IRIE (Ivan Ristić Improved ETA) is a numerical weather prediction model, developed from previous models that have their roots in Belgrade in the former Yugoslavia. In the 70s of the 20th century, the first modern computer code for solving atmospheric motion equations (Limited Area Primitive Equation Model - LAPEM) was written by Fedor Messinger and Zavisa Janjić. A period of improvement followed and the Hydrometeorological Institute and Belgrade University model (HIBU) was created, and then the highly regarded ETA model. ETA model became officially operational at NCEP on 8 June 1993. At the end of the 20th century, the development of the IRIE model, based on the basic equations of the ETA model, began. Over the next 20 years, work on improving the IRIE model was almost daily with real-time feedback from meteorologists and continuous support from Antonio Sanò (iLMeteo-Italy) and Jesper Thiesen (ConWX-Denmark). It was necessary to reach the quality of the official models, which were worked on by large teams of scientists and programmers. IRIE model followed all the latest scientific developments, as reported by its successful users such as iLMeteo, ConWX, EMD, Femax, Weathernews and Vestas. As part of the development of the IRIE model, the most significant achievements are related to clouds and precipitation, i.e. cloud microphysics and convection.

IRIE MODEL START-UP MANUAL

Download IRIEv1.0.0_weather2.tar.gz (317 MB) and extract it with following command: tar -xzpf IRIEv1.0.0_weather2.tar.gz
NOTE: The next step can be skipped, except for the part to set the number of processors to be used, since the model domain is set to Europe. It is recommended that you first compile the model to ensure all necessary files are created, then change the model domain and recompile.

How to set the model domain and number of processors to be used for the IRIE model:
In the file IRIE-1.0.0/eta/install/set_parmeta set the size of the model domain with the change of IM and JM
(Note: JM MUST be an odd number!)
Set the number of CPUs to be used with INPES and JNPES
INPES X JNPES = total number of CPUs to be used
In the file IRIE-1.0.0/eta/namelist/ETAIN.in update ONLY these model domain parameters: TLM0D, TPH0D, DLMD, DPHD, DT
For more detailed instructions on how to set up the model domain, please read: http://etamodel.cptec.inpe.br/~retamod/Download/model/SHORT_HOW2RUN_EtaModel.pdf
You can also use Domain Wizard: https://esrl.noaa.gov/gsd/wrfportal/DomainWizard.html

IRIE model postprocessing uses the COPYGB program for GRIB conversion from rotated E-GRID to regular LAT-LON.
This is easy to do with COPYGB, but it will require you to define the grid to be used on the command line.
Edit the file IRIE-1.0.0/eta/runs/outjob_special.in if you need to change the command line according to your model domain:
set gridlat='255 0 401 201 30000 -20000 128 63200 46400 166 166 64'

How to compile an IRIE model:
Go to the IRIE-1.0.0/eta/install directory
Edit the set_flags file if you need to change compiler options (itype_fort, MPI_FC, MPIRUN_ARG...)
to use PGI as compiler set itype_fort=1
to use GFORTRAN as compiler set itype_fort=2
to use IFORT as compiler set itype_fort=3
MPIRUN_ARG is set for OpenMpi, if using MPICH set MPIRUN_ARG = ""
Before a new compilation, run the ./clean_all script.
To compile the IRIE model, run the ./KOMPAJLIRAJ_SVE script.
The compilation process will take several minutes depending on the compiler you are using.
If IRIE-1.0.0/eta/exe contains these files, the compilation process was successful:
• copygb.x
• etafcst_all.x
• etapost_new.x
• etatopo.exe
• initbc.exe
• initbc_new.exe
• quilt.x
If not, look for some errors in IRIE-1.0.0/eta/install/eta_install.log

How to include the time-dependent GFS data needed to run the IRIE model:
The IRIE model uses GFS data that must be downloaded externally with a forecast time step of 6 hours.
Current GFS data can be found at the following link: https://www.ftp.ncep.noaa.gov/data/nccf/com/gfs/prod/
For an explanation of GFS file names, read: https://www.nco.ncep.noaa.gov/pmb/products/gfs/

Choose where you want and create a directory or directories to put the GFS files in: full_path_to_data_GFS/data_GFS.HH{00,06,12,18}
GFS files must be renamed like this: gblav.YYMMDDHHFH (grib1) or gblav2.YYMMDDHHFH (grib2)
where YY(year)MM(month)DD(day)HH(hour) is the start date and FH is the forecast hour.
Example: 00 UTC GFS data from 17.10.2022 and 12-hour forecast: gblav.2210170012 or gblav2.2210170012
NOTE: GFS files with equal HH are in the same directory

Go to the IRIE-1.0.0/newFTP directory
To set the path to the GFS files, run the following command: ./update_paramavn full_path_to_data_GFS/data_GFS. (point "." after GFS is important)
To run IRIE model preprocessing, run the gfs_GRIB1.sh or gfs_GRIB2.sh script depending on the grib files you are using.
The script requires arguments to describe the start date and the length of the forecast in hours like this: ./gfs_GRIB1.sh YY MM DD HH FH
Example: ./gfs_GRIB1.sh 22 10 17 00 24

How to run the IRIE model:
Go to the IRIE-1.0.0/eta/run directory
Edit the file common.sh if you need to change the number of processors (ncpus) or the length of the forecast in hours (model_length).
To run the IRIE model, run the ./KRENI script.

If the previous processes were done correctly, the final output from the model in grib1 format at an hourly timestep can be found in the IRIE-1.0.0/eta/runs directory:
latlon_000
latlon_001
latlon_002
...
These grib files can later be used to create different weather maps, also as boundary files for some other models such as wave and current models.

"Think Globally, Model Locally!"

Weather2/IRIE is licensed under the MIT License
Copyright (c) 2022 Weather2