Characterizing Convection Schemes Using Their Responses to Imposed Tendency Perturbations

This repository contains the data and scripts required to reproduce the results of the paper "Characterizing Convection Schemes Using Their Responses to Imposed Tendency Perturbations" submitted to the Journal of Advances in Modeling Earth Systems (JAMES).

Brief description of project

An SCM intercomparison project based on the Linear Response Function framework of Kuang (2010), where we examine the temperature and moisture responses to small convective tendency perturbations.

Participating Single-Column Models (SCMs): WRF, LMDZ, CNRM, UM, SCAM

Convection schemes tested:

1. WRF Kain-Fritsch
2. WRF New-Tiedtke
3. WRF New-Simplified-Arakawa-Schubert
4. WRF Betts-Miller-Janjic
5. WRF Zhang-McFarlane
6. CNRM PCMT scheme
7. UM Simplified-Betts-Miller
8. UM Mass-Flux scheme (Gregory & Rowntree)
9. SCAM Zhang-McFarlane
10. LMDZ modified-Emanuel scheme + Cold pool formulation

Repository structure

• /data directory: contains the data from the SCMs, in csv format
  • /data/[model_name]/ directories: contain the data of the individual models. There are four sub-directories (five for WRF) for each model:
    • REF directory: contains data for RCE mean state, including Temperature (T) and Relative Humidity (RH)
    • matrix_X_raw directory: contains raw data for the T and q responses to dT/dt and dq/dt perturbations
    • matrix_M_inv directory: contains the post-processed (normalized and standardized) M-1 matrix data
    • response_profiles directory: contains the post-processed response profiles (vertical column) for perturbation at two levels (850 and 650 hPa)
    • (for WRF only) pbl_mp_sensitivity directory: data for WRF PBL and MP sensitivity tests. There are three sub-directories in this folder:
      • mean_states directory: contains data for the RCE mean state sensitivity to PBL and MP schemes
      • response_profiles directory: contains data for the sensitivity of T and q responses to PBL and MP schemes
      • response_profiles_non_idealized directory: contains data to compare sensitivity of T and q responses to PBL and MP schemes between idealized and non-idealized setups
• /scripts directory: contains the python scripts to post-process and plot the figures in the paper
  • /scripts/plot_inidividual_matrix/ directory: contains scripts to plot the M-1 matrix and the 2-levels response profiles for individual SCMs
    • plot_matrix.py : script to post-process and plot the matrix and response profiles for selected SCM using the raw data in the matrix_X_raw folders. Option available to save the (post-processed) outputs as csv files in the matrix_M_inv and response_profiles folders of the selected SCM
  • /scripts/plot_figures/ directory: contains scripts to plot the Figures in the paper
    • plot_rce_mean_states_all_models.py : script to plot Figure 1 (RCE mean state T and RH of all SCMs)
    • plot_anomaly_profiles.py : script to plot Figure 2 - 3 (vertical profiles of T and q responses for selected SCMs)
    • plot_matrix_all_models.py : script to plot Figures 4 - 7 (M-1 matrices of all SCMs)
    • plot_rh_q_correlation_matrix.py : script to plot Figure 8 (correlation matrix of RH vs. q')
    • plot_pbl_mp_sensitivity_mean_states.py : script to plot Figures 9 - 10 (sensitivity of RCE mean state to PBL and MP schemes)
    • plot_pbl_mp_sensitivity_responses.py : script to plot Figures 11 - 12 (sensitivity of T and q responses to PBL and MP schemes)
    • plot_pbl_mp_sensitivity_responses_non_idealized.py : script to plot Figures A1 - A2 (sensitivity of T and q responses to PBL and MP schemes comparison between idealized and non-idealized setups)