Based on meteorological observations, traffic flow data and information of traffic accidents caused by fog or agglomerate fog along the expressways in Jiangsu Province and Anhui Province in China from 2012 to 2021, key impact factors including meteorological conditions, road hidden dangers and traffic flow conditions are integrated to establish the prediction model for risk levels of expressway agglomerate fog-related accidents. This model takes the discrimination of the occurrence conditions of agglomerate fog as the starting term, and determines the hazard levels of agglomerate fog-related accidents by introducing the probability prediction value of meteorological conditions for fog-related accident as the disaster-causing factor. On this basis, the hourly road traffic flow and the location of road sections with a hidden danger of agglomerate fog are taken as traffic and road factors to construct the correction scheme for the hazard levels, and the final predicted risk level of agglomerate fog-related accident is obtained. The results show that for the criteria of disaster-causing factor classification threshold, 72.3% of fog-related accidents correspond to a hazard of a medium level or above, and 86.2% of the road traffic flow conditions are consistent with the levels of the traffic factor defined based on parametric indexes. For risk level prediction, six out of the seven agglomerate fog-related accidents correspond to the level of higher risk or above, which can help provide meteorological support for traffic safety under severe weather conditions. Moreover, the model takes into account the impacts of traffic flow and the road environment, which is conducive to further improving the reliability of the risk assessment results.

Lightning-caused fires (LCFs) are an important forest disturbance closely associated with meteorological conditions. However, reliable non-fire samples are difficult to define because the absence of recorded fires does not necessarily indicate unsuitable fire conditions, limiting conventional binary classification. The purpose of this study is to develop a positive-only framework for assessing the meteorological compatibility of LCFs using verified historical fire records. Using 984 verified LCF records from the Daxing’anling forest region, China, a Positive-Only Spatiotemporal Conditional Normalizing Flow with Context-Mismatch Learning (PO-STCNF-CM) was developed to characterize LCF-associated meteorological conditions across spatial and seasonal contexts. The model incorporates same-day event weighting and context-mismatch learning, and a Meteorological Compatibility Index (MCI) converts density-based anomaly scores into an interpretable 0–100 scale. GMM provided strong likelihood fitting, while ablation analysis showed that spatiotemporal conditioning improved neural density fitting and context-mismatch learning strengthened contextual discrimination. For the held-out records, PO-STCNF-CM produced a mean MCI of 53.761, a median MCI of 57.169, and an 80% coverage error of 0.015. The proposed framework provides an interpretable approach for characterizing LCF-associated meteorological conditions without requiring artificial non-fire samples.