This study is the first to use the observation data of a fog monitor, a visibility meter, and an automatic weather station to carry out a comprehensive observation experiment from the perspective of microphysics on a severe sea fog process in Beilun District, China, from 14 to 15 June 2021. The results show the following: (1) Temperature is closely related to nucleation, condensation growth, and other processes. The decrease (increase) in temperature is the main reason for the enhancement (weakening) of nucleation and the growth of condensation (evaporation of droplets), which leads to an increase (or decrease) in microphysical quantities, such as droplet number concentration and liquid water content. (2) The average droplet number spectral distribution roughly conforms to the Gamma distribution, and the spectral distribution of the fog process presents a ”multi-peak” structure, with peak diameters of 6 μm, 12 μm, 16 μm, 24 μm, and 44 μm. Droplets with a diameter of less than 16 μm account for 75% of the droplet size distribution. (3) During this sea fog process, three microphysical parameters, namely, number concentration, liquid water content, and average diameter, are all positively correlated in pairs, but the positive correlation between the number concentration and the average diameter is weak. This shows that the condensation nucleation and the condensation growth of droplets are the main processes in this sea fog process and that the collision process occurs but is not the dominant process. The sea fog comprehensive observation experiment provides an important demonstration of the microphysics research of sea fog in the eastern coastal areas of China and provides more reference information for sea fog research and equipment comparisons between different regions. At the same time, it also provides an essential scientific basis for the short-term forecast of sea fog in the future and for the optimization of the microphysical parameters of related models.

Wind direction time series exhibit angular periodic discontinuity, multi-scale non-stationary fluctuations and abrupt wind shifts, which hinder the precision of short-term forecasting for wind turbine yaw control. In this paper, a dual-branch forecasting framework based on Variational Mode Decomposition (VMD) and Transformer is developed to address the above drawbacks, with a hysteresis gating and zoning residual compensation module embedded for targeted error correction. First, sine–cosine encoding is adopted to eliminate the numerical discontinuity between 0° and 360° for wind direction angular data, and valid meteorological input features are screened to discard redundant covariates. Second, the sine–cosine-encoded wind direction sequence is decomposed into multiple band-limited intrinsic mode functions (IMFs) via VMD, extracting frequency-specific features that reduce non-stationarity and facilitate subsequent Transformer modeling. The standard Transformer encoder serves as the normal branch to capture long-range temporal dependencies across the whole time series, while a lightweight multilayer perceptron (MLP) constitutes the compensation branch to learn prediction deviations between baseline predictions and ground-truth values. The hysteresis gating unit activates residual compensation based on historical prediction errors and angular variation, without requiring access to the current ground-truth value, and compensation intensity is adaptively adjusted via the zoning strategy; relevant coefficients are optimized by random search. Verified on a real wind farm dataset consisting of 10,421 15 min sampling points, the proposed model achieves the lowest MAE of 9.64° among six benchmark models. For the improved genuine mutation samples (angle change > 70°), the model achieves a mean improvement of 6.02°. Ablation experiments verify that VMD preprocessing, the MLP compensation branch, and the hysteresis gating mechanism play indispensable roles in forecasting performance. The proposed framework can support accurate yaw control of wind turbines, and the decomposition–compensation workflow can also be generalized to other periodic non-stationary forecasting tasks.