Management Implications
Butomus umbellatus (flowering rush) is an invasive plant species with an adaptive growth form capable of growing in aquatic and wetland habitats. This requires resource managers to control B. umbellatus in a variety of environments, and resource managers therefore need multiple control strategies. In most aquatic environments, B. umbellatus is controlled with the use of herbicides labeled for use in aquatic sites; however, in some locations herbicides may be prohibited or may not be a feasible control option due to high rates of water exchange. In areas such as these, mechanical control in the form of clipping and harvesting may be a suitable control option. However, mechanical removal requires frequent repetition at 2-wk intervals, if control is the ultimate objective. Monthly or bimonthly clipping had varying degrees of success for control of B. umbellatus tissues and propagules. Two potential negative effects of mechanical control are spawn disturbance of fish or other desirable aquatic fauna and spreading B. umbellatus propagules in hydrosoil; areas targeted for clipping would need to have floating booms placed around them to capture floating propagules for containment and later disposal.
Introduction
Flowering rush (Butomus umbellatus L.) is an invasive plant in the northern United States and southern Canada (Anderson et al. Reference Anderson, Zeis and Alam1974; Kliber and Eckert Reference Kliber and Eckert2005) that impedes water flow, reduces human uses of water resources, and reduces biodiversity in aquatic and wetland habitats (Bellaud Reference Bellaud, Gettys, Haller and Bellaud2009). Butomus umbellatus is native to Europe and Asia and first entered the United States in 1928 (Muenscher Reference Muenscher1930; Tutin et al. Reference Tutin, Heywood, Burgess, Moore, Valentine, Walter and Webb1980). After introduction, B. umbellatus most likely expanded across North America via horticultural activities (Les and Mehrhoff Reference Les and Mehrhoff1999).
Butomus umbellatus has two biotypes (triploid and diploid) in its home and invaded ranges (Hroudova et al. Reference Hroudova, Krahulcova, Zakravsky and Jarolimova1996; Kliber and Eckert Reference Kliber and Eckert2005), suggesting multiple introductions to North America. While the diploid biotype can reproduce sexually, vegetative propagation via rhizome fragments and buds appears to be the primary means of dispersal and colonization for both B. umbellatus biotypes (Hroudova et al. Reference Hroudova, Krahulcova, Zakravsky and Jarolimova1996). Carter et al. (Reference Carter, Madsen and Ervin2018) found that while initial B. umbellatus propagule size affected final plant biomass over a growing season, it did not affect propagule survivorship, suggesting the smallest rhizome fragments or individual rhizome buds are sufficient to establish new colonies. In fact, rhizome fragments and buds can sprout leaves and roots while free floating in water (GT, personal observation), which further prepares propagules for invasion of new sites while in transit. Butomus umbellatus has an adaptive growth form (Sarbu et al. Reference Sarbu, Smarandache, Paraschiv and Mihai2009) and can grow as a wetland plant, a shallow-water emergent plant (0 to 1.2 m), or as a completely submersed plant (1.2 to 6 m; Rice et al. Reference Rice, Reddish, Dupuis and Mitchell2010), which facilitates survival in diverse environments with fluctuating water and light conditions (Carter et al. Reference Carter, Madsen and Ervin2018; Madsen et al. Reference Madsen, Woolf and Wersal2017). Additionally, because most of the starch content of B. umbellatus is stored in the belowground structures (Marko et al. Reference Marko, Madsen, Smith, Sartain and Olson2015) and because the species can produce hundreds of rhizome buds per square meter (Madsen et al. Reference Madsen, Wersal and Marko2016c), it is also capable of survival and regrowth after many control activities targeting individual colonies.
Much has been learned about the biology and ecology of B. umbellatus in its invaded range over the past two decades (Carter et al. Reference Carter, Madsen and Ervin2018; Eckert et al. Reference Eckert, Massonnet and Thomas2000; Gunderson et al. Reference Gunderson, Kapuscinski, Crane and Farrell2016; Madsen et al. Reference Madsen, Wersal and Marko2016c; Marko et al. Reference Marko, Madsen, Smith, Sartain and Olson2015); however, few effective control measures have been found to reduce B. umbellatus colonies. Several small-scale mesocosm and laboratory studies have identified herbicide chemistries that are effective for controlling B. umbellatus (Madsen et al. Reference Madsen, Sartain, Turnage and Marko2016a, Reference Madsen, Turnage and Getsinger2016b; Poovey et al. Reference Poovey, Mudge, Thum, James and Getsinger2012, Reference Poovey, Mudge, Getsinger and Sedivy2013; Wersal et al. Reference Wersal, Poovey, Madsen, Getsinger and Mudge2014), but few of these have been further substantiated with field trials. In fact, the contact herbicide diquat is the only chemical control strategy with documented operational use for the control of B. umbellatus (Madsen et al. Reference Madsen, Sartain, Turnage and Marko2016a; Turnage et al. Reference Turnage, Alcott and Guetter2018). While control is evident with large-plot submersed treatments of diquat, repeated treatments may be required (Madsen et al. Reference Madsen, Sartain, Turnage and Marko2016a; Turnage et al. Reference Turnage, Alcott and Guetter2018), and in locations where threatened or endangered species are present, herbicide applications may be prohibited.
Resource managers need alternatives to chemical control methods for B. umbellatus management in areas that restrict chemical options and for management of herbicide resistance in treated populations. At this time, no biological control options have been identified for B. umbellatus. Madsen et al. (Reference Madsen, Woolf and Wersal2017) investigated the use of annual drawdown as an alternative to chemical control of B. umbellatus and found that physical (benthic barriers) and mechanical (hand pulling and excavation) control methods were ineffective. The fact that B. umbellatus was already present at the test sites suggests that drawdown is an ineffective control measure (Madsen et al. Reference Madsen, Woolf and Wersal2017).
Using mechanical harvesting to manage aquatic nuisance plants has not been investigated for B. umbellatus. Clipping, done often enough, should reduce nutrient stores in the rhizome complex of B. umbellatus comparable to repeated contact herbicide applications. Therefore, the purpose of this study was to evaluate short- (over a growing season) and long-term (across years) efficacy of mechanical harvesting on B. umbellatus.
Materials and Methods
Experiments were conducted twice (2016 and 2017) at the R.R. Foil Plant Research Center’s Aquatic Plant Research Facility at Mississippi State University in 378-L (100-gal) mesocosms. Six treatments were used over a 4-mo period: a nontreated reference, plants that received two submersed diquat (Harvester® Aquatic Herbicide, Applied Biochemists, 1200 Bluegrass Lakes Parkway, Alpharetta, GA 30004) applications (0.19 ppmv ai) 1 mo apart, plants clipped twice per month (biweekly, eight clippings), plants clipped every month (four clippings), plants clipped every other month (bimonthly, two clippings), and plants clipped once per growing season (Table 1). Mesocosm water treated with diquat was in contact with plants for 12 h, then mesocosms were drained and refilled. Pots (3.78 L) filled with sand amended with fertilizer (Osmocote® 19-6-12 fertilizer, Scotts-Sierra Horticultural Products, 14111 Scottslawn Road Marysville, OH 43041) at a rate of 2 g L−1 of sand were planted with two 8-cm-long (3-inch-long) B. umbellatus rhizome segments. Ten pots of B. umbellatus were placed in each mesocosm. Mesocosms were filled with water to a volume of 216 L (16-inch depth), and plants were given 1.5 mo to become established. Each treatment was replicated in three mesocosms. Two extra mesocosms were planted for harvesting pretreatment data for a total of 20 mesocosms and 200 pots. Pretreatment plants (20 pots) were harvested the day before the first treatments were administered. Harvested plant tissues were separated into above- and belowground biomass, placed in labeled paper bags, and dried in a forced-air oven at 70 C for 5 d. Rhizome bud number was recorded before drying. Biomass weights were recorded after drying.
Table 1. Diquat and clipping treatments.
Mechanically treated plants were clipped approximately 5 cm (2 inches) above the sediment surface. Clipped biomass was removed from the mesocosms to simulate mechanical control from harvesting-boat operations in field settings. At 16 wk after initial treatment (WAT), half the pots in each mesocosm were harvested in the same manner as pretreatment samples. The remaining pots were harvested at 52 WAT.
Samples were analyzed statistically using a mixed model to determine whether significant differences existed in mean biomass and bud density. Treatment was considered a fixed variable, while year of experiment (2016 and 2017) was considered a random variable. Differences in means were further separated using a Tukey’s post hoc test. All statistical tests were conducted at the P ≤ 0.05 significance level in the statistical software R using the ‘lmerTest’ and ‘emmeans’ packages (R Core Team 2017).
Results and Discussion
At 16 WAT, aboveground biomass of B. umbellatus was only reduced by diquat treatments (80%) and biweekly clippings (95%) when compared with reference plants (Figure 1). However, control of aboveground biomass at 16 WAT was not significantly different between all clipping and diquat treatments. At 52 WAT, only biweekly clippings had reduced B. umbellatus aboveground biomass (95%) compared with reference plants, but monthly clippings had the same level of control as biweekly clippings.
Figure 1. Butomus umbellatus aboveground biomass, belowground biomass, and rhizome bud density at 16 and 52 wk after initial treatment (WAT). Solid lines represent pretreatment levels of each plant metric. Error bars are 1 standard error of the mean. Bars sharing the same letter are not statistically different from one another at the P = 0.05 level of significance. Time periods (16 and 52 WAT) were analyzed separately.
Butomus umbellatus belowground biomass was reduced 71% by diquat, 76% with monthly clippings, and 91% with biweekly clippings when compared with reference plants at 16 WAT. Clipping every other month (bimonthly) had the same level of control on B. umbellatus belowground biomass as biweekly clippings at 16 WAT. At 52 WAT, only monthly (84%) and biweekly (99%) clipping reduced belowground tissues compared with reference plants; however, the two sequential diquat applications caused the same level of control as both monthly and biweekly clippings, while bimonthly clippings had the same level of control as diquat and monthly clippings (Figure 1).
Reduction of B. umbellatus rhizome bud density was only achieved by diquat (66%), monthly (70%), and biweekly (81%) clippings compared with reference plants at 16 WAT. Rhizome bud density was still controlled by diquat (54%), monthly (78%), and biweekly (89%) clippings compared with reference plants at 52 WAT. Bimonthly clippings had the same level of control over B. umbellatus rhizome buds as diquat at 52 WAT (Figure 1).
These data are the first to provide evidence that mechanical control of B. umbellatus is possible. In locations where herbicide use is restricted, mechanical harvesting may be the only option to control B. umbellatus. However, there are concerns that should be considered before initiating a mechanical harvesting program. For example, the timing of clipping/harvesting would need to be such that it did not coincide with spawning of desirable fish species, as the cutter head would likely kill or injure any fish or other desirable aquatic fauna it contacts, or kill these same organisms by trapping them in harvested vegetation (Booms Reference Booms1999; Mikol Reference Mikol1985; Wile Reference Wile1978). Also, the cutter head could disturb spawning sites, as vegetation below the cutter head would move in such a way as to disturb the habitat around it. Additionally, if B. umbellatus is growing in a heterogeneous plant stand with desirable native species, clipping/harvesting activities may need to be avoided or timed when B. umbellatus would be controlled and desirable plant species avoided (e.g., B. umbellatus is the only sprouted species present or the tallest species).
Similar studies examining mechanical control of other invasive aquatic and wetland plant species have also shown that control was enhanced with higher-frequency clipping events (Derr Reference Derr2008; Wile Reference Wile1978). While effective at high intensity (every 2 wk), mechanical control via clipping/harvesting may be too time-consuming and expensive (Bryant Reference Bryant1970, Reference Bryant1974; Charudattan Reference Charudattan and Riches2001; Haller Reference Haller, Gettys, Haller and Bellaud2009) to see wide use among resource managers dealing with B. umbellatus infestations. Culpepper and Decell (Reference Culpepper and Decell1978) found that less than 9,072 kg (10 tons) of waterhyacinth [Eichhornia crassipes (Mart.) Solms] and less than 4,536 kg (5 tons) of hydrilla [Hydrilla verticillata (L. f.) Royle] could be removed per hour from a body of water via clipping/harvesting; which was lower than their minimum stated management goal of 72,575 kg (80 tons) h−1. Culpepper and Decell (Reference Culpepper and Decell1978) also found that the pace for clipping/harvesting operations is usually set by the time needed to transport clipped vegetation to shore and/or disposal sites rather than the rate at which the harvester operates.
If B. umbellatus populations are large, mechanical clipping/harvesting may not provide control, as harvester boats may not be capable of clipping an entire infestation in 2 wk due to their slow speeds (Culpepper and Decell Reference Culpepper and Decell1978) and need to periodically offload clipped vegetation (Culpepper and Decell Reference Culpepper and Decell1978; Haller Reference Haller, Gettys, Haller and Bellaud2009; Newroth Reference Newroth1979; Unmuth et al. Reference Unmuth, Sloey and Lillie1998). Additionally, harvesting operations can release vegetative propagules from clipped aquatic plants, which can drift away in water currents and infest new sites (Culpepper and Decell Reference Culpepper and Decell1978; Haller Reference Haller, Gettys, Haller and Bellaud2009). Placing floating booms around B. umbellatus sites targeted for clipping may help to contain dislodged propagules for later collection and disposal.
Mechanical harvesting could be effectively used as part of an integrated management approach for B. umbellatus in areas where herbicide use may not reach an appropriate concentration exposure time. Similarly, clipping/harvesting could be useful as a short-term control option for small B. umbellatus colonies until resource managers can identify other more suitable control methodologies for their management areas. Regardless, these data indicate that both chemical and mechanical options can be effective alone, and possibly together, to manage this troublesome plant. Future studies should investigate the use and timing of clipping/harvesting as part of an integrated control strategy using chemical, physical, and mechanical control techniques on B. umbellatus.
Author ORCIDs
Lee Turnage, 0000-0003-3239-6663
Acknowledgments
We would like to thank Sam Hansen, Steven Geary, Nick Bailey, Mary Nunenmacher, Cory Shoemaker, Wesley Presnall, and Zay Speed for helping with this study. This publication is a contribution of the Mississippi Agricultural and Forestry Experiment Station and the Mississippi State University Cooperative Extension Service. Mention of a trade name does not constitute an endorsement of that product by the USDA or MSU over others not listed. No conflicts of interest have been declared.
