Appearance of a population of the mangrove rail Rallus longirostris Boddaert, 1783 (Rallidae) in salt marshes invaded by the exotic tanner grass Urochloa arrecta (Poaceae) and its disappearance after plant management

Marcos R. Bornschein UNESP Campus Experimental do Litoral Paulista: Universidade Estadual Paulista Julio de Mesquita Filho Campus Experimental do Litoral Paulista Bianca L. Melchiori UNESP Campus Experimental do Litoral Paulista: Universidade Estadual Paulista Julio de Mesquita Filho Campus Experimental do Litoral Paulista Larissa Teixeira UNESP Campus Experimental do Litoral Paulista: Universidade Estadual Paulista Julio de Mesquita Filho Campus Experimental do Litoral Paulista Bianca L. Reinert In memorian Giovanna Sandretti-Silva (  giovanna.sandretti@unesp.br ) Universidade Estadual Paulista: Universidade Estadual Paulista Julio de Mesquita Filho


Introduction
The introduction of alien species has intensi ed with globalization in recent decades (Meyerson and Mooney 2007). Exotic species are de ned as species that are not native to an ecosystem and that cause or are likely to cause economic, environmental, and/or human health damage (Catling 2005). They can change the composition of ecosystems rapidly and profoundly (Hobbs et al. 2009) and, through their direct and indirect effects, contribute substantially to species extinction (Vitousek et al. 1997; Bellard et al. 2006). Consequently, biological invasions are considered the second most common cause of biodiversity loss (Simberloff 2007). However, the effects of invasive species are not all negative, and the "native good, alien bad" dichotomy has been questioned (Goodenough 2010). Exotic species can bene t native species through habitat modi cation, trophic subsidy, pollination, competitive release, and predator release mechanisms (Overton et al. 2014). Understanding and studying the responses of native species to invasive alien species is essential for understanding impacts and deciding conservation actions  (Doody 2001). They are regularly ooded by tides, have rapid sediment accumulation, and include transitions to non-tidal vegetation in the absence of human interference (Doody 2001). They occur in temperate areas across the globe, are more extensive in the northern hemisphere, and have seagrass Spartina spp. as the most common plant species (Doody 2001). Recently, some marshes on parts of the South American Atlantic coast have been recognized as salt marshes-speci cally, subtropical salt marshes (Bornschein et al. 2017). They are characterized by the dominance of the crinum lily Crinum americanum L. and the California bulrush Schoenoplectus californicus (C. A. Mey.) Soják; therefore, the presence of smooth cordgrass Spartina alterni ora Loisel is rare (Bornschein et al. 2017). Subtropical salt marshes occur in association with mangroves, distributed in Brazil to a small extent from the estuaries of the central-south coast of São Paulo to the north coast of Santa Catarina (Bornschein et al. 2017).
Tanner grass Urochloa arrecta (Hack. ex T. Durand & Schinz) Morrone & Zuloaga invasion and domination of South American salt marshes is the greatest threat to the conservation of endemic birds in southern Brazil and was described only in 1995 regarding the marsh antwren Formicivora acutirostris (Reinert et al. 2007). Environments completely invaded by exotic grass are no longer occupied by this bird, which is why the impact is considered an area suppressor (Reinert et al. 2007). In contrast, marshes in California, USA, were invaded by a hybrid species of Spartina that bene ted populations of the threatened Ridgway's rail Rallus obsoletus (taxonomy according to Maley and Brum eld [2013]). The hybrid form increases the survival rate of individuals of this bird by providing refuge against predators during extreme tides that inundate native vegetation, particularly during winter when native vegetation enters senescence (Overton et al. 2014). Hybrid plant management programs reduced the survival rate of R. obsoletus, and the plans for its conservation suggested offering refuges against high tides resulting from rises in sea level (Overton et al. 2014).
From 2006-2022, salt marshes in southern Brazil have been studied as part of a long-term project aiming to monitor and conserve F. acutirostris-an endangered (EN) species at risk of extinction in Brazil (Ordinance #444 of the Brazilian Ministério do Meio Ambientes, December 17, 2014). In 2012, F. acutirostris conservation efforts encompassed a challenging program that aimed to eradicate U. arrecta (Bornschein 2013), which was impacting a local population of the mangrove rail Rallus longirostris Boddaert, 1783. The Rallus longirostris is generally restricted to mangroves (Vieira 2015) and distributed across small portions of the Paci c coastal region, Central America, and northern South America, and along a vast stretch of the Atlantic coast in South America (Maley et al. 2016). A signi cant extent of the species' geographic distribution occurs along the Atlantic coast of Brazil-a country in which it was considered at risk of extinction (Vieira 2015). In this article, we report on the appearance and distributional expansion of R. longirostris on subtropical salt marshes invaded by U. arrecta and its disappearance after the eradication of this plant. We also discuss the possible causes of this appearance and disappearance.

Study Species
The target species was the mangrove rail Rallus longirostris, which is considered a separate species from the North American R. obsoletus and the clapper rail R. crepitans (Maley and

Study Areas and Field Time
We worked in the Guaratuba Bay estuary, RAMSAR Site Guaratuba, in the municipality of Guaratuba, on the southern coast of Paraná, in southern Brazil. Speci cally, we studied four areas: Jundiaquara Island (c. 25°52'25"S, 48°45'32"W; 11.30 ha), the con uence of the Claro and São João Rivers ("Continente"; c. onward, we worked in the areas for 3-8 days per month, every month. We did not work in the areas for six straight months in 2020 due to the COVID-19 pandemic (March-August). Fieldwork was carried out by 2-7 people (usually three). We accessed the places by boat. On each eldwork day, we worked from dawn until 12 a.m. or 1 p.m., and for a further 2-3.5 h in the afternoon, before dusk. We speci cally studied F. acutirostris in the sampling effort (see Bornschein et al. 2015), but in all the areas described, we noted the presence, abundance, and nesting aspects of the other local birds.

Studied Environments
The study areas are estuaries ooded daily by high tides and classi ed as estuarine marshes (Doody 2001

Local Impact and Management of Urochloa arrecta
In the studied areas, there were patches of the alien African species tanner grass Urochloa arrecta (or Brachiaria subquadrippara) as the dominant species (Fig. 1B). Native species co-occur with alien species but at low frequencies. Urochloa arrecta has become a dominant species due to the accumulation of stolons that shade and crush native vegetation, causing the death of native plants ( Fig. 2B; Reinert et al. [2007]). A subtropical salt marsh vegetation structure, with space on the ground and in the vertical column of vegetation for the movement of birds, no longer exists in areas dominated by this alien species ( Fig. 2A; Reinert et al. [2007]). Urochloa arrecta also advances over the water as oating banks of vegetation that are sometimes ripped out by oods and transferred to other areas previously free of its presence (Reinert et al. 2007).
Urochloa arrecta was mechanically managed until its complete local eradication, without the use of herbicides, by clear-cutting vegetation with brush cutters and stacking plant biomass (Bornschein 2013). The stacked biomass was contained with stakes so that the high tides could not move it (Fig. 3). We obtained management permissions from the Instituto Ambiental do Paraná (#357/11) and Instituto Água e Terra (#12.20).
These piles contained biomass mixed up to six times. Biomass from the interior of the pile, already dead, was moved to the edge of the pile, and that from the edge drawn inward toward the interior (Bornschein 2013). Up to six inspections were carried out in the managed areas to remove rooted and sprouted fragments of the alien species, which were pulled out by hand and placed in piles (Bornschein 2013). Urochloa arrecta does not form seed banks locally, and within 10 months of the start of management, the native vegetation covered the land again, free from the presence of alien species (Bornschein 2013). We delimited and measured polygons of areas invaded by U. arrecta and managed them using the Google Earth Pro program (7.3.3.7786) and the advances and reductions in invaded areas shown in the program's historical orbital images.

Results
We recorded Rallus longirostris for the rst time in 2007, observing and recording two individuals vocalizing in duet in a patch dominated by U. arrecta in Riozinho (Patch 1; Table 1; Fig. 4). In 2010, this patch of U. arrecta covered 1.23 ha (Table 1). It was 43 m away from Patch 2 and 80 m away from Patch 3 of U. arrecta, with 0.22 ha and 0.17 ha, respectively (Table 1; Fig. 4B), in which we did not record R. longirostris. This bird became regular in Patch 1 of U. arrecta, where it also nested. We still observed the bird in the vicinity of Patch 1, on subtropical salt marshes free of exotic grass. We even observed R. longirostris on almost every eld day during 2008-2012 in Patch 1, when the tide was low, feeding on small shrimp thrown onto the riverbanks by boat waves. Two photographs on the WikiAves website  Table 1; Fig. 4B). We saw only one couple and listened to them regularly. We observed three nests with eggs in subsequent years on a mass of U. arrecta stolons. We also observed lone individuals or couples of R. longirostris on subtropical salt marshes up to 441 m distant from patches of U. arrecta (in this case, on Jundiaquara Island; 22'14"S, 48°45'32"W). In 2012, we started managing U. arrecta in Patch 1. In 2015, we still recorded R. longirostris in this patch (Table 1), and on that occasion, we observed R. longirostris feeding on Leptuca mordax chased away by the management brush cutters. The bird was nesting on top of a biomass pile of managed vegetation (but the eggs were preyed upon). In 2016, Patch 1 of U. arrecta was practically eliminated, and since then, R. longirostris has not been recorded at this location. Also in 2016, we started to record R. longirostris in a (then) 0.57-ha Patch 2 of U. arrecta (Table 1), but only a few times a year, suggesting either a tiny population or irregular occurrence. In 2021, we eradicated U. arrecta from Patch 2, and since the beginning of 2021, we have not recorded any R. longirostris in it (Table 1). We started managing the (then) 0.16-ha Patch 4 of U. arrecta in 2016 (Table 1). In 2017, we practically eradicated this patch of U. arrecta, and we no longer recorded R. longirostris at the site (Table 1).
In Patch 3 of U. arrecta, which reached 0.23 ha, we recorded no R. longirostris (Table 1). In 2015, this patch of U. arrecta was eradicated. On Folharada Island, where U. arrecta was absent, we also recorded no R. longirostris.

Discussion
This long-term study allowed us to verify the 1) appearance of a population of R. longirostris occupying and nesting in patches dominated by the exotic U. arrecta; 2) increase in its geographic distribution, with the occupation of other patches dominated by this plant; and 3) disappearance after the eradication of the plant. This is a further case of a native species bene ting from an alien species invasion (see Maley and Brum eld 2013) and evidence of a preference for an altered over a natural environment for nesting.
This suggests that i) the subtropical salt marsh structurally impedes the nesting of R. longirostris or that ii) there is an ecological impediment to nesting in this environment not invaded by U. arrecta.
Regarding Hypothesis "i", the most abundant bird at the study sites was the blackish rail Pardirallus nigricans, which is similar in size to R. longirostris (Dunning 2008) and nests in the subtropical salt marshes (MRB per. obs.). Locally, these two birds build nests as baskets made from fragments of native herbaceous plants of similar size, supported over the herbaceous vegetation (n = 4 nests of R. longirostris; n = c. 150 nests of P. nigricans). Thus, subtropical salt marshes are environments in which herbaceous plants support the construction of relatively large bird nests. This, however, leads to the possibility that there could be disputes over reproductive sites between these birds on subtropical salt marshes (Hypothesis ii), with P. nigricans dominating over R. longirostris. We suggest that R. longirostris could have occupied the patches dominated by U. arrecta as vacant nesting niches, eliminating disputes with P. nigricans.
Pardirallus nigricans does not seem to use areas dominated by exotic U. arrecta, possibly due to the high density of vegetation (Fig. 2B), which limits its movements and access to food (MRB per. obs.). This bird also apparently does not nest in patches invaded by exotic species, perhaps due to the reduced structural wren-like rushbird Phleocryptes melanops, the yellow-chinned spinetail Certhiaxis cinnamomeus, the many-colored rush tyrant Tachuris rubrigastra, and the Brazilian tanager Ramphocelus bresilia (MRB per. obs.).
The principle of equal opportunity (MacArthur 1972), which predicts that the occupation of environments by species depends on the relationship between the resources of that habitat and the pressure to use them, meaning that individuals up to a certain point prefer less competitive environments, could be a factor in the choice of use of the environment by R. longirostris and P. nigricans and their consequent nesting in patches of U. arrecta by the former. The occupation of environments depends on competition (Cody 1985), and few species can occupy resource-poor habitats, such as patches of U. arrecta with impoverished ora and a simpli ed vegetation structure, which leads to reduced competition (Cody 1985). Rallus longirostris could have bene ted from occupation of the patches containing the exotic plant. Although R. longirostris and P. nigricans coexist in vegetation free of exotic plants without apparent mutual aggression, even when they are side by side (MRB per. obs.), the similarity between these species suggests competition, which can happen silently, without obvious aggression (MacArthur 1984).
Bene ts were veri ed for the congeneric R. obsoletus following the invasion of a hybrid form of Spartina grass, which increased the survival rate of individuals by offering refuge against predators (Overton et al. 2014   brush cutters and piling up the biomass, which was stacked with bamboo supports to prevent it from being carried away by high tides (since the water usually almost reached the tops of the highest piles).
The managed area was inspected for the manual removal of sprouts of exotic grass up to six times, and the biomass piles were turned over equally up to six times to ensure the death of sprouts. Photograph: Marcos R. Bornschein