Sediment transport event analysis on the western Adriatic continental shelf

ARTICLE IN PRESS

Continental Shelf Research 27 (2007) 431–451 www.elsevier.com/locate/csr

Sediment transport event analysis on the western Adriatic continental shelf A.M.V. Faina,, A.S. Ogstonb, R.W. Sternbergb a

Environmental Science Associates, 225 Bush Street, Suite 1700, San Francisco, CA 94104, USA b School of Oceanography, University of Washington, Box 357940, Seattle, WA 98195, USA

Received 7 December 2004; received in revised form 20 March 2005; accepted 24 March 2005 Available online 16 January 2007

Abstract The sediment-transport mechanisms that contribute to and redistribute the modern sediment deposits on the western Adriatic continental shelf were evaluated utilizing data collected from two instrumented benthic tripods deployed at 12-m water depth, one in the northern Adriatic basin on the Po River subaqueous delta, and the other in the central Adriatic basin on the Pescara River shelf. Sediment-resuspension events driven by cold, northeasterly Bora winds dominate the along-shelf transport climatology at both tripod locations, but at the Po delta site, the southwesterly Scirocco wind events also play a significant role. At the Pescara shelf site, interaction between Bora wind-driven currents and the Western Adriatic Coastal Current strongly contributes to the resuspension and advection of suspended sediment. Interannual variability of the forcing mechanisms (including strength, frequency, and relative mix of Bora and Scirocco wind events) is evident in the three winters of data collected on the Po River subaqueous delta. In both types of wind events, and throughout all years of data collection, the net along-shelf sediment transport is significantly larger than the net acrossshelf transport at the 12-m sites. This may be characteristic of low-energy environments, where sediment resuspension and transport occurs in such shallow water that it is not subjected to strong downwelling features characteristic of higherenergy environments. r 2006 Elsevier Ltd. All rights reserved. Keywords: Sediment flux; Sediment resuspension; In situ measurements; Nearbed sediment concentration; Adriatic sea; Po river; Pescara river

1. Introduction Over the past 30 years, observational studies of nearbed sediment transport have occurred on many continental shelves. Winter storms and associated waves, as well as river input, have been found to be the dominant forcing mechanism of sediment flux in Corresponding author. Tel.: +1 415 962 8490; fax: +1 415 896 0332. E-mail address: [email protected] (A.M.V. Fain).

0278-4343/$ - see front matter r 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.csr.2005.03.007

high-energy environments (e.g., Drake and Cacchione, 1985; Sherwood et al., 1994; Ogston and Sternberg, 1999). The strength of storms, the forcing of upwelling and downwelling conditions, and timing of storms with respect to river discharge are important factors in determining the amount of sediment transported along and across the continental shelf, and ultimately the fate of particles discharged to the ocean. The majority of these studies have been undertaken on relatively short time-scales, monthly to seasonal, and in high-energy

ARTICLE IN PRESS 432

A.M.V. Fain et al. / Continental Shelf Research 27 (2007) 431–451

environments. In studies on highly energetic coastal areas such as the Eel, Russian, and Columbia River continental shelves, episodic events have been found to dominate the transport regime (Drake and Cacchione, 1985; Sternberg, 1986; Sherwood et al., 1994; Ogston and Sternberg, 1999). From 1995 to 2000, one of the few long-term comprehensive shelf studies (STRATAFORM) occurred on the high-energy northern California shelf, offshore of the Eel River. Scientists involved in this study showed that there was significant interannual variability in terms of sediment resuspension, nearbottom sediment flux, and dispersal direction on the Eel continental shelf (Ogston et al., 2004). The magnitude and distribution of sediment-suspension and transport events showed a strong seasonality, controlled primarily by winter storms and episodic floods (Ogston and Sternberg, 1999), the concurrence of which provided conditions for strong across-shelf transport due to both downwelling processes and fluid-mud formation (Ogston et al., 2000; Traykovski et al., 2000). In the lower-energy environment of the Ebro continental shelf, off the southeast coast of Spain, storms cause significant sediment transport in shallower water, i.e., on the inner shelf, and minimal sediment transport in deeper areas. On the Ebro shelf, sediment transport has been studied in a number of short deployments (Palanques et al., 2002; Puig et al., 2001) and alongshelf sediment transport was found to be an order of magnitude greater than across-shelf during a 3month period (Palanques et al., 2002). Although lower-energy settings have not been studied extensively, processes controlled by the interaction between storms and floods that impact the offshelf sediment flux may differ from those on higher energy shelves. As a contrast to the studies on the Eel River shelf, one component of the present study (EuroSTRATAFORM) has focused on sediment transport and accumulation processes in the Adriatic Sea. The region is characterized as an epicontinental basin with a relatively low-energy wave environment and spatially variable forcing by wind and river discharge. Studies of the delivery and redistribution of sedimentary deposits were undertaken along the western Adriatic continental shelf from December 2000 to May 2003. Bottom boundary layer (BBL) instrumentation was deployed on the Po River subaqueous delta to provide long-term monitoring of nearbed flow and sediment transport from 2000 to 2003. Additionally, BBL instrumentation was

deployed on the Pescara River shelf, adjacent to one of the largest Apennine Rivers as part of a multiinvestigator Po and Apennine Sediment Transport and Accumulation (PASTA) observational and modeling experiment. This focused study was designed to explore the regional variations of transport processes during winter 2002–2003. One important component of the overall project is directed at understanding the pathways and mechanisms of sediment transport that both create and redistribute the modern sediment deposit along the western Adriatic continental shelf. The dominance of particular sediment-transport pathways and mechanisms are dependent on the interactions of specific storm systems with sediment supplied from rivers to the coastal zone. Not only is understanding the long-term variability of a transport system relevant for ecological and environmental studies, but it is also relevant for geologic studies of the ultimate fate of particulate matter in the marine environment and the resulting depositional structure. In this paper, sediment-transport events along the western Adriatic shelf are described during various storm and river flow conditions. The main objective is to characterize spatial differences in frequency, duration, and magnitude of forcing mechanisms of sediment resuspension, and their implication for sediment deposition. The temporal (interannual) variability of transport processes on the Po River subaqueous delta is also addressed. The characteristics of transport events provide a basis to discuss implications for redistribution of modern sedimentary deposits in this low-energy system and to compare to areas of differing wave energy and river input. 2. Study area and event forcing The Adriatic Sea is a semi-enclosed basin consisting of three morphologic domains: the northern, central and southern basins (Fig. 1). This analysis examines transport processes on the western side of the northern and central basins. The northern basin, into which the Po River flows, has shallow water depths and low slopes, while the central basin, into which the Apennine rivers flow, reaches depths of 4200 m and has steeper slopes (Cattaneo and Trincardi, 1999). Multiple rivers discharge freshwater and sediment on both the east and west sides of the Adriatic Sea, but only on the western Adriatic has a significant nearshore mud

ARTICLE IN PRESS A.M.V. Fain et al. / Continental Shelf Research 27 (2007) 431–451

Fig. 1. Map of the northern and central basins of the Adriatic Sea showing the locations of the two study sites offshore of the Po and Pescara rivers. The Porto Tolle (PT) wind station, located on the subareal Po delta and maintained by E.N.E.L. S.p.A., is marked by +, Po River discharge gauging station (GS) at Pontelagoscuro is marked by &, and the Ancona wave buoy (AB) is marked by D. The Po River is the major contributor of freshwater and sediment to the northern basin, and the combined Apennine rivers act as a distributed source in the central basin.

wedge formed (Correggiari et al., 2001). The Po subaqueous delta is at the northern end of this rapidly formed mud deposit which stretches 600 km along the Italian coastline in a relatively continuous, narrow band (Correggiari et al., 1996). The Po River is the largest Italian river and is the primary source of freshwater and sediment to the Adriatic Sea (Fig. 1). It drains an area of over 70,000 km2, an area one-quarter the size of Italy (Nelson, 1970). The drainage basin receives input from both the Apennine Mountains to the south and from the Alps to the north and west. The Po carries a sediment load of 1.4  107 tons/year and has an average discharge at Pontelagoscuro of 1500 m3/s (Cattaneo et al., 2003). The Po River flows into the sea via five distributaries with more than half of the river flowing through the Pila mouth (Nelson, 1970). The Po hydrograph tends

433

toward two seasonal high-discharge peaks (autumn and spring) and contributes up to one-third of the total annual freshwater input into the Adriatic Sea (Kourafalou, 1999). The river has been channelized by construction of artificial levees (Marchetti, 2002), which has resulted in reduced overbank flow and rapid delivery of sediment. Along the 12-m isobath the seabed sediment on the Po subaqueous delta is characterized as mixed silt and clay with a small amount of sand (Palinkas et al., 2005). The Apennine rivers drain into the central basin of the Adriatic Sea, and although individually small, provide significant amounts of freshwater and sediment as a more distributed source. The Apennine Mountains are very steep, easily eroded, and close to the coast, causing a great potential for high sediment discharge (Milliman, 2001; Syvitski and Kettner, 2007). The Apennine rivers south of the Po River basin and north of the Gargano Penninsula have a total annual sediment discharge of approximately 3.2  107 tons/year (Frignani and Langone, 1991), larger than the input from the Po River, and are characterized by episodic discharges. The Pescara River, one of the largest Apennine rivers, has been estimated to have a sediment discharge of 9  105 tons/year and an average flow of 61 m3/s (Milliman and Syvitski, 1992). Dams, borrow pits, and breakwaters have reduced the water and sediment flow in the Pescara River during the last few decades (Vittori et al., 2001). Offshore of the Pescara River, along the 12-m isobath, the seabed sediment is generally coarse silt to fine sand (Passega et al., 1967; George et al., 2007). Thermohaline forcing drives the general circulation pattern of the Adriatic Sea. This results from freshwater input in the north and exchange with saltwater from the Mediterranean Sea in the south, causing northward flow along the eastern Adriatic, and southward flow along the western Adriatic that is boundary intensified and is known as the Western Adriatic Coastal Current (WACC; Zavatarelli et al., 2002). Strong storm winds generate short-term circulation features that can enhance or override the general cyclonic circulation pattern. Two prevalent storm patterns exist in the Adriatic basin called Bora and Scirocco. Bora winds are cold and come from the northeast causing downwellingfavorable currents along the western Adriatic coast, while Scirocco winds are warmer and come from the southeast causing upwelling-favorable currents along the western Adriatic coast (Kourafalou, 1999). Studies have shown that Bora wind events

ARTICLE IN PRESS 434

A.M.V. Fain et al. / Continental Shelf Research 27 (2007) 431–451

enhance the WACC, and may cause dense-water formation in the northern Adriatic (Hendershott and Rizzoli, 1976). Bora winds cause a cyclonic gyre in the northern Adriatic and a small anti-cyclonic gyre south of the Po region (Paklar et al., 2001). The double gyre system has been documented in various modeling (Paklar et al., 2001), in situ (Zore-Armada and Gacic, 1987), and remote sensed studies (Sturm et al., 1992; Orlic et al., 1994). The Bora winds blow across the main Adriatic axis (via the Alp mountain gaps in Slovenia and Croatia) and thus waves generated under Bora winds are predicted to have relatively short wavelength due to their short fetch. The Scirocco winds blow along the main Adriatic axis and have been shown to advect low-salinity Po waters towards the shallow regions of the northern Adriatic and can reverse the southward baroclinic WACC (Orlic et al., 1994; Kourafalou, 1999). These events push water against the northern Adriatic coast resulting in storm surges and flooding in northern Italian cities such as Venice (Orlic et al., 1994; Pirazzoli and Tomasin, 2002). The Scirocco winds produce significant swell waves due to their long fetch along the basin axis. 3. Methods 3.1. River, wave, and wind data River, wave, and wind stations have been maintained along the western Adriatic by various Italian governmental and private organizations, and data were generally available for the sampling period of this study (December 2000–May 2003). Daily mean river discharge values were obtained at Pontelagoscuro (approximately 75 km upstream of the Po River mouth, Fig. 1) and at Sta. Teresa (approximately 19 km upstream of the Pescara River mouth). The Po River water discharge data was obtained from ‘‘Ufficio Idrografico del Magistrato per il Po (Parma)’’ and the Pescara River discharge from the ‘‘Servizio Idrografico di Pescara’’. A wave buoy offshore of Ancona in approximately 55-m water depth (431370 0000 N, 131510 0000 E, Fig. 1) provided wave height and period data every 3 h for most of the sampling period (envirtech.org). The Porto Tolle wind station, at 10m height near the apex of the Pila and Tolle distributaries of the Po River, provided hourly direction and magnitude of winds for the entire sampling period (ENEL S.p.A.). Additionally, the Local Area Model Italy (LAMI) 7-km, non-hydro-

static, model provided values of wind direction and magnitude throughout the western Adriatic Sea every 3 h from October 2002 to May 2003. 3.2. Shelf tripods Data from BBL tripods deployed at two sites were used in this analysis, one located on the Po River subaqueous delta with deployments between January 2001 and May 2003, and the other located 300 km to the south on the Pescara River shelf with deployments from October 2002 to May 2003 (Fig. 1). Both tripods were located at 12-m depth to evaluate the near-shore sediment transport processes along the western Adriatic continental shelf. The latter deployments (winter 2002–2003) were part of the joint PASTA experiment that included eight additional instrument packages deployed along the western Adriatic continental shelf. The Po tripod deployments occurred over three winters. The first series of Po delta deployments, designated PoI, began in January 2001 and continued through April 2002, thus documenting conditions through most of the winter 2000–2001 and the winter 2001–2002 (Fig. 2). The second series of Po delta deployments, designated PoII, continued from October 2002 to February 2003, documenting the third winter period (winter 2002–2003). Although redeployed in February 2003, the PoII tripod was not recovered probably due to fishing in the area. At the Po delta, the initial deployment site (PoI) was located at the edge of a mussel farm to provide safety from fishing activity. Concerns over the impacts of local biological processes on the PoI data were addressed in the later PASTA deployments (PoII) by moving the tripod site 2.9 km southwest of the initial Po tripod location. Although the two tripod locations are different, they are located at the same water depth and in a region where the bathymetric contours do not vary significantly and thus the resuspension processes should be similar at both sites. On the Pescara shelf, the BBL tripod was located 5.3 km southeast of the Pescara River mouth, an area predicted to be in the region of influence of both flood-induced plumes and the WACC. Depth at the site was 12 m to be consistent with the depth of the Po delta study site. This tripod was located as the innermost site of a shore-normal transect made up with another benthic tripod at 20-m depth and an instrumented mooring at 50-m depth (Puig et al.,

ARTICLE IN PRESS A.M.V. Fain et al. / Continental Shelf Research 27 (2007) 431–451

435

Fig. 2. Time-series of the major forcing for resuspension events in the Adriatic Sea, the defined winter periods (see Section 4.1), and the instrument deployment periods: Po River discharge at Pontelagoscuro (a), Pescara River discharge at Sta. Teresa (b), wave height offshore of Ancona (c), and Porto Tolle wind speed (d) from September 2000 to June 2003.

2007). The Pescara shelf deployments occurred from October 2002 to May 2003. The timeline for the complete data set for this study is shown in Fig. 2. Additional results from the experiment relating to sediment dynamics on the shelf can be found in Nittrouer et al. (2004), Sherwood et al. (2004), Puig et al. (2007), Traykovski et al. (2007). The tripods generally had similar capabilities of measuring nearbed current flows, suspended-sediment profiles, water-column currents, pressure, temperature and salinity; however specific sensor types varied between the tripods (Table 1). For example, the early Po River tripod (deployments at the PoI site, Table 1) contained two electromagnetic (EM) current meters while the later Po River tripod and the Pescara tripods (deployments at the PoII and Pescara sites) (Table 1) each contained one EM current meter and one acoustic doppler velocimeter (ADV). In this study, response of the EM current meter and the ADV is assumed to be similar in

terms of direction and magnitude, a valid assumption as hourly averaged values are used. Current meters, regardless of type, were positioned to measure at 30 and 100 cm above the bed (cmab). Optical backscatter sensors (OBS) were located at three elevations above the bed for PoI (12, 30, 100 cmab) and five elevations for the PoII and Pescara tripods (13, 30, 80, 100, 200 cmab). The combination of currents and suspended-sediment concentration (SSC) at 30 cmab are the primary data focused on in this study as they provide the most complete data record and are readily comparable between times and locations. The elevations of the other ancillary instruments including the ADCP, pressure sensor, temperature sensors and CTD on each tripod are summarized in Table 1. Measurements were made on an hourly basis with sensor output sampled at a 2 Hz rate over a 360 or 480 s burst for both the Po and Pescara tripods (Table 1). The current meter and the OBS measurements are the primary data considered in the following

ARTICLE IN PRESS 436

A.M.V. Fain et al. / Continental Shelf Research 27 (2007) 431–451

Table 1 Sensors mounted on the bottom boundary layer tripods used in this study, their elevations in cm above the seabed (cmab), and sampling schedule Site (location)

Deployment dates

Instrument

Nominal elevation above the bed (cmab)

Hourly sampling scheme

PoI BBL tripod (441530 2800 N, 121320 600 E)

January 2001–April 2002

PoII BBL tripod (441520 800 N, 121300 2500 E)

October 2002–February 2003

EM current meters (2) OBS (3) ADCP (300 kHz) EM current meter ADV OBS (5) ADCP (300 kHz) EM current meter ADV OBS (5) ADCP (300 kHz)

30, 100 12, 30, 100 0.5-m bins 3 to 30 100 13, 30, 80, 100, 0.5-m bins 3 to 30 100 13, 30, 80, 100, 0.5-m bins 3 to

480-s 480-s 480-s 480-s 360-s 480-s 480-s 480-s 360-s 480-s 480-s

Pescara BBL tripod (421270 4600 N, 141150 5600 E)

October 2002–May 2003

analysis because they relate most directly to forcing mechanisms (currents, wave-generated oscillatory flows), sediment response, and sediment flux. Additionally, the auxiliary winds and river flow data are used to help classify the various resuspension events. 3.3. Data processing 3.3.1. Currents Current observations were averaged over the burst duration collected each hour and were rotated to local bathymetry (PoI-301T; PoII-241T; Pescara3151T) for analysis purposes. Currents are shown relative to the bathymetry in all figures (across- and along-shelf). The wave-orbital velocity was calculated using the root-mean-square (RMS) of the velocity fluctuations around the mean velocity in each hourly burst. The wave height and period was evaluated from the spectrum of pressure within each burst. In each hourly burst, the dominant period was identified in the spectrum, and the significant wave height determined from the spectral energy. 3.3.2. Suspended-sediment concentration SSC was determined from the OBS measurements that were calibrated with surficial seabed sediment both before and after each deployment. Calibrations were performed by mixing specific concentrations of suspended sediment in a tank and determining a relationship between the concentration and the sensor response. Additionally, the SSC data from the tripods was compared with in situ SSC from water-column profiling studies. The values of SSC compared favorably, although no

11 mab

200 11 mab

200 11 mab

(2-Hz) (2-Hz) (2-Hz) (2-Hz) (2-Hz) (2-Hz) (2-Hz) (2-Hz) (2-Hz) (2-Hz) (2-Hz)

burst burst burst burst burst burst burst burst burst burst burst

high concentration in situ data were obtained (profiling studies could not be conducted during energetic resuspension events). Each of the OBS was adjusted to a different gain setting; therefore the OBS range and resolution response to SSC differed accordingly. During more energetic events throughout the PoII, winter 2002–2003 deployment, one or more OBS reached its maximum output value for a period of time, creating a clipped signal. At these times of maximum response, the SSC and thus the results of flux calculations are underestimates. No evidence was seen of the very-high-concentration turnaround of OBS response that has been documented in other studies (e.g., Kineke et al., 1996; Ogston et al., 2000). Although mitigated with antifouling agents, periods of biofouling occurred in this shallow environment as is evident in the data stream during spring, summer, and autumn. Limited signal drift due to biofouling appeared as a relatively slow, predictable increase of the background SSC and was removed from the data series. At times OBS data had an unrecoverable biofouling signal that appeared as rapidly fluctuating noise or such high levels of background response that no event response signal was evident. To prevent overestimates of sediment flux, the unrecoverable biofouling signal periods were not used in this analysis. 3.3.3. Resuspension event identification Sediment-resuspension events within the winter storm period were identified by periods of high SSC and wave-orbital velocity (Urms) that lasted 12 or more hours. During these events, the SSC exceeded 200 mg/L at 30 cmab, 20 mg/L at 100 cmab, and the

ARTICLE IN PRESS A.M.V. Fain et al. / Continental Shelf Research 27 (2007) 431–451

437

Urms exceeded 9 cm/s. The combination of SSC and Urms exceedance was used to prevent misidentification of storm events when other processes caused an enhanced SSC signal (e.g. trawling, fish congregation, biofouling). The sediment-resuspension events were categorized as being due to Bora, Scirocco, or Mixed wind conditions. Both modeled and measured wind magnitude and direction were used to classify wind conditions. The Porto Tolle winds were used as the primary indicator of storm magnitude and wind direction for winters 2000–2001 and 2001–2002. The Porto Tolle and LAMI winds were used to indicate wind conditions during winter 2002–2003. During the period of this study, examination of Porto Tolle wind records and LAMI modeled winds in the Po region showed the same general trends for both observed and modeled wind characteristics for the larger events, although the specific magnitude and direction could be quite variable, especially for smaller events. The Bora and Scirocco events were identified by resuspension that coincided with winds blowing from the northeast (0–90 1T) and southeast (90–180 1T), respectively, for 12 or more hours. Mixed events were identified as events where wind directions varied considerably, and included winds from the land, northwest and/or southwest.

One aspect of interannual variability is the length of time over which energetic storm events occur each year, herein defined as the winter storm period. The vast majority of seabed reworking occurs in this period, and the remainder of this paper deals only with this period. Winters in the Adriatic Sea were characterized from nearbed time-series data by increased frequency of high sediment concentrations, wave-orbital velocities, and along-shelf currents. Specifically, the winter period was defined when strong along-shelf currents (410 cm/s) and enhanced wave-orbital velocities (45 cm/s) recurred within 15 days at both study sites (Table 2). Based on these constraints the winter periods documented were 114 and 118 days for the winters 2000–2001 and 2001–2002 (PoI deployments only), and 154 days for winter 2002–2003 (PoII and Pescara deployments) (Table 2). The average duration of the three winters is 128 days. The three winter periods identified during this study are: 22 December 2000–14 April 2001 (winter 2000–2001), 10 November 2001–7 March 2002 (winter 2001–2002), and 4 November 2002–6 April 2003 (winter 2002–2003) (Table 2, Fig. 2).

3.4. Sediment flux

4.2. River flow, wave, and wind conditions

Hourly sediment flux at both 30 and 100 cmab was calculated from tripod data as the product of the hourly averaged current and SSC. In this study, the calculated sediment flux is generally presented from the 30-cmab elevation except in one case where EM current meter problems were experienced at that level (i.e., the PoI deployment from January 2002 to April 2002). In that case, the currents at 100 cmab and the SSC at 30 cmab were used to estimate the flux at the 30-cmab level. The currents at 100 cmab showed slightly higher magnitudes and thus flux estimates are overestimated. The resulting cumulative flux is only slightly impacted, as a limited number of large events occurred in this time period. The net, or total, sediment flux within a time period or an event was determined by summing the hourly flux components over the period of the event, including the positive or negative directions and thus the net flux could result in a small value if currents changed direction through the event.

River flow varied considerably over the three winters of data collection. For example, the mean Po River discharge was high during the first winter (2000–2001) and third winter (2002–2003) (Fig. 2a) with average discharge levels of 2436 and 2219 m3/s, respectively (Table 2). Discharge over the second winter (2001–2002) was relatively low with average discharge level of 906 m3/s. Maximum discharge events in the Po region and associated sediment discharge surrounding the study period also experienced wide extremes. In October 2000, just prior to the first tripod deployment, a 100-year flood event of the Po River occurred in which the mean daily discharge at Pontelagoscuro peaked at almost 10,000 m3/s. Additionally, one of the largest Po River discharge events of the decade occurred in November 2002 when instrumentation was deployed, with a mean daily discharge of almost 8,000 m3/s. In contrast, during the low-discharge winter (2001–2002), the Po River peaked at a mean daily discharge of less than 3000 m3/s (Fig. 2a).

4. Results 4.1. Winter storm period

ARTICLE IN PRESS 438

A.M.V. Fain et al. / Continental Shelf Research 27 (2007) 431–451

Table 2 Mean currents, wave-orbital velocity, and SSC calculated at 30 cmab for winters at Po delta and Pescara shelf sites. Additionally, mean Po and Pescara River flow during each winter is included

Time period of selected winters Number of days of winter Number of days of data Along-shelf velocity (cm/s)a Across-shelf velocity (cm/s)a Wave-orbital velocity (Urms; cm/s) Percentage of time Urms410 cm/s Number of days Urms410 cm/s SSC (mg/L) River flow (m3/s) a

Winter 2000–2001 (PoI)

Winter 2001–2002 (PoI)

Winter 2002–2003 (PoII)

Winter 2002–2003 (Pescara)

Dec 22–Apr 14 114 78 2.8 1.4 4.5 10 13 90 2436

Nov 10–Mar 7 118 118 0.6 1.1 3.9b 10 12 51 906

Nov 4–Apr 6 154 109 3.8 0.7 6.5 23 35 141 2219

Nov 4–Apr 6 154 154 5.0 0.6 6.0 22 34 127 52

Negative values indicate southward and seaward directions for along-shelf and across-shelf, respectively. The 100 cmab data were used to determine wave-orbital velocity.

b

River flow in the Pescara River, as with most of the Apennine rivers, is significantly lower than the Po River, and more episodic (Fig. 2b). The daily average flow is generally between 40 and 60 m3/s. Winter flood peaks tend to reach values of 100 m3/s. During the PASTA experiment (winter 2002–2003), three floods occurred which exceeded any of the previous three years records. On 8 December 2002, the river flow reached 336 m3/s; on 25 December 2002 it reached 121 m3/s; and on 25 January 2003 it reached 364 m3/s. Wave conditions for the complete 3-year period measured from the wave buoy at Ancona in water depth of 55 m illustrate the annual variation of wave characteristics for this region (Fig. 2c). Wave heights are significantly greater during the winter months throughout the Adriatic Sea, with a maximum significant wave height of 4.2 m at 55-m depth (Ancona wave buoy) and of 2.9 m at 12-m depth (at the Po delta and Pescara shelf sites). Comparison of significant wave height calculated from pressure variation at the 12-m measurement sites on a year-to-year basis shows considerable variability (Fig. 3a). For example, at the Po delta site a 1.0-m wave height was exceeded 5% of the time during the first two winters (2000–2001 and 2001–2002) and 21% of the time during the third winter (2002–2003). At the Pescara River site, a 1.0m wave height was exceeded 19% of the time during winter 2002–2003 (similar to the 21% value for the same winter at the Po River site). Nearbed wave-orbital velocities observed at the Po delta site followed a similar pattern as the wave height. The average wave-orbital velocity was 4.5

and 3.9 cm/s during winters 2000–2001 and 2001–2002, respectively, while the average waveorbital velocity was 6.5 cm/s during winter 2002–2003 (Table 2). In comparison, the Pescara shelf had an average wave-orbital velocity of 6.6 cm/s during winter 2002–2003, similar to the Po delta site during the same winter. Exceedance values of wave-orbital velocity over each winter at the Po and Pescara sites also show year-to-year variability (Fig. 3b). A nominal threshold value of 10 cm/s was chosen as a typical value of wave-orbital velocity (Urms) that can resuspend unconsolidated fine-grained sediment (Komar and Miller, 1975). The percentage of time that the wave-orbital velocity exceeded the threshold was calculated during the three winter periods. The percent exceedance values were 10%, 10%, and 23% for the three winters at the Po delta, respectively, and was 22% at the Pescara shelf during winter 2002–2003 (Table 2). Thus at the Po delta, threshold exceedance varied by a more than a factor of two between the lower energy winters (2000–2001 and 2001–2002) and the higher-energy winter (2002–2003). Winds primarily came from the northeast (Bora) direction (43% of the observations throughout the study period) at the Porto Tolle gauging station, and secondarily from the southeast (Scirocco) direction at 22% of the time (Fig. 2d). Wind speeds during the Bora events were generally much greater than during the Scirocco events. The LAMI modeled winds, during storm events, showed similar patterns of Bora and Scirocco events, although detailed winds differed between those gauged and modeled. Wind events were short, lasting

ARTICLE IN PRESS A.M.V. Fain et al. / Continental Shelf Research 27 (2007) 431–451

439

Fig. 3. Percent exceedance of significant wave height (a) and wave-orbital velocity (b) from pressure and velocity observations (100 cmab), respectively, on the Po subaqueous delta and the Pescara River shelf during the winter periods.

approximately one to two days. Both the Po River and Pescara River discharge peaks were not necessarily correlated with peaks in winds. Po River floods were long relative to the wind events, due to the large drainage basin. Flood discharge levels last on the order of weeks, and thus are not necessarily associated with individual sediment-resuspension events. The Pescara flood peaks are presently controlled by dams, and did not appear associated with wind events over this study. 4.3. Along- and across-shelf currents In general, the time-series of along-shelf and across-shelf velocity show small tidal oscillations, with major wind-driven current velocity events superimposed (see Figs. 4a and b, Figs. 5a and b, and Figs. 6a and b for the time series during winters 2000–2001 and 2001–2002 at the PoI site, winter 2002–2003 at the PoII site, and winter 2002–2003 at the Pescara site, respectively). Averaged over the winter period, the along-shelf velocity at the Po delta was small and directed to the north during all winters (Table 2). In contrast, the mean currents on the Pescara continental shelf were directed towards the south (Table 2, Fig. 6). At both sites, the winteraveraged across-shelf currents were low and variable between the years (Table 2), and probably within the level of error for compass and current meters.

During resuspension events at both the Po and Pescara sites, the along-shelf currents were significantly enhanced and became greater than the across-shelf currents (Figs. 4b, 5b and 6b). At the Pescara shelf site the mean southerly current (WACC) was enhanced by the Bora storm events. The strongest along-shelf currents occurred during, or shortly after, the highest wave-orbital velocities that correspond with the strongest winds (Fig. 6). At times, currents driven by the WACC and enhanced by wind-driven forcing were of a magnitude to be sufficient (410 cm/s) for sediment resuspension to occur. The tidal range and currents in this microtidal system were higher in the Po region than in the Pescara region. During the deployment between October 2002 and February 2003, the mean tidal range at the Po site was 64 cm, while at the Pescara tripod it was 30 cm. Over the same time period, the significant tidal currents at the Po site were 6.8 cm/s, while at the Pescara site they were 4.9 cm/s. 4.4. Sediment suspension and flux Between the storm events, within the winter periods, the SSC near the bed was generally less than 50 mg/L. This value is consistent with that observed by Fox et al. (2004). During storms, the observed nearbed SSC was often 10–60 times greater than that observed between storms. During

ARTICLE IN PRESS 440

A.M.V. Fain et al. / Continental Shelf Research 27 (2007) 431–451

Fig. 4. Time-series observations from the Po delta during the two winters at the PoI site, showing along-shelf velocity (a), across-shelf velocity (b), nearbed wave-orbital velocity (c), suspended-sediment concentration (d), and along-shelf and across-shelf sediment flux (e). The beginning of the events analyzed in this study are marked with vertical lines; Bora events are marked by solid lines, Scirocco events marked by large dashed lines, and Mixed events marked by small dashed lines.

winter 2000–2001 and 2001–2002, the maximum concentration observed at the sensor located nearbed (12–13 cmab, data not shown) was 2.0 g/ L, and during the winter 2002–2003, it was greater than 3.1 g/L (tripod settling caused this sensor to be closer to the seabed than the designed 13 cm height). During the Pescara deployment the maximum concentration was 3.0 g/L. Only in winter 2002–2003 at the PoII site were the measurements at 13 and 30 cmab OBS sensors limited by instrument gain settings, and peak concentrations measured are considered minimums. The cumulative sediment flux at 30 cmab for the sampling period that surrounds the winter desig-

nations illustrates the net direction of nearbed sediment transport and the contrast between along-shelf and across-shelf components. During the winters of 2000–2001 and 2001–2002 the net flux was slightly southward with a zero or slight offshore component (Fig. 4e) at the PoI site. During the winter of 2002–2003, the net flux was moderately northward with no offshore component at the PoII study site (Fig. 5e), but flux estimates are minimum estimates due to the OBS sensor clipping. Based on observations during the prior 2 years, the maximum SSC would likely have reached approximately twice the value of the clipped maximum, thus flux estimates are likely

Fig. 5. Time-series observations from the Po delta during the one winter at the PoII site, showing along-shelf velocity (a), across-shelf velocity (b), nearbed wave-orbital velocity (c), suspended-sediment concentration (d), and along-shelf and across-shelf sediment flux (e). Events are marked as in Fig. 4. Note the large storms denoted with arrows. These are the largest storms observed in the three winters of this study.

ARTICLE IN PRESS A.M.V. Fain et al. / Continental Shelf Research 27 (2007) 431–451

441

ARTICLE IN PRESS 442

A.M.V. Fain et al. / Continental Shelf Research 27 (2007) 431–451

Fig. 6. Time-series observations from the one winter at the Pescara River shelf site, showing along-shelf velocity (a), across-shelf velocity (b), nearbed wave-orbital velocity (c), suspended-sediment concentration (d), and along-shelf and across-shelf sediment flux (e). Events are marked as in Fig. 4.

more than 50% of their actual value (most of the data is not clipped, only the peaks). Clipped peaks occurred during both northward and southwardoriented events, and thus the net flux direction is thought to be uninfluenced by the instrument limitations. On the Pescara shelf, the cumulative flux was strongly southward and slightly onshore (Fig. 6e). Over all winters at both study sites it is noteworthy that the along-shelf sediment flux was at least an order of magnitude greater than the across-shelf flux.

4.5. Transport event climatology The prevalent storm systems in the Adriatic basin, Boras and Sciroccos, represent major forcing for inner-shelf flows, upwelling and downwelling conditions, and wave generation. For these reasons, the identification and comparison of individual storm events within the time series from the Po delta and Pescara shelf provides input for understanding the characteristics of sediment transport in the region. For comparison purposes, the resulting

ARTICLE IN PRESS A.M.V. Fain et al. / Continental Shelf Research 27 (2007) 431–451

443

Table 3 Event summary at Po delta and Pescara shelf sites. Data was evaluated at 30 cmab except where noted and includes SSC, currents, waves, and flux Event type: B-Bora, S- Winter 2000–2001 Scirocco, M-Mixed (PoI)

Event recurrence (days) Number of events

B S M Event duration (days) B S M Mean concentration (mg/L) B S M Mean current speed (cm/s) B S M Mean wave-orbital velocity (cm/s) B S M B Net along-shelf flux (g/cm2)a S M Net across-shelf flux (g/cm2)a B S M

6 6 5 1 2.2 1.2 1.5 219 183 274 9.6 8.6 9.5 9.2 10.3 14.2 1568 213 24 366 219 18

Winter 2001–2002 (PoI)

Winter 2002–2003 (PoII)

Winter 2002–2003 (Pescara)

13 7 1 1 1.7 0.8 1.0 263 205 148 14.8 6.0 4.0 13.0b 12.6b 8.6b 1647 2 28 540 75 8

9 8 3 2 1.3 2.0 0.7 354c 354c 371c 13.5 7.0 4.0 14.0 22.4 13.7 1709c 364c 81c 97c 125c 24c

12 9 3 1 1.5 1.2 1.4 681 306 738 19.2 9.1 12.1 17.2 13.4 16.7 12958 599 593 294 83 44

a

Minimum estimates due to instrument limitations. The 100 cmab data was used for determining wave-orbital velocity for comparison. c Negative values indicate southward and landward directions for along-shelf and across-shelf, respectively. b

climatology of Bora, Scirocco and Mixed events during winters are summarized in Table 3 including their associated currents, wave-orbital velocity, and suspended-sediment flux characteristics. This comparison shows that 9–13 wind-driven resuspension events occurred during each winter. Events occurred approximately every 6–13 days. The longest Bora events lasted 3–4 days and sediment transport conditions during events lasted, on average, about 1–2 days. Bora wind events occurred more frequently than Scirocco events. At the Po delta site, nearbed current speed, averaged over events, varied from 4 to 15 cm/s (Table 3) although current directions fluctuated within storms resulting in low event-averaged currents (Figs. 5a and b). Net current magnitude and direction also varied from storm to storm and from year to year (Table 3). At the Po delta, Bora events exhibited faster event-averaged currents than the Scirocco events. At the Pescara shelf site, current speed was substantially stronger than at the Po delta. In addition, currents were consistently

directed southward, regardless of wind direction, resulting in higher event-averaged currents. This southerly dominance is probably the result of the strong WACC along the shelf (Orlic et al. 1994; Kourafalou, 1999) with winds tending to enhance or diminish the WACC but only occasionally causing along-shelf flow reversal to the north (Fig. 6a and b). Wave-orbital velocities during events tended to be rather strong, averaging from 9 to 23 cm/s (Table 3). These averages are generally above the threshold of grain motion and the correlation between peaks in orbital velocity and SSC are clearly visible in the time series of Fig. 4-6. The mean wave-orbital velocities during Bora, Scirocco, and Mixed events were generally similar. During winter 2002–2003, mean Urms during Bora events were slightly less at the Po delta and Pescara shelf sites, but during Scirocco events, significantly larger mean and peak Urms were observed on the Po delta than the Pescara shelf (Table 3). SSC and net sediment flux also increased significantly during events. In general, event-averaged

ARTICLE IN PRESS 444

A.M.V. Fain et al. / Continental Shelf Research 27 (2007) 431–451

SSC at 30 cmab on the Po delta varied from 140 to 360 mg/L during events, although concentrations in winter 2002–2003 are underestimated due to instrument gain limitations in that period. At the Pescara site, event-averaged SSC was 680 mg/L during Bora events. Based on comparison of smaller events when instruments were well within their operating range, events at the Po delta were likely equal to or greater in concentration than those on the Pescara shelf. While wave-orbital velocity, current speed and SSC are of similar magnitude during Bora, Scirocco and Mixed events, observations of sediment flux show significant differences. In the along-shelf direction, Bora wind events dominate the sediment flux by an order of magnitude when compared to Scirocco events and by an even greater amount when compared to Mixed events. On the Pescara shelf, southerly sediment flux during the winter of 2002–2003 appears to be approximately eight times greater than the equivalent winter data at the Po location but estimates of SSC are low due to clipping at 0.4 g/l. If the SSC peaks were in fact as large (or larger as the most energetic events of the record occurred in winter 2002–2003) than those in the prior two winters (PoI deployments), then the sediment flux for the two areas would be of the same order. When comparing events that did not exceed the instrument maximum at the PoII site, SSC appears to be of similar magnitude as events at the PoI site for similar wave-orbital velocities (see Figs. 4d and 5d). In all cases, along-shelf flux dominates over the across-shelf flux (Figs. 4e, 5e, and 6e).

5. Discussion Storm events are the major mechanisms that cause sediment to be resuspended and advected on the Po River subaqueous delta and on the Pescara River continental shelf. Sediment transport at both locations is primarily in the along-shelf direction. The observed resuspension events occur near simultaneously at the two sites, but the direction and strength of the associated winds and their effect on the parameters that impact sediment flux in the different regions varies considerably. In addition, the forcing appears to change interannually. The transport of sediment during resuspension events controls the redistribution of sediment from its initial flood deposition site to its ultimate site of long-term accumulation. Thus, the event-driven sediment transport has a major influence on the

formation and maintenance of sedimentary deposits on the Adriatic shelf. 5.1. Net sediment flux on the Po delta and the Pescara shelf during winters The net sediment flux on the western Adriatic shelf over the winter periods shows significant variations both spatially and temporally. Records from the Po and Pescara regions show that the sediment flux is strongly storm-driven along the 12-m isobath, as seen in the stepped structure of the cumulative sediment flux diagram (Figs. 4e, 5e, and 6e). Even though the largest events in winter 2002–2003 at the PoII site are underestimated due to OBS sensor clipping, it can be estimated that the order of magnitude of the sediment flux in events is comparable between the Po subaqueous delta and the Pescara River shelf based on observations during the prior two winters, yet the forcing and directional structure during individual events varies distinctly. This results in significantly different net flux as observed in the overlapping period of study (October 2002–February 2003) (Figs. 5b and 5c). The net flux in the across-shelf direction is minimal at 12-m water depth at both study areas. On the Po River subaqueous delta, there was a net southward along-shelf flux during the first two winters (Fig. 4e) yet individual events were directed both to the north and to the south. During winter 2002–2003, when both the Po delta and the Pescara shelf were concurrently instrumented, data from the Po delta show a northward net flux (Fig. 5b) that is inconsistent with the view that the net flux in this area is towards the south. At the Po delta there was only a partial winter in 2002–2003 record and flux magnitudes are underestimates. Therefore, the record is not necessarily representative of the entire winter. Other observations in the region also show net northward flux in the early part of the winter, but overall southward flux for the entire winter period (Traykovski et al., 2007). In addition, currents in the Po region are highly variable due to small-scale spatial variability and gyres that occur in the vicinity of the Po River subaqueous delta (Cushman-Roisin et al., 2001). Thus, the sediment flux direction may depend on location relative to the morphology of the delta and the distributary mouths. In general, the observed net sediment flux on the Po River subaqueous delta is consistent with seabed studies that show the sediment deposit to be a result

ARTICLE IN PRESS A.M.V. Fain et al. / Continental Shelf Research 27 (2007) 431–451

of rapid initial deposition near the river mouth during high discharge events (Fox et al., 2004; Palinkas et al., 2005; Milligan et al., 2007) a portion of which is subsequently transported in the alongshelf direction. This creates a relatively continuous band of fine-grained sediment along the coastline at water depths between 0 and 30 m, which extends from the Po delta southward for a distance of over 70 km (Cattaneo et al., 2003). Little sediment accumulation is found at water depths greater than 30 m (Palinkas et al., 2005), consistent with the limited offshore sediment flux observed. On the Pescara River continental shelf, the net along-shelf sediment flux shows strong persistent southward transport with very little across-shelf transport during the entire winter deployment (Fig. 6e). At this location, all major sediment flux events were directed toward the south. The net magnitude of along-shelf flux is high relative to the Po subaqueous delta region due to the consistent directionality forced by the WACC in this region. As seen in Puig et al. (2007), two transport pathways appear active: one in shallow water dominated almost completely by along-shelf transport due to the WACC as shown in this study, and one in deeper water (412 m depth) with both along and offshelf transport due to Ekman veering. The delimiting factor between the two zones appears to be the water depth at which the surface and bottom Ekman layers merge (Puig et al., 2007). 5.2. Characteristic storm-driven transport events on the western Adriatic shelf The sediment-flux data (Figs. 4e, 5e and 6e) show that the majority of transport occurs during storm events and are generally forced by Bora and Scirocco winds. In the following, the distinct Bora and Scirocco events are discussed to elucidate the difference between their impacts on the redistribution of the modern mud deposit in both the northern and central Adriatic basins. 5.2.1. Resuspension and transport events on the Po River subaqueous delta The resuspension events at the Po River subaqueous delta show a variety of forcing, strength, and direction. Both the Bora and Scirocco wind events play a significant role in the redistribution of sediment in this region. Although individual events vary significantly, on average both types of wind events had similarly large wave-orbital velocities

445

while the average net flux for event type was quite different. The Bora events were the major contributor to the net sediment flux throughout the 3 years of winter records due to the relatively strong along-shelf currents associated with them. Bora events were more frequently directed towards the south, but individual events transported sediment both north and south along bathymetry, probably controlled by the factors that generate the complex circulation features in the northern Adriatic such as stratification, strength of the thermohaline circulation, and specific wind direction (e.g., Orlic et al., 1994). Scirocco events were less frequent throughout the record, but can be of great magnitude, as was seen in winter 2002–2003 (Figs. 5 and 6), and thus they play a significant, but secondary, role in the along-shelf redistribution of sediment at the Po River subaqueous delta. In general, mean currents were significantly smaller during Scirocco events resulting in less sediment in the current boundary layer, and less net flux (Table 3). The currents during Scirocco events tend to be oriented more offshore and to the north. During the most energetic events (both Bora and Scirocco), event-averaged SSC exceeded 360 mg/L at 30 cmab and peak concentration values during the events reached 1.3 g/L (Fig. 4d and 5d). Although these are extremely high SSC values, they do not reach levels that would suggest fluid mud, or gravity-driven processes. The maximum concentrations observed at the PoII study site were in excess of 3.2 g/L, but this sensor was extremely close to the seabed (the tripod had settled into the seabed approximately 10–15 cm—original sensor elevation was 13 cmab) and at times was buried. At a site 12 km south of the Po delta locations discussed in this paper, Traykovski et al. (2007) reported thin fluid mud flows (5–10 cm thick) during three of the most energetic events seen in the 3-year record. Two of these events were exceptionally large Scirocco events and one was a Bora (15–20 November, 25–28 November, and 5–10 December 2002, respectively) (Fig. 5). The short wave periods in the fetch-limited Adriatic basin may support very thin fluid mud layers that are below the sensor elevations used in this study, or are localized to specific areas of readily available sediment supply to the wave boundary layer. These gravitydriven events may control the across-shelf location of the ultimate deposition site (Friedrichs and Scully, 2007), but do not appear to be significant relative to the net sediment flux in this

ARTICLE IN PRESS 446

A.M.V. Fain et al. / Continental Shelf Research 27 (2007) 431–451

study and would not be relevant to the along-shelf redistribution. 5.2.2. Resuspension and transport events on the Pescara River shelf During the winter of observation on the Pescara River shelf, 13 events were observed, most of which were forced by Bora winds (Table 3). On average, wave-orbital velocities during events were similar or slightly higher than those at the Po River site, and mean currents were significantly higher, resulting in slightly higher bed shear stress. At the Pescara shelf site, Bora events contribute more to the net sediment flux than the Scirocco events, not only due to their number, but also their magnitude. Wave-orbital velocities for the Bora events were higher than for the Scirocco events. It appears that the western coast of the central basin is shadowed from the strength of the Scirocco winds. In contrast, at the Po delta site, winter 2002–2003 was dominated by two strong Scirocco events that were the peak events in the 3-year record (Fig. 5) and waveorbital velocity was higher for the Scirocco events. Both Bora and Scirocco events at the Pescara shelf caused net southward sediment flux, even though the winds associated with Scirocco events may cause instances of advection of sediment to the north (Fig. 6a). The Scirocco wind conditions caused a weakening or stopping of the southward WACC currents, but in general, nearbed current reversal did not occur throughout the events. Consistent with modeling studies (Orlic et al., 1986, 1992), the nearbed currents during Bora events were strongly enhanced over the non-event condition, causing strong currents capable of transporting sediment in the southward along-shelf direction. Thus, the net sediment flux during Scirocco resuspension events is generally southward due to the persistent WACC in the region (but much less than during Bora events) due to smaller wave-orbital velocities and weaker mean currents. During events of both types, the across-shelf flux component was only 10–20% of the along-shelf component. Peak SSC observed on the Pescara shelf reached 3.0 g/L at 13 cmab. When comparing with events that did not exceed the instrument maximum at the PoII site, SSC appears to be of similar magnitude for similar wave-orbital velocities (see Figs. 5d and 6d) at the two sites. During the largest events it is difficult to compare the Po and Pescara absolute flux magnitude due to the sensor limitations of the

OBS on the PoII tripod deployment. As at the Po delta site, no high-concentration fluid-mud events were apparent in the record on the Pescara shelf. Resuspension events on the Pescara River shelf exhibit different characteristics to those at the Po subaqueous delta. Although the events are almost always concurrent (the basin is small enough that storm-driven events are felt in both areas), the forcing and resuspension regimes between the two sites are different for each event. In general, while the overall energy at the Pescara and Po sites is similar, as reflected in the average wave-orbital velocities (Table 2), the mean advective currents are stronger at the Pescara site (Table 2). Yet at the Po delta, the supply per unit length of shoreline of unconsolidated sediment from river discharge is higher than on the Apennine coastline (Palinkas and Nittrouer, 2007). The overall result is that the sediment flux in storm-driven transport events at both ends of the system are of the same order of magnitude, but the forcing and net sediment flux in each event is quite different between the sites (Figs. 5e and 6e). 5.2.3. Implications for redistribution of sediment on the western Adriatic shelf The sediment redistribution on the western Adriatic shelf primarily occurs during winter storm events, and thus in the northern basin the relative timing of Po River floods and the winter storms affects the initial sediment deposition. Due to the large drainage basin for the Po River, floods tend to last weeks while local storms occur over days and thus sediment frequently is discharged into the marine system under conditions of relatively low energy. Distinct flood deposits develop which are modified by subsequent transport events. For example, sediment budget calculations suggest that 25–47% of the sediment load discharged from the Po River in the flood of December 2000 was retained on the delta front (Palinkas et al., 2005). Subsequent to deposition (days to weeks), our observations show that the surface of the recently deposited sediment is resuspended during storm events. Although much of the sediment is seen to move towards the south in the storm events, adding to the narrow band of modern sediment that extends from the Po delta to the Gargano Penninsula (Correggiari et al., 1996), transport events can also move sediment towards the north. The lack of a consistent along-shelf direction of transport is reflected in the near-field deposit mapped by

ARTICLE IN PRESS A.M.V. Fain et al. / Continental Shelf Research 27 (2007) 431–451

Palinkas and Nittrouer (2007) in which the highest accumulation rates appear to extend not only to the south of the distributary mouths but also somewhat to the north. Circulation patterns in the northern basin (persistent CCW gyre north of the delta) may limit the extent that sediment may move toward the north, and the increasingly focused structure of the WACC to the south would accommodate the continuous lateral band leading away from the Po delta towards the south. The energy in the Adriatic is relatively low due to the nature of the confined basin. The flux of suspended sediment is low (on the order of 1–5 kg/cm2/year) (Figs. 4e, 5e and 6e) relative to the quantity of sediment discharged by the Po River (15 million tons/year). This may explain why a significant portion of the Po River sediment discharge can be found in the near field of the Po delta on 100-year time scales (Palinkas and Nittrouer, 2007). Bora winds strongly affect processes in both the northern and central Adriatic basin. The WACC is enhanced by the Bora winds to differing degrees at the two sites. Near the Po River, the WACC is influenced by topography (the delta protrusion into the northern basin) and gyres are generated that impact the net flux direction at specific locations on the delta front. To the south of the Po River delta, the WACC becomes stronger, narrower, and more focused, creating strong persistently southward currents in the winter period. The WACC also contributes to strong sediment fluxes that spread the deposits from the multiple discharge points of the Apennine rivers. On the central Adriatic shelf, the multiple Apennine rivers discharge at distinct points along the coast; yet the modern mud deposit remains laterally continuous (Correggiari et al., 2001; Cattaneo et al., 2003). The strong southward flux observed near the Pescara River mouth rapidly redistributes sediment toward the south that is discharged following the patterns of circulation (Orlic et al., 1992) and the resulting sediment deposition (Correggiari et al., 2001). Although these data show very little offshelf sediment flux at 12-m water depth, sediment must be delivered to the outer region where Ekman transport can carry it offshore to the foreset region of the clinoform as modern sediment accumulation is evident (Palinkas et al., 2005). This may be a result of episodic input or slow diffusion from the inner shelf region to the region where Ekman transport can move it to the foreset region.

447

5.3. Interannual variability of storm-driven transport events Data from the three winters of observation at the Po River subaqueous delta provide a glimpse into the interannual variability of sediment resuspension and flux in this environment. Although the timeseries is not long or complete enough for a thorough analysis of the interannual variability, the range of data provides a picture of how distinctly different each winter storm period can be and how that impacts the resuspension and transport of sediment. For example, in comparison (Tables 2 and 3), winter 2000–2001 had high river discharge and frequent small storms equally distributed between forcing from the Bora and Scirocco directions. Winter 2001–2002 had lower river discharge, and had less frequent storms, most of which were Boras. Winter 2002–2003 had moderate to high river discharge and frequent large and small storms, again mostly Boras. The resulting net sediment flux is controlled by the combination of these factors. To exhibit the variability between the years, a value of wave-orbital velocity (rms) of 10 cm/s is chosen as a typical critical erosion velocity for the sediment in the Po deposit, and the number of hourly averaged observations that exceeded this value were summed for each winter. The threshold exceedance at the Po delta site summed to 13 days in the first winter, 12 days in the second winter and 35 days in the third winter (Table 2). Not only does the duration of energetic conditions in the basin vary interannually, but the river flow also has a widely fluctuating hydrograph. Over the course of the study, the average winter Po River flow varied by over a factor of two, and two large floods were experienced in the first and third winters. The resuspension events observed at 12-m water depth saw little effect of river discharge, but rather were associated with wave-orbital velocity. But the large river discharge events supply a significant amount of sediment to shallow water depths that can be transported past the instrumented sites. The resulting net sediment flux per winter period tended towards the south and slightly offshore in the first two winters, while in the third winter the net flux was towards the north with little across-shelf net transport. This significant variation between the first two winters and the third winter (Figs. 4e and 5e) could generate differing interpretations of the dominant transport mechanisms on the Po delta if only a single year is studied.

ARTICLE IN PRESS 448

A.M.V. Fain et al. / Continental Shelf Research 27 (2007) 431–451

In addition, the mix of Bora and Scirocco events was variable between the three winters. On the Po delta, this may not have a large impact on the alongshelf transport regime, as the transport direction is variable for both types of events. On the Pescara River shelf, the mix of events will have a large impact on the net transport because the Bora events dominate the net sediment flux due to their greater wave-orbital velocity and enhanced WACC. On longer time scales, studies have revealed a declining trend in both strength and frequency of Bora events as shown by statistical analysis on wind data from 1951 to 1996 (Pirazzoli and Tomasin, 2002). On the Po delta, this declining trend would impact the overall depositional environment by retaining more of the Po River sediment in the shallow waters of the delta and these seasonal flood deposits would be less likely to be redistributed and spread in the alongshelf direction before consolidation or burial. On the Pescara River shelf this declining trend would cause a significant decrease in the net transport toward the south. Although the 3-year record is too short to evaluate long-term environmental change, perhaps these can be seen in interpretation of the stratigraphic structure. The highly fluctuating transport regime on the western Adriatic shelf emphasizes the limitations on interpreting a transport regime and connections to longer-term depositional features based on a single winter of data collection. 5.4. Comparison with other shelves Near-bottom sediment fluxes during multiple winters have been studied on the high-energy northern California shelf (Drake and Cacchione, 1985, 1986; Sherwood et al., 1994; Ogston and Sternberg, 1999) and relatively the low-energy shelves of the western Adriatic and the Ebro (Palanques et al., 2002). A comparison between these environments is challenging because processes are evaluated at different locations on the shelves (e.g., 12-m water depth in the western Adriatic, and 60-m water depth on the Eel shelf in northern California), but some similarities and differences can be noted. Studies on the Ebro continental shelf (NW Mediterranean) provide a comparison at similar water depths as the present study, while studies on the Eel River shelf provide a contrasting high-energy environment where the shelf mud deposit is in deeper water. The percentage of time that the erosion threshold is exceeded in a winter

period in combination with the water depth can be used to characterize regions considered high- and low-energy. Studies conducted on the Ebro delta have shown that the erosion threshold is regularly exceeded during winter conditions at inner shelf depths at 12 m (Palanques et al., 2002), but infrequently at depths 60 m (Cacchione et al., 1990; Palanques et al., 2002). In these studies, it was found that the along-shelf transport in shallow water on the inner shelf significantly exceeds the across-shelf transport. This is similar to the situation found at both the Po and Pescara sites, and may be characteristic of lowenergy environments. In all of these regions, shallow water depth, wave mixing and coastal currents create a relatively unstratified water column in the region where sediment is frequently resuspended through the winter storm period. The shallow water depth can prevent Ekman veering as the bottom Ekman layer becomes as thick or thicker than the water depth, and downwelling is limited (Puig et al., 2007) and thus the effects of downwelling conditions on enhancing offshelf sediment flux will be minimal. This leads to a relatively strong along-shelf flux in comparison to the across-shelf flux, and thus a narrow along-shelf oriented distribution of modern mud may be a characteristic of low-energy shelves. In contrast to the low-energy environments, the mid-shelf deposit offshore of the Eel River (60-m depth) is subject to relatively long winter periods and high percentage of time conditions are above erosion threshold (5-year average of 53 days/winter of erosion threshold exceedance; Guerra, 2004). For comparison, in this study the erosion threshold was exceeded on average only 14 days/winter on the Po delta at 12-m depth. The Eel mid-shelf mud deposit is in a region that exhibited strong downwelling effects and where along-shelf and across-shelf fluxes are of equal magnitude (Ogston et al, 2004; Guerra, 2004). The along-shelf flux on the Po modern mud deposit is an order of magnitude greater than the across-shelf flux, and on the Pescara deposit where the coastal current is well formed, is two orders of magnitude greater than the across-shelf flux. It appears not only that fine-grained sediment transport in low-energy systems is compressed into shallower water relative to the high-energy systems, but also that the processes active in the different shelf environments control the orientation of sediment transport. In low-energy systems, the transport is dominantly along-shelf whereas in

ARTICLE IN PRESS A.M.V. Fain et al. / Continental Shelf Research 27 (2007) 431–451

high-energy systems, a strong across-shelf component occurs in addition to the along-shelf transport. 6. Conclusions In the relatively low-energy epicontinental Adriatic Sea, sediment resuspension in the BBL occurs primarily in response to winter storm events, as is the case in higher energy environments. On the western Adriatic shelf near-bed observations from the Po subaqueous delta and the Pescara River shelf tripods show that Bora wind-driven events predominantly control the sediment transport regime, but Scirocco wind events also play an important role. At the Po delta, sediment transport induced by Bora winds is highly variable in terms of magnitude and direction, probably due to the interaction between bathymetry and the wind-driven circulation that generates gyres and mesoscale circulation features in the northern Adriatic. On the Pescara shelf, although the magnitude may vary amongst events, the sediment transport direction is always along isobath towards the south. The magnitudes can be enhanced by advective currents in the WACC. During Scirocco events on the Pescara shelf, the WACC is reduced and at times reversed, but all observed resuspension events had a net southward sediment flux. In low-energy environments, the sediment transport regime is compressed into shallow water and sediment resuspension and transport appears to occur primarily in the along-shelf direction, with little across-margin transport. This dominance of along-shelf transport is thought to be due to the well-mixed nature of the shallow water, allowing little Ekman transport to be induced. For example, on the western Adriatic continental shelf and the Ebro margin, both relatively low-energy environments, the along-shelf winter flux is one to two orders of magnitude greater than the across-shelf flux. On the high-energy Eel River shelf the acrossand along-shelf sediment fluxes are of the same order of magnitude. This difference may partially explain the resulting patterns of modern sediment accumulation on the inner shelf. In the Adriatic Sea, a narrow deposit extending from the Po River 600 km along-shelf occurs (Correggiari et al., 1996), while on the Eel River shelf, a mid-shelf bulls-eye pattern contained within 50 km of the river mouth occurs (Sommerfield and Nittrouer, 1999). There is significant interannual variability in frequency and magnitude of resuspension events in

449

the western Adriatic Sea. In the three winters of observation there was an almost three-fold difference in the number of days the sediment-resuspension threshold was exceeded between the least energetic and most energetic winters. The interannual variability of Po River flow and mix of wind event type may impact longer-term stratigraphic records, as the signatures of rapid deposition and sediment redistribution are preserved or obliterated. Acknowledgments This study has been supported by the Office of Naval Research, Marine Geology and Geophysics Program, Grant No. N00014-02-1-0082, as part of the EuroSTRATAFORM program, coordinated by C. Nittrouer. Special thanks to the many EuroSTRATAFORM scientists, ISMAR-CNR-Italy and specifically to Stefano Miserocchi, Fabio Trincardi, Pere Puig, Cindy Palinkas and the captain and crew from the M/V Sarom VIII, R/V Urania, R/V Garcia del Cid, and R/V Seward Johnson II for making this work possible. Two anonymous reviewers provided valuable input to the earlier version of this manuscript. References Cacchione, D.A., Drake, D.E., Losada, M., Medina, A., 1990. Bottom-boundary layer measurements on the continental shelf off the Ebro River, Spain. Marine Geology 95, 179–192. Cattaneo, A., Correggiari, A., Langone, L., Trincardi, F., 2003. The Late-Holocene Gargano subaqueous delta, Adriatic shelf: Sediment pathways and supply fluctuations. Marine Geology 193, 61–92. Cattaneo, A., Trincardi, F., 1999. Late Quaternary transgressive record in the Adriatic epicontinental sea: basin widening and facies partitioning. SEPM Special Publication 64, 127–146. Correggiari, A., Field, M.E., Trincardi, F., 1996. Late Quaternary transgressive large dunes on the sediment-starved Adriatic shelf. In: De Batist, M., Jacobs, P. (Eds.), Geology of Siliciclastic Shelf Seas. Geological Society Special Publication No. 117, pp. 155–169. Correggiari, A., Trincardi, F., Langone, L., Roveri, M., 2001. Styles of failure in Late Holocene highstand prodelta wedges on the Adriatic Shelf. Journal of Sedimentary Research 71 (2), 218–236. Cushman-Roisin, B., Gacic, M., Poulain, P.-M., Aregiani, A., 2001. Physical Oceanography of the Adriatic Sea. Kluwer Academic Publishers, Netherlands, 304pp. Drake, D.E., Cacchione, D.A., 1985. Seasonal variation in sediment transport on the Russian River shelf, California. Continental Shelf Research 4 (5), 495–514. Drake, D.E., Cacchione, D.A., 1986. Field observations of bed shear stress and sediment resuspension on continental shelves,

ARTICLE IN PRESS 450

A.M.V. Fain et al. / Continental Shelf Research 27 (2007) 431–451

Alaska and California. Continental Shelf Research 6 (3), 415–429. Fox, J.M., Hill, P.S., Milligan, T.G., Boldrin, A., 2004. Flocculation and sedimentation on the Po River Delta. Marine Geology 203, 95–107. Frignani, M., Langone, L., 1991. Accumulation rates and 137Cs distribution in sediments off the Po River delta and the Emilia-Romagna coast (northwestern Adriatic Sea, Italy). Continental Shelf Research 11, 525–542. Friedrichs, C.T., Scully, M.E., 2007. Modeling deposition by wave-supported gravity flows on the Po river Prodelta: from seasonal floods to prograding clinoforms. Continental Shelf Research 27, 322–337. George, D.A., Hill, P.S., Milligan, T.G., 2007. Flocculation, heavy metals (Cu, Pb, Zn) and the sand–mud transition on the Adriatic continental shelf, Italy. Continental Shelf Research 27, 475–488. Guerra, J., 2004. Interannual variability of nearbed sediment flux and associated physical processes on the Eel River shelf, northern California—USA. Ph.D. Thesis, University of Washington, Seattle, 114pp. Hendershott, M.C., Rizzoli, P., 1976. The winter circulation of the Adriatic Sea. Deep-Sea Research 23, 353–370. Kineke, G.C., Sternberg, R.W., Trowbridge, J.H., Geyer, W.R., 1996. Fluid–mud processes on the Amazon continental shelf. Continental Shelf Research 16, 667–696. Komar, P.D., Miller, M.C., 1975. Sediment threshold under oscillatory waves. In: Proceedings of the 14th Conference on Coastal Engineering, pp. 756–775. Kourafalou, V.H., 1999. Process studies on the Po River plume, North Adriatic Sea. Journal of Geophysical Research 104 (C12), 29963–29985. Marchetti, M., 2002. Environmental changes in the central Po Plain due to fluvial modifications and anthropogenic activities. Geomorphology 44, 361–373. Milliman, J.D., 2001. Delivery and fate of fluvial water and sediment to the sea: a marine geologist’s view of European rivers. Scientia Marina 65 (2), 121–132. Milliman, J.D., Syvitski, J.P.M., 1992. Geomorphic/tectonic control of sediment discharge to the ocean: The importance of small mountainous rivers. The Journal of Geology 100, 525–544. Milligan, T.G., Hill, P.S., Law, B.A., 2007. Flocculation and the loss of sediment from the Po River plume. Continental Shelf Research 27, 309–321. Nelson, B.W., 1970. Hydrography, sediment dispersal, and recent historical development of the Po river delta, Italy. In: Morgan, J.P., Shaver, R.H. (Eds.), Deltaic Sedimentation Modern and Ancient, Society of Economic Paleontologists and Mineralogists. Tulsa, Oklahoma, pp. 152–184. Nittrouer, C.A., Miserocchi, S., Trincardi, F., 2004. The PASTA Project: investigation of Po and Apennine Sediment Transport and Accumulation. Oceanography 17 (4), 46–57. Ogston, A.S., Cacchione, D.A., Sternberg, R.W., Kineke, G.C., 2000. Observations of storm and river flood-driven sediment transport on the northern California continental shelf. Continental Shelf Research 20, 2141–2162. Ogston, A.S., Guerra, J.V., Sternberg, R.W., 2004. Interannual variability of nearbed sediment flux on the Eel River shelf, northern California. Continental Shelf Research 24, 117–136. Ogston, A.S., Sternberg, R.W., 1999. Sediment-transport events on the northern California continental shelf. Marine Geology 154, 69–82.

Orlic, M., Kuzmic, M., Vucak, Z., 1986. Wind-curl currents in the Northern Adriatic and formulation of bottom friction. Oceanologica Acta 9 (4), 425–431. Orlic, M., Gacic, M., La Violette, P.E., 1992. The currents and circulation of the Adriatic Sea. Oceanologica Acta 15 (2), 109–124. Orlic, M., Kuzmic, M., Pasaric, Z., 1994. Response of the Adriatic Sea to the bora and sirocco forcing. Continental Shelf Research 14 (1), 91–116. Paklar, G.B., Isakov, V., Koracin, D., Kourafalou, V., Orlic, M., 2001. A case study of bora-driven flow and density changes on the Adriatic Shelf (January 1987). Continental Shelf Research 21, 1751–1783. Palanques, A., Puig, P., Guillen, J., Jimenez, J., Gracia, V., Sanchez-Arcilla, A., Madsen, O., 2002. Near-bottom suspended sediment fluxes in the microtidal low-energy Ebro continental shelf (NW Mediterranean). Continental Shelf Research 22, 285–303. Palinkas, C.M., Nittrouer, C.A., Wheatcroft, R.A., Langone, L., 2005. The use of 7Be to identify event and seasonal sedimentation near the Po River delta, Adriatic Sea. Marine Geology 222–223, 95–112. Palinkas, C.M., Nittrouer, C.A., 2007. Modern sediment accumulation on the Po shelf, Adriatic Sea. Continental Shelf Research 27, 489–505. Passega, R., Rizzini, A., Borghetti, G., 1967. Transport of sediments by waves, Adriatic coastal shelf, Italy. The American Association of Petroleum Geologists Bulletin 51 (7), 1304–1319. Pirazzoli, P.A., Tomasin, A., 2002. Recent evolution of surgerelated events in the northern Adriatic area. Journal of Coastal Research 18 (3), 537–554. Puig, P., Palanques, A., Guillen, J., 2001. Near-bottom suspended sediment variability caused by storms and nearinertial internal waves on the Ebro mid continental shelf (NW Mediterranean). Marine Geology 178, 81–93. Puig, P., Ogston, A.S., Guillen, J., Fain, A.M.V., Palanques, A., 2007. Sediment transport processes from the topset to the foreset of a crenulated clinoform (Adriatic Sea). Continental Shelf Research 27, 452–474. Sherwood, C.R., Butman, B., Cacchione, D.A., Drake, D.F., Gross, T.F., Sternberg, R.W., Wiberg, P.L., Williams III, A.J., 1994. Sediment-transport events on the northern California continental shelf during the 1990–1991 stress experiment. Continental Shelf Research 14 (10/11), 1063–1099. Sherwood, C.R., Carniel, S., Cavaleri, L., Chiggiato, J., Himangshu, D., Doyle, J., Harris, C., Niedoroda, A., Pullen, J., Reed, C., Russo, A., Sclavo, M., Signell, R., Traykovski, P., Warner, J., 2004. Sediment dynamics in the Adriatic Sea investigated with coupled models. Oceanography 17 (4), 46–57. Sommerfield, C.K., Nittrouer, C.A., 1999. Modern accumulation rates and a sediment budget for the Eel Shelf: a flooddominated depositional environment. Marine Geology 154, 227–241. Sternberg, R.W., 1986. Transport and accumulation of riverderived sediment on the Washington continental shelf. Journal of the Geological Society London 143, 943–956. Sturm, B., Kuzmic, M., Orlic, M., 1992. An evaluation and interpretation of CSCS-derived patterns on the Adriatic shelf. Oceaologica Acta 15, 13–23.

ARTICLE IN PRESS A.M.V. Fain et al. / Continental Shelf Research 27 (2007) 431–451 Syvitski, J.P.M., Kettner, A.J., 2007. On the flux of water and sediment into the Northern Adriatic Sea. Continental Shelf Research 27, 296–308. Traykovski, P., Wiberg, P.L., Geyer, W.R., 2007. Observations and modeling of wave-supported sediment gravity flows on the Po prodelta and comparison to prior observations from the Eel Shelf. Continental Shelf Research 27, 375–399. Traykovski, P., Geyer, W.R., Irish, J.D., Lynch, J.F., 2000. The role of wave-induced density-driven fluid mud flows for crossshelf transport on the Eel River continental shelf. Continental Shelf Research 20, 2113–2140.

451

Vittori, E., Berti, D., Lalli, F., Miozzi, M., Miscione, F., Porfidia, B., Serva, L., 2001. Environmental effects of channel-harbor modifications at Pescara River mouth (Adriatic Sea, Italy). Geological Society of America Abstracts with Programs 33 (6), 275. Zavatarelli, M., Pinardi, N., Kourafalou, V.H., Maggiore, A., 2002. Diagnostic and prognostic model studies of the Adriatic Sea general circulation: seasonal variability. Journal of Geophysical Research 107 (C1), 3004. Zore-Armada, M., Gacic, M., 1987. Effects of bora on the circulation in the North Adriatic. Annales Geophysicae 5, 93–102.