This is a DNN written with PyTorch to Emulate the gravity wave drag (GWD, both zonal and meridional ) in the WACCM Simulation.
Sample output data from WACCM emulation. It is 3D global output from the WACCM model, on the original model grid.
However, the demo data here is one very small part of the WACCM output due to storage limit of Github. NN trained on this Demodata will not work.
load 3D WACCM data and reshaping them to the NN input.
train.py train the files and generate the weights for NN.
NN-pred.py load the weights and do prediction.
replace original GWD scheme in WACCM with this emulator.
a. the emulator can be trained offline
b. training the emulator online