Globes in Figure 3: quickplot_track.py saved as CUMULO.ps and pred.ps and one of the .eps files
Figure 7/8 (Joint densities): dfplots.py saved as joint*.eps in the "fancy" section
Figure 9 (CloudSat map): CS_class_map.py saved as CS_allctypes.png
Figure 10 (prediction maps): ESACCI_regional.py saved as *_allctypes.svg
Figure 11 (time series): ESACCI_regional.pysaved as *_temporal.svg
Figure 12 (differences): ESACCI_regional.pysaved as *_diff.eps, using the twodegreegrid flag
###
Fig. S1 (IResNet boxplots): compare_pkls.py saved as source_varbox_byclass.png
Fig. S2 (RF boxplots): dfinteract.pysaved as stats.eps
Fig. S3a/S3b (prediction maps): ESACCI_regional.py saved as _Ci.eps/_Cu.eps
Fig. S3c/S3c (variable maps): inputmaps.py saved as _relice.eps/ptop.eps
Fig. S4a/b/c (feature importances): permimportance.py testpm.eps/trainpm.eps/fimp.eps
Fig. S4d (correlation): dfinteract.pysaved as ctcorr*inputset.eps
Fig. S5 (cell resolutions): coarsecompare.py saved as *coarsecomp.svg
Fig. S6/S7 (map temporal): ESACCI_regional.py *_CLOUDTYPE.eps for instantaneous and daily mean data
Fig. S8 (monthly mean): ESACCI_regional.py *allctypes.svg for monthly mean data