Chemical Control of Invasive Phragmites and the Seedbank

GBF attended a Ontario Phragmites Working Group (OPWG) conference where we were introduced to Graham MBJ Howell, M.Sc. Candidate , at the University of Waterloo. Mr. Howell agreed to provide us this summary of the research he is conducting under the guidance of Dr. R. C. Rooney, . Thank you to Mr. Howell, Dr. Rooney and the University of Waterloo team for their work and helping to educate us.
Invasive Phragmites. Why this plant is a problem.

European Common Reed (Phragmites australis) is invading Canadian wetlands, threatening their biodiversity and integrity. So-called “Phrag” can reach over 3 m in height, forming dense, shady, impenetrable stands that are like trying to walk through wicker furniture. This excludes native plants and also reduces habitat for wetland birds, amphibians and reptiles. Phrag is a clonal grass with perennial roots, but annual shoots. Consequently, it is capable of regenerating rapidly from control efforts that only tackle the above-ground part of the plant and it is challenging to eradicate once it becomes established.

What about herbicide management solutions?

Because tackling the shoots has little effect (GBF note: Selective cut method does help control Phragmites in water, where herbicides illegal) , the most successful treatment method is to spray the plant with herbicide in autumn to kill the roots, and then to mow, burn, or flatten the shoots later that winter to eliminate the standing dead plant litter and open up the canopy for native plants to recolonize. Typically, the major source of colonists is the seedbank: the repository of seeds resting dormant in the soil and waiting for the right conditions to sprout.

Of course, there are concerns (including by GBF) that using herbicides in a wetland may have undesirable side effects. Although the herbicide used in Phrag control is only meant to affect adult plants, we wanted to test whether there was a risk that the herbicide application would affect the seedbank .

To test this, we looked at soil samples from sites that had been sprayed with a 5% glyphosate solution and from an adjacent area that was not sprayed: our control plot. Both plots were located in Big Creek National Wildlife Area, in Long Point, Ontario.

    We also thought the density of Phrag might influence the kind of vegetation that sprouts from the seedbank, because really dense stands would have less exposed soil for seeds to land in. Alternatively, the less dense Phrag patches might let more of the herbicide slip through the canopy to land on the soil, potentially affecting germination rates.

    To test the effect of Phrag density, we collected samples from high and low density Phrag patches in both the areas that were sprayed with herbicide and our untreated control patches. We washed the soil from the collected seeds and planted them in germination trays in the University of Waterloo greenhouse with 12 hours of light per day and daily watering.

GBF note:
What you can do:

In Ontario, there are NO approved herbicides to use over water. GBF teaches non-chemical treatment, or selective cutting which is necessary to help Georgian Bay coastal wetlands.

1. Learn more about and how you can help stop invasive Phragmites, in your own community.

2. Email us at if you want to be a Phragbuster.

3. Visit our Community Phragbuster page.
Find and work with others in your community, or promote your group's efforts and get more volunteers.
Join one or register your group

The results

After seven months of growing, we got our answer:

  • Glyphosate treatment does not influence the diversity or number of seedlings that germinate from the seedbank samples. This was as we expected, because glyphosate is a post-emergent herbicide, meaning it should not affect seeds. But it reassures us that glyphosate residue does not affect seedlings that would sprout after Phrag control treatment. This is an important finding, because usually we rely on the seedbank to recolonize the sites after Phrag control treatment. If glyphosate were reducing the viability of the seedbank, it would mean poor recolonization by native species.
  • More surprising, we found that the number of seedlings that sprout and the diversity of species in the seedbank did not differ among our low and high Phrag density patches. This is contrary to our expectation that denser stands of Phrag would have less seeds reaching the seedbank. We think this might be because we collected samples from a long, skinny berm feature that had other plant communities growing on either side of it. Because of the shape of the Phrag patch on the berm, we think penetration of seeds from the neighboring communities was less affected by Phrag density. We are following up with another trial, using seedbank samples collected from the interior of a large patch of Phrag. It is too early to say for certain, but our early analysis indicates that seeds are less abundant and that there are fewer species represented in the seedbank from this large Phrag patch.
  • A final important lesson from this experiment was that the seedbank actually contains quite a bit of viable Phrag seed. Conventional wisdom was that Phrag is primarily dispersed vegetatively and that even though each seed head contains thousands of seeds, most of them were not viable. Initially, we thought the seedlings were a native grass, but as they developed the telltale “hairy ligule” became visible and the seedlings sent up runners characteristic of Phrag growth patterns. The ramifications of this finding is that most Phrag control efforts will likely require some follow-up treatment in subsequent years, as even though adult plants are eradicated by control efforts, it is able to reinvade from the seedbank. If left unchecked, these seedlings will quickly grow to dominate the wetland once more.
  • Thank you from Georgian Bay Forever for sharing your research.
    Here are some more research projects on invasive Phragmites and management controls.