List_of_Figures.txt

This list points to the script used for the presented Figures in the stochastic and deterministic ensmeble ML paper.

Main paper:

Fig1: summary schematic (no .py code needed)
Fig2: Median coefficient of determination R² of the vertical profiles of specific humidity and temperature tendency for ensembles and individual models --> deterministic_analysis/real_geo_R_2_all_networks_updt.ipynb
Fig3: Spread-Skill figure for specific humidity tendency, temperature tendency, cloud liquid water tendency and cloud ice water tendency of different ensembles --> uncertainty_quantification/real_geo_Spread_skill_diagrams_all_networks_updt.ipynb
Fig4: Probability Integral Transform diagram of different ensembles for cloud liquid water tendency in the planetary boundary layer   --> uncertainty_quantification/real_geo_PIT_composite_all_networks_updt.ipynb
Fig5: Mean continous rank probability score of ensembles for the vertical tendency profiles  --> CRPS_analysis/real_geo_CRPS_overview_updt.ipynb
Fig6: Latitutude - Longitude Map of CRPS of different ensembles for cloud ice water tendency in the upper troposphere --> CRPS_analysis/real_geo_CRPS_lat_lon_plots.ipynb
Fig7: Simulated zonal average of precipitation percentiles and histogram of different parameterizations inCESM2 simulations --> online_evaluation/global_analysis_online_ANN_lin_boost_precip_updt.ipynb
Fig8: Latitude-Longitude plots of peak of diurnal precipitation in Local Solar Time --> online_evaluation/global_analysis_online_ANN_lin_boost_precip_updt.ipynb


Supporting information:

FigS1: Median coefficient of determination R² of the vertical profiles of cloud liquid water and and cloud ice water tendency for ensembles and individual models --> deterministic_analysis/real_geo_R_2_all_networks_updt.ipynb
FigS2: Latitude - Longitude map of R² for specific humidity tendency in the plantary boundary layer for different ensembles and DNN 1 --> deterministic_analysis/real_geo_R_2_all_networks_updt.ipynb
FigS3: Median coefficient of determination R2 for the remaining 8 2D SP output variables of different ensembles and individual models 
variables --> deterministic_analysis/real_geo_R_2_all_networks_updt.ipynb
FigS4: Median mean absolute error (MAE) for vertical profiles of specific humidity tendency, temperature tendency, cloud liquid water tendency and cloud ice water tendency of different ensembles and individual models  --> deterministic_analysis/real_geo_MAE_all_networks_update.ipynb
FigS5: Median MAE for the remaining 8 2D SP output variables of different ensembles and individual models 
variables --> deterministic_analysis/real_geo_MAE_all_networks_update.ipynb
FigS6: Aggregated Continuous Rank Probability Score (CRPS) for different ensembles  --> CRPS_analysis/real_geo_CRPS_overview_updt.ipynb
Fig S7: Latitude - Longitude Maps of CRPS for cloud liquid water tendency in the plantary boundary layer of ensmebles --> CRPS_analysis/real_geo_CRPS_lat_lon_plots.ipynb
FigS8: Latitude - Longitude Maps of CRPS for surafce specific humidity tendency of ensembles  --> CRPS_analysis/real_geo_CRPS_lat_lon_plots.ipynb
FigS9: Latitude - Longitude Maps of CRPS for surafce specific temperature tendency of ensembles  --> CRPS_analysis/real_geo_CRPS_lat_lon_plots.ipynb
FigS10: Probability Integral Transform diagram of different ensembles for cloud ice water tendency in the upper troposphere   --> uncertainty_quantification/real_geo_PIT_composite_all_networks_updt.ipynb
FigS11: Probability Integral Transform diagram of different ensembles for surface specific humidity tendencies   --> uncertainty_quantification/real_geo_PIT_composite_all_networks_updt.ipynb
FigS12: Probability Integral Transform diagram of different ensembles for surface temperature tendencies   --> uncertainty_quantification/real_geo_PIT_composite_all_networks_updt.ipynb
FigS13: CRPS vs. magnitude of static latent space perturbation \alpha --> latent_perturbation_tuning/Tuning_static_gaussian_noise_VED_1.ipynb
FigS14: R² vs. magnitude of static latent space perturbation \alpha --> latent_perturbation_tuning/Tuning_static_gaussian_noise_VED_1.ipynb
FigS15: (1-R²)+PIT_distance vs. magnitude of static latent space perturbation \alpha --> latent_perturbation_tuning/Tuning_static_gaussian_noise_VED_1.ipynb
FigS16: Mean Root Mean Squared Error (RMSE) of specific humidity below 200 hPa of ensemble parameterizations and individual models in simulations with CESM2 --> online_evaluation/real_geo_ANN_boost_zonal_averages.ipynb
FigS17: Mean Root Mean Squared Error (RMSE) of temperature below 200 hPa of ensemble parameterizations and individual models in simulations with CESM2 --> online_evaluation/real_geo_ANN_boost_zonal_averages.ipynb
FigS18: Zonal mean temperature of ensemble parameterizations and benchmark paramaterizations in simulations with CESM2 --> online_evaluation/real_geo_ANN_boost_zonal_averages.ipynb
FigS19: Zonal mean specific humidity of ensemble parameterizations and benchmark paramaterizations in simulations with CESM2 --> online_evaluation/real_geo_ANN_boost_zonal_averages.ipynb
Fig20: Precipitation histograms of different parameterizations in simulations of CESM2 --> online_evaluation/global_analysis_online_ANN_lin_boost_precip_updt.ipynb
Fig21: Latitude - Longitude plot of median precipitation of different parameterizations in simulations of CESM2 --> online_evaluation/global_analysis_online_ANN_lin_boost_precip_updt.ipynb
Fig22: Selected regions for diurnal cycle analysis of precipitation in simulations of CESM2 --> online_evaluation/global_analysis_online_ANN_lin_boost_precip_updt.ipynb
Fig23: Diurnal cycle of precipitation in selected regions in simulations of CESM2 --> online_evaluation/global_analysis_online_ANN_lin_boost_precip_updt.ipynb