Microphysics experiments for neural network radiation scheme
Authors/Creators
- 1. National Institute of Meteorological Sciences, Korea Meteorological Administration
Description
Reference: Song, H.-J., & Kim. P. S. (2022). Effects of cloud microphysics on the universal performance of neural network radiation scheme. Geophysical Research Letters, 49, e2022GL098601. https://doi.org/10.1029/2022GL098601
Source codes (WRF model) for the 2D idealized squalline simulation using the neural network (NN) radiation scheme based on the real-case training
Control.tar.gz: Control run using the RRTMG-K radiation and WDM7 microphysics schemes
NN.tar.gz: NN radiation emulator (60-fold faster than the control run)
NN+CP.tar.gz: NN radiation emulator along with the Compound Parameterization (CP) that allows return to the RRTMG-K when the predicted error of heating rate exceeds a certain threshold (three times slower than the emulator)
For the modification of microphysics scheme, user should change "mp_physics" in "test/em_squall2d_x/namelist.input in the WRF model. In this study, we considered "mp_physics" of 26, 2, 5, 6, 7, 8, 9, 10, 11, 13, 16, 17, 19, 28, and 50.