Abstract
Seasonal human mobility plays a crucial role in economic growth, labor market dynamics, and public health management. Traditional mobility models often rely on geographic distance, which fails to capture the influence of destination popularity, such as family gatherings or economic opportunities. To address this, we introduce the effective intervening opportunity (EIO) model, which replaces geographic distance with a more accurate "effective distance" that integrates both destination appeal and proximity. We use mobility flow data from 400 million mobile users over 6 years to simulate China's chunyun-the world's largest annual human migration during the spring festival. The EIO model outperforms traditional models, accurately predicting mobility patterns not only during chunyun but also for other holidays, such as Labour Day and National Day. Our findings demonstrate that incorporating destination popularity into mobility models enhances their predictive power, offering a more accurate representation of seasonal human mobility dynamics.