rwgen.RainfallModel.simulate

RainfallModel.simulate(output_types='point', output_subfolders='default', output_format='txt', catchment_metadata=None, grid_metadata=None, epsg_code=None, cell_size=None, dem=None, simulation_length=30, n_realisations=1, timestep_length=1, start_year=2000, calendar='gregorian', random_seed=None, run_simulation=True, apply_shuffling=False, weather_model=None)

Simulate stochastic time series realisation(s) of NSRP process.

Parameters:

  • output_types (str or list of str) – Types of output (discretised) rainfall required. Options are 'point', 'catchment' and 'grid'.

  • output_subfolders (str or dict) – Sub-folder in which to place each output type. If 'default' then dict(point='point', catchment='catchment', grid='grid') is used for a spatial model and dict(point='') for a point model (i.e. output to self.output_folder). If None then all output files are written to self.output_folder.

  • output_format (str) – Flag indicating output file format for point and catchment output. Current option is txt. Gridded output will be written in NetCDF format.

  • catchment_metadata (geopandas.GeoDataFrame or str) – Geodataframe containing catchments for which output is required (or path to catchments shapefile). Optional.

  • grid_metadata (dict or str) – Specification of output grid to use for both gridded output (optional). This grid is also used to support catchment output. Dictionary keys use ascii raster header keywords, e.g. dict(ncols=10, nrow=10, ...). Use xllcorner and yllcorner, as well as lowercase for each keyword. If None then a grid is defined to encompass catchment locations using the cell_size argument. The path to an ascii raster file to use as a template for the grid can be given instead.

  • epsg_code (int) – EPSG code for projected coordinate system used for domain (required if catchment or grid output is requested).

  • cell_size (float) – Cell size to use if grid is None but a grid is needed for gridded output and/or catchment output.

  • dem (xarray.DataArray or str) – Digital elevation model (DEM) [m] as data array or ascii raster file path. Optional but recommended.

  • simulation_length (int) – Number of years to simulate in one realisation (minimum of 1).

  • n_realisations (int) – Number of realisations to simulate.

  • timestep_length (int) – Timestep of output [hr]. Default is 1 (hour).

  • start_year (int) – Start year of simulation.

  • calendar (str) – Flag to indicate whether gregorian (default accounting for leap years) or 365-day calendar should be used.

  • random_seed (int) – Seed to use in random number generation.

  • run_simulation (bool) – Flag for whether to run simulation. Setting to False may be used to update self.simulation_args to allow self.postprocess() to be run without self.simulate() having been run first (i.e. reading from existing simulation output files).

  • apply_shuffling (bool) – Indicates whether to run model with or without shuffling following Kim and Onof (2020) method. # TODO: Provide explanation of method

  • weather_model (object) – Instance of WeatherModel. # TODO: Expand explanation

Notes

Though gridded output is calculated (if output_types includes 'grid') it is not yet available to write (under development).

All locations in self.point_metadata will be written as output currently. # TODO: To be changed

Catchment shapefile or geodataframe assumed to have an ID field that can be used to identify catchments. Both point and catchment metadata are assumed to have a Name field for use as a prefix in file naming.

Updates self.simulation_args in preparation for post-processing.