2.1. Study system
Southern Biscayne Bay is a shallow (b 3 m) subtropical estuary located at the southeastern tip of the Florida peninsula (Fig. 1 ). The extensive seagrass communities in southern Biscayne Bay are dominated by dense T. testudinum. S. filiforme and H. wrightii are also found throughout this area in lower abundance and with patchy distributions. This study was conducted at four shoals: Biscayne Channel, No Name Shoal, East Featherbed Bank, and Cutter Bank (Fig. 1 ), which
are characterized by thriving seagrass communities, but have been impacted by multiple vessel grounding injuries over the past two decades. Vessel groundings remove seagrass and excavate sediment in discrete patches on shallow seagrass shoals. Existing vessel grounding injuries selected at each shoal as study sites were unvegetated patches surrounded by a dense seagrass community, and ranged from 40 to 60 m2.
2.2. Experimental design
Seagrass and fish community surveys, herbivory assays, and elemental analysis of seagrass leaf tissue were conducted during the Spring of 2009 through the Winter of 2010.
2.3. Seagrass community surveys
Seagrass community composition was documented in undisturbed seagrass habitat in the vicinity (within 100 m) of each herbivory assay site. Seagrass and macroalgae abundance was estimated within randomly placed 0.25-m2 PVC quadrats according to a modified Braun–Blanquet cover-abundance scale (Fourqurean et al., 2001 ). The number of quadrats varied per site (18–268), as did the timing and frequency of monitoring efforts. For each site, quadrat data from all available monitoring events were pooled for analysis. At No Name Shoal, data from a single monitoring event conducted in March 2011 were available (111 quadrats). At Biscayne Channel, East Featherbed Bank, and Cutter Bank, four monitoring events were conducted at each site
during the 20-month period from December 2009 through July 2011, providing 71, 400, and 1070 total quadrats for analysis, respectively.
2.4. Fish community surveys
Two fish surveys were conducted at each assay site in October 2009 (Fall09) and in February–March 2010 (Win10), respectively (4 sites × 2 events × 2 surveys per site per event = 16 fish surveys) using a modification to the roving diver technique (Schmitt and Sullivan, 1996). For each survey, a snorkeler slowly surveyed one half of a 50-diameter circle (total survey area approximately 980 m2), as delimited with a 25-m transect tape attached to a central stake, for 30 min. The snorkeler counted fish observed within the survey area and identified them to species (when possible) or genus. Water temperature and salinity were measured for each survey using a YSI Model 30 instrument. Fish taxa expected to directly consume seagrass were categorized as seagrass herbivores (Ferreira and Floeter, 2004; Floeter, 2004; Valentine and Duffy, 2006). Note that while sea urchins are known to be important consumers of seagrass biomass, few sea urchins were observed at the study sites. As such, the role of sea urchins in removing seagrass biomass was not evaluated in this study, and was suspected to be minimal.
2.5. Herbivory assays
Herbivory pressure was examined through herbivore exclusion in a year-long study conducted in 2009–2010. A full factorial design was employed, with the following factors and levels: location (Biscayne Channel, E. Featherbed, Cutter Bank, No Name Shoal); seagrass species (T. testudinum, H. wrightii, S. filiforme); time (spring, summer, fall, winter), and herbivore exclusion (no cage, partial cage, full cage). Planting units (PUs) designed to mimic the growth form and
appearance of natural seagrass (sensu Hay, 1981; Kirsch et al., 2002) were used as experimental units in the herbivory assays. PUs consisted of freshly-harvested seagrass leaf material inserted into simulated short shoots (binder clips) affi xed to a simulated rhizome (labeled semi-rigid 30-cm plastic strip cut from cable tie). For deployment, the “ rhizome” of the PU was gently pressed into bottom sediments and held in place with a U-shaped sod staple. This approach
wa s b en efi cial for three reasons. It enabled the standardization of plant leaf tissue biomass used in each PU across species. It also eliminated consideration of non-photosynthetic and belowground tissue biomass that may mask leaf tissue biomass loss. Lastly, impacts to donor beds were minimized by removing only leaf material but leaving belowground tissues intact.
Seagrass shoots for use in constructing PUs were collected from a single donor site: Pelican Shoal (Fig. 1 ). Collecting donor material from a single location ensured that a sufficient supply of all three seagrass species was available for all assays over the year-long study. Shoots were collected on the morning of each assay deployment. Intact green leaves were selected from the collection for PUs; age of leaf tissue was not considered. Leaf tissue for each PU was
spun in a salad spinner to remove excess water, gently wiped clean of epiphytes, and patted dry. Leaf tissue biomass was standardized across the three seagrass species, and approximately 1.5 g of fresh seagrass leaf material was used for each PU. After weighing, the leaf tissue for each PU was carefully clipped into the set of binder clips attached to each plastic rhizome. PUs were then placed into individual zip lock bags flooded with fresh seawater, for transport to the study site in seawater-filled coolers. Herbivore exclusion cages were constructed of 0.635-cm black plastic aquaculture mesh. This mesh size was chosen as appropriate for excluding juvenile and adult fish of species known to directly consume seagrasses in the subtropical western Atlantic such as pinfishes and parrotfishes (Valentine and Heck, 1999). Full cages were formed
of cylinders approximately 40 cm in height with a flat top approximately 30 cm in diameter. To control for possible influences of the cage structures on the PUs not related to herbivory, we also deployed partial cages as “cage controls”. Partial cages had two sides and a top, and were open on two sides. The partial cages were approximately 30 cm × 30 cm × 30 cm in size. Full and partial cages had lengths of lead-core line cable-tied to the bottom edges that rested on the
sediment to help keep them upright. U-shaped sod staples were inserted through mesh into the sediments to further stabilize the cages. A cage contained a single PU. Ten replicate PUs per seagrass species were randomly assigned to each of the three herbivore exclusion treatments, for a total of 90 PUs (per assay deployment, per site). Each PU with its respective caging treatment was randomly placed at 0.5 m intervals in rows in existing grounding injury features at each location. Placing PUs in the unvegetated injuries was intended to mimic a transplanting scenario. The configuration (number and length of rows) of the array varied according to the injury shape. In each assay, PUs were deployed for a period of 72 h. Upon retrieval, the PUs were disassembled in the lab. Plant material was again spun and gently patted dry, then reweighed. Change in biomass was Fig. 1. Location of study sites in southern Biscayne Bay, Biscayne National Park, Florida, USA. A.S. Bourque, J.W. Fourqurea n / Journal of Experiment al Marine Biology and Ecology 445 (2013) 29– 37 31 expressed as the percent biomass loss per planting unit, determined by calculating the proportional loss in wet weight of each planting unit following the assay. Herbivory assays were conducted at each site four times over a one-year period in 2009– 2010: in May– June 2009 (Spr09), August 2009 (Sum09), November 2009 (Fall09), and February– March 2010 (Win10). These time periods were selected in order to coincide with warm (summer), cold (winter), and intermediate (spring, fall) water temperatures in this subtropical system. During each event, water temperature and salinity were measured at each site using a YSI Model 30 instrument.
2.6. Seagrass elemental composition
Seagrass leaf tissue was collected from each study site and the donor site for elemental content (total carbon = C, total nitrogen =N, total phosphorus = P) analyses. Leaf tissue was collected in triplicate (one replicate = six T. testudinum leaves, thirty S. filiforme leaves, or thirty H. wrightii leaves) during each of the four seasonal assays, from all seagrass species present at the study site at that time (S. filiforme and H. wrightii were not observed at every assay). Seagrass leaves were gently scraped to remove epiphytes, dried at 80 °C, and ground to a fine powder in a ball mill grinder. Total phosphorus content was determined through a dry-oxidation, acid hydrolysis extraction followed by colorimetric analysis of phosphate concentration in the
extract (Fourqurean et al., 1992). Total carbon and total nitrogen were determined using a CHN analyser. Elemental content was calculated on a dry weight basis (mass of element/dry weight of sample) × 100%. Elemental ratios were calculated as molar ratios.