained the purpose of HL_DATA -basically the main "working directory" and HL_ARC (Archive) already. HL_DATA and HL_ARC also contain the climate files (at least this is where you will put them after they're generated and after you receive them from ecgate).
You could also try to attach your config. file in this forum, if something doesn't work properly

the SCRATCH folder might be the part of a root path to both of your $HL_ARC or $HL_DATA or one of them e.g. in our system, $SCRATCH points to a separate data storage which is mounted to the platform. So it depends on your choice, it might as well be disks on the same platform etc.

Some fine-tuning - in HIRLAM grid box boundaries lay on the coordinates announced, not the centres, thus e.g. nlon is (east-west)/deltax + 1


NSChile)
NLON=198
NLAT=260
NLEV=65
SOUTH=-40.90
WEST=-82.70
NORTH=-15.00
EAST=-63.00
POLAT=-90.00
POLON=0.0
NPBPTS=2
NBNDRY=8
GPHALO=8
NDTIME=600
HDFSET=regular
HIRES=0.025
;;

You may need to make the time step 600 s smaller when you start running, as your grid size is not more than 11 km, and you are over the complex orography of Andes. Would guess something like 300 could work. But you can try: the run will crash if the time step is too large and you can restart with a smaller value.

Started creation of the first two climate files for December and January. Hope you have boundary files ready (what will you use there?) so you can try starting from these. I'll start more climate generation runs later. This procedure nowadays takes quite a lot of time, due to the new lake and orography fields we are preparing.

By the way, the Hirlam 7.4beta1 code will be tagged next week. You might perhaps try it instead of the 7.3. The only essential difference is the default resolution with 65 levels (you have that already) and FLake lake scheme. I prepare climate files so that the lake depth is included. However, to really use the lake scheme you need also initial values of the in-lake prognostic fields. For this, you would need a couple of additional climate files which I can provide if you are interested. No idea how good the values are for your domain, but they surely exist globally, as the lake depths also.

This is your domain, plotted surface elevation in metres. O.K.?

Thanks Laura! Well I do have some ECWF boundary conditions (but this are for 2000) certainly I'll use this as a test for Hirlam. As I mention I need probably Met fields for December to March of 2006. That Topography look fine for me and the dimensions seems correct.

I'm going to post my config.---- file just to discuss about it, and certainly I have doubts about how to run the model and check if it's working. There are kind of things that I did not get from
https://hirlam.org/trac/wiki/HirlamSystemDocumentation/Install/Installation

O.K., I'll create all climate files for this domain and copy them somewhere for you to pick them. Hope someone will help you with the configuration definitions.

Hi,

I just talked to Felix, seems that he had some simple problems related with perl-tk missing etc.
everything should be fine in some time.

Please download hirlam.fmi.fi/tmp/chicli.tar.gz and have a try. Just for training, you could try to draw (with grads?) some fields from there. Files cl*0000 are the 12 monthly climate files. Six files cl*0100* contain FLake cold start parameters, too, so in principle you can start with FLake if you start experiment exactly at those days, e.g. 1st of December, February ... Lakes.dat.gz contains lake depths for your domain. All files should reside in $HL_DATA (stored in $HL_ARC/ClimateFiles) during your experiment..

Thank you Laura!! I've got the climate files now, and Martynas is helping me how to get the appropiate configuration to run the code.

Now I need the boundary conditions (ECWF) in order to complete the run. I'd like to get this boundary conditions for December-January-February-March of 2006. Is it possible? Does anybody have this fields data? The domain for the run is the following:

NSChile)
NLON=198 (=2^6*3+2*(NPBPTS+1) )
NLAT=260
NLEV=65
SOUTH=-41.00
WEST=-82.80
NORTH=-15.00
EAST=-63.00
POLAT=-90.00
POLON=0.0
NPBPTS=2
NBNDRY=8
GPHALO=8
NDTIME=600 # What this number means? Time step.
HDFSET=regular
HIRES=0.025
;;

Any help or suggestions, will be very useful!

Laura Rontu wrote:
Laura and others,

just to clarify that with the recent HIRLAM versions the requirement of nlon being that of 2**k* 3**m * 5**n + 2*(npbdpts+1) is no longer there IF one choose in scripts/Env_expdesc the value of LHHITER as yes, (use of iterative Helmholtz solver as met.no had been applying for many years). I mentioned this in earlier post

hirlam.org/index.php?option=com_kunena&I...15&limitstart=15#253

I'll be back to address Felix's question about LBC in a later post.

Hi,

Just to update some things.
Felix now has a working demo setup runing on his cluster with gfortran+mpi like configuration with some demo boundaries for HL7 domain, also, the climate files for his domain are available (generated by Laura).
So the thing left is to use the climate files and the boundary files for his desired domain. The confguration setup itself is more or less ok.

Felix,

please report if you're still facing any problems with the experiment.

Felix Carrasco wrote:
Felix,

I've now prepared the ECMWF lateral boundary data for the first 10 days of your specified period, together with the observation data for the entire period, which are downloadable from
hirlam.org/portal/download/chilean11/lbc/lbc2005120.tar.gz
(ca 300 mb)
hirlam.org/portal/download/chilean11/obs/ob200512.tar.gz
(ca 24 mb)
hirlam.org/portal/download/chilean11/obs/ob200601.tar.gz
hirlam.org/portal/download/chilean11/obs/ob200602.tar.gz
hirlam.org/portal/download/chilean11/obs/ob200603.tar.gz
hirlam.org/portal/download/chilean11/obs/ob200604.tar.gz

After downloading these data, do gunzip to place the data in your $HL_DATA
directory.

Please also download suggested scripts in
hirlam.org/portal/download/chilean11/scripts/Env_expdesc
hirlam.org/portal/download/chilean11/scripts/Env_domain
hirlam.org/portal/download/chilean11/scripts/Climate

In Env_expdesc, I suggested ANALYSIS=NOUA to avoid going through upper air data assimilation, but you will perform surface analysis to update surface parameters including SST, that's why you need the observation data.

Please test using the data and if they are usable, I can prepare the LBC data for the rest of your episodes and place them at the same place.