**2. Materials and methods**

#### **2.1 Study site and sampling**

The study was carried out at the Guánica Dry Forest which is situated in the semiarid region of Southwestern Puerto Rico (**Figure 1A**–**C**). Trees in this area are dwarfed, and the vegetation is located between 0 and 150 m from the coastline. The mean annual precipitation of this zone is 750 mm and exhibits a bimodal distribution. Even though monthly rainfall is highly erratic, most studies have documented that almost 50% of the annual precipitation occurs between September and November [11, 12]. Historical data also demonstrates that the forest also exhibits two periods of predominant drought that start in January and June. The data presented here correspond to samples that were taken during July 2011 (wet) and January 2012 (dry) representing the months of higher (wet) and lower (dry) rainfall [13], therefore allowing us to measure maximum and minimum response of the substrates microbiome.

sampling depth was not fixed; it ranged from 0 to 8 cm. We selected three (3) trees from three species, previously tagged and studied, that grow from 100 m to approximately 300 m from the coast. The tree species selected complied with the following characteristics: (a) they were interspersed within the study area, (b) each tree formed an island that was isolated from other trees by exposed rock, and (c) that their litter and belowground substrate originated from their own residue decomposition [14]. The three species selected were *Tabebuia heterophylla (DC.)* Briton (facultative deciduous), *Pisonia albida* (Heimerl) Briton ex Standal, (obligate decid-

*Location and description of the study site. (A) The Island of San Juan Puerto Rico forms part of the Greater Antilles and is bordered by the Caribbean Sea. (B) The Guánica Dry Forest Biosphere Reserve found in the southwestern area of Puerto Rico. (C) Picture of the landscape of the study site. Here we observe the coastal area of the forest where trees are dwarfed and have established their growth in the cracks and crevices of the rocky*

*First Insights into the Resilience of the Soil Microbiome of a Tropical Dry Forest…*

*DOI: http://dx.doi.org/10.5772/intechopen.90395*

Each tree was used as a sampling unit supported by the very high heterogeneity in the vegetation structure of the site and the actual physical separation of the trees.

uous) and *Ficus citrifolia* Mill., (facultative deciduous) (**Figure 1D**).

*substrate. (D) Representation of tree species used in this study.*

**Figure 1.**

**47**

The soils of the area are described as isohyperthermic Calcic Lithic Petrocalcids of the Pitahaya-Limestone outcrop-La Covana Association which consist of shallow, well-drained, very slowly permeable soils that formed or were deposited in material that weathered from limestone bedrock (USDA NRCS 2008). The depth of the substrate varies according to ground relief and among seasons [14]. The low stature woody vegetation grows on a rocky calcareous substrate where plants establish their roots in the holes, cracks, and crevices, of the rocky material accumulating water and very shallow soil substrates. Due to the high variability of the surface area soil

*First Insights into the Resilience of the Soil Microbiome of a Tropical Dry Forest… DOI: http://dx.doi.org/10.5772/intechopen.90395*

#### **Figure 1.**

the Caribbean point towards progressively drier periods, with precipitation loss between 10 and 50% [3]. There is limited information regarding the diversity of soil microbial communities in these ecosystems, and it needs to be assessed in order to establish baseline information that is crucial to help elucidate the degree of the ecosystems resilience to the proposed precipitation changes that are affecting these

The intrinsic effects of vegetation are strong influencers of soil properties. Due to the confounding factor of plant species and plant roots sharing the same area, there is very little information on the effect of specific plant species on microbial diversity and soil enzymatic activities [4, 5]. Seasonally dry ecosystem, such as the Guánica Dry Forest, can serve as a model system to better understand the impact of seasonal variations and tree species effect on microbial community composition and activity. In this forest the trees are growing in the cracks of the calcareous platform, forming individual islands of leaf litter and organic matter under similar climatic conditions [5, 6]. The canopies of the trees are close to ground level limiting the transfer of leaf litter between neighboring trees, thus forming individual islands of fertility. These individual islands of fertility prevent belowground competition for resources [6]. Given that plant species vary in their effects on soil properties [7–9], one of our objectives was to determine how responsive the soil microbial diversity is to plant species effects. Our second objective was to understand to what degree the soil microbial diversity shows resilience to rainfall variability and dry periods. The study was conducted during the rainiest period of 2011 (August) and the driest period of 2012 (January). We selected three tree species (a pantropical species and two native species) that are highly distributed in the forest [10]. We hypothesize that both rainfall and plant species will regulate of

The study was carried out at the Guánica Dry Forest which is situated in the semiarid region of Southwestern Puerto Rico (**Figure 1A**–**C**). Trees in this area are dwarfed, and the vegetation is located between 0 and 150 m from the coastline. The mean annual precipitation of this zone is 750 mm and exhibits a bimodal distribution. Even though monthly rainfall is highly erratic, most studies have documented

The soils of the area are described as isohyperthermic Calcic Lithic Petrocalcids of the Pitahaya-Limestone outcrop-La Covana Association which consist of shallow, well-drained, very slowly permeable soils that formed or were deposited in material that weathered from limestone bedrock (USDA NRCS 2008). The depth of the substrate varies according to ground relief and among seasons [14]. The low stature woody vegetation grows on a rocky calcareous substrate where plants establish their roots in the holes, cracks, and crevices, of the rocky material accumulating water and very shallow soil substrates. Due to the high variability of the surface area soil

that almost 50% of the annual precipitation occurs between September and November [11, 12]. Historical data also demonstrates that the forest also exhibits two periods of predominant drought that start in January and June. The data presented here correspond to samples that were taken during July 2011 (wet) and January 2012 (dry) representing the months of higher (wet) and lower (dry) rainfall [13], therefore allowing us to measure maximum and minimum response of the

ecosystems.

*Microorganisms*

modify the soil microbiome.

**2. Materials and methods**

**2.1 Study site and sampling**

substrates microbiome.

**46**

*Location and description of the study site. (A) The Island of San Juan Puerto Rico forms part of the Greater Antilles and is bordered by the Caribbean Sea. (B) The Guánica Dry Forest Biosphere Reserve found in the southwestern area of Puerto Rico. (C) Picture of the landscape of the study site. Here we observe the coastal area of the forest where trees are dwarfed and have established their growth in the cracks and crevices of the rocky substrate. (D) Representation of tree species used in this study.*

sampling depth was not fixed; it ranged from 0 to 8 cm. We selected three (3) trees from three species, previously tagged and studied, that grow from 100 m to approximately 300 m from the coast. The tree species selected complied with the following characteristics: (a) they were interspersed within the study area, (b) each tree formed an island that was isolated from other trees by exposed rock, and (c) that their litter and belowground substrate originated from their own residue decomposition [14]. The three species selected were *Tabebuia heterophylla (DC.)* Briton (facultative deciduous), *Pisonia albida* (Heimerl) Briton ex Standal, (obligate deciduous) and *Ficus citrifolia* Mill., (facultative deciduous) (**Figure 1D**).

Each tree was used as a sampling unit supported by the very high heterogeneity in the vegetation structure of the site and the actual physical separation of the trees. The minimum distance between any two trees was about 1 m, and the maximum was approximately 30 m. We collected one soil sample of each sampling unit (tree) during the months of the study. Soil samples were sieved in the field with a 2 mm mesh and placed in plastic bags. Samples were then placed on ice, taken to the laboratory, and frozen until they were sent to the Molecular Research Facility at Lubbock Texas. Total soil DNA extraction and 454 pyrosequencing were completed at the Molecular Research Facility. The molecular research facility reported all results as OUT tables.

PAST3 statistical program [22]. Nonparametric Kruskal-Wallis was calculated using the JMP10 statistical software to evaluate differences between diversity indexes as affected by tree species and rainfall. In the book "Microbial Source Tracking: Methods applications and case studies," Cao et al. (2011), page 278, discusses that "common multivariate techniques used for the examination of microbial community structure include cluster analysis, principle components analysis (PCA), correspondence analysis (CA), and nonmetric multidimensional scaling (NMDS). All of these techniques belong to a group called indirect gradient analysis, which aims to reveal community similarities among sites or samples through grouping or ordering the sites or samples into either dendrograms or on a two (2D) or three-dimensional (3D) plot." On the other hand, they also mention that "direct gradient analysis such as canonical correspondence analysis (CCA), aims to correlate the overall multivariate community profile with environmental variables." To identify the influence of soil physicochemical characteristics on the bacterial community, canonical correspondence analysis (CCA) was performed using the OTU tables of each community and the soil physicochemical characteristics. Canonical correspondence analysis is a site/species matrix where each site has given values for one or more environmental variables. The ordination axes are linear combinations of the 169 environmental variables. It is a gradient analysis that shows species abundances as a response to an environmental gradient. Environmental variables are plotted as correlations with site scores. I reported two types of scaling. Type 1 emphasizes the relationship between sampling sites and environmental variables, and Type 2 emphasizes relationships between species and environmental variables [21, 22]. Indicator species analysis (ISA) was performed to identify the bacterial species responsible for changes in soil microbial communities between tree species and sampling periods. ISA analysis was completed in R using the IndVal script, where R calculates the indicator value d of species as the product of the relative frequency and relative

*First Insights into the Resilience of the Soil Microbiome of a Tropical Dry Forest…*

*DOI: http://dx.doi.org/10.5772/intechopen.90395*

**3.1 OTU abundance at the Guánica Dry Forest as affected by sampling period**

**Phyla Dry S.D. Wet S.D. p value**

*Bacteroidetes* **2.80 0.94 7.60 23.26 0.01** *Chloroflexi* **2.60 0.99 0.75 0.58 0.01** *Cyanobacteria* 0.08 0.15 0.00 0.00 0.47 *Actinobacteria* **37.65 6.57 16.93 6.77 <0.0001** *Verrucomicrobia* **0.32 0.30 0.08 0.15 0.02** *Spam (candidate division)* 0.07 0.15 0.00 0.00 0.47 *Planctomycetes* **2.26 0.68 0.68 0.40 0.01** *Nitrospirae* **0.56 0.39 0.18 0.21 0.03**

Sequencing data revealed 17 predominant bacterial phyla for this forest (**Table 1** and **Figure 2**). The phyla with the highest relative abundance were *Proteobacteria*

**Bacterial phyla**

average abundance [23].

**and tree species**

**3. Results**

**49**
