Input Data

Example files for a spatiotemporal NSRP model are given in the examples/stnsrp/input folder in the RWGEN root folder. In the description below it is assumed that we are located in this folder (current directory).

The minimal requirements for setting up a spatiotemporal rainfall model using gauge data are:

  1. A csv file containing gauge metadata - see gauge_metadata.csv

  2. Files containing gauge time series (one file per gauge) - see gauge_data directory

A DEM is optional but good if available.

A polygon shapefile containing catchment/sub-catchment boundaries is needed if catchment-average outputs are requested.

Gauge Metadata File

The gauge metadata csv file has mandatory columns (data types) of:

  1. Point_ID (integer) - unique integer for each gauge

  2. Easting (integer or float) - units = metres

  3. Northing (integer or float) - units = metres

  4. Name (string or integer) - to help identify the relevant gauge time series file

It is assumed that the Name field corresponds with a gauge time series file name (after .csv is added to the Name entry). For example, Burgkunstadt corresponds with the file ./gauge_data/Burgkunstadt.csv. Avoid spaces and special characters.

Elevation (integer or float) in metres is not technically mandatory but good to include if possible (depending on region).

Other fields in the gauge metadata csv file are not currently needed or used (but can be added for reference).

Gauge Time Series Files

The example uses csv format for time series files. Required columns are:

  1. DateTime - dd/mm/yyyy hh:mm format (i.e. '%d/%m/%Y %H:%M', e.g. 01/01/2005 00:00) or yyyy-mm-dd hh:mm format (i.e. '%Y-%m-%d %H:%M', e.g. 2005-01-01 00:00)

  2. Value (float) - mm/timestep

Missing data can be represented as -999 (the code will set any negative values as missing).

All gauge time series files should be placed in the same directory, e.g. ./gauge_data

DEM

Supplying a DEM is optional but useful if “ungauged” locations are to be sampled or grid-based / catchment-average outputs are required (depending on region to some extent).

If supplying a DEM, the gauge metadata file should contain an Elevation column.

The DEM should be in ascii raster format and it should not contain any missing values (for now at least).

DEM resolution can be relatively high (e.g. 100 m in this example of a ~6000 km**2 overall domain). While a 100 m resolution DEM might be supplied as input, internally the model simulation can take place (and produce outputs) at a coarser resolution (e.g. 1 km).

Unless output on a specific grid is required, precise DEM extent does not matter but it should not be much larger than it really needs to be.

Catchment(s) Shapefile

A polygon shapefile of catchment(s) is needed if catchment-average output is requested (otherwise it is not required).

Each polygon is treated independently by the code, so the shapefile can contain any polygons of interest (i.e. it does not matter if they overlap). Output time series will be generated for each polygon in the shapefile.

It is currently assumed that the shapefile contains the following fields:

  1. ID (integer) - identifier for catchment / sub-catchment

  2. Name (string) - for use in writing output files

Other fields may be present in the shapefile - they will be ignored.