Pangu-Weather: A 3D High-Resolution Model for Fast and Accurate Global Weather Forecast

TL;DR

Pangu-Weather presents a data-driven approach to global, medium-range weather forecasting that surpasses traditional numerical weather prediction (NWP) in accuracy while maintaining high spatial resolution. The method introduces a 3D Earth-specific Transformer (3DEST) and a hierarchical temporal aggregation strategy, trained on decades of ERA5 reanalysis data to predict weather states across 13 pressure levels plus the surface. Results show AI-based forecasts outperform ECMWF IFS and previous AI models across all factors and lead times from 1 hour to 7 days, with strong capabilities for extreme-event tracking and ensemble forecasting, and fast inference enabling real-time applications. The work demonstrates that incorporating full 3D atmospheric structure and multi-timescale aggregation can close the gap with, and even surpass, physics-based NWP, offering practical benefits for timely warnings and probabilistic forecasting.

Abstract

In this paper, we present Pangu-Weather, a deep learning based system for fast and accurate global weather forecast. For this purpose, we establish a data-driven environment by downloading $43$ years of hourly global weather data from the 5th generation of ECMWF reanalysis (ERA5) data and train a few deep neural networks with about $256$ million parameters in total. The spatial resolution of forecast is $0.25^\circ\times0.25^\circ$, comparable to the ECMWF Integrated Forecast Systems (IFS). More importantly, for the first time, an AI-based method outperforms state-of-the-art numerical weather prediction (NWP) methods in terms of accuracy (latitude-weighted RMSE and ACC) of all factors (e.g., geopotential, specific humidity, wind speed, temperature, etc.) and in all time ranges (from one hour to one week). There are two key strategies to improve the prediction accuracy: (i) designing a 3D Earth Specific Transformer (3DEST) architecture that formulates the height (pressure level) information into cubic data, and (ii) applying a hierarchical temporal aggregation algorithm to alleviate cumulative forecast errors. In deterministic forecast, Pangu-Weather shows great advantages for short to medium-range forecast (i.e., forecast time ranges from one hour to one week). Pangu-Weather supports a wide range of downstream forecast scenarios, including extreme weather forecast (e.g., tropical cyclone tracking) and large-member ensemble forecast in real-time. Pangu-Weather not only ends the debate on whether AI-based methods can surpass conventional NWP methods, but also reveals novel directions for improving deep learning weather forecast systems.

Figures (13)

• Figure 1: A showcase of Pangu-Weather's forecast results. Top: Pangu-Weather claims significant advantages over operational IFS (NWP) and FourCastNet (AI-based) in terms of forecast accuracy (i) of different factors ($500\mathrm{hPa}$ geopotential, Z500, and $850\mathrm{hPa}$ temperature, T850) and (ii) with respect to different months in year. Middle: visualization of Pangu-Weather's $3$-day forecast of 2m temperature (T2M) and 10m wind speed at 00:00 UTC, September 1st, 2018, with comparison to the ERA5 ground-truth. Bottom: Pangu-Weather produces more accurate tracking for two tropical cyclones in 2018, i.e., Typhoon Kong-rey (2018-25) and Yutu (2018-26). Specifically, Pangu-Weather predicts the correct path of Yutu (i.e., it goes to the Philippines) $48$ hours earlier than the ECMWF-HRES forecast.
• Figure 2: An overview of the 3D Earth-specific transformer (3DEST). Based on the standard encode-decoder design, we (i) adjust the shifted-window mechanism and (ii) apply an Earth-specific positional bias -- see the main texts for details.
• Figure 3: The motivation of using an Earth-specific positional bias. Left: the horizontal map corresponds to an uneven spatial distribution on Earth's sphere. Middle: the geopotential height is closely related to the latitude. Right: the mean wind speed and temperature are closely related to the height (formulated as pressure levels).
• Figure 4: The curves showing cumulative forecast errors when one performs up to $7$-day forecast with the base lead time being $1$ hour, $3$ hour, $6$ hours, and $24$ hours, respectively. The statistics are performed in the March 2018 subset.
• Figure 5: The comparison of forecast accuracy in terms of latitude-weighted RMSE (lower is better) and ACC (higher is better) of four upper-air variables at the pressure level of $500\mathrm{hPa}$. Here, T, Q, U and V stand for temperature, specific humidity, $u$-component and $v$-component of wind speed, respectively.