Large eddy simulation of turbulence and solute transport in a forested headwater stream

Abstract

The large eddy simulation (LES) module of the Virtual StreamLab (VSL3D) model is applied to simulate the flow and transport of a conservative tracer in a headwater stream in Minnesota, located in the south Twin Cities metropolitan area. The detailed geometry of the stream reach, which is ∼135 m long, ∼2.5 m wide, and ∼0.15 m deep, was surveyed and used as input to the computational model. The detailed geometry and location of large woody debris and bed roughness elements up to ∼0.1 m in size were also surveyed and incorporated in the numerical simulation using the Curvilinear Immersed Boundary approach employed in VSL3D. The resolution of the simulation, which employs up to a total of 25 million grid nodes to discretize the flow domain, is sufficiently fine to directly account for the effect of large woody debris and small cobbles (on the streambed) on the flow patterns and transport processes of conservative solutes. Two tracer injection conditions, a pulse and a plateau release, and two cross sections of measured velocity were used to validate the LES results. The computed results are shown to be in good agreement with the field measurements and tracer concentration time series. To our knowledge, the present study is the first attempt to simulate via high-resolution LES solute transport in a natural stream environment taking into account a range of roughness length scales spanning an order of magnitude: from small cobbles on the streambed (∼0.1 m in diameter) to large woody debris up to ∼3 m long.

Field and Numerical Investigation of Transport Mechanisms in a Surface Storage Zone

“…The complex hydrodynamics that are generated by flow interactions with geomorphic features in rivers (i.e., confluences, bars, riffles, pools, flow obstructions, meander bends, etc.) have significant effects on biogeochemical processes and nutrient cycles, which maintain the properties and spatial distribution of fluvial ecosystems (Jackson et al, ; Khosronejad et al, ; Patil et al, ). Energetic, large‐scale, unsteady, coherent structures are continuously generated by local shear, pressure imbalances, and flow separation, and they exert a fundamental control on transport and deposition of contaminants along the channel (Mignot et al, , ).…”

Reply to Comment by Sookhak Lari, K. and Davis, G. B. on “ ‘Large Eddy Simulation of Turbulence and Solute Transport in a Forested Headwater Stream’: Invalid Representation of Scalar Transport by the Act of Diffusion”

“…In equation (3) of Khosronejad et al (), we presented a transport equation (in curvilinear coordinates) of a passive scalar (i.e., NaCl) in a turbulent stream flow with a Reynolds number of Re ~ 25,000. This equation reads as follows: where ψ , ρ , U j , μ , σ * , μ t , G jk , and J are the solute concentration in volume fraction, density of stream water, contravariant volume flux, molecular diffusion coefficient for the scalar, inverse of the turbulent Schmidt number (=0.75), subgrid scale (SGS) dynamic eddy viscosity, contravariant metric tensor components, and Jacobian of geometric transformation from physical to curvilinear coordinate system, respectively.…”

“…In equation (3) of Khosronejad et al (2016), we presented a transport equation (in curvilinear coordinates) of a passive scalar (i.e., NaCl) in a turbulent stream flow with a Reynolds number of Re~25,000. This equation reads as follows:…”

“…We clarify that a molecular diffusion coefficient of μ = 1.4286 × 10 À6 Pa s was used for the solute in that study. In what follows we attempt to further elucidate the importance of turbulence diffusion in contrast to that of molecular diffusion and that our work in Khosronejad et al (2016) correctly represents the convection and diffusion of solute in Eagle Creek.…”

“…To our knowledge, Khosronejad et al () was the first attempt to simulate via high‐resolution LES solute transport in a natural stream environment taking into account a range of roughness length scales spanning an order of magnitude: from small cobbles of ∼0.1 m in diameter to large woody debris up to ∼3 m long. Sookhak Lari and Davis are primarily concerned with the molecular diffusion coefficient we used for the solute transport in equation (3) of Khosronejad et al (; i.e., equation below). We posit that the comments stem from our accidental use of the same symbol ( μ ) for two different parameters (the dynamic viscosity of water and the molecular diffusion coefficient of the tracer), which caused Sookhak Lari and Davis to think we had incorrectly used the dynamic viscosity of the fluid as a representation of molecular diffusion of the tracer.…”