Ancient Port Structures

The main elements of a port are its breakwater(s) to reduce wave action inside a protected basin, where quays or jetties, with some mooring devices, are available for loading/unloading ships. Hence, a breakwater and a quay have to be built using available construction materials and methods, and a basin has to be dredged and maintained at adequate depth.

For a general overview of ancient and modern port structures, please refer to “Ancient Port Structures, An engineer’s perspective“.

The oldest known port structure (in 2019) is the wadi al-Jarf breakwater in the Gulf of Suez (ca. 2570 BC, Khufu-Chéops). This structure is ca. 325 m long and ca. 6 m wide. It is made of cobbles and clay[1]. The port of Byblos (Lebanon) is from the same period, but it is located inside a natural cove with no known port structures[2]. Between 2400 and 2000 BC, a 4 m deep dock of 215 x 35 m was built with fired mudbrick at Lothal (India) at the outlet of River Sabarmati.

Anchorages more or less sheltered by offshore ridges were used as natural shelters on the Levantine coast in the 2nd millennium BC: Arwad, Sidon, Sarepta, Tyre. In Yavne-Yam (Israel) a 100 m x 50 m stone rampart may have been built to improve the shelter[3].

A series of Minoan ports were found on the north coast of Crete: Kydonia (Chania), Knossos and Amnissos (near Iraklio), Mallia, Ag. Nikolaos, Istron, Pachia Ammos, Tholos, Pseira, Mochlos, which are usually quite small.

The very large port on Pharos island might also date around 2000 BC and its more than 2 km long main breakwater might be seen as an ancestor of the typical Phoenician breakwater structure with two ashlar vertical walls with interspace filled with rubble[12].

Early Phoenicians gradually improved their natural shelters by adding breakwater structures on top of the offshore ridges, like at Sidon on the “Languette rocheuse” mentioned by Poidebard and Lauffray in 1951, and at other places (Arwad, Batroun, Zire)[7]. Corings show that Sidon’s inner port was already existing in the 17-15th c. BC thanks to this artificially improved reef[14].

At Kommos (Crete) a shipshed located at some distance from the coast, and including 6 galleries of 37 x 5.60 m, is dated Late Minoan (ca. 1400 BC)[4]. A possible Minoan slipway with 2 galleries of ca. 5 x 40 m is located at Nirou Khani (Crete). Mycenaean ports on the Peloponnesus also date from this period: Epidauros, Egina, Asini, Tiryns, Gytheion, Pylos.

Next are the following port structures, all located in ancient Phoenicia:

    • Dor (Israel, ca. 1000 BC) with a 35 m shallow water quay made of large ca. 0.7 x 0.5 x 2 m ashlar headers facing the sea[5],
    • Tabbat el-Hammam breakwater (Syria, ca. 900 BC) 200 x 15 m[8],
    • Sidon (Lebanon, ca. 800-600 BC) north breakwater 230 m long with headers up to 5 m[2],
    • Tyre (Lebanon, ca. 800-600 BC) north breakwater 70 x 12 m with 0.5 x 0.4 x 2 m headers[9],
    • Athlit breakwater (Israel, ca. 700 BC) 130 x 10 m, with 0.6 x 0.45 x 2 m headers[6].

These breakwaters were all made with ashlar headers ca. 0.5-1 x 0.5-1 x 1-5 m. These pioneering breakwaters consist of two ashlar vertical walls with interspace filled with rubble. Moreover, this type of structure was still built much later in the 3rd c. BC (e.g. Amathous in Cyprus[10] 380 m with 3 m headers) and even in the 4th c. AD (Seleucia Pieria, 120 m with 5 m headers[15]). They re-emerged in the 18th c. when international sea-borne trade asked for them again[13].

The first rubble mound breakwater was possibly built on Delos island in the 8th c. BC[16], but the Samos breakwater (ca. 530 BC) described by Herodotos (Hist, 3, 44-60) is more famous. This type of structure was widely used for breakwaters in water depths deeper than a few meters where positioning of ashlar headers by divers was difficult but dumping loose rock over board barges was easy. This construction method was described later on by Pliny the Younger at Centumcellae (103 AD). The largest ancient breakwater of this type is at Portus Claudius (ca 60 AD). This construction method is still used very often nowadays.

A major evolution was the introduction of ‘Puteolanus pulvis’ (often translated by ‘pozzolana’) for hardening concrete under water. This enabled large blocks of hundreds of cubic meters of concrete to be constructed under water by pouring concrete into prefabricated wooden caissons. The first known use is at Agrippa’s naval base of Portus Iulius, near Pozzuoli, in 37 BC, and the most famous is at Caesarea Maritima (Israel) built between 21 and 10 BC[11]. Also this construction method is still used by modern engineers.

Some of these breakwaters have been luckily preserved and survived two millennia of wave attack, but most of the ancient breakwaters were destroyed by wave action and remains are found under water as “submerged breakwaters”. Careful examination of historical Google Earth images enables us to see quite a few breakwater remains in shallow waters.
As the process of destruction of breakwaters by waves was not all that clear, further analysis was undertaken by the author, focussed on the worst possible wave conditions, considering that they will eventually occur in the long term. In other cases, an approach based on a “design wave” must be used.

Vitruvius‘s “de Architectura” dated around 20 BC, is the major ancient text left about maritime construction methods. Unfortunately, no drawings  are available, so that his descriptions are not all that clear to us. The three of his methods are considered here in some detail with help of various sketches prepared by previous architects and engineers.
See: OLESON, J. et al. “Building for Eternity”, 2014 .
See also: .
and our list of ancient Greek terms on maritime structures.

A question might be asked why the ancient engineers did not invent reinforced concrete, e.g. by means of chains placed inside the mortar. As steel is subject to corrosion and therefore to increase of its volume, that induces cracking of the concrete, the ancients may not have found it such a good idea (NB: the oldest modern reinforced concrete structures are around one century old and are not in a good condition today, e.g. Tour Perret in Grenoble, France). Another part of the answer might be that as the ancients had vaults, they did not use overhanging structures that require reinforced concrete. However, massive structures like walls and towers needed to be reinforced at their base in order to provide internal cohesion. It appears that courses of bonding tiles were used for this purpose. It can be shown from available testing results that the initial shear strength of lime mortar on tiles and bricks is somewhat larger than on natural stones. Hence, each course of tiles placed inside the stone masonry acts like a modern tie beam made of reinforced concrete.

Pilae are massive piles (opus pilarum), which are made of stone or concrete (opus caementicium) which have been used as a base for arched structures like aqueducts and bridge piers. Many of them can still be seen on Google Earth pictures and some, like the one at Nisida, have been studied in detail. It is proposed here that several alignments of maritime pilae may have been the base of arched breakwaters.

Pierced stones can be used as mooring devices when the hole has a horizontal axis. Holes with a vertical axis are believed to be used for derricks like those used onboard ships.

Defensive chains strechting across a harbour entrance are mentioned by several ancient authors, including Vitruvius who explains that chains are suspended by means of machinery placed inside towers located on each side of the harbour entrance. Consideing the forces involved, the length and the weight of the chain was obviously limited.

Silting up of harbours was always a major concern and that is still the case for modern port engineers. One should remember that waves are the driving force of the so called “littoral drift” (longshore sand transport along the coast). As the aim of breakwaters is to reduce wave penetration into the port, sand will settle down. Hence structures including arches are not efficient to stop waves while letting sand passing through. That simply does not work!

Around 65% of the natural tombolos found on the coasts of the Mediterranean Sea are ancient settlements. A tombolo is a sandy isthmus connecting the mainland at a right angle to an offshore island or obstacle. A bell-shaped salient occurs when the offshore island is smaller, or further away from the initial coastline, than for a tombolo. If we define L as the length of the island and D as the distance from the initial shoreline, a tombolo occurs if the ratio L/D > 0.65.


[1] TALLET, P., 2015: Khufu-Khéops is therefore a precursor, not only for his Great Pyramid, but also for his maritime works.

[2] CARAYON, N., 2012, “Les ports phéniciens du Liban – Milieux naturels, organisation spatiale et infrastructures”, Archaeology and History in Lebanon, 36-37 (2012-2013), (p 1-137).

[3] GALILI, E., et al, 1993, “Underwater surveys and rescue excavations along the Israeli coast”, IJNA, 1993, 22.1, (p 61-77).

[4] BLACKMAN, D. & RANKOV, B., 2013, “Shipsheds of the Ancient Mediterranean”, Cambridge University Press, p 10.

[5] ARKIN SHALEV, E., 2019, “The Iron Age Maritime Interface at the South Bay of Tel Dor: results from the 2016 and 2017 excavation seasons”, International Journal of Nautical Archaeology, 48.2, (p 439-452).
Headers are long blocks placed with the smallest section towards the outer side of the wall. Stretchers are placed with their large side to the outer side.

[6] HAGGI, A., 2005, “Underwater excavation at the Phoenician harbor at Athlit, 2002 season”, R.I.M.S. News, report N° 31, Haifa, 2005.

[7] VIRET, J., 2005, “Les « murs de mer » de la côte levantine”, Méditerranée, N°104, (p 15-24). This paper is very informative, even if we do not completely agree with its conclusion.

[8] BRAIDWOOD, R., 1940, “Report on two sondages on the coast of Syria, south of Tartous”, in: Syria. Tome 21 fascicule 2, 1940. pp. 183-226.

[9] NOUREDDINE, I., 2010, “New Light on the Phoenician Harbor at Tyre”, Near Eastern Archaeology 73:2–3 (2010). See also his 2018 publication: Archaeological Survey of the Phoenician Harbour at Tyre, Lebanon, BAAL 18, 2018, (pp 95-112).

[10] Navis II Project:

[11] GALILI, E., et al., 2021, “Archaeological and Natural Indicators of Sea-Level and Coastal Changes: The Case Study of the Caesarea Roman Harbor”, Geosciences 2021, 11, 306, (26 p).
OLESON, J., BRANDON, C., HOHLFELDER, R., JACKSON, M., 2014, “Building for Eternity – The history and Technology of Roman Concrete Engineering in the Sea”, Oxbow Books, (327 p).

[12] JONDET, G., 1916, “Les ports submergés de l’ancienne île de Pharos”, Mémoires présentés à l’institut égyptien, Tome IX, Le Caire, (121 p).
SAVILE, L., 1940, Presidential address of Sir Leopold Halliday Savile, K.C.B. on 6/11/1940, Journal of the institution of Civil Engineers 15, No 1, November 1940, (pp 1-26).
WEILL, R., 1916, “Les ports antéhelléniques de la côte d’Alexandrie et l’Empire crétois”, Bulletin de l’Institut Français d’Archéologie Orientale, Tome XVI.

[13] ALLSOP, PIERSON & BRUCE, 2017, “Orphan breakwaters-what protection is given when they collapse?”, ICE Coastal Structures and Breakwaters, Liverpool.

[14] MARRINER, N., 2009, “Géoarchéologie des ports antiques du Liban”, edt. L’Harmattan, (262 p).

[15] PAMIR, H., 2014, “New Researches and New Discoveries in the Harbours of Seleucia Pieria”, Harbors and Harbor Cities in the Eastern Mediterranean, BYZAS 19, (p 177-198).

[16] FLEMMING, N., 1980, “Cities under the Mediterranean”, in: “Archaeology under Water”, edt. Keith Muckelroy, McGraw-Hill Book Co, (p 162-177).