Trees and shrubs are invading natural grassland ecosystems around the world, and woody plant encroachment has been reported as a driver of declines in grassland-dependent birds (Andersen and Steidl, 2023; Coppedge et al., 2004; Lautenbach et al., 2020; Silber et al., 2024). Although these declines are often attributed to habitat loss from land-use change, some grassland-dependent species are disappearing even within protected areas. Legally protected areas of native grassland in the extreme south of Brazil (law no. 12.727/2012; Brancalion et al., 2016) generally have their management practices suppressed (i.e., grazing and fire). This has resulted in the gradual encroachment of shrubs, with the predominance of some species such as Baccharis uncinella and B. dracunculifolia (Blanco et al., 2014; Boldrini, 2009; Oliveira and Pillar, 2004). These species are typical of southern grasslands and characterize the initial stages of invasion by woody vegetation in grasslands in the absence of disturbances, such as grazing and fire (Oliveira and Pillar, 2004). This process is particularly concerning in the Brazilian Pampa, where protected areas cover only about 3% of the biome, with just 0.54% under strict protection, highlighting the limited extent of effective conservation and the importance of appropriate management to maintain grassland structure and biodiversity (Overbeck et al., 2023; Brasil, 2024).

Woody encroachment can generate profound structural changes in grassland habitats, with important consequences for bird communities. Increased shrub cover has been associated with higher rates of nest predation, as woody plants provide perches and movement corridors for predators such as birds of prey and snakes (Graves et al., 2010; Klug et al., 2010; Mason et al., 2005; With, 1994). In addition, denser or taller vegetation can attract mammals seeking shelter or shade, further increasing predation pressure. These structural changes also reduce the availability of resources required by grassland specialists. For instance, the regionally endangered Culicivora caudacuta relies heavily on grass inflorescences for nest construction (Jacoboski et al., 2023; Silva and Silva, 2021), making the progressive loss of herbaceous vegetation particularly detrimental. Over time, woody encroachment can therefore reduce habitat quality, survival, and reproductive success of grassland-dependent birds (Archer et al., 2017; Sirami and Monadjem, 2012).

In this sense, when conducted with appropriate frequency and intensity, management practices in grassland ecosystems, such as the use of prescribed fire and grazing, are essential for the conservation of grassland biodiversity (Andrade et al., 2015; Fontana et al., 2016; Jacoboski et al., 2017; Veldman et al., 2015). The mechanical removal of woody vegetation is a management method that has been used to restore several ecosystems around the world (Hamilton et al., 2004), since the encroachment of woody vegetation has become one of the most widespread processes of grassland habitat degradation due to changes in management regimes and climate change (Archer et al., 2017; Stevens et al., 2017). This removal method is used to promote the restoration of vegetation to levels close to its original state with the aim of improving habitat quality and conserving species dependent on such environments (Archer et al., 2017; Hamilton et al., 2004). However, because it is a mechanical method, the frequency, intensity, and scale of this type of intervention must be carefully evaluated. Different woody vegetation removal methods can influence grassland birds through their distinct effects on vegetation structure, soil disturbance, and patterns of regrowth. Manual removal tends to eliminate individual shrubs while leaving the surrounding herbaceous layer relatively intact. However, because this method acts locally and generates fine-scale gaps, woody species can recolonize rapidly, reducing structural heterogeneity and potentially disadvantaging species that rely on a mosaic of open microhabitats (Codesido et al., 2009). In contrast, low-intensity mechanized removal clears shrubs more uniformly and over larger patches, slowing woody regrowth and temporarily expanding open areas and herbaceous resources, conditions that typically favor grassland specialists (Coria et al., 2015).

Empirical studies reflect these contrasts. In Argentina, different removal intensities produced markedly different responses in bird communities: manual shrub removal led to declines in bird richness and abundance (Codesido et al., 2009), whereas low-intensity mechanical cutting maintained bird diversity (Coria et al., 2015). Likewise, selective mechanical cutting did not affect overall bird diversity but increased the abundance of the endangered Gubernatrix cristata, demonstrating that woody vegetation removal, when carefully implemented, can enhance habitat conditions for species of conservation concern (Rebollo et al., 2025).

Such studies, however, have only assessed species diversity in general terms, without assessing the responses of bird species according to their relationship to vegetation type. Therefore, we aimed to evaluate the response of the bird community of a legally protected grassland area in southern Brazil to the removal of woody vegetation by evaluating grassland-dependent and woody vegetation-dependent bird species separately. In doing so, we provide the first attempt to evaluate the effect of woody vegetation removal on grassland birds in southern South America. We addressed two main hypotheses. First, there will be an increase in grassland bird richness and abundance after woody vegetation removal, given that there would be an increase in the amount of habitat available for grassland bird species (Andersen and Steidl, 2023). Furthermore, the opposite will occur with birds associated with woody vegetation, with a reduction in richness and abundance after removal. Second, there will be changes in species composition between sampling periods, since both the absence of management and the suppression of vegetation will alter the structure of the local vegetation, which also affects bird occurrence.

A total of 55 bird species was recorded during the sampling. Of these, 14 are grassland-dependent species, 17 are associated with shrubland, and 19 are forest-dependent (Azpiroz et al., 2012; Stotz et al., 1996) (Table 1). Zonotrichia capensis was the only species recorded in all sampling periods; it is a common species typical of shrubland environments, occurring in open areas with secondary vegetation. Eighteen species were recorded in only one sampling period, most of which are associated with forest vegetation, such as Cyanocorax caeruleus and Turdus albicollis. The fact that these species were recorded only once indicates that many of these species do not actually use the area, but just perch on trees while moving through, as observed several times during the sampling.

Notably, four grassland bird species (sensuAzpiroz et al., 2012), Ammodramus humeralis, Emberizoides herbicola, Rhynchotus rufescens and Sicalis luteola, that had been recorded four years before the removal of woody vegetation, but not recorded the year prior to removal, were recorded again after the removal.

Grassland bird species richness differed significantly between before and after woody vegetation removal (Table S1). Initially, there was a significant reduction in grassland bird richness one year before removal (1YBWVR) compared to the first sampling period (4YBWVR). Subsequent to vegetation removal, grassland species richness increased significantly and continued to do so until one year after removal (1YAWVR). At two years after removal (2YAWVR), however, grassland bird richness decreased significantly to values similar to pre-removal. Forest-dependent bird species richness showed the opposite behavior, increasing significantly one year before removal (1YBWVR), then decreasing after removal and increasing significantly again two years after removal (2YAWVR) (Fig. 1, Table S3). Shrub-associated bird species richness did not differ significantly among sampling periods (Table S2).

Fig. 1.

Boxplots demonstrating variations in bird species richness according to vegetation and sampling period. Different letters indicate significant differences between sampling periods. The dashed red line indicates the time of woody vegetation removal. Caption: 4YBWVR (four years before woody vegetation removal), 1YBWVR (one year before woody vegetation removal), 1MAWVR (one month after woody vegetation removal), 2MAWVR (two months after woody vegetation removal), 3MAWVR (three months after woody vegetation removal), 4MAWVR (four months after woody vegetation removal), 1YAWVR (one year after woody vegetation removal), 2YAWVR (two years after woody vegetation removal).

Results for species abundance showed a trend similar to that for the results for species richness. There was a significant increase in the abundance of grassland birds immediately after removal (1MAWVR), which continued until a significant decrease two years after woody removal (2YAWVR) (Table S4). Shrub-associated bird species abundance decreased significantly only during the first month following woody vegetation removal, with no significant differences observed in subsequent sampling periods (Table S5). The abundance of forest-dependent bird species increased significantly one year before woody vegetation removal (1YBWVR), then decreased significantly in the first three months after removal followed by a significant increase again two years after removal (1YAWVR), reaching values similar to those found one year before removal (Fig. 2, Table S6).

Fig. 2.

Boxplots demonstrating variations in the abundance of bird species according to their relationship with vegetation and sampling period. Different letters indicate significant differences between sampling periods. The dashed red line indicates the time of woody vegetation removal. Caption: 4YBWVR (four years before woody vegetation removal), 1YBWVR (one year before woody vegetation removal), 1MAWVR (one month after woody vegetation removal), 2MAWVR (two months after woody vegetation removal), 3MAWVR (three months after woody vegetation removal), 4MAWVR (four months after woody vegetation removal), 1YAWVR (one year after woody vegetation removal), 2YAWVR (two years after woody vegetation removal).

No significant differences in species composition were found among sampling periods for any of the vegetation-dependent bird groups, except for grassland birds, for which the result was marginally significant (PERMANOVA, F(6,14) = 1.56, p = 0.059). In the NMDS, it is possible to observe that the periods 3MAWVR and 4MAWVR were the most distinct, indicating that species composition changed during these periods, likely due to a higher availability of resources for grassland birds. (Fig. 3A). For birds associated with shrub and forest vegetation, no significant differences were found (PERMANOVA, F(7,16) = 1.02, p = 0.444 and PERMANOVA, F(7,13) = 1.14, p = 0.272, respectively).

Fig. 3.

Nonmetric multidimensional scaling (NMDS) based on Bray-Curtis dissimilarity matrix, representing the composition of bird communities between the different sampling periods. Each point represents a point-count and the colors and symbols indicate the different periods. The polygons delimit the groupings of samples from each period, closer points indicate more similar communities. (A) Grassland birds; (B) Shrub birds; (C) Forest birds. Caption: 4YBWVR (four years before woody vegetation removal), 1YBWVR (one year before woody vegetation removal), 1MAWVR (one month after woody vegetation removal), 2MAWVR (two months after woody vegetation removal), 3MAWVR (three months after woody vegetation removal), 4MAWVR (four months after woody vegetation removal), 1YAWVR (one year after woody vegetation removal), 2YAWVR (two years after woody vegetation removal).

Consistent with our first hypothesis, grassland bird richness and abundance increased following woody vegetation removal. Conversely, species associated with woody vegetation showed a decline, confirming the expected trade-off. This results indicate that the positive responses of grassland birds were transient: richness and abundance increased up to one year after removal but declined significantly two years later. This temporal pattern suggests an immediate but short-lived effect of management, likely driven by the rapid regeneration of woody vegetation, which reduces the availability of open-habitat structure. Together, these findings highlight both the effectiveness and the limitations of woody vegetation removal as a management strategy for grassland bird communities.

After removal, some portions of the cleared area already showed a significant increase in shrub density two years after the removal of woody vegetation, reaching levels similar to that recorded prior to removal. This suggests an accelerated woody encroachment process in the study area, possibly driven by the proximity of source areas for these species. In this case, the increase in shrub density improves the habitat for woody-dependent birds, since the increased density of woody plants makes new foraging opportunities, hiding places, and nesting sites available (Andersen and Steidl, 2019; Archer et al., 2017; Coppedge et al., 2004).

Aligned with our second hypothesis, temporal changes in species composition demonstrate that woody vegetation removal reshapes not only community structure but also the identity of species occupying the experimental area at different successional stages. This pattern is consistent with BealNeves et al. (2020), who showed that grassland bird communities vary according to the time elapsed since the last fire, with the highest richness and diversity occurring under intermediate post-fire intervals, when habitat structural heterogeneity is greatest. Similarly, woody vegetation removal operates as a disturbance that resets vegetation structure, opening a temporal window in which grassland species encounter favorable conditions for foraging, reproduction, and shelter before woody vegetation begins to regenerate.

Our results indicate that targeted reductions in woody vegetation can effectively reestablish grassland habitat conditions and promote favorable responses among grassland‐dependent bird species. This pattern aligns with experimental evidence demonstrating that decreases in woody cover improve habitat quality and demographic performance in grassland bird communities. For example, Thompson et al. (2016) reported positive, although delayed, responses to large-scale tree removal, underscoring the importance of temporal scale in post-treatment dynamics. Similarly, experimental removal of tree rows and shelterbelts has led to increased grassland bird densities and nesting activity (Ellison et al., 2013; Tack et al., 2017). Together, these studies highlight that the scale and timing of interventions, as well as the need for recurrent management to prevent rapid regrowth, are key factors driving the magnitude and persistence of avian responses (Coffman et al., 2014; Thompson et al., 2016). In line with this evidence, we observed an initial increase in grassland species richness and activity immediately after removal, followed by a decline associated with accelerated shrub regeneration and the absence of additional management actions.

The broader literature, together with our findings, indicates that strategies aimed at reducing woody vegetation can enhance both the structural and functional quality of grassland habitats (Andersen and Steidl, 2023; Silber et al., 2024). Field observations corroborated this pattern: three months after removal, we recorded high diversity of fruiting grasses, and species such as Emberizoides herbicola, which bred in the area, and Sicalis luteola, observed foraging in small flocks, returned to the site. These responses suggest improved resource availability and habitat suitability. Enhanced resource conditions may also contribute to higher reproductive success and reduced predation risk (Archer et al., 2017; Jacoboski et al., 2023; Sirami and Monadjem, 2012), particularly given that proximity to woody vegetation can negatively affect nest survival for some species (Andersen and Steidl, 2023; Graves et al., 2010).

It is important to highlight that the study area is embedded in a grassland-dominated matrix that includes unmanaged patches, particularly within legally protected areas. These unmanaged areas are subject to strong and continuous shrub encroachment pressure. Consequently, part of the temporal variation observed before and after vegetation removal reflects landscape-scale processes operating in the surroundings, especially the rapid advance of woody vegetation in unmanaged grassland areas. Considering the estimates by Sühs et al. (2020) that unmanaged grassland vegetation in high-altitude grasslands in southern Brazil will be completely replaced by woody vegetation within 30 years, the removal of this type of vegetation should be considered a priority in conservation strategies for grassland bird species in areas where shrubland is dense (Silber et al., 2024), and especially in areas where threatened species occur. Although the present study was conducted on a small scale, the results demonstrate that woody vegetation removal benefits grassland bird species due to an increase in habitat availability and a greater supply of resources, although this was not measured directly. However, the declining trend observed two years after vegetation removal, both in richness and abundance, demonstrates that specific actions are not sufficient to maintain the grassland bird community for the long term. Continuous or cyclical management is necessary to prevent the reestablishment of woody vegetation. Maintaining a minimum level of vegetation disturbance is essential to maintaining the herbaceous-grassland component of vegetation, which benefits grassland-dependent birds (Fontana et al., 2016; Jacoboski et al., 2017, 2022). Therefore, we recommend that, in addition to managing woody vegetation in protected areas within grassland ecosystems, this management should be carried out continuously and at appropriate intensities and scales to allow the maintenance of vegetation mosaics in the landscape to improve habitat for grassland birds and maximize species diversity.