8000 GitHub - gardenia111/WRF-DL
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content

gardenia111/WRF-DL

 
 

Repository files navigation

This github is the code implementation corresponding to the article "WRF-ML v1.0: A Bridge between WRF v4.3 and Machine Learning Parameterizations and its Application to Atmospheric Radiative Transferr"

Please contact x7zhong@gmail.com if you have any questions about this code.

Code description:

Build_WRF

Contains the libraries necessary for building WRF.

Build_WRF/LIBRARIES/dl-inference-plugin

Contains the library necessary for building and running WRF_DL

dl_inference

Contains the files used for building the library in dl-inference-plugin, and the python script to run ML models within WRF_DL

dl_inference/model

ML-based radiation emulators are saved.

WRF

Contains the files that are different from the original WRF v4.3 files and used for implementing the WRF-ML coupler. To run the WRF coupled with ML-based radiation schemes, you need to add files that did not exist or overwrite the existing WRF files.

example

Contains the examples about how to call Python script from Fortran code.

Instructions

Setting environment variables

add the path to the WRF ML inference library for example: ##WRF DL INFERENCE LIB export WRF_DL_INFER_DIR=$DIR/dl-inference-plugin export LD_LIBRARY_PATH=$WRF_DL_INFER_DIR/lib:$LD_LIBRARY_PATH

modifications to the WRF configure.wrf

After running the ./configure when compiling WRF, modify the WRF configure.wrf file as: add -L${path_to_dl-inference-plugin/lib} and "-lDL_inference_plugin" following LIB_EXTERNAL image

ML-based parameterization

Add ML-based parameterization schemes' related modules in the main/depend.common file, e.g.: image

add infer_init in the main/wrf.F to initialize the services of ML model inference

profiling mode (optional)

This part is used for profiling

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Fortran 87.1%
  • Python 6.7%
  • Makefile 3.3%
  • Perl 1.4%
  • Shell 1.1%
  • C 0.4%
0