This directory contains matlab and shell scripts used to generate figures for: "A Randomized Dormant Ensemble Kalman Filter" by Mohamad El Gharamti which was submitted to Monthly Weather Review.
- Scalar 1D Cases:
- gen_Fig_1.m: A matlab script that generates the Bayesian solution for the RD-EnKF
- gen_Figs_3_4.m: This matlab script runs the 1D scalar example for both linear and nonlinear cases
- Lorenz-96 Attractor:
- run_l96_exps.sh: A bash shell script that runs all L96 experiments (designed for HPC architecture)
- gen_Figs_5_6_7.m: A matlab script to generate L96 results with perfect modeling conditions
- Time-series plot with and without adaptive inflation
- Localization sensitivity
- Ensemble correlations plot
- gen_Fig_8.m: Produces L96 results with model errors with and without adaptive inflation
- gen_Fig_9.m: Rank histograms for 3 different experiments
- Flood Prediction Using HydroDART:
- gen_Fig_10.m: Generates the stream network on the flooding domain including gauges and major cities and rivers
- gen_Fig_11.m: Generates hydrographs comparing EnKF and RD-EnKF at three different gauge locations
- gen_Fig_12.m: Generates CRPSS boxplot summary for all gauges
- HydroDARTdiags.m: An all purpose diagnostic function to HydroDART (generates hydrograph, maps, etc)
- gauges_2_indices_subset.m: A utility function to read gauge IDs and their associated DART indicies
- obs_increment_eakf.m: Computes observation increments for the analysis
- move_forward.m: Integrates the scalar model forward in time
- rh.m: Finds ensemble bins to plot rank histograms
RouteLink.nc containing network geometry for WRF-Hydro & all other figures Fig*.pdf
HydroDART (WRF-Hydro+DART) Python and Shell scripts using the main DART repo here.