Florida State University Libraries

Electronic Theses, Treatises and Dissertations The Graduate School

2017 Fate of the Mesophotic Coral Ecosystem (MCE) in the Northeastern Gulf of Mexico after the Deepwater Horizon Incident: Impacts, Restoration, Conservation, and Hazards Mauricio G. Silva-Aguilera

Follow this and additional works at the DigiNole: FSU's Digital Repository. For more information, please contact [email protected]

FLORIDA STATE UNIVERSITY

COLLEGE OF ARTS AND SCIENCES

FATE OF THE MESOPHOTIC CORAL ECOSYSTEM (MCE) IN THE NORTHEASTERN

GULF OF MEXICO AFTER THE DEEPWATER HORIZON INCIDENT: IMPACTS,

RESTORATION, CONSERVATION, AND HAZARDS.

MAURICIO SILVA-AGUILERA

A Dissertation submitted to the Department of Earth, Ocean, and Atmospheric Science in partial fulfillment of the requirements for the degree of Doctor of Philosophy

2017 Mauricio Silva-Aguilera defended this dissertation on November 17, 2017.

The members of the supervisory committee were:

Ian R. MacDonald Professor Directing Dissertation

Janie L. Wulff University Representative

Markus Huettel Committee Member

Amy Baco-Taylor Committee Member

Mariana Fuentes Committee Member

The Graduate School has verified and approved the above-named committee members, and certifies that the dissertation has been approved in accordance with university requirements.

To Martina and Diego, the reason for everything.

To follow the path:

Look to the master,

Follow the master,

Walk with the master,

See through the master,

Become the master.

Zen Proverb

iii TABLE OF CONTENTS

List of Tables ...... vi List of Figures ...... vii Abstract ...... x

1. INTRODUCTION ...... 1

1.1 General Aspects ...... 1 1.2. Geological Characteristic of the Gulf of Mexico ...... 2 1.2.1 Formation ...... 2 1.2.2 Depositional and Sea Level History of the GoM ...... 3 1.3 Study Site ...... 4 1.4 Thesis Objectives ...... 5 2. CORAL INJURIES OBSERVED AT MESOPHOTIC REEFS AFTER THE DEEPWATER HORIZON OIL DISCHARGE ...... 7

2.1 Introduction ...... 7 2.2 Methods...... 8 2.2.1 Study Site ...... 8 2.2.2 Sample Collection ...... 9 2.3.3 Image Analysis...... 12 2.2.4 PAH Analysis...... 14 2.3 Results ...... 14 2.3.1 Coral Injury ...... 14 2.3.2 Coral Injury Frequency and Distribution ...... 18 2.3.3 PAH results ...... 20 2.4 Discussion ...... 25 2.5 Conclusions ...... 31

3. HABITAT SUITABILITY MODELING FOR MESOPHOTIC CORAL IN THE NORTHEASTERN GULF OF MEXICO ...... 32

3.1 Introduction ...... 32 3.2 Methods...... 34 3.2.1 Study Site ...... 34 3.2.2 Survey ...... 36 3.2.3 Environmental Data ...... 37 3.3.4 Dynamic Environmental Variables ...... 38 3.3.5 Modeling ...... 40 3.3 Results ...... 41 3.3.1 Model Performance ...... 42 3.3.2 Taxon-specific Model Results ...... 43 3.3.3 Independent Test Models ...... 45 3.4 Discussion ...... 46 iv 3.6 Conclusions ...... 51

4. SUBMARINE CHANNELS, MASS MOVEMENTS, AND OIL INDUSTRY: A POTENTIAL HAZARD FOR THE MESOPHOTIC CORAL ECOSYSTEM (MCE) IN EASTERN GULF OF MEXICO ...... 53

4.1 Introduction ...... 53 4.2 Methods...... 56 4.2.1 Study Site ...... 56 4.2.2 Geophysical Data ...... 56 4.2.3 Piston Core Samples ...... 58 4.2.4 Geophysical Data Analysis ...... 58 4.2.5 Piston Core Analysis...... 59 4.3 Results ...... 60 4.3.1 Geomorphological Description ...... 60 4.3.2 Sediment Data ...... 63 4.3.3 Channels and Evidence for Mass Movement ...... 67 4.3.4 Stable and Radio Isotopes ...... 71 4.4 Discussion ...... 75 5.5 Conclusions ...... 81

5. CONCLUSIONS ...... 83

APPENDIX ...... 87 A: SUPPLEMENTARY MATERIAL FOR CHAPTER 3 ...... 87

References ...... 94 Biographical sketch ...... 115

v LIST OF TABLES

Table 1. Summary of visited sites during the research expedition in September 2011. Site names in parentheses are abbreviations used in text. Distance indicates proximity to DWH discharge site ...... 10

Table 2. Visual scale of coral injury adapted from White et al. (2012) ...... 13

Table 3. Summary of images collected and analyzed from primary study sites. Only photos with corals visible were analyzed ...... 13

Table 4. Frequency of injured coral taxa (levels 1–4), by taxa, in AAR and RTR study sites. The 1997–1999 columns refer to results from the MAPTEM study, while 2011 is the post-DWH discharge observation...... 19

Table 5. Abundance of tall (>0.5 m) growth form coral colonies ( Bebryce spp. excepted) for all injury stages in both study sites...... 20

Table 6. Total polycyclic aromatic carbon (tPAH) values in parts per billion (PPB) dry weight for invertebrate tissues and sediments collected during surveys in 2010 and 2011 and from baseline samples collected in 2001 (Continental Shelf Associates and Texas A&M University, 2001). Reported as number of samples: mean value (standard deviation)...... 25

Table 7. Summary of the environmental variables used to develop the model ...... 39

Table 8. Summary statistics (average and SD) of training and test AUC and gain, suitable area (% of the total, and total in square km), and jackknife test of variables for each model. Variable abbreviation is as follows: Rug: rugosity, B_u: bottom u, Hurr: hurricane winds, CDOM: chromophoric dissolved organic matter, DoSed: dominant sediments, TPI500: topographic index position 500m, OS: oxygen saturation, SST: sea surface temperature, LoSed: loose sediment, UnIso: backscatter unsupervised iso-classification. Var: variables; All: all variables; Sel: selected variables...... 45

Table 9. Summary of sediments samples taken from each core for each analysis...... 60

Table 10. Summary of channel metrics ...... 60

Table 11. Summary table of stable and radio isotopes of PC04A and PC04B in the study area. .72

Table 12. Statistic summary of variables used to perform the model ...... 92

Table 13. Pearson correlation used to evaluate significance of individual variables in the coral suitability model. Variables with positive Pearson correlation >0.4, or negative Pearson correlation <-0.4 were selected as main drivers of habitat suitability……………………...... 93

vi LIST OF FIGURES

Figure 1. Map of the location of the study sites in relation of the oil discharge (DWH). Alabama Alps Reef (AAR) and Roughtongue Reef (RTR) are the principal mesophotic reef sites. Additional sites visited during 2010 and/or 2011 included Coral Trees Reef (CTR), Talus Block (TBR), Yellowtail Reef (YTR), and Madison Swanson South Reef (MSSR). Area covered by oil based on normalized observations of oil per unit area (5×5 km2 grid) imaged by Satellite Aperture Radar (Lessard and DeMarco, 2000) from Garcia-Pineda et al. (2013)...... 10

Figure 2. Photographic documentation of coral injuries. Alabama Alps Reef (left) and Roughtongue Reef (right). Yellow crosses indicate navigation waypoints occupied by the ROV. Each red star represents an image sample analyzed for photo-documentation of coral health. See Table 2 for photograph inventory by site. Blue circles represent the area within which randomly located photographs were collected during the MAPTEM surveys...... 11

Figure 3. Healthy an unhealthy tissue observed on Hypnogorgia pendula (above) and Swiftia exserta (below) colonies. Insets show details of healthy and damaged tissue to distinguish differences in the same colony...... 17

Figure 4. Injury levels (1–4) observed on Hypnogorgia pendula colonies...... 18

Figure 5. Distribution of injury levels 0 (left), 1 (center), and 2–4 (right) and coral colony abundances within Alabama Alps Reef (upper panels) and Roughtongue Reef (lower panels). Dot size indicates abundance of injured corals ( Bebryce spp. excluded)...... 21

Figure 6. Damage frequency by injury levels (Table 2) for MAPTEM and 2011 surveys for study sites. Upper chart: all coral colonies taxa except Bebryce spp. Lower chart: Bebryce spp. alone...... 22

Figure 7. Box and whisker plots of Kruskal–Wallis test for frequency at Alabama Alps Reef (upper) and Roughtongue Reef (lower) study sites comparing intermediate to extreme injuries for MAPTEM and the 2011 for tall growth forms (Bebryce spp. excluded). Plot key as follows: blue square—outliers, bar—greatest value excluding outliers, red diamond—median, upper box limit—quartile 75%, middle line—mean (quartile 50%), and lower box limit—quartile 25% ...23

Figure 8. PAHs from 2000 to 2011. Sediment core and tissue collections for PAH analysis. Location symbols indicate sample type and year. Grab sample locations reported in a 2001 study (Continental Shelf Associates and Texas A&M University, 2001) are plotted for comparison. ..24

Figure 9. Map of aerial dispersant application (red lines) and oil burning sites (flames) over floating oil layers discharged by the DWH blowout (NOAA, 2013)...... 29

Figure 10. Areas of floating oil over Gulf of Mexico between July 23 and July 26 of 2010 as interpreted by NOAA NESDIS (NOAA, 2013) and subsequent dissipation after TS Bonnie passed over the region. Blue dotted line represents the complete trajectory of TS Bonnie. Daily

vii location of TS Bonnie is plotted using the hurricane symbol for July 24th (yellow) and July 25th (red)...... 30

Figure 11. Extent of the area mapped by USGS (Gardner et al., 2001; 2002 and 2003) showing the region of potential mesophotic habitats between Mississippi and Florida analyzed in the coral habitat suitability model. A. Alabama-Mississippi Pinnacle Trend; B. Head of Desoto Canyon; and C. Florida Middle Grounds...... 35

Figure 12. Octocoral and antipatharian colonies found in the mesophotic coral ecosystem (MCE) of the study area. A) Hypnogorgia pendula. B) Swiftia exserta C) Stichopathes sp. D) Ellisella sp. E) Thesea cf nivea . F) Thesea cf rubra . G) Placogorgia sp . H) Placogorgia cf tenuis . I) Bebryce sp. J) Nicella sp. K) Antipathes cf. atlantica ...... 37

Figure 13. General model of habitat suitability for all mesophotic corals in the study area using the selected parameters. Warm areas (yellow to red) show the predicted locations were corals are likely to be found (high habitat suitability), while cold areas (dark green to blue) indicates low probability of finding mesophotic corals (low habitat suitability). A) model build with MAPTEN data set. B) Model build with the NOAA data set...... 44

Figure 14. Responses curves of the most important variables showing the relationship between performance of predictors and the “general” habitat suitability model for all mesophotic octocorals and black corals in the study area. Black lines indicate how the model prediction changes as the environmental variable varies taking in consideration all the other variables. Grey curves characterize the model’s response using only that variable ...... 45

Figure 15. Multi-beam bathymetry map of the study site area, which is located Northeast Gulf in the continental slope and shelf off shore of between Mississippi and Florida...... 57

Figure 16. Locations of piston cores collected inside and outside of channels. Note proximity of energy platforms installed in or adjacent to potentially unstable slope regions...... 59

Figure 17. 3D view of the channel system of the study area. Transverse profiles in different sections of each channel allows to see the progression from V shape (steep slopes) to U shape (gradual slopes)...... 63

Figure 18. Display of the sub-bottom profiles transects collected progressively along the downslope course of the A-channel. Inset shows associated bathymetry and transect lines...... 68

Figure 19. Multibeam bathymetry of upper slope of Desoto Canyon with indications of sediment movements. Cross section of upper slope subbottom profile and multibeam profile ...... 69

Figure 20. High resolution scan images of the sediment cores retrieve and analyzed of the study area. Length-width relationship not in scale...... 70

Figure 21. Scan image of PC03A showing different stratigraphic facies along the core ...... 71

viii Figure 22. Stable isotopes ( ¹³C and ¹⁵N) profiles analyzed for the study area...... 73

Figure 23. Left: ¹³C versus C/N ratio relationship of sediment cores. Right: stable ( ¹³C) and radiocarbon isotope ages (Δ¹⁴C age BP) correlation. Both figures show three different sources of sediment (carbon) in the study area ...... 74

Figure 24. Radiocarbon depth profiles of PC04 (outside of the channel) and PC04-B (inside of the channel). Left: Δ¹⁴C‰; Right: Δ¹⁴C age in years BP...... 75

Figure 25. Detail of sinuous channels and submarine landslide areas showing locations of piston cores, energy platforms, and pipelines. Note that several energy platforms have been installed in or adjacent to potentially unstable slope regions. Bathymetry compiled from NOAA and BOEM datasets ...... 80

Figure 26. Map of the study area including coral records from the MAPTEM data and the 2014 cruise (MAPTEM) that were used in this study to create the habitat suitability model...... 87

Figure 27. Map of the study area including independent coral records NOAA used to create the habitat suitability model ...... 88

Figure 28. Responses curves of selected variables showing the relationship between predictors behavior and the “general” habitat suitability model for all mesophotic octocorals and black corals in the study area. Black lines indicate how the model’s prediction changes as the environmental variable varies taking in consideration all the other variables. Grey curves characterize the model’s response using only that variable ...... 89

Figure 29. “ By Taxon ” models of habitat suitability for mesophotic corals in the study area using the selected parameters. Warm areas (yellow to red) show the predicted locations were corals are likely to be found ...... 90

Figure 30. Independent models of habitat suitability for mesophotic corals compiled using NOAA coral records and subset of significant parameters...... 91

ix ABSTRACT

The Gulf of Mexico, one of the most geomorphologically complex oceanic basin, is also one of the most prolific hydrocarbon reserves in the world. It hosts a varied range of marine communities from shallow reefs and estuarine communities in coastal areas, to deep-sea and chemosynthetic communities in greater depths. In some particular areas in water depth between 50 to 120 m are hosted the mesophotic reef ecosystem, also known as the twilight communities, where sunlight is almost extinguished but still enough to support some photosynthesis. Mesophotic ecosystems are mainly characterized by the presence of both light dependent and independent corals, which are the bioengineering support from a vast variety of invertebrate and vertebrate that compound the mesophotic coral ecosystem. One of these mesophotic coral ecosystem, the Pinnacle Reef Trend, was under the influence of floating oil after the Deepwater Horizon incident released more than 4.1 million barrels of crude oil and 500 T of gas to the environment. Visible injuries in over 400 octocoral and antipatharian colonies were quantified in the aftermath of the DWH oil discharge. Observations were made in September 2011 at water depths of about 65 to 75 m in the Pinnacle Reefs area offshore of Mississippi and Alabama, Gulf of Mexico, using a digital macro camera deployed from an ROV to examine the coral populations of two principal sites: Alabama Alps Reef (AAR) and Roughtongue Reef (RTR). Observed taxa (identifications provisional), listed in order of injury frequency, included the following: Hypnogorgia pendula (Duchassaing & Michelotti, 1864), Bebryce spp ., Thesea nivea (Deichmann, 1936), Swiftia exserta (Ellis & Solander, 1786), Antipathes atlantica (Gray, 1857), Stichopathes sp ., and Ellisella barbadensis (Duchassaing & Michelotti, 1864). The most conspicuous injuries observed were a biofilm, often with a clumped or flake-like appearance, that covered sea-fan branches. Extreme injuries were characterized by bare skeleton, broken, and missing branches. Comparing the 2011 results to previous photo surveys in the same study sites between 1997 and 1999, we found, in 2011, significantly elevated occurrences of injury covering more than 10% of colony area among taxa with growth forms > 0.5 m. We hypothesize that Tropical Storm Bonnie facilitated and accelerated the mixing process of dispersant-treated hydrocarbons into the water column, resulting in harmful contact with coral colonies at mesophotic depths. Analysis of sediment PAH concentrations at AAR and RTR found levels elevated above pre-DWH discharge values, but orders of magnitude below toxicity thresholds established for

x fauna in estuarine sediments. The PAH concentrations measured in octocoral and echinoderm tissue samples from AAR and RTR were low compared to detection thresholds (10 - 100 ppb). Our findings indicate that coral injuries observed in 2011 resulted from an acute, isolated event rather than ongoing natural processes.

Maximum Entropy Modeling was applied to predict the spatial extent of mesophotic azooxanthellate octocorals and antipatharians within the mesophotic area located between Mississippi (Pinnacle Trend Area) and the mid continental shelf and upper slope of Florida, eastern Gulf of Mexico. Habitat prediction models were generated using geo-referenced, coral-presence records obtained by compiling photographic samples with co-located geophysical data, oceanographic variables, and atmospheric variables. Resulting models were used to predict the extent of suitable habitat in the study area. An independent set of presence-records was used to test the model performance. Results (general and by taxon) predict that suitable areas for MCE exceed 400 km 2, which occur along carbonate mounds and paleo-shoreline ridges (hard substrata and high ruggedness) with lower amounts of fine sediments and surrounding waters rich in CDOM and upwelling currents (w). The model results significantly exceeded (>0.5) random output and predicted that ruggedness and CDOM are the most important variables associated with coral habitat. Areas of hard substrate within the study area that were not identified as coral habitat by the model suggest that mesophotic sea fans and sea whips depend as much on the chemical and physical conditions (e.g. currents that transport oxygen and food) as on hard substrata for settlement.

Finally, three submarine channels that incise into the continental slope in Northeastern Gulf of Mexico have been identified and described. The slope-source channels, seemingly formed after a mass wasting events, are 65 km long and 2.5 km width on average. They show a transition of their transversal profile from V-shape in the head to U-shape in the toe, high sinuosity in the upper ¾ of the channel, showing accumulation of sediments in the turn-sections of the channels, and accumulation of sediment on the walls due to collapsing of unconsolidated sediments. Sedimentary facies obtained by sediment cores show the evidence of disturbed interval of sedimentation and erosion, with erratic periods of laminar sedimentation, erosion, turbidities, disturbed by coarse fractions of sediment. C14 dating of sediments cores from the inside of the channel suggest that

xi hydrodynamic erosion is continuously occurring inside the channel, exposing older sediments to the surface. However, we do not discard the possibility that mass wasting processes are still occurring in the area. The presence of coarse sediment fractions typically found in the continental shelf, beside the occurrence of turbidite homogenous fractions demonstrate that sediment wasting from the upper slope could be still occur. This could lead to a potential disaster in the area because of the presence of oil industry structures. Massive turbidity flows and submarine slumps have been described to occur in this area and other places.

xii

CHAPTER 1 INTRODUCTION

1.1 General Aspects

The Gulf of Mexico (GoM) is a relatively small sedimentary ocean basin located on the passive margin of North America between the Yucatan block (actual Mexico) and the North American continental plate, which is divided from the Atlantic and Caribbean Ocean by the Florida-Bahamas platform (Galloway, 2008). One of the most distinctive characteristics of the GoM basin are its geomorphological and bathymetric complexity and diversity of environments, which is a reflection of its complex geological history since formation.

Today, the Gulf of Mexico displays a geomorphology shaped during the Late Quaternary (Briantt et al., 1991) characterized by the prominent Mississippi fan that rests on the abyssal plain of the basin (>3500 m depth) connecting a broad continental shelf, especially in the Florida Platform area and the northwestern GoM. The continental slope is the most complex and diverse geomorphologically speaking. The eastern GoM slope represents the abrupt transition from the coastal to the abyssal plain, with a short but almost vertical slope (>90º) called the Florida Escarpment. The north-east section is characterized by the presence of Desoto Canyon, a partially filled portion (by Appalachian Mountain sediments) of the South Georgia Rift Basin and the Georgia Seaway. During the Mesozoic and Cenozoic Era, when low sea level periods allowed rivers from the North America plate filled with sediments and merge the Suwanee and Florida blocks to the continental plate (Klitgord and Hutchinson, 1988; Redfern, 2001; Hine, 2013). This area is also characterized by the presence of erosional channels and platforms mixed with salt- domes produced by the migration of halite to the surface of the seafloor. The continental slope west of the Mississippi Canyons is the longest in extension and the most geomorphologically complex. This area is called the Sigsbee Escarpment, characterized by the deformation of the seafloor by the movement of salt diapirs, which buoyancy makes move up through deep layers of sediments; allowing gas and oil to migrate from the reservoirs towards the sea floor, being released to superficial sediments layers and the water column, offering an energy source for chemosynthetic communities. (Rowan and Ratliff, 2012).

This diversity of geomorphological settings across the entire range of depths, generates multiple habitats for marine fauna. In this context, we will focus in the coral communities located in the mesophotic or twilight zone, located in clear water between 30 to 150 m depth (Locker et al., 2010). Surface waters down to the Mesophotic Coral Ecosystem (MCE) (Hinderstein et al., 2010) in northwestern GoM were exposed to oil and the subsequent application of dispersant after the Deepwater Horizon oil discharge in April 2010 (NOAA, 2013; USGS, 2011). One of the main concerns during that time was that floating oil could reach and affect the MCE in the area under the influence of the spill. Therefore, it was important to quantify the level and extent of damage in this region. Analyzing a series of images obtained after the disaster and comparing them to previous studies and images obtained nearly two decades earlier, made it possible to evaluate and quantify possible injuries to mesophotic octocorals and antipatharians, the ecological engineers of the MCE in the area. In addition, we used mathematical modeling to establish the areas where MCE can actually occur within the MCE potential range. Finally, we provide evidence of the likelihood of potential disasters that could happen in the area below of the MCE using geological records of mass wasting processes in the continental slope. This raises safety concerns for expansion of oil and gas production in the region

1.2. Geological Characteristics of the Gulf of Mexico

1.2.1 Formation

The origin of the GoM is dated to the Late Triassic (237 201 Ma), starting with the detachment of the North American Plate from the African and South American continental plates, which comprises the mega-continent Pangea (Mancini et al., 2001; Galloway 2008). The break split Pangea into the two super-continents Gondwana (Africa and South America) and Laurentia (North America and Europe). The created space due to the extension between both continental plates giving place to the proteo-Atlantic Ocean, called the Tethys Sea. At the same time, a series of rift basins formeded diagonally to the coast of Laurentia, in the actual location of Georgia, called the South Georgia Rift Basin (Klitgord and Hutchinson, 1988). As the seafloor spreading continued, the Florida Block also separated from Laurentia moving towards the North America plate, forming the North Atlantic and Caribbean Ocean. The space in between the Florida Platform and North America plate (South Georgia Rift Basin) was completely flooded with seawater,

2 creating the Georgia Seaway or Suwannee Channel (Salvador, 1991). This seaway was eventually filled with carbonate sediments transported by rivers from the Appalachian Mountains, closing the Seaway and forming the region of the Gulf of Mexico that today is known as Desoto Canyon (Salvador, 1991; Pindell and Kennan, 2009). About 160 Ma, seafloor spreading continued moving the Yucatan Block towards the south of the early Gulf of Mexico basin, closing it and finally forming the Gulf of Mexico basin as is know today. The original Gulf of Mexico basin extended beneath modern land areas. (Salvador, 1991; Mancini et al., 2001).

Today, the GoM marine basin is smaller than when it was when formed, reaching current inland areas as far as Oklahoma. The thickest sedimentary stratigraphy is roughly 10-15 km in the middle of the abyssal plain (Salvador, 1991). Continuous changes on the sea level, changes of location of the Mississippi Delta, and changes in sedimentation rates due to inter-glaciation periods formed various depositional centers and chapped the intricate geomorphology of the GoM (Mancini et al., 2001, Galloway 2008., Galloway et al., 2011).

1.2.2 Depositional and Sea Level History of the GoM

Salt tectonism, changes in sedimentary rates, and changes in sea level are mainly responsible for the modern geomorphology and of the GoM. During the last several million years, the GoM has experienced major fluctuations of sea level, which has produced significant shifts in shoreline position (above and bellow currents levels). Sea-level records from the Gulf of Mexico from the Last Glacial Maximum (20,000 YBP) are over 35m higher than records from equatorial regions in the same period of time (Sims et al., 2007). Although changes in sedimentation rates in the Mississippi Delta could affect changes in sea-level by adding large volumes of sediment and displacing water, paleo-oceanographic and radiocarbon sediments record from the Mississippi fan do not support this hypothesis (Stuiver and Braziunas, 1993; Stuiver et al., 1998; Sims et al., 2007).

So far, the most plausible explanations for the sea-level changes in the region are glacio- hydro-isostasy processes resulting from the formation and disappearance of the Laurentide Ice Sheet formation (15,000 YBP - 6,000 YBP) (Sims et al., 2007)

3 After the formation of ice sheets in North America, sea-level in the GoM decreased more than 100 m in comparison to present levels. At this depth, shorelines in the northern GoM were located on the continental shelf break, allowing accumulation of carbonate sediments and the formation of coastal coral reefs (Continental Shelf Associates and University, 2001; Donoghue, 2011). After this period, the retreat and melt of ice sheets produced a pulse of fresh water and terrigenous sediments, increasing the sea-level by an average of 6 mm/year, with periods of rapid rise beyond 40 mm/year (Donoghue, 2011). During this time, scleractinian reefs where drowned and covered by sediments.

Today, these ancient reefs and paleoshorelines form part of the MCE called The Pinnacles Reefs trend, which comprises a series of carbonate structures of different sizes and morphologies (Continental Shelf Associates and University, 2001; Peccini and MacDonald, 2008). The feature hosts diverse invertebrate and fish fauna. The most distinctive and diverse groups associated with these structures are azooxanthellate octocorals, antipatharians, and scleractinan corals, which also provide structural habitat for mobile fauna. (Peccini and MacDonald, 2008).

1.3 Study Site

This study is located in the northeastern portion of the Gulf of Mexico offshore of Mississippi and Florida. The area in question extends between 60 and 110 m depth in the continental shelf and shelf break, where a series of hard ground pinnacles, mounds and platforms host a substantial mesophotic coral ecosystem (MCE). This area is connected to the deep basin through a series of submarine channels and erosional platforms that run from the shelf break to the deep basin.

The first part of this study (Chapter 2) focused on the impact of the Deep Water Horizon (DWH) oil incident (2010) over the MCE in The Pinnacles Reef trend. This region extends from offshore of Mississippi/Alabama towards Florida between 50 to 150 m depth range. This region is characterized by the presence of sessile fauna attached to a numerous biogenic rocky outcrops that originated when the area was located in shallower water, forming long extension of scleractinian reefs, which after a de-glaciation period were inundated due the rising of the sea level and now are located in a greater depth.

The second part of this study (Chapter 3) is focused on the same area, The Pinnacles Reef trend, and other two less explored and studied MCE sites in the Gulf of Mexico; the Florida Middle 4 Ground reef system (Smith et al., 2006), and the area in between this reef and The Pinnacle Reef trend, which is called the “Upper Desoto Canyon” (Locker et al., 2010) in order to provide a better understanding on where the MCEs are located and which variables are responsible for the allocation and survival of the MCE in the area.

Finally, in Chapter 4 we focused our efforts on the region just below the MCE zone, which is characterized by a series of erosional channels and platforms that connects the coastal areas to the deep basin. This area has the potential to generate mass movement events that may affect oil and gas infrastructure that could lead to a new disaster similar to the DWH incident, again exposing the MCE to pollutants.

1.4 Thesis Objectives

The transformations that have occurred to the Gulf of Mexico in addition to inter-glaciation periods, sea level changes, salt tectonics, natural and unnatural hydrocarbon release, mass wasting processes have helped to shape a variety of habitat setting along the its depth range. For this reason, it is important to identify the different geomorphological settings and the differences among them. Understanding the geomorphology and the communities that resides on them, evaluate the extension of those locations (and its own biological communities) and evaluate possible risks that could modify/affect them.

Before the DWH incident, relatively few studies have focused on the MCE in northeast Gulf of Mexico (Continental Shelf Associates and University, 2001; Peccini and MacDonald, 2008). Those studies are the only reports of the baseline of the MCE prior to the DWH incident. Similarly, few studies provided information on the effect of crude oil on octocorals and antipatharians in mesophotic and deep areas. In addition, due the potential extension of the MCE in the Gulf of Mexico, more of the area remains still unexplored, including the area below the floating oil slicks during and after the DWH that could have been affected by it. Providing new insights over the real extent of the MCE and having a better understanding which oceanographic and geomorphological factors will positively influence the health and survival of this communities will provide an exceptional tool for restoration, management and exploration of the MCE. Furthermore, it is important to evaluate and quantify risk factors that could set of a similar or even worse disasters than the DWH accident. 5

The main goals of this study are

(1) Provide a careful evaluation and quantification of the health state of the MCE octocorals and antipatharians in the area close to the DWH incident and compare the results with previous studies before the oil spill.

(2) To create a habitat suitability model of the MCE in the northeastern GoM in order to elucidate in which areas of the potential MCE region it is most likely to find MCEs.

(3) to evaluate the potential risk for future incidents like the DWH accident in the area due to mass wasting events.

CHAPTER 2 CORAL INJURIES OBSERVED AT MESOPHOTIC REEFS AFTER THE DEEPWATER HORIZON OIL DISCHARGE Manuscript published on Deep Sea Research Part II: Topical Studies in Oceanography Volume 129, July 2016, Pages 96–107 doi: http://dx.doi.org/10.1016/j.dsr2.2015.05.013

2.1 Introduction

The Pinnacles Reef trend is a region of the Gulf of Mexico that extends from offshore of Mississippi eastward toward the Florida Panhandle; it encompasses numerous rocky outcrops colonized by sessile reef fauna in the 50–150 m depth range (Gittings et al., 1992a; Gittings et al., 1992b). High-relief hard grounds on the northeastern Gulf of Mexico shelf are widely understood to have a biogenic origin. They formed when reef organisms were inundated to greater depths because of the rising sea level during the last deglaciation. The Pinnacles Reef trend comprises one of four well-known mesophotic coral ecosystem (MCE) sites in the Gulf of Mexico; the others being the Flower Garden Banks and other hard grounds offshore of Texas(Continental Shelf Associates and University, 2001), the Florida Middle Ground reef system (Smith et al., 2006b), and Pulley Ridge (Locker et al., 2010). Together, they are part of 178,867 km 2 of substrata that could support MCE in the northern Gulf of Mexico (Continental Shelf Associates and Texas A&M University, 2001).

The Pinnacles Reefs comprise a series of spatially distinct, carbonate features of different sizes and morphologies, which serve as important destinations for commercial and recreational fisheries (Dennis and Bright, 1988; Weaver et al., 2002). They host diverse invertebrate and fish fauna. Continental Shelf Associates and Texas A&M University (2001) reported 40 invertebrate taxa for the Pinnacle Reef area offshore of Mississippi–Alabama. The most diverse groups are azooxanthellate octocorals, sponges, antipatharians, and hermatypic corals, all of which also provide structural habitat for invertebrates and fish. Gorgonian octocorals and antipatharian black corals have both sea-fan and sea-whip morphologies: i.e. a single attachment point called a holdfast, from which branches may or may not bifurcate to form a flexible proteinaceous skeleton (Bayer, 1961). Large branching colonies have a flabellate shape, along which colonies of polyps are distributed. Corals on the Pinnacles Reefs are predominantly heterotrophic suspension feeders that depend on both plankton and detritus provided by local currents for nutrition (Bayer, 1961). Octocorals in the Pinnacles Reefs have been shown to grow so that their fans are permanently oriented to maximize exposure to local currents (Peccini and MacDonald, 2008).

On April 20th 2010, an explosion on the drilling platform Deepwater Horizon (DWH) and the subsequent well blowout discharged more than 4.1 million barrels of crude oil (McNutt et al., 2012) and over 500,000 T of hydrocarbon gas (Joye et al., 2011a) into the water column over a span of 84 days. The ocean surface over the Pinnacle Reefs was covered by oil released by the DWH blowout for approximately 35 days on average depending on location; floating oil received repeated applications of dispersant, including in areas directly above Pinnacles Reefs to the north and northeast of the wellhead (ERMA, 2016; NOAA, 2013; Operational Scene Advisory, 2010, 2011). The combination of surface oil and dispersant application raised concerns that sessile organisms such as octocorals in this vicinity might have suffered injury as a result of exposure to these chemical contaminants. Similar effects have been reported for deep-sea communities of octocorals in the vicinity of the Macondo well as a result of the DWH discharge (White et al., 2012).

The objectives of the present study were to characterize and quantify pathologies observed among octocorals and antipatharian corals on mesophotic reefs situated below the surface oil discharged during the DWH incident and to compare these results with data collected before the oil spill. Sampling efforts were undertaken in September 2010 and September 2011 to obtain sediment and tissue samples (2010 and 2011) and to conduct extensive high-resolution photographic inspections (2011 only). The present study provides a descriptive and quantitative analysis of these data and compares the results to historic data from the region. Photographic data collected in 1997–1999, which were reanalyzed for this study, provided a quantitative, pre- discharge baseline for the natural frequency of coral injuries in the Pinnacle Reefs habitat.

2.2 Methods

2.2.1 Study Sites

Between 1997 and 1999, a series of photographic surveys were conducted for the “Mississippi–Alabama Pinnacle Trend Ecosystem Monitoring” (MAPTEM) program, visiting 8 nine MCE sites in total (Continental Shelf Associates and Texas A&M University, 2001). This effort was accomplished using a SeaRover ROV outfitted with a still camera. The frequency of coral injury had not been previously assessed for these data.

Post-discharge assessments of DWH oil impacts were conducted during August 2010 on the RV Nancy Foster and September 2011 on the MV Holiday Chouest. Over the course of these two post-discharge cruises, a total of six sites were visited (Fig. 1; Table 1): four were within ~100 km distance of the DWH site offshore from Mississippi and Alabama; two presumably unaffected sites were located offshore from Florida. This study focuses on photographs taken during 2011 at two mesophotic sites also surveyed by MAPTEM expeditions: Alabama Alps Reef (AAR) and Roughtongue Reef (RTR), because they provide a good opportunity for before–after comparison of still images. Surveys also took place on ‘secondary sites’ Talus Block (TBR), Yellowtail Reef (YTR), Coral Trees Reef (CTR), and Madison Swanson South Reef (MSSR), but with less effort towards still images. Data on sediment hydrocarbon concentrations presented herein are taken from the 2010 effort, and data on hydrocarbon concentrations in coral tissues are taken from both the 2010 and 2011 efforts. On both cruises, an industrial-class remotely operated vehicle was used to survey the bottom area and collect samples.

2.2.2 Sample Collection

2.2.2.1 Photography Survey

Baseline photographic data from AAR and RTR were collected during the MAPTEM project (Continental Shelf Associates and Texas A&M University, 2001). Photographic surveys were conducted using a ROV outfitted with a still camera illuminated by a 150 watt-second electronic strobe. Camera and strobe were triggered by shipboard control. The camera and strobe were installed on the front of the ROV looking down at an angle of approximately 70 degrees. The camera covered an area of 0.3 m 2. The overall location for photographic survey work was determined prior to each cruise using a digital elevation model to focus effort on the reef crest and not the surrounding seabed. A 200 m diameter circular area was divided into eight equal sectors, each containing 16 randomly chosen points; photographs were then taken to document the sessile community at each of the 16 locations (Fig. 2).

Table 1. Summary of visited sites during the research expedition in September 2011. Site names in parentheses are abbreviations used in text. Distance indicates proximity to DWH discharge site.

Distance Reef Site (km) Longitude Latitude Depth (m) Area (km²) Alabama Alps (AAR) 57 -88.33924 29.253668 74 0.276 Talus Block (TBR) 87 -87.76679 29.320955 130 0.023 Yellowtail (YTR) 109 -87.59169 29.450339 64 0.119 Roughtongue RTR) 109 -87.57581 29.439161 66 0.140 Coral Tree (CTR) 231 -86.13945 29.486935 88 0.143 Madison Swanson (MSSR) 266 -85.67931 29.186576 73 0.402

During the 2011 field effort, a digital still camera (AquaSLR) in a deep-sea housing was mounted on a frame that carried two high-output LED lamps; the array was deployed by the ROV manipulator arm. In preparation for photograph collection, the ROV performed a reconnaissance of each site and identified the regions of the two reefs where octocoral colonies were found in noticeable abundance. During reconnaissance, the ROV transited at 0.5–1 m above the bottom in transects across hard ground areas in the respective reefs, while recording position events at 1 min intervals and noting the occurrence of corals (Fig. 2). Generally, sea fans and other octocorals were found concentrated on rocky promontories within the reef crest. Each locality where corals were observed was then revisited to determine presence or absence of injured corals. Coral localities were photographed with the digital camera to document the condition of corals present there. Injured corals were often photographed from several angles to facilitate later evaluation. For detailed photo-documentation, the ROV manipulator arm was used to aim the camera at subjects of interest. The close placement of the camera with abundant lighting produced high resolution images. The camera was also deployed in transit mode in conjunction with the video cameras on

11 the ROV to document coral health characteristics over larger areas. The locations of images taken and the dive-tracks of the ROV at AAR and RTR are summarized in Fig. 2.

2.2.2.2 Coral and Sediment Collection

Coral tissues for taxonomic identification and chemical analysis were collected using the ROV manipulator arm. Individual colonies were targeted for collection in the video display. Branches were broken off from the main colony, stored temporarily in the ROV collection box, and carried to the surface at the end of each dive. Species diagnoses were achieved using scanning electron microscopy of sclerites, polyps, and branches in conjunction with photographic and taxonomic keys.

Sub-samples for PAH analyses were packed in methanol-rinsed foil and stored frozen at temperature of −20 °C. Sediment samples were collected on two occasions by two different sediment sampling devices. In August 2010, sediments were collected from Alabama Alps and Roughtongue Reef using 6.2 cm diameter push-cores. Additionally, in September 2010 sediment samples were collected on a separate cruise using a 0.04 m2 Young-modified Van Veen benthic grab sampler deployed from the ship deck (Cooksey et al., 2014). Samples for the analysis of contaminants were sub-sampled from composited surface sediment contained in the grab samples (the upper 2–3 cm) and stored frozen.

2.2.3 Image Analysis

Digital images from the 2011 survey and the MAPTEM program were carefully reviewed for indications of injury to corals, including the following pathologies: loose tissue, broken branches, bare skeleton, necrosis, and areas covered by strands of mucus or biofilm overgrowth. We then focused our analysis on those pathologies indicative of polyp mortality, and quantified the extent to which each injured colony was affected. Each coral colony exhibiting biofilm overgrowth, broken branches, bare skeleton, or necrosis was classified according to the extent of injury across the colony into five categories, following the scale proposed by White et al. (2012), where level-0 indicates no evident injury and level-4 is the most extensive (Table 2). Corals exhibiting limited signs of stress, such as retracted polyps or mucus production, were classified as level-1.

12 Table 2. Visual scale of coral injury adapted from White et al. (2012).

Level Description Injury Area (%) 0 Uninjured Less than 1% 1 Mild 1% to 10% 2 Intermediate 10% to 50% 3 Severe 50% to 90% 4 Extreme Over 90%

A dataset of 185 images from the 2011 survey was analyzed for evidence of coral injuries. Images of poor quality, low resolution, or insufficient luminosity, as well as duplicated images, were discarded from analysis. For baseline comparison, 580 images (those containing corals) were analyzed from MAPTEM dataset (Table 3). For MAPTEM images and the 2011 images, “Image J” software was used to calculate the area of the coral colony and the relative proportion of injury. The nonparametric one-way analysis of variance by ranks test (Kruskal and Wallis, 1952) was used to test for statistical differences comparing the presence of injured corals between the MAPTEM and 2011 studies. Data were grouped by year in eight different categories: a group of all levels of injury (levels 1–4), a group of mild-intermediate injuries (level 1), a group of intermediate to extreme injuries (levels 2–4) and each 0–4 separately. The dependent variable was the number of injured colonies per photograph, and the independent variable was the sampling year.

Table 3. Summary of images collected and analyzed from primary study sites. Only photos with corals visible were analyzed.

Site Date Study ID Taken Analyzed 1997 CSA-TAMU 187 77 1998 CSA-TAMU 164 71 AAR 1999 CSA-TAMU 188 70 2011 This study 633 104 1997 CSA-TAMU 181 139 1998 CSA-TAMU 171 124 RTR 1999 CSA-TAMU 144 99 2011 This study 741 81 Total AAR-RTR 2409 765

13 2.2.4 PAH Analysis

Tissue samples of benthic invertebrates were analyzed for total polycyclic aromatic hydrocarbon (tPAH) by Alpha Analytical Laboratories in Mansfield, MA. Analyses were conducted pursuant to the Analytical Quality Assurance Plan for Mississippi Canyon 252 (Deepwater Horizon) NRDA Version 3.0 (NOAA, 2011) using the ToxPAH50 methodology. Sediment samples were analyzed for total petroleum hydrocarbon and tPAH, which were measured by Battelle Laboratories using EPA Method 8270-SIM (semi-volatile organic compounds by gas chromatography/mass spectrometry with selective ion monitoring). The tPAH values were calculated using protocols listed in the NOAA PAH-51 method. The tPAH data, as well as additional data on abiotic environmental variables in sediment (including grain size, total organic carbon, latitude–longitude, and water depth) were downloaded from the Environmental Response Management Application (ERMA) Gulf Response website (NOAA, 2013).

2.3 Results

2.3.1 Coral Injury

Coral samples collected by ROV and identified by morphological techniques included the following taxa: Hypnogorgia pendula, Bebryce spp. Thesea nivea , Swiftia exserta , Antipathes atlantica , Stichopathes sp., and Ellisella barbadensis . This section will compare the characteristics of injury to these corals at AAR and RTR prior to the DWH discharge with characteristics of injury observed at the same reefs in 2011 after the discharge. The numerical analysis comparing frequency of injury pre- and post-discharge and spatial distribution of post-discharge injury is presented separately.

2.3.1.1 Pre-discharge Conditions

Between 1997 and 1999, the mesophotic corals at AAR and RTR had relatively few cases of injury overall; furthermore, the types of injuries observed differed qualitatively from post- discharge injuries, and injuries occurred in different taxa.

In pre-discharge conditions, the most injured taxon was Bebryce , especially in low-growth- form colonies (<20 cm above attachment point). Bebryce colonies were frequently found covered

14 by a thin layer of fluffy brown sediment. There were no examples of retracted polyps or dead tissue. In some cases, baseline photographs showed a light mucus layer surrounding functional polyps. In the most severe cases, Bebryce was covered by sediment, leaving only terminal branches exposed.

Among the taller-growth-form species >50 cm above attachment point), such as H. pendula and S. exserta, broken branches, dead tissue, bare skeleton, and hydroid overgrowth were not observed in pre-discharge images. Injury to coral colonies occurred at mild and intermediate injury levels (levels 1 and 2), characterized by small branch areas that were covered by a thin layer of sediment and mucus, but polyps were functional. The few cases of severe injury in H. pendula and S. exserta which were observed indicated gross mechanical impact that toppled the colony or buried it under sediment. Some colonies of T. nivea had indications of minor stress, such as mucus secretion and branches partly covered by sediment. Injuries to whip corals Stichopathes sp., E. barbadensis , and the black coral A. atlantica were rarely observed. Colonies of Stichopathes sp. and E. barbadensis were occasionally found with discolored soft tissue and polyps lacking close to the attachment point (holdfast). Some A. atlantica had small patches of green biofilm covering distal branches in the upper portion of the colony or a layer of sediment covering the basal portion of the coral.

2.3.1.2 Post-discharge Condition

Several coral pathologies were observed, and these differed among taxa. In the case of T. nivea and Bebryce spp., injured branches were typically covered by hydrozoans and/or biofilm; denuded branches were not observed in these species. Patches of healthy and unhealthy polyps, characterized by retracted polyps or areas covered by mucus and biofilm material were observed in some cases in these two taxa when injury only covered less than 10% of the colony (level 1).

Taxa with growth forms taller than approximately 0.5 m, such as H. pendula, S. exserta , and T. nivea , typically displayed the most severe injury levels, while those growing close to the substratum showed fewer severe injuries in the post-discharge study. Among H. pendula there was biofilm overgrowth with hydroids or greenish algal material. In severe examples, H. pendula colonies lost some or all branches. Broken branches were sometimes observed in a semi-detached

15 state (Fig. 3a). S. exserta colonies were commonly observed with intermediate to severe injuries (levels 2–3); extreme injury levels were also seen (level 4). In general, injured corals exhibited sharp gradients between healthy tissues with extended polyps and necrotic tissues with brown or green biofilm overgrowth or skeletal branches denuded of tissue. The living tissue near dead patches of the colony did not show biofilm material or overgrowth with hydroids, but polyps were often retracted (Fig. 3b).

Among the three taller taxa, H. pendula was found with exposed skeleton from necrotic loss of soft tissue. The full spectrum of injury was evident in this taxon. The lowest degree of injury (level 1) was characterized by the secretion of mucus and locally retracted polyps. In more pronounced cases of injury, biofilm overgrowth and tissue necrosis were present on soft tissue. hydroid colonization was visible on necrotic tissue and bare skeleton. Finally, (level 4), branches weighted down by hydroids broke apart from the colony (Fig. 4). A similar range was observed in S. exserta colonies and, to a lesser degree, in Stichopathes sp. colonies. In these specimens, presence of brown biofilm overgrowth was evident along with disintegration of smaller branches. In the case of T. nivea and Bebryce spp., which are low-growing taxa, mucus secretion and abnormal tissue coloration were common indicators of stress at both study sites. Injuries to A. atlantica and Bebryce consisted of debris overlain on branches. Epifauna on Bebryce in particular made it difficult to assess minor injury levels.

Among the three taller taxa, H. pendula was found with exposed skeleton from necrotic loss of soft tissue. The full spectrum of injury was evident in this taxon. The lowest degree of injury (level 1) was characterized by the secretion of mucus and locally retracted polyps. In more pronounced cases of injury, biofilm overgrowth and tissue necrosis were present on soft tissue. Hydroid colonization was visible on necrotic tissue and bare skeleton. Finally, (level 4), branches weighted down by hydroids broke apart from the colony (Fig. 4). A similar range was observed in S. exserta colonies and, to a lesser degree, in Stichopathes sp. colonies. In these specimens, presence of brown biofilm overgrowth was evident along with disintegration of smaller branches. In the case of T. nivea and Bebryce spp., which are low-growing taxa, mucus secretion and abnormal tissue coloration were common indicators of stress at both study sites. Injuries to A. atlantica and Bebryce consisted of debris overlain on branches. Epifauna on Bebryce in particular made it difficult to assess minor injury levels. 16

Figure 3. Healthy and unhealthy tissue observed on Hypnogorgia pendula (above) and Swiftia exserta (below) colonies. Insets show details of healthy and damaged tissue to distinguish differences in the same colony.

Figure 4. Injury levels (1–4) observed on Hypnogorgia pendula colonies.

2.3.2 Coral Injury Frequency and Distribution

Six taxa of octocoral and antipatharian corals were found to have both tissue and skeleton injuries. The following octocoral taxa are sorted according to prevalence of injury: H. pendula, which showed the greatest frequency of injured colonies at both sites, comprising 40.3% at AAR and 42.5% at RTR of injured colonies (levels 1–4), followed by Bebryce spp. (16.5% and 18.0%), T. nivea (15.4% and 13.5%), and S. exserta (17.8% and 19.5%); see Table 4 and Table 5. Details of species' frequency by injury level are reported in Supplementary Tables 1 and 2. Review of the data shows that Bebryce spp., the low-growing species, exhibited an injury frequency that was variable year-by-year during the pre-discharge observations (Table 4).

18 Table 4. Frequency of injured coral taxa (levels 1–4), by taxa, in AAR and RTR study sites. The 1997–1999 columns refer to results from the MAPTEM study, while 2011 is the post-DWH discharge observation.

Site AAR RTR Species 1997 1998 1999 2011 1997 1998 1999 2011 Hypnogorgia pendula 0 0 0 97 8 0 3 85 Swiftia exserta 1 8 6 43 14 7 20 39 Thesia nivea 0 0 0 37 4 0 0 27 Bebryce sp. 7 34 14 40 107 26 64 36 Antipathes sp. 0 11 2 15 2 0 1 9 Placogorgia sp. 2 0 1 0 25 0 1 0 Elisella sp. 0 0 2 0 22 16 9 0 Stichopathes sp. 0 0 4 9 1 3 3 4 Total injured (n) 10 53 29 241 183 52 101 200

Coral colonies with intermediate or greater degrees of injury (levels 2–4) were more abundant at AAR and RTR post-discharge (150 cases) than in the pre-discharge surveys. This was the case for T. nivea and S. exserta . In addition, injured colonies of H. pendula , S. exserta , and to a lesser degree Stichopathes sp. were frequently found with high amounts of hydroids and biofilm growing over bare skeleton, and broken branches were observed on the sea floor. H. pendula was the only taxon found in the four different levels of injury at both sites. In contrast, healthy colonies with lower stress indicators (<10% injury) were more frequent at RTR. Healthier colonies were represented by all coral taxa found at both sites, but mainly by small colonies (<20 cm) of T. nivea and Bebryce spp. (Fig. 5). However, a large number of Bebryce colonies were observed displaying mild to intermediate injuries in RTR in 1997 and 1999. Bebryce is a problematic taxon for assessment because the overgrowth of hydroids and epiphytes occurs in these species with uncertain impact on coral health. In addition, colonies of low profile such as Bebryce can be affected by sedimentation and sediment re-suspension to a higher degree than coral species growing above the substrata. Injury quantification for this taxon is therefore presented separately (Fig. 6).

19 Table 5. Abundance of tall (>0.5 m) growth form coral colonies ( Bebryce spp. excepted) for all injury stages in both study sites.

Injury level 0 1 2 3 4 Frequency n % n % n % n % n % 1997 238 96.0 8 3.2 1 0.4 0 0.0 1 0.4 1998 190 78.2 35 14.4 17 7.0 1 0.4 0 0.0 AAR 1999 127 81.4 20 12.8 7 4.5 2 1.3 0 0.0 2011 264 52.3 91 18.0 71 14.1 42 8.3 37 7.3 1997 405 68.9 104 17.7 60 10.2 14 2.4 5 0.9 1998 449 89.6 19 3.8 28 5.6 5 1.0 0 0.0 RTR 1999 259 71.9 52 14.4 24 6.7 16 4.4 9 2.5 2011 358 64.2 98 17.6 60 10.8 9 1.6 33 5.9

A Kruskal–Wallis test was performed for each treatment, showing significant differences (p<0.05) between post-discharge observations and all pre-discharge years for injury levels 2–4 combined ( Bebryce spp. excluded). A Friedman “a posteriori” test (multiple comparisons between p-values and z-values) showed that the numbers of intermediate, severe, and extreme injuries (levels 2–4 combined) found in 2011 were significantly higher than the numbers of injuries found in the pre-discharge samples (MAPTEM). (Fig. 7)

2.3.3 PAH Results

Oil was found at detectable levels in invertebrate tissues and sediments from both sites: AAR and RTR (Fig. 8; Table 6). The sensitivities to total PAHs, dispersants, and mixtures of oil and dispersants have not been assessed for deep-water sea fans, but shallow corals exhibit adverse reactions to low concentrations of dispersed oil over 3–6 d time periods (National Research Council, 2005). At AAR and RTR, baseline tPAH levels in sediments were available from sampling that pre-dated the DWH discharge (Continental Shelf Associates and Texas A&M University, 2001). The values found post-discharge exceeded baseline values at both sites by a

Figure 6. Damage frequency by injury levels (Table 2) for MAPTEM and 2011 surveys for study sites. Upper chart: all coral colonies taxa except Bebryce spp. Lower chart: Bebryce spp. alone .

Figure 7. Box and whisker plots of Kruskal–Wallis test for frequency at Alabama Alps Reef (upper) and Roughtongue Reef (lower) study sites comparing intermediate to extreme injuries for MAPTEM and the 2011 for tall growth forms ( Bebryce spp. excluded). Plot key as follows: blue square—outliers, bar—greatest value excluding outliers, red diamond—median, upper box limit— quartile 75%, middle line—mean (quartile 50%), and lower box limit—quartile 25%

factor of 5 in 2010 and a factor of 2 in 2011. tPAH values from the tissues of invertebrates were relatively high compared to tPAH values from the sediments. The highest tPAH values were found in octocoral tissues at AAR. Sediment PAH values from AAR, closer to the wellhead, also were higher than sediment tPAH values at RTR, which is farther from the wellhead on the eastern side of the northern Gulf of Mexico. The highest post-discharge tPAH concentrations found in

23 sediments (at AAR), however, were 25–33% of the lowest values reported within 3 km of DWH (NOAA, 2013).

24 Table 6. Total polycyclic aromatic carbon (tPAH) values in parts per billion (PPB) dry weight for invertebrate tissues and sediments collected during surveys in 2010 and 2011 and from baseline samples collected in 2001 (Continental Shelf Associates and Texas A&M University, 2001). Reported as number of samples: mean value (standard deviation).

AAR RTR Sample type 2000 2010 2011 2000 2010 2011 Sediment Grabs 10: 26 (n.a.) n.a. n.a. 12: 10 (n.a.) Push cores n.a. 3: 101 (43.9) 10: 47 (20.9) n.a. 6: 62 (52.6) 10: 19 (8.8) Tissue Cnidaria n.a. 1: 150 (n.a.) 23: 367 (807.5) n.a. 4: 224 (267.0) 37: 50 (84.3) Porifera n.a. n.a. na n.a. 1: 27 (n.a.) 4: 17 (9.6) Echinodermata* n.a. 8: 56 (23.4) 4: 34 (32.3) n.a. 1: 228 (n.a.) n.a. *basket star

2.4 Discussion

This study found a large number of octocoral and antipatharian corals in September 2011 that exhibited pathologies consistent with injury resulting from acute impact (see for example Chan et al., 2012), but did not observe coral predators, fishing gear, or evidence for sedimentation that would suggest the cause of the injuries. Similar severity and frequency of injuries to corals with growth forms >0.5 m had not been noted during repeated surveys of the Pinnacle Reefs coral habitats prior to the DWH discharge (Continental Shelf Associates and Texas A&M University, 2001). One objective of the 2001 report was to assess health and condition of hard bottom communities. Our review of the pre-discharge data did find minor (level 1) injuries to a low- growth-form taxa (e.g. Bebryce spp.), which might have been caused by regularly occurring processes. However, injuries to the tall-growth-form taxa observed in at the AAR and RTR study sites in 2011 were significantly more frequent and severe than anything reported in 1997, 1998, or 1999 at the same sites.

Pathologies similar to those reported in this study were observed among deep-sea corals thought to have been exposed to crude oil and dispersant-treated oil from the DWH discharge via a deep-water plume (Hsing et al., 2013 and Rooker et al., 2012). These authors reported the presence of flocculated biofilm material, mucus secretion, necrosis, bare skeleton, and branches colonized by hydroids on several gorgonians and other coral taxa at a 1370 m coral community in the lease block MC292.

25 The total numbers of injured colonies documented at AAR and RTR exceeded those reported by White et al. (2012) and Hsing et al. (2013). This reflects the larger size of the shallow reef habitats surveyed and an overall higher abundance of corals in the shallower mesophotic marine setting. According to White et al. (2012), the observed injury to the deep-sea corals was attributed to the DWH oil and gas discharge close to the study sites. This evaluation was corroborated by analysis of biofilm material, which was composed of dead coral tissue and residual hydrocarbon compounds, the biomarker signature of which matched DWH oil. We hypothesize, based on the characteristics and severity of the damage found in coral colonies after the DWH incident and the relative scarcity of coral injury reported in the pre-discharge MAPTEM study, that injury observed among corals in the vicinity of AAR and RTR was caused either directly or indirectly by the DWH oil discharge and/or response activities, which included extensive use of dispersants directly over these affected sites.

Coral injuries attributable to stress from chronic pollution, oil spill incidents, and bioaccumulation of hydrocarbon compounds (PCB's, PAH's and NPEs) have been reported previously by Loya and Rinkevich (1980), Guzman et al., 1994 and Guzman et al., 1991, Vogt (1995), Poulsen et al. (2006), Cailleaud et al. (2007). Chan et al. (2012) provide a detailed description of how heavy-metal bioaccumulation produces similar injuries to those reported in the study. McClanahan et al. (2004) propose that necrosis is an extreme response of the soft coral immune system to stresses caused by external factors such as temperature changes, pollution, or diseases. Pollutant intake by corals can result from direct contact and be further exacerbated by feeding behavior that selects floating particles, as has been suggested by Anthony (2000). Mitra et al. (2012) reported evidence that mesozooplankton in the Gulf of Mexico exposed to PAH's related to the DWH blowout eventually entered the marine food chain, transferring PAH's to higher trophic levels. So the literature amply supports pathological responses to stress in this group and indicates possible pathways for exposure.

Other known and documented natural sources of pathology in octocoral colonies include disease, predation, bottom-contact fishing impacts, landslides, and changes to the ocean temperature (i.e. coral bleaching). In shallow Caribbean reefs, fungal infections observed on sea fans are taxon-specific (Toledo-Hernández et al., 2008 and Zuluaga-Montero et al., 2010);

26 however, in this study we observed similar injuries among several different octocoral and antipatharian taxa. Fishing gear is more likely to impact octocoral taxa indiscriminately than disease. It is possible that recreational fishing lines could inflict sub-lethal abrasions to soft corals that would make the sea fans more vulnerable to infections or predation. However, few of the injuries observed in 2011 or previous MAPTEM ROV imagery showed any evidence of fishing gear. Predators to deep water octocorals include flamingo tongue snails, gastropods, asteroid sea stars, and butterflyfish. However, few coral predators were observed in either MAPTEM or 2011 surveys. The Pinnacles Reefs may experience periods of high sedimentation associated with the Mississippi outflow, which may also result in injuries similar to what we observed, including exposed, dead skeleton and hydroid growth. However, historical studies of Pinnacles reefs have suggested that MCEs are resistant to adverse impacts from sediment, and may even use sediment as a food source (Continental Shelf Associates and Texas A&M University, 2001).

Prior to the DWH blowout, Peccini and MacDonald (2008) measured soft coral orientation in relation to bottom currents at both AAR and RTR, examining photographs of 365 sea fans at AAR, RTR, and three other sites in the Pinnacle Reef area. They did not report any notable coral injury at that time. This is also consistent with our analysis of the MAPTEM photography collection, of which Sites 1 and 7 correspond to RTR and AAR, respectively. Photographic analysis of coral communities reveals that some injured colonies were present in that period, but the severity and extent of coral injury does not match those reported in the 2011 study. This is corroborated by the Kruskal–Wallis test results, which found statistical differences (p<0.05) between samples, and the Friedman test demonstrated that post-discharge samples (2011) were significantly higher in terms of the numbers of injuries found in all the injury levels in the pre- discharge samples (MAPTEM).

Both sites, AAR and RTR, were situated under layers of floating oil from the DWH oil discharge. Benthic organisms could be exposed to this oil through trophic pathways or by direct contact. Gin et al. (2001) proposed a model of how an oil spill could affect food chain interactions in marine environments. The model assumes that there are two ways that an organism interacts with hydrocarbon compounds: (1) consumption of contaminated prey items and (2) direct

27 incorporation by tissues of dissolved oil from surrounding water or sediments. Measurable concentrations of hydrocarbons in octocoral tissues, more than 1 year after oil had disappeared from the surface waters of the Gulf, were well below the bioeffect concentrations established for estuarine sediments (Long et al., 1995).

One of the palliative measures taken to mitigate the floating oil was the use of dispersants; this could have increased chances of direct contact of oil with coral. Chemical dispersants are surfactants with both lipophilic and hydrophilic properties that serve to break the oil into smaller droplets, increasing surface area for volatilization, dissolution, or degradation (Lessard and DeMarco, 2000). The repeated application of dispersants near and directly over the AAR and RTR sites (Fig. 9) increased the probability of exposure of corals at these sites to oil and/or dispersants by facilitating mixture of floating oil into sub-surface water depths. Dispersant application demonstrates the presence of thick oil layers near the sites, because thicker accumulations of surface oil were specifically targeted by aerial dispersant missions. Moreover, oil dispersants such as Corexit 9500 (Nalco Holding Company), used to respond to the DWH blowout, have deleterious effects in different marine organisms. Goodbody-Gringley et al. (2013) described the effect of different concentrations of Corexit mixed with crude oil on larval behavior of coral Montastraea faveolata and Porites asteroides. Both coral larvae changed their settlement behavior. The degree of impact of dispersants depends on its chemical concentrations and which species are affected. In extreme cases, oil and dispersant produce complete failure of settlement and total mortality of coral larvae, especially when they are exposed to high concentration (25 ppm) of Corexit alone. Similar results have been obtained by DeLeo et al. (2016) where colonies of Paramuricea biscaya, Callogorgia americana and Leiopathes glaberrima exposed to treatments of dispersant alone and oil-dispersant mixtures were harmed more than those exposed to oil alone.

Another mechanism oil spill responders used to decrease the amount of oil on the surface water was burning the thicker layers of oil (Fig. 9). When oil is burned and combusted in ocean environments, large amounts of soot particles (black carbon) are dispersed into broad areas by ocean currents and wind. Perring et al. (2011) reported that 4% (1.35±0.72×106 Kg.) of burned oil was released to the atmosphere as black carbon. The particles of black carbon can eventually reach

Figure 9. Map of aerial dispersant application (red lines) and oil burning sites (flames) over floating oil layers discharged by the DWH blowout (NOAA, 2013).

the surface of the water by rain or sedimentation. Weber et al. (2012) suggest that disease and mortality on coral colonies can be attributed to particle sedimentation. Organic-rich material can produce tissue degradation within a day, and inorganic particles produce the same effect after approximately 6 days. White et al. (2012) found dark droplets and particles in biofilm material that was removed from injured deep-water corals close to the DWH site.

An additional event that could have contributed to floating oil reaching the bottom was the passage of Tropical Storm (TS) Bonnie over the Gulf of Mexico between July 22 and July 26. Synthetic Aperture Radar (Lessard and DeMarco) images of the floating oil obtained from NOAA Satellite and Information Service (NESDIS) showed that the floating oil over the mesophotic Pinnacles Reefs almost disappeared from the ocean surface over the Pinnacles on July 26 (Fig. 10). Rain and wind would have accelerated oil and water mixing rates on the ocean surface, removing and breaking the floating oil into small particles, introducing and dissolving it into the water column. Finally, the oil dispersed in the water column would have been mixed downward to the pinnacles because of the large waves generated by TS Bonnie. Such effects would have been more deleterious to coral health than storms under normal conditions. Continental Shelf Associates and Texas A&M University (2001) reported an increase of surface and bottom (97 cm/s) currents 29 in the Pinnacle Reefs after the passage of hurricanes Earl and George in 1998. However, no occurrence of coral pathologies was noted by researchers who subsequently surveyed gorgonian colonies (e.g. Peccini and MacDonald, 2008). So it is unlikely that coral injuries reported here were caused by the TS Bonnie alone.

Analysis of the TPH and tPAH concentrations in sediments and tissues post-discharge (2010 and 2011) does show a slight increase above pre-discharge levels, but this increase may not in itself be significant for coral health because the concentrations are far below toxic-effect levels for test organisms. Chronic-effect levels of TPH and tPAH for the octocoral taxa have not been developed, nor have effect levels been established for dispersants, so deleterious impact cannot be

30 entirely ruled out. However, the low concentrations suggest that injury did not occur due to exposure from residual pools of hydrocarbons in the reef environment.

2.5 Conclusion

We conclude that widespread coral injuries observed at two mesophotic communities within the Pinnacles Reefs in the Gulf of Mexico during 2011 were qualitatively different and significantly greater in frequency than anything noted prior to the DWH discharge. A characteristic of this injury was patchy damage affecting portions of the sea fan colony. Anomalous injuries were observed in coral taxa with a growth form >0.5 m; the species H. pendula sustained the most injury at both sites. In more severe levels, biofilm overgrowth covered portions of the colony, and patches of necrotic tissue were observed. The most severe levels of injury were characterized by bare, broken, and missing branches.

The cause of these injuries is hypothesized to be exposure to oil and/or dispersant from the DWH discharge. Proximity of the sites to dispersant applications and, in the case of the western- most site, burning operations increased the risk of harmful impact. Passage of TS Bonnie over the region would have cause the upper water column to mix turbulently downward, which would have exposed corals to material floating near the surface, particularly on promontories of the reefs and for taxa with relatively higher relief above the bottom. Because the concentrations of hydrocarbons in sediments and animal tissues collected in 2011 were well below chronic bio-effect levels, exposure must have occurred as an acute event

31 CHAPTER 3 HABITAT SUITABILITY MODELING FOR MESOPHOTIC CORAL IN THE NORTHEASTERN GULF OF MEXICO Manuscript accepted in Marine Ecology Progress Series. DOI:https://doi.org/10.3354/meps12336

3.1 Introduction

In the Gulf of Mexico, there are 4 areas where mesophotic coral ecosystems (MCEs) extend from 30 to 100 m depths. They are characterized by the presence of light-dependent/independent corals and associated fauna (Kahng et al., 2010) and are designated as follows: the Pinnacles Reefs, the Flower Garden Banks and other hard-ground features offshore of Texas (Continental Shelf Associates and University, 2001; Locker et al., 2010; Wright et al., 2004), the Florida Middle Ground reef system (Smith et al., 2006a), and Pulley Ridge (Locker et al., 2010). However, these well-studied areas constitute a small fraction of the approximately 179 000 km2 of substrata that could potentially support MCEs in the northern Gulf of Mexico (Locker et al. 2010).

The Pinnacles Reefs and the Florida Middle Ground mesophotic reefs comprise a broad array of carbonate features within the 30−100 m isobath, including flat-topped mounds measuring hundreds of meters across, fields of boulder-sized fragments, and intermediate aggregates, all of which are important habitat for commercial and recreational fish species (Dennis and Bright, 1988; Weaver et al., 2002) The invertebrate fauna of the MCE is also diverse and abundant, with octocorals, hermatypic corals, sponges, and antipatharians numbering at least 40 taxa (Continental Shelf Associates and University, 2001).

In 2010, more than 4.1 million barrels of crude oil were discharged into the waters of the northeast Gulf of Mexico following loss of well control in the Macondo Prospect exploration site and the sinking of the Deepwater Horizon (DWH) drillship (McNutt et al., 2012). In addition, at least 500 000 t of hydrocarbon gas were released into the water column during the 87 d event (Joye et al., 2011a). Potential exposure of the MCE in the Pinnacles Reefs and a portion of the Florida Middle Ground occurred via oil slicks that covered the region for approximately 35 d (MacDonald et al., 2015; Valentine et al., 2014). Response operations treated the floating oil with repeated 32 aerial application of dispersants and by corralling the floating oil and burning it (Rufe et al. 2011; and see NOAA ERMA Deepwater Gulf Response: http:// gomex.erma.noaa.gov/). Concerns for injury to octocorals and other sessile organisms of the MCE arise from organismal exposure to surface oil, dispersant treatments, and soot from burning. DeLeo et al. (2016) demonstrated that the exposure of deep-sea corals to different combinations of oil and dispersant produce a decline in coral health. Several studies found frequent injuries in octocorals and antipatharians in mesophotic reefs within the impacted area (Etnoyer et al. 2016, Silva et al. 2016). The pathology of these injuries was similar to what has been described in corals from deep-sea habitats that were known to be impacted by oil from the DWH discharge (White et al. 2012, Hsing et al. 2013, Fisher et al. 2014). Thus, there is well-founded concern regarding the health of regionally significant portions of the Gulf of Mexico’s MCE and a need to better define the extent of this resource.

While vulnerable marine ecosystems do occur at the scale of ocean basins, conservation efforts are generally undertaken within national jurisdictions where the national legal framework supports the establishment of marine protected areas (Elliott & Crowder 2005). Anthropogenic pressure on marine ecosystems, coastal areas, and the deep sea has been increasing (Davies et al. 2007). One of the biggest challenges in design and management of marine protected areas is estimating the spatial extent of habitats that can support taxa of concern (Rogers et al. 2007). This need increases in the case of corals because the composition of associated reef communities can change over relatively short distances (Rogers 1999). Determining the extent of ecosystem injury and verifying the efficacy of recovery strategies would benefit from a more comprehensive appraisal of the available habitat for mesophotic corals.

Habitat suitability models can be used to produce maps that predict continuous-coverage habitat in data-sparse areas, including the potential MCE area of the Gulf of Mexico. These models combine taxon presence/absence data with relevant environmental variables to statistically predict the distribution of species and to impart better understanding of which environmental factors are most important for the survival of species of interest (Guisan & Zimmermann 2000). Rengstorf et al. (2013) reviewed some of the approaches to this problem used by previous studies and cited the following methods: multidimensional envelopes of environmental factors (e.g. BIOCLIM), parametric regression and non-parametric smoothing procedures (Guisan et al. 2002), algorithms

33 based on training sets, and hybrid combinations of these methods (Elith et al. 2006, 2011, Elith & Leathwick 2009). Data that verify the non-occurrence of modeled taxa in a region of interest, i.e. absence data, are problematic for museum collections and for surveys that could not comprehensively cover large regions; the alternative is to simulate absence through use of randomly selected background points (Pearce & Boyce 2006). The literature includes several studies where machine learning methods such as Maxent (Phillips et al. 2006) have performed well for presence-only datasets (Elith et al. 2006, Rengstorf et al. 2013, Georgian et al. 2014). Accordingly, this approach was used in the present study, which also provides a comparison with Maxent habitat models developed for deep-sea corals.

The aim of this study was to characterize and quantify the habitat suitability niches for the most conspicuous octocorals and antipatharian corals within the 30−100 m isobaths in relation to the available substrata, bottom topography, and oceanographic parameters situated between Mississippi and Florida. It builds upon a broad array of available data. Results can be used to formulate new strategies of restoration and conservation, and to inform management of mesophotic corals in the Gulf of Mexico.

3.2 Methods

3.2.1 Study Site

The present coral habitat suitability model was compiled within a region defined by the extent of a USGS bathymetric mapping program, which covers an area of 6228 km 2 and encompasses the following areas (Gardner et al. 2001b, 2002, 2003): (1) the Pinnacle Trend located between Mississippi and Alabama; (2) the Head of Desoto Canyon offshore of Panama City, Florida; and (3) The West Florida Shelf (Fig. 11). The MCE in this area comprises multiple structures dating from Quaternary carbonates to Holocene sandstones (Locker et al. 2010). The area has been previously investigated during the multidisciplinary ‘Mississippi-Alabama Pinnacle Trend Ecosystem Monitoring’ (MAPTEM) program (Brooks & Giammona 1991, Continental Shelf Associates & Texas A&M University 2001).

Figure 11. Extent of the area mapped by the US Geological Survey (Gardner et al. 2001b, 2002, 2003), showing the region of potential mesophotic habitats between Mississippi and Florida analyzed in the coral habitat suitability model. (A) Alabama−Mississippi Pinnacle Trend Alabama Alps Reef area; (B) Alabama−Mississippi Pinnacle Trend Roughtongue Reef area; (C) Head of Desoto Canyon; (D)Florida Middle Grounds Coral Trees area; and (E) Florida Middle Grounds Madison Swanson Reef area.

35 3.2.2 Survey

Photographic survey records of sea fans and sea whips were collected during 3 consecutive years (1997−1999) by the MAPTEM report (Continental Shelf Associates & Texas A&M University 2001) and Peccini & MacDonald (2008); they provide a baseline of the marine resources of the region, including the gross anatomical condition of mesophotic sea fans and sea whips in the Pinnacle Trend before the DWH oil spill. In the case of sites that were not visited during the MAPTEM program, the presence data were augmented with records from a survey expedition completed in 2014 (2014 Natural Resource Damage Assessment [NRDA] Mesophotic Expedition, see Fig. S1 in the Supplement data

The MAPTEM survey efforts were accomplished using a SeaRover remotely operated vehicle (ROV) outfitted with a forward-facing still camera oriented at a 45° angle and illuminated by a 150 Ws strobe. The 2014 NRDA expedition utilized an ROV (‘Global Explorer’) equipped with a similarly oriented digital camera and LED lamps. The survey sites were located on several carbonate structures and the adjacent seabed within 2 main sites: Alabama Alps Reef (AAR) and Roughtongue Reef (RTR) (from the MAPTEM collection); and Coral Trees Reef (CTR) and Madison Swanson Reef (MSSR) (from the 2014 NRDA Mesophotic Expedition). Following the method ology described by Peccini & MacDonald (2008), each survey examined a circular region, nominally 200 m diameter, within which approximately 100 randomly chosen points were photographed in order to document the sessile community and local geomorphology. Although the camera systems differed, the survey methodology utilized in the 1997−1999 surveys was replicated in the 2014 survey. In total, 1975 photographic records of sea fans and sea whips were used to construct a presence database, from which the model was compiled. We developed morphological descriptions for 7 coral taxa (Fig. 12), which were known to be among the endemic taxa of the region (Etnoyer et al. 2016). This made it possible to classify potential habitat preferences by taxon .

Figure 12. Octocoral and antipatharian colonies found in the mesophotic coral ecosystem (MCE) of the study area: (A) Hypnogorgia pendula, (B) Swiftia exserta, (C) Stichopathes sp., (D) Ellisella sp., (E) Thesea cf. nivea , (F) Thesea cf. rubra, (G) Placogorgia sp., (H) Placogorgia cf. tenuis, (I) Bebryce sp., (J) Nicella sp., (K) Antipathes cf. atlantica

3.2.3 Environmental Data

We considered 32 environmental variables to build the model and test for factors that were significant in the observed patterns of coral presence. All variables were selected based on environmental factors thought to influence coral settlement, feeding, growth, and survival (Table 1). These variables include bottom current data because octocorals in the Pinnacle Reefs have been shown to grow so that their fans are oriented to maximize exposure to local currents (Peccini & MacDonald 2008). Bathymetric data were obtained from the USGS database (Gardner et al. 2001b, 2002, 2003) and gridded at their native resolution (8 m) using ArcGIS. Slope, curvature, and surface rugosity were calculated from the bathymetry layer using the Benthic Terrain Modeler Tool (Wright 2012) in ArcGIS. Slope was measured in degrees using the 8-cell method (3 × 3 cell neighborhood; Burrough & McDonell 1998). Surface rugosity, sometimes referred to as surface roughness (Hobson 1972), measures the topographical complexity of the seabed; areas with greater

37 complexity often exhibit higher levels of diversity (Kostylev et al. 2001). Finally, we calculated the topographic position index (TPI) (Weiss 2001) using the ArcGIS Land Facet Corridor Designer Tool (Jenness 2006). TPI quantifies the elevation of points relative to surrounding features and is considered a measure of preference for exposure to topographically intensified currents (Wilson et al. 2007). TPI layers were calculated at increasing window sizes (10, 50, 100, 250, 500, and 750 m) in order to consider a range of potentially important geomorphologies within the study area. Seafloor locations containing hard substrata were derived from the backscatter raster and processed using the iso-cluster unsupervised classification from the Image Processing Toolbox, in ArcGIS. Four sediment classes were designated based on photograph records of the seafloor in the area: fine sediment areas, coarse-sand areas, gravel sandy rocky areas, and exposed rock. The habitat classification layer was obtained using the Benthic Terrain Modeler Tool (Wright 2012), which classifies a given location as belonging to 1 of the following 7 categories: (1) high reef or ridge, (2) reef base, (3) low reef or ridge, (4) reef or ridge slope, (5) steep slope, (6) gentle slope, and (7) deep depression. In addition, dominant bottom sediment types and seabed sediment Folk code variables were obtained from the Gulf of Mexico Data Atlas (Jenkins 2011a,b) to further characterize potential habitat sites.

3.2.4 Dynamic Environmental Variables

All nutrients (phosphate, nitrate, silicate, dissolved oxygen, oxygen saturation, and apparent oxygen utilization) were obtained from the Gulf of Mexico Data Atlas, and normalized on a 5 × 5 km grid (Garcia et al. 2010a,b). These data were imported into ArcGIS; they were then scaled and re-projected to match the finer resolution and geographic projection of the bathymetry, but retained the source binning. Water column and atmospheric factors showed significant variations across the study area. To include model potential effects, synoptic datasets of sea surface temperature (SST day/night average), chlorophyll a (chl a), diffuse attenuation coefficient, fluorescence, photosynthetically active radiation (PAR), and chromatic dissolved organic material (CDOM) were acquired from the ERDDAP database (NOAA CoastWatch; http:// coastwatch.noaa.gov). Water column temperature, salinity, and bottom currents (u, v, and w over the 30−60 m depth range) were obtained from the Hybrid Coordinate Ocean Model (HYCOM, https://hycom.org; Dukhovskoy et al. 2015). All oceanographic data were downloaded as monthly (satellite data) or dai (HYCOM) files in NetCDF format (*.nc), which were standardized in 38 Table 7. Summary of the environmental variables used to develop the model. See Table S1 in the Supplement for statistics of each variable. Variables shown in bold were selected after Spearman correlation analysis (Table S2 in the Supplement) to perform a simplified model with selected variables. TPI: topographic position index (subscripts indicate grid cell size); SST: sea surface temperature; CDOM: chromophoric dissolved organic matter; PAR: photosynthetically active radiation; dV: digital value; na: not applicable

39 MatLab. Historical hurricane data over the study area were obtained from NOAA (Knapp et al. 2010) and displayed in ArcGIS in order to create a hurricane wind intensity raster, weighting wind intensity as the primary factor. All atmospheric and oceanographic layers were scaled, resampled, and re-projected to match with the bathymetric data.

3.2.5 Modeling

Maximum entropy modeling routines of Maxent were used to evaluate the extent of habitat suitable for mesophotic sea fans and sea whips in the eastern Gulf of Mexico (Phillips et al. 2006). The method was chosen after consideration of the available occurrence data for MCE organisms. Some methods for modeling habitat suitability require both presence and absence data. When reliable absence data are available, presence/absence methods may perform better than presence-only models (Reiss et al. 2011, Rengstorf et al. 2013). However, reliable absence data are rarely available, or are poorly controlled. In such cases, presence-only models like Maxent have been shown to be more robust and consistent (Elith et al. 2006, Reiss et al. 2011, Yesson et al. 2012) because they utilize pseudo-absence (background) data rather than true absence data and have consistently outperformed other presence-only techniques (Elith et al. 2006, Elith & Leathwick 2009, Tittensor et al. 2009, Tong et al. 2013). Maxent assigns non- negative probability values to each background pixel of the study area such that their total sums to 1. Furthermore, presence-only modeling results have produced results consistent with traditional presence/absence methods in shallow corals (Bridge et al. 2012, Couce et al. 2013) and deep-water corals (Davies & Guinotte 2011, Howell et al. 2011, Tracey et al. 2011, Yesson et al. 2012, Rengstorf et al. 2013, Taylor et al. 2013).

Models were created using the default Maxent parameters that have been shown to optimize model performance (i.e. convergent threshold = 10−5, number of background points = 10 000, default prevalence = 0.50; see Phillips & Dudik 2008). In preliminary trials, the model was tested using different regularization multipliers (β = 1, 3, 5, 7, 11, and 13; default = 1), following procedures proposed by Georgian et al. (2014). However, after inspection, the best performance was obtained in default mode. The number of maximum iterations was increased to 5000 to ensure convergence. Pearson’s rank correlation was calculated for all of the environmental layers following Yesson et al. (2012) This identified uncorrelated variables within each group and 40 produced a simplified set of inputs, which avoided over-specification of the model (see Table S2 in the Supplement).

A jackknife procedure was employed to calculate the percent contribution of variables to each model. Because Maxent is robust regarding auto-correlated inputs (Phillips et al. 2006, Elith et al. 2011), 2 models were constructed. One model used all available variables, while an alternative model removed autocorrelated variables prior to analysis. This avoided biasing the results based on preconceived perceptions of variable importance.

A cross-validation routine was used to verify model performance. Data were randomly partitioned so that 75% of occurrences were used for calibrating and running the model, and 25% were used to evaluate model performance. As the model was constructed, the test gain, which is the improvement in the model performance compared with random, was used to assess the contribution of the individual variables and the entire variable set. The model was performed with 50 replications (number of runs), and averaged for the final result. The receiver operating characteristic measures performance of the model at any threshold using the area under the curve (AUC) value, which ranges from 0 (worse than random model) to 1 (ideal model), and includes a random prediction of 0.5 (Phillips et al. 2006). Target AUC values of 0.9 were surpassed by the final models. Sampling bias was not anticipated in the data due to the randomization of photo collection. To verify the habitat prediction, we evaluated results of the 2 models created using the online coral-presence dataset (NOAA National Database for Deep-Sea Corals and Sponges, https://deepseacoraldata.noaa.gov) , which compiles various coral presence records from the study area (see Fig. S2 in the Supplement)

3.3 Results

The single or ‘general’ model of sea fan and sea whip distribution of the potential mesophotic ecosystem encompassed the entire 6228 km2 area mapped in the USGS bathymetric survey. A visual inspection of the suitable habitat of mesophotic corals revealed that most of the predictions of where coral may occur were located within well-known mesophotic reefs such as the AAR, RTR, CTR, and MSSR, but also extended to several tall (> 7 m relief), intermediate (3−7 m relief), and, to a lesser degree, small (< 3 m relief) carbonate mounds, ridges, and paleo-shoreline 41 structures adjacent to the main reefs. It also appears that suitable habitat for mesophotic corals occurs in some flat zones (Fig. 13)

3.3.1 Model Performance

The general model performance significantly exceeded the random model prediction (AUC > 0.5). The performance of the model configured with all variables (test/training AUC): mean ± SD 0.978 ± 7 × 10−5/0.975 ± 8 × 10−4) was marginally better than the model configured with selected variables (AUC 0.963 ± 4 × 10−5/0.960 ± 0.001) (Table 2). Overall, the jackknife test of importance indicated that surface rugosity was the best predictor to explain the distribution of mesophotic sea fans and sea whips in results for both models (Table 2), with 19.9 and 19.1% contribution to the all-variables models, versus selected-variables models, respectively. Bottom eastward currents (13.1%), hurricane winds (11.1%), CDOM (17.4%), and dominant sediments (16.1%) were secondary contributors to predict habitat suitability for both models (all and selected variables), respectively (Table 2, Fig. S3 in the Supplement). Within the general model for all variables, slightly positive values on the u component of bottom currents (0 to 0.25 m s−1) provided favorable conditions for octocorals and antipatharians in the study area. Additionally, locations with CDOM values around 39.1 would host high densities of octocorals and antipatharians (Table 2). Other good contributors for both models were depth, bottom- upward currents (w), bottom salinity and temperature, slope, and nitrate concentrations.

Performance of each environmental variable is represented in the response curves (Fig. 4), which show the response of the suitability index plotted against change in the variable. Some variables showed a better response to the logistic prediction when the model takes advantage of all environmental variables available to build the habitat prediction with the model. This was the case for PAR, chl a, CDOM, SST, oxygen saturation, slope, curvature, and loose and dominant sediment fractions, all of which produced improved model performance when they interact with all variables. In contrast, bottom currents (u, v, and w), bottom salinity and temperature, back scatter (unsupervised isoclassification), and sediment type showed better individual performance in the marginal response curves (Fig. 14).

42 3.3.2 Taxon-specific Model Results

Similar to the general model, suitability models that were independently evaluated for all coral taxa showed a good performance reflected in AUC values for training and test statistics. All test AUC values were > 0.97 with the exception of Ellisella sp. (mean ± SD test AUC: 0.876 ± 0.111), and all test gain values were over 3.2, with the same excep tion for Ellisella sp. (test gain: 2.1). All by-taxon models significantly outperformed the random model (Table 2). When the habitat suitability model was applied independently to all coral taxa, the same environmental factors played fundamental roles in affecting species or taxon distribution, but with a different order of contribution (Table 2). For the antipatharians Antipathes atlantica and Stichopathes sp., and the octocorals Swiftia exserta, Bebryce spp., Hypnogorgia pendula, and Placogorgia sp., bottom u was the most important environmental factor (>14.3%). For Ellisella sp., rugosity (34.3%) was the most relevant factor; for Thesea nivea and Nicella sp., loose sediment (29.3 and 21.2%) was the most important environmental variable in the model (Table 2; Fig. S3).

Likewise, other environmental variables contributed to different degrees to the taxon’s suitability model. These include bottom salinity, bathymetry, water temperature (surface and bottom), dominant and loose sediment fractions, current components (u, v, w), and TPI. Predictions of potential habitat area indicated that Ellisella sp. had the largest potential area (660 km2) for the study area, followed by Bebryce sp. (186 km2), and S. exserta (168 km2). H. pendula, which is among the most conspicuous coral taxa in the MCE, was predicted to occur across only 80.9 km2 within the study area (Table 2). With regard to geomorphology of predicted habitat for mesophotic ocotocorals and antipatharians, the following preferences were indicated for all coral taxa and the general model: small and tall carbonate mounds, with high preference for reef tops and slopes, and adjacent hard substrata with small fractions of fine sediments. There was also high preference for areas with negative w current (down-welling) and moderate u (eastward) and v (north ward) current, high CDOM, and relatively low bottom temperature for the area (18°C; Fig. 3; Fig. S4 in the Supplement)

Figure 13. General model of habitat suitability for all mesophotic corals in the study area using the selected variables and the MAPTEM data set. Warm colored areas (yellow to red) show the predicted locations where corals are likely to be found (high habitat suitability), while cold colored areas (dark green to blue) indicate a low probability of finding mesophotic corals (low habitat suitability). Habitat areas are listed in Fig. 1. See complementary Fig. 27 for model build with the NOAA data set

44 Table 8. Summary statistics (mean ± SD) of training and test models for area under the curve (AUC) and gain, suitable area (% of the total, and total in km2), and jackknife test of variables used in each model. Variable abbreviations are as follows: Rug: rugosity; B_u: bottom u (eastward velocity); Hurr: hurricane winds; CDOM: chromophoric dissolved organic matter; DoSed: dominant sediments; TPI 500: topographic position index at 500 m; OS: oxygen saturation; SST: sea surface temperature; LoSed: loose sediment; UnIso: backscatter unsupervised isoclassification; Var: variables; All: all variables; Sel: selected variables .

Var Sample Training Test Suitable habitat JackKnife test of importance Species/Group size Gain AUC Gain AUC (%) (km²) Variable 1 Variable 2 Variable 3 All corals MAPTEM All 865 2.4 ±0.004 0.963 ±7E-05 2.5 ±0.02 0.963 ±8E-04 0.065 404.81 Rugg. (19.8%) B_u (13.2%) Hurr. (11.6%) All corals MAPTEM Sel 865 2.3 ±0.006 0.960 ±4E-05 2.4 ±0.03 0.959 ±0.002 0.072 448.40 Rugg. (18.4%) CDOM (17.6%) DoSed (16.7%) All corals Test NOAA All 185 3.1 ±0.037 0.983 ±0.001 3.1 ±0.51 0.960 ±0.018 0.018 112.10 Rugg. (31.0%) TPI 500 (28.6%) OS (6.6%) All corals Test NOAA Sel 185 3.1 ±0.027 0.981 ±0.001 3.1 ±0.38 0.958 ±0.027 0.019 118.33 Rugg. (32.2%) TPI 500 (27.3%) OS (9.2%) Antipathes atlantica All 230 3.2 ±0.017 0.986 ±0.001 3.2 ±0.11 0.980 ±0.005 0.021 130.78 B_u (14.3%) STT (13.7%) Depth (9.8%) Bebryce spp. All 345 3.1 ±0.015 0.982 ±4E-04 3.2 ±0.12 0.980 ±0.005 0.030 186.83 B_u (25.7%) CDOM. (14.6%) TPI 500 (13.9%) Ellisella sp. All 19 2.9 ±0.346 0.974 ±0.011 2.1 ±1.85 0.876 ±0.111 0.106 660.15 Rugg. (34.3%) LoSed (16.8%) UnIso. (13.9%) Hypnogorgia pendula All 144 3.7 ±0.009 0.993 ±1E-16 3.8 ±0.25 0.992 ±0.001 0.013 80.96 B_u (17.8%) Rugg. (16.9%) CDOM (15.1%) Nicella sp. All 8.8 2.9 ±0.488 0.995 ±0.002 3.7 ±0.18 0.972 ±0.037 0.023 143.24 MoSed (21.2%) DoSed (21.1%) Nitrate (15.8%) Placogorgia sp. All 138 3.8 ±0.017 0.993 ±1E-04 4.0 ±0.17 0.992 ±0.001 0.012 74.73 B_u (20.2%) Rugg. (13.5%) TPI 500 (12.6%) Stichopathes sp. All 206 3.4 ±0.025 0.989 ±4E-04 3.5 ±3.22 0.987 ±0.005 0.019 118.33 B_u (20.4%) CDOM (13.5%) Rugg. (12.8%) Swiftia exserta All 312 3.2 ±0.007 0.984 ±4E-05 3.3 ±0.02 0.984 ±7E-04 0.027 168.15 B_u (33.0%) CDOM (18.0%) Rugg. (17.4%) Thesea nivea All 64 3.8 ±0.069 0.992 ±0.002 4.1 ±0.50 0.989 ±0.014 0.007 43.59 LoSed (29.3%) Rugg.(21.9%) CDOM (10.4%)

Figure 14. Response curves of the most important variables showing the relationship between performance of predictors and the‘general’ habitat suitability model for all mesophotic octocorals and black corals in the study area. Black lines indicate how the model prediction changes as the environmental variable varies taking all other variables into consideration. Grey curves characterize the model’s response using only that variable

3.3.3 Independent Test Models

The suitability model created using an independent set of mesophotic coral records (NOAA) performed similarly to the general models that used all variable and selected-variable approaches. The AUC value for the test model based on all variables was 0.96, versus 0.95 for the test model generated with selected variables; both exceeded our target significance level of 0.90. The total extension of the suitable area for these models was less than the general models (>118 45 km2). A jackknife test of importance of the variables showed that both test models were prmarily driven by the same variables: rugosity (31.0 and 32.2%); TPI500 m (28.6 and 27.3%), and oxygen saturation (6.6 and 9.2%) (Table 2; Fig. S5 in the Supplement).

3.4 Discussion

The results obtained from suitability models for mesophotic corals in the Eastern Gulf of Mexico indicate that there are 405 km2 of habitat suitable for mesophotic octocorals and antipatharians within that portion of the continental shelf between Mississippi and Florida for which detailed bathymetric mapping has been completed. Previous exploration efforts have focused on 2 well-known reefs: AAR and RTR, which are sites that have been explored over the past 2 decades (Peccini & MacDonald 2008, Etnoyer et al. 2016, Silva et al. 2016). We used a series of random photographic samples of those 2 locations from the MAPTEM program and expanded to 2 other wellknown reefs in the Florida Middle Ground (MSSR and CTR) randomly sampled in 2014 to construct alltaxa and by-taxon suitability models for octocorals and antipatharians. For model validation, we used an independent set of coral records compiled by the NOAA Deep Coral Data Portal. Although the results confirm the suitable habitat for large and high carbonate structures (e.g. AAR and RTR), the results also suggest that mesophotic octocorals and antipatharians will also be found on many other medium sized and small carbonate mounds, high reflectivity platforms, and paleo-shoreline structures like small ridges (Gardner et al. 2001a). The resulting suitable niche prediction for all models is located between 100 and 50 m depth. A similar depth range was used by Locker et al. (2010) and Bridge et al. (2012) to report areas with the potential to host MCEs in waters of the Gulf of Mexico and the mesophotic area in the Great Barrier Reef in Australia, respectively. Therefore, other factors besides depth range may play an important role in the formation of MCEs in the Gulf of Mexico.

High suitability indices from the general model, using all and selected variables, were driven primarily by geomorphological complexity of the terrain, expressed in several bathymetry- derived factors, principally surface rugosity and, to a lesser degree, TPI. In this case, high suitability indices are associated with high rugosity values, which are a measure of the complexity (‘bumpiness’) of the seafloor. This relationship be tween rugosity and coral diversity, especially hermatypic corals, has been well documented for shallow-water corals (Zainul Fuad 2010, Dustan 46 et al. 2013) and deep reef assemblages (Howell et al. 2011, Rengstorf et al. 2013). High surface rugosity values in this area cannot be attributed to reef formation by corals in the present. Rather, these carbonate structures are evidence of relict shallow reef habitat, which provide hard substrata suitable for colonization by sessile organisms such as corals adapted to the 30−100 m depths (Locker et al. 2010, Donoghue 2011, Reich et al. 2013). Model results displayed some binning in areas where fine-scale variables, such as surface rugosity, were relatively monotonic (Fig. 3). Under these conditions, the climatic variables, which were gridded at a 5 × 5 km scale, tended to dominate. Setting higher display thresholds might reduce this apparent effect; however, it is not entirely an artifact because it might inform habitat potential for shipwrecks or other artificial structures located in such regions.

High suitability indices were further correlated with positive indices in TPI500 m, especially for the model formulated from the NOAA data set. High positive values of TPI500 m represent topography elevated above surrounding areas. In the by-taxon models, TPI500 m did make a relevant contribution to habitat suitability (third place), particularly for sea fans (e.g. Bebryce sp. and Placogorgia spp.). For other taxa, TPI played a less important role in each model. This was the case for Hypnogorgia pendula, which also prefers high negative and high positive values of TPI50 m (fine TPI). A possible explanation is that finescale TPI detects smaller features, such as crests, depressions, small mounds, and reef-tops, which some taxa prefer. Large-scale TPI identifies larger features like ridges, mounds, flat areas, and slopes. Therefore, when suitability models predict potential habitat in areas characterized by positive values in a broad TPI raster, the models are predicting that suitable areas for corals will occur in larger structures that are elevated in comparison with their surroundings. In contrast, when suitable habitats are predicted in high or negative zones, this indicates that a specific taxon may prefer to settle in a par ticular zone (e.g. crest, depression) within a major structure (e.g. ridge). Among the taxa modeled individually, Bebryce were characterized by abundant, but small, colonies that were frequently observed among loose rubble and sediment. Taxon-specific preferences for TPI may be aliased with preference for additional variables as well.

In the general model, the large-scale TPI raster contributes more to the model because it includes all coral records in the model. However, when the model was constructed by taxon, some

47 of the corals preferred flat tops or reef crests (e.g. H. pendula), while others would also be found in depressions and slopes within the reef (e.g. Antipathes atlantica ). This situation is clearer in the habitat classification layer, which is derived from bathymetry, slope, surface rugosity, and fine and broad TPI layers. A possible explanation of coral preference for high grounds is that they can take advantage of currents, which transport nutrients, oxygen, and food (Thiem et al. 2006, Peccini & MacDonald 2008), while also reducing sediment deposition or re-suspension (C. Rogers 1990, A. Rogers 1999). For almost all individual suitability models (by taxon), currents, especially the eastward component (u), were the major contributor to suitable habitat in the study area (Table 2). In addition, coral-suitable areas were correlated with slightly negative values of vertical speed currents (w), and low values of northward current (v) for lateral transport. This situation has been documented by Peccini & MacDonald (2008), who demonstrated that octocorals in the Pinnacle Reefs area grow with their fans permanently oriented to local currents to maximize exposure to lateral transport and suspended material. Mesophotic octocorals and antipatharians are predominantly azooxanthellate corals and heterotrophic suspension feeders that depend on both plankton and detritus provided by local currents for nutrition (Bayer 1961, Fabricius & De’ath 2008).

Primary productivity in the water column, expressed as chl a, was mostly irrelevant in all models. This was also reported by Bridge et al. (2012), who identified high suitability areas for mesophotic corals in areas of the Australian Great Reef Barrier where chl a is low. In this study, seafloor chl a concentration or another measurement of food availability might be a better predictor of suitable areas for heterotrophic corals. However, Georgian et al. (2014) used ‘export productivity’ to predict a suitable niche for Lophelia pertusa in waters of the Gulf of Mexico. In the same vein, suitable areas for corals were characterized by fairly high values of CDOM, which corresponds to the dissolved organic carbon fraction that absorbs light (Rochelle-Newall & Fisher 2002). This applied especially to Bebryce sp., H. pendula , Placogorgia sp., and Stichopathes sp.

In the suitability model for the remainder of the coral taxa in this study, as well as in the general model, the CDOM variable was the second in importance. It was third in importance for the independent test model. Although CDOM cannot always be interpreted as an indication of particulate organic matter (POM), it could potentially be used as an indication of the food source

48 for heterotrophic organisms. Evidence suggests that CDOM values are proportional to the POM concentration in highly productive waters due to the transformation in phase from particulate to dissolved fraction, which is mediated by physical, chemical, and biological processes (Stedmon & Markager 2001). In the northern Gulf of Mexico, particulate organic carbon supply to the seafloor is mediated by lateral currents and downwelling (vertical transportation) of detritus and marine snow, especially in areas close to the Mississippi Delta (Rowe et al. 2008), which could be a source of food for mesophotic corals. Nevertheless, the proximity to the Mississippi River does not just provide a source of potential food; eutrophication and oxygen values can also be affected by runoff water, which potentially affects settlement and survival rates of mesophotic coral. All nutrient variables (nitrate, phosphate, and silicate) were highly correlated with each other, with oxygen (dissolved oxygen, oxygen saturation, and oxygen utilization), and with salinity (Table S1). However, the influence of the river and these variables is most pronounced in the area near the Mississippi Delta (west of AAR), while reduced towards Florida (Garcia et al. 2010b). High suitability areas are almost absent approaching the Mississippi River, where high nutrient concentrations (e.g. nitrate) values are found. Several studies have attributed high rates of coral diseases to highly eutrophic waters (Hallock & Schlager 1986, Bruno et al. 2003). High nutrient concentrations increase the rate of epizootics on octocorals, and may be associated with decreased autotrophic function in zooxanthellate hermatypic corals due to turbidity. Although the diffuse attenuation coefficient and fluorescence values increase with proximity to the Mississippi River, while PAR decreases, these variables do not seem to be major factors in either the general suitability model or the taxon-specific models. While it is evident that sea fan and sea whip corals need reasonable oxygen concentrations for their survival, oxygen saturation contributed in a meaningful way only to the models generated from the independent set of coral records from NOAA. Although dissolved oxygen and oxygen saturation decrease inproximity to the Mississippi River (because oxygen consumption increases due to biological and chemical processes), the values are still sufficient to support MCEs to the east of the Mississippi River,despite the nutrient loading.

Two of the most important environmental variables were bottom reflectivity and sediment type. High values of bottom reflectivity are usually interpreted as hard substrate available for colonization by sessile organisms such as corals. However, sand and gravel deposits can also return

49 high values in the backscatter image. Furthermore, high backscatter values can also be associated with biogeneous substrata (e.g. carbonates) or hard seabed covered by fine sediments, which are not always suitable for colonization due to biogeochemical process (methane release/oxidation). For example, Tseng et al. (2011) demonstrated that high concentrations of suspended sediments negatively affect feeding behavior, colony expansion, and mucus formation by octocorals. In our models, we used raw backscatter values and a backscatter-derived raster called un supervised isoclassification. This technique separates pixel intensity values into classes, which subsequently can be associated with previously identified sediment types. This seems to be a better approach than binary interpretation (e.g. soft vs. hard) commonly used in this type of analysis. In our general model, suitable areas for mesophotic corals are correlated with gravel and exposed carbonate classes (see Fig. 4, response curves). If we look independently by taxa, some octocorals (e.g. H. pendula) will prefer areas of exposed rock as dominant substrate, whereas other taxa (e.g. Stichopathes sp. and A. atlantica ) seem to prefer softer substrata (e.g. gravel with sandy fractions). All models (general, by-taxon, and test) failed to identify numerous smaller reefs that we a priori expected would be part of the suitable habitat for mesophotic octocorals and antipatharians. However, when the predicted suitable area is compared with new data from an exploration conducted in 2014 (data under analysis), the new records within the Pinnacle Reef area are consistent with the habitat suitability prediction made by our model. A drawback of heavily weighting habitat suitability predictions on such widely used factors as substrate type can be a tendency to over-predict the area of habitat. By using a variety of environmental factors, the models developed for the present work incorporate more marginal differences that may favor individual taxa or influence development of a mesophotic coral community.

We acknowledge that habitat suitability models are highly dependent on the quantity and quality of taxon-presence records for robust habitat prediction, as well as the resolution of the environmental layers. The differences between the general and test models were not the result of the number of environmental variables used to construct the models. This is reflected in the stability of the model on the importance assigned to each variable (jackknife test) in the model. Surface rugosity was always the most important contributor to the models. Although there was some difference in the order of the following variables, the main difference was their contribution percentage, slightly changing the order of the variables; however, the same variables still played a

50 relevant role in the model. This fact was more evident in the test model, where the order of the 3 most important contributors to the models built with all or with only selected variables, was always the same (rugosity, TPI500 m, and oxygen saturation), varying only by the respective percentage of their contribution to the model. Other statistics like AUC and Gain values were almost identical between models using selected or all environmental variables. The main difference between models is indicated by the predicted area (10th percentile) between the general and test models. The general model predicted a suitable habitat area of 405 km2, while the suitable habitat predicted by the test model did not exceed 118 km2 (Table 2). This difference can be attributed to source and number of records for each data set. While the MAPTEM/ 2014 dataset relies on a large number of samples (865 records) obtained from 4 well explored reefs (AAR, RTR, CTR, and MSSR), the NOAA dataset is composed of only 185 coral records distributed in the whole study area (6227.8 km2). This reveals how the algorithm depends upon the number of samples. Additionally, the number of records, distribution, and reliability of the presence data are relevant when we are trying to avoid over-prediction of suitable habitat in the model. It is important to use a set of records which has a high density of samples within similar values of the environmental variables rather than have presence records in a broader range of values within each variable. That approach could lead to over-estimation of the real extent of the suitable habitat for a given taxon. Therefore, it is imperative to continue to build habitat suitability models that include updated presence records and new and reliable environmental variables. This will allow us to increase the biogeographical extent of a given taxon or community in order to protect and eventually quantify the extent of corals susceptible to harm by natural or anthropogenic disasters.

3.5 Conclusions

We proposed 8 habitat suitability models (1 general and 7 by taxon) for mesophotic octocorals and antipatharians to predict their occurrence throughout a potential area located in the Northern Gulf of Mexico between Mississippi and Florida. All models significantly out-performed a randomized prediction of suitable areas, and also indicated that the distribution of octocorals and antipatharians is predominantly driven by surface rugosity of the substrata. Other variables like TPI, CDOM, currents, sediments, and hurricane wind density, among others, could influence the occurrence of these mesophotic corals. Our results differed from other coldwater coral suitability models with respect to which predictor factors played a more relevant role in the model, and, for 51 instance, the extent of the suitable area. For reef- forming corals like Lophelia pertusa, bathymetry and substrate type are the most important pre dictive variables (Rengstorf et al. 2013, Georgian et al. 2014); in contrast, deep-sea octocoral suitability models are more sensitive to oxygen saturation and calcite concentration.

The habitat suitability maps provided in this study can be used as a baseline for future exploration in the area in order to verify the true extent of MCEs in the eastern Gulf of Mexico, and in turn, to provide an accurate estimation of MCEs that could be impacted by natural (heat stress, hurricanes) or anthropogenic (oil spills, eutrophication, ocean acidification) events for resource management planning and marine conservation strategies for MCEs in the Gulf of Mexico. In addition, this study improves the basis for evaluating the potential extent of injured octocorals and antipatharians resulting from the DWH oil discharge

CHAPTER 4 SUBMARINE CHANNELS, MASS MOVEMENTS, AND OIL INDUSTRY: A POTENTIAL HAZARD FOR THE MESOPHOTIC CORAL ECOSYSTEM (MCE) IN EASTERN GULF OF MEXICO.

4.1 Introduction

Submarine canyon and channels are sedimentary structures formed by hydrodynamic erosion of the seafloor (Miall 2000). These structures are the result of various processes in which water flow erodes, transports, and deposits sediments. These processes can be the result of changes on the sea level, geological history, high-energy processes mediated by water (e.g. storms, floods, tides, waves, and currents).

In this sense, submarine canyons and channels play a fundamental role in the atmospheric- land-ocean coupling, connecting the land and coastal environments with the deep-sea; moving, recycling and storing sediments (organic and inorganic compounds) from the continents to the ocean (Normark, 2003). At the same time, canyons and channel systems influence ocean circulation by steering currents from the continental shelf towards the deep ocean and vice versa (Galy et al. 2008)

The origin of submarine canyon-channel system is one of the most controversial and unclear topics for marine geologist and oceanographers. (Miall, 2000). However, the agreement is that the main driver on submarine channel and canyon formation directly or indirectly involves sediment movements (e.g. turbidity currents, gravity flows, submarine slides, and slumps)

Submarine mass movements events in all its forms represent the main mechanics of sediment transportation from coastal areas to deep ocean basins (Masson et al. 2006). The length of canyons and the thickness of canyons deposits range from meters to kilometers length and from centimeter to hundred of meter thick (Masson et al. 2006). They can occur in active and passive margins, in all physiographic environments form coastal areas to deep ocean zones, but are most often found in the continental rise and slope (Masson et al. 2006). 53

Triggering factors for mass movements on the continental slope are numerous and variable. However, it is frequently found that the combination of several events result in slope failure. In passive continental margins, one of the most common triggers of slope failure is high sedimentation rates mediated by high sediment supply by rivers (Casas et al., 2003). Large-scale movements like slide or slumps can eventually disintegrate and mix with water creating turbidity currents, which erodes the seafloor in its way down (Carter et al 2012).

Submarine mass movements often occurs in low gradients (<16˚), which could be indicative of excess pore pressure (Kvalstad et al. 2005, Talling et al. 2014). An increase in sediment pore pressure lowers the vertical effective stress of sediments in the slope, which reduces shear strength and undermines slope stability (Duncan and Sheaham, 2012). Although the increment on hydrostatic pressure of the water column should not make the slope more susceptible to fail, a fast and unusual increment of sea level (e.g. sea level rise by global warming) could lead to an intensification of pore pressure in low-permeable sediments (Smith et al., 2013). In the same fashion, a rapid decrease on hydrostatic pressure (e.g Glaciation, extreme low tides) could increase pore pressure due the expansion of free grass trapped into sediments (Christian et al, 1997).

High sedimentation rates, especially in high productivity areas and coastal zones influenced by rivers (*e.g. Mississippi River), where sediments present low permeability (Flemings et al 2008), facilitates the increment of sediment pore pressure due to the low capacity of sediments to dissipate interstitial fluids (Duncan & Sheaham, 2012).

Smaller submarine features in the slope and shelf break (e.g. gullies) close to areas with high sedimentation are an indication of sediment removal mediated by high sedimentation periods and currents related to sea-level fall (Alonso y Maldonado 1990; Ercilla & Casas 2012). These features can become mayor structures by enlarging their extension by erosion and the removal of accumulated sediment, thereby forming the head of channels and canyons.

To a lesser degree, areas where the continental slope begins in relatively shallow waters (<200 m) and are under the seasonal impact of hurricanes (e.g. Gulf of Mexico coast, Bea et al

54 1983) can also be susceptible to suffer slope failure due the strong currents (< 10 ms-1) generated by hurricane winds. Also, it has been proposed that dissociation of gas hydrates due global warming increased of the frequency of submarine landslides and lead to canyon formation (Maslin 2004; Owen et al. 2007). Although these two triggering factor seem to have a lower probability of generating a massive sediment displacement by mass wasting, changes of sea level and water temperature could initiate these processes.

Due the extension and amount of sediment removed by mass wasting processes, it has been suggested that submarine mass movement modify the marine landscape by creating new features, exposing hard grounds suitable to be colonized and covering others with sediment. In addition, large scale submarine landslides could play an important role in global climate by increasing atmospheric concentration of methane in areas where free gas and hydrates can be exposed to water (Kennett, 2006).

Although detection of submarine canyons in the sedimentary record is important for hydrocarbon prospecting, contemporary features are directly relevant to installation of pipelines and other submarine infrastructures because they are potential hazards (Pipper 1999; Masson et al. 2006; Barley 2009). The occurrence of these events have been primarily detected thanks to the damage caused to submarine infrastructures that lays on the over and below the sea-floor such as pipelines and telecommunication cables (Masson et al. 2006)

Because the northeastern Gulf of Mexico is an area potentially susceptible to submarine mass movement triggering mechanisms (sharp shelf break, high sedimentation periods, storms, gas hydrate accumulation), and because submarine landslides have been previously reported from the central Gulf in the Sigsbee area and Mississippi Canyon (McAdoo et al., 2000), it is important to investigate the extent and frequency of these events, especially due the fact that this region is subject to increasing levels of energy development. There is the potential for \a huge environmental impact to the marine system and human population if an exploration well structures or production platform were impacted by submarine mass movements such as the Taylor offshore platform destroyed by a submarine landslide triggered by the hurricane Ivan (Daneshgar et al., 2015). The aim of this manuscript is to characterize from a geomorphological and geological point of view a

55 series a submarine channels and submarine mass movement scars located in the continental shelf offshore Alabama and Mississippi were existing and future energy industry structures are located. In addition, we also try to elucidate whether this submarine features were originated through mass movement processes, which the factors could have triggered the mass movement and the frequency of those events. Finally, we tried to identify critical areas where submarine mass movemt events could be a hazard for fauna and human facilities placed on the seafloor of coastal and deep ocean in the Gulf of Mexico.

4.2 Methods

4.2.1 Study Site

The study area is located in the continental slope and shelf edge of northeastern Gulf of Mexico between Mississippi and Florida where it forms part of the north wall of Desoto Canyon. By definition, Desoto Canyon corresponds to a partially filled indention of the north and south Florida shelf, which were part of the South Georgia Rift Basin and the Georgia Seaway. (Citation). Desoto canyon, unlike other submarine canyons, it is gently-sloped and narrow submarine feature that limits with the continental shelf and slope of Florida and the Florida Escarpment in its southern wall. One of the main characteristics of the Desoto canyon is a series of submarine channels and mass movement scars on incised along the edge of the continental shelf and slope, which end with debris depositions in the deep basin. The area was briefly described by McAdoo et al. (2000) as a submarine landslide area, but this study mainly focused on Sigsbee Escarpment in the central Gulf, where the Mississippi and Green Canyon are regional mas wasting structures

4.2.2 Geophysical Data

Multi-beam bathymetry and sub-bottom profiler data and was collected in the continental shelf, slope, and deep basin between Mississippi and Florida, during 2000 to 2013 under a series of oceanographic expeditions (Gardner et al., 2002; Gardner et al., 2003; Gardner et al., 2001b; NOAA, 2012), and Deep-C Consortium. Primarily, multi-beam bathymetry data was collected by the United States Geological Service in the continental shelf and upper slope of Mississippi to Florida on board the R/V Ocean Surveyor, using a MB echo sounder Kongsberg EM-1002. Subsequently, multibeam data of Desoto Canyon and continental slope were collected under the 56 expedition “Gulf of Mexico 2012” on board the NOAA ship R/V Okeanos using a Kongsberg Simrad EM-302 echo sounder (30-34 KHz). (Fig. 15)

Figure 15. Multi-beam bathymetry map of the study site area, which is located Northeast Gulf in the continental slope and shelf off shore of between Mississippi and Florida.

Sub-bottom profiler (SBP) data were collected on the “Deep-C Geomorphology expedition “ (June, 2013) on board the R/V Weatherbird II (Florida Institute of Oceanography) using a high- resolution chirp sonar (EdgeTech 2400-Dw-424, 1-10 kHz) deployed on the MILET tow camera platform. A series of transects was conducted across the channels using two different surveys modes: 1) MILET kept on a standard altitude (approx. 100 -150 m above the bottom) which provides just SBP record, and 2) MILET 3m above the bottom in order to conduct a photo survey and record high resolution SBP data.

57 4.2.3 Piston Core Samples

Piston core sediments samples were collected during the “Deep-C Geomorphology cruise” (June, 2013) on board the R/V Weatherbird II, using a 3 m barrel piston core (effective penetration) with a release system that allows the piston core fall free at 10m above the seafloor for maximum penetration. Piston core locations were previously chosen in key locations order to characterize sedimentary facies in different geomorphology scenarios: inside and outside of canyons, head, mouth and fan of the channels, in mass movement scars and debris flow. In addition, a control piston core sample was taken on the Florida Escarpment, where evidence of mass movement processes were observed on the multibeam bathymetry map (Fig 16). Once the piston core was onboard, the butyrate liner with the sediment sample was retrieve, sealed, labeled, and split, and stored at 2°C in freezer onboard. Once in land, piston core samples were transported and deposited for posterior analysis to the Antarctic Marine Geology Research Facility at Florida State University.

4.2.4 Geophysical Data Analysis

Multi-beam raw data was filtered of disturbances and noise created by the vessel’s movement (e.g pitch, roll, and yaw) which creates artifacts on the seafloor using Caris HIPS and SIPS software. Afterwards, raw data was uploaded into D-Magic and FM-GT in order to be processed and gridded prior to using it on Fledermaus, which was then used to calculate slope gradient, rugosity, and geomorphologic profile on related variables. The resulting data were transformed into ArcGIS *.ascii files and uploaded to ArcMap (v.10.3.1) in order to provide a geomorphological description of the seafloor. Subbotom profiler data was filtered, processed, and plotted on Triton software (Perspective Trail version).

Figure 16. Locations of piston cores collected inside and outside of channels. Note proximity of energy platforms installed in or adjacent to potentially unstable slope regions.

4.2.5 Piston Core Analysis

Piston cores samples were in first place analyzed in the Antarctic Marine Geology Research Facility (Florida State University). Core sections of each piston core sample were split in half using an electric stationary saw. One half of the piston core sample was reserved to obtain high resolution line scan image and x-ray radiographies for core sediment characterization. The other half of the core was used to grab sediment subsamples for stable and radio carbon analysis (Chanton Lab – FSU), mercury and grain size analysis (Brooks Lab – USF). Samples sediments were chosen in concordance to core length and/or predetermined sediment fractions. Stable isotopes samples were treated with 10% HCl (carbonates removal), rinsed, freeze-dried, and grounded. Sediments fractions were analyzed for percent organic carbon (%C), percent of nitrogen (%N), δ13C and Δ14C. using a Carlo Erba elemental analyzer coupled to a Delta XP Thermo Finnigan isotope ratio mass spectrometer. Sediments for C analysis were combusted at FSU and sent as purified CO2 (water vapor and non-condensable gases were removed by cryogenic separation on a vacuum line)

59 to the National Ocean Sciences Accelerator Mass Spectrometry Facility (NOSAMS). Samples of CO2 were prepared as graphite targets and analyzed by accelerator mass spectrometry (Table 9)

Table 9. Summary of sediments samples taken from each core for each analysis. Core ID NT1000 PC-01B PC-01A PC-02B PC-03B PC-03A PC-04A PC-04B PC-06B PC-06A Sample C¹³/C¹⁴ Hg /φ C¹³/C¹⁴ Hg /φ C¹³/C¹⁴ Hg /φ C¹³/C¹⁴ Hg /φ C¹³/C¹⁴ Hg /φ C¹³/C¹⁴ Hg /φ C¹³/C¹⁴ Hg /φ C¹³/C¹⁴ Hg /φ C¹³/C¹⁴ Hg /φ C¹³/C¹⁴ Hg /φ 1-2 2-4 1-2 2-4 1-2 2-4 1-2 2-4 1-2 2-4 1-2 2-4 1-2 2-4 1-2 2-4 0-1 n.a. 1-2 2-4 7-8 8-10 6-7 n.a. 7-8 n.a. 7-8 8-10 7-8 8-10 7-8 n.a. 7-8 8-10 4-5 5-7 1-2 2-4 7-8 n.a. 15-16 16-18 10-11 n.a. 13-14 n.a. 13-14 14-16 13-14 14-16 14-15 n.a. 15-16 16-18 10-11 11-13 7-8 n.a. 13-14 n.a. 25-26 26-28 15-16 n.a. 20-21 n.a. 20-21 21-23 20-21 21-23 20-21 21-23 20-21 21-23 15-16 16-18 13-14 n.a. 20-21 n.a.

) 45-46 46-48 32-33 n.a. 27-28 n.a. 30-31 31-33 30-31 31-33 40-41 n.a. 50-51 51-53 20-21 21-23 20-21 n.a. 41-42 n.a. m c

( 65-66 66-68 45-46 n.a. 40-41 41-43 50-51 51-53 57-58 58-60 47-48 48-50 110-111 111-113 30-31 31-33 35-36 n.a. 55-56 56-58

e 105-106 106-108 120-121 121-123 140-141 68-69 69-71 66-67 67-69 64-65 n.a. 180-181 181-183 36-37 37-39 65-66 n.a. 125-126 n.a. t c e

l 125-126 126-128 205-206 206-208 240-241 241-243 114-115 115-117 90-91 91-93 77-78 n.a. 250-251 251-253 44-45 45-47 128-129 129-131 148-149 149-151 l o c

162-163 163-165 n.a. n.a. n.a. n.a. 140-141 141-143 103-104 104-106 113-114 n.a. n.a. n.a. 60-61 61-63 165-166 166-168 n.a. n.a. s e l 225-226 226-228 n.a. n.a. n.a. n.a. 148-149 149-151 154-155 155-157 160-161 161-163 n.a. n.a. 120-121 121-123 191-192 n.a. n.a. n.a. p

m n.a. n.a. n.a. n.a. n.a. n.a. 161-162 162-164 174-175 175-177 n.a. n.a. n.a. n.a. 220-221 221-223 260-261 261-263 n.a. n.a. a S n.a. n.a. n.a. n.a. n.a. n.a. 190-191 191-193 192-193 193-195 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 202-203 203-205 230-231 231-233 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 215-216 216-218 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 262-263 263-265 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 267-268 268-270 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.

4.3 Results

4.3.1 Geomorphological Description

In total, three major submarine channels and a complex series of erosion platforms and mass movement scars were present on the continental slope of Desoto Canyon. These are described below. For descriptive purposes, these channels are designated as channels A, B1 and B2 or “twin channels” (Figure17). Basic geomorphological parameters such as length, width, depth, perimeter, area, and volume are summarized in Table 10.

Table 10. Summary of channel metrics . Length Width Perimeter Area Volume Channel ID (km) (km) (km) (km²) (km³) Channel A 60.23 3.2 232,474 147.04 5.11 Channel B1 74.58 2.85 153,120 149.19 4.70 Channel B2 73.98 2.34 166,939 174.19 4.05

All three channels progressively change their V-shaped form to U-shaped while they continue down slope. The A channel, which started as two V-shaped erosional incisions in the seafloor, joins together at about 1660m depth, forming a much wider and deeper, U-shaped structure (3.1 km-wide and 100 m depth), with flat bottom and sediment accumulation in the edge of its walls. The last section of these channels is oriented to the south and it extends for 23,5 km

60 until both channels reach a depth of 2350 m. In this last section, both channels increase in width (from 1.5 km in section 2 to 2.5 km), but with decreasing profile depth (~70 m). Although this section is also characterized by high sinuosity, it lacks the accumulation of material seen in the other sections. At the end of both channels, there is a marked decrease of depth profiles (~15 m), but both channels retain their U-shape and flat bottom. There is also an increment on width to 4.5 (B1) to 6.4 km (B2) at the channel mouth (Fig. 17).

Channel A This erosional feature is located 172 km south –southwest of Pensacola, FL. It is characterized by its Y-form, which is composed by two “parent” channels separated 4.5 km in the head of the channel (480 m depth); they then join together at 1660 m depth after 29km from their origin. The A channel ends at 2212 m depth, where the submarine fan with transported sediments from shallow areas starts to develop. Total length of the channel is 60.2 km and the average width of 3.2 km. The cross section profile of channel A alters progressively with downslope distance is characterized by a narrow V shape incision of approximately 80 m depth and 2.1 km width (average slope= 7.5º). Once the channels join to form a single feature, its shape progressively changes from a V-shaped to U-shaped until it reaches the end of the channel. As the profile of the channel changes to a U-shaped form, channel width and depth increases (3.5 km wide and 100 m depth). U-shaped section of the A channel is characterized by an evident and narrow channel thalweg line. Intrusion and accumulation of sedimentary material on the east edge of the channel wall and steeper slopes on the west wall (>10º). The last and deepest section of the A channel is characterized by decrease of the channel profile depth (< 60 m), flattening of the bottom, and increasingly gentle wall slopes (>2º). The channel walls disappear first on the east wall at 2020 m depth and last on the west edge at 2123 m depth, giving the way to sediment accumulation in the center of the channel (mouth) and the formation of the noteworthy submarine fan (Fig.17). Carbon dating of the upper sediment layers and interpretation of the x-ray and XRF scanning will illustrate the frequency of transport events. Interpretation of core laminations and anomalies together with the sub-bottom profiles may reveal the downslope extension turbidity flows and address the following research question: Do flow events affect the entire length of a channel, or do channels form as a result of sequential flows that activate portions of the channels at different times.

61 Channels B1 and B2 These two erosional structures are located 158 km south-southwest of Pensacola, FL. Their main characteristic is two erosional features running down-slope parallel to each other and following the same shape and sinuosity. The average distance between the two channels is 2.2 km and total length of 73.8 and 74.5 km respectively. The head of these channels is located at 950 m depth, and both start as minor incisions on the seafloor (< 30 m depth of channel profile) and 1.5 to 2 km wide. The first section (20 km) of these channels is oriented south and it is characterized by and smooth transition from V-shape to U-shape un profile, channel B2 (eastern channel) starts to get deeper on its profile than channel B1 (60 m and 30 m respectively), with steeper slopes (>8º) on its walls. Channels B1 and B2 shows steeper slopes on the west walls and more tendency to accumulate material on the east wall. Some areas of this section appear to be more sediment accumulation in the form of sediment deposition and wall collapse. What is noteworthy are change of slope in the corresponding channel walls and the geomorphology of these features. At about 1600 m depth, both channels abruptly change the orientation from south to southwest. These final sections extends for about 13.5 km and are characterized by high sinuosity, steep slopes, flat bottom, and U-shaped profiles in both channels. Sediment accumulations are mainly located on wall edges in highly sinuous areas.

Analyses of core profiles from these channels can address how slope steers the direction that channels flows follow. Specifically, what causes the abrupt change in course from an apparent cross-slope flow where the slope is steepest, to a southerly, downslope direction where the overall slope is actually decreasing.

Figure 17. 3D view of the channel system of the study area. Channel A in the upper panel and channels B1 and B2 in the lower panel. Cross section profiles of each channel allows to see the progression from V shape (steep slopes) to U shape (gradual slopes).

4.3.2 Sediment Data

In total 12 sediments cores were retrieved during the 2013 campaign. Location, depth, and core length are summarized in Table 1. Figure 20 shows an example of a piston core with obvious anomalies indicating turbidity flows. Of the 12 sediments cores, 6 piston cores samples were taken in pairs on one of the “twin channels” (one inside and one outside of the channel); an additional 4 piston cores were taken in the slump/slide area closer to the Desoto Canyon mouth. In addition, one core was recovered at the mouth/fan area of the Y channel (the second piston core attempt outside the channel did not triggered). A control sediment core was recovered from the site NT100 over the Florida Escarpment, where evidences of mass movement processes were not observable on the multi-beam bathymetry and backscatter map. Descriptions of the individual cores are as follows:

NT1000 The piston core retrieved on the control site shows the clear and constant pattern of undisturbed laminar sedimentation pattern along the sediment core. Laminar sediment stratigraphy

63 on NT1000 were characterized by alternation of light and dark grey sand bands, presumably hemipelagic carbonates, with the appearance of thin layers of black and brown silt/clay in some section of the core. Lamination patterns were similar in the whole extension of the sediment core, consisting on millimeters (just perceptible on x-ray radiographs) to few centimeter bands. It is no appreciable signal of disturbance or anomalous sediment deposition alongside the core.

PC01A –PC01B Both sediments cores were retrieved from the area in between of the Florida Escarpment (control site) and the area where the mass movement scars have been observed on the bathymetry map. The area is characterized by a short and narrow U-shaped channel close to the head of Desoto Canyon. Both sediment cores retrieved in the area were characterized by two main stratigraphic styles: laminar sedimentation and turbidite deposition. In general, laminar deposits just extend few centimeters in both cores and is predominate composed by grey sand and sand/silt sediments, similar to the control sites. Turbidite stratigraphic layers in both cores are dominant along the sediment core. The face is mainly characterized by homogeneous deposition in coloration and sediment type, highly compacted and high content of interstitial water. Sand/clay turbidites are light grey in coloration with the appearance of clastic components and some snail shells and sea urchin spines. Clay/silt turbide were more homogenous in color than sand turbide, mainly composed by dark grey components with absence of clay and organic remains.

PC02A-PC02B These sediment cores were retrieved from the area highly exposed to mass movement processes in the Desoto Canyon. The area is characterized by large mass wasting scars, presumabling submarine slides and turbitiy flows. Both cores were retrieved from the deepest mass waste scars between 1920 to 2100m deep. Sediment core stratigraphies were characterized by laminar deposition of sand and clay in the first 10 cm of the core, followed by sandy-silt turbides with clastic and calcareous animal parts intrusion in approximately the first 85 cm approximately, transiting to dark clay-silt tubides between pproximately 100 to 200cm. This section was homogenous in color and texture but x-ray radiographs showed differences in density along the stratigraphic facie. The deepest stratigraphic sections of the core (below 200 cm) is characterized by laminar deposition of light and dark grey silt with some section of brown clay. The last 20cm

64 of piston core PC02B presented a transition from 2.5 cm of brown clay to 4.5 cm of light grey sand and followed by another layer of brown clay. A similar pattern was observed in the last 5 cm of PC02A core (Fig. 17).

PC03A – PC03B This sediment cores were retrieved from the deepest section of the right twin channel. PC03A was collected immediately inside the channel and PC03B outside of it. PC03A presented the greatest variety of sediment stratigraphies and sedimentation/erosion processes among the complete set of piston cores. It starts with some laminar depositions for in the first 10 cm. Laminar deposition alternates with layers of light and dark grey sand, with dark silt and brown clay. A short and heterogeneous turbidite layer appears at 10cm depth, characterized by the mixture of sand and silt with intrusion of calcareous animal structures like bivalve shell fragments, gastropods shells, and equinoderms spines. At approximetally15cm depth, approximately appears the first erosion signal on the core. The light grey turbidite abruptly changed to dark turbidite facie that extended for approximately 35cm. This stratigraphic layer was composed of a well compacted layer or dark silt - silt/clay sediments with low content of interstitial water. Clastic and animal part intrusion were not perceptible on the scan or visual inspection. Between 47 and 50cm appear another erosional section with irregular edges between the previous section, a thin (1 cm) brown layer and a new stratigraphy facie of light grey coarse and fine sand with irregular intrusion of brown clay and dark grey silt. Below the 82cm, the core showed a new erosional section of irregular edges between the previous facie and a new layer of grey fine sand disturbed by dark grey silt and coarse sand. Bellow 200 cm, sediment patter tends to change to laminar sedimentation with short sections of disturbed sediments and homogenous turbidite (Figure 7).

PC03B presents almost along the whole core laminar sedimentation. The first section of the core shows a thick layer of brown clay with no intrusion of foreign material, followed by laminar sedimentation of light and dark grey coarse sand until 75 cm depth. Below that, a gradual transition from sand to dark silt starts until a homogenous facie of dark silt, apparently take place until the 175 cm depth. Below that, alternation of dark clay, coarse and fine sand and brown clay layers is observe in the core. X-ray radiographs shows high homogeneity in terms of density on the intermediate layer of dark silt (90 to 175 cm).

65 PC04A – PC04B These to sediment cores were retrieved from the same channel of PC03A and B, but in an intermediate point between the head and the mouth of the channel, at 1800 m depth approximately. PC04A was retrieve in a stable area outside of the channel and presents similar sedimentation pattern than PC03B, but sediment stratigraphies were dominated by dark silt and brown clay layer along core. The first 25 cm showed an alternation between brown clay and dark silt layers of less than 1cm, with some appearances of fine sand layer, moving gradually to dark silt dominance with few facies of clay and sand. X ray radiographs confirm similar sedimentation pattern, showing banding of high and low density layers along the core. PC04B show high homogeneity of sediment and coloration along the core. First 15 cm are dominated by brown clay facie with not intrusion of estrange elements and other types of sediments. Below this, a homogenous stratigraphic layer of coarse sand with some intrusion of dark grey silt occurred to 45 cm depth. This chaotic layer did not show a clear pattern of sedimentation. Small blocks of dark silt and calcareous valves and shells were found in this section. Bellow 45cm depth, there was a gradual but rapid transition from fine grey sand to dark homogenous silt until 121cm depth. X-ray radiograph did not show any difference in density along this stratigraphic layer, but it indicated some spots of higher density, which could be small sand patches or shells trapped in the turbidite. Between 121 to 127cm spears a layer of grey fine sand that limits in both edges with a thin layer (<1cm) of coarse sand interrupted by another turbidite layer of silt and fine sand. Below this section, a more consistent and stable laminar sedimentation patter starts to be more evident. This statigraphic layer is also dominated by dark and light grey silt alternated with brown clay and fine sand. (Fig. 20)

PC06A – PC06B Both sediment cores were retrieved in the shallower area of the twin channels, close to the mouth of the channel in the shelf break. This area is characterized by step slope (<10°) which lead to slope failure and mass movements. Both cores were retrieved 5 km below the head of the channel at 1130 m depth, in an area with an evident slide scar. Both sediment cores presented chaotic sediment stratigraphies, dominated by coarse light grey sand for the first 90 cm. depth. This layer in both sediment cores present intrusion of small carbonates pieces, shells, bivalves, and echinoderm spines. In some section, there was also appreciable accumulation of dark pockets of

66 sediment and liquid, which were also noticeable in the x-ray radiograph. Below this section, sediment stratigaphies gradually change to fine sand and in the deepest part of core PC06B (<180 cm) to a mixture of fine sand and silt in a laminated sedimentation pattern (Fig. 21)

4.3.3 Channels and Evidence for Mass Movements

In total, five chirp sonar transects were completed to characterize part of the study area: four transect profiles were acquired on the head-ward end of the A channel (Figure 4). A second line was completed over and outside the disturbance field formed by a significant mass movement in the shelf break over the channels B1 and B2 (Figure 4). All sub-bottom profiles transecting areas outside of the channel showed a clear and continuous stratification of sediment layers between 0 to 20 m below the sea floor. However, approaching the channel edge, this stratification pattern altered. Line A-A’ showed the complete disappearance of three sedimentary facies (green, blue, and purple) between the edge of the west wall of the canyon and the base, where they then reappeared. A similar pattern was observed in transect B-B’, but in the eastern wall of the channel, accompanied by lesser sediment accumulation on the channel bottom and a deformation of sedimentary facies on the west wall. Likewise in line B-B’, the western wall of the channel in chirp transect C-C’ showed an evident deformation and tensional depression on all sedimentary facies, which appear displaced towards the bottom of the channel. In the same sub-bottom profile, one stratigraphic layer (dark) appears to approach the edge of canyon and completely to disappear in the west wall of the canyon. This layer seems to sporadically appear and mix with newer sediment facies on the bottom of the canyon, where newer sediment layer get thicker, apparently accumulated from sediment remove from the east wall of the channel. stratigraphic layers on the west wall of the canyon. Furthermore, the top sediment facies seem to be thinner in comparison to others chirp profiles. (Fig.18)

67 Figure 18. Display of the sub-bottom profiles transects collected progressively along the downslope course of the A-channel. Inset shows associated bathymetry and transect lines.

The fifth chirp sonar profile was taken over a massive mass movement scar in the upper slope close to the shelf-break, over the canyons B1 and B2 (Figure 19). The slope failure starts about 100 m depth, just below of the mesophotic reef area called “Talus Block Reef”. Swath bathymetry data show an extensive slump scar of 3,1 km wide and 4.3 km long, with and average depth of 20 m in relation to the surrounding seafloor. The mass movement scar presents irregular walls and bottom, forming channels of about 100 m width which merge at 300 m depth into a single structure that apparently is the head of the twin channels.

The sub-bottom profiles were collected with a towed vehicle that was flown at an altitude of 2 to 3 m above the seafloor. This provides exceptional resolution and bottom penetration for deep-sea sediment profiling. However, some processing is required to correct vehicle motion caused by heave of the ship. With the completion of this processing, the profiles offer an a detailed overview of strata in Desoto Canyon, particularly when interpreted in conjunction with the sediment cores. These data can be analyzed to address the question of timing and continuity in channel formation and for potential identification of marker strata that may have been disturbed at synchronous intervals over much of the slope.

Figure 19. Multibeam bathymetry of upper slope of Desoto Canyon with indications of sediment movements. Cross section of upper slope subbottom profile and multibeam profile

Figure 20. High resolution scan images of the sediment cores retrieve and analyzed of the study area. Length-width relationship not in scale.

Figure 21. Scan image of PC03A showing different stratigraphic facies along the core.

4.3.4 Stable and Radio Isotopes

Stable isotopes values of both sediment cores fluctuates from ¹³C= -20.5‰ (PC04B, 2cm) to ¹³C= -26.67‰ (PC04B, 61cm) and ¹⁵N = 1.88‰ (PC04B, 61cm) to ¹⁵N = 5.06‰ (PC04B, 31cm). Values of radiocarbon Δ¹⁴C varied from -242.44 ‰ (2cm, PC04B) which correspond to 2,230 ±20 years BP, to 982.63 ‰ (61cm, PC04B) with an age rated to 32,560 ±100 years BP. (Table 11)

71 Table 11. Summary table of stable and radio isotopes of PC04A and PC04B in the study area.

When we compare the variability of stable isotopes ( ¹³C and ¹⁵N) concentration (‰) in relation to sediment depth. Both carbon and nitrogen isotopes present higher values at surface tending to decrease with depth. PC04A show higher values for both isotopes at the surface layer (2 cm) but drastically decreasing within the next 45 cm depth and stabilizing around ¹³C = -26‰ and ¹⁵N= -2.5‰. However, PC04B (inside the channel) shows a different behavior in both isotopes. Like PC04A, this core presents higher values of both stable isotopes at surface layer of the sediment core ( ¹³C ≈ -20‰; ¹⁵N≈ -4.5‰). Still, these values are relatively constant until 37 cm sediment depth. In the case of carbon 13, isotopes values fluctuate between ¹³C =-20.50‰ and -21.69‰, and decreasing to ¹³C=-26.66‰ (45 cm) to stabilize around that value until the last sample of the core (221 cm). Similarly, nitrogen values were stable the first 21 cm of sediments around ¹⁵N ≈ -4.4‰, then got slightly elevated around ¹⁵N ≈ -5‰ between 31 to 37 cm sediment depth, decreasing and getting stable to ¹⁵N ≈ -2‰ (45 cm) until the end of the core. (Fig. 22)

Figure 22. Stable isotopes ( ¹³C and ¹⁵N) profiles analyzed for the study area.

The origin of organic and inorganic nitrogen and carbon content on sediments samples can be determined by the relation between ¹³C over C/N ratio. For PC04-A and B, there area three 73 main groups: high values of carbon 13 ( ¹³C≈-21‰) and low C/N ratio (≈7), composed by surface sediments layers of both sediment cores plus sediment samples of the first 37 cm of sediments of core PB04B. Second group composed by sediments samples of PC04A between 7 to 251 cm with values of ¹³C≈-25.5‰ and C/N≈15; and last group composed by the deepest samples of of core PC04B (45 to 221 cm) with values of ¹³C≈-27‰ and C/N≈18. (Fig. 23 left.) Same situation is observed when it is compared radiocarbon ages (Δ¹⁴C age BP) values against the correspondent ¹³C values. The relationship between both carbon isotopes indicates that there are three different groups (sources) of carbon content on sediments in the area which are related with the age of the sediment sample (Fig. 23 right).

Radiocarbon (Δ¹⁴C) profiles values in both sediment cores indicate that surface intervals (2cm) values of ¹ ⁴C were higher (Δ¹⁴C >-242.55‰) in comparison to underlying sediment layers that were more depleted (Δ¹⁴C ≈-975) (Figure 10 left). This contrast with radiocarbon age (BP) profiles for both sediment cores. As expected, PC04A present newer sediment in the surface interval Δ¹⁴C =3580 ±25 YBP, increasing to Δ¹⁴C =27600±70 YBP at 111 cm until reach a maximum of Δ¹⁴C =28350±80 YBP at 251 cm below the sediment surface. In contrast, sediments of PC04B were notoriously older at deeper intervals. Surface sediment layer inside the channel was younger Δ¹⁴C = 2230±20 YBP. The oldest sediments were found at 61 cm below the sediment 74 surface with that sediments outside on Δ¹⁴C = 32560±100 YBP. The deeper interval (221 cm) in PC04B was dated to be Δ¹⁴C = 30940±90 YBP (Fig 24 right).

Δ¹⁴C age (years BP) Δ¹⁴C (‰) 0 10000 20000 30000 40000 -1000 -800 -600 -400 -200 0

100

150 Depth (cm) Depth 200

250 PC-04A 300 PC-04B

Figure 24. Radiocarbon depth profiles of PC04 (outside of the channel) and PC04-B (inside of the channel). Left: Δ¹⁴C‰; Right: Δ¹⁴C age in years BP.

4.4 Discussion

Submarine landslides and channel geomorphology of the United States continental slope have been extensively documented for the Pacific, Atlantic, and Gulf of Mexico continental slope. (McAdoo et al., 2000) identified four major areas where submarine landslides occurs: Oregon, California, Atlantic continental margin and Gulf of Mexico. Of those four areas, the Atlantic coast and the Gulf of Mexico are characterized to be tectonically passive margins with a long extension of the continental shelf, where a high number of submarine canyons and channels connect the coastal and deep ocean environments. In general, submarine channels are initiated on the continental shelf and upper slope and their origin is associated to mass movement processes and

75 the subsequent erosion (McGregor, 1983). (Twichell and Roberts, 1982) identified a series of channels and canyon in the Atlantic slope between Baltimore and the Hudson River (>200km) establishing differences in the geomorphology of the canyon and channel is related to the processes that give origin to these structures. Longer, deeper and low sinuosity channels are associated to fluvial erosion as a continuation of a river valley. In the other hand, shorter and more sinuous channels can be connected to erosion by currents and sediments. In the case of the Gulf of Mexico, The Mississippi Canyon is the only sedimentary structure that can be related to an extension of a fluvial body, in this case the Mississippi River.

However, that series of channels incised in the continental slope of the Desoto Canyon have to have a complete different origin. The sinuosity of the channels, changes in the channel profile structure, accumulation of sediments in certain areas inside the channels, and the presence of submarine gullies in the head portion of the channel close to the continental shelf suggest that submarine erosion and mass wasting processes have played a mayor factor in their formation. However, the presence of gullies on channels along the Atlantic coast seems to differ to the existence of gullies on the series of channels in the study area. The presence of gullies on submarine channels in the Atlantic coast are mainly located in the first third of the channel and always bellow the shelf break with the exception of massive canyons (e.g. Baltimore canyon, (Obelcz et al., 2014; Twichell and Roberts, 1982; Vachtman et al., 2013). In the series of channels studied in this work gullies are located in the head of the channels and over the shelf break, and in lesser degree along the channel in areas in and outside of the channel’s incision. The presence of gullies associated to canyon and channels walls in depths below 1000 m indicates that their origin is bound to submarine erosion (Twichell and Roberts, 1982). Gullies identified on the series of channel in northeastern GoM were predominately located in the upper slope and head of the channels, which suggest that subarea erosion and and mass wasting processes are the main precursors of the the channel formation, as has been previously documented by (Ryan et al., 1978; Twichell and Roberts, 1982; McAdoo et al., 2000; Vatchman et al., 2013). The few gullies located close to the walls in the middle and lower section to the channel can be attributable to collapses on the channel wall and/or hydrodynamic erosion in at outside of the channel.

76 In terms of the geomorphology of the channels, (Menard, 1955) and later Stanley and Moore (1983) classify channels on passive margins in relationship to their location to the continental shelf break two categories: Shelf and Slope Sources channels. Shelf-source channels are generally a continuation of a river valley, extending it origin several kilometer towards the coastline from the shelf break. Instead, slope –source channels generally start immediately below the shelf break due slope failure of unconsolidated sediments or/and changes in the ocean sea-level in inter-glaciation periods (McGregor, 1983; McAdoo et al., 2000; Tripsanas et al., 2007; Obletz et al., 2014).

In slope-source channels, as the channel of this study, the length and sinuosity of the channel are associated to the extension and steepness of the continental slope. Submarine channels offshore of the central Atlantic coast trend directly downslope towards the abyssal plain, in a linear shape, avoiding any evident sinuosity on its morphology. The northeastern Gulf of Mexico channels are characterized by high sinuosity and changes on its horizontal profile shape along of it. In general, the first ¾ of the channel length are characterized by high sinuosity and V-shaped transverse profile. This shape is characteristic of channels where large accumulation on sediments occur along sections of the channel where sediment transport is stopped due sinuosity, and collapse of the channel wall, creating terraces and slump scars along the channel. Clark et al. (1992) attribute the geomorphology and sinuosity of a channel to the grain size of sediment transported in channels. High sinuosity channels are usually linked to fine grain size sediments, while coarse grain sediments trend to generate lower sinuosity channels. In similar fashion, the degree of inclination of the slope and bathymetry gradient influence the morphology of the channel. Sediment cores inside of of the channels shows a disparity of their sediment faces, with despair intervals of carbonate coarse sand and silt and mud, erosional faces, and tubidite accumulation. This is the evidence that different processes of sediment transportation continuously change according the environmental conditions. Turbidities mass flow from upper run off downslope thru the channel, homogenizing fine grain sediment fractions. The appearance of coarse fractions of sediment in areas of sediment accumulation, could be related to slump and slides processes in the upper slope due high rates of sedimentation, slope failure of unconsolidated sediments and erosion of the channel walls. In addition, hydrodynamic erosion by currents that run parallel to the

77 coastline in the slope is easier in silty and muddy substrates, resulting in more sinuous channels (Peakall et al., 2007; Kleinhans, 2010; Verhagen et al., 2013).

The last section of the studied channels was characterized by less more straight channel toward the abyssal plain, a U-shape in transverse profile, broader in width but less deep in terms of the incision on the seafloor. A difference of the upper section of the channel, sediment accumulation mostly occurs in the thalweg section, avoiding the channel walls. There is not evidence of erosion or collapse of the channels walls in the sub-bottom profile and multi-beam bathymetry data, which leads us to infer that hydrodynamic erosion by currents is not a key processes in the channel formation, and mass wasting processes must be reduced to turbidite flows. Sediment cores and sun-bottom profile data shows an homogenous transition of sediment faces, dominated by trubidite of fine sediments with not erosional faces or coarse sand from the continental shelf. Similar outcomes has been commonly reported for submarine canyons and channels along the Pacific and Atlantic coast of North America and other (Menard, 1955; Bergantino, 1971; McGregor, 1983; McAdoo et al., 2000; Tripsanas et al., 2007 Obelcz et al., 2014).

However, it is unclear if submarine landslides and its triggering factors are the responsible for the creation of the series of channels in northeastern Gulf of Mexico. Initially Bergantino (1971) described the Mississippi/Alabama shelf and slope as a part of the Upper Mississippi fan, predominated by Pleistocene sediments deposited on the shelf and later transported downslope thru the canyon and marine slumps events. Although he also identifies the Desoto slope as a deposition area of the Mississippi river sediments, its structure appears to be unmodified and certainly stable. Later McAdoo et al (2000) reviewed triggering factors of marine landslides in the continental slope of the Gulf of Mexico, identifying three main areas as submarine landslide critical zone: Mississippi canyon, Green Canyon and the Sigsbee Platform. Although they briefly mention the area close to the head of the DeSoto canyon as an area with potential condition for mass wasting processes, they main focus was in the in the Mississippi Canyon due its historical depositional characteristics.

78 The formation and geomorphology of slope-source submarine channels and canyons along passive margins are generally linked to low sea-level periods, where shorelines were deeper than modern days and de-glaciation generate an impute of water and sediment into the marine system (Catuneanu, 2006). Sea level history of Gulf of Mexico have suffered changes sue eustatic variations in a global scale (Donoghue, 2011) Over the last several million years, coastal areas of the Gulf of Mexico had experimented changes on the sea level above and below the present levels, which bring dramatic changes on the position of the shoreline within the basin. Sea level records show the last minimum level sea level occurred about 20,000 year BP, locating the shoreline approximately 120 m below modern location (Siddall et al., 2003, Donoghue, 2011; Sims, 2014). This low sea-level records are linked to the Last Glacial Maximum, where maximum accumulation of ice in the North America Wisconsinan ice sheet occurred. However, this minimum sea level recorded for the Gulf of Mexico is still 35 meter higher in comparison to sea-lever record time- equivalent in other similar locations (Sims et al., 2007) This difference has been linked to the hypothesis of sediment loading by the Mississippi. Santschi and Rowe (2008) determinate the average sedimentation rates for the Gulf of Mexico is 9±2cm/ky, with peaks of 15 cm/ky, rating older sediments at 13cm depth with 27,100±210 YBP. Tripsanas et al. (2007) reports report radiocarbon ages for sediment cores inside of Bryan Canyon area with maximum ages of over 51,770 YPB at 685 cm of core depth. Our maximum value found inside of the channel was 32, 650 YBP at 61 cm core depth. Similar values were found by Tripsanas et al (2007) at 421 cm core depth. This discordance between same sediment ages at different position in sediment facies could be explained by hydrodynamic erosion or mass wasting processes such tubidity flows are happening inside of the channel. This processes can be removing (or impeding) the accumulation of newer sediment faces inside of the channel. This is supported by the analysis of multi-beam and sub-bottom profile data, which shows areas with large sediment deposition downslope, and areas with submarine landslide scars and sediment debri after mass wasting events have occur in the upper slope.

After the DWH incident, investigation of the impact of oil spills in the Gulf of Mexico has been mainly focused on the discharge of oil, gas and dispersant application resulting from well explosion (Camilli et al., 2010; Joye et al., 2011b; McNutt et al., 2012). Nevertheless, another incident similar in characteristics but without the exposure of the DWH yet not well studied demonstrate that natural processes such as submarine landslides and storm could lead to oil incidents similar in magnitude to the DWH. The history of the Taylor Energy site points to another scenario that could generate an oil spill of national significance. The final landfall of Hurricane Ivan in 2004 caused a turbidity flow and slope failure along the Mississippi Delta that toppled a production platform at ~100 m water depth in the MC-20 lease block. This initiated discharge that has continued, despite strenuous response efforts, to the present day (Daneshgar Asl et al., 2015; MC-20 Response Information Center, 2016). This event suggests a risk profile for an accidental 80 spill that would be potentially more devastating than the DWH spill, particularly if it were it to happen further offshore.

As noted above, the large sediment load and salt tectonic deformation of Gulf of Mexico slope contributes to the geologic record of repeated major slope failures (Dugan and Stigall, 2010). The increasing practice of water injection to foster sub-seafloor production compounds the potential for fault activation (Ellsworth, 2013). A turbidity flow that occurred without warning could destroy a production platform. This could potentially lead to an uncontrolled flow from dozens of wells and the slide debris would make it very challenging to drill relief wells as in the case of the Taylor platform. Hydrocarbon discharge might, therefore, continue for years. Sinuous channels are a case of evident risk that may not be fully considered when exploration wells, production platforms, and pipelines are installed. Although industry regulations require thorough hazard analysis and geotechnical surveys, sinuous channels might be missed without a regional scale, high-resolution mapping. We note that several production platforms in the DeSoto Canyon region have been constructed in proximity to channels (Fig.25). Additionally, sinuous canyons represent potential areas of higher biomass and diversity (De Leo et al., 2010), which are potentially subject to impact from exploration and drilling activities. These factors impel further investigation of the channel morphologies, their formation and recent geological history, and evidence for material transport within the canyon confines.

4.5 Conclusions

A series of submarine channels and erosional platforms located in the continental slope in Northeastern Gulf of Mexico, below of the MCE in the continental shelf between Mississippi and Florida have been identifies and described. The channels, three in total, are slope-source structures, which start right below of the shelf break apparently originated after a mass wasting event occur, this is supported by the slump and slides scars found in the shelf break over the head of the channels.

In average, the channels are 65 km long and 2.5 km with, with an incision on the seafloor of 20 meter in average. All channels show a transition od their transversal profile from V-shape in the head to U-shape in the toe, finishing in a small channel fan in the abyssal plain of the basin. 81 All channels present high sinuosity in the upper ¾ of the channel, showing accumulation of sediments in the turn-sections of the channels, and accumulation of sediment on the walls of it due collapsing of unconsolidated sediments. Last quarter of the channel present high accumulation of sediments on the thalweg section.

Sedimentary faces obtained by sediment cores show the evidence of disturbed interval of sedimentation and erosion, with erratic periods of laminar sedimentation, erosion, turbidities of small sediment fractions, this interrupted by erosional faces and coarse fractions of sediment, characteristic from the continental shelf. We assume that the appearance or turbidite, erosional and coarse sediment faces could be explained by turbidity flows, slumps and slides, and hydrodynamic erosion.

Sediment ages along the inside of the canyon suggest that hydrodynamic erosion is continuously occurring inside the channel, exposing older sediments to the surface. However, we do not discard the possibility of mass wasting processes are still occurring in the area. The presence of coarse sediment fractions typically found in the continental shelf, beside the occurrence of turbidite homogenous fractions demonstrate that sediment wasting from the upper slope could still occurring. This, could lead to potential disaster in the area because the presence of oil industry structures. Massive turbidity flows and submarine slumps have been describe to the area and other places. We recommend to the authority and oil industry further studies in order to prohibit the installation of submarine oil structures such us pipelines and rigs in places with high potential of mass wasting processes.

82 CHAPTER 5 CONCLUSIONS `The complexity and variety of processes that have took place during the long geological history of the Gulf of Mexico have shaped and transformed the seafloor geomorphology of the basin, as well created and provided the necessary conditions that make the Gulf of Mexico one of the most prolific oil and gas reservoirs in the world. This is especially evident in Northern Gulf of Mexico, where the seafloor geomorphology is completely intricate and assorted, founding areas with rugged topography such as the Sigsbee Escarpment, submarine canyons, salt domes, submarine channels, erosional platforms, mass wasting vestiges, relict paleo shorelines, among other geomorphologic scenarios. In addition to this, several inter-glacial periods modified the sea level, changing the location of the shoreline meters below and above the present location, and loading with terrigenous sediments the ocean system also helped to modify the marine geomorphology of the Gulf, especially in the continental shelf and slope.

This diversity of geomorphologic setups have allowed the possibility of hosting different ecosystems along the whole depth rage that the basin offers. Deep ocean chemosynthetic communities such as cold seep that relies on oil and gas as source of energy which is released to the water columns from deep reservoirs thousands of meter below the ocean sea floor. or soft sediments communities, deep-sea, mesophotic shallow water soft and reef forming coral communities, whom rely on primary productivity that is transferred to upper levels of the food net in different ways.

The mesophotic coral ecosystem (MCE) in northeastern Gulf of Mexico, as other structures that provide hard substrate to be colonized by soft and reef corals, are of the last eustatic rise of the sea level, which at the time were large reef aggregations of zooxanthellate corals that were unable to survive at the new depth. (Gardner et al., 2001a). With the melting of the ice sheet after the glaciation period, large volumes of freshwater and terrigenous sediments were deposited by the Mississippi river into the Gulf, rising the sea level more that 120m to the actual level and killing and covering with clay and sand the submareal reef fields.

83 This MCE, specially the area called The Pinnacle Reef, were under the influence of floating oil and the dispersant application after the 2010 Deepwater Horizon oil disaster. After an exhaustive revision, it was concluded that coral injuries observed in the Pinnacle Reef trend in 2011 were unequivocally unusual and significantly vaster in terms of frequency, characteristic and severity of the injuries in comparison to data obtained in the same area prior to the DWH incident. The most characteristic injuries were patchy damage affecting portions of the sea fan colony, retracted polyps and mucus secretion. Advanced levels of injuries were characterized by biofilm overgrowth, necrotic tissue, bare and broken branches that eventually fall apart of the colony due the overgrowth weight.

A possible explanation for the exposure of mesophotic corals to oil released by the DWH discharge is the proximity of the sites to the incident and the floating oil and subsequent dispersant applications to the slicks. The trajectory of the TS Bonnie over the area would eventually have cause turbulent mix in the surface, sinking floating oil particles which could have reach corals and other sessile organism in the area.

However, we were just able to document coral injuries in just two reefs, which together represent less than 0.0002% of the potential of mesophotic habitat in Northeaster Gulf of Mexico. The long extension of the mesophotic area in Gulf of Mexico and the insufficient resources available for deep sea exploration made almost impossible to visit every single square meter of potential MCE. Furthermore, it is unlikely to find a well stablished mesophotic coral community in the whole extension of the mesophotic because the geomorphological, oceanographic, and biological conditions for recruitment and survival rate of mesophotic coral will not be available in the whole extension of the mesophotic zone.

To resolve this problem, we proposed eight habitat suitability models for mesophotic corals in order to predict their incidence along the potential mesophotic area located in the between Mississippi and Florida. In addition, we aimed to identify which environmental variables are necessaries to allow the occurrence of mesophotic gorgonians and antipatharians in the area. All predictive models exceed their capability of predicting suitable area for mesophotic gorgonians

84 and antipatharians. MCE distribution is primarily determined by sea-floor rugosity, topographic index position, CDOM, currents, loose sediments, hurricane wind density.

This suitability maps for mesophotic corals are an exceptional tool that ca be used as a guidance for exploration of new MCE areas different from the already well know reefs such AAR or MSSR in order to extend the knowledge of MCE’s in the eastern Gulf of Mexico. Knowing the real extension of the MCE’s is the first step in order to offer a precise approximation of MCE’s that could be impacted catastrophic events. The protection, management, and conservation of MCE’s will also depend on the accuracy of the extension and location where this ecosystem occurs.

Finally, we focused our efforts in another and completely different marine structure, which is adjacent to the mesophotic reefs ecosystem in Northeastern Gulf of Mexico. These structure are a series a submarine channels and erosional platforms incised in the continental slope, connecting the continental shelf just below of the Pinnacle reefs to the abyssal plane of the basin. Although the creation of these sedimentary and erosional structures is not clear, slopes-sourced submarine channels origin is usually associated to slope failure and subsequent mass wasting events due rapid accumulation of unconsolidated sediments in the shelf break and/or changes in the hydrostatic pressure over the seafloor. This two types of events can occur separately or simultaneously when eustatic sea level rise occur during de-glaciation events, as it happened 20,000 years ago.

This series of submarine channels and erosional platforms are slope-sourced structures, starting right below of the shelf break, where submarine landslide scars were observed on the head of these channels. The three submarine channels are 65km long and 2.5km width in average, with an incision on the seafloor of up to 200m. All channels show a transition on their transversal profile from V-shape (head) to U-shape (toe), with a small channel fan in the at the end of each structure. High sinuosity and sediment accumulation by transport and wall collapse was also appreciated on the first sections of the channels, giving pass to high accumulation of sediments on the thalweg section close to the end of each structure.

The sediment cores analyzed showed the evidence of disturbed sedimentation intervals, passing from laminar sedimentation, to turbidite, slumps of fine o very coarse sediment fractions,

85 and erosional facies, which followed an unpredictable rate of occurrence from core to core. We infer the appearance of turbidite and coarse sediment facies could be linked to turbidity flows, slumps and slides events; and the erosional facies related to hydrodynamic erosion by currents.

Radiocarbon dating of core sediments obtained inside of the canyon suggest that hydrodynamic erosion by currents is continuously occurring inside the channel, removing upper sediments layers and uncovering older sediments facies to the surface. Still, mass wasting processes are regularly happening in inside and outside of the channels in the erosional platforms of the continental slope, which is evident by finding coarse sediment fractions, characteristically found in the continental shelf, at 1200m depth inside of the channel. Furthermore, the occurrence of turbidite inside and outside of the channel validate the idea that sediment-wasting processes could still be occurring.

The mass wasting events in northern Gulf of Mexico might generate disaster of unprecedented proportions in the area due the crowded presence of oil industry structures, from oil rigs to pipelines that cross the channel and the continental slope towards coastal areas in the Gulf. Is well known that turbidity flows and submarine slumps and slides have been the cause of many disasters in different parts of the world, braking telegraph lines in the past and fiber optic line in present days, generating tsunamis, and destroying oil structures in that leads to oil spills disasters. Understanding the triggering factors, frequency and areas more susceptible to mass wasting processes in the Gulf of Mexico it is crucial for planning which locations should be avoided for placing submarine structures such oilrigs and pipelines. Prevention, protection, and planning should be the strategy in the area in order to minimize the risk of oil disaster in the area, which is right below of the MCE already subject to contamination after the DWH incident.

86 APPENDIX A

SUPPLEMENTARY MATERIAL FOR CHAPTER 3

Figure 26. Map of the study area including coral records from the MAPTEM data and the 2014 cruise (MAPTEM) that were used in this study to create the habitat suitability model.

Figure 27. Map of the study area including independent coral records NOAA used to create the habitat suitability model.

Figure 30. Independent models of habitat suitability for mesophotic corals compiled using NOAA coral records and subset of significant parameters.

Table 12. Statistic summary of variables used to perform the model.

Layer MIN MAX MEAN STD Depth -360.688 -35.6741 -96.6578 40.0265 Aspects -1 359.9999 188.8434 75.3649 Ruggedness 0 0.0787 0 0.0001 Ruggosity 1 1.2572 1.0002 0.0012 Slope (3) 0 36.603 0.5376 0.7751 Slope (5) 0 36.603 0.5376 0.7751 TPI 50m -46.9451 106.2572 0 1.0209 TPI 100m -34.9448 73.3246 0 1.0236 TPI 250m -24.8008 39.9363 -0.0002 1.0239 TPI 500m -21.3843 32.5621 -0.0003 1.0236 Backscatter 19.0014 253.9601 185.7836 12.6537 Dom. Sed. -99 3000 183.2104 555.0895 Loose Sed. 3 322 167.651 111.3046 Backs. Un. Iso. 1 10 5.2519 2.4965 T bottom 18.4737 22.0865 19.5826 0.6189 S bottom 35.9659 36.3683 36.2199 0.0629 U bottom -0.0676 0.3979 0.0085 0.0651 V bottom -0.0768 0.0836 0.0033 0.0185 W bottom -0.0002 0.00004 -0.00009 0 Hurr. winds 37.2851 50.6197 42.5331 3.1452 STT 17.552 20.988 19.1077 0.5031 Fluorecense 0.0089 0.0195 0.0112 0.0019 Diff. Att. Coef. 0.0571 0.3278 0.0996 0.0454 Chl. A 0.3367 3.4844 0.8654 0.5768 CDOM 38.6087 39.4638 38.9761 0.1411 PAR 38.5579 39.4638 38.9709 0.145 Silicate 3.8605 8.0678 6.1876 1.0837 Phosphate 0.2017 0.3797 0.2903 0.0509 Nitrate 0.52 3.4464 1.4359 0.851 Oxygen utilization 0.2757 0.3792 0.3142 0.024 Oxygen saturation 92.9918 95.0188 94.224 0.4899 Dissolved oxygen 4.6261 4.7699 4.7196 0.0361

92 Table 13. Pearson correlation used to evaluate significance of individual variables in the coral suitability model. Variables with positive Pearson correlation >0.4, or negative Pearson correlation <-0.4 were selected as main drivers of habitat suitability.

Variables Depth Aspects Ruggedness Ruggosity (3) Slope (5) Slope TPI50m TPI100m TPI250m TPI500m Backscatter Dom.Sed. LooseSed. Backs.Un. Iso. Tbottom bottom S Ubottom bottom V bottom W winds Hurr. STT Fluorecense Diff.Coef. Att. Chl.A CDOM PAR Silicate Phosphate Nitrate Oxygenutilization Oxygensaturation oxygen Dissolved Depth 1.00 -0.03 0.01 -0.14 -0.27 -0.27 0.02 0.04 0.09 0.16 0.15 0.08 0.36 0.20 0.26 -0.04 -0.27 0.10 0.21 0.35 -0.56 0.16 0.13 0.14 0.14 0.21 0.16 0.15 0.09 -0.16 0.15 0.03 Aspects 1.00 0.00 -0.05 -0.09 -0.09 0.00 0.00 0.01 0.01 0.04 -0.08 -0.07 -0.13 0.03 0.24 -0.08 -0.04 -0.11 0.00 0.05 -0.20 -0.17 -0.19 -0.02 -0.04 -0.30 -0.31 -0.27 0.02 0.00 0.18 Ruggedness 1.00 0.40 0.32 0.32 0.16 0.17 0.14 0.11 0.05 0.01 0.01 0.06 -0.02 -0.02 -0.01 -0.02 0.00 -0.01 0.00 -0.01 -0.01 -0.01 0.00 0.00 0.00 0.00 -0.01 0.00 0.00 0.01 Ruggosity 1.00 0.77 0.77 0.07 0.10 0.12 0.09 -0.01 0.05 0.01 0.09 -0.02 -0.07 0.16 0.01 -0.05 -0.07 0.09 0.04 0.03 0.03 -0.01 0.00 0.08 0.08 0.08 0.03 -0.03 -0.07 Slope (3) 1.00 1.00 0.04 0.06 0.08 0.06 0.03 0.04 0.00 0.18 -0.09 -0.13 0.21 -0.04 -0.05 -0.11 0.15 0.02 0.01 0.02 -0.02 -0.02 0.10 0.11 0.10 0.04 -0.05 -0.09 Slope (5) 1.00 0.04 0.06 0.08 0.06 0.03 0.04 0.00 0.18 -0.09 -0.13 0.21 -0.04 -0.05 -0.11 0.15 0.02 0.01 0.02 -0.02 -0.02 0.10 0.11 0.10 0.04 -0.05 -0.09 TPI 50m 1.00 0.88 0.55 0.33 0.01 0.00 0.00 0.02 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 TPI 100m 1.00 0.78 0.52 0.02 0.00 0.01 0.03 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 TPI 250m 1.00 0.86 0.04 0.01 0.02 0.07 0.00 0.00 -0.02 -0.01 0.00 0.01 -0.02 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 TPI 500m 1.00 0.06 0.01 0.04 0.10 0.01 0.00 -0.03 -0.01 0.00 0.01 -0.03 0.01 0.00 0.00 0.01 0.01 -0.01 -0.01 -0.01 0.00 0.00 0.00 Backscatter 1.00 -0.07 0.06 0.70 -0.26 0.11 -0.46 -0.49 0.13 0.25 -0.31 -0.61 -0.53 -0.55 -0.03 -0.08 -0.54 -0.54 -0.62 -0.30 0.35 0.61 Dom. Sed. 1.00 0.18 0.07 -0.05 -0.13 0.11 0.07 -0.11 0.01 0.12 0.24 0.29 0.28 0.06 0.06 0.26 0.27 0.27 0.07 -0.08 -0.23 Loose Sed. 1.00 0.25 0.13 -0.11 0.00 -0.02 0.10 0.13 -0.32 0.06 0.01 0.03 0.03 0.07 0.17 0.16 0.10 -0.16 0.15 0.02 Backs. Un. Iso. 1.00 0.06 -0.06 -0.01 -0.14 0.05 0.07 -0.20 -0.08 -0.11 -0.10 0.01 0.05 -0.01 -0.01 -0.05 -0.09 0.09 0.09 T bottom 1.00 0.25 0.23 0.44 -0.01 -0.27 -0.19 0.33 0.15 0.18 -0.03 0.08 0.15 0.15 0.24 0.34 -0.35 -0.36 S bottom 1.00 -0.12 -0.06 -0.15 -0.06 0.19 -0.13 -0.12 -0.14 0.10 0.10 -0.51 -0.51 -0.35 0.32 -0.28 0.08 U bottom 1.00 0.54 -0.41 -0.11 0.44 0.48 0.44 0.46 0.09 0.13 0.47 0.46 0.55 0.26 -0.31 -0.54 V bottom 1.00 0.10 0.05 0.11 0.56 0.53 0.56 0.15 0.20 0.49 0.49 0.58 0.28 -0.32 -0.55 W bottom 1.00 0.14 -0.61 -0.12 -0.19 -0.18 0.08 0.07 0.07 0.07 -0.05 -0.24 0.24 0.17 Hurr. winds 1.00 -0.28 -0.06 0.09 0.07 0.32 0.35 0.08 0.06 -0.08 -0.63 0.62 0.41 STT 1.00 0.25 0.38 0.35 0.18 0.12 0.00 0.02 0.21 0.51 -0.52 -0.43 Fluorecense 1.00 0.93 0.96 0.22 0.29 0.84 0.85 0.94 0.35 -0.41 -0.86 Diff. Att. Coef. 1.00 0.99 0.33 0.37 0.76 0.77 0.88 0.35 -0.40 -0.81 Chl. A 1.00 0.31 0.37 0.80 0.81 0.91 0.35 -0.41 -0.83 CDOM 1.00 0.93 0.08 0.10 0.18 0.23 -0.25 -0.22 PAR 1.00 0.16 0.19 0.27 0.25 -0.27 -0.29 Silicate 1.00 1.00 0.94 -0.04 -0.03 -0.65 Phosphate 1.00 0.95 0.01 -0.08 -0.68 Nitrate 1.00 0.29 -0.36 -0.86 Oxygen utilization 1.00 -1.00 -0.73 Oxygen saturation 1.00 0.78 Dissolved oxygen 1.00

REFERENCES

Ai F, Strasser M, Preu B, Hanebuth TJJ, Krastel S, Kopf A (2014) New constraints on oceanographic vs. seismic control on submarine landslide initiation: a geotechnical approach off Uruguay and northern Argentina. Geo-Marine Letters 34:399-417

Anderson JB, Wallace DJ, Simms AR, Rodriguez AB, Milliken KT (2014a) Variable response of coastal environments of the northwestern Gulf of Mexico to sea-level rise and climate change: Implications for future change. Marine Geology 352:348-366

Anderson JB, Wallace DJ, Simms AR, Rodriguez AB, Milliken KT (2014b) Variable response of coastal environments of the northwestern Gulf of Mexico to sea-level rise and climate change: Implications for future change. Marine Geology 352:348-366

Anderson KS, Graham SA, Hubbard SM (2006) Facies, architecture, and origin of a reservoir- scale sand-rich succession within submarine canyon fill: Insights from Wagon Caves Rock (Paleocene), Santa Lucia Range, California, USA. Journal of Sedimentary Research 76:819-838

Anderson OF, Guinotte JM, Rowden AA, Clark MR, Mormede S, Davies AJ, Bowden DA (2016) Field validation of habitat suitability models for vulnerable marine ecosystems in the South Pacific Ocean: Implications for the use of broad-scale models in fisheries management. Ocean & Coastal Management 120:110-126

Austin M (2007) Species distribution models and ecological theory: A critical assessment and some possible new approaches. Ecological Modelling 200:1-19

Baeten NJ, Laberg JS, Vanneste M, Forsberg CF, Kvalstad TJ, Forwick M, Vorren TO, Haflidason H (2014a) Origin of shallow submarine mass movements and their glide planes- Sedimentological and geotechnical analyses from the continental slope off northern Norway. Journal of Geophysical Research-Earth Surface 119:2335-2360

Baeten NJ, Laberg JS, Vanneste M, Forsberg CF, Kvalstad TJ, Forwick M, Vorren TO, Haflidason H (2014b) Origin of shallow submarine mass movements and their glide planes- Sedimentological and geotechnical analyses from the continental slope off northern Norway. Journal of Geophysical Research-Earth Surface 119:2335-2360

Balsam WL, Beeson JP (2003) Sea-floor sediment distribution in the Gulf of Mexico. Deep-Sea Research Part I-Oceanographic Research Papers 50:1421-1444

Bayer F (1961) The shallow water Octocorallia of the West Indian region. Book 1St Edition. Martinus Nijhoff (Holland)

Bergantino RN (1971a) Submarine Regional Geomorphology of the Gulf of Mexico. Geological Society of America Bulletin 82:741-752

Bergantino RN (1971b) Submarine Regional Geomorphology of the Gulf of Mexico. Geological Society of America Bulletin 82:741-752

Braithwaite CJR (2016) Coral-reef records of Quaternary changes in climate and sea-level. Earth-Science Reviews 156:137-154

Bridge T, Beaman R, Done T, Webster J (2012) Predicting the Location and Spatial Extent of Submerged Coral Reef Habitat in the Great Barrier Reef World Heritage Area, Australia. PLoS One 7

Brooks J, Giammona CE (1991) Mississippi-Alabama continental shelf ecosystem study: data summary and synthesis. OCS Study MMS 91-0062. U.S. Department of the Interior, Minerals Management Service, Gulf of Mexico OCS Regional Office, New Orleans. Executive Summary, Book I

Bruno J, Petes L, Hettinger A (2003a) Nutrient enrichment can increase the severity of Caribbean coral diseases. Ecological Society of America Annual Meeting Abstracts 88:50-50

Bruno JF, Petes LE, Harvell CD, Hettinger A (2003b) Nutrient enrichment can increase the severity of coral diseases. Ecology Letters 6:1056-1061

Bryn P, Berg K, Forsberg CF, Solheim A, Kvalstad TJ (2005) Explaining the Storegga Slide. Marine and Petroleum Geology 22:11-19

Burrough PA, McDonell RA (1998) Principles of Geographical Information Systems. Oxford University Press, New York

Canals M, Danovaro R, Heussner S, Lykousis V, Puig P, Trincardi F, Calafat AM, de Madron XD, Palanques A, Sanchez-Vidal A (2009) CASCADES IN MEDITERRANEAN SUBMARINE GRAND CANYONS. Oceanography 22:26-43

Canals M, Lastras G, Urgeles R, Casamor JL, Mienert J, Cattaneo A, De Batist M, Haflidason H, Imbo Y, Laberg JS, Locat J, Long D, Longva O, Masson DG, Sultan N, Trincardi F, Bryn P (2004) Slope failure dynamics and impacts from seafloor and shallow sub-seafloor geophysical data: case studies from the COSTA project. Marine Geology 213:9-72

Cecchini S, Taliana D, Giacomini L, Herisson C, Bonnemaire B (2011) Submarine geo-hazards on the eastern Sardinia-Corsica continental margin based on preliminary pipeline route investigation. Marine Geophysical Research 32:71-81

Chanton J, Zhao TT, Rosenheim BE, Joye S, Bosman S, Brunner C, Yeager KM, Diercks AR, Hollander D (2015) Using Natural Abundance Radiocarbon To Trace the Flux of Petrocarbon to the Seafloor Following the Deepwater Horizon Oil Spill. Environmental Science & Technology 49:847-854

Chaytor JD, ten Brink US, Solow AR, Andrews BD (2009) Size distribution of submarine landslides along the US Atlantic margin. Marine Geology 264:16-27

95 Choi YH, Wang Y (2004) Dynamics of carbon sequestration in a coastal wetland using radiocarbon measurements. Global Biogeochemical Cycles 18

Clarke AJ, Van Gorder S (2016) Subinertial canyon resonance. Geophysical Research Letters 43:3872-3879

Coffin RB, Osburn CL, Plummer RE, Smith JP, Rose PS, Grabowski KS (2015) Deep Sediment- Sourced Methane Contribution to Shallow Sediment Organic Carbon: Atwater Valley, Texas- Louisiana Shelf, Gulf of Mexico. Energies 8:1561-1583

Continental Shelf Associates, University TAM (2001) Mississippi/Alabama Pinnacle Trend Ecosystem Monitoring, Final Synthesis Report. Book OCS Study MMS 2001 080, U.S. Department of the Interior, Geological Survey, Biological Resources Division, USGS BSR 2001 0007 and Minerals Management Service, Gulf of Mexico OCS Region, New Orleans, LA,

Couce E, Irvine PJ, Gregorie LJ, Ridgwell A, Hendy EJ (2013) Tropical coral reef habitat in a geoengineered, high-CO2 world. Geophysical Research Letters 40:1799-1804

Crutchley GJ, Gorman AR, Fohrmann M (2007) Investigation of the role of gas hydrates in continental slope stability west of Fiordland, New Zealand. New Zealand Journal of Geology and Geophysics 50:357-364

Davies AJ, Guinotte JM (2011) Global Habitat Suitability for Framework-Forming Cold-Water Corals. Plos One 6

Davies AJ, Roberts JM, Hall-Spencer J (2007) Preserving deep-sea natural heritage: Emerging issues in offshore conservation and management. Biological Conservation 138:299-312

DeLeo DM, Ruiz-Ramos DV, Baums IB, Cordes EE (2016) Response of deep-water corals to oil and chemical dispersant exposure. Deep Sea Research Part II: Topical Studies in Oceanography

Dennis GD, Bright TJ (1988) Reef fish assemblages on hard bank in the Northwestern Gulf of Mexico. Bulletin of Marine Science 43:280-307

Donoghue JF (2011a) Sea level history of the northern Gulf of Mexico coast and sea level rise scenarios for the near future. Climatic Change 107:17-33

Donoghue JF (2011b) Sea level history of the northern Gulf of Mexico coast and sea level rise scenarios for the near future. Climatic Change 107:17-33

Doughty CL, Quattrini AM, Cordes EE (2014) Insights into the population dynamics of the deep-sea coral genus Paramuricea in the Gulf of Mexico. Deep-Sea Research Part Ii-Topical Studies in Oceanography 99:71-82

Dukhovskoy DS, Leben RR, Chassignet EP, Hall CA, Morey SL, Nedbor-Gross R (2015) Characterization of the uncertainty of loop current metrics using a multidecadal numerical simulation and altimeter observations. Deep Sea Research Part I: Oceanographic Research Papers 100:140-158

96 Dustan P, Doherty O, Pardede S (2013) Digital Reef Rugosity Estimates Coral Reef Habitat Complexity. Plos One 8

El Morjani ZEA, Ebener S, Boos J, Abdel Ghaffar E, Musani A (2007a) Modelling the spatial distribution of five natural hazards in the context of the WHO/EMRO Atlas of Disaster Risk as a step towards the reduction of the health impact related to disasters. International journal of health geographics 6:8-8

El Morjani ZEA, Ebener S, Boos J, Ghaffar EA, Musani A (2007b) Modelling the spatial distribution of five natural hazards in the context of the WHO/EMRO Atlas of Disaster Risk as a step towards the reduction of the health impact related to disasters. International Journal of Health Geographics 6

Elith J, Graham CH, Anderson RP, Dudik M, Ferrier S, Guisan A, Hijmans RJ, Huettmann F, Leathwick JR, Lehmann A, Li J, Lohmann LG, Loiselle BA, Manion G, Moritz C, Nakamura M, Nakazawa Y, Overton JM, Peterson AT, Phillips SJ, Richardson K, Scachetti-Pereira R, Schapire RE, Soberon J, Williams S, Wisz MS, Zimmermann NE (2006) Novel methods improve prediction of species' distributions from occurrence data. Ecography 29:129-151

Elith J, Kearney M, Phillips S (2010) The art of modelling range-shifting species. Methods in Ecology and Evolution 1:330-342

Elith J, Leathwick JR (2009) Species Distribution Models: Ecological Explanation and Prediction Across Space and Time. Annual Review of Ecology Evolution and Systematics 40:677-697

Elith J, Phillips SJ, Hastie T, Dudik M, Chee YE, Yates CJ (2011) A statistical explanation of MaxEnt for ecologists. Diversity and Distributions 17:43-57

Elliott AN, Crowder LB (2005) Marine Conservation Biology: The Science of Maintaining the Sea's Biodiversity. Island Press, Washington, DC.

Ellwood BB, Balsam WL, Roberts HH (2006a) Gulf of Mexico sediment sources and sediment transport trends from magnetic susceptibility measurements of surface samples. Marine Geology 230:237-248

Ellwood BB, Balsam WL, Roberts HH (2006b) Gulf of Mexico sediment sources and sediment transport trends from magnetic susceptibility measurements of surface samples. Marine Geology 230:237-248

Ellwood BB, Balsam WL, Roberts HH (2006c) Gulf of Mexico sediment sources and sediment transport trends from magnetic susceptibility measurements of surface samples. Marine Geology 230:237-248

Ercilla G, Casas D (2012) Submarine Mass Movements: Sedimentary Characterization and Controlling Factors, Earth Sciences. In: Dar DIA (ed) InTech

97 Etnoyer PJ, Wickes LN, Silva M, Dubick JD, Balthis L, Salgado E, MacDonald IR (2016) Decline in condition of gorgonian octocorals on mesophotic reefs in the northern Gulf of Mexico: before and after the Deepwater Horizon oil spill. Coral Reefs 35:77-90

Fabricius KE, De'ath G (2008) Photosynthetic symbionts and energy supply determine octocoral biodiversity in coral reefs. Ecology 89:3163-3173

Fisher CR, Demopoulos AWJ, Cordes EE, Baums IB, White HK, Bourque JR (2014a) Coral Communities as Indicators of Ecosystem-Level Impacts of the Deepwater Horizon Spill. Bioscience 64:796-807

Fisher CR, Hsing P-Y, Kaiser CL, Yoerger DR, Roberts HH, Shedd WW, Cordes EE, Shank TM, Berlet SP, Saunders MG, Larcom EA, Brooks JM (2014b) Footprint of Deepwater Horizon blowout impact to deep-water coral communities. Proceedings of the National Academy of Sciences of the United States of America 111:11744-11749

Folk RL (1965) The Petrology of Sedimentary Rocks. Hemphill Publishing Company, Austin, Texas

Gagne H, Lajeunesse P, St-Onge G, Bolduc A (2009) Recent transfer of coastal sediments to the Laurentian Channel, Lower St. Lawrence Estuary (Eastern Canada), through submarine canyon and fan systems. Geo-Marine Letters 29:191-200

Galloway WE (2008) DEPOSITIONAL EVOLUTION OF THE GULF OF MEXICO SEDIMENTARY BASIN. Sedimentary Basins of the United States and Canada 5:505-549

Galloway WE, McGilvery TA (1995a) Facies of a submarine canyon fill reservoir complex, Lower Wilcox group (Paleocene), central Texas coastal plain. Turbidites and Associated Deep- Water Facies:1-23

Galloway WE, McGilvery TA (1995b) Facies of a submarine canyon fill reservoir complex, Lower Wilcox group (Paleocene), central Texas coastal plain

Galy V, Beyssac O, France-Lanord C, Eglinton T (2008) Recycling of Graphite During Himalayan Erosion: A Geological Stabilization of Carbon in the Crust. Science 322:943-945

Garcia HE, Locarnini RA, Boyer TP, Antonov JI, Baranova OK, Zweng MM, Johnson DR (2010a) Dissolved Oxygen, Apparent Oxygen Utilization, and Oxygen Saturation. Book Volume 3. S, Levitus.

NOAA Atlas NESDIS 70, World Ocean Atlas 2009

Garcia HE, Locarnini RA, Boyer TP, Antonov JI, Baranova OK, Zweng MM, Johnson DR (2010b) Nutrients (phosphate, nitrate, silicate). Book Volume 4. S. Levitus

NOAA Atlas NESDIS 71., World Ocean Atlas 2009

98 Gardner JV, Dartnell P, Sulak KJ (2002) Multibeam Mapping of the Pinnacles Region, Gulf of Mexico: U.S. Geological Survey Open-File Report 02-006.

Gardner JV, Dartnell P, Sulak KJ, Calder BR, Hellequin L (2001a) Physiography and Late Quaternary-Holocene Processes of Northeastern Gulf of Mexico Outer Continental Shelf off Mississippi and Alabama. Marine Environmental Sciences Consortium of Alabama, Gulf of Mexico Science and International Journal

Gardner JV, Hughes Clarke JE, Mayer LA, Dartnell P (2003) Bathymetry and Acoustic Backscatter of the Mid and Outer Continental Shelf, Head of De Soto Canyon, Northeastern Gulf of Mexico—Data, Images, and GIS: U.S. Geological Survey Open-File Report 03-007.

Gardner JV, Mayer LA, Hughes Clarke JE, Dartnell, Peter., Sulak KJ (2001b) Cruise Report: RV Moana Wave cruise M1-01-GM; the bathymetry and acoustic backscatter of the mid shelf to upper slope off Panama City, Florida, northeastern Gulf of Mexico.

Gennesseaux M, Mauffret A, Pautot G (1980) SUBMARINE LANDSLIDES ON THE CONTINENTAL-SLOPE NEAR NICE AND THE BREAKING OF SUBMARINE CABLES IN THE LIGURIAN SEA (WESTERN MEDITERRANEAN. Comptes Rendus Hebdomadaires Des Seances De L Academie Des Sciences Serie D 290:959-962

Georgian S, Shedd W, Cordes E (2014) High-resolution ecological niche modelling of the cold- water coral Lophelia pertusa in the Gulf of Mexico. Mar Ecol Prog Ser 506:145-161

Georgian SE, DeLeo D, Durkin A, Gomez CE, Kurman M, Lunden JJ, Cordes EE (2016a) Oceanographic patterns and carbonate chemistry in the vicinity of cold-water coral reefs in the Gulf of Mexico: Implications for resilience in a changing ocean. Limnology and Oceanography 61:648-665

Georgian SE, DeLeo D, Durkin A, Gomez CE, Kurman M, Lunden JJ, Cordes EE (2016b) Oceanographic patterns and carbonate chemistry in the vicinity of cold-water coral reefs in the Gulf of Mexico: Implications for resilience in a changing ocean. Limnology and Oceanography 61:648-665

Gittings SR, Bright TJ, Schroeder WW, Sager WW, Laswell JS, Rezak R (1992) INVERTEBRATE ASSEMBLAGES AND ECOLOGICAL CONTROLS ON TOPOGRAPHIC FEATURES IN THE NORTHEAST GULF OF MEXICO. Bulletin of Marine Science 50:435- 455

Greene HG, O'Connell VM, Brylinsky CK (2011a) Tectonic and glacial related seafloor geomorphology as possible demersal shelf rockfish habitat surrogates-Examples along the Alaskan convergent transform plate boundary. Continental Shelf Research 31:S39-S53

Greene HG, O'Connell VM, Brylinsky CK (2011b) Tectonic and glacial related seafloor geomorphology as possible demersal shelf rockfish habitat surrogates-Examples along the Alaskan convergent transform plate boundary. Continental Shelf Research 31:S39-S53

99 Greene HG, O'Connell VM, Brylinsky CK (2011c) Tectonic and glacial related seafloor geomorphology as possible demersal shelf rockfish habitat surrogates-Examples along the Alaskan convergent transform plate boundary. Continental Shelf Research 31:S39-S53

Grilli ST, Taylor O-DS, Baxter CDP, Maretzki S (2009a) A probabilistic approach for determining submarine landslide tsunami hazard along the upper east coast of the United States. Marine Geology 264:74-97

Grilli ST, Taylor ODS, Baxter CDP, Maretzki S (2009b) A probabilistic approach for determining submarine landslide tsunami hazard along the upper east coast of the United States. Marine Geology 264:74-97

Guisan A, Edwards TC, Hastie T (2002) Generalized linear and generalized additive models in studies of species distributions: setting the scene. Ecological Modelling 157:89-100

Guisan A, Zimmermann NE (2000) Predictive habitat distribution models in ecology. Ecological Modelling 135:147-186

Haflidason H, Lien R, Sejrup HP, Forsberg CF, Bryn P (2005a) The dating and morphometry of the Storegga Slide. Marine and Petroleum Geology 22:123-136

Haflidason H, Lien R, Sejrup HP, Forsberg CF, Bryn P (2005b) The dating and morphometry of the Storegga Slide. Marine and Petroleum Geology 22:123-136

Haflidason H, Sejrup HP, Nygard A, Mienert J, Bryn P, Lien R, Forsberg CF, Berg K, Masson D (2004a) The Storegga Slide: architecture, geometry and slide development. Marine Geology 213:201-234

Haflidason H, Sejrup HP, Nygard A, Mienert J, Bryn P, Lien R, Forsberg CF, Berg K, Masson D (2004b) The Storegga Slide: architecture, geometry and slide development. Marine Geology 213:201-234

Hallock P, Schlager W (1986) Nutrients excess and the demise of coral reefs and carbonate platforms. Palaios 1:389-398

Harbitz CB, Lovholt F, Bungum H (2014a) Submarine landslide tsunamis: how extreme and how likely? Natural Hazards 72:1341-1374

Harbitz CB, Lovholt F, Bungum H (2014b) Submarine landslide tsunamis: how extreme and how likely? Natural Hazards 72:1341-1374

Harbitz CB, Lovholt F, Bungum H (2014c) Submarine landslide tsunamis: how extreme and how likely? Natural Hazards 72:1341-1374

Hinderstein LM, Marr JCA, Martinez FA, Dowgiallo MJ, Puglise KA, Pyle RL, Zawada DG, Appeldoorn R (2010) Theme section on “Mesophotic Coral Ecosystems: Characterization, Ecology, and Management”. Coral Reefs 29:247-251

100 Hobson RD (1972) Surface roughness in topography: quantitative approach. In: Chorley RJ (ed) In: Spatial analysis in geomorphology, New York, New York, USA.

Howell KL, Holt R, Endrino IP, Stewart H (2011) When the species is also a habitat: Comparing the predictively modelled distributions of Lophelia pertusa and the reef habitat it forms. Biological Conservation 144:2656-2665

Hsing P-Y, Fu B, Larcom EA, Berlet SP, Shank TM, Govindarajan AF, Lukasiewicz AJ, Dixon PM, Fisher CR (2013) Evidence of lasting impact of the Deepwater Horizon oil spill on a deep Gulf of Mexico coral community. Elementa-Science of the Anthropocene 1:000012-Article No.: 000012

Janocko M, Cartigny MBJ, Nemec W, Hansen EWM Marine and Petroleum Geology 41:222- 249

Jaynes ET (1957a) Information Theory and Statistical Mechanics. 106:620-630

Jaynes ET (1957b) Information Theory and Statistical Mechanics. II. 108:171-190

Jenkins C (2011a) Dominant Bottom Types and Habitats. In: Gulf of Mexico Data Atlas Stennis Space Center (MS): National Coastal Data Development Center

Jenkins C (2011b) Seabed Sediment Folk Codes. In: Gulf of Mexico Data Atlas. Stennis Space Center (MS): National Coastal Data Development Center

Jenness J Topographic Position and Landforms Analysis. Poster presentation. Proc ESRI User Conference

Jenness J (2006) Topographic Position Index (tpi_jen.avx) extension for ArcView. Jenness Enterprises.

Joye SB, MacDonald IR, Leifer I, Asper V (2011) Magnitude and oxidation potential of hydrocarbon gases released from the BP oil well blowout. Nature Geoscience 4:160-164

Kahng SE, Garcia-Sais JR, Spalding HL, Brokovich E, Wagner D, Weil E, Hinderstein L, Toonen RJ (2010) Community ecology of mesophotic coral reef ecosystems. Coral Reefs 29:255-275

Keevil GM, Peakall J, Best JL, Amos KJ (2006a) Flow structure in sinuous submarine channels: Velocity and turbulence structure of an experimental submarine channel. Marine Geology 229:241-257

Keevil GM, Peakall J, Best JL, Amos KJ (2006b) Flow structure in sinuous submarine channels: Velocity and turbulence structure of an experimental submarine channel. Marine Geology 229:241-257

101 Keller G (1985) EOCENE AND OLIGOCENE STRATIGRAPHY AND EROSIONAL UNCONFORMITIES IN THE GULF-OF-MEXICO AND GULF-COAST. Journal of Paleontology 59:882-903

Kostylev VE, Todd BJ, Fader GBJ, Courtney RC, Cameron GDM, Pickrill RA (2001) Benthic habitat mapping on the Scotian Shelf based on multibeam bathymetry, surficial geology and sea floor photographs. Marine Ecology Progress Series 219:121-137

Kvalstad TJ, Andresen L, Forsberg CF, Berg K, Bryn P, Wangen M (2005a) The Storegga slide: evaluation of triggering sources and slide mechanics. Marine and Petroleum Geology 22:245-256

Kvalstad TJ, Andresen L, Forsberg CF, Berg K, Bryn P, Wangen M (2005b) The Storegga slide: evaluation of triggering sources and slide mechanics. Marine and Petroleum Geology 22:245-256

L'Heureux J-S, Longva O, Steiner A, Hansen L, Vardy ME, Vanneste M, Hallidason H, Brendryen J, Kvalstad TJ, Forsberg CF, Chand S, Kopf A (2012a) Identification of Weak Layers and Their Role for the Stability of Slopes at Finneidfjord, Northern Norway. Submarine Mass Movements and Their Consequences 31:321-330

L'Heureux J-S, Longva O, Steiner A, Hansen L, Vardy ME, Vanneste M, Hallidason H, Brendryen J, Kvalstad TJ, Forsberg CF, Chand S, Kopf A (2012b) Identification of Weak Layers and Their Role for the Stability of Slopes at Finneidfjord, Northern Norway. Submarine Mass Movements and Their Consequences 31:321-330

L'Heureux J-S, Longva O, Steiner A, Hansen L, Vardy ME, Vanneste M, Hallidason H, Brendryen J, Kvalstad TJ, Forsberg CF, Chand S, Kopf A (2012c) Identification of Weak Layers and Their Role for the Stability of Slopes at Finneidfjord, Northern Norway. Submarine Mass Movements and Their Consequences 31:321-330

Lamarche G, Mountjoy J, Bull S, Hubble T, Krastel S, Lane E, Micallef A, Moscardelli L, Mueller C, Pecher I, Woelz S (2016) Submarine Mass Movements and Their Consequences: Progress and Challenges. Submarine Mass Movements and Their Consequences 41:1-12

Larsson AI, Jarnegren J, Stromberg SM, Dahl MP, Lundalv T, Brooke S (2014) Embryogenesis and Larval Biology of the Cold-Water Coral Lophelia pertusa. Plos One 9

Lastras G, Canals M, Urgeles R, De Batist M, Calafat AM, Casamor JL (2004a) Characterisation of the recent BIG'95 debris flow deposit on the Ebro margin, Western Mediterranean Sea, after a variety of seismic reflection data. Marine Geology 213:235-255

Lastras G, Canals M, Urgeles R, De Batist M, Calafat AM, Casamor JL (2004b) Characterisation of the recent BIG'95 debris flow deposit on the Ebro margin, Western Mediterranean Sea, after a variety of seismic reflection data. Marine Geology 213:235-255

Lawler JJ, Wiersma YF, Huettmann F (2011) Using Species Distribution Models for Conservation Planning and Ecological Forecasting

102 Lee HJ (2009a) Timing of occurrence of large submarine landslides on the Atlantic Ocean margin. Marine Geology 264:53-64

Lee HJ (2009b) Timing of occurrence of large submarine landslides on the Atlantic Ocean margin. Marine Geology 264:53-64

Lee HJ (2009c) Timing of occurrence of large submarine landslides on the Atlantic Ocean margin. Marine Geology 264:53-64

Lee HJ, Locat J, Desgagnes P, Parsons JD, McAdoo BG, Orange DL, Puig P, Wong FL, Dartnell P, Boulanger E (2007a) Submarine mass movements on continental margins. In: Nittrouer CA, Austin JA, Field ME, Kravitz JH, Syvitski JPM, Wiberg PL (eds) Continental Margin Sedimentation: from Sediment Transport to Sequence Stratigraphy, Book 37

Lee HJ, Locat J, Desgagnes P, Parsons JD, McAdoo BG, Orange DL, Puig P, Wong FL, Dartnell P, Boulanger E (2007b) Submarine mass movements on continental margins. In: Nittrouer CA, Austin JA, Field ME, Kravitz JH, Syvitski JPM, Wiberg PL (eds) Continental Margin Sedimentation: from Sediment Transport to Sequence Stratigraphy, Book 37

Lee HJ, Locat J, Desgagnés P, Parsons JD, McAdoo BG, Orange DL, Puig P, Wong FL, Dartnell P, Boulanger E Submarine Mass Movements on Continental Margins.213-274

Leynaud D, Mienert J, Vanneste M (2009a) Submarine mass movements on glaciated and non- glaciated European continental margins: A review of triggering mechanisms and preconditions to failure. Marine and Petroleum Geology 26:618-632

Leynaud D, Mienert J, Vanneste M (2009b) Submarine mass movements on glaciated and non- glaciated European continental margins: A review of triggering mechanisms and preconditions to failure. Marine and Petroleum Geology 26:618-632

Leynaud D, Mienert J, Vanneste M (2009c) Submarine mass movements on glaciated and non- glaciated European continental margins: A review of triggering mechanisms and preconditions to failure. Marine and Petroleum Geology 26:618-632

Leynaud D, Mienert J, Vanneste M (2009d) Submarine mass movements on glaciated and non- glaciated European continental margins: A review of triggering mechanisms and preconditions to failure. Marine and Petroleum Geology 26:618-632

Li W, Alves TM, Wu S, Voelker D, Zhao F, Mi L, Kopf A (2015) Recurrent slope failure and submarine channel incision as key factors controlling reservoir potential in the South China Sea (Qiongdongnan Basin, South Hainan Island). Marine and Petroleum Geology 64:17-30

Lidz BH, Reich CD, Shinn EA (2003) Regional Quaternary submarine geomorphology in the Florida Keys. Geological Society of America Bulletin 115:845-866

Lindhorst K, Krastel S, Papenberg C, Heidarzadeh M (2014a) Modeling Submarine Landslide- Generated Waves in Lake Ohrid, Macedonia/Albania. Submarine Mass Movements and Their Consequences: 6th International Symposium 37:497-506

103 Lindhorst K, Krastel S, Papenberg C, Heidarzadeh M (2014b) Modeling Submarine Landslide- Generated Waves in Lake Ohrid, Macedonia/Albania. Submarine Mass Movements and Their Consequences: 6th International Symposium 37:497-506

Lindhorst K, Krastel S, Papenberg C, Heidarzadeh M (2014c) Modeling Submarine Landslide- Generated Waves in Lake Ohrid, Macedonia/Albania. Submarine Mass Movements and Their Consequences: 6th International Symposium 37:497-506

Locat J (2001) Instabilities along ocean margins: a geomorphological and geotechnical perspective. Marine and Petroleum Geology 18:503-512

Locat J, Lee H (2009) Submarine Mass Movements and Their Consequences: An Overview

Locker SD, Armstrong RA, Battista TA, Rooney JJ, Sherman C, Zawada DG (2010) Geomorphology of mesophotic coral ecosystems: current perspectives on morphology, distribution, and mapping strategies. Coral Reefs 29:329-345

MacDonald IR, Garcia-Pineda O, Beet A, Daneshgar Asl S, Feng L, Graettinger G, French- McCay D, Holmes J, Hu C, Huffer F, Leifer I, Muller-Karger F, Solow A, Silva M, Swayze G (2015) Natural and unnatural oil slicks in the Gulf of Mexico. Journal of Geophysical Research: Oceans 120:8364-8380

Masson DG, Harbitz CB, Wynn RB, Pedersen G, Lovholt F (2006) Submarine landslides: processes, triggers and hazard prediction. Philosophical Transactions of the Royal Society a- Mathematical Physical and Engineering Sciences 364:2009-2039

Mayall M, Lonergan L, Bowman A, James S, Mills K, Primmer T, Pope D, Rogers L, Skeene R (2010) The response of turbidite slope channels to growth-induced seabed topography. Aapg Bulletin 94:1011-1030

McAdoo BG, Pratson LF, Orange DL (2000a) Submarine landslide geomorphology, US continental slope. Marine Geology 169:103-136

McAdoo BG, Pratson LF, Orange DL (2000b) Submarine landslide geomorphology, US continental slope. Marine Geology 169:103-136

McAdoo BG, Pratson LF, Orange DL (2000c) Submarine landslide geomorphology, US continental slope. Marine Geology 169:103-136

McGregor BA Submarine canyon and slope processes of the U.S. Atlantic continental margin.

McNutt MK, Camilli R, Crone TJ, Guthrie GD, Hsieh PA, Ryerson TB, Savas O, Shaffer F (2012) Review of flow rate estimates of the Deepwater Horizon oil spill. Proceedings of the National Academy of Sciences 109:20260-20267

Meckler AN, Ren HJ, Sigman DM, Gruber N, Plessen B, Schubert CJ, Haug GH (2011) Deglacial nitrogen isotope changes in the Gulf of Mexico: Evidence from bulk sedimentary and foraminifera-bound nitrogen in Orca Basin sediments. Paleoceanography 26

104 Menard HW (1955a) 39:236

Menard HW (1955b) Deep-sea channels, topography, and sedimentation. AAPG Bulletin 39:236-255

Miall AD (2000) Principles of Sedimentary Basin Analysis. Springer-Velag, Berlin

Monk J, Ierodiaconou D, Versace Vl, Bellgrove A, Harvey E, Rattray A, Laurenson L, Quinn Gp (2010) Habitat suitability for marine fishes using presence-only modelling and multibeam sonar. Mar Ecol Prog Ser 420:157-174

Mosher DC, Moscardelli L, Shipp RC, Chaytor JD, Baxter CDP, Lee HJ, Urgeles R (2010) Submarine Mass Movements and Their Consequences, Vol 28

Mulcan A, Mitsova D, Hindle T, Hanson H, Coley C (2015) Marine Benthic Habitats and Seabed Suitability Mapping for Potential Ocean Current Energy Siting Offshore Southeast Florida. 3:276-298

Muller-Karger FEUSMM, Service Gulf of Mexico OCS Region. University of South Florida, Science. DoM (2000) Remote sensing study of upwelling in the northeastern Gulf of Mexico and the effects of Hurricanes Earl and Georges annual report: year 2., New Orleans, La : U.S. Dept. of the Interior, Minerals Management Service, Gulf of Mexico OCS Region

Nixon MF, Grozic JLH (2006a) A simple model for submarine slope stability analysis with gas hydrates. Norwegian Journal of Geology 86:309-316

Nixon MF, Grozic JLH (2006b) A simple model for submarine slope stability analysis with gas hydrates. Norwegian Journal of Geology 86:309-316

NOAA (2012) Multibeam Sonar Bathymetry Data collected aboard Okeanos Explorer (EX1202Leg1). NOAA National Centers for Environmental Information. In: (NOAA) NOAA (ed). DOC/NOAA/NESDIS/NCEI > National Centers for Environmental Information, NESDIS, NOAA,U.S. Department of Commerce DOC/NOAA/NESDIS/NGDC > National Geophysical Data Center, NESDIS, NOAA, U.S. Departmentof Commerce

NOAA (2013) ERMA Deepwater Gulf Response. NOAA Office of Response and Restoration., http://gomex.erma.noaa.gov/

NOAA Coastal Watch http://coastwatch.noaa.gov/.

Normark WR, Carlson PR (2003) Giant submarine canyons: Is size any clue to their importance in the rock record? Geological Society of America Special Papers 370:175-190

Obelcz J, Brothers D, Chaytor J, ten Brink U, Ross SW, Brooke S (2014) Geomorphic characterization of four shelf-sourced submarine canyons along the US Mid-Atlantic continental margin. Deep-Sea Research Part Ii-Topical Studies in Oceanography 104:106-119

105 Olsen PM, Kolden CA, Gadamus L (2015) Developing Theoretical Marine Habitat Suitability Models from Remotely-Sensed Data and Traditional Ecological Knowledge. 7:11863-11886

Ottaviani D, Lasinio GJ, Boitani L (2004) Two statistical methods to validate habitat suitability models using presence-only data. Ecological Modelling 179:417-443

Owen M, Day S, Maslin M (2007a) Late Pleistocene submarine mass movements: occurrence and causes. Quaternary Science Reviews 26:958-978

Owen M, Day S, Maslin M (2007b) Late Pleistocene submarine mass movements: occurrence and causes. Quaternary Science Reviews 26:958-978

Pearce JL, Boyce MS (2006) Modelling distribution and abundance with presence-only data. Journal of Applied Ecology 43:405-412

Peccini MB, MacDonald IR (2008) Correspondence of sea fan orientations with measured currents on hard bottom habitats of the Mississippi/Alabama continental shelf. Continental Shelf Research 28:302-308

Perov G, Bhattacharya JP (2011) Pleistocene shelf-margin delta: Intradeltaic deformation and sediment bypass, northern Gulf of Mexico. Aapg Bulletin 95:1617-1641

Phillips SJ, Anderson RP, Schapire RE (2006) Maximum entropy modeling of species geographic distributions. Ecological Modelling 190:231-259

Phillips SJ, Dudik M (2008) Modeling of species distributions with Maxent: new extensions and a comprehensive evaluation. Ecography 31:161-175

Pini GA, Ogata K, Camerlenghi A, Festa A, Lucente CC, Codegone G (2012a) Sedimentary Melanges and Fossil Mass-Transport Complexes: A Key for Better Understanding Submarine Mass Movements? Submarine Mass Movements and Their Consequences 31:585-594

Pini GA, Ogata K, Camerlenghi A, Festa A, Lucente CC, Codegone G (2012b) Sedimentary Melanges and Fossil Mass-Transport Complexes: A Key for Better Understanding Submarine Mass Movements? Submarine Mass Movements and Their Consequences 31:585-594

Previati M, Scinto A, Cerrano C, Osinga R (2010) Oxygen consumption in Mediterranean octocorals under different temperatures. Journal of Experimental Marine Biology and Ecology 390:39-48

Quattrini AM, Etnoyer PJ, Doughty C, English L, Falco R, Renion N, Rittinghouse M, Cordes EE (2014) A phylogenetic approach to octocoral community structure in the deep Gulf of Mexico. Deep-Sea Research Part Ii-Topical Studies in Oceanography 99:92-102

Reich CD, Poore RZ, Hickey TD (2013) The Role of Vermetid Gastropods in the Development of the Florida Middle Ground, Northeast Gulf of Mexico. Journal of Coastal Research:46-57

106 Reiss H, Cunze S, Koenig K, Neumann H, Kroencke I (2011) Species distribution modelling of marine benthos: a North Sea case study. Marine Ecology Progress Series 442:71-86

Rengstorf AM, Yesson C, Brown C, Grehan AJ (2013) High-resolution habitat suitability modelling can improve conservation of vulnerable marine ecosystems in the deep sea. Journal of Biogeography 40:1702-1714

Roberts JM, Brown CJ, Long D, Bates CR (2005) Acoustic mapping using a multibeam echosounder reveals cold-water coral reefs and surrounding habitats. Coral Reefs 24:654-669

Roberts JM, Wheeler AJ, Freiwald A (2006a) Reefs of the deep: The biology and geology of cold-water coral ecosystems. Science 312:543-547

Roberts JM, Wheeler AJ, Freiwald A (2006b) Reefs of the deep: The biology and geology of cold-water coral ecosystems. Science 312:543-547

Rochelle-Newall EJ, Fisher TR (2002) Chromophoric dissolved organic matter and dissolved organic carbon in Chesapeake Bay. Marine Chemistry 77:23-41

Rodriguez AB, Simms AR, Anderson JB (2010a) Bay-head deltas across the northern Gulf of Mexico back step in response to the 8.2 ka cooling event. Quaternary Science Reviews 29:3983- 3993

Rodriguez AB, Simms AR, Anderson JB (2010b) Bay-head deltas across the northern Gulf of Mexico back step in response to the 8.2 ka cooling event. Quaternary Science Reviews 29:3983- 3993

Rogers AD (1999a) The biology of Lophelia pertusa (LINNAEUS 1758) and other deep-water reef-forming corals and impacts from human activities. International Review of Hydrobiology 84:315-406

Rogers AD (1999b) The biology of Lophelia pertusa (LINNAEUS 1758) and other deep-water reef-forming corals and impacts from human activities. International Review of Hydrobiology 84:315-406

Rogers AD, Baco A, Griffiths H, Hall-Spencer JM, Gjerde KM (2007) Corals on seamounts. In: Pitcher TJ, Morato T, Hart PJB, Clrck MR, Haggan N, Santos RS (eds) Seamounts: Ecology, Fisheries and Conservation. Blackwell Publishing Oxford

Rogers CS (1990) Responses of coral reefs organisms to sedimentation. Marine Ecology Progress Series 62:185-202

Romero IC, Schwing PT, Brooks GR, Larson RA, Hastings DW, Ellis G, Goddard EA, Hollander DJ (2015) Hydrocarbons in Deep-Sea Sediments following the 2010 Deepwater Horizon Blowout in the Northeast Gulf of Mexico. Plos One 10

Rosenheim BE, Pendergraft MA, Flowers GC, Carney R, Sericano JL, Amer RM, Chanton J, Dincer Z, Wade TL (2016) Employing extant stable carbon isotope data in Gulf of Mexico

107 sedimentary organic matter for oil spill studies. Deep-Sea Research Part Ii-Topical Studies in Oceanography 129:249-258

Rowe GT, Morse J, Nunnally C, Boland GS (2008) Sediment community oxygen consumption in the deep Gulf of Mexico. Deep-Sea Research Part Ii-Topical Studies in Oceanography 55:2686-2691

Santschi PH, Rowe GT (2008) Radiocarbon-derived sedimentation rates in the Gulf of Mexico. Deep-Sea Research Part Ii-Topical Studies in Oceanography 55:2572-2576

Sawyer DE, Flemings PB, Nikolinakou M (2014) Continuous deep-seated slope failure recycles sediments and limits levee height in submarine channels. Geology 42:15-18

Schnellmann M, Anselmetti FS, Giardini D, McKenzie JA (2005a) Mass movement-induced fold-and-thrust belt structures in unconsolidated sediments in Lake Lucerne (Switzerland). Sedimentology 52:271-289

Schnellmann M, Anselmetti FS, Giardini D, McKenzie JA (2005b) Mass movement-induced fold-and-thrust belt structures in unconsolidated sediments in Lake Lucerne (Switzerland). Sedimentology 52:271-289

Silva M, Etnoyer PJ, MacDonald IR (2016) Coral injuries observed at mesophotic Reefs after the Deepwater Horizon oil discharge. Deep Sea Research Part II: Topical Studies in Oceanography

Simms AR, DeWitt R, Rodriguez AB, Lambeck K, Anderson JB (2009) Revisiting marine isotope stage 3 and 5a (MIS3-5a) sea levels within the northwestern Gulf of Mexico. Global and Planetary Change 66:100-111

Simms AR, Lambeck K, Purcell A, Anderson JB, Rodriguez AB (2007a) Sea-level history of the Gulf of Mexico since the Last Glacial Maximum with implications for the melting history of the Laurentide Ice Sheet. Quaternary Science Reviews 26:920-940

Simms AR, Lambeck K, Purcell A, Anderson JB, Rodriguez AB (2007b) Sea-level history of the Gulf of Mexico since the Last Glacial Maximum with implications for the melting history of the Laurentide Ice Sheet. Quaternary Science Reviews 26:920-940

Sionneau T, Bout-Roumazeilles V, Biscaye PE, Van Vliet-Lanoe B, Bory A (2008) Clay mineral distributions in and around the Mississippi River watershed and Northern Gulf of Mexico: sources and transport patterns. Quaternary Science Reviews 27:1740-1751

Sionneau T, Bout-Roumazeilles V, Flower BP, Bory A, Tribovillard N, Kissel C, Van Vliet- Lanoe B, Serrano JCM (2010) Provenance of freshwater pulses in the Gulf of Mexico during the last deglaciation. Quaternary Research 74:235-245

Smith MD, Zhang J, Coleman FC (2006) Effectiveness of marine reserves for large-scale fisheries management. Can J Fish Aquat Sci 63:153-164

108 Stedmon CA, Markager S (2001) The optics of chromophoric dissolved organic matter (CDOM) in the Greenland Sea: An algorithm for differentiation between marine and terrestrially derived organic matter. Limnology and Oceanography 46:2087-2093

Stephens D, Diesing M (2014) A Comparison of Supervised Classification Methods for the Prediction of Substrate Type Using Multibeam Acoustic and Legacy

Grain-Size Data. 9:e93950

Sultan N, Cochonat P, Canals M, Cattaneo A, Dennielou B, Haflidason H, Laberg JS, Long D, Mienert J, Trincardi F, Urgeles R, Vorren TO, Wilson C (2004a) Triggering mechanisms of slope instability processes and sediment failures on continental margins: a geotechnical approach. Marine Geology 213:291-321

Sultan N, Cochonat P, Canals M, Cattaneo A, Dennielou B, Haflidason H, Laberg JS, Long D, Mienert J, Trincardi F, Urgeles R, Vorren TO, Wilson C (2004b) Triggering mechanisms of slope instability processes and sediment failures on continental margins: a geotechnical approach. Marine Geology 213:291-321

Sultan N, Cochonat P, Foucher JP, Mienert J (2004c) Effect of gas hydrates melting on seafloor slope instability. Marine Geology 213:379-401

Sultan N, Gaudin M, Berne S, Canals M, Urgeles R, Lafuerza S (2007a) Analysis of slope failures in submarine canyon heads: An example from the Gulf of Lions. Journal of Geophysical Research-Earth Surface 112

Sultan N, Gaudin M, Berne S, Canals M, Urgeles R, Lafuerza S (2007b) Analysis of slope failures in submarine canyon heads: An example from the Gulf of Lions. Journal of Geophysical Research-Earth Surface 112

Talling PJ, Clare M, Urlaub M, Pope E, Hunt JE, Watt SFL (2014) Large Submarine Landslides on Continental Slopes Geohazards, Methane Release, and Climate Change. Oceanography 27:32-45

Talling PJ, Geohazards and Sedimentology Research Group NOCSUK Large Submarine Landslides on Continental Slopes: Geohazards, Methane Release, and Climate Change. Oceanography 27

Tappin DR (2010a) Submarine mass failures as tsunami sources: their climate control. Philosophical Transactions of the Royal Society a-Mathematical Physical and Engineering Sciences 368:2417-2434

Tappin DR (2010b) Submarine mass failures as tsunami sources: their climate control. Philosophical Transactions of the Royal Society a-Mathematical Physical and Engineering Sciences 368:2417-2434

109 Taylor ML, Yesson C, Agnew DJ, Mitchell RE, Rogers AD (2013) Using fisheries by-catch data to predict octocoral habitat suitability around South Georgia. Journal of Biogeography 40:1688- 1701 ten Brink US, Lee HJ, Geist EL, Twichell D (2009a) Assessment of tsunami hazard to the US East Coast using relationships between submarine landslides and earthquakes. Marine Geology 264:65-73 ten Brink US, Lee HJ, Geist EL, Twichell D (2009b) Assessment of tsunami hazard to the US East Coast using relationships between submarine landslides and earthquakes. Marine Geology 264:65-73 ten Brink US, Lee HJ, Geist EL, Twichell D (2009c) Assessment of tsunami hazard to the US East Coast using relationships between submarine landslides and earthquakes. Marine Geology 264:65-73 ten Brink US, Lee HJ, Geist EL, Twichell D (2009d) Assessment of tsunami hazard to the US East Coast using relationships between submarine landslides and earthquakes. Marine Geology 264:65-73

Thiem O, Ravagnan E, Fossa JH, Berntsen J (2006) Food supply mechanisms for cold-water corals along a continental shelf edge. Journal of Marine Systems 60:207-219

Tittensor DP, Baco AR, Brewin PE, Clark MR, Consalvey M, Hall-Spencer J, Rowden AA, Schlacher T, Stocks KI, Rogers AD (2009) Journal of Biogeography 36:1111-1128

Tong R, Purser A, Guinan J, Unnithan V (2013) Modeling the habitat suitability for deep-water gorgonian corals based on terrain variables. Ecological Informatics 13:123-132

Tracey DM, Rowden AA, Mackay KA, Compton T (2011) Habitat-forming cold-water corals show affinity for seamounts in the New Zealand region. Marine Ecology Progress Series 430:1- U59

Tripsanas EK, Bryant WR, Slowey NC, Bouma AH, Karageorgis AP, Berti D (2007a) Sedimentological history of Bryant Canyon area, northwest Grulf of Mexico, during the last 135 kyr (Marine Isotope Stages 1-6): A proxy record of Mississippi River discharge. Palaeogeography Palaeoclimatology Palaeoecology 246:137-161

Tripsanas EK, Bryant WR, Slowey NC, Bouma AH, Karageorgis AP, Berti D (2007b) Sedimentological history of Bryant Canyon area, northwest Grulf of Mexico, during the last 135 kyr (Marine Isotope Stages 1-6): A proxy record of Mississippi River discharge. Palaeogeography Palaeoclimatology Palaeoecology 246:137-161

Tseng L-C, Dahms H-U, Hsu N-J, Hwang J-S (2011) Effects of sedimentation on the gorgonian Subergorgia suberosa (Pallas, 1766). Marine Biology 158:1301-1310

110 Tubau X, Canals M, Lastras G, Rayo X, Rivera J, Amblas D (2015) Marine litter on the floor of deep submarine canyons of the Northwestern Mediterranean Sea: The role of hydrodynamic processes. Progress in Oceanography 134:379-403

Twichell DC, Roberts DG (1982) MORPHOLOGY, DISTRIBUTION, AND DEVELOPMENT OF SUBMARINE CANYONS ON THE UNITED-STATES ATLANTIC CONTINENTAL- SLOPE BETWEEN HUDSON AND BALTIMORE CANYONS. Geology 10:408-412

Urgeles R, Camerlenghi A (2013a) Submarine landslides of the Mediterranean Sea: Trigger mechanisms, dynamics, and frequency-magnitude distribution. Journal of Geophysical Research- Earth Surface 118:2600-2618

Urgeles R, Camerlenghi A (2013b) Submarine landslides of the Mediterranean Sea: Trigger mechanisms, dynamics, and frequency-magnitude distribution. Journal of Geophysical Research- Earth Surface 118:2600-2618

Urgeles R, Camerlenghi A (2013c) Submarine landslides of the Mediterranean Sea: Trigger mechanisms, dynamics, and frequency-magnitude distribution. Journal of Geophysical Research- Earth Surface 118:2600-2618

Urgeles R, Leynaud D, Lastras G, Canals M, Mienert J (2006a) Back-analysis and failure mechanisms of a large submarine slide on the ebro slope, NW Mediterranean. Marine Geology 226:185-206

Urgeles R, Leynaud D, Lastras G, Canals M, Mienert J (2006b) Back-analysis and failure mechanisms of a large submarine slide on the ebro slope, NW Mediterranean. Marine Geology 226:185-206

Urgeles R, Leynaud D, Lastras G, Canals M, Mienert J (2006c) Back-analysis and failure mechanisms of a large submarine slide on the ebro slope, NW Mediterranean. Marine Geology 226:185-206

Urlaub M, Talling PJ, Zervos A, Masson D (2015a) What causes large submarine landslides on low gradient (< 2 degrees) continental slopes with slow (similar to 0.15 m/kyr) sediment accumulation? Journal of Geophysical Research-Solid Earth 120:6722-6739

Urlaub M, Talling PJ, Zervos A, Masson D (2015b) What causes large submarine landslides on low gradient (< 2 degrees) continental slopes with slow (similar to 0.15 m/kyr) sediment accumulation? Journal of Geophysical Research-Solid Earth 120:6722-6739

USGS (2011) BP Deepwater Horizon Oil Spill: Incident Specific Preparedness Review. Final Report. https://www.uscg.mil/foia/docs/DWH/BPDWH.pdf

Vachtman D, Mitchell NC, Gawthorpe R (2013) Morphologic signatures in submarine canyons and gullies, central USA Atlantic continental margins. Marine and Petroleum Geology 41:250- 263

111 Vanoudheusden E, Sultan N, Cochonat P (2004a) Mechanical behaviour of unsaturated marine sediments: experimental and theoretical approaches. Marine Geology 213:323-342

Vanoudheusden E, Sultan N, Cochonat P (2004b) Mechanical behaviour of unsaturated marine sediments: experimental and theoretical approaches. Marine Geology 213:323-342

Vanoudheusden E, Sultan N, Cochonat P (2004c) Mechanical behaviour of unsaturated marine sediments: experimental and theoretical approaches. Marine Geology 213:323-342

Vanoudheusden E, Sultan N, Cochonat P (2004d) Mechanical behaviour of unsaturated marine sediments: experimental and theoretical approaches. Marine Geology 213:323-342

Vanoudheusden E, Sultan N, Cochonat P (2004e) Mechanical behaviour of unsaturated marine sediments: experimental and theoretical approaches. Marine Geology 213:323-342

Vladimir E. Kostylev, Brian J. Todd, Gordon B. J. Fader, R. C. Courtney, Gordon D. M. Cameron, Richard A. Pickrill (2001) Mar Ecol Prog Ser 219:121-137

Voelker D, Scholz F, Geersen J (2011) Analysis of submarine landsliding in the rupture area of the 27 February 2010 Maule earthquake, Central Chile. Marine Geology 288:79-89

Vogel JS, Southon JR, Nelson DE, Brown TA (1984) PERFORMANCE OF CATALYTICALLY CONDENSED CARBON FOR USE IN ACCELERATOR MASS- SPECTROMETRY. Nuclear Instruments & Methods in Physics Research Section B-Beam Interactions with Materials and Atoms 5:289-293

Wagner AJ (2009) Oxygen and Carbon Isotopes and Coral Growth in the Gulf of Mexico and Caribbean Sea as Environmental and Climate Indicators. . Doctoral dissertation, Texas A&M University,

Wagner B, Ortlepp S, Kenig F, Doran PT, Melles M (2010) Palaeoenvironmental implications derived from a piston core from east lobe Bonney, Taylor Valley, Antarctica. Antarctic Science 22:522-530

Weaver DC, Dennis GD, Sulak KJ (2002) Community Structure and Trophic Ecology of Fishes on the Pinnacles Reef Tract. USGS

Weinbauer MG, Velimirov B (1996) Relative Habitat Suitability and Stability of the Mediterranean Gorgonian Coral Eunicella Cavolini (Coelenterata: Octocorallia). Bulletin of Marine Science 58:786-791

Weiss A Topographic Position and Landforms Analysis. Poster presentation. Proc ESRI User Conference

Wheeler AJ, Beyer A, Freiwald A, de Haas H, Huvenne VAI, Kozachenko M, Roy KOL, Opderbecke J (2007) Morphology and environment of cold-water coral carbonate mounds on the NW European margin. International Journal of Earth Sciences 96:37-56

112 White HK, Hsing P-Y, Cho W, Shank TM, Cordes EE, Quattrini AM, Nelson RK, Camilli R, Demopoulos AWJ, German CR, Brooks JM, Roberts HH, Shedd W, Reddy CM, Fisher CR (2012) Impact of the Deepwater Horizon oil spill on a deep-water coral community in the Gulf of Mexico. Proceedings of the National Academy of Sciences of the United States of America 109:20303-20308

Wilson MFJ, O'Connell B, Brown C, Guinan JC, Grehan AJ (2007) Multiscale terrain analysis of multibeam bathymetry data for habitat mapping on the continental slope. Marine Geodesy 30:3- 35

Woodroffe CD, Webster JM (2014) Coral reefs and sea-level change. Marine Geology 352:248- 267

Wright DJ, Lundblad ER, Larkin EM, Rinehart RW Oregon State University Davey Jones Locker Seafloor Mapping/Marine GIS Lab (project background at http://dusk.geo.orst.edu/esri04/p1433_ron.html). In: Joshua Murphy, Lori Cary-Kothera, Draganov K (eds). NOAA Coastal Services Center

Yamada Y, Kawamura K, Ikehara K, Ogawa Y, Urgeles R, Mosher D, Chaytor J, Strasser M (2012) Submarine Mass Movements and Their Consequences. Submarine Mass Movements and Their Consequences 31:1-12

Yesson C, Taylor ML, Tittensor DP, Davies AJ, Guinotte J, Baco A, Black J, Hall-Spencer JM, Rogers AD (2012) Global habitat suitability of cold-water octocorals. Journal of Biogeography 39:1278-1292

Yuan F, Wang L, Guo Z, Shi R (2012a) A refined analytical model for landslide or debris flow impact on pipelines. Part I: Surface pipelines. Applied Ocean Research 35:95-104

Yuan F, Wang L, Guo Z, Shi R (2012b) A refined analytical model for landslide or debris flow impact on pipelines. Part I: Surface pipelines. Applied Ocean Research 35:95-104

Yuan F, Wang L, Guo Z, Xie Y (2012c) A refined analytical model for landslide or debris flow impact on pipelines - Part II: Embedded pipelines. Applied Ocean Research 35:105-114

Yuan F, Wang L, Guo Z, Xie Y (2012d) A refined analytical model for landslide or debris flow impact on pipelines - Part II: Embedded pipelines. Applied Ocean Research 35:105-114

Yuan F, Wang LZ, Guo Z, Shi RW (2012e) A refined analytical model for landslide or debris flow impact on pipelines. Part I: Surface pipelines. Applied Ocean Research 35:95-104

Yuan F, Wang LZ, Guo Z, Xie YG (2012f) A refined analytical model for landslide or debris flow impact on pipelines - Part II: Embedded pipelines. Applied Ocean Research 35:105-114

Zainul M (2010) Coral Reef Rugosity and Coral Biodiversity. Master, International Insitute for Geo-information Science and Earth Observation Encshede,

Zakeri A (2008) Journal of Offshore Mechanics and Arctic Engineering 131:014001-014001

113 Zakeri A (2009a) Submarine debris flow impact on suspended (free-span) pipelines: Normal and longitudinal drag forces. Ocean Engineering 36:489-499

Zakeri A (2009b) Submarine debris flow impact on suspended (free-span) pipelines: Normal and longitudinal drag forces. Ocean Engineering 36:489-499

Zavala-Hidalgo J, Gallegos-Garcia A, Martinez-Lopez B, Morey SL, O'Brien JJ (2006) Seasonal upwelling on the Western and Southern Shelves of the Gulf of Mexico. Ocean Dynamics 56:333- 338

Zhang M, Huang Y, Bao Y (2016a) The mechanism of shallow submarine landslides triggered by storm surge. Natural Hazards 81:1373-1383

Zhang M, Huang Y, Bao YJ (2016b) The mechanism of shallow submarine landslides triggered by storm surge. Natural Hazards 81:1373-1383

Zitter TAC, Grall C, Henry P, Ozeren MS, Cagatay MN, Sengor AMC, Gasperini L, de Lepinay BM, Geli L (2012a) Distribution, morphology and triggers of submarine mass wasting in the Sea of Marmara. Marine Geology 329:58-74

Zitter TAC, Grall C, Henry P, Ozeren MS, Cagatay MN, Sengor AMC, Gasperini L, de Lepinay BM, Geli L (2012b) Distribution, morphology and triggers of submarine mass wasting in the Sea of Marmara. Marine Geology 329:58-74

Zuckerberg B, Huettmann F, Frair J (2011) Proper Data Management as a Scientific Foundation for Reliable Species Distribution Modeling. Predictive Species and Habitat Modeling in Landscape Ecoloogy: Concepts and Applications:45-70

114 BIOGRAPHICAL SKETCH

CURRENT POSITION Ph.D . Candidate at Florida State University. Earth, Ocean and Atmospheric Science Department. Major Professor Dr. Ian MacDonald.

EDUCATION Academic Degree : B.A. Ocean Science (Highest Honors) Arturo Prat University. Major : Marine Biology, Arturo Prat University. Thesis : “Temporal Variability of the Biomass Size Spectra of soft bottom benthic community of Patillos Bay (20° 46' S; 70° 14' W)” Advisor Dr. Eduardo Quiroga.

HONORS AND AWARDS 2011-2016: Latin American - Caribbean Scholarship at Florida State University. (tuition support) 2011-2015: Fulbright CONICYT Scholarship to obtain Ph.D. degree in the United States of America (4 year full support) 2007: Academic Excellency, Highest Honor to Best the Graduate of 2007 generation, Arturo Prat University. 2004: Undergraduate Research Grant, Benthos Laboratory, to finance Undergraduate Thesis.

PROJECT LIST Ecosystem Impacts of Oil and Gas Inputs to the Gulf-2 (ECOGIG-2). Year 5-7 Consortia Grants (RFP-IV) Role: Graduate Student - PhD Level, GoMRI Scholar Deep-sea to Coast Connectivity in the Eastern Gulf of Mexico (DEEP-C). Year 2-4 Consortia Grants (RFP-I) Role: Graduate Student - PhD Level. Ecosystem Impacts of Oil and Gas Inputs to the Gulf (ECOGIG). Year 2-4 Consortia Grants (RFP-I) Role: Graduate Student - PhD Level, GoMRI Scholar Coast Watch: Remote Sensing and Verification Sampling of Oil Spill Impact on Florida Coast. Year One Block Grant - Florida Institute of Oceanography. Role: Graduate Student - PhD Level

ASSISTANT TEACHING 2015: Teaching Assistant for Environmental Science, Laboratory Module (EVR1001L). Instructor Professor Dr. William Landing. Florida State University 2015: Teaching Assistant for Elementary Oceanography (OCE1001). Instructors Professor Dr. Ian MacDonald and Professor Dr. Allan Clarke. Florida State University

115 2004 – 2007: Teaching Assistant for Benthos. Instructor Professor Raul Soto. Ocean Science Department, Arturo Prat University. 2006: Teaching assistant for Biology of Marine Decapods. Mecesup Project: Instructor Professor Dr. Enrique Dupré and Dr. Guillermo Guzmán. 2004 – 2005: Teaching Assistant for Zoology of Marine Invertebrates. Instructor Professor Raul Soto. Ocean Science Department, Arturo Prat University. 2004: Teaching assistant for Biology of Invertebrates. Instructor Professor Raul Soto. Ocean Science Department, Arturo Prat University. 2003 – 2004: Teaching assistant for Zoology I. Instructor Professor Raul Soto. Ocean Science Department, Arturo Prat University. 2002 – 2003: Teaching assistant for Parasitology. Instructor Professor Raul Soto. Ocean Science Department, Arturo Prat University.

COURSES PARTICIPATION 2005: International Course “Molecular Technics applied to study of ecology of Microbial community on Natural Environments” Professors: Ph.D. Ramón Rosello-Mora, IMADEA, CSIC- University of Baleares Island, Spain; Ph.D. Roberto Vidal Álvarez, ICBM, Medicine Faculty, University of Chile and M.Sc. Rubén Moraga Mamani, Ocean Science department, Arturo Prat University. 2004: “Molecular Biology and Genomics technics applied to Study of Microorganisms. Professor Ph.D. Roberto Vidal Alvarez, ICBM, Medicine Faculty, University of Chile and M.Sc. Rubén Moraga Mamani, Ocean Science department, Arturo Prat University.

CONFERENCES PARTICIPATION • Garcia Pineda, O. M., MacDonald, I., Shedd, W., & Silva, M. (2015). Transience and Persistence of Natural Hydrocarbon Seepage in Mississippi Canyon, Gulf of Mexico. In Gulf of Mexico Oil Spill & Ecosystem Science Conference 2015. Houston, TX. Poster. • Johansen, C., Todd, A. C., Silva, M., Shedd, W., & MacDonald, I. R. (2015). Variability and quantification of oil and gas bubble release from natural seeps in the Gulf of Mexico. In Gulf of Mexico Oil Spill & Ecosystem Science Conference 2015. Houston, TX. • Silva-Aguilera, M. G., Etnoyer, P., & MacDonald, I. (2015). Coral injuries observed at Mesophotic Coral Communities following the Deepwater Horizon oil discharge. In Gulf of Mexico Oil Spill & Ecosystem Science Conference 2015. Houston, TX. • Silva-Aguilera, M. G., Etnoyer, P., & MacDonald, I. (2015). Monitoring Recovery of Mesophotic Corals: 2011-2014. In Gulf of Mexico Oil Spill & Ecosystem Science Conference 2015. Houston, TX. Poster. • MacDonald, I. R., Garcia-Pineda, O., Silva, M., Johansen, C., Reddy, C., & Shedd, W. (2014). Post-Accident Discharge Forensics at the Deepwater Horizon Site, Gulf of Mexico. In Gas in Marine Sediments – 12th International Conference. Taipei, Taiwan.

116 • Silva M., Locker, S., & MacDonald, I. (2014) Geomorphological characterization of the channel system in Desoto Canyon using Chirp Sonar and Piston Core. In Gulf of Mexico Oil Spill & Ecosystem Science Conference 2013. Mobile, AL. • Johansen, C., Shedd, W., Abichou, T., Garcia-Pineda, O. G., Silva, M., & MacDonald, I. R. (2013). Dynamics of hydrocarbon vents: Focus on primary porosity. In Gulf of Mexico Oil Spill & Ecosystem Science Conference 2013. New Orleans, LA. • Johansen, C., Shedd, W., Abichou, T., Garcia-Pineda, O., Silva, M., & MacDonald, I. R. (2012). Dynamics of hydrocarbon vents: Focus on primary porosity. In AGU Annual Meeting. San Francisco, CA, December 6, 2012: Poster. • Silva M., Locker, S., & MacDonald, I. (2013) High resolution characterization of the seafloor of the channel system in Desoto Canyon, Gulf of Mexico; using sub-bottom profiler. “All hand” meeting workshop. Tallahasse, FL. Poster. • Silva M., Etnoyer, P.J., & I. MacDonald (2012) Coral injuries observed at Mesophotic Reef Communities following the Deepwater Horizon oil discharge. Conference: Oils spill & ecosystem science conference. New Orleans, LA. Poster. • Silva M. I. MacDonald & K. Sulak. (2012) Post-spill assessment of Mesophotic reef communities. Gulf of Mexico Conference: Oils spill & ecosystem science conference. New Orleans, LA. Poster • Silva M. & I. MacDonald. (2012) Erosion channels at Desoto Canyon. Deep C Consortium: “All hand” meeting workshop. Tallahasse, FL. Poster • Silva M, Quiroga E, Guzmán G. (2007). Variability patterns of biomass size spectra on of soft bottom benthic community of Patillos Bay (20°46´S; 70°14´W). XII Lantin America Congress of Ocean Science, Florianopolis, Brazil. Poster • Contreras, P, M. Silva, R. Soto. (2007). Invertebrates diversity associated to Aulacomya atra banks on Patillos Bay, North of Chile. XXVII Chilean Congress of Ocean Science.

PUBLICATIONS • Silva M, MacDonald IR (2017) Habitat suitability modeling for mesophotic coral in the northeastern Gulf of Mexico. Mar Ecol Prog Ser 583:121-136. https://doi.org/10.3354/meps12336 • Baco A., Morgan N., Roark B., Silva M., Shamberger K.E. & Miller K. (2017) Defying Dissolution, Discovery of Deep-Sea Scleractinian Coral Reefs in the North Pacific. Scientific Reports 7, Article number: 5436. doi:10.1038/s41598-017-05492-w • Silva, M. (In Prep) Submarine channels and oil industry: a potential threat for the Mesophotic and Deep Coral Ecosystem in Northeastern Gulf of Mexico. • Johansen, C., Marty, E., Silva, M., Natter, M., Hill, J., Viso, R., Lobodin, V.V., Diercks, A.R., M. Woolsey, M., Woolsey, A., Macelloni, L., Maskimova, E.V., Shedd, W., Joye, S., Abrams, M., & MacDonald, I.R (In Review). A hydrocarbon budget for a Gulf of Mexico natural seep site: GC600. Submitted to PlosOne.

117 • Garcia-Pineda, O., MacDonald, I., Silva, M., Shedd, W., Daneshgar Asl, S., & Schumaker, B. (2016). Transience and persistence of natural hydrocarbon seepage in Mississippi Canyon, Gulf of Mexico. Deep Sea Research Part II: Topical Studies in Oceanography, 129, 119–129. • Etnoyer, P.J., Wickes, L.N., Silva, M., Dubick, J. D. Balthis, L., Salgado, E., & Macdonald, I. R. (2016). Decline in condition of gorgonian octocorals on mesophotic reefs in the northern Gulf of Mexico: before and after the Deepwater Horizon oil spill. Coral Reefs 35(1). DOI: 10.1007/s00338-015-1363-2 • Silva, M., Etnoyer, P.J., & MacDonald I.R. (2016) Coral injuries observed at Mesophotic Reefs after the Deepwater Horizon oil discharge. Deep Sea Research Part II: Topical Studies in Oceanography. DOI: 10.1016/j.dsr2.2015.05.013 • MacDonald, I. R., Garcia-Pineda, O., Beet, A., Daneshgar Asl, S., Feng, L., Graettinger, G., Silva, M., et al. (2015). Natural and unnatural oil slicks in the Gulf of Mexico. J. Geophys. Res. Oceans, 120(12), 8364–8380.

RESEARCH CRUISES (Most relevant)

• The Mesophotic Reef Cruise (2014). Purpose: To evaluate DWH oil spill damage over the fish and coral communities. Chief Scientists: Dr. Peter Etnoyer. Ship: R/V Walton Smith. • AT26-13 Cruise. (2014) Purpose: To evaluate long term effects of the DWH oil spill and monitoring of chemosynthetic communities of the Gulf of Mexico. Chief scientists: Dr. Samantha Joye. Ship: R/V Atlantis. HOV Alvin. • Long-term effects of the DWH oil spill (Leg 1 and 2, 2012) Purpose: To evaluate the effects over deep sea ecosystems near to the DWH oil spill. Seafloor and water column mapping over seep sites. Chief scientists: Ian MacDonald (Leg 1) and Charles Fisher (Leg 2). Ship: R/V Falkor. • Geomorphology and benthic ecology cruise- Deep C (Fall-2011, Spring-2012, Summer- 2013, Fall-2013, and Spring-2013 campaigns) Purpose: Geomorphology characterization and photo survey in Desoto Canyon. Chief Scientist: Dr. Ian MacDonald. Ship: R/V Weatherbird II. • Deep Sea Coral Shakedown. (2012) Purpose: Seep sea coral evaluation and seafloor mapping at the Florida escarpment and Desoto Canyon. Chief Scientist: Dr. Peter Etnoyer. Ship: R/V Falkor. • The Mesophotic Reef Cruise (2011). Purpose: To evaluate DWH oil spill damage over the fish and coral communities. Chief Scientist: Dr. Kenneth Sulak. Ship: M/V Holiday Chouest. • Seep-Ox Cruise. (2006) Purpose: to visit the first Methane Seep area discovered in Chile. Chief Scientist: Dr. Javier Sellanes. Ship: R/V Agor Vidal Gormaz.

118

RESEARCH EXPERIENCE • 2012 - 2017: Researcher assistant. ECOGIC. Observation and Remote Sensing Team. • 2012 - : Researcher assistant. Deep C Consortium. Geomorphology and Habitat Classification Team. • 2011 – 2012: Researcher on project: Evaluation of DWH over the Mesospheric Reef communities in the Gulf of Mexico. NOAA-NRDA. • 2008 – 2010: Environmental and Oceanographic Manager, Aquatecma Labs. Chiloé- Chile. • 2003 – 2008: Researcher assistant. “Environmental Vigilance Program to Riles and Quality of Water Column of Arica, Iquique, Tocopilla and Mejillones in influenced areas for CORPESCA S.A.- Benthos Sub-project” Ocean Science Department, Arturo Prat University. • 2005 – 2008: Researcher assistant. “Vigilance program of quality water column recipient at influenced area for organic waste of Camanchaca S.A:” Ocean Science Department, Arturo Prat University. • 2004 – 2007: Researcher assistant “Study of marine environment influenced by Patillos Port” Ocean Science Department, Arturo Prat University. • 2007: Researcher assistant. “Description of management area of Iquique Port” Ocean Science Department, Arturo Prat University. • 2007: Researcher assistant. “Knowledge of Marine communities of rocky shores” Ocean Science Department, Arturo Prat University. • 2002 – 2006: Researcher assistant.“CELTA, Monitoring of marine communities of rocky shores in Punta Patache area” Ocean Science Department, Arturo Prat University. • 2006: Researcher assistant.“Megafauna of the First Methane Seep Area Discovered off Chile: Trophodynamics, Biological and Aspects Relationship with the Patagonian tooth- fish (Dissostichus eleginoides) Fishery off Concepcion (~36°s)”, Fondecyt Project Nº: 1061217 • 2006: Researcher assistant. “Collahuasi” Ocean Science Department, Arturo Prat University. • 2005: Researcher assistant. “Study of Base line in discharge zone of marine emissary of Camanchaca S:A:” Ocean Science Department, Arturo Prat University. • 2003: Researcher assistant. “Study of environmental impact of El Morro Marine” Ocean Science Department, Arturo Prat University. • 2003: Researcher assistant. “Environmental vigilance Program to coast ecosystem of Patache” Ocean Science Department, Arturo Prat University.

119