Shuting Yu, Shunjiang Ni, Shifei Shen
The profound uncertainty surrounding emerging infectious diseases, as exemplified by the cross-seasonal dynamics of COVID-19, poses significant challenges to conventional forecasting efforts. Traditional models often rely heavily on granular population flow data, limiting their applicability across city, national, and global scales. This study aims to shift the paradigm from deterministic prediction to multi-scale scenario projection. We propose an extended SEQIHRS compartmental model integrating key epidemic drivers. The core innovation is the introduction of a space-induced immunity factor ( \({k}_{0}\) ), which characterizes transmission barriers across geographical scales without requiring explicit mobility data. A periodic skewness function captures asymmetric seasonal effects, and a composite intervention framework simulates adaptive control measures. Three empirical datasets from Hong Kong, Mainland China, and the Global scale were utilized to calibrate and validate the model's structural plausibility. The model successfully reproduced the observed epidemic trends across diverse spatiotemporal settings. The close alignment between simulated curves and surveillance data demonstrates that the model can capture complex transmission patterns under various parameterizations, confirming its robustness as a scenario-generating engine. Compared to traditional approaches, this framework offers three breakthroughs: (1) It achieves seamless multi-scale simulation (local to global) via the \({k}_{0}\) factor, bypassing the need for complex population flow data; (2) It incorporates uncertainty through interpretable parameters, allowing users to explore a range of plausible outcomes rather than a single forecast; (3) It provides a practical tool for pandemic preparedness. By adjusting pathogen parameters, spatial scales, and intervention intensities, policymakers can conduct "what-if" analyses to support adaptive emergency responses.