Alien Species Introductions Harm Native Ecosystems

Human-mediated introductions of species outside their native range, so-called alien species, have encompassed all taxonomic groups and geographic regions1. A subset of alien species has inflicted substantial harm on native biodiversity, exacerbating its decline alongside other major anthropogenic drivers1,2,3,4; these are commonly referred to as invasive alien species5. Not all alien species harm native biodiversity, and the magnitude of their impacts varies across different alien populations6. However, such context dependence of impacts remains poorly understood7,8, and for most taxonomic groups it is unclear which species are the most harmful and which biological traits and ecological factors determine impact severity. For instance, while islands are generally more susceptible to invasion9 and more vulnerable to anthropogenic pressures than continents10,11, it is unclear whether alien species' impacts on insular biodiversity are consistently more pronounced than those on the mainland. Studies have shown that global extinction risk posed by invasive predators on native species is higher on islands12,13, although it is yet to be determined if the same pattern holds for other negative impact types and magnitudes (reduction in performance of individuals, reduction in population size, local extinction14). Similarly, theory predicts that higher trophic levels are more vulnerable to environmental alterations15,16,17, but few studies have investigated if invasive alien species cause stronger negative impacts to native species positioned high in the food chain18,19.

Native species can also benefit from the introduction of alien species20,21. Yet, positive impacts are less often documented than their negative counterparts20. While most scientists acknowledge the existence of positive impacts22, there is a controversy over whether they are overlooked or their extent has been over- or underemphasized in comparison with negative impacts23,24,25,26. As far as we know, there has been no quantitative, systematically collected and taxonomically controlled study in support of any of these claims, so that the alleged bias of focusing on negative impacts or overstating their magnitude has never been rigorously tested. Additionally, no in-depth investigation on the factors that determine the positive impacts of alien species on native biodiversity has been conducted. A persistent challenge in this regard has been the lack of a transparent and comprehensive framework for measuring and evaluating negative and positive impacts, as well as for effectively comparing their frequencies and magnitudes21,27,28,29.

Here, we employ the International Union for Conservation of Nature’s (IUCN) Environmental Impact Classification for Alien Taxa (EICAT) framework14,30 and the recently developed EICAT+ framework31 to systematically assess negative and positive impacts of introduced large mammalian herbivores (LMH) (LMH; Cetartiodactyla, Perissodactyla, Proboscidea) on native biodiversity on a global scale. While many LMH face dramatic population declines and range contractions because of global change32, there is a growing call to introduce them outside their native range for conservation purposes, such as trophic rewilding, assisted colonization and climate change mitigation33,34,35. The EICAT(+) frameworks consider native biodiversity as the entity of conservation concern36 and classify alien species’ impacts by their direction, i.e. distinguishing whether they pose harms or offer benefits to local populations of native species. These frameworks also classify impact magnitude into Minimal, Minor, Moderate, Major and Massive levels (see Methods). For this study, assessed impact magnitudes have been further categorized as “weak” or “strong” based on whether they involved individual-level (Minor or lower) or population-level (Moderate or higher) changes to native species. This dichotomous variable is referred to as “impact magnitude” hereafter.

Additionally, the mechanisms through which these impacts were caused have been classified as “direct” or “indirect” (mechanism type). Finally, each impact has been attributed a confidence (low, medium or high) to express the uncertainty associated with the accuracy of the assigned impact magnitude. The combined use of EICAT(+) enables us to conduct a standardized, comparable and taxonomically controlled bidirectional impact assessment needed to address the above controversy. Furthermore, we used the assessed impact data to investigate to what extent insularity and trophic position shape the magnitude of both negative and positive impacts experienced by native species.

Under the assumption that the introduction of species outside their native range mostly disrupts established eco-evolutionary dynamics, we hypothesize that (1) negative impacts caused by introduced LMH on native species occur at a higher frequency and with greater impact magnitude compared to positive impacts. We name this hypothesis the “Harm Dominance Hypothesis”. We also hypothesize that (2) both negative and positive impacts of introduced LMH are greater in magnitude on islands and on native species positioned higher in the trophic chain. The hypothesis regarding the influence of insularity and trophic position in amplifying negative impacts stems from circumstantial evidence from previous research19,37. Conversely, the hypothesis of greater positive impacts on islands and higher trophic level stems from the rarely tested assumptions that introduced species can restore functions of extinct insular species34,38, or serve as an important novel food resource for native consumers positioned directly above in the trophic chain39,40. Finally, we hypothesize that (3) due to their salience, negative and positive impacts of higher magnitude (strong impacts), such as local extinctions, have been identified first, and thus the reported impact magnitudes across studies would decline over time.

Here, we show that negative impacts of LMH are more common, and of higher magnitude, than positive impacts, while both are greater on islands and at higher trophic levels. We also observe that reported impact magnitudes decline over time only for positive impacts. We conclude that caution is necessary when considering the intentional introduction of LMH for conservation purposes, such as rewilding or assisted colonization, without a rigorous evaluation of their multifaceted impacts on native communities.

Frequency of negative and positive impacts

We found 303 reports describing 1616 negative and 405 positive impacts for native species that could be classified under EICAT or EICAT + , from 29 of the 66 listed alien LMH species. Negative and positive impacts were caused by 28 and 21 LMH species, respectively (Fig. 1, Supplementary Data 1). About two thirds of alien LMH species (20 out of 29) caused simultaneously both negative and positive impacts, although for species having bidirectional impacts, we detected 3.7 times more negative than positive impact observations overall (1489 vs. 399, Fig. 1). When comparing these LMH species individually, the trend remained largely consistent (paired sign test: n = 20, p < 0.001), with records of negative impacts (mean = 74.5 ± 64.6 SD) outnumbering positive impacts (mean = 20 ± 17.3 SD) in all species except two (Bos taurus and Boselaphus tragocamelus, Fig. 1).

Species having exclusively negative impacts in their alien ranges were the Aoudad (Ammotragus lervia), the American bison (Bison bison), the Wapiti (Cervus canadensis), the Asian elephant (Elephas maximus), the Guanaco (Lama guanicoe), the Gemsbok (Oryx gazella), the Mouflon (Ovis orientalis), and the Javan deer (Rusa timorensis). Only one species, the Indian hog deer (Axis porcinus), had exclusively positive impacts, Fig. 1).

A great majority (n = 27; 93%) of alien LMH species for which impacts are reported caused strong impacts (positive or negative). Almost all the 28 species with negative impacts caused strong impacts (93%, n = 26). By contrast, among the 21 species causing positive impacts, only 71% (n = 15) caused strong impacts (Fig. 1). Observations of negative impacts outnumbered those of positive impacts across all levels of magnitude (Fig. 1, Supplementary Fig. 1) and confidence (Supplementary Fig. 2).

The predominant impact magnitude observed was Moderate (MO and MO + , Supplementary Fig. 1., Supplementary Data 1), with native population decline documented at 52% (840/1616) and native population increase at 45% (184/407). Across the five impact magnitude levels, confidence was mostly categorized as low and medium, while a high confidence was less frequently assigned (Supplementary Fig. 2). This trend was consistent for both negative and positive impacts, except for cases where alien LMH increased the size of native populations (MO + ). These cases were assigned with significantly higher confidence compared to instances of native population decreases (MO) (z-test, z = −4.1, p < 0.001, Supplementary Fig. 2).

Overall, alien LMH caused negative impacts mostly through direct mechanisms (direct = 78%, denoted by black labels in Fig. 2), while the opposite trend was observed for positive impacts (indirect = 85%, denoted by green labels in Fig. 2, Supplementary Table 1). The most frequently recorded mechanism for negative impacts was direct “grazing, herbivory, or browsing” (n = 982 impacts), followed by indirect “chemical, physical, or structural impacts on ecosystems” (n = 314 impacts), and direct “bio-fouling or other direct physical disturbances” (n = 296 impacts). Conversely, “indirect impact through interactions with other taxa” (n = 275 impacts) was the predominant mechanism through which positive impacts were caused, followed by indirect “chemical/physical/structural impact on the ecosystem” (n = 74 impacts).

Negative impacts of alien LMH were more frequently documented on islands (68% of all reports of negative impacts), whereas positive impacts showed a more even distribution (51% from islands, 49% from mainland) (Fig. 3B). Impacts from alien LMH affected four trophic levels: decomposers, producers, primary consumers, and secondary consumers. The trophic level most frequently impacted, both negatively (74%) and positively (59%), was producers (Fig. 3D).

Predictors of impact magnitude

Among the 511 models obtained, each representing different combinations of variables and their interactions with impact direction, 12 exhibited a ΔAICc <6 in relation to the best model (Table 1). Thus, we performed model averaging and estimated the relative importance (sum of Akaike weights) of each factor and interaction within the selected set of models (Table 2).

After averaging models across the 13 best-fitting candidates, predictors (both factors and interactions) demonstrating sufficient explanatory power (relative importance > 0.5, Table 2) aligned with those featured in the model characterized by the lowest AICc (best-fitting model, Table 1). We therefore selected the best-fitting model (Tables 1, 3) as the most supported model for further analyses, pairwise comparisons and data visualizations.

Globally, alien LMH species exhibited a higher probability to cause strong negative impacts than positive impacts (Table 3, Fig. 3A, Supplementary Table 2). Moreover, both negative and positive impacts were stronger in insular locations compared to mainland locations (Table 3, Fig. 4A, Supplementary Table 2). Over the years, we detected a non-significant overall decrease in the probability of causing strong impacts, with a steeper decline in the magnitude of positive compared to negative impacts (Table 3, Fig. 4B). Regardless of impact direction, alien LMH species were more likely to cause strong impacts on insular than on mainland locations (Fig. 3B, Supplementary Table 2), through indirect compared to direct impact mechanisms (Fig. 3C, Supplementary Table 2), and on secondary consumers compared to primary consumers (Fig. 3D, Supplementary Table 2). Conversely, the effects on other trophic levels (producers and decomposers) were indistinguishable (Fig. 3D, Supplementary Table 2).

Confidence in assigning impact magnitude

The complete model including Confidence, Direction, Year and their 2-way interactions as predictors of impact magnitude strongly outperformed all simpler models (ΔAICc = 6.45 from the second-best model). According to this model (Table 4), strong impacts were assigned with higher confidence than weak impacts, regardless of impact direction (Fig. 5A, Supplementary Table 2), but the rise in confidence with impact magnitude was steeper in positive than in negative impacts (Fig. 5B, Supplementary Table 2). Moreover, the probability of causing strong impacts decreased significantly faster over the years for impacts classified with high and medium confidence than for those classified with low confidence (Fig. 5C).

The introduction of LMH outside their native range has both harmed and benefited local native biodiversity, but negative consequences have largely surpassed positive outcomes, both in frequency and magnitude. Here, we comprehensively compared the negative and positive impacts of alien species and identified factors determining their magnitude. By leveraging the methodological advances of the EICAT(+) frameworks, we systematically tested hypotheses that were previously only supported anecdotally for negative impacts and never tested for positive impacts. This enabled us to provide a rigorous and detailed examination of how species that have established alien populations impact native biodiversity, demonstrating that the magnitude of both their negative and positive impacts is influenced by common factors such as insularity and trophic position.