Datasets for radiation emulator in a cloud resolving model

Training sets used in the following reference to the RRTMG-K radiation emulator in a cloud resolving model

The structure of input and output variables is slightly changed from the GRL 2020 study.

Full datasets, without sub-sampling, are provided. 

3s_full.i and 3s_full.o are minimum and maximum values for input and output, respectively.  

See more details in readme file. 

Reference: Roh, S., and H.-J. Song, 2020, Evaluation of neural network emulations for radiation parameterization in cloud resolving model. Geophysical Research Letters, 47, e2020GL089444, doi:10.1029/2020GL089444