I'm using the water quality module to simulate a wastewater discharge containing a conservative pollutant to a river .
I want to see how the concentration will decrease downstream as the pollutant is transported away from the discharge point.
I am getting good results with instantaneous mass injections, but injections of greater durations are not producing results that illustrate reductions in concentration downstream.
The cell size = 1m, the total channel length =1500m.
The time step interval is calculated by the software by ensuring the Courant and Peclet constraints are satisfied. Hand calculations using equations 19-34 and 19-35 from the user's manual estimate the timestep to be 0.09s. This is the timestep value that satisfies the Courant condition. The peclet timestep and the user assigned maximum allowable timestep were greater values (0.12s and 10s respectively). (The model code selects the smaller of these three values.)
The dispersion coefficients used are computed by the software. A quick hand calculation using equations 19-1 and 19-2 from the user's manual estimates the dispersion coefficient to be 2.5m2/s.
The channel geometry consists of identical trapezoidal cross-sections with a constant slope of 0.0015, so the velocities at each computation cell face are also identical. Therefore, the dispersion coefficients are constant throughout the channel also.
The dx, dt and dispersion values for the two runs are identical, the only difference is duration of mass injection.
Why is it that the instantaneous mass injection simulation produces results that illustrate dispersion but the mass injection with a given duration (say 1hr) seems to illustrate no dispersion but mainly advection?
The second figure does show diffusion of the conservative pollutant. Notice that the duration at the peak concentration becomes shorter downstream. There is diffusion in upstream and downstream directions. Notice also that the rates of increase and of decrease in concentration each become lower as the slug moves downstream.
There is a plateau because at any instant, in the middle of the slug, diffusion is occurring away from the point of flow in both directions and to the point of flow from both directions, simultaneously. Concentration does not go down until the clear water at the leading or tail edge of the slug diffuses to that point in the slug.
So is the HECRAS water quality module unsuitable for observing decreases in concentrations in receiving waters from a constant pollutant discharge?
Or does the distance between the observation points just need to be great enough to observe a greater decrease in concentration?
The diffusion will also greatly depend on the hydraulic flow regimes in the channel?
I do not think it is a matter of suitability of the water quality module. If it is a continuous injection of a conservative constituent in a single channel with uniform flow, numbers would look pretty boring downstream. The flow, cross section, and injection rate tell you enough for the calculation.
But if there is complex topography, additional inflows, a reactive constituent, or other complexities, 2D transport can answer interesting questions. For instance, if a flood washes out an open sludge pond, how concentrated is the pollutant at the edge of a flood eddy down river? 2D should be quite useful for the numbers you cannot calculate by hand.
I guess I was expecting the concentration graphs of the mass injection with specified duration to be more "squashed" as the injection moved downstream, ie with lower concentrations but increased distance between the rising and falling limbs such as in the sketch below.
Obviously the mass injected into the system would be retained as it is a conservative pollutant.
Just an update.
I have come to realise that HEC-RAS computes longitudinal dispersion only and assumes that transverse and vertical mixing occurs instantly in a control volume (water quality cell).
Therefore the software is not suitable for determining mixing lengths in a channel.