Use Case 2: Probability distributions¶

In the following RCAT is applied to calculate standard empirical probability distribution functions. Similar to Use Case 1 most changes will be done in the configuration file, src/config/config_main.ini.

Create hourly PDF statistics and visualize the results¶

In the first example PDF’s based on hourly data for historical and scenario simulations will be calculated for precipitation and T2m. Output statistics are then compared in line plots for specified regions.

STEP 1: Data input¶

Under MODELS section configure for two arome simulations – historic (arome_his) and future scenario (arome_scn). Thus, different years are specified, however, in the example here months 6,7,8 are specified so that only data for June, July and August is extracted.

arome_his = {
     'fpath': '/nobackup/rossby21/rossby/joint_exp/norcp/NorCP_AROME_ECE_ALADIN_1985_2005/netcdf',
     'start year': 1990, 'end year': 1994, 'months': [6,7,8]
     }
arome_scn = {
     'fpath': '/nobackup/rossby21/rossby/joint_exp/norcp/NorCP_AROME_ECE_ALADIN_RCP85_2080_2100/netcdf',
     'start year': 2090, 'end year': 2094, 'months': [6,7,8]
     }

We’re looking at climate change signal in the model, so the OBS section can be left as is.

STEP 2: Variables¶

Set output directory under the SETTINGS section.

The key variables defines which variables to analyze along with some options regarding that particular variable. Since only models will be analyzed here, ‘obs’ is set to None. Further, models will be kept at their respective grids, thus ‘regrid to’ is also set to None. Statistics is configured for T2m (tas) and precipitation (pr) with hourly data as input (‘freq’ set to ‘1H’).

Specify regions the regions key for which statistics will be selected for.

output dir = /nobackup/rossby22/sm_petli/analysis/test_pdf_analysis

variables = {
    'pr': {'freq': '1H', 'units': 'mm', 'scale factor': None, 'accumulated': True, 'obs': None, 'regrid to': None},
    'tas': {'freq': '1H', 'units': 'K', 'scale factor': None, 'accumulated': False, 'obs': None, 'regrid to': None},
    }

regions = ['Scandinavia']

STEP 3: Select statistics¶

Under STATISTICS pdf will be specified. A list of bins will be used in the pdf. If not given here these bins will be defined automatically in RCAT by taking minimum and maximum of input data. This can be quite crude and not so representative, so it is suggested to define them here under the pdf key. They are specified in a dictionary where the keys are input variables and the values are the respective bin definitions. The bin definition is a list/tuple with start, stop and step values. For example, for precpitation a list of bins starting with 0 and ending with 50 using a step of 1 is defined here.

stats = {
   'pdf': {'bins': {'pr': (0, 50, 1), 'tas': (264, 312, 1)}}
    }

Default options for pdf can be seen in the default_stats_config function in the statistics module.

STEP 4: Plotting¶

Under PLOTTING, validation plot should be set to True to enable plotting. Plotting of pdf’s will be line plots only (regions should therefore be specified). We only specify linewidths to be 2.5.

validation plot = True

map configure = {}
map grid setup = {}
map kwargs = {}

line grid setup = {}
line kwargs = {'lw': 2.5}

STEP 5: Configure cluster¶

The number of nodes to be used in the selected SLURM cluster is set to 20 (increase if needed) and a walltime of 2 hours.

cluster type = slurm
nodes = 20
cluster kwargs = {'walltime': '02:00:00'}

STEP 6: Run RCAT¶

To run the analysis run from terminal (see Run RCAT in RCAT Configuration):

python $HOME/git/rcat/src/RCAT_main.py -c config_main.ini

Output statistics files will be located in the sub-folder stats under the user-defined output directory.

Calculate PDF’s for daily maximum values instead¶

Imagine one would like to do the same kind of statistical analysis as above, however, with a different temporal resolution and/or time statistic on the input data. For example, let’s assume that pdf’s should be calculated for daily maxmimum data instead. How can this be achieved?

This can be done during RCAT runtime, using an option in the stats property (under SETTINGS) called resample resolution. It is specified by a list/tuple with two locations; the first index represents the time resolution sought after and the second location the statistic used for each sample in the resampling. In the example here data is resampled into daily maximum values:

stats = {
   'pdf': {'bins': {'pr': (0, 50, 1), 'tas': (264, 312, 1)}, 'resample resolution': ['D', 'max']}
    }

When set, run RCAT again.