Skip to content

How To Install The Model

Jump to bottom
Luiz Flávio edited this page Oct 21, 2021 · 1 revision

BRAMS 6.0 - INSTALL GUIDE

Before You install the model BRAMS-6.0 You must have the prerequisites installed. See the document BRAMS_6.0_-_INSTALL_PREREQUISITES (extension md or html) to see how to install the requisites.

  1. Building PATHS and linking compilers.

    In order to compiler the BRAMS code using the compilers and libraries You install in prerequisites make the commands bellow.

    export PATH={YOUR_DIR}/bin:$PATH
export LD_LIBRARY_PATH={YOUR_DIR}/lib:$LD_LIBRARY_PATH
sudo ln -s /usr/bin/gfortran {YOUR_DIR}/bin/gfortran
sudo ln -s /usr/bin/gcc {YOUR_DIR}/bin/gcc

    Notice: See the prerequisites document to use the correct {YOUR_DIR} .

    Please, check if Your path have in first part {YOUR_DIR} and if the correct library is in first part of LD_LIBRARY_PATH.

    echo $PATH
echo $LD_LIBRARY_PATH

    Please check if Fortran and gcc version is the correct.

    gfortran --version
gcc --version

    The results must be something like. See that in this case we use 8.4.0.

    $ gfortran --version
GNU Fortran (Ubuntu 8.4.0-4ubuntu1) 8.4.0
Copyright (C) 2018 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
$ gcc --version
gcc (Ubuntu 8.4.0-4ubuntu1) 8.4.0
Copyright (C) 2018 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

  2. Configure the model

    Now You must configure the model passing all libreries You will be use. The {YOUR_BIN_AREA} is the folder you want to put the binary and namelists of model.

    ./configure --program-prefix=BRAMS_6.0 --prefix={YOUR_BIN_AREA} --enable-jules    --with-chem=RELACS_TUV --with-aer=SIMPLE --with-fpcomp={YOUR_DIR}/bin/mpif90    --with-cpcomp={YOUR_DIR}/bin/mpicc --with-fcomp=gfortran --with-ccomp=gcc --with-netcdff={YOUR_DIR} --with-netcdfc={YOUR_DIR} --with-wgrib2={YOUR_DIR}

Bellow an example of use, where {YOUR_BIN_AREA} is /home/oscar.

./configure --program-prefix=BRAMS_6.0 --prefix=/home/oscar --enable-jules    --with-chem=RELACS_TUV --with-aer=SIMPLE --with-fpcomp=/opt/gnu8/bin/mpif90    --with-cpcomp=/opt/gnu8/bin/mpicc --with-fcomp=gfortran --with-ccomp=gcc --with-netcdff=/opt/gnu8 --with-netcdfc=/opt/gnu8 --with-wgrib2=/opt/gnu8

  1. Make and Make install

    The make command will create the brams-6.0 executable. After creation it is necessary to run the make install command so that the basic files for the run are copied to the area set in {YOUR_BIN_AREA}

    make
make install

    The output of BRAMS's model is presented with colors in terminal. If You save the output log (by ">&" use) may be a problem if You want to edit the file. Is possibles to make a filter to extract all special symbols used to color the output. To make the filter just

    make filter
make install-filter

    BRAM's models needs a lot of input files. One of them is the Initial and boundary conditions (IC). This files is created from a global model data. Two diferents data input must be created, from NCEP model GFS or from ERA5 data. To create the IC data You must create the PRE-BRAMS utility.

    make pre-brams
make install-pre-brams

  2. Download tables, fixed files an a test case

    A series of fixed files and tables are required to run the model. You can get them by downloading the compressed file available in the ftp area. This file has 16GB and may take a while to download. It depends on your network speed. We advise you to check the checksum of the file to make sure that the download has not broken it.

    cd {YOUR_BIN_AREA}
cd ..
wget http://ftp.cptec.inpe.br/pesquisa/bramsrd/BRAMS-6.0/test_set_data/MD5SUM
wget http://ftp.cptec.inpe.br/pesquisa/bramsrd/BRAMS-6.0/test_set_data/brams6.0_test_bin.tar.xz
md5sum brams6.0_test_bin.tar.xz
cat MD5SUM
tar -xvf brams6.0_test_bin.tar.xz

  3. Running a small meteorological case

    The BRAMS Model is a limited area model and therefore requires input data from outputs from a global model. In the test case presented here, these data are already ready and available. For "expert" cases a new document shows how to get this data and generate it. The BRAMS model runs in 3 distinct phases:

    i) MAKESFC - Creates the surface files using the input files that contain global them and adjusts them to the specifications chosen by the user and arranged in the input namelist. The BRAMS namelist is called RAMSIN.

graph TD;
    NDVI[(ndvi)]-- datafix_model/NDVI_MODIS_GRADS -->BRAMS;
    TOPOGRAPHY[(topography)]-- datafix_model/topo1km -->BRAMS;
    SST[(sst)]-- datafix_model/sst_week -->BRAMS;
    VEGETATION[(vegetation)]-- datafix_model/GL_OGE_INPE -->BRAMS;
SOIL[(soil)]-- datafix_model/GL_FAO_INPE -->BRAMS
RAMSIN(RAMSIN_SFC)-.->BRAMS
style RAMSIN fill:#f9f,stroke:#333,stroke-width:4px
BRAMS-- dataout/SFC -->TOPFILES[(top_OQ3g)]
BRAMS-- dataout/SFC -->SFCFILES[(sfc_OQ3g)]
BRAMS-- dataout/SFC -->SSTFPFX[(sst_OQ3g)]
BRAMS-- dataout/SFC -->NDVIFPFX[(ndv_OQ3g)]
Loading

ii) MAKEVFILE - From the global data the phase creates the files with the boundary conditions and initial conditions interpolating them to the specifications chosen by the user and arranged in THE RAMSIN.

graph TD;
    IC[(IC)]-- datain/GRADS -->BRAMS;
    RAMSIN(RAMSIN_VFL)-.->BRAMS
style RAMSIN fill:#f9f,stroke:#333,stroke-width:4px
BRAMS-- dataout/IVAR -->IVAR[(ivar)]
Loading

iii) INITIAL - This is the phase of model time integration where prediction data will be produced as specified in RAMSIN.

graph TD;
    RAMSIN(RAMSIN_INI)-.->BRAMS
    TABLES(tables...)-. tables -->BRAMS
    style RAMSIN fill:#f9f,stroke:#333,stroke-width:4px
    style TABLES fill:#bbf,stroke:#f66,stroke-width:2px,color:#fff
    TOPFILES[(top_OQ3g)]-- dataout/SFC -->BRAMS
    SFCFILES[(sfc_OQ3g)]-- dataout/SFC -->BRAMS
    SSTFPFX[(sst_OQ3g)]-- dataout/SFC -->BRAMS
    NDVIFPFX[(ndv_OQ3g)]-- dataout/SFC -->BRAMS
    IVAR[(ivar)]-- dataout/IVAR -->BRAMS
    UMID[(GL_SM.GPNR)]-- datain/UMID -->BRAMS
    VAR(variables.csv)-.->BRAMS
    BRAMS-- dataout/POST -->POST[(brams-meteo...grads)]
    style POST fill:#2bf,stroke:#ff0,stroke-width:2px,color:#ff0
    BRAMS-- dataout/HIS -->HIS[(history)]
    BRAMS-- dataout/ANL -->ANA[(analisys)]
Loading

WARNING: The first two phases (i and ii) should only be rotated with 1 processor!

The test case is small enough to run on a laptop or desktop with 3 processors and with 8GB memory. It covers an area as shown in the figure below.

This area is configured in RAMSIN with 60x60x45 points (x,y,z), 20 km of horizontal resolution (surface) and is centered at 22.6S, 45.2W. The test only allows it to run for 24 hours of forecast. After the end of the round, one can verify the behavior of the model by comparing the results that were produced and are in {YOUR_BIN_AREA}/dataout/POST with those stored in the {YOUR_BIN_AREA}/dataout/POST_COMPARE area. To look at the outputs (GRADS/CTL format) you need to know how to use grads software

WARNING: Make sure your environment is tuned to use the packages installed in the pre-requirements (PATH and LD_LIBRARY_PATH) To look at the outputs (GRADS/CTL format) you need to know how to use grads software..

cd {YOUR_BIN_AREA}
mpirun -np 1 brams-6.0 -f RAMSIN_SFC_MID
mpirun -np 1 brams-6.0 -f RAMSIN_VFL_MID
mpirun -np 3 brams-6.0 -f RAMSIN_INI_MID

A series of logs will be shown on the screen in each of the 3 phases. Be aware of errors reported in these logs. Often there are non-existent files that have not been properly downloaded in the download or the system has some limitation. In the INITIAL phase you can follow the integrations by looking at the screen. It shows each timestep performed and other important information. See for example the figure below that shows part of these messages.

  1. How to see the data outputs

    The output of model as You read above is in GRADS file. May be used another formats but we will show how in expert users guide. If You are using Ubuntu Linux or some Debia derivated Linux You can install grads in a simple way, using apt-get command

    sudo apt-get install grads

    To install grads right from source, please, see the information on site GrADS Downloads. Pay attention on the necessary libraries explained on GrADS Supplibs

    To learn about grads we recommend to read the document Grads Manual from NCEP.

  2. How to filter the log output

    Supposed You run the model and put the output in a file. See example:

    cd {YOUR_BIN_AREA}
mpirun -np 3 brams-6.0 -f RAMSIN_INI_MID >& output.txt

    If You open the file output.txt using a specific editor (for example: vi) You will see all escape characters used for color the output. Example bellow ( a small peace of output log read by vi editor):

    +--------------+--------------+
|  variable    |    grid 1    |
+--------------+--------------+
^[[30m^[[47m|NNXP          |          0060|^[[0m
^[[0m|NNYP          |          0060|^[[0m
^[[30m^[[47m|NNZP          |          0045|^[[0m
^[[0m|NXTNEST       |          0000|^[[0m
^[[30m^[[47m|NSTRATX       |          0001|^[[0m
^[[0m|NSTRATY       |          0001|^[[0m

    This may appear a little bit confused. To solve the problema You must only run the filter (created in item 3 above) that comes with BRAMS model. Try

    filter output.txt

    And then, You can see the same file with this look

+--------------+--------------+
|  variable    |    grid 1    |
+--------------+--------------+
|NNXP          |          0060|
|NNYP          |          0060|
|NNZP          |          0045|
|NXTNEST       |          0000|
|NSTRATX       |          0001|
|NSTRATY       |          0001|

  1. How to generate Initial and Boundary conditions (IC) from global data

    The program "pre" created in item 3 above is the program that get data from global input files and converted them to IC for BRAMS. First of all You must edit a namelist file named pre.nml. See the example bellow

    $ARGS_INPUT

!!!!! DATE !!!!!

init_year  = 2020,
init_month = 08,
init_day   = 18,
init_hour  = 0,
final_year = 2020,
final_month= 08,
final_day = 19,
final_hour = 00,

!!!!! TIME STEP !!!!!!!

step  = 6, !Timestep in hours

!!!!! ATMOS !!!!!
atmos_type   = 1, !0=DP, 1=GFS Grib2, 2=ERA5
atmos_prefix ='gfs.t00z.pgrb2.0p25.f',
atmos_sufix  ='.2020081800.grib2',
atmos_idir   ='./datain/GFS/',
levels = 23,
initial_latitude = -70., !initial latitude for domain of model (-90 to 90)  
final_latitude  = 20., !final latitude for domain of model(-90 to 90)
initial_longitude = 250., !initial longitude for domain of model (0 to 360)
final_longitude = 358., !Final longitude for domain of model (0 to 360)

!!!!! CHEM !!!!!! 
chem_type     = 1, !0 = no Chem, 1 = CAMS 
chem_idir  = "./datafix_cams/",
chem1_prefix ='',
chem1_sufix  ='-CAMS-EC-2010-2019-AMS',

!!!!! OUTPUT !!!!!
out_type   = 2, !0=text, 1=VFM, 2=Grads
out_prefix = 'IC',
out_sufix  = '',
out_dir    = './datain/GRADS/',

$END

    The namelist is prepared for test case You have downloaded in item 4. After You edit it You must run the pre.

    ./pre

    A lot of information will be show in screen and some files with IC will be write in the out_dir You fill in namelist._

      $ls datain/GRADS -latr
  rw-r--r--   1   lufla   lufla      2 KiB   Mon Sep 13 14:23:59 2021    IC2021011318.ctl 
  rw-r--r--   1   lufla   lufla      2 KiB   Wed Oct  6 10:15:40 2021    IC2020081800.ctl 
  rw-r--r--   1   lufla   lufla    297 MiB   Wed Oct  6 10:15:41 2021    IC2020081800.gra 
  rw-r--r--   1   lufla   lufla      2 KiB   Wed Oct  6 10:15:43 2021    IC2020081806.ctl 
  rw-r--r--   1   lufla   lufla    297 MiB   Wed Oct  6 10:15:44 2021    IC2020081806.gra 
  rw-r--r--   1   lufla   lufla      2 KiB   Wed Oct  6 10:15:46 2021    IC2020081812.ctl 
  rw-r--r--   1   lufla   lufla    297 MiB   Wed Oct  6 10:15:47 2021    IC2020081812.gra 
  rw-r--r--   1   lufla   lufla      2 KiB   Wed Oct  6 10:15:49 2021    IC2020081818.ctl 
  rw-r--r--   1   lufla   lufla    297 MiB   Wed Oct  6 10:15:51 2021    IC2020081818.gra 
  rw-r--r--   1   lufla   lufla      2 KiB   Wed Oct  6 10:15:53 2021    IC2020081900.ctl 
  rw-r--r--   1   lufla   lufla    297 MiB   Wed Oct  6 10:15:54 2021    IC2020081900.gra

    This files will be used as inut for BRAMS´Model.

  2. Download GFS Data for model

    The BRAMS comes with a script in bash that get data from GFS from NCEP site. To get the data just run the script passing yera, month, day and output folder

    ./getGFS.sh 2021 10 02 ./datain/GFS/

    In the example above gfs data is downloaded for 02Oct2021 and the data will be put in ./datain/GFS/ directory.

Have Fun with BRAMS-6.0!

Clone this wiki locally