Effects of nonlinear observation operators for visible and infrared radiances in ensemble data assimilation

Description

Dataset to accompany the first revision of the manuscript "Effects of nonlinear observation operators for visible and infrared radiances in ensemble data assimilation" for the QJRMS.

Contains - experiments (one folder per experiment)  each contains   - diagnostics (DART format) for each assimilation time    - obs_seq.final: assimilation output     - obs_seq.final-linear: contains linear posterior as "prior" ("evaluate" at +1s after analysis)    - obs_seq.final-evaluate: at analysis time (includes DART clamping), at 1s after analysis time: "nonlinear posterior"  - config file DART-WRF for each experiment (python format)

- code    assim_tools_mod.f90: DART code modification to compute the linear posterior

- other data:  see v1 of this repository!    nature run initial conditions (WRF format)    forecast ensemble initial conditions (WRF format)    script to read "obs_seq.final"-files into pandas.DataFrame (python format)

Experiment naming:VIS ... visible reflectance 0.6 micrometer assimilationWV73 ... infrared 7.3 micrometer assimilationobs10 ... observation density: 1 observation per 10x10 km of domain areaobs30 ... 1 observation per 30x30 kmloc10 ... localization radius: 10 kminf0 ... no prior nor posterior covariance inflation in the assimilationsec0 ... no sampling error correctionreject ... not assimilating observations with a small first-guess departure <= 0.03

Experiment names used in Figures

Section 3.1- Figure 2a: exp_v1.23_P2_rr+1_VIS_obsi211_loc20_inf0- Figure 2b: exp_v1.23_P2_rr+1_VIS_obsi360_loc20_inf0

Section 3.2.1 - Figure 3: exp_v1.23_P2_rr+1_VIS_obs30_loc14_inf0

Section 3.2.2- Figure 4: exp_v1.23_P2_rr+1_VIS_obs10_loc20, exp_v1.23_P2_rr+1_VIS_obs10_loc20_reject

Section 3.3- Figures 5a,c: exp_v1.23_P2_rr+1_T2M_obs10_loc20_inf0_sec0- Figures 5b,5d,7a: exp_v1.23_P2_rr+1_VIS_obs10_loc20_inf0_sec0- Figure 6a,6b,7b: exp_v1.23_P2_rr+1_WV73_obs10_loc20_inf0_sec0