Eurasian watermilfoil forecast reveals Quebec lakes at risk

Most weeds perish when they are broken up. One species behaves differently: each snapped-off piece can float away, establish roots and develop into a completely separate plant.

This trait has enabled an invasive submerged weed to colonise hundreds of lakes throughout southern Quebec.

Cold conditions have been the only real barrier to its expansion. Further north, lakes remain ice-covered for longer, while summers are still too brief for the plant to become established.

However, a new projection indicates that this safeguard is weakening, while pinpointing the lakes most at risk of being invaded next.

A weed worth watching

The species is Eurasian watermilfoil, Myriophyllum spicatum. For decades, it has gradually overtaken lakes and rivers, forming dense beds just beneath the water’s surface.

These beds cause the damage. They prevent sunlight from reaching below, displace native vegetation and, as they decompose, remove oxygen from the water. Cutting the plant does not solve the issue, because it regrows from the resulting fragments.

The research team was headed by Grace Fedirchuk, a biologist at the University of Quebec at Rimouski (UQAR).

She aimed to establish both where the plant is already found and where it may emerge next. Her team therefore worked to predict its future spread.

Warming in the north

For many years, cold weather kept this plant at bay. Northern lakes froze deeply and remained cool well into the summer, seldom becoming warm enough for the weed to establish itself.

Climate change is weakening that restraint. As mean temperatures rise, lakes that formerly stayed cold far into summer are warming earlier and retaining that warmth for longer. This trend is not confined to Quebec.

Studies have projected that warm-water plants in other regions will move towards the poles as the climate warms, occupying areas that have become newly suitable, as one paper on a floating weed described.

Building the forecast

To anticipate the invasion, the team supplied a model with thousands of records showing where the plant had and had not been found.

Every record included around two dozen local variables, including temperature, rainfall, lake size and the population nearby.

The model identifies the conditions that support invasion, then calculates the probability of invasion in other lakes. The researchers used it across approximately 12,000 watershed areas in Quebec.

They modelled six scenarios, combining varying projections for warming and population growth up to 2100.

A prediction of this kind relies greatly on the standard of the underlying data. An earlier study of the same species found that water temperature and human access were among the strongest indicators of where it would occur.

Local data wins

This study produced an unexpected finding. The researchers ran the model in three forms, using data from Quebec, across North America and from the plant’s native European range.

They then assessed which performed best, and the Quebec-based version came out ahead - by a substantial margin.

A model trained solely on Quebec records made much more accurate predictions than the continental and European models, despite those datasets being considerably larger. The outcome offers an important lesson.

A species moving into unfamiliar territory will not inevitably act as it does in places where it has been established for many decades. The model had to learn from the leading edge of its range.

A driving force

When the researchers examined the conditions with the greatest influence, one pattern stood out.

The clearest signals were the maximum heat a lake reaches and the length of time it remains warm, followed by rainfall and water retention.

One human influence also mattered: the number of people living around a lake. More populated shorelines bring more docks, more boat launches and more boating activity.

That is precisely how the weed travels from lake to lake. A single broken stem can be enough.

Pieces can attach themselves to propellers and trailers, endure an overland journey, then take root in another lake. An earlier model linked this pattern closely to boat movements and road access.

Mapping the spread

In the warmest scenarios, the territory in which the weed could establish expands by almost five times, representing a broad advance into Quebec waters that currently remain unsuitable for it.

The largest increases are concentrated in two regions. Western Quebec becomes newly vulnerable, as does an extended corridor along the St. Lawrence.

These locations are not arbitrary. They are areas where increasing heat overlaps with intensive human activity.

The model identified these as the two factors most likely to leave a lake susceptible to invasion.

Tangible next steps

Before this work, there was no Quebec-wide view of where Eurasian watermilfoil might spread as the climate became warmer.

There is now, based on the finding that local records outperform broad datasets when forecasting a plant that is expanding its range.

This alters the options available to managers. Rather than pursuing the weed once it has arrived, they can monitor the forecast and attempt to prevent it from establishing in the first place.

Prevention has always been the least expensive means of combating an aquatic invader, and a more precise map makes that approach achievable.

The cold conditions that protected Quebec’s northern lakes for so long are receding, and the province can now identify where the weed is most likely to spread next.