Management Implications
Butomus umbellatus (flowering rush) is an invasive emergent monocot commonly found along the U.S.–Canadian border that can grow rapidly and cause numerous water-use issues as a monotypic population. Within the United States, there are two cytotypes of the species, a diploid and triploid, which have slightly different reproductive strategies, with diploid reproducing both sexually and asexually and triploid plants reproducing only asexually. The current study sought to describe the life history, resource allocation, and phenology of diploid B. umbellatus. Diploid B. umbellatus in this study produced 8,000 rhizome bulbils m−2 on average, with one site having more than 15,000 rhizome bulbils m−2. Not only does diploid B. umbellatus produce large numbers of propagules, but up to 30% of total plant starch is found in rhizome bulbils, which will allow for longevity in the sediment. A successful management program will need to indirectly target the belowground biomass and starch storage of this plant. Because rhizomes and rhizome bulbils are at or below the sediment surface, they are difficult to directly treat with a herbicide or to harvest. Management programs should target aboveground biomass before July to remove new growth. Additional in-season applications should occur as plants regrow to limit rhizome bulbil formation. Repeat applications should reduce year-to-year recruitment, although management of B. umbellatus will be long term, as the propagule bank will need to be exhausted to completely control this species.
Introduction
Flowering rush (Butomus umbellatus L.) is an invasive perennial monocot primarily found in aquatic and wetland habitats along the U.S.–Canadian border. Butomus umbellatus was first documented in the St Lawrence River, which facilitated spread into the Great Lakes region by the mid-1950s (Bellaud Reference Bellaud, Gettys, Haller and Bellaud2009; Gunderson et al. Reference Gunderson, Kapuscinski, Crane and Farrell2016). Butomus umbellatus is thought to have had multiple introductions into North America due to two separate cytotypes, a triploid and diploid, that have been identified with at least six amplified fragment-length polymorphism genotypes (Anderson et al. Reference Anderson, Zeis and Alam1974; Gaskin et al. Reference Gaskin, Andreas, Grewell, Haefliger and Harms2021). Both cytotypes of B. umbellatus are capable of vegetative reproduction through fragmentation of the rhizome and production of rhizome bulbils developed along the rhizome and leaf axils (Hroudová and Zákravský Reference Hroudová and Zákravský1993; Thompson and Eckert Reference Thompson and Eckert2004). The diploid cytotype can also sexually reproduce through self-compatible flowers on an inflorescence containing 20 to 50 flowers (Thompson and Eckert Reference Thompson and Eckert2004). Current information about B. umbellatus biology and management largely centers on the triploid cytotype and its invasion within lakes, whereas the diploid cytotype and impacted wetlands remain understudied.
Within aquatic systems, plants must overcome unfavorable conditions such as low light availability and limited CO2 concentrations, which can impact photosynthesis and growth (Ralph et al. Reference Ralph, Durako, Enríquez, Collier and Doblin2007). Butomus umbellatus can grow both emergent and submersed leaves (often together) and can be found in water as deep as 6 m but is usually found at depths of <1.3 m (Carter et al. Reference Carter, Madsen and Ervin2018; Madsen et al. Reference Madsen, Wersal and Marko2016). Triploid B. umbellatus is typically found growing in lake systems, whereas diploid B. umbellatus is more commonly found in wetland systems with variable hydrology (Gunderson et al. Reference Gunderson, Kapuscinski, Crane and Farrell2016). Alongside reproductive differences, the triploid cytotype is more widespread within the United States, with populations found from Minnesota westward in lake or reservoir systems (Gaskin et al. Reference Gaskin, Andreas, Grewell, Haefliger and Harms2021; Liu et al. Reference Lui, Thompson and Eckert2005). Conversely, diploid B. umbellatus has been documented from Minnesota eastward and is prolific especially around the Great Lakes region within wetlands (Trebitz and Taylor Reference Trebitz and Taylor2007). Differences in geography and growing environment can impact a plant’s life history, resource allocation, and phenology.
Emergent wetland plants are subject to the environment, which can impact life-history characteristics such as growth and resource accumulation (MacNeill et al. Reference MacNeill, Mehrpouyan, Minow, Patterson, Tetlow and Emes2017; Scofield et al. Reference Scofield, Ruuska, Aoki, Lewis, Tabe and Jenkins2009). Phenology is studied to understand and generate timelines for critical plant processes such as peak biomass and resource allocation patterns that can provide managers with important information to optimize the timing of control activities (Clarke et al. Reference Clarke, Wersal and Turnage2023; Wersal et al. Reference Wersal, Cheshier, Madsen and Gerard2011, Reference Wersal, Madsen and Cheshier2013). Triploid B. umbellatus phenology has been documented in the Detroit Lakes system in Minnesota, where it was observed that aboveground biomass peaks in late summer and rhizomatic biomass peaks in late fall (Madsen et al. Reference Madsen, Wersal and Marko2016; Marko et al. Reference Marko, Madsen and Sartain2015). Due to differences in genetics between cytotypes and depths at which diploid B. umbellatus plants grow, characterizing phenology and resource allocation could yield important information for managing invaded sites. Whereas phenology focuses on timing of biomass changes or reproduction, resource allocation aims to understand locations within the plant where starch and other carbohydrates are typically stored (Clarke et al. Reference Clarke, Wersal and Turnage2023; Haram and Wersal Reference Haram and Wersal2023; Wersal et al. Reference Wersal, Cheshier, Madsen and Gerard2011, Reference Wersal, Madsen and Cheshier2013). Management targeting critical tissues and times of low resource allocation can reduce growth, prevent reproduction, impact nutrient acquisition, and limit annual recruitment (Scofield et al. Reference Scofield, Ruuska, Aoki, Lewis, Tabe and Jenkins2009).
By combining environmental data with trends in plant growth and reproduction, a detailed understanding of phenological timing can be developed for a plant species. Previous studies on B. umbellatus phenology focused on the triploid cytotype within the Detroit Lakes, MN (Madsen et al. Reference Madsen, Wersal and Marko2016; Marko et al. Reference Marko, Madsen and Sartain2015), but there is a paucity of data describing life history, resource allocation, and phenological timings for diploid B. umbellatus. Documenting diploid B. umbellatus phenology would permit management strategies to be optimized if growth of both cytotypes are similar. Therefore, the objectives of this study were to (1) evaluate the phenology of diploid B. umbellatus by determining times of peak biomass and starch allocation patterns from three field populations over two growing seasons in Ohio (one population) and New York (two populations) and (2) relate B. umbellatus growth to air temperature, water temperature, and water depth. It is hypothesized that diploid B. umbellatus will have similar phenological timings with respect to peak aboveground biomass in late summer (August or September) and peak rhizome biomass in late fall (November to December). It is also hypothesized that the starch allocation within the plant will be higher in the belowground structures, similar to the pattern documented in triploid B. umbellatus.
Materials and Methods
Sites chosen for this study were selected based on knowledge of historical B. umbellatus infestations, access for sampling, and in close proximity to the Great Lakes. The B. umbellatus populations in these three sites were previously determined to be diploid B. umbellatus genotype G4 by Gaskin et al. (Reference Gaskin, Andreas, Grewell, Haefliger and Harms2021). Subsequent genetic testing by the resource managers associated with each site reconfirmed the cytotype.
Study Site Description
Mentor Marsh, OH (41.73649°N, 81.29879°W) is a large emergent marsh complex managed by the Cleveland Museum of Natural History (CMNH) that receives hydrologic input directly from Lake Erie. Mentor Marsh historically was used as a dumping site for salt mine tailings, which caused widespread ecosystem degradation and native plant loss. The open niche space was quickly filled by [Phragmites australis (Cav.) Trin. ex Steud] haplotype M, which is an aggressive invader (Guo et al. Reference Guo, Lambertini, Nguyen, Zhen Li and Brix2014; M Yeager, personal communication, March 6, 2023). In 2013, CMNH began large-scale habitat restoration efforts to manage P. australis in the marsh, and the disturbance from the restoration work likely allowed B. umbellatus to expand its range within the marsh. Currently, CMNH’s goal has been to reduce B. umbellatus populations by hand pulling the plants directly. More recently, B. umbellatus invaded and has begun to spread in the wetland system.
Tonawanda Wildlife Management Area, NY (Tonawanda WMA; 43.10590°N, 78.48118°W) is a large emergent wetland complex managed by New York State Department of Environmental Conservation that undergoes seasonal water-level management through on-site water control structures. The wetland is naturally flooded in the spring (April) and can be manually or naturally drained before summer and, depending on water availability, reflooded in the fall. Historically, this site was part of glacial Lake Tonawanda until the lake naturally drained through Niagara Falls (Calkin and Brett Reference Calkin and Brett1978). Today, there are numerous dikes to control water flow within the system to create a waterfowl management area through seasonal flooding and draining. Butomus umbellatus was first discovered in Tonawanda WMA in 2009 and has since continued to spread throughout the marsh (Kennedy et al., Reference Kennedy, Appleby, Palermo, Bonk and Mahoney2018; Roster et al., Reference Roster, McMahon and Kahan2011; U.S. Geological Survey 2023). Tonawanda WMA staff currently aim to control P. australis and water chestnut (Trapa spp.) through herbicide applications and hand pulling, respectively.
Unity Island, NY (42.92999°N, 78.90453°W) is a wetland adjacent to the upper Niagara River near Buffalo, NY. The wetland was created in 2018 through beneficial use of dredged material from the Buffalo River federal navigation channel. Butomus umbellatus was first recorded within the Unity Island site in 2019. Butomus umbellatus is common in the mainstem upper Niagara River, where it can grow in at least 3 m of water (Gunderson et al. Reference Gunderson, Kapuscinski, Crane and Farrell2016). It is thought that propagules recruited to the Unity Island site dispersed from the adjacent upper Niagara River, and establishment was likely aided by general site disturbance during restoration efforts. Butomus umbellatus continues to spread and be an aggressive invader at this site.
Biomass
Sampling was conducted once per month at all three sites between May and November over 3 yr (2021 through 2023; n = 420 samples per site). During each sampling event, 20 samples were harvested from each study site using a polyvinyl chloride (PVC) coring device designed to remove the plant and soil within the area of the device (0.018 m2) (Madsen et al. Reference Madsen, Wersal and Woolf2007; Wersal and Madsen Reference Wersal and Madsen2018). Before sample collection, geographic coordinates, plant height, water depth, and presence of emergent leaves and inflorescences were recorded. Data loggers (HOBO pendants, Onset Computer Corporation, Bourne, MA) were deployed at two locations at each site to collect air and water temperature every 30 min throughout the sampling season each year. Plant samples were rinsed, placed into labeled plastic bags, and shipped overnight in a cooler on ice to Minnesota State University, Mankato, MN, for further processing.
In the lab, samples were washed and separated into tissue types described by Hroudová and Zákravský (Reference Hroudová and Zákravský1993), with aboveground tissue (leaves), rhizome and roots, inflorescences, rhizome bulbils, and vegetative bulbils formed at the base of the inflorescence (Hroudová and Zákravský Reference Hroudová and Zákravský1993; Hroudová Reference Hroudová1989). In the 3-yr span of sampling, bulbils formed at the base of the inflorescence were not recorded in any sample regardless of site. Thus, only rhizome bulbils were collected as a form of vegetative reproduction. The number of rhizome bulbils and leaves from each collected sample were also counted and recorded. Sorted tissues were put into separate paper bags and placed in a drying oven at 48 C for at least 72 h, until dry. Once dried, samples were weighed to the nearest 0.001 g then dry weight (DW) was divided by the area of the PVC coring device (0.018 m2) to determine grams of dry weight per square meter (g DW m−2) for each tissue. Densities of rhizome bulbils and ramets were calculated in a similar manner.
Starch Allocation
Tissue samples from each site and month were consolidated into sets of five; thus, tissue samples 1 to 5, 6 to 10, 11 to 15, and 16 to 20 became starch samples 1, 2, 3, and 4 respectively. Samples were then placed in a food processor until the tissue was roughly cut. Rough-cut biomass was then ground using a Cyclone Sample Mill 3010-014 (UDY Corporation, Fort Collins, CO) and sieved through a #40 mesh screen (1 mm). After the biomass samples were ground, 50 to 55 mg of the sample was transferred to a plastic centrifuge tube for the starch analysis. Starch percent dry weight (% DW) of each sample set was determined using the amylase/amyloglucosidase method via the commercially available STA-20 starch assay kit (Sigma Aldrich, St Louis, MO). Recovery, accuracy, and precision were assessed using several repeated measures with every starch extraction assay. Wheat (Triticum aestivum L.; 89% purity) and corn (Zea mays L.; 93% purity) starch standards were used to assess recovery, which was 90.0% ± 1.8 SE and 89.6% ± 1.4 SE for wheat and corn, respectively. The starch extraction assay accuracy was demonstrated by a five-point (dilution) standard curve that was constructed for every assay, and all absorbance points were consolidated and produced an r2 = 0.98. Finally, the precision of the assays was determined by duplicating three B. umbellatus samples for every extraction and computing the percent difference between repeated samples, which on average was 10.0% ±1.9 SE (Wersal et al. Reference Wersal, Cheshier, Madsen and Gerard2011).
Statistical Analysis
Monthly averages for tissue biomass, starch content, and environmental variables were calculated and pooled across study sites (N = 13,727 data points). Study site was included as a random variable in the construction of the regression models to account for its influence on model variance. Due to the high variability in the data set, an average and standard deviation were calculated per biomass and starch variable each month, then all data points that were ±1 SD were removed (4,500 of 13,727 data points) (Aguinis et al. Reference Aguinis, Gottfredson and Joo2013; Osborne and Overbay Reference Osborne and Overbay2019). Linear regression models were fit to determine the relationship between air temperature, water temperature, and water depth and diploid B. umbellatus biomass. Relationship strength was defined as no relationship (0 to 0.1), weak relationship (0.1 to 0.4), moderate relationship (0.4 to 0.6), strong relationship (0.6 to 0.9), or a perfect relationship (0.9 to 1; Dancey and Reidy Reference Dancey and Reidy2004), using absolute values. Kruskal-Wallis analyses were conducted between sites to determine site-specific differences in rhizome bulbil density and leaf density. If a difference was observed, a Dunn’s all-pairwise comparison was used to separate site-specific data. All analyses were conducted at α = 0.05 significance level using R statistical software v. 4.4.0 (R Core Team 2017). Packages used for analyses are as follows: dplyr (Wickham et al. Reference Wickham, Vaughan and Girlich2024), tidyr (Wickham et al. Reference Wickham, François, Henry, Müller and Vaughan2023), ggpubr (Kassambra Reference Kassambara2023a), rstatix (Kassambra Reference Kassambara2023b), lme4 (Bates et al. Reference Bates, Maechler, Bolker and Walker2015), lmerTest (Kuznetsova et al. Reference Kuznetsova, Brockhoff and Christensen2017), and MuMIn (Bartoń Reference Bartoń2024).
Results and Discussion
Overall, aboveground biomass peaked from July to September and comprised 50% to 70% of peak total biomass. Peak total B. umbellatus biomass was 1,970.40, 2,394.07, and 2,399.28 g DW m−2 for plants sampled at Mentor Marsh, Tonawanda, and Unity Island, respectively. Aboveground biomass in Mentor Marsh peaked at 1,410.18 ± 1,167.22, 624.10 ± 192.98, and 390.72 ± 49.11 g DW m−2 July 2021, August 2022, and July 2023, respectively (Figure 1). Aboveground biomass at Tonawanda peaked at 1,303.04 ± 98.77, 543.64 ± 29.26, and 574.47 ± 26.94 g DW m−2 in July 2021, June 2022, and July 2023, respectively (Figure 2). Aboveground biomass at Unity Island peaked at 1,006.31 ± 155.65, 1,306.55 ± 191.03, and 1,141.65 ± 61.33 g DW m−2 in September 2021, September 2022, and August 2023, respectively (Figure 3). Aboveground tissues on average stored less than 2.5% starch across all sampling sites and years, with the highest concentrations of starch occurring in May to June (Figures 1–3). Inflorescence tissue at each site was less than 10% of the total biomass and peaked between August and October. Inflorescence starch was more variable throughout the year, with low points around 0.6% and the highest concentration at 7% (Figures 1–3). Leaf density in Mentor Marsh peaked at 3,011 ± 796, 3,708 ± 780, and 1,963 ± 371 ramets m−2 in July to August for all sampling years (Figure 4). Leaf density at Tonawanda peaked at 2,400 ± 223, 1,560 ± 114, and 1,453 ± 185 ramets m−2 in June to July for each sampling year (Figure 4). Leaf density at Unity Island peaked at 2,061 ± 254, 2,629 ± 318, and 2,793 ± 271 ramets m−2 in June to August for each sampling year (Figure 4). Overall, leaf density was not different (P = 0.93) between Mentor Marsh and Unity Island, although B. umbellatus had fewer (P < 0.01) leaves per square meter at Tonawanda when compared with the other sampling sites.
Figure 1. Mean (±1 SE) biomass and starch content in diploid Butomus umbellatus harvested from Mentor Marsh, OH. Scaling for y axes varies based on the collected values for each measurement associated with the axis title.
Figure 2. Mean (±1 SE) biomass and starch content in diploid Butomus umbellatus harvested from Tonawanda, NY. Scaling for y axes varies based on the collected values for each measurement associated with the axis title.
Figure 3. Mean (±1 SE) biomass and starch content in diploid Butomus umbellatus harvested from Unity Island, NY. Scaling for y axes varies based on the collected values for each measurement associated with the axis title.
Figure 4. Mean (±1 SE) leaf density and rhizome bulbil density in diploid Butomus umbellatus harvested from each study site. Scaling for y axes varies based on the collected values for each measurement associated with the axis title.
Peak rhizome biomass occurred September to November and comprised 40% to 50% of total biomass. Peak rhizome biomass was 1,506.57 ± 488.3, 2,025.24 ± 128.82, and 936.41 ± 170.54 g DW m−2 for plants sampled at Mentor Marsh, Tonawanda, and Unity Island, respectively. Rhizome biomass in Mentor Marsh peaked at 593.27 ± 75.44, 540.94 ± 96.91, and 222.63 ± 33.08 g DW m−2 in September 2021, October 2022, and November 2023, respectively (Figure 1). Rhizome biomass at Tonawanda peaked at 2,025.24 ± 128.82, 372.28 ± 44.03, and 371.65 ± 32.78 g DW m−2 in September 2021, September 2022, and October 2023, respectively (Figure 2). Rhizome biomass at Unity Island peaked at 936.41 ± 170.54, 911.60 ± 94.47, and 442.44 ± 34.25 g DW m−2 in October of all years (Figure 3). Rhizomes stored a large proportion of total starch (16% to 28%) across sites and years, with peak storage between October and November. Low points in rhizome starch storage occurred from May to July.
Rhizome bulbil biomass was always less than 600 g DW m−2 for all sites over the course of the study (Figure 4). Rhizome bulbil densities peaked in the fall from September to November at 3,000 to 16,000 rhizome bulbils m−2 for each sampling site. Rhizome bulbil density in Mentor Marsh peaked at 15,333 ± 3,153, 11,757 ± 2,368, and 3,845 ± 967 bulbils m−2, respectively, for 2021, 2022, and 2023 (Figure 4). Rhizome bulbil density at Tonawanda peaked at 3,026 ± 442, 3,422 ± 429, and 2,124 ± 306 bulbils m−2, respectively for 2021, 2022, and 2023 (Figure 4). Rhizome bulbil density at Unity Island peaked at 12,802 ± 2,023, 11,422 ± 2,582, and 9,379 ± 1,490 bulbils m−2, respectively, for 2021, 2022, and 2023 (Figure 4). Rhizome bulbil densities at Mentor Marsh and Unity Island were not different (P = 0.34); however, fewer (P < 0.01) bulbils were produced on an annual basis at Tonawanda when compared with the other sampling sites. Even though rhizome bulbils accounted for a small proportion of total biomass, there were high amounts of stored starch that ranged from 15% to 30% throughout most of the year at each sampling site (Figures 1–3).
Rhizome starch content had strong relationships with air temperature (r2 = 0.52) and water temperature (r2 = 46) (Table 1). Other significant, though weak, relationships were found, including a positive relationship between aboveground biomass and air temperature (r2 = 0.17), a negative relationship between rhizome bulbil biomass and air temperature (r2 = 0.18), and a positive relationship between leaf density and air temperature (r2 = 0.17). Both water and air temperature are important factors that drive photosynthesis, plant development, and phenological timing of plants (Henne et al. Reference Henne, Hu and Cleland2007). During this study, when water temperature approached 25 C, there was a general decrease in the production of aboveground, rhizome, and inflorescence biomass, as well as a shift in the peak rhizome starch content. Water depth has an influence on the temperature that plants will experience, in that deeper water will disperse heat energy and thus result in cooler temperatures, while shallow water will result in higher temperatures (Erwin Reference Erwin2009). Temperature is a notable stressor on plants, mainly due to the production of reactive oxygen species (ROS), which inhibits gas movement, resulting in potential changes of phenological patterns in the plant (Hassanuzzaman et al. Reference Hassanuzzaman, Nahar, Mahabub, Roychowdhury and Fujita2013; Yamamoto et al. Reference Yamamoto, Aminaka, Yoshioka, Khatoon, Komayama, Takenaka, Yamashita, Nijo, Inagawa, Morita, Sasaki and Yamamoto2008). The creation of ROS causes oxidative stress in plants, which can lead to damage in the leaf and shoot tissues, thus reducing photosynthetic potentials (Marchland et al. Reference Marchand, Mertens, Kockelbergh, Beyens and Nijs2005). Heat stress can occur from modest increases (1 to 2 C) in temperature. In warm, dry years, heat stress may reduce aboveground biomass and change timing of starch storage and growth of rhizomes.
Table 1. Results of the linear regression analyses between the plant and environmental metrics of diploid Butomus umbellatus pooled across all study sites.
Deeper water (1 m) is not ideal for B. umbellatus growth, as a previous study reported that water depth had a negative relationship with triploid B. umbellatus growth (Madsen et al. Reference Madsen, Wersal and Marko2016). For diploid B. umbellatus, only plant height was related to water depth in the current study, and the relationship was positive (Table 1). Differences in growth response relating to water depth between the cytotypes may be due to the characteristics of the water bodies they were growing in. Triploid B. umbellatus is often found in areas with deeper and more stable hydrology (i.e., lake, reservoirs, and slow-moving rivers). In contrast, the sites sampled in this study for diploid B. umbellatus were shallow wetland complexes with fluctuating hydrology. Mentor Marsh had little standing water throughout the growing season, but water depth at Tonawanda WMA and Unity Island fluctuated between 10 and 60 cm (Figure 5). Differences in phenology demonstrate that B. umbellatus is influenced by the flooding regime that is present within an invaded wetland system (Pigliucci and Kolodynska Reference Pigliucci and Kolodynska2002). Wetlands likely offer the most optimal environment for B. umbellatus, as water levels often remain below 1 m, which allows for increased biomass and production of high densities of rhizome bulbils due to receding water levels (Hroudová Reference Hroudová1989; Madsen et al. Reference Madsen, Wersal and Marko2016c).
Figure 5. Water depth (cm) and mean air temperature (C) for study sites where diploid Butomus umbellatus samples were harvested. Scaling for y axes varies based on the collected values for each measurement associated with the axis title.
Diploid B. umbellatus can reproduce and spread vegetatively via rhizome fragmentation, rhizome bulbils, and inflorescence bulbils formed at the base of the inflorescence, although there were no inflorescence bulbils observed during this study. Diploid B. umbellatus has the potential to produce >15,000 rhizome bulbils m−2. In contrast, triploid B. umbellatus produces fewer but larger reproductive structures (600 rhizome buds m−2) (Marko et al. Reference Marko, Madsen and Sartain2015). Diploid B. umbellatus also reproduces sexually through flowering and seed production. Butomus umbellatus can produce 20 to 50 flowers per inflorescence and approximately 200 seeds per flower, with each seed having roughly 31% viability (Eckert et al. Reference Eckert, Massonnet and Thomas2000). Individual plants can produce multiple inflorescences, thus increasing the amount of seeds produced per square meter as well. Considering both reproductive strategies, diploid B. umbellatus could produce up to 18,000 individuals m−2 (15,000 rhizome bulbils m−2 with 3,000 viable seeds per inflorescence) during the reproductive season (fall) under optimal conditions, which is almost 30-fold higher than triploid B. umbellatus.
Overall, many of the life-history traits between triploid and diploid B. umbellatus are similar. Rhizome biomass and peak periods are similar (Marko et al. Reference Marko, Madsen and Sartain2015). Aboveground biomass in triploid biomass reached a peak of just over 500 g DW m−2, with diploid B. umbellatus producing between 600 and 1,100 g DW m−2 (Marko et al. Reference Marko, Madsen and Sartain2015). Both cytotypes had aboveground biomass peak between July and August, with the rhizome and rhizome bulbils reaching peak production October to November. Management strategies for both cytotypes should be similar and focus on two goals: (1) long-term biomass reduction and (2) reduction in vegetative propagules. Butomus umbellatus management is challenging, as rhizomes and rhizome bulbils combined store up to 60% of total starch and present a unique challenge to management, as these structures cannot be reached directly with herbicides (Parsons et al. Reference Parsons, Baldwin and Lubliner2019; Turnage et al. Reference Turnage, Wersal and Madsen2017). However, rhizome biomass and starch content decrease as aboveground biomass increases in June and July. Management programs should target aboveground biomass before July to remove new growth and further deplete starch content in the rhizome tissue. Additional herbicide applications should be utilized, as plants can recover, which ultimately exhausts starch reserves and limits rhizome bulbil formation (Bellaud Reference Bellaud, Gettys, Haller and Bellaud2009; Erwin Reference Erwin2009). Repeat applications should be sufficient to reduce year-to-year recruitment, although management of B. umbellatus will be long term, as the propagule bank will need to be exhausted to completely manage this species. Sites invaded by diploid B. umbellatus may see slightly different trends in production, with those sites further south potentially experiencing earlier biomass peaks. Monitoring populations closely would provide important information for when management should occur.
In addition to site-specific management, large-scale prevention programs should be developed to identify water bodies that have a high probability of invasion. Diploid B. umbellatus posesses some differences to its triploid counterpart due to the differences in reproductive output and environmental response (Banerjee et al. Reference Banerjee, Harms, Mukherjee and Gaskin2020; Gebhart and Wersal Reference Gebhart and Wersal2023). There is also evidence of differential diploid B. umbellatus growth based on site-specific characteristics. These site-specific characteristics should suggest that populations of B. umbellatus may need a particular approach when aiming for management and control. High reproductive output, especially through vegetative bulbils, increases the likelihood of diploid B. umbellatus spread, especially through aquatic systems that are associated with rivers. Based on the current distribution in the Great Lakes region, diploid B. umbellatus has access to the largest lakes and numerous river systems in North America that can move seeds, rhizome bulbils, and rhizome fragments. With the ability to produce almost 18,000 individuals m−2 during peak production, diploid B. umbellatus is an invasive plant cytotype that poses a high potential to expand to most of North America. However, a differential response to environmental conditions may not allow diploid B. umbellatus to expand westward due to increased temperatures seen in the western United States (Levin Reference Levin1983). Further temperature and phenology studies could be conducted to create a coherent profile for diploid B. umbellatus reproduction and life-history strategies within the United States.
Acknowledgments
We would like to thank the following personnel for assistance in the field or laboratory: Ian Knight, Blake DeRossette, Logan Byers, Rebecca Cantrell, Jacob Crestani, Stuart Davis, Sam Fellows, Jeremy Jones, Megan Landis, Amber Mack, Kaylie Malloy, Anna Mathis, Jessie McDonald, Charlotte Moore, Abby Penko, Aden Ricketts, James Rubish, Ryan Wilkinson, Heidi Kennedy, Denise Appleby, Danielle Meyers, Brandan King, Braeden Schmidt, Grace C. Costello, Rebecca G. Giangreco, and Jane E. Clark. We would also like to thank many people in the Aquatic Weed Science Lab at MNSU for their help processing samples and offering support throughout this time. The manuscript was reviewed in accordance with U.S. Army Engineer Research and Development Center policy and approved for publication. Citation of trade names does not constitute endorsement or approval of the use of such commercial products. The content of this work does not necessarily reflect the position or policy of the U.S. government and no official endorsement should be inferred.
Funding statement
The research was supported by the Great Lakes Restoration Initiative U.S. Army Corps Buffalo District through the U.S. Army Engineer Research and Development Center under cooperative agreement W912HZ2120036.
Competing interests
The authors declare no competing interests.
Open access
