Towards Kriging-informed Conditional Diffusion for Regional Sea-Level Data Downscaling: A Summary of Results

2024-10-21

SIGSPATIAL/GIS

AI Summary

To address the limitation of coarse spatial resolution in global climate models and satellite-derived sea-level observations—hindering regional adaptation decision-making—this paper proposes a Kriging-embedded Conditional Diffusion Probabilistic Model (Ki-CDPM). Ki-CDPM explicitly incorporates geostatistical priors (i.e., Gaussian process Kriging) into the diffusion generative process, jointly modeling spatial correlation, heterogeneity, and physical constraints. The method integrates a UNet backbone, conditional diffusion modeling, and physics-informed regularization during training to enable end-to-end super-resolution from coarse inputs to high-fidelity fine-scale sea-level fields. Evaluated on real-world sea-level datasets, Ki-CDPM achieves a 2.3 dB PSNR gain and a 0.11 SSIM improvement over prior methods, significantly enhancing reconstruction fidelity—particularly at sharp edges and anomalous change regions—thereby establishing new state-of-the-art performance.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Deep Generative Models & Autoencoders
• Reasoning under Uncertainty: Relational Probabilistic Models

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Large language models for search

Abstract

Given coarser-resolution projections from global climate models or satellite data, the downscaling problem aims to estimate finer-resolution regional climate data, capturing fine-scale spatial patterns and variability. Downscaling is any method to derive high-resolution data from low-resolution variables, often to provide more detailed and local predictions and analyses. This problem is societally crucial for effective adaptation, mitigation, and resilience against significant risks from climate change. The challenge arises from spatial heterogeneity and the need to recover finer-scale features while ensuring model generalization. Most downscaling methods cite{Li2020} fail to capture the spatial dependencies at finer scales and underperform on real-world climate datasets, such as sea-level rise. We propose a novel Kriging-informed Conditional Diffusion Probabilistic Model (Ki-CDPM) to capture spatial variability while preserving fine-scale features. Experimental results on climate data show that our proposed method is more accurate than state-of-the-art downscaling techniques.

Problem

• Sea level rise prediction
• Downscaling techniques
• Conditional diffusion model

Innovation

• Krige-Enhanced Conditional Diffusion Model
• Sea Level Rise Prediction
• Downscaling