NMC_MET_TOOLS

NMC-MET Verification Project

• Provide basic verification products for station and gridding observation.
• Provide Object-Oriented space verifications products.
• Experimental verification products for real-time QPF-operation in National Meteorological Center.
• Numerical Models: ECMWF/GFS/WARMS/GRAPES/PWAFS
• Python-based project running on Linux-platform.

1 Setup MET_TOOLS using docker-container or local-compiled programms

main executable tools used from MET_TOOLS:

• POINT_STAT for station verification
• GRID_STAT for grid verification
• ENSEMBLE_STAT for ensemble forecast verification
• MODE for Object-based space verification
• WAVELET

2 Setting-up observation and forecast data in NETCDF-format

• Micaps r24-88 precipitation data
• CMISS Micaps r1jm data ?
• CMISS QPE data
• CMORPH QPE data from dss050.1

Download single-file: wget -c ftp://ftp.cpc.ncep.noaa.gov/precip/CMORPH_V0.x/RAW/0.25deg-DLY_00Z/2018/201806/CMORPH_V0.x_RAW_0.25deg-DLY_00Z_20180602.gz Download directory: wget -c -r -np -k -L -p ftp://ftp.cpc.ncep.noaa.gov/precip/CMORPH_V0.x/RAW/0.25deg-DLY_00Z/2018/201807/

• reforming CMISS- or CMORPH- QPE data to NETCDF4-format using ascii2nc by python-script

3 Setting-up precipitation forecast data from several models in NETCDF-format

• ECMWF
• NCEP
• ECMWF_EPS
• GRAPES-GFS/MESO/3KM
• WARMSII, RMAPES, etc

4 Verification Products and post-progress

• Traditional verification scores
• MODE-generated postscript
• Wavelet decomposition for grid-verification
• Ensemble verification
• Setting up met-viewer(plan)

5 Experimental MODE verification products for ECMWF precipition forecast initialized on 2018062712UTC

6 Experimetnal Rscript-generated figs from MET_TOOLS

• Rscript plot_mpr.R # for *mpr.txt
• Rscript plot_cnt.R # for _cnt.txt *.stat
• convert convert -density 300 mpr_plots.pdf mpr_plots.png # format-transition using ImageMagick command-tools

7 Setup met-viewer for MET_TOOLS output

• setup met-viewer container

git clone https://github.com/NCAR/container-dtc-metviewer

#From container-dtc-metviewer/METViewer, build METViewer image.

cd ./METViewer

# setup environmental variants in .bashrc

docker build -t metviewer . # may take a long time

export MYSQL_DIR=/path/for/mysql/tables 

export METVIEWER_DIR=/path/for/metviewer/output 

export METVIEWER_DATA=/path/for/data

#From container-dtc-metviewer, start the containers:

• #It opens up a shell in the docker environment and point to METViewer home directory

   docker-compose run --rm --service-ports metviewer
  

• #It starts the containers in the background

   docker-compose up -d 
  

  #To open a shell in the docker environment

   docker exec -it metviewer_1 /bin/bash
  

• Still have bugs: can not linkup with MySQL - VSDB data.