Vallés et al. (2025) SERGHEI v2.1: a Lagrangian model for passive particle transport using a two-dimensional shallow water model (SERGHEI-LPT)

Identification

• Journal: Geoscientific model development
• Year: 2025
• Date: 2025-10-17
• Authors: Pablo Vallés, Mario Morales‐Hernández, Volker Roeber, Pilar García‐Navarro, Daniel Caviedes‐Voullième
• DOI: 10.5194/gmd-18-7399-2025

Research Groups

• I3A, University of Zaragoza, Zaragoza, Spain
• E2S Chair HPC-Waves, University of Pau, Anglet, France
• Institute of Bio- and Geosciences: Agrosphere (IBG-3), Forschungszentrum Jülich, Jülich, Germany
• Simulation and Data Lab. Terrestrial Systems, Jülich Supercomputing Centre (JSC), Forschungszentrum Jülich, Jülich, Germany

Short Summary

This paper introduces SERGHEI v2.1, a new Lagrangian particle transport (LPT) model coupled with a 2D shallow water model, designed to simulate passive particle advection and turbulent diffusion. The study evaluates the accuracy and computational efficiency of various numerical schemes, concluding that the online Euler method offers the best compromise for large-scale applications.

Objective

• To present the capabilities of the new Lagrangian particle transport (LPT) model and provide a comprehensive analysis of the accuracy and computational efficiency of its numerical schemes and implementation for particle transport.

Study Configuration

• Spatial Scale:
  • Steady circular vortex: 100 m radius.
  • Transient dam break: 36 m length, 3.6 m width.
  • Channel with cavities: 0.02 m × 0.02 m and 0.2 m × 0.2 m grid resolutions.
  • Arnás catchment: 5 m resolution, 1.1 × 10^5 computational cells.
• Temporal Scale:
  • Steady circular vortex: Particle completed four full revolutions.
  • Transient dam break: States analyzed up to 20 s; Runge–Kutta offline updated every 5, 10, 20, or 50 hydrodynamic time steps.
  • Arnás catchment: Simulation run times of 317.75 s and 177.02 s for two precipitation events, which lasted approximately 10 hours and 15 hours, respectively.

Methodology and Data

• Models used:
  • SERGHEI v2.1 (Simulation Environment for Geomorphology, Hydrodynamics, and Ecohydrology in Integrated form).
  • Lagrangian Particle Transport (LPT) model for passive particles.
  • Two-dimensional (2D) Shallow Water Equations (SWE) model for hydrodynamics (explicit upwind finite-volume scheme, augmented Roe–Riemann solver).
  • Numerical schemes for LPT: Online first-order Euler method, online fourth-order Runge–Kutta method, and offline fourth-order Runge–Kutta method.
  • Turbulent diffusion modeled using a random-walk approach with velocity-dependent anisotropic diffusivity coefficients.
• Data sources:
  • Analytical solutions for the steady circular vortex test case.
  • Experimental data from laboratory setups (e.g., dam break by Soares-Frazão and Zech, 2008).
  • Real precipitation event data (hyetographs) and topography for the Arnás catchment.

Main Results

• The online Euler method demonstrated the best compromise between accuracy and computational efficiency across various test cases, especially for large numbers of particles (e.g., 10^5 particles: Euler 21.22 s vs. Runge–Kutta online 121.93 s).
• The implemented trajectory algorithm effectively handles dry–wet interfaces and obstacles, preventing unphysical particle movements.
• Incorporating a random-walk model for turbulent diffusion significantly improved the realism of particle trajectories and dispersion, particularly in complex geometries like channels with cavities.
• Coarser grid resolutions in the hydrodynamic model can hinder particle entry into small features (e.g., cavities), a limitation partially mitigated by the random-walk dispersion model.
• Application to the Arnás catchment showed that particles are primarily transported through main gullies, with longer and less intense precipitation events leading to greater travel distances and shorter stop times for some particles (up to 10 km).

Contributions

• Development and validation of SERGHEI v2.1, a novel Lagrangian particle transport (LPT) model integrated into the performance-portable SERGHEI framework, capable of simulating passive particle movement in 2D shallow water flows with turbulent diffusion.
• A comprehensive comparative analysis of numerical schemes for LPT, establishing the online Euler method as the most suitable for large-scale, computationally efficient simulations while maintaining high accuracy.
• Implementation of a robust particle trajectory algorithm that accurately handles complex hydrological features such as dry–wet interfaces and topographical obstacles.
• Demonstration of the model's utility in ecohydrological applications, providing insights into particle transport dynamics, hydrological pathways, and potential accumulation zones in real-world catchments.

Funding

• UPPA-UNIZAR Research Grant PI-PRD/2022-03
• Erasmus+ KA103 “IBERUSC” (2021-1-ES01-KA130-HED-000004265)
• Government of Aragón (Fluid Dynamics Technologies, T32_23R)
• Universidad de Zaragoza (project JIUZ2023-IA-04)
• Project PID2022-137334NB-I00 funded by MCIN/AEI/10.13039/501100011033 and ERDF/EU
• Institutional support from Forschungszentrum Jülich
• Article processing charges covered by Forschungszentrum Jülich

Citation

@article{Vallés2025SERGHEI,
  author = {Vallés, Pablo and Morales‐Hernández, Mario and Roeber, Volker and García‐Navarro, Pilar and Caviedes‐Voullième, Daniel},
  title = {SERGHEI v2.1: a Lagrangian model for passive particle transport using a two-dimensional shallow water model (SERGHEI-LPT)},
  journal = {Geoscientific model development},
  year = {2025},
  doi = {10.5194/gmd-18-7399-2025},
  url = {https://doi.org/10.5194/gmd-18-7399-2025}
}