Single-layer vs. Multi-layer Simulations in Hydrodynamic and Water Quality Modeling
The selection of a single layer vs. multiple layers for simulating hydrodynamic and water quality depends upon a range of factors. Most significant are the conceptual model of the system, the constituents that need to be modeled, the physical characteristics of the system, and computing resources. The importance of using multiple layers may not always be apparent, however, and we often come across situations where modelers may benefit from changing the layering scheme.
Coastal waters are complex environments characterized by rich biological diversity and natural resources. As the population of coastal communities increases, the deterioration of the coastal environment becomes a critical issue. Estuaries are coastal waters where the freshwater of the river mixes with the saline water from the ocean. Salinity variations in estuaries are often so large that they significantly affect circulation. This impacts the two-directional net flows, which are seaward in the surface layer and landward in the bottom layer. In turn, this frequently controls the long-term transport of pollutants.
A common mistake is modeling salinity in a one-layer hydrodynamic simulation. Salinity affects the density of the water, and different layers may have different salinity concentrations. In these cases, the denser saline water from the bottom layer moves differently than the upper layer. To clearly show an example of this effect, we created a simple demonstration model with multiple vertical layering schemes.
Our simple demonstration model replicates a rectangular tank (dimensions 1 x 2 x 0.5 m) with 400 horizontal cells and several different vertical layer scenarios (1, 2, 5, 10, and 200 layers). At the model start time, the saline water is completely banked up in the right half of the tank. The characteristics of the tank are described in Figure 1 below.
This demonstration model uses a standard sigma vertical layering scheme without any external forcing data. As the model simulation starts, the saline water starts moving from the right-hand side to the left-hand side of the tank, and the freshwater moves in the opposite direction, due to differences in density. We observed that process in all the models (Figure 2 and 3), except the single-layer model, underscoring that single-layer models should be avoided when modeling salinity in a hydrodynamic model.
This movement of saline and fresh water is smoother in models with a greater number of layers, which demonstrates how increasing the number of layers can provide a more realistic simulation of the system. The entire simulation time lasts only one minute. If the simulation is run for longer, it is found that after approximately 5-6 minutes, the model largely reaches a steady state in the 200-layer model.
Figure 2. Animation shows the vertical slice of how salinity affects the density of the water in (top) the 1-layer model (top) and 200-layer model (bottom)
Figure 3. Animation shows the vertical slice of how salinity affects the density of the water in four different vertical-layer models
Although increasing the number of vertical layers can provide a more realistic simulation of the natural processes, it also increases the computational requirements of the model. A modeler must make a judicious selection of the appropriate number of vertical layers for their model. In any case, we strongly advise against using a single-layer model to simulate situations where density can vary due to other factors (salinity, temperature).
You can test this effect yourself by downloading EEMS and running it in demo mode with the 2D tank model, which is available here.
