Pedoforms Microclimate and Seasonal Forest Structure in Médio Vale do Paraíba

In the Rio de Janeiro State, the Atlantic Forest is restricted to small fragments, responsible for the maintenance of environmental services. These fragments occur on steep slopes dominated by concave and convex pedoforms, capable of changing the local microclimate conditions. To preserve these communities, it is necessary to determine the level of influence of these pedoforms on forests structure and species distribution. Thus, we performed a phytosociological study and installed a set of four sensors to determine the monthly mean of air temperature and relative humidity in concave and concave pedoforms. The forest structure and the microclimate averages are compared by the Levene’s T-test. We observed differences between microclimatic average, importance value indexes (IVIs) and total height, indicating that microclime variations can influence the structure and composition of the tree community. However, these structure differences are punctual and insufficient to determine specific communities.


INTRODUCTION AND OBJECTIVES
In the Valley of Paraíba do Sul, a region located between the Mantiqueira Mountains at the north and the Mar Mountains at the southeast (Almeida, 1964), the Atlantic Forest was submitted to few interventions until the early 19th century (Dean, 1996), when the fertility of the soils in the region boosted the conversion of forests into coffee crops (Dean, 1996). As this region has a sloping relief where mounds, hillocks, and hills predominate with convex, concave and linear pedoforms (Silva, 2008), the removal of the forest cover led to fast soil degradation, which resulted in the decline of the coffee production in a few decades. With the collapse of the productive system, crops were gradually replaced with pastures, which hindered natural regeneration and worsened erosion in the region. Currently, it is possible to observe the silting of the surface water system, erosion of soil surface horizons, and the reduction of their productive capacity, as well as habitat fragmentation. Studies that determine the conservation status of the forest fragments in the region are essential for the establishment of actions for recovery and preservation.
In this context, phytosociological inventories are important tools as they allow the analysis, interpretation, and description of the plant community structure. They also provide knowledge on species richness and diversity (Felfili et al., 2011). In spite of their applications, most of the phytosociological studies in the region are merely descriptive (Carvalho et al., 2007;Cysneiros et al., 2015;Dan et al., 2010;Ivanauskas et al., 1999;Ivanauskas et al., 2000;Medeiros et al., 2016). Therefore, the influence of environmental variables on the dynamics of ecosystems should be further discussed, especially about the structure and floristic composition of Semideciduous Seasonal Forest arboreal communities.
Our hypotheses are that geomorphological variations, such as the different concave and convex pedoforms, have different microclimate conditions, capable of influencing the floristic composition and the structuring of the forest communities in the Paraíba do Sul Valley, being the verification of this phenomenon important for understanding ecological relations and applications of resources in future forest restoration projects.

MATERIALS AND METHODS
The study area is located in the municipality of Pinheiral, Rio de Janeiro State, in Médio Vale do Paraíba do Sul, Southeastern Brazil. The studied hillsides are located between the coordinates 22°33'S/44°01'W and 22°32'S/44°02'W in a fragment locally known as Mata do Peixoto (with 204 ha), which belongs to the Bom Sucesso Farm. The original vegetation cover is located in a transition between submontane and montane semideciduous seasonal forests (IBGE, 2013;Oliveira-Filho & Fontes, 2000). The relief of the region is wavy to strongly wavy, with altitudes varying from 360 to 720 m.a.s.l., with a predominance of steep hillsides, where linear, concave, and convex pedoforms occur (Casseti, 2005). The regional climate is classified (Köppen, 1948) as Cwa -temperate climate with dry winter and rainy summer and Am -rainy tropical climate with dry winter.
To determine the temperature (°C) and relative humidity (%) in each pedoform, four HOBO U12 Temp/RH/Light digital thermo-hygrometer sensors were installed, two apparatuses in each feature of the pedoforms, to increase local representativeness and consistent statistical data. The devices were installed in polyvinyl chloride (PVC) supports (one meter high), where they were fixed with the help of plastic clamps. In the upper portion of the structure was installed a 25 × 25 cm translucent glass slide, which protected the unit from direct rain. With the help of the HOBOware software, the devices were programmed to register the variables cited at noon every day, when the pedoforms receive solar radiation simultaneously. Data collection was performed monthly, totaling 60 monthly samples per pedoform.
For the phytosociological inventory, we selected concave and convex pedoforms. In each pedoform set, we allocated 54 plots (100 m² each), in a total of 0.54 ha sampled in each type of pedoforms, where we measured individual trees with DBH ≥ 5 cm. All samples were herborized following traditional botanical techniques (IBGE, 2013) and deposited in the Herbarium of the Department of Botany (RBR) of the Federal Rural University of Rio de Janeiro (UFRRJ). We classified the species according to their occurrence in the pedoforms as generalist (species with common occurrence in the concave and convex pedoforms) and exclusive (species that occurred only in one pedoform type). We calculated the phytosociological parameters following the form proposed by Felfili et al. (2011). All analyses were performed in Microsoft Excel (2007). We expressed floristic diversity using the Shannon index (Magurran, 2011) and evenness with the Pielou (1984) index, which we compared with results found in similar formations with acknowledged biological diversity. We analyzed the horizontal and vertical structure of each pedoform using DBH and height distribution charts. The values of total height, total basal area, and abundance of the five main species sampled and microclimate data were submitted to a normality test (Shapiro-Wilk test / R × 64 2.15.3) and homogeneity of variance test (Cochran and Bartlett / R × 64 2.15.3). According to Zar (1996), environmental variables do not usually meet the theoretical assumption due to the broad number of factors present. However, the same author affirms that T-tests are robust enough to find significant differences between treatments. Hence, we used the Levene's T-test for independent samples, at the significance level of 95%, calculated in the SPSS 21 software.
The botanical nomenclature followed APG III (2009). Names, synonyms, and occurrence of the species in the Rio de Janeiro State were checked in the Species List of the Flora of Brazil (Forzza, 2005).

RESULTS
The relative humidity was higher in concave pedoforms (12.52% higher), being January (16.35%), February (14.27%), July (15.01% higher), August (15%) and October (20.94%) the months with the highest percentage differences. The air temperature had lower mean values in concave pedoforms, with significant differences between all evaluated months and the annual average (2.9 °C), especially in December (2.49 °C) and January (4.13 °C). The significant differences between the averages of said variables in 2014 were analyzed by the Levene's T-test and are presented in Table 1. In the phytosociological inventory, we recorded 1,843 individual trees (912 in concave and 931 in convex pedoforms) which were distributed in 39 families, 97 genera, and 126 species. Table 2 show the summary of phytosociological parameters for the 12 main species that occur in concave and convex pedoforms. The Shannon index was 3.81 for Mata do Peixoto, 3.63 for concave pedoforms, and 3.78 for convex pedoforms. Regarding the horizontal and vertical structure of the fragment, we found the same J-shaped data distribution pattern for DBH and height variables for both pedoforms. Figure 1 shows the distribution of individuals according to DBH and height in different pedoforms. When we compared the structural averages, the Levene's T-test did not show significant differences between the average values of the horizontal structure (basal area and DBH). However, the vertical structure (total height) showed different averages between pedoforms (Table 3). The average values of basal area, total height, and the abundance of the main generalist species suggest punctual differences as shown in Table 4.

DISCUSSION
Empirical studies and forest succession models predict changes in abiotic conditions during the succession process (Lebrija-Trejos et al., 2010;Oliver & Larson, 1990), promoted by the development of pioneer species that favor the germination of others belonging to different ecological groups (Valiente-Banuet & Verdú, 2007).
According to the progression of this succession, the variables "temperature" and "relative humidity of the air" are changed, influencing biogeochemical processes, such as content and complexity of organic matter, degree of decomposition to microbial activity (Biasi et al., 2008;Raich et al., 2006;Wagai et al., 2008), altering the availability of nutrients that influence the development of plants.
In regions such as Médio Vale do Paraíba, where different pedoforms occur, this process can be facilitated, since the microclimatic conditions found, especially in concave pedoforms, favor the occurrence of specific environmental conditions, capable of influencing the community's ecological dynamics.
The different microclimatic pattern between pedoforms can be attributed to the shape of the feature, which in concave ones restrains the direct incidence of the winds and favors the shorter period of sunshine, besides promoting the convergence of surface waters to a central drainage line (Casseti, 2005),  (Casseti, 2005), besides being exposed to higher solar radiation and dissecting by the action of direct winds, favoring higher temperature means and lower relative humidity. When we analyzed the phytosociological and structural parameters of the main generalist species, we observed some differences between pedoforms, such as the significant variations between the vertical structure among pedoforms, the basal area and abundance of Cupania oblongifolia, the variation in the vertical structure and abundance of Apuleia leiocarpa, and the difference in the abundance of Astrocaryum aculeatissimum, being observed better establishment of their populations in concave pedoforms.
Considering the classification of those species by ecological groups (Carvalho, 2003;Lorenzi, 1998Lorenzi, , 2002Lorenzi, , 2009, it is possible to associate those pedoforms with differentiated microclimate conditions to promote the establishment of secondary and climax species. On the other hand, the significantly higher abundance of Piptadenia gonoacantha, Brosimum guianense and Guapira opposites in convex pedoforms suggests the microclimate conditions of those pedoforms favored the establishment of pioneer species.
Although it is probable that the existing microclimatic differences favored the occurrence of specific species and genera, in addition to populations with different behavior between pedoforms, it is worth mentioning that these variations are subtle and punctual, not being possible to determine the existence of communities of different tree species between pedoforms.
For practical purposes, knowledge about the microclimatic conditions differentiated between pedoforms and their influence on the behavior of species allows the technician, manager or owner to formulate specific conservation and preservation strategies for the Paraíba Valley forest fragments. Such strategies range from choosing appropriate species for seedling production, the ideal point of the slope for planting, or simply choosing areas with higher resilience for isolation, such as concave pedoforms.
In relation to the diversity of species, Miranda & Diógenes (1998) found for tropical forests average values of Shannon-Wiener index that vary from 1.5 to 3.5, and rarely surpass 4.5 nats. ind -1 . Kurtz & Araújo (2000) found for the Atlantic Forest of the Rio de Janeiro State values that vary from 1.69 to 4.4 nats. ind -1 , indicating that the Shannon index calculated for Mata do Peixoto (H' = 3.81), in concave (H' = 3.63) and convex pedoforms (H' = 3.78), suggests high diversity in the study area, in particular when compared with other studies carried out in semi-deciduous seasonal forests of Southeastern Brazil (Table 5). Corroborating the high values of the Shannon index, the Pielou evenness index calculated for the entire fragment (J' = 0.821), and for concave (J' = 0.798) and convex pedoforms (J' = 0.817), can also be considered high. The closer this index is to one, the higher is the diversity and species abundance in the sampling units (Magurran, 1988).
According to Harper (1990), tropical forests have a higher frequency of individuals in the classes of small diameters. This behavior was also found by Guariguata et al. (1997) in Costa Rica and by Lima et al. (2013) in the Brazilian Amazon, being indicative of stability of ecological processes in tropical forests, and pointing to the existence of a "stock community" made up of different species responsible for the replacement of dead individuals (Scolforo et al., 1998). For both pedoforms were found this standard, as well as for studies by Dan et al. (2010), Gonzaga et al. (2008), Carvalho et al. (2007), Reis et al. (2007), Coraiola & Péllico Netto (2003), Silva et al. (2003), Rodrigues et al. (2003), Meira-Neto & Martins (2002), Botrel et al. (2002) in seasonal Forests of Rio de Janeiro, Minas Gerais and São Paulo.
The joint analysis of the horizontal and vertical structure of the concave pedoform points out to a dense understory, where predominate individuals with low diameter, followed by a dense subcanopy, and a continuous canopy, surpassed by a small number of emerging large-sized individuals. However, despite the structural similarity between pedoforms, we detected a small number of individuals in the canopy of convex pedoforms, which promoted a slight discontinuity of this stratum.
The low number of individuals in the canopy and the high number of individuals in the understory of convex were possibly responsible for the low average of total height and by the significant difference between pedoforms ( Table 2). The cause of high density of thin individuals in the understory can be attributed to a greater luminosity promoted by the low density of individuals in the canopy (Carvalho & Nascimento, 2009;Parrotta, 1993). Such structural characteristics result from the convex edge effect due to its exposure to dry winds, higher insulation time and unfavorable microclimatic conditions (an inverse concave condition is observed).
The results of the number of individuals per species indicated a small number of species concentrated values close to 50% of the total sampled. In this context, it is worth emphasizing the predominance of the species Apuleia leiocarpa, Pseudopiptadenia contorta, and Guapira opposita, which showed a large number of individuals, regardless of the pedoform studied. When compared with the main studies carried out in seasonal forests in Minas Gerais and São Paulo (Botrel et al., 2002;Carvalho et al., 2007;Coraiola & Péllico Netto, 2003;Gonzaga et al., 2008;Meira-Neto & Martins, 2002;Reis et al., 2007;Rodrigues et al., 2003;Silva et al., 2003), Mata do Peixoto shows different structure and floristic composition, especially for the absence Copaifera langsdorffii, species cited by the authors with greater abundance in these formations. These floristic differences can be associated with the occurrence of ecotonal areas in those states between seasonal forests and Cerrado formation, because C. langsdorfii is typical of the latter biome.
Among the other species with high importance value in Mata do Peixoto, Cupania oblongifolia, Astronium graveolens, and Astrocaryum aculeatissimum stand out for their broad distribution in forests of the state, particularly in formations at the late natural regeneration stage (Borém & Oliveira-Filho, 2002;Carvalho & Nascimento, 2009;Neves & Peixoto, 2008). On the other hand, Piptadenia gonoacantha and Nectandra membranacea mostly occur in forests at early regeneration stage, but large-sized individuals were also found in communities at intermediate regeneration stage, both in dense rainforests (Cysneiros et al., 2015;Gandra et al., 2011;Guedes-Bruni et al., 2006) and semi-deciduous seasonal forests (Dan et al., 2010).
Likewise in the present study, Dan et al. (2010), Pinto Sobrinho et al. (2010), Guedes-Bruni et al. (2006) and Borém & Oliveira-Filho (2002) observed that Pseudopiptadenia contorta is one of the main species of the communities, evidencing its broad distribution in different forest formations in the Rio de Janeiro State. On the other hand, Apuleia leiocarpa shows in Mata do Peixoto a high IVI due to the large number of individuals with small basal area, a structural characteristic different from that found in other populations in the state. In their studies on dense rainforests, Cysneiros et al. (2015), Carvalho &Nascimento (2009) andCarvalho et al. (2007) pointed A. leiocarpa as a poorly relevant species at the lower strata of communities, as it occurs mainly in the canopy (Neves & Peixoto, 2008), and consequently has a large basal area. This structural difference for the species in Mata do Peixoto can be mainly explained by logging exploitation, which is common in seasonal forests of the Rio de Janeiro State, as most fragments of the formation are located in private areas.

CONCLUSIONS
The presence of species typical of ecosystems at late stage succession in concave pedoforms, as well as the largest number of pioneer and initial secondary species in convex pedoforms, indicates that, problably, the studied fragment has two different successional stages, one more incipient (convex) and another intermediate to late (concave).
This condition is result of higher annual mean relative humidity and lower temperature in concave pedoforms in 9 -10 relation to convex pedoforms, which may have favored the processes of resilience and the advancement of succession stages.
Despite exclusives species occurrences, differences between IVIs and Total Height of some species, the microclimate variations do not promoted significant differences in the structure and composition of tree community between pedoforms, being these differences punctual and insufficient to determine specific communities. However, they can help the development of natural regeneration.
The generalist species showed good development regardless of the pedoform, and therefore are good options for the reforestation of degraded and disturbed areas of the seasonal semideciduous forest of Rio de Janeiro. Regarding specific species, they are indicated for enrichment plantations according to their ecological group and the successional stage of the community.
The information obtained with floristic and structural analysis of the fragment contributed to the knowledge on the remains of the regional flora, as well as the amplitude of structural variations. Such information, associated with the differentiated microclimatic characteristics of each pedoform, will allow the effective application of a resource for the recovery, enrichment or preservation of seasonal semi-deciduous forests.