Benchmarking configuration¶

Plot configuration details¶

For each variable, a configuration file with the extension .ini is used to specify. The INI files are expected to be located in a directory which is specified in the main configuration file (.cfg). The plot configuration file specifies for each variable

which diagnostics should be applied to a certain variable

how plots for a particular diagnostic should look like (e.g. colorbars, limits ...)

which observational datasets should be used

An INI file has two major parts:

Global plot options

Observation specific plot options (for each used observational dataset)

Global plot options¶

The global plot options have the following structure (example below):

[OPTIONS]
map_difference =  True
map_seasons    =  True
map_season_difference = False
reichler_plot  =  True
gleckler_plot   =  True
hovmoeller_plot   =  False
regional_analysis = True
global_mean    = True
vmin           =  0.
vmax           =  8.
dmin           =  -1.
dmax           =  1.
units          =  $mm/day$
label          =  Daily evaporation
cticks         = [0.,2.,4.,6.,8.,10.]
nclasses       = 8
preprocess     = True
interpolation  = conservative
targetgrid     = t63grid
projection     = robin
region_file    = /home/m300028/shared/data/CMIP5/evap/evspsbl/merged/dummy_mask2.nc
region_file_varname = regmask

map_difference [True,False]: use diagnostic to plot difference between models and observations
map_seasons [True,False]: use diagnostic to plot climatological monthly mean or seasonal mean maps of models and observations
map_season_difference [True,False]: same as map_seasons, but for difference between models and observations.
reichler_plot [True,False]: Summarize error skill score for this variable at the end of the section for this variable.
gleckler_plot [True,False]: Use this variable in the Portraet Diagram at the end of the report.
hovmoeller_plot [True,False]: Generate a hovmoeller plot for the variable, for both observations and models.
regional_analysis [True,False]: Perform regional analysis (statistics and correlation) of observations and models per variable.
region_file: Name of netCDF file which contains the rasterized region IDs (user needs to ensure that the same geometry as the target grid is provided)
region_file_varname: name of variable in region_file, which shall be read to identify regions; note that the data is interpreted as integer values.
global_mean: generate a global mean plot for this variable (see XXXX)
vmin: minimum plotting limit for data
vmax: maximum plotting limit for data
dmin: minimum plotting limit for difference plot
dmax: maximum plotting limit for difference plot
units: string to specify units of the variable. This is used for automatic labeling of plots. Note that all text can be used which can also be used for labelling in matplotlib. In particular the usage of $ is usefull to render text using latex (e.g. $frac{a}{b}$ will plot you the a/b in a nice way).
label: label text to be used for the variable
cticks: tick labels for colormap

Observation specific plot options¶

Below the global options, one can include an arbitrary number of observations.

Each observation is specified by a block of configuration parameters, like in the following example.:

[CLARASAL]
obs_file =  #get_data_pool_directory() + 'data_sources/CMSAF/CLARA-SAL/DATA/SAL_all_t63.nc'#
obs_var  =  sal
scale_data = 0.01
gleckler_position = 2
add_to_report = True
valid_mask = land

[Observation_Identifier] : str: unique identified for the observation. Will be used e.g. in plots as labels
obs_file : str: name of observation file. Here the user can either specify a full path name to a file or, like shown in the example above, execute a python command that is used to construct the filename. In the above example, the hash (#) is used to identify a python command. If the value of obs_file starts and ends with a hash, then the string in between is executed like you would execute a python command. Here, the routine get_data_pool_directory() is called, which returns a path name and then the remaining path to the observational data file is appended.
obs_var : str: name of variable in observation file
scale_data : float: scaling factor to be applied on data of the file. This is e.g. usefull if the netCDF file does not contain an own scale_factor attribute or if you want to apply simple conversions (e.g. from kg/m**2 s to mm/day for precipitation). The data is multiplied by the scaling factor.
gleckler_position : int: [1,2,3,4] position of the observational dataset in the Portraet diagram. Up to four different datasets can be shown at once. The meaning if the numbers is as follows: 1=top, 2=bottom, 3=left, 4=right
add_to_report : str: add this observation to the report [True,False]
valid_mask : str: [land,ocean,global]; specifies if a mask shall be applied to the dataset and model. If ‘land’, then all ocean areas are masked if ‘ocean’, then all land areas are masked. For any other options, the whole globe is used.