Native Seeds in Restoration: Garlic Mustard Part 3 - Monitoring Results & Discussion

This is the third in a series of blog posts describing an RNPP research project that took place 2022-2024. 

To recap, the goal of this project was to trial a seed mix to reintroduce native plants in areas that have been treated for garlic mustard in riparian areas. Garlic mustard (Alliaria petiolata) is an invasive plant with allelopathic properties that prevents other plant seeds from germinating. Though it is widespread in the northern Willamette Valley, garlic mustard has a limited presence (net treatment area less than 100 acres) in the Rogue Basin. When allowed to overtake forest understories, garlic mustard greatly reduces the availability of native plants that are vital to wildlife, including deer, elk and pollinators (ODA, 2015). 

The timeline so far:

• 2012: Garlic mustard is discovered in the Rogue basin and herbicide treatment begins.

• Summer 2022: RNPP receives a grant in collaboration with the Jackson/Josephine County Cooperative Weed Management Area from the Oregon Department of Agriculture (ODA). Two sites with garlic mustard, one which has been herbicide treated since 2012, are selected for demonstrating the native seed mix. See our first blog post in this series for an overview of the demonstration sites and site history.

• Fall 2022: Seeds are applied to the two demonstration plots. See second blog post for a detailed description of our treatment strategy, including species used in the seed mix.

• Winter 2022-23: One demonstration plot is inadvertently destroyed by landowner.

• Spring/Summer 2023: First round of monitoring to document native plant emergence at remaining plot (ODFW)

• Spring/Summer 2024: Second round of monitoring at ODFW demonstration plot.

Monitoring Protocols

Although the native seed demonstration area was destroyed at the John Day Road site, we were able to return to the ODFW demonstration area for monitoring in 2023 and 2024. This graphic shows the layout of the plot:

The “treatment” areas were raked and then seeded at a rate of 85 seeds/square foot (no other treatments). The “control” areas were raked but not seeded. The seeding rate was a bit high for restoration projects, in an effort to overcome the allelopathy of garlic mustard plants/future seedlings in the demonstration area. We also used a divoting method for seeding. More on that in the Part 2 blog post linked earlier.

To monitor the demonstration plot, we created 20 transects; 10 in the treatment areas and 10 in the control areas. Along each transect, two 1x1 meter square areas (subplots) were selected at random for monitoring. Figure 2 shows the location of subplots in the first 5 control and 5 treatment transects. 

This monitoring method allowed us to get good representative data in a reasonable amount of time, to assess how well the reseeding worked compared to non-seeded areas across the entire plot. 

In each subplot, both control and treatment, our monitoring protocol was the same:

1. Assess ground cover percentages (how much bare ground vs. wood, moss or litter)

2. Assess plant cover by species 

3. Take a photo looking down at the plot

Monitoring was conducted in June of 2023 and June 2024. 

Results

The first-year results in 2023 were encouraging. We saw Heracleum maximum seedlings emerging from the divots where we had seeded in the treatment areas of the plot. Heracleum, or cow parsnip, is a large native plant in the carrot family. The native plant with the highest cover in the plots was the grass Elymus glaucus, blue wildrye, which had been seeded in this area previously but was also included in the new seed mix because if its ability to establish quickly. We also found a native brome species, Bromus carinatus, which we had seeded into the plot.

Overall, this first year we found a significant difference in native plant cover in the treatment area versus the control as shown in Figure 3. 

Figure 3: Percent native cover in treatment and control plots, 2023

However, it also seemed that garlic mustard was doing just fine in both the control and treatment plots, with no evidence showing that the natives were starting to compete with the invasive just yet. 

Things took a different turn in 2024. This year, we returned to find only a few surviving but stunted Heracleum maximum, and no other seeded native plants growing in the plot, aside from some Elymus glaucus. Garlic mustard cover had increased due to herbicide spraying being skipped in 2023. 

In 2023 this graph shows that the more garlic mustard there was, the fewer native plants. In 2024 the relationship is flat, ie there were not many natives regardless of garlic mustard cover. 

Overall, these data seem to add to the evidence that garlic mustard suppresses growth of native plants. Although first year results were encouraging, any gains were lost by 2024.

Another interesting observation is that although Acer macrophyllum (bigleaf maple) seedlings were found in both years, they appeared to be first-year seedlings. In other words, the seedlings we saw in 2023 did not survive into the second year. Overall the site has very few young maple or other trees. The mature overstory of large trees, mostly bigleaf maple and Oregon ash, surely provides an abundant seed source which would be creating many seedlings and saplings in a healthy forest. In this way the garlic mustard infestation seems to be having a similar impact to a blackberry infestation, where very few young trees (or understory forbs and grasses) are able to penetrate the dense blackberry thicket. In this situation, allelopathy may be playing a bigger role than shading. 

Discussion

We found that in the ODFW demonstration site, seeding an area treated for garlic mustard with a 100% native plant mix was unsuccessful. It seems likely that the surviving Heracleum maximum plants will not continue to grow to maturity due to the affects of the allelopathic garlic mustard. Although Elymus glaucus appears to be surviving at the site, there are no flowering plants aside from garlic mustard to provide forage for pollinators, and other aspects of a dynamic riparian forest ecosystem (tree regeneration, diverse understory) are not present. It is difficult to say how much of this effect is caused by garlic mustard itself, and how much from the intensive herbicide treatment that has been ongoing for over a decade at the site. 

However, we did learn that the two native plants surviving, Heracleum maximum and Elymus glaucus, have at least some resistance to the allelopathic chemistry of garlic mustard. We also saw some native shrubs and one other native forb (Agastache urtricifolia) growing in the demonstration plot, which were not seeded. 

If this study were to be repeated, we would alter the following strategies:

1. Improved landowner communication/marking of demonstration plot to avoid destruction of seeded area

2. More comparison plots with different herbicide, manual control methods of garlic mustard utilized post-seeding

3. More seeded plots using different seeding methods (broadcast, divots, seed drilling)

4. More diversity in seed mix, focusing on grasses, shrubs, mint- and carrot-family species (which may share similar competitive features to Heracleum maximum and Agastache urtricifolia)

Conclusion

Invasive plants, especially allelopathic ones, present one of the most difficult puzzles in ecological restoration. While the herbicide treatment of garlic mustard along the Rogue River has reportedly prevented a more widespread invasion of this plant into riparian forests, there doesn’t seem to be an end in site for spraying this plant in heavily invaded areas like the ODFW site. Meanwhile, plant biodiversity at this site has plummeted and there are domino effects for the ecosystem: There are likely to be great reductions in the number of pollinators, browsing/foraging animals and birds at this site than in a healthy, functioning riparian forest. In another decade or two, there are likely to be fewer overstory trees as the existing mature trees die off and are not replaced by the next generation. 

Taking on this study was a risk, knowing the challenging nature of reseeding an area that is heavily invaded by an allelopathic plant and still undergoing active treatment to reduce the invasive population. However, we felt that the greater risk was not to try. By undertaking this demonstration project, we gained some understanding of how garlic mustard interacts with other plants, and some ideas for improving this type of study in the future. We also highlighted the need for improved native seed availability; for example we were not able to trial shrub seeds because they were not available at the time.