ISSN 1000-0526
CN 11-2282/P
CN 11-2282/P
Analysis of sea fog characteristics and forecast evaluation of classification decision tree models at Yangshan Port
- Article
- Figures
- Metrics
- Preview PDF
- Reference
- Related
- Cited by
- Materials
Abstract:
Based on data from automatic weather stations, buoy observation stations, and ERA5 reanalysis spanning from 2015 to 2023, this study analyzes the characteristics of sea fog at Yangshan Port. Classification decision tree models were trained and validated using a comprehensive case database of sea fog events, and compared and verified their forcast results with those of the EC model. The results indicate that 2016 recorded the highest number of foggy days, with spring and early summer being the peak season, followed by winter. During dense fog events, the dominant wind directions ranged from northeast to north and southeast. Southeast winds prevailed during non-precipitation periods, while north winds dominated during precipitation. Monthly wind patterns transitioned from predominantly northerly in winter to northeasterly and southeasterly in spring. In the development stage of sea fog, southeast winds were dominant; during the mature stage, northeast winds prevailed; and during dissipation, north winds dominated. Fog events accompanied by precipitation were more frequent and longer-lasting. The classification decision tree models identified the temperature-dewpoint spread as a key factor in the formation of various sea fog types. Decision tree models demonstrated a lower miss rate and higher prediction performance than the EC model, particularly in forecasting the formation and duration of advection fog, while providing valuable insights into frontal and radiation fog events.
Keywords:
Project Supported:
Clc Number:
Mobile website
