Evaluation of Florida Bay Seagrass After Hurricane

EVALUATION OF FLORIDA BAY SEAGRASS

AFTER HURRICANE IRMA

AUBREY LYNN LORIA

MICHAEL STEINBERG, COMMITTEE CHAIR JULIA CHERRY NICHOLAS MAGLIOCCA

A THESIS

Submitted in partial fulfillment of the requirements For the degree of Master of Science in the Department of Geography in the Graduate School of The University of Alabama

TUSCALOOSA, ALABAMA

2019

ABSTRACT

When Hurricane Irma made US landfall on the Florida Keys on September 10, 2017, the category 4 storm became the first major hurricane to hit the state since 2005. This study investigated the impacts that Irma had on seagrasses in Florida Bay – an estuary at the southern outlet of the Florida Everglades that is crucial to the environment and economy of South Florida but that is threatened by anthropogenic impacts on water quality. A massive seagrass die-off in

2015 left the bay with hypersaline water and sulfide-saturated sediment that was held in place by the bay’s network of carbonate mud banks. The promise of rainfall and storm surge from a major hurricane offered a chance to refresh the bay and begin the process of recovering from the die- off. By comparing seagrass density and community composition from after the 2015 die-off to after Irma, I found that seagrass cover has become more uniform and that species richness has increased throughout the study area. Seagrass density was lost in some shallow sections of the study area, but pioneer species Halodule wrightii had begun colonizing the disturbed areas, indicating that water quality had changed enough to support the less salt-tolerant species. The positive changes in seagrass density, distribution, and diversity after Hurricane Irma indicate that the storm may have flushed out the hypersaline water left after the 2015 die-off that had been limiting seagrass recovery for two years.

DEDICATION

Dedicated to the memory of my hero, teacher, and grandfather, Don, who meant to make me an artist but accidentally made me a scientist. Don passed away two weeks after I completed the field research for this thesis, but the love and pride he showed for his grandchildren has continued to encourage me in the months since.

iii

LIST OF ABBREVIATIONS AND SYMBOLS

D Density using the Braun-Blanquet Cover-Abundance scale

PSU Practical Salinity Units

RKB Rabbit Key Basin

TKB Twin Key Basin

< Less than

= Equal to

≤ Less than or equal to

ACKNOWLEDGMENTS

This thesis simply could not have been done without all the help I received from my incredible mentors, family, and support system. First, I thank Dr. Michael Steinberg (Committee

Chairman), for accepting my request to join his lab and encouraging me to revisit the work of his former graduate student, Cynthia “Jo” McGinnis. Thank you to Jo for providing the groundwork that makes this comparison so interesting and for making a trip to come speak with me at the

University of Alabama while I was putting together my thesis proposal.

Thank you to Everglades National Park (ENP), South Florida Natural Resources Center

(SFNRC), and the Florida Bay Interagency Science Center (FBISC) for providing logistical support for this thesis. More specifically, thank you to Matt Patterson of the SFNRC for not only coordinating this thesis with NPS and FBISC, but also for making sure I never had any question go unanswered, I never went without any equipment that he could provide, and I never felt alone when I was so far from home. Thank you to Dr. Jed Redwine for advice on several methodologies. Thank you to Troy Mullins for helping me acquire GIS datasets.

Thank you to Eduarda and Mima for providing me with free room and board (respectively) for nine weeks while I conducted my research in South Florida. It was wonderful getting to spend that time with both of you.

Finally, thank you to my parents for my life, my sister for my ambition, and Stirton and

Lita for my courage. Thank you all for getting me here.

CONTENTS ABSTRACT ...... ii

DEDICATION ...... iii

LIST OF ABBREVIATIONS AND SYMBOLS ...... iv

ACKNOWLEDGMENTS ...... v

LIST OF TABLES ...... viii

LIST OF FIGURES ...... ix

CHAPTER 1 INTRODUCTION ...... 1

CHAPTER 2 BACKGROUND ...... 5

2.1 Florida Bay Ecosystem ...... 5

2.2 Florida Bay Hydrology ...... 7

2.3 Plants Studied...... 11

2.4 Hurricane Effects ...... 13

CHAPTER 3 MATERIALS AND METHODS ...... 17

3.1 Study Sites ...... 17

3.2 Sampling Method ...... 19

3.3 Spatial Analysis ...... 22

CHAPTER 4 RESULTS ...... 24

CHAPTER 5 DISCUSSION ...... 38

CHAPTER 6 CONCLUSION...... 41

REFERENCES ...... 42

APPENDIX ...... 47

Accuracy Assessment ...... 47

vii

LIST OF TABLES

Table 1(a-b): Braun-Blanquet assessment of seagrass density (D) for all samples, separated by species and basin. 1a contains pre-Irma data and 1b contains post-Irma. Adapted from Fourqurean et al. (2001)...... 28 Table 2: Root Mean Square Error for each map by basin (when applicable) ...... 47

viii

LIST OF FIGURES

1. Everglades National Park, Florida, USA. Park extent (green outline) and Florida Bay extent (black dashed line) are shown ...... 4

2. Annual mean volume transports (m3 s−1) for measured channels in Florida Bay (red arrows and numbers), through the tidal passes between the Florida Keys (blue arrows and numbers), and from mean river discharge (black arrows and numbers) (Lee et al. 2016) ...... 10

3. Thalassia testudinum (top), Halodule wrightii (middle), and Syringodium filiforme (bottom). Illustrations are by Phillips and Menez (1988) and photographs are by Ken Dunton (University of Texas)...... 11

4. Hurricane Irma's track in relation to Florida Bay and South Florida ...... 13

5. A wrack line composed entirely of dried seagrass, mostly Thalassia testudinum, left by Hurricane Irma. This photo was taken over 300 meters from Flamingo Beach, FL...... 15

6. Florida Bay salinity (PSU) one month before Hurricane Irma, one month after, and five months after (SFWMD 2018)...... 16

7. Twin Key and Rabbit Key Basins and their position within Florida Bay...... 18

8. Bathymetry of Rabbit Key and Twin Key Basins. Contour interval: 0.25 m. Digital Elevation Model provided by National Centers for Environmental Information (NCEI). Vertical error estimated by NCEI to be no more than 15 cm ...... 19

9. Location of sampling sites in RKB and TKB...... 20

10. Total seagrass cover defined by broad cover categories. Bare sediment is 0% cover; Sparse cover is <10%; Moderate cover is 10-50%; dense cover is >50%...... 27

11. Halodule wrightii distribution before (top) and after (bottom) Hurricane Irma ...... 30

12. Syringodium filiforme distribution before (a) and after (b) Hurricane Irma ...... 31

13. Proportion (%) of sites in each basin that had significant positive or negative change in density ...... 32

14. Direction and intensity of change in total seagrass cover after Hurricane Irma...... 33

15. Thalassia testudinum cover before Hurricane Irma (top) and after Hurricane Irma (bottom)...... 34

16. Halodule wrightii density before (top) and after (bottom) Hurricane Irma...... 35

17. Syringodium filiforme density before (top) and after (bottom) Hurricane Irma...... 36

18. Total Seagrass Cover by Braun-Blanquet value for before (top) and after (bottom) Hurricane Irma ...... 37

19. Aerial photography of the study area from one month after Hurricane Irma. Study basins are outlined in yellow. The yellow star on the eastern edge of TKB marks the location of sediment blow-outs...... 39

CHAPTER 1 INTRODUCTION

Florida Bay is a saltwater portion of Everglades National Park positioned between the southern river outlets, the Gulf of Mexico, and the Florida Keys (Figure 1). It supports a tremendous variety of fauna, many of which are critical to the South Florida economy and ecosystem. Not only is Florida Bay home to threatened and endangered species, including four protected species of sea turtles and the endangered American crocodile (Crocodylus acutus) and manatee (Trichechus manatus laterostris) (Florida Museum of Natural History), but it is also a popular fishing destination that contributes over $800 million annually in economic activity and nearly 9,000 jobs to South Florida (Fedler 2009). The sport fish behind this spending, as well as the previously mentioned protected species, all depend, at least in part, on the seagrass communities of Florida Bay for survival (Tilmant 1989; Fourqurean and Robblee 1999).

Hurricane Irma made US landfall through the Lower Florida Keys, with the eye of the hurricane passing within 60 kilometers of the western extent of Florida Bay. Throughout the southeastern US, the storm decimated crops and demolished coastal developments, leaving an estimated $50 billion worth of damage in its wake (Cangialosi et al. 2018). The damage to land structures and vegetation was obvious but what interested South Florida ecologists was more difficult to observe; what effects did Hurricane Irma have on seagrass beds in Florida Bay?

The anthropogenically modified hydrology and polluted waters of Florida Bay are severely limiting to the wetland and benthic communities within it. The past 100 years of coastal development and hydrologic alteration to the Everglades have caused the “chronic state of low

1 flow” in the northern and eastern sections of the bay that contribute to elevated salinity in the estuarine basins (Brandt et al. 2016), while increased evaporation from rising sea surface temperatures leads to a deficit of freshwater in the other subregions (Short and Neckles 1998;

Nuttle et al. 2000; Kelble et al. 2007). Prolonged periods of elevated salinity have been observed to cause monocultures of Thalassia testudinum, which provide less suitable habitat for bay fish species and are susceptible to widespread mortality (Thayer and Chester 1989; Bennet et al.

2005). The combination of high sea surface temperatures, hypoxia, and hypersalinity are believed to be responsible for die-offs of T. testudinum like the one in 2015 (Borum et al. 2005;

Koch et al. 2007; Hall et al. 2016; National Park Service 2016).

In the months before Hurricane Irma struck in 2017, the basins that had been impacted by the 2015 die-off still had water of higher salinity than average ocean water (South Florida Water

Management District 2018) and it was unclear whether the seagrass was recovering (Matt

Patterson, personal communication to the author, November 2017). Hurricanes are known to reduce salinity in Florida Bay and are also expected to play a role in removal of sediment buildup, increasing flow between basins (Zieman et al. 1989; Rudnick et al. 2005; Kelble et al. 2007).

It was predicted that total seagrass cover would increase after Irma and that community composition would become more diverse as the influx in freshwater reduced bay salinity and increased circulation, making conditions more suitable for the two target seagrass species

(Halodule wrightii and Syringodium filiforme) that are less salt-tolerant than T. testudinum.

Braun-Blanquet field survey methods were used to evaluate coverage and composition of seagrass communities in two Florida Bay basins nearly a year after Hurricane Irma, at which point colonization of scoured sediments could be seen. The basins used in this study had been

2 previously assessed using these same field-survey methods to map the extent of the 2015 die-off, the data from which was compared to the post-Irma survey to detect changes in seagrass status.

The purpose of this project is to record the changes in seagrass beds in Florida Bay after

Hurricane Irma. Submersed aquatic vegetation is considered a keystone species in Florida Bay and, as such, is monitored by multiple agencies throughout South Florida (Brandt et al. 2016).

Visualization of the cover and community composition of seagrass beds in Florida Bay after

Hurricane Irma provide insight into the environmental conditions of the ecosystem and can be used to track changes as conditions in the bay change, either as a result of restoration or of climate change. Visualization of areas where there is no seagrass cover may also help guide restoration projects in the future.

Figure 1: Everglades National Park, Florida, USA. Park extent (green outline) and Florida Bay extent (black dashed line) are shown.

CHAPTER 2 BACKGROUND 2.1 Florida Bay Ecosystem

Resting at the southern outlet of the Everglades, Florida Bay is a subtropical body of water that is known for its diverse plants and wildlife. The bay comprises nearly one third of

Everglades National Park and the Marjory Stoneman Douglas Wilderness, giving it the highest level of protection assigned to US federal lands (National Park Service 2009). Florida Bay is a popular destination for ecotourists while also serving as a place for both work and recreation for locals. Sport fish such as tarpon (Megalops atlanticus), bonefish (Albula vupes), red drum

(Sciaenops ocellata), and spotted seatrout (Cynoscion nebulosus) live within the bay, making the site popular for recreational fishing (Tabb and Manning 1961; Tilmant 1989; Powell 2003). In addition, a rich community of piscivorous birds (e.g., wading birds, magnificent frigatebirds, and osprey) and several protected vertebrate species, like American crocodiles (Crocodylus acutus) and green sea turtles (Chelonia mydas), hunt within the bay or graze upon its seagrass beds

(Fourqurean and Robblee 1999). When seagrass beds significantly decrease in density, the bay fauna decrease in abundance and diversity in response (Fourqurean and Robblee 1999).

Florida Bay is important to people from all walks of life, so the widespread impacts from seagrass die-offs are an economic concern as much as an ecologic concern. Changes in climate and Florida Bay hydrology throughout the 20th and 21st century have caused accelerated evaporation and low circulation between basins, resulting in periods of hypersalinity, often at the end of the dry season (November to April) (Fourqurean and Robblee 1999; Koch et al. 2007; Lee

5 et al. 2016), the results of which have twice led to disastrous, widespread seagrass die-offs (Hall et al. 2016). In basins with the lowest amounts of freshwater input, salinity has sometimes reached 70 PSU, which is twice the average concentration of ocean water (Kelble et al. 2007).

While seagrasses are adapted to tolerate saline conditions, the mechanisms that control their osmotic balance reduce photosynthetic efficiency (Touchette 2007) and increase their oxygen demand for respiration (Koch et al. 2007).

The combination of elevated water temperature, decreased oxygen production from photosynthesis, increased oxygen demand in plant tissues, and low circulation in the water column creates hypoxic conditions, which encourages hydrogen sulfide production by anaerobic microbes (Borum et al. 2005; Koch et al. 2007). Recent experiments have also found that, at night, oxygen-deprived root tissue in Thalassia testudinum – the dominant seagrass species in

Florida Bay and the only species to survive long-term exposure to salinity over 40 practical salinity units (PSU) – is especially susceptible to hydrogen sulfide toxicity via intrusion from porewater when exposed to hypersaline conditions (Johnson et al. 2018). The decay of plant tissues in anoxic conditions causes further hydrogen sulfide production, which results in a positive feedback loop of decomposing seagrass leading to the mortality of the surrounding seagrass (Carlson et al. 1994)

Hydrogen sulfide creates a “rotten egg” smell and yellow “fog” in the water (Hall et al.

2016; National Park Service 2016). The smell, along with floating masses of dead seagrass and fish hurt the tourism industry and the decline in sport fish populations hurt the recreational fishing industry. As stated by the Everglades Marine Team, “when Florida Bay is unhealthy, the impacts are felt across South Florida” (National Park Service 2016).

The literature suggests that precipitation from extreme storms can reduce hypersaline conditions and remove sediment and detritus buildup in Florida Bay, relieving some of the chemical stress on seagrass (Davis et al. 2004; Rudnick et al. 2005; Steward et al. 2006). This would improve seagrass productivity, allow the less salt-tolerant seagrass species to populate the seagrass communities, and provide higher-quality habitat for marine life (Thayer and Chester

1989; Bennett et al. 2005; Touchette 2007). However, these weather events also cause physical stress on plants from wave activity and scouring and may reduce light availability to submerged vegetation with sediment suspension and algal blooms from the released sediment nutrients

(Fonesca et al. 1998; Fourqurean et al. 1999; Davis et al. 2004; Rudnick et al. 2005; Steward et al. 2006; Carlson et al. 2010; Cole et al. 2018).

2.2 Florida Bay Hydrology

Starting in 1881, the Florida Everglades underwent over 100 years of hydrologic alterations in order to reclaim wetlands to accommodate the growing population of South Florida

(Light and Dineen 1994). The most intensive alterations, the Central & South Florida Project, were completed in the 1950s-1970s; the project installed a network of canals, levees, pumps, and water conservation areas with the goal of preventing floods and augmenting the local water supply (Light and Dineen 1994; Ogden et al. 2005). Because of the hydrologic diversion, 6.4 billion kilograms of water bypass the Everglades, leaving the system with only 30% of its original flow (USFWS 1999). The reduced flow is often credited for the high salinity and increased salinity variability in northern sections of Florida Bay and influencing community composition (Bennett et al. 2005; Rudnick et al. 2005).

While these developments allowed massive population growth and urbanization in South

Florida, by the early 1980s public concern for the ecosystem inspired changes in water

7 management, starting with Florida Governor Graham’s launch of the “Save Our Everglades” program in 1983 (Light and Dineen 1994). In 2000, the Comprehensive Everglades Restoration

Plan (CERP) set the guidelines for an interagency effort to restore the Everglades ecosystem functions by imitating historic hydrology patterns (USACE and SFWMD 1999).

One of the many challenges of Everglades restoration is determining what the historic hydrology patterns were; the quality and quantity of water in the freshwater Everglades and

Florida Bay were not well documented before the flow was diverted, so restoration efforts depend upon statistical models and paleoecologic proxy data to estimate historic conditions

(Rudnick et al. 2005). The simulated historic conditions calculated by Rudnick et al. (2005) were found to deliver 2.5-4 times as much freshwater to estuaries and wetlands as modern flow, resulting in salinities in modern Florida Bay that are 5.3-20.1 PSU higher than before the anthropogenic changes were made.

The complex network of mudbanks in Florida Bay restricts flow between basins, creating a patchwork of variable water quality throughout the bay. Because of current hydrologic conditions, only the northeastern section of the bay functions as a true estuary, where freshwater from the Everglades mixes with saline water from the Gulf of Mexico; the rest of the bay functions more like a tropical coastal lagoon, where most of the water flows in to the bay from the Gulf of Mexico and precipitation is the primary source of freshwater (Rapozo 2001; Melo and Lee 2012). Flow over shallow banks and through channels is primarily wind-driven and only provide weak exchange between basins (Lee et al. 2006, 2008, 2016). As shown in Figure 2 (Lee et al. 2016), basins in the western bay have greater exchange than interior and northeastern basins, giving western basins a residence time of about 1 month compared to the 6-12-month residence time in other sections (Lee et al. 2006, 2008, 2016). This restricted horizontal flow

8 makes it unclear how restoration of freshwater inflow from the Everglades may affect the bay; approximately 75% of all freshwater runoff from the Everglades flows into the northeast section of the bay, where the low outflow stores it for an average residence time of one year (Lee et al.

2008). Nuttle et al. (2000) argue that, while increasing runoff from the Everglades would have little effect on salinity in the central and west regions of Florida Bay – where seagrass die-offs have occurred in the past – and would increase the variability of salinity in the southern section, the effect of freshwater runoff in the bay could be greater if inflow is increased during periods when evaporation exceeds precipitation. Similarly, Lee et al. (2006, 2016) suggest that increasing flow to McCormick Creek during the dry season would reduce salinity in the north- central subregion of the bay, possibly preventing hypersaline events.

Figure 2: Annual mean volume transports (m3 s−1) for measured channels in Florida Bay (red arrows and numbers), through the tidal passes between the Florida Keys (blue arrows and numbers), and from mean river discharge (black arrows and numbers) (Lee et al. 2016).

2.3 Plants Studied

The seagrass species most commonly found in the Florida Bay are Thalassia testudinum, Halodule wrightii and Syringodium filiforme and they are the only species evaluated in this research. Thalassia testudinum has wide, long, green leaves which grow off an extensive underground rhizome. Syringodium filiforme produces cylindrical leaves that are similar in length, if not longer and are a lighter shade of green than that of T. testudinum. Halodule wrightii produces the shortest leaves of the three, which are flat and dark green to almost black in color (Figure 3).

Thalassia testudinum is the most abundant seagrass in Florida Bay and is considered the dominant species in these communities (Fourqurean et al. 2001). The size and low nutrient requirements of this species allow it to out-compete the other species in most conditions. It is also the most salt- tolerant species investigated; studies have found Figure 3: Thalassia testudinum (top), Halodule wrightii (middle), and that the optimal salinity range of Thalassia Syringodium filiforme (bottom). Illustrations are by Phillips and Menez testudinum is likely between 20 and 40 PSU (1988) and photographs are by Ken Dunton (University of Texas).

(Touchette 2007) and that in periods of prolonged salinity over 40 PSU, a monoculture of

Thalassia testudinum is observed (Bennett et al. 2005). Thalassia testudinum can be out competed by S. filiforme in environments with low light availability or high phosphorous, which is a limiting nutrient in Florida Bay and increases in concentration from east to west in this system (Fourqurean et al. 1992). Halodule wrightii is the least sensitive to fluctuations in salinity of the three species and is competitively excluded by the others in all environments expect waters with high nutrient concentrations that experience widely fluctuating salinity or in the wake of disturbance (Robblee et al. 1991).

While T. testudinum is dominant in succession, this species grows slowly (the median age of single shoots of T. testudinum in Florida Bay is greater than 5 years) and typically reproduces clonally, making it slow to repopulate areas defoliated by extreme weather (Robblee et al. 1991;

Peterson and Fourqurean 2001; Fourqurean et al. 2003). Halodule wrightii is considered a pioneer species, which becomes established quickly in bare sediment if water quality is suitable

(Peterson and Fourqurean 2001). Considering the rates and requirements for growth of the three species, it was expected that H. wrightii and S. filiforme, which could not tolerate the hypersaline conditions present during the pre-Irma study, will exhibit new distributions after Irma. Halodule wrightii was expected to be especially abundant in areas that were disturbed by the storm energy where the dominant successional species – Thalassia testudinum – would not have had time to reestablish after Hurricane Irma.

2.4 Hurricane Effects

Hurricane Irma made U.S. landfall at Cudjoe Key as a category 4 hurricane on September

10, 2017. Florida Bay and Everglades National Park were located to the east of Irma’s path, where the strongest winds of the storm were directed (Figure 4). While traveling past the western

Figure 4: Hurricane Irma's track in relation to Florida Bay and South Florida.

13 edge of Florida Bay, the storm’s maximum recorded wind speed was 115 knots and minimum pressure was 931 mb. Maximum inundation in the Lower Keys was nearly 2.5 m above the normal tide line. Throughout the Keys, rainfall totals of 10-15 inches (25-38 cm) were reported

(Cangialosi et al. 2018).

In the United States, Irma caused 10 direct human deaths and 82 indirect human deaths.

The Middle and Lower Keys sustained the most damage, where an estimated 25% of all homes were destroyed, while 90% of all homes were damaged. Irma cost an estimated $50 billion in damages to structures and agriculture in the U.S. alone (Cangialosi et al. 2018).

Hurricanes and other extreme weather events are known to cause damage to natural and man-made structures alike, but the short- and long-term effects of tropical cyclones (TCs) on submerged aquatic vegetation can depend on the bathymetry, as well as the size, windspeed, and traveling speed of the storm and water quality in the weeks after the storm passes (Craighead and

Gilbert 1962; Davis 2004; Carlson et al. 2010). It is uncommon for storm surge and shear strength to damage fully submerged seagrasses, especially in a protected bay, but the sediments disturbed by the storm can limit seagrass photosynthesis for several days and the resultant deposition can bury leaves, either of which can cause seagrass mortality or reduced productivity

(Carlson et al. 2010). Because nutrients (especially phosphorous) accumulate in the carbonate bay sediments, suspension of sediment also makes these nutrients available to algae, causing blooms (Rudnick et al. 2005).

In Florida Bay, some basins are shallow enough that storm surge and winds can scour seagrass beds, removing sediment, seagrass root networks, and detritus from the bay floor (Davis et al. 2004). Wrack lines

– lines made of buoyant material that clings to structures and is often used Figure 5: A wrack line composed entirely of dried seagrass, to determine the peak water level of a mostly Thalassia testudinum, left by Hurricane Irma. This photo was taken over 300 meters from Flamingo Beach, FL. flood – that could be seen across

South Florida after Irma made it clear that seagrass had indeed been removed from the bay floor by the storm (Figure 5).

While hurricanes cause devastation to structures and natural features, TCs are important for regulating water quality and circulation in Florida Bay, with extreme weather events providing a hydrologic pulse to the region (Davis et al. 2004; Rudnick et al. 2005; Steward et al.

2006). Buildup of carbonate sediment banks reduce flow between basins, contributing to the hypersaline and hypoxic conditions during periods of high evaporation and low inflow to the bay; storm surge is thought be able to remove enough sediment to restore flow while also transporting excess, sediment-bound nutrients out of the bay (Rudnick et al. 2005). Freshwater in

Florida Bay is almost entirely acquired as rainfall, meaning that TC precipitation can also relieve salt stress in confined basins (Nuttle et al. 2000). In 2018, the South Florida Water Management

District (SFWMD) reported salinities of 30-45 practical salinity units (PSU) throughout most of the bay the month before Hurricane Irma, whereas one month after Irma most of the bay was

15 recorded at 0-20 PSU (Figure 3). Five months after Irma, most of the bay still had salinity concentrations below 30 PSU (SFWMD 2018), which is below the average salinity of ocean water (35 PSU) and conducive to growth for all three of the target seagrass species (Fourqurean et al. 2002; Touchette 2007).

Figure 6: Florida Bay salinity (PSU) one month before Hurricane Irma, one month after, and five months after (SFWMD 2018).

CHAPTER 3

MATERIALS AND METHODS

3.1 Study Sites

This study used two adjacent hydrogeologic basins – Rabbit Key Basin (RKB) and Twin

Key Basin (TKB) – that were impacted by the seagrass die-off in 2015 (McGinnis 2017). The basins are located in the western subregion of Florida Bay (Figure 7), where precipitation is the primary source of freshwater and most inflow comes from the Gulf of Mexico (Nuttle et al.

2000; Lee et al. 2016). Despite their close proximity, the basins differ substantially in water quality and community compositions (Fourqurean et al. 2003), so all analysis of seagrass data in this investigation will consider the basins individually.

Twin Key Basin was the first basin surveyed. The 65 km2 basin was turbid and affected by abundant microalgae but the visibility was not impacted enough to prevent sampling. TKB experiences hypersalinity more frequently than its western neighbor, RKB, due to its main source of inflow coming from the bay’s most hypersaline basins (Lee et al. 2006, 2016). RKB is smaller and shallower than TKB (47 km2) (Figure 8), but its position in the bay provides RKB with direct water exchange with the Gulf of Mexico and the associated regulation of salinity (Lee et al. 2016). RKB also had better visibility than TKB, most likely due to lower water column nutrient concentrations limiting microalgal populations (Fourqurean et al. 1992). The southern

17 edge of TKB had the greatest depth in the study area (nearly 3 meters below mean sea level) but, overall, the majority of the study area is 1-2 meters deep (Figure 8).

Figure 7: Twin Key and Rabbit Key Basins and their position within Florida Bay.

Figure 8: Bathymetry of Rabbit Key and Twin Key Basins. Contour interval: 0.25 m. Digital Elevation Model provided by National Centers for Environmental Information (NCEI). Vertical error estimated by NCEI to be no more than 15 cm.

3.2 Sampling Method

In the pre-Irma, post die-off study, 35 study sites were randomly generated for each study basin (McGinnis 2017). Of those 35 sites, 30 sites in TKB and 20 sites in RKB were surveyed pre-Irma. For this Post-Irma study, researchers navigated to the same locations surveyed in the pre-Irma study using a vessel-mounted GPS with waypoints uploaded using the DNRGPS application by the Minnesota Department of Natural Resources. Of the sites that had Pre-Irma data, all but two sites were surveyed – two sites in RKB were in water too shallow for the research vessel that was larger than the one used for the Pre-Irma survey – resulting in a total of

48 study sites (Figure 9).

For each study site, once the vessel reached the waypoint, a team member dropped a soft- weighted buoy to mark the spot. A pair of snorkelers then attached two 30-meter transect tapes to the buoy and swam in different, random directions. Each diver haphazardly dropped a 0.25 m²

PVC quadrat at 5-meter intervals (5, 10, 15, 20, and 25 meters from the buoy) and analyzed the seagrass cover within the quadrat where it landed.

Figure 9: Location of sampling sites in RKB and TKB.

Density was assessed using the Braun-Blanquet cover-abundance method, which ranks shoot density on an ordinal scale; a 5 on the scale represents 75-100% cover, 4 represents 50-

75%, 3 represents 25-50%, 2 represents 10-25%, and 1 represents less than 10%. If seagrass is found in the transect but there is too little for it to provide cover, a ranking of 0.5 is used to

20 indicate that that there were several shoots of seagrass or 0.1 indicates that there was one shoot of seagrass. Each new snorkeler was trained by an experienced team member to ensure that

Braun-Blanquet values assigned by different researchers corresponded to the same approximate density throughout the study. Snorkelers recorded the Braun-Blanquet value describing cover- abundance of the three target-species at each of the 10 transects upon their return to the vessel.

The Braun-Blanquet method is used in this study for two reasons: the pre-Irma survey utilized this method and changing the metric would make the validity of comparisons between datasets questionable (Fonesca et al. 1998), and because it is a rapid, easily used method which is invaluable when the funding, resources, and weather limits the number of days that can be spent collecting data. If a more time-intensive metric, such as shoot density, had been found for each site, a much smaller number of sites would have been used in this study and, as a result, interpolation would have been less robust. This survey method is also preferred by National

Parks officials over more destructive methods, such as measuring biomass or collecting dredge samples, both of which kill the seagrass sampled (Matt Patterson, personal communication to author, November 2017). In comparison, Braun-Blanquet surveying is minimally destructive, with the anchor of the boat inflicting the only obvious damage to seagrass. Other less direct methods of quantifying seagrass distribution involve remote sensing, which was not appropriate for this research because the type of benthic vegetation visible in remotely sensed imagery is nearly impossible to determine for Florida Bay. Not only is the water often too turbid to see the bottom, but several macroalgae species can be seen throughout the study area, some of which drift freely along the sea floor, making it possible for images taken hours apart to have different distributions of submersed aquatic vegetation (Matt Patterson, personal communication to author, November 2017). The community structure of seagrass also limits the utility of remote

21 sensing in this environment, as seagrass can grow in multiple stories and all three species can be present at a site in a nearly-even distribution (Fourqurean et al. 2002); spectral signatures could be used to identify species in areas with more homogenous composition but, for heterogenous communities like those in the western Florida Bay, it would not have been useful. In the context of this research, the only practical survey method was Braun-Blanquet.

3.3 Spatial Analysis

Change analysis for each species was performed in IBM® SPSS®. While the Braun-

Blanquet cover-abundance method is useful for rapidly approximating density in the field, the ordinal nature of the data requires use of non-parametric tests for determining significant change.

The Kruskal-Wallis test was chosen to determine which sites had significant (P=0.05) difference in species density and total density – the sum of the three species densities in each sample – between the two study periods. The test was performed for all four density variables for each of the 48 sites (192 Kruskal-Wallis calculations total). Tests used the 20 sample values describing each density variable – ten samples from the pre-Irma survey and ten samples from the post-Irma survey – a total of 3,840 data points. The recorded Kruskal-Wallis test statistics indicated whether a variable had the same mean rank across the two study periods; test statistics of P<0.05 meant that the data distribution between the two time periods were significantly different.

Sites that had significant change in total density were plotted in ArcMap. Maps indicating zones of significant change in density were created using only these sites; the difference in medians from before and after Hurricane Irma was used to indicate if the change was positive or negative and inverse Distance Weighting (IDW) was used to create a raster for these points.

Density of seagrass and density of individual species was mapped by plotting the median

Braun-Blanquet score for the density variable at the coordinates assigned to each site. A raster

22 surface was created for each species by basin using Inverse Distance Weighting (IDW) interpolation with a cell size of 45 meters. This cell size was selected because it produced cells of the same area as a circle with a radius of 25 meters – the furthest sampling distance from the buoy at each site. Other IDW settings were kept at default, with a distance power of 2 and a 12- point search parameter. The rasters were processed to the extent of each basin using appropriate polygons. After extraction by a mask of the same basin polygons, the raster data was classified by Braun-Blanquet value. Maps are shown with an overlay outlining locations of shallow banks, where intertidal communities are likely to differ from benthic seagrass meadows and where sampling was sometimes impossible in the 25-foot NPS research vessel. The shallow banks polygon layer is provided by the Florida Fish and Wildlife Conservation Commission-Fish and

Wildlife Research Institute.

The validity of statistical relationships between points from ordinal data is a topic of debate in statistics (Allen and Seaman 2007). The ranges of percent cover that each Braun-

Blanquet class represents are not all equal and the distribution of the data is not normal, leading to my assertion that it is not appropriate to derive statistical relationships from this data as if it were interval data. For these reasons, Kriging – an interpolation method that uses the variance of the data to create maps that are weighted by the statistical significance of data over distance – was not used, despite its reputation as a more robust interpolation method. Because of the ordinal nature of the Braun-Blanquet values, spatial interpolation should only be used to determine the boundaries between the classes; the area between boundaries cannot be interpolated to non- integer numbers (Feoli and Orloci 2012). While IDW interpolation may oversimplify the geospatial patterns present, it produces a visualization of general trends in the study area that may be useful for natural resource management.

CHAPTER 4 RESULTS

After Hurricane Irma, overall density of seagrass in the study area became more contiguous, with fewer patches of bare sediment (no seagrass) and sparse cover (<10% seagrass cover) observed (Figure 10). Distribution of the three seagrass species increased in both basins in all but one variable; S. filiforme in Twin Key Basin (Table 1). The pioneer species Halodule wrightii experienced the greatest change in distribution, as it was found in 25% more samples in

RKB and 6.3% more samples in TKB after Irma. Thalassia testudinum grew in the greatest densities, providing dense cover (>50% seagrass cover) at 40.6% of samples in RKB and 18.3% of samples in TKB.

Twin Key Basin was overwhelmingly dominated by T. testudinum (Table 1). Only 0.3% of samples in this basin reported more than 10% cover from any other species, which in this case was Halodule wrightii. While density of H. wrightii was low in TKB, its distribution after Irma spanned most of the basin (Figure 11). Syringodium filiforme, however, was almost completely absent in TKB, where only one pre-Irma sample at one site reported its presence. With a Braun-

Blanquet value of 0.1, the single shoot of S. filiforme found in TKB was not enough to change the site’s median S. filiforme density, so no map of S. filiforme density in TKB was produced.

T. testudinum was also the most abundant seagrass species in RKB, but H. wrightii and S. filiforme appeared concurrently in greater densities in this basin than in TKB (Table 1). Both H. wrightii and S. filiforme were distributed across the basin, with few patches where they were absent (Figures 11, 12).

Kruskal-Wallis tests determined significant change in density of seagrasses in both basins. The difference between post-Irma and pre-Irma median densities for each site provided information on the direction of change at these significant sites (Figures 13, 14). While positive and negative changes were seen throughout the study area, Halodule wrightii was the only species to not experience significant negative change after Hurricane Irma. Syringodium filiforme experienced equal proportions of positive and negative change while T. testudinum and total seagrass density in both basins had at least 10% more sites experience positive change than negative change.

Thalassia testudinum density became more uniform throughout the study area, as areas of sparse cover and areas of dense cover both approached more moderate densities (Figure 15).

Density of H. wrightii increased in both basins, though it provided no more than 10% cover in most of the study area (Figure 16). Basin-wide density of S. filiforme did not change noticeably, though the patterns of density shifted after Hurricane Irma (Figure 17). Changes in individual species’ density influenced the total seagrass cover, which became more uniform after Irma

(Figures 10, 18).

The accuracy of each density interpolation, calculated by Root Mean Squared Error

(RMSE), is presented in the Appendix. Error was measured using the difference between interpolated values and a set of observations not used to create the interpolated maps; approximately 10% of the sites from each basin were reserved for use in the RMSE. The most accurate interpolation was the post-Irma map of total cover in RKB (0.035 RMSE) and the least accurate interpolation was the pre-Irma map of total cover in TKB (1.517 RMSE). Because of the “patchy” nature of seagrass beds in the study area, this amount of error is expected; the interpolation maps represent the median density of each species, but bare patches and unusually

25 dense patches should be anticipated (Pramova et al. 2013). More reliable maps, made by classifying total cover into the broader categories of “Bare,” representing no cover, “Sparse,” representing less than 10% cover, “Moderate,” representing 10-50% cover, and “Dense,” representing more than 50% cover, are shown for before and after Irma (Figure 10).

Figure 10: Total seagrass cover defined by broad cover categories. Bare sediment is 0% cover; Sparse cover is <10%; Moderate cover is 10-50%; dense cover is >50%.

Table 3(a-b): Braun-Blanquet assessment of seagrass density (D) for all samples, separated by species and basin. 1a contains pre- Irma data and 1b contains post-Irma. Adapted from Fourqurean et al. (2001).

Figure 11: Halodule wrightii distribution before (top) and after (bottom) Hurricane Irma.

Figure 12(a-b): Syringodium filiforme distribution before (a) and after (b) Hurricane Irma.

Figure 13: Proportion (%) of sites in each basin that had significant positive or negative change in density.

Figure 14: Direction and intensity of change in total seagrass cover after Hurricane Irma.

Figure 15: Thalassia testudinum cover before Hurricane Irma (top) and after Hurricane Irma (bottom). 34

Figure 16: Halodule wrightii density before (top) and after (bottom) Hurricane Irma.

Figure 17: Syringodium filiforme density before (top) and after (bottom) Hurricane Irma.

Figure 18: Total Seagrass Cover by Braun-Blanquet value for before

(top) and after (bottom) Hurricane Irma.

CHAPTER 5 DISCUSSION

Hurricane Irma produced strong western winds over Florida Bay, which directed powerful waves towards the northwest within both basins that may have caused the decrease in seagrass density visible in the west and northwest sections of both basins (Figure 13). In TKB,

T. testudinum cover increased after Hurricane Irma in the north-northeast portion of the basin.

This region was severely impacted by the seagrass die-off in 2015, so positive change in cover indicates that conditions became more hospitable for T. testudinum in this part of the bay (Matt

Patterson, personal communication to author, November 2018). The northern section of TKB receives flow from Whipray Basin and other north-central basins of that experience some of the most extreme, chronic hypersalinity in Florida Bay (Lee et al. 2006, 2016), so regulation of salinity and improved flow due to the influx of freshwater from Hurricane Irma may have had greater impact in this area than in other parts of the study area.

Other changes in cover may be explained by the bathymetry of these basins. RKB is confined from the southwest by Ninemile Bank, which is believed to have shielded much of the basin from the most intense wave action. Conversely, the eastern inlets to TKB are less protected by banks and are shallow, leaving this region exposed to scouring waves. As seen in Figure 17, this region did not have uniform dense cover before Irma, meaning that existing sediment could have been removed or suspended by wave action. As a result, sediment experienced a “blow-out” at this location, possibly taking the seagrass with it; aerial photography from one month after

Irma shows plumes of suspended sediment going from the banks and narrow channels into TKB

(Figure 18). If physical removal by wave energy did not cause the decline in T. testudinum density on the eastern edge of TKB, then the reduced light penetration from suspended sediment and algal blooms is another possible cause (Short and Neckles 1999; Rudnick et al. 2005). This area is also noteworthy because, while T. testudinum density significantly decreased, H. wrightii has increased in density there after Irma (Figure 15). The transition from T. testudinum to H. wrightii aligns with the literature, which identifies H. wrightii as a successful pioneer species after disturbance (Peterson and Fourqurean 2001). This also indicates that water quality has changed to be more suitable to H. wrightii since the pre-Irma survey; H. wrightii has been found to decline in water with salinity over 30 PSU and to be absent over 40 PSU (Bennett et al. 2005;

Touchette 2007).

Figure 19: Aerial photography of the study area from one month after Hurricane Irma. Study basins are outlined in yellow. The yellow star on the eastern edge of TKB marks the location of sediment blow-outs.

The distribution of H. wrightii and S. filiforme also changed, with H. wrightii (Figure 8) becoming denser throughout the center and southeastern portion of the basin and S. filiforme

(Figure 9) becoming denser in the northern portion of the basin. These changes in density could have been caused by improved nutrient or light availability, which is evidenced by the inverse relationship between T. testudinum and S. filiforme in this basin; S. filiforme has been found to out-compete T. testudinum in waters with low light or high phosphorous concentrations, which is the limiting nutrient for primary production in the western section of the bay (Fourqurean et al.

1992).

The change in median total density of both basins (Figure 10), makes it clear that after

Hurricane Irma, both basins showed less area that had bare sediment or sparse cover (less than

10% cover, or 0-1 in Braun-Blanquet values, as mapped). The post-Irma map also shows more uniform coverage, indicating that cover is more contiguous after Irma. More contiguous, dense cover can stabilize sediments more efficiently, preventing the resuspension of sediment and nutrients, which lead to algae blooms (Rudnick et al. 2005; Hansen and Reidenbach 2017).

The data shows that there are sections of the bay that experienced a significant loss of seagrass but there are other areas that not only experienced a significant increase in seagrass density, but also demonstrated higher species richness when water quality became more hospitable to H. wrightii and S. filiforme, which forms more suitable habitat for fish and other bay fauna (Bennett et al. 2005). This research contributes new data to the literature on Florida

Bay ecology by providing evidence that tropical cyclones may help accelerate recovery after seagrass die-off events. Factors that may have influenced seagrass recovery include 1) dilution of hypersaline water by precipitation and storm surge and 2) removal of built-up sediments from banks and channels, allowing more water exchange between basins.

CHAPTER 6 CONCLUSION

The precipitation-dependency and low flow between basins in Florida Bay makes the region prone to water deficit, leading to variable salinity. In cases of prolonged hypersalinity, interaction between environmental stressors can cause wide-spread seagrass die-offs (Borum et al. 2005; Koch et al. 2007; Johnson et al. 2018). After an extensive die-off event in 2015, hypersaline water that was high in hydrogen sulfide concentration and low in dissolved oxygen due to the decomposing seagrass remained trapped in the confined basins of the bay. When

Hurricane Irma passed Florida Bay in 2017, scientists from multiple agencies in South Florida were interested to see the positive and negative effects that the event might have on the hydrologically challenged system.

While the storm caused decline in seagrass density in some parts of the study area, seagrass cover in the majority of the study area became more uniform and more diverse. This research shows that, throughout the study basins, the overall seagrass density increased and there was less area with bare sediment. I also found that the non-dominant species Halodule wrightii and Syringodium filiforme began growing in greater densities in some areas where they had already been established and had colonized some patches of sediment that had been left bare after Hurricane Irma. There are still portions of both basins that show a significant reduction in seagrass density after Irma but the positive influence the storm had on seagrass communities after a die-off event is an ecological silver lining to this devastating storm.

REFERENCES

Allen, I.E., Seaman, C.A. 2007. Likert Scales and Data Analyses. Quality Progress, July 2007.

Bennett, R., Johnson, D., Browder, J., McDonald, A., Madden, C., Rudnick, D., Robblee, M. 2005. Examining interactive effects of salinity and seagrass habitat on higher trophic level species for the development of Florida Bay Minimum Flows and Levels technical criteria. Paper presented at 2005 Florida Bay and Adjacent Marine Systems Science Conference, University of Florida, Duck Key, FL, 11-14 December pp. 156-157. Borum, J., Pedersen, O., Greve, T.M., Frankovich, T.A., Zieman, J.C., Fourgurean, J.W., Madden, C.J. 2005. The potential role of plant oxygen and sulfide dynamics in die-off events of the tropical seagrass, Thalassia testudinum. Journal of Ecology 93: 148-158. Bosence, D. 1989. Biogenic carbonate production in Florida Bay. Bulletin of Marine Science 44: 419-433. Brandt, L.A., Browder, J.A., Charkiss, M., Frederick, P., Gaiser, E., Gawlik, D., Geiger, S., Kelble, C., Kelly, S., Kline, J., Kotun, K., Lorenz, J., Madden, C., Mazzotti, F.J., Rodgers, L., Rodusky, A., Rudnick, D., Sharfstein, B., Sobszak, R., Trexler, J., Volety, A. 2016. System-wide Indicators for Everglades Restoration. 2016 Report. Unpublished Technical Report. Cangialosi, J.P., Latto, A.S., Berg, R. 2018. National Hurricane Center tropical cyclone report: Hurricane Irma. National Oceanic and Atomspheric Administration. Carlson, P.R., Jr., Yarbro, L.A., Barber, T.R. 1994. Relationship of sediment sulfide to mortality of Thalassia testudinum in Florida Bay. Bulletin of Marine Science 54: 733-746. Carlson, P. R., Yarbro, L. A., Kaufman, K. A., Mattson, R. A. 2010. Vulnerability and resilience of seagrasses to hurricane and runoff impacts along Florida’s west coast. Hydrobiologia 649: 39- 53. Cole, A.M., Durako, M.J., Hall, M.O. 2018. Multivariate analysis of water quality and benthic macrophyte communities in Florida Bay, USA reveals hurricane effects and susceptibility to seagrass die-off. Frontiers in Plant Science 9: Article 630. Craighead, F. C., Gilbert, V. C. 1962. The effects of Hurricane Donna on the vegetation of Southern Florida. The Quarterly Journal of the Florida Academy of Sciences 25: 1-28. Davis, S. E., Cable, J. E., Childers, D. L., Coronado-Molina, C., Day, J. W., Hittle, C. D., Madden, C. J., Reyes, E., Rudnick, D., Sklar, F. 2004. Importance of Storm Events in

Controlling Ecosystem Structure and Function in a Florida Gulf Coast Estuary. Journal of Coastal Research 20: 1198-1208. DNRGPS Application [Computer software]. 2014. Minnesota Department of Natural Resources. Retrieved from https://www.dnr.state.mn.us/mis/gis/DNRGPS/DNRGPS.html Fedler, T. 2009. The economic impact of recreational fishing in the Everglades Region. Bonefish & Tarpon Trust. https://www.bonefishtarpontrust.org/downloads/research- reports/stories/everglades-economics-report.pdf Feoli, E., Orloci, L (eds). 2012. Computer assisted vegetation analysis. Page 313. Vol. 11. Springer Science & Business Media. Florida Museum of Natural History. 2018. Threatened and endangered species: Florida Bay. University of Florida. Online. https://www.floridamuseum.ufl.edu/southflorida/regions/floridabay/endangered-species/ Fonseca, M.S., Kenworthy, W.J., Thayer, G.W. 1998. Guidelines for the conservation and restoration of seagrasses in the United States and adjacent waters. NOAA Coastal Ocean Program Decision Analysis Series No. 12. NOAA Coastal Ocean Office, Silver Spring, MD. 222 Fourqurean, J.W., Zieman, J.C., Powell, G.V.N. 1992. Phosphorous limitation of primary production in Florida Bay: Evidence from C:N:P ratios of the dominant seagrass Thalassia testudinum. Limnology and Oceanography 37: 162-171. Fourqurean, J. W., Robblee, M. B. 1999. Florida Bay: A history of recent ecological changes. Estuaries 22: 345-357. Fourqurean, J.W., Willsie, A., Rose, C. D., Rutten, L. M. 2001. Spatial and temporal pattern in seagrass community composition and productivity in south Florida. Marine Biology 138: 341- 354. Fourqurean, J.W., Durako, M.J., Hall, M.O., Hefty, L.N. 2002. Seagrass distribution in South Florida: A multi-agency coordinated monitoring program, pages 497-522. In J. W. Porter and K.G. Porter (eds.), The Everglades, Florida Bay, and Coral Reeds of the Florida Keys. CRC Press, Boca Raton, Florida. Fourqurean, J.W., Boyer, J.N., Durako, M.J., Hefty, L.N., Peterson, B.J. 2003. Forecasting responses of seagrass distributions to changing water quality using monitoring data. Ecological Applications 13: 474-498. Hall, M.O., Furman, B.T., Merello, M., Durako, M.J. 2016. Recurrence of Thalassia testudinum seagrass die-off in Florida Bay, USA: initial observations. Marine Ecology Progress Series 560: 243-249. Hansen, J.C.R., Reidenbach, M.A. 2017. Turbulent mixing and fluid transport within Florida Bay seagrass meadows. Advances in Water Resources 108: 205-215.

Johnson, C.R., Koch, M.S., Pedersen, O., Madden, C.J. 2018. Hypersalinity as a trigger of seagrass (Thalassia testudinum) die-off events in Florida Bay: evidence based on shoot meristem O2 and H2S dynamics. Journal of Experimental Marine Biology and Ecology 504: 47-52. Kelble, R.C., Johns, E.M., Nuttle, W.K., Lee, T.N., Smith, R.H., Ortner, P.B. 2007. Salinity patterns of Florida Bay. Estuarine, Coastal and Shelf Science 71: 318-334. Koch, M.S., Schopmeyer, S.A., Holmer, M., Madden, C.J., Kyhn-Hansen, C. 2007. Thalassia testudinum response to the interactive stressors hypersalinity, sulfide, and hypoxia. Aquatic Botany 87: 104-110. Lee, T.N., Johns, E., Melo, N., Smith, R.H., Ortner, P., Smith, D. 2006. On Florida Bay hypersalinity and water exchange. Bulletin of Marine Science 79: 301-327. Lee, T.N., Melo, N., Johns, E., Kelble, C., Smith, R.H., Ortner, P. 2008. On water renewal and salinity variability in the northeast subregion of Florida Bay. Bulletin of Marine Science 82: 83- 105. Lee, T.N., Melo, N., Smith, N., Johns, E.M., Kelble, C.R, Smith, R.H., Ortner, P.B. 2016. Circulation and water renewal of Florida Bay, USA. Bulletin of Marine Science 92: 153-180. Light, S. S. and J. W. Dineen. 1994. Water control in the Everglades: A historical perspective. p. 47–84. In S. M. Davis and J. C. Ogden (eds.) Everglades, the Ecosystem and its Restoration. St. Lucie Press, Delray Beach, FL, USA. McGinnis, C.J. 2017. Mapping 2015 seagrass die-off in Florida Bay using interdisciplinary methods. Masters Thesis, University of Alabama. Melo, N., Lee, T.N. 2012. Water circulation and renewal in Florida Bay is influenced by flows from the Southwest Florida Shelf and tidal passes. Tropical Connections: South Florida’s Marine Environment. (IAN Press: Cambridge, MD.) National Park Service. 2009. The Marjory Stoneman Douglas Wilderness: An Introduction. Online. https://www.nps.gov/ever/upload/WildernessLoResSecure.pdf National Park Service. 2016. 2015 Florida Bay Seagrass Die-Off. Online. https://www.nps.gov/ever/learn/nature/upload/seagrass-Dieoff_final_web_hi_res.pdf National Park Service. 2018. Florida Bay Monitoring Program. Online. https://www.nps.gov/ever/learn/nature/flbaymon.htm Nuttle, W.K., Fourqurean, J.W., Cosby, B.J., Zieman, J.C., Robblee, M.B. 2000. Influence of net freshwater supply on salinity in Florida Bay. Water Resources Research 36:1805-1822. Ogden, J.C., Davis, S.M., Jacobs, K.J., Barnes, T., Fling, H.E. 2005. The use of conceptual ecological models to guide ecosystem restoration in South Florida. Wetlands 25: 795-809. Peterson, B.J., Fourqurean, J.W. 2001. Large-scale patterns in seagrass (Thalassia testudinum) demographics in south Florida. Limnology and Oceanography 46: 1077-1090.

Powell, A.B. 2003. Larval abundance, distribution, and spawning habits of spotted seatrout (Cynoscion nebulosus) in Florida Bay, Everglades National Park, Florida. Fishery Bulletin 101: 704-711. Pramova, E., Chazarin, F., Locatelli, B., Hoppe, M. 2013. Climate change impact chains in tropical coastal areas. Policy Brief. Inventory of Methods for Adaptation to Climate Change. Online. Accessed Mar 03, 2018. Rapozo, C. 2001. Defining Florida Bay. Florida Bay Education Project. Online. http://www.floridabay.org/pub/newspaper/defining.shtml. Robblee, M.B., Barber, T.R., Carlson, P.R., Jr., Durako, M.J., Fourqurean, J.W., Muehlstein, L.K., Porter, D., Yarbro, L.A., Zieman, R.T., Zieman, J.C. 1991. Mass mortality of the tropical seagrass Thalassia testudinum in Florida Bay (USA). Marine Ecology Progress Series 71: 297- 299. Rudnick, D.T, Ortner, P.B., Browder, J.A., Davis, S.M. 2005. A conceptual ecological model of Florida Bay. Wetlands 25: 870-883. Short, F.T., Neckles, H.A. 1999. The effects of global climate change on seagrasses. Aquatic Botany 63: 169-196. South Florida Water Management District. 2018. Ecological effects of Hurricane Irma: ecosystem reports. https://www.sfwmd.gov/science-data/irma-eco-effects. Steward, J. S., Virnstein, R.W., Lasi, M.A., Morris, L.J., Miller, J.D., Hall, L.M., Tweedale, W.A. 2006. The impacts of the 2004 hurricanes on hydrology, water quality, and seagrass in the central Indian River Lagoon, Florida. Estuaries and Coasts 29: 954-965. Tabb, D.C., Manning, R.B. 1961. A checklist of the flora and fauna of northern Florida Bay and adjacent brackish waters of the Florida mainland collected during the period July 1957 through September 1960. Bulletin of Marine Science of the Gulf and Caribbean 11: 552-649. Thayer, G.W., Chester, A.J. 1989. Distribution and abundance of ﬁshes among basin and channel habitats in Florida Bay. Bulletin of Marine Science 44: 200-219. Tilmant, J.T. 1989. A history and an overview of recent trends in the fisheries of Florida Bay. Bulletin of Marine Science 44: 3-33. Touchette, B. Seagrass-salinity interactions: Physiological mechanisms used by submersed marine angiosperms for a life at sea. 2007. Journal of Experimental Marine Biology and Ecology 350: 194-215. University of Texas. Texas seagrass guide. http://texasseagrass.org/TxSeagrasses.html USACE and SFWMD. 1999. Central and Southern Florida Project comprehensive review study final integrated feasibility report and programmatic environmental impact statement. US Army

Corps of Engineers, Jacksonville District, Jacksonville, FL, USA and South Florida Water Management District, West Palm Beach, FL, USA. USFWS. 1999. South Florida multi-species recovery plan. US Fish and Wildlife Service Southeast Region, Atlanta, GA, USA Zieman, J.C., Fourqurean, J.W., Iverson, R.L. 1989. Distribution, abundance, and productivity of seagrasses and macroalgae in Florida Bay. Bulletin of Marine Science 44: 292-311.

APPENDIX

Accuracy Assessment

The accuracy of each interpolated map was performed using 10% of the data points, which were reserved during interpolation. Root Mean Squared Error of each site is reported in Table 3. Error was calculated as the difference between observed values and the nearest boundary of the range used to define the predicted values.

Table 4: Root Mean Square Error for each map by basin (when applicable).

RABBIT KEY BASIN TWIN KEY BASIN T. TESTUDINUM pre-Irma 1.051855741 1.041633333 T. TESTUDINUM post-Irma 0.637965908 0.777281588 H. WRIGHTII pre-Irma 0.360555128 N/A H. WRIGHTII post-Irma 1.457909634 0.057735027 S. FILIFORME pre-Irma 0.070710678 N/A S. FILIFORME post-Irma 0.320156212 N/A TOTAL COVER pre-Irma 0.285745516 1.517124473 TOTAL COVER post-Irma 0.035355339 0.777281588