This post is dominated by my experience supporting regulatory review, including a failed attempt to revise waterway policy - largely due to challenges of inter-agency jurisdictions.
Setting a target for flushing rates, particularly when modelling can be easily manipulated, gives a tendency towards poor water quality outcomes. An objective to maximise flushing, while admittedly rather vague, should be the target.
Evaluating flushing time is useful, in that it is something which can be physically measured (e.g. rhodamine dye testing) and it isn’t subject to as much conjecture as WQ modelling.
Hard flushing time criteria (e.g. 4-day, 7-day or 10-day e-folding times) are appallingly abused. They are equivalent to saying ‘about head height’.
Development of modelling techniques for evaluating flushing occurred through the 1980s and 1990s, converging towards some relatively standard techniques (e.g. e-folding time). In Western Australia a shift to unofficial acceptance criteria occurred through the early 2000s. For small craft harbours, a 10-day e-folding time for ‘worst case’ (tide only) was typically used, although this remained highly susceptible to manipulation of model parameters - the same waterway, modelled two different ways (i.e. ¼ grid scale, but otherwise identical) will give a different answer. Sometimes not much different…
Following investigation of a couple of harbours with groundwater-related WQ issues, someone decided to get smart and incorporate all sorts of other mixing parameters (not me, but we all make mistakes!) but retain the SAME 10-day e-folding criteria. This was inappropriate, as the ability to make a teensy adjustment (driving time series, boundary conditions, dispersion coefficient) to get different outcomes makes using any criteria very challenging.
Once a modelled flushing time became an excuse for plain old poor layout, things had ‘crossed over’.
I don’t think that numerical modelling shouldn’t be done… it’s just that it’s so horribly wobbly as a tool, that it needs to be very carefully used and interpreted. Following international practice, and Western Australian EPA guidance:
Make sure that the water body is designed with good flushing in mind (there’s the old USEPA 1985 guidance, as well as more recent work by PIANC
Get an understanding of WQ pressures
Undertake numerical modelling, to demonstrate flushing performance across a range of conditions. There’s difficulty getting hard criteria here, but something where there was Tf<4 days for >90% of the year, and Tf>15 days for <1% of the year would be much better than a waterway with 8<Tf<10 days for 100% of the year. The bigger problem is understanding how Tf varies with the modelling approach.
Have a WQ Management Plan that incorporates monitoring, tolerance criteria and the ability to detect trends.
So... some take-home messages:
A 10-day flushing time is not a suitable target
Modelled flushing times are almost meaningless (comparative only in themselves)
Water quality is not just about algal blooms
For those of you who aren’t aware… we attempted to develop a major revision to WAPC DC1.8 about 15 years ago, incorporating EPA’s Water Quality Management Framework. It had a lot of effort involved, but ultimately failed due to the difficulty of getting cross-jurisdictional agreement.