MITgcmtools-py
Hand crafted python tools to play with the MITgcm's output. Not a coherent package, but this might help others. All tools here were used in order to perform the data analysis for the upcoming paper "Vertical heat fluxes beneath idealized sea ice leads in large-eddy simulations: comparison with observations from the SHEBA experiment" by Bourgault, Straub, Nadeau, Duquette and Tremblay, to be published in JPO somewhere in 2020.
This repo and "package" contains three modules. I tried to stay general when writing them, but they might be quite specific to my configuration. Example : my model was cartesian, I haven't tested any of this with other geometries.
gendata.py
This file defines functions and a cli to help generate input data of simple shape for the MITgcm. The way I used it is the following: define a DataPreset, SurfacePreset or TimePatternPreset then call it through the gendata cli in a bash script that was called just before running mitgcmuv. I found this system more scalable than transporting a gendata.m file in each of my run folders, as the main shape generation is centralized and version-controlled.
I have written all available "presets" to work on normalized coordinates (0 .. 1). One can see the available presets by calling gendata list.
io.py
Defines helper functions to open full run folders in one call (through xmitgcm). Also, adds a functionality to read partial levels (when diagnostics is used with a levels field), but xmitgcm is in the process of implementing this.
quantities.py
Defines a lot of the diagnostics I computed from the raw output of the model (I was always using diagnostics). Also provides a cli called diagnose that is sadly undertested, but was created in order to do this computation direclty of the computation server, so to be able to copy only a subset of the data. "quantities" are registered with a special decorator so that diagnose is aware of them. Those that can use xgcm for coordinate handling. Finally, a power density spectrum calculation once was working but moving to a different configuration broke it and it hasn't be repaired. Sorry.
Other files
common.py defines a few things I use (or not) in the other files. jmd95.py replicates the Jacket and McDougall (1995) density formula that is used in the MITgcm. This file is directly a syntax-modified copy of the python function available in MITgcmutils that ships with the MITgcm.