Climatic Data

[lorismichel]

How to access the scripts of verena:

• open a linux terminal of your choice (the windows terminal doesn't do)
• type "ssh -XYC yourusername@fog.ethz.ch"
• to do computation: being on fog, type "ssh -XYC yourusername@so4.ethz.ch"

Run some sample script:

• cd /home/bverena/python/scripts/
• activate python environment by executing:
• source startup
• conda activate keras_env or conda activate missing_env (try)
• execute my script by running: python plottest.py

a figure should pop up with the gap-free data that I am using (for a sample month) in the first row and the corresponding data with the deleted data in the second row. if nothing pops up, please look at the end of the script for instructions (you can view the script using vim by typing "vim plottest.py" and ":q" to quit vim again). you can set the month to look at and the fraction missing in the script as you like.

other files:

• "era_missing_xr.py" the script to run a full iteration of deleting and gapfilling the data and saving both to disk -"imputethis.py" containing the class and all the methods for a general iterative imputation.

How to access the data

• ssh username@fog.ethz.ch
• cd /net/exo/landclim/data/dataset/ERAS_deterministic/recent/0.25deg_lat-lon_1m/original`

files are .nc (use package raster in R or python, maybe Verena is sending us some example code)

Soil moisture corresponds to swvl1

• produce a loading file
• check a few variables and try to plot them on a world map
• search for missing data, access the mask

[lorismichel]

• @lorismichel producing some R code with raster to load data from nc files.
• @JeffNaef studying the folders (how many variables etc)
• @spohnm open a new overleaf document with graphs + questions to Verena