The Forced Component Estimation Statistical Method Intercomparison Project (ForceSMIP) will utilize climate model ensembles to develop and assess statistical methods for isolating the forced response from individual realizations of the climate system. These methods will then be applied to observations to produce a community estimate of the forced climate response.
This repository contains functions, scripts, and notebooks associated with the ForceSMIP project <https://sites.google.com/ethz.ch/forcesmip/> and hackathon. The notebooks directory contains example methodologies for obtaining the forced climate response using ForceSMIP data (via Jupyter Notebooks). The scripts directory also contains methodologies (but not in Jupyter Notebook format). Methods include dynamical adjustment, linear inverse models, linear regression models (e.g., ridge, PLS, multiple linear), neural networks, and low frequency component analysis.
- Free software: MIT license