SynxFlow-based urban pluvial flood simulation and sensitivity evaluation in the central urban area of Shenzhen

Shenzhen, located in southern Guangdong Province, is a core city within the Guangdong–Hong Kong–Macao Greater Bay Area in China. This study aims to assess the efficacy of SynxFlow, a newly developed open-source hydrodynamic model, by conducting an urban pluvial flood simulation and sensitivity evaluation in the central urban area of Shenzhen across six rainfall return periods (2-, 5-, 10-, 20-, 50-, and 100-year). The research methodology employs SynxFlow's 2D hydrodynamic engine, which utilizes a Godunov-type shock-capturing scheme to solve the Shallow Water Equations (SWE) to robustly simulate complex inundation dynamics. Model validation was rigorously conducted using observed depths from six flood events and media-reported data from the “9·7 Rainstorm” event. The results demonstrate SynxFlow’s robust performance, characterized by high precision in depth estimation (average deviation < 0.04 m) and strong spatial consistency (90.24% capture rate). Comparative simulations with the widely used Lisflood-FP model showed broad agreement. Additionally, a slight systematic deviation was observed, with SynxFlow generally yielding marginally lower inundation depths than Lisflood-FP. This discrepancy, corresponding to an average bias of −0.021 m, is likely attributable to differences in the drainage modules of the two models. This study addresses a key gap in the application of the emerging SynxFlow model through a benchmark application in a high-density urban setting. Results indicated significant sensitivity to rainfall intensity: the inundated area expanded from 4.66% (2-year) to 13.74% (100-year), while the 99.5th percentile depth rose from 0.72 m to 1.88 m. Notably, the inundation extent exhibits a non-linear response to rainfall intensity, with the expansion being most rapid between 2- and 20-year return periods. Ultimately, this study underscores the critical role of integrating explicit drainage modules within open-source hydrodynamic frameworks, offering a scalable and cost-effective pathway for enhancing flood resilience planning in high-density mega-cities. • Validates SynxFlow for high-density urban flood modeling in Shenzhen. • Benchmarks SynxFlow against Lisflood-FP showing strong consistency. • Flood sensitivity follows a logarithmic trend, rising steepest at 2- to 20-year return periods.