- Email: [email protected]
Propagule pressure governs establishment of an invasive herb
Acta Oecologica 68 (2015) 18e23
Contents lists available at ScienceDirect
Acta Oecologica journal homepage: www.elsevier.com/locate/actoec
Original article
Propagule pressure governs establishment of an invasive herb Satu Ramula a, b, *, Miia Jauni b, Tapio van Ooik b a b
€gen 9, 10600 Ekena €s, Finland Aronia Coastal Zone Research Team, Åbo Akademi University and Novia University of Applied Sciences, Raseborgsva Section of Ecology, Department of Biology, University of Turku, 20014 Turku, Finland
a r t i c l e i n f o
a b s t r a c t
Article history: Received 29 January 2015 Received in revised form 14 July 2015 Accepted 14 July 2015 Available online xxx
The success of plant invasions may be limited by the availability of propagules and/or of suitable microsites, with microsite availability being affected by, for example, disturbance and interspecific competition. A mechanistic understanding of the contributions of propagule pressure and microsite limitation to plant invasions is therefore required to minimise future invasions. Here, we investigated the relative roles of propagule pressure, the availability of microsites, and their interaction on the establishment of an invasive herb, Lupinus polyphyllus, in two geographic regions representing different climate and growth conditions in Finland (a more productive southern region and a harsher central region). We carried out a field experiment in 14 L. polyphyllus populations, in which we manipulated both propagule pressure and disturbance. In a complementary greenhouse experiment, we manipulated propagule pressure and interspecific competition. Seedling establishment of L. polyphyllus was higher in the more productive southern region than in the harsher central region. The number of L. polyphyllus seedlings increased with increasing propagule pressure regardless of disturbance or interspecific competition. However, the number of L. polyphyllus seedlings per sown seed (relative establishment) tended to decrease with increasing propagule pressure, indicating that the positive effect of propagule pressure on early invasion is partially counteracted by density-dependent mortality at high seed densities. Our results highlight the dominant role of propagule pressure over disturbance and interspecific competition in the establishment of L. polyphyllus, suggesting that the early stage of invasion is limited by the availability of propagules rather than the availability of suitable microsites. © 2015 Elsevier Masson SAS. All rights reserved.
Keywords: Density dependence Germination Invasive species Invasion success Lupinus polyphyllus Non-native species Seed number
1. Introduction The success of plant invasions may be restricted by several factors, such as the availability of propagules or of suitable microsites. The former, also known as propagule pressure, has frequently been found to be positively associated with successful plant invasions (reviewed by Colautti et al., 2006). The availability of suitable microsites may, in turn, be determined by disturbance, cooccurring plant species, or geographic location (as a proxy for different growth conditions and productivity levels). For example, disturbance may increase both the amount of resources available per individual plant and the number of accessible microsites, thus increasing the probability of plant recruitment and, consequently, community invasibility (Davis et al., 2000). On the other hand, strong interspecific competition may reduce community
* Corresponding author. Section of Ecology, Department of Biology, University of Turku, 20014 Turku, Finland. E-mail address: satu.ramula@utu.fi (S. Ramula). http://dx.doi.org/10.1016/j.actao.2015.07.001 1146-609X/© 2015 Elsevier Masson SAS. All rights reserved.
invasibility (Crawley, 1986) because fewer microsites and resources are available per individual plant. Efforts to minimise future invasions, therefore, require a mechanistic understanding of the relative contributions of propagule pressure and microsite limitation to plant invasions. The significant role of disturbance in plant invasions was supported by a recent meta-analysis that revealed that non-native plant species are more abundant and diverse at disturbed sites than at undisturbed sites (Jauni et al., 2015). However, previous studies have also questioned the importance of disturbance in plant invasions. For example, a global, empirical study reported that disturbance is a poor predictor of plant invasions in general, although it tends to slightly increase the richness and cover of nonnative plant species (Moles et al., 2012). Moreover, Buckley et al. (2007) showed theoretically that disturbance has a negligible effect on plant invasions when the level of disturbance is equal between invader occupied and unoccupied sites because in such a situation invasibility depends on propagule pressure only. These somewhat contrasting findings about the importance of
S. Ramula et al. / Acta Oecologica 68 (2015) 18e23
disturbance in plant invasions may be because of interactive effects between disturbance and other factors that contribute to plant invasions. In other words, the importance of disturbance to plant invasions may vary depending on community properties, such as propagule pressure and productivity. Both theoretical and empirical studies indeed suggest that interactive effects between propagule pressure and disturbance (or other factors that affect microsite availability) should not be ignored (e.g., Huston, 2004; Buckley et al., 2007; Eschtruth and Battles, 2009). In general, one might expect microsite availability to play a more prominent role in determining community invasibility under conditions of high, rather than low, propagule pressure because of stronger intraspecific competition for resources (Lockwood et al., 2005). In order to make such generalisations, though, and to determine the relative roles of propagule pressure and the availability microsites in plant invasions, it is necessary to carry out experimental studies that explicitly manipulate these two factors under different conditions at multiple sites. So far, such large-scale studies conducted in the invaded range are rare. Here, we investigated the relative roles of propagule pressure, microsite limitation, and their interaction in the establishment of an invasive herb, Lupinus polyphyllus Lindl. We carried out an experiment in 14 L. polyphyllus populations, in which we manipulated propagule pressure and disturbance in two geographic regions of Finland, a more productive southern region and a harsher central region. In a complementary greenhouse experiment, we manipulated propagule pressure and interspecific competition. We hypothesised that propagule pressure and disturbance would increase the establishment of L. polyphyllus, while interspecific competition would decrease it. We also hypothesised that seedling establishment would increase with high propagule pressure and disturbance, and with high propagule pressure and less interspecific competition (i.e. we predicted an interaction between propagule pressure and disturbance, and between propagule pressure and interspecific competition). Moreover, we predicted that community susceptibility to invasion would be greater at disturbed sites in the southern region than at disturbed sites in the central region because of more favourable growth conditions (i.e. better resource availability). 2. Materials and methods 2.1. Study species L. polyphyllus (Fabaceae) is a perennial herb which originates from North America and stands 50e100 cm high. In its invaded range, L. polyphyllus often inhabits frequently disturbed habitats, such as road verges and wastelands (Fremstad, 2010), suggesting that disturbance is essential for its persistence. The species primarily reproduces from seed (Fremstad, 2010; S. Ramula, personal observation), although clonal reproduction is also possible at least in the invaded range (Rapp, 2009). An individual L. polyphyllus plant can produce hundreds of seeds which are dispersed ballistically up to a few metres from the parent plant. 2.2. Propagule pressure and disturbance We examined the role of propagule pressure in the establishment of L. polyphyllus in the presence and absence of disturbance in 14 populations. Seven of the populations were located in southwestern Finland (hereafter, “southern region”) and seven were located about 500 km to the north, in central Finland (hereafter, “central region”). Generally, the southern region can be considered more productive than the central region, as indicated by the longer growing season (Tuhkanen, 1980; Grytnes et al., 1999) and about
19
40% greater seed production (per plant) of L. polyphyllus in southern Finland (Ramula, 2014). Most study populations (11 out of 14) inhabited road verges or wastelands, while the rest inhabited forest understories close to roads. All study sites were characterised by light, sandy soils. In each L. polyphyllus population, we established 14 plots of 1 m2 which we further divided into four subplots of 0.5 0.5 m. The plots were at least 0.5 m apart in each study population. In half of the plots, we simulated mechanical disturbance by breaking the soil surface of the whole plot with a hoe (about 20e30 hits per plot), whereas the rest of the plots were left undisturbed. The purpose of this treatment was to simulate smallscale natural disturbances that are caused by, for example, trampling and grazing. The mechanical disturbance did not remove the existing vegetation but created small openings of bare soil (about a 29% increase in the amount of bare soil), thus increasing the number of litter-free microsites available. The disturbance treatment had also the potential to damage the aboveground and belowground parts of the existing plants, possibly reducing their competitive ability. To minimise differences in vegetation and underlying abiotic conditions, disturbed and undisturbed plots were paired, and each pair was located only a few metres from the nearest L. polyphyllus individual (i.e. the plots were in locations that were suitable for the study species). L. polyphyllus seeds were sown in the study plots in late July 2011. Within each plot, four sowing densities (7, 15, 30, and 60 seeds) were randomly allocated to the subplots. Seeds were collected, cleaned, and sown on the same day in the field, ensuring that all sown seeds at a given site were from the same L. polyphyllus population, and that only fully developed seeds were sown. Fully developed seeds of L. polyphyllus tend to have viability close to 100% with little variation among populations (mean ± SD ¼ 97.7% ± 2.1%, n ¼ 21 populations; unpublished data). We revisited the study plots at the end of May 2012 and recorded the number of L. polyphyllus seedlings per subplot (hereafter, “absolute establishment”). We also calculated relative establishment as the number of L. polyphyllus seedlings in May divided by the number of seeds sown per subplot. To control for potential background germination from the soil seed bank, we observed seedling establishment from three unsown control plots (0.5 0.5 m) per population in 2012, but no seedlings were detected in these plots during the experiment. Unfortunately, one population in the southern region was lost due to vandalism and was therefore excluded from the analyses (leaving 13 populations; 6 in the southern region and 7 in the central region). 2.3. Propagule pressure and interspecific competition To examine the role of propagule pressure in the establishment of L. polyphyllus in the presence of a competing species, we conducted a greenhouse experiment in which we manipulated propagule pressure and interspecific competition. We used two different competitors: the perennial grass Elymus repens (L.) Gould (Poaceae) and the perennial herb Trifolium pratense L. (Fabaceae), which both co-occur with L. polyphyllus in Finland. E. repens is a common native weed, while T. pratense is a common native species that belongs to the same family as L. polyphyllus. For both competitors, we used commercial seed material, while for L. polyphyllus, we collected seeds from the 14 study populations in July 2011. First, to assess the effect of propagule pressure in the absence of interspecific competition, we sowed the L. polyphyllus seeds into 8 8cm pots at five different densities (2, 4, 8, 16, and 32 seeds per pot) in February 2012. Then, to explore how interspecific competition affects seedling establishment under different degrees of propagule pressure, the seeds of L. polyphyllus and one competitor were sown in the following density combinations: 1) 2 L. polyphyllus and 32 competitor seeds, 2) 4 L. polyphyllus and 16 competitor seeds, 3) 8
20
S. Ramula et al. / Acta Oecologica 68 (2015) 18e23
L. polyphyllus and 8 competitor seeds, 4) 16 L. polyphyllus and 4 competitor seeds, and 5) 32 L. polyphyllus and 2 competitor seeds. Each sowing density was replicated five times with each competitor (either E. repens or T. pratense) for each of the 14 L. polyphyllus populations. Altogether, there were 75 pots per population and 1050 pots in total. After sowing, the pots were kept in a cold (þ4 C) environment for four weeks in order to break seed dormancy, after which the temperature was gradually raised to þ 20 C. We counted the number of L. polyphyllus, E. repens, and T. pratense seedlings (hereafter, “absolute establishment”) once per week for 12 weeks, from the beginning of March to the end of May 2012, at which point new seedlings rarely emerged. All three species germinated at slightly different times, with T. pratense germinating about five weeks after the seed sowing followed by L. polyphyllus and E. repens that both began to germinate about seven weeks after the seed sowing. We calculated the relative establishment of L. polyphyllus as the maximum number of L. polyphyllus seedlings observed during the experiment (often in May) divided by the number of seeds sown per pot. For all three species, most established seedlings also survived to the end of the experiment (average survival 82e97%), indicating the absence of density-dependent mortality due to selfthinning. Although we used commercial seed material for the two competitor species, the establishment of E. repens was extremely low (mean ± SD ¼ 0.7% ± 3.6%), whereas T. pratense established well (mean ± SD ¼ 89.0% ± 21.3%). Since non-germinated seeds in the soil may inhibit the germination and establishment of other seeds (Wardle et al., 1991), we nevertheless included the data on poorly established E. repens in the analyses. 2.4. Statistical analyses To examine the effect of propagule pressure on the absolute and relative establishment of L. polyphyllus in the presence and absence of disturbance in the field, we constructed two generalised linear mixed-effects models (function ‘glmer’ in the lme4 package in R 2.15.2; R Development Core Team, 2013) which used, respectively, the absolute and relative establishment of L. polyphyllus seedlings per subplot as the response variable. Region (southern or central), propagule pressure (number of L. polyphyllus seeds sown as a continuous variable), and soil disturbance (disturbed or undisturbed), together with all their two-way interactions and a threeway interaction, were included as fixed explanatory variables in the models. To account for the spatial relatedness of the plots, plotpair nested within population, which was further nested within region, was considered a random effect. We also included a quadratic term of propagule pressure in the model to examine whether the effect of propagule pressure on seedling establishment was non-linear. Similarly, we constructed generalised linear mixed-effects models to examine the effect of propagule pressure on the absolute and relative establishment of L. polyphyllus in the presence and absence of interspecific competition in the greenhouse. Origin of seeds (southern or central region), propagule pressure (as linear and quadratic effects), and competition type (interspecific competition with E. repens, interspecific competition with T. pratense, or intraspecific competition), together with all their two-way interactions and a three-way interaction, were included as fixed effects. The total number of seeds sown per pot (including both L. polyphyllus and the competitor) was used as a covariate to account for different initial seed densities in the pots, and population, nested within the origin of seeds, was included as a random effect. We used a likelihood ratio test (LR) fit with maximum likelihood to examine the significance of the fixed variables (Pinheiro and Bates, 2000). In other words, we simplified the models starting
by removing non-significant interaction terms. The random effects of population and plot-pair were not tested as they were not our primary interest. The negative binomial distribution was used for absolute seedling establishment (i.e. the number of seedlings) estimated from the field data, while the Poisson distribution was used for absolute establishment estimated from the greenhouse data, and the binomial distribution was always used for relative seedling establishment (because the data were binary). Overdispersion in the models (dispersion factor > 1.5) was corrected by including a random term of plot and pot in the analyses based on field and greenhouse data, respectively. The goodness of fit was confirmed by visual examination of the residuals for each model. 3. Results 3.1. Propagule pressure and disturbance In the field experiment, the absolute establishment of L. polyphyllus increased with increasing propagule pressure regardless of geographic region and disturbance (Table 1; Fig. 1a), with the quadratic term of propagule pressure being statistically significant (slope: 0.0006, SE ¼ 0.0001, Table 1). The relative establishment (the number of seedlings per sown seed) of L. polyphyllus was generally low (mean ± SD ¼ 0.12 ± 0.13), and decreased in a marginally significant manner with increasing propagule pressure (Table 1, Fig. 1b). The establishment of L. polyphyllus tended to be higher in the southern region than in the central region (mean ± SD ¼ 4.23 ± 4.51 vs. 2.25 ± 3.77 for absolute establishment and 0.16 ± 0.14 vs. 0.09 ± 0.13 for relative establishment), but did not differ between disturbed and undisturbed plots (Table 1). 3.2. Propagule pressure and interspecific competition The greenhouse experiment yielded results similar to those of the field experiment, in that the absolute establishment of L. polyphyllus increased non-linearly with increasing propagule pressure (Fig. 2) and decreased with the total number of seeds sown per pot (slope: 0.0065, SE ¼ 0.003, Table 2). Relative establishment was low also in the greenhouse (mean ± SD ¼ 0.13 ± 0.19), but, unlike in the field experiment, was not associated with propagule pressure (Table 2). For both absolute and relative establishment, there was a significant interaction between competition and region (Table 2). The seeds collected from
Table 1 Results from generalised linear mixed-effects models used to examine the effect of propagule pressure on the absolute and relative establishment of Lupinus polyphyllus seedlings in the presence and absence of disturbance in the field (n ¼ 13 populations). Plot-pair, nested within population and region, was included as a random effect in the models. The likelihood ratio-test (LR) was used to assess the significance of the fixed explanatory variables (P 0.058 in bold), df and ddf denote the degrees of freedom in the nominator and in the denominator, respectively. Explanatory variables
Absolute establishment (no. Seedlings) LRdf,
Propagule pressure Quadratic propagule pressure Disturbance (disturbed, undisturbed) Region (southern, central) Propagule pressure disturbance Propagule pressure region Disturbance region Prop. pressure disturbance region
ddf
37.661,10 12.821,10 0.221,10 3.691,10 0.001,13 0.101,13 0.011,13 0.011,14
Relative establishment (no. Seedlings per sown seed)
P
LRdf,
<0.001 <0.001 0.642 0.055 0.967 0.755 0.958 0.926
3.591,9 2.241,9 0.791,9 3.801,9 0.051,12 0.001,12 0.001,12 0.121,13
ddf
P 0.058 0.134 0.373 0.051 0.810 0.968 0.993 0.735
S. Ramula et al. / Acta Oecologica 68 (2015) 18e23
21
establishment only). 4. Discussion
Fig. 1. Propagule pressure (i.e. the number of seeds sown per subplot) in relation to the (A) absolute and (B) relative establishment of Lupinus polyphyllus seedlings (mean ± SE) based on the field experiment (n ¼ 13 populations). Absolute establishment refers to the number of seedlings, while relative establishment refers to the number seedlings per sown seed.
Fig. 2. Propagule pressure in terms of the number of seeds sown per pot in relation to the absolute establishment (the number of seedlings) of Lupinus polyphyllus (mean ± SE) based on the greenhouse experiment (n ¼ 14 populations).
the southern region had the lowest absolute and relative establishment when they were grown together with E. repens, while competition type did not affect the establishment of seeds collected from the central region (Fig. 3, results shown for relative
We observed that propagule pressure contributed to the invasion success of L. polyphyllus, with the number of established seedlings increasing non-linearly with increasing propagule pressure under both field and greenhouse conditions. Such a strong positive relationship between propagule pressure and invasion success is not surprising and has been reported in numerous studies (e.g., Lockwood et al., 2005; Von Holle and Simberloff, 2005; Colautti et al., 2006; Fensham et al., 2013). A more interesting finding in the present study is the minor roles of disturbance and interspecific competition in the establishment of L. polyphyllus as both these factors are expected to affect the availability of microsites either directly or through altered resource levels, and have been shown to contribute to community invasibility (e.g., Eschtruth and Battles, 2009; Kempel et al., 2013; Houseman et al., 2014). For example, a study of American hemlock forests revealed that the effects of canopy disturbance interacted with those of propagule pressure in increasing invasibility, with the cover of invasive plant species being highest at disturbed sites under high propagule pressure and lowest at undisturbed sites under low propagule pressure (Eschtruth and Battles, 2009). Here, disturbance did not contribute to the establishment of L. polyphyllus either on its own or through an interaction with propagule pressure, suggesting that the establishment of the study species may not be limited by microsite availability. The proportion of bare ground in the study plots prior to disturbance was on average 14% ± 16% (SD), confirming that there were indeed microsites available at the study sites even in the absence of simulated disturbance. As a consequence, a more dramatic disturbance event (e.g., clearing the whole litter layer) or more frequent disturbance events might have been required to alter community invasibility in the present study. Alternatively, the microsites that were available after disturbance (29% ± 4.2% (SD) bare ground in the disturbed plots) may have been unsuitable for the establishment of L. polyphyllus for various reasons. For example, soil moisture may have been considerably lower in the disturbed plots than in the vegetated control plots, making microsites unsuitable for germination. Moreover, as we did not observe possible changes in the cover of other plant species in the study plots, we cannot exclude the possibility that our disturbance treatment may have promoted the spread of the existing herbaceous plant species rather than may have increased the availability of microsites for L. polyphyllus. However, given multiple study sites used here, it is unlikely that existing herbaceous species would have always been able to colonise microsites more efficiently than L. polyphyllus. Consistent with our finding, the role of disturbance in plant community invasibility has been questioned previously, as disturbance per se explains little variation (<7%) in the cover or richness of non-native plant species (Moles et al., 2012). The present study together with previous studies (Moles et al., 2012; Fensham et al., 2013) suggest that disturbance may not be the primary mechanisms that determines community invasibility. Similar to disturbance, competition type (interspecific vs. intraspecific seedling competition) did not interact with propagule pressure in the present study, although it did interact with the origin of seeds (southern or central region). The greenhouse experiment revealed that the establishment of L. polyphyllus seeds collected from the southern region was reduced when the species was grown together with E. repens (but not when it was grown alone or together with T. pratense), whereas interspecific competition had no effect on the establishment of seeds collected from the central region. These region-dependent outcomes for interspecific
22
S. Ramula et al. / Acta Oecologica 68 (2015) 18e23
Table 2 Results from generalised linear mixed-effects models used to examine the effect of propagule pressure on the absolute and relative establishment of Lupinus polyphyllus seedlings under different types of competition (intraspecific vs. interspecific with two different competitors, Elymus repens and Trifolium pratense) in the greenhouse (n ¼ 14 populations). Population, nested within the origin of seeds, was included as a random effect in the models. The likelihood ratio-test (LR) was used to assess the significance of the fixed explanatory variables (P < 0.05 in bold); all main factors were not tested because of significant interactions, df and ddf denote the degrees of freedom in the nominator and in the denominator, respectively. Explanatory variables
Absolute establishment (no. Seedlings) LRdf,
Propagule pressure Quadratic propagule pressure Competition (Elymus, Trifolium, intraspecific) Origin of seeds (southern, central) Total seeds per pot Propagule pressure competition Propagule pressure origin of seeds Competition origin of seeds Prop. pressure competition origin of seeds
ddf
171.711,12 39.871,12 Not tested Not tested 6.801,12 0.072,15 0.951,15 7.712,15 0.052,17
Fig. 3. The effect of competition on the relative establishment (the number of seedlings per sown seed) of Lupinus polyphyllus (mean ± SE) under greenhouse conditions (n ¼ 14 populations), with seeds originating from two geographic regions (southern or central). Elymus refers to interspecific competition with E. repens, Trifolium refers to interspecific competition with T. pratense, and Lupinus refers to intraspecific competition. Significant differences between different types of competition in the southern region are indicated by different letters (P < 0.05, Tukey's test).
competition might be because of different germination times of L. polyphyllus seeds originating from southern and central regions. Seeds from the southern region germinated on average a week later in the greenhouse than seeds collected from the central region (results not shown), which might have affected the strength of
Relative establishment (no. Seedlings per sown seed)
P
LRdf,
<0.001 <0.001
1.661,12 1.881,12 Not tested Not tested 1.741,12 1.602,15 0.811,15 8.412,15 0.132,17
0.009 0.963 0.331 0.021 0.978
ddf
P 0.197 0.170
0.187 0.450 0.369 0.015 0.946
asymmetric competition during the experiment. In other words, the later germination may have coincided with the germination of E. repens more, leading to the lower establishment of L. polyphyllus. The overall negligible effect of competition on the establishment of L. polyphyllus in the present study could be partly due to the favourable conditions in the greenhouse, which enabled seedlings to establish even under conditions of strong interspecific competition. On the other hand, the seedling establishment and population growth rates of L. polyphyllus are robust to habitat type and soil ~ ber and Ramula, 2013; Ramula, 2014) and therefore, it is pH (So possible that interspecific competition has a minor effect on the invasion success of this species also in the field. However, it should be noted that our greenhouse experiment focused on seedling competition, which is restricted to the situation where plants (regardless of their invasiveness status) are colonising open microsites from seed and does not necessarily apply to other types of interspecific competition. For example, invasive species that establish from seed often compete for resources with existing (adult) plants that already have an initial size advantage. In such situations, the outcome of competition often depends on the strength of asymmetric competition, with strong interspecific competition having more severe consequences for the invader than less intense competition (Weiner et al., 2001). Despite the minor roles of disturbance and interspecific competition in the establishment of L. polyphyllus, we did find that seedling establishment tended to be higher in the more productive southern region than in the less productive central region (Table 1). In other words, plant communities in southern Finland seemed to be more susceptible to L. polyphyllus invasions than those in the central parts of the country. The higher seedling establishment of the southern region might reflect the climate preference of the study species. This explanation is also supported by the fact that both the seed production of individual L. polyphyllus plants as well as the long-term population growth rates of L. polyphyllus populations are about 40e46% higher in the southern region than in the central region (Ramula, 2014). Although invasion success often increases with increasing propagule pressure, density dependence resulting from intraspecific competition during the life-cycle of colonising individuals may suppress the overall role of propagule pressure in invasion success (Poulsen et al., 2007). For example, high propagule pressure resulted in a reduction in both germination and seedling survival (when measured at the relative scale) for the invasive grass Microstegium vimineum (Warren et al., 2012). In the present study, high propagule pressure tended to reduce the relative seedling establishment of L. polyphyllus in the field, indicating density-
S. Ramula et al. / Acta Oecologica 68 (2015) 18e23
dependent mortality. For L. polyphyllus, density dependence also plays a role later in the life-cycle when it reduces the survival and fecundity of individual plants under field conditions (Ramula, 2014). This density dependence in later life stages, together with density-dependent establishment, may partially cancel out the absolute increase in invasion success that results from increased propagule pressure in L. polyphyllus. However, longer-term observations covering the whole life-cycle of the study species are required to quantify the overall benefit (if any) of high propagule pressure for invasion success. In summary, the present study, conducted in multiple plant populations across different environments, revealed that propagule pressure plays a dominant role in increasing the absolute establishment, and thus invasion success, of the perennial herb L. polyphyllus. However, under field conditions, the relative establishment of L. polyphyllus tended to decrease with increasing propagule pressure, probably because of density-dependent processes. In contrast to expectations, both disturbance and interspecific competition had only a negligible effect on the seedling establishment of L. polyphyllus on their own or through interactions with propagule pressure. Overall, these findings suggest that L. polyphyllus invasions are limited by the availability of propagules rather than the availability of suitable microsites. Acknowledgements ~ ber for establishing the sowing plots, three We thank Virve So anonymous reviewers for their thoughtful comments, and the Emil Aaltonen Foundation for funding. References Buckley, Y.M., Bolker, B.M., Rees, M., 2007. Disturbance, invasion and re-invasion: managing the weed-shaped hole in disturbed ecosystems. Ecol. Lett. 10, 809e817. Colautti, R.I., Grigorovich, I.A., MacIsaac, H.J., 2006. Propagule pressure: a null model for biological invasions. Biol. Invas 8, 1023e1037. Crawley, M.J., 1986. The population biology of invaders. Phil. Trans. R. Soc. B 314, 711e731. Davis, M.A., Grime, J.P., Thompson, K., 2000. Fluctuating resources in plant communities: a general theory of invasibility. J. Ecol. 88, 528e534. Eschtruth, A., Battles, J., 2009. Assessing the relative importance of disturbance, herbivory, diversity, and propagule pressure in exotic plant invasion. Ecol. Monogr. 79, 265e280.
23
Fensham, R.J., Donald, S., Dwyer, J.M., 2013. Propagule pressure, not fire or cattle grazing, promotes invasion of buffel grass Cenchrus ciliaris. J.Appl. Ecol. 50, 138e146. Fremstad, E., 2010. NOBANISeInvasive Alien Species Fact SheeteLupinus Polyphyllus. eFrom: Online Database of the European Network on Invasive Alien SpecieseNOBANIS. www.nobanis.org. Grytnes, J.A., Birks, H.J.B., Peglar, S.M., 1999. Plant species richness in Fennoscandia: evaluating the relative importance of climate and history. Nord. J. Bot. 19, 489e503. Houseman, G.R., Foster, B.L., Brassil, C.E., 2014. Propagule pressure-invasibility relationship: testing the influence of soil fertility and disturbance with Lespedeza cuneata. Oecologia 174, 511e520. Huston, M.A., 2004. Management strategies for plant invasions: manipulating productivity, disturbance and competition. Divers. Distrib. 10, 167e178. Jauni, M., Gripenberg, S., Ramula, S., 2015. Non-native plant species benefit from disturbance: a meta-analysis. Oikos 124, 122e129. Kempel, A., Chrobock, T., Fischer, M., Rohr, R.P., van Kleunen, M., 2013. Determinants of plant establishment success in a multispecies introduction experiment with native and alien species. Proc. Nat. Acad. Sci. USA. 110, 12727e12732. Lockwood, J.L., Cassey, P., Blackburn, T., 2005. The role of propagule pressure in explaining species invasions. Trends Ecol. Evol. 5, 223e228. Moles, A.T., Flores-Moreno, H., Bonser, S.P., Warton, D.I., Helm, A., Warman, L., Eldridge, D.J., Jurado, E., Hemmings, F.A., Reich, P.B., et al., 2012. Invasions: the trail behind, the path ahead, and a test of a disturbing idea. J. Ecol. 100, 116e127. Pinheiro, J.C., Bates, D.M., 2000. Mixed-effect Models in S and S-plus. Springer, New York. Poulsen, J.R., Osenberg, C.W., Clark, C.J., Levey, D.J., Bolker, B.M., 2007. Plants as reef fish: fitting the functional form of seedling recruitment. Am. Nat. 170, 167e183. R Development Core Team, 2013. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. Ramula, S., 2014. Linking vital rates to invasiveness of a perennial herb. Oecologia 174, 1255e1264. Rapp, W., 2009. Invasive Plant Management in Glacier Bay National Park and Preserve Summer 2009. Field Season Report. Invasive Species Program, Glacier Bay National Park and Preserve, National Park Service, U.S. Department of the Interior, Gustavus, AK, p. 164. ~ ber, V., Ramula, S., 2013. Seed number and environmental conditions do not So explain seed size variability for the invasive herb Lupinus polyphyllus. Plant Ecol. 214, 883e892. Tuhkanen, S., 1980. Climatic parameters and indices in plant geography. Acta Phytogeogr. Suec. 67, 1e110. Von Holle, B., Simberloff, D., 2005. Ecological resistance to biological invasion overwhelmed by propagule pressure. Ecology 86, 3212e3218. Wardle, D.A., Ahmed, M., Nicholson, K.S., 1991. Allelopathic influence of nodding thistle (Carduus nutans L.) seeds on germination and radicle growth of pasture plants. New Zeal. J. Agr. Res. 34, 185e191. Warren, R.J., Bahn, V., Bradford, M.A., 2012. The interaction between propagule pressure, habitat suitability and density-dependent reproduction in species invasion. Oikos 121, 874e881. Weiner, J., Griepentrog, H.-W., Kristensen, L., 2001. Supression of weeds by spring wheat Triticum aestivum increases with crop density and spatial uniformity. J. Appl. Ecol. 38, 784e790.
Contents lists available at ScienceDirect
Acta Oecologica journal homepage: www.elsevier.com/locate/actoec
Original article
Propagule pressure governs establishment of an invasive herb Satu Ramula a, b, *, Miia Jauni b, Tapio van Ooik b a b
€gen 9, 10600 Ekena €s, Finland Aronia Coastal Zone Research Team, Åbo Akademi University and Novia University of Applied Sciences, Raseborgsva Section of Ecology, Department of Biology, University of Turku, 20014 Turku, Finland
a r t i c l e i n f o
a b s t r a c t
Article history: Received 29 January 2015 Received in revised form 14 July 2015 Accepted 14 July 2015 Available online xxx
The success of plant invasions may be limited by the availability of propagules and/or of suitable microsites, with microsite availability being affected by, for example, disturbance and interspecific competition. A mechanistic understanding of the contributions of propagule pressure and microsite limitation to plant invasions is therefore required to minimise future invasions. Here, we investigated the relative roles of propagule pressure, the availability of microsites, and their interaction on the establishment of an invasive herb, Lupinus polyphyllus, in two geographic regions representing different climate and growth conditions in Finland (a more productive southern region and a harsher central region). We carried out a field experiment in 14 L. polyphyllus populations, in which we manipulated both propagule pressure and disturbance. In a complementary greenhouse experiment, we manipulated propagule pressure and interspecific competition. Seedling establishment of L. polyphyllus was higher in the more productive southern region than in the harsher central region. The number of L. polyphyllus seedlings increased with increasing propagule pressure regardless of disturbance or interspecific competition. However, the number of L. polyphyllus seedlings per sown seed (relative establishment) tended to decrease with increasing propagule pressure, indicating that the positive effect of propagule pressure on early invasion is partially counteracted by density-dependent mortality at high seed densities. Our results highlight the dominant role of propagule pressure over disturbance and interspecific competition in the establishment of L. polyphyllus, suggesting that the early stage of invasion is limited by the availability of propagules rather than the availability of suitable microsites. © 2015 Elsevier Masson SAS. All rights reserved.
Keywords: Density dependence Germination Invasive species Invasion success Lupinus polyphyllus Non-native species Seed number
1. Introduction The success of plant invasions may be restricted by several factors, such as the availability of propagules or of suitable microsites. The former, also known as propagule pressure, has frequently been found to be positively associated with successful plant invasions (reviewed by Colautti et al., 2006). The availability of suitable microsites may, in turn, be determined by disturbance, cooccurring plant species, or geographic location (as a proxy for different growth conditions and productivity levels). For example, disturbance may increase both the amount of resources available per individual plant and the number of accessible microsites, thus increasing the probability of plant recruitment and, consequently, community invasibility (Davis et al., 2000). On the other hand, strong interspecific competition may reduce community
* Corresponding author. Section of Ecology, Department of Biology, University of Turku, 20014 Turku, Finland. E-mail address: satu.ramula@utu.fi (S. Ramula). http://dx.doi.org/10.1016/j.actao.2015.07.001 1146-609X/© 2015 Elsevier Masson SAS. All rights reserved.
invasibility (Crawley, 1986) because fewer microsites and resources are available per individual plant. Efforts to minimise future invasions, therefore, require a mechanistic understanding of the relative contributions of propagule pressure and microsite limitation to plant invasions. The significant role of disturbance in plant invasions was supported by a recent meta-analysis that revealed that non-native plant species are more abundant and diverse at disturbed sites than at undisturbed sites (Jauni et al., 2015). However, previous studies have also questioned the importance of disturbance in plant invasions. For example, a global, empirical study reported that disturbance is a poor predictor of plant invasions in general, although it tends to slightly increase the richness and cover of nonnative plant species (Moles et al., 2012). Moreover, Buckley et al. (2007) showed theoretically that disturbance has a negligible effect on plant invasions when the level of disturbance is equal between invader occupied and unoccupied sites because in such a situation invasibility depends on propagule pressure only. These somewhat contrasting findings about the importance of
S. Ramula et al. / Acta Oecologica 68 (2015) 18e23
disturbance in plant invasions may be because of interactive effects between disturbance and other factors that contribute to plant invasions. In other words, the importance of disturbance to plant invasions may vary depending on community properties, such as propagule pressure and productivity. Both theoretical and empirical studies indeed suggest that interactive effects between propagule pressure and disturbance (or other factors that affect microsite availability) should not be ignored (e.g., Huston, 2004; Buckley et al., 2007; Eschtruth and Battles, 2009). In general, one might expect microsite availability to play a more prominent role in determining community invasibility under conditions of high, rather than low, propagule pressure because of stronger intraspecific competition for resources (Lockwood et al., 2005). In order to make such generalisations, though, and to determine the relative roles of propagule pressure and the availability microsites in plant invasions, it is necessary to carry out experimental studies that explicitly manipulate these two factors under different conditions at multiple sites. So far, such large-scale studies conducted in the invaded range are rare. Here, we investigated the relative roles of propagule pressure, microsite limitation, and their interaction in the establishment of an invasive herb, Lupinus polyphyllus Lindl. We carried out an experiment in 14 L. polyphyllus populations, in which we manipulated propagule pressure and disturbance in two geographic regions of Finland, a more productive southern region and a harsher central region. In a complementary greenhouse experiment, we manipulated propagule pressure and interspecific competition. We hypothesised that propagule pressure and disturbance would increase the establishment of L. polyphyllus, while interspecific competition would decrease it. We also hypothesised that seedling establishment would increase with high propagule pressure and disturbance, and with high propagule pressure and less interspecific competition (i.e. we predicted an interaction between propagule pressure and disturbance, and between propagule pressure and interspecific competition). Moreover, we predicted that community susceptibility to invasion would be greater at disturbed sites in the southern region than at disturbed sites in the central region because of more favourable growth conditions (i.e. better resource availability). 2. Materials and methods 2.1. Study species L. polyphyllus (Fabaceae) is a perennial herb which originates from North America and stands 50e100 cm high. In its invaded range, L. polyphyllus often inhabits frequently disturbed habitats, such as road verges and wastelands (Fremstad, 2010), suggesting that disturbance is essential for its persistence. The species primarily reproduces from seed (Fremstad, 2010; S. Ramula, personal observation), although clonal reproduction is also possible at least in the invaded range (Rapp, 2009). An individual L. polyphyllus plant can produce hundreds of seeds which are dispersed ballistically up to a few metres from the parent plant. 2.2. Propagule pressure and disturbance We examined the role of propagule pressure in the establishment of L. polyphyllus in the presence and absence of disturbance in 14 populations. Seven of the populations were located in southwestern Finland (hereafter, “southern region”) and seven were located about 500 km to the north, in central Finland (hereafter, “central region”). Generally, the southern region can be considered more productive than the central region, as indicated by the longer growing season (Tuhkanen, 1980; Grytnes et al., 1999) and about
19
40% greater seed production (per plant) of L. polyphyllus in southern Finland (Ramula, 2014). Most study populations (11 out of 14) inhabited road verges or wastelands, while the rest inhabited forest understories close to roads. All study sites were characterised by light, sandy soils. In each L. polyphyllus population, we established 14 plots of 1 m2 which we further divided into four subplots of 0.5 0.5 m. The plots were at least 0.5 m apart in each study population. In half of the plots, we simulated mechanical disturbance by breaking the soil surface of the whole plot with a hoe (about 20e30 hits per plot), whereas the rest of the plots were left undisturbed. The purpose of this treatment was to simulate smallscale natural disturbances that are caused by, for example, trampling and grazing. The mechanical disturbance did not remove the existing vegetation but created small openings of bare soil (about a 29% increase in the amount of bare soil), thus increasing the number of litter-free microsites available. The disturbance treatment had also the potential to damage the aboveground and belowground parts of the existing plants, possibly reducing their competitive ability. To minimise differences in vegetation and underlying abiotic conditions, disturbed and undisturbed plots were paired, and each pair was located only a few metres from the nearest L. polyphyllus individual (i.e. the plots were in locations that were suitable for the study species). L. polyphyllus seeds were sown in the study plots in late July 2011. Within each plot, four sowing densities (7, 15, 30, and 60 seeds) were randomly allocated to the subplots. Seeds were collected, cleaned, and sown on the same day in the field, ensuring that all sown seeds at a given site were from the same L. polyphyllus population, and that only fully developed seeds were sown. Fully developed seeds of L. polyphyllus tend to have viability close to 100% with little variation among populations (mean ± SD ¼ 97.7% ± 2.1%, n ¼ 21 populations; unpublished data). We revisited the study plots at the end of May 2012 and recorded the number of L. polyphyllus seedlings per subplot (hereafter, “absolute establishment”). We also calculated relative establishment as the number of L. polyphyllus seedlings in May divided by the number of seeds sown per subplot. To control for potential background germination from the soil seed bank, we observed seedling establishment from three unsown control plots (0.5 0.5 m) per population in 2012, but no seedlings were detected in these plots during the experiment. Unfortunately, one population in the southern region was lost due to vandalism and was therefore excluded from the analyses (leaving 13 populations; 6 in the southern region and 7 in the central region). 2.3. Propagule pressure and interspecific competition To examine the role of propagule pressure in the establishment of L. polyphyllus in the presence of a competing species, we conducted a greenhouse experiment in which we manipulated propagule pressure and interspecific competition. We used two different competitors: the perennial grass Elymus repens (L.) Gould (Poaceae) and the perennial herb Trifolium pratense L. (Fabaceae), which both co-occur with L. polyphyllus in Finland. E. repens is a common native weed, while T. pratense is a common native species that belongs to the same family as L. polyphyllus. For both competitors, we used commercial seed material, while for L. polyphyllus, we collected seeds from the 14 study populations in July 2011. First, to assess the effect of propagule pressure in the absence of interspecific competition, we sowed the L. polyphyllus seeds into 8 8cm pots at five different densities (2, 4, 8, 16, and 32 seeds per pot) in February 2012. Then, to explore how interspecific competition affects seedling establishment under different degrees of propagule pressure, the seeds of L. polyphyllus and one competitor were sown in the following density combinations: 1) 2 L. polyphyllus and 32 competitor seeds, 2) 4 L. polyphyllus and 16 competitor seeds, 3) 8
20
S. Ramula et al. / Acta Oecologica 68 (2015) 18e23
L. polyphyllus and 8 competitor seeds, 4) 16 L. polyphyllus and 4 competitor seeds, and 5) 32 L. polyphyllus and 2 competitor seeds. Each sowing density was replicated five times with each competitor (either E. repens or T. pratense) for each of the 14 L. polyphyllus populations. Altogether, there were 75 pots per population and 1050 pots in total. After sowing, the pots were kept in a cold (þ4 C) environment for four weeks in order to break seed dormancy, after which the temperature was gradually raised to þ 20 C. We counted the number of L. polyphyllus, E. repens, and T. pratense seedlings (hereafter, “absolute establishment”) once per week for 12 weeks, from the beginning of March to the end of May 2012, at which point new seedlings rarely emerged. All three species germinated at slightly different times, with T. pratense germinating about five weeks after the seed sowing followed by L. polyphyllus and E. repens that both began to germinate about seven weeks after the seed sowing. We calculated the relative establishment of L. polyphyllus as the maximum number of L. polyphyllus seedlings observed during the experiment (often in May) divided by the number of seeds sown per pot. For all three species, most established seedlings also survived to the end of the experiment (average survival 82e97%), indicating the absence of density-dependent mortality due to selfthinning. Although we used commercial seed material for the two competitor species, the establishment of E. repens was extremely low (mean ± SD ¼ 0.7% ± 3.6%), whereas T. pratense established well (mean ± SD ¼ 89.0% ± 21.3%). Since non-germinated seeds in the soil may inhibit the germination and establishment of other seeds (Wardle et al., 1991), we nevertheless included the data on poorly established E. repens in the analyses. 2.4. Statistical analyses To examine the effect of propagule pressure on the absolute and relative establishment of L. polyphyllus in the presence and absence of disturbance in the field, we constructed two generalised linear mixed-effects models (function ‘glmer’ in the lme4 package in R 2.15.2; R Development Core Team, 2013) which used, respectively, the absolute and relative establishment of L. polyphyllus seedlings per subplot as the response variable. Region (southern or central), propagule pressure (number of L. polyphyllus seeds sown as a continuous variable), and soil disturbance (disturbed or undisturbed), together with all their two-way interactions and a threeway interaction, were included as fixed explanatory variables in the models. To account for the spatial relatedness of the plots, plotpair nested within population, which was further nested within region, was considered a random effect. We also included a quadratic term of propagule pressure in the model to examine whether the effect of propagule pressure on seedling establishment was non-linear. Similarly, we constructed generalised linear mixed-effects models to examine the effect of propagule pressure on the absolute and relative establishment of L. polyphyllus in the presence and absence of interspecific competition in the greenhouse. Origin of seeds (southern or central region), propagule pressure (as linear and quadratic effects), and competition type (interspecific competition with E. repens, interspecific competition with T. pratense, or intraspecific competition), together with all their two-way interactions and a three-way interaction, were included as fixed effects. The total number of seeds sown per pot (including both L. polyphyllus and the competitor) was used as a covariate to account for different initial seed densities in the pots, and population, nested within the origin of seeds, was included as a random effect. We used a likelihood ratio test (LR) fit with maximum likelihood to examine the significance of the fixed variables (Pinheiro and Bates, 2000). In other words, we simplified the models starting
by removing non-significant interaction terms. The random effects of population and plot-pair were not tested as they were not our primary interest. The negative binomial distribution was used for absolute seedling establishment (i.e. the number of seedlings) estimated from the field data, while the Poisson distribution was used for absolute establishment estimated from the greenhouse data, and the binomial distribution was always used for relative seedling establishment (because the data were binary). Overdispersion in the models (dispersion factor > 1.5) was corrected by including a random term of plot and pot in the analyses based on field and greenhouse data, respectively. The goodness of fit was confirmed by visual examination of the residuals for each model. 3. Results 3.1. Propagule pressure and disturbance In the field experiment, the absolute establishment of L. polyphyllus increased with increasing propagule pressure regardless of geographic region and disturbance (Table 1; Fig. 1a), with the quadratic term of propagule pressure being statistically significant (slope: 0.0006, SE ¼ 0.0001, Table 1). The relative establishment (the number of seedlings per sown seed) of L. polyphyllus was generally low (mean ± SD ¼ 0.12 ± 0.13), and decreased in a marginally significant manner with increasing propagule pressure (Table 1, Fig. 1b). The establishment of L. polyphyllus tended to be higher in the southern region than in the central region (mean ± SD ¼ 4.23 ± 4.51 vs. 2.25 ± 3.77 for absolute establishment and 0.16 ± 0.14 vs. 0.09 ± 0.13 for relative establishment), but did not differ between disturbed and undisturbed plots (Table 1). 3.2. Propagule pressure and interspecific competition The greenhouse experiment yielded results similar to those of the field experiment, in that the absolute establishment of L. polyphyllus increased non-linearly with increasing propagule pressure (Fig. 2) and decreased with the total number of seeds sown per pot (slope: 0.0065, SE ¼ 0.003, Table 2). Relative establishment was low also in the greenhouse (mean ± SD ¼ 0.13 ± 0.19), but, unlike in the field experiment, was not associated with propagule pressure (Table 2). For both absolute and relative establishment, there was a significant interaction between competition and region (Table 2). The seeds collected from
Table 1 Results from generalised linear mixed-effects models used to examine the effect of propagule pressure on the absolute and relative establishment of Lupinus polyphyllus seedlings in the presence and absence of disturbance in the field (n ¼ 13 populations). Plot-pair, nested within population and region, was included as a random effect in the models. The likelihood ratio-test (LR) was used to assess the significance of the fixed explanatory variables (P 0.058 in bold), df and ddf denote the degrees of freedom in the nominator and in the denominator, respectively. Explanatory variables
Absolute establishment (no. Seedlings) LRdf,
Propagule pressure Quadratic propagule pressure Disturbance (disturbed, undisturbed) Region (southern, central) Propagule pressure disturbance Propagule pressure region Disturbance region Prop. pressure disturbance region
ddf
37.661,10 12.821,10 0.221,10 3.691,10 0.001,13 0.101,13 0.011,13 0.011,14
Relative establishment (no. Seedlings per sown seed)
P
LRdf,
<0.001 <0.001 0.642 0.055 0.967 0.755 0.958 0.926
3.591,9 2.241,9 0.791,9 3.801,9 0.051,12 0.001,12 0.001,12 0.121,13
ddf
P 0.058 0.134 0.373 0.051 0.810 0.968 0.993 0.735
S. Ramula et al. / Acta Oecologica 68 (2015) 18e23
21
establishment only). 4. Discussion
Fig. 1. Propagule pressure (i.e. the number of seeds sown per subplot) in relation to the (A) absolute and (B) relative establishment of Lupinus polyphyllus seedlings (mean ± SE) based on the field experiment (n ¼ 13 populations). Absolute establishment refers to the number of seedlings, while relative establishment refers to the number seedlings per sown seed.
Fig. 2. Propagule pressure in terms of the number of seeds sown per pot in relation to the absolute establishment (the number of seedlings) of Lupinus polyphyllus (mean ± SE) based on the greenhouse experiment (n ¼ 14 populations).
the southern region had the lowest absolute and relative establishment when they were grown together with E. repens, while competition type did not affect the establishment of seeds collected from the central region (Fig. 3, results shown for relative
We observed that propagule pressure contributed to the invasion success of L. polyphyllus, with the number of established seedlings increasing non-linearly with increasing propagule pressure under both field and greenhouse conditions. Such a strong positive relationship between propagule pressure and invasion success is not surprising and has been reported in numerous studies (e.g., Lockwood et al., 2005; Von Holle and Simberloff, 2005; Colautti et al., 2006; Fensham et al., 2013). A more interesting finding in the present study is the minor roles of disturbance and interspecific competition in the establishment of L. polyphyllus as both these factors are expected to affect the availability of microsites either directly or through altered resource levels, and have been shown to contribute to community invasibility (e.g., Eschtruth and Battles, 2009; Kempel et al., 2013; Houseman et al., 2014). For example, a study of American hemlock forests revealed that the effects of canopy disturbance interacted with those of propagule pressure in increasing invasibility, with the cover of invasive plant species being highest at disturbed sites under high propagule pressure and lowest at undisturbed sites under low propagule pressure (Eschtruth and Battles, 2009). Here, disturbance did not contribute to the establishment of L. polyphyllus either on its own or through an interaction with propagule pressure, suggesting that the establishment of the study species may not be limited by microsite availability. The proportion of bare ground in the study plots prior to disturbance was on average 14% ± 16% (SD), confirming that there were indeed microsites available at the study sites even in the absence of simulated disturbance. As a consequence, a more dramatic disturbance event (e.g., clearing the whole litter layer) or more frequent disturbance events might have been required to alter community invasibility in the present study. Alternatively, the microsites that were available after disturbance (29% ± 4.2% (SD) bare ground in the disturbed plots) may have been unsuitable for the establishment of L. polyphyllus for various reasons. For example, soil moisture may have been considerably lower in the disturbed plots than in the vegetated control plots, making microsites unsuitable for germination. Moreover, as we did not observe possible changes in the cover of other plant species in the study plots, we cannot exclude the possibility that our disturbance treatment may have promoted the spread of the existing herbaceous plant species rather than may have increased the availability of microsites for L. polyphyllus. However, given multiple study sites used here, it is unlikely that existing herbaceous species would have always been able to colonise microsites more efficiently than L. polyphyllus. Consistent with our finding, the role of disturbance in plant community invasibility has been questioned previously, as disturbance per se explains little variation (<7%) in the cover or richness of non-native plant species (Moles et al., 2012). The present study together with previous studies (Moles et al., 2012; Fensham et al., 2013) suggest that disturbance may not be the primary mechanisms that determines community invasibility. Similar to disturbance, competition type (interspecific vs. intraspecific seedling competition) did not interact with propagule pressure in the present study, although it did interact with the origin of seeds (southern or central region). The greenhouse experiment revealed that the establishment of L. polyphyllus seeds collected from the southern region was reduced when the species was grown together with E. repens (but not when it was grown alone or together with T. pratense), whereas interspecific competition had no effect on the establishment of seeds collected from the central region. These region-dependent outcomes for interspecific
22
S. Ramula et al. / Acta Oecologica 68 (2015) 18e23
Table 2 Results from generalised linear mixed-effects models used to examine the effect of propagule pressure on the absolute and relative establishment of Lupinus polyphyllus seedlings under different types of competition (intraspecific vs. interspecific with two different competitors, Elymus repens and Trifolium pratense) in the greenhouse (n ¼ 14 populations). Population, nested within the origin of seeds, was included as a random effect in the models. The likelihood ratio-test (LR) was used to assess the significance of the fixed explanatory variables (P < 0.05 in bold); all main factors were not tested because of significant interactions, df and ddf denote the degrees of freedom in the nominator and in the denominator, respectively. Explanatory variables
Absolute establishment (no. Seedlings) LRdf,
Propagule pressure Quadratic propagule pressure Competition (Elymus, Trifolium, intraspecific) Origin of seeds (southern, central) Total seeds per pot Propagule pressure competition Propagule pressure origin of seeds Competition origin of seeds Prop. pressure competition origin of seeds
ddf
171.711,12 39.871,12 Not tested Not tested 6.801,12 0.072,15 0.951,15 7.712,15 0.052,17
Fig. 3. The effect of competition on the relative establishment (the number of seedlings per sown seed) of Lupinus polyphyllus (mean ± SE) under greenhouse conditions (n ¼ 14 populations), with seeds originating from two geographic regions (southern or central). Elymus refers to interspecific competition with E. repens, Trifolium refers to interspecific competition with T. pratense, and Lupinus refers to intraspecific competition. Significant differences between different types of competition in the southern region are indicated by different letters (P < 0.05, Tukey's test).
competition might be because of different germination times of L. polyphyllus seeds originating from southern and central regions. Seeds from the southern region germinated on average a week later in the greenhouse than seeds collected from the central region (results not shown), which might have affected the strength of
Relative establishment (no. Seedlings per sown seed)
P
LRdf,
<0.001 <0.001
1.661,12 1.881,12 Not tested Not tested 1.741,12 1.602,15 0.811,15 8.412,15 0.132,17
0.009 0.963 0.331 0.021 0.978
ddf
P 0.197 0.170
0.187 0.450 0.369 0.015 0.946
asymmetric competition during the experiment. In other words, the later germination may have coincided with the germination of E. repens more, leading to the lower establishment of L. polyphyllus. The overall negligible effect of competition on the establishment of L. polyphyllus in the present study could be partly due to the favourable conditions in the greenhouse, which enabled seedlings to establish even under conditions of strong interspecific competition. On the other hand, the seedling establishment and population growth rates of L. polyphyllus are robust to habitat type and soil ~ ber and Ramula, 2013; Ramula, 2014) and therefore, it is pH (So possible that interspecific competition has a minor effect on the invasion success of this species also in the field. However, it should be noted that our greenhouse experiment focused on seedling competition, which is restricted to the situation where plants (regardless of their invasiveness status) are colonising open microsites from seed and does not necessarily apply to other types of interspecific competition. For example, invasive species that establish from seed often compete for resources with existing (adult) plants that already have an initial size advantage. In such situations, the outcome of competition often depends on the strength of asymmetric competition, with strong interspecific competition having more severe consequences for the invader than less intense competition (Weiner et al., 2001). Despite the minor roles of disturbance and interspecific competition in the establishment of L. polyphyllus, we did find that seedling establishment tended to be higher in the more productive southern region than in the less productive central region (Table 1). In other words, plant communities in southern Finland seemed to be more susceptible to L. polyphyllus invasions than those in the central parts of the country. The higher seedling establishment of the southern region might reflect the climate preference of the study species. This explanation is also supported by the fact that both the seed production of individual L. polyphyllus plants as well as the long-term population growth rates of L. polyphyllus populations are about 40e46% higher in the southern region than in the central region (Ramula, 2014). Although invasion success often increases with increasing propagule pressure, density dependence resulting from intraspecific competition during the life-cycle of colonising individuals may suppress the overall role of propagule pressure in invasion success (Poulsen et al., 2007). For example, high propagule pressure resulted in a reduction in both germination and seedling survival (when measured at the relative scale) for the invasive grass Microstegium vimineum (Warren et al., 2012). In the present study, high propagule pressure tended to reduce the relative seedling establishment of L. polyphyllus in the field, indicating density-
S. Ramula et al. / Acta Oecologica 68 (2015) 18e23
dependent mortality. For L. polyphyllus, density dependence also plays a role later in the life-cycle when it reduces the survival and fecundity of individual plants under field conditions (Ramula, 2014). This density dependence in later life stages, together with density-dependent establishment, may partially cancel out the absolute increase in invasion success that results from increased propagule pressure in L. polyphyllus. However, longer-term observations covering the whole life-cycle of the study species are required to quantify the overall benefit (if any) of high propagule pressure for invasion success. In summary, the present study, conducted in multiple plant populations across different environments, revealed that propagule pressure plays a dominant role in increasing the absolute establishment, and thus invasion success, of the perennial herb L. polyphyllus. However, under field conditions, the relative establishment of L. polyphyllus tended to decrease with increasing propagule pressure, probably because of density-dependent processes. In contrast to expectations, both disturbance and interspecific competition had only a negligible effect on the seedling establishment of L. polyphyllus on their own or through interactions with propagule pressure. Overall, these findings suggest that L. polyphyllus invasions are limited by the availability of propagules rather than the availability of suitable microsites. Acknowledgements ~ ber for establishing the sowing plots, three We thank Virve So anonymous reviewers for their thoughtful comments, and the Emil Aaltonen Foundation for funding. References Buckley, Y.M., Bolker, B.M., Rees, M., 2007. Disturbance, invasion and re-invasion: managing the weed-shaped hole in disturbed ecosystems. Ecol. Lett. 10, 809e817. Colautti, R.I., Grigorovich, I.A., MacIsaac, H.J., 2006. Propagule pressure: a null model for biological invasions. Biol. Invas 8, 1023e1037. Crawley, M.J., 1986. The population biology of invaders. Phil. Trans. R. Soc. B 314, 711e731. Davis, M.A., Grime, J.P., Thompson, K., 2000. Fluctuating resources in plant communities: a general theory of invasibility. J. Ecol. 88, 528e534. Eschtruth, A., Battles, J., 2009. Assessing the relative importance of disturbance, herbivory, diversity, and propagule pressure in exotic plant invasion. Ecol. Monogr. 79, 265e280.
23
Fensham, R.J., Donald, S., Dwyer, J.M., 2013. Propagule pressure, not fire or cattle grazing, promotes invasion of buffel grass Cenchrus ciliaris. J.Appl. Ecol. 50, 138e146. Fremstad, E., 2010. NOBANISeInvasive Alien Species Fact SheeteLupinus Polyphyllus. eFrom: Online Database of the European Network on Invasive Alien SpecieseNOBANIS. www.nobanis.org. Grytnes, J.A., Birks, H.J.B., Peglar, S.M., 1999. Plant species richness in Fennoscandia: evaluating the relative importance of climate and history. Nord. J. Bot. 19, 489e503. Houseman, G.R., Foster, B.L., Brassil, C.E., 2014. Propagule pressure-invasibility relationship: testing the influence of soil fertility and disturbance with Lespedeza cuneata. Oecologia 174, 511e520. Huston, M.A., 2004. Management strategies for plant invasions: manipulating productivity, disturbance and competition. Divers. Distrib. 10, 167e178. Jauni, M., Gripenberg, S., Ramula, S., 2015. Non-native plant species benefit from disturbance: a meta-analysis. Oikos 124, 122e129. Kempel, A., Chrobock, T., Fischer, M., Rohr, R.P., van Kleunen, M., 2013. Determinants of plant establishment success in a multispecies introduction experiment with native and alien species. Proc. Nat. Acad. Sci. USA. 110, 12727e12732. Lockwood, J.L., Cassey, P., Blackburn, T., 2005. The role of propagule pressure in explaining species invasions. Trends Ecol. Evol. 5, 223e228. Moles, A.T., Flores-Moreno, H., Bonser, S.P., Warton, D.I., Helm, A., Warman, L., Eldridge, D.J., Jurado, E., Hemmings, F.A., Reich, P.B., et al., 2012. Invasions: the trail behind, the path ahead, and a test of a disturbing idea. J. Ecol. 100, 116e127. Pinheiro, J.C., Bates, D.M., 2000. Mixed-effect Models in S and S-plus. Springer, New York. Poulsen, J.R., Osenberg, C.W., Clark, C.J., Levey, D.J., Bolker, B.M., 2007. Plants as reef fish: fitting the functional form of seedling recruitment. Am. Nat. 170, 167e183. R Development Core Team, 2013. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. Ramula, S., 2014. Linking vital rates to invasiveness of a perennial herb. Oecologia 174, 1255e1264. Rapp, W., 2009. Invasive Plant Management in Glacier Bay National Park and Preserve Summer 2009. Field Season Report. Invasive Species Program, Glacier Bay National Park and Preserve, National Park Service, U.S. Department of the Interior, Gustavus, AK, p. 164. ~ ber, V., Ramula, S., 2013. Seed number and environmental conditions do not So explain seed size variability for the invasive herb Lupinus polyphyllus. Plant Ecol. 214, 883e892. Tuhkanen, S., 1980. Climatic parameters and indices in plant geography. Acta Phytogeogr. Suec. 67, 1e110. Von Holle, B., Simberloff, D., 2005. Ecological resistance to biological invasion overwhelmed by propagule pressure. Ecology 86, 3212e3218. Wardle, D.A., Ahmed, M., Nicholson, K.S., 1991. Allelopathic influence of nodding thistle (Carduus nutans L.) seeds on germination and radicle growth of pasture plants. New Zeal. J. Agr. Res. 34, 185e191. Warren, R.J., Bahn, V., Bradford, M.A., 2012. The interaction between propagule pressure, habitat suitability and density-dependent reproduction in species invasion. Oikos 121, 874e881. Weiner, J., Griepentrog, H.-W., Kristensen, L., 2001. Supression of weeds by spring wheat Triticum aestivum increases with crop density and spatial uniformity. J. Appl. Ecol. 38, 784e790.