Biological invasions are a major factor of global change that cause negative impacts on native biodiversity and ecosystem functioning (Pyšek and Richardson, 2010; Simberloff et al., 2013; Katsanevakis et al., 2014a; Vergés et al., 2016). In the marine environment, alien species (i.e., organisms introduced outside their natural range) can become invasive and cause numerous ecological changes such as the loss of native genotypes, degradation of habitats, changes in trophic interactions, and displacement of native species (Molnar et al., 2008; Albins, 2013; Giakoumi, 2014; Vergés et al., 2014, 2016). Furthermore, marine invasive species may have negative socio-economic impacts on fisheries and recreational activities, e.g., decline of commercial fish stocks (Bax et al., 2003; Katsanevakis et al., 2014b). Although marine biological invasions have been recognized as one of the major threats of marine ecosystems (Halpern et al., 2008; Molnar et al., 2008), the uptake of management actions to mitigate their impacts falls short (Thresher and Kuris, 2004; Giakoumi et al., 2016).

Marine protected areas (MPAs) form the cornerstone of marine conservation. They have emerged as a prominent management tool for the conservation and recovery of marine ecosystems and their ecosystem services (Lubchenco and Grorud-Colvert, 2015). The level of protection in MPAs varies from fully protected, where all extractive activities are prohibited, to partially protected, where some extractive activities occur with varying restrictions (Lubchenco et al., 2016). The ecological effects of fully protected areas (also called “no-take zones” or “marine reserves”) include increases in sizes of organisms, density and biomass of commercially exploited species and whole assemblages, reproductive potential, species richness, live cover of benthic organisms, and restoration of trophic interactions (e.g., Guidetti and Sala, 2007; Selig and Bruno, 2010; Edgar et al., 2014). Despite the availability of a large amount of information on MPAs, the effects of protection on marine alien species remain largely unknown (Burfeind et al., 2013; Ardura et al., 2016; Giakoumi et al., 2016).

Different mechanisms could explain how MPAs can control the presence and reduce the impacts of alien species within their borders. First, vectors of alien species are expected to be fewer in MPAs, because many human uses, such as aquaculture and marine traffic, are forbidden or restricted (Ardura et al., 2016). Secondly, based on the “biotic resistance hypothesis” (i.e., ecosystems with high species richness are more resistant to invaders than those with low biodiversity; Levine and D'Antonio, 1999), the high native species richness within MPAs could prevent the penetration and settlement of alien species. Thirdly, the restoration of top-down regulation processes (e.g., restoration of top predators' populations) in MPAs could help control the population of some alien species inside MPAs (e.g., Mumby et al., 2011 but see also Hackerott et al., 2013). On the other hand, several mechanisms could support the opposite argument, that MPAs favor the spreading of alien species. For example, according to the “biotic acceptance hypothesis” ecosystems can accommodate the establishment of aliens and their coexistence with native species, and based on a rich-get-richer pattern, areas with high native species richness could support high numbers of alien species (Stohlgren et al., 2006; Fridley et al., 2007). Moreover, harvested alien species could be enhanced in MPAs, mainly because they would benefit from harvesting bans and restrictions applying within the MPA (Burfeind et al., 2013). Finally, MPAs may have no effect on alien species as most of them have dispersal mechanisms that expand beyond the MPA boundaries (Burfeind et al., 2013).

Assessing the effects of MPAs on invasive species is crucial for the effective management of existing MPAs as well as for the optimal site selection for future MPAs (Giakoumi et al., 2016). If MPAs prove to have no effect or favor the settlement and expansion of invasive species, then the location of new MPAs in impacted areas should either be avoided (Boudouresque and Verlaque, 2005), or complemented with other management measures for the mitigation of invasive species' impacts (Thresher and Kuris, 2004). Here, building on previous efforts (Burfeind et al., 2013), we aim to synthesize all available information on this important topic at a global scale and assess the effects of MPAs on alien/invasive species following a meta-analytical approach.

The first study providing quantitative data on alien species inside and outside MPAs was published in 1995 (Grosholz and Ruiz, 1995). Yet, the first study that specifically aimed at investigating the performance of alien species in MPAs comparing to unprotected sites was only published 10 years later (Byers, 2005). Eleven out of the 17 studies explicitly refer to this subject providing different conclusions. Some highlight the negative (e.g., Ardura et al., 2015) or positive (e.g., Klinger et al., 2006) effects of MPAs on alien/invasive species, whereas other studies did not find any difference between MPAs and unprotected sites (a summary is provided in Table S2).

The largest proportion of the available literature provides data on alien molluscs (40%, n = 10), followed by algae (28%, n = 7) (Figure 1A). Studies have been conducted in only 7 out of 62 marine biogeographic provinces (Figure 1B). The majority of the studies (70%) come from the Mediterranean Sea and the Tropical Northwestern Atlantic (more specifically the Caribbean Sea).

Most species were detected in their host environment throughout the 20th century, while the vast majority of MPAs were established between 1980s and 2000s (Table S1). On some occasions, alien species were totally absent from the MPA while present in the broader study area; this was the case for: (1) Mytilus galloprovincialis in the Western Cape of South Africa (Malherbe and Samways, 2014), and (2) Drupa albolabris, Nerita tessellata, Saccostrea cucullata, and Semiricinula tissoti in Moorea, French Polynesia (Ardura et al., 2015). All studies used visual census to estimate the density and/or biomass, or percent cover, or catch rate, or colony size of alien species inside vs. outside MPAs, apart from one that used trapping methods (Thresher et al., 2003). The majority of the alien species (~70%) have been reported to have negative impacts on native species and habitats in their host environment while few species (25%) were harvested in the study region (Table S1).

When we estimated the effect of protection on the density or biomass of individual alien species, we found that protection had a significant negative effect on 50% of the species (Figure 2). A significant positive effect was detected on 33% of the species, whereas for the remaining 12% no statistically significant effect was found. All species showing negative responses to protection were mollusc species in Moorea Island, French Polynesia (Marquesas) and in Western Cape, South Africa.

Our review demonstrates the scarcity of published evidence on a crucial topic: the effects of MPAs on alien/invasive species. The available data support the hypothesis that alien species respond negatively to protection; the density of most alien species was found to be greater outside than inside MPAs (as shown in Figure 2). Our findings differ from the conclusions of the precedent review article on this topic (Burfeind et al., 2013), according to which alien species do equally well or better within MPAs. This difference in results may be explained by the different methodological approaches and datasets used in each study. However, Burfeind et al. (2013) argue that their conclusions are based on too few data and caution is needed in interpreting their results. We also argue that our results should be considered with caution and that more data on various taxonomic groups across different biogeographic regions are needed to draw robust conclusions. Below, we provide recommendations for advancing knowledge on this important research topic.

Currently, studies based on robust impact assessment designs aiming to investigate the relationships between MPAs and invasive species are scarce (but see Byers, 2005; Klinger et al., 2006). Quantitative studies estimating specific biological and ecological variables (e.g., density, biomass, species richness) should be conducted by sampling alien/invasive species inside MPAs (both in zones of full and partial protection) and outside (control) areas. In most cases, data are not available to compare communities before and after the invasion, so the implementation of BACI (Before-After-Control-Impact) design (Schmitt and Osenberg, 1996) cannot be applied. Yet, replication of control—impact studies using the same sampling protocol and comparing specific variables through time is important, because biological invasions are a dynamic threat (Strayer et al., 2006). Appropriate experimental designs are needed to investigate the mechanisms explaining how MPAs can control the presence and reduce the impacts of alien species within their borders.

Most of the available information (~80% of the species assessed) refers to molluscs and algae. More studies are needed across various taxonomic groups as their responses to protection might differ significantly. In addition, studies within taxonomic groups should examine whether alien species of similar traits have similar responses to protection. Functional traits have been associated with the invasiveness of alien marine species (e.g., Elleouet et al., 2014; Samaha et al., 2016) as well as to the response of native species to protection (e.g., Claudet et al., 2010). Hence, alien/invasive species of similar traits are likely to display similar patterns and tendencies within MPAs. Such information may have important implications for the selection of appropriate management actions as well as for comparing different management options (see Bremner, 2008).

Our results show that information on the effects of protection on alien/invasive species is available for only 11% of the marine biogeographic regions, even though biological invasions are considered a global threat affecting almost the entire ocean (Bax et al., 2003; Halpern et al., 2008). The response of alien/invasive species to protection is likely to be place-specific (Giakoumi et al., 2016). Therefore, the investigation of the effects of MPAs on alien/invasive species in different marine bioregions is a prerequisite for making informed decisions on the uptake of appropriate mitigation measures at a local and regional scale.

The relationships between MPAs and invasive species are likely to be complex. MPA characteristics, such as location, size, level of protection (full vs. partial) and age, as well as dynamic processes within an MPA, such as disturbance regime and magnitude of invasion pressure, can affect the response of alien/invasive species to protection (Burfeind et al., 2013). Unfortunately, the available data do not allow us to disentangle the influence of each factor on the variables under study. Current information is limited and evidence provided is often inconsistent. For example, both studies conducted to assess the performance of the invasive lionfishes (Pterois volitans, P. miles) in the Caribbean Sea found that the species presented lower density and biomass in the MPAs than unprotected sites (Mumby et al., 2011; Hackerott et al., 2013). However, one study attributed the lower values within the MPAs to the restoration of predatory relationships (more and larger top predators—groupers—controlled the densities of lionfishes within the MPA; Mumby et al., 2011), whereas the other study attributed the lower densities to targeted removal (by spearfishing) within MPAs (Hackerott et al., 2013). Another study examining the effects of MPAs on alien fishes in the Mediterranean Sea, found that the density of alien species did not show any response to protection (Guidetti et al., 2014). Nevertheless, this study lacked appropriate impact assessment design that would allow pairwise comparisons of control and impact sites. Appropriate sampling designs should be applied to investigate not only the effect size of protection on alien/invasive species comparing sites inside and outside MPAs, but also the effects of factors influencing the relationships between MPAs and alien/invasive species.

In conclusion, our review demonstrated the scarcity of evidence on the effects of protection on alien/invasive species. The tropicalization of temperate marine environments (i.e., the increase in the proportion of warm-water species as the ocean warms) can have community-wide effects and threaten iconic ecosystems (Vergés et al., 2016). As biological invasions are a dynamic threat spreading rapidly, especially in human-dominated marine environments (Zenetos et al., 2012; Galil et al., 2014), urgent action is required to arrest or mitigate this threat. Assessing whether MPAs—the main management tool currently applied for ocean protection—can control the spread and/or mitigate the impacts of invasive species is critical. Our results indicate that MPAs can have a negative effect on alien/invasive species. Yet, more empirical studies on various species and taxonomic groups in different marine biogeographic regions are necessary to draw robust conclusions.

SG conceived the idea, conducted the literature review, performed the analyses, and led the writing of the manuscript and all aspects of the project. AP contributed to the literature review, produced the figures, and revised the manuscript.