TC-Pairs Tool: General

TC-Pairs Functionality

The TC-Pairs tool provides position and intensity information for tropical cyclone forecasts in Automated Tropical Cyclone Forecast System (ATCF) format. Much like the Point-Stat tool, TC-Pairs produces matched pairs of forecast model output and an observation dataset. In the case of TC-Pairs, both the model output and observational dataset (or reference forecast) must be in ATCF format. TC-Pairs produces matched pairs for position errors, as well as wind, sea level pressure, and distance to land values for each input dataset.

TC-Pairs Usage

View the usage statement for TC-Pairs by simply typing the following:

At a minimum, the input ATCF format -adeck (or -edeck) source, the input ATCF format -bdeck source, and the configuration -config file must be passed in on the command line.

The -adeck, -edeck, and -edeck options can be set to either a specific file name or a top-level directory which will be searched for files ending in ".dat".

TC-Stat Tool: General

TC-Stat Functionality

The TC-Stat tool reads the .tcst output file(s) of the TC-Pairs tool. This tools provides the ability to further filter the TCST output files as well as summarize the statistical information. The TC-Stat tool reads .tcst files and runs one or more analysis jobs on the data. TC-Stat can be run by specifying a single job on the command line or multiple jobs using a configuration file. The TC-Stat tool is very similar to the Stat-Analysis tool. The two analysis job types are summarized below:

• The filter job simply filters out lines from one or more TCST files that meet the filtering options specified.
• The summary job operates on one column of data from TCST file. It produces summary information for that column of data: mean, standard deviation, min, max, and the 10th, 25th, 50th, 75th, and 90th percentiles, independence time, and frequency of superior performance.
• The rirw job identifies rapid intensification or weakening events in the forecast and analysis tracks and applies categorical verification methods.
• The probrirw job applies probabilistic verification methods to evaluate probability of rapid inensification forecasts found in edeck's.

TC-Stat Usage

View the usage statement for TC-Stat by simply typing the following:

At a minimum, you must specify at least one directory or file in which to find TCST data (using the -lookin path command line option) and either a configuration file (using the -config config_file command line option) or a job command on the command line.

The TCPairs use case utilizes the MET TC-Pairs tool.

Optional: Refer to the MET Users Guide for a description of the MET tools used in this use case.
Optional: Refer to the METplus Config Glossary section of the METplus Users Guide for a reference to METplus variables used in this use case.

There is an example of using TC-Pairs to match Extratropical Cyclones as well as Tropical Cyclones

Let's take a look at the data to understand why the configuration file is set up the way it is. If you recall, the {INPUT_BASE} is set in tutorial.conf as ${METPLUS_DATA}, so we will look there:

The output from this command will return four separate directories, each with their own collection of track files. Those four directories are:

An example of the track files found in each of these directories is:

Let's look at the output file:

You will notice that there are two models configurations included in the first file for Hurricane Florence, CTRL and MYNN and the line type written out is TCMPR indicate it is the matched pair information. If you open up the third file, you'll notice that there is only the CTRL model listed.

You will see that in the files are in the ATCF file format. Also note: in the files in the CTRL directory, the model listed in the 5th column is HLAT. Similarly, in the H19C directory, the model is CTRL; in the H19M directory it is MYNN, and in the MYNN directory, it is HLMY. While this is confusing in some ways, it does demonstrate that the TC-Pairs tool does not look at the directory structure and naming convention of the input files, but rather the model information included in the ATCF "a-decks", which is traditional naming convention for files with forecast tracks.

The TC_PAIRS_ADECK_TEMPLATE includes MODEL in the path, so we will leave MYNN and CTRL in the list.

Review output

You will notice that now all models are included in the appropriate files. Also note that all of the forecasts leads from 000000 to 1260000 are included in the 10th column.

The TCPairs use case utilizes the MET TC-Pairs tool.

Optional: Refer to the MET Users Guide for a description of the MET tools used in this use case.
Optional: Refer to the METplus Config Glossary section of the METplus Users Guide for a reference to METplus variables used in this use case.

Change to the ${METPLUS_TUTORIAL_DIR}

1. See what TCPairs use cases are available in met_tool_wrapper directory and run one

There is an example of using TC-Stat with no delineation between Extratropical Cyclones as well as Tropical Cyclones

This use case is focused on summarizing the ASPEED data including data from 24 hours before landfall. Keep in mind the summary information will be written out. Also, because the -dump_row flag was set, output will be placed in the file {TC_STAT_OUTPUT_DIR}/tc_stat_summary.tcst

2. Run the use case:

3. Review the output and input files:

Look at the output written to the directory specified as TC_STAT_OUT_DIR

You will notice this looks very similar to output from TCPairs, confirmed by noting the LINE_TYPE is TCMPR. While this is helpful, the outcome of the SUMMARY job is part of standard output from the tool, unless the -out flag is used when calling the tool. Standard output will be found in the location specified by the TC_STAT_OUTPUT_TEMPLATE. For this example, the file will be named job.out

Open the standard output file in ${METPLUS_TUTORIAL_DIR}/output/TCStat/, with the filename job.out. Inside, you'll find:

The headers indicate that this information includes details about the distribution of data as well as measures to generate plots of Frequency of Superior Performance. The description of each column can be found at: https://met.readthedocs.io/en/latest/Users_Guide/index.html. The other notable result is that most columns are listed as NA. This is because the -column setting is listed as 'ASPEED' with quotes around the column name. The quotes appears to be an error, probably a remnant of a previous METplus version. Let's change that and see what happens.

4. Copy METplus config file, modify and re-run

5. Review output

The values of ASPEED and TK_ERR are now available. Let's take one more step to demonstrate additional capability.

6. Copy METplus config file, modify and re-run

7. Review output

Open standard output file in ${METPLUS_TUTORIAL_DIR}/output/TCStat_run3 under the name job.out and scroll down to the bottom.

There are now 48 lines with lines summarizing AMAX_WIND and TK_ERR across 24 STORM_NAMEs. The information for FSP columns is zero because there is only data for 1 model being summarized.

METplus Use Case: Track and Intensity TCMPR (Tropical Cyclone Matched Pair) Plotter

This is a wrapper to the MET plot_tcmpr.R, based on R-project Statistical package Rscript.  This Rscript will be deprecated in a future METplus release.

Review: Take a look at the following settings.

The default image resolution for the plot_tcmpr.R Rscript is set to 300, for print quality, but quite large for display purposes. The following is an R config file that is passed to the Rscript and changes the default and reduces the image size. It is a simple one line file, but go and take a look.

The plot_tcmpr.R script knows to read this TCMPRPlotterConfig_Customize R config file since it is defined in the TCMPRPlotter.conf file.

Note: If you modify the TCMPR_PLOTTER_CONFIG_FILE variable and leave it undefined (not set), then no configuration file will be read by the R script and the default setting of img_res=300 will be used.

Optional, Refer to the TC-Stat tool example section of the MET Users Guide for a description of plot_tcmpr.R 
If you wish, refer to the Format Information for TCMPR output line type table, for a tabular description of the TC-Pairs output TCST format. After running this Track and Intensity example, you can view the TC-Pairs TCST output files generate by this example under your ${METPLUS_TUTORIAL_DIR}/output/track_and_intensity/tc_pairs directory while comparing to the information in the table.

The tcmpr defaults are used for any conf settings left as, unset, For example as: XLAB = 
The current settings in the conf files are ok and no changes are required. However, after running the exercise feel free to experiment with these settings.

The R script that Track and Intensity runs is located in the MET installation; share/met/Rscripts/plot_tcmpr.R. The usage statement with a short description of the options for plot_tcmpr.R can be obtained by typing:

with no additional arguments. The usage statement can be helpful when setting some of the option in METplus, below is a description of these settings. 

Run METplus: Run Track and Intensity use case.

Examples: Run the track and intensity plotting script
Generates plots using the MET plot_tcmpr.R Rscript.

Example 1: 
In this case, the TCMPRPlotter.conf file is configured to generate MEAN and MEDIAN plot types with TCMPR_PLOTTER_PLOT_TYPES.
Run the following command:

The following directory lists the files generated from running these use cases:

Example 2: 
Copy the TCMPRPlotter.conf file to the user_config directory:

Open the new conf file with an editor and change the list of plot types to only include BOXPLOT.

Change this line (around line 65):

to

Close the file and rerun METplus:

The following directory lists the additional files generated from running this use case:

To view the png images generated by running these examples, use the display command. 
cd to your OUTPUT_BASE/tcmpr_plots directory.

End of Session 6

Congratulations! You have completed Session 6!