Supplementary Material: Detailed Scene List and Metadata for the FYDD Sea Fog Dataset (2022–2023)

This dataset serves as the Supplementary Material for the research article: "DSST-Fog: A Dual-Stream Transformer Model for Sea Fog Nowcasting" To ensure reproducibility and address the reviewers' requests regarding data transparency, we provide the complete scene list of the FengYun Deep-learning Dataset (FYDD) used in our experiments. The list details the exact timestamps, satellite sources, and pixel-level statistics for all 12,847 samples. Dataset Specifications: Temporal Range: January 2022 – April 2023. Temporal Resolution: 15 minutes. Data Source: Primarily FY-4B AGRI (to maintain strict 15-min intervals) and FY-4A. Observation Window: Strictly filtered to UTC Daytime (00:00 – 08:15) to ensure the availability of visible spectral channels required by the dual-stream architecture. Spatial Grid: Data statistics are calculated based on the projected 0.1°×0.1° analysis grid (12,000 total pixels per frame). File Content: filename: The standardized identifier for the satellite data file. timestamp(utc): Observation time in UTC. fog pixels: Number of identified fog pixels within the study area (0.1° grid). dataset split: Designation of the sample (Train / Val / Test).