A physics-guided machine learning framework for enhancing dust storm visibility prediction in arid and semi-arid regions
Authors:
Open Access
- 3 March 2026
- journal article
- research article
- Published by Springer Nature in Scientific Reports
- Vol. 16 (1)
Abstract
Dust sharply degrades visibility in arid and semi-arid regions, yet operational forecasting remains challenged by near-surface process errors in numerical weather prediction (NWP) and the poor generalization of purely data-driven models. We present a physics-guided machine learning (PGML) framework that post-processes European Centre for Medium-Range Weather Forecasts (ECMWF) forecasts to predict five ordinal visibility grades over the Kumtag Desert. A dust-lifecycle feature library (emission, vertical mixing, transport, wet scavenging) is coupled with an ordinal LightGBM architecture. On an independent test period, the model attains quadratic weighted kappa (QWK) of 0.26 (0–24 h), 0.17 (24–48 h), and 0.18 (48–72 h), with mean absolute error (MAE) of 0.48–0.56; gains versus data-only baselines increase with forecast horizon. Ablation experiments show that physics priors can effectively improve visibility prediction accuracy—reducing MAE by up to 10% and sustaining QWK beyond 24 h by constraining non-physical drift. Accordingly, the PGML visibility predictions show improved performance relative to data-only baselines. SHAP analysis reveals a forecast-horizon-dependent mechanistic shift from emission/surface-layer dynamics to stability-controlled vertical mixing, consistent with dust dynamics. The framework offers an interpretable, transferable paradigm for physics-constrained environmental forecasting.Keywords
Funding Information
- Xinjiang Uygur Autonomous Region Department of Education (XJ2024G083)
- Shandong Provincial Department of Science and Technology
This publication has 51 references indexed in Scilit:
- Deep learning ensembles for accurate fog-related low-visibility events forecastingNeurocomputing, 2023
- Early warning of low visibility using the ensembling of machine learning approaches for aviation services at Jay Prakash Narayan International (JPNI) Airport PatnaSN Applied Sciences, 2023
- Deep learning models for visibility forecasting using climatological dataInternational Journal of Forecasting, 2022
- Current challenges of improving visibility due to increasing nitrate fraction in PM2.5 during the haze days in Beijing, ChinaEnvironmental Pollution, 2021
- Development of a new visibility parameterization based on the measurement of fog microphysics at a mountain site in KoreaAtmospheric Research, 2019
- Analysis of a severe dust storm and its impact on air quality conditions using WRF-Chem modeling, satellite imagery, and ground observationsAir Quality, Atmosphere & Health, 2019
- An intercomparison of long‐term planetary boundary layer heights retrieved from CALIPSO, ground‐based lidar, and radiosonde measurements over Hong KongJournal of Geophysical Research: Atmospheres, 2017
- The Persian Gulf summertime low‐level jet over sloping terrainQuarterly Journal of the Royal Meteorological Society, 2011
- An examination of the physical and optical properties of aerosols collected in the IMPROVE programAtmospheric Environment, 2007
- Wind modification and bed response during saltation of sand in air1991