Regular articles
Special Issues



Caribbean Naturalist
    CANA Home
    Range and Scope
    Board of Editors
    Staff
    Editorial Workflow
    Publication Charges
    Subscriptions

Other EH Journals
    Northeastern Naturalist
    Southeastern Naturalist
    Urban Naturalist
    Eastern Paleontologist
    Eastern Biologist
    Journal of the North Atlantic

EH Natural History Home

Potential Impact of Coastal Development on Nearshore Bacterial Diversity, Southwest Puerto Rico
Gary E. Schultz Jr., Jeffrey J. Kovatch, and Heidi Hertler

Caribbean Naturalist, No. 42

Download Full-text pdf: complete with cover or manuscript only (Accessible only to subscribers. To subscribe click here.)

 

Site by Bennett Web & Design Co.
Caribbean Naturalist 1 G.E. Schultz Jr., J.J. Kovatch, and H. Hertler 22001177 CARIBBEAN NATURALIST No. 42N:1o–. 2422 Potential Impact of Coastal Development on Nearshore Bacterial Diversity, Southwest Puerto Rico Gary E. Schultz Jr.1,*, Jeffrey J. Kovatch1,†, and Heidi Hertler2,3 Abstract - The biodiversities of nearshore bacterial communities along a coastline development gradient were examined using pyrosequencing and classical ecological diversity indices. Relative abundance of 16S rRNA gene sequences was determined from 9 sampling stations along a nearshore 3-km transect in southwest Puerto Rico that has both anthropogenically developed and undeveloped areas. Cyanobacteria and Proteobacteria were the numerically dominant phyla, and the relative abundance of either phylum was inversely proportional to the other along the transect. Land development appeared to be related to diversity, with lower diversities observed at the developed stations. Multidimensional scaling based on Bray-Curtis similarities indicated partial recovery of the relatively pristine community a short distance beyond the developed region. This study shows that the bacterial diversity of a nearshore marine environment may be influenced by proximity to coastal development and lays the foundation for future studies into the diversity of microbial assemblages in coastal areas. Introduction In recent years, we have become more aware of the ecological importance of coastal ecotones. The complexity of these coastal regions, however, makes them difficult to examine because they may be impacted by freshwater streams and rivers, overland inputs, tides, and human development. Several recent studies have examined the water-column microbial diversity of coastal systems using nextgeneration sequencing (Campbell et al. 2011, Clasen and Shurin 2015, Crespo et al. 2013, Dong et al. 2014, Fortunato et al. 2012, Ortega-Retuerta et al. 2013, Thompson et al. 2011, Villa-Costa et al. 2012). Our study seeks to add to this work by examining whether proximity to human development influences the diversity and composition of surface-water bacterial communities within a longshore current along a sub-tropical coastline. La Parguera, a small coastal village located in the sub-tropical dry zone of southwest Puerto Rico, was chosen for study because it offers a region adjacent to a coastal area in varying stages of development with little input from streams or rivers. In the developed areas, runoff is brought into the water primarily from the storm sewer system, and human and industrial waste is treated at a municipal 1Department of Biological Sciences, Marshall University, Huntington, WV 25755, USA. †Deceased. 2Inter American University of Puerto Rico, Center for Environmental Education, Conservation and Research, San German, PR 00683, USA. 3Current address - School for Field Studies, Center for Marine Resource Studies, 1 West Street, South Caicos, Turks and Caicos Islands. *Corresponding author - schultzga@marshall.edu. Manuscript Editor: Dawn A.T. Phillip Caribbean Naturalist G.E. Schultz Jr., J.J. Kovatch, and H. Hertler 2017 No. 42 2 water treatment facility (Hertler et al. 2009). Other sources of water and other materials to the coast are leaking pipes or materials generated in the coastal waters themselves due to recreation, boating, or other activities. Lack of stream input simplifies an extremely complex system by removing confounding influences that may be brought from further inland. There is also a significant amount of existing data on water quality, sediment quality, and the seagrass system present at Parguera from previous studies of this area (Environmental Quality Board 1972, Hertler 2002, Hertler et al. 2009). There is great functional redundancy in the microbial world. Therefore, it seems that the microbial community of a particular ecosystem might be determined through chance—the ecosystem would be colonized by whatever well-suited microbes happened to be there or brought there first. There is evidence, however, that clear patterns of microbial community structure exist, similar to patterns of macro-organisms (Martiny et al. 2006). Thompson et al. (2011) found similarity among the compositions of bacterial communities associated with coastlines in Latin America many thousands of kilometers apart despite the fact that large changes in a microbial community can potentially occur in a matter of hours (Kirchman 2008). Each region, however, seemed to have its own signature community as well, indicating that local forces may influence the bacterial community. These forces may stem, in part, from human development on the coast. Since bacteria form the basis of all biogeochemical cycles, it is important to understand the factors that affect bacterial communities. Thus, we need to know what bacteria are present in coastal systems and how the community behaves as it flows through these coastal areas. As human development continues along the world’s coastlines, it is vital to understand if this development influences bacterial communities. Several studies have examined the bacterial communities of estuarine, beach, and coastal sediments and showed that those bacterial communities are impacted by anthropogenic influences and appear to be indicators of contaminant stress (Piccini and Garcia-Alonzo 2015; Störmer et al. 2013; Sun et al. 2012, 2013). Yet the response of coastal water-column bacterial communities to anthropogenic stressors is understudied, and, as Sale et al. (2014) stated, managers of coastal seas must be aware of and cope with all variables that affect coastal waters. Therefore, we examined the bacterial community in the longshore current for detectable changes in the diversity, composition, and community structure of the bacterial community as the current flows past human development. This paper examines the transitions in taxonomy and diversity of the bacterial community along a transect paralleling the coast of southwest Puerto Rico that includes undeveloped ecosystems as well as areas of human development (Fig. 1). In addition, this work provides a valuable companion survey and comparison study to the survey of Latin America coastal bacterioplankton communities begun by Thompson et al. (2011) aimed at the characterization of the bacterioplankton diversity in Latin America. Field-site Description Our sampling stations were located just offshore from La Parguera. There is little freshwater influence because the Palmarejo Mountains to the north produce a rainCaribbean Naturalist 3 G.E. Schultz Jr., J.J. Kovatch, and H. Hertler 2017 No. 42 shadow effect that keeps annual precipitation totals low, between 800 and 1000 mm of rain (Ewel and Whitmore 1973). Therefore, streams are small, intermittent, and dependent on rainfall events. The closest river along this section of the coast is the Rio Loco. The Rio Loco enters Bahia Noroeste ~20 km east and up-current from our site and generally has a small discharge (3.5 m3/s [121.9 ft3/s] in June 2010; United States Geological Survey 2017). A salt flat and mangrove forest system form the riparian zone of the region surrounding La Parguera. The average tidal range is less than 0.25 m; wave action is minimal and suppressed by the mangrove forest. The shoreline faces south and a long-shore current runs from east to west at 5–10 cm/sec (Avila et al. 1979). We utilized stations sampled during previous studies (Hertler 2002; Hertler et al. 2003, 2009) and retained site designations. Station depth varied from 1 to 2 m (Hertler et al. 2003). Stations 1 and 2 were up-current of the development in La Parguera. We considered station 1 to be relatively pristine in this study as it was east of all La Parguera development and adjacent to an undeveloped, naturally vegetated hill. An extensive mangrove system was present on the terrestrial side. Station 2 was also up-current of La Parguera; however, we did not consider this station to be so pristine because it was adjacent to an unvegetated hill with many small houses. The hillside was steep with exposed rock. There were no mangroves in this area. Stations 3 and 4 were adjacent to the established portion of La Parguera that had been developed for over 40 years and had a well-vegetated landscape. Station 3 had a fringe mangrove system between the developed area and the sea. The mangrove system had been modified and cut to fit boats, docks, and various buildings. Station 4 was near a public boat ramp and a ferry to the University of Puerto Rico Marine Laboratory on Isla Magueyes. Mangroves here have also been cut extensively for development. Station 5 marked the beginning of the newer development in which extensive land development had occurred over the past 15 Figure 1. Map of stations sampled in La Parguera (17°58'N, 67°03'W), southwest Puerto Rico. The longshore current moves from east to west. Caribbean Naturalist G.E. Schultz Jr., J.J. Kovatch, and H. Hertler 2017 No. 42 4 years. The hillside was heavily developed and there were several boat ramps in this area. Mangroves were present. Station 14 was adjacent to a sparsely vegetated area near a dock used by local fishermen. Station 11 was just west of La Parguera by a marina in a well-used boat channel. Station 6 was farther west of La Parguera. There was little residential development and mangroves were present at this site, but the local waste treatment facility was located there. Finally, Station 22 (the second relatively pristine station) was located farther west and down current from Parguera (Fig. 1). There were extensive mangrove forests and no built development near this site. Methods Sampling We collected samples by boat on 28 June 2010 between the hours of 0900 and 1200. Rain had not fallen for the previous 96 hours (Bio-Optical Oceanography Laboratory 2017), so stream input and overland run-off were minimal at the time of sampling. Winds were light and variable, and there was no boat traffic except for at station 4. All samples were collected by hand 0.5 m below the water’s surface from the forward, windward side of the boat. We gathered samples for total suspended solids, chlorophyll-a, and nutrients using clean, acid-washed polyethylene bottles. These samples were taken to the Inter American University of Puerto Rico (IAUPR) laboratory for processing. We collected samples for DNA analysis using sterile 1-L glass bottles. Following collection, DNA samples were immediately placed on ice in a cooler in the dark and promptly taken back to the IAUPR laboratory for processing. We used a YSI-85 meter (YSI, Inc., Yellow Springs, OH, USA) to measure temperature, dissolved oxygen, conductivity and salinity on site. Total suspended solids and turbidity Whatman GF/F glass-fiber filters (0.7 μm, nominal) were combusted at 450 °C for 4 h, then rinsed with deionized water, dried for 24 h, and desiccated for 6 h before being weighed. We then filtered 1 L of each water sample through a preweighed filter. Filters were dried at 105 °C and then weighed (APHA 1995). We subtracted the pre-determined mass of the filters from the dried mass to obtain the dry mass of particulate material captured on each filter. The same filters were weighed again after being ashed at 450 °C for 4 h to determine the inorganic and organic fractions (Crompton 1989). We measured turbidity according to standard methods using a Turner nephelometer (APHA 1995). Dissolved nutrients Total nitrate (4500-NO3- E. Cadmium Reduction Method; APHA 1995), total nitrite (4500-NO2- B. Colorimetric Method; APHA 1995), and total phosphate (4500-P E. Ascorbic Acid Method; APHA 1995) concentrations within the samples were analyzed at IAUPR’s Center for Environmental Education Conservation and Research (CECIA) on the day we collected them. Caribbean Naturalist 5 G.E. Schultz Jr., J.J. Kovatch, and H. Hertler 2017 No. 42 Chlorophyll-a We determined chlorophyll-a concentration by filtering 1 L of the water sample in the dark through an ashed Whatman filter. Filters were immediately wrapped in aluminum foil and frozen. We performed extraction of chlorophyll-a in subdued light using 10 ml of 90% acetone. Samples were refrigerated overnight before centrifugation at 3000 rpm for 20 min (procedure modified from Strickland and Parson [1972]). We then analyzed the solution using a Turner Designs TD-700 fluorometer for chlorophyll-a (10200 H; APHA 1995). DNA extraction and pyrosequencing Samples for DNA analysis were filtered through a 0.22-μm Whatman Nuclear Track-Etch membrane and placed in a sterile 15-L Falcon tube (Fisher Scientific, Hampton, NH, USA). We sent filters on ice overnight to the Research and Testing Laboratory (RTL; Lubbock, TX, USA) for DNA extraction and pyrosequencing. Bacterial tag-encoded FLX amplicon pyrosequencing (bTEFAP) was performed at the RTL (Dowd et al. 2008). bTEFAP was analyzed on the genome sequencer FLX instrument using titanium protocols and reagents (Roche, Indianapolis, IN; Dowd et al. 2008). PCR primers 104F (5’-GGA CGG GTG AGT AAC ACG TG-3’) and 530R (5’-GTA TTA CCG CGG CTG CTG-3’), located in the V2-V3 hypervariable regions of 16S rRNA gene, were used for FLX amplicon pyrosequencing. Data analysis Sequencing data analyses were performed at RTL in 2 major stages: the quality checking and read denoising stage, and the diversity stage. After sequencing, denoising and chimera checking involved quality trimming to clean up poor-quality read ends, followed by classifying reads into clusters using USEARCH (Edgar 2010). Reads not joining a cluster were removed from analysis. The data were then checked for chimeras using B2C2 (Gontcharova et al. 2010) and any potential chimeras were removed. Denoised and quality-checked files were placed in FASTA formatted sequence files and then underwent further quality-control screening involving removal of poor reads, reads with low quality tags, and reads that were not at least 250 bp. FASTA files that passed the quality checking and read denoising stage were put through the analysis pipeline to determine each read’s taxonomic information. First, USEARCH was used to cluster files into operational taxonomic unit (OTU) clusters with 0% divergence. The resulting seed sequences were then compared against an NCBI database using BLASTN+ (KrakenBlast). The outputs were then compiled, and a NET and C# data-reduction analysis was performed (Callaway et al. 2010; Dowd et al. 2005, 2008; Sen et al. 2009; Suchodolski et al. 2009). Sequences with identity scores of >97% were resolved at the species level, >95% at the genus level, > 90%at the family level, >85% at the order level, >80% at the class level, and ≥77% at the phylum level. Sequences with identity scores of less than 77%, were discarded. The relative abundance of each organism was then presented at each taxonomic level (Schultz et al. 2013). Every sequence resolved to at least phylum level was considered a distinct OTU. Throughout this work, we based calculations on the OTU data to ensure the most accurate results. Caribbean Naturalist G.E. Schultz Jr., J.J. Kovatch, and H. Hertler 2017 No. 42 6 Correlations We determined Pearson correlations using the statistical package R (R Development Core Team 2007). Diversity and community similarity All calculations were performed on the OTU data regardless of taxonomic level assigned. We determined Beta diversity between stations using the Bray- Curtis similarity coefficient (Bray and Curtis 1957), which has been suggested to accurately reflect true community similarity (Bloom 1981, Faith et al. 1987). Bray- Curtis similarity and multidimensional scaling were performed using the vegan library (Oksanen et al. 2007) in the statistical package R (R Development Core Team 2007). We rarified individual samples (n = 9) to 9500 reads using step-sizes of 250 reads, random selection without replacement, and 10 iterations per sample using the SAS IML.We performed this procedure because comparisons of diversity estimates should be performed on samples with similarly sized sequencing depths (Gihring et al. 2012). We used the rarified data sets to calculate richness (SRare), Shannon’s diversity index (H'; Shannon and Weaver 1963), Shannon’s evenness (E; Pielou 1969, 1975), Simpson’s diversity index (D; Simpson 1949), and Simpson’s evenness (DE; Simpson 1949) (Table 1). Results Taxonomy along the transect Five hundred and eighty-one unique genera representing 25 unique phyla were represented across the 9 samples taken. At the phylum level, either Cyanobacteria or Proteobacteria made up the largest proportion of bacteria in all samples (Fig. 2). The top 5 phyla, Proteobacteria, Cyanobacteria, Bacteroidetes, Actinobacteria, and either Chloroflexi (stations 4, 11, and 6) or Firmicutes (stations 1, 2, 3, 5, 14, and 22), together made up 99% of the bacterial community in all samples. Planctomycetes and Fusobacteria were also seen in all samples, with the remaining 17 phyla seen sporadically among the samples. The phylum Proteobacteria was mostly made up of Alpha- and Gammaproteobacteria. Alphaproteobacteria made up from 59.1% to 78.7% of the total Proteobacteria (stations 2 and 14, respectively), and Gammaproteobacteria made up from 18.1% to 37.3% (stations 4 and 2, respectively). Beta-, Delta-, and Epsilonproteobacteria each made up less than 5% of the Proteobacteria at any particular station. Of the total bacterial community, Alphaproteobacteria made up from 18.5 to 42.3%, Gammaproteobacteria made up from 5.7 to 24.9%, and Beta-, Delta-, and Epsilonproteobacteria never made up more than 1.5% of the total community. At the genus level, either Silicibacter (8.5–23.9%) or Cyanobacterium (12.4– 57.4%) made up the largest proportion of bacteria at all sites except station 14, where Candidatus Pelagibacter was the second most abundant genus (11.4%). We examined the genus level to investigate whether particular genera changed from rare to abundant over the length of the transect. For consistency with previous Caribbean Naturalist 7 G.E. Schultz Jr., J.J. Kovatch, and H. Hertler 2017 No. 42 studies,we considered all bacteria present at an abundance of >1.0% at each station to be abundant and those present at an abundance of less than 1% to be rare (Campbell and Kirchman 2013, Campbell et al. 2011). The number of abundant genera seen varied from 9 at station 5 to 16 at station 1. The abundant genera made up between 80% and 85% of the total abundance at every station. Eighteen unique genera were considered abundant in at least 1 station. Synechococcus was the only genus that moved from being classified as rare (with a 0.4% relative abundance at station 1) to abundant (reaching a high of 3.5% relative abundance at station 6). Water quality parameters Physical and chemical data are presented in Table 2. Overall, the water quality characteristics were variable, but clear patterns were evident among stations. Temperature varied from 29.6 °C at station 2 to 31.4 °C at station 22. Dissolved oxygen varied from 54.9% at station 1 to 75.4% at station 11. Conductivity varied from 56.1 mS/cm at station 1 to 58.4 at station 22. Salinity varied from 33.7 ppt at station 1 Table 1. Diversity and operational taxonomic unit (OTU) richness estimates. Sample rarefactions to 9500 reads for individual samples were used to calculate comparable richness of OTUs (SRare), Shannon’s (H') and Simpson’s (1/D) diversity indices and Shannon’s (E) and Simpson’s (De) evenness. Estimates are means (n = 10 iterations) and standard deviations (in parentheses). Observed richness (Sobs) is not directly comparable due to dissimilar read sizes among samples. Station Reads Sobs SRare H' E D De 1 9542 239 238.6 3.37 0.62 14.68 0.062 (0.516) (0.00098) (0.000315) (0.0193) (0.000175) 2 11823 214 200.2 2.79 0.53 5.52 0.028 (2.90) (0.00929) (0.00159) (0.0524) (0.000426) 3 18673 268 217.8 2.50 0.47 4.02 0.018 (6.12) (0.0117) (0.00199) (0.0445) (0.000488) 4 16458 263 226.5 2.88 0.53 6.84 0.030 (3.78) (0.0086) (0.00225) (0.0775) (0.000732) 5 19826 242 190.3 2.44 0.47 3.90 0.021 (5.46) (0.0173) (0.00412) (0.0673) (0.000759) 14 14916 188 162.5 2.29 0.45 3.44 0.022 (2.64) (0.0165) (0.00308) (0.0441) (0.000393) 11 17466 206 171.0 2.46 0.48 3.86 0.023 (4.76) (0.0188) (0.00319) (0.0549) (0.000619) 6 12230 205 192.4 2.66 0.51 4.33 0.023 (2.37) (0.0109) (0.00228) (0.0458) (0.000371) 22 18607 284 233.0 3.05 0.56 7.93 0.030 (3.65) (0.0173) (0.00362) (0.127) (0.000798) Caribbean Naturalist G.E. Schultz Jr., J.J. Kovatch, and H. Hertler 2017 No. 42 8 Table 2. Water quality parameters for the 9 sampling stations. Sampling stations are given in order from east to west along the shoreline. No temperature was recorded for station 4. DO = dissolved oxygen, Cond. = conductivity, TSS = total suspended solids, and Chl-a = chlorophyll-a. Temp DO Cond. Salinity Turbidity TSS Chl-a PO4 NO2 NO3 Station (°C) (%) (mS/cm) (ppt) pH (NTU) (mg/l) (μg/l) (mg-P/l) (mg-N/l) (mg-N/l) 1 29.7 54.9 56.1 33.7 8.14 3.77 9.0 2.19 0.064 0.006 0.012 2 29.6 66.0 56.6 34.1 8.21 1.46 12.2 1.56 0.085 0.008 0.014 3 29.9 68.2 57.1 34.4 8.21 2.47 11.9 3.11 0.079 0.006 0.017 4 - 68.8 57.8 34.4 8.22 10.23 23.9 5.84 0.151 0.023 0.030 5 30.4 70.2 57.4 34.2 8.20 4.76 16.1 3.81 0.099 0.007 0.018 14 30.4 68.9 57.7 34.3 8.16 3.46 8.4 9.11 0.101 0.008 0.020 11 30.3 75.4 57.6 34.3 8.16 3.68 10.1 8.58 0.108 0.007 0.016 6 30.7 66.2 57.7 34.2 8.08 5.65 9.1 9.97 0.099 0.018 0.013 22 31.4 66.5 58.4 34.2 8.15 1.36 8.8 1.56 0.071 0.007 0.017 Figure 2. Relative abundances of the 6 most abundant bacterial phyla for 9 sampling stations. Caribbean Naturalist 9 G.E. Schultz Jr., J.J. Kovatch, and H. Hertler 2017 No. 42 to 34.4 ppt at stations 3 and 4. pH varied from 8.08 at station 6 to 8.22 at station 4. Turbidity, total suspended solids (TSS), PO4, NO2, and NO3 were highest at station 4 (10.23 NTU, 23.9 mg/L, 0.151 mg-P/L, 0.023 mg-N/L, and 0.03 mg-N/L, respectively). These parameters were lowest at station 22 (turbidity), station 14 (TSS), station 1 (PO4), stations 1 and 3 (NO2), and station 1 (NO3). Chlorophyll-a varied from 1.56 μg/L at stations 1 and 3 to 9.97 μg/L at station 6. Correlations between human development and water quality parameters Water quality measurements were plotted against station, and linear and polynomial curves were tested for goodness of fit. The effect of human development on water quality can be seen by a polynomial curve fitted against the data (see Supplemental File 1, available online at http://www.eaglehill.us/CANAonline/suppl-files/ C164-Schultz-s1). With the 2 relatively pristine stations at both ends of the transect, a second-order curve should show any significant relationship between water quality and station. If no significant relationship exists or if a linear relationship better describes the relationship, then human development could be said to have little effect on that parameter. Of the 11 parameters measured, 7 have significant r2 values (P < 0.05) when a polynomial regression is performed (Table 3). These include temperature, dissolved oxygen, conductivity, salinity, pH, chlorophyll -a, and phosphate. In all cases, the second-order polynomial curve is a better fit than a linear model as determined by a larger r2 value (see Supplemental File 1, available online at http://www.eaglehill.us/CANAonline/suppl-files/C164-Schultz-s1). If station 4 is omitted as an outlier due to it being the only station impacted by the passage of a boat, the relationship between human development and nitrate is also seen to be significant (Table 3). Correlations between bacterial genera and water quality parameters Correlations between genera percent abundances and physical parameters for each station were calculated for the 18 most abundant genera. A P-value < 0.05 was Table 3. R2 values for polynomial curves between water quality parameters and stations along transect. An asterisk indicates a significant relationship (P < 0.05). DO = dissolved oxygen, TSS = total suspended solids, and Chl-a = chlorophyll-a. R2 values Parameter All stations Station 4 omitted Temp (°C) 0.8946* 0.8867* DO (%) 0.7593* 0.7965* Conductivity (mS/cm) 0.8388* 0.9305* Salinity (ppt) 0.6791* 0.7453* pH 0.5057* 0.4184* Turbidity (NTU) 0.2096 0.1178 TSS (mg/L) 0.3548 0.3398 Chl-a (μg/L) 0.4714* 0.4812* PO4 (mg-P/L) 0.5158* 0.7751* NO3 (mg-N/L) 0.3555 0.5057* NO2 (mg-N/L) 0.0854 0.2021 Caribbean Naturalist G.E. Schultz Jr., J.J. Kovatch, and H. Hertler 2017 No. 42 10 considered a strong correlation. Percent dissolved oxygen concentration (DO), conductivity, salinity, and chlorophyll-a were the only physical factors that correlated with the relative abundance of any genus. These parameters were also significantly correlated with development (Table 3). DO was strongly and negatively correlated with the relative abundance of Candidatus Thiobios , Citreicella, Nautella, Fluviicola, Tamlana, Neptunomonas, and Thiomicrospira (r < - 0.6, P < 0.05), and strongly and positively correlated with Cyanobacterium (r > 0.8, P < 0.05). Conductivity was strongly positively correlated with Candidatus Thiobios, Tamlana, Neptumonas, and Thiomicrospira (r < -0.6, P < 0.05). Salinity was strongly negatively correlated with Candidatus Thiobios, Tamlana, Fluviicola, Neptumonas, and Thiomicrospira (r < -0.7, P < 0.05), and strongly positively correlated with Cyanobacterium (r > 0.7, P < 0.05). The only other strong correlation seen was a positive relationship between Synechococcus and chlorophyll-a (r > 0.7, P < 0.05). Similarity (β diversity) between stations We determined the similarity among the various stations at the OTU-level using the Bray-Curtis similarity index. Similarity was lowest between station 1 (undeveloped) and the mid-transect (developed) stations (Table 4). Similarity was highest between adjacent or nearly adjacent developed stations (Table 4). Multidimensional scaling analysis supported these findings (Fig. 3). Undeveloped and up-current station 1 did not cluster with any of the other stations. The 3 least developed stations (1, 2, and 22) lay outside the cluster formed by the developed stations (Fig. 3). However, station 4 also did not cluster with the developed stations. This result may be due to the large amount of suspended material in the channel at station 4. A boat passed through the station immediately before sampling, and TSS at station 4 was higher than at other stations (Table 2). Diversity measurements Sample rarefaction decreased richness estimates (comparable richness) at every station (Table 1) due to removal of OTUs from analysis. Comparable OTU richness was highest at the undeveloped stations 1 and 22 (Table 1). Alpha diversity, as measured by Shannon and Simpson’s diversity indices, was also highest at the undeveloped stations 1 and 22 (Table 1, Fig. 4). Comparable richness was lowest at the developed stations 11 and 14, whereas diversity indices were lowest at the Table 4. Bray-Curtis similarity coefficients for bacterial diversity. The similarity between bacterial communities at each station is shown. Identical communities have a value of 1. Station 1 2 3 4 5 14 11 6 2 0.69 1 3 0.60 0.90 1 4 0.77 0.91 0.82 1 5 0.57 0.87 0.97 0.80 1 14 0.51 0.81 0.91 0.73 0.93 1 11 0.54 0.84 0.94 0.76 0.96 0.97 1 6 0.54 0.84 0.92 0.77 0.94 0.93 0.95 1 22 0.75 0.89 0.82 0.97 0.79 0.74 0.76 0.77 Caribbean Naturalist 11 G.E. Schultz Jr., J.J. Kovatch, and H. Hertler 2017 No. 42 developed stations 5 and 14 (Table 1, Fig. 4). The comparable OTU richness and alpha diversity at station 4 was the highest among the developed stations. As is typical, the diversity function for all of the stations sampled suggested that the actual richness or asymptotic richness was higher than observed (Fig. 5). Proteobacteria vs. Cyanobacteria Bacteria from the phylum Proteobacteria (70%) dominated station 1, with a large fraction of Cyanobacteria and Bacteroidetes present (15% and 14%, respectively). As the current passed stations adjacent to human developments, Proteobacteria percentages decreased, while Cyanobacteria increased to as high as 65% of the total. Upon reaching the undeveloped area down current (station 22), the percentage of Proteobacteria was again highest and that of Cyanobacteria had fallen (Fig. 6A). A linear regression between the relative abundance of these 2 phyla (Fig. 6B) shows Figure 3. Multidimensional scaling diagram showing the degree of similarity between bacterial communities along the longshore transect. Community similarity (Bray-Curtis) was calculated with relative abundance of operational taxonomic units (OTUs). Stations are divided by degree of development and whether Cyanobacteria or Proteobacteria dominate relative abundance. Dim = dimension. Caribbean Naturalist G.E. Schultz Jr., J.J. Kovatch, and H. Hertler 2017 No. 42 12 a significant, negative relationship (n = 8, F = 278.58, P < 0.05, r2 = 0.98). To examine the impact of this relationship on bacterial community diversity, we plotted Simpson’s inverse diversity against the abundance ratio of the 2 phyla (Fig. 7). A strong linear relationship (r2 = 0.96; P < 0.05) was evident (Fig. 7). Figure 4. Shannon’s (blue) and Simpson’s (red) diversity indices across the east (station 1) to west (station 22) shoreline transect. Figure 5. Rarefaction curves for bacterial richness from all samples. Individual samples were rarified to 9500. All rarefaction curves were generated with a step-size of 300 reads, no replacement, and 10 iterations. Caribbean Naturalist 13 G.E. Schultz Jr., J.J. Kovatch, and H. Hertler 2017 No. 42 Discussion Comparison to other nearshore bacterial communities In other pyrosequencing studies of coastal marine bacterial communities, Proteobacteria numerically dominated nearshore bacterial communities while Figure 6. Relative abundances of Cyanobacteria and Proteobacteria, shown as a function of sampling stations (A), demonstrate an inverse relationship between the two phyla (B) (n = 8, F = 278.58, P less than 0.05, r2 = 0.98). Caribbean Naturalist G.E. Schultz Jr., J.J. Kovatch, and H. Hertler 2017 No. 42 14 Cyanobacteria were detected at low relative abundances or not at all (Ortega- Retuerta et al. 2013, Rodriguez-Mora et al. 2015, Thompson et al. 2011, Villa- Costa et al. 2012). In this study, however, Cyanobacteria were dominant at 6 of the 9 stations, with Proteobacteria making up the next most dominant phylum (Fig. 2). Proteobacteria was the most abundant phylum at the other 3 stations, with Cyanobacteria the next most abundant. A recent study of the free-living bacteria in coastal water of Igoumenitsa Gulf, a semi-enclosed bay in Greece, also showed a Proteobacteria-dominated bacterial community (Mezitia et al. 2015). In more directly comparable bacterial communities seen along Latin American coasts, this same pattern of a high Proteobacteria relative abundance coupled with a lower relative abundance for Cyanobacteria was observed, including in a Puerto Rico sample (Thompson et al. 2011). However, as these were all single samples obtained from a single site at each coast (Thompson et al. 2011) and since samples collected for this study included sites with both greater cyanobacterial abundance and sites with greater Proteobacteria abundance, no conclusions can be reached. These differences illuminate, however, the need for more studies to fully characterize coastal regions over larger temporal and spatial scales. Comparing diversity estimates alongside comparisons of taxonomy is also informative. However, comparisons of diversity estimates between studies should only be attempted on samples with similarly sized sequencing depths (Gihring et al. 2012). In addition, Lundin et al. (2012) found that estimates of alpha diversity at sequencing depths of at least 5000 agree with deeper sequencing. Under these criteria (studies with sequencing depths > 5000 reads), appropriate comparisons can be made with 454 pyrosequencing studies performed in the northwest Mediterranean (Pommier et al. 2010) and with the study performed in the Igoumenitsa Figure 7. Simpson’s diversity plotted against the ratio of Proteobacteria abundance to Cyanobacteria abundance (n = 8, F = 177.36, P less than 0.05, r2 = 0.96). Caribbean Naturalist 15 G.E. Schultz Jr., J.J. Kovatch, and H. Hertler 2017 No. 42 Gulf in Greece (Meziti 2015). The overall bacterial community diversity of the nearshore systems included in the present study was lower (Table 1) than that found in the northwest Mediterranean, where the Shannon’s index of coastal surface water was over 5 (Pommier et al. 2010). This value is almost twice as high as at any of the 9 stations in the current study (Table 1). Shannon’s index of the Igoumenitsa Gulf ranged from ~ 2 to 3, which is on the same order as the values in the present study (Table 1). The NW Mediterranean study station was moderately impacted by allocthonous contributions (Pommier et al. 2010), while the study stations in Greece were in a semi-enclosed bay strongly impacted by human development. From these limited observations, one may speculate that changes in bacterial community diversity are potential indicators of stress to an ecosystem. However, care must be taken when comparing sites outside the same geographic region. A site described as moderately impacted in one region may be seen as relatively pristine in another. Potential impact of development on diversity and similarity Water quality data that was collected during this study showed a relationship to development and to bacterial community composition (Table 3, Fig. 3). We also know from data collected between 1996 and 2002 that the sediments collected nearest to marinas and new development had metal concentrations of copper, nickel, and zinc that exceeded values reported to cause biological impairment (Hertler et al. 2009), which is indicative of development impacts. In such a heterogeneous region, it is impossible to measure all physico-chemical factors that might affect bacterial community composition. These, or other water quality parameters, have the potential to influence the bacterial community that flows past these developed areas. Thus, these factors, as well as the water quality parameters we tested, may be changed by and/or cause changes in bacterial community composition. The differences in diversity and richness, the clustering of stations in the MDS plot, and the order of curves in the rarefaction plot all indicate that the observed changes in the bacterial community were associated with development. Further studies are needed to elucidate the details of these interactions. The 2 most diverse stations were the relatively pristine stations 1 and 22. The lowest Shannon’s diversity value was seen at station 14, which is in the developed section near a dock used by local fishermen. In addition, richness and diversity increase with distance away from station 14, located within the human-impacted region (Table 1, Fig. 4). This decline in diversity may be caused by some factor(s) present in the developed area. Alternatively, approaching the developed area may have minimal impact on diversity, but diversity may decrease with distance away from undeveloped areas. In either case, Bray-Curtis similarities along the transect support the hypothesis that in this region, bacterial diversity is lower near human development relative to undeveloped areas (Table 1). The Bray-Curtis similarity between station 1 (undeveloped) and the down-current stations decreases with distance from station 1 (Table 4) until station 14, then increases to station 22. This pattern of decreasing similarity between stations from undeveloped to developed Caribbean Naturalist G.E. Schultz Jr., J.J. Kovatch, and H. Hertler 2017 No. 42 16 and then increasing to undeveloped may indicate that the bacterial community is impacted by the differing water quality in the developed region. It is also informative to note that multidimensional scaling broadly clusters station 1 with the undeveloped stations, but clearly farther away from the developed clusters than station 22. Because station 1 is up-current of the developed stations, factors in the developed area do not affect station 1. However, even though station 22 is not associated with development, the flow brings the affected water down-current, and thus, station 22 is likely to be affected by the development. Station 4 was more diverse than the other developed areas and deserves special consideration. Water quality in station 4 was quantitatively different from the other stations along the transect (Table 2). A boat passed through this station immediately before sampling, and this disturbance may have affected the results. The turbidity values and total suspended solids values were highest at station 4. The large concentration of particles in the water column here may provide an explanation for the increase in diversity seen at this station relative to the other developed stations. Next-generation pyrosequencing has shown that the particle-attached bacterial community is typically more rich and more diverse than the free-living community (Crespo et al. 2013, Yung, et al. 2016). Therefore, the turbid water of station 4 may have been richer and more diverse due to a larger particle-attached community. Alternatively, there may be an unknown intrinsic difference between station 4 and the other stations. Regardless of whether the data were reduced through rarefaction, the slopes of the rarefaction curves show again that the stations farthest from development (1 and 22) were the most taxonomically rich and diverse stations (Fig. 5), with station 14 being the least rich and least diverse station. The lower bacterial diversity seen in the developed stations relative to the undeveloped stations may be due to several potential loss mechanisms. Typical loss mechanisms may include poor water quality, predation, and viral lysis. Predation and viral lysis have been shown to have potentially significant impacts on bacterial community structure in aquatic systems (e.g., Berdjeb et al. 2011, Pernthaler 2005, Šimek et al. 2007, Zhang et al. 2007), but it is unknown whether either changed significantly in this area. Changes in the physical water quality parameters were shown to correlate with development along the transect (Table 3; see Supplemental File 1, available online at http://www.eaglehill.us/CANAonline/ suppl-files/C164-Schultz-s1). Thus, one or some combination of these 3 loss mechanisms might be the driver of the observed changes in diversity. The bacterial community evenness of the longshore current was also highest in the 2 undeveloped regions (0.62 and 0.56, stations 1 and 22 respectively; Table 1). A site with lower evenness means that the organisms present, and their associated genes, are not evenly distributed. If a site has fewer of a particular species present, then their associated set of genes would be fewer as well. Thus, by definition there would be fewer genes available to the community in the less even site. Even if all the species are present in both sites, there will be distinct differences in the number of genes between sites. It is possible that this difference in genetic potential may Caribbean Naturalist 17 G.E. Schultz Jr., J.J. Kovatch, and H. Hertler 2017 No. 42 affect the community’s ability to react to changes in the environment and therefore lead to a less efficient utilization of the resources available to the community. If such a decreased efficiency exists, other trophic levels as well as down-current microbial communities may be impacted. Further experiments and observations are necessary to determine the extent of such potential impacts. Proteobacteria and Cyanobacteria The relationship between Proteobacteria and Cyanobacteria (Fig. 6) warrants future investigation. Proteobacteria are typically the numerically dominant bacteria seen in nearshore regions of Latin America and elsewhere (Thompson et al. 2011). In this study, however, Cyanobacteria dominated the developed regions, which have historically poorer water quality (Hertler et al. 2009). This relationship between the relative abundance of Proteobacteria and Cyanobacteria was very strong, with abundance of the 2 phyla almost perfectly negatively correlated (Fig. 6). This intimate negative relationship appears to impact bacterial community diversity as well (Fig. 7). Simpson’s inverse diversity demonstrates a strong linear relationship with the ratio of Proteobacteria abundance to Cyanobacteria abundance (Fig. 7). Examining the changes in the ratio of Proteobacteria relative abundance to Cyanobacteria relative abundance may shed light on the stress being placed upon the ecosystem. Although the nature of this relationship is currently undefined, it appears that small differences in water quality parameters may be quickly taken advantage of by Cyanobacteria or Proteobacteria at the expense of the other. Care must be taken in these speculations, as correlations between Proteobacteria and Cyanobacteria may be due to autocorrelation biases inherent in the data; as particular bacteria increase in relative abundance, by necessity the relative abundance of other bacteria must decrease in order for relative abundance to sum to 100% (Friedman and Alm 2012). It is important to note, however, that when either Proteobacteria or Cyanobacteria lost relative abundance, despite the presence of many other phyla, it was always the other that gained. Future studies are necessary to more fully elucidate the importance of this relationship. Rare cells and seed-bank It is currently unknown what role rare cells may play in the functioning of this ecosystem or in ecosystems in general. Nemergut et al. (2011) found that most bacteria demonstrate a limited distribution within habitat types. The long “tail” of low-abundance species might constitute a “seed-bank” of species that could become more numerous if conditions changed (Pedrós-Alió 2006, 2007). This situation was demonstrated in the English Channel where a deep read of sequences contained virtually all bacterial sequences seen in greater abundances in the water over a 6-year period (Caporaso et al. 2011). This scenario is also plausible for this system and others that contain waters impacted by different factors that may act as sources of bacteria species (e.g., mangroves, seagrass beds, currents, tides, surface runoff, etc.). In this system, however, Synechococcus is the only genus that became abundant after being rare (from 0.36% at station 1 to as high as 3.41% at station 6). Although the bacterial community contained within the longshore current of this Caribbean Naturalist G.E. Schultz Jr., J.J. Kovatch, and H. Hertler 2017 No. 42 18 study may change and rare bacteria may become abundant as the current continues along the coast, the seed-bank hypothesis does not seem to be relevant at the temporal and spatial scales of this study. Limitations and weaknesses This study relies on single samples collected on one day. In a coastal system, these samples may not be representative of the system at other times or conditions. In addition, with the data collected, it is impossible to pinpoint what factors associated with human development cause these changes in bacterial community composition and diversity. However, whether the community is representative of all conditions or not, it is clear that under the present conditions the bacterial community in proximity to developed areas is different and less diverse than the bacterial community in undeveloped areas. This study also shows that the community composition and diversity recover quickly even over this short (several km) distance and time scale. Conclusions Along this portion of the coast of Puerto Rico, there is little freshwater input (Ewel and Whitmore 1973). The water and associated bacterial community found in the longshore current along the coast are therefore primarily offshore (marine) in origin except during rain events when there may be inputs from intermittent streams and overland runoff. Since these samples were taken during a time of little rainfall and thus little freshwater input, the numerically dominant bacteria in the longshore current will be constrained to those most capable of thriving in the conditions found as the current flows along the coast. Despite limitations to this study, the data show that in this coastal, nearshore current located adjacent to human development, the numerically dominant bacteria are somewhat different from those seen in other Latin American coastal systems (Thompson et al. 2011). Cyanobacteria dominate the system near the developed areas (Fig. 2), and where they are not the most abundant phylum, they are the second most abundant phylum. In other areas of Latin America, Cyanobacteria make up a much smaller fraction of the total community (Thompson et al. 2011). In the undeveloped stations of this study (1 and 22), however, Proteobacteria dominate, which is more similar to other systems in Latin America (Thompson et al. 2011). This change in community demonstrates a potential impact of human development on the bacterial community of nearshore waters. Small changes in environmental factors may affect the abundance and diversity of the bacterial community. Despite the small relative changes in the physical parameters of the system (Table 2), diversity declined and changes in the dominant taxa of the microbial community occured as the community passed the developed region (Figs. 3, 4, 5, and 6). These results may indicate that bacteria are very sensitive to small changes in water quality and/or to other changes potentially associated with human development not measured in this study. These differences in community structure and diversity could indicate a system under stress and may ultimately change the function and appearance of the system. Caribbean Naturalist 19 G.E. Schultz Jr., J.J. Kovatch, and H. Hertler 2017 No. 42 Acknowledgments This material is based upon work supported by Marshall University and the National Science Foundation under Cooperative Agreement No. OIA-1458952. The authors thank Graciela Ramirez from CECIA for support. The authors also thank Clarence Stringer for invaluable assistance during the sample-collection process. This publication is dedicated to Dr. Jeff Kovatch, who died unexpectedly while this article was under review. Dr. Kovatch was a great friend and colleague with a true love for science and the environment. Literature Cited American Public Health Association (APHA). 1995. American Water Works Association, Water Environment Federation, Standard Methods for the Examination of Water and Wastewater. Washington, DC. Avila, M.L.H., J.A. Suarez-Caabro, and G.C. Goldman. 1979. A historical review of the physical and biological characteristics of the ocean near Puerto Rico, relative to an OTEC Power Plant. University of Puerto Rico, Center for Energy and Environmental Research, Mayaguez, PR, USA. Berdjeb, L., T. Pollet, I. Domaizon, and S. Jacquet. 2011. Effect of grazers and viruses on bacterial community structure and production in two contrasting trophic lakes. BMC Microbiology 11:88. Available online at doi:10.1186/1471-218U-11-88. Bio-Optical Oceanography Laboratory 2017. Monthly climatological summary for June 2010. University of Puerto Rico Mayagüez, PR, USA. Available online at http://biooptics. uprm.edu/docs/davisclimatology/NOAAMOJUN2010.txt. Accessed in 2017. Bloom, S.A. 1981. Similarity indices in community studies: Potential pitfalls. Marine Ecology Progress Series 5:125−128. Bray, J.R., and J.T. Curtis. 1957. An ordination of upland forest communities of southern Wisconsin. Ecological Monographs 27:325–349. Callaway, T.R., S.E. Dowd, T.S. Edrington, R.C. Anderson, N. Krueger, N. Bauer, P.J. Kononoff, and D.J. Nisbet. 2010. Evaluation of bacterial diversity in the rumen and feces of cattle fed different levels of dried distillers grains plus solubles using bacterial tag-encoded FLX amplicon pyrosequencing. Journal of Animal Science 88:3977−3983. Campbell, B.J., and D.L. Kirchman. 2013. Bacterial diversity, community structure and potential growth rates along an estuarine salinity gradient. The ISME Journal 7:210–220. Campbell, B.J., L. Yu, J.F. Heidelberg, and D.L. Kirchman. 2011. Activity of abundant and rare bacteria in a coastal ocean. Proceedings of the National Academy of Sciences 108:12776–12781. Caporaso J.G., K. Paszkiewicz, D. Field, R. Knight, and J.A. Gilbert. 2011. The Western English Channel contains a persistent microbial seed bank. The ISME Journal 6:1089– 1093. Available online at doi:10.1038/ismej.2011.162. Clasen J.L., and J.B. Shurin. 2015. Kelp forest size alters microbial community structure and function on Vancouver Island, Canada. Ecology 96:862–872. Crespo, B.G., T. Pommier, B. Fernández-Gómez, and C. Pedrós-Alió. 2013. Taxonomic composition of the particle-attached and free-living bacterial assemblages in the Northwest Mediterranean Sea analyzed by pyrosequencing of the 16S rRNA. Microbiology Open doi:10.1002/mbo3.92 Crompton, T.R. 1989. Analysis of Seawater. Butterworths, Boston, MA. Pp. 1–30. Dong H.P., Y.G. Hong, S. Lu, and L.Y. Xie. 2014. Metaproteomics reveals the major microbial players and their biogeochemical functions in a productive coastal system in the northern South China Sea. Environmental Microbiology Reports 6:68–-695. Caribbean Naturalist G.E. Schultz Jr., J.J. Kovatch, and H. Hertler 2017 No. 42 20 Dowd, S.E., J. Zaragoza, J.R. Rodriguez, M.J. Oliver, and P.R. Payton. 2005. Windows. NET network distributed basic local alignment search toolkit (W.ND-BLAST). BMC Bioinformatics 6:93−106. Dowd, S.E., T.R. Callaway, R.D. Wolcott, Y. Sun, T. McKeehan, R.G. Hagevoort, and T.S. Edrington. 2008. FLX amplicon pyrosequencing (bTEFAP). BMC Microbiology 8:125−132. Edgar, R.C. 2010. Search and clustering orders of magnitude faster than BLAST. Bioinformatics 26:2460−2461. Environmental Quality Board. 1972. A water quality study of La Parguera and nearby Bahia Fosforescente. Office of the Governor, San Juan, PR, USA. Ewel, J.J., and J.L. Whitmore. 1973. The ecological life zones of Puerto Rico and the US Virgin Islands.USDA Forest Service, International Institute of Tropical Forestry, Research Publication. ITF-18, Rio Pedras, PR, USA. Faith, D.P., P.R. Minchin, and L. Belbin. 1987. Compositional dissimilarity as a robust measure of ecological distance. Vegetation 69:57−68. Fortunato, C.S., L. Herfort, P. Zuber, A.M. Baptista, and B.C. Crump. 2012. Spatial variability overwhelms seasonal patterns in bacterioplankton communities across a river-toocean gradient. The ISME Journal 6:554–563. Friedman, J., and E.A. Alm. 2012. Inferring correlation networks from genomic survey data. PLOS Computational Biology doi:10.1371/journal.pcbi.1002687. Gihring, T.M., S.J. Green, and C.W. Schadt. 2012. Massively parallel rRNA gene sequencing exacerbates the potential for biased community-diversity comparisons due to variable library sizes. Environmental Microbiology 14:285−290. Gontcharova, V., E. Youn, R.D. Wolcott, E.B. Hollister, T.J. Gentry, and S.E. Dowd. 2010. Black Box Chimera Check (B2C2): A Windows-based software for batch depletion of chimeras from bacterial 16S rRNA gene datasets. The Open Microbiology Journal 10:47–52. Hertler, H. 2002. The implication of resource management in La Parguera, Puerto Rico. Ph.D. Dissertation. Drexel University, Philadelphia, PA, USA. Hertler, H., J. Spotila, and D.A Kreeger. 2003. Effects of houseboats on organisms of the La Parguera Reserve, Puerto Rico. Environmental Monitoring and Assessment 98(1– 3):319–407. Hertler, H., A.R. Boettner, G.I. Ramírez-Toro, H. Minnigh, J. Spotila, and D. Kreeger. 2009. Spatial variability associated with shifting land use: Water quality and sediment metals in La Parguera, Southwest Puerto Rico. Marine Pollution Bulletin 58:672–678. Kirchman, D. L. 2008. Microbial Ecology of the Oceans. 2nd Edition. Oxford University Press, New York, NY, USA. Lundin, D., I. Severin, J.B. Logue, O. Östman, A.F. Andersson, and E.F. Lindström. 2012. Which sequencing depth is sufficient to describe patterns in bacterial α- and β-diversity? Environmental Microbiology Reports 4:367−372. Martiny, J.B., B.J. Bohannan, J.H. Brown, R.K. Colwell, J.A. Fuhrman, J.L. Green, M.C. Horner-Devine, M. Kane, J.A. Krumins, C.R. Kuske, P.J. Morin, S. Naeem, L. Ovreås, A.L. Reysenbach, V.H. Smith, and J.T. Staley. 2006. Microbial biogeography: Putting microorganisms on the map. Nature Reviews Microbiology 4(2):102–112. Meziti, A., K.A. Kormas, M. Moustaka-Gounic, and H. Karayanni. 2015. Spatially uniform but temporally variable bacterioplankton in a semi-enclosed coastal area. Systematic and Applied Microbiology 38:358–367. Caribbean Naturalist 21 G.E. Schultz Jr., J.J. Kovatch, and H. Hertler 2017 No. 42 Nemergut, D.R., E.K. Costello, M. Hamady, C. Lozupone, L. Jiang, S.K. Schmidt, N. Fierer, A.R. Townsend, C.C. Cleveland, L. Stanish, L., and R. Knight. 2011. Global patterns in the biogeography of bacterial taxa. Environmental Microbiology 13(1):135–144. Available online at doi:10.1111/j.1462-2920.2010.02315.x. Oksanen, J., R. Kindt, P. Legendre, and R.B. O’Hara. 2007. Vegan: Community ecology package for R, version 1.9-25. R Foundation for Statistical Computing, Vienna, Austria. Available online at http://cran.r-project.org. Ortega-Retuerta, E., F. Joux, W.H. Jeffrey, and J.F. Ghiglione. 2013. Spatial variability of particle-attached and free-living bacterial diversity in surface waters from the Mackenzie River to the Beaufort Sea (Canadian Arctic). Biogeosciences 10:2747–2759. Available online at doi:10.5194/bg-10-2747-2013. Pedrós-Alió, C. 2006. Marine microbial diversity: Can it be determined? Trends in Microbiology 14:257−263. Pedrós-Alió, C. 2007. Dipping into the rare biosphere. Science 315:192−193. Pernthaler, J. 2005. Predation on prokaryotes in the water column and its ecological implications. Nature Reviews Microbiology 3:537−546. Piccini, C., and J. García-Alonso. 2015. Bacterial diversity patterns of the intertidal biofilm in urban beaches of Río de la Plata. Marine Pollution Bulletin 91:476–82. Available online at doi:10.1016/j.marpolbul.2014.08.039. Pielou, E.C. 1969. An Introduction to Mathematical Ecology. Wiley-Interscience, New York, NY, USA. Pielou, E.C. 1975. Ecological diversity. Wiley, New York, NY, USA. Pommier, T., P.R. Neal, J.M. Gasol, M. Coll, S.G. Acinas, and C. Pedrós-Alió. 2010. Spatial patterns of bacterial richness and evenness in the NW Mediterranean Sea explored by pyrosequencing of the 16S rRNA. Aquatic Microbial Ecology 61:221–233. Available online at doi:10.3354/ame01484. R Development Core Team. 2007. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Available online at https:// www.R-project.org. Rodriguez-Mora, M.J., M.I. Scranton, G.T. Taylor, A.Y. Chistoserdov, and P. Sobecky. 2015. The dynamics of the bacterial diversity in the redox transition and anoxic zones of the Cariaco Basin assessed by parallel tag sequencing. FEMS Microbiology Ecology 91:fiv088. Available online at doi:10.1093/femsec/fiv088. Sale, P.F., T. Agardy, C.H. Ainsworth, B.E. Feist, J.D. Bell, P. Christie, O. Hoegh-Guldberg, P.J. Mumby, D.A. Feary, M.I. Saunders, T.M. Daw, S.J. Foale, P.S. Levin, K.C. Lindeman, K. Lorenzen, R.S. Pomeroy, E.H. Allison, R.H. Bradbury, J. Corrin, A.J. Edwards, D.O. Obura, Y.J. Sadovy de Mitcheson, M.J. Samoilys, and C.R.C. Sheppard. 2014. Transforming management of tropical coastal seas to cope with challenges of the 21st century. Marine Pollution Bulletin 85:8–23. Available online at doi:10.1016/j. marpolbul.2014.06.005. Schultz, G.E., Jr., J.J. Kovatch, and E.M. Anneken. 2013. Bacterial diversity in a large, temperate, heavily modified river, as determined by pyrosequencing. Aquatic Microbial Ecology 70:69–179. Available online at doi:10.3354/ame01646. Sen, R., H.D. Ishak, D. Estrada, S.E. Dowd, E. Hong, and U.G. Mueller. 2009. Generalized antifungal activity and 454-screening of Pseudonocardia and Amycolatopsis bacteria in nests of fungus-growing ants. Proceedings of the National Academy Sciences USA 106(42):17805–17810. Shannon, C.E., and W. Weaver. 1963. The Mathematical Theory of Communication. University of Illinois Press, Urbana, IL, USA. Caribbean Naturalist G.E. Schultz Jr., J.J. Kovatch, and H. Hertler 2017 No. 42 22 Šimek, K., M.G. Weinbauer, K. Hornák, J. Jezbera, J. Nedoma, and J.R. Dolan. 2007. Grazer and virus-induced mortality of bacterioplankton accelerates development of Flectobacillus populations in a freshwater community. Environmental Microbiology 9:789−800. Simpson, E.H. 1949. Measurement of diversity. Nature 163:688. Störmer, R., A. Wichels, and G. Gerdts. 2013. Geo-chip analysis reveals reduced functional diversity of the bacterial community at a dumping site for dredged Elbe sediment. Marine Pollution Bulletin 77:113–122. Available online at doi:10.1016/j.marpolbul. 2013.10.022. Strickland, J.D.H., and T.R. Parson. 1972. A Practical Handbook of Seawater Analysis, 2nd Edition. Bulletin of Fisheries Research Board of Canada, No. 67. Ottawa, ON, Canada. Suchodolski, J.S., S.E. Dowd, E. Westermarck, J.M. Steiner, R.D. Wolcott, T. Spillmann, and J.A. Harmoinen. 2009. The effect of the macrolide antibiotic tylosin on microbial diversity in the canine small intestine as demonstrated by massive parallel 16S rRNA gene sequencing. BMC Microbiology 9:210−225. Sun, M.Y., K.A. Dafforn, M.V. Brown, and E.L. Johnston. 2012. Bacterial communities are sensitive indicators of contaminant stress. Marine Pollution Bulletin 64:1029–1038. Available online at doi:10.1016/j.marpolbul.2012.01.035. Sun, M.Y., K. A. Dafforn, E.I. Johnston, and M.V. Brown. 2013. Core sediment bacteria drive community response to anthropogenic contamination over multiple environmental gradients. Environmental Microbiology 15:2517–2531. Avaialble online at doi:10.1111/1462-2920.12133. Thompson, F.L., T. Bruce, A. Gonzalez, A. Cardoso, M. Clementino, M. Costagliola, C. Hozbor, E. Otero, C. Piccini, S. Peressutti, R. Schmieder, R. Edwards, M. Smith, L.R. Takiyama, R. Vieira, R. Paranhos, and L.F. Artigas. 2011. Coastal bacterioplankton community diversity along a latitudinal gradient in Latin America by means of V6 tag pyrosequencing. Archives of Microbiology 193:105–114. United States Geological Survey. 2017. USGS 50129254 Rio Loco at Las Latas nr La Joya nr Guanica, PR. Available online at https://nwis.waterdata.usgs.gov/nwis/monthly?site_ no=50129254&por_50129254_123739=2377867,00060,123739,2006-12,2016- 02&format=html_table&date_format=YYYY-MM-DD&rdb_ compression=file&submitted_form=parameter_selection_list. Accessed in 2017. Vila-Costa, M., J.M. Gasol, S. Sharma, and M.A. Moran. 2012. Community analysis of high- and low-nucleic acid-containing bacteria in NW Mediterranean coastal waters using 16S rDNA pyrosequencing. Environmental Microbiology 14(6):1390–1402. Available online at doi:10.1111/j.1462-2920.2012.02720.x. Yung, C.-M., C.S. Ward, K.M. Davis, Z.I. Johnson, and D.E. Hunt. 2016. Insensitivity of diverse and temporally variable particle-associated microbial communities to bulk seawater environmental parameters. Applied and Environmental Microbiology 82:3431–3437. Zhang, R., M.G. Weinbauer, and P.Y. Qian. 2007. Viruses and flagellates sustain apparent richness and reduce biomass accumulation of bacterioplankton in coastal marine waters. Environmental Microbiology 9:3008−3018.