A System Evolutionary Game Analysis of Tripartite Collaborative Disaster Governance Under Dynamic Reward–Punishment in the Guangdong–Hong Kong–Macao Greater Bay Area

Disaster governance in the Guangdong–Hong Kong–Macao Greater Bay Area (GBA) spans multiple administrative regions, institutions, and platforms, and hazards such as typhoons, storm surges, and urban waterlogging generate strong cross-boundary spillovers that no single city can manage alone. To explain the persistent problems of stakeholder hesitation, underinvestment, and free-riding, this study treats collaborative disaster governance as a multi-actor system and develops a tripartite evolutionary game model linking the regional coordinating regulatory authorities, local emergency management actors, and social participation actors. Static, linear dynamic, and nonlinear dynamic reward–punishment regimes are embedded into the payoff matrix and examined through stability analysis and numerical simulation. The results show that, without incentives, the system settles into an imbalanced “regulatory authorities-driven, weak collaboration” equilibrium; static mechanisms improve targeted behavior but erode the regulatory authorities’ willingness to lead; and linear dynamic mechanisms induce strategy oscillations in cross-regional settings. Under a parameter-adjusted nonlinear dynamic mechanism that combines incremental incentives with progressive constraints, the tested system converges toward “active leadership–proactive collaboration–active participation” under the assumed cost, benefit, and initial-condition settings. These findings provide a basis for designing differentiated subsidies, performance-based rewards, and liability recovery in GBA disaster governance.