Welcome to the ncreper wiki!

Running ncreper.bin

The ncreper.bin program can be used to preprocess NetCDF files and create new fields like wind speed from wind components, rain rate from accumulated rain, and return periods.

The ncreper.bin program is in the nctools module which can be loaded on the met.no ppi using

  module load nctools

The binary reads input from stdin. For instance:

  cd ncreper/work
  ncreper.bin < ncreper.in

Program input

The input format is based on headers, data and commands, and may look like this:

  # ?
  # exit
  #
   NCREPER V1.0 [0]
  #
   INPUT FILE (NETCDF) [1]
      input.nc
  #
   OUTPUT FILE (NETCDF) [1]
      output.nc
  #
   NUMBER OF DATA SLICES [1]VFL
      10
  #
   RETPER FILE (NETCDF) [*]
    meps_return_level_annual.nc
  #
   ZERO RETPER LEVEL [1]
      0.0D0
  #
  # IMPORT VARIABLES FROM RETPER FILE [*]
  #  land_area_fraction
  #
  #
   SPEED FROM VARX, VARY, MAX [1]VFLR
      wind_speed_10m      x_wind_10m       y_wind_10m        1000.0
      wind_speed_of_gust  x_wind_gust_10m  y_wind_gust_10m   1000.0
  #
   NEW VAR, OLD VAR, DIFFERENTIATE OVER HOURS, MIN, MAX [*]
      precipitation_60min     precipitation_amount_acc 1   0.0D0   1.0D15
      precipitation_120min    precipitation_amount_acc    120/60  0 99999
      precipitation_180min    precipitation_amount_acc    180/60  0 99999
      precipitation_360min    precipitation_amount_acc    360/60  0 99999
      precipitation_720min    precipitation_amount_acc    720/60  0 99999
      precipitation_1440min   precipitation_amount_acc   1440/60  0 99999
  #
   RETPER DIM[1]
      return_period
  #
   NEW VAR, OLD VAR, RETPER VAR [*]
      wind_speed_10m_return_period          wind_speed_10m          wind_speed_10m_return_level
      wind_speed_of_gust_return_period      wind_speed_of_gust      wind_speed_of_gust_return_level
      precipitation_60min_return_period     precipitation_60min     precipitation_amount_acc_1h_return_level
      precipitation_180min_return_period    precipitation_180min    precipitation_amount_acc_3h_return_level
      precipitation_360min_return_period    precipitation_360min    precipitation_amount_acc_6h_return_level
      precipitation_720min_return_period    precipitation_720min    precipitation_amount_acc_12h_return_level
      precipitation_1440min_return_period   precipitation_1440min   precipitation_amount_acc_24h_return_level
  #
  # PERCENTILE DIM, PERCENTILE [1]
  #   ensemble_member 95
  #
  # NEW VAR, OLD DIM VAR
  #    precipitation_60min_return_period_p95    precipitation_60min_return_period
  #    precipitation_60min_p95                  precipitation_60min

  #
   KEEP VARS [*]VFLR
      surface
      height0
      height7
      time
      forecast_reference_time
      x
      y
      latitude
      longitude
      projection_lambert
      projection_regular_ll
      land_area_fraction
  #
      precipitation_60min
      precipitation_60min_return_period    
      precipitation_180min
      precipitation_180min_return_period    
      precipitation_360min
      precipitation_360min_return_period    
      precipitation_720min
      precipitation_720min_return_period    
      precipitation_1440min
      precipitation_1440min_return_period    
  #
  #    precipitation_60min_p95
  #    precipitation_60min_return_period_p95
  #
      wind_speed_10m  wind_speed_10m_mb0
      wind_speed_10m_return_period wind_speed_10m_return_period_mb0
      wind_speed_of_gust  wind_speed_of_gust_mb0
      wind_speed_of_gust_return_period wind_speed_of_gust_return_period_mb0

Comment-lines are preceded by #. The square brackets indicate how many lines the data body should consist off, the trailing & indicates that the header is mandatory, and the VFLRM indicates the data preprocessor options.

This program is very memory intensive. If you run out of memory, the program will stop, typically when reading a variable. In this case you have to request more memory from the PPI, or you have to slice the data more (NUMBER OF DATA SLICES [1]VFL).

You may import environment variables using the import command, for instance

  import file

and later use the variable in the data body, for instance

  #
   INPUT FILE (NETCDF) [1]VFLR &
     data/$file
  #
   OUTPUT FILE (NETCDF) [1]
     output/nora3_$file
  #

Note that variable substitution is on available for headers with the V option enabled.

Local installation

To install ncreper locally, you first have to log onto the PPI,

  ssh ppi-clogin-b1

The you have to download the source code from github and compile

  git clone https://github.com/FrankThomasTveter/ncreper.git
  cd ncreper
  make
  ls ncreper/*.bin

Updating the nctools module on PPI

The ncreper.bin software is part of the nctools module. You may install a new module by creating a new module file and directory in,

  cd /modules/centos7/user-apps/nctools
  cp -rf 0.34 XXX
  # cp   ~/ncreper.bin   XXX/bin/
  cd /modules/MET/centos7/user-modules/nctools
  cp 0.34 XXX
  emacs XXX

where XXX is the new version you have created.