Scalable coarse-to-fine spatial downscaling

2026-06-29

AI Summary

This study addresses the computational inefficiency and limited scalability of traditional spatial downscaling methods when applied to large datasets. To overcome these challenges, the authors propose a scalable coarse-to-fine spatial downscaling approach (CF-DS) that synthesizes multiscale local models to effectively satisfy aggregation constraints without requiring covariance matrix inversion or likelihood evaluation. The CF-DS method achieves a favorable balance between predictive accuracy and computational efficiency, delivering accuracy comparable to area-to-point kriging while substantially reducing computation time. The approach is successfully demonstrated in a case study on downscaling electricity consumption across the Tokyo metropolitan area. The associated algorithms have been implemented in the R package spCF, and their performance and validity are confirmed through Monte Carlo simulations.

Abstract

This study proposes coarse-to-fine downscaling (CF-DS), a scalable spatial downscaling method extending coarse-to-fine spatial modeling. Unlike conventional spatial-statistical downscaling methods such as area-to-point kriging, CF-DS does not require covariance matrix inversion or likelihood evaluation. Instead, it represents latent spatial processes through the synthesis of multi-scale local models, substantially reducing computational cost while approximately satisfying the aggregation constraint. Monte Carlo experiments show that CF-DS achieves predictive accuracy comparable to area-to-point kriging with dramatically shorter computation times, particularly for large datasets. An application to downscaling electricity consumption in the Tokyo metropolitan area further demonstrates its practical usefulness. The results suggest that CF-DS provides an efficient alternative for large-scale spatial downscaling problems. CF-DS is implemented in an R package spCF (https://cran.r-project.org/web/packages/spCF/index.html).

Problem

• spatial downscaling
• scalability
• computational cost
• large-scale data
• aggregation constraint

Innovation

• coarse-to-fine downscaling
• spatial downscaling
• multi-scale local models
• computational efficiency
• aggregation constraint