A Python wrapper to setup and run the hydrologic modelling framework Raven.


Keywords
ravenpy, raven
License
MIT
Install
pip install ravenpy==0.7.8

Documentation

RavenPy

PyPI Conda-Forge Build status Documentation Status Coveralls

A Python wrapper to setup and run the hydrologic modelling framework Raven.

RavenPy is a Python wrapper for Raven, accompanied by utility functions that facilitate model configuration, calibration, and evaluation.

Raven is an hydrological modeling framework that lets hydrologists build hydrological models by combining different hydrological processes together. It can also be used to emulate a variety of existing lumped and distributed models. Model structure, parameters, initial conditions and forcing files are configured in text files, which Raven parses to build and run hydrological simulations. A detailed description about modeling capability of Raven can be found in the docs.

RavenPy provides a Python interface to Raven, automating the creation of configuration files and allowing the model to be launched from Python. Results, or errors, are automatically parsed and exposed within the programming environment. This facilitates the launch of parallel simulations, multi-model prediction ensembles, sensitivity analyses and other experiments involving a large number of model runs.

The code is currently undergoing a number of changes to support semi-distributed watersheds, so expect some API changes over the next versions.

Features

  • Download and compile Raven with pip
  • Configure, run and parse Raven outputs from Python
  • Parallel simulations over parameters, models or watersheds
  • Utility command to create grid weight files
  • Extract physiographic information about watersheds
  • Algorithms to estimate model parameters from ungauged watersheds
  • Exposes outputs (flow, storage) as xarray.DataArray objects

Install

Please see the detailed installation docs.

Acknowledgements

RavenPy's development has been funded by CANARIE and Ouranos and would be not be possible without the help of Juliane Mai and James Craig.

This package was created with Cookiecutter and the Ouranosinc/cookiecutter-pypackage project template.