MOVEROPLOT

Moveroplot is a tool for the production of various verification graphics including line graphs, maps, histograms, and reliability diagrams based on verification results produced by Movero.

Installation

This section describes the installation for users. If you intend to contribute, follow the instructions in CONTRIBUTING.md.

1. Clone a specific code tag of the repository:

git clone --depth 1 --branch <tag> https://github.com/MeteoSwiss-APN/movero-plot

2. Create a virtual environment:

cd movero-plot
./tools/setup_env.sh

This will create a conda environment named moveroplot (change with -n) and install the pinned runtime and development dependencies in requirements/environment.yml.

3. Activate the environment and install package:

conda activate moveroplot
pip install --no-deps .

4. The executable is now installed as

   $CONDA_PREFIX/bin/moveroplot

To determine the executable, the Movero script movero_plot uses

    $(conda info --base)/envs/moveroplot/bin/moveroplot

Usage

The primary command for moveroplot follows the structure

moveroplot [OPTIONS] MODEL_VERSION(S)
moveroplot [OPTIONS] MODEL_VERSION(S)

To see the available options, run:

moveroplot -h

The following options are listed below:
Options:
  --plot_type TEXT            Specify the type of plot to generate: [total,
                              time, station, daytime, ensemble].
  --debug                     Add debug comments to command prompt.
  --lt_ranges TEXT            Specify the lead time ranges of interest. Def:
                              19-24
  --plot_params TEXT          Specify parameters to plot.
  --plot_scores TEXT          Specify scores to plot.
  --plot_cat_params TEXT      Specify categorical parameters to plot.
  --plot_cat_thresh TEXT      Specify categorical scores thresholds to plot.
  --plot_cat_scores TEXT      Specify categorical scores to plot.
  --plot_ens_params TEXT      Specify parameters to ensemble plots.
  --plot_ens_scores TEXT      Specify scores to ensemble plots.
  --plot_ens_cat_params TEXT  Specify categorical parameters to ensemble
                              plots.
  --plot_ens_cat_scores TEXT  Specify categorical scores to ensemble plots.
  --plot_ens_cat_thresh TEXT  Specify categorical scores thresholds to
                              ensemble plots.
  --input_dir PATH            Specify input directory.
  --output_dir TEXT           Specify output directory. Def: plots
  --colors TEXT               Specify the plot color for each model version
                              using matploblib's color coding
  --relief                    Add relief to maps.
  --grid                      Add grid to plots.
  -V, --version               Show the version and exit.
  -v, --verbose               Increase verbosity; specify multiple times for
                              more.
  -h, --help                  Show this message and exit.

moveroplot efficiently processes user inputs to construct a plot_setup dictionary, which is pivotal in organizing the plotting process. This dictionary is structured with two primary keys: 'model_versions' and 'parameter'.

• 'model_versions': This key maps to a list encompassing the model versions to plot.
• 'parameter': This key connects to a nested dictionary. Within this nested structure, each parameter serves as a key linked to its corresponding scores to plot.

To offer a clearer understanding, the image below illustrates the potential parameters and their associated scores and their thresholds:

In the subsequent stages, plot_setup is channeled into distinct plotting pipelines. There, the source files are retrieved, parsed and plotted. Ultimately, all plots are saved in the <output_dir>/ directory as PNG files.

Usage Examples

Example Command plotting Station, Time, Total and Daytime Scores:

moveroplot C-1E_ch/C-2E_ch --lt_ranges 07-12,19-24,61-72 --input_dir /scratch/osm/movero/wd/2022s4 --plot_type station,time,daytime,total
--plot_cat_params TOT_PREC12,TOT_PREC6,CLCT,T_2M,TD_2M,FF_10M,VMAX_10M6
--plot_cat_thresh 0.1,1,10:0.2,1,5:0.2,0.5,2:2.5,6.5:0,15,25:0,15,25:-5,5,15:-5,5,15:2.5,5,10:2.5,5,10:5,12.5,20:5,12.5,20
--plot_cat_scores FBI,MF/OF,POD,FAR,THS,ETS
--plot_params TOT_PREC12,TOT_PREC6,TOT_PREC1,CLCT,GLOB,DURSUN12,DURSUN1,T_2M,T_2M_KAL,TD_2M,TD_2M_KAL,RELHUM_2M,FF_10M,FF_10M_KAL,VMAX_10M6,VMAX_10M1,DD_10M,PS,PMSL
--plot_scores ME,MMOD/MOBS,MAE,STDE,RMSE,COR,NOBS

Example Command plotting Ensemble Scores:

moveroplot C-1E_ch/C-2E_ch --lt_ranges 07-12,19-24,61-72 --input_dir /scratch/osm/movero/wd/2022s4 --plot_type ensemble
--plot_ens_params TOT_PREC12,TOT_PREC6,CLCT,T_2M,TD_2M,FF_10M,VMAX_10M6
--plot_ens_scores OUTLIERS,RANK,RPS,RPS_REF
--plot_ens_cat_params TOT_PREC12,TOT_PREC6,CLCT,T_2M,TD_2M,FF_10M,VMAX_10M6
--plot_ens_cat_thresh 0.1,0.2,2.5,0,0,2.5,5
--plot_ens_cat_scores REL,RES,BS,BS_REF,BSS,BSSD,REL_DIA

Plotting Pipeline and Output

Plotting Multiple Model Versions

moveroplot offers the option to visualize multiple results from distinct model versions within a single plot or image, depending on the plot type. This can be achieved through the use of specific delimiters: a slash (/) signifies combined plotting, while a comma (,) indicates separate plots.

Example:

Input: C-1E_ch/C-2E_ch,C-1E_alps

Interpretation: Display results of C-1E_ch and C-2E_ch into one combined plot. Plot the results of C-1E_alps separately.

Spatial Verification

Relevant File: station_scores.py

Note: Each station score image is consistent in its LT range. The number of plots per image can vary (model versions along columns, scores along rows).

Note: Invalid Atab files are ignored.

Time Series of Verification Scores

Relevant File: time_scores.py

Remark: The order specified in --plot_scores and --plot_cat_scores is crucial. Two plots are assigned per page. Each threshold and parameter initialize a new page.

Remark: Model versions and scores can be displayed in the same plot using / in the input.

Diurnal Cycle of Verification Scores

Relevant File: daytime_scores.py

Remark: The order specified in --plot_scores and --plot_cat_scores is crucial. Two plots are assigned per page. Each threshold and parameter initialize a new page.

Remark: Model versions and scores can be displayed in the same plot using / in the input

Total scores for all lead times

Relevant File: total_scores.py

Remark: The order specified in --plot_scores and --plot_cat_scores is crucial. Four plots are assigned per page. Each threshold and parameter initialize a new page.

Remark: Model versions and scores can be displayed in the same plot using / in the input

Ensemble scores

Relevant File: ensemble_scores.py

Remark: The order specified in --plot_ens_scores and --plot_ens_cat_scores is crucial. Each threshold and parameter initialize a new page.

Remark: Model versions and scores can be displayed in the same plot using / in the input.

Remark: RANK, REL_DIA and line plots are saved in separate images.

Regular Line Plots

RANK

Reliability Diagram