DIRESA, a distance-preserving nonlinear dimension reduction technique based on regularized autoencoders

TL;DR

This paper introduces DIRESA, a novel autoencoder-based dimension reduction method that preserves distances in the latent space, enabling efficient and interpretable analysis of large climate datasets with improved performance over existing techniques.

Contribution

The paper presents DIRESA, a distance-regularized Siamese autoencoder architecture that effectively preserves nonlinear distances in reduced dimensions, outperforming PCA, UMAP, and variational autoencoders.

Findings

DIRESA reduces storage needs for climate datasets.

Latent components reveal physical climate variability modes.

Outperforms PCA, UMAP, and variational autoencoders in distance preservation.

Abstract

In meteorology, finding similar weather patterns or analogs in historical datasets can be useful for data assimilation, forecasting, and postprocessing. In climate science, analogs in historical and climate projection data are used for attribution and impact studies. However, most of the time, those large weather and climate datasets are nearline. This means that they must be downloaded, which takes a lot of bandwidth and disk space, before the computationally expensive search can be executed. We propose a dimension reduction technique based on autoencoder (AE) neural networks to compress the datasets and perform the search in an interpretable, compressed latent space. A distance-regularized Siamese twin autoencoder (DIRESA) architecture is designed to preserve distance in latent space while capturing the nonlinearities in the datasets. Using conceptual climate models of different complexities, we show that the latent components thus obtained provide physical insight into the dominant modes of variability in the system. Compressing datasets with DIRESA reduces the online storage and keeps the latent components uncorrelated, while the distance (ordering) preservation and reconstruction fidelity robustly outperform Principal Component Analysis (PCA) and other dimension reduction techniques such as UMAP or variational autoencoders.