An Interpretable Model of Climate Change Using Correlative Learning

TL;DR

This paper introduces an interpretable neural network model that predicts years based on climate data and uses a novel correlative learning algorithm to identify patterns indicative of climate change over time.

Contribution

It presents a new neural network approach combined with Alopex, a stochastic correlative learning algorithm, for interpreting climate change indicators from temperature and precipitation data.

Findings

The model accurately predicts the year from climate data.

Patterns associated with specific years evolve over time.

Alopex effectively identifies climate change indicators.

Abstract

Determining changes in global temperature and precipitation that may indicate climate change is complicated by annual variations. One approach for finding potential climate change indicators is to train a model that predicts the year from annual means of global temperatures and precipitations. Such data is available from the CMIP6 ensemble of simulations. Here a two-hidden-layer neural network trained on this data successfully predicts the year. Differences among temperature and precipitation patterns for which the model predicts specific years reveal changes through time. To find these optimal patterns, a new way of interpreting what the neural network has learned is explored. Alopex, a stochastic correlative learning algorithm, is used to find optimal temperature and precipitation maps that best predict a given year. These maps are compared over multiple years to show how temperature and precipitations patterns indicative of each year change over time.