Factors influencing flashflood deposit preservation in shallow marine sediments of a hyperarid environment

Marine Geology 411 (2019) 22–35

Contents lists available at ScienceDirect

Marine Geology journal homepage: www.elsevier.com/locate/margo

Factors influencing flashflood deposit preservation in shallow marine sediments of a hyperarid environment

T

⁎

Alysse Mathalona,b, , Beverly Goodman-Tchernovc, Paul Hilla, Ákos Kálmánb,c, Timor Katzd a

Dalhousie University, Oceanography Department, 1355 Oxford Street, P.O.B 15000, Halifax, Nova Scotia B3H4R2, Canada Interuniversity Institute of Eilat, P.O.B 469, Eilat 88103, Israel c University of Haifa, Department of Marine Geosciences, Leon Charney School of Marine Sciences, Mt. Carmel, Haifa 3498838, Israel d Israel Oceanographic and Limnological Research, Management and National Institute of Oceanography, P.O.B 8030, Haifa 31080, Israel b

A R T I C LE I N FO

A B S T R A C T

Editor: Michele Rebesco

Flashfloods entering marine basins can introduce large quantities of terrigenous sediment that deposit on the shallow seafloor. Over time, biological and physical disturbances can induce post-depositional sediment transport to deeper depths. In the hyperarid desert region surrounding Eilat, Israel, flashfloods discharge into the northern Gulf of Eilat-Aqaba, Red Sea episodically. Their deposits may leave useful records of flood events in the marine sediment record. What is poorly understood, however, are the processes of post-depositional alteration of flashflood sediments and the resulting effects on the preservation of the deposits. To address this knowledge gap, short sediment cores (~15–30 cm length, 13 m water depth) were collected at bi/tri-monthly intervals following two different flashflood events in the Gulf of Eilat-Aqaba. Concurrently, mechanisms for resuspension, transport, and vertical mixing of flashflood sediments were investigated by measuring near-bottom water currents and suspended sediment concentrations, photographing demersal fish activities, and measuring the depths and magnitudes of bioturbation using fluorescent sediment tracers. The aim was to record the changing grain size of surface flashflood sediments over time, and identify correlative physical and biological factors. The results show that the original fine-grained flashflood deposits coarsened rapidly following each event. Of the mechanisms recorded, currents generally were too weak to entrain sediment, but did influence the transport of sediment in suspension to deeper depths. Demersal fish resuspended sediment when present, and bioturbation was strongest in the upper 2 cm of the seabed. Despite the rapid reworking of surficial sediments and dissipation of surface flood layers, irregular lenses of fine sediment persisted within the shallow seafloor (down core). Deposition within seafloor holes and depressions, and rapid burial by biologically constructed sediment mounds and ensuing flashflood deposits are proposed as mechanisms for this patchy, localized preservation of flood layers. These findings are useful for identification and interpretation of isolated fine-grained layers in older sedimentary deposits, and for understanding the dynamics influencing the marine sediment stratigraphy in arid, coastal environments.

Keywords: Flashflood deposit Shelf stratigraphy Bioturbation Resuspension Transport Red Sea

1. Introduction Sediment layers formed on the seafloor that originate from shortterm events such as flashfloods or earthquakes are known as ‘event layers’. These layers attract interest because their presence in sedimentary strata can provide valuable information about the frequencies and magnitudes of past meteorological, oceanographic, and geological events (Mulder et al., 2001; Postma, 2001; Lamb and Mohrig, 2009). Accurate interpretation of event layers in the sediment record, however, requires an understanding of the post-depositional processes that

affect layer preservation in marine sediments. Post-depositional processes can include burial, sediment reworking, and removal by biological and physical processes. Preservation of flashflood deposits in marine environments can serve as a direct link between terrigenous sediment sources and marine sediment depositional sinks (Mulder et al., 2001; Bentley and Nittrouer, 2003; Lamb and Mohrig, 2009; Kniskern et al., 2014). The present study explores the physical and biological factors that influence flashflood deposit preservation in shallow marine sediments off the coast of Eilat, Israel, in the Gulf of Eilat-Aqaba (GOA), Red Sea. This environment is

⁎

Corresponding author at: 53 Millcroft way, Thornhill, Ontario L4J6P2, Canada. E-mail addresses: [email protected] (A. Mathalon), [email protected] (B. Goodman-Tchernov), [email protected] (P. Hill), [email protected] (T. Katz). https://doi.org/10.1016/j.margeo.2019.01.010 Received 18 June 2018; Received in revised form 16 January 2019; Accepted 28 January 2019 Available online 31 January 2019 0025-3227/ © 2019 Elsevier B.V. All rights reserved.

Marine Geology 411 (2019) 22–35

A. Mathalon et al.

Fig. 1. A) Local map of the study area showing the sampling site, ~200 m offshore of the Kinnet Canal outlet at 13 m water depth. B) Regional map; the sampling location is marked with a red circle. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

The links between the properties of sediment layers deposited during floods and the floods that delivered them are complicated by post-depositional physical and biological reworking (Bentley and Nittrouer, 2003; Wheatcroft and Drake, 2003; Kniskern et al., 2014; Steiner et al., 2016). The preservation or destruction of event layers is mitigated by the competition between rates of sediment accumulation, which preserves layers, and the biological and physical processes that lead to sediment reworking (Howard, 1975; Wheatcroft, 1990; Wheatcroft and Drake, 2003). The presumed dominant physical forces and transport mechanisms that can cause sediment resuspension and removal in the GOA are bottom shear stresses due to currents and windgenerated surface gravity waves (Yahel et al., 2002). The main biological processes that have the potential to alter surface sediments are biological resuspension from demersal fish (Yahel et al., 2002) and benthic organisms, and bioturbation (sediment mixing by benthic organisms) (Black et al., 2012; Steiner et al., 2016). Previous studies quantified current measurements from the northern GOA. Yahel et al. (2002) measured average and maximum current speeds of 4.6 cm s−1 and 9.6 cm s−1 at 1 m above the bottom in water depths between 8 and 15 m, near the fringing reefs on the northwestern shore. From that 1.5-year study, current speeds were calculated to have the capacity to move surface sand sediments < 4% of the time. Yahel et al. (2002) observed that rare storms associated with southern winds produced waves > 2 m at the coast, causing sediment resuspension and increased turbidity, though the depth of this influence was not directly measured. Typically, however, winds prevail from the north in the GOA (Katz et al., 2015; Israel National Monitoring Program, 2017), and are associated with small waves < 0.3 m high

characterized by low mean sediment accumulation rates, high episodic sedimentation rates, and rich benthic and demersal communities. Eilat, Israel is located on the coast of the most northern point of the GOA (Fig. 1). The climate in this region is hyperarid (average precipitation < 30 mm year−1), but sporadic rainfall can induce flashfloods, most frequently in autumn and spring. Some of these floods ultimately flow into the GOA. At times flood waters can be so concentrated with sediment that they are denser than the relatively dense seawater of the GOA, resulting in hyperpycnal flows along the bottom (Katz et al., 2015). Although highly episodic, with years at times passing between events, flashfloods account for the primary source of terrigenous sediment to the GOA (Pittauerová et al., 2014; Katz et al., 2015; Steiner et al., 2016). Despite the importance of these floods to the sediment budget of the GOA, and more broadly to the ecology of the surrounding catchments, the record of flood frequency and magnitude is limited to the past 24 years (Pittauerová et al., 2014; Katz et al., 2015). Flashfloods occur in arid locations with low vegetal coverage, and in areas surrounded by mountains that drain into small- to medium-sized rivers (Milliman and Syvitski, 1992; Mulder et al., 2001, 2003). Where floods discharge into the ocean, their associated deposits have been preserved within marine sediment records in environments characterized by high mean sediment accumulation rates (Mulder et al., 2003; Kniskern et al., 2014; Walsh et al., 2014) and in deep sea, low energy environments (Mulder et al., 2001; Postma, 2001). However, little is known about the mechanisms of flashflood deposit preservation within shallow marine sediments located in arid environments, where overall sediment accumulation rates are low.

23

Marine Geology 411 (2019) 22–35

A. Mathalon et al.

rivers into the GOA. During dry periods, the Kinnet returns a small amount of pumped, clear seawater into the gulf. There are several ephemeral rivers that enter the GOA from the eastern and western shores (Fig. 1a). Sediments on the shelf are mostly fine sands. While there are no direct wave measurements on the northern shelf, waves exceeding 0.5 m are rare, and waves > 1 m only occur during infrequent southern storms (Yahel et al., 2002). Measured current speeds are low, averaging at 4.6 cm s−1, 1 m above the bottom in ~10 m water depth (Yahel et al., 2002). The GOA is a long, narrow and deep waterbody (max. depth = 1850 m) with steep lateral slopes. This morphology derives from its positioning on the Dead Sea transform fault system, on the boundary between the Arabian and African plates (Ben-Avraham, 1985) at the northeastern upper extension of the Red Sea (Fig. 1b). The GOA has annual surface water temperatures of 20–28 °C and surface water salinity that is nearly constant between 40 and 41 ppt (Biton and Gildor, 2011a). The eastern and western edges of the GOA bear fringing coral reef ecosystems to depths deeper than 100 m in some locations (Fricke and Schuhmacher, 1983; Shlesinger et al., 2018). The northern shore is sand-dominated, and it has a narrow continental shelf that is approximately 1.7 km wide with a mean slope of 3° (Katz et al., 2015). The experiments and observations were carried out on the northern shelf of the GOA, approximately 200 m seaward of the Kinnet Canal outlet, in 13 m water depth (Fig. 1a).

with 2–3 s periods (Yahel et al., 2002). Relatively low-energy currents and infrequent wind waves suggest that biological processes may play an important role in the redistribution of flood sediments. Within the coral reef regions of the GOA, demersal fish were identified as dominant contributors to sediment resuspension relative to wave and current resuspension (Yahel et al., 2002). The measured rate of sediment resuspension by demersal fish was > 36 events m−2 day−1, and when fish were excluded from an area of the reef, the volume of resuspended sediment decreased by 26–86%. The effects that demersal fish had on sediment resuspension in the shallow GOA reefs suggest that their activities may affect preservation of flood layers on the northern shelf. Bioturbation, defined as the activity of benthic organisms that alter sediment deposits from their primary structures on millimeter to meter scales (Gerino et al., 1998), occurs nearly ubiquitously in marine sediments. The average mixing depth from bioturbation (Lb) is relatively constant globally at ~10 cm, but its magnitude varies depending on environmental conditions (Boudreau, 1994). Bioturbation transforms initial sedimentary structures within the seabed, while sedimentation can bury sediments to depths greater than Lb, where they are no longer modified extensively by physical and biological processes, and are thus preserved (Bentley et al., 2006). Wheatcroft (1990) defined ‘preservation potential’ as the probability that sediment layers will remain recognizable prior to burial. Bioturbation has been observed to smear or erase event layers within the sediments when the rate at which layers are removed or reworked exceeds the rate at which they are buried (Wheatcroft and Drake, 2003; Bentley et al., 2006; Steiner et al., 2016). Generally, bioturbation rates are faster closer to the sediment surface, and decrease with increasing depth within the seabed until the rates are negligible (Boudreau, 1994). Therefore, preservation potential is greater for event layers that are buried quickly beyond the depth where bioturbation is most active (Bentley et al., 2006). Katz et al. (2015) observed significantly higher proportions of siltand clay-sized grains within recently deposited flood sediments on the shallow seafloor in the GOA compared to in the ambient sediments unaffected by flood events for at least one year. For example, a flashflood that entered the GOA in February 2013 brought in a minimum of 21,000 tons of sediment, 92% of which was fine, silt-clay material. In a sediment core taken from the shallow GOA 52 days after a flashflood in January 2013, which was the third flood event within two months, the top 4–5 cm comprised 81% silt-clay material by mass, while the underlying sediment had 6% fine sediments by mass. Intriguingly, one year following a flashflood that entered the GOA in January 2010, the silt-clay flood deposit was no longer recognizable on the surface of the shallow seafloor (Katz et al., 2015). The goal of this study is to elucidate the factors that control the preservation of flashflood deposits in the shallow GOA. To this end, the time evolution of flashflood deposits was documented in order to link the transformation of the deposits to the physical and biological processes that disturb them. This information is most valuable for understanding post-flood sedimentological processes in general, in the GOA in particular, and for enabling better interpretations of paleoflood layers in the sedimentary record. A suite of observations was made on the shallow shelf, offshore of the dominant drainage basin of flashfloods to the GOA from Eilat, over a period of time that included multiple floods.

3. Materials and methods 3.1. Bi/tri-monthly core collections To track the changing grain size character of flashflood deposits on the shallow shelf, short sediment push cores (30 cm maximum length, 4.5 cm diameter) were collected by divers in a 50 m2 area in 2–3 month intervals for 14 months following a flashflood that entered the GOA in May 2014, and for 12 months following two flashfloods on September 15th and October 25th, 2015. Initial cores were taken within 10 days of flashflood events. The standard procedure was to take two cores within a 1 m2 area at each collection. Single cores were collected at the beginning of the study, prior to the commencement of this sampling sequence, and occasionally between the scheduled bi- and tri-monthly collection periods, which allowed for additional analyses. Each core was subsampled in 1 cm vertical intervals immediately following collection, or cores were placed in a 4 °C refrigerator until sliced. Throughout the study, care was taken to minimize disturbance of the sediments during coring, transport and subsampling. Nonetheless, some alterations in the vertical distribution of the grain size compared to the ambient conditions might have occurred because of compaction and smearing of the sediment during the extraction process. All cores were extracted, transported and subsampled in a consistent manner. A 1–2 mL subsample was isolated from the center of each 1 cm thick slice for grain size analysis. Each subsample was treated with 30% hydrogen peroxide to digest organic matter. Subsamples were homogenized, passed through a 2000-μm sieve, and grain size was measured using a Beckman Coulter Counter LS 13 320 Laser Diffraction Particle Size Analyzer (detailed methods in Goodman-Tchernov et al., 2009). 3.2. Bottom boundary layer observations

2. Regional setting

A mooring station was deployed with a Sea-bird SBE19 plus CTD (Conductivity, Temperature, Depth instrument) placed on the seafloor, attached to three Campbell Scientific Optical Backscatter (OBS-3+) Turbidity Sensors positioned on a rope at 0.1, 2 and 8 m above the bottom (mab). The CTD provided time-series data of salinity, temperature, and water depth, and the OBS were used to generate timeseries of suspended particle concentrations. Measurements were averaged over 1 s at 5-min interals during 5 deployments from April 2016 to January 2017.

Eilat is surrounded by a desert mountain range composed of ~500million-year old magmatic-metamorphic rock that underlies an eroding ~60-million-year old suite of sedimentary sandstone, carbonate, and volcanic rocks (Beyth et al., 2011). The mountains are dissected by networks of wadis (dry riverbeds) that make up the Arava drainage basin, and lead into the northern GOA via the manmade draining channel named the Kinnet Canal (Fig. 1a). The Kinnet drains several 24

Marine Geology 411 (2019) 22–35

A. Mathalon et al.

Fig. 2. Diagram representing the steps taken to transplant a sediment core with a fluorescent tracer into the seafloor. (1) Begin with a tracer-tagged core. (2) Place the plate on the seafloor, and hammer three poles into the center holes of the plate. The black arrow on the plate references its consistent orientation. (3) Hammer the transplant device, shown in panel 2, into one of the corner holes of the plate. (4) Remove the inner tube of the transplant device. (5) Remove the bottom plug of a core and place it within the seabed, inside the outer tube of the transplant device. (6) Remove the outer tube of the transplant device and the top plug of the core. (7) Remove the core liner and the plate. (8) Tracer is positioned at the surface of the seafloor.

seafloor. Cores were recovered after 3 different time intervals.

The OBS measured the intensity of backscattered infrared light (850 ± 5 nm) in mV, which was converted to particle concentration (g m−3) using lab-generated calibration curves. Each OBS sensor had both high and low sensitivity channels, which were calibrated in the laboratory by exposing the sensors to volumes of seawater with known concentrations of flood sediments (collected from a nearby riverbed). Increasing increments of 2.5 mL of flood sediment were mixed with 1 L of seawater, then introduced to a large tank with the OBS sensors. Following each sediment addition, once the OBS readings stabilized, 200 mL of tank water were filtered onto pre-weighed glass microfiber filters to obtain the concentration of suspended solids in g m−3. The calibration provided upper detection limits of ~600 and 2300 g m−3 for the high and low sensitivity channels, respectively (Appendix A). Periods in the OBS time-series that showed progressive increases in optical backscatter approaching scheduled biweekly cleanings of the sensors were flagged as contaminated by fouling and were not included in the reported observations. A Nortek 2.0 MHz Aquadopp Acoustic Doppler Current Profiler (ADCP) was placed on the seafloor to track current velocities at heights between 0.15 and 1.5 mab. Each current measurement was averaged over a 60s period every 10 or 20 min from April 2016 to February 2017, over 7 deployments. Daily data on wind speed and direction were obtained from the National Monitoring Program at the Gulf of Eilat (Israel National Monitoring Program, 2017).

3.4.1. Tracking surface sediment movement with tracers Vertical redistribution and removal of surficial sediments were tracked using luminophores, which are sediments that were dyed to fluoresce under Ultra Violet (UV) light. Each tracer disc had 1.5 g of green luminophores (< 63 μm), 2 g of orange luminophores (125–250 μm), 1.5 g of surface sediment from the study region, and water. Tracer discs were homogenized and frozen. From the study site, 27 short cores ranging in length from 22.5 to 26.5 cm were collected by divers and brought back to the lab, where a tracer disc was placed on the surface of each. Cores were then transferred into a salt water pool and left uncovered for up to 6 days, to keep organisms alive. 3.4.2. Core transplantation in seafloor A total of 27 cores with tracer discs at their surfaces were transplanted into the seafloor in May 2016. Three cores were emplaced at 9 different sites on the seafloor that were spaced 10 m apart. At each site, three poles were placed in predetermined positions on the seafloor, so that a custom-made plate fit over them. The plate had four larger holes (diameter = 6.5 cm) at its corners that marked the locations of the core transplants (3 of these holes were used). A core transplant device was created to emplace tracer-tagged sediment cores into the seafloor. It consisted of an inner tube with a bottom, that was driven into the seabed to create the hole for the core, and an outer metal, bottomless tube (height = 33 cm, diameter = 5.6 cm) that prevented collapse of the walls of the core hole prior to emplacement of the tracer-tagged cores. The core transplantation procedure is shown in Fig. 2.

3.3. Monitoring of sediment resuspension by demersal fish Time lapse photographs were taken of the near-bottom environment to observe if demersal fish were resuspending sediments. A GoPro Hero 4 was mounted 60 cm above the seafloor on a pole that was pushed into the sediment. The camera was deployed 8 times during daylight hours, taking pictures in 20 or 30 s intervals for a duration of 1–2 h per deployment. Three of the deployments were in June, three in July, and the remaining two were in September 2016. The number of photographs were counted in which demersal fish were present and local resuspension of near bed sediments occurred. Local resuspension was defined as the occurrence of observable, temporary sediment plumes immediately above the seabed.

3.4.3. Core recovery and tracer analysis One core was recovered from each site after 1 week (t1), 3 weeks (t2) and 6 weeks (t3). The process for core recovery is shown in Fig. 3. When transplanting and recovering the cores, the cores were kept as vertical as possible, in order to maximize the recovery of the transplanted sediment and tracer. Upon recovery, cores were sectioned at 1-cm vertical resolution (Fig. 4a) and dried immediately upon returning to the laboratory, or they were placed in a 4 °C refrigerator for later sectioning. A subsample of 15 g was collected from the center of each centimeter slab that contained tracer in each core. The subsamples were disaggregated with a mortar and pestle to separate sediment grains. Each subsample was homogenized, and ~0.01 g were placed on three respective 1.2 cm2 pieces of tape. Each piece of tape was photographed in a dark room under UV light using a binocular microscope (1× magnification).

3.4. Bioturbation An in situ experiment using fluorescent sediment tracers was set up to measure mixing and removal of surficial sediments. Cores with fluorescent tracer discs on their surfaces were transplanted into the 25

Marine Geology 411 (2019) 22–35

A. Mathalon et al.

Fig. 3. Diagram representing the core recovery process. (1) Return to the three poles positioned in the seafloor. (2) Place the plate over the three poles and then insert the core liner into the corner hole of the plate, where the core had initially been transplanted. (3) Plug the top of the core, and pull it out with the sediment in it. (4) Plug the bottom of the core.

were formed (Appendix B). Coefficients of determination (R2) were 0.94 and 0.99. The calibration accounted for 99% of the original 1.5 g of green luminophores in the tracer. 3.5. Transects, measurements and grain size of biological mounds and holes Sediment mounds and holes created by burrowing organisms were present throughout the seafloor at the study site. These features could influence the deposition and fate of the fine-grained flood deposits. Therefore, six transects were conducted to count the number of mounds and holes (diameter > 5 cm) over an area of 9 m × 1 m. Measurements were carried out on the widths and heights of 13 mounds and depths and diameters of 6 holes that were selected randomly. To explore the possibility that biological mounds and holes within the seafloor influenced the preservation of flashflood deposits, three short cores were collected in May 2017 from a flat part of the seafloor, from within a hole (depth = ~15 cm) and from within a mound (height = ~15 cm). Coring was two months after a flashflood entered the GOA in March 2017, and six months after a flood had entered the GOA in late October 2016. Grain size analyses were carried out on subsamples from each centimeter of each core. 4. Results 4.1. Grain size profiles from bi/tri-monthly core collections Cores taken immediately following flashfloods were consistently characterized by primarily fine sediments (75% to nearly 100% of grains were smaller than 63 μm) in the upper few centimeters of the cores, with coarser, sandy sediments (< 27% of grains were smaller than 63 μm) below. Fine layers ranged in thickness from 1 to 5 cm (Fig. 5, Panels: 1, 12–15, 27–28). With the passage of time after a flooding event, the surface sediments coarsened (Fig. 5). Coarsening started from the surface and progressed downwards with time. In some cores distinct sediment layers, 1–6 cm thick, with higher proportions of finer sediments (~65–99% of grains were smaller than 63 μm) were found between 10 and 25 cm depth (Fig. 5, Panels: 7, 10, 12–15, 19, 20). In other cores grain size distributions were more uniform and coarser below 5 cm (Fig. 5, Panels: 6, 18, 24–26). In many cases, replicate cores had grain size profiles that differed markedly (Fig. 5, Panels: 6–8, 10–11, 14–15, 16–17, 21–24).

Fig. 4. A) Photographs of the 1-cm section horizons of a core with a tracer after recovery. B) Photograph taken under a binocular microscope of a subsample from a centimeter of a core on a 1.2 cm2 piece of tape (left), and after Otsu's thresholding has been applied (right).

Green luminophores (< 63 μm) were primarily analyzed as the fluorescence of the orange luminophores (125–250 μm) was weaker and harder to detect. The area of fluorescence (pixels)/picture was measured using MatLab for image analysis. Particle edge detection was achieved using Otsu's thresholding (Fig. 4b) (Otsu, 1979). The average area of fluorescence/picture was determined within each centimeter (3 pieces of tape per cm), and profiles of tracer distributions were established for each core. The accuracy in analyzing luminophore concentrations using this method was assessed by creating two calibration curves from the top centimeter of two tracer-tagged cores that were transplanted into the seafloor and immediately recovered. The samples were dried, and 15 g from the center of each layer was subsampled, disaggregated and homogenized. The subsample from each core was split and diluted 7 times with sediment collected from the site. After each split and dilution, the resulting sediment-tracer mixture was subsampled onto three 1.2 cm2 pieces of tape. The areas of fluorescence/picture from the three pieces of tape were averaged for each split, and two calibration curves

4.2. Bottom boundary layer dynamics Near-bottom current velocities averaged 3.7 ± 2.5 cm s−1 (Fig. 6b), and they were not correlated with suspended particle concentrations (Fig. 6a). Throughout much of the time-series, near-bottom currents traveled mostly southward (83% of time) and eastward (75% of the time), and had an average displacement value of 0.013 km day−1 to the south. From late October to early February, during a period of cooling of the water column, average flow displacement increased to 0.029 km day−1, and the directionality shifted to only southerly (Fig. 6e). During this time, elevated wind speeds from the south occurred almost simultaneously with an increase in water density (caused by cooling), enhanced water currents, and a slight increase in 26

Marine Geology 411 (2019) 22–35

A. Mathalon et al.

Fig. 5. Grain size distributions as % volume of sediment < 63 μm in cores taken at 10 m or 13 m water depth, in front of the outlet of the Kinnet Canal. Flood cores are labelled and are represented with triangles, and the label “mo.” on the remaining cores represents the number of months passed since the last flashflood event. Each panel is numbered in the bottom right corner, and replicate cores are listed in the titles in the parentheses.

possible to see that interaction and therefore these observations were quantified separately as fish presence/absence and suspension event presence/absence. From six deployments in June and July, demersal fish were present in a maximum of 3.5% of useable photographs per deployment, and local resuspension events occurred in a maximum of 7.0% of useable photographs per deployment. The respective number of useable photographs per deployment were: 21, 83, 81, 10, 155, and 171. From two deployments in September, demersal fish were observed in 15.1% and 100% of useable photographs per deployment, and local resuspension events were present in 5% and 100% of useable photographs per deployment (Fig. 7). The number of respective useable photographs per deployment were: 311 and 201.

suspended particle concentrations (Fig. 6). In the middle of the record, a short-lived peak in suspended particle concentration occurred simultaneously with a drop in salinity. This peak coincided with a flashflood that entered the GOA via the Kinnet Canal on October 28th, 2016 (Fig. 6a, d (in the grey box)). Throughout the observational period, additional short-lived peaks in suspended sediment concentrations were recorded between May and June, and from late July to early August. 4.3. Demersal fish activity Due to generally poor visibility, only presence or absence of demersal fish was recorded. Presence or absence of suspension events, based on the identification of temporary, concentrated sediment ‘puffs’ rising from the seafloor, was also recorded. While it was apparent that these puffs were likely caused by the fish foraging, it was not always

4.4. Tracer records The majority of tracer distributions decreased consistently from the Fig. 6. Physical time-series showing (a) suspended particle concentration measured from the high sensitivity channel of an OBS sensor in 5 min intervals. Gaps in the record indicate periods where the sensor was affected by bio-fouling. Lower panels show the following: (b) current speeds, measured in 10–20 min intervals, (c) wind velocities, presented in 3 h intervals (vectors point in the direction from which the wind originated) (d) water salinity and temperature, measured in 10–20 min intervals, and (e) average N-S current displacements, measured per day. Measurements within the grey rectangle show a recorded flashflood event on Oct. 28th, 2016. Mab means ‘meters above bottom’.

27

Marine Geology 411 (2019) 22–35

A. Mathalon et al.

Fig. 7. Presence of demersal fish and resuspension events from GoPro camera photographs, in 2016.

Fig. 8. Profiles showing the distributions of the luminophore tracer down to 10 cm below the seafloor in cores that were in the seafloor for a maximum of 6 weeks. The top row shows the cores with tracer maximums found ≤2 cm below the seafloor (surface tracers) and the bottom row shows cores with tracer maximums found 3–5 cm below the seafloor (deep tracers). Error bars represent the standard deviation from the three luminophore areas measured/picture/cm. Note that the x-axis range varies per core, and each panel is numbered in the bottom right corner.

(R2015b, MathWorks)). This test revealed that significantly less tracer was recovered from cores with tracer maxima found at depths ≤2 cm. To test if the variance in tracer distributions varied between the cores with shallower and deeper maxima, the average tracer concentration in each of the 5 cm surrounding the tracer peak were compared (n = 55) using Barlett's test of equal variance (p = 0) (Matlab (R2015b, MathWorks)). This test revealed that tracers in cores with depth maxima of ≤2 cm had lower variance than tracer distributions in cores where tracers were found deeper within the seabed, presumably indicating greater mixing of the tracer away from the peak.

tracer maxima within each core (Appendix C). In three cores there was evidence of a secondary deeper peak in tracer concentration (Fig. 8, Panels: 3, 7, 10). The average tracer penetration depth was 4.4 ± 1.8 cm, which stabilized after t1 (1 week), and the maximum depth interval that contained tracer was 9 cm. There was no apparent effect of time on variations in tracer distribution or recovery (ShapiroWilks normality tests, p > 0.05; 1-way ANOVA, p > 0.05) (Matlab (R2015b, MathWorks)). Tracer distribution was based on the number of centimeters down a core that contained tracer, and tracer recovery was based on the total tracer area/picture (averaged for each cm) calculated per core. The depths of peak tracer concentrations varied from 0 to 7 cm within the cores, which was interpreted to be a result of differential emplacement depths of the initial tracers within the seabed, despite the care taken to position core tops at the sediment surface. To determine if tracer recovery differed among cores with shallower (≤2 cm) and deeper (3–5 cm) depth maxima, the total tracer area/picture (averaged for each cm) calculated per core in the two groups of cores were compared (n = 11) using a 2-sample t-test (p = 0.0148) (Matlab

4.5. Concentration, measurements and grain size profiles of biological mounds and holes The seabed at the study site was covered by biogenic sediment mounds and holes (Fig. 9). An average of 1.27 mounds m−2, and 0.94 holes m−2 were counted, and measurements of the widths and heights of 13 mounds and depths and diameters of 6 holes were recorded (Table 1). The dimensions of some of these features were 5 or 28

Marine Geology 411 (2019) 22–35

A. Mathalon et al.

5.2. Interpretation of time evolution of flood deposits The initial thicknesses of fine-grained flashflood deposits varied spatially among cores from 1 cm to 5 cm (Fig. 5, Panels: 1, 12–15, 27–28), which can be attributed to the heterogeneity in the seafloor topography caused by pre-existing biogenic mounds, holes and depressions. Additionally, locally variable biological reworking immediately following flashflood events may have contributed to variable initial thicknesses of flood deposits. 5.2.1. Rapid reworking of surficial sediments Flashflood sediments coarsened at the surface over time, but the rates and magnitudes of coarsening were spatially and temporally variable (Fig. 5). The winnowing recognized in the deposits over time cannot be attributed solely to entrainment and resuspension by currents because current speeds were generally too low. For example, the average grain size of surface sediments (top 3 cm) measured immediately following a flashflood was 14 μm, while they reached average values of 187 μm 3 months later. Respective current velocities required to resuspend such sediments should be at least 13 cm s−1 and 41 cm s−1, 25 cm above the seafloor (assuming non-cohesive sediments with a dry density of 2.65 g cm−3, water temperature of 25 °C, salinity of 40 ppt, and level bottom; Wiberg and Smith, 1987). The current speeds measured at the study site exceeded the minimum velocity to resuspend fine flood sediments only 0.28% of the time during this observational record. Coarsening of fine sediments at the surface may have resulted from a combination of sediment mixing by bioturbation by infauna that transported fine flood sediments downwards and coarse sediments from below the deposit upwards (Fig. 8), and from biological resuspension (Fig. 7). Fine sediments could then be carried away by the slow bottom currents (Fig. 6b). Rapid reworking and removal of flood deposits has been observed in multiple marine environments. Because there are no parallel examples from similarly hyperarid settings, comparisons are made to other flooddominated environments. One such location was on the shallow seafloor off the Waipaoa River in Poverty Bay, New Zealand. There, flood signatures were altered by waves and currents, which resuspended and redistributed sediments, in addition to active bioturbation that mixed surface sediments. These effects resulted in inconsistent stratigraphic layering on the inner continental shelf (Kniskern et al., 2014; Walsh et al., 2014). Flood deposits were better preserved in ‘depocenters’ (between 20 and 70 m water depth) where high concentrations of flood sediment accumulated and remained due to less active biological and physical reworking (Kniskern et al., 2014; Walsh et al., 2014). The removal of a fine-grained flood deposit from the surface of the seafloor was also observed on the continental shelf offshore of the Eel River in California. In 1995, three floods of the Eel River discharged large quantities of sediment to the adjacent continental shelf and created a distinct silt-clay deposit that was 5–10 cm thick. In < 6 months, the upper portion of the flood deposit thinned and coarsened as a result of a combination of biological mixing, primarily in the upper 3–5 cm of the seabed, sediment resuspension and transport, and from sedimentation of coarser particles subsequent to the flood events (Bentley and Nittrouer, 2003). Preferential removal of fine particles from the sediment-water interface also was observed in the grain size profiles within this study (e.g. in Fig. 5, Panels: 16–26), which created diffusive-like profiles of fine particle concentrations decreasing towards the surface.

Fig. 9. Photograph showing seafloor topography at the study site in June 2017.

Table 1 Measurements of biological mounds and holes taken in Feb. and Apr., 2017. Number

Mound height (cm)

1 2 3 4 5 6 7 8 9 10 11 12 13

Mound width (cm)

12 19 14.5 20 16 14.5 13.75 8.75 17.5 20.5 15 16 11

36 90 35 44 44 51 46 41 46 80 38 40 7

Hole depth (cm) 18 21.5 9 9 15 7

Hole diameter (cm) 27.5 60 10 15 18 5

more times higher and deeper than the observed thicknesses of flashflood deposits. Grain size profiles of fine sediments differed within the cores collected from the flat seafloor, from within a mound and from within a hole two and six months after flashflood events (Fig. 10). In the core from the flat seafloor, two fine-grained layers 5 cm and 2 cm thick were present between the surface and 5 cm, and 7–8 cm depth. Both at the surface and between the layers the sediment was coarser. In the core from the mound, coarser sediment overlay a distinct fine-grained layer 3 cm thick that appeared between 15 and 17 cm below the surface, and in the core from the hole, fine sediments increased with depth down the core.

5. Discussion 5.1. Recent flashflood history The documented record of flashfloods entering the GOA via the Kinnet Canal outlet begins in 1994. Between 1997 and 2012 there was a prolonged drought, with only two floods entering the GOA, one in 2006 and another in 2010. In the fall of 2012 three flood events occurred within 25 days of each other, and after less than two months, two more flashfloods occurred within 10 days of each other in 2013. Thirteen months later in March 2014 another flashflood entered the GOA, such that by the time the data collection began for the present study, in May 2014, flood sediments could have been present within the seafloor from the previous 1.5 years.

5.2.2. Patchy preservation Some cores in this study had fine-grained sediment layers that were 1–6 cm thick between 10 and 25 cm down core, which are interpreted to be buried flashflood deposits from the high flood frequency period in the winter of 2012–13 (Example: Fig. 11). Buried flood layers were not observed in sediment cores that were collected in the study area both a year before and during the 2012–2013 rainy season (Katz et al., 2015), 29

Marine Geology 411 (2019) 22–35

A. Mathalon et al.

Fig. 10. Distributions of the % volume of sediment < 63 μm within cores taken from the flat seafloor, within a mound, and within a hole.

in Southwestern France. Distinct alternations between marine flood sedimentation and inter-flood sedimentation were not evident, which was attributed to low sedimentation rates and/or active bioturbation. Inconsistent sediment stratigraphy with respect to flood layers also was observed on the shallow seafloor of Poverty Bay, New Zealand, due to high rates of biological mixing (Kniskern et al., 2014). Additionally, Gerino et al. (2007) found high local variability in bioturbation activities in the shallow waters of the Venice Lagoon, suggesting local heterogeneity in biological mixing may be common in marine environments. The benthic environment in the shallow GOA is characterized by low sediment accumulation rates and active bioturbation, which helps to explain the lack of uniform preservation of flashflood deposits and the heterogeneity of grain size profiles throughout the shallow seafloor. The preservation potential of flashflood deposits was therefore dependent on local rates and depths of post-depositional mixing, removal and burial.

5.3. Physical forces & transport mechanisms Current speeds measured by the ADCP generally were not strong enough to resuspend sediments in 13 m water depth (Fig. 6b). Current directions on the northern shelf were primarily southeasterly, which caused preferential transport and redistribution of resuspended flood sediments in that direction. To enhance southward transport, resuspended sediments that were subjected to southerly currents theoretically traveled farther compared to suspended particles that were transported by northerly currents, due to the mean slope of the seafloor of ~3° (Katz et al., 2015). Over time, the net direction resuspended flood sediments traveled was southwards and potentially off of the shelf and into the adjacent deep basin. Offshelf transport presumably increased in the fall and winter months when current velocities were larger and were oriented consistently to the south (Fig. 6e). This seasonal change can be explained by the occurrence of dense water formation, which begins when cooler surface water from the Red Sea travels through the Straits of Tiran into the GOA. As this water reaches the northern gulf, it undergoes additional cooling from the atmosphere in the fall and winter, subducts due to its increased density, and flows along the bottom back out towards

Fig. 11. Distributions of the % volume of sediment < 63 μm within replicate cores taken 6 months following two consecutive flashflood events in September and October 2015.

therefore it is concluded that the buried flood deposits that were observed in samplings from this study are not remnants from flood events prior to 2012. The average sediment accumulation rate on the shelf is estimated at 1.3 mm y−1 at 16 m depth (Goodman-Tchernov et al., 2016). With that sedimentation rate, it would take ~15 years to bury sediments to 2 cm depth. It is therefore assumed that the flashflood layers found at depth within the seabed were buried by local sedimentation rates that were substantially larger than the mean accumulation rate. Other cores had more uniform and coarser grain size profiles that showed no evidence of lower fine-grained flood deposit preservation, and in many cases, replicate cores had grain size profiles that were different. Baumann et al. (2017) explored the sediment record (1.5 m cores, ~2000 years) of marine flooding events in back barrier marshes 30

Marine Geology 411 (2019) 22–35

A. Mathalon et al.

population density. Bioturbation has been shown to increase particle resuspension by a factor of 3–8 (Davis, 1993). As discussed, biological resuspension in the shallow coral reef habitat of the GOA is mostly attributable to fish activity. However, it is possible that in the sanddominated environment of the northern shelf, benthic organisms may have greater influence on resuspension compared to the reefs (Black et al., 2012).

the mouth of the Red Sea (Manasrah et al., 2004; Biton and Gildor, 2011b; Pittauerová et al., 2014). Transport of flood sediments to the deep may also be enhanced by the narrowness of the GOA's continental shelf (Milliman and Syvitski, 1992). Current and wave stresses on the seabed affect where sediments accumulate and where event layer preservation occurs (Kniskern et al., 2014; Walsh et al., 2014). For example, on the continental shelf off the Eel River in California, both recent and historical flood sediments appeared in the sediment record in the middle of the continental shelf due to transportation by wave-induced suspension and seaward flowing currents (Sommerfield and Nittrouer, 1999). In order to get insight into the locations on the seafloor with highest sediment accumulation rates and minimal biological and physical reworking, it is important to sample multiple depths on and off of the continental shelf both immediately after a flood and shortly after (~1 year) (e.g. Sommerfield and Nittrouer, 1999; Kniskern et al., 2014; Walsh et al., 2014). Peaks in suspended sediment concentrations recorded by the OBS from May–June and from late July to early August (Fig. 6a) are not interpreted to be locally resuspended sediment, but instead they may be particles that originated from shallower depths or land-based activities, as currents were too weak to resuspend bottom sediments at this depth. Alternatively, suspended sediment concentrations could have been elevated by biological resuspension or bioturbation activities.

5.5. Influences of bioturbation on the integrity of flashflood deposits Because the initial thicknesses of flashflood deposits varied between 1 and 5 cm, the preservation potential from one flashflood event was spatially heterogeneous at the same water depth and general location. Initial flood layers in the shallow GOA were thinner than the universal mixed layer depth, so over time they were likely destroyed by bioturbation, and consequently failed to become preserved as ‘historical layers’ within the seabed (coined by Berger et al., 1979, from Bentley et al., 2006; Wheatcroft and Drake, 2003). The results from the bioturbation experiment support the observations from the bi/tri-monthly grain size profiles, where flood deposits on the seafloor were removed from the surface and mixed vertically within the seabed over time. The distributions of the luminophore tracers within the majority of the cores decreased from the locations of the tracer maxima (Fig. 8). This profile shape is similar to those in experimental cores from Bentley and Nittrouer (2003), as well as from Gerino et al. (1998), who deployed an in situ bioturbation experiment using luminophore tracers over 2–3 weeks in a low energy marine environment. An additional deeper tracer peak was present within three core profiles between 3 and 6 cm below the seafloor (Fig. 8, Panels: 3, 7, 10). Secondary deeper peaks in luminophore tracers were observed in previous studies, where they were attributed to bioadvection of particles into the seabed by infauna (Gerino, 1990; Gerino et al., 2007). The primary result from the bioturbation experiment was that differential tracer depth maxima had the strongest influence on vertical mixing and removal of the tracer, which potentially masked the effect of time. Tracers found below the surface of the cores upon recovery likely resulted from initial transplantation below the ambient seafloor, followed by rapid burial by surrounding sediment into the created depressions. Alternatively, tops of cores with subsurface luminophore maxima could have been buried rapidly by sediment redistributed by organisms. The results from the bioturbation experiment imply that flashflood deposits thicker than 2 cm have greater preservation potential compared to thinner deposits, as the top 2 cm were exposed to the strongest mixing and removal. Measurements taken from cores with tracers initially positioned at or near the seafloor surface (≤2 cm depth) showed that significantly less tracer was recovered and the tracer was more uniformly distributed from the tracer maxima compared to cores with tracers emplaced deeper in the seafloor (3–5 cm) (Fig. 8). Black et al. (2012) conducted a survey of macro infauna at ~23 m water depth on the northern shelf of the GOA, and found 84% of benthic species in the top 3 cm of the seabed. As well, bioturbation rates have been calculated to be 1–3 orders of magnitude stronger in the upper half of the mixed layer depth (Lb) compared to in the lower half (Bentley et al., 2006). Both observations are consistent with the results of this study.

5.4. Biological resuspension The coarsening of surface sediments that occurred during the year following flood events and the inconsistent presence of fine sediments at depth within the seabed (Fig. 5) suggest that surface flood sediments undergo removal. Data obtained from the GoPro camera revealed that demersal fish were present episodically at the study site. These fish, which during regular dives were often observed to actively resuspend sediment, were associated in the photos with the formation of sediment plumes above the seafloor. The enhanced presence of demersal fish in the GoPro footage from September compared to earlier summer months could have been due to clearer visibility and a higher concentration of useable photographs. Alternatively, that result may indicate that there are seasonal fluctuations in fish abundance. If fish were more abundant in September, however, they did not cause systematic increases in turbidity measured by the OBS sensors (Fig. 6a). The sources of variability in turbidity throughout the observational record are unclear and data on fish abundance are qualitative, so conclusions regarding the presence or absence of correlation between fish and sediment resuspension are not possible. Both during dives and within the photographs, Forsskal's goatfish (Parupeneus forsskali) were observed. Yahel et al. (2002) identified these benthivorous fish to be the primary sediment-resuspending fish in the shallow coral reefs of the GOA. They were commonly observed to create sediment plumes > 2 m2 in cross-section, as well as occasionally observed to dig trenches 10–15 cm deep and > 10 cm long within a single resuspension event (Yahel et al., 2002). Yahel et al. (2002, 2008) showed that demersal fish have the capacity to rework the upper centimeters of the seabed within days. The measured rate of fish induced sediment resuspension in the shallow coral reefs of the GOA translated to reworking of surficial sediments over a period of only 5 days (Yahel et al., 2002). As well, Yahel et al. (2008) measured fish resuspension events at ~90 m depth in the Saanich Inlet in Canada, where current speeds were similar to those measured in the present study. They found that reworking of surficial sediments occurred within 2.5 days. These findings shed light on how important fish-induced biological resuspension can be in influencing the preservation of event beds at the sediment surface. Bioturbation also can enhance resuspension, by increasing the roughness height of the sediments (Graf and Rosenberg, 1997) and by breaking down the cohesion between silt-clay sediments (Davis, 1993). The strength of resuspension by bioturbation depends on species and

5.6. Influences of mounds and holes on preservation of flashflood deposits Since biogenic mounds and holes are abundant in this environment (~1 m−2, respectively), and were at times 5 or more times higher and deeper compared to the thicknesses of flood deposit layers, their presence likely contributed to differential flashflood deposition and subsequent burial. The variation in the grain size profiles from the flat seafloor, from within a mound, and from within an adjacent hole (Fig. 10) reinforced that there is diversity in local rates of sediment accumulation, 31

Marine Geology 411 (2019) 22–35

A. Mathalon et al.

resuspension and mixing. The coarsening that occurred at the surface and between the two fine-grained layers in the top 10 cm of the core taken from the flat seafloor provides further evidence that surface flood sediments become winnowed if they are exposed to resuspension and removal. The fine-grained deposit that was buried and preserved 15 cm down the core taken from the mound provides evidence that biological mounds can enhance preservation if they form (or spill sideways) over flashflood deposits, keeping the layers at depth (distance) from the surface, where they are protected from biological and physical reworking. Many mound-building organisms process sediments from below the seafloor surface, and therefore expel mostly coarser, sandsized sediments in this environment. Below the mound, the fine sediment layer was located at an equivalent, pre-burial depth similar to the fine-grained layer at the surface of the core taken from the flat seafloor, which may indicate that these sediments originated from the same source (Fig. 10). The formation rate of biogenic mounds (and the rate they spill sideways from their peaks) is likely much faster than the calculated mean sediment accumulation rate on the shallow shelf of 1.3 mm year−1 (Goodman-Tchernov et al., 2016). This helps explain how 1.5-year-old flashflood layers are preserved decimeters below the seafloor. Fine sediments accumulated with depth within the core taken from the hole, indicating that fine sediments can become trapped within seafloor depressions (Fig. 10). This process was investigated by Yager et al. (1993), who found that fine sediments accumulated within pits formed in the seafloor at rates faster than on the flat seafloor, and that particle flux was positively correlated with increasing pit size. The diversity in the three grain size profiles suggest that biological activities differ locally within the seafloor, and have varying impacts on the preservation of flashflood deposits at a single water depth.

Fig. 12. Diagrams depicting how the preservation potential of an event layer can be reduced if bioturbation and resuspension rates exceed sedimentation rates.

off of the continental shelf to find depocenters of flashflood deposits after known events occurred would be useful in identifying locations for where flashflood deposits may be preserved.

6. Conclusion Recent flashfloods discharging into the GOA delivered distinct finegrained deposits to the shallow shelf. The initial thicknesses of these deposits varied between 1 and 5 cm. Over the course of a year following a flashflood event, the surface flood sediments coarsened from the top down due to the combined effects of bioturbation, resuspension and sediment transport. In general, currents were too weak to resuspend flood sediments at 13 m depth. Water flow preferentially traveled in a southeasterly direction, which influenced the locations to which resuspended particles were transported. Dense water formation in the fall and winter increased southward currents and the transport of resuspended flood sediments to deeper depths. Demersal fish were associated with the formation of sediment plumes at the seafloor, causing sediment resuspension. Bioturbation was most active in the top 2 cm of the seabed, so burial to depths below 2 cm improved the preservation potential of flood layers/lenses. Due to variability in seafloor relief, in addition to locally variable rates of sediment resuspension, mixing and burial processes dominated by biology, lenses of fine-grained flashflood deposits were identified decimeters below the seafloor in some locations. Three mechanisms account for their presence. The first is the deposition of fine flood sediments in holes or depressions in the seafloor, thereby reducing the deposit's exposure to resuspension and increasing the potential for differential burial. The second is burial by multiple flashflood events within less than a year, which could introduce sufficient new sediment to bury earlier fine-grained deposits before they are destroyed by mixing and resuspension. The third mechanism is deep and rapid burial of flood sediments by the construction of sediment mounds by burrowing organisms. With regard to the reconstruction of a record of the frequencies and magnitudes of flashflood events in marine sediments in the GOA, deeper depths on the shelf or the deep basin may be better suited than the shallow shelf, assuming that flood sediments can be identified after transport. The findings from this study are relevant to consider when interpreting the marine sediment record, especially in locations with low sediment accumulation rates, and in locations characterized by active processes of sediment reworking.

5.7. Preservation of event layers in biologically active environments The preserved lenses of fine flashflood deposits below the seafloor are likely attributable to enhanced sediment deposition in holes and depressions, and to rapid post-depositional burial by biogenic mounds or by consecutive flashfloods. The avenues for a distinct flashflood deposit to be preserved as a continuous, uninterrupted layer within the seafloor in this environment are: (1) an initial flood deposit is uniformly substantially thicker than Lb, (2) a deposit is buried quickly and deeply by consecutive floods, bringing an entire portion of the primary deposited flood sediments to depths below Lb (Wheatcroft and Drake, 2003; Bentley et al., 2006). Consecutive flooding events contributing large quantities of sediment in short periods of time would be the most efficient means of uniformly preserving flood deposits in this environment (Bentley et al., 2006). In the oligotrophic GOA, sediment accumulation rates on the shelf between flashfloods are small. Therefore, even if a flood deposit does become buried to depths below exposure to physical and biological reworking, if there is net erosion of sediments, then the bioturbation depth can eventually penetrate deeper into the seabed, thereby reducing preservation potential (Fig. 12) (Kniskern et al., 2014). Better understanding of the mechanisms of flashflood deposit preservation in shallow marine environments will require quantitative measurements of biological and physical sediment resuspension. Highfrequency, year round observations of demersal organism (mostly fish) species, abundance and activities are required, as are measurements of waves (height and period), currents and turbidity. Deployment of sediment traps would provide time-integrated evidence of resuspension (Katz and Crouvi, 2018). Quantification of local, short-timescale bioturbation rates and sediment accumulation rates at a single water depth would increase understanding of the preservation potential of flashflood deposits. It is also relevant to survey and analyze the formation of biological mounds and holes and depressions on the seafloor to get more information on rates and mechanisms of local sediment accumulation and removal. Finally, sampling at deeper water depths on and 32

Marine Geology 411 (2019) 22–35

A. Mathalon et al.

contributing a hand in sampling flashflood events, Moty Ohevia for constructing materials for the bioturbation experiment, and Drs. Magali Gerino and Martin Solan for their advice in creating and analyzing the luminophore tracers. As well, we thank the wonderful people from the Interuniversity Institute of Eilat who dived with us and offered their support throughout this project. This research was funded by the Schulich Foundation and the Natural Sciences and Engineering Research Council of Canada.

Acknowledgements We thank Dr. Bernard Boudreau and Dr. Stephanie Kienast for their useful feedback and discussions. We also thank Asaph Rivlin for helpful discussions about the oceanographic instruments, and Dr. Derya Akkaynak for her Matlab coding contributions. We thank members of the Haifa University Laboratory, and lab managers, Dr. Eli Shemesh and Dr. Nimer Taha, for their technical guidance, Dr. Gal Eyal for Appendix A. OBS calibration curves

Fig. 13. Calibration curves for both channels of the three OBS sensors.

Appendix B. Luminophore concentration calibration curves

Fig. 14. Calibration curves of the 1st cm of cores with tracers on their surfaces that were transplanted into the seafloor and immediately recovered. Error bars represent the standard deviation from the three luminophore areas measured/picture/cm.

33

Marine Geology 411 (2019) 22–35

A. Mathalon et al.

Appendix C. All bioturbation core profiles

Fig. 15. Profiles showing the distributions of the luminophore tracer down to 10 cm below the seafloor in cores that were recovered at three time intervals of 1 week (t1), 3 weeks (t2), and 6 weeks (t3). As well, there are the profiles of three control cores (capped on top and bottom) that were organism free, which were emplaced at t1 and removed at t3. Note that the x-axis range varies per core.

Gerino, M., Frignani, M., Mugnai, C., Bellucci, L.G., Prevedelli, D., Valentini, A., Castelli, A., Delmotte, S., Sauvage, S., 2007. Bioturbation in the Venice Lagoon: rates and relationship to organisms. Acta Oecol. 32 (1), 14–25. https://doi.org/10.1016/j. actao.2007.02.003. Goodman-Tchernov, B.N., Dey, H.W., Reinhardt, E.G., McCoy, F., Mart, Y., 2009. Tsunami waves generated by the Santorini eruption reached Eastern Mediterranean shores. Geology 37 (10), 943–946. https://doi.org/10.1130/G25704A.1. Goodman-Tchernov, B.N., Katz, T., Shaked, Y., Qupty, N., Kanari, M., Niemi, T., Agnon, A., 2016. Offshore evidence for an undocumented tsunami event in the ‘low risk’ Gulf of Aqaba-Eilat, Northern Red Sea. PLoS One 11 (1), e0145802. https://doi.org/10. 1371/journal.pone.0145802. Graf, G., Rosenberg, R., 1997. Bioresuspension and biodeposition: a review. J. Mar. Syst. 11 (3–4), 269–278. https://doi.org/10.1016/S0924-7963(96)00126-1. Howard, J.D., 1975. The sedimentological significance of trace fossils. In: Frey, R.W. (Ed.), The Study of Trace Fossils. Springer, Berlin, pp. 131–146. Israel National Monitoring Program at the Gulf of Eilat (NMP) meteorological station, 2017. Meteo-tech. http://www.meteo-tech.co.il/eilat-yam/eilat_about_en.asp. Katz, T., Crouvi, O., 2018. Sediment flux dynamics over the shallow (25 m depth) shelf of the Mediterranean Sea along the Israeli coast. Mar. Geol. 406, 1–11. https://doi.org/ 10.1016/j.margeo.2018.09.004. Katz, T., Ginat, H., Eyal, G., Steiner, Z., Braun, Y., Shalev, S., Goodman-Tchernov, B.N., 2015. Desert flash floods form hyperpycnal flows in the coral-rich Gulf of Aqaba, Red Sea. Earth Planet. Sci. Lett. 417, 87–98. https://doi.org/10.1016/j.epsl.2015.02.025. Kniskern, T.A., Mitra, S., Orpin, A.R., Harris, C.K., Walsh, J.P., Corbett, D.R., 2014. Characterization of a flood-associated deposit on the Waipaoa River shelf using radioisotopes and terrigenous organic matter abundance and composition. Cont. Shelf Res. 86, 66–84. https://doi.org/10.1016/j.csr.2014.04.012. Lamb, P.M., Mohrig, D., 2009. Do hyperpycnal-flow deposits record river-flood dynamics? Geology 37, 1067–1070. https://doi.org/10.1130/G30286A.1. Manasrah, R., Badran, M., Lass, H.U., Fennel, W., 2004. Circulation and winter deep water formation in the northern Red Sea. Oceanologia 46 (1), 5–23. Milliman, J.D., Syvitski, J.P.M., 1992. Geomorphic/tectonic control of sediment discharge to the ocean: the importance of small mountainous rivers. J. Geol. 100 (5), 525–544. https://doi.org/10.1086/629606. Mulder, T., Migeon, S., Savoye, B., Jouanneau, J.M., 2001. Twentieth century floods recorded in the deep Mediterranean sediments. Geology 29 (11), 1011–1014. https:// doi.org/10.1130/0091-7613(2001)029<1011:TCFRIT>2.0.CO;2. Mulder, T., Syvitski, J.P.M., Migeon, S., Faugères, J.C., Savoye, B., 2003. Marine hyperpycnal flows: Initiation, behaviorand related deposits. A review. Mar. Pet. Geol. 20 (6–8), 861–882. https://doi.org/10.1016/j.marpetgeo.2003.01.003. Otsu, N., 1979. A threshold selection method from gray-level histograms. IEEE Trans. Syst. Man Cybern. 9 (1), 62–66. https://doi.org/10.1109/TSMC.1979.4310076. Pittauerová, D., Kirchner, G., Garbe-Schönberg, D., Herut, B., Nishri, A., Fischer, H.W., 2014. Radionuclides and recent sedimentation and mixing rates in Northern Gulf of Eilat/Aqaba, Red Sea. J. Mar. Syst. 139, 1–8. https://doi.org/10.1016/j.jmarsys. 2014.04.017. Postma, G., 2001. Physical climate signatures in shallow- and deep-water deltas. Glob.

References Baumann, J., Chaumillon, E., Schneider, J.L., Jorissen, F., Sauriau, P.G., Richard, P., Schmidt, S., 2017. Contrasting sediment records of marine submersion eventsrelated to wave exposure, Southwest France. Sediment. Geol. 353, 158–170. https://doi.org/ 10.1016/j.sedgeo.2017.03.009. Ben-Avraham, Z., 1985. Structural framework of the Gulf of Elat (Aqaba), NorthernRed Sea. J. Geophys. Res. 90 (10), 703–726. https://doi.org/10.1029/JB090iB01p00703. Bentley, S.J., Nittrouer, C.A., 2003. Emplacement, modification, and preservation of event strata on a flood-dominated continental shelf: Eel shelf, Northern California. Cont. Shelf Res. 23 (16), 1465–1493. https://doi.org/10.1016/j.csr.2003.08.005. Bentley, S.J., Sheremet, A., Jaeger, J.M., 2006. Event sedimentation, bioturbation, and preserved sedimentary fabric: field and model comparisons in three contrasting marine settings. Cont. Shelf Res. 26 (17–18), 2108–2124. https://doi.org/10.1016/j. csr.2006.07.003. Berger, H., Ekdale, A.A., Bryant, P.B., 1979. Selective preservation of burrows in deep sea carbonates. Mar. Geol. 32, 205–230. https://doi.org/10.1016/00253227(79) 90065-3. Beyth, M., Eyal, Y., Garfunkel, Z., 2011. The geology of the Elat sheet, explanatory notes. In: Rep No. GSI/22/2011. Geological Survey of Israel, Jerusalem 27 pp. Biton, E., Gildor, H., 2011a. Stepwise seasonal restratification and the evolution of salinity minimum in the Gulf of Aqaba (Gulf of Eilat). J. Geophys. Res. 116 (C08022). https://doi.org/10.1029/2011JC007106. Biton, E., Gildor, H., 2011b. The general circulation of the Gulf of Aqaba (Gulf of Eilat) revisited: the interplay between the exchange flow through the Straits of Tiran and surface fluxes. J. Geophys. Res. 116 (C8). https://doi.org/10.1029/2010JC006860. Black, K.D., Calder, L.A., Nickell, T.D., Sayer, M.D.J., Orr, H., Brand, T., Angel, D., 2012. Chlorophyll, lipid profiles and bioturbation in sediments around a fish cage farm in the Gulf of Eilat, Israel. Aquaculture 356–357, 317–327. https://doi.org/10.1016/j. aquaculture.2012.04.049. Boudreau, B.P., 1994. Is burial velocity a master parameter for bioturbation? Geochim. Cosmochim. Acta 58 (4), 1243–1249. https://doi.org/10.1016/0016-7037(94) 90378-6. Davis, W.R., 1993. The role of bioturbation in sediment resuspension and its interaction with physical shearing. J. Exp. Mar. Biol. Ecol. 171 (2), 187–200. https://doi.org/10. 1016/0022-0981(93)90003-7. Fricke, H.W., Schuhmacher, H., 1983. The depth limits of Red Sea Stony Corals: an ecophysiological problem (a deep diving survey by submersible). Mar. Ecol. 4 (2), 163–194. https://doi.org/10.1111/j.1439-0485.1983.tb00294.x. Gerino, M., 1990. The effects of bioturbation on particle redistribution in Mediterranean coastal sediment. Preliminary results. Hydrobiologia 207 (1), 251–258. https://doi. org/10.1007/BF00041463. Gerino, M., Aller, R.C., Lee, C., Cochran, J.K., Aller, J.Y., Green, M.A., Hirschberg, D., 1998. Comparison of different tracers and methods used to quantify biotrubation during a spring bloom: 234-thorium, luminophores and chlrophyll a. Estuar. Coast. Shelf Sci. 46, 531–547. https://doi.org/10.1006/ecss.1997.0298.

34

Marine Geology 411 (2019) 22–35

A. Mathalon et al.

Wheatcroft, R.A., Drake, D.E., 2003. Post-depositional alteration and preservation of sedimentary event layers on continental margins, I. The role of episodic sedimentation. Mar. Geol. 199 (1–2), 123–137. https://doi.org/10.1016/S0025-3227(03)00146-4. Wiberg, P.L., Smith, J.D., 1987. Calculations of the critical shear stress for motion of uniform and henetogeneous sediments. Water Resour. Res. 23 (8), 1471–1480. https://doi.org/10.1029/WR023i008p01471. Yager, P.L., Nowell, A.R.M., Jumars, P.A., 1993. Enhanced deposition to pits: local food source for benthos. J. Mar. Res. 51 (1), 209–236. https://doi.org/10.1357/ 0022240933223819. Yahel, R., Yahel, G., Genin, A., 2002. Daily cycles of suspended sand at coral reefs: a biological control. Limnol. Oceanogr. 47 (4), 1071–1083. https://doi.org/10.4319/ lo.2002.47.4.1071. Yahel, G., Yahel, R., Katz, T., Lazar, B., Herut, B., Tunnicliffe, V., 2008. Fish activity: a major mechanism for sediment resuspension and organic matter remineralization in coastal marine sediments. Mar. Ecol. Prog. Ser. 372, 195–209. https://doi.org/10. 3354/meps07688.

Planet. Chang. 28 (1–2), 93–106. https://doi.org/10.1016/S0921 8181(00)00067-9. Shlesinger, T., Grinblat, M., Rapuano, H., Amit, T., Loya, Y., 2018. Can mesophotic reefs replenish shallow reefs? Reduced coral reproductive performance casts a doubt. Ecology 99 (2), 421–437. https://doi.org/10.1002/ecy.2098. Sommerfield, C.K., Nittrouer, C.A., 1999. Modern accumulation rates and a sediment budget for the Eel shelf: a flood-dominated depositional environment. Mar. Geol. 154 (1–4), 227–241. https://doi.org/10.1016/S0025-3227(98)00115-7. Steiner, Z., Boaz, L., Shani, L., Shimon, T., Omer, P., Revital, B., Jonathan, E., 2016. The effect of bioturbation in pelagic sediments: lessons from radioactive tracers and planktonic foraminifera in the Gulf of Aqaba, Red Sea. Geochim. Cosmochim. Acta 194, 139–152. https://doi.org/10.1016/j.gca.2016.08.037. Walsh, J.P., Corbett, D.R., Kiker, J.M., Orpin, A.R., Hale, R.P., Ogston, A.S., 2014. Spatial and temporal variability in sediment deposition and seabed character on the Waipaoa River margin, New Zealand. Cont. Shelf Res. 86, 85–102. https://doi.org/10.1016/j. csr.2014.07.001. Wheatcroft, R.A., 1990. Preservation potential of sedimentary layers. Geology 18 (9), 843–845. https://doi.org/10.1130/0091-7613(1990)018<0843:PPOSEL>2.3.CO;2.

35