Introduction
Over recent decades, global human transportation networks have led to the establishment of once geographically restricted species in new ecosystems (Hulme 2009; Hulme et al. 2008; Roderick & Navajas 2015; Sinclair et al. 2020). Introduced taxa may act as “ecosystem engineers” by changing the structure of an ecosystem through alteration of abiotic and biotic factors (Jones et al.1994); such non-native taxa may then be called invasive because of their negative impact on ecosystem services and native species (Richardsonet al. 2011). Invasive plants are particularly pervasive and can act as ecosystem engineers by changing habitat structure, soil chemistry (Ehrenfeld 2003), nutrient cycling (Weidenhamer & Callaway 2010) and microclimates (Ruckli et al. 2013). For example, the Hawaiian invasion of the firetree Myrica faya increases nitrogen availability in a generally nitrogen-limited system (Vitousek et al. 1987). One consequence of ecosystem engineers is invasional meltdown, in which modification driven by one invasive facilitates the establishment of other non-native taxa (Green et al. 2011; Simberloff & Von Holle 1999). Major compositional shifts in the community often occur following introduction of multiple non-native taxa, resulting in trophic cascades and shifts in biotic interactions (Borges et al. 2006; Cucherousset et al. 2012; Wainrightet al. 2021). Through changed abiotic and biotic factors, invaded ecosystems may then function as a habitat “sink” for native taxa in a heterogenous landscape; in this case, endemic species would migrate from less-disturbed, productive “source” habitat into invaded “sink” habitat where reproductive output is diminished below population sustainability (Pulliam 1988). By studying source-sink dynamics across adjacent invaded and non-invaded habitat, we can start to quantify how alien invasion impacts ecosystem structure and function at a landscape scale.
To understand source-sink dynamics caused by invasion, it is important to go beyond single taxa and study entire biological communities. However, source-sink dynamics can produce misleading results when quantifying communities using classic biodiversity metrics. This may explain the contrasting results of studies examining how invasions impact biotic communities. For instance, the response of arthropod communities to plant invasion shows mixed effects, with many studies detecting abundance and richness decreases while others show increases (Litt et al. 2014). In a source-sink context, species presence and abundance values in a habitat would not indicate a productive system. Through changes in biotic interactions associated with invasion (Bezemer et al. 2014) such as increased predation risk, lack of the most nutritionally beneficial prey items or a higher interaction with parasites (Mattos & Orrock 2010; Suarez et al. 2000), invaded habitat may serve an ecological trap. Understanding the new network of biotic interactions between native and non-native taxa following invasion will provide insight into functional shifts that occur in ecosystems and indicate if invasion can induce an ecological trap for native taxa.
In our study, we use a recent invasion of kāhili ginger (Hedychium gardnerianum ) in Hawai‘i to measure how the establishment of an invasive ecosystem engineer influences biotic interactions. Kāhili ginger is one of the world’s worst invasive species and has invaded the Azores, Madeira, Jamaica, Réunion, New Zealand and Hawai‘i, while also expanding in South and Central America, Australia and Southern Africa (Pereira et al. 2021). Native to the Himalayas, H. gardnerianum was brought to Hawai‘i in 1940 and is found in wet forest throughout the Hawaiian Islands, where it can form almost impenetrable stands up to 3 m in height (Santos et al. 1992; Vorsino et al. 2014). It is a very aggressive, shade-tolerant plant, and can invade and establish in intact native rain forest habitat, displacing native understory vegetation and altering composition of soil microbial decomposers (Kao-Kniffin & Balser 2008; Minden et al. 2010). Compositional differences between native forest and ginger-invaded sites have been noted using invertebrate communities in New Zealand, though the impact was context dependent (Yeates & Williams 2001). Changes in the abundance of fungivores and decomposers is the most consistent result of the invasion (Bassett 2014). The evidence of compositional shifts caused by ginger most certainly causes differences in the interactions between native and non-native taxa. Because of the global prevalence of ginger, understanding such impacts on biotic interactions are crucial to mitigate the costs associated with invasion.
Our current study uses the generalist endemic spider (Pagiopalus , Philodromidae) (Suman 1967) to assess whether modified sites could serve as sink habitat. While detailed demographic studies are essential to confidently identify source-sink dynamic, this is logistically challenging in most systems. Instead, by assessing shifts in biotic interactions, we can ask if major shifts in ecological function occur for native taxa; because such changes will likely carry a cost, altered biotic interactions may then indicate suboptimal habitat. We utilize metabarcoding to compare dietary communities and parasite loads in spiders across sites in ginger-invaded habitat and native-forest. Through the diet of a generalist predator, we not only can assess biotic interactions between predator and prey, but we also obtain a window into the arthropod community composition. The study focuses on a site on the island of Maui, specifically the mesic forest of Waikamoi on East Maui. Here, a sharp boundary exists between the ginger invasion and the native forest due to the efforts of the Nature Conservancy of Hawai‘i in protecting their lands (The Nature Conservancy of Hawai‘i 2011). The adjacency of native forest and ginger-invaded sites creates the possibility of connectivity between metapopulations. This provides an optimal system to assess differences in the quality of neighboring habitat and ask if ginger habitat serves as a “sink” by changing the biotic interactions of an endemic generalist predator found in both habitats.
In pursuit of this question, we have four hypotheses. First, becausePagiopalus are generalist predators capable of eating a wide diversity of prey times, we expect to find establishment of the spider in ginger, reflected in similar abundances across sites. Second, the altered environmental conditions in ginger will result in arthropod communities differing from native forest sites; we expect, then, to detect compositional differences in the diets between spiders collected in ginger and native forest. Third, ginger sites will host more non-native prey items, and this will be reflected in the diet ofPagiopalus collected in ginger habitat while the diets ofPagiopalus in native forests will instead consist of mostly native prey. Lastly, given the hypothesis that the novel environments created by ginger will serve as sink habitat by increasing diversity of non-natives, we hypothesize a higher interaction with non-native parasites indicative of poor habitat quality.
Methods