Su et al. (2026) SynxFlow-based urban pluvial flood simulation and sensitivity evaluation in the central urban area of Shenzhen

Identification

• Journal: Journal of Hydrology Regional Studies
• Year: 2026
• Date: 2026-03-13
• Authors: Peng Su, Wei Xu, Xiaodong Ming, Zhihui Gu, Guisong Xia
• DOI: 10.1016/j.ejrh.2026.103325

Research Groups

• State Key Laboratory of Earth Surface Processes and Disaster Risk Reduction, Faculty of Geographical Science, Beijing Normal University, Beijing, China
• Academy of Disaster Reduction and Emergency Management, Ministry of Emergency Management and Ministry of Education, Beijing, China
• Key Laboratory of Environmental Change and Natural Disaster of Ministry of Education, Beijing Normal University, Beijing, China
• Risk Control & Technology Department, Ping An P&C Insurance Company of China, Shenzhen, Guangdong, China
• School of Architecture & Urban Planning, Shenzhen University, Shenzhen, Guangdong, China
• School of Artificial Intelligence, Wuhan University, Wuhan, Hubei, China

Short Summary

This study assesses the efficacy of SynxFlow, a newly developed open-source hydrodynamic model, for urban pluvial flood simulation and sensitivity evaluation in the central urban area of Shenzhen. It demonstrates SynxFlow's robust performance and highlights the critical role of integrated drainage modules for enhancing flood resilience planning in mega-cities.

Objective

• To evaluate the applicability and efficacy of the SynxFlow hydrodynamic model for urban pluvial flood simulation and sensitivity analysis in the central urban area of Shenzhen across six rainfall return periods (2-, 5-, 10-, 20-, 50-, and 100-year).

Study Configuration

• Spatial Scale: Central urban area of Shenzhen, China, encompassing Nanshan, Futian, and Luohu districts, covering an area of 342.76 square kilometers. The Digital Surface Model (DSM) resolution was 5 meters, and validation simulations were conducted at 1 meter resolution around monitoring sites.
• Temporal Scale: Six synthetic rainfall events with return periods of 2, 5, 10, 20, 50, and 100 years, each with a fixed duration of two hours. Simulations were run for a total of three hours. Model validation utilized six observed flood events in 2024 and the "9⋅7 Rainstorm" event in 2023.

Methodology and Data

• Models used:
  • SynxFlow: An open-source, Python-based, GPU-accelerated 2D hydrodynamic model that solves the Shallow Water Equations (SWE) using a Godunov-type shock-capturing scheme and features a built-in drainage module.
  • Lisflood-FP: A widely used simplified hydrodynamic model, employed for comparative simulations.
• Data sources:
  • Digital Surface Model (DSM): 5-meter resolution, derived from elevation points and contour lines obtained from the Resource and Environmental Science Data Platform, combined with urban building data.
  • Rainfall Input: Synthetic rainfall events generated using Shenzhen’s local Intensity–Duration–Frequency (IDF) formula and the Chicago Design Storm method, based on 30 years (1994–2023) of continuous rainfall records.
  • Drainage Capability: Subdistrict-level drainage parameters derived from one-hour surface water accumulation thresholds specified in the Shenzhen Meteorological Disaster Risk Warning Guidelines (2023 edition).
  • Model Validation Data:
    • Observed inundation depths from six urban pluvial flood events at three monitoring stations (Luohu District Committee, Shangbu–Hongli Road intersection, Guangshen Expressway Bridge) provided by the Water Authority of Shenzhen Municipality.
    • Media reports (collected via Baidu search engine) describing 123 inundation locations and approximate severity during the "9⋅7 Shenzhen Rainstorm" event in 2023.

Main Results

• SynxFlow demonstrated robust performance, with an average deviation of less than 0.04 meters in depth estimation and a 90.24% capture rate for media-reported flood points, indicating high spatial consistency.
• Comparative simulations with Lisflood-FP showed strong agreement in water depth distribution, with SynxFlow generally simulating marginally lower inundation depths (average bias of -0.021 meters), attributed to differences in drainage modules.
• Urban pluvial flood hazards consistently escalated with increasing rainfall return periods:
  • The inundated area expanded from 4.66% (2-year return period) to 13.74% (100-year return period) of the study area.
  • The 99.5th percentile depth increased from 0.72 meters (2-year) to 1.88 meters (100-year).
  • Areas with deep inundation (>1.0 meters) rose sharply from 0.28% in the 2-year scenario to 1.56% in the 100-year scenario.
• The inundation extent exhibited a non-linear, logarithmic response to rainfall intensity, with the most rapid expansion occurring between the 2- and 20-year return periods.
• A distinct north–south gradient in flood sensitivity was observed, with low-lying southern subdistricts (e.g., Nanshan, Futian) showing significantly higher sensitivity to rainfall return periods compared to northern high-elevation areas, primarily influenced by topography.

Contributions

• Provided the first benchmark application and comprehensive evaluation of the emerging open-source SynxFlow model in a high-density urban environment, addressing a critical gap in its practical application.
• Conducted a rigorous comparative analysis against the established Lisflood-FP model, verifying SynxFlow's reliability and highlighting the impact of integrated drainage modules.
• Introduced a novel, evidence-based framework for drainage parameter selection, utilizing official disaster warning guidelines, which offers a robust solution for parameterizing drainage in data-scarce urban regions.
• Underscored the critical importance of integrating explicit drainage modules within open-source hydrodynamic frameworks for enhancing flood resilience planning and management in mega-cities.
• Delivered an efficient, reliable, and accessible framework for urban flood hazard assessment, with actionable insights for Shenzhen and transferable strategies for other flood-prone cities globally.

Funding

• National Natural Science Foundation of China (grant no. U22B2011)
• Ministry of Science and Technology of the People's Republic of China (grant no. 2023YFC3008505)

Citation

@article{Su2026SynxFlowbased,
  author = {Su, Peng and Xu, Wei and Ming, Xiaodong and Gu, Zhihui and Xia, Guisong},
  title = {SynxFlow-based urban pluvial flood simulation and sensitivity evaluation in the central urban area of Shenzhen},
  journal = {Journal of Hydrology Regional Studies},
  year = {2026},
  doi = {10.1016/j.ejrh.2026.103325},
  url = {https://doi.org/10.1016/j.ejrh.2026.103325}
}