By the time zebra mussels are visible on your dock, eradication is already off the table.

Invasive species management has a narrow window, and almost every lake misses it. Not through negligence — through the simple fact that visual detection requires a population large enough to be noticed by chance, and by then it is established.

Everything after that point is containment. Which is expensive, permanent, and much less satisfying than the alternative.

What eDNA actually is

Every organism sheds genetic material into its environment continuously — skin cells, mucus, waste, gametes. In water, that material stays suspended and detectable for a period of days.

Environmental DNA sampling means filtering a volume of water, capturing whatever genetic material is in it, and then testing for a specific species using a marker unique to that organism. If the target has been in that water recently, its DNA is on the filter.

The sensitivity is the point. A molecular assay can detect a species at densities far below anything netting, trapping, angling, or visual survey will find. You are looking for the organism's traces rather than the organism.

Two different tests

People conflate these and they answer different questions.

A targeted assay tests for one species. You choose the target — zebra mussel, quagga mussel, starry stonewort, hydrilla, round goby, northern snakehead — and get a detection or non-detection. Fast, cheaper, unambiguous. Use it when you have a specific suspicion.

A broad survey sequences everything present and returns a community profile. Use it when you do not know what to look for, or when you want a biodiversity baseline.

The usual sequence: survey first to see what you have, then targeted assays to watch for the specific things you are worried about.

What it will not tell you

Be clear about the limits, because eDNA is frequently oversold.

It reports presence, not abundance. A detection does not tell you whether there are ten individuals or ten thousand. It does not tell you where in the lake they are — only that the DNA was in the water you sampled. It cannot distinguish a living population from DNA that arrived on a boat hull yesterday, or from a dead organism. And it says nothing about life stage or whether the population is reproducing.

Regulatory action generally requires a physical specimen. Treat a detection as the trigger for a targeted search, not as the conclusion.

Sample where introductions happen

The most common mistake is sampling the middle of the lake. Invasives arrive on trailers, and the water nearest the ramp is where the first shed DNA will be.

Sample the boat launch, the marina, the inlet, and any high-traffic shoreline. Timing matters too — detection probability rises when the target is spawning or actively growing, so a mid-winter sample for a warm-water species is a weak test.

And contamination discipline is not optional. Gloves on, no contact between the filter and any surface, and do not sample downstream of a boat you have just handled.

The result nobody thinks to buy

Non-detection is worth paying for, and it is the most undervalued result in the category.

A dated record establishing that a species was absent as of a given date is what tells you when it arrived. That matters for grant applications, for management plans, for state reporting, and occasionally for questions of liability.

You cannot establish a baseline retroactively. The season to document that your lake is clean is the season before it is not.