Caesarea Underwater Archaeological Park

The Old Caesarea Diving Center publishes a maximum depth and a time for each track. Track A, the sunken harbour's inner and middle basins, runs to four metres over 40 minutes and is the route it calls ideal for snorkelling: Herodian harbour ruins, granite and marble column shafts, the wreck of a steamer, and whatever Mediterranean fish are working the quay that morning. Track B follows the base of the main mole and the subsidiary breakwater to six metres over 45 minutes and takes in the ancient anchors. Track C crosses the great breakwater itself, at eight metres over 50 minutes, past basalt slabs, ancient concrete blocks, piers and a large pile of collapsed blocks. Track D is the deepest and the shortest, 30 minutes at nine metres over the harbour entrance and the northern breakwater, where the ruins of the two entrance towers and ashlar blocks lie.

The layout underwater follows the layout of the port, so the trails read as a plan rather than as a scatter of interesting objects. The enclosed anchorage was about 20 hectares in three basins. The inner one was dug inland and is now completely silted up, marked in the water by a round tower of ashlar blocks 13 metres across, laid with their narrow sides facing out, which may belong to the Hellenistic fortifications of Straton's Tower rather than to Herod's works at all. The middle basin sits in a natural bay and is still the working anchorage. The external basin, the largest, is the one the two breakwaters made.

The main breakwater arcs from south-east to north-west for about 500 metres and is 60 to 70 metres wide at the base. Estimates of the original width of the entrance channel between it and the northern mole range from 30 to 80 metres, which is a useful reminder of how much of this is still being argued from fragments. At the southern end of the mole stood the largest tower, the Drusion, which may have doubled as a lighthouse, and outside the entrance stood twin towers linked at the top by a platform built to carry sculpture that has not survived.

A guided dive with a Dive Master or instructor lasts about 45 minutes to a maximum of eight metres and opens with a talk on the harbour before anyone gets wet. Independent divers can rent gear and take the map out themselves.

• Track A: inner and middle basins, 4 m, 40 minutes, snorkellers
• Track B: base of the main mole and the subsidiary breakwater, 6 m, 45 minutes
• Track C: over the great breakwater, 8 m, 50 minutes
• Track D: harbour entrance and northern breakwater, 9 m, 30 minutes

The city was never lost. The harbour was. Its site slipped out of knowledge entirely, and the first underwater excavations here were headed by A. D. Link in 1960, at the point where recreational diving equipment was new enough to make the work possible at all. Galili, Salamon, Gambash and Zviely, writing in Geosciences in 2021, date the modern research programme from there.

Sustained work began in 1975, when the Center of Maritime Studies at the University of Haifa started on the harbour under Avner Raban and kept at it until his death in 2004. The same 2021 paper puts the Haifa programme at 1975 to 2005, headed by Raban throughout. It is his survey work, more than anything else, that the numbered signs are built on.

Raban proposed the underwater park in 1992. Nothing happened for thirteen years, and the project was only revived after his death; the guide text the diving centre publishes, written by Sarah Arenson, is drawn from his publications and dedicated to his memory. That gap is worth knowing about, because it explains the character of the place: this is a research trail turned into a visitor route, not a visitor attraction that acquired some archaeology.

The breakwater foundations are the real thing and are where they fell. They were made of concrete poured into wooden caissons prefabricated on shore, in sizes on the order of 14 by 7 metres and 15 by 11 metres; in places the timber survives, and in others the impression of the cross beams is still visible in the set concrete. The quays and superstructures above them used ashlar blocks reaching seven metres long and over ten tonnes, some joined with iron clamps fixed by lead poured in situ. The subsidiary breakwater Josephus calls the Prokumatia is also down there, about six metres wide and built deliberately never to rise above the water, so that it broke the winter waves before they reached the main mole.

The six anchors on Track B are a different case, and the guide says so plainly. They were collected by the excavators and placed together on the seafloor to show the anchor developing from the Bronze Age to the modern era, each set as it was designed to sit. One of the six, the composite wooden anchor of about 2,000 years ago, has a reconstructed shaft and arms; only its lead stock and assembling piece are original. This is an exhibit, arranged, in the middle of a site that is otherwise not arranged at all.

The most recent thing on the map is also the most obviously modern. In 1950 the State of Israel built a new quay for the fishermen of neighbouring Kibbutz Sdot Yam, and sank the metal hull of an old steamer, once used to carry illegal Jewish immigrants to Palestine after the Second World War, filled with cement, at the northern edge of the new breakwater. The guide notes without much sympathy that this modern concrete quay, at the time of writing only fifty years old, was already disintegrating.

In between sits the medieval reuse. Muslim and then Crusader builders took marble and granite columns out of the Roman ruins and laid them into a sea wall, a square tower and a jetty. The construction is poor: cemented rubble reinforced with column shafts. Some of the column drums a snorkeller sees on Track A are Roman stone in a medieval wall, not Roman architecture standing where it was put up.

The harbour began sinking almost as soon as it was finished. By the first century CE the head of the breakwater was already under water, which is known from the ships that wrecked on it, and subsidence today is put at five to seven metres. Quays and a pierced mooring stone that were built at water level now sit under five to seven metres of sea.

Why is the open question, and this site is a good place to watch a scholarly disagreement in progress. The park's own guide lists the candidates: a geological fault line running some 150 metres west of the present shoreline, earthquakes, possible tsunami waves, an unstable sandy bottom and the sheer weight of the structures. Galili and colleagues, in Geosciences in 2021, rejected most of that list, finding no conclusive evidence of active faults across the harbour, no reliable historical earthquake account and no unequivocal tsunami signal, and argued instead for geotechnical failure: winter storm waves and currents scouring, washing away or liquefying the sand under the foundations.

A 2025 remote sensing survey of the harbour entrance by Gambash, Arkin-Shalev, Wood, Nantet and Gambin, published in the Journal of Marine Science and Engineering, records that the debate now favours high-energy winter storms over tectonics, tsunami or sea level rise. That survey also puts numbers on the entrance a diver crosses on Track D: the northern breakwater's extremity at roughly 7.6 metres, the northern end of the southern breakwater averaging 7.3 metres, and structures outside the entrance at 7.4 to 8 metres. It reads the two sides of the entrance as architecturally different, with large ashlars up to three metres on the north and a rubble scatter on the south that may be a later attempt at repair.

The repair attempts are visible in the archaeology too. A trench dug across the entrance channel found that in the sixth century a rampart of rubble fill was laid right across it, in an effort to bring the decaying port back into use. It did not work.

In March 2021 the Haifa district planning and building committee approved for deposit a plan for a marine national park at Caesarea, the first in Israel. The plan covers roughly 1,096 dunams of sea west of the terrestrial national park, south of Aqueduct Beach and north of Kibbutz Sdot Yam, and was drawn up with the Israel Nature and Parks Authority and the Israel Antiquities Authority.

The parks authority sets out what it changes. Fishing stops in the core of the park and the movement of motorised craft inside it is reduced, both for the ecology and for the safety of the divers, snorkellers and swimmers who are already there. The plan also allows a floating jetty and operational areas for the diving club already working inside the terrestrial park. Before this, on the planning authority's own account, most Israeli coastal antiquity sites had only their land halves anchored as national parks, leaving the submerged part unprotected.

The scale is worth holding next to the figures on the destination page. The harbour itself covered about 200 dunams; the marine park drawn over and around it is more than five times that. What a visitor gets from the change is not a new exhibit but a quieter, better policed patch of water above a site that spent most of two millennia with nothing formal defending it.