Learning General Causal Structures with Hidden Dynamic Process for Climate Analysis

TL;DR

This paper introduces a unified framework for uncovering both observable-to-observable causal relations and latent driving forces in climate data, enabling better understanding and forecasting of climate dynamics.

Contribution

It proposes a novel method that jointly learns causal structures among observed variables and latent processes, with theoretical guarantees even in nonparametric settings.

Findings

Theoretical identifiability of causal and latent structures from time-series data.

CaDRe achieves competitive forecasting accuracy on real climate datasets.

Recovered causal graphs align with domain expertise, providing interpretable insights.

Abstract

Understanding climate dynamics requires going beyond correlations in observational data to uncover their underlying causal process. Latent drivers, such as atmospheric processes, play a critical role in temporal dynamics, while direct causal influences also exist among geographically proximate observed variables. Traditional Causal Representation Learning (CRL) typically focuses on latent factors but overlooks such observable-to-observable causal relations, limiting its applicability to climate analysis. In this paper, we introduce a unified framework that jointly uncovers (i) causal relations among observed variables and (ii) latent driving forces together with their interactions. We establish conditions under which both the hidden dynamic processes and the causal structure among observed variables are simultaneously identifiable from time-series data. Remarkably, our guarantees hold even in the nonparametric setting, leveraging contextual information to recover latent variables and causal relations. Building on these insights, we propose CaDRe (Causal Discovery and Representation learning), a time-series generative model with structural constraints that integrates CRL and causal discovery. Experiments on synthetic datasets validate our theoretical results. On real-world climate datasets, CaDRe not only delivers competitive forecasting accuracy but also recovers visualized causal graphs aligned with domain expertise, thereby offering interpretable insights into climate systems.