A Station-Based Evaluation of Machine Learning-based Weather Forecasting Models in Northern Norway
Abstract
Recent machine learning weather prediction (MLWP) models have demonstrated remarkable forecasting skill on global reanalysis-based benchmarks. However, their performance remains unclear in challenging environments such as Northern Norway, where narrow fjords and rapidly changing weather result in highly variable local wind conditions. In this case study, we evaluate FourCastNet3 (FCN3), GraphCast, and ECMWF High Resolution Forecast (HRES) for wind speed forecasting using multi-year station observations from Northern Norway, focusing on their relative performance, generalization beyond the training period, and performance under high-wind conditions. Our results show that HRES slightly outperforms FCN3 and GraphCast, with an overall RMSE of 2.89 m s−1, compared to 2.96 m s−1 for FCN3 and 2.94 m s−1 for GraphCast. Notably, the MLWP models maintain comparable performance beyond their respective training periods, with no clear evidence of noticeable degradation. FCN3 performs best under high-wind conditions, although all models substantially underestimate strong winds. Our findings suggest that MLWP has become competitive with NWP for local wind, but further refinements are still needed to capture complex terrain better.
Related
- Station-specific variability and short-term forecasting of a particulate air-quality index using machine learning in Charleroi, Belgium
- Hybrid machine learning framework for drought forecasting: integration of LSBoost, RQNN, CNN-RNN, CNN-LSTM and advanced data decomposition
- Advancing Near-Term Water Quality Forecasting with Explainable Machine Learning and Probabilistic Uncertainty Quantification
- Weather forecasting in a changing climate: the rise of AI and Machine learning?
- AGST-WindFormer: A Forecasting Model for Multi-Station Near-Surface Winds
BibTeX
@misc{chen2026stationbased,
title = {{A Station-Based Evaluation of Machine Learning-based Weather Forecasting Models in Northern Norway}},
author = {Siyan Chen and Lars Uebbing and Eirik Mikal Samuelsen and Georgios Leontidis and Arnt-Børre Salberg and Sébastien Lefèvre and Robert Jenssen and Kristoffer Wickstrøm},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.10564},
eprint = {2609.10564},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.10564},
}
RIS
TY - UNPB
TI - A Station-Based Evaluation of Machine Learning-based Weather Forecasting Models in Northern Norway
AU - Siyan Chen
AU - Lars Uebbing
AU - Eirik Mikal Samuelsen
AU - Georgios Leontidis
AU - Arnt-Børre Salberg
AU - Sébastien Lefèvre
AU - Robert Jenssen
AU - Kristoffer Wickstrøm
PY - 2026
DA - 2026/09/01
PB - arXiv
DO - 10.48550/arxiv.2609.10564
UR - https://arxiv.org/abs/2609.10564
AB - Recent machine learning weather prediction (MLWP) models have demonstrated remarkable forecasting skill on global reanalysis-based benchmarks. However, their performance remains unclear in challenging environments such as Northern Norway, where narrow fjords and rapidly changing weather result in highly variable local wind conditions. In this case study, we evaluate FourCastNet3 (FCN3), GraphCast, and ECMWF High Resolution Forecast (HRES) for wind speed forecasting using multi-year station observations from Northern Norway, focusing on their relative performance, generalization beyond the training period, and performance under high-wind conditions. Our results show that HRES slightly outperforms FCN3 and GraphCast, with an overall RMSE of 2.89 m s^−1, compared to 2.96 m s^−1 for FCN3 and 2.94 m s^−1 for GraphCast. Notably, the MLWP models maintain comparable performance beyond their respective training periods, with no clear evidence of noticeable degradation. FCN3 performs best under high-wind conditions, although all models substantially underestimate strong winds. Our findings suggest that MLWP has become competitive with NWP for local wind, but further refinements are still needed to capture complex terrain better.
ER -