<%BANNER%>

Event Bed Preservation Potential in a St. Vincent Island Salt Marsh in Florida


PAGE 1

1 EVENT BED PRESERVATION POTENTIAL IN A ST. VINCENT ISLAND SALT MARSH IN FLORIDA By LISA MARIE ERICKSON MERTZ A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2007

PAGE 2

2 Copyright 2007 by Lisa Marie Erickson Mertz

PAGE 3

3 This is dedicated to Little Miss Megan and Herm es. I will always love you both. Megan, youve always made me want to be a better person so y oud have someone to look up to. I have come to discover that you are the one I look up to; there have been times that your existence has made it all worthwhile. Hermes, you may not be with me now but you were when I needed you most.

PAGE 4

4 ACKNOWLEDGMENTS I gratefully acknowledge the lo gistical support of the St. Vi ncent National Wildlife Refuge for granting access to the island and aid in sampling. I also thank the Gulf Coast Association of Geological Sciences for the grant and opportunity to present research they helped to fund. Additional thanks go to the Department of Soil a nd Water Science at the University of Florida for use of their coring equipment. A big tha nk you goes to Dr. Samuel Bentley at Louisiana State University for allowing me to use his biot urbation model and all the assistance he provided in its application. Most importantly, I thank Dr. John Jaeger for taking me on as a student; budgeting me into his projects because I wanted to do non-funded research; teaching me so much about so many things; and being a great pe rson to know, work with, and work for. Special thanks go to the following: my pare nts for pushing me to reach for the northern lights and supporting all my endea vors; Daniel Gorman for jumpi ng in and staying to the very end; Rebekah Wagner, my soul-mate, this is a ll her fault and would have never happened if it werent for her; KillerD and Chocolate, who rock; Mike Mertz for a good beginning; MaryLea Hart for sparking the interest that became my life for three years; Mari sa Martinovich for her assistance and keeping me compa ny through the monotony of wet siev ing; and all my family and friends for just being there.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........7 LIST OF FIGURES................................................................................................................ .........8 ABSTRACT....................................................................................................................... ..............9 CHAPTERS 1 INTRODUCTION..................................................................................................................11 2 BACKGROUND....................................................................................................................17 Salt Marshes................................................................................................................... .........17 Conseptual Model............................................................................................................... ....18 Numerical Model................................................................................................................ ....18 Study Area..................................................................................................................... .........19 St. Vincent Island............................................................................................................19 Tropical Cyclones and Storms A ffecting the Florida Panhandle....................................20 Major hurricanes......................................................................................................21 Minor hurricanes......................................................................................................22 3 METHODS........................................................................................................................ .....24 Sample Collection.............................................................................................................. .....24 Core Logging................................................................................................................... .......25 Lithology...................................................................................................................... ...........26 Gamma Spectroscopy.............................................................................................................27 Grain Size..................................................................................................................... ..........28 Event Determination............................................................................................................ ...29 Modeling....................................................................................................................... ..........29 4 RESULTS........................................................................................................................ .......33 Lithology...................................................................................................................... ...........33 Grain Size Analysis............................................................................................................ ....35 Gamma Bulk Density.............................................................................................................36 Magnetic Susceptibility........................................................................................................ ..36 X-radiograph Pixel Intensity...................................................................................................37 Chronology..................................................................................................................... ........37 Lead-210....................................................................................................................... ...37 Cesium-137..................................................................................................................... .38

PAGE 6

6 5 DISCUSSION..................................................................................................................... ....50 Lithology and Evidence of Storm Bedding............................................................................50 Modeling Preservation Potential of Storm Beds....................................................................52 6 CONCLUSIONS....................................................................................................................60 LIST OF REFERENCES............................................................................................................. ..61 BIOGRAPHICAL SKETCH.........................................................................................................65

PAGE 7

7 LIST OF TABLES Table page 5-1 Constant BBM Parameters.................................................................................................54 5-2 Variable by Event BBM Parameters..................................................................................55

PAGE 8

8 LIST OF FIGURES Figure page 1-1 Effects of Hurricane Isabel (2003) Hatters Village, North Carolina................................14 1-2 Preservation Potential of an Event Bed.............................................................................15 1-3 Hurricane Landfalls and St. Vincent Island.......................................................................16 3-1 Lithology of Cores......................................................................................................... ....31 4-1 Percent Sand of Cores T2C1 and T3C2.............................................................................39 4-2 Percent Sand and Percent Organi c Matter for Cores T2C1 and T3C2..............................40 4-3 Textural Analysis of Cores T2C1 and T3C2.....................................................................41 4-4 Gamma Bulk Density of Cores..........................................................................................43 4-5 Residual Bulk Density of cores T2C1 and T3C2...............................................................45 4-6 Pixel Intensity of Cores T2C1 and T3C2...........................................................................46 4-7 Total Activity of Lead-210 and Total Activ ity of Lead -210 Corrected for Grain Size....47 4-8 Total Corrected Activity of Cesium-137 and Cesium-137 Activity Corrected for Grain Size..................................................................................................................... ......49 5-1 Physical Properties of Cores T2C1 and T3C2...................................................................56 5-2 Modeling Results for Cores T2C1 and T3C2....................................................................58

PAGE 9

9 Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science EVENT BED PRESERVATION POTENTIAL IN A ST. VINCENT ISLAND SALT MARSH IN FLORIDA By Lisa Marie Erickson Mertz May 2007 Chair: John M. Jaeger Major Department: Geological Sciences Tropical cyclones affect the Gu lf of Mexico coastline annua lly; however, the historical record of these events is limited to about 400 y ears. To establish the occurrence frequency and the importance of cyclones in creating strata, this record needs to be validated and lengthened by examining coastal sediments for the preservation potential of storm be dding. St. Vincent Island National Wildlife Refuge is located in Apalachic ola Bay, FL, and provides a research area with minimal human impact and frequent occurrence of tropical cyclone landfalls, making it ideal to study the natural processes controll ing event bed preservation. Co res were collected in a salt marsh to establish preservation potential of storm bedding. The preservation potential of a storm event in coastal sediments is related to three factors: biologic mixing de pth and intensity, storm layer thickness, and sediment accumulati on rate in the coastal environment. Storm deposition can be detected by changes in bulk density, magnetic susceptibility, or lithology. Radioisotopes were used to quantif y the following processes: mixing depth and intensity by 234Th and 7Be; accumulation rates by 210Pb and 137Cs. Mixed depths are 3 cm in coastal marshes and sedimentation rates are <5 mm y-1. Comparison of transit time of a storm layer through the surface mixed layer to the dissipation tim e of the bed through mixing reveals that beds >3 cm thick are likely preserved. Th e most important control on preservation appears

PAGE 10

10 to be the initial thickness of the storm bed, which is expected to be highly variable for each storm and is controlled by the antecedent topography of the beach/dune system, where the presence of topographic lows (blow-outs) allows for more overwash and thicker beds.

PAGE 11

11 CHAPTER 1 INTRODUCTION In 2002 the United States Census Bureau reporte d that the population of Florida was just under 16 million and had grown 24% since 1990, making it the fastest growing state in the US. By the year 2030, the population is expected to be about 26 million. Most of the population growth is expected to be in th e larger cities located along the co asts (e.g., Tampa/ St. Petersburg, Miami, Jacksonville). As a peninsular state, se parating the Atlantic Ocean and Gulf of Mexico, Florida has 2,170 kilometers of coastline which is annually subjected to tropical cyclones and storms. Along this coastline are the structures (buildings and homes) creating the cities where the majority of Floridas populati on resides. Insurance companies base hurricane insurance rates on the frequency of landfalls with in the region the insured structur e is located. The historical record of tropical cyclones affecting the Florid a Gulf of Mexico coastline is limited to human accounts (about 400 years). It has been proposed that this record may be extended backward by identifying and analyzing event-bedding in coastal sediments (Donnelly et al., 2001; Lui & Fearn, 1993, 2000, & 2002), which in turn may facilita te the ability to make future predictions about tropical storm and cyclone activity. Tropical cyclones and tropical storms release a great deal of energy along coastlines damaging buildings and property, but also resuspending and rewo rking sediments that may be redeposited anywhere water or wind may carry th em. For example, Figure 1-1 shows the effects of Hurricane Isabel (2003) along the North Carolina coastline; th e storm wind, waves, and tides dramatically altered the coastline, moving buildings, trees, and sediment. Coastal sedimentary environments such as this act as long-term recorders of environmental conditions (Boggs, 2001), so event bedding, such as hurricane deposits, have the potential to be preserved within coastal sedimentary records. However, the preserva tion potential of any

PAGE 12

12 depositional event is environmentally dependent varying according to bioturbation intensity, depth of sediment mixing (biological and physical ), burial rate, and orig inal event bed thickness (Wheatcroft and Drake, 2003; Bentley and Shermet, 2003). Preservation potential of any event bed is a function of how rapidly subsequent sedimentation buries it through the potentially rapi d-mixed surface layer (Fig. 1-2). This mixing acts to degrade the intensity of the event si gnal (Wheatcroft, 1990). Consequently, the most important variables of preser vation potential to measure w ould include the following: 1) sediment burial rate, 2) mixing layer thickness, 3) bioturbation intensity, and 4) original event layer thickness. In order to better resolve the mo st important characterist ics that favor storm bed preservation, it is possible to establish a set of criteria favoring preservation based on paleocyclone studies (e.g., Davi s et al., 1989; Donnelly et al., 2001; Liu and Fearn, 2000) that have been performed on a number of different coastal depositional enviro nments (e.g., subtidal, intertidal, and supratidal). An ideal environmental setting would fit the following criteria: 1) high sedimentation rates that quickly bury event beds and prevent sediment mixing; 2) regular cyclone activity resulting in likely production of event bedding; and 3) unaffected by frequent tidal and sea level fluctuations that may erode/mix event beds after deposition (Wheatcroft and Drake, 2003). Although, Wheatcroft and Drake ( 2003) studied event bedding on a continental shelf, the basic concept of thei r study and its criteria should be applicable to any coastal environment. Salt marshes along the northern Gulf of Mexico meet two of these crit eria: regular tropical cyclone activity (Fig. 1-3 A) and negligible eff ects of tidal and sea leve l changes (barrier island salt marshes remain at sea level if supplied with sediment). Also, the marsh vegetation traps sediment within the root masses and stalks aiding in the depos ition of an event-bed within the

PAGE 13

13 salt marsh. Salt marshes typically are composed of organic-rich sa ndy muds (Allen, 2000). Therefore, the deposition of non-local sedime nt resuspended during storm events may be detected by changes in bulk density, magnetic su sceptibility, or lithol ogy (e.g., Donnelly et al., 2001). The preservation potential of event beds in salt marshes has not been determined or quantified, either in the Gulf of Mexico or elsewhere. This study is an evaluation of storm-bed preservation potential in a salt marsh located on St. Vincent Island, Florida (Fig. 1-3 B). This project was designed to test th e hypothesis that intense biological mixing in salt marshes (Frey & Bassan, 1985) may make detection of storm deposits the sedimentary record nearly impossible; unless the deposited layer is thicker than the de pth of mixing. To test this hypothesis, several cores were collected in a salt ma rsh fronted by a low berm next to a small beach (Fig. 1-3 B) and analyzed for lithology, bulk density, magnetic suscep tibility, organic conten t, and grain size and any abrupt changes within to determine possible preservation potential. Numerical modeling of the specific event-beds (Bentley & Sheremet, 2003) was used to examine what conditions are required for preservation of these beds.

PAGE 14

14 Figure 1-1. Effects of Hurricane Isabel (2003), Hatters Villag e, North Carolina. Aerial photographs taken in 1998 and just after the landfall of Hurri cane Isabel in 2003. Images courtesy of National Oceanographi c and Atmospheric Association and the National Geodetic Survey.

PAGE 15

15 Figure 1-2. Preservation Potential of an Even t Bed. This is dependent upon dissipation time (TD) of the event bed and the surface mixed layer (SML) transit time. TD is a function of the event layer thickness (Ls) and mixing intensity and biodiffusion (Db). TT is a function of the event layer thickness (Ls), the depth of mixing (Lb) and the sediment accumulation rate ( ). So, if TD=10 and TT=1 then TD>TT, event bedding is dissipated and not preserved (t op); if TD=1 and TT=100 then TD<
PAGE 16

16 Figure1-3. Hurricane Landfalls a nd St. Vincent Island. A) Hurrica ne landfalls along the coast of Florida from 1885-1984. In the panhandle regi on of Florida, 56% of landfalls have occurred in the Apalachicola Bay area. B) Digital orthophoto quarter quad of St. Vincent Island. Location of the depositional environment (salt marsh) chosen for study. Inset is a blow-up of Marsh B showing core collection location. A B

PAGE 17

17 CHAPTER 2 BACKGROUND Salt Marshes Salt marshes are an upper intertidal to suprat idal muddy environment typical of temperate latitudes that host a range of halophilic pl ant and biological communities (Frey and Bassan, 1978; Woodroffe, 2002). Gulf Coast and southeas tern U.S. marshes typically consist of topographically lower regions adjo ining tidal creeks. The dominant vegetation consists of the cordgrasses Spartina alterniflora and Spartina cynosuroides that generate thick, dense root masses (Edwards and Frey, 1977). The topographi cally elevated high marsh consists of salt tolerant plants along with Spartina spp Gulf Coast marsh sedimentary facies include the following scenarios: root mats, organic oozes grading into clays; vegetati on mats, coarse to medium organic fibers, dark or silty clays; or vegetation mats, and firm clays (Edwards and Frey, 1977; Frey and Bassan, 1985). The typical grain size for the lower salt mars h is in the silt range (4-63 m) and sediments usually have high organic content, whereas the high marsh tends to be sandier. Spartin a decomposes slowly, leaving thick mats of organic matter that ca n bind sediment and limit resuspension (Frey & Basan, 1978). Peat layers are also common with in salt marshes, depending on the decomposition rate of the vegetation. Combined sea-level ri se and autocompaction of marsh sediments are generally balanced, maintaining marsh elevations (Allen, 2000). Mixing of southeastern U.S. marsh sediment by plant roots and macroinvertebrate faunas often completely destroys any evidence of primar y depositional fabric (E dwards and Frey, 1977). Bioturbation causes mixing that penetrates down 10 to 100 cm in the sediment (Koretsky et al., 2002). Also, "biologic ingestion, digestion, and egestion" may cause changes in grain-size and

PAGE 18

18 chemical composition in the sediments, which may degrade or alter any depositional bedding (Koretsky et al., 2002). Conseptual Model Wheatcrofts (1990) conceptual model of ev ent-bed preservation was adapted for this study (Fig. 1-2). This model basically states that a portion of an event bed will be preserved if 1) the event-bed is thicker than th e depth of mixing or 2) the sedi mentary burial rate is fast enough to bury the event bed before it can be mixe d beyond recognition of analytical techniques. Dissipation time (TD) is the time it takes for an event bed to be dissipated or destroyed by bioturbation. According to Wheatcroft and Dr ake (2003), the destruc tion and dissipation of sedimentary fabric occurs at a faster rate th an the destruction and di ssipation of sedimentary textures and physical properties. The surface mixed layer (SML) is the thickness from the sediment surface to the greatest depth of rapid bi ologic mixing. Transit time (TT) is the amount of time an event bed spends within the surface mixed layer. TD is a direct function of the mixing intensity as represented by a biodiffusion coefficien t (Db). An increase in mixing intensity (Db) leads to a faster dissipation time, TD. The SML tr ansit time TT is a dir ect function of the event layer thickness (Ls), depth of mixing (Lb), and sedimentation/burial rate ( ). If the dissipation time is shorter than SML transit time, the even t bed is destroyed and not recorded in the sedimentary record (Fig. 1-2). However, if SML transit time is much less than the dissipation time, a portion of the original event bed will be buried and preserved within the sedimentary record. For example, if it takes 5 years to dissi pate the bed, it would be preserved if the transit time was only 1 year, but completely destroye d if it took 10 years to transit the SML. Numerical Model The Bentley Bioturbation Model (Bentley & Sh eremet, 2003) was used to quantitatively examine event beds for this study. It assumes that depositional fabric is irreversibly transformed

PAGE 19

19 into bioturbated fabric leaving behind some preserved depositional fabric (q). The model uses Eq. 2-1 and 2-2 (Bentley & Sheremet, 2003). The equation used is determined by the depth in sedimentation over time compaired to the depth of mixing (see Eq. 2-1 and 2-2). q = exp[ (o/ ) ( (exp(z)-1) / o+ ) ] ;if z( t) Lb (Equation 2-1) q = exp[ (o/ ) [ ( exp(z) exp(z t ) ) / o ] + [ (exp(z t)-1) / ( o+ ) ] ]; if z( t) < Lb (Equation 2-2) where q is the preservation quotient (%), o is the sediment surface bioturbation rate (biodiffusion coefficient) (cm2 yr-1), is the constant cont rolling the exponential attenuation of o with depth in the sediment, o is sediment accumulation rate prior to the event (cm yr-1), is the sediment accumulation rate following the event (cm yr-1), is the rate of supplemental instantaneous sedime ntation due to event bed deposition (cm yr-1), z is depth in the sediment (cm), t is the change in time (yr) through each run of the model, and Lb is the depth of biological mixing (cm) Beta is determined empirically by Eq. 2-3 (Bentley & Sheremet, 2003). The initial value for z is assumed to be 0.001* o (i.e., 0.1%) as the biological mixing rate at Lb is much, much less than at the surface where mixing is rapid (Boudreau, 1986). Chemical and physical variability within the sediment (lateral and temporal), erosion, and consolidation are not accounted for in this model. Following every mo del run, the output is q, the percent preserved, with dz, the depth. z= o exp(z) (Equation 2-3) where z is very close to zero at the depth Lb an d is forced to zero for all depths greater than Lb Study Area St. Vincent Island St. Vincent Island is an uninhabited 50 km2 barrier island along the northwestern Florida panhandle (Fig. 1-3 B). In 1968 th e U.S. Fish and Wildlife Servic e (FWS) purchased St. Vincent Island from the Loomis estate in accordance with the Migratory Bird Conservation Act, creating the St. Vincent National Wildlife Refuge (S VNWR) (SVNWR Narrative Report, 1868; SVNWR Management Review, 2000). Prio r to this sale, the island was owned privately by four other families or companies. For the most part it wa s treated as a wildlife sanctuary and was managed

PAGE 20

20 in that manner. The most significant huma n impact occurred from logging in 1940-1945 and 1960-1965 when nearly 128.7 km of gravel roads were built (Fig. 1-3 B), creating breaks among the natural communities and disrupted the natural hydrology of the island (SVNWR Narrative Report 1968; SVNWR Management Revi ew 2000). However, near the marsh study area, roads were not construc ted and human interference and influence has been minimal, making this portion of St. Vincent Isla nd an ideal location for this study. It is a triangular shaped island 14.5 kilometers long and 6.4 kilometers wide. Elevation ranges from 0.9 to 3.5 meters above m ean sea level. There were 11 km2 of estuarine marsh in 2001 (SVNWR Narrative Report, 2001). The main salt marsh, Mallard Slough, is located in the northeast area of the island near the study area. Water flows east to southeast through the Big Bayou into Mallard Slough. Vegeta tion of the marsh consists of Distichlis spicata Spartina bakeri and Juncus spp Throughout the Holocene, sediment has been supplied to the island mainly from the Apalachicola River and consists largely of quartz sand (Campbell, 1986). Some sediment is supplied by the longsho re current which travels east to west in this portion of the Gulf of Mexico. Tidal data fo r the island proper are unavailable. In nearby Apalachicola, FL, however, mean tidal range is 0.34 meters with a diurnal range of 0.51 m; maximum water level is 3.53 m above the mean high; and the minimum water level is -0.52 m below the mean low [NOAA Center for Operational Oceanographic Products and Services (CO-OP) station # 8728690 (29O43.6'N, 84O58.9'W)]. Tropical Cyclones and Storms Affecting the Florida Panhandle Hurricanes impact land with wind, waves, rain, and storm surges, which elevate sea levels sufficiently to inundate supratidal land. Florida, being a penins ula, is impacted by hurricanes from both the Atlantic and Gulf of Mexico. Th e written historical reco rd of tropical cyclones and tropical storms is limited to about 400 year s and for the Apalachicola Bay area this record

PAGE 21

21 extends only to the 1800s. Apalachicola Bay has had a significant number of recorded hurricane landfalls (Fig. 1-3 A); 56% of the hurricanes that hit the Flor ida panhandle from 1885 to 1984 occurred in the Apalachicola Bay Area (Davis et al., 1989). As an area that was and continues to be somewhat sparsely populated, some storm details (precipitation, wi nd speeds, barometric pressure, etc.) are not known, esp ecially concerning storms that occurred before the 1950s. For this study, only the hurricanes (major and minor) that were likely to have a direct impact on St. Vincent Island were researched extensiv ely and described in detail below. Major hurricanes In 1886 (June), Apalachicola-Tallahassee ( unnamed) was classified as extreme (classification prio r to 1970: mean wind speed of 220 km hr-1) initially. With the Saffir/Simpson Scale, however, it would be a major (category thr ee or higher) hurricane. Little information is known about this event except for its floodi ng high tides (Williams & Duedall, 2002). In 1975 (September), Eloise (Category 3) made landfall just west of Panama City Florida. Tides in Panama City were measured 3.7 4.9 mete rs above average. Winds were sustained at 200 km hr-1 with 250 km hr-1 gusts; about 38.1 cm of rain fell at Eglin Air Force Base. Winds and tides caused the most damage along the coast (Williams & Duedall, 2002). Eloise is ranked 22nd on the Costliest U.S. Hurricanes and 44th on the Most Intense Hurricanes in the United States. Both lists are from 1900-2000 (NOAA, 2003). In 1985 (August September), Hurricane Elena (Cat egory 3) approached the west coast of Florida from Venice to Pensacola. Although ne ver making landfall in Florida the waves and storm surge (2.1 2.7 meters) caused major evacua tions (one million people) of coastal areas (Williams & Duedall, 2002). In Apalachicola 28. 7 cm of rain fell and tides were 1.4 2.8 m above MHHW and the pressure was 998 millibars. Winds in Carrabelle were measured at 201.2 km hr-1 as the eyewall passed 16.1 to 22.5 km sout h of the island trav eling west (SVNWR

PAGE 22

22 Narrative Report, 1985). Elena is ranked 14th on the Costliest U.S. Hurricanes and 56th on the Most Intense Hurricanes in the United States Both lists are from 1900-2000 (NOAA, 2003). In 1995 (Late September to Early October), Hurricane Opal (Category 5 in Gulf of Mexico; Category 3 at landfall) made landfall between Destin a nd Panama City Florida with 200 km hr-1 winds and gusts of 230 km hr-1. The storm surge was 4.6 meters and affected the coast from Alabama to Cedar Key, Florida (Williams & Duedall, 2002). Minor hurricanes In 1972 (June), Agness landfall was near Apalach icola and Port St. Joe, Florida. This Category One hurricane had peak winds of 140 km hr-1, a maximum surge of 2.1 m, and a reported 32.3 cm of rainfall. At Apalachicola there was 8.6 cm of rainfall, a 2.0 meter storm surge, 90 km hr-1 winds, and 987 millibars of pressure. This large storm was 1609 km in diameter (1852 km circulation envelope) (Williams & Duedall, 2002; Simpson & Herbert, 1973). In 1985 (November 21), Hurricane Kate (Category 2) made landfall near Port St. Joe, was not as destructive as Elena, and damage was at tributed mostly to the winds and storm surge (maximum at 12.9 meters) (Williams & Dueda ll, 2002). The maximum water level ever measured in Apalachicola occurred on November 21, 1985 (NOAA, 2003). There was a storm surge of about 2 meters (Keen & Stone, 2000). Locals claim that during the storm the western half of St. Vincent Island was completely s ubmerged. This would qualify as sheet overwash large enough to carry sand into th e backbarrier marshes of the is land (e.g., Donnelly et al., 2001). According to the SVNWR 1985 Narrative Report, Ka tes eyewall moved acr oss the islands long axis at about 4:45 pm and brought with it a 2.4 3.1 meter surge. In 1995 (June 5), Allison (Category 1 in Gu lf of Mexico; Weak Category 1 to strong tropical storm at landfall), the earl iest storm ever to hit Florida, made landfall just east of St. George Island. Apalachicola experienced the most damage with extreme tides, 12.7 cm of rain,

PAGE 23

23 and maximum winds of 120 km hr-1. Sustaine d winds at St. George Island were 100 km hr-1; Apalachicola reported 60 km hr-1 winds with 70 km hr-1 gusts. The storm surge in Franklin County was 1.2-1.8 meters (Pasch, 1996). (July August) Category 1 Erin entered the Gulf of Mexico by crossing the Florida Peninsula then traveled west roughly pa rallel to the Panhandle coastline before making her second landfall in Pe nsacola (Williams & Duedall, 2002). Erin is number 18 on the Costliest U.S. Hurricanes 1900-2000 list (NOAA, 2003). Apalachicola had gusts of 90 km hr-1 while St. George Island reported gusts of 118.5 km hr-1 (Rappaport, 1995).

PAGE 24

24 CHAPTER 3 METHODS Sample Collection Two sites were chosen for this study: Mars h A and Marsh B (Fig. 1-3 B). Marsh A is located on a small spit in the nor thwestern entrance of Big Bayou (F ig. 1-3 B). This site was chosen based on the ability for storm surges to completely inundate the spit with water from either side, either Big Bayou or St. Vincent Soun d. It also represents a more protected marsh environment with smaller fetches on each side of the marsh. Marsh B is located along the southwestern beach of the Island. This site wa s chosen to replicate to coastal conditions of Donnelly and others research in hurricane storm deposition Ne w England salt marshes (2001). It also represents a more exposed marsh enviro nment with the larger Apalachicola Bay fetch; also West Pass is near-by creating a c onnection with the Gulf of Mexico. Four long cores (~ 1 m) were collected fr om Marsh A in two two-core shore-normal transects. Core one of each transect was in th e sub-tidal region of the ma rsh and core two in the supra-tidal region of the marsh (about 5 meters ap art). Two short cores were also collected. Short cores were about 10 centimeters (cm) long and collected near core two of transect two. In Marsh B four two-core shore-normal transects were collected. Transects were started 20 meters (m) from high tide (determined by wrack line) an d the second core was taken 30 m from high tide. Again, two short cores were collected near co re two (C2) of transect two (T2) (Fig. 1-3 B). Long cores were collected using a modified piston corer and tripod. The coring device consists of a four-foot long, 7.5 cm ID aluminum barrel with a sharpened stainless steal cutting head and an aluminum driving head with handles The core barrel has a 7.5 cm OD CAB liner. The CAB liner was gently pulled out of the barre l and capped resulting in an undisturbed core. Each short core was collected by pushing a 45 cm piece of CAB into the ground and digging it

PAGE 25

25 out. Core collection took two days (a day at each site). After initial measurements of short-lived radioisotopes 234Th and 7Be of a short core from each site, it was decided that the Marsh A cores would not be examined further. The core s from Marsh A did not contain any excess radioisotopic activity (234Th, 7Be, 210Pb, 137Cs), most likely because of the very high sand content. Consequently, chr onologies could not be estab lished for Marsh A cores. Core Logging The long cores all underwent nondestructive analysis on the Geotek multi-sensor core logger (MSCL) to determine bulk density and magne tic susceptibility. Prio r to running cores, a calibration standard was created and measured to determine the calibration constants needed to process data collected by the MSCL. The core logger measures bulk density by gamma attenuation at a rate of 0.5 cm every 10 seconds. The MSCL uses to Eq. 3-1 to dete rmine bulk density by gamma attenuation; gamma attenuation is measured by the number of gamma photons that pass through the width of the core at each sampling point and the count time in that sampling interval. = 1 / d*ln(I0/I) (Equation 3-1) where = sediment bulk density, = the Compton attenuation coefficient, d = the sediment thickness, I0 = the gamma source intensity, a nd I = the measured intensity through the sample Gamma bulk density was corrected for compaction in cores T2C1 and T3C2. This was done by taking the measured density and subtracti ng out a model fit to the data. The model was an exponential decay with Eq. 3-2. y = yo + A ( 1exp-bx) (Equation 3-2) where yo, A, and b were changed to fit the measured data and x was the depth Magnetic susceptibility was determined by a loop created field and how the field is affected by the core for 10 s per one-half cm; the da ta were integrated over 10 cm of core length. Magnetic Susceptibility was measured by the Ba rtington MS2 meter; sample time was measured

PAGE 26

26 in hertz (HZ) where 1 s = 1 Hz and 10 s = 0.1 Hz. Magnetic susceptibility was processed and corrected for both mass and volume. Volume specific magnetic susceptibility (K) is dimensionless and was calculated by the MSCL software using Eq. 3-3, where Kuncorrected is the raw data collected by the MSCL and Krelative is determined by Eq. 3-4, which is a correc tion factor for the loop size. Because the MSCL collects data in cgs units Eq. 35 is used to convert to SI units. K = Kuncorrected / Krelative (Equation 3-3) Krelative = 4.8566(d/D1)2 3.0163(d/D1) 0.6448 (Equation 3-4) where d = core diameter and D1 = loop diameter K (SI units) = 4 *10-6K (cgs units) (Equation 3-5) Mass specific magnetic susceptibility was calcu lated using Eq. 3-6. Equation 3-7 was used to convert from cgs to SI units. = K/ (Equation 3-6) where, = mass specific magnetic susceptibility, K = magnetic susceptibility corrected for loop size, and = sediment density (kg m-3) (SI units) = 4 *10-3 (cgs units) (Equation 3-7) Lithology Cores were split lengthwise into one-third and two-third thick lengths. The thinner or onethird thick lengths were all run through the MSCL for digital imaging and then x-rayed for the detection of sedimentary structur es and for bulk density. Images from the MSCL were processed to determine the red, green, and blue absorption a nd respective ratios to aid in the determination of lithologic changes. The images and digitally scanned x-rays were spliced together in Adobe Photoshop creating a continuous image of each core; maki ng it easier to understand the lithology and observe other visual patterns and sedimentary structures within th e cores, as well as the amount

PAGE 27

27 of bioturbation and the general sedimentology of each core (Figs. 3-1 A and B). Down-core pixel intensity (300 dpi) for select core x-radiographs was determined along a center transect of the core. The data were then smoothed by a 35 point running mean or 40 point running mean averaging the data over fracti ons of a centimeter for cores T3C2 and T2C1, respectively. Smear slides were created at the various color changes within core T3C2 to examine the composition and look for diatoms to determine if the marsh has been fresh or salt water dominated, or had experienced a sa linity change, and/or if it was ev er an environment other than marsh. Diatoms were identified by Dr. Paul Ciesilski at the University of Florida. Gamma Spectroscopy The two short cores collected at each ma rsh were sub-sampled on the day of their collection into one centimeter (cm) intervals. Upon return to the lab these intervals were oven dried and powdered for gamma spectroscopy, to measure short-lived isotopes Thorium-234 (234Th) (half life = 22.3 days) and Berylium-7 (7Be) (half life = 53 days) as a method of determining mixing rates or mixing coefficients and the sedimentation/ accumulation rates of each marsh. Cores T2C1 (transect 2, core 1; proximal) a nd T3C2 (transect 3. core 2; distal) were chosen for higher resolution because they appe ared, from digital imaging, x-radiography and measurements of 210Pb, 137Cs, and 226Ra, to have the least amount of disturbance from bioturbation. Samples were collected at 2-cm intervals, homogenized, freeze dried, powdered, and packed for gamma spectroscopy. Once these initial samples were counted it was decided that additional samples from the high resolution cores (T2C1 and T3C2) would be beneficial for better solving chronologies. Additional samples were prepared and counted. The counting efficiency of the germanium detector used for this study was established using NIST-calibrated sediment standards. Precision of the appara tus was not determined because there was not

PAGE 28

28 enough sample mass to run samples in triplicate or duplicate; also due to time constraints the time was not taken to run samples multiple times. The constant initial concen tration (CIC) model (Benninge r et al., 1979) was used to establish chronologies for each core using the excess 210Pb data and with some assumptions about the sedimentation rates core chronologies were determined. The constant rate of supply (CRS) model (Appleby & Oldfield, 1992) was at tempted for chronology; however, this model did not perform well because of th e variable activity due to grai n size effects. Peak activity (1963) and first-appearance (early 1950s) of 137Cs were used as a check for, and to back up the 210Pb data from the CIC model. All radioisot ope data were plotted against cumulative mass rather than depth to remove the effect of autocompaction, which is common to salt marsh sediments. One gram was removed from each of the samples used for the radiometric analysis for loss on ignition (LOI) to determine the organic matter concentration. LOI was completed by heating each sample to 550OC for approximately 4 hours and rewe ighing the sample to determine mass changes. The remainder of each sample used fo r radioisotopic analyses was weighed, wet sieved (to separate the sand (>63 ) from the silt and clay (<63 )), and dried to de termine percent mass of each grain-size. Both these analyses were completed to determine whether organic content and grain size are related, or if one is more prevalent and wh ich should be used to normalize radioisotope data. Grain Size Long cores (thick half) T2C1 and T3C2 were sub-sectioned at one cen timeter intervals. These cores were chosen because of the high re solution radioisotope work, which was done due to the low degree of visual bioturbation. After homogenizing the sub-samples, approximately 10 g of each sample was treated with 5 ml of 30% molar hydrogen peroxide to oxidize/digest the

PAGE 29

29 organic matter. Samples were then wet sieved w ith deionized water at 1m m to remove any large pieces of organic matter; organic matter was discarded and any sand that remained in the 1 mm sieve was oven dried and stored. The remaining sample was wet sieved with deionized water at 63 to separate sand, silt, and cla y. The silt and clay fraction wa s allowed to settle out, then transferred to whirl-pak plastic ba gs and put into cold room storag e for possible further analysis. The sand fraction was treated again with 5 ml of 30% hydrogen peroxide to remove any remaining organics; some interval s needed to be treated twice to complete organic removal. Once all organic material was removed the sand was oven dried. Sand samples were then run on a settling column to determine sorting, mean grain-size, and modal gr ain-size (Boggs, 2001). Many sand samples were inadequate in mass to be run on the settling column. Event Determination Events were identified by comparing grai n-size, gamma bulk dens ity, and x-ray pixel intensity data. Distinct peaks in all three vari ables for T2C1 and T3C2 were the criterion for identifying to be event beds. The events we re then given a preser vation quotient which was determined from the x-radiographs and based on the amount of apparent mixing (Fig 3-1 A and B). This quotient was a percent range (0 -20, 20-40, 40-60, 60-80, and 80-100) where 100% has experienced no mixing and 0% is completely mixed (Bromley, 1996). This was done for the three event beds that could be found in both high resolution cores. These three events were then modeled using the Bentley Bioturbation Model. Modeling Conceptual modeling compiled the results of gamma-ray attenuation bulk density, grain size analysis, and x-radiography pixel intensity to fulfill all necessa ry variables except dissipation time.

PAGE 30

30 Numerical Modeling was done using the Bentle y Bioturbation Model (see Models chapter) in MatLab Student 6.5 Version 13. Model variab les and parameters (biodiffusion coefficient, mixing depth, and sedimentation ra te) were collected and determin ed from radioisotope and xradiograph data from cores T2C1 and T3C2. Original event bed thickness was the only unknown variable. Each event bed was modeled separately rather than modeling a whole core at once. Simulations were repeated changing only the event bed thickness, time of event, and duration of simulation until the model output patterned th e preservation quotient of each event.

PAGE 31

31 Figure 3-1. Lithology of Cores. A) Digital im agery and x-radiography, for visual examination, of salt marsh proximal to the beach, collect ed 20 meters from high tide (determined from wrack line) The sediments are co mprised of organic-rich sandy muds. A

PAGE 32

32 Figure 3-1. Continued B) Core Lithology Digi tal imagery and x-radiography, for visual examination, of salt marsh distal to the b each were collected 30 meters form high tide (determined from wrack line). The sedime nts are comprised of organic-rich sandy muds. B

PAGE 33

33 CHAPTER 4 RESULTS Lithology Marsh B strata that are closest to the beach (Fig. 3-1 A) are best represented by a core collected along Transect Two (core T2C1). Two di stinct lithofacies are observed in this core. There is a sharp contact at a bout 36 cm depth representing the division of the two units, Unit A and Unit B. Unit A (0 36 cm) is a mottled tan grey/green sandy mud and contains two beds. The top bed of the unit is two-cm-thick, organicdominated with abundant woody plant material. From 2 to 7 cm is a bed of sand with a diffuse bottom boundary that is greener in color. The interval between 12 and 20 cm contains abundant r oots. The top 20 cm appear to show a higher degree of bioturbation containing many burrows while 20 36 cm contains only two large burrows. The second lithofacies, Unit B (36 78.5 cm), is a sandy mud. The bed from 36 50 cm is mottled brown and very dark grey with plant roots and small burrows. From 50 to 78.5 cm the bed is a dark grey grading into a light to very light grey. The unit is faintly to well laminated; most contacts between beds and laminations are sharp. In the other three beach-proximal cores (T1C1, T3C1, T4C1), similar lithological units are observed, although the thickness of each varies betw een cores. In Core T1C1, Unit A has a sharp contact at about 15 cm depth. The top 15 cm contains abundant woody plant material and burrows. The contact between Unit A and Unit B is at approximately 27 cm. The top of Unit B contains less bioturbation than in T2C1.

PAGE 34

34 In Core T3C1, the contact between Unit A a nd Unit B occurs at about 37 cm; however, it is not as sharp a contact as obs erved in T2C1. At 23 cm depth (Unit A) the amount of apparent bioturbation and plant materi al decreases significantly. Core T4C1 is vastly different from the other cores proximal to the beach. This core does not appear to have Unit A or Unit B. A sandy sh ell hash constitutes the upper 16 cm of the core where grey-brown muddy sand begins to mix w ith the shell hash. The amount of sandy shell hash decreases with depth to about 40 cm wher e the grey-brown muddy sand begins to dominate the remainder of the core. Marsh strata that are farthest from the beach (Fig. 3-1 B) are best represented by a core collected along Transect Three (core T3C2), whic h has three observable lithofacies; Unit A, Unit B, and Unit C. The contact between the two upper units is sharp at ~ 36cm depth. Unit A (0 36 cm) is a muddy sand with visible burrows fr om 0 20 cm. The top six centimeters are mottled dark grey green and contain more or ganic material including roots and woody plant fragments. The interval from 6 22 cm is mottled greenish-tan to brown; 22 36 cm is a mottled greenish-grey. There is a sandier bed with a sh arp bottom contact and a diffused top contact at approximately 26 cm depth. The interval from 30 32 cm is a muddier bed with diffuse boundaries and more brown coloring. The second lithofacies, Unit B (36 ~87 cm), is a gradational black to dark grey to light grey. There are fairly well-preserved, faintly to well laminated sand beds, and some are very pronounced. Burrows are preserved in sandy mud fr om 36 42 cm, while fine plant material (roots?) are found throughout the unit. At 42 cm is a fairly sharp contact that marks a change to muddy sand. At 58 cm there is a bed consisting of a series of whitish sandy lenses and pods. Below a diffuse contact at about 69 cm, the texture returns to sandy mud.

PAGE 35

35 The third lithofacies, Unit C (87 ~90 cm), is a brown to very dark brown mottled muddy sand with plant material. There is no apparent bedding. In the other three beach-distal cores (T1C2, T2 C2, T4C2), similar lithological units are observed, although again the thickness of each va ries between cores. In T1C2, the contact between Unit A and Unit B is much more diffuse than in T3C2 and is determined to be at approximately 35 cm based on the x-radiograph. There is abundant woody plant material to 15 cm depth, which is also the depth where biogenic traces decrease dramatically. A whitish sandy lenticular bed in Unit B is at ~ 51 cm. Unit C st arts at approximately 80 cm and as in T3C2 it continues to the bottom of the core. In core T2C2, the Unit A/Unit B contact is at about 35 cm, but is not sharp. There is a large burrow feature from the top of the core to 43 cm. Smaller burrows predominantly occur to a depth of 19 cm. Woody plant material appe ars to disappear around 15 cm while finer plant material (roots ?) continues to about 22 cm. The whitish sand bed of Unit B is around 50 cm but does not have the textural contrast seen in other cores. Unit C begins at about 80 cm and has a much more varied lithology than the other distal cores. The contact between Unit A and Unit B in core T4C2 has been slightly mixed, but can be seen between 40 and 45 cm. Plant material is observed throughout Unit A but the woodier material is found only to a depth of about 30 cm Evidence of bioturbation is apparent through the entirety of Unit A but dramatically decrease s at about 30 cm. The wh ite lenticular bed of Unit B occurs between 53 and 62 cm. Unit C can ba rely be seen in the bottom centimeter of the core Grain Size Analysis Given the large similarities in lithology among cores, detailed textur al analyses were only performed on the two main cores from the proxima l/distal transects (T2C1 and T3C2). The mass

PAGE 36

36 percent sand ranges between 0 and about 51 %, indicating that the ma rsh strata are silt and clay dominated (i.e., sandy mud) (Fig. 4-1). As percent organic matter increases percent sand decreases (Fig. 4-2). Because depositional features, such as bedding, s een in the x-radiographs indicate transport and deposition, the textural char acteristics of the sand fraction can provide information on the mode and strength of transport (i.e., wind, ove rwash) and possibly the source (i.e., beach, terrestrial sand ridges). However, due to amoun t (> 1 gram) of sand needed to for textural analysis sections of the cores we re not analyzed (Fig. 4-3). The median, mean, and modal size of the sand-fraction only range from 1.5 to 2.5 phi (f ine to medium sand). Sorting ranges between 0.5 and 1.0 phi, averaging 0.75 (moderately to modera tely well sorted). Skewness ranges from 0.01 to about 0.5 phi with an average of 0.25 and 0.4 phi for T2C1 and T3C2 respectively. The data are finely (positively) skewed (Fig. 4-3 A). Textural analysis has no significance difference between each interval (Fig. 4-3 B). Gamma Bulk Density Gamma-ray attenuation (GRA) bulk density showed an increase down core in all cores. There appears to be no significant difference between the two sets of data (90O rotation of core) collected for each core (Fig. 4-4). The GRA density range is from 0.8 (indicating partial dewatering of the core) to 1.8 g cm-3. The variation with depth a ppears to track lithology and bedding, so to better isolate this variation, the effect of auto compaction was removed by fitting an exponential model to the GRA density data. Subtracting the differe nce between the model and the data resulted in the density residuals (Fig. 4-5). Magnetic Susceptibility The range of magnetic susceptibility is 0.0 36 *10 SI units; most data are between 0.0 and 5.0 *10 SI units. As with density, there is no significant difference between the two sets of

PAGE 37

37 data (90O rotation) collected for each core. There also appears to be a correlation of the peaks between the cores. However, there is not a strong correlation between the peaks and bedding observed in the x-radiographs. As collected by the Geotek MSCL, magnetic susceptibility values are smoothed over about a 5 to 8 cm window making it a much lower resolution proxy than bulk density, core imagery, and x-radiogra phy. Consequently, it was not used for more detailed modeling of bed preservation. X-radiograph Pixel Intensity X-radiograph positive images reflect variations in sediment density at sub-mm resolution, making them ideal for examining subtle variations in texture. Due to this high resolution, the resultant digitized pixel intensity data were extrem ely variable and difficult to interpret. So, the down-core pixel intensity data were smoothed us ing a running mean. A 40 and 35 pixel running mean (~2mm @ 300 dpi resolution) was used for T2C1 and T3C2 respectively. Relatively low pixel intensities represent increases in sand (i.e., sand beds) (Fig. 46). The notable low intensities range in thickness from a fraction of a centimeter to up to 5 cm. Down-core, the low intensities appear to merge together creating thic ker noisier sections of sand increases (Fig. 4-6). Chronology Lead-210 Because sediment texture can play a contro lling role on particle-reactive radioisotope activities, the activity data are plotted as both dpm (disintegrations per minute) per gram total sediment (Fig. 4-7 A) and per gram mud mass (Fig. 4-7 B) only. When corrected for mud content, there is a large difference between the corrected and uncorrected activity for T2C1, yet little difference between the mud corrected and un corrected data for T3C2. The data used for chronological measurements were the mud-corrected activities. Because the 226Ra data used to calculate supported 210Pb activity were highly variable, it was assumed that the total 210Pb

PAGE 38

38 activity at the bottom of the cores represented the average 226Ra supported 210Pb activity. These activity data subtracted from the total 210Pb data to determine the amount of unsupported, or excess, 210Pb. The first appearance of excess ac tivity is at about 35 and 37 g cm-2 cumulative mass (approximately 23 and 27cm depth) in T2C1 a nd T3C2 respectively (Fig. 4-7 B). Because excess 210Pb activity represents 4-5 half-lives of accumulation (~100-110 years), the resulting mean sedimentation rate based on this first appearance is around 0.2 cm yr-1. Using this sedimentation rate the surface mixed layer is dete rmined to be no thicker than 10 cm depth, as this is the apparent depth of mixing in the 210Pb data (Fig.4-7 B). With depth, activity of 210Pb appears to be increasing until about 10 cm when activity begins to decrease for the remainder of the core. It is possible however, the depth of mixing be much shallower than the 10 cm depth deduced by the higher 210Pb activity. If the surface mixed layer is shallower than 10 cm there is either a change in the 210Pb source or a sedimentation rate change resulting in a misleading mixing depth profile. Cesium-137 As in the case of 210Pb activities, 137Cs activities are susceptible to a grain size influence, and were also corrected to reflect the mass of mud only. Yet, correcting for mud does not significantly affect the data (Fig. 4-8). Becau se mud-corrected activity data were used for 210Pb, they were also used for 137Cs. First appearance of 137Cs occurs at 25 in core T3C2 and at 35 g cm-2 depth in core T2C1. There also is a 137Cs peak at 17 in core T3C2 and at 35 g cm-2 in core T2C1. 137Cs was introduced into the atmosphere in th e mid 1950s and was rele ased in the largest quantities in the early-mid 1960s. The de pth of the first appearance and peak 137Cs activity was divided by the years elapsed between these dates and core collection (48 and 38 years respectively) to determine the 137Cs-based sedimentation rate for the marsh. That sedimentation

PAGE 39

39 rate is about 0.5 cm yr-1 for the beach-proximal core and ~0.2-0.3 cm yr-1 for the beach-distal core. The potential downward transport (diffusi on or advection) of cesium was not accounted for in these calculations and could be substant ial given the coarseness and higher potential permeability of these cores. Downward transpor t would lead to higher calculated sedimentation rates, which may explain the discrepancy between sediment accumulation rates based on the two radioisotopes. Figure 4-1. Percent Sand of Cores T2C1 and T3C 2. As a textural analyses on cores T2C1 and T3C2; mass percent sand with depth in core as determined by wet sieving. Events I, II, and III visual preserved thicknesses are shaded in blue. These are the same event beds modeled using the Bentley Bioturbation Model.

PAGE 40

40 Figure 4-2. Percent Sand and Percent Organic Ma tter for Cores T2C1 and T3C2. A lithological examination of salt marsh cores using per cent sand and percent organic matter. Loss on ignition (LOI) was used to determine th e percent organic matter while wet sieving was used to determine percent sand. Data were collected from sediments used in gamma spectroscopy as they were already prepared for LOI.

PAGE 41

41 Figure 4-3. Textural Analysis of Cores T2C1 and T3C2. A) Each data point is representative of the sand at that particular depth interval; stat istics with depth in core. Statistics were determined from data collected using a settling column. A

PAGE 42

42 Figure 4-3. Continued. B) Increm ental percent of grain size (Phi) for each depth interval containing enough mass sand to analyze usi ng a settling column (top). Cumulative percent of grain size (Phi) for each dept h interval containing enough mass sand to analyze using a settling column (bottom). Data in red represents depth intervals that may be event beds; blue represents sand from Unit A; green represents sand from Unit B. B

PAGE 43

43 Figure 4-4. Gamma Bulk Density of Cores. Lith ological examination of salt marsh cores using gamma bulk density; data are from use of th e GeoTek Multi-Sensor Core Logger. A) The cores proximal to the beach, collected 20 meters from high tide (determined from wrack line). The sediments are compri sed of organic-rich sandy muds. A

PAGE 44

44 Figure 4-4. Continued. B) The co res distal to the beach, collect ed 30 meters from high tide (determined from wrack line). The sedime nts are comprised of organic-rich sandy muds. B

PAGE 45

45 Figure 4-5. Residual Bulk Density of cores T2C1 and T3C2. A lithological examination of salt marsh cores using residuals of gamma bulk de nsity; gamma bulk density data is from use of the GeoTek Multi-Sensor Core Logger. Auto compaction was removed by fitting an exponential model to the density data and removing the difference between the two. This resulted in the density resi duals used to better isolate variations.

PAGE 46

46 Figure 4-6. Pixel Intensity of Cores T2C1 and T3C2. A lithological examination of salt marsh cores using pixel intensity of x-radiographs. Decreases in intensity represent increases in sand attributed to sand beds and po ssibly event beds. Smoothing data, using a running mean, averaged the data over fractions of centimeters.

PAGE 47

47 Figure 4-7. Total Activity of Lead-210 and To tal Activity of Lead -210 Corrected for Grain Size. A) Data were determined by gamma spectroscopy. Particle-reactive radioisotopes, such as lead-210, adhere more readily to silts and clays than larger grain sizes. By correcting for this a more accurate activity is determined. Activity data are plotted against cumulative mass to correct for auto compaction. A

PAGE 48

48 Figure 4-7. Continued. B) Data were determined by gamma spectroscopy and corrected for grain-size. Excess activity was determined by subtracting the average total activity below the assumed first appearance. Firs t appearance is assumed to be ~1900 AD. Activity data are plotted against cumulative mass to correct for auto compaction. B

PAGE 49

49 Figure 4-8. Total Corrected Activ ity of Cesium-137 and Cesium-137 Activity Corrected for Grain Size. Data were determined by ga mma spectroscopy. The mean detection limit of the germanium crystal used in gamma spectroscopy is about 0.1 0.2 dpm g-1. Activity data are plotted against cumulative mass to correct for auto compaction.

PAGE 50

50 CHAPTER 5 DISCUSSION Lithology and Evidence of Storm Bedding X-radiographs and visual obser vations of the cores indicate that they are moderately bioturbated. The catchment basin for the mars h was not large enough to collect measurable amounts of excess 234Th (derived from seawater) or 7Be (derived from atmospheric fallout) to determine biological mixing rates from these shor t-lived radioisotopes. Mixing is attributed mainly to microfauna and roots. Had there been a significant amount of macrofaunal mixing (i.e., fiddler crabs), the core logging likely would have s hown differences between the 90O rotations of the cores. Because the data remained the same within error for the two orientations, I assume that there is no significant macrofauna l burrows or cavities within the cores. In addition, x-radiography did not e xpose any larger cavities. The 0.2 -0.5 cm y-1 burial rates determined by 210Pb and 137Cs appear to be roughly equal to the rate at which accommodation space is created by the rise in relative s ea level for this area of the Gulf of Mexico, which is about 0.3 cm y-1 (Thieler and Hammar-Klose, 1999). Most salt marshes accumulate sediment at a rate equal to the relative rate of sea-level rise (Allen, 2000). The lithofacies of the cores are typical of sa lt marshes from the southeastern U.S., having less organic matter and more detrital sediment than what is seen in colder climates (Frey and Bassan, 1985). The overall decreasing trend of organic matter up core may be due to the increase of erosion and encroachment of the b each on the south side of the marsh. The higher organic content in T3C2 may be a direct result of its distance from the beach. In essence, the organic input into all the cores may be the same but the cores proximal to the beach have more clastic input diluting the organic matter.

PAGE 51

51 Differences in sand content between beach-pr oximal and beach-distal cores can also be attributed their locations with respect to the beach (Fig. 1-2), which is assumed to be the dominant source of clastic sedime nt. There is abundant vegetati on separating the marsh from the tidal creeks to the west, so the mechanism to be st explain the relatively high mean mass percent sand (~30%) throughout the marsh (located <5 m from south shore) is either aeolian or overwash. Due to the apparent l ack of a continuous input of sand into Marsh B overwash is the more likely mechanism. However, during a stor m, increased wind speeds could deposit an event bed with aeolian mechanism. Overwash is a mo re likely process controlling sand deposition at these coring sites, given that there is no evid ence of bedload sand tran sport in x-radiographs (e.g., sharp erosional contacts, cros s-lamination, graded bedding). Initial event be d thickness is a function of the volume of sediment availabl e for transport, shoreline vegetation, dune morphology, distance from the shore, and storm surge. Statistically (cumulative and incremental per cent, mean, mode, sorti ng), the source of sand is the same throughout both cores (Fig. 4-3). Th e mean and modal sizes of the sand fraction are the same as the mean and modal grain size of the south beach (Hart, 2003). The shell hash at the top of T4C1, however, may be an overwash fa n from a storm that occurred after 1954 as indicated by 137Cs activity found beneath the shell hash. The lack of vegetation on top of this shell hash suggests a more recent event such as the 1985 Hurricane Kate whose eyewall passed directly over the island, 1995 category five Hu rricane Opal, or 1998 Hurricane Earl with the 2.5 meter storm surge. Based on diatoms identified in the smear slid es (P. Ciesielski, pe rsonal communication), it was determined that this marsh was once freshw ater dominant (Unit B) and became saltwater dominant (Unit A). It is possibl e that at one time this was a coastal pond location that has in

PAGE 52

52 filled; depending on island and coastal morphology. The overall down-core increase in sand starting at 44 cm depth (Fig. 4-1) may be due to the marsh being freshwater dominant. At the transition between Units A and B found in all core s, 35 to 40 cm depth, it is possible that a storm event between 116 and 88 years ago (betw een 1886 and 1915) caused a change in geomorphology allowing daily tidal inundations to extend inland as they do today. Six separate storms occurred in this time frame but the 1886 was the only major hurricane. The storms discussed in the Study Area se ction of Chapter 2 provided the initial chronology for possible event bedding. Using th e determined sedimentation rate and storm history approximate Events were determined with in the cores. These Events were confirmed by comparing and correlating the three main proxi es (percent sand, gamma bulk density residuals, and pixel intensity) as shown by Figure 5-1. It is possible that the corr elating peaks of these proxies are event beds or temporary changes in se dimentation. However, it is less likely these peaks represent sedimentation changes due to the short duratio n of deposition, lack of sharp contacts, and lack of significan t change in aerial photographs. It is more difficult to find and correlate peaks in the distal co res and the correlated peaks in the proximal cores are shallower than those in the distal cores. Modeling Preservation Pote ntial of Storm Beds In order for an event layer to be preserve d, transit time of the storm layer through the surface mixed layer must be faster than the di ssipation time due to mixing. From Eq. 2-1, the transit time can be calculated for the marsh study areas given estimates of the mixed depth (Lb), storm bed thickness (Ls), and burial rates (i.e., sedimentation rate) (Fig. 1-3). For this study, the thickness of the surface mixed layer was calcula ted to be approximately 3.0 cm based on 210Pb profiles and x-radiographs, as shorter-lived tracers (e.g., 234Th) were not effective. If surface sediments have been rapidly mixed, the 210Pb profile may have a near-surface interval of

PAGE 53

53 constant activity that corresponds in x-radiographs with abundant biogenic traces (Figs. 3-1, 3-2, 4-4, and 4-7 B) (Benninger et al., 1979). For the marshes, Ls was estimated to be 3-50 cm matched to depths of known hurricanes (Event I 1974/1985; Event II 1886; Event III prehistoric) established partially with 210Pb/137Cs geochronology. The compli cating fact is that the thickness of a storm layer would need to be m easured directly following a hurricane for an accurate estimate of Ls, as what is measured in the core s is the minimum preserved bedding thickness as some unknown quantity at the top of the bed has been mixed (Fig. 3-1, 3-2, 4-4, and 4-7 B). Using the above data and a sedimentation rate of ~0.2-0.5 cm yields a transit time of 0-30 years (Lb > ~1 cm) for the salt marshes. The fast transit time suggests that an event layer should be preserved, but does not account for dissipation of the event layer due to physical and biologic mixing. Since time series cores are not available for this study, the dissip ation time for the cores can be estimated based on proba ble biodiffusion coefficients (Db) can be calculated from (a) 210Pb activity data (Aller et al ., 1980) or (b) data from prev ious studies. Maximum mixing coefficients are ~2 cm2 y-1 from the 210Pb activity data from these marsh cores, but these rates would represent much slower mixing averaged over several decades (Nittr ouer et al., 1984). For salt marshes, short-lived tracers ideally suited for calculating Db (Aller et al., 1980) were not detected in this study, but x-radiography reve als substantial mixing and burrowing. Thus Db is assumed to be ~ 10 cm2 y-1 based on estimates from other salt marshes (Boudreau, 1994). Wheatcroft and Drake (2003) report Db values ranging from 10 to 100 cm2 y-1 for continental margin sediments where sediments are more bi ologically active and correspond to dissipation times of 3-5 years. Thus, it is estimated that the dissipation time for coastal marshes is ~5 years. Given transit times of 0-10 years and a SML of 10 cm for this salt marsh, it is clear that very

PAGE 54

54 large (> 5cm thick) beds will be preserved, but that beds 1 cm thick will be mixed and not preserved. Original event bed thickness was then estimated by numerically modeling the event-beds found using the Bentley Bioturbatio n Model (BBM). Estimation of original even t bed thickness was used to validate the conceptual model. Th is same numerical model was used to quantify preservation potential of the salt marsh sedime nts (Fig. 5-2). Although created to model event beds created on continental shelves the BBM was accurately and successfully used in this salt marsh as I was able to reproduce the ev ent beds found in the Marsh B cores. The values for parameters in the BBM we re established usin g the results form 210Pb/137Cs chronology, x-radiographs, core logging, and esti mated from literature when necessary. The BBM parameters used for this specific salt ma rsh, and all events modeled, were defined as in Table 5-1. Table 5-1. Constant BBM parameters determined by cores collected in Marsh B on St. Vincent Island, Fl. BBM Constant Parameter Value o and (mm2 y-1) (sedimentation rate prior to and after event, respectively) 0.025 Lb (cm) (depth of mixing or surf ace mixed layer thickness) 3 o (cm2 yr-1) (biodiffusion coefficient) 6.733 (controlling constant of o) 2.3026 (cm yr-1) (supplemental rate of accumulation) 0 The value for was also used for o because it was assumed the sedimentation rate did not change prior to or following each event. Ze ro was used for the supplemental rate of accumulation ( ) to portray an instantaneous deposition as would be expected with an event bed

PAGE 55

55 due to a tropical cyclone. The pa rameters that varied from even t to event were defined as in Table 5-2. Table 5-2. Variable by event BBM parameters de termined by cores collected in Marsh B on St. Vincent Island, Fl. BBM Variable Parameter t (years) (length of time model is run) t (year) (time event occurs) Ls (cm) (event bed original thickness) T2C1 Event I 14.17 10 4.5 Event II 4.17 2.75 2.25 Event III 12.5 2.08 6.5 T3C2 Event I 0.417 0.25 2 Event II 8.33 4.17 2.25 Event III 16.67 2.5 10.417 2 1.5 The values in Table 5-2 were determined by trial and error. The model was run multiple times until the output (depth of sedimentation and percent of event preserved) best fit the preservation as determined from the x-radi ographs (Fig. 5-2). The amount of time, t, used for a particular run of the model was actually input as mont hs because that is the unit of each timestep. The year or time-step the event is deposited in is represented by t an d is also entered into the model as a month. The initial thickness of each event was thicker in the proximal core compared to the thickness in the distal core. Th is was expected as an overwash fan thins with its extent. However, Event II modelin g resulted in the same original thickness in both cores T2C1 and T3C2. This may be attributed to the differe nce in the number of time-steps used for each or that in actuality they do not represent the same event. Event III in core T3C2 wound up being two separate events that app ear to be one when looking at the proxy data (Fig. 5-2 B). Modeling these events produced results that are at odds with the theory of event preservation in that there are event beds pres erved which originated with thicknesses less than the depth of mixing. According to the theory, these events should not be recognizable within the

PAGE 56

56 sedimentary record. One explanation for this di screpancy is that as a coastal environment the salt marsh is very dynamic and the assumed steady st ate parameter values us ed are inaccurate for the entirety of the marsh and even each core. It also may be attri buted to the lack of macrofaunal mixing. Figure 5-1. Physical Properties of Cores T2C1 and T3C2. Lithological examination of salt marsh cores using textural analysis (percen t sand), residuals of gamma bulk density, and pixel intensity of x-radiographs. These ar e the three main proxies used to identify event beds. Event beds I, II, and III (used in Bentley Bioturbation Model) are shaded in blue. A) The cores proximal to the b each, collected 20 meters from high tide (determined from wrack line). The sedime nts are comprised of organic-rich sandy muds. A

PAGE 57

57 Figure 5-1. Continued. B) The cores distal to the beach, coll ected 30 meters from high tide (determined from wrack line). The sedime nts are comprised of organic-rich sandy muds. B

PAGE 58

58 Figure 5-2. Modeling Results for Cores T2C1 and T3 C2. A) Results from the BBM for Events I, II, and III in the proximal core T2C1. Th e cores proximal to the beach, collected 20 meters from high tide (determined from wr ack line). Plots represent the minimum and maximum preservation from the x-radiographs with the results of the BBM for each given event. A

PAGE 59

59 Figure 5-2. Continued. B) Results from the BBM for Events I, II, and III in the distal core T3C2. The cores distal to the beach, collected 30 meters from high tide (determined from wrack line). Plots represent the minimu m and maximum preservation determined from the x-radiographs with the result from the BBM for each event. B

PAGE 60

60 CHAPTER 6 CONCLUSIONS The depositional environments on St. Vincent Island offer an advantageous environment for studying the preservation of storm deposits in the coastal stratigraphi c record due to the frequent occurrence of large storms, relatively low mixing rates (i.e., long dissipation times) and fast sedimentation rates (a few mm y-1). Event beds are likely to be preserved within the salt marsh sedimentary record. This study found evidence of event be d preservation within the salt marsh investigated. They are not however, repr esented solely by a change in grain size, but rather by other changes with in the sediments as well (i .e., lithologic changes). The ultimate control on or primary function of preservation potential is the initial or original thickness of the event la yer and the bioturbation factors pr esent. Event beds are more likely to be found proximal to the beach and less likely to be found with increasing distance. Initial or original event be d thickness is dependant on co stal morphology and vegetation. This study also indicates pixel intensity (i.e., x-radiography) to be an extremely useful proxy of preservation; perhaps the ideal proxy for th is type of research. It is extremely high resolution with minimal destruction to the cores. It also offers the capability of locating very fine beds / laminations within the stratigraphic record. The Bentley Bioturbation Model can be appl ied to coastal marshes. The Bentley Bioturbation Model, which was developed for co ntinental shelf environments, is applicable within coastal marsh environments. It successfully reproduced the chosen event beds found within T2C1 and T3C2. However, it should be used with caution as th e dynamics of coastal environments may create inaccurate results for this steady state model. It should be tested in other sedimentary environments, es pecially in the coastal realm.

PAGE 61

61 LIST OF REFERENCES Allen, J.R.L. (2000) Morphodynamics of Holocene Salt Marshes: A Review Sketch from The Atlantic and Southern North Sea Coasts Of Europe. Quaternary Science Review 19 11551231. Allen, E.A., & Curran, H.A., (1974) Biogenic Sedi mentary Structures Produced By Crabs in Lagoon Margin And Salt-Marsh Environmen ts Near Beaufort, North-Carolina. Journal of Sedimentary Petrology 44 (2), 538-548. Aller, R.C., Benninger, L.K., & Cochran, J.K. (1 980) Tracking Particle-Asso ciated Processes in Nearshore Environments by use of 234TH/238U Disequilibrium. Earth and Planetary Science Letters 47 161-175. Appleby, P.G., & Oldfield, F. (1992) Application of Lead-210 to Sedimentation Studies. In: Uranium-series Disequilibrium: Applica tions to Earth, Marine, and Environmental Sciences, Vol. 0 (Ed. By I.M. and R.S. Harmon) Clarendon Press, Oxford, pp.731-778. Benninger, L.K., Aller, R.C., Cochran, J.K., & Turekian, K.K. (1979) Effects of Biological Sediment Mixing on the 210Pb Chronology and Trace Metal Distribution in a Long Island Sediment Core. Earth and Planetary Science Letters 43 241-259. Bently, S.J., & Sheremet, A. (2003) New M odel for the Emplacement, Bioturbation, and Preservation of Fine-Scaled Sedimentary Strata. Geology 31 (8), 725-728. Boggs, S. (2001) Principals of Sedimentology and Stratigraphy Third Edition. Upper Saddle River, Prentice Hall, pp.59-74. Boudreau, B.P. (1994) Is Burial Velocity a Master Parameter for Bioturbation? Geochimica et Cosmochimica Acta 58 1243-1249. Bromley, R.G. (1996) Trace Fossils Biology, Ta phononmy, and Applications Second Edition: London, Chapman & Hall, pp.225-229. Campbell, K. (1986) St. Vincent Island (St. Vincent National Wildlife Refuge), Florida. Geological Society of America Centenni al Field Guide Southeastern Section 351-353. Collins, E.S., Scott, D.B., & Gayes, P.T. (1999) Hurricane Records on the South Carolina Coast: Can They be Detected in the Sediment Record? Quaternary International 56 15-26. Cutshall, N.H., Larsen, I.L., & Olsen, C.R. ( 1983) Direct Analysis of 210Pb in Sediment Samples: Self-Adsorption Corrections Nuclear Instrumentation and Methods 206, 309312. Davis, R.A. (1995) Geologic impact of Hurrican e Andrew on Everglades Coast of Southwest Florida. Environmental Geology 25 (3), 143-148.

PAGE 62

62 Davis, R.A., Knowles, S.C., & Bland, M.J. (1989) Role of Hurrican es in the Holocene Stratigraphy of Estuaries Examples from the Gulf Coast of Florida. Journal of Sedimentary Petrology 59 (6), 1052-1061. Donnelly J.P., Bryant, S.S., Butler, J., Do wling, J., Fran, L., Hausmann, N., Newby, P., Shurman, B., Stern, J., Westover, K., & Webb, T. III. (2001) 700 Yr Sedimentary Record of Intense Hurricane Landfalls in Southern New England. GSA Bulletin 113 (6), 714-727. Dukat, D.A., & Kuehl, S.A. (1995) Non-Stea dy State 210-Pb Flux and the Use of 228Ra/226Ra as a Geochronometer on the Amazon Continental Shelf Marine Geology 125, 329-350. Edwards, J.M., & Frey, R.W. (1977) Substrate Characteristics within a Holocene Salt Marsh, Sapelo Island, Georgia. Senchenberg. Marit. 9 215-259. Frey, R.W., & Basan, P.B. ( 1978, 1985) Coastal Salt Marshes: in, Davis, R.A., ed., Coastal Sedimentary Environments: New York, Springer-Verlag, pp.225-301. Frey, R.W., & Howard, J.D. (1969) A Profile of Biogenic Struct ures in a Holocene Barrier Island-Salt Marsh Complex, Georgia. Transactions of the Gulf Coast Association of Geological Sciences 19 427-444. Green, M.A., & Aller, R.C. (2001) Early Diagen esis of Calcium Carbonate in Long Island Sound Sediments: Benthic Fluxes of Ca+2 and Mi nor Elements During Seasonal Periods of Net dissolution. Journal of Marine Research 59 769-794. Harris, P.T., Heap, A.D., Bryce, S.M., Porter-Smith, R., Ryan, D.A., & Heggie, D.T. (2002) Classification of Australian Clastic Co astal Depositional Environments Based upon a Quantitative Analysis of Wave, Tidal, and River Power. Journal of Sedimentary Research 72 855-870. Hart, M.R. (2003) Evaluating the Preservati on of Hurricane Deposits in Florida Coastal Sediments. University of Florida, Masters Thesis. Katz, L.C., 1980. Effects of Burrowing by th e Fiddler Crab, Uca-Pugnax (Smith). Estuarine and Coastal Marine Science 11 (2), 233-237. Keen, T.R., & Stone, G.W. (2000) Anomalous Re sponse of Beaches to Hurricane Waves in a Low-Energy Environment, Northeast Gulf of Mexico, U.S.A. Journal of Coastal Research 16 (4), 1100-1110. Koretsky, C.M., Meile, C., & Van Cappellen, P. (2002) Quantifying Bioirrigation Using Ecological Parameters: A Stochastic Approach. Geochemistry Transactions 3 (3), 17-30. Liu, K., & Fearn, M.L. (1993) Lake-Sediment R ecord of Late Holocene Hurricane Activities from Coastal Alabama. Geology 21 793-796.

PAGE 63

63 Liu, K., & Fearn, M.L. (2000) Reconstruction of Prehistoric Landf all Frequencies of Catastrophic Hurricanes in Northwestern Florida from Lake Sediment Records. Quaternary Research 54 238-245. Liu, K., & Fearn, M.L. (2002) Lake Sediment Evidence of Coastal Geologic Evolution and Hurricane History from Western Lake Florida. Quaternary Research, 57 (3) 429-431. McCraith, B.J., Gardner, L.R., Wethey, D.S., & Moore, W.S., 2003. The Effect of Fiddler Crab Burrowing on Sediment Mixing and Radionuclid e Profiles Along a Topographic Gradient in a Southeastern Salt Marsh. Journal of Marine Research 61 (3), 359-390. Nittrouer, C.A., DeMaster, D.J., McKee, B.A., Cu tshall, N.H., & Larsen, I.L. (1984) The Effect of Sediment Mixing on Pb-210 Mixing Rate s for the Washington Continental Shelf. Marine Geology 54 210-222. National Oceanic Atmospheric Administration s National Weather Service (2003) National Hurricane Center, http://www.nhc.noaa.gov/index.shtml updated regularly. Rappaport E. N. (1995) Hurricane Erin 31 July 6 August 1995. National Hurricane Center Preliminary Report, 26 November 1995 http://www.nhc.noaa.gov/1995erin.html last updated 02 January 1999. Risi, J.A. (1998) Event Sedimentation from Hurricane Andrew Along the Southwest Florida coast, University of Miami, Ph.D. Dissertation, pp.199. Rose, P.R., & Lidz, B. (1977) Diagnostic Forami nifera Assemblages of Shallow-Water Modern Environments : South Florida and the Baha mas. Comparative Sedimentology Laboratory, Division of Marine Geology and Geophysic Rosenstiel School of Marine and Atmospheric Science, The University of Miami, pp.56. Simpson & Herbert. (1973) Monthly Weather Review. 101 (4). Thieler, E.R., & Hammar-Klose, E.S. (1999) Nati onal Assessment of Coastal Vulnerability to Sea-Level Rise: Preliminary Results for the U.S. Gulf of Mexico Coast. U.S. Geological Survey Open-File Report 00-179 http://pubs.usgs.gov/of/2000/of00179/textonly/textindex.html last updated 06 Sep 2001. Wheatcroft, R.A. (1990) Preservation Po tential of Sedimentary Event Layers. Geology 18 843845. Wheatcroft, R.A., & Drake, D.E. (2003) Post -Depositional Alteration and Preservation of Sedimentary Event Layers on Con tinental Margins, I. The Role of Episodic Sedimentation. Marine Geology, 199 123-137. Wheeler, A.J., Orford, J.D., & Dardis, O., 1998. Saltmarsh Deposition and its Relationship to Coastal Forcing Over the Last Centur y on the North-West Coast of Ireland. Geologie en Mijnbouw 77 (3-4), 295-310.

PAGE 64

64 Williams, J.M., & Duedall, I.W. (2002) Flor ida Hurricanes and Tropi cal Storms 1871-2001: Gainesville, University Press of Florida, pp.88-100. Woodroffe, C.D. (2002) Coasts: Form, Process, & Evolution: Cambridge University Press, pp. 378-432.

PAGE 65

65 BIOGRAPHICAL SKETCH Lisa Marie Erickson Mertz was born May 27, 1979, to Curtis and Jaclyn Erickson. Growing up in Minnesota with a younger sister, Lisa Marie always enjoyed the outdoors. A honor graduate of Proctor Senior High Sc hool in 1997, she was involved in drama and volleyball. She continued a volleyball career at the University of Wisconsin River Falls while studying geology and hydrogeology. In 2001, she be gan her graduate studies under Dr. John M. Jaeger in the University of Floridas Departme nt of Geological Scien ces. Her interests in sedimentology lead her to a salt marsh and its history of tropical cycl ones. Lisa Marie is currently working with an environmental a nd engineering consulting firm in western Washington.


Permanent Link: http://ufdc.ufl.edu/UFE0013200/00001

Material Information

Title: Event Bed Preservation Potential in a St. Vincent Island Salt Marsh in Florida
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0013200:00001

Permanent Link: http://ufdc.ufl.edu/UFE0013200/00001

Material Information

Title: Event Bed Preservation Potential in a St. Vincent Island Salt Marsh in Florida
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0013200:00001


This item has the following downloads:


Full Text





EVENT BED PRESERVATION POTENTIAL IN A ST. VINCENT ISLAND SALT MARSH
IN FLORIDA




















By

LISA MARIE ERICKSON MERTZ


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA

2007

































Copyright 2007

by

Lisa Marie Erickson Mertz

































This is dedicated to Little Miss Megan and Hermes. I will always love you both. Megan, you've
always made me want to be a better person so you'd have someone to look up to. I have come to
discover that you are the one I look up to; there have been times that your existence has made it
all worthwhile. Hermes, you may not be with me now but you were when I needed you most.









ACKNOWLEDGMENTS

I gratefully acknowledge the logistical support of the St. Vincent National Wildlife Refuge

for granting access to the island and aid in sampling. I also thank the Gulf Coast Association of

Geological Sciences for the grant and opportunity to present research they helped to fund.

Additional thanks go to the Department of Soil and Water Science at the University of Florida

for use of their coring equipment. A big thank you goes to Dr. Samuel Bentley at Louisiana

State University for allowing me to use his bioturbation model and all the assistance he provided

in its application. Most importantly, I thank Dr. John Jaeger for taking me on as a student;

budgeting me into his projects because I wanted to do non-funded research; teaching me so much

about so many things; and being a great person to know, work with, and work for.

Special thanks go to the following: my parents for pushing me to reach for the northern

lights and supporting all my endeavors; Daniel Gorman for jumping in and staying to the very

end; Rebekah Wagner, my soul-mate, this is all her fault and would have never happened if it

weren't for her; KillerD and Chocolate, who rock; Mike Mertz for a good beginning; MaryLea

Hart for sparking the interest that became my life for three years; Marisa Martinovich for her

assistance and keeping me company through the monotony of wet sieving; and all my family and

friends for just being there.









TABLE OF CONTENTS

page

A C K N O W L E D G M E N T S ..............................................................................................................4

LIST O F TA B LE S ......... ..... .............. ................................................................. 7

LIST OF FIGURES .................................. .. ..... ..... ................. .8

ABSTRAC T .........................................................................................

CHAPTERS


1 INTRODUCTION ............... .............................. ........................ ..... 11

2 B A CK G R O U N D .......................................... ................ ......................... .... 17

S alt M arsh es ................... ...................1...................7..........
C o n sep tu al M o d el .......................................................................................................1 8
Num erical M odel ................................................................ ..... ...... ........ 18
Stu dy A rea ................... ...................1...................9..........
St. V incent Island ........................................ ..... .........................19
Tropical Cyclones and Storms Affecting the Florida Panhandle ..................................20
M major hurricanes ......................................... ....................... ....... 21
M inor hurricanes ......................... ............ ......................... 22

3 METHODS .........................................24

S am ple C collection ......................................................................24
C o re L o g g in g .................................................................................2 5
L ith o lo g y ................... ...................2...................6..........
Gamma Spectroscopy ..................... ..................................... 27
G ra in S iz e .............. .... ...............................................................2 8
E vent D term nation ......................................................................................................29
M o d e lin g ................... ...................2...................9..........

4 R E SU L T S .............. ... ................................................................33

L ith o lo g y ................... ...................3...................3..........
G rain Size A naly sis ................................................................35
G am m a B ulk D density ....................................................... 36
M magnetic Su sceptibility ................................................................36
X -radiograph P ix el Inten sity ............................................................................................. 37
Chronology .........................................37
L e a d -2 1 0 ..........................................................................................................................3 7
Cesium -137 ............... ...... ...................... .. ................ 38









5 DISCU SSION ......... .. ....... .... .... .. ........ ...................... ........ 50

Lithology and Evidence of Storm Bedding ........................................ ........................ 50
M modeling Preservation Potential of Storm Beds ........................................ .....................52

6 C O N C L U SIO N S ................. ......................................... ........ ........ ..... .... ...... .. 60

L IST O F R E F E R E N C E S .............................................................................. ...........................6 1

B IO G R A PH IC A L SK E T C H .............................................................................. .....................65













































6









LIST OF TABLES

Table page

5-1 Constant BBM Param eters........................................................... .. ............... 54

5-2 V variable by Event BB M Param eters...................................................................... ...... 55









LIST OF FIGURES


Figure pe

1-1 Effects of Hurricane Isabel (2003), Hatters Village, North Carolina ..............................14

1-2 Preservation Potential of an Event Bed .............. .... ............................ ............. 15

1-3 Hurricane Landfalls and St. Vincent Island............... ........ ...... ..... .. ............ 16

3-1 Lithology of Cores ................... ...................................... .. ............. 31

4-1 Percent Sand of Cores T2C1 and T3C2...................................................... ..................39

4-2 Percent Sand and Percent Organic Matter for Cores T2C1 and T3C2 .............................40

4-3 Textural Analysis of Cores T2C1 and T3C2 .............. ......... .................. .............. .41

4-4 Gam m a Bulk D density of Cores ......... .................................... ................. ............... 43

4-5 Residual Bulk Density of cores T2C1 and T3C2......................................45

4-6 Pixel Intensity of Cores T2C1 and T3C2 .............................. ... .... ............... 46

4-7 Total Activity of Lead-210 and Total Activity of Lead -210 Corrected for Grain Size.... 47

4-8 Total Corrected Activity of Cesium-137 and Cesium-137 Activity Corrected for
G rain S iz e .......................................................4 9

5-1 Physical Properties of Cores T2C1 and T3C2................................ .............56

5-2 Modeling Results for Cores T2C1 and T3C2 .............................. ..............58









Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

EVENT BED PRESERVATION POTENTIAL IN A ST. VINCENT ISLAND SALT MARSH
IN FLORIDA

By

Lisa Marie Erickson Mertz

May 2007

Chair: John M. Jaeger
Major Department: Geological Sciences

Tropical cyclones affect the Gulf of Mexico coastline annually; however, the historical

record of these events is limited to about 400 years. To establish the occurrence frequency and

the importance of cyclones in creating strata, this record needs to be validated and lengthened by

examining coastal sediments for the preservation potential of storm bedding. St. Vincent Island

National Wildlife Refuge is located in Apalachicola Bay, FL, and provides a research area with

minimal human impact and frequent occurrence of tropical cyclone landfalls, making it ideal to

study the natural processes controlling event bed preservation. Cores were collected in a salt

marsh to establish preservation potential of storm bedding. The preservation potential of a storm

event in coastal sediments is related to three factors: biologic mixing depth and intensity, storm

layer thickness, and sediment accumulation rate in the coastal environment.

Storm deposition can be detected by changes in bulk density, magnetic susceptibility, or

lithology. Radioisotopes were used to quantify the following processes: mixing depth and

intensity by 234Th and 7Be; accumulation rates by 210Pb and 137Cs. Mixed depths are <3 cm in

coastal marshes and sedimentation rates are <5 mm y1. Comparison of transit time of a storm

layer through the surface mixed layer to the dissipation time of the bed through mixing reveals

that beds >3 cm thick are likely preserved. The most important control on preservation appears









to be the initial thickness of the storm bed, which is expected to be highly variable for each storm

and is controlled by the antecedent topography of the beach/dune system, where the presence of

topographic lows ("blow-outs") allows for more overwash and thicker beds.









CHAPTER 1
INTRODUCTION

In 2002 the United States Census Bureau reported that the population of Florida was just

under 16 million and had grown 24% since 1990, making it the fastest growing state in the US.

By the year 2030, the population is expected to be about 26 million. Most of the population

growth is expected to be in the larger cities located along the coasts (e.g., Tampa/ St. Petersburg,

Miami, Jacksonville). As a peninsular state, separating the Atlantic Ocean and Gulf of Mexico,

Florida has 2,170 kilometers of coastline which is annually subjected to tropical cyclones and

storms. Along this coastline are the structures (buildings and homes) creating the cities where

the majority of Florida's population resides. Insurance companies base hurricane insurance rates

on the frequency of landfalls within the region the insured structure is located. The historical

record of tropical cyclones affecting the Florida Gulf of Mexico coastline is limited to human

accounts (about 400 years). It has been proposed that this record may be extended backward by

identifying and analyzing event-bedding in coastal sediments (Donnelly et al., 2001; Lui &

Fearn, 1993, 2000, & 2002), which in turn may facilitate the ability to make future predictions

about tropical storm and cyclone activity.

Tropical cyclones and tropical storms release a great deal of energy along coastlines

damaging buildings and property, but also resuspending and reworking sediments that may be

redeposited anywhere water or wind may carry them. For example, Figure 1-1 shows the effects

of Hurricane Isabel (2003) along the North Carolina coastline; the storm wind, waves, and tides

dramatically altered the coastline, moving buildings, trees, and sediment.

Coastal sedimentary environments such as this act as long-term recorders of environmental

conditions (Boggs, 2001), so event bedding, such as hurricane deposits, have the potential to be

preserved within coastal sedimentary records. However, the preservation potential of any









depositional event is environmentally dependent, varying according to bioturbation intensity,

depth of sediment mixing (biological and physical), burial rate, and original event bed thickness

(Wheatcroft and Drake, 2003; Bentley and Shermet, 2003).

Preservation potential of any event bed is a function of how rapidly subsequent

sedimentation buries it through the potentially rapid-mixed surface layer (Fig. 1-2). This mixing

acts to degrade the "intensity" of the event signal (Wheatcroft, 1990). Consequently, the most

important variables of preservation potential to measure would include the following: 1)

sediment burial rate, 2) mixing layer thickness, 3) bioturbation intensity, and 4) original event

layer thickness. In order to better resolve the most important characteristics that favor storm bed

preservation, it is possible to establish a set of criteria favoring preservation based on

paleocyclone studies (e.g., Davis et al., 1989; Donnelly et al., 2001; Liu and Fearn, 2000) that

have been performed on a number of different coastal depositional environments (e.g., subtidal,

intertidal, and supratidal). An ideal environmental setting would fit the following criteria: 1)

high sedimentation rates that quickly bury event beds and prevent sediment mixing; 2) regular

cyclone activity resulting in likely production of event bedding; and 3) unaffected by frequent

tidal and sea level fluctuations that may erode/mix event beds after deposition (Wheatcroft and

Drake, 2003). Although, Wheatcroft and Drake (2003) studied event bedding on a continental

shelf, the basic concept of their study and its criteria should be applicable to any coastal

environment.

Salt marshes along the northern Gulf of Mexico meet two of these criteria: regular tropical

cyclone activity (Fig. 1-3 A) and negligible effects of tidal and sea level changes (barrier island

salt marshes remain at sea level if supplied with sediment). Also, the marsh vegetation traps

sediment within the root masses and stalks aiding in the deposition of an event-bed within the









salt marsh. Salt marshes typically are composed of organic-rich sandy muds (Allen, 2000).

Therefore, the deposition of non-local sediment resuspended during storm events may be

detected by changes in bulk density, magnetic susceptibility, or lithology (e.g., Donnelly et al.,

2001).

The preservation potential of event beds in salt marshes has not been determined or

quantified, either in the Gulf of Mexico or elsewhere. This study is an evaluation of storm-bed

preservation potential in a salt marsh located on St. Vincent Island, Florida (Fig. 1-3 B). This

project was designed to test the hypothesis that intense biological mixing in salt marshes (Frey &

Bassan, 1985) may make detection of storm deposits the sedimentary record nearly impossible;

unless the deposited layer is thicker than the depth of mixing. To test this hypothesis, several

cores were collected in a salt marsh fronted by a low berm next to a small beach (Fig. 1-3 B) and

analyzed for lithology, bulk density, magnetic susceptibility, organic content, and grain size and

any abrupt changes within to determine possible preservation potential. Numerical modeling of

the specific event-beds (Bentley & Sheremet, 2003) was used to examine what conditions are

required for preservation of these beds.










S1998













---I "
September 19, 2003






Jr










Figure 1-1. Effects of Hurricane Isabel (2003), Hatters Village, North Carolina. Aerial
photographs taken in 1998 and just after the landfall of Hurricane Isabel in 2003.
Images courtesy of National Oceanographic and Atmospheric Association and the
National Geodetic Survey.









Sand


-----


% Sand


m m- m -m -


S----


Dissipation Time (T) = f(L,, D)
SML Transit Time (TT) = f(L,, Lb, w)

Figure 1-2. Preservation Potential of an Event Bed. This is dependent upon dissipation time
(TD) of the event bed and the surface mixed layer (SML) transit time. TD is a
function of the event layer thickness (Ls) and mixing intensity and biodiffusion (Db).
TT is a function of the event layer thickness (Ls), the depth of mixing (Lb) and the
sediment accumulation rate (co). So, if TD=10 and TT=1 then TD>TT, event bedding
is dissipated and not preserved (top); if TD=1 and TT=100 then TD< preservation is possible (bottom). Adapted from Wheatcroft, 1990.


Lb












m Lb
























Hurricane landfall 1885 to 1984.
Davis et al, 1989


A















B ......
T
Marsh B



Figurel-3. Hurricane Landfalls and St. Vincent Island. A) Hurricane landfalls along the coast of
Florida from 1885-1984. In the panhandle region of Florida, 56% of landfalls have
occurred in the Apalachicola Bay area. B) Digital orthophoto quarter quad of St.
Vincent Island. Location of the depositional environment (salt marsh) chosen for
study. Inset is a blow-up of Marsh B showing core collection location.









CHAPTER 2
BACKGROUND

Salt Marshes

Salt marshes are an upper intertidal to supratidal muddy environment typical of temperate

latitudes that host a range of halophilic plant and biological communities (Frey and Bassan,

1978; Woodroffe, 2002). Gulf Coast and southeastern U.S. marshes typically consist of

topographically lower regions adjoining tidal creeks. The dominant vegetation consists of the

cordgrasses Spartina alterniflora and Spartina cynosuroides that generate thick, dense root

masses (Edwards and Frey, 1977). The topographically elevated high marsh consists of salt

tolerant plants along with Spartina spp.

Gulf Coast marsh sedimentary facies include the following scenarios: root mats, organic

oozes grading into clays; vegetation mats, coarse to medium organic fibers, dark or silty clays; or

vegetation mats, and firm clays (Edwards and Frey, 1977; Frey and Bassan, 1985). The typical

grain size for the lower salt marsh is in the silt range (4-63 tm) and sediments usually have high

organic content, whereas the high marsh tends to be sandier. Spartina decomposes slowly,

leaving thick mats of organic matter that can bind sediment and limit resuspension (Frey &

Basan, 1978). Peat layers are also common within salt marshes, depending on the decomposition

rate of the vegetation. Combined sea-level rise and autocompaction of marsh sediments are

generally balanced, maintaining marsh elevations (Allen, 2000).

Mixing of southeastern U.S. marsh sediment by plant roots and macroinvertebrate faunas

often completely destroys any evidence of primary depositional fabric (Edwards and Frey, 1977).

Bioturbation causes mixing that penetrates down 10 to 100 cm in the sediment (Koretsky et al.,

2002). Also, "biologic ingestion, digestion, and egestion" may cause changes in grain-size and









chemical composition in the sediments, which may degrade or alter any depositional bedding

(Koretsky et al., 2002).

Conseptual Model

Wheatcroft's (1990) conceptual model of event-bed preservation was adapted for this

study (Fig. 1-2). This model basically states that a portion of an event bed will be preserved if 1)

the event-bed is thicker than the depth of mixing or 2) the sedimentary burial rate is fast enough

to bury the event bed before it can be mixed beyond recognition of analytical techniques.

Dissipation time (TD) is the time it takes for an event bed to be dissipated or destroyed by

bioturbation. According to Wheatcroft and Drake (2003), the destruction and dissipation of

sedimentary fabric occurs at a faster rate than the destruction and dissipation of sedimentary

textures and physical properties. The surface mixed layer (SML) is the thickness from the

sediment surface to the greatest depth of rapid biologic mixing. Transit time (TT) is the amount

of time an event bed spends within the surface mixed layer. TD is a direct function of the mixing

intensity as represented by a biodiffusion coefficient (Db). An increase in mixing intensity (Db)

leads to a faster dissipation time, TD. The SML transit time TT is a direct function of the event

layer thickness (Ls), depth of mixing (Lb), and sedimentation/burial rate (co). If the dissipation

time is shorter than SML transit time, the event bed is destroyed and not recorded in the

sedimentary record (Fig. 1-2). However, if SML transit time is much less than the dissipation

time, a portion of the original event bed will be buried and preserved within the sedimentary

record. For example, if it takes 5 years to dissipate the bed, it would be preserved if the transit

time was only 1 year, but completely destroyed if it took 10 years to transit the SML.

Numerical Model

The Bentley Bioturbation Model (Bentley & Sheremet, 2003) was used to quantitatively

examine event beds for this study. It assumes that depositional fabric is irreversibly transformed









into bioturbated fabric leaving behind some preserved depositional fabric (q). The model uses

Eq. 2-1 and 2-2 (Bentley & Sheremet, 2003). The equation used is determined by the depth in

sedimentation over time compared to the depth of mixing (see Eq. 2-1 and 2-2).

q = exp[ (-ao/p) ((exp(-pz)-1) / coo+Q) ] ;if z(At) > Lb (Equation 2-1)
q = exp[ (-ao/p) [ (exp(-pz) exp(-zAt) ) / coo ] + [ (exp(-pzAt)-1) / (oo+D) ] ];
if z(At) < Lb (Equation 2-2)
where q is the preservation quotient (%), ao is the sediment surface bioturbation rate
(biodiffusion coefficient) (cm2 yr-1), 0 is the constant controlling the exponential
attenuation of ao with depth in the sediment, coo is sediment accumulation rate prior to the
event (cm yr1), co is the sediment accumulation rate following the event (cm yr-1), Q is the
rate of supplemental instantaneous sedimentation due to event bed deposition (cm yr-1), z
is depth in the sediment (cm), At is the change in time (yr) through each run of the model,
and Lb is the depth of biological mixing (cm)

Beta is determined empirically by Eq. 2-3 (Bentley & Sheremet, 2003). The initial value

for az is assumed to be 0.001 ao (i.e., 0.1%) as the biological mixing rate at Lb is much, much

less than at the surface where mixing is rapid (Boudreau, 1986). Chemical and physical

variability within the sediment (lateral and temporal), erosion, and consolidation are not

accounted for in this model. Following every model run, the output is q, the percent preserved,

with dz, the depth.

az=ao exp(-pz) (Equation 2-3)
where az is very close to zero at the depth Lb and is forced to zero for all depths greater
than Lb
Study Area

St. Vincent Island

St. Vincent Island is an uninhabited 50 km2 barrier island along the northwestern Florida

panhandle (Fig. 1-3 B). In 1968 the U.S. Fish and Wildlife Service (FWS) purchased St. Vincent

Island from the Loomis estate in accordance with the Migratory Bird Conservation Act, creating

the St. Vincent National Wildlife Refuge (SVNWR) (SVNWR Narrative Report, 1868; SVNWR

Management Review, 2000). Prior to this sale, the island was owned privately by four other

families or companies. For the most part it was treated as a wildlife sanctuary and was managed









in that manner. The most significant human impact occurred from logging in 1940-1945 and

1960-1965 when nearly 128.7 km of gravel roads were built (Fig. 1-3 B), creating "breaks

among the natural communities and disrupted the natural hydrology of the island" (SVNWR

Narrative Report 1968; SVNWR Management Review 2000). However, near the marsh study

area, roads were not constructed and human interference and influence has been minimal,

making this portion of St. Vincent Island an ideal location for this study.

It is a triangular shaped island 14.5 kilometers long and 6.4 kilometers wide. Elevation

ranges from 0.9 to 3.5 meters above mean sea level. There were 11 km2 of estuarine marsh in

2001 (SVNWR Narrative Report, 2001). The main salt marsh, Mallard Slough, is located in the

northeast area of the island near the study area. Water flows east to southeast through the Big

Bayou into Mallard Slough. Vegetation of the marsh consists of Distichlis spicata, Spartina

bakeri, and Juncus spp. Throughout the Holocene, sediment has been supplied to the island

mainly from the Apalachicola River and consists largely of quartz sand (Campbell, 1986). Some

sediment is supplied by the longshore current which travels east to west in this portion of the

Gulf of Mexico. Tidal data for the island proper are unavailable. In nearby Apalachicola, FL,

however, mean tidal range is 0.34 meters with a diurnal range of 0.51 m; maximum water level is

3.53 m above the mean high; and the minimum water level is -0.52 m below the mean low

[NOAA Center for Operational Oceanographic Products and Services (CO-OP) station #

8728690 (29043.6'N, 84058.9'W)].

Tropical Cyclones and Storms Affecting the Florida Panhandle

Hurricanes impact land with wind, waves, rain, and storm surges, which elevate sea levels

sufficiently to inundate supratidal land. Florida, being a peninsula, is impacted by hurricanes

from both the Atlantic and Gulf of Mexico. The written historical record of tropical cyclones

and tropical storms is limited to about 400 years and for the Apalachicola Bay area this record









extends only to the 1800s. Apalachicola Bay has had a significant number of recorded hurricane

landfalls (Fig. 1-3 A); 56% of the hurricanes that hit the Florida panhandle from 1885 to 1984

occurred in the Apalachicola Bay Area (Davis et al., 1989). As an area that was and continues to

be somewhat sparsely populated, some storm details (precipitation, wind speeds, barometric

pressure, etc.) are not known, especially concerning storms that occurred before the 1950s. For

this study, only the hurricanes (major and minor) that were likely to have a direct impact on St.

Vincent Island were researched extensively and described in detail below.

Major hurricanes

In 1886 (June), Apalachicola-Tallahassee (unnamed) was classified as extreme

(classification prior to 1970: mean wind speed of 220 km hr-) initially. With the Saffir/Simpson

Scale, however, it would be a major (category three or higher) hurricane. Little information is

known about this event except for its flooding high tides (Williams & Duedall, 2002).

In 1975 (September), Eloise (Category 3) made landfall just west of Panama City Florida.

Tides in Panama City were measured 3.7 4.9 meters above average. Winds were sustained at

200 km hr-1 with 250 km hr-1 gusts; about 38.1 cm of rain fell at Eglin Air Force Base. Winds

and tides caused the most damage along the coast (Williams & Duedall, 2002). Eloise is ranked

22nd on the Costliest U.S. Hurricanes and 44th on the Most Intense Hurricanes in the United

States. Both lists are from 1900-2000 (NOAA, 2003).

In 1985 (August September), Hurricane Elena (Category 3) approached the west coast of

Florida from Venice to Pensacola. Although never making landfall in Florida the waves and

storm surge (2.1 2.7 meters) caused major evacuations (one million people) of coastal areas

(Williams & Duedall, 2002). In Apalachicola 28.7 cm of rain fell and tides were 1.4 2.8 m

above MHHW and the pressure was 998 millibars. Winds in Carrabelle were measured at 201.2

km hr- as the eyewall passed 16.1 to 22.5 km south of the island traveling west (SVNWR









Narrative Report, 1985). Elena is ranked 14th on the Costliest U.S. Hurricanes and 56th on the

Most Intense Hurricanes in the United States. Both lists are from 1900-2000 (NOAA, 2003).

In 1995 (Late September to Early October), Hurricane Opal (Category 5 in Gulf of

Mexico; Category 3 at landfall) made landfall between Destin and Panama City Florida with 200

km hr-1 winds and gusts of 230 km hr-1. The storm surge was 4.6 meters and affected the coast

from Alabama to Cedar Key, Florida (Williams & Duedall, 2002).

Minor hurricanes

In 1972 (June), Agnes's landfall was near Apalachicola and Port St. Joe, Florida. This

Category One hurricane had peak winds of 140 km hr-1, a maximum surge of 2.1 m, and a

reported 32.3 cm of rainfall. At Apalachicola there was 8.6 cm of rainfall, a 2.0 meter storm

surge, 90 km hr-1 winds, and 987 millibars of pressure. This large storm was 1609 km in

diameter (1852 km circulation envelope) (Williams & Duedall, 2002; Simpson & Herbert, 1973).

In 1985 (November 21), Hurricane Kate (Category 2) made landfall near Port St. Joe, was

not as destructive as Elena, and damage was attributed mostly to the winds and storm surge

(maximum at 12.9 meters) (Williams & Duedall, 2002). The maximum water level ever

measured in Apalachicola occurred on November 21, 1985 (NOAA, 2003). There was a storm

surge of about 2 meters (Keen & Stone, 2000). Locals claim that during the storm the western

half of St. Vincent Island was completely submerged. This would qualify as sheet overwash

large enough to carry sand into the backbarrier marshes of the island (e.g., Donnelly et al., 2001).

According to the SVNWR 1985 Narrative Report, Kate's eyewall moved across the island's long

axis at about 4:45 pm and brought with it a 2.4 3.1 meter surge.

In 1995 (June 5), Allison (Category 1 in Gulf of Mexico; Weak Category 1 to strong

tropical storm at landfall), the earliest storm ever to hit Florida, made landfall just east of St.

George Island. Apalachicola experienced the most damage with extreme tides, 12.7 cm of rain,









and maximum winds of 120 km hr-1. Sustained winds at St. George Island were 100 km hr-1;

Apalachicola reported 60 km hr- winds with 70 km hr-1 gusts. The storm surge in Franklin

County was 1.2-1.8 meters (Pasch, 1996). (July August) Category 1 Erin entered the Gulf of

Mexico by crossing the Florida Peninsula then traveled west roughly parallel to the Panhandle

coastline before making her second landfall in Pensacola (Williams & Duedall, 2002). Erin is

number 18 on the Costliest U.S. Hurricanes 1900-2000 list (NOAA, 2003). Apalachicola had

gusts of 90 km hr-1 while St. George Island reported gusts of 118.5 km hr-1 (Rappaport, 1995).









CHAPTER 3
METHODS

Sample Collection

Two sites were chosen for this study: Marsh A and Marsh B (Fig. 1-3 B). Marsh A is

located on a small spit in the northwestern entrance of Big Bayou (Fig. 1-3 B). This site was

chosen based on the ability for storm surges to completely inundate the spit with water from

either side, either Big Bayou or St. Vincent Sound. It also represents a more protected marsh

environment with smaller fetches on each side of the marsh. Marsh B is located along the

southwestern beach of the Island. This site was chosen to replicate to coastal conditions of

Donnelly and other's research in hurricane storm deposition New England salt marshes (2001).

It also represents a more exposed marsh environment with the larger Apalachicola Bay fetch;

also West Pass is near-by creating a connection with the Gulf of Mexico.

Four long cores (- 1 m) were collected from Marsh A in two two-core shore-normal

transects. Core one of each transect was in the sub-tidal region of the marsh and core two in the

supra-tidal region of the marsh (about 5 meters apart). Two short cores were also collected.

Short cores were about 10 centimeters (cm) long and collected near core two of transect two. In

Marsh B four two-core shore-normal transects were collected. Transects were started 20 meters

(m) from high tide (determined by wrack line) and the second core was taken 30 m from high

tide. Again, two short cores were collected near core two (C2) of transect two (T2) (Fig. 1-3 B).

Long cores were collected using a modified piston corer and tripod. The coring device

consists of a four-foot long, 7.5 cm ID aluminum barrel with a sharpened stainless steal cutting

head and an aluminum driving head with handles. The core barrel has a 7.5 cm OD CAB liner.

The CAB liner was gently pulled out of the barrel and capped resulting in an undisturbed core.

Each short core was collected by pushing a 45 cm piece of CAB into the ground and digging it









out. Core collection took two days (a day at each site). After initial measurements of short-lived

radioisotopes 234Th and Be of a short core from each site, it was decided that the Marsh A cores

would not be examined further. The cores from Marsh A did not contain any excess

radioisotopic activity (234Th, 7Be, 210Pb, 137Cs), most likely because of the very high sand

content. Consequently, chronologies could not be established for Marsh A cores.

Core Logging

The long cores all underwent non-destructive analysis on the Geotek multi-sensor core

logger (MSCL) to determine bulk density and magnetic susceptibility. Prior to running cores, a

calibration standard was created and measured to determine the calibration constants needed to

process data collected by the MSCL.

The core logger measures bulk density by gamma attenuation at a rate of 0.5 cm every 10

seconds. The MSCL uses to Eq. 3-1 to determine bulk density by gamma attenuation; gamma

attenuation is measured by the number of gamma photons that pass through the width of the core

at each sampling point and the count time in that sampling interval.

p = 1 / [td*ln(Io/I) (Equation 3-1)
where p = sediment bulk density, t = the Compton attenuation coefficient, d = the
sediment thickness, Io = the gamma source intensity, and I = the measured intensity
through the sample

Gamma bulk density was corrected for compaction in cores T2C1 and T3C2. This was

done by taking the measured density and subtracting out a model fit to the data. The model was

an exponential decay with Eq. 3-2.

y = yo + A ( 1- exp-bx) (Equation 3-2)
where yo, A, and b were changed to fit the measured data and x was the depth

Magnetic susceptibility was determined by a loop created field and how the field is

affected by the core for 10 s per one-half cm; the data were integrated over 10 cm of core length.

Magnetic Susceptibility was measured by the Bartington MS2 meter; sample time was measured









in hertz (HZ) where 1 s = 1 Hz and 10 s = 0.1 Hz. Magnetic susceptibility was processed and

corrected for both mass and volume.

Volume specific magnetic susceptibility (K) is dimensionless and was calculated by the

MSCL software using Eq. 3-3, where Kuncorrected is the raw data collected by the MSCL and

Krelative is determined by Eq. 3-4, which is a correction factor for the loop size. Because the

MSCL collects data in cgs units Eq. 3-5 is used to convert to SI units.

K = Kuncorrected/ Krelative (Equation 3-3)

Krelative = 4.8566(d/D1)2 3.0163(d/D1) 0.6448 (Equation 3-4)
where d = core diameter and D = loop diameter

K (SI units) = 47t*10-6K (cgs units) (Equation 3-5)

Mass specific magnetic susceptibility was calculated using Eq. 3-6. Equation 3-7 was used

to convert from cgs to SI units.

x= K/p (Equation 3-6)
where, x= mass specific magnetic susceptibility, K = magnetic susceptibility corrected for
loop size, and p = sediment density (kg m-3)

X (SI units) = 47t*10-3 X (cgs units) (Equation 3-7)

Lithology

Cores were split lengthwise into one-third and two-third thick lengths. The thinner or one-

third thick lengths were all run through the MSCL for digital imaging and then x-rayed for the

detection of sedimentary structures and for bulk density. Images from the MSCL were processed

to determine the red, green, and blue absorption and respective ratios to aid in the determination

of lithologic changes.

The images and digitally scanned x-rays were spliced together in Adobe Photoshop

creating a continuous image of each core; making it easier to understand the lithology and

observe other visual patterns and sedimentary structures within the cores, as well as the amount









of bioturbation and the general sedimentology of each core (Figs. 3-1 A and B). Down-core

pixel intensity (300 dpi) for select core x-radiographs was determined along a center transect of

the core. The data were then smoothed by a 35 point running mean or 40 point running mean

averaging the data over fractions of a centimeter for cores T3C2 and T2C1, respectively.

Smear slides were created at the various color changes within core T3C2 to examine the

composition and look for diatoms to determine if the marsh has been fresh or salt water

dominated, or had experienced a salinity change, and/or if it was ever an environment other than

marsh. Diatoms were identified by Dr. Paul Ciesilski at the University of Florida.

Gamma Spectroscopy

The two short cores collected at each marsh were sub-sampled on the day of their

collection into one centimeter (cm) intervals. Upon return to the lab these intervals were oven

dried and powdered for gamma spectroscopy, to measure short-lived isotopes Thorium-234

(34Th) (half life = 22.3 days) and Berylium-7 (7Be) (half life = 53 days) as a method of

determining mixing rates or mixing coefficients and the sedimentation/accumulation rates of

each marsh.

Cores T2C1 transectt 2, core 1; proximal) and T3C2 transectt 3. core 2; distal) were

chosen for higher resolution because they appeared, from digital imaging, x-radiography and

measurements of 210Pb, 137Cs, and 226Ra, to have the least amount of disturbance from

bioturbation. Samples were collected at 2-cm intervals, homogenized, freeze dried, powdered,

and packed for gamma spectroscopy. Once these initial samples were counted it was decided

that additional samples from the high resolution cores (T2C1 and T3C2) would be beneficial for

better solving chronologies. Additional samples were prepared and counted. The counting

efficiency of the germanium detector used for this study was established using NIST-calibrated

sediment standards. Precision of the apparatus was not determined because there was not









enough sample mass to run samples in triplicate or duplicate; also due to time constraints the

time was not taken to run samples multiple times.

The constant initial concentration (CIC) model (Benninger et al., 1979) was used to

establish chronologies for each core using the excess 210Pb data and with some assumptions

about the sedimentation rates core chronologies were determined. The constant rate of supply

(CRS) model (Appleby & Oldfield, 1992) was attempted for chronology; however, this model

did not perform well because of the variable activity due to grain size effects. Peak activity

(1963) and first-appearance (early 1950s) of 137Cs were used as a check for, and to back up the

210Pb data from the CIC model. All radioisotope data were plotted against cumulative mass

rather than depth to remove the effect of autocompaction, which is common to salt marsh

sediments.

One gram was removed from each of the samples used for the radiometric analysis for loss

on ignition (LOI) to determine the organic matter concentration. LOI was completed by heating

each sample to 5500C for approximately 4 hours and reweighing the sample to determine mass

changes. The remainder of each sample used for radioisotopic analyses was weighed, wet sieved

(to separate the sand (>63[t) from the silt and clay (<63[t)), and dried to determine percent mass

of each grain-size. Both these analyses were completed to determine whether organic content

and grain size are related, or if one is more prevalent and which should be used to normalize

radioisotope data.

Grain Size

Long cores (thick half) T2C1 and T3C2 were sub-sectioned at one centimeter intervals.

These cores were chosen because of the high resolution radioisotope work, which was done due

to the low degree of visual bioturbation. After homogenizing the sub-samples, approximately 10

g of each sample was treated with 5 ml of 30% molar hydrogen peroxide to oxidize/digest the









organic matter. Samples were then wet sieved with deionized water at 1mm to remove any large

pieces of organic matter; organic matter was discarded and any sand that remained in the 1 mm

sieve was oven dried and stored. The remaining sample was wet sieved with deionized water at

63[ to separate sand, silt, and clay. The silt and clay fraction was allowed to settle out, then

transferred to whirl-pak plastic bags and put into cold room storage for possible further analysis.

The sand fraction was treated again with 5 ml of 30% hydrogen peroxide to remove any

remaining organic; some intervals needed to be treated twice to complete organic removal.

Once all organic material was removed the sand was oven dried. Sand samples were then run on

a settling column to determine sorting, mean grain-size, and modal grain-size (Boggs, 2001).

Many sand samples were inadequate in mass to be run on the settling column.

Event Determination

Events were identified by comparing grain-size, gamma bulk density, and x-ray pixel

intensity data. Distinct peaks in all three variables for T2C 1 and T3C2 were the criterion for

identifying to be event beds. The events were then given a preservation quotient which was

determined from the x-radiographs and based on the amount of apparent mixing (Fig 3-1 A and

B). This quotient was a percent range (0-20, 20-40, 40-60, 60-80, and 80-100) where 100% has

experienced no mixing and 0% is completely mixed (Bromley, 1996). This was done for the

three event beds that could be found in both high resolution cores. These three events were then

modeled using the Bentley Bioturbation Model.

Modeling

Conceptual modeling compiled the results of gamma-ray attenuation bulk density, grain

size analysis, and x-radiography pixel intensity to fulfill all necessary variables except

dissipation time.









Numerical Modeling was done using the Bentley Bioturbation Model (see Models chapter)

in MatLab Student 6.5 Version 13. Model variables and parameters (biodiffusion coefficient,

mixing depth, and sedimentation rate) were collected and determined from radioisotope and x-

radiograph data from cores T2C1 and T3C2. Original event bed thickness was the only unknown

variable. Each event bed was modeled separately rather than modeling a whole core at once.

Simulations were repeated changing only the event bed thickness, time of event, and duration of

simulation until the model output patterned the preservation quotient of each event.









T1C1 T2C1 T3C1 T4C1




i I I F









IW
,"













., Unit A
Unit B
Unit C





A

Figure 3-1. Lithology of Cores. A) Digital imagery and x-radiography, for visual examination,
of salt marsh proximal to the beach, collected 20 meters from high tide (determined
from wrack line) The sediments are comprised of organic-rich sandy muds.









T1C2 T2C2 T3C2 T4C2



I


Figure 3-1. Continued B) Core itoog Digita imagery and -radiogra for visua

IIm .
E
a. II



dI I
SI I
I. I





'],
SI.I

















Unit A
Unit E
Unit C

Figure 3-1. Continued B) Core Lithology Digital imagery and x-radiography, for visual
examination, of salt marsh distal to the beach were collected 30 meters form high tide
(determined from wrack line). The sediments are comprised of organic-rich sandy
muds.









CHAPTER 4
RESULTS

Lithology

Marsh B strata that are closest to the beach (Fig. 3-1 A) are best represented by a core

collected along Transect Two (core T2C1). Two distinct lithofacies are observed in this core.

There is a sharp contact at about 36 cm depth representing the division of the two units, Unit A

and Unit B.

Unit A (0 36 cm) is a mottled tan grey/green sandy mud and contains two beds. The

top bed of the unit is two-cm-thick, organic-dominated with abundant woody plant material.

From 2 to 7 cm is a bed of sand with a diffuse bottom boundary that is greener in color. The

interval between 12 and 20 cm contains abundant roots. The top 20 cm appear to show a higher

degree of bioturbation containing many burrows, while 20 36 cm contains only two large

burrows.

The second lithofacies, Unit B (36 78.5 cm), is a sandy mud. The bed from 36 50 cm

is mottled brown and very dark grey with plant roots and small burrows. From 50 to 78.5 cm the

bed is a dark grey grading into a light to very light grey. The unit is faintly to well laminated;

most contacts between beds and laminations are sharp.

In the other three beach-proximal cores (T1C1, T3C1, T4C1), similar lithological units are

observed, although the thickness of each varies between cores. In Core T1C1, Unit A has a sharp

contact at about 15 cm depth. The top 15 cm contains abundant woody plant material and

burrows. The contact between Unit A and Unit B is at approximately 27 cm. The top of Unit B

contains less bioturbation than in T2C1.









In Core T3C1, the contact between Unit A and Unit B occurs at about 37 cm; however, it

is not as sharp a contact as observed in T2C1. At 23 cm depth (Unit A) the amount of apparent

bioturbation and plant material decreases significantly.

Core T4C 1 is vastly different from the other cores proximal to the beach. This core does

not appear to have Unit A or Unit B. A sandy shell hash constitutes the upper 16 cm of the core

where grey-brown muddy sand begins to mix with the shell hash. The amount of sandy shell

hash decreases with depth to about 40 cm where the grey-brown muddy sand begins to dominate

the remainder of the core.

Marsh strata that are farthest from the beach (Fig. 3-1 B) are best represented by a core

collected along Transect Three (core T3C2), which has three observable lithofacies; Unit A, Unit

B, and Unit C. The contact between the two upper units is sharp at 36cm depth. Unit A (0 -

36 cm) is a muddy sand with visible burrows from 0 20 cm. The top six centimeters are

mottled dark grey green and contain more organic material including roots and woody plant

fragments. The interval from 6 22 cm is mottled greenish-tan to brown; 22 36 cm is a mottled

greenish-grey. There is a sandier bed with a sharp bottom contact and a diffused top contact at

approximately 26 cm depth. The interval from 30 32 cm is a muddier bed with diffuse

boundaries and more brown coloring.

The second lithofacies, Unit B (36 -87 cm), is a gradational black to dark grey to light

grey. There are fairly well-preserved, faintly to well laminated sand beds, and some are very

pronounced. Burrows are preserved in sandy mud from 36 42 cm, while fine plant material

(roots?) are found throughout the unit. At 42 cm is a fairly sharp contact that marks a change to

muddy sand. At 58 cm there is a bed consisting of a series of whitish sandy lenses and pods.

Below a diffuse contact at about 69 cm, the texture returns to sandy mud.









The third lithofacies, Unit C (87 -90 cm), is a brown to very dark brown mottled muddy

sand with plant material. There is no apparent bedding.

In the other three beach-distal cores (T1C2, T2C2, T4C2), similar lithological units are

observed, although again the thickness of each varies between cores. In T1C2, the contact

between Unit A and Unit B is much more diffuse than in T3C2 and is determined to be at

approximately 35 cm based on the x-radiograph. There is abundant woody plant material to 15

cm depth, which is also the depth where biogenic traces decrease dramatically. A whitish sandy

lenticular bed in Unit B is at 51 cm. Unit C starts at approximately 80 cm and as in T3C2 it

continues to the bottom of the core.

In core T2C2, the Unit A/Unit B contact is at about 35 cm, but is not sharp. There is a

large burrow feature from the top of the core to 43 cm. Smaller burrows predominantly occur to

a depth of 19 cm. Woody plant material appears to disappear around 15 cm while finer plant

material (roots ?) continues to about 22 cm. The whitish sand bed of Unit B is around 50 cm but

does not have the textural contrast seen in other cores. Unit C begins at about 80 cm and has a

much more varied lithology than the other distal cores.

The contact between Unit A and Unit B in core T4C2 has been slightly mixed, but can be

seen between 40 and 45 cm. Plant material is observed throughout Unit A but the woodier

material is found only to a depth of about 30 cm. Evidence of bioturbation is apparent through

the entirety of Unit A but dramatically decreases at about 30 cm. The white lenticular bed of

Unit B occurs between 53 and 62 cm. Unit C can barely be seen in the bottom centimeter of the

core

Grain Size Analysis

Given the large similarities in lithology among cores, detailed textural analyses were only

performed on the two main cores from the proximal/distal transects (T2C1 and T3C2). The mass









percent sand ranges between 0 and about 51 %, indicating that the marsh strata are silt and clay

dominated (i.e., sandy mud) (Fig. 4-1). As percent organic matter increases percent sand

decreases (Fig. 4-2).

Because depositional features, such as bedding, seen in the x-radiographs indicate transport

and deposition, the textural characteristics of the sand fraction can provide information on the

mode and strength of transport (i.e., wind, overwash) and possibly the source (i.e., beach,

terrestrial sand ridges). However, due to amount (> 1 gram) of sand needed to for textural

analysis sections of the cores were not analyzed (Fig. 4-3). The median, mean, and modal size of

the sand-fraction only range from 1.5 to 2.5 phi (fine to medium sand). Sorting ranges between

0.5 and 1.0 phi, averaging 0.75 (moderately to moderately well sorted). Skewness ranges from -

0.01 to about 0.5 phi with an average of 0.25 and 0.4 phi for T2C1 and T3C2 respectively. The

data are finely (positively) skewed (Fig. 4-3 A). Textural analysis has no significance difference

between each interval (Fig. 4-3 B).

Gamma Bulk Density

Gamma-ray attenuation (GRA) bulk density showed an increase down core in all cores.

There appears to be no significant difference between the two sets of data (900 rotation of core)

collected for each core (Fig. 4-4). The GRA density range is from 0.8 (indicating partial

dewatering of the core) to 1.8 g cm-3. The variation with depth appears to track lithology and

bedding, so to better isolate this variation, the effect of auto compaction was removed by fitting

an exponential model to the GRA density data. Subtracting the difference between the model

and the data resulted in the density residuals (Fig. 4-5).

Magnetic Susceptibility

The range of magnetic susceptibility is 0.0 36 *10 SI units; most data are between 0.0

and 5.0 *10 SI units. As with density, there is no significant difference between the two sets of









data (900 rotation) collected for each core. There also appears to be a correlation of the peaks

between the cores. However, there is not a strong correlation between the peaks and bedding

observed in the x-radiographs. As collected by the Geotek MSCL, magnetic susceptibility

values are smoothed over about a 5 to 8 cm window making it a much lower resolution proxy

than bulk density, core imagery, and x-radiography. Consequently, it was not used for more

detailed modeling of bed preservation.

X-radiograph Pixel Intensity

X-radiograph positive images reflect variations in sediment density at sub-mm resolution,

making them ideal for examining subtle variations in texture. Due to this high resolution, the

resultant digitized pixel intensity data were extremely variable and difficult to interpret. So, the

down-core pixel intensity data were smoothed using a running mean. A 40 and 35 pixel running

mean (-2mm @ 300 dpi resolution) was used for T2C1 and T3C2 respectively. Relatively low

pixel intensities represent increases in sand (i.e., sand beds) (Fig. 4-6). The notable low

intensities range in thickness from a fraction of a centimeter to up to 5 cm. Down-core, the low

intensities appear to merge together creating thicker noisier sections of sand increases (Fig. 4-6).

Chronology

Lead-210

Because sediment texture can play a controlling role on particle-reactive radioisotope

activities, the activity data are plotted as both dpm disintegrationss per minute) per gram total

sediment (Fig. 4-7 A) and per gram mud mass (Fig. 4-7 B) only. When corrected for mud

content, there is a large difference between the corrected and uncorrected activity for T2C 1, yet

little difference between the mud corrected and uncorrected data for T3C2. The data used for

chronological measurements were the mud-corrected activities. Because the 226Ra data used to

calculate supported 210Pb activity were highly variable, it was assumed that the total 210Pb









activity at the bottom of the cores represented the average 226Ra supported 210Pb activity. These

activity data subtracted from the total 210Pb data to determine the amount of unsupported, or

excess, 210Pb.

The first appearance of excess activity is at about 35 and 37 g cm-2 cumulative mass

(approximately 23 and 27cm depth) in T2C1 and T3C2 respectively (Fig. 4-7 B). Because

excess 210Pb activity represents 4-5 half-lives of accumulation (-100-110 years), the resulting

mean sedimentation rate based on this first appearance is around 0.2 cm yr-1. Using this

sedimentation rate the surface mixed layer is determined to be no thicker than 10 cm depth, as

this is the apparent depth of mixing in the 210Pb data (Fig.4-7 B). With depth, activity of 210Pb

appears to be increasing until about 10 cm when activity begins to decrease for the remainder of

the core. It is possible however, the depth of mixing be much shallower than the 10 cm depth

deduced by the higher 210Pb activity. If the surface mixed layer is shallower than 10 cm there is

either a change in the 210Pb source or a sedimentation rate change resulting in a misleading

mixing depth profile.

Cesium-137

As in the case of 210Pb activities, 137Cs activities are susceptible to a grain size influence,

and were also corrected to reflect the mass of mud only. Yet, correcting for mud does not

significantly affect the data (Fig. 4-8). Because mud-corrected activity data were used for 210Pb,

they were also used for 137Cs. First appearance of 137Cs occurs at 25 in core T3C2 and at 35 g

cm-2 depth in core T2C1. There also is a 137Cs peak at 17 in core T3C2 and at 35 g cm-2 in core

T2C1. 137Cs was introduced into the atmosphere in the mid 1950s and was released in the largest

quantities in the early-mid 1960s. The depth of the first appearance and peak 137Cs activity was

divided by the years elapsed between these dates and core collection (48 and 38 years

respectively) to determine the 137Cs-based sedimentation rate for the marsh. That sedimentation











rate is about 0.5 cm yr' for the beach-proximal core and -0.2-0.3 cm yr' for the beach-distal


core. The potential downward transport (diffusion or advection) of cesium was not accounted


for in these calculations and could be substantial given the coarseness and higher potential


permeability of these cores. Downward transport would lead to higher calculated sedimentation


rates, which may explain the discrepancy between sediment accumulation rates based on the two


radioisotopes.

Percent Sand (%)
T201 T132
0 '0 20 30 43 6 0 10 20 30 40 W
0 -vEvent


E vent I









Ev ent I


0 -



50 -
EEvvent II

5 Event III




70

7r


8-"





Figure 4-1. Percent Sand of Cores T2C1 and T3C2. As a textural analyses on cores T2C1 and
T3C2; mass percent sand with depth in core as determined by wet sieving. Events I,
II, and III visual preserved thicknesses are shaded in blue. These are the same event
beds modeled using the Bentley Bioturbation Model.










55

50 T2C1

45

40


*
S30

25



15

10

5


0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80
55

o T3C2

45

40






b r=0.11
20 *
0.
15.

10
5

0
0 5 10 15 20 25 30 35 40 45 50 55 60 85 70 75 80
Percent Sand

Figure 4-2. Percent Sand and Percent Organic Matter for Cores T2C 1 and T3C2. A lithological
examination of salt marsh cores using percent sand and percent organic matter. Loss
on ignition (LOI) was used to determine the percent organic matter while wet sieving
was used to determine percent sand. Data were collected from sediments used in
gamma spectroscopy as they were already prepared for LOI.










T2C1


-0.6

5
10
15
20
25
30
35
40
45
50
55
00


B5
70
75
50
a5
I0
95

-0.5

5
10
15
20
25
30
35
40
45
50
55
05
70
75
80
an

95


0.0 0.6

l l I


'4'


Inadequate Maese of Sland
to Prfbrm An"lyml


cL.,t!i
-'-C- -i


Fr


rD




fw


T3C2
0.0 0.5 1.0 1.5 2.0 2.5 3
" I I l . l .


Mmdlan
Mom
Mdin
Mode
Sorting
Sklwnar


Inadequate MRse of Bhnd
to Prtbrin Anmihil


-". *'...''
..



r ,
0' ,," S....
I -,


Figure 4-3. Textural Analysis of Cores T2C1 and T3C2. A) Each data point is representative of
the sand at that particular depth interval; statistics with depth in core. Statistics were
determined from data collected using a settling column.


--- S ^ ^
-- ----












T3C2


-2 -1 0 1 2 3 4 -2 -1 0 1 2 3


-*- Evet BedSands
- UitB Snds
- UitA Sads










icrnDeplh
5an Deplh







/A










4
B


Figure 4-3. Continued. B) Incremental percent of grain size (Phi) for each depth interval
containing enough mass sand to analyze using a settling column (top). Cumulative
percent of grain size (Phi) for each depth interval containing enough mass sand to
analyze using a settling column (bottom). Data in red represents depth intervals that
may be event beds; blue represents sand from Unit A; green represents sand from
Unit B.












Gamma Bulk Density (g cm-3)
T2C1 T3C1 T4C1


0.8 1.2 1.6 2.0 0.8 1.2 1.6 2.0





25


25

30

35




50 V
4g


55

00

05
70

75
60

85

90


0.8 1.2 1.6 2.0 0.8 1.2 1.6 2.0






























Gamma Buk Denslt
Gamma Buk Denatly at O0 Rotation
A


Figure 4-4. Gamma Bulk Density of Cores. Lithological examination of salt marsh cores using
gamma bulk density; data are from use of the GeoTek Multi-Sensor Core Logger. A)
The cores proximal to the beach, collected 20 meters from high tide (determined from
wrack line). The sediments are comprised of organic-rich sandy muds.


T1C1














Gamma Bulk Density (g cm-3)
T2C2 T3C2


-- GOmm Buk Danlty
Gamma Bulk Denilty aI 0 Rotlilon

T4C2


4.0
a4









Is
Is
3


Is











M
75
U

aa








IS:
30

a


41

'a







H:



10:
U

Uo







'a


1.2 1.8 2.0 0.A 1.2 1S 2.0 0d 12 1.8 2.0 0. 12 1.8 2.0


Figure 4-4. Continued. B) The cores distal to the beach, collected 30 meters from high tide
(determined from wrack line). The sediments are comprised of organic-rich sandy
muds.


T1C2










Rualdual Bulk Denlity (g cmr3)


T2C1


T3C2


0.8 1.0 1.2 IA 1.8 1.9 2.0 0. 0.8 1.0 1.2 I.A B 1.A 2.0


- Rsluab af Buk Dandly
- uk- ekDandly
Bk Dandly at G00 Rolatan


Figure 4-5. Residual Bulk Density of cores T2C1 and T3C2. A lithological examination of salt
marsh cores using residuals of gamma bulk density; gamma bulk density data is from
use of the GeoTek Multi-Sensor Core Logger. Auto compaction was removed by
fitting an exponential model to the density data and removing the difference between
the two. This resulted in the density residuals used to better isolate variations.


45
Cl




70






75
75









Pixel Intensity


250 200 1w 100
. .. | . m .m


250 200 160 100


-% i
N.


P4




.'




Is

^.


Doa smouhd by 40 point ruing moin.


Dae smoolhed by
35 pdnt mnrikg meaL


Figure 4-6. Pixel Intensity of Cores T2C1 and T3C2. A lithological examination of salt marsh
cores using pixel intensity of x-radiographs. Decreases in intensity represent increases
in sand attributed to sand beds and possibly event beds. Smoothing data, using a
running mean, averaged the data over fractions of centimeters.


T2C1


T3C2











Taotl 210PbAt


0 2 4 8 U 10 12 14
. . . . . .


0 2 4 8 I 1I 12 14


I




11




14




s,




I3,






4 ,
81








U
81


* dam Mud
da le


Figure 4-7. Total Activity of Lead-210 and Total Activity of Lead -210 Corrected for Grain
Size. A) Data were determined by gamma spectroscopy. Particle-reactive
radioisotopes, such as lead-210, adhere more readily to silts and clays than larger
grain sizes. By correcting for this a more accurate activity is determined. Activity
data are plotted against cumulative mass to correct for auto compaction.









10Pb Actvty (dpm (1l mud)


T2C1
o0 4 B 1 10 12 14
........ m..... ..... ..... ..... ..... n..


0.


a.



10
IS


20



38

i"

I33
U-


T3C2
0 2 4 8I 10 12 14
S......... ....n....n....n.... ..... .. .


M
- -18mm0A
,,Mr III )


G Adr*


Figure 4-7. Continued. B) Data were determined by gamma spectroscopy and corrected for
grain-size. Excess activity was determined by subtracting the average total activity
below the assumed first appearance. First appearance is assumed to be -1900 AD.
Activity data are plotted against cumulative mass to correct for auto compaction.


N
E---I=Ao
N


4"1







an
-I



'











Total m137 um Acvity


O I 1.0







I





i i
UJ






I,-
NI






"I








1j ,



:1


TMC
1I
----


0D M 1 ,
--I--.------ I.


--I


-4-I
-1
Mel


T3C2
0IO 1.0 1.
......- .... ...


1 2 UD
.|... i......


I I

i'1


i'4I


i'i
i'1


* aiMg
4 mullud
N6MrdMOML"


Figure 4-8. Total Corrected Activity of Cesium-137 and Cesium-137 Activity Corrected for
Grain Size. Data were determined by gamma spectroscopy. The mean detection limit
of the germanium crystal used in gamma spectroscopy is about 0.1 0.2 dpm g1.
Activity data are plotted against cumulative mass to correct for auto compaction.


L - - - - - - - j









CHAPTER 5
DISCUSSION

Lithology and Evidence of Storm Bedding

X-radiographs and visual observations of the cores indicate that they are moderately

bioturbated. The catchment basin for the marsh was not large enough to collect measurable

amounts of excess 234Th (derived from seawater) or Be (derived from atmospheric fallout) to

determine biological mixing rates from these short-lived radioisotopes. Mixing is attributed

mainly to microfauna and roots. Had there been a significant amount of macrofaunal mixing

(i.e., fiddler crabs), the core logging likely would have shown differences between the 900

rotations of the cores. Because the data remained the same within error for the two orientations,

I assume that there is no significant macrofaunal burrows or cavities within the cores. In

addition, x-radiography did not expose any larger cavities.

The 0.2 -0.5 cm y-1 burial rates determined by 210Pb and 137Cs appear to be roughly equal

to the rate at which accommodation space is created by the rise in relative sea level for this area

of the Gulf of Mexico, which is about 0.3 cm y-1 (Thieler and Hammar-Klose, 1999). Most salt

marshes accumulate sediment at a rate equal to the relative rate of sea-level rise (Allen, 2000).

The lithofacies of the cores are typical of salt marshes from the southeastern U.S., having

less organic matter and more detrital sediment than what is seen in colder climates (Frey and

Bassan, 1985). The overall decreasing trend of organic matter up core may be due to the

increase of erosion and encroachment of the beach on the south side of the marsh. The higher

organic content in T3C2 may be a direct result of its distance from the beach. In essence, the

organic input into all the cores may be the same but the cores proximal to the beach have more

plastic input diluting the organic matter.









Differences in sand content between beach-proximal and beach-distal cores can also be

attributed their locations with respect to the beach (Fig. 1-2), which is assumed to be the

dominant source of plastic sediment. There is abundant vegetation separating the marsh from the

tidal creeks to the west, so the mechanism to best explain the relatively high mean mass percent

sand (-30%) throughout the marsh (located <5 m from south shore) is either aeolian or

overwash. Due to the apparent lack of a continuous input of sand into Marsh B overwash is the

more likely mechanism. However, during a storm, increased wind speeds could deposit an event

bed with aeolian mechanism. Overwash is a more likely process controlling sand deposition at

these coring sites, given that there is no evidence of bedload sand transport in x-radiographs

(e.g., sharp erosional contacts, cross-lamination, graded bedding). Initial event bed thickness is a

function of the volume of sediment available for transport, shoreline vegetation, dune

morphology, distance from the shore, and storm surge.

Statistically (cumulative and incremental percent, mean, mode, sorting), the source of sand

is the same throughout both cores (Fig. 4-3). The mean and modal sizes of the sand fraction are

the same as the mean and modal grain size of the south beach (Hart, 2003). The shell hash at the

top of T4C 1, however, may be an overwash fan from a storm that occurred after 1954 as

indicated by 137Cs activity found beneath the shell hash. The lack of vegetation on top of this

shell hash suggests a more recent event such as the 1985 Hurricane Kate whose eyewall passed

directly over the island, 1995 category five Hurricane Opal, or 1998 Hurricane Earl with the 2.5

meter storm surge.

Based on diatoms identified in the smear slides (P. Ciesielski, personal communication), it

was determined that this marsh was once freshwater dominant (Unit B) and became saltwater

dominant (Unit A). It is possible that at one time this was a coastal pond location that has in









filled; depending on island and coastal morphology. The overall down-core increase in sand

starting at 44 cm depth (Fig. 4-1) may be due to the marsh being freshwater dominant. At the

transition between Units A and B found in all cores, 35 to 40 cm depth, it is possible that a storm

event between 116 and 88 years ago (between 1886 and 1915) caused a change in

geomorphology allowing daily tidal inundations to extend inland as they do today. Six separate

storms occurred in this time frame but the 1886 was the only major hurricane.

The storms discussed in the Study Area section of Chapter 2 provided the initial

chronology for possible event bedding. Using the determined sedimentation rate and storm

history approximate Events were determined within the cores. These Events were confirmed by

comparing and correlating the three main proxies (percent sand, gamma bulk density residuals,

and pixel intensity) as shown by Figure 5-1. It is possible that the correlating peaks of these

proxies are event beds or temporary changes in sedimentation. However, it is less likely these

peaks represent sedimentation changes due to the short duration of deposition, lack of sharp

contacts, and lack of significant change in aerial photographs. It is more difficult to find and

correlate peaks in the distal cores and the correlated peaks in the proximal cores are shallower

than those in the distal cores.

Modeling Preservation Potential of Storm Beds

In order for an event layer to be preserved, transit time of the storm layer through the

surface mixed layer must be faster than the dissipation time due to mixing. From Eq. 2-1, the

transit time can be calculated for the marsh study areas given estimates of the mixed depth (Lb),

storm bed thickness (Ls), and burial rates (i.e., sedimentation rate) (Fig. 1-3). For this study, the

thickness of the surface mixed layer was calculated to be approximately 3.0 cm based on 210Pb

profiles and x-radiographs, as shorter-lived tracers (e.g., 234Th) were not effective. If surface

sediments have been rapidly mixed, the 210Pb profile may have a near-surface interval of









constant activity that corresponds in x-radiographs with abundant biogenic traces (Figs. 3-1, 3-2,

4-4, and 4-7 B) (Benninger et al., 1979). For the marshes, Ls was estimated to be 3-50 cm

matched to depths of known hurricanes (Event I 1974/1985; Event II 1886; Event III pre-

historic) established partially with 210Pb/137Cs geochronology. The complicating fact is that the

thickness of a storm layer would need to be measured directly following a hurricane for an

accurate estimate of Ls, as what is measured in the cores is the minimum preserved bedding

thickness as some unknown quantity at the top of the bed has been mixed (Fig. 3-1, 3-2, 4-4, and

4-7 B).

Using the above data and a sedimentation rate of -0.2-0.5 cm yields a transit time of 0-30

years (Lb > -1 cm) for the salt marshes. The fast transit time suggests that an event layer should

be preserved, but does not account for dissipation of the event layer due to physical and biologic

mixing. Since time series cores are not available for this study, the dissipation time for the cores

can be estimated based on probable biodiffusion coefficients (Db) can be calculated from (a)

210Pb activity data (Aller et al., 1980) or (b) data from previous studies. Maximum mixing

coefficients are -2 cm2 y-1 from the 210Pb activity data from these marsh cores, but these rates

would represent much slower mixing averaged over several decades (Nittrouer et al., 1984). For

salt marshes, short-lived tracers ideally suited for calculating Db (Aller et al., 1980) were not

detected in this study, but x-radiography reveals substantial mixing and burrowing. Thus Db is

assumed to be 10 cm2 y-1 based on estimates from other salt marshes (Boudreau, 1994).

Wheatcroft and Drake (2003) report Db values ranging from 10 to 100 cm2 y-1 for continental

margin sediments where sediments are more biologically active and correspond to dissipation

times of 3-5 years. Thus, it is estimated that the dissipation time for coastal marshes is -5 years.

Given transit times of 0-10 years and a SML of 10 cm for this salt marsh, it is clear that very









large (> 5cm thick) beds will be preserved, but that beds < 1 cm thick will be mixed and not

preserved.

Original event bed thickness was then estimated by numerically modeling the event-beds

found using the Bentley Bioturbation Model (BBM). Estimation of original event bed thickness

was used to validate the conceptual model. This same numerical model was used to quantify

preservation potential of the salt marsh sediments (Fig. 5-2). Although created to model event

beds created on continental shelves the BBM was accurately and successfully used in this salt

marsh as I was able to reproduce the event beds found in the Marsh B cores.

The values for parameters in the BBM were established using the results form 210Pb/137Cs

chronology, x-radiographs, core logging, and estimated from literature when necessary. The

BBM parameters used for this specific salt marsh, and all events modeled, were defined as in

Table 5-1.

Table 5-1. Constant BBM parameters determined by cores collected in Marsh B on St. Vincent
Island, Fl.
BBM Constant Parameter Value
o, and co (mm2 y-1)
(sedimentation rate prior to and after event, respectively)
Lb (cm) 3
(depth of mixing or surface mixed layer thickness)

to (cm2 y1) 6.733
(biodiffusion coefficient)

1 2.3026
(controlling constant of ao)
Q (cm yr-) 0
(supplemental rate of accumulation)

The value for co was also used for oo because it was assumed the sedimentation rate did not

change prior to or following each event. Zero was used for the supplemental rate of

accumulation (Q) to portray an instantaneous deposition as would be expected with an event bed









due to a tropical cyclone. The parameters that varied from event to event were defined as in

Table 5-2.

Table 5-2. Variable by event BBM parameters determined by cores collected in Marsh B on St.
Vincent Island, Fl.
BBM At (years) Ls (cm)
Variable (length of time t (year) (event bed original
Parameter model is run) (time event occurs) thickness)
T2C1
Event I 14.17 10 4.5
Event II 4.17 2.75 2.25
Event III 12.5 2.08 6.5
T3C2
Event I 0.417 0.25 2
Event II 8.33 4.17 2.25
Event III 16.67 2.5 2
10.417 1.5

The values in Table 5-2 were determined by trial and error. The model was run multiple

times until the output (depth of sedimentation and percent of event preserved) best fit the

preservation as determined from the x-radiographs (Fig. 5-2). The amount of time, At, used for a

particular run of the model was actually input as months because that is the unit of each time-

step. The year or time-step the event is deposited in is represented by t and is also entered into

the model as a month. The initial thickness of each event was thicker in the proximal core

compared to the thickness in the distal core. This was expected as an overwash fan thins with its

extent. However, Event II modeling resulted in the same original thickness in both cores T2C 1

and T3C2. This may be attributed to the difference in the number of time-steps used for each or

that in actuality they do not represent the same event. Event III in core T3C2 wound up being

two separate events that appear to be one when looking at the proxy data (Fig. 5-2 B).

Modeling these events produced results that are at odds with the theory of event

preservation in that there are event beds preserved which originated with thicknesses less than

the depth of mixing. According to the theory, these events should not be recognizable within the










sedimentary record. One explanation for this discrepancy is that as a coastal environment the

salt marsh is very dynamic and the assumed steady state parameter values used are inaccurate for

the entirety of the marsh and even each core. It also may be attributed to the lack of macrofaunal

mixing.

Physical Propertlis T2C1

Percent Sad Resduls of Gamma Buk Demnty (g cm") Pkl Irtenslly
0 10 a SO 40 a o0. o0 1. 1.2 IA 1 00 210 0 a 100 m

50








5 Event I
20


Event III

55





so







Figure 5-1. Physical Properties of Cores T2C1 and T3C2. Lithological examination of salt
marsh cores using textural analysis (percent sand), residuals of gamma bulk density,
and pixel intensity of x-radiographs. These are the three main proxies used to identify
70 S












event beds. Event beds I, II, and III (used in Bentley Bioturbation Model) are shaded
in blue. A) The cores proximal to the beach, collected 20 meters from high tide
muds.
90



95 A

Figure 5-1. Physical Properties of Cores T2C1 and T3C2. Lithological examination of salt
marsh cores using textural analysis (percent sand), residuals of gamma bulk density,
and pixel intensity of x-radiographs. These are the three main proxies used to identify
event beds. Event beds I, II, and III (used in Bentley Bioturbation Model) are shaded
in blue. A) The cores proximal to the beach, collected 20 meters from high tide
(determined from wrack line). The sediments are comprised of organic-rich sandy
muds.











Physical Proportios T3C2


Percent Sand Reiduals of Gamma Bdk Demnty (g omn) Pixel hnteilty
10 20 0 40 00 0. 0D 1u 1.2 IA IA 2 as 00 Is Io
I-- - - . i. i. .. . . i. I i . i . i . l .i ,


E





E tiI
~~~~t ll~ -


iI


-d

C,
C-
S


Figure 5-1. Continued. B) The cores distal to the beach, collected 30 meters from high tide
(determined from wrack line). The sediments are comprised of organic-rich sandy
muds.


Event III


vLY eILe II













revsvion (%i Event I


2


3






0
I
s-

U
B"


Bently Bioturbation Model T2C1

Prusevdion(Y.)- Eventl
0 W 20 30 I E w] W n w IOm 1


-. R.es. d Ihi bmm IunX-Rahdao
------ Rmd Nfsmun fatrX-ilSh i
-.- m.il.lalPm.ealn irimMldal
PrOIvliaon (P) EBfnt III
o0 a 1I 3 4050 D 70 W 110


Figure 5-2. Modeling Results for Cores T2C1 and T3C2. A) Results from the BBM for Events I,
II, and III in the proximal core T2C 1. The cores proximal to the beach, collected 20
meters from high tide (determined from wrack line). Plots represent the minimum and
maximum preservation from the x-radiographs with the results of the BBM for each
given event.













Presrvelon (%) Event I
0 10 2 30 40 s s0 7 0 0 100


Bently Bloturbatlon Model T3C2

Preservelon (%) Bent II
0 10 20 3 40 0 00 70 0 90110
32 1 I ' I a


Qal Modded Eert
Bad TRLdsu 225 an.


I37


-.... P1i mullnim himX-a oM
--. Poln PMWisaiaHn irm odl
Prnsrvllon (%) Bent III
0 10 20 30 40 50 0 80 90 O100


i I t Mod erat
B-d .ll.. 2 m.
--.... -- --- ...









U d9 1ma cm.


Figure 5-2. Continued. B) Results from the BBM for Events I, II, and III in the distal core T3C2.
The cores distal to the beach, collected 30 meters from high tide (determined from
wrack line). Plots represent the minimum and maximum preservation determined
from the x-radiographs with the result from the BBM for each event.









CHAPTER 6
CONCLUSIONS

The depositional environments on St. Vincent Island offer an advantageous environment

for studying the preservation of storm deposits in the coastal stratigraphic record due to the

frequent occurrence of large storms, relatively low mixing rates (i.e., long dissipation times) and

fast sedimentation rates (a few mm y-1). Event beds are likely to be preserved within the salt

marsh sedimentary record. This study found evidence of event bed preservation within the salt

marsh investigated. They are not however, represented solely by a change in grain size, but

rather by other changes within the sediments as well (i.e., lithologic changes).

The ultimate control on or primary function of preservation potential is the initial or

original thickness of the event layer and the bioturbation factors present. Event beds are more

likely to be found proximal to the beach and less likely to be found with increasing distance.

Initial or original event bed thickness is dependant on costal morphology and vegetation.

This study also indicates pixel intensity (i.e., x-radiography) to be an extremely useful

proxy of preservation; perhaps the ideal proxy for this type of research. It is extremely high

resolution with minimal destruction to the cores. It also offers the capability of locating very

fine beds / laminations within the stratigraphic record.

The Bentley Bioturbation Model can be applied to coastal marshes. The Bentley

Bioturbation Model, which was developed for continental shelf environments, is applicable

within coastal marsh environments. It successfully reproduced the chosen event beds found

within T2C1 and T3C2. However, it should be used with caution as the dynamics of coastal

environments may create inaccurate results for this steady state model. It should be tested in

other sedimentary environments, especially in the coastal realm.









LIST OF REFERENCES


Allen, J.R.L. (2000) Morphodynamics of Holocene Salt Marshes: A Review Sketch from The
Atlantic and Southern North Sea Coasts Of Europe. Quaternary Science Review, 19, 1155-
1231.

Allen, E.A., & Curran, H.A., (1974) Biogenic Sedimentary Structures Produced By Crabs in
Lagoon Margin And Salt-Marsh Environments Near Beaufort, North-Carolina. Journal of
Sedimentary Petrology, 44(2), 538-548.

Aller, R.C., Benninger, L.K., & Cochran, J.K. (1980) Tracking Particle-Associated Processes in
Nearshore Environments by use of 234TH/238U Disequilibrium. Earth and Planetary
Science Letters, 47, 161-175.

Appleby, P.G., & Oldfield, F. (1992) Application ofLead-210 to Sedimentation Studies. In:
Uranium-series Disequilibrium: Applications to Earth, Marine, and Environmental
Sciences, Vol. 0 (Ed. By I.M. and R.S. Harmon) Clarendon Press, Oxford, pp.731-778.

Benninger, L.K., Aller, R.C., Cochran, J.K., & Turekian, K.K. (1979) Effects of Biological
Sediment Mixing on the 210Pb Chronology and Trace Metal Distribution in a Long Island
Sediment Core. Earth and Planetary Science Letters, 43, 241-259.

Bently, S.J., & Sheremet, A. (2003) New Model for the Emplacement, Bioturbation, and
Preservation of Fine-Scaled Sedimentary Strata. Geology, 31(8), 725-728.

Boggs, S. (2001) Principals of Sedimentology and Stratigraphy Third Edition. Upper Saddle
River, Prentice Hall, pp.59-74.

Boudreau, B.P. (1994) Is Burial Velocity a Master Parameter for Bioturbation? Geochimica et
Cosmochimica Acta, 58, 1243-1249.

Bromley, R.G. (1996) Trace Fossils Biology, Taphononmy, and Applications Second Edition:
London, Chapman & Hall, pp.225-229.

Campbell, K. (1986) St. Vincent Island (St. Vincent National Wildlife Refuge), Florida.
Geological Society ofAmerica Centennial Field Guide Sulwet ni\i Section, 351-353.

Collins, E.S., Scott, D.B., & Gayes, P.T. (1999) Hurricane Records on the South Carolina Coast:
Can They be Detected in the Sediment Record? Quaternary International, 56, 15-26.

Cutshall, N.H., Larsen, I.L., & Olsen, C.R. (1983) Direct Analysis of 210Pb in Sediment
Samples: Self-Adsorption Corrections. Nuclear Instrumentation and Methods, 206, 309-
312.

Davis, R.A. (1995) Geologic impact of Hurricane Andrew on Everglades Coast of Southwest
Florida. Environmental Geology, 25(3), 143-148.









Davis, R.A., Knowles, S.C., & Bland, M.J. (1989) Role of Hurricanes in the Holocene
Stratigraphy of Estuaries Examples from the Gulf Coast of Florida. Journal of
Sedimentary Petrology, 59(6), 1052-1061.

Donnelly J.P., Bryant, S.S., Butler, J., Dowling, J., Fran, L., Hausmann, N., Newby, P.,
Shurman, B., Stern, J., Westover, K., & Webb, T. III. (2001) 700 Yr Sedimentary Record
of Intense Hurricane Landfalls in Southern New England. GSA Bulletin, 113(6), 714-727.

Dukat, D.A., & Kuehl, S.A. (1995) Non-Steady State 210-Pb Flux and the Use of 228Ra/226Ra
as a Geochronometer on the Amazon Continental Shelf. Marine Geology, 125, 329-350.

Edwards, J.M., & Frey, R.W. (1977) Substrate Characteristics within a Holocene Salt Marsh,
Sapelo Island, Georgia. Senchenberg. Marit., 9, 215-259.

Frey, R.W., & Basan, P.B. (1978, 1985) Coastal Salt Marshes: in, Davis, R.A., ed., Coastal
Sedimentary Environments: New York, Springer-Verlag, pp.225-301.

Frey, R.W., & Howard, J.D. (1969) A Profile of Biogenic Structures in a Holocene Barrier
Island-Salt Marsh Complex, Georgia. Transactions of the Gulf Coast Association of
Geological Sciences, 19, 427-444.

Green, M.A., & Aller, R.C. (2001) Early Diagenesis of Calcium Carbonate in Long Island Sound
Sediments: Benthic Fluxes of Ca+2 and Minor Elements During Seasonal Periods of Net
dissolution. Journal ofMarine Research, 59, 769-794.

Harris, P.T., Heap, A.D., Bryce, S.M., Porter-Smith, R., Ryan, D.A., & Heggie, D.T. (2002)
Classification of Australian Clastic Coastal Depositional Environments Based upon a
Quantitative Analysis of Wave, Tidal, and River Power. Journal of Sedimentary Research,
72, 855-870.

Hart, M.R. (2003) Evaluating the Preservation of Hurricane Deposits in Florida Coastal
Sediments. University of Florida, Master's Thesis.

Katz, L.C., 1980. Effects of Burrowing by the Fiddler Crab, Uca-Pugnax (Smith). Estuarine and
Coastal Marine Science, 11(2), 233-237.

Keen, T.R., & Stone, G.W. (2000) Anomalous Response of Beaches to Hurricane Waves in a
Low-Energy Environment, Northeast Gulf of Mexico, U.S.A. Journal of Coastal
Research, 16(4), 1100-1110.

Koretsky, C.M., Meile, C., & Van Cappellen, P. (2002) Quantifying Bioirrigation Using
Ecological Parameters: A Stochastic Approach. Geochemistry Transactions, 3(3), 17-30.

Liu, K., & Fearn, M.L. (1993) Lake-Sediment Record of Late Holocene Hurricane Activities
from Coastal Alabama. Geology, 21, 793-796.









Liu, K., & Feam, M.L. (2000) Reconstruction of Prehistoric Landfall Frequencies of
Catastrophic Hurricanes in Northwestern Florida from Lake Sediment Records.
Quaternary Research, 54, 238-245.

Liu, K., & Feam, M.L. (2002) Lake Sediment Evidence of Coastal Geologic Evolution and
Hurricane History from Western Lake, Florida. Quaternary Research, 57(3) 429-431.

McCraith, B.J., Gardner, L.R., Wethey, D.S., & Moore, W.S., 2003. The Effect of Fiddler Crab
Burrowing on Sediment Mixing and Radionuclide Profiles Along a Topographic Gradient
in a Southeastern Salt Marsh. Journal ofMarine Research, 61(3), 359-390.

Nittrouer, C.A., DeMaster, D.J., McKee, B.A., Cutshall, N.H., & Larsen, I.L. (1984) The Effect
of Sediment Mixing on Pb-210 Mixing Rates for the Washington Continental Shelf.
Marine Geology, 54, 210-222.

National Oceanic Atmospheric Administration's National Weather Service (2003) National
Hurricane Center, http://www.nhc.noaa.gov/index.shtml, updated regularly.

Rappaport E. N. (1995) Hurricane Erin 31 July 6 August 1995. National Hurricane Center
Preliminary Report, 26 November 1995, http://www.nhc.noaa.gov/1995erin.html, last
updated 02 January 1999.

Risi, J.A. (1998) Event Sedimentation from Hurricane Andrew Along the Southwest Florida
coast, University of Miami, Ph.D. Dissertation, pp.199.

Rose, P.R., & Lidz, B. (1977) Diagnostic Foraminifera Assemblages of Shallow-Water Modem
Environments : South Florida and the Bahamas. Comparative Sedimentology Laboratory,
Division of Marine Geology and Geophysic, Rosenstiel School of Marine and
Atmospheric Science, The University of Miami, pp.56.

Simpson & Herbert. (1973) Monthly Weather Review. 101(4).

Thieler, E.R., & Hammar-Klose, E.S. (1999) National Assessment of Coastal Vulnerability to
Sea-Level Rise: Preliminary Results for the U.S. Gulf of Mexico Coast. US. Geological
Survey Open-File Report 00-179, http://pubs.usgs.gov/of/2000/of00-
179/textonly/textindex.html, last updated 06 Sep 2001.

Wheatcroft, R.A. (1990) Preservation Potential of Sedimentary Event Layers. Geology, 18, 843-
845.

Wheatcroft, R.A., & Drake, D.E. (2003) Post-Depositional Alteration and Preservation of
Sedimentary Event Layers on Continental Margins, I. The Role of Episodic Sedimentation.
Marine Geology, 199, 123-137.

Wheeler, A.J., Orford, J.D., & Dardis, O., 1998. Saltmarsh Deposition and its Relationship to
Coastal Forcing Over the Last Century on the North-West Coast of Ireland. Geologie en
Mijnbouw, 77(3-4), 295-310.









Williams, J.M., & Duedall, I.W. (2002) Florida Hurricanes and Tropical Storms 1871-2001:
Gainesville, University Press of Florida, pp.88-100.

Woodroffe, C.D. (2002) Coasts: Form, Process, & Evolution: Cambridge University Press, pp.
378-432.









BIOGRAPHICAL SKETCH

Lisa Marie Erickson Mertz was born May 27, 1979, to Curtis and Jaclyn Erickson.

Growing up in Minnesota with a younger sister, Lisa Marie always enjoyed the outdoors. A

honor graduate of Proctor Senior High School in 1997, she was involved in drama and

volleyball. She continued a volleyball career at the University of Wisconsin River Falls while

studying geology and hydrogeology. In 2001, she began her graduate studies under Dr. John M.

Jaeger in the University of Florida's Department of Geological Sciences. Her interests in

sedimentology lead her to a salt marsh and its history of tropical cyclones. Lisa Marie is

currently working with an environmental and engineering consulting firm in western

Washington.