Sediment Supply to Coastal Zones

There are three main sources of sediment supply to the coastal zone: (1) the coastal landforms themselves, notably sea cliffs; (2) the land inland from the coast; and (3) material transported from offshore (Summerfield Reference Summerfield1991: 324–25). The volume and nature of the sediments are controlled by geology, tectonics, climate, oceanography, and the topography of the sediment source area (Wells Reference Wells, Stein and Farrand2001: 155). On open coastlines, high wave energy can erode coastal landforms, causing recession and liberating sediment to be transported and reworked. The amount of sediment thus produced depends on wave energy and the degree of consolidation of the cliffs or other landforms suffering erosion.

By far the greatest sediment load, however, is brought to the coast from inland by rivers. On average, mid-latitude coastlines receive more than 90% of their sediment from fluvial sources. Where drainage systems are long and broad (e.g., the Nile, Tigris, Euphrates, Maeander), the material transported to the shore will be predominantly fine grained as coarser clasts are deposited well upstream. By contrast, short, steep drainages carry coarser materials to the coast. In Greece, there are relatively few large, perennial rivers with broad drainage basins (to name two, the Peneios and the Spercheios). Most rivers have a seasonal flow that is heaviest in the winter months as a result of rains that trigger high-energy, swift-flowing torrents capable of transporting much coarse material. As a result, a typical stony beach will consist of a mix of coarser and finer material.

Free sediment is transported in the interface of coast and inshore waters in several ways (Summerfield Reference Summerfield1991: 325; Wells Reference Wells, Stein and Farrand2001: 155–57). Tides and wave action move sediment continuously and more or less perpendicularly into and away from shore. Wind-driven waves can be significant agents of erosion and sediment transport on open coastlines, but because the Mediterranean is a virtually tideless sea, the effect of tides on coastal processes is minimal. Where offshore winds are strong and persistent, substantial amounts of airborne sand-sized and smaller grains can be deposited onshore (aeolian transport). Yet the dominant movement of sediments along Mediterranean coasts is due to oceanic currents (Fig. 5.2). Generally, the fine clay- to silt-sized sediments escape offshore in suspension and are carried away by oceanic currents. The coarser (sand-sized and greater) material ordinarily remains in the littoral zone, to be entrained in longshore currents, which transport material parallel to shore through a process called longshore drift (also known as littoral drift). The most important contributor to this material is alluvial sediment, but sediments from the erosion of coastal landforms as well as material previously entrained in ocean currents can also be present. The rate and volume of longshore transport along a coastline are controlled by current velocity and wave energy, as well as the angle at which waves strike the coast. The impact of longshore drift on coastal configuration depends on the presence or absence of features of coastal topography and inshore bathymetry that act to impede or facilitate sediment movement. Sediment remains in transit until a sediment sink drains it offshore or an impediment causes deposits to form against it and build up over time. Thus, on linear coasts, longshore drift may proceed with little interruption or effect on coastal configuration. Where coasts have irregular configurations, however, as is so often the case in Greece, longshore sediments become trapped against headlands, offshore islands, delta formations, and other prominent features (Fig. 5.2). These deposits contribute to the development of landforms such as barrier islands, spits, and sandbars, behind which lagoons form. The elongated form of these deposits reflects the prevailing direction and strength of the longshore currents. The long and complex history of longshore-derived landforms of coastal Elis shows these interactions quite clearly (Kraft et al. Reference Kraft, Rapp, Gifford and Aschenbrenner2005). Coastal locations downcurrent from significant sediment traps are characteristically starved of sediment, and these features will be absent. Littoral sediments can also be drained offshore by deep-sea canyons.

The overall contribution of sedimentation to coastline configuration can be measured in terms of a sediment budget: if the rate of sediment delivery exceeds the capacity for sediment transport away from the coastal zone, accretion will occur; if not, sediment will remain in transit until it reaches a sediment trap that captures it or conducts it out to sea. It is important to note that even at the local scale, erosion and sedimentation can co-occur. In the region of the lower Acheron River valley of southwestern Epirus, the tectonic structure of the linear Ionian coastline promotes coastal erosion and efficient longshore transport, while within the once-broad embayment at the mouth of the Acheron River, a sharply progradational sequence since the Neolithic period has been documented (Besonen et al. Reference Besonen, Rapp, Jing, Wiseman and Zachos2003).

5.2 Movement of sediments along Mediterranean coasts. After Wells Reference Wells, Stein and Farrand2001: 156, fig. 6.3. Courtesy of University of Utah Press.

Coastal Cliffs and Shore Platforms

Coastal cliffs are steep, often vertical, slopes that rise above the shoreline. Fault-derived cliffs may plunge precipitously into deep water, suffering little erosion since wave energy against them is minimized by reflection of swell waves. In shallower waters, cliffs are susceptible to basal erosion by the action of breaking waves. Particularly on open coasts, wave action undercuts sea cliffs, cutting a notch around mean sea level. Over time, an overhang forms that becomes increasing unstable and eventually collapses into the sea. This material, along with other gravity-entrained colluvium from the slope, is available for reworking and transport in the littoral zone. This erosional process forms recessive shorelines: as the coastal cliffs retreat, horizontal shore platforms are left behind at the basal level where wave cutting occurred. These wave-cut or intertidal platforms are common in Greece where high wave energies combine with easily eroded limestone strata. Once formed, they are subject to weathering and abrasion. The rate of cliff erosion varies over time: in particularly stormy years, wave energy and weather can accelerate coastal erosion. If the process continues, however, eventually the cliff will be protected because wave energy is expended crossing the platform.

The sequence of shore platforms and sea cliffs may leave physical traces on land or underwater, particularly where coasts are uplifted or submerged, respectively. The Corinthia in southern Greece preserves examples of both (Hayward Reference Hayward, Williams and Bookidis2003: 16–17; Pirazzoli et al. Reference Pirazzoli, Stiros, Arnold, Laborel, Laborel-Deguen and Papageorgiou1994; Wells 2001: 173–75, fig. 6.7). The northern Corinthian plain, bordering the Gulf of Corinth, comprises a stair-stepping sequence of risers and treads that represent Pleistocene coastal cliffs and shore platforms, subsequently uplifted and subjected to faulting and erosion. Beginning in the Neolithic period, these uplifted features were attractive locations for repeated human habitation because they afforded expansive views and defensive possibilities (Tartaron et al. Reference Tartaron, Pullen and Noller2006: 496). To the east, the Corinthian port at Kenchreai on the Saronic Gulf is partially submerged as a result of approximately 2.3 meters of tectonic subsidence over the last two millennia. As many as five submerged notches, cut by wave action, dissolution, and bioerosion, have been documented in and around Kenchreai, each representing a paleoshoreline formed during a period of relative sea-level stability (Noller et al. Reference Noller, Wells, Reinhardt and Rothaus1997).

Beaches

A beach is a shore built of unconsolidated sediment (Hamblin and Christiansen Reference Hamblin and Christiansen1998: 398). Beach sediments form at the dynamic interface of land and sea, where the shoreline is constantly exposed to wave action and littoral currents. Most beaches consist of sand- or pebble-sized clasts (thus, sand or pebble beaches) because wave energy is usually sufficient to remove silt and clay from the littoral zone. Waves and currents sort beach sediments both vertically and laterally, and round and polish the clasts. These well-known processes produce a predictable set of sedimentary structures that allow the identification of fossil beach deposits (Wells Reference Wells, Stein and Farrand2001: 158, table 6.1).

Beaches are composed of both organic and inorganic material, but in Greece, carbonate pebble beaches are especially common because of the pervasive carbonate geology (limestone and other biogenic carbonates). Beachrock, intertidal sediment indurated by calcite or aragonite cement, forms readily because of the abundance of carbonate in solution in near-shore waters (Wells Reference Wells, Stein and Farrand2001: 158). Submerged or uplifted beachrock can indicate the approximate lateral positions of former shorelines, particularly in the nearly tideless Mediterranean.

The progradation of sediment seaward is sometimes manifest as sets of sandy, shore-parallel beach ridges and intervening swales. William Tanner (Reference Tanner1995) identifies four types, based on their origin. The most common are swash-built ridges, which occur where there is a constant source of sandy sediments, but longshore currents are insufficient to remove them. The sediment is instead reworked by modest storm waves (swash) into narrow mounds with low relief along the shoreline. The lower-lying surface between the new ridge and the adjacent landward ridge floods seasonally, filling with sandy near-shore sediment (e.g., backswamp or marshy deposits), thus forming the ridge-swale set.³ This is a process of gradual accretion, with each ridge representing an old beach and the ridge system marking the advance of the shoreline seaward over time. Formation can be relatively rapid, however: Tanner (Reference Tanner1995: 159–60) suggests an average interval of around 50 years for sandy swash-built ridges on ocean coasts. A striking example can be found in Greece at the mouth of the Acheron River in Epirus (Besonen et al. Reference Besonen, Rapp, Jing, Wiseman and Zachos2003: 216), where concentric accretionary beach ridges and swales have formed in historical times in an embayment well sheltered from longshore currents (Fig. 5.3).

Beach deposits can also grow outward to join with an offshore island by a narrow neck known as a tombolo. This typically occurs because the offshore island behaves as a natural breakwater, creating a wave shadow along the coast behind it, where sediment begins to build outward from the shore. Subsequently, longshore drift moves sediments onto and around the tombolo (Hamblin and Christiansen Reference Hamblin and Christiansen1998: 399; Fig. 5.4). Tombolos can enhance the natural qualities of an anchorage by creating a double-harbor configuration, and in historical times they were emulated by built causeways. The island of Mochlos off Crete's northern coast is believed to have been connected to the mainland by a tombolo in the Bronze Age (Branigan 1991: 97), and the same may have been true of the southern Cretan harbor at Kommos (Shaw Reference Shaw2006: 55).

5.3 View of the sequence of ridges and swales at the mouth of the Acheron River, Epirus. Courtesy of Nikopolis Project archives.

Estuaries and Tidal Landforms

An estuary is a partially enclosed coastal body of water that communicates with the open sea, but is protected from the action of open ocean waves by its topography or some form of barrier that absorbs and refracts wave energy (e.g., sand bars or barrier islands). Usually the term is applied to inlets into which rivers or streams flow, causing fresh and marine waters to intermingle (Fig. 5.5). These transitional zones between river and ocean environments experience both marine processes (tides, waves, saline water influx) and fluvial processes (freshwater influx, sedimentation), contributing diverse nutrients that make estuaries among the world's most productive natural habitats.

5.4 View of tombolo, Paximadi Cape, Euboea. Courtesy of Tim Bekaert.

5.5 View of an estuary in South Carolina, USA. Courtesy of the National Oceanic and Atmospheric Administration.

Estuaries trap fine-grained sediments, leading to the formation of a variety of landforms such as tidal mudflats, tidal marshes and swamps, and tidal inlets. In the Aegean, where tidal currents are weak, mud and clay are deposited on shallow mudflats in the lower intertidal zone. In the upper (landward) intertidal zone, continued deposition of mud can cause vertical accretion and eventually the formation of a brackish marsh above the normal high tide level. These different landforms exhibit characteristic sedimentary structures and contents (Wells Reference Wells, Stein and Farrand2001: 158–59, 162). Mudflat sediments typically consist of a mix of silt and clay with a high iron content and in situ molluscan fauna. Tidal marshes are rich in peat, comprising very fine-grained clay and silt with a high organic content. In tectonically stable contexts, marshes build up and out, gradually infilling the estuarial basin.

Estuaries and their associated landforms are highly significant to the study of ancient coastlines. Because of their micro- and macrofossil content, as well as datable organic material, they provide key evidence for coastal evolution. Equally important is the fact that estuaries were a far more prominent feature of coastal landscapes in the Bronze Age than they are today, offering quiet environments ideal for anchorage; these would have been desirable harbor locations throughout the Mediterranean (Raban Reference Raban, Laffineur and Basch1991). Great estuaries at major river mouths are obvious targets for Bronze Age harbors, but because of the tendency of estuaries to fill in over time, smaller estuarial harbors might only be revealed by systematic geomorphological prospection.

Barrier Islands and Lagoons

Lagoons are tidal inlets of shallow marine or brackish water that are separated from the sea by a barrier. Mediterranean lagoons are typically sheltered by barrier islands, which are elongated offshore sand ridges extending parallel to the shoreline and separated from it by the lagoon (Summerfield Reference Summerfield1991: 330). Barrier islands range from a few meters in width and a few hundred meters in length to long islands a few kilometers in width and hundreds of kilometers in length. Lagoon waters are replenished by tidal inlets that perforate the barrier at irregular and migrating intervals. In response to changes in sediment supply, relative sea level, and climate, barriers may form, retreat landward, or disappear.

Lagoons fill with coarser sediment predominantly supplied by longshore drift. Because lagoons are shallow, they are susceptible to infilling and are strongly influenced by precipitation and evaporation. When evaporation rates are high, salinity in a lagoon can exceed that of the sea, and stranded lagoons can form salt pans or salt lakes. The Akrotiri Salt Lake on the southern tip of Cyprus was probably part of a lagoonal system in the Bronze Age that was later isolated from the sea (Blue 1997: 35–38). Salt deposits were surely exploited in the Bronze Age for subsistence and trade.

The western coast of the Peloponnese is a prominent example of lagoon and barrier formation. On the coast of Elis, coastal landform change has been dominated for the last 8,000 years by longshore redistribution of sediments from the Alpheios and Peneus rivers into an extensive Holocene barrier accretion and lagoonal system (Kraft et al. Reference Kraft, Rapp, Gifford and Aschenbrenner2005: 4; Fig. 5.6). In these settings, both sedimentation from fluvial inputs and erosion and littoral transport from high wave energy have shaped Elean and Messenian shorelines into long, sandy beach strands (hence, Homer's “sandy Pylos”). Hans-Jeorg Streif (1964, cited in Kraft et al. Reference Kraft, Rapp, Gifford and Aschenbrenner2005: 5) identifies three major alluvial terraces in the Alpheios River system and correlates them with three episodes of coastal barrier accretion at the delta. These are roughly dated to EBA, LBA, and Classical to modern (in several subphases).

Deltas

Deltas form where rivers deposit sediment directly into a standing body of water, such as a lake or the ocean, more rapidly than it can be redistributed by coastal processes. The sediment load deposited by deltas into seas and estuaries creates progradational complexes of river sediments reworked by littoral or estuarine processes. There are two essential components of a river delta: (1) the delta front, comprising the shoreline and the gently sloping, submerged offshore zone; and (2) the delta plain, an extensive lowland area landward of the delta front, made up of active and abandoned distributary channels fanning out over the plain (Summerfield Reference Summerfield1991: 333–36). The terrain between the channels is occupied by floodplains, levees, tidal flats, marshes, and lakes. All of these features can be recovered in a geological core, and organic material is usually available to establish a chronometric framework for the sequence.

5.6 Example of a lagoon and barrier system on coastal Elis. Kraft et al. Reference Kraft, Rapp, Gifford and Aschenbrenner2005: 14, fig. 5. Courtesy of the Trustees of the American School of Classical Studies at Athens.

The structure and associated landforms of a delta depend on the interaction of the sediment-carrying stream with ocean currents and waves. Because in Mediterranean embayments, fetch – the distance wind can travel unimpeded by landforms – is limited and tidal effects are minimal, deltas are subject to limited modification by coastal processes (Summerfield Reference Summerfield1991: 333). Thus, on a continuum of delta types dominated by tides, waves, and rivers, most Mediterranean deltas are fluvially dominated (Summerfield Reference Summerfield1991: 334–35, table 13.4, fig. 13.22). Tectonic and climatic factors also play an important role in delta morphology (Summerfield Reference Summerfield1991: table 13.5). If a delta region is tectonically stable, the delta plain will aggrade as it progrades; if subsiding, it will form overlapping sedimentary lobes as it progrades; and if rising, river distributaries will cut down into and rework previously deposited sediments. Precipitation and temperature control the type and amount of the vegetation cover. Once rooted, plants trap sediments and contribute to the formation of peat.

The evolution of Mediterranean river deltas since the mid-Holocene deceleration of marine transgression has been the work of both natural and human agents. As sea-level rise slowed, river sediments began to aggrade (build vertically) and prograde (build laterally seaward) in sheltered and shallow marine embayments. Wave energy was low, and thus erosion and littoral transport were not prominent mechanisms. Instead, the deltaic material prograded gradually toward the open sea in a landscape of multiple active and abandoned river channels, distributary levees and swamps, brackish to saline lagoons, and isolated ponds or lakes ranging from saline through freshwater (Fig. 5.7). At Troy, it was only as the delta coast approached the Dardanelles that littoral currents and increased wave action deposited sands in nearshore shoals and beaches (Kraft, Rapp et al. Reference Kraft, Kayan, Brückner, Rapp, Wagner, Pernicka and Uerpmann2003: 164), and the same is true of the Maeander River, across the mouth of which a long, arcuate barrier ridge has developed in recent times (Brückner Reference Brückner, Wagner, Pernicka and Uerpmann2003). Kraft and colleagues (Kraft, Kayan et al. Reference Kraft, Kayan, Brückner, Rapp, Wagner, Pernicka and Uerpmann2003: 367, fig. 4) cite the delta of the Spercheios River, emptying into the Gulf of Malia on Greece's eastern coast, as a modern analogue where similar landforms produced by the progradation of the river delta can be observed. There, a complex landscape of meandering levees and backswamps, distributional channels, and pervasive coastal marshes has formed as the delta gradually progrades into the deep, sheltered Gulf of Malia. Because of the absence of high wave energy and littoral drift at the current delta shoreline, no coastal barriers have formed.

Coastal Sand Dunes

Coastal sand dunes can form in a range of settings, including deltas, coastal plains, and embayments. They are aeolian landforms that develop where certain conditions exist: coastlines where waves and currents interact with abundant loose, sand-sized sediments available for transport; persistent onshore or alongshore winds blowing for at least part of the year at 16 kilometers per hour or more; and flat or low-relief terrain immediately inland of the coastline (Bauer and Davidson-Arnott Reference Bauer and Davidson-Arnott2002; Martínez et al. Reference Martínez, Psuty, Lubke, Martínez and Psuty2004; Pye Reference Pye1983). Dune formation occurs when winds blow dry sand particles landward from the beach; the main sources of the sand are exposed offshore sandbars and river-mouth and other backshore sediments. Objects inland from the coast, such as plants, logs, or human constructions, interrupt the wind flow, causing sand to be deposited in drifts around them. These drifts act as barriers to sand movement, and grow over time to landforms ranging from small hillocks to vast dune systems tens of meters in elevation.

Sediment supply is the key limiting factor to dune formation. Fluvial systems are noteworthy for the large amount of sediment that they can contribute to dune formation: the Nile River has supplied sufficient sands to result in the formation of vast belts of dune ridges (El Banna and Frihy Reference El Banna and Frihy2009). In Greece, coastal dunes are common, but because of the prevalence of rocky (as opposed to sandy) shorelines and mountainous coastal topography, long, continuous costal dune systems such as are found on French or Dutch coastlines are rare. Instead, coastal dunes tend to occur in isolated embayments between promontories and cliffs, or as more extensive but discontinuous coastal dune systems, for example in the western Peloponnese or on Euboea's eastern coast (Heslenfeld et al. Reference Heslenfeld, Jungerius, Klijn, Martínez and Psuty2004: 338; Spanou et al. Reference Spanou, Verroios, Dimitrellos, Tiniakou and Georgiadis2006: 235–37, fig. 1). They tend to exist where barrier islands or wave-dominated depositional landforms occur, often as integral elements of barrier and lagoon systems (Kraft et al. Reference Kraft, Rapp, Gifford and Aschenbrenner2005).

5.7 Formation of the Scamander plain, Troy. Kraft, Kayan et al. Reference Kraft, Kayan, Brückner, Rapp, Wagner, Pernicka and Uerpmann2003: 365, fig. 2. Courtesy of Springer Science+Business Media.

Coastal dunes perform several functions beneficial to the stability of coastlines and potentially to humans as well. They shield low-lying coastlines from violent storm winds and waves, and inhibit coastal erosion. They protect against saltwater intrusion into wetlands and dry lands behind them, and they support a diverse flora and fauna (Spanou et al. Reference Spanou, Verroios, Dimitrellos, Tiniakou and Georgiadis2006; Sýkora et al. Reference Sýkora, Babalonas and Papastergiadou2003). Thus, they form an integral part of coastal ecology and the resources to be found there.

The Anthropogenic Contribution

The human role in coastline formation has been conspicuous in two processes: the acceleration of sedimentation to prograding river deltas, and the creation of artificial harbors. It has long been noted that the shift to settlement in sedentary agropastoral communities in the Mediterranean coincided with mid-Holocene delta formation and shoreline progradation in the estuaries that had resulted from the late Pleistocene–early Holocene marine transgression. Human agency in soil loss – caused by stripping vegetation cover through forest clearing, agriculture, and grazing – has been implicated in increased sediment load to streams and the resulting expansion of deltas and related coastal landforms. But it is not easy to demonstrate a primary human role in this process, particularly for periods as remote as the Neolithic and Bronze Ages. There are three related problems: contemporaneity, causality, and degree of impact (Halstead Reference Halstead, Halstead and Frederick2000). The palynological, geoarchaeological, and archaeological data used to assess human impact on sediment load to coastal areas can often be dated only approximately, with the result that the contemporaneity of the impacts seen in these data sets is highly uncertain. Even when the chronological correlation is fairly secure, human causality, as opposed to a variety of climatic and other environmental changes, can be difficult to establish. Finally, the magnitude of the human impact, given the level of population and the specific activities in which communities were engaged, must be evaluated. Can it be demonstrated that Bronze Age population levels in and around coastal drainage systems were sufficiently large and their subsistence practices sufficiently destructive to account for significantly accelerated delta formation and coastline progradation?

In Greece, the most detailed documentation of putative human agency in landscape destabilization resulting in erosion and catastrophic soil loss comes from the southern Argolid. The regional surveys of the Argolid Exploration Project in the 1970s to the 1990s amassed a large body of geological and archaeological data that seemed to indicate human agency in certain episodes of massive Holocene soil erosion (Runnels Reference Runnels1995, 2000; van Andel et al. Reference van Andel, Runnels and Pope1986, 1990; Zangger Reference Zangger1994a). One of these (the Pikrodafni alluvium) was dated broadly by pottery sherds to the end of EH II, and was concentrated in valleys where FN–EH II settlement was extensive (van Andel et al. Reference van Andel, Runnels and Pope1986: 113). The Pikrodafni alluvium is dominated by debris flows: “…chaotic beds of ill-sorted, largely angular boulders, cobbles, and pebbles, surrounded and supported by a matrix of finer material” (van Andel et al. Reference van Andel, Runnels and Pope1986: 111), likely the result of sheet erosion of slopes made vulnerable to soil loss when vegetation cover was devastated by drought, fire, or clearing. Although a change of climate to drier conditions that reduce vegetation may instigate sheet erosion and result in debris flows, the investigators attributed the Pikrodafni alluvium to careless slope clearance and the “…eventual failure of EH agriculture to contain the loss of soil” (van Andel et al. Reference van Andel, Runnels and Pope1986: 117). This conclusion has been questioned on the grounds of chronological and causal ambiguity as outlined above (e.g., Bintliff Reference Bintliff, Bell and Boardman1992; Butzer Reference Butzer2005; Endfield Reference Endfield, Sinclair, Slater and Gowlett1997; Moody Reference Moody, Kardulias and Shutes1997, 2000; Whitelaw Reference Whitelaw, Halstead and Frederick2000). This debate highlights the problems and underscores the need to assess each event on its own merits before wider inferences about regional land–human relationships can be made.

In general, it is unlikely that the effect of human subsistence activities was such that large natural harbors were strongly affected in the Bronze Age, and particularly it is doubtful that humans made a greater contribution to sedimentation than the combination of climatic oscillations and natural sediment transport after the mid-Holocene stabilization of eustatic sea level. This conclusion is supported by much geomorphological research that shows that rapid infilling of major natural embayments and estuaries in the eastern Mediterranean has been a phenomenon of much more recent times (Brückner Reference Brückner, Wagner, Pernicka and Uerpmann2003: 122–25; Kraft, Kayan et al. Reference Kraft, Rapp, Kayan and Luce2003; Raban Reference Raban, Laffineur and Basch1991). On the other hand, small natural inlets and harbors based on barrier and lagoon systems must always have been more susceptible to both anthropogenically and naturally induced sedimentation (Kraft et al. Reference Kraft, Rapp, Gifford and Aschenbrenner2005), migrating and going in and out of practical use with much greater frequency.

The creation of artificial harbors in the Bronze Age Aegean is unlikely, but cannot be ruled out entirely. This kind of harbor is an artificial estuary or lagoon, where breakwaters have been constructed to reduce wave energy, creating a quiet and sheltered environment in which vessels may operate. At the same time, by minimizing the normal forces of waves and littoral currents, artificial harbors promote the net accumulation of sediment. A universal feature of artificial harbors is that they must be maintained by dredging or by some means of flushing sediment from the harbor basin. All of the processes that occur in natural estuaries also occur in artificial harbors, but they may be accelerated due to intensive human presence. Because the maintenance of harbors responds to the ebb and flow of political and economic conditions, the eventual infilling and abandonment of artificial harbors are all but inevitable (Wells Reference Wells, Stein and Farrand2001: 171).

Artificial harbors leave distinctive signatures in the geoarchaeological record, which allow them to be distinguished from natural harbors. Nick Marriner and Christophe Morhange (2007: 175–77) have identified a fairly consistent geomorphological sequence repeated throughout the eastern Mediterranean, which they term the “Ancient Harbour Parasequence” (AHP). The AHP comprises the depositional history of the harbor basin, from the natural pre-harbor state to postabandonment, with the following surfaces (boundaries) and deposits (facies):

1. (1) The Maximum Flooding Surface (MFS) marks the maximum marine transgression circa 6000 BP. It forms the lower boundary of the sediment archive, and laterally the farthest landward position of the coast. The deposit is characterized by coarse sand and pebbles.

2. (2) After 6000 BP, beach sands began to aggrade naturally, overlaying the MFS with little or no human contribution. Net sediment supply increased as the coastline prograded and the basin began to fill.

3. (3) The Harbour Foundation Surface (HFS) marks the incipient human modification of the basin, in the form of built harborworks, to create a sheltered harbor basin. A sharp transition from coarse beach sands to fine-grained silts and clays characterizes the sedimentology of the harbor basin. In most of the Mediterranean, this surface postdates the Bronze Age. Human exploitation of natural low-energy basins in the Bronze Age is rarely measurable on the basis of granulometry, but can sometimes be detected in subtle patterns of molluscan micro- and macrofossil assemblages (Marriner and Morhange Reference Marriner and Morhange2007: 176).

4. (4) The Ancient Harbour Facies (AHF) refers to the stratigraphic sequence of deposits during active use of an artificial harbor. Enhanced harbor engineering through time is evident in increasingly fine deposits (silts and plastic clays) through the Roman period, and remnants of harbor architecture (moles, breakwaters, quays, etc.), artifacts, and other anthropogenic debris are often present. The AHF may generate a diagnostic assemblage of macro- and microfauna, as well as a strong geochemical signature from human pollutants.

5. (5) The Harbour Abandonment Surface (HAS) records the initial (semi-)abandonment of the harbor basin, often after the Late Roman period. It corresponds to the deterioration or abandonment of maintenance of harbor infrastructure, and is marked by a transition from fine-grained harbor silts and clays to coarse sands and gravels.

6. (6) The Harbour Abandonment Facies (HAF) registers a return to “natural” conditions after the harborworks have deteriorated to the point that the basin is exposed to higher-energy wave action and the formation of coarse-grained sand and gravel beaches.

The Ancient Harbour Parasequence framework has been applied by this investigative group to ancient harbors including Beirut (Marriner et al. Reference Marriner, Morhange and Saghieh-Beydoun2008), Sidon and Tyre (Marriner and Morhange Reference Marriner and Morhange2006), and Marseille (Morhange et al. Reference Morhange, Blanc, Bourcier, Carbonel, Prone, Schmitt-Mercury, Vivent and Hesnard2003), and they have also fitted the geostratigraphy of other harbors, such as Caesarea Maritima (Reinhardt and Raban Reference Reinhardt and Raban1999), into this scheme (Marriner and Morhange Reference Marriner and Morhange2007: 172–74, fig. 29).

Field Techniques

Geomorphological survey involves the initial mapping of coastal landforms, typically starting from maps (geological, topographic, soils), images (satellite, aerial photographs), and archaeological information (chronological, distributional) on sites and features in the coastal zone. These documents may reveal a wealth of information, and form a baseline for the research design. High-resolution satellite images and low-altitude aerial photographs often yield visual evidence for coastal landforms, such as barrier and lagoon systems, or evidence for ancient harbors, such as uplifted harbor basins or submerged harbor installations (Fig. 5.10).

Subsequent ground truthing, both on land and under water, allows natural and anthropogenic features to be studied firsthand in order to form provisional hypotheses about the morphology, genesis and developmental sequences, and chronology of coastal landforms. In many cases, landforms may be visible, for example, dune ridges, infilled lagoons, estuaries, river and stream features, beach ridges, and fossil sea cliffs, but they must be constrained chronologically and present morphologies should not be extrapolated to the past without detailed analysis. Indications of sea-level change may be apparent in submerged buildings, stranded harbor basins, wave-cut notches on fossil sea cliffs, erosion benches (platforms), and other terrestrial and underwater features. Road cuts, irrigation ditches, foundation trenches, and eroded coastal cliffs are common targets of opportunity providing windows onto past depositional and erosional sequences. There are two essential principles that must guide observations about sea-level change, however. First, the critical consideration is relative sea level; that is, the relationship between land and sea at any location is affected by erosional, depositional, and tectonic processes, and not merely by the absolute changes suggested by global or regional eustatic sea-level curves. Consequently, the second principle is that each coastal location has its own relative sea-level curve, determined by local tectonic controls as well as distinct erosional and depositional characteristics. Rapp and Kraft (1994: 73) emphasize that composite regional sea-level curves, or those borrowed from “nearby” localities, usually lead to errors in interpretation.

5.10 Aerial photograph of submerged harbor remains near Naples. Reprinted from Marriner and Morhange Reference Marriner and Morhange2007: 151, fig. 13, with permission of Elsevier.

Observations of visible coastal landforms and archaeological features often form the basis for speculation regarding the layout and date of ancient harbors. This tradition is particularly strong with respect to purported Minoan harbors on Crete (Chryssoulaki 2005; Hadjidaki 2004; Marinatos Reference Marinatos1926; Raban Reference Raban, Laffineur and Basch1991; Shaw Reference Shaw, Hardy, Doumas, Sakellarakis and Warren1990). Typically, the evidence might include artifact concentrations, foundations of Minoan-style walls extending into the sea, anchors found in the nearby sea bed, or cuttings in bedrock that cannot be dated. The arguments for the existence of these harbors often cite the prominence of Minoan seafaring in the Bronze Age (Chryssoulaki 2005; Hadjidaki 2004), an important consideration but not a directly relevant form of evidence. Thus, although reasonable hypotheses may be derived from archaeological observations and general historical arguments, little can be said with real conviction, and little detailed information can be obtained, in the absence of a full geoarchaeological assessment. Marriner and Morhange (2007: 162) cite the cautionary tale of the harbor at Kition-Bamboula on Cyprus, which two researchers depicted as a cothon based on modern engravings and field observations, before a geoarchaeological investigation revealed this to be erroneous (Morhange et al. Reference Morhange, Goiran, Bourcier, Carbonel, Le Campion, Rouchy and Yon2000).

Geophysical survey employs nondestructive techniques to detect remotely subsurface features on land and under water. In terrestrial archaeological contexts, extensive use has been made of ground-penetrating radar, magnetometry, electromagnetics, and resistivity (Kvamme Reference Kvamme2003; Sarris and Jones Reference Sarris and Jones2000). Each instrument and technique has its optimal use conditions and detection characteristics. Because harbors and other coastal features are often buried in sediment, geophysical surveys can frequently reveal the outlines of buried features, provided that their properties (magnetism, conductivity, density, etc.) contrast sufficiently with the surrounding matrix, and that they are located vertically within the detection limits of the geophysical instrument. An excellent example is the buried harbor and urban area of Portus, where Simon Keay and colleagues integrated the results of magnetometry, resistance survey, ground-penetrating radar, and electrical resistivity tomography to reveal much of the plan of Rome's principal imperial port (Keay et al. Reference Keay, Earl, Hay, Kay, Ogden and Strutt2009). These results were confirmed and expanded with programs of coring and excavation.

Comparable techniques can be used to detect features on the sea floor and buried in marine sediments beneath it, typically by towing geophysical equipment on a rig attached to a small boat. Side-scan sonar and sub-bottom profiler surveys are used to create bathymetric maps that measure depths of sediment and bedrock, giving indications of the shape and orientation of ancient harbor basins over time (Lafferty et al. Reference Lafferty, Quinn and Breen2006; Papatheodorou et al. Reference Papatheodorou, Hasiotis, Geraga, Lyberis, Ferentinos, Bassiakos, Aloupi and Facorellis2001). Marine magnetic surveys are well established and have often yielded spectacular results where underwater features with strong magnetic signals are present. A magnetic survey off the coast of Caesarea Maritima in Israel – Herod's Roman harbor – has clarified a number of questions concerning its construction and use. The magnetic data revealed the hydraulic concrete foundations for the ruined harbor moles by detecting the volcanic ash (pozzolana) in the concrete as a strong magnetic anomaly (Boyce et al. Reference Boyce, Reinhardt, Raban and Pozza2004; Fig. 5.11). These results allowed new interpretations of the shape and construction of the now-submerged outer harbor, defined by the moles and additional breakwater structures. Detected in the same survey were numerous low-relief mounds with elevated magnetic signals, on the sea bed beyond the outer harbor. These features, subsequently investigated by jet probing and excavation, turned out to be ships’ ballast piles, identified by a mixture of igneous and metamorphic boulders and fired pottery exhibiting strong magnetic properties (Boyce et al. Reference Boyce, Reinhardt and Goodman2009). It was even possible, through careful mapping and examination, to hypothesize the formation process of the ballast piles by ships anchored around a designated anchorage point (Boyce et al. Reference Boyce, Reinhardt and Goodman2009: 1524). Ballast piles associated with Bronze Age pottery have recently been discovered off the Saronic Gulf coast at Kalamianos (Tartaron et al. 2011; see Chapter 7).

5.11 Plan of submerged remains at Caesarea Maritima. From Boyce et al. Reference Boyce, Reinhardt, Raban and Pozza2004: 132, fig. 8. Reproduced with permission of Blackwell Publishing Ltd.

Geophysical techniques are nondestructive and non-intrusive, yet they can often supply rapid and reliable information about the location, depth, and nature of buried or submerged features without excavation. They may play the key role of guiding subsequent coring and excavation strategies, and more generally helping to form hypotheses and research questions. A geophysical project will more than pay for itself if it saves the investigator from a misguided research design. There are also disadvantages to geophysical surveys. There is no chronological control on the stratigraphy and archaeological features they detect, and the interpretation of anomalies often involves a high degree of subjectivity if their forms and signals are not transparently characteristic of known features. Ground truthing, in the form of excavation, coring, or other means of direct examination, is usually required to verify the identity of potential archaeological features. The entire enterprise of archaeological geophysics relies on a high level of expertise and expensive equipment that may not be easily available or affordable. Nevertheless, geophysical surveys are indispensable for the investigation of buried harbors and submarine features.

Coastal stratigraphy is the central concept of the geoarchaeological method, which entails direct observation of the geological record to locate, characterize, and reconstruct physical evidence of natural and anthropogenic processes that contribute to the long-term evolution of coastal environments. As we have seen, observations of the modern surface can be misleading or ambiguous in their relationship to the past, but all major processes that generate change (cited above) leave a record in local sediments that can be decoded by careful analysis. Ideally, one would excavate a number of sections of the vertical and horizontal dimensions of coastal and underwater sediments, but this approach is rarely feasible in terms of resources (including time and money) and permissions. Thus, geological core drilling⁷ is the method of choice to obtain sediment samples for analysis. The decision about where to sample and how many cores to take is a critical aspect of the research design. Although it is obvious that enhancing the resolution of the study, that is, by increasing the number of cores and decreasing the spatial interval at which they are taken, should lead to higher confidence in the results, given the usual constraints a well-rationalized sampling design is of the utmost importance. Moreover, no coring program, no matter how large or well designed, will provide all the answers hoped for, but will generate new questions that can only be addressed with further research. A perfect illustration is the decades of coring work dedicated to reconstructing the marine embayment and coastal paleogeographies of ancient Troy (Kraft, Kayan et al. Reference Kraft, Kayan, Brückner, Rapp, Wagner, Pernicka and Uerpmann2003; Kayan et al. Reference Kraft, Kayan, Brückner, Rapp, Wagner, Pernicka and Uerpmann2003). Hundreds of cores have now been taken and analyzed from the plain around Hisarlik, yet although the general outlines of the evolution of this coastal landscape are reasonably well understood, the details of interpretation continue to be debated and new questions remain to be addressed.

Many macroscopic attributes of the core sample are described in the field, using one of a number of recording systems (e.g., Folk Reference Folk1980). In the lower Acheron River valley of southern Epirus, Mark Besonen and colleagues recorded lithology, grain-size distribution, color when wet using the Munsell soil color chart, sediment consistency, plant and animal macrofossils, pedogenic characteristics (structure, sesquisoxide/reduction mottling, and calcium carbonate filaments or nodules), and chance finds such as pottery fragments (Besonen et al. Reference Besonen, Rapp, Jing, Wiseman and Zachos2003: 209). The depth of the core varies with the equipment used and the nature of subsurface layers encountered. Hand-augering is the conventional technique, because the instrument is portable and widely owned by earth science departments and private geological firms. With hand-operated equipment, it is not uncommon to reach impenetrable stony layers well before the maximum length of the auger. More elaborate power corers generally do not have this limitation (Kayan et al. Reference Kayan, Öner, Uncu, Hocaoǧlu, Vardar, Wagner, Pernicka and Uerpmann2003: 382–84), but their use is far more expensive and they may not be readily available in country. Once the attributes of the core are described, all or part of each sediment column is packed in aluminum foil or some other protective material for transport to the laboratory.

Laboratory Techniques

The core that reaches the laboratory preserves a stratified, and thus diachronic, record of the sedimentary sequence at a specific location. It is a storehouse of information, but decisions must be made about the tests that will be performed and the materials that will be tested, as well as the resolution at which the examinations are to be made. The tighter the sampling interval, the higher the resolution of the data and the better the reconstruction of coastal stratigraphy and coastal history (including chronology); Marriner and Morhange (2007: 165) recommend five centimeters or less. Time, money, and expertise for analysis are costs that should be resolved before coring begins.

After initial description of sedimentary structures, the cores are separated into subsamples to be used for analyses of different proxy materials. These analyses broadly involve sedimentology, biostratigraphy, and geochemistry.

Sedimentological analysis allows the investigator to characterize the sedimentary structures of a core sample and to identify the environmental facies that are represented in the stratified deposit.

The structural characteristics of the sediment provide important clues to the depositional environment and the source of the material that formed the deposit. Typically, subsamples are analyzed for color, grain size, microfossils, organic matter, and calcium carbonate content; organic material is extracted for radiocarbon (usually AMS) dating and sherds or other cultural material are noted (Jing and Rapp Reference Jing, Rapp, Wiseman and Zachos2003: 159, fn. 4).

Granulometry – the analysis of grain size, shape, sphericity, roundness, and sorting of clastic particles in a sediment – can indicate the mechanisms of transport and deposition and distinguish between high- and low-energy formational environments in a coastal area. Sediment texture refers in part to the range of particle sizes present; these are determined by a process of wet and dry sieving to separate the gravel-, sand-, silt-, and clay-sized fractions (Rapp and Hill Reference Rapp and Hill1998: 22–23, figs. 2.2, 2.3). These fractions are then weighed and their relative proportions are plotted graphically and statistically. The resulting ratios can indicate different kinds of coastal environments; for example, predominance of gravel and sand may indicate a fossil beach; predominant sand may indicate littoral deposits such as sandbars, barrier reefs, and islands; and predominant silts and clays may indicate an artificial harbor basin. The association between grain size distribution and paleolandform is not always straightforward, however. Rapp and Hill (1998: 38–39) identify five factors influencing grain size distribution: (1) the type of source rock and the original size of grains; (2) the type of transporting medium; (3) abrasion and solution during transportation; (4) sorting of size fractions before deposition; and (5) the depositional environment. Thus, the sources of sediment (e.g., fluvial, aeolian, longshore), the distance of transport, and the reworking of sediments in the coastal environment are some of the variables in working back to coastal paleoenvironments. Artifacts that may be present as clasts in the sand- and gravel-sized fractions are of great interest when recovered from cores.

Cores are also sampled for biogenic material, which can provide conclusive evidence for the depositional environments in which particular species lived. Biostratigraphy entails the identification of fossil organisms and the study of their temporal and spatial distribution in order to record the changing abundance and species composition, and thereby to reconstruct the depositional environments (biofacies) that these organisms characterize. The main fossil types are marine mollusks, ostracods, and foraminifera, each being both abundant in coastal waters and specific in its environmental tolerances to depth, salinity, and temperature.⁸ For this reason, they are excellent indicators of depositional environment and accordingly are used widely in paleoenvironmental reconstruction of coastal regions.

There are different advantages and disadvantages to these fossil types. Marine mollusks tend to be abundant, with known environmental tolerances by species. In situ shells are useful for radiocarbon dating. Mollusk shells are often the primary component of coastal middens that may be detected during archaeological or geomorphological survey. Ostracods are microcrustaceans that leave a calcite bivalve carapace, the morphological characteristics of which can provide taxonomic and phylogenetic information (Marriner and Morhange Reference Marriner and Morhange2007: 170). They tend to be ubiquitous in both fresh and marine waters, their carapaces preserve well, and because of their small size a great number can often be extracted from a core subsample. They are excellent indicators of paleoenvironment because their composition, population density, and population diversity vary as a function of water temperature, water salinity, water depth, and anthropogenic impacts (Marriner and Morhange Reference Marriner and Morhange2007: 170). As a result, ostracod species are strongly correlated with very specific depositional environments.

Foraminifera are single-celled organisms with tests divided into chambers that accumulate during growth. They are among the most ubiquitous shelled organisms, often yielding more than one million living specimens per cubic meter, and they live in all marine habitats from intertidal to deepest ocean, from tropics to poles. Their mineralized tests preserve well, so a small sediment sample is likely to provide a large and statistically valid assemblage. They do not, however, live in freshwater environments, and only a few species live in brackish contexts. Further, individual species are highly specific to environment and are quick to respond to environmental change. Thus, foraminifera are highly useful for determining a range of paleoenvironmental characteristics, including sea-level change, paleoclimate, temperature, salinity, carbonate chemistry, diet, and nutrient conditions (Marriner and Morhange Reference Marriner and Morhange2007: 170).

In the laboratory, fossil organisms are extracted and assigned to general ecological assemblages or biofacies, such as freshwater, brackish lagoon, marine lagoon, coastal, and marine. These categories are determined, as explained above, on the basis of the environmental tolerances of the species present, but also by geochemical analysis to determine the stable isotope ratios of oxygen and carbon in the shells of ostracods and foraminifera. The ratio of the two stable isotopes of oxygen, ¹⁶O and ¹⁸O, in ocean waters is a function of global evaporation rates, which in turn are proxy measures for long-term climate fluctuations as well as water temperature and salinity (for principles, see Rapp and Hill Reference Rapp and Hill1998: 104–105). This ratio is also recorded in the shells of fossil marine and freshwater organisms because the shell is formed by the precipitation of carbonates from the surrounding water. The ¹⁸O/¹⁶O ratio, expressed as variance from present mean sea water (δ¹⁸O in parts per thousand [‰]), is determined mainly by the temperature and salinity of water; negative values record progressively depleted ¹⁸O and consequently warmer and less saline water, while positive values reflect enriched ¹⁸O and colder, more saline water (Deith Reference Deith, Bailey and Parkington1988). Similarly, the ratio of the two stable isotopes of carbon, ¹³C and ¹²C, imparts information about temperature and salinity as well as the presence of carbon from decomposed plants. When stable isotope ratios are combined with the species’ ecological preferences and tolerances, the depositional environment can often be determined conclusively (e.g., Goodman et al. Reference Goodman, Reinhardt, Dey, Boyce, Schwarcz, Sahoğlu, Erkanal and Artzy2009).

The ultimate aim of geoarchaeological analysis of paleocoastal environments is to identify the changing coastal environments of deposition, or facies, over time. Facies are defined as the characteristics of a rock unit (in this case, a sedimentary deposit) that reflect the condition of its origins and differentiate it from adjacent units. Each sedimentary facies identified in the core has an essentially uniform character that reflects its origin and distinguishes it from adjacent units laterally and vertically (Rapp and Kraft Reference Rapp, Kraft and Kardulias1994: 72). Biofacies derived from microfaunal analysis are correlated with sedimentary facies resulting from the granulometric analysis (Fig. 5.12). The chronometric framework for the sedimentary sequence is usually provided by radiocarbon determinations on organic material present in the core, including shells, charcoal, wood fragments, peat, charred roots, and other plant remains. Datable sherds can also provide chronological information. When a number of sediment cores are taken across a coastal landscape, the sedimentary units can be correlated to construct a broad record of changing coastal landforms over time.

The work of Besonen and colleagues in the lower Acheron River valley provides an example of facies designations (Besonen et al. Reference Besonen, Rapp, Jing, Wiseman and Zachos2003). On the basis of geoarchaeological fieldwork using the techniques described above and others (Besonen et al. Reference Besonen, Rapp, Jing, Wiseman and Zachos2003: 209), they identified 14 distinct sedimentary facies on the modern landscape, which correlate to paleoenvironments that existed at one time or another in the Holocene (Table 5.1). They divide the facies into two broad depositional systems: the fluvial depositional system, consisting of the sedimentary environments landward of the shoreline, and the deltaic nearshore system, comprising those seaward of the shoreline but within the marine embayment (Besonen et al. Reference Besonen, Rapp, Jing, Wiseman and Zachos2003: 212–16). These data, combined with an array of radiocarbon dates from core material, allowed for the reconstruction of coastline configuration and coastal environments over time (Besonen et al. Reference Besonen, Rapp, Jing, Wiseman and Zachos2003: figs. 6.13–6.15). Because the terminology used to name these facies is not standard across the Mediterranean, it can take some effort to work out the correlations from one site to another.

5.12 Correlation of biofacies and sedimentary facies in the Ancient Harbour Parasequence. Reprinted from Marriner and Morhange Reference Marriner and Morhange2007: 177, fig. 31, with permission of Elsevier.

Table 5.1. Sedimentary facies in the lower Acheron valley (Besonen et al. Reference Besonen, Rapp, Jing, Wiseman and Zachos2003: 213–16)