Scalable Changepoint Detection for Large Spatiotemporal Data on the Sphere

TL;DR

This paper introduces a scalable Bayesian framework for detecting changepoints in large-scale spherical spatiotemporal data, combining advanced statistical models with efficient computational techniques to handle high-dimensional environmental datasets.

Contribution

It presents a novel approach that models spatially dependent changepoints using a multinomial probit model with Gaussian processes, leveraging SPDE and spherical harmonics for scalable inference.

Findings

Demonstrates high accuracy and robustness in simulation studies

Achieves significant computational efficiency improvements

Successfully identifies atmospheric event changepoints in real data

Abstract

We propose a novel Bayesian framework for changepoint detection in large-scale spherical spatiotemporal data, with broad applicability in environmental and climate sciences. Our approach models changepoints as spatially dependent categorical variables using a multinomial probit model (MPM) with a latent Gaussian process, effectively capturing complex spatial correlation structures on the sphere. To handle the high dimensionality inherent in global datasets, we leverage stochastic partial differential equations (SPDE) and spherical harmonic transformations for efficient representation and scalable inference, drastically reducing computational burden while maintaining high accuracy. Through extensive simulation studies, we demonstrate the efficiency and robustness of the proposed method for changepoint estimation, as well as the significant computational gains achieved through the combined use of the MPM and truncated spectral representations of latent processes. Finally, we apply our method to global aerosol optical depth data, successfully identifying changepoints associated with a major atmospheric event.