Species composition, community and population dynamics of two gallery forests from the Brazilian Cerrado domain

Abstract Background To understand the impacts of global changes on future community compositions, knowledge of community dynamics is of crucial importance. To improve our knowledge of community composition, biomass stock and maintenance of gallery forests in the Brazilian Cerrado, we provide two datasets from the 0.5 ha Corrego Fazendinha Gallery Forest Dynamics Plot and the Corrego Fundo Gallery Forest Dynamics Plot situated in the Bom Despacho region, Minas Gerais, Southeastern Brazil. New information We report diameter at breast height, basal area and height measurements of 3417 trees and treelets identified during three censuses in both areas.


Introduction
Although the Brazilian Cerrado is a hotspot of biodiversity (Mendonça et al. 2008, Myers et al. 2000 and holds carbon stocks of nearly 300 Mg per hectare (Batlle-Bayer et al. 2010, Paiva et al. 2011, its species richness, diversity and biomass are still threatened by habitat loss, fragmentation, biological invasion and climate change (e.g., Lapola et al. 2013, Jantz et al. 2015, Rossi et al. 2014. Within the Cerrado domain, gallery forests accompany the borders of rivers, creeks and streams, forming important corridors for wildlife among patches of remaining vegetation (Silveira et al. 2014) that also protect aquatic ecosystems from substrate input, reducing water temperatures and erosion of river banks (Monteiro et al. 2016, Londe andSilva 2014). Furthermore, gallery forests have the highest above ground biomass per hectare in the Cerrado domain (Moreira-Burger and Delitti 1999). Worldwide, these forests are threatened by human activities, including domestic livestock, which prevent tree seedling establishment, and the construction of dams and weirs, which cause flooding or interference with natural stream flow (FAO 2010). Long-term monitoring studies, so-called community dynamics, are necessary to outline and understand the impacts of these disturbances on vegetation communities and on carbon stocks (Couvet et al. 2011, Fidelis et al. 2012, Pocock et al. 2015.
Therefore, the aim of this data paper is to make available data from forest dynamics from two gallery forest dynamics plots from the Bom Despacho region, Minas Gerais, Southeastern Brazil, to increase knowledge about community composition, biomass stock and maintenance of such forests in the Brazilian Cerrado.
Within the municipality, two study sites were selected within properties owned by the ArcelorMittal Bioflorestas company. The Corrego Fazendinho Gallery Forest, situated 5.5 km west of Quartel Geral center, covers approximately 50 ha on both sides of the upper 5 km of Fazendinho Creek (Fig. 1). It is surrounded on all sides by eucalypt plantations from ArcelorMittal Bioflorestas.
The second study site, the Corrego Fundo Gallery Forest, is situated approximately 10 km southeast of the Corrego Fazendinho Gallery Forest (Fig. 1). It is a forest remnant that flanks the complete upper Fundo Creek. The mean width of the gallery forest is approximately 80 m. On its northern side, the forest adjoins native Cerrado vegetation belonging to the ArcelorMittal Bioflorestas legal reserve, while cattle pastures are found beyond its southern limit.

Sampling methods
Sampling description: Within each of the gallery forests, two plots of 50 x 50 m were delimitated and divided into 25 subplots of 10 × 10 m (24 plots in the second plot from the Corrego Fundo Gallery Forest). All plots are situated at the northern part of the gallery forests (see Fig. 1).
Within these plots, three censuses in four-year intervals of all trees with a diameter at breast height (dbh) greater than 3.2 cm were carried out (Table 1). Trees fulfilling the inclusion criterion were tagged and identified. Tree diameter (dbh) was measured and

Localization of study sites
Species composition, community and population dynamics of two gallery forests ... basal area was calculated; for multiple stem individuals, we calculated basal area at breast height for all shoots, summed these, and calculated from that the pooled dbh. Specimens not recognized during fieldwork were collected, deposited in the Herbarium of the Federal University of Viçosa (VIC) and identified with the help of material from the VIC or by consultation of specialists and literature sources (Lorenzi 1992). Species names were verified using the Taxonomic Name Resolution Service (TNRS) proposed by Boyle et al. (2013); species classification follows the Angiosperm Phylogeny Group III guidelines (APG III 2009).
Diversity indices as well as Jaccard similarity between study sites were computed using EstimateS (Colwell and Coddington 1994). Mortality and recruitment rates, as well as gains and losses of the basal area, were calculated according to Sheil (1995).

Geographic coverage
Description: This study was carried out in the Counties Quartel Geral and Dores do Indaía, Bom Despacho region, Minas Gerais, Brazil (Fig. 1).
Gallery forest plots census histories. BA is basal area.

Taxonomic coverage
Description: Altogether, 3413 trees and treelets belonging to 158 species, 96 genera and 41 families were detected in both study sites during all censuses. Thirty species were identified to genus level only, two to family level, and three species remain unidentified.
With a total of 1862 trees and treelets from 114 species (70 genera, 35 families) from three censuses, species richness and stem density in the Corrego Fazendinha Gallery Forest was higher than in the Corrego Fundo Gallery Forest (1551 trees and treelets, 89 species, 67 genera, 35 families, Table 1). Forty-five species occur in both study sites, yielding a Jaccard similarity between the study sites of 0.28. More than 75% of species and around 80% of basal area belongs to common species.
Due to higher species richness, diversity is also higher in the Corrego Fazendinha Gallery Forest than in the Corrego Fundo Gallery Forest. While the basal area increased from the first to the third census in the Corrego Fazendinha Gallery Forest, it declined in the Corrego Fundo Gallery Forest (Tables 1, 2).      Table 7.
Corrego Fazendinha Gallery Forest ranking by species accoridng to basal area (BA) and number of individuals (N), data from the third census.  Table 8.
Corrego Fundo Gallery Forest ranking by species according to basal area (BA) and number of individuals (N), data from the third census. The recruitment rate in the Corrego Fazendinha Gallery Forest exceeded the mortality rate during 2007 and 2011; but mortality was higher than recruitment in the period from 2011 to 2015 (Table 9). In the Corrego Fundo Gallery Forest, mortality exceeded recruitment during both observed periods. Further, gains of basal area were higher than losses in the Corrego Fazendinha Gallery Forest, indicating an increase in carbon stock, while losses in the Corrego Fundo Gallery Forest outpaced its gains (Table 9).  Mendonça Machado and de Oliveira-Filho 2010), differences in successional stages (Magurran 2011) or neutral factors such as ecological drift and stochasticity (Hubbell 2001). Large numbers of individuals and basal area belonging to common species indicates that the high beta-diversity is due to large number of species represented by few individuals only, which is typical for tropical forests (Condit 2000). High beta-diversity increases the importance for the protection of biotic ressources and highlights the demand for further research to understand underlying determinants.
Although study sites were sampled three times during similar periods, forest dynamics show large differences between study sites. Mortality and recruitment rates between 1.5 and 3 % are within the expectations for undisturbed alluvial or gallery forests (Higuchi et al. 2008, Fontes andTeles Walter 2011). Causes for the elevated mortality rate during the second observation period in Corrego Fazendinha Gallery Forest remain unknown, as external disturbances were not registered during the field campaigns, but may be related to extreme water deficits between 2012 and 2015 in Brazil (Getirana 2016). These findings indicate the importance to give once continuity and to amplify these kind of studies, to come to a better understanding of the drivers of forest dynamics that influence the maintanence of biodiversity as well as that ecosystem services such as carbon sequestration in biomass.