PeakWeather: MeteoSwiss Weather Station Measurements for Spatiotemporal Deep Learning

2025-06-16

AI Summary

High-quality in-situ benchmark data for meteorological AI modeling in complex terrain remains scarce. Method: We introduce the first deep learning–oriented, nationwide Swiss meteorological benchmark dataset—comprising high-density, sub-hourly (10-minute) multi-variable surface observations from 302 stations over >8 years, integrated with physics-informed terrain indices derived from a digital elevation model (DEM) and high-resolution numerical weather prediction (NWP) forecasts as physically consistent baselines. Contribution/Results: The dataset uniformly supports spatiotemporal forecasting, graph neural network modeling, missing-data imputation, and virtual sensing. Experiments demonstrate significant performance gains for state-of-the-art models across nowcasting, station-level gap-filling, and virtual sensing tasks. As the first open-source, reproducible, and physics-embedded observational meteorological benchmark, it addresses a critical data gap for machine learning–driven meteorology in complex terrain.

Technology Category

• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Search and Optimization: Metareasoning and Metaheuristics
• Machine Learning: Graph-based Machine Learning

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Accurate weather forecasts are essential for supporting a wide range of activities and decision-making processes, as well as mitigating the impacts of adverse weather events. While traditional numerical weather prediction (NWP) remains the cornerstone of operational forecasting, machine learning is emerging as a powerful alternative for fast, flexible, and scalable predictions. We introduce PeakWeather, a high-quality dataset of surface weather observations collected every 10 minutes over more than 8 years from the ground stations of the Federal Office of Meteorology and Climatology MeteoSwiss's measurement network. The dataset includes a diverse set of meteorological variables from 302 station locations distributed across Switzerland's complex topography and is complemented with topographical indices derived from digital height models for context. Ensemble forecasts from the currently operational high-resolution NWP model are provided as a baseline forecast against which to evaluate new approaches. The dataset's richness supports a broad spectrum of spatiotemporal tasks, including time series forecasting at various scales, graph structure learning, imputation, and virtual sensing. As such, PeakWeather serves as a real-world benchmark to advance both foundational machine learning research, meteorology, and sensor-based applications.

Problem

• Providing high-quality weather data for spatiotemporal deep learning
• Enhancing weather prediction accuracy using machine learning techniques
• Supporting diverse meteorological tasks with comprehensive dataset

Innovation

• High-frequency surface weather dataset
• Spatiotemporal deep learning applications
• Includes topographical and NWP data