Code for analyzing lower tropospheric stability on the NCAR supercomputing system. It provides tools to access datasets stored on glade, and requires that the user be running code on a computer with access to glade. This package provides the code required to reproduce the results of my paper.
Data used:
- CESM1-CAM5
- CESM1-WACCM
- CESM2-CAM6
- CESM2-WACCM
Planned
- ERA-Interim
- ERA5
- CESM Large Ensemble
The first step is to create a Parameters object for each of the configurations desired. The Parameters object is passed around to various parts of the code, and keeps track of begin/end times, source dataset, lat/lon subsets, and variables needed.
source = 'cesm1-cam5' # or cesm1-waccm, cesm2-cam6, cesm2-waccm
params = lts.Parameters(source)
To see which variables are available, you can use the method params.list_available_variables(). (not yet implemented)
You can see the currently selected parameters by calling
params.display().
Choose a save location large enough to store the subsetted data. I made directories
for each of the sources, so I use
params.save_location = '/glade/scratch/dwatkins/' + params1.source + '/'
The variables parameter must be a list, and has the near-natural-language names of variables in it.
params.variables = ['air_temperature', '2m_temperature', 'surface_downward_longwave', 'eastward_wind', 'northward_wind', 'condensed_ice_path', 'condensed_water_path', 'sensible_heat_flux', 'sea_level_pressure']
Next, by calling get_data(), datasets are read from glade, subsetted, and saved locally.
Within get_data or load_data, I want lts to get computed. There could also be an option to warn if overwriting existing data. I need to test all the variable options.
Next steps:
- build a function to load locally saved data
- make sure I can load in the cloud data
- make functions for histograms and timeseries analysis
- make sure all the 4 datasets can load