This program plots multiple
Taylor diagrams based on comparison
statistics
between a list of model output and
a data set.
This program has three modes
depending on a command line option:
Mode 1:
for one variable, plot multiple Taylor diagrams,
one per statistical component (the default mode)
Mode 2:
for one variable, plot all statistical components
on one Taylor diagram.
Mode 3:
for multiple variables, plot the total space-time
components
on one Taylor diagram.
For one variable (e.g. Sea Surface
Temperature), comparison statistics
between observation data and
models are read from a netCDF file.
This file contains, for each
model, comparison statistics for all
statistical components.
The component list is the
following:
- 1. Total
space-time field
- 2. Zonal
annual mean
- 3. Zonal
monthly mean
- 4. Zonal
monthly anomaly: Monthly anomaly to zonal annual
mean
- 5.
Annual mean map
- 6.
Monthly map anomaly: Monthly anomaly to
annual mean map
- 7.
Longitudinal anomaly: Spatial anomaly to zonal
monthly mean
- 8.
Annual mean longitudinal anomaly: Annual mean
longitudinal anomaly to zonal mean
Additional points may be shown on
the diagrams along with models, like
- all
models mean or median
-
additionnal data-set
Each model is represented on the
diagrams as
- either a
one-or-two letter code, optionally with a specific
color.
- or a
coloured dot with a one-or-two letter code. The dot
color is
taken from a color scale varying from purple to blue
and indicates
the model bias.
By default, model names are
extracted from input NetCDF file.
Code-letters are derived from
these names and predefined colors are
automatically assigned to
them.
An optional parameter file allows
the user to change the model list
and specify their attributes. (see
usage section below)
Output files:
Each Taylor diagram is saved
in a file. Default format is PostScript.
Alternate option is GIF.
Respective file name extension is '.ps' or '.gif'
When running in mode 1, each file
name is after the component short name
(first listed above) with all
white space replaced by underscores.
For example:
Total_space-time_field
Directory where these files are
created is by default the current working
directory. This may be specified
differently, using a command line option.
If the specified directory does
not exist, it is created. If output files
already exist in that directory,
they are overwritten.
When running in mode 2, output
image file name is after the variable name
from input NetCDF file and is in
current working directory.
Path and name may be specified
differently using a command line option.
When running in mode 3, output
image file name is after the FIRST variable name
from (FIRST) input NetCDF file and
is in current working directory.
Path and name may be specified
differently using a command line option.
Usage: plot_taylor_diagrams.py
[options] [stats-filename]
or: plot_taylor_diagrams.py
(--mix_variables | -v)
[stats-filename
stats-filename2]
|
stats-filename and stats-filename2 are
NetCDF input file names.
If none
is given, <stats-filename> defaults to
'model_vs_data_comp_stats.nc'
in
current working directory.
Options are:
*** General options ***
--help
-h print out this
help
--output=<pathname>
-o <pathname>
Depending on whether single or multiple image file to
be output,
specify output filename or output directory
pathname.
If multiple output files, file names are fixed (not
an option).
In any case, default directory is current working
directory.
For single output file, default filename depends on
chosen mode
option (see below).
--gif
-g Output file format is
GIF. File extension is '.gif'
--portrait
-p Output image
orientation is 'portrait'. (Default is
'landscape')
--models=<model-list>
-m
<model-list>
Comma-separated list of all models/composite-var to
show
on the diagram. Composite variables may be 'Median'
or 'Mean',
for example.
By default, all variables present in the input NetCDF
file are shown.
Listed names must be consistent with variable ids in
input NetCDF file.
That is, when <name> appears in the list,
<name>_Correlation and
name>_StandardDeviation must exist as variables in
the input file.
--param-file=<filename>
-f <filename>
Specify a model parameter file that contains python statements
defining model list and plot parameters for each models.
Listed models must be consistent with variable ids in input file.
See paragraph 'Model parameter file' below for more details.
This option is incompatible with --models or -m option above.
If neither model list nor model parameter file is specified on the
command line, the program looks for a parameter file named
'model_dictionary.py' in current working directory.
--symbol-size=<size>
-i=<size>
Adjust size of points and associated symbols on the
diagrams.
<size> is an integer value in the range [1,
100]. Default is 50.
--title=<title>
-t <title>
Add a title to the top of each Taylor diagram
--units=<units>
-u <units>
Add the variable units on the standard deviation
axis.
*** Mode options ***
--superimpose
-s Selector for mode 2.
All components are superimposed on the same
Taylor diagram. One different color of standard
palette is used
for each component. Only one image file is
output.
Default is mode 1 where each component is plot
separately
on a Taylor diagram.
--mix-variables
-v Selector for mode 3.
Main components (Total space-time field)
of all variables are on the same Taylor diagram.
One different color of standard palette is used for
each variable.
Only one image file is output.
*** Miscelaneous (mode
dependendant) options ***
--components=<component-list>
(valid only for mode 1 and 2)
-C
<component-list>
Choose a subset of components to plot.
<component-list> is a list of numbers
corresponding to ranks
in available component list above (from 1 to 8).
List order is meaningless.
--bias
(valid only for mode 1)
-b For the main
statistical component, models are represented on
the
diagram with a dot whose color is function of model
bias.
By default, models are represented with only a
coloured letter-code.
--color
(valid only for mode 1)
-c If no option --bias or
-b is used, represent models on the diagrams
with coloured code-letter instead of black
code-letter.
--discrete
(valid only for mode 1)
-d When using --bias or
-b option, bias is represented using a limited
number of colors instead of a continuous color
scale.
--red-blue
(valid only for mode 1)
-r When using --bias or
-b option, bias is represented using a standard
blue to red palette centered on value 0.
Model parameter file:
A model parameter file is an optional python script that lists all models and
composite vars used for comparison. It describes also, for each model/
composite-var, the letter-code and the color that represents this model/var
on the Taylor diagram.
By default, the program looks for that file under the name
'model_dictionary.py' in current working directory. User may specify
a different one using option '-f'.
File structure:
It
defines two Python variables named:
'model_dictionary' (mandatory, for models)
and
'other_dictionary' (optional, for additionnal
variables).
These variables are Python dictionnaries with,
- as
keys: model or other variable names
- as
values: inner two-item dictionnaries defining
code-letter and color for each model/variable
Code-letter is a string of one or more letter(s),
typically 1 or 2.
Color is a number ranging from 241 to 255. It is an
index
into
the default color palette.
color: 241 Black
color: 242 Red
color: 243 Green
color: 244 Blue
color: 245 Yellow
color: 246 Cyan
color: 247 Magenta
color: 248 Orange
color: 249 Marroon
color: 250 Navy
color: 251 Brown
color: 252 Grey
color: 253 Pale green
color: 254 Rose
color: 255 Light blue
The color
parameter is relevant only when using mode 1
(default)
whith -c
or --color option.
Example:
Let's say we deal with two models 'Model_one' and
'Model_two'
and
one additionnal data-set named 'Data'
'model_dictionary' variable will be defined as the
following:
model_dictionary = {\
"Model_one": {
"code_letter": "A", "color": 241 }, \
"Model_two": {
"code_letter": "2", "color": 242 }, \
}
'other_dictionary' variable will be defined as the
following:
other_dictionary = {\
"Data": { "code_letter":
"D", "color": 244 }, \
}
|