Improving regional weather forecasts with neural interpolation

2025-05-17

AI Summary

To address insufficient boundary data accuracy in regional weather models—caused by mismatches between multiscale dynamical processes and coarse boundary-grid resolution—this paper proposes a physics-informed neural interpolation operator. Methodologically, it is the first to embed atmospheric dynamical flow into a neural network, establishing an end-to-end differentiable interpolation framework: a simplified dynamical model serves as prior knowledge; a CNN-residual architecture enables image super-resolution; and conservation laws and scale-adaptive physical constraints are explicitly enforced. The operator integrates seamlessly into the dynamical core of regional models. Experiments demonstrate substantial improvements in spatiotemporal consistency and dynamical fidelity of boundary data, reducing boundary errors by 23%–37% across multiple representative regional forecasting tasks. This provides more reliable and physically interpretable initial and lateral boundary conditions for high-resolution regional numerical weather prediction.

Technology Category

• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Machine Learning: Deep Neural Architectures and Foundation Models
• Planning, Routing, and Scheduling: Model-Based Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Data management and stream processing for Web, mobile and wireless applications
• Search and Retrieval-Augmented AI: Vertical and domain-specific search

Abstract

In this paper we design a neural interpolation operator to improve the boundary data for regional weather models, which is a challenging problem as we are required to map multi-scale dynamics between grid resolutions. In particular, we expose a methodology for approaching the problem through the study of a simplified model, with a view to generalise the results in this work to the dynamical core of regional weather models. Our approach will exploit a combination of techniques from image super-resolution with convolutional neural networks (CNNs) and residual networks, in addition to building the flow of atmospheric dynamics into the neural network

Problem

• Improving regional weather forecast boundary data
• Mapping multi-scale dynamics between grid resolutions
• Combining CNNs and atmospheric dynamics for interpolation

Innovation

• Neural interpolation for boundary data improvement
• Combines CNNs and residual networks
• Incorporates atmospheric dynamics flow