Management Implications
Humulus japonicus (Japanese hop), a vine originating from East Asia, poses a threat to wetland and riparian ecosystems by smothering vegetation and decreasing biodiversity. Current management of H. japonicus includes a combination of mechanical and chemical control. However, there is limited research on H. japonicus ecology, its interactions with other species, and its role as an invader in wetlands. Therefore, the objectives of this study were to quantify the invasion of H. japonicus; determine the effects of invasion on Phalaris arundinacea (reed canarygrass) cover, plant species richness, and community composition; and examine environmental preferences, specifically shade tolerance, of H. japonicus.
Key findings include an increase in H. japonicus cover between years and a corresponding decrease in P. arundinacea cover, indicating the replacement of P. arundinacea. The replacement of a dense, rhizomatous, perennial grass by an annual vine has potential implications for ecosystem functions, including soil stability, and H. japonicus invasion has the potential to increase erosion in riparian areas. Additionally, the invasion of H. japonicus reduced species richness and altered species composition. Increased canopy cover reduced H. japonicus cover, meaning that restoring canopy cover and planting fast-growing trees could effectively manage H. japonicus populations.
Introduction
The introduction of invasive species in wetland ecosystems has detrimental impacts, including altering ecosystem function (Vitousek et al. Reference Vitousek, D’Antonio, Loope, Rejmanek and Westbrooks1997) and reducing biodiversity (Zedler and Kercher Reference Zedler and Kercher2004). Wetlands are especially vulnerable to biotic homogenization (Price et al. Reference Price, Spyreas and Matthews2018), which occurs when local, native species are replaced by few generalist, non-native ones across many sites (McKinney and Lockwood Reference McKinney and Lockwood1999). With an increase in anthropogenic activity, ecosystems are more likely to be invaded by non-native species that replace native species (Lázaro-Lobo and Ervin Reference Lázaro-Lobo and Ervin2021), resulting in biotic homogenization.
For floodplain forest ecosystems, biotic homogenization occurs at an even faster rate than nearby upland forests (Johnson et al. Reference Johnson, Amatangelo, Townsend and Waller2016). This has led to biological homogenization at a regional scale, with the introduction of the invasive reed canarygrass (Phalaris arundinacea L., Poaceae) (Price et al. Reference Price, Spyreas and Matthews2018). A common wetland invader, P. arundinacea is a long-lived perennial grass that spreads aggressively due to its expansive network of underground rhizomes (Lavergne and Molofsky Reference Lavergne and Molofsky2004). Invasion by P. arundinacea is widespread throughout much of North America, including Alaska, Canada, and the temperate, conterminous United States, and invasion negatively affects plant diversity and floristic quality, resulting in changes in community composition (Spyreas et al. Reference Spyreas, Wilm, Plocher, Ketzner, Matthews, Ellis and Heske2010).
Despite the prolific spread of P. arundinacea in wetlands, one emerging invasive plant that has been observed to invade and cover wetland vegetation, including P. arundinacea, is the Japanese hop (Humulus japonicus Siebold & Zucc., Cannabaceae) (personal observation). An annual vine, H. japonicus has been observed to form dense monotypic mats (Guyon and Cosgriff Reference Guyon and Cosgriff2022) that outcompete plant communities with species of conservation concern (EMPPO 2019). The vine overgrows all vegetation, even overtopping shrubs and trees and suppressing their regeneration (Guyon and Cosgriff Reference Guyon and Cosgriff2022; Kim and Kim Reference Kim and Kim2009). For species growing underneath H. japonicus, limited light and high moisture leads to death and rapid decomposition (Kim and Kim Reference Kim and Kim2009), resulting in decreases in biodiversity, species richness, and functional richness (Fried et al. Reference Fried, Carboni, Mahaut and Violle2019). As H. japonicus continues its invasion in wetlands, increases in its range distribution and local abundance are likely to contribute to biotic homogenization across vulnerable ecosystems. Additionally, as H. japonicus replaces both native and invasive vegetation whose roots stabilize soil, ecosystems are more vulnerable to erosion, which can alter ecosystem functions (Cellone et al. Reference Cellone, Carol and Tosi2016).
Humulus japonicus was introduced from eastern Asia to Europe and North America for medicinal and ornamental purposes (EMPPO 2019). In North America, H. japonicus has established in eastern Canada and the United States, west to North Dakota and south to Kansas (USDA-NRCS 2023), making it a major species of concern. Humulus japonicus easily colonizes disturbed, flood-prone ecosystems that have open canopy gaps (Guyon and Cosgriff Reference Guyon and Cosgriff2022) and is found in wetlands, as it is mainly dispersed by water along rivers (EMPPO 2019) and grows best on riversides with alluvial soils (Balogh and Dancza Reference Balogh, Dancza, Tokarska-Guzik, Brock, Brundu, Child, Daehler and Pyšek2008; EMPPO 2019). Because H. japonicus grows best in full light, restoring canopy cover is the most effective long-term treatment for controlling invasion in forested floodplains, wetlands, and riparian zones that are subject to repeated or ongoing colonization pressure (Guyon and Cosgriff Reference Guyon and Cosgriff2022). Measures taken to prevent the establishment and spread of H. japonicus in these types of areas include preserving native tree cover and avoiding disturbances that allow for its establishment (Fried et al. Reference Fried, Mahaut, Pinston and Carboni2018).
While research on H. japonicus has been conducted in Europe (e.g., Balogh and Dancza Reference Balogh, Dancza, Tokarska-Guzik, Brock, Brundu, Child, Daehler and Pyšek2008; Georgescu et al. Reference Georgescu, Popa, Luchian, Moisescu, Milos and Savulescu2021; Urziceanu et al. Reference Urziceanu, Cîșlariu, Nagodă, Nicolin, Măntoiu and Anastasiu2022) and Asia (e.g., Ju et al. Reference Ju, Kim, Lee, Lee, Kim, Nam and Kang2006; Kim and Kim Reference Kim and Kim2009), research on H. japonicus in the North America is limited, especially compared with other invasive wetland plant species such as P. arundinacea or the common reed [Phragmites australis (Cav.) Trin. ex Steud.]. With its status as an emerging species of concern and as such an aggressive invader, H. japonicus should be further researched to fill a current gap in the literature. There is a need for observational field studies of its ecology as an invasive species, its interaction with other plants, and its role as an invader in wetlands. Studying its invasion dynamics and patterns within a field setting rather than an experimental setting allows for species interactions to be observed under natural environmental conditions. More effective management strategies for H. japonicus invasion can be developed through invasion monitoring.
We conducted an observational study in a restored floodplain forest from October 2022 to October 2023 to track the spatiotemporal dynamics of H. japonicus invasion relative to gradients in canopy cover and species composition. Our specific objectives were (1) to quantify the invasion or decline of H. japonicus by measuring its cover along the same transects between years, (2) to investigate the change in P. arundinacea cover in response to H. japonicus invasion, (3) to determine the effects of H. japonicus invasion on plant species richness and community composition, and (4) to examine the relationship between H. japonicus cover and canopy cover in a field setting.
Materials and Methods
Study Site
The Joslin wetland mitigation site is a 6.1-ha floodplain wetland located along the Rock River in Henry County, northwest Illinois, USA (41.5542°N, 90.1835°W). Previously agricultural land, the site is now a compensatory mitigation wetland, meaning that it has been restored as a wetland to provide compensation for losses of aquatic resources (Matthews et al. Reference Matthews, McIntyre, Peralta and Rodgers2020; USACE, USEPA 2008). Topography and hydrology are similar across the site, and the soil type is Sawmill silty clay loam (fine-silty, mixed, superactive, mesic Cumulic Endoaquolls; Matthews et al. Reference Matthews, McIntyre, Peralta and Rodgers2020).
The restoration site was designed by the Illinois Department of Transportation and restored in 1998 as an experiment to test the effectiveness of five reforestation methods: balled-and-burlapped tree plantings, bare-root tree plantings, seedling plantings, acorn plantings, and passive restoration (Matthews et al. Reference Matthews, McIntyre, Peralta and Rodgers2020). Fifteen years after restoration, in treatments with less expensive methods (i.e. passive restoration and acorn plantings), P. arundinacea invaded and dominated the vegetation cover. In treatments with more costly methods (i.e. bareroot and balled-and-burlapped tree plantings), there was less P. arundinacea invasion due to greater canopy cover, resulting in greater plant species richness (Matthews et al. Reference Matthews, McIntyre, Peralta and Rodgers2020). While P. arundinacea previously dominated large areas of the Joslin field site, a new, emerging invader, H. japonicus, was observed to have invaded many of the P. arundinacea patches by 2021 (personal observation). Invasion by H. japonicus has resulted in decreased P. arundinacea abundance.
Ten transects extending roughly east to west were established and sampled in 2022. Transects were evenly split between forested and open areas based on the initial tree-planting treatments described by Matthews et al. (Reference Matthews, McIntyre, Peralta and Rodgers2020; Figure 1). Seven 1 m by 1 m quadrats were surveyed per transect, with the midpoint of each transect positioned at the boundary of H. japonicus expansion, marking the edge of H. japonicus patches as it replaces P. arundinacea stands (H. japonicus cover >20%). From the midpoint, two halves of a 24-m transect were established, with 12 m extending into H. japonicus growth and 12 m beyond the H. japonicus invasion front, into an area where it has the potential to expand. Quadrats were placed every 4 m, with the midpoint (0 m), 4 m, 8 m, and 12 m at the center of each quadrat (Figure 2). A marking flag was placed in the center of the midpoint quadrat (0 m) and labeled with the transect number. Additionally, a Garmin GPS (Garmin International, Inc., Olathe, KS, USA) unit was used to mark the ends and middle of each transect.
Map of field site in Joslin, IL, USA divided by restoration treatment: (A) balled-and-burlapped tree plantings, (B) bare-root tree plantings, (C) seedling plantings, (D) acorn plantings, and (E) passive restoration, or seedbank. Ten transects were located at the Joslin field site in October 2022, June 2023, and October 2023 across the different treatments. Transects were split between treatments, resulting in open and forested canopies.
Quadrat and transect setup for vegetation and canopy cover sampling. Quadrats were sampled every 4 m, extending roughly east and west, into and out of Humulus japonicus patches. Quadrat 0 is at the invasion front of H. japonicus, whereas negative quadrats were sampled within H. japonicus patches, and positive quadrats were sampled in patches where H. japonicus had not yet invaded.
Vegetation and Canopy Cover Sampling
Field sampling was conducted on October 5, 2022, June 6, 2023, and October 2, 2023. Within each quadrat, vegetation percent cover was estimated for each individual plant species as the percentage of ground within a quadrat covered by the species, rounded to the nearest 5%. Additionally, tree canopy cover, measured as the percentage of occupied overhead area using a spherical densiometer, was collected in each quadrat in October 2022 and October 2023. Additionally, in October 2023, light intensity readings were obtained by collecting a 15-s average reading for each quadrat using a Li-Cor Quantum Sensor (model LI-190; LI-COR Biosciences, Lincoln, NE, USA). Li-Cor readings were collected during times without cloud cover to ensure consistency, although some variability in sky conditions may still have been present. Densiometer data were used to explore the relationship between H. japonicus cover and shade, while Li-Cor data were used to support measurements made via densiometer.
Statistical Analysis
All statistical analysis was conducted in R v. 4.3.1 (R Core Team 2023). Figures were visualized with the collected raw data using the ggplot2 package (Wickham Reference Wickham2016). Light intensity recorded using the Li-Cor sensor was related to canopy cover from the densiometer using a Kendall rank correlation due to nonnormality of the light intensity data. Additionally, a linear mixed-effects model was conducted to determine whether canopy cover varied from October 2022 to October 2023. The response variable was canopy cover in October 2023, while the fixed effect was canopy cover in October 2022, and the random effect was transect.
To determine the effects of year (October 2022 and October 2023), canopy cover, distance along the transect (indicated by quadrat), and transect on H. japonicus cover, P. arundinacea cover, and species richness, separate linear mixed-effect models were conducted using the Kenward-Roger approximation method with the lme4 package in R (Bates et al. Reference Bates, Mächler, Bolker and Walker2015).
Initially, the effect of position along the transect was tested using three alternative variables: quadrat as a continuous numerical variable, a categorical variable split into negative (−12 m, −8 m, −4 m), neutral (0 m), and positive quadrats (4 m, 8 m, 12 m), and using H. japonicus cover as a proxy for distance along the transect. The best variable for position along the transect (quadrat) was determined for each response variable (H. japonicus cover, P. arundinacea cover, and species richness) based on Akaike information criterion values. Quadrat as a categorical variable was determined to be the best predictor of H. japonicus cover. Humulus japonicus cover was found to be the best predictor of both P. arundinacea cover and species richness. Then, for each response variable, linear mixed-effects models were constructed with year, canopy cover, and distance along the transect (or its proxy) as fixed effects and transect as a random effect. Model selection was done through backward elimination (P-value for removal ≥ 0.05) using the lmerTest package (Kuznetsova et al. Reference Kuznetsova, Brockhoff and Christensen2017). All models were validated using QQ and residual plots, as well as formal tests of normality (Shapiro-Wilk). For models explaining P. arundinacea cover and species richness, shade and H. japonicus cover values were standardized, while P. arundinacea cover and species richness values were log + 1 transformed to reduce heteroskedasticity. Effect plots were created to understand interactions between variables, and points were extrapolated from the linear model predictions using raw data.
To determine whether species composition differed in quadrats invaded by H. japonicus (negative quadrats) versus quadrats not invaded by H. japonicus (positive quadrats), a permutational multivariate analysis of variance (PERMANOVA) was conducted. Analyses were done separately for October 2022 and October 2023. A Bray-Curtis dissimilarity index between quadrats was calculated for each analysis, and additionally, a nonmetric multidimensional scaling (NMDS) ordination with two dimensions was conducted using the MetaMDS function of the R package vegan (1,000 random starts) (Oksanen et al. Reference Oksanen, Simpson, Blanchet and Kindt2022). NMDS plots were created using ggplot2 and base R (Wickham Reference Wickham2016). Because we were interested in determining the effect of H. japonicus on species composition, H. japonicus was excluded from the species list and analysis. However, some quadrats contained only H. japonicus. Therefore, to ensure at least one species was present in each transect/quadrat combination, a dummy species was included in the analysis and given a small percentage of cover (0.01%) in all quadrats.
Transects acted as a random effect, while H. japonicus cover and shade were fixed effects. To account for this design, transect was the first term in the model, so variation among transects would be accounted for before testing other terms, and permutations were restricted so that quadrats were freely permutated within transects, but not across transects (permutations = 999). Terms were tested in order starting with transect, H. japonicus cover, and shade. To ensure that each step used the same permutations, set.seed was used. The model was analyzed using the adonis2 function within the vegan package (Oksanen et al. Reference Oksanen, Simpson, Blanchet and Kindt2022).
Results and Discussion
Invasion of Humulus japonicus and Replacement of Phalaris arundinacea
There was a positive correlation between the readings from the Li-Cor sensor and the readings made using the spherical densiometer (Kendall’s tau = 0.30, P-value <0.001), indicating that increased canopy led to decreased light availability to the herbaceous layer. Additionally, canopy cover in October 2022 was a significant predictor of canopy cover in October 2023 (ANOVA with transect included as random effect; F = 31.92, df = 1, P-value < 0.001). There was not a significant difference in canopy cover in October 2022 compared with October 2023, meaning that any changes in H. japonicus cover, P. arundinacea cover, and species richness were not due to changes in canopy cover.
To determine changes in H. japonicus and P. arundinacea cover from October 2022 to October 2023, we compared the cover of each species, separately, under all explanatory variables. The model explaining H. japonicus cover included fixed effects of year, canopy cover, quadrat position (categorical), date:canopy cover interaction, and a canopy cover:quadrat interaction. The model explaining P. arundinacea cover included fixed effects of canopy cover, quadrat position (H. japonicus cover), and a canopy cover:quadrat interaction. The random effect of transect had significant impacts on both H. japonicus and P. arundinacea (Tables 1 and 2). Humulus japonicus cover was significantly impacted by transect, year, canopy cover, and quadrat (Table 1), while P. arundinacea cover was significantly impacted by transect and quadrat, which was represented by H. japonicus cover (Table 2).
ANOVA results for Humulus japonicus cover. a
a Marginal R2 = 0.43.
b Mean square.
c P-values in bold indicate statistical significance.
ANOVA results for log + 1–transformed Phalaris arundinacea cover. a
a Marginal R2 = 0.27.
b Mean square.
c P-values in bold indicate statistical significance.
Humulus japonicus percent cover was high in October 2022, much lower in June 2023, and greatest in October 2023 (Figure 3A). Mean H. japonicus cover greatly increased from 7.76% in June 2023, the beginning of the growing season, to 43.9% in October 2023, the end of the same growing season. Generally, H. japonicus cover increased at each quadrat along the transect (Figure 3A), meaning that it grew in areas where it had previously colonized as well as into areas where it had yet to colonize in 2022. From October 2022 to October 2023, mean H. japonicus cover increased from 32.4% to 43.9%, indicating an expansion of invasion of H. japonicus across the site. Quadrat as a categorical variable significantly impacted H. japonicus cover (Figure 3A). In positive quadrats, H. japonicus mean cover increased from 9.33% to 21.8% in 2023, meaning that H. japonicus has continued to spread throughout the field site. Even in negative quadrats, H. japonicus grew more densely in 2023, increasing its mean percent cover from 57.2% to 65.9%.
Standard error of the mean (SEM) of (A) Humulus japonicus percent cover and (B) Phalaris arundinacea percent cover in quadrats across all transects in October 2022, June 2023, and October 2023.
For both October 2022 and October 2023, areas at the site with higher canopy cover had lower H. japonicus cover (Figure 4). However, there was a significant interaction between year and canopy cover, indicating that H. japonicus cover was greater under low canopy cover in October 2023 compared with October 2022 (Table 1; Figure 4A). This may be due to the colonization of areas under low canopy cover that had not yet been invaded by H. japonicus in October 2022 but had been colonized by October 2023. The interaction between canopy cover and quadrat was also significant (Table 1; Figure 4A). In the negative quadrats, H. japonicus cover decreased more rapidly as canopy cover increased compared with the positive quadrats. Because the positive quadrats were beyond the H. japonicus invasion front in October 2022, it is possible that H. japonicus had not yet reached equilibrium with the environmental conditions, such as shade, in the positive quadrats. Given more time, the relationship between H. japonicus cover and canopy cover is expected to be similar in both positive and negative quadrats. Another reason for this interaction could be due to the presence of P. arundinacea present in the positive quadrats in October 2022. Because P. arundinacea was on the other side of the invasion front, it may have hindered the spread of H. japonicus due to competition, regardless of canopy cover. Because P. arundinacea develops dense stands and a thick layer of thatch, it inhibits seedling establishment by other species (Thomsen et al. Reference Thomsen, Brownell, Groshek and Kirsch2012). However, based on its ability to invade established stands of P. arundinacea, we speculate that, given sufficient light, H. japonicus invasion into other, native vegetation would be even more rapid.
Effect plot showing Humulus japonicus cover in relation to canopy cover percentage in (A) both October 2022 and October 2023 and (B) negative (invaded in 2022), neutral, and positive quadrats (not invaded in 2022).
Because H. japonicus was observed invading P. arundinacea stands at the Joslin field site, changes in P. arundinacea cover were examined from October 2022 to October 2023. As H. japonicus cover increased from October 2022 to October 2023, P. arundinacea cover decreased, suggesting the invasion of H. japonicus and the replacement of P. arundinacea. Phalaris arundinacea mean cover was stable at 19.3% in October 2022 and 19.3% in June 2023, then decreased to 16.5% in October 2023, although the change between years was not significant (Table 2; Figure 3B). Humulus japonicus cover, which differed along the transect due to the study design, negatively affected P. arundinacea cover. Phalaris arundinacea was found in quadrats within the H. japonicus invasion front in June 2023. However, by October 2023, P. arundinacea mean cover had decreased in these negative quadrats, declining from 10.6% to 5.03% (Figure 3B). Thus, P. arundinacea, a perennial grass, established earlier in the growing season, while H. japonicus, an annual vine, established later but was able to outgrow P. arundinacea by the end of the growing season. In quadrats beyond the H. japonicus front, P. arundinacea cover decreased from 26.0% in October 2022 to 25.4% in October 2023, suggesting that H. japonicus was able to expand its invasion, albeit only slightly (Figure 3B).
There was also a significant interaction between canopy cover and quadrat (Table 2; Figure 4B). As H. japonicus cover in quadrats decreased, the relationship between P. arundinacea cover and canopy cover changed from a positive to negative relationship. This means that for quadrats beyond the H. japonicus invasion, P. arundinacea cover decreased with increased canopy cover, consistent with previous studies demonstrating the shade intolerance of P. arundinacea. However, for quadrats within H. japonicus growth, P. arundinacea cover increased with increased canopy cover, which may be due to the decreased competitive ability of H. japonicus in shaded but occupied quadrats.
The invasion of H. japonicus across the study site has resulted in the replacement of P. arundinacea. Before the invasion of H. japonicus around 2015, P. arundinacea previously dominated open-canopy areas that had been planted with acorns or left unplanted in 1998 (Matthews et al. Reference Matthews, McIntyre, Peralta and Rodgers2020; Spyreas et al. Reference Spyreas, Wilm, Plocher, Ketzner, Matthews, Ellis and Heske2010). Humulus japonicus was present at the site by 2015 and still present in 2020 (Charles Reference Charles2021), although the domination of H. japonicus and intrusion into P. arundinacea stands was not observed until 2022. These previous observations demonstrate a decrease in P. arundinacea cover as H. japonicus has invaded throughout the site.
Our study suggests that the invasion of H. japonicus in this restored wetland is ongoing. Our results show an increase in H. japonicus cover throughout the site between years, which corresponded with a slight decrease in P. arundinacea cover. However, P. arundinacea was able to persist as an invader at the site, which may be due to temporal niche partitioning, when species’ niches are separated by time (Carothers et al. Reference Carothers, Jaksić and Jaksic1984). Its rapid growth and early establishment in June, compared with the later establishment of H. japonicus in the growing season, allows it to survive, even in October, when H. japonicus has completely covered it. It is important to note that its invasion dynamics, specifically its interactions with P. arundinacea and response to canopy cover, may be different for sites that are at equilibrium, where H. japonicus spread has stabilized.
One concern about the replacement of P. arundinacea by H. japonicus is the alteration of ecosystem functions. Phalaris arundinacea was introduced into the United States for soil stabilization and erosion control to provide aid for susceptible ecosystems (Lavergne and Molofsky Reference Lavergne and Molofsky2004). It has a dense root system and spreads laterally via rhizomes (Apfelbaum and Sams Reference Apfelbaum and Sams1987; Lavergne and Molofsky Reference Lavergne and Molofsky2004), while H. japonicus is an annual plant with a shallow root system (Pannill et al. Reference Pannill, Cook, Hairston-Strang and Swearingen2009). Previous work has found that under low water-flow conditions, P. arundinacea produced adventitious roots that protected soils from erosion by higher, swifter flows (Ree Reference Ree1976). The replacement of dense roots that protect waterways from erosion by shallow roots that are easily washed away by flooding events leaves soil bare and increases erosion in wetlands. Erosion has detrimental consequences for wetlands, including wetland loss (Cellone et al. Reference Cellone, Carol and Tosi2016) and sediment mobilization (Castillo et al. Reference Castillo, Rubio-Casal, Luque, Nieva and Figueroa2002).
Effects of Humulus japonicus Invasion on Species Richness and Plant Community Composition
The model explaining species richness included fixed effects of year and quadrat position (H. japonicus cover). Species richness was significantly impacted by year, quadrat, and the random effect of transect (Table 3). Transects were established in both forested and open areas, and species richness differed among transects, with higher species richness along forested transects compared with open transects, defined by the original tree-planting treatments. In October 2022, species richness averaged 3.09 species per quadrat in forested transects and 1.54 species per quadrat in open transects, while in October 2023, species richness averaged 4.23 species per quadrat in forest transects and 1.63 species per quadrat in open transects. Species richness also differed significantly between years (Table 3). From October 2022 to October 2023, species richness increased from an average of 2.31 to 2.93 species per quadrat (Figure 5). Humulus japonicus cover, which differed along the transect, also significantly affected species richness (Table 3). Species richness was greater where there was less H. japonicus cover, which includes quadrats beyond the H. japonicus invasion front in October 2022 (Figure 5).
ANOVA results for log + 1–transformed species richness. a
a Marginal R2 = 0.31.
b Mean square.
c P-values in bold indicate statistical significance.
Standard error of the mean (SEM) of species richness in quadrats across all transects in October 2022 and October 2023. Species richness data do not include either Humulus japonicus or Phalaris arundinacea.
Interestingly, there was an increase in species richness from year to year, even with the increase in H. japonicus. The increase in richness might be explained by interannual variation in weather and the general spatial and temporal variability of annual ground cover vegetation inherent to floodplain ecosystems, which can result in large fluctuations in species richness (Jonas et al. Reference Jonas, Buhl and Symstad2015). Species richness increased with higher canopy cover, most likely due to the shade intolerance of both H. japonicus and P. arundinacea. Additionally, species richness was reduced in areas where H. japonicus cover was greater, due its strong competitive abilities that allow it to be a dominant species (Balogh and Dancza Reference Balogh, Dancza, Tokarska-Guzik, Brock, Brundu, Child, Daehler and Pyšek2008).
Not only did H. japonicus invasion affect species richness, it also affected species composition. There were differences in species composition along transects, as half the transect was within the H. japonicus invasion front and the other half was beyond the invasion front. The PERMANOVA indicated that there was a significant difference in species composition between transects and along transects in both October 2022 and October 2023 (Table 4). Transect significantly impacted species composition due to the location of the transect in forested versus open areas. Species composition was also significantly impacted by H. japonicus cover (Table 4). Similarly, the NMDS plots show slight differences between quadrats invaded by H. japonicus and quadrats not invaded by H. japonicus (Figure 6). There is a tendency for non-invaded quadrats to appear on the right of NMDS axis 1 and for invaded quadrats to appear on the left, which suggests some differentiation of species composition. Phalaris arundinacea tended to appear on the left of the NMDS plots with invaded quadrats, while other species, such as Sicyos angulatus L. (oneseed bur cucumber), [Pilea pumila (L.) A. Gray] (Canadian clearweed), and Galium aparine L. (stickywilly; Figure 6A), and P. pumila, Morus alba L. (white mulberry), Bidens frondosa L. (devil's beggartick), and [Toxicodendron radicans (L.) Kuntze] (eastern poison ivy; Figure 6B), appeared on the right of the plots, in quadrats not invaded by H. japonicus or P. arundinacea, which form monospecific stands.
Permutational multivariate analysis of variance (PERMANOVA) results for vegetation community composition differences between plots invaded and not invaded by Humulus japonicus.
a Sum of squares.
b Values in bold indicate statistical significance.
Non-metric multidimensional scaling (NMDS) for permutational multivariate analysis of variance (PERMANOVA) in (A) October 2022 (stress = 0.22) and (B) October 2023 (stress = 0.18). (A) Species with relative cover >5% are shown in the plot (GAL.APA, Galium aparine; PER.PEN, Polygonum pensylvanicum L. (Pennsylvania smartweed); PHA.ARU, Phalaris arundinacea; PIL.PUM, Pilea pumila; SIC.ANG, Sicyos angulatus). (B) Species with relative cover >4% are shown in the plot (BID.FRO, Bidens frondosa; MOR.ALB, Morus alba; PHA.ARU, Phalaris arundinacea; PIL.PUM, Pilea pumila; TOX.RAD, Toxicodendron radicans).
Humulus japonicus Control Methods
Understanding the negative relationship between H. japonicus abundance and canopy cover helps inform H. japonicus management. Our results were consistent with previous knowledge about the ecology of H. japonicus, suggesting that it has a preference for sunlit conditions, often being found in areas with gaps in canopy cover (Pannill et al. Reference Pannill, Cook, Hairston-Strang and Swearingen2009; Pasiecznik Reference Pasiecznik2022). Therefore, the findings from this study support previous recommendations for H. japonicus management by increasing canopy cover by planting native trees (Guyon and Cosgriff Reference Guyon and Cosgriff2022; Pannill et al. Reference Pannill, Cook, Hairston-Strang and Swearingen2009). However, due to the climbing nature of H. japonicus as a vine, trees need to be fast growing to establish canopy cover quickly. Even more importantly, H. japonicus, if established early, could overtop and possibly kill planted trees. A previous study by Guyon and Cosgriff (Reference Guyon and Cosgriff2022) looked at controlling H. japonicus by planting two fast-growing tree species, eastern cottonwood (Populus deltoides W. Bartram ex Marshall) and American sycamore (Platanus occidentalis L.), in combination with herbicidal treatments. They determined that planting containerized trees may be viable for long-term H. japonicus control, when combined with herbicide application for 3 to 5 yr to allow native trees to grow tall enough and escape getting overtopped. Additionally, monitoring invaded and restored sites is essential for long-term H. japonicus control. Monitoring should occur later in the growing season to be effective, because H. japonicus reaches peak cover near the end of the growing season. If monitored too early, H. japonicus dominance will be underestimated.
Protecting wetland ecosystems is extremely important, especially for ecosystem resilience against future invasions. To manage H. japonicus populations, mechanical control is often used, although this is only effective for small, concentrated populations (Guyon and Cosgriff Reference Guyon and Cosgriff2022). Chemical control, including both pre- and postemergence herbicides, can be effective (Pannill et al. Reference Pannill, Cook, Hairston-Strang and Swearingen2009; Steffen and Edgin Reference Steffen and Edgin2017), although this is only short term (Guyon and Cosgriff Reference Guyon and Cosgriff2022). Additionally, for ecosystems with frequent flooding, such as floodplain forests, the effects of herbicides may be short-lived and affect native species (Guyon and Cosgriff Reference Guyon and Cosgriff2022). Biological control, which is only effective when there is high host specificity, is difficult and not feasible for H. japonicus control due to its similarity to the native common hop (Humulus lupulus L.) (Guyon and Cosgriff Reference Guyon and Cosgriff2022; Pannill et al. Reference Pannill, Cook, Hairston-Strang and Swearingen2009; Steffen and Edgin Reference Steffen and Edgin2017). Prescribed burns are ineffective and may even stimulate H. japonicus growth (Steffen and Edgin Reference Steffen and Edgin2017). However, H. japonicus is an annual invasive vine, not a perennial, so controlling its seedbank, which remains viable in soil for up to 3 yr (Pannill et al. Reference Pannill, Cook, Hairston-Strang and Swearingen2009; Urziceanu et al. Reference Urziceanu, Cîșlariu, Nagodă, Nicolin, Măntoiu and Anastasiu2022), could potentially control H. japonicus populations.
In managing H. japonicus, consequences of its removal, such as increased soil erosion in riparian wetlands, need to be mitigated. Humulus japonicus populations must be eradicated and native vegetation, including grasses and sedges that spread laterally through rhizomes to effectively hold soil must be reestablished (Zuazo and Pleguezuelo Reference Zuazo, Pleguezuelo, Lichtfouse, Navarrete, Debaeke, Véronique and Alberola2009). Fast-growing species, including trees, are also effective at erosion control and H. japonicus management due to fast growth of both aboveground and belowground biomass (Burylo et al. Reference Burylo, Rey, Mathys and Dutoit2012). However, care must be taken when removing invasive species and reestablishing native vegetation in invaded ecosystems. The removal of invasive species often creates a “weed-shaped hole” that allows reinvasion by the same invader or different invaders (Buckley et al. Reference Buckley, Bolker and Rees2007). Several studies (e.g., Hulme and Bremner Reference Hulme and Bremner2005; Magnoli et al. Reference Magnoli, Kleinhesselink and Cushman2013; Pavlovic et al. Reference Pavlovic, Leicht-Young, Frohnapple and Grundel2009) have found an increase in invasive, rather than native, plant species, after the removal of an invasive plant species. Therefore, there is the need to quickly reestablish native species in wetlands not only to restore biodiversity, but also to prevent reinvasion by invasive species.
Future Directions
Increased anthropogenic disturbances will only facilitate the spread of invasive species into new ecosystems (Zimmermann et al. Reference Zimmermann, Brandt, Fischer, Welk and Von Wehrden2014). Especially for emerging invaders, limiting the introduction and dispersal of the species is more cost-effective and efficient relative to post-invasion removal (Davies and Sheley Reference Davies and Sheley2007). Identifying suitable habitat and, for species that are dispersed via water, determining dispersal corridors along waterways are important for preventing spread to new ecosystems (Urziceanu et al. Reference Urziceanu, Cîșlariu, Nagodă, Nicolin, Măntoiu and Anastasiu2022). Tools such as species distribution models (Srivastava et al. Reference Srivastava, Lafond and Griess2019) and invasion risk maps (Rodríguez-Merino et al. Reference Rodríguez-Merino, García-Murillo, Cirujano and Fernández-Zamudio2018; Urziceanu et al. Reference Urziceanu, Cîșlariu, Nagodă, Nicolin, Măntoiu and Anastasiu2022) can be used to mitigate the effects of current and future H. japonicus invasions.
An important part of invasive species management is restoring native vegetation while controlling invasive plant populations (Catford Reference Catford2016; Funk et al. Reference Funk, Cleland, Suding and Zavaleta2008). Therefore, there is a need for further studies focusing on the interactions between H. japonicus and species that it coexists with, both native and invasive species, such as P. arundinacea. With rapid global change affecting ecosystems and species interactions (Kuebbing and Nuñez Reference Kuebbing and Nuñez2015), it is even more important to study how these interactions shift under variable environmental conditions to further inform management and restoration. Continued research on H. japonicus is essential for a complete understanding of its ecology, role as an invader, and interactions with native species.
Data availability
All data are available through the University of Illinois Data Bank (https://doi.org/10.13012/B2IDB-6760644_V1 ).
Acknowledgments
Thanks to Brian Charles, Luca Lee, and Matt Finzel for help with vegetation sampling and plant identification; Christopher Evans, Greg Spyreas, David Zaya, Wes Bollinger, David Junga, Chelsea Peterson, and Jacob Ridgway for comments on the article; and David Zaya and Becky Fuller for help with statistical analysis.
Author contributions
Both authors contributed to the study conception and design. AHH conducted the study and wrote the first draft of the manuscript. JWM supervised the research and edited the final manuscript.
Funding statement
This project was supported by National Institute of Food and Agriculture, U.S. Department of Agriculture, under Hatch project 1018621.
Competing interests
The authors have no relevant financial or non-financial interests to disclose.
Open access
