AGST-WindFormer: A Forecasting Model for Multi-Station Near-Surface Winds
- 1College of Intelligent Systems Science and Engineering, Harbin Engineering University, Harbin 150001, China
Abstract
Accurate forecasts of near-surface winds are important for low-altitude aviation, unmanned aerial vehicle operations, and local weather services. Compared with single-station forecasting, multi-station forecasting can exploit complementary spatial information to improve wind-field representation. However, pronounced spatial heterogeneity and complex inter-station dependencies make this task inherently a spatially structured multivariate time-series forecasting problem. To address these issues, we propose AGST-WindFormer, an Adaptive-Graph Spatiotemporal Gated Wind Transformer for directly forecasting the horizontal wind components U and V. A shared Transformer captures station-wise temporal dependencies, while node embeddings generate an adaptive static graph shared across samples. A feature-wise gate then integrates the temporal and graph-aggregated representations, and a multi-output head predicts future U and V values. Experiments used hourly Meteostat observations from 854 stations during 2022–2025. The model achieved a mean wind-speed RMSE of 1.507 ± 0.009 m/s over the 10-h forecast horizon and a wind-direction MAE of 23.5 ± 0.1° for observations with wind speed ≥ 2 m/s. Relative to the strongest baseline for each metric, errors decreased by 1.1–3.2%. Ablations identified cross-station aggregation and nonuniform learned edges as the principal contributors. These results indicate that AGST-WindFormer can effectively exploit the spatiotemporal information contained in multi-station near-surface wind observations and improve forecasting accuracy across multiple consecutive future time steps.
Related
- Station-specific variability and short-term forecasting of a particulate air-quality index using machine learning in Charleroi, Belgium
- A Station-Based Evaluation of Machine Learning-based Weather Forecasting Models in Northern Norway
- Station-specific thunderstorm forecasting using climate predictors with seasonal decomposition analysis in Bangladesh’s most vulnerable regions
- KiloDA: Reconstructing kilometer-scale near-surface wind states from sparse station observations
- Mesoscale Modeling of Titan’s Shangri‐La Region: The Impact of Surface Properties on Winds
BibTeX
@article{li2026agstwindformer,
title = {{AGST-WindFormer: A Forecasting Model for Multi-Station Near-Surface Winds}},
author = {Chun Li and Yuankang Ye and Shihang Chen and Feng Gao},
journal = {Atmosphere},
year = {2026},
month = oct,
doi = {10.3390/atmos17100980},
url = {https://doi.org/10.3390/atmos17100980},
note = {CC BY},
}
RIS
TY - JOUR
TI - AGST-WindFormer: A Forecasting Model for Multi-Station Near-Surface Winds
AU - Chun Li
AU - Yuankang Ye
AU - Shihang Chen
AU - Feng Gao
PY - 2026
DA - 2026/10/26
JO - Atmosphere
DO - 10.3390/atmos17100980
UR - https://doi.org/10.3390/atmos17100980
AB - Accurate forecasts of near-surface winds are important for low-altitude aviation, unmanned aerial vehicle operations, and local weather services. Compared with single-station forecasting, multi-station forecasting can exploit complementary spatial information to improve wind-field representation. However, pronounced spatial heterogeneity and complex inter-station dependencies make this task inherently a spatially structured multivariate time-series forecasting problem. To address these issues, we propose AGST-WindFormer, an Adaptive-Graph Spatiotemporal Gated Wind Transformer for directly forecasting the horizontal wind components U and V. A shared Transformer captures station-wise temporal dependencies, while node embeddings generate an adaptive static graph shared across samples. A feature-wise gate then integrates the temporal and graph-aggregated representations, and a multi-output head predicts future U and V values. Experiments used hourly Meteostat observations from 854 stations during 2022–2025. The model achieved a mean wind-speed RMSE of 1.507 ± 0.009 m/s over the 10-h forecast horizon and a wind-direction MAE of 23.5 ± 0.1° for observations with wind speed ≥ 2 m/s. Relative to the strongest baseline for each metric, errors decreased by 1.1–3.2%. Ablations identified cross-station aggregation and nonuniform learned edges as the principal contributors. These results indicate that AGST-WindFormer can effectively exploit the spatiotemporal information contained in multi-station near-surface wind observations and improve forecasting accuracy across multiple consecutive future time steps.
SN - 2073-4433
ER -