Efforts to restore ecosystems invaded by non-native plants often aim to reverse invasion impacts, yet removal can trigger new community shifts across trophic levels. We examined how removal of Rhamnus cathartica (common buckthorn) influenced plants, arthropods, pollinators and rodents in a post-industrial forest preserve in western New York, USA. Across 18 plots representing restored, buckthorn tree, and buckthorn shrub stands, we quantified vegetation structure, leaf litter biomass and faunal abundance. Buckthorn removal promoted a tenfold increase in herbaceous cover and species richness, yielding more diverse and structurally complex understories. These changes coincided with higher arthropod and pollinator abundance, but also a three- to fivefold increase in the invasive ant Myrmica rubra (European fire ant). Contrary to expectations, M. rubra flourished in restored plots, responding to thicker and more stable leaf litter rather than canopy shade. Pollinator abundance declined with increasing M. rubra density, suggesting interference or avoidance effects, whereas detritivores and rodents responded primarily to vegetation and litter structure. Our single-site design isolates general mechanisms—litter-mediated ant responses and structure-driven pollinator and arthropod gains—that are testable across temperate forests. The results indicate that R. cathartica removal reshaped multitrophic community structure – enhancing plant and arthropod recovery but simultaneously facilitated a secondary invasion. Restoration success may therefore depend on eliminating the initial invader while recognizing that restoration itself can act as an invasion-promoting disturbance.

Study Site

Tifft Nature Preserve is a 107-ha post-industrial forest-wetland complex near the eastern shore of Lake Erie in Buffalo, NY, USA (42.84651, −78.85951). The area was originally part of a broad floodplain and wetland system south of the Buffalo River mouth, later converted to farmland and used as a transshipment facility for iron ore, lumber and coal. It subsequently served as a dumping site for industrial and municipal waste before being restored as a nature preserve in the 1970s. The site has alkaline soils (pH 7.6–7.8) high in calcium (Spiering 2009). The woodland canopy is dominated by Populus deltoides (eastern cottonwood) with some Salix spp. (willow) and Ailanthus altissima (tree-of-heaven). The understory is dominated by Rhamnus cathartica (with Frangula alnus [syn. Rhamnus frangula] in wetter areas), Reynoutria spp. (Japanese knotweed), Cornus spp. (dogwood) and Lonicera spp. (honeysuckle), with limited natural regeneration of native woody species (Labatore et al. 2017). Odocoileus virginianus (white-tailed deer) are overabundant at Tifft and in surrounding urban and suburban areas (Spiering 2009; Booth-Binczik and Hurst 2018), and invasive M rubra has colonized mesic areas in the preserve (Warren II et al. 2019a)). As part of a site restoration plan implemented between 2010 and 2013, about 2,000 saplings of 31 native tree species were planted across 40 ha of forested areas following mechanical removal of R. cathartica.

Target species

Rhamnus cathartica is a woody species introduced to North America during European colonization in the 1800s for medicinal and ornamental purposes (Kurylo and Endress 2012). In its native range, it occurs primarily in open habitats and forest edges, but in North America it vigorously invades forest understories, forming dense monotypic stands that monopolize light and space (Knight et al. 2007; Klionsky et al. 2011). These invasions are associated with declines in native plant richness and abundance and contribute to altered community composition (Mascaro and Schnitzer 2007; Knight et al. 2007; Warren et al. 2017). Beyond vegetation effects, the species produces nitrogen-rich leaves that decompose rapidly, accelerating nutrient cycling and releasing nitrogen into the soil (Heneghan et al. 2006; Klionsky et al. 2011). It is often found in moist soils, though this may reflect habitat preference rather than direct hydrological effects (Frappier et al. 2003; Heneghan et al. 2006; Knight et al. 2007). Collectively, these traits restructure microhabitats and food webs, reducing the abundance of both native plants and animals (Mascaro and Schnitzer 2007; Knight et al. 2007; Grunzweig et al. 2015). For these reasons, management frequently targets the species for removal, which can increase light availability, reduce competition, and alter soil nutrient and moisture conditions (Wragg et al. 2021; Lamb et al. 2022; Schuster et al. 2024).

            Myrmica rubra, native to Europe and parts of Asia, has established invasive populations in North America, particularly in the northeastern United States and southeastern Canada (Wheeler 1908; Groden et al. 2005; Wetterer 2011). At our study site, it has colonized mesic areas within the preserve (Warren II et al. 2019a, b). Myrmica rubra prefers moist, shaded habitats such as woodlands and gardens but can also persist in disturbed areas, provided they do not become excessively dry (Groden et al. 2005; Chen and Adams 2018; Warren II et al. 2019a). Unlike most native ants, M. rubra forms large, interconnected (polydomous) colonies, which confer a competitive advantage over native species that occupy smaller, isolated nests  (Garnas et al. 2014; Warren II et al. 2019b; Goodman and Warren II 2019). Its invasion has been associated with declines of up to 90% in native ant populations, as well as reductions in some native arthropods on which it preys (Reznikova and Panteleeva 2001; Gammans et al. 2018; Goodman and Warren II 2019).

Study plots

In June 2021, we established 18 plots (100 m² each; 10 × 10 m; mean distance between plots = 75m) at Tifft, distributed across three treatment types (six plots per treatment): (1) ‘restored’ plots – R. cathartica was mechanically removed and mulched in 2020 and resprouts chemically treated in 2020 and early 2021 with 5% Garlon 3A (the stumps were targeted rather than broadscale application); (2) ’buckthorn shrub‘ plots – R. cathartica was mechanically removed in 2012 but not re-treated, resulting in dense regrowth of plants by 2021; and (3) ‘buckthorn tree’ plots – *R. cathartica *was never removed and has matured into tree-dominated stands with little understory (Fig. S2). All plots contained native trees planted between 2010 and 2013 as part of restoration efforts following initial R. cathartica removal. Approximately 90% of the surviving native trees were planted individuals (primarily Acer spp., but also included Quercus, Tilia and Betula spp.), with limited volunteer native tree recruitment (P. deltoides).

Floral and faunal surveys

Stem diameter (diameter at breast height, DBH) and stem counts of shrub-sized woody plants (DBH < 10 cm) were measured along a 14-m diagonal transect in each plot using a caliper. Tree-sized woody plants (DBH ≥ 10 cm) were measured with a DBH tape. Herbaceous-layer vegetation was sampled in four 0.25 m² square subplots per plot, positioned 1 m from each corner.

Pollinators were collected three times (June, July, and August 2021) using blue vane traps (SpringStar LLC, Woodinville, WA, USA). Each trap consisted of a plastic container (15 cm diameter × 15 cm high) with a blue polypropylene screw funnel and cross vanes. Traps were filled with water and a drop of detergent (to reduce surface tension) and suspended 1.2 m above the ground using a shepherd’s hook. Traps remained deployed for seven consecutive days per sampling period. Tree-dwelling arthropods were collected three times (June, July, and August 2021; n = 54 total) using a 1 m² canvas beat sheet placed beneath low-hanging branches and trunks of trees in each plot. A PVC stick was used to shake 4 branches for 15 seconds each to dislodge arthropods onto the canvas for collection.* *Leaf litter arthropods were sampled three times (June, July, and August 2021; n = 54 total) by collecting all leaf litter and decaying organic material at the litter-soil interface (Edgar 1992). Samples were processed using Berlese-Tullgren funnels under tungsten bulbs for seven days, with extracted arthropods preserved in alcohol. Myrmica rubra foragers were collected arboreally in the beat-sheet sampling and terrestrially in the leaf-litter samples.

Rodents were live-trapped twice (June and July 2021) using Sherman traps (H.B. Sherman Inc., Tallahassee, FL, USA). Four traps were placed 1 m from each corner of each plot at dusk and checked at dawn. Traps were baited with a high-quality seed mixture and freeze-dried mealworms. Captured rodents were transferred to a cloth bag, identified to genus, and released. Permission for handling rodents was granted by the New York State Department of Environmental Conservation (License to Collect), the Buffalo Science Museum, and the Buffalo State Institutional Animal Care and Use Committee.     

            Fauna were identified to order (pollinators to family) and flora to family (Bland and Jaques 1978; Brown 2020; Tredici 2020; Neal et al. 2023)