The density of species sampled (1,323 individuals ha− 1) was higher than in a 9-year-old restored area with exotic and native species in Mogi-Guaçu, State of São Paulo (809 individuals ha− 1) (Colmanetti and Barbosa, 2013). The richness of species sampled was higher than those of surveys in forest remnants of the Seasonal Semideciduous Forest in the Zona da Mata region (Prado Júnior et al., 2010) and in a fragment of a Legal Reserve (Lopes et al., 2011), both in the Minas Gerais state with 73 and 86 species, respectively. The density of individuals and richness of species indicate successful restoration as found in restored areas and/or preserved fragments of Seasonal Semideciduous Forest in the same region. This is due to the increase in the number of species over the course of secondary succession in tropical forests by partition of resources between a great number of species (Aidar et al., 2001; Rodrigues et al., 2004). This improves nutrient cycling, soil fertility and shading, following a successional facilitation model (Connel and Slatyer, 1977). The recovery of diversity in secondary forests in Zona da Mata follows that of other studies (Guariguata and Ostertag, 2001; Martin et al., 2013; Rozendaal et al., 2019), showing a similar pattern between secondary forest plots and propagation sources. In addition, the higher species richness in older secondary than in primary forests is due to the coexistence of P and LS species (Bongers et al., 2009; Huston and Smith, 1987). This indicator of tree species richness in planted areas was similar to the reference areas present in the surroundings, showing that efforts have been made so that naturally diverse areas are restored from a large set of species, which is more efficient for permanent forest restoration (Rodrigues et al., 2009; Gardon et al., 2020). However, at least 20 years may be necessary for the recovery of species richness in areas of assisted and unassisted natural regeneration. This coincides with the recovery of the richness of animal species in abandoned tropical areas, showing a strong relationship between flora and fauna (Dunn, 2004; Londe et al., 2020).
The predominance of species with the Zoo dispersion syndrome followed by Ane and Auto syndromes is important in the maintenance, spatial distribution and frequency of species (Tabarelli et al., 2002; Talora and Morellato, 2000) as found in two Semideciduous Seasonal Forest remnant with Zoo, Ane and Auto dispersion of 78%, 20%, and 2% (Lopes et al., 2011) and 68, 22 and 9%, respectively (Prado Junior et al., 2012). The predominance of species with Zoo dispersion syndrome showed the presence of dispersing animals and availability of food and shelter (Miranda Neto et al., 2012) reduction in the number of native species in the 45-year-old restored area was possibly due to the most of tree species belonged to the successional class ST, which need specific environmental conditions such as more shaded areas (Turchetto et al., 2017). The individuals planted, mostly IS species, influenced the succession and indicates that the forest is heading towards an intermediate successional stage, due to the lower percentage of P species (Silva et al., 2004). The largest number of Zoo forest species reveals plant communities in advanced stages of succession in good condition (Mazer and Wheelwright, 1993), ensuring greater protection and supply of resources for animals. High values of Zoo syndrome indicate an area closer to naturalness, confirming an advanced stage of restoration (Londe et al., 2020). This suggests that adequate proportions of this group of plants have been used in the restored area and that there are also zoochoric dispersal species in natural regeneration areas. High values for zoo-dispersal tree groups show the progression of ecological succession, the participation of local species and the attraction of fauna, contributing to the diversity of species and accelerating the regeneration of native vegetation (Wunderle, 1997).
The high values of H 'and J' (3.46 and 0.74, respectively) agrees with results for other areas in Minas Gerais state with greater diversity than that of a Deciduous Forest in Montes Claros state, Minas Gerais (H'= 3.3) (Neri et al., 2007; Santos et al., 2007) and lower than in a Semideciduous Seasonal Forest in Viçosa, Minas Gerais (H'= 4.25) (Marangon et al., 2007). The estimated values of the H ’and J’ indices varies with factors such as the plot size and the inclusion criterion for the succession of vegetation and the sampling method (Higuchi et al., 2012), as well as the different degrees of naturalness of the forests in the various study sites (Tang et al., 2020). The high J’ was similar to that in a well-preserved vegetation fragment in a Seasonal Semideciduous Forest representative of the original condition in the Uberaba, Minas Gerais state (J '= 0.73) (Dias Neto et al., 2009) and higher than that of restored areas (J'= 0.89) (Colmanetti and Barbosa, 2013). The floristic results in the Semideciduous Seasonal Forest area were similar to the present study but factors such as succession stage, degree of disturbance and site characteristics can affect them (Gonçalves and Souza, 2014). The highest VIs of G. guidonia and A. cunninghamiana was due to the diametric distribution in the inverted J-pattern, i.e., a large number of individuals in the smaller diameter classes and that G. Guidonia being an LS species and well adapted to shaded environments (Lorenzi, 2002). The current intermediate-advanced succession stage of the study area allowed the propagation and development of this species, which did not occur in the initial successional stage after planting. In addition, G. guidonia takes longer to enter reproductive age because it is an IS (Lorenzi, 2002). Archontophoenix cunninghamiana is an exotic and invasive species in the understory of forest fragments or mature restored forests and this palm coexists with other native species tolerant to shade (Dislich et al., 2002; Miranda Neto et al., 2012). The presence of this species in a high-density value in a few plots makes an imbalance distribution of this species in the area. The lower number of shade-tolerant exotic species compared to that of heliophytes increases the impact of the first species threatening best-preserved environments and the conservation value (Martin et al., 2009).
The average height of arboreal individuals was greater than that of a forest fragment in good conservation stage in the Minas Gerais state (Campos et al. 2006). The average tree height greater than 12 m represents forests in advanced succession stage in the Atlantic Forest of the Minas Gerais state (CONAMA Resolution 392/2007) and its structural diversity shows a succession stage as found for primary forests. This is due to the establishment of multilayer strata at early stages of forest vegetation development, allowing the forest system to efficiently use nutrients, light, water and other resources (Danescu et al., 2016; Potzelsberger and Hasenauer, 2015). The basal area was higher than that of fragments of seasonal semi-deciduous forest in the region (Braga et al., 2011), with 20.01 m² ha− 1 and initial forest with 23 m² ha− 1 (Colmanetti and Barbosa, 2013). The increase in the basal during the succession did not reach the values found in primary forests. The presence of very large and probably centennial trees in primary forests can explain its greater variation of the basal area compared to old secondary forests (Lucas-borja et al., 2016), contributing to an increasing basal area. The basal area may indicate success in forest restoration, despite the age difference of the plantations (Braga et al., 2011; Colmanetti and Barbosa, 2013).