9/26 Meeting Notes

[panalluri]

• Choose a single region
• Pass the whole region into a CNN (FCNNs don't mind the size of image)
• Switch from NN to CNN
• Do not use halo method yet
• Do not use dictionaries --> build xarray datasets instead (label region and type of input var with numbers for now, and later use word names to index)
• Replace monthly averages with every 30th snapshot from the large dataset
• Define Q_coarse and Q_star appropriately
• Train and test on single region with CNN
• Use Zotero for paper readings
• Read one more paper this week (Guillaumin, A. P., & Zanna, L. (2021). Stochastic-deep learning parameterization of ocean momentum forcing. Journal of Advances in Modeling Earth Systems, 13 (9), e2021MS002534.)
• Use method in paper above for training dataset creation
• Create IDP (fill out form, and ask Dhruv for feedback)