[Paper Review] HydroPol2D -- Distributed Hydrodynamic and Water Quality Model: Challenges and Opportunities in Poorly-Gauged Catchments

Marcus N. Gomes, César Ambrogi Ferreira Lago+4 more|arXiv (Cornell University)|Apr 21, 2023

Hydrology and Watershed Management StudiesEnvironmental Science3 citations

TL;DR

HydroPol2D is a distributed 2D hydrodynamic and water quality model designed to simulate stormwater runoff quantity and pollutant transport in poorly gauged, urbanized catchments. It integrates Green-Ampt infiltration, build-up and wash-off processes, and momentum-based flow dynamics, demonstrating strong performance in simulating TSS concentrations (456 ± 260 mg·L⁻¹·km⁻²) and a high first-flush effect (89% ± 10% of total TSS washed in 30% of runoff volume) for a 1-year return period event.

ABSTRACT

Floods are one of the deadliest natural hazards and are fueled by excessive urbanization. Urban development decreases infiltration by reducing pervious areas and increases the accumulation of pollutants during dry weather. During wet weather events, there is an increase in the levels of pollution concentrations and stormwater runoff that eventually reach creeks and rivers. Polluted stormwater runoff may be sources of water supply. Modeling the quantity and quality dynamics of stormwater runoff requires a coupled hydrodynamic module capable of estimating the transport and fate of pollutants. In this paper, we evaluate the applicability of a distributed hydrodynamic model coupled with a water quality model (HydroPol2D). First, the model is compared to GSSHA and WCA2D in the V-Tilted catchment, and the limitation of the critical velocity of WCA2D is investigated. We also applied the model in a laboratory wooden board catchment, focusing on the validation of the numerical approach to simulate water quality dynamics. Then, we apply HydroPol2D in the Tijuco Preto catchment, in Sao Carlos - Brazil, and compare the modeling results with the full momentum solver of the Hydrologic Engineering Center - River System Analysis (HEC-RAS). The model implementation, the governing equations, and the estimation of input data are discussed, indicating the challenges and opportunities of the application of distributed models in poorly-gauged catchments. For a 1-yr return period of rainfall and antecedent dry days and assuming an uncertainty of $40\%$ in the water quality parameters, the results indicate that the maximum concentration of total suspended solids (TSS), the maximum load and the mass of the pollutant washed in $30\%$ of the volume are, $456~\pm~260~\mathrm{mg.L^{-1}.km^{-2}}$, $\mathrm{2.56 \pm 0.4~kg.s^{-1}.km^{-2}}$, and $\mathrm{89\%~\pm~10\%}$, respectively.

Motivation & Objective

• To develop a distributed hydrodynamic and water quality model suitable for data-scarce, urbanized catchments.
• To evaluate the impact of critical velocity limitations in existing models like WCA2D on simulation accuracy.
• To validate the water quality module using laboratory-scale experimental data from a wooden board catchment.
• To apply HydroPol2D in a real-world, poorly gauged urban catchment (Tijuco Preto, Brazil) with limited observed data.
• To quantify the spatial-temporal dynamics of surface runoff, pollutant concentration, and load under varying rainfall and dry-day conditions.

Proposed method

• Implementation of a 2D depth-averaged shallow water equation solver with explicit finite volume method for hydrodynamic simulation.
• Coupling of Green-Ampt infiltration model to simulate surface runoff generation in urban impervious and pervious surfaces.
• Incorporation of build-up and wash-off modules for total suspended solids (TSS), using empirical coefficients calibrated against lab data.
• Use of the method of lines and adaptive time-stepping for numerical stability and performance in complex terrain.
• Application of uncertainty quantification (±40%) in pollutant parameters to assess model robustness under data scarcity.
• Comparison with HEC-RAS full momentum solver and GSSHA/WCA2D in benchmark cases to validate performance and accuracy.

Experimental results

Research questions

• RQ1 How does limiting flow velocity to critical values affect model performance in flood simulation?
• RQ2 Can the HydroPol2D water quality module accurately simulate TSS build-up and wash-off dynamics in a controlled laboratory setting?
• RQ3 To what extent can HydroPol2D simulate runoff quantity and quality in a real, poorly gauged urban catchment with limited data?
• RQ4 What are the key pollutant transport dynamics, particularly the first-flush effect, under a 1-year return period rainfall and antecedent dry day scenario?
• RQ5 How do uncertainties in wash-off parameters affect the reliability of TSS load and concentration estimates?

Key findings

• Limiting flow velocity to critical values significantly reduces model performance, indicating that such constraints are not suitable for accurate flood simulation.
• The water quality module of HydroPol2D was successfully calibrated and validated against laboratory data from a wooden board catchment, confirming its ability to simulate TSS build-up and wash-off processes.
• For a 1-year return period event with 5 antecedent dry days, the maximum TSS concentration at the outlet was 456 ± 260 mg·L⁻¹·km⁻², normalized per km².
• The maximum TSS load at the outlet was 2.56 ± 0.4 kg·s⁻¹·km⁻², indicating high pollutant export during peak flow.
• The wash-off ratio coefficient was 0.63 ± 0.11, and the event mean concentration (EMC) was 131.59 ± 16 mg·L⁻¹, confirming significant pollutant mobilization.
• 89% ± 10% of the total TSS mass was washed out in just 30% of the total runoff volume, demonstrating a strong first-flush phenomenon in the Tijuco Preto catchment.