Impacts of invasive Asian (Amynthas hilgendorfi) and European (Lumbricus rubellus) earthworms in a North American temperate deciduous forest
- First Online:
- Cite this article as:
- Greiner, H.G., Kashian, D.R. & Tiegs, S.D. Biol Invasions (2012) 14: 2017. doi:10.1007/s10530-012-0208-y
- 644 Views
Current understanding of earthworm invasions in North America is founded on studies of European species belonging to a single family (Lumbricidae); the ecological effects of taxa from other regions are largely unknown, despite many reports of established populations. Amynthas (Megascolecidae), a genus of invasive Asian earthworm, has increasingly been documented in North American regions lacking native earthworms. We present results from complimentary field and laboratory experiments designed to (1) evaluate potential impacts of Amynthas hilgendorfi on forest–floor nutrient cycling, leaf-litter decomposition, and soil structure, (2) compare these impacts to those of a better-understood invasive European species, Lumbricus rubellus, and (3) test for interactive effects between these species. While each species increased litter-decomposition rates in laboratory mesocosms, the effect of L. rubellus was greater than that of A. hilgendorfi. Each species also increased concentrations of mineral forms of soil nitrogen and phosphorus in the laboratory, and the increases caused by A. hilgendorfi were greater than those of L. rubellus. A. hilgendorfi increased mean soil aggregate size in the field while L. rubellus did not. Additionally, we determined the growth rate of A. hilgendorfi in the field and found that at 1.35 mg AFDM/day, the rate was greater than most published values for invasive European species. Treatment effects were stronger in laboratory mesocosms than field enclosures. No interactive effects between the two species were observed. These results suggest that the effects of A. hilgendorfi can be significant and, like those of European species, are undesirable from the perspective of meeting conservation goals.
KeywordsMegascolecidaeLumbricidaeBiological invasionSoil ecologyAsian jumping wormExotic species
Most forest ecosystems of northern North America were free of earthworms from the end of the Wisconsinian glaciation until European settlement of the continent (Reynolds 1994). At this time, exotic earthworms belonging to the family Lumbricidae were first introduced via the dumping of ballast soil and the transport of plants rooted in European soil (Hendrix and Bohlen 2002; Callaham et al. 2006), and subsequently spread by other human activities including timber harvest (Costello et al. 2011), their use as fishing bait (Keller et al. 2007) and through the vermicomposting industry (Callaham et al. 2006). Though earthworm-free areas still exist (Hendrix and Bohlen 2002; Holdsworth et al. 2007), established communities of European lumbricids have now been found in most US states and several Canadian provinces of North America (Reynolds and Wetzel 2008), including deciduous forest communities consisting of plant and animal species that evolved in the absence of earthworms (Hale et al. 2005a).
Understanding of North American earthworm invasions originates largely from studies of exotic European species belonging to the family Lumbricidae. As ecosystem engineers (sensu Jones et al. 1994) and detritivores, exotic European earthworms can fundamentally alter the ecosystems they invade (Hale et al. 2005a; Hendrix et al. 2006). In earthworm-free temperate deciduous forests, leaf litter and duff accumulates and constitutes a thick insulating upper organic layer (Hale et al. 2005b). This upper litter layer retains moisture (Hale 2007), functions as substrate for microbes and fungi (McLean and Parkinson 2000), provides habitat for arthropods (Migge-Kleian et al. 2006), amphibians (Maerz et al. 2009) and birds (Mattsson 2001), and promotes distinctive plant communities (Gundale 2002; Hale et al. 2005a, 2006). Additionally this habitat is where most forest–floor–nutrient cycling occurs (Bohlen et al. 2004a). Litter decomposition in earthworm-free temperate deciduous forests is primarily fungally mediated, and relatively slow due to the absence of large native detritivores (Suarez et al. 2006). However, European earthworms consume litter and other detritus, mix it into the mineral layer below (Lavelle 1998), and convert discrete duff and soil layers into a single compact layer of topsoil with very different ecological properties.
Consumption of litter and soil organic matter by exotic earthworms alters nutrient cycling within invaded forests (e.g., Alban and Berry 1994; Bohlen et al. 2004b; Groffman et al. 2004) with consequences for terrestrial and aquatic ecosystems (Costello and Lamberti 2008, 2009). Rates of nitrogen (N) mineralization, leaching, and denitrification, have been shown to increase after earthworm invasion (Haimi and Huhta 1990; Scheu and Parkinson 1994; Burtelow et al. 1998; Costello and Lamberti 2009). Similarly, earthworms can accelerate phosphorus (P) transformation, resulting in the relocation of P from the litter layer to mineral soil beneath (Suarez et al. 2004). These earthworm-mediated changes to soil–nutrient cycling can fundamentally alter the function of the forest floor (Groffman and Bohlen 1999).
While much is known of how European earthworms belonging to the family Lumbricidae influence soil ecosystems within temperate deciduous forests, very few studies have examined the potential effects of earthworms from other origins and families. Of particular concern are earthworms of Asian origin belonging to the family Megascolecidae and the genus Amynthas, which have invaded North America (Burtelow et al. 1998; Reynolds and Wetzel 2008; Snyder et al. 2011; Zhang et al. 2010), including northern temperate regions that lack native earthworm populations (Greiner et al. 2010). Unlike the most ecologically significant invasive European earthworms in temperate North America, Amynthas species can have an annual life cycle in which juveniles emerge from cocoons in the spring, rapidly accumulate biomass as they reach sexual maturity, then reproduce and die in the autumn of the same year (Reynolds 1978; Burtelow et al. 1998; Callaham et al. 2003; Greiner et al. 2010). Amynthas species reported in North America are generally epi–endogeic (sensu Bouché 1977), and compared to epigeic European species, Amynthas are large (commonly in excess of 10 cm in length; Greiner et al. 2010), can exhibit greater dietary flexibility (Zhang et al. 2010), and live at greater densities than common European invaders (Hale 2007). Because of these differences, the ecological effects of Amynthas have the potential to affect soil ecosystems differently than better-studied European species.
An established population of Amynthas hilgendorfi in a riparian forest in the Laurentian Great Lakes region provided the opportunity to assess the potential impacts of this species on a temperate–forest–soil ecosystem, and compare these impacts to the better-studied European invasive earthworm, Lumbricus rubellus, a species known to be a driver of ecological change (Hale et al. 2006). Field and laboratory experiments were targeted at understanding the effects of these two species on fundamental environmental variables including mineralized soil nutrients, organic-matter decomposition, soil organic matter, and soil structure. Interactive effects of A. hilgendorfi and L. rubellus were also evaluated. Finally, since rapid growth often correlates to greater invasion success of a species (Sakai et al. 2001), repeated sampling of the A. hilgendorfi population in the field was performed to calculate a growth rate during their growing season. The field experiment was designed to detect impacts of A. hilgendorfi in an environmentally realistic setting while the laboratory experiment allowed for earthworm effects to be isolated from other forest processes.
A field experiment was conducted from May to August 2009 on the Oakland University Biological Preserve near Rochester, MI, USA. The experiment was conducted within the riparian forest of Galloway Creek, the soils of which were largely silty loam (Sloan Series) (USDA NRCS 2009). The forest canopy was moderately open, and consisted mostly of Acer saccharum (Sugar maple), Acer rubrum (Red maple), Populus deltoides (Eastern cottonwood), and Tilia americana (American basswood). Phalaris arundinacea (Reed canary grass), was the dominant ground cover. An established earthworm community was present and consisted mostly of L. rubellus, L. terrestris, Apporectodea spp., and A. hilgendorfi.
Organic-matter decomposition was evaluated with a litter-bag approach (after Tiegs et al. 2009). Litterbags (1-cm mesh size) contained 3.75 ± 0.25 g of P. deltoides leaves (petioles removed), a common riparian species in the region. This genus has moderate decomposition rates (Tiegs et al. 2009) and was chosen so that litter material would be present at the end of the experiment, but enough decomposition would occur so that potential treatment effects could be detected. The leaves had senesced naturally and were gathered the previous autumn before being air dried and stored in the lab over winter. The leaves were moistened before being placed in the litterbags to render them soft and flexible. Four litterbags were introduced into each enclosure in May by placing them flat on the soil surface. They were removed in August on day 82 and the P. deltoides leaf material remaining within the litterbags was cleaned with water, air-dried, and weighed to the nearest 0.001 g. Decomposition was expressed as percent mass loss.
Mineral nutrient availability
At the end of the experiment four 5-g soil samples were collected from the upper 3.5 cm of each field enclosure using an 8-cm wide coring tool to determine concentrations of mineral N and P. To extract mineral N and P, each sample was mixed with 50 mL 1 M KCl, shaken vigorously for 1 min by hand and allowed to equilibrate for approximately 24 h (adapted from Robertson et al. 1999). The supernatant was then filtered through a Whatman GF/F (mean pore size 0.7 μm) and frozen until being analyzed for ammonium (NH4+), nitrate (NO3−), and exchangeable P using standard colorimetric procedures (APHA 1998) on a technicon auto analyzer II according to methods described in Davis and Simmons (1979). Total mineral N (Nmin) was calculated by summing the NH4+ and NO3− concentrations for each replicate.
Relative soil organic matter
At the end of the experiment, four soil samples (3 g each) were taken from the upper 3.5 cm of each enclosure using a coring tool to determine soil organic matter (SOM). Each sample was dried at 60 °C for 24 h and stored in a desiccator before being re-weighed to determine dry mass. Samples were then ashed at 500 °C for 4 h, and the ash was weighed to the nearest 0.001 g and subtracted from the dry mass to determine AFDM. Percent organic matter for each sample was calculated by dividing AFDM by dry mass. Since samples were not acidified prior to ashing, carbonate was not removed, and therefore measured SOM values are considered relative.
Soil aggregate size distribution
The size of soil aggregates was determined by taking four 175-cm3 soil core samples from the upper 3.5 cm of soil from each field enclosure on day 84, air drying in the lab and serial sieving for 1 min to separate aggregates into four size classes: >2, 1–<2, 0.5–<1, and <0.5 mm (adapted from Larney 2008). Each size class for each sample was weighed individually to the nearest 0.001 g and divided by the mass of the total sample to determine the proportion of each size class.
Above–ground vegetation biomass
At the close of the field experiment, biomass of above–ground herbaceous vegetation was determined by cutting vegetation just above the soil surface with shears, drying at 60 °C, and then weighing to the nearest 10 g.
Amynthas hilgendorfi growth rate
Amynthas hilgendorfi individuals were sampled from the population on the Oakland University Biological Preserve approximately once or twice per week from the time they were first found as hatchlings on 15 April 2009 to the time when widespread adult mortality was apparent on 12 October 2009. Using the ‘flip and strip’ method (Hale 2007; Costello et al. 2011), 10 individuals were gathered on each sampling date and preserved in 70 % ethanol. The length of each individual was measured to the nearest 0.5 mm. AFDM was determined using the allometric relationship developed by Greiner et al. (2010) for A. hilgendorfi. Assuming AFDM is 14 % of fresh weight (Greiner, unpublished data), AFDM values were later converted to fresh weight in order to compare them to published growth rates of other earthworm species. A growth rate for A. hilgendorfi was calculated by dividing the difference in biomass from the initial sampling date to the saturation phase (sensu Drake 2005) by the number of days lapsed. The resulting growth rate was then compared to published growth rates (expressed as fresh weight) for other invasive earthworm species.
Forty mesocosms were assembled by adding 500 g of homogenized, earthworm-free forest soil to 1–L plastic terraria (17.5 × 14.5 × 4 cm). The soil was obtained from the A-horizon of an oak-dominated temperate forest in Southeast Michigan, sieved through a 1-cm mesh screen to remove rocks and debris, and mixed thoroughly in the lab. Mesocosms were randomly assigned to one of four earthworm treatments: 10 A. hilgendorfi individuals, 10 L. rubellus individuals, five individuals of each species (the ‘combined treatment’), or no earthworms added. Densities were similar to those used in mesocosms in Hale et al. (2008), which were derived from field data. Based on subsamples, average individual earthworm biomass per mesocosm was 121 and 94 mg AFDM for A. hilgendorfi and L. rubellus, respectively. Mesocosms were placed in an 18 °C temperature chamber with a 12 h light–dark cycle for the 6-week duration of the experiment. The mesocosms received a weekly misting of water in order to maintain a water content of approximately 30 %.
For the laboratory experiment, 1.25 ± 0.25 g dry mass of T. americana leaves picked directly from trees was added to the soil surface of each mesocosm at the beginning of the experiment. This genus has relatively rapid decomposition rates in the presence of earthworms (Holdsworth et al. 2008) and was deemed appropriate for the lab experiment, which was of a shorter duration than the field experiment. At the conclusion of the experiment on day 42, the remaining detectable leaf material was picked from the mesocosms, cleaned, dried and weighed in the same manner as the leaves used in the field experiment.
Mineral nutrient availability
At the end of the experiment, four 5-g samples were taken from the top 2 cm of each mesocosm and analyzed for mineral N and P using the same methods as for the field experiment.
Relative soil organic matter (SOM)
A 3-g soil sample was taken from the top 2 cm of each mesocosm at the end of the 6-week-long incubation period and AFDM was determined for each sample with the same methods as the field experiment.
Randomized-block analysis of variance (ANOVA; model III two-factor analysis of variance without replication) was used to test for treatment effects in the field experiment. The mean of the four subsamples was used for analysis of data obtained from the field experiment for leaf-mass loss, soil nutrients, SOM, and soil-aggregate analyses. One-way ANOVA was used to test for treatment effects in the laboratory mesocosms. Tukey comparisons were used to identify differences between specific treatments for all response variables that yielded significant (P < 0.05) ANOVA results. Shapiro–Wilks tests were used to test for normality, and when needed, non-percent data were natural-log transformed to improve normality. Data expressed as percentages were square-root arcsine transformed. All statistics were performed using Systat (version 12.0).
ANOVA summary (as F and P values) for all response variables measured in the laboratory experiment
Soil exchangeable P
Soil nutrient mineralization
ANOVA summary (as F and P values) for all response variables measured in the manipulative field experiment. L. rubellus increased leaf-mass loss and A. hilgendorfi increased the proportion of large soil aggregates compared to reference enclosures
Source of variation
Soil exchangeable P
Soil aggregates (mm)
Relative soil organic matter
All earthworm treatments increased SOM in laboratory mesocosms relative to controls (Table 1). A. hilgendorfi, L. rubellus, and combined species mesocosms did not differ with regards to SOM, each containing approximately 5.5 % organic matter, while control mesocosms contained approximately 4.2 %. Differences in relative SOM were not observed among treatments within field enclosures (Table 2).
Soil-aggregate size partitioning
Above–ground herbaceous vegetation biomass
In the field enclosures, above–ground biomass of vegetation did not differ among treatments (Table 2). The amount of vegetation present was highly variable among plots and ranged from 75 to 440 g dry mass per enclosure; P. arundinacea comprised at least 90 % of the biomass for all plots.
Amynthas hilgendorfi growth rate
Using manipulative field and lab approaches, we compared the ecological effects of a well-studied European and largely unknown Asian invasive earthworm in a temperate forest setting. Though effects were generally not detected in the field, laboratory results indicated that both species impact leaf-litter decomposition and mineral nutrient availability, though to different degrees. These impacts, coupled with the increase in soil-aggregate size in the presence of A. hilgendorfi in field enclosures, suggest that the effect of this species can be significant, even relative to L. rubellus, a well-known driver of ecological change in earthworm-invaded forests. In general these results are consistent with the small body of research that has been conducted on Amynthas species in other North American forests (e.g., Burtelow et al. 1998; Snyder et al. 2011; Zhang et al. 2010) in that A. hilgendorfi altered soil nutrient cycling and structure.
Different results were observed in laboratory mesocosms than in field plots. In laboratory mesocosms, all earthworm treatments increased the concentration of soil Nmin, and the A. hilgendorfi and combined species treatments increased exchangeable P. Differences in mineralized N and P were not observed among treatments in the field experiment. Key differences between the field enclosures and the laboratory mesocosms that may have resulted in lack of treatment effects include numerous pathways for nutrient loss. Riparian soils, like those of the field study, are subject to episodic flooding, creating conditions that favor denitrification (Costello and Lamberti 2009) and nutrient leaching (Gregory et al. 1991), potentially masking earthworm effects. Plant uptake may have concealed treatment effects in the field. Lastly, heterogeneous distributions of soil nutrients within our experimental blocks could have masked what may have been subtle earthworm effects.
Mass loss of senesced P. deltoides litter in the field was greatest in L. rubellus treatments and did not differ between A. hilgendorfi and control plots. In laboratory mesocosms, however, A. hilgendorfi treatments did exhibit significant leaf-mass, though not to the extent of mesocosms containing L. rubellus. The feeding habits of invasive Amynthas in North America are largely unknown, though a study by Zhang et al. (2010) found that the congener A. agrestis, abundant in the southeastern United States, fed primarily on SOM, and on leaf litter to a lesser degree, and displayed superior dietary flexibility compared to L. rubellus. Our findings suggest that the effects of A. hilgendorfi on litter mass loss are less than those of L. rubellus, perhaps because of a preference by A. hilgendorfi for other types of organic matter. The significant effect of A. hilgendorfi on the mass loss of T. americana leaves in the laboratory experiment may have been an experimental artifact associated with lack of choice for other types of organic matter. Additionally, the discrepancy between the field and lab experiments may be explained by the use of different litter types; that is, perhaps A. hilgendorfi feeds more effectively on the fresh litter of T. americana than the senesced litter P. deltoides.
Since all earthworm treatments in the laboratory exhibited leaf-mass loss, the increase in SOM in all earthworm treatments is likely reflective of the incorporation of leaf-material into the soil through earthworm castings. This is consistent with the findings of Snyder et al. (2009), who found that A. corticis effectively incorporated C from litter into earthworm casts, protecting it from mineralization. SOM values did not differ among treatments in the field experiment, despite L. rubellus’ significant contribution to leaf-mass loss. The presence of a background earthworm population, including endogeic consumers of SOM, may have hindered detection of treatment effects in the field.
The size partitioning of soil aggregates differed between the two earthworm species in the field experiment. A. hilgendorfi increased the proportion of aggregates of the largest class size (>2 mm) while L. rubellus’ soil aggregate profile was similar to controls. We did not assess the water stability of the aggregates. However, we did observe that the uppermost soil horizon in A. hilgendorfi field enclosures consisted almost entirely of earthworm casts and were visibly more granular in comparison to enclosures without A. hilgendorfi, a finding consistent with the field observations of Snyder et al. (2009, 2011). This granular condition of the soil persisted throughout the summer, suggesting that A. hilgendorfi casts are stable over time frames of weeks–months.
Though previous studies found invasive earthworms to be associated with declines in forest understory vegetation, particularly when L. rubellus is present (e.g., Gundale 2002; Hale et al. 2006), we did not observe differences in above–ground biomass of herbaceous vegetation among treatments in field enclosures. Our field site contained established exotic earthworm populations, whereas other (e.g., Hale et al. 2006) investigations were conducted in earthworm-free sites. The plant communities at our site may have been resistant to our earthworm treatments given that earthworm populations were present prior to the start of the experiment. Furthermore, previous studies tended to focus on native plant species (Larson et al. 2010; Hale et al. 2006; Gundale 2002), and most of the plant biomass in our study consisted of an exotic European species (P. arundinacea) which may be less influenced by the presence of earthworms.
A characteristic shared by many successful invasive species is rapid growth (Sakai et al. 2001), and we found A. hilgendorfi growth rates were greater than those published for many common European invaders: 9.65 mg fresh weight (FW)/day. This is substantially greater than rates calculated for L. terrestris (1.43 mg FW/day, from Bohlen et al. 1999), Dendrodrilus rubidus (1.91 mg FW/day, from Arnold et al. 2008), and Eisenia fetida (6.81 mg FW/day, from Suthar 2009), and similar to the widespread and ecologically significant invader L. rubellus (9.42 mg FW/day, from Klok 2007). If these growth rates are standardized, by correcting for average adult weight (using values from Greiner et al. 2010), then L. rubellus and E. fetida growth rates become significantly faster than that of A. hilgendorfi, which is perhaps not surprising given both species’ ability to consume large quantities of organic matter (Hale et al. 2006; Suthar 2009). Nevertheless, in absolute terms, A. hilgendorfi accumulates biomass rapidly, and this coupled with the high densities in which Amynthas spp. can thrive (Burtelow et al. 1998; Greiner and Tiegs, personal observation), suggests that A. hilgendorfi has the potential to have ecosystem-level impacts. Though growth rates were taken from published values of studies performed under different environmental conditions, and are therefore not ideal comparisons, they do offer a context for evaluating A. hilgendorfi growth relative to other species, and indicate that A. hilgendorfi posess a trait that is common among successful invaders.
We observed no evidence that A. hilgendorfi and L. rubellus displayed interactive effects on the response variables measured in the field or laboratory experiment. In all instances, the combined-species treatment was similar to one or both of the single-species treatments. Other observations suggest that A. hilgendorfi may displace other earthworm species and become the only surface-dwelling species present (Nidia Arguedas, personal communication), or, in the instance of Zhang et al. (2010), A. agrestis hindered the feeding of L. rubellus. Nevertheless, what is clear is that both species can impact forest ecosystems, and the presence of either is probably undesirable in the context of resource management and meeting conservation goals.
Here we presented the first examination of A. hilgendorfi in a temperate deciduous forest that lacks native earthworms. These studies demonstrated that, like European earthworms, A. hilgendorfi can significantly alter soil processes, and the impacts of A. hilgendorfi rival those of ecologically important European invaders. As invasive Asian earthworms continue to be introduced throughout temperate ecosystems, the importance of understanding the impacts and invasion dynamics of these species will become increasingly important if conservation and management goals are to be met.
Dave Costello provided valuable input regarding experimental design, field methods, and earthworm ecology in general. We thank Mac A. Callaham, Jr. for assistance with A. hilgendorfi identification, and Cassandra Belcher, Keith Berven, Ashley Burtner, Tim Campbell, Josh Martin, Crystal Moon, Jeff Stephens, and Andrew Stonehouse for assistance with field work and sample processing. George Gamboa provided feedback on an earlier draft of this manuscript. Catherine Starnes, Janis Bills, and Rita Perris provided administrative support, and Sheryl Hugger allowed for use of laboratory equipment. This project was funded by the Oakland University Provost’s Graduate Student Research Award given to HGG and the Oakland University Summer Research Fellowship given to Cassandra R. Belcher.