Spatiotemporal Similarity-Based Approach for Analyzing the Relationship Between Sea Fog Occurrence and Sea Level Pressure Distributions

This study presents a novel methodology for analyzing the relationship between sea level pressure (SLP) distributions and sea fog occurrences, focusing on a spatiotemporal similarity-based approach. Using SLP data from 2001 to 2019 and visibility observations from Baengnyeong Island (BYI), Yellow Sea, the proposed framework quantifies the connection between atmospheric pressure patterns and sea fog formation. The methodology integrates three key components: defining temporal and spatial domains, calculating weighted similarities, and validating the results using sea fog occurrence data. The temporal domain was set to a 7-hour period, determined by analyzing visibility trends prior to sea fog events. This period captures the critical atmospheric changes leading to fog formation. Spatially, a 2D weighted map was constructed using Pearson correlation coefficients between SLP variations at BYI and other locations in the study area. This weighting emphasizes regions with strong correlations, ensuring the analysis focuses on areas most relevant to sea fog dynamics. The Spatiotemporal Similarity Measure (STSM) method was then applied to compare reference SLP maps from 2017–2019 with historical SLP data from 2001–2015. By identifying historical cases with high similarity to reference conditions, the study examined the likelihood of sea fog occurrences under similar atmospheric setups. These similarities were categorized into thresholds, and their connection to sea fog events was evaluated using Probability of Detection (POD) and False Alarm Ratio (FAR) metrics. The results demonstrate that higher SLP similarity corresponds to increased POD and decreased FAR, validating the effectiveness of the STSM method. This approach highlights the role of recurring SLP patterns in sea fog formation and underscores the utility of historical data in improving sea fog forecasting.