**3.3 Relationship between the relative abundance of taxa and enzyme activities**

Axis 1 of the canonical correspondence analysis for bacterial community explained 80% of the variation (**Figure 3**). Two groups were segregated with regard

#### **Figure 3.**

*Canonical correspondence analysis (CCA) of bacterial phyla demonstrating the effect of wet/dry periods at the Guánica Dry Forest. Blue symbols represent wet periods and red symbols represent dry periods. Triangles, circles and plus sign represent tree species (Tabebuia heterophylla, Ficus citrifolia and Pisonia albida), respectively.*

#### **Figure 4.**

*Canonical correspondence analysis (CCA) of bacterial phyla and soil enzyme activities (Phosdi = phosphodiesterasae, ArylSul = aryl sulphatase, APho = alkaline phosphatase, AcidPho = acid phosphatase, Bcosi = β-glucosaminidase, and Bglu = β-glucosidase). Vectors represent enzyme activities.*

abundance due to tree species (**Table 2**). There were no significant differences in any of the diversity indexes with regard to tree species (**Table 3**). Sampling period exhibited an effect on bacterial species richness (P = 0.05) and on equitability (P = 0.0243). but not on the Shannon index. We found higher bacterial species richness during the wet period and higher equitability during the dry period

**3.2 Evaluation of indicator species analysis (ISA) for this forest ecosystem**

indicator species were found for the dry period (**Table 4**).

*Bacterial indicator species analysis (ISA) at the Guánica Dry Forest.*

The identification of species associated or indicative of groups of samples is a common aspect of ecological research [24]. Indicator species analysis (ISA) identified several bacterial (**Table 4**) species responsible for changes in soil microbial communities. Out of 185 bacterial OTUs, 10 served as indicator species for *Ficus citrifolia* during the wet period (**Table 4**). A total of 31 bacterial OTUs were identified as indicator species for *Pisonia albida* during the wet period. No bacterial

(**Table 3**).

**52**

**Table 4.**

*Microorganisms*

to dry and wet periods (**Figure 3**). Samples that correspond to the wet period were associated with acid phosphatase, alkaline phosphatase, β-glucosaminidase, β-glucosidase, and arylsulfatase (**Figure 4**), whereas samples corresponding to the dry period are associated with phosphodiesterase (**Figure 4**). Microbial taxa associated with wet samples were Bacteroidetes, Proteobacteria, and Tenericutes. Microbial taxa associated with dry sampling points were *Actinobacteria*, *Planctomycetes*, *Acidobacteria*, *Verrucomicrobia*, *Cyanobacteria*, and *Chloroflexi*.

under study have an accumulation of organic matter mainly composed of new and old leaf litter due to the harsh environmental conditions; this bacterial phylum is associated with the degradation of more complex substrates of SOM including lignin, which could explain their predominance in this forest soils [37–39]. Our data suggest that *Actinobacteria* were highly associated with phosphodiesterase at the Guánica forest. Other studies have also found that *Actinobacteria* populations (characterized by TRLFP) correlated with P content in semiarid environments [40]. Other studies have also demonstrated the production of phosphatases by cultivable soil *Actinobacteria* [41–43] serving as a further evidence of the trend observed in this study. I also found that Actinobacteria were the most abundant phyla during the dry period which is consistent with other reports. A similar trend was also found for a semiarid, high desert grassland north of Flagstaff, Arizona [44] and in another study, which evaluated African tropical forest soils and Chinese forest soils [45]. *Bacteroidetes*, the third most abundant phylum in this forest, are widely distributed in different habitats ranging from Antarctic ice, lakes, the gut of animals, and terrestrial environments [46]. Additionally, this phylum has the ability to withstand extreme desiccation conditions including droughts and UV light [47]. They have been found in fine dust traveling thousands of kilometers [48] and inside microaggregates; authors explain that in semiarid agroecosystems this could be a protection strategy employed by these microbes to endure extreme environmental conditions [48]. I found that *Bacteroidetes* were highly associated with the activities of two soil enzymes evaluated (acid phosphatase and alkaline phosphatase). These enzymes mineralize organic P forms into inorganic P forms. Another study provided evidence suggesting that environmental *Bacteroidetes* specialize in the mineralization of high molecular weight organic matter making them a key compartment for carbon fluxes and budgets in ecosystems [49]. *Bacteroidetes* are oligotrophic and are commonly associated with substrates rich in organic matter [50] as is the forest area where I based my study. The high abundance of *Bacteroidetes* could be associ-

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

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

ated with the rich organic matter present in the sampling area.

**55**

**4.2 Influence of tree species on the bacterial populations in this forest**

Even though the most abundant bacterial phyla identified under all tree species were the same (*Proteobacteria*, *Actinobacteria*, and *Bacteroidetes*), some were specific for each tree species. Indicator species analysis (ISA) revealed that specific bacterial species were present under *Ficus citrifolia* and *Pisonia albida* only during the wet period. This trend suggests that these species assemblages may play an important role in the soil ecosystem processes under these specific tree species. One species was *Devosia* spp. whose OUT frequency was 14 (ind val = 0.62; p = 0.009) in soils collected under *Ficus citrifolia. Devosia* spp. forms part of the α-*Proteobacteria*; this genus is a non-rhizobia nodulating, nitrogen fixer [51]. Three different species of *Brevundimonas* spp. also occurred in high frequency under *Ficus citrifolia* during the wet period as indicated by ISA. This genus is actually known to produce phosphodiesterase, a group of enzymes involved in the degradation of organophosphorus [52]. During the wet period, 32 different OTUs were identified exclusively for *Pisonia albida*, and at least 7 of them had a frequency of 18. This high frequency suggests that indicator species may be playing an important role in the soil dynamics of *Pisonia albida*. For instance, three of the most frequent species are nitrogen fixers (*Microlunatus* sp., *Rhodospirillacease* spp*.*, and *Paenibacillus*), and *Pisonia albida* was the tree species with the highest total available nitrogen in this study (data not shown here). Even though we did not cultivate the indicator species nor have information regarding the physiology of the indicator species identified for
