The Diversity of Vascular Epiphytes Is Influenced by Forest Fragment and Landscape Characteristics?

Driélli de Carvalho Vergne; Beatriz Rezzieri Marchezini

DOI: 10.22167/2675-6528-202602841

Article derived from a Final Course Work (TCC), with content based on the student’s original work and adapted to the editorial format of the E&S Magazine with the support of the ResumeAI tool, an artificial intelligence solution developed by Instituto Pecege for textual synthesis and organization.

Summary

The Atlantic Forest faces serious threats from deforestation and forest fragmentation, resulting in habitat loss, species loss, and alteration of local biodiversity. A study investigated whether landscape and forest fragment characteristics influenced the richness and abundance of vascular epiphytes in remnants of the Atlantic Forest. For this purpose, 11 forest fragments were selected in the South of the State of Minas Gerais. In these locations, vascular epiphytes were sampled in circular plots, and landscape metrics were calculated using satellite imagery and geoprocessing software. Data analysis employed clustering techniques, principal component analysis (PCA), and ordinary least squares (OLS). The results indicated that the richness and abundance of vascular epiphytes were not homogeneously distributed among the groups identified by clustering. Terrain slope influenced epiphyte richness, and terrain slope and proximity to water bodies affected both richness and abundance. It was concluded that landscape characteristics, such as terrain slope and proximity to water, are capable of influencing the richness and abundance of vascular epiphytes in forest fragments of the Atlantic Forest.

Keywords: Diversity; Landscape Ecology; Vascular Epiphytes; Forest Fragmentation.

1. Introduction

The Atlantic Forest, recognized for its vast biodiversity, has faced intense eradication threats due to high deforestation rates (Wagner et al., 2019a). In recent years, the original cover of this biome in Brazil has been reduced to only thirteen percent (INPE, 2017), with remnants limited to isolated forest fragments in the landscape. The study was conducted in an area with natural vegetation characterized as Montane Semideciduous Atlantic Forest, which presents a combination of deciduous and perennial species at altitudes above five hundred meters (Veloso et al., 1991). The local climate is classified as Cwa and Cwb, with hot, rainy summers and dry, cold winters (Köppen, 1948; Alvares et al., 2013). The predominant landscape is agricultural and highly fragmented (Olivetti et al., 2015).

Forest degradation directly leads to habitat loss, alteration of ecosystem functions, and decrease in local diversity (Kaimowitz, 1996; Houghton, 2003). These factors also modify the abundance and distribution of species populations in these habitats (Hanski et al., 1996). The conversion of forest areas into smaller, non-continuous fragments, surrounded by habitats different from the original, known as matrices (Forman and Godron, 1986), acts as a filter that hinders the movement of different species across the landscape.

Vascular epiphytes represent one of the most susceptible groups to diversity loss due to forest fragmentation, as this process results in habitat loss and causes changes in the structure and composition of the tree species community, in addition to altering local microclimatic conditions (Werner et al., 2005). Epiphytes use tree species, called phorophytes, as mechanical support for the development of their life cycle, without contact with the soil (Benzing, 1990). Among the twenty thousand plant species described in the Atlantic Forest, it is estimated that three thousand to four thousand are vascular epiphytes, which corresponds to about fifteen to twenty-five percent of the flora (Kersten, 2010; Myers et al., 2000).

These plants contribute significantly to the functioning of the ecosystem, aiding in forest dynamics and biogeochemical cycles (Nadkarni, 1992; Oliveira, 2004). They promote environmental heterogeneity and niche availability, serving as a source of food and shelter for various groups of species (Cruz-Angón and Greenberg, 2005; Henwood et al., 2014; Teixeira et al., 2002). Furthermore, epiphytes are sensitive to climate change, pollution, and ecosystem damage (Lugo and Scatena, 1992), being considered indicator species of forest conservation status (Zotz and Bader, 2011). Their biomass and photosynthesis can be equivalent to those of their own phorophytes (Benzing, 1990), highlighting their great importance for the dynamics of tropical forests (Kersten, 2006).

Despite the recognized ecological importance of epiphytes and the increasing fragmentation of the Atlantic Forest, there is still a gap in understanding how landscape and forest fragment characteristics directly influence the richness and abundance of these communities in such altered environments. Understanding these factors is crucial for the development of effective conservation strategies.

Given the relevance of epiphytes for maintaining biodiversity and ecosystem functioning in a biome under intense anthropic pressure, this study is justified by the need to investigate the patterns of their diversity in fragmented landscapes. Thus, this work aimed to verify whether landscape and forest fragment characteristics influence the richness and abundance of vascular epiphytes in remnants of the Atlantic Forest.

2. Material and Methods

The study, of a quantitative and descriptive nature, aimed to verify the influence of landscape characteristics and forest fragments on the richness and abundance of vascular epiphytes. The methodology combined field surveys, geoprocessing, and multivariate statistical analyses. The research was conducted in 11 fragments of Seasonal Semideciduous Montana Atlantic Forest, in the South of Minas Gerais, Brazil, within a 50-kilometer radius of the municipality of Alfenas. This region, with altitudes between 794 and 1139 meters and Cwa and Cwb climate (Köppen, 1948; Alvares et al., 2013), is predominantly agricultural and fragmented, with only 11% of original forest cover (Olivetti et al., 2015).

For the collection of epiphytic data, five circular plots with a radius of 8 meters were established in each fragment. The center of each plot was a host tree with at least one epiphyte. The plots were demarcated at a minimum of 30 meters from the fragment edge to minimize edge effects.

In each plot, vascular epiphytes were identified to the species level on host trees with a diameter at breast height (DBH) equal to or greater than 5 centimeters. DBH and tree height were measured, and the presence and coverage of lianas and bamboo were recorded. Identification was carried out in the field or by collection and cultivation for subsequent taxonomic determination.

From the field data, the richness (number of species) and abundance (number of individuals) of epiphytes per fragment were calculated. Additionally, the effective number of species (SPIE) was determined, which is the inverse of the Simpson Diversity Index, reflecting the diversity and evenness of the community.

The characteristics of the landscape and fragments were obtained from Sentinel-2 and Landsat-8 satellite images, processed with QGIS and Fragstats. To classify the landscape and determine the matrix type, 1-kilometer radius buffers were delimited from the center of the fragments, quantifying the soil cover composition.

Landscape metrics included area, shape (“shape index”), perimeter, core area, edge-to-area ratio, and degree of isolation. To characterize composition and topography, the percentage of core area in 100, 500, and 1000 m buffers, mean slope and dominant slope direction, mean altitude, and distances to planted forest, water bodies, conservation units, pastures, and roads were calculated.

The statistical analysis began with agglomerative hierarchical clustering (Euclidean distance, complete linkage) to group fragments based on richness, abundance, SPIE, and landscape metrics. An F-test assessed inter- and intra-group variability.

To investigate the influence of landscape and fragment characteristics on epiphyte richness and abundance, multiple linear regression using Ordinary Least Squares (OLS) was employed. Previously, a Principal Component Analysis (PCA) selected independent predictor variables with lower correlation, avoiding multicollinearity. Quantitative variables for PCA were selected based on ecological relevance (composition, spatial configuration, and topography), including percentage of core area in 100 and 1000-meter buffers, average slope and dominant slope direction, average fragment altitude, and distances to water bodies, roads, and conservation units. Analyses were performed in the SPYDER software, using the Python language.

After the correlation analysis of the variables with the highest weight in the PCA, proximity to water bodies and terrain slope in the fragment were selected for the multiple regression model. This procedure ensured that the OLS model was built with predictor variables that minimized multicollinearity, allowing for precise evaluation of their effects on the response variables.

3. Results and Discussion

The investigation into the diversity of vascular epiphytes in fragments of the Atlantic Forest revealed complex patterns influenced by intrinsic characteristics of the forest remnants and the configuration of the surrounding landscape. The findings indicated that the richness and abundance of these species are not homogeneously distributed among fragments, suggesting that environmental and landscape factors play a crucial role in structuring these communities. The methodology employed cluster analysis, principal components analysis, and regression to unravel these relationships, providing an in-depth understanding of the mechanisms that shape epiphyte diversity in fragmented environments.

Clustering analysis

The initial cluster analysis, focused on richness, total abundance, and effective number of species (SPIE), allowed for the formation of six distinct groups of forest fragments. Species richness was the variable that contributed most to the formation of these groups, followed by total abundance, while SPIE, although significant, had a smaller contribution. The variability between groups was statistically smaller than the variability within them, as indicated by the F values of 200.53 for richness (p=0.000009), 61.64 for total abundance (p=0.000172), and 19.82 for SPIE (p=0.002606), confirming the heterogeneity among the clusters.

It was observed that the Matão and Nanette fragments formed the most similar group, joining at a Euclidean distance of less than 0.5. In contrast, the cluster that brought together Açude 2, Fragment N, Campos Gerais, Fragment I, and São Thomé represented the largest grouping, indicating considerable homogeneity in the epiphyte community among these locations. On the other hand, the Monte Belo and Zoo fragments remained isolated from the other groups, connecting to them only at a distance close to 5.0, which suggests that these fragments possess characteristics of richness and abundance that significantly distinguish them from the others.

The inclusion of landscape metrics in cluster analysis reorganized the fragments into six new groups, highlighting the influence of these characteristics on cluster formation. The percentages of core area within 1000-meter and 500-meter buffers were the landscape metrics that most contributed to this formation, with F values of 81.05 (p=0.000088) and 16.8 (p=0.003816), respectively. Other influential metrics included average terrain slope, distance to pastures, roads, and conservation units, slope direction, and average fragment altitude, in addition to total epiphyte abundance and richness.

The analysis revealed that fragments such as Açude 1 and Paraguaçu, which initially appeared isolated when considering only the dependent variables, formed new groupings (Açude 1 and Zoo in Cluster 1; Paraguaçu and Nanette in Cluster 4) when landscape metrics were included. This demonstrates that landscape similarity can group fragments that would otherwise be considered distinct. However, the Monte Belo fragment remained isolated in all analyses, reinforcing that its diversity and landscape characteristics classify it as the most unique among the studied fragments.

When analyzing the clustering based on the number of individuals per species, the fragments were distributed into five groups, confirming the non-uniformity in the distribution of individuals. Cluster 2, composed of Açude 1, Nannete, Zoo, São Thomé, Açude 2, and Matão, presented the largest number of fragments and the highest abundance, with 624 individuals. Cluster 1, formed by Campos Gerais and Fragment N, stood out as an isolated group, indicating a distinct species composition. The Monte Belo fragment, with 342 individuals distributed across 23 species, remained isolated, corroborating its compositional uniqueness.

Cluster analysis by taxonomic families resulted in six groups, with Cluster 5 being the most similar, containing five fragments and twenty families. The Zoo and Paraguaçu fragments showed a composition dominated by the Orchidaceae and Polypodiaceae families, grouping together in Cluster 3. Açude 1 and Monte Belo, in turn, remained isolated, connecting to the others at a distance of 50, which suggests taxonomic uniqueness. The Monte Belo fragment stood out for housing the largest number of individuals belonging to the Araceae family.

The Orchidaceae and Polypodiaceae families were the only ones with representatives in all sampled fragments, which is consistent with their requirements for shaded and humid environments for establishment and survival. Orchidaceae species are frequently found in Atlantic Forest forests, representing about 45% of the epiphytic species identified in the study (Kersten et al., 2001; Borgo et al., 2002; Kersten, 2010). The Polypodiaceae family, in turn, is recognized as the most expressive among pteridophytes (Benzing, 1990), which justifies its wide distribution and presence in all analyzed fragments.

Principal Component Analysis (PCA)

To avoid multicollinearity and select the most relevant predictor variables for the regression models, a Principal Component Analysis (PCA) was performed. Quantitative independent variables were selected based on ecological relevance criteria, covering composition, spatial configuration, and topography, such as percentage of core area in different buffers, average terrain slope, average altitude, slope aspect, and distances to water bodies, roads, and conservation units. The first principal component (PC1) explained 37.21% of the total data variance, while the second principal component (PC2) explained 30.44% of the variance.

In PC1, the distance to protected areas exerted a significant positive influence, with a weight of 0.51. In contrast, the average altitude of the fragment (-0.51), the distance to water bodies (-0.41), and the terrain slope (-0.38) were the variables with the greatest negative influence. This analysis was fundamental to identify the main environmental gradients that structure the landscape and that potentially affect epiphyte communities, allowing for a more robust selection of variables for subsequent regression models.

Ordinary Least Squares (OLS)

From the first principal component (PC1) generated for the landscape structure, linear regressions were performed with the dependent variables richness, abundance, and SPIE. However, no significant linear relationship was found between landscape structure (PC1) and richness (R² = 0.080; p-value = 0.39), abundance (R² = 0.000; p-value = 0.99), or SPIE (R² = 0.116; p-value = 0.30). This suggests that the overall landscape structure, as synthesized by PC1, does not directly explain the variation in these epiphyte diversity metrics in a linear manner.

For the construction of the multiple regression model using Ordinary Least Squares (OLS), the quantitative independent variables that showed the greatest weight in the PCA analysis and lower multicollinearity were selected. The chosen variables were the distance to the nearest body of water and the average slope of the terrain in the fragment. The preliminary correlation analysis between the selected independent variables revealed that, although some correlations were high, the final choice sought to minimize redundancy for the regression model.

The results of the multiple regression model indicated that landscape characteristics influenced the epiphyte community. The average terrain slope in the fragment exerted a significant negative influence on epiphyte richness (coefficient = -0.6191, p-value = 0.037). This suggests that areas with steeper slopes, which are generally more exposed to light and wind action, tend to have a less humid microclimate, which disfavors the establishment and development of vascular epiphytes. Epiphyte richness, therefore, is favored in areas with gentler slopes, where atmospheric humidity is higher.

Regarding the abundance of epiphytes, the average slope of the terrain also showed a significant negative influence (coefficient = -10.4497, p-value = 0.013). Additionally, the distance to the nearest water body demonstrated a significant positive influence on abundance (coefficient = 0.0511, p-value = 0.028). This relationship corroborates the importance of local humidity for epiphytes, which depend directly on atmospheric moisture for their life cycle, as they do not have contact with the soil (Kersten, 2010). Less steep areas and those close to water bodies create more favorable microclimatic conditions, promoting greater abundance.

The analysis of the multiple linear regression residuals provided additional inferences about the relative performance of the fragments. The Açude 1, Monte Belo, and Zoo fragments showed positive performance in the residuals. The Açude 1 fragment exhibited a more balanced performance for richness, abundance, and SPIE. Monte Belo, in turn, showed high performance for abundance, but low for SPIE, indicating a large number of individuals concentrated in few species (342 individuals distributed among 23 species), which suggests lower diversity.

The Zoo fragment stood out with high richness and SPIE, but low abundance, being the fragment with the highest SPIE (25 species). This indicates that the Zoo has a more diverse and balanced epiphyte community compared to the other fragments studied. The other fragments mostly showed negative performance in the residuals for the analyzed variables, with isolated exceptions for SPIE in Matão, abundance in Paraguaçu, and SPIE and richness in Nannete. This negative performance may indicate that these areas are under some stress that directly affects the abundance and richness of local epiphytes, and that fragments like Monte Belo and Paraguaçu may be undergoing a process of biodiversity loss, characterized by the concentration of many individuals in few species.

In summary, the results confirmed that the diversity of vascular epiphytes is influenced by landscape and forest fragment characteristics. Terrain slope and proximity to water bodies were identified as determining factors, negatively impacting epiphyte richness and abundance in steeper areas and positively impacting abundance in locations near water, respectively. These findings highlight the importance of microclimatic conditions and environmental heterogeneity for maintaining epiphyte diversity in fragmented Atlantic Forest remnants, responding to the study’s objective of verifying the influence of these characteristics.

4. Conclusion

The study sought to verify the influence of landscape characteristics and forest fragments on the richness and abundance of vascular epiphytes in remnants of the Atlantic Forest. It was found that the distribution of richness and abundance of these species did not occur homogeneously among the analyzed fragments. The findings indicated that the average slope of the terrain in the fragment exerted a significant negative influence on the richness and abundance of epiphytes, suggesting that steeper areas, with greater exposure and lower humidity, hinder their establishment. In contrast, proximity to water bodies showed a significant positive influence on abundance, highlighting the dependence of these plants on local atmospheric humidity. It was also observed that families such as Orchidaceae and Polypodiaceae were present in all fragments, reinforcing the need for shaded and humid environments for their survival. Cluster analysis, by incorporating landscape metrics, revealed that the percentage of core area in different buffers was a relevant factor in cluster formation, reorganizing the similarity among fragments.

The main contribution of this work lies in the identification of specific landscape factors, such as terrain slope and proximity to water bodies, which are determinants for the diversity of vascular epiphytes in fragmented Atlantic Forest environments. These results are crucial for the development of more effective conservation strategies, which should prioritize not only the extent of forest cover but also the preservation of areas with hydrological connectivity and topographic integrity. Although the general landscape structure, synthesized by a principal component, did not show a direct linear relationship with epiphyte richness and abundance, the detailed analysis of individual variables revealed the complexity of these interactions. It is suggested that future studies deepen the investigation into the specific microclimates generated by these topographic and hydrological characteristics, as well as the resilience of different epiphyte species to environmental variations in highly fragmented landscapes.

Bibliographic References

Alvares, C.A.; Stape, J.L.; Sentelhas, P.C.; Gonçalves, J.L.M.; Sparovek, G. 2013. Köppen climate classification map for Brazil. Meteorologische Zeitschrift 22(6): 711–728.

Benzing, D.H. 1990. Vascular epiphytes. Cambridge: Cambridge University Press.

Forman e Godron, 1986 [Referência completa não encontrada no documento original]

Hanski et al., 1996 [Referência completa não encontrada no documento original]

Houghton, R.A. 2003. Why are estimates of the terrestrial carbon balance so different? Global Change Biology 9: 500–509.

INPE. 2017. Atlas dos remanescentes florestais da Mata atlântica—período 2015-2016—Relatório técnico, Fundação SOS Mata Atlântica—Instituto Nacional de Pesquisas Espaciais.

Kaimowitz, D. 1996. Livestock and Deforestation in Central America in the 80s and 90s: A Policy Perspective. Cifor, Jakarta, Indonesia.

Kersten, R.A. 2010. Epífitas vasculares – Histórico, participação taxonômica e aspectos relevantes, com ênfase na Mata Atlântica. Hoehnea 37(1): 9-38.

Köppen, W.P. 1948. Climatologia. México: Fundo de Cultura Econômica.

Myers, N.; Mittermeier, R.A.; Mittermeier, C.G.; Fonseca, G.B.; Kent, J. 2000. Biodiversity hotspots for conservation priorities. Nature 403: 853-858.

Nadkarni, N. M. 1992. Biomass and nutrient dynamics of epiphytic literfall in a Neot

Olivetti, D.; Mincato, R.L.; Ayer, J.E.B.; Naves Silva, M.L.; Curi, N. 2015. Spatial and temporal modeling of water erosion in dystrophic red latosol (oxisol) used for farming and cattle raising activities in a sub-basin in the South of Minas Gerais. Ciência e Agrotecnologia 39(1): 58-67.

Veloso, H.P.; Rangel-Filho, A.L.R.; Lima, J.C.A. 1991. Classificação da vegetação brasileira, adaptada a um sistema universal. Ibge.

Wagner, F.H.; Sanchez, A.; Tarabalka, Y.; Lotte, R.G.; Ferreira, M.P.; Aidar, M.P.M.; Gloor, E.; Phillips, O.L.; Aragão, L.E.O.C. 2019b. Using the U-net convolutional network to map forest types and disturbance in the Atlantic rainforest with very high resolution images. Remote Sensing in Ecology and Conservation 5(4): 360-375.

Werner, F.A.; Homeier, J.; Gradstein, R. 2005. Diversity of vascular epiphytes on isolated trees in the montane forest belt of southern Ecuador. Ecotropica 11: 21-49.

Article originating from the Final Course Work of the Specialization in Data Science and Analytics of the MBA USP/Esalq

To learn more about the course, click here and access the MBX Academy platform