Hydrobiologia (2010) 649:39–53 DOI 10.1007/s10750-010-0257-0
SEAGRASS ECOSYSTEMS
Vulnerability and resilience of seagrasses to hurricane and runoff impacts along Florida’s west coast Paul R. Carlson Jr. • Laura A. Yarbro • Kristen A. Kaufman • Robert A. Mattson
Published online: 2 May 2010 Ó Springer Science+Business Media B.V. 2010
Abstract Many climate change models predict increasing frequency and severity of tropical cyclones (hurricanes) in the Atlantic Ocean, Caribbean Sea, and Gulf of Mexico. To assess this potential threat to seagrass communities in Florida’s Big Bend region, we performed a habitat change analysis based on aerial seagrass surveys performed prior to, and after, the extremely active Atlantic cyclone seasons of 2004 and 2005. To provide a regional context for changes in the Big Bend region, we also compared impacts there with changes in three other West Florida estuaries. Our analysis showed that storm impacts on seagrasses varied along Florida’s west coast. Physical disturbance caused minor losses in parts of Charlotte Harbor and the Big Bend region. However, heavy rainfall in Florida and Georgia associated with Frances and
Guest editors: M. Holmer & N. Marba` / Dynamics and functions of seagrass ecosystems P. R. Carlson Jr. (&) L. A. Yarbro Fish and Wildlife Research Institute, Fish and Wildlife Commission, 100 Eighth Avenue Southeast, Saint Petersburg, FL 33701, USA e-mail:
[email protected] K. A. Kaufman Southwest Florida Water Management District, 7601 US Highway 301, Tampa, FL 33637, USA R. A. Mattson St. Johns River Water Management District, 4049 Reid Street, Palatka, FL 32178, USA
Jeanne combined with winter rains to cause complete loss of 1,500 ha of seagrasses and thinning of another 1,700 ha in the vicinity of the Suwannee River mouth. In Tampa Bay, Sarasota Bay, and Charlotte Harbor, despite localized losses, total seagrass area actually increased between 2004 and 2006. On the other hand, Tampa Bay, Sarasota Bay, and Charlotte Harbor all showed significant, and more pronounced, declines in seagrass cover as the result of another major rainfall and runoff event: the 1997–1998 El Nino event. Our results indicate that light stress, likely caused by suspended sediments, phytoplankton blooms, and dissolved organic matter, resulted in seagrass losses extending up to 40 km from the mouth of the Suwannee River. We conclude that water quality impacts, especially if they are persistent, can be more damaging than physical impacts of moderate (Category 1–3) tropical cyclones. We also conclude that runoff-related impacts on seagrasses vary depending on the timing, volume, and persistence of storm runoff in relation to normal seasonal runoff patterns and seagrass growth in each estuary. Keywords Seagrass Storm impacts El Nino Hurricanes Optical water quality Florida Big Bend
Introduction In 2004, 15 tropical storms formed in the North Atlantic Ocean, 9 became hurricanes, and 6 became
123
40
major hurricanes, causing over $45 billion in damage in the United States (Franklin et al., 2006). Tropical cyclone activity in the North Atlantic Ocean in 2005 was even more intense, with 27 tropical storms, 15 hurricanes, and 7 major hurricanes (Beven II et al., 2008). Five hurricanes, including Katrina—the third deadliest hurricane in US history—caused over $100 billion in damages in 2005. Since 1995, cyclone activity in the North Atlantic has increased (Webster et al., 2005) leading to concerns that climate change will result in more numerous and/or more intense tropical cyclones. Although the Intergovernmental Panel on Climate Change acknowledges disagreement and uncertainty regarding the relationship between climate change and tropical cyclones (IPCC, 2007), a comprehensive review by Holland & Webster (2007) concluded that trends in sea surface temperature and numbers of tropical cyclones in the Atlantic Ocean have increased substantially as the result of global warming. With over 1,900 km of coastline facing the Atlantic Ocean and Gulf of Mexico and 3,500 km of tidal shoreline, Florida is extremely vulnerable to hurricane impacts. Of the many storms in 2004 and 2005, Bonnie, Charley, Frances, Ivan, and Jeanne had the greatest impacts on Florida in 2004. In 2005, Katrina passed over Florida en route to Louisiana causing damage to Biscayne Bay and Florida Bay. Dennis and Wilma also hit Florida in 2005. In addition to human resources, natural coastal communities such as seagrasses, mangroves, and coral reefs are also vulnerable to hurricane impacts (Tilmant et al., 1994). All of the seagrass communities along Florida’s west coast were impacted by the 2004 and 2005 hurricanes to varying degrees depending on location and storm path, intensity, and size. As noted for these and other hurricanes in the Gulf of Mexico and Caribbean Sea (Ball et al., 1967; Rodriguez et al., 1994; Van Tussenbroek et al., 2008), direct physical impacts on seagrass communities included sediment erosion or deposition as well as defoliation or removal from sediments (blowouts). Indirect physical impacts included turbidity and light stress caused by resuspended sediments (Preen et al., 1995; Campbell & McKenzie, 2004). Seagrasses are also vulnerable to rapid and pronounced declines in salinity and optical water quality caused by heavy rainfall and runoff (Ridler et al., 2006; Steward et al., 2006).
123
Hydrobiologia (2010) 649:39–53
The present study was undertaken to determine the impacts of the 2004 and 2005 hurricanes on seagrass communities along Florida’s west coast in an effort to forecast the potential impacts of stronger and more numerous tropical storms and hurricanes. To distinguish between direct physical impacts and longerterm impacts related to storm runoff, we compared seagrass cover changes in areas close to river mouths with areas with similar coastline orientation located farther from river mouths in Florida’s Big Bend region. When we found that runoff-related impacts of the 2004 and 2005 storms appeared to exceed direct, physical disturbance, we also compared changes in seagrass cover caused by the 2004 and 2005 hurricanes with changes caused by the 1998 El Nino event. This event delivered over 50 cm of rainfall to West Florida during the winter of 1998 with minimal wind or physical disturbance.
Study areas The west coast of Florida has three large estuaries with significant amounts of seagrass: Tampa Bay, Sarasota Bay, and Charlotte Harbor. However, these estuarine seagrass beds are dwarfed by seagrass beds in Florida’s Big Bend region located in the northeastern Gulf of Mexico (Fig. 1). While Tampa Bay, Sarasota Bay, and Charlotte Harbor have approximately 11,450, 3,990, and 7,430 ha of seagrass, respectively, Florida’s Big Bend has more than 106,000 ha of mapped, nearshore seagrass beds (this study) and an equal or greater amount of offshore seagrass in water too deep for conventional aerial photography and mapping (Iverson & Bittaker, 1986; Thompson & Phillips, 1987). Rivers discharging into the Big Bend region include the Suwannee, Steinhatchee, Fenholloway, Econfina, Aucilla, St. Marks, and Ochlockonee rivers. Freshwater runoff from the Suwannee River overwhelms inputs by the other river systems. According to statistics assembled by the NOAA Coastal Geospatial Data Project (NOAA, 2007), the Suwannee River watershed encompasses 26,000 km2 in Florida and Georgia, and annual mean discharge of the Suwannee River is approximately 300 m3 s-1. The estuary is not well bounded, and published areal estimates vary between 100 and 165 km2. However, our estimates indicate that the area of Suwannee
Hydrobiologia (2010) 649:39–53
41
Fig. 1 Seagrass study areas, West Florida Coast, showing tracks of 2004 and 2005 tropical cyclones. Big Bend regions are St. Marks West (MKW), St. Marks East (MKE), Aucilla (AUC), Econfina (ECO), Fenholloway (FEN), Keaton Beach (KTB), Steinhatchee North (STN), Steinhatchee South (STS), Horseshoe Beach West (HBW), Horseshoe Beach East (HBE), and Suwannee (SUW)
Sound and Horseshoe Cove are 70 and 140 km2, respectively, so we have used a value of 220 km2 as a conservative estimate for the total area of the Suwannee estuary. Tomasko et al. (2005) reported that Tampa Bay has a watershed area of 5,896 km2 and an estuarine surface area of 959 km2. The largest rivers discharging into Tampa Bay are the Manatee, Hillsborough, Little Manatee, and Alafia rivers, with a total mean annual discharge of 41 m3 s-1. Sarasota Bay has an estuarine area of 135 km2, a drainage basin area of 389 km2, and a much lower mean
annual river discharge (0.07 m3 s-1). The Charlotte Harbor watershed covers 8,549 km2, the estuary area is 337 km2, and mean annual river discharge is 59 m3 s-1. Solely on the basis of the ratio of watershed area to estuary area, runoff impacts should be greatest for the Suwannee River estuary (118), followed by Charlotte Harbor (25.4), Tampa Bay (6.1), and Sarasota Bay (2.9). There are also significant differences in the timing of runoff into west Florida estuaries. Schmidt et al. (2001) reported that Florida has two distinct seasonal
123
42
patterns of rainfall and runoff. North Florida and the Panhandle receive much of their rain from cold fronts in winter. South Florida receives most of its rainfall from tropical and convectional storms in summer and fall. The middle portion of the state has characteristics of both patterns (Kelly & Gore, 2009). Summer rainfall (July–September) delivers 49, 36, and 47%, respectively, of annual total runoff to Charlotte Harbor, Sarasota Bay, and Tampa Bay (NOAA, 2007). However, summer runoff comprises only 22 and 23% of annual totals for the Suwannee River and gauged streams entering the northern portion of the Big Bend region. For the Suwannee River and other Big Bend tributaries, winter runoff dominates, providing 30 and 34%, respectively, of annual total runoff. Only 15, 20, and 19% of annual total runoff for Charlotte Harbor, Sarasota Bay, and Tampa Bay, respectively, comes in winter.
Methods Rainfall data for long-term monitoring sites were obtained from the NOAA National Climate Data Center (NCDC). We used data from the United States Historical Climatology Network (HCN) Serial Temperature and Precipitation Data. To estimate rainfall in the Big Bend watershed, we used the unweighted mean of monthly rainfall values from sites at Inverness, Lake City, Madison, and Tallahassee, FL. For the Tampa Bay watershed, we used the mean of rainfall values from Bartow, Inverness, Ocala, St. Leo, and Tarpon Springs, FL. For Sarasota Bay, we used the mean of values from Bartow, Inverness, Fort Myers, and Tarpon Springs. Finally, for the Charlotte Harbor watershed, we used the mean rainfall at sites in Bartow and Fort Myers. To calculate total rainfall for the 1998 El Nino event, we summed the mean monthly values described above for the period from October 1997 through March 1998. To calculate total rainfall for the 2004 hurricane season, we summed mean monthly values for the period from July 1 through December 31, 2004. The effects of the 1998 El Nino event and the 2004– 2005 hurricanes on river discharge were evaluated for the Suwannee, Hillsborough, and Peace Rivers. The Suwannee is the largest river in the Big Bend region, the Hillsborough is the largest river discharging into
123
Hydrobiologia (2010) 649:39–53
Tampa Bay, and Peace River is the largest river discharging into Charlotte Harbor. No single large river discharges into Sarasota Bay, but runoff patterns probably resemble those for Tampa Bay and Charlotte Harbor. Discharge data were obtained from the United States Geological Survey (USGS) online stream flow database (http://waterdata.usgs.gov/fl/nwis/current). Estuarine and watershed area statistics were obtained from the National Oceanic and Atmospheric Administration (NOAA) Coastal Geospatial Data Project (http://coastalgeospatial.noaa.gov). Storm event rainfall data were obtained from the NOAA National Hurricane Center archives (http://www.nhc.noaa.gov) and from the University of North Carolina Southeast Regional Climate Center (www.sercc.com/past_tropical). There is no long-term water quality monitoring program for the entire Big Bend region. However, the Florida Department of Agriculture and Consumer Services Shellfish Environmental Assessment Section (SEAS) graciously provided biweekly salinity data for several sites in the southern part of the region near the mouth of the Suwannee River. Interpretations of gains, losses, and density changes in seagrass beds were made from time series of aerial photographs for all estuaries. In fall 2001, seagrass imagery for the Big Bend region was acquired using a film camera at a scale of 1:24,000 from Waccasassa Bay to the mouth of the Ochlockonee River (Fig. 1). Diapositive, natural-color, transparencies were scanned, rectified, and georeferenced to provide digital files with a resolution of 30 cm from which seagrass beds greater than 0.25 ha in size were mapped. In spring 2006, the Florida Department of Transportation acquired imagery of the area using a Zeiss DMC digital camera. Approximately 1,326 frames were acquired at a scale of approximately 1:12,000 and a resolution of 15 cm. Positional accuracy for both datasets was estimated to be ± 5 m. For the purposes of change analysis, we divided the Big Bend region into 11 smaller regions as shown in Fig. 1. Regions range in size from approximately 10,000 ha for the Suwannee (SUW) region to 26,900 ha for the Steinhatchee South (STS) region. The total area of nearshore benthic habitat mapped in 2001 and 2006 was 196,000 ha. After our 2001–2006 change analysis showed significant changes in seagrass cover, we found two earlier mapping efforts for the region that we hoped would allow us to assess the impact of the 1998 El Nino
Hydrobiologia (2010) 649:39–53
event in the Big Bend region. The most recent project, carried out by the U. S. Geological Survey National Wetlands Research Center, was based on 1:24,000 scale, natural color aerial photographs obtained in fall 1992 and winter 1993. Unfortunately, poor water quality through much of the Big Bend left large areas uninterpretable. In its stead, we used the 1984 mapping project based on 1:40,000 scale aerial photography flown by the U. S. Minerals Management Service (Continental Shelf Associates, 1985). This dataset also has limitations: georeferencing is not good in some places, and delineation of habitat polygons appears to be coarse. We plan to re-interpret both datasets, but, for this study, we used the 1984 MMS dataset to describe seagrass distribution in the Big Bend region prior to the 1998 El Nino event. Flight navigation was performed with Loran-C, technology which is much less accurate than global position system (GPS). To improve imagery georeferencing, large targets were deployed at 50 ground control points prior to aerial photography (Continental Shelf Associates, 1985). Seagrass cover data from ground control points were used to groundtruth photo-interpretation. Seagrass cover in Tampa Bay, Sarasota Bay, and Charlotte Harbor has been mapped from aerial photography approximately every 2 years since 1988. Between 1988 and 2002, various contractors working for the Southwest Florida Water Management District have used aerial film cameras to acquire natural-color imagery at a scale of 1:24,000. Photointerpreters delineated seagrass beds greater than 0.25 ha from 900 9 900 diapositives. Since 2004, digital imagery has been acquired using a Zeiss DMC digital aerial camera. Tomasko et al. (2005) have provided an excellent summary of seagrass cover trends in Tampa Bay, Sarasota Bay, and Charlotte Harbor between 1950 and 2002. Tomasko (2000) also provides a more detailed description of mapping methods for these estuaries. For the Big Bend region, we classified habitats in the 2001 and 2006 aerial imagery data sets using the Florida Land Use Land Cover Classification System. Five benthic habitat types were mapped: unvegetated sand or mud bottom in deep water (5400, 5700), unvegetated tidal flats (6510), patchy seagrass beds (9113), and continuous seagrass beds (9116). Photointerpretation contractors provided the maps to us as Arcmap shapefiles. We determined the accuracy of each set of maps by visiting 75–200 sites, depending
43
on the estuary and year, and comparing mapped habitat classifications with the actual communities present. Habitat change analyses were performed using Arcmap software. It should be noted that the FLUCCs habitat code 6510 is represented in tables as ‘‘Shallow Bare’’ habitat, and codes 5400 and 5700 are combined as ‘‘Deep Bare’’ habitat. However, the distinction is based on photo-interpretation rather than on actual bathymetric data. ‘‘Shallow Bare’’ habitats are often described in metadata files as ‘‘Tidal Flats’’ because they are generally close to shore and have an optical signature suggesting shallow water depth. In general, ‘‘Shallow Bare’’ habitats are less than one meter deep.
Results Storm rainfall and river runoff On Florida’s west coast, the 2004 storm season began with Bonnie, a tropical storm with maximum wind speeds of 26 m/s (Table 1). Although wind damage was light, the storm dropped 4–8 cm of rain. Charley was the most intense storm to hit west Florida with winds of 67 m3 s-1 when it struck Charlotte Harbor. Because it was a compact storm, tropical storm-force winds extended only 50 km from its center. Rainfall from Charley was greatest from Sarasota Bay southward. Frances made landfall on Florida’s east coast, and wind speeds decreased as it traversed the state. It recurved in the Gulf of Mexico and made a second landfall in the Big Bend region (Fig. 1), bringing 18–26 cm to west Florida watersheds. Ivan, though a strong and deadly storm, caused minimal wind damage and rainfall in the Big Bend and southwest Florida. Jeanne also made landfall on the east coast and recurved along the west Florida coast, delivering 14–17 cm of rain along most of the west Florida coast. Both Frances and Jeanne also traveled across the upper portion of the Suwannee River watershed in the adjacent state of Georgia. Winter 2004 rainfall fell within historical ranges, but spring 2005 rainfall for all four study areas was nearly twice historical average values. In 2005, Dennis made landfall very close to where Ivan hit in 2004. However, Dennis brought more rain (6–15 cm) to west Florida watersheds and estuaries than Ivan did. Rainfall totals for tropical storms and hurricanes in summer and fall
123
44
Hydrobiologia (2010) 649:39–53
Table 1 Characteristics of hurricanes and tropical storms affecting West Florida in 2004 and 2005 Storm characteristics
Bonnie 2004
Charley 2004
Frances 2004
Ivan 2004
Jeanne 2004
Dennis 2005
Maximum Saffir-Simpson category
TS
4
4
5
3
4
Saffir-Simpson category—landfall
TS
4
2
3
3
3
Maximum wind speed (m/s)
26
67
64
75
54
67
Barometric pressure at landfall (mb)
1,002
941
960
946
950
946
Landfall date
8/12/2004
8/13/2004
9/5/2004
9/16/2004
9/26/2004
7/10/2005
Initial landfall location
Apalachicola
Sanibel
Stuart, FL
Pensacola
Fort Pierce
Navarre, FL
Rainfall—Big Bend watershed (cm)
4.4
5.6
25.8
2.8
15.2
15.2
Rainfall—Tampa Bay watershed (cm)
8.6
7.3
20.9
0.3
16.6
6.2
Rainfall—Sarasota Bay watershed (cm)
3.9
18.8
19.8
0.2
13.7
7.2
Rainfall—Charlotte Harbor (cm)
7.9
19.9
17.8
0.2
14.2
11.2
Sources: Sallenger et al., (2006); NOAA National Hurricane Center; NOAA National Climate Data Center. Storm total rainfall data are centimeters. Note: Frances and Jeanne made landfall twice–once on the Florida east coast and again in the Big Bend region on Florida’s west coast
2004 ranged from 98 to 106 cm across all regions along Florida’s west coast (Table 2). Mean rainfall totals for the 1998 El Nino event ranged from 89 cm in the Big Bend watershed to 109 cm for the Tampa Bay and Sarasota Bay watershed (Table 2), roughly comparable to the 2004–2005 totals. However, El Nino event rainfall totals represented a greater departure from average conditions than the 2004–2005 storm totals. Storm rainfall in summer and fall 2004 represented only 116–147% of normal rainfall from July through December. However, fall and winter rainfall during the El Nino event ranged from 189 to 318% of longterm average values for fall and winter rainfall. Both the 1998 El Nino event and the 2004–2005 hurricanes caused river discharges higher than longterm averages (Fig. 2), but the relative impact of these two types of storm patterns differed among estuaries. For the 1998 El Nino event, Suwannee River discharge began to increase in fall 1997, reaching 840 m3 s-1 in winter 1998 (2.5 times long-term average values). Hillsborough and Peace River discharges in winter 1998 were seven and three times seasonal average values, respectively. River flows returned to long-term averages or lower values by summer 1998. In the 2004–2005 storm event, summer rains and Tropical Storm Bonnie raised flows for all three rivers to long-term average, or higher, values in summer 2004. In fall 2004, Frances and Jeanne raised flows in the Suwannee, Peace, and Hillsborough Rivers to 2.4, 2.8, and 1.8 times average
123
values. Dennis and other lesser storms in 2005 increased flows in the Suwannee and Peace Rivers, but not in the Hillsborough River. The duration of elevated discharges also varied among rivers and between events. For the 1998 El Nino event, discharge of the Suwannee, Peace, and Hillsborough Rivers was higher than long-term average values in fall 1997 and winter 1998 but returned to normal levels by summer 1998. The storms of 2004 and 2005 raised Suwannee River flows for over 10 months, Peace River flows were above average for 18 months, but Hillsborough River discharge was elevated for only 6 months. Seagrass mapping The accuracy of seagrass maps based on Big Bend 2001 imagery was evaluated from 185 points assessed on site in 2002 and 2003. Mapped classification and field data agreed for 162 points (approximately 86%). Of the 23 points where mapped habitat type and field data did not agree, four points were covered by drift macroalgae rather than seagrass at the time of sampling. Because drift algae are not fixed to the bottom, there is no way to know whether they were present at the time of aerial photography. Another 12 sites had extremely sparse or dead seagrass at the time of sampling and were virtually unmappable. When the comparison was adjusted to account for sites with algae, very sparse seagrass and dead seagrass, the overall mapping accuracy was 94%.
Hydrobiologia (2010) 649:39–53
45
Table 2 Comparison of rainfall (cm) in West Florida watersheds for 1998 El Nino and 2004 hurricane events with seasonal averages, 1997–2006 (excluding storm events) Storm/rainfall event Watershed
Fall 1997 ? Winter 1998
Big Bend
Seasonal averages
Event exceedance
Summer and Fall 2004
Fall and Winter
Summer and Fall
Fall 1997 ? Winter 1998 (%)
Summer and Fall 2004 (%) 147
89
103
47
70
189
Tampa Bay
109
106
36
74
302
143
Sarasota Bay
109
98
34
75
318
132
Charlotte Harbor
103
100
32
87
318
116
All watersheds
103
102
37
76
274
133
1000
840
345 279 223
594
506
467
449 345 291
318 242
242 224
318 242
223
176
223 179
Discharge (m3/sec)
143 113
100
78
73 58
61
58
58
58
46 37
35 29 24
25 20
20 18
24
16 12
12
7
6
Suwannee--Storm Period Suwannee--Average
4
24
20
18
11
10 6
36
16 12
7
6 5
5 4
4
Hillsborough-Storm Period Hillsborough-Average Peace-Storm Period Peace-Average
1 1997
1998
1998
1998
2004
2004
2005
2005
2005
2005
Oct-Dec
Jan-Mar
Apr-Jun
Jul-Sep
Jul-Sep
Oct-Dec
Jan-Mar
Apr-Jun
Jul-Sep
Oct-Dec
Fig. 2 Long-term mean and storm discharge for Suwannee, Hillsborough, and Peace Rivers
The accuracy of seagrass maps made from Big Bend 2006 imagery was assessed from 190 sites sampled in 2007. One hundred twenty-five sites classified as continuous seagrass were correctly classified. Thirty-one sites mapped as bare bottom had little or no seagrass present. Ten sites were classified as continuous seagrass and had low seagrass abundance. However, those 10 sites had macroalgae abundance scores which might well have contributed to their mapping signature. Significant and unexplainable differences between mapped classification and ground-truth data occurred in only 9 of 190 sites, representing a potential mapping error of less than 5%. Similar mapping accuracy results were obtained in Tampa Bay, Sarasota Bay, and Charlotte Harbor. Overall mapping accuracy estimates for
2004, 2006, and 2008 mapping efforts were 97, 98, and 92%, respectively. Mapping accuracy for the 1985 Big Bend aerial imagery was difficult to assess because few of the 50 ground-truth sites fell within our nearshore mapping area. Habitat maps for the Big Bend region based on 1984, 2001, and 2006 imagery are shown in Fig. 3. Corresponding habitat area data and change calculations are shown in Table 3. In 1984, seagrass covered approximately 101,340 ha of the Big Bend region. In 2001, the total area of continuous seagrass was approximately 95,000 ha, and patchy seagrass covered another 5,280 ha for a total of 106,000 ha. In 2006, the total area of continuous seagrass declined to 91,000 ha, roughly a 4.5% decline. The decline was apparently offset by an increase in patchy seagrass area to
123
46
Hydrobiologia (2010) 649:39–53
Fig. 3 Big Bend seagrass cover in 1984, 2001, and 2006
4,100 ha. The total area of all seagrass declined from approximately 106,100 ha in 2001 to 105,000 ha in 2006, a decline of 1,100 ha. However, the total area change statistics include the Fenholloway region where visual photointerpretation is made difficult by a large plume of darkly stained water (Fig. 1) which covers approximately 2,700 ha. In 1984, the plume area was interpreted as a mixture of patchy and continuous seagrass. In 2001, the area was considered too dark to interpret. In 2006, approximately half of the plume area was interpreted as patchy seagrass and half was considered to be uninterpretable. For that reason, the entire Fenholloway region was excluded from the area calculations reported below. Total seagrass cover in the other Big Bend regions increased from 82,734 ha in 1984 to 84,800 in 2001 and declined to 83,090 ha in 2006. However, regional changes were more complex (Table 3). Apparent gains in seagrass cover between 1985 and 2001 occurred in the northern half of the Big Bend region. The St. Marks West, St. Marks East, Aucilla, Econfina, and Keaton Beach regions gained a total of almost 5,300 ha of seagrass between 1984 and 2001. The southern Big Bend regions—Steinhatchee North, Steinhatchee South, Horseshoe Beach West, Horseshoe Beach East, and Suwannee—lost over 3,200 ha of seagrass during the same period.
123
Between 2001 and 2006, excluding Fenholloway data, there was a net loss of seagrass (294 ha) in the northern half of the Big Bend region (Fig. 3, Table 3). During the same interval, seagrass losses in the southern half of the Big Bend region (over 1,400 ha) were even greater. Combined losses of 1,550 ha in the Horseshoe East and Horseshoe West regions were only slightly offset by gains of 12, 77, and 51 ha in the Steinhatchee North, Steinhatchee South, and Suwannee regions. For the interval between 1984 and 2006, the spatial pattern of seagrass net gains and losses is striking: a net gain of approximately 5,000 ha in the northern half of the Big Bend is slightly greater than the net loss of 4,643 ha in the southern half. Changes in deep and shallow unvegetated habitat mirrored the changes in total seagrass cover. The spatial pattern and magnitude of Big Bend seagrass loss were even more pronounced when only continuous seagrass beds were considered (Table 3). Between 1984 and 2001, the area of continuous seagrass beds increased by over 19,000 ha in the northern half of the Big Bend region and declined by 3,226 ha in the southern half. For the period 2001– 2006, continuous seagrass cover declined in both the northern half (1,239 ha) and the southern half (3,032 ha) of the Big Bend region. Between 1984
95
Patchy seagrass
478
5,632
Patchy seagrass
Continuous seagrass
-343
Net seagrass change
18
-396
414
18
-36
2,453 1,690
-762
-58
-1,633
5,921
721
277
2,117
6,317
307
258
2,153
3,864
1,070
316
3,786
St. Mark East
14
-75
90
4
-18
1,826 734
-1,092
34
-768
9,860
464
178
1,960
9,935
374
174
1,978
8,109
1,466
140
2,746
Aucilla
8
-48
57
-43
35
9,493 921
-8,571
133
-1,054
11,490
113
206
198
11,538
57
250
163
2,046
8,628
117
1,217
Econfina
1,395
2
1,392
3
-1,413
5,503 -2,794
-8,298
124
2,686
15,764
1,446
182
1,477
15,761
54
179
2,890
10,258
8,351
55
204
Fenway
8
6
1
7
-14
4,323 644
-3,679
124
-768
15,419
494
173
3,794
15,412
493
166
3,808
11,089
4,172
42
4,577
Keaton Beach
12
-143
154
18
-30
-682 -792
-110
104
689
9,119
647
148
6,107
9,262
492
130
6,137
9,944
602
26
5,448
Stein North
77
-299
376
2
-79
-288 -135
153
33
102
8,138
1,388
248
14,508
8,437
1,012
246
14,587
8,725
859
213
14,485
Stein South
-1,397
-770
-627
14
1,383
-1,989 -274
1,714
-17
291
8,498
1,182
127
7,663
9,268
1,809
113
6,280
11,257
94
130
5,989
Horseshoe West
-158
-1,689
1,530
219
-60
-1,475 -1,105
369
911
194
1,167
1,963
1,243
5,296
2,856
433
1,025
5,356
4,330
64
113
5,162
Horseshoe East
51
-131
182
-62
11
-757 -919
-163
650
269
54
340
652
7,749
185
158
714
7,738
942
321
63
7,468
Suwannee
-316
-4,269
3,953
249
49
19,654 -729
-20,383
2,266
-1,519
91,062
9,237
3,730
57,624
95,330
5,284
3,481
57,576
75,676
25,667
1,216
59,094
All regions
Regions are arranged from northwest to southeast (refer to Fig. 1). Fenholloway (Fenway), Stein North (Steinhatchee North), and Stein South (Steinhatchee South) are abbreviated. Data are area in hectares
383
-726
Continuous seagrass
69
Bare shallow
Patchy seagrass
271
Bare deep
Change 2001–2006
1,246 1,302
56
Patchy seagrass
Continuous seagrass Net seagrass change
227
Bare shallow
Bare deep
-1,527
296
Bare shallow
Change 1984–2001
6,756
Bare deep
2006 habitat area
6,359
227
Bare shallow
Continuous seagrass
6,485
Bare deep
5,112
39
Patchy seagrass
Continuous seagrass 2001 habitat area
0
8,012
St. Mark West
Bare shallow
Bare deep
1984 habitat area
Habitat type
Table 3 Changes in Big Bend benthic habitat area by habitat type and by region 1984–2006
Hydrobiologia (2010) 649:39–53 47
123
48
Hydrobiologia (2010) 649:39–53
Table 4 Changes in seagrass cover resulting from 1998 El Nino event and 2004–2005 hurricanes Bay/region
1998 El Nino
2004–2005 Hurricanes
Before 1984
After 2001
Change
Before 2001
After 2006
Change
Imagery year Big Bend Econfina
10,674
11,595
921
11,595
11,603
8
Big Bend Stein South
9,584
9,450
-134
9,449
9,526
77
Big Bend Horseshoe
17,008
14,709
-2,299
14,708
13,204
-1,504
Imagery year
1996
1999
2004
2006
Tampa Bay
10,897
10,057
-840
10,941
11,457
516
Charlotte Harbor
7,781
7,261
-520
7,344
7,435
91
Sarasota Bay
4,183
3,744
-439
3,742
3,990
248
All data are seagrass cover in hectares. For this table, the Big Bend Horseshoe region includes Horseshoe Beach West, Horseshoe Beach East, and Suwannee regions
and 2006, continuous seagrass cover increased by 18,102 ha in the northern half of the Big Bend region (excluding the Fenholloway region), and it declined by 8,222 ha in the southern half of the region. Changes in seagrass cover for Tampa Bay, Sarasota Bay, and Charlotte Harbor are shown in Table 4. Between 2002 and 2006, seagrass cover in Tampa Bay, Sarasota Bay, and Charlotte Harbor increased by 900 (8.5%), 270 (7.4%), and 50 ha (1%), respectively (Table 4). However, significant seagrass losses occurred in all three estuaries as a result of the 1998 El Nino event. Tampa Bay seagrass cover declined by 840 ha (7.7%), Sarasota Bay seagrass cover declined by 440 ha (10.5%), and Charlotte Harbor seagrass cover declined by 520 ha (6.7%). The net change in total seagrass cover for three Big Bend regions is also shown in Table 4. For this analysis, the boundaries of the Econfina and Steinhatchee South regions were not changed, but data from the Horseshoe West (3), Horseshoe East (2), and Suwannee (1) regions were combined as one region called Horseshoe. Total seagrass cover in the Econfina region did not change significantly between 2001 and 2006. Total seagrass cover in the Steinhatchee South region increased slightly, and the combined losses of the Horseshoe region totaled over 1,500 ha, almost 11%.
Discussion Seagrasses are vulnerable to both direct and indirect impacts of hurricanes. Severe hurricanes, categories 4
123
and 5 on the Saffir-Simpson scale, can cause massive seagrass losses by direct, physical disturbance or removal. Rodriguez et al. (1994) reported that Hurricane Hugo, a category 5 hurricane that struck Puerto Rico in 1989, uprooted over 1 km2 of seagrass along the south shore of Vieques Island and buried another 1– 2 km2 with sand. Poiner & Peterkin (1996) reported that Cyclone Sandy destroyed approximately 180 km2 of seagrass in the Gulf of Carpentaria along the coast of northern Australia. Van Tussenbroek et al. (2008) reported that Hurricane Wilma, a category 4 storm, destroyed the nearshore fringe of seagrass beds in the Puerto Morelos Lagoon of the Yucatan peninsula. However, seagrass communities have also demonstrated considerable resilience to direct, physical impacts of hurricanes. Hurricane Donna, a category 4 storm, passed directly over Florida Bay in 1960, and Ball et al. (1967) reported severe damage to coral reefs of the Florida Keys. However, Craighead & Gilbert (1962) reported that, despite huge wrack lines of seagrass leaf blades and mud deposits along the southern shore of the Florida mainland, seagrass communities in Florida Bay were not severely damaged. Similar observations were made for Hurricane Andrew, a category 5 storm which passed over Biscayne Bay in 1992. Tilmant et al. (1994) reported that, despite extensive damage to mangrove communities, the effects of Andrew on Biscayne Bay seagrass beds were limited. Byron & Heck (2006) found that Hurricanes Ivan and Katrina had little effect on seagrasses in Alabama despite their disastrous impacts on land.
Hydrobiologia (2010) 649:39–53
Our change analyses show significant losses of seagrass in the Suwannee, Horseshoe East, and Horseshoe West regions of Florida’s Big Bend between 2001 and 2006 that likely resulted from the 2004 and 2005 cyclones. However, the spatial pattern of seagrass gains and losses during this period strongly suggest that direct physical disturbance was not the primary cause of seagrass loss. First, seagrass losses occurred in the Horseshoe region while seagrass cover increased in other regions (Econfina and Steinhatchee North) with similar shoreline orientation and exposure to wave action. Second, seagrass cover increased or remained stable for most regions except Horseshoe East, Horseshoe West, and St. Marks West—three regions which are close to the Suwannee and Ochlockonee Rivers. Although there was a small increase in total seagrass cover in the Suwannee region between 2001 and 2006, continuous seagrass cover in the Suwannee region declined by 131 ha during this period. Ironically, the reason that seagrass loss in the Suwannee region between 2001 and 2006 was not more severe is that the Suwannee region had already lost over 900 ha of seagrass prior to 2001. Of the 1,263 ha of seagrass present in the Suwannee region in 1984, only 343 ha (27%) remained in 2001. Our Big Bend seagrass change analysis also indicates that seagrass loss has been an ongoing problem for a long time. During the 17-year interval between 1984 and 2001, the Suwannee, Horseshoe East, and Horseshoe West regions lost almost 2,300 ha of seagrass (Table 4). For the 22 year period between 1984 and 2006, combined seagrass losses totaled 3,800 ha or 22%. During the same time period, seagrass cover in the Aucilla, Econfina, and Keaton Beach regions increased or remained stable. The spatial linkage of seagrass loss to the Suwannee, and to a lesser extent, Ochlockonee Rivers is very strong. The spatial pattern of Big Bend seagrass loss, extending up to 40 km from the mouth of the Suwannee River and up to 15 km from the mouth of the Ochlockonee River, strongly suggests that the observed losses were caused by lowered salinity and/ or higher turbidity, color, and phytoplankton biomass caused by storm discharge from these rivers. These indirect impacts of severe, moderate, and even minor storms are caused by turbidity and runoff, and these impacts can be more severe than physical losses.
49
Preen et al. (1995) documented the loss of approximately 1,000 km2 in Hervey Bay (Queensland, Australia) as the result of a cyclone and two river floods in 1992. They reported that seagrasses at depths less than 10 m were uprooted while deeper beds died as the result of light stress from turbidity. They found substantial recovery of deep seagrass beds after 2 years but little recovery of shallower inshore grass beds. Campbell & McKenzie (2004) found that flooding of the Mary River (Queensland, Australia) in 1999 increased turbidity and water column nutrients which persisted for 6 months and caused major losses of seagrasses in the northern Great Sandy Strait. More recently, in Florida, Steward et al. (2006) found that the combined effects of Charley, Frances, Ivan, and Jeanne on the Indian River Lagoon lowered salinities from 30 psu to less than 15 psu, raised color (CDOM) values from 10 pcu to 100 pcu, and increased turbidity from 3 ntu to over 15 ntu. Ridler et al. (2006), working farther south in the Loxahatchee River (Florida, US) estuary, found that lowered salinities caused dramatic declines in the occurrence, density, and biomass of Syringodium filiforme with little or no recovery after 1 year. In Tampa Bay, Morrison et al. (2006) found that the 1998 El Nino event and 2004–2005 hurricanes had similar impacts on water clarity. For most segments of Tampa Bay, variations in water clarity were linked to water column chlorophyll concentrations which were, in turn, linked to rainfall. Color, driven by CDOM was also related to rainfall in all bay segments. Impacts of both El Nino and the 2004–2005 hurricanes persisted for months. Salinity data collected by the Florida Department of Agriculture and Consumer Services in the Suwannee, Horseshoe East, and Horseshoe West regions suggest that salinity excursions did not cause seagrass loss in Florida’s Big Bend region (Fig. 4). Biweekly measurements showed that surface water salinity in the Suwannee region fell below 10 psu for up to 2 weeks in fall 2005 and 4 weeks in spring 2006. Although similar, short excursions below 10 psu were noted for Horseshoe East, surface salinity did not fall below 14 psu in the Horseshoe West region in 2004, 2005, or 2006. Despite consistently higher salinities, the Horseshoe West region experienced the greatest amount of seagrass loss (1,397 ha) between 2001 and 2006. Despite longer periods of lower
123
50
Hydrobiologia (2010) 649:39–53
Fig. 4 Surface water salinity in Suwannee, Horseshoe East, and Horseshoe West regions. Data provided by M. Kuhman, Florida Department of Agriculture and Consumer Services
salinity, a small increase in seagrass cover (51 ha) occurred in the Suwannee region over the same time period. We conclude that seagrass losses in the Horseshoe East and Horseshoe West regions were likely caused by turbidity and CDOM associated with Suwannee river discharge or phytoplankton blooms stimulated by riverborne nutrients. This conclusion is consistent with the studies of Bledsoe & Phlips (2000) who found that phytoplankton abundance was strongly
123
limited by CDOM in the Suwannee River itself. They also found that CDOM was responsible for up to 50% of water-column light attenuation at sites in the Suwannee Estuary, that phytoplankton abundance increased as dilution and flocculation reduced CDOM concentrations in the estuary, and that nutrient availability limited phytoplankton production at their Gulf of Mexico sites on all sampling dates. A more detailed analysis of the relationship between seagrass loss and bathymetry might provide supporting
Hydrobiologia (2010) 649:39–53
evidence for our conclusion. However, bathymetric surveys for this region are over 100 years old, and no digital bathymetric models currently exist for the Big Bend region. Florida’s Big Bend seagrasses are potentially vulnerable to far-field impacts of river discharge, as well. Far-field impacts of river runoff on seagrass beds also result from light stress caused by turbidity, CDOM, and phytoplankton blooms, but these effects may occur tens of kilometers from the coastline. In this study, by focusing on the near shore habitat that could be mapped with conventional aerial photography, we only mapped a fraction (ca. 10%) of the total seagrass habitat in the Big Bend region. Seagrasses occur at depths up to 20 m and distances up to 50 km offshore (Iverson & Bittaker, 1986; Thompson & Phillips, 1987). The latter study used aerial photography, benthic sleds, and divers to delineate approximately 233,000 ha of dense seagrass beds, 498,000 ha of sparse seagrass beds, and 280,000 ha of patchy seagrass beds. They found that near shore seagrass beds were dominated by Thalassia testudinum, S. filiforme, and Halodule wrightii while deeper beds offshore were composed of Halophila decipiens, Halophila engelmannii and algal species. The shallow slope of the West Florida Shelf, less than 1 m/ km, means that, if turbidity, CDOM, or phytoplankton cause the ecological compensation depth of seagrasses to decrease by 10 cm, thousands of hectares of seagrass might die undetected. Differences in runoff volume alone could not explain why Big Bend seagrass losses were more severe than in other estuaries. Storm related river discharge was high but not remarkable for the period July 2004 to December 2005. Mean discharge for the Suwannee, Hillsborough, and Peace Rivers was 141, 162 and 190%, respectively, of long-term mean discharge values. The number of months during this period when discharge for those same rivers exceeded long-term mean values was 9, 12, and 18, respectively. If the volume or duration of runoff alone were an important factor affecting seagrass loss, Tampa Bay and Charlotte Harbor would have lost more seagrass than the Big Bend. However, differences in nitrogen loads among estuaries might have played a role in seagrass loss patterns. Bledsoe & Phlips (2000) estimated the annual nitrogen load of the Suwannee River to be 3.8 9 106 kg N year-1. Asbury & Oaksford (1997)
51
estimated in-stream nitrogen and phosphorus loads for the Suwannee River to be 9.7 9 106 kg and 1.3 9 106 kg, respectively, for the Suwannee River and 1.1 9 106 kg and 1.3 9 105 kg for the Ochlockonee River. Tomasko et al. (2005) estimated nitrogen loads for Tampa Bay, Sarasota Bay, and Charlotte Harbor to be 4.7 9 106 kg N year-1, 4.9 9 105 kg N year-1, and 1.6 9 106 kg N year-1, respectively. When these values are divided by the surface area of each estuary, the area-specific annual nitrogen loads are 4.8 9 103 kg N km-2 year-1, 3.6 9 103 kg N km-2 year-1, and 4.8 9 103 kg N km-2 year-1 for Tampa Bay, Sarasota Bay, and Charlotte Harbor, respectively. In contrast, the areaspecific nitrogen load for the Suwannee estuary is four times higher than for other estuaries (17.2 9 103 kg N km-2 year-1). The vulnerability of seagrasses in Florida estuaries to large runoff events might also depend on timing of the events. As noted earlier, the Suwannee River drainage basin receives most of its rainfall in winter while southern estuaries receive most of their rainfall in summer. As a result, elevated rainfall totals in summer and fall can extend the length of time when Suwannee River discharge is high and estuarine optical water quality is poor. Higher summer rainfall in southern estuaries might have proportionally less effect because normal seasonal patterns deliver high rainfall and runoff in summer, anyway. However, El Nino events might have a strong effect on seagrasses in southern estuaries because most of the rainfall comes in fall and winter and extends normal periods of poor water quality by several months. In contrast, although an El Nino event can bring higher than average rainfall and runoff to Florida’s Big Bend region in winter, it might not necessarily prolong the duration of poor water clarity conditions as it does for southwest Florida estuaries. At this time, the recovery potential of seagrass beds in the Suwannee Estuary is uncertain. Fourqurean & Rutten (2004) reported that some seagrass sites damaged by Hurricane Georges (1998) recovered within 1 year while other, more severely eroded sites, showed no recovery after 3 years. Ridler et al. (2006) saw no recovery of S. filiforme in a year of post-hurricane monitoring. However, Halophila johnsonii beds increased considerably in size during the same period. Campbell & McKenzie (2004) found that recovery of Zostera capricornii from
123
52
flood-related seagrass losses in Queensland required 24–31 months. Seagrass mapping data collected by one of our authors, Kaufman, show that it took 5 years for Tampa Bay to regain 2,075 ha of seagrass lost in the 1998 El Nino event. Charlotte Harbor lost 1,284 ha of seagrass, and seagrass area measured in 2008 (17,374 ha) is still less than the amount measured in 1996 (19,226 ha). Sarasota Bay seagrass cover declined from 10,332 ha in 1996 to 9,247 ha in 1999, and rebounded to 9,243 ha in 2004. These data suggest that, under favorable conditions, seagrasses can recover in 5 years or less. However, the slow recovery in Charlotte Harbor shows that, if poor conditions persist after a major climatic event, seagrass recovery might be delayed for years. Seagrass losses along Florida’s west coast associated with the 1998 El Nino event and the 2004– 2005 hurricanes demonstrate the vulnerability of seagrasses to runoff-related impacts, regardless of the cause. These losses might presage severe impacts of climate change on seagrasses. Not only do climate change models forecast more numerous, and more intense, tropical storms, but sea levels are forecast to rise 50–150 cm by 2100 (Rahmstorf, 2007). The intensity and variability of rainfall in the southeast United States is also likely to increase (Wang et al., 2008). The only realistic actions that can mitigate impacts of climate change on west Florida seagrass communities are those that increase the transparency of coastal waters by reducing suspended sediment and nutrient loads, and phytoplankton biomass. An aggressive, and successful program has been launched to reduce nutrient loading to Tampa Bay (Greening & Janicki, 2006), and similar programs are underway for Sarasota Bay and Charlotte Harbor. In contrast, the entire Big Bend region receives runoff from two rivers draining very large watersheds: the Suwannee and Ochlockonee (in the Florida Panhandle) Rivers. These are the two largest rivers emptying into the northeastern Gulf of Mexico with annual mean flows of 300 and 100 m3/s, respectively. Although the two watersheds are mostly rural and forested, they also contain a number of chicken, swine, and cattle farms that produce considerable amounts of nitrogen and phosphorus. The Florida Department of Environmental Protection (FDEP) has recently identified the lower portion of the Suwannee River and the estuarine area adjacent to the river mouth as
123
Hydrobiologia (2010) 649:39–53
‘‘impaired’’ water bodies based on elevated chlorophyll concentrations in the water column (Hallas and Magley, 2008). Under the Total Maximum Daily Load (TMDL) program administered by FDEP, action plans will now be developed to reduce nutrient loads to the lower Suwannee river and estuary. No nutrient impairment has yet been identified for the Ochlockonee River, but recent seagrass losses in the St. Marks West region require further investigation. Ultimately, the protection and restoration of Big Bend seagrass beds will require reduction of nutrient loads for these two major rivers and other streams draining into the northeastern Gulf of Mexico. Acknowledgments Several agencies and organized provided support for this research, including the Suwannee River Water Management District, U. S. Environmental Protection Agency Gulf of Mexico Program, U. S. National Oceanographic and Atmospheric Agency Office of Ocean and Coastal Resource Management, and Florida Department of Environmental Protection Coastal Management Program. We also thank Mr. Michael Kuhman of the Florida Department of Agriculture and Consumer Services for the use of his salinity data. Two anonymous reviewers provided substantial and helpful suggestions.
References Asbury, C. E. & E. T. Oaksford, 1997. A comparison of drainage basin inputs with instream nutrient loads for seven rivers in Georgia and Florida, 1986–90. USGS Water Resources Investigations Report 97-4006. Ball, M. M., E. A. Shinn & K. W. Stockman, 1967. Geologic effects of Hurricane Donna in South Florida. Journal of Geology 75: 583–597. Beven II, J. L., L. A. Avila, E. S. Blake, D. P. Brown, J. L. Franklin, R. D. Knabb, R. J. Pasch, J. R. Rhome & S. R. Stewart, 2008. Annual Summary Atlantic Hurricane Season of 2005. Monthly Weather Review 136: 1109– 1173. Bledsoe, E. L. & E. J. Phlips, 2000. Relationships between phytoplankton standing crop and physical, chemical, and biological gradients in the Suwannee River and Plume Region, USA. Estuaries 23: 458–473. Byron, D. & K. L. Heck Jr., 2006. Hurricane effects on seagrasses along Alabama’s Gulf Coast. Estuaries and Coasts 29: 939–942. Campbell, S. J. & L. J. McKenzie, 2004. Flood related loss and recovery of intertidal seagrass meadows in southern Queensland, Australia. Estuarine and Coastal Shelf Science 60: 477–490. Continental Shelf Associates. 1985. Florida Big Bend Seagrass Habitat Study Photographic Atlas. Continental Shelf Associates, Jupiter, FL. 17 pp.
Hydrobiologia (2010) 649:39–53 Craighead, F. C. & V. C. Gilbert, 1962. The effects of Hurricane Donna on the vegetation of southern Florida. Quarterly Journal of the Florida Academy of Science 25: 1–28. Fourqurean, J. W. & L. M. Rutten, 2004. The impact of Hurricane Georges on soft-bottom, backreef communities: site- and species-specific effects in south Florida seagrass beds. Bulletin of Marine Science 75: 239–257. Franklin, J. L., R. J. Pasch, L. A. Avila, J. L. Beven II, M. B. Lawrence, S. R. Steware & E. S. Blake, 2006. Annual summary Atlantic hurricane season of 2004. Monthly Weather Review 134: 981–1025. Greening, H. & A. Janicki, 2006. Toward reversal of eutrophic conditions in a subtropical estuary: water quality and seagrass response to nitrogen loading reductions in Tampa Bay, Florida, USA. Environmental Management 38: 163– 178. Hallas, J. F. & W. Magley. 2008. Nutrient and Dissolved Oxygen TMDL for the Suwannee River, Santa Fe River, Manatee Springs (3422R), Fanning Springs (3422S), Branford Spring (3422J), Ruth Spring (3422L), Troy Spring (3422T), Royal Spring (3422U), and Falmouth Spring (3422Z). Florida Department of Environmental Protection, Tallahassee, FL. 117 pp. Holland, G. J. & P. J. Webster, 2007. Heightened tropical cyclone activity in the North Atlantic: natural variability or climate trend? Philosophical Transactions of the Royal Society 365: 2695–2716. International Panel on Climate Change, 2007. Climate Change 2007: Synthesis Report Summary for Policy Makers. Iverson, R. L. & H. F. Bittaker, 1986. Seagrass distribution and abundance in eastern Gulf of Mexico waters. Estuarine, Coastal and Shelf Science 22: 577–602. Kelly, M. H. & J. A. Gore, 2009. Florida river flow patterns and the Atlantic multidecadal oscillation. River Research and Applications 24: 598–616. Morrison, G., E. T. Sherwood, R. Boler & J. Barron, 2006. Variations in water clarity and chlorophyll a in Tampa Bay, Florida, in response to annual rainfall, 1985–2004. Estuaries and Coasts 29: 926–931. National Oceanographic and Atmospheric Agency, 2007. National Coastal Assessment and Data Synthesis (http:// coastalgeospatial.noaa.gov/). Poiner, I. R., & C. Peterkin, 1996. Seagrasses. In L. P. Zann & P. Kailola (eds), The State of the Marine Environment Report for Australia. Technical Annex 1: 40–45. Preen, A. R., W. J. L. Long & R. G. Coles, 1995. Flood and cyclone related loss, and partial recovery, of more than 1000 km2 of seagrass in Hervey Bay, Queensland, Australia. Aquatic Botany 52: 3–17. Rahmstorf, S., 2007. A semi-empirical approach to projecting future sea-level rise. Science 315: 368–370. Ridler, M. S., R. C. Dent & D. A. Arrington, 2006. Effects of two hurricanes on Syringodium filiforme, Manatee Grass,
53 within the Loxahatchee River Estuary, southeast Florida. Estuaries and Coasts 29: 1019–1025. Rodriguez, R. W., R. M. T. Webb & D. M. Bush, 1994. Another look at the impact of Hurricane Hugo on the shelf and coastal resources of Puerto Rico, USA. Journal of Coastal Research 10: 278–296. Sallenger, H. F., H. F. Stockdon, L. Fauver, M. Hansen, D. Thompson, C. W. Wright & J. Lillycrop, 2006. Hurricanes 2004: an overview of their characteristics and coastal change. Estuaries and Coasts 29: 880–888. Schmidt, N., E. K. Lipp, J. B. Rose & M. A. Luther, 2001. ENSO influences on seasonal rainfall and river discharge in Florida. Journal of Climate 14: 615–628. Steward, J. S., R. W. Virnstein, M. A. Lasi, L. J. Morris, J. D. Miller, L. M. Hall & W. A. Tweedale, 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. Thompson, M. J. & N. W. Phillips, 1987. Assessment of hurricane damage in the Florida Big Bend seagrass beds. OCS study MMS 87-0001. New Orleans, Louisiana. Tilmant, J. T., R. W. Curry, R. Jones, A. Szmant, J. C. Zieman, M. Flora, M. B. Robblee, R. W. Snow & H. Wanless, 1994. Hurricane Andrew’s effects on marine resources. Bioscience 43: 230–237. Tomasko, D. A., 2000. Status and trends of seagrass coverage in Tampa Bay, with reference to other southwest Florida estuaries. pp. 171–177. In Greening, H. S. (ed.), Seagrass management: it’s not just nutrients! Proceedings of a Symposium, St. Petersburg, Florida, August 22–24, 2000. Tampa Bay Estuary Program, St. Petersburg, Fl. Tomasko, D. A., C. A. Corbett, H. S. Greening & G. E. Raulerson, 2005. Spatial and temporal variation in seagrass coverage in Southwest Florida: assessing the relative effects of anthropogenic nutrient load reductions and rainfall in four contiguous estuaries. Marine Pollution Bulletin 50: 797–805. Van Tussenbroek, B. I., M. Santos, J. van Dijk, S. N. Alcaraz & M. Calderon, 2008. Selective elimination of rooted plants from a tropical seagrass bed in a back-reef lagoon: a hypothesis tested by Hurricane Wilma (2005). Journal of Coastal Research 24: 278–281. Wang, H., R. Fu & W. Li, 2008. Intensification of summer rainfall variability in the southeastern United States in recent decades. Science and Technology Infusion Climate Bulletin. NOAA’s National Weather Service 33rd NOAA Annual Climate Diagnostics and Prediction Workshop, Lincoln, NE, 20-24 October 2008. Webster, P. J., G. J. Holland, J. A. Curry & H.-R. Chang, 2005. Changes in tropical cyclone number, duration, and intensity in a warming environment. Science 309: 1844– 1846.
123
