/project_causal_climate_dynamics

Causal discovery of drivers of the summer Himalayan precipitation

Primary LanguageJupyter Notebook

Causal discovery of drivers of the summer Himalayan precipitation

This code uses a causal discovery and causal inference methods to analyse linkages among ENSO, circulation fields, and the summer monsoon precipitation over the Himalayas.

Requirements

Main requirements

  • Python >= 3.9
  • Tigramite >= 5.0

The rest of requirements can be installed by using 'requirements_file.txt' by creating a virtual environment as follows:

$ conda create --name <env> --file <this file>

The code has been tested on platform: osx-64.

Instructions

To run the code you can use a jupyter notebook: 'Causal analysis of climate dynamics.ipynb'. The notebook uses 'config.yml' with the basic setup for:

  • preprocessing of climate indices
  • plotting maps
  • statistical testing
  • lagged cross-correlation
  • causal discovery using PCMCI algorithm and causal inference implemented in Tigramite
  • plotting causal graphs
  • bootstrapping