Pangu-Weather beats ECMWF-HRES on typhoon tracks near Hainan

Hohai University researchers found that Pangu-Weather's long-range forecasts of typhoon tracks for Guangdong and Hainan beat ECMWF-HRES. Its winds were too weak until corrected.

  • Not rated an organization

    The authors report forecast skill on past typhoons and no use by a forecasting or coastal agency.

Depth runs from 1, a convenience or time saved, to 5, a life or a livelihood preserved.

Status
Reported and intended benefits.
Evidence

foundstudiedreportedclaimed

An independent peer-reviewed study supports the benefit.

Compared with

ECMWF-HRES forecasts of the same typhoon tracks. Measured.

The system’s part
Not isolated. The abstract compares Pangu-Weather's tracks with ECMWF-HRES without method details or error figures. The gain in wave heights depended on a separate correction of the winds, so the model's share of it is not isolated.
Limits
Not yet checked.
Costs and harms
Not yet checked.
Later
Not yet checked.

Download data Cite Suggest a correction

Benefit type
Improved Task Accuracy
System
Pangu-Weather (盘古气象大模型) with a tropical cyclone tracking algorithm
Made by
Huawei Cloud
Where
Guangdong Province and Hainan Island, China
Stage
Tested in a laboratory
Date
, when the source was published
Kind
Aggregate result
Broad benefit
Estimating the Quantifiable Characteristics of Products, Events, or Information
Last revised
31 October 2026

Sources

  1. Peer-reviewed study, by authors with no tie to the system.

    Hu Z, Feng X, Feng W, et al. Research on prediction of typhoon tracks and typhoon wave heights based on Pangu-Weather [基于盘古气象大模型的台风路径及台风浪波高预报研究]. Journal of Hohai University (Natural Sciences) 2026, 54(3):28–36, 160.

    The results indicate that Pangu-Weather can achieve long-term typhoon track prediction, and the results are superior to those of the high-resolution numerical atmospheric model (ECMWF-HRES), but Pangu-Weather underestimates the typhoon wind field.

Details

AI system

Workflow

A tracking algorithm for tropical cyclones extracted typhoon tracks from Pangu-Weather's forecasts. A direct correction method improved the forecast winds before they drove the SWAN wave model.

Follow-up

Conditions for replication

Pangu-Weather's typhoon winds must be corrected before they drive a wave model. Measured wave heights are needed to check the forecasts.

Entry details

Sector

Weather forecasting / tropical cyclones.

Purpose

Huawei Cloud built Pangu-Weather to produce global weather forecasts for the medium range with a deep learning model.

Not yet checked: model, release, provider, role, configuration, inputs and outputs, autonomy, human oversight, AI contribution, beneficiaries, reach, distribution and access, duration, every measurement question, every money question, ownership and control, dependencies, cross-border connections, geopolitical implications, replication attempts, availability, institutional decision, grounds, action and result, success criterion, every evidence question, observed or intended, alternative assessments, related entries, other identifiers, category mappings.

Cite this record

Peter H. Vartanian. (2026). Pangu-Weather beats ECMWF-HRES on typhoon tracks near Hainan (AIBD-0765). Artificial Intelligence Benefits Database, edition of 31 October 2026. https://petervartanian.xyz/aibd/records/AIBD-0765.html (archived at https://doi.org/10.7910/DVN/AADSII, version 4.0)

BibTeX Data Field definitions