ABSTRACT The evaluation of emergency plans is crucial for building a high‐quality emergency management system. However, existing methods often rely on qualitative judgement and lack systematic, quantitative tools. This study developed a quantitative evaluation method, the PMC‐AE index, which combined the Policy Modeling Consistency (PMC) index with an autoencoder (AE) and integrated text‐mining techniques (e.g., term frequency, co‐word network and cluster analysis). We applied this method to 484 meteorological disaster emergency plans from the national to county levels in Guangdong Province, China. To examine hierarchical policy transmission, the framework was further extended by introducing three indicators: coverage, degree of detail and grey relational analysis (GRA). The results revealed that although the lower‐level plans were more detailed and operational, vertical linkage remained weak, exhibiting a pattern of ‘high coverage—low detail—high correlation’. A high degree of homogeneity was found between municipal and county plans (mean GRA values of 0.729 and 0.760, respectively), indicating insufficient adaptation to local contexts and thus limiting their practicality. The proposed method provides a novel pathway for the scientific evaluation and optimisation of emergency plans. Trial Registration N/A.
Cao et al. (Wed,) studied this question.