The bird records obtained are consistent with previous records (Rossetti and Giraudo 2003; Ronchi Virgolini et al. 2011; Magnano et al. 2019; eBird 2020; Frutos et al. 2020 a, b). Our results show that bird communities present in the levee forests of the Delta and Paraná Islands Ramsar site wetlands are not homogeneous; rather, the assemblages differ in bird community ecological attributes (richness, abundance and diversity).
The maximum shrub cover in IF and herb cover in OF (Fig. 6) might be due to the greater light availability at the lower vegetation levels, reaching the lower forest strata due to the open canopy structure. OF might be considered edge environments favored by abiotic factors, especially light, which generates microclimatic variations that increase plant diversity (Nieves Vele and Orellana Peralta 2018; Granados Sánchez 2006); in turn, these conditions are considered optimal for bird diversity (Baker et al. 2002; Rannestad et al. 2015). It is known that the development of the different vertical vegetation strata is enhanced when a disturbance of the forest canopy increases light availability (Anderson et al. 1978), with solar radiation being a factor determining the natural regeneration of forest vegetation (Viana and Jardim 2013).
In a closed canopy forest, most of the light is absorbed by the tree canopy, reducing the development of the understory vegetation. In these forests, we found the lowest bird abundance value in autumn, and in general, the lowest value of diversity across seasons (Table 2). There may be a relationship among the type of resources offered by the CF, a decrease of those resources in cool seasons and preferences for these resources by the species present there. To explain this relationship, a species composition analysis might be necessary to detect species preferences for different physiognomies (Horlent et al. 2003). Overall, our results show that CF had the lowest richness, abundance and diversity values in their bird communities, whereas the highest richness and diversity values, but not of abundance, were detected in OF (Table 1). Similar results have also been reported for the El Kilombero alluvial plain in Tanzania, the largest wetland in East Africa (Rannestad et al. 2015). Other studies have highlighted the key role of edge habitats (Baker et al. 2002; Miller et al. 2004) in bird community structuring. Considering the levee open forests have the most diverse of bird community, and is located close to the water, we can suggest that they behave as edge environments. This information is relevant to our results and suggests that OF would have characteristics of edge environments.
Intermediate cover forests showed differences only in the abundance of birds in summer, with the lowest value (Table 2). Despite the increased presence of shrubs, which is generally associated with increased bird richness (Martin 1993; Godoi et al. 2017), this pattern was not evident.
The recent history of livestock production activity in the ISFNP (https://www.argentina.gob.ar/parquesnacionales) and further livestock removal may have favored the expansion of shrubs, generating higher cover values in this site. However, no differences in vegetation variables (Fig. 5) were found for any of the plant cover strata between sites. By contrast, there were seasonal differences possibly related to abiotic/climatic factors, in the cover of the lower strata (Zerda et al. 2010).
Results obtained in bird assemblages support the structural variation present inside the different levee forests. The overall environmental heterogeneity of these forests is mainly influenced by the flood pulses (Miller et al. 2004; Carvalho et al. 2013; Lorenzón et al. 2019). Such heterogeneity has been associated with the increase or decrease in richness, abundance or diversity of birds among homogeneous environmental units of the mosaic of the floodplain environment (Robledano Aymerich et al. 1992; Ronchi Virgolini et al. 2010; Lorenzón et al. 2016). We suggest that the effects of spatial heterogeneity within the forest unit operate in a similar way as it occurs at a greater spatial scale. Therefore, we propose that these dynamics are repeated at a smaller habitat scale, i.e. microhabitat differences for birds among physiognomic variables are generated within the forest environment.
In a highly heterogeneous environment, niche availability is high; therefore, species richness is high while evenness is low, reducing diversity (Rosenzweig 1995; Symonds and Johnson 2008). Other authors argue that a greater niche availability drives greater evenness (Cotgreave and Harvey 1994). However, community structure depends on landscape structure and spatial scale (Drobner et al. 1998; Horlent et al. 2003). Based on our richness-abundance relationship analisys, we could suggest that more heterogenous ambient could host more diversity, independent on the evenness of bird community. An example of this is the richness record in OF, where ambients heterogeneity is evident with differents of vegetation strata. However, OF were not always able to differentiate themselves with a higher richness value in all seasons. Our results show the existence of dynamism in the arrangement of the community, which is not easily identifiable, doe by influence of a combination of factors, like seasonality and microspatial scale.
The differential use of microhabitats has been mentioned for other taxa, such as beetles which use wood remains (Anderson et al. 1978, Stain et al. 2014), fish and invertebrates, macrophytes and plancton community dynamic in lotic environments associated with levees (Agostinho et al. 2000; Tockner et al. 2000 a, b; Thomaz et al. 2007). The same pattern has been described for density and composition of flying insects associated with floodplains (Langhans 2005) and for bird assemblages between environmental units associated with islands (Miller et al. 2004; Hamza et al. 2015; Magnano et al. 2019).
At the global scale, riparian forests are affected by natural disturbances characteristic of the fluvial environments, along with the effect of anthropogenic activities in the basins, such as agroindustry, dams, hunting activities or livestock production, contributing to heterogeneity of these environments at different scales: landscape and microhabitat (Jansen and Robertson 2001; Tockner et al. 2010; Sica et al. 2016). It has been demonstrated that those activities generate habitat fragmentation, loss of gallery forests and redistribution of animal species (Briceño Vanegas 2015; Schondube et al. 2018). Some authors claim that the heterogeneity contributed by those activities favors the increase of environmental diversity through species turnover (Bó et al. 2010; Frutos et al. 2020 a, b), or that knowledge about this topic is still scarce (Mitchell et al. 2010); other authors suggest the importance of maintaining the interior of a single environmental unit in riparian forests as heterogeneous as possible to ensure the increase of richness of species with different habitat requirements (Magnano et al. 2019). Our results show no differences in the vegetation structure between the study sites, despite their different management histories. However, we found differences in the arrangement of birds associated with the sites. This result suggests that, despite the past anthropogenic activities in the study sites, spatial heterogeneity in the floodplains is still high, possibly due to the dynamics contributed by the pulses. Thus, anthropogenic production activities in the natural riparian environments in some cases might be considered to favor bird diversity (Frutos et al. 2020 a, b); however, they are not a necessary requirement to maintain a heterogeneous and beneficial structure for the bird dynamics in riparian environments.
Beta diversity and species turnover from one area to another have a substantial influence on general species richness in a region (Vellend 2001; Ronchi-Virgolini et al. 2011). The variation of the assemblage between sites (Table 2), independently of the variation contributed by the physiognomies of the forest, could be related to the floristic composition. This variation would provide specific perches, shelters, or feeding sites for the birds and influence the decision to choose one site over another. But the seasonal differences in bird assemblage between sites also might be influenced by the migratory effect. Seasonal variations in the bird assemblages, caused by migrations, have been widely studied (Capllonch et al. 2008; Carvalho et al. 2013; Capllonch 2018). The studied wetlands are located in the biological fluvial corridor of the Parana River. The relative position of the national parks studied in the great corridor and their proximity to migratory routes might be additional variables that influence the preference for one site or another and, therefore, in the arrangement of birds community between sites.
In bird assemblages of highly heterogeneous riparian environments, seasonal changes are very evident (Ronchi-Virgolini et al. 2013; Lorenzón et al. 2019), which may explain the high seasonal variation of our results, regardless of microspatial variation. However it should be noted that the migratory condition would not allow us to predict habitat selection (Horlent et al. 2003); therefore, the effect of migration on forest preference would rather be random.
The variation detected within levee forests shows that a deeper analysis is needed for differences in composition among vertical strata of the understory (De Stefano et al. 2012). Thus, the differential niche offer that optimizes resources for birds could be more clearly evidenced, considering that the turnover rate in plant composition is high in these environments. Therefore, deeper analyses of the differences in the bird assemblage using a functional classification criterion, i.e., guild, may reveal subtle aspects that are poorly detectable from a general perspective (Root 1967; Milesi et al. 2002; Farías et al. 2007; Ronchi-Virgolini et al. 2011). Guild-based analyses would contribute more specific information to properly guide conservation efforts in these wetland environments, where variation at different scales are clear.