Tedla et al. (2026) Modelling groundwater flow and heat transport in freezing–thawing riverbanks: a fully coupled numerical framework

Identification

• Journal: Advances in Water Resources
• Year: 2026
• Date: 2026-11-01
• Authors: Hailay Zeray Tedla, Erik van Rooijen, Eliisa Lotsari
• DOI: 10.1016/j.advwatres.2026.105562

Research Groups

• Water and Environmental Engineering, Department of Built Environment, Aalto University, Espoo, Finland
• Rijkswaterstaat, Water, Verkeer en Leefomgeving, Utrecht, The Netherlands

Short Summary

This study presents a new fully coupled two-dimensional finite-volume framework to simulate transient groundwater flow and heat transport with phase change in freezing-thawing riverbanks. It reveals how advective heat transport, soil stratification, and bank geometry control hydraulic head evolution and moisture redistribution, which are critical for bank erosion processes.

Objective

• How does advective groundwater heat flux control hydraulic head evolution under dynamic river-stage and groundwater boundary conditions, with particular focus on cryosuction-driven and thaw-induced moisture redistribution?
• How do soil stratification and bank geometry modulate the spatial pattern and magnitude of coupled hydraulic head dynamics during freezing and thawing?

Study Configuration

• Spatial Scale: Two-dimensional (2D) numerical model domain (e.g., 4 m × 3 m for verification, hypothetical riverbank profiles up to 3 m high and 30 m wide). Structured mesh with uniform spatial resolution of 0.1 m in both horizontal (Δx) and vertical (Δz) directions.
• Temporal Scale: Transient simulations over periods up to 2,000,000 seconds (approximately 23 days). Adaptive time steps used, with initial steps of 0.005 s for freezing and 0.01 s for thawing. Verification simulations used 0.1 s time steps.

Methodology and Data

• Models used:
  • Fully coupled two-dimensional finite-volume numerical framework.
  • Mixed-form Richards equation for variably saturated groundwater flow.
  • Van Genuchten–Mualem formulation for soil hydraulic characteristics.
  • Transient energy balance equation for heat transport, accounting for conduction, advection, and latent heat exchange.
  • Zhang et al. (2016) Soil Freezing Characteristic Curve (SFCC) formulation.
  • Kulik (1978) formulation for cryosuction-corrected pressure head.
  • Implicit Newton–Raphson iterative scheme for solving the highly nonlinear coupled problem.
• Data sources:
  • Model verification against the Celia et al. (1990) benchmark for variably saturated groundwater flow.
  • Model verification against the Stefan analytical solution for the temperature model.
  • Hypothetical riverbank profiles (homogeneous and heterogeneous sand-silt-clay-sand) constructed based on field observations from the Pulmankijoki River, a sub-arctic river (Lotsari et al., 2020).
  • Soil hydraulic parameters from Carsel and Parrish (1988).
  • Thermal parameters after Farouki (1981) and de Vries (1963).
  • Fourier-series representation for time-varying temperature boundary conditions.

Main Results

• A new fully coupled numerical framework successfully resolved transient groundwater–heat transport with phase change in freezing–thawing riverbanks, demonstrating stable numerical performance (Nash–Sutcliffe efficiencies of 0.95 for hydraulic head and 0.97 for temperature, with mass and energy balance errors below 10⁻⁶ tolerance).
• Cryogenic suction generated upward hydraulic gradients that drove nonlinear moisture migration toward the advancing freezing front, leading to a 2–3 orders of magnitude reduction in hydraulic conductivity in upper layers.
• Advective heat transport significantly influenced freezing and thawing front dynamics, with a magnitude comparable to conduction heat transfer alone.
• During thawing, domain-averaged hydraulic heads exceeded pre-freezing values by approximately 120% under elevated groundwater forcing and 108% under elevated river-stage forcing, indicating a period of peak hydraulic head and increased erosion susceptibility.
• Layered soil profiles exhibited 30–40% slower hydraulic recovery during thawing compared to homogeneous profiles, sustaining transient perched saturation zones.
• Steep-slope configurations (35°) amplified thawing-phase domain-averaged hydraulic heads by approximately 17% compared to gentle slopes (15°).
• Sensitivity analysis identified saturated hydraulic conductivity (Ks), thermal conductivity (λ), and porosity (φ) as the dominant controls on hydraulic head evolution, with sensitivities ranging from 31% to 42%.

Contributions

• Development of the first fully coupled model to jointly resolve dynamic river-stage, thermal, and groundwater conditions across sloped, stratified seasonally frozen riverbanks.
• Quantification of the coupled advective–conductive thermal control on hydraulic head evolution in heterogeneous bank profiles.
• Elucidation of the compounding interaction among boundary conditions, soil stratigraphy, and bank geometry on riverbank erosion susceptibility.
• Identification of thawing conditions as the period of peak hydraulic head, highlighting its critical role in driving bank erosion processes in seasonally frozen environments.

Funding

No specific funding projects, programs, or reference codes were explicitly listed in the provided paper text.

Citation

@article{Tedla2026Modelling,
  author = {Tedla, Hailay Zeray and Rooijen, Erik van and Lotsari, Eliisa},
  title = {Modelling groundwater flow and heat transport in freezing–thawing riverbanks: a fully coupled numerical framework},
  journal = {Advances in Water Resources},
  year = {2026},
  doi = {10.1016/j.advwatres.2026.105562},
  url = {https://doi.org/10.1016/j.advwatres.2026.105562}
}