Tombolos & Salients

Tombolos have been inhabited by people since prehistoric times and around 65% of the natural tombolos found on the coasts of the Mediterranean Sea are ancient settlements.

Our aim in this section is to describe some parameters defining the existence and the size of tombolos and salients. Much research on tombolos and salients has been done in the past 50 years because of its interest for coastal protection works like detached breakwaters, which are man-made structures placed parallel to the existing initial shoreline (see some references hereafter).


Strictly speaking, a tombolo is a sand spit connecting the mainland to an offshore island or obstacle. A sand spit is often generated on each side of the island. In the lee of a small island, both sand spits will join as a sandy isthmus, possibly leaving a triangular marsh area near the initial coastline. A large island may generate two separate sand spits and a large marsh area (Giens, Orbetello). In this study, we will call “tombolo” the whole mushroom-like feature including the sandy isthmus and the island.

Small tombolo with a single isthmus, Cap Serrat, Tunisia (Google Earth, 2019)
Large tombolo with a double isthmus, Orbetello, Italy (Google Earth, 2019)

The Cap Serrat tombolo is small with a 160 m island, and the Orbetello tombolo is one of the largest with a 12 km island, about 80-times larger, but the general geomorphological shapes are similar.

A rocky cape is obviously not a tombolo, as it does not have a sandy isthmus.

Wave crests nearing the coast are often more or less parallel to the shoreline (say with an angle smaller than 10°), and this will generate the typical symmetrical shape of the tombolo shown in the pictures above. If wave crests have a more oblique incidence (say between 10° and 45°), the tombolo will have an asymmetrical shape. In case wave crests approach the coastline with an angle larger than 45°, a sand spit is generated. A true sand spit has a free end usually turning around like a hook. However, in a few cases, such a sand spit may encounter an island and this may technically speaking be called a tombolo, but we will not do so in this study because no wave diffraction is involved in that process (Cadiz in Spain, Chesil Beach in UK).

This sand spit is not part of a tombolo because waves do not approach from the left, Cadiz, Spain (Google Earth, 2019)

A bell-shaped salient occurs when the offshore island is smaller, or further away from the initial coastline, than for tombolos. In the case of oblique wave crests, the salient will be asymmetrical. It must be noted also that some alluvial fans (fluvial sediment deposit at the river outlet) have a similar aspect, but they usually have no offshore island and no wave diffraction effects.

Summarising, a tombolo or a salient can exist only if:

  • the initial coastline is a sandy beach (or shingle, pebbles),
  • wave crests approach the initial coastline with an angle smaller than 45°,
  • an island (or obstacle) yields an area sheltered from waves and generating wave diffraction,
  • no currents flow between the island and the initial coastline.

L is the length of the island and D is the distance from the initial shoreline. For the tombolo, we define b as the smallest width of the isthmus, and for the salient, b is the width at the inflexion point, and d is the distance from the tip of the salient to the initial shoreline.
Each tombolo is thus geometrically defined by L/D which is a constant, as long as the sea level is constant.


When offshore wave crests approach the coast with a small angle (usually less than 10°, but in any case, less than 45°), they are subjected to refraction which tends to align the wave crest parallel to the coastline. If wave crests encounter an obstacle like a small island, they are subjected to diffraction which tends to turn the waves around the island.
When breaking on a sloping coastline, waves release their energy in turbulence and in longshore currents which may transport sediment (usually fine sand with a median diameter D50 of 0.2 mm to one or more millimetres). Shoreline forms are thus created by wave refraction and diffraction and longshore sand transport. The combination of these physical phenomena is complex but now well modelled in some physical scale models and in numerical so-called “one-line models” reproducing the shoreline.

Wave diffraction pattern behind a detached breakwater, Xilxes, Spain (Google Earth, 2015)

The diffraction pattern is clearly shown on the picture above where breaking wave crests are always at a small angle with the shoreline. This implies that some longshore sand transport is ongoing from right to left, towards the tombolo. If you would walk from right to left on the beach shown on this picture, you would first see waves refracting during their approach to the shore (bottom right of the picture), then you would enter the curved shoreline where wave refraction and diffraction are combined (the shoreline shape is close to that of a logarithmic spiral). Finally, you would arrive in the lee of the detached breakwater and enjoy maximum shelter from offshore waves. It has been shown that a current system exists which forms a circulation cell turning in clockwise direction (in the particular case shown on the picture above, Mory & Hamm, 1997).

Mediterranean tombolos

Using Google Earth, we searched the 45 000 km of the Mediterranean Sea coasts to detect tombolos and salients, and over 120 sites were listed. This list includes around 75 natural tombolos and around 20 natural salients generated by offshore islands, islets or reefs. In addition, around 25 sites with man-made single or multiple detached breakwaters were listed.

Let’s first have a look at the relationships between L and D, and b and D for natural tombolos:

L/D for natural tombolos (for L < 1000 m)
b/L for natural tombolos (for L < 1000 m)

Both figures show much scatter which will unfortunately not be reduced much in the following analyses.

For L/D, it may be seen that, very schematically, L/D = 1 to 2, with many exceptions.
For b/L, it may be seen that b/L = 0.05 to 0.7 with much scatter. A kind of average at b/L = 0.3 corresponds very well to results given by Rosen (1982, fig. 8) who gives:  a = 0.33 L, with a = (L-b)/2.

Let’s now have a look at a relationship between the dimensionless parameters L/D and b/L:

L/D for tombolos and for salients, including both natural and artificial man-made structures (for L/D < 2)

Again, much scatter is found.

We can distinguish “fat” tombolos with b/L = 0.5 to 1, and “umbilical” tombolos with b/L < 0.05. Between these extremes, a wide area is covered by “normal” tombolos, where artificial tombolos nicely mix-up with natural ones, without any difference between single and multiple detached breakwaters.
The demarcation line between tombolos and salients is L/D = 0.65. This result is in full agreement with several other researchers (Mangor, 2020 and Sunamura, 1987) and in line with other researchers who give wider ranges (van Rijn, 2013, Bricio, 2008, Ming, 2000).

Tombolo:L/D > 0.65
Salient:L/D < 0.65

but beware !!
Tombolos may connect and disconnect the island or detached breakwater (i.e. become a salient) depending on wave conditions., e.g. storm waves may change a tombolo into a salient, not only due to waves turning around the breakwater’s roundheads, but also due to waves overtopping the breakwater and/or wave energy passing through the porous breakwater. This is particularly true for umbilical tombolos with small b/L < 0.05 where the demarcation line between tombolos and salients may range between 0.65 and 1.5.

Mediterranean salients

As mentioned above, salients are found for L/D smaller than 0.65, and we may have a look at the relationship between the dimensionless parameters L/D and d/D:

L/D for salients, including both natural and artificial man-made structures (for L/D < 0.65)

Strictly speaking, a tombolo is formed when d/D = 1 and L/D > 0.65, that is in the upper right of the figure above. We can distinguish “weak” salients with small d/D, and “developed” salients with larger d/D. The tags show the values of d/b and a trend is visible as shown by the oblique lines.

Umbilical tombolo,
Çiçek island, Turkey (Google Earth, 2019)
Umbilical tombolo,
Porto Pollo, Sardinia (Google Earth, 2019)

The large scatter found in the above results for tombolos and salients is a bit frustrating and we intuitively feel that the wave climate must have some effects on the shape of tombolos and salients, e.g. one third of the umbilical tombolos are located in areas sheltered from offshore waves … but two thirds are not!

Effects of wave climate

The above-mentioned tombolo-shape changes due to periodical storms make any assessment of a Google Earth picture somewhat tricky because we do not know if it was taken just before or just after a storm. However, Google Earth usually provides at least half a dozen pictures taken at different times over the past 20 years and we are able to estimate some movements of tombolos and salients.

Same salient moved NE,
Altafulla, Spain (Google Earth, 9/3/2017)
Salient moved SW,
Altafulla, Spain (Google Earth, 29/9/2016)

Altafulla is located around 10 km NE of Tarragona (Spain) and features a beautiful salient with L/D = 0.44 and b/L = 1.05, sheltered by a 105 m detached breakwater. It seems that during the winter 2016-2017 the tip of the salient moved ca. 40 m from SW to NE, showing that the waves came mainly from SE in September 2016 and from SW in March 2017.

Although the tip of the Altafulla salient moved during a winter season, and although it is observed also that b varied from 85 m to twice this value during a period of 5 years, the overall shape did not change much and we may consider that all tombolos and salients found on Google Earth are in a state of long-term dynamic equilibrium.

A first step to find out how the wave climate interacts with the tombolo- and salient shapes, would be to measure the “Dean number” for each site [De = H0/T/Vf  with H0: offshore wave height (m), T: wave period (s), Vf: fall velocity of sediment in water (m/s)]. This parameter is dimensionless and includes the effect of both waves and coastal sediment. As we do not have the value of the Dean number for each tombolo site, we used a more regional approach, comparing shapes of tombolos and salients located in various parts of the Mediterranean area (eastern Spain, southern France, western Italy, south Aegean, Levant, North Africa) and on the Pacific Ocean between Vancouver and Cape Horn.

All tombolos and salients for various specific areas.

The figure above does not show any groups of tombolo-shapes (in terms of L/D versus b/L) for the selected areas, and we have to conclude that local wave climate differences (if any) seem to have no effect.

Moreover, Sunamura’s model-test results (their fig. 3 & 4) can be re-interpreted as showing no influence at all of their “K” parameter including wave effects.

In other words, the wave climate (and perhaps even the sediment grain size) does not seem to affect the final equilibrium shape of tombolos and salients. It might then be considered that waves affect only the speed of shape evolutions, and not the final equilibrium state, but this does sound very realistic as we know that wave parameters (e.g. wave steepness) and sediment grain size both have an influence on the beach slopes.

In any case, a much finer approach is probably needed and the last word has certainly not yet been said …


MANGOR, K., 2020, “Detached breakwaters – MarineSpecies Introduced Traits”, Wiki,
For single detached breakwaters:
Tombolo: L/D > 0.9 to 1.0
Salient: L/D < 0.6 to 0.7

VAN RIJN, L., 2013, “Design of hard coastal structures against erosion”,
For multiple detached breakwaters:
Tombolo: L/D > 2
Salient: L/D = 0.5 to 2
Weak salient: L/D = 0.2 to 0.5
no effect: L/D < 0.2

BRICIO et al., 2008, “Geometric Detached Breakwater Indicators on the Spanish Northeast Coastline”, Journal of Coastal Research, Vol. 24, N° 5, (p 1289-1303).
For 27 multiple detached breakwaters on the Catalan coast:
Tombolo: L/D > 1.3
Salient: L/D = 0.5 to 1.3
Limited response: L/D < 0.5

MING , D., & CHIEW, Y., 2000, “Shoreline Changes behind Detached Breakwater”, Journal of Waterway, Port, Coastal, and Ocean Engineering, ASCE, (p 63-70).
From scale model tests, with interesting result on salient area.
Tombolo: L/D > 1.25

MORY, M. & HAMM, L., 1997, “Wave height, setup and currents around a detached breakwater submitted to regular or random wave forcing”, Coastal Engineering, Vol. 31, (p 77-96).

ROSATI, J., 1990, “Functional Design of Breakwaters for Shore Protection: Empirical Methods”, Technical Report CERC-90-15, US Army Corps of Engineers, (49 p).
Nice overview of all formulations, but no final answers.

SUNAMURA, T. & MIZUNO , O., 1987, “A study on depositional shoreline forms behind an island”, Ann. Rep., Inst. Geosci., Univ. Tsukuba, N° 13, (p 71-73).
For 23 natural islands in Japan:
Tombolo: L/D > 0.67
Salient: L/D = 0.3 to 0.67
Limited response: L/D < 0.3

SUH, K. & DALRYMPLE, R., 1987, “Offshore Breakwaters in Laboratory and Field”, Journal of Waterway, Port, Coastal, and Ocean Engineering, ASCE, (p 105-121).
Scale model tests and some field data.

ROSEN, D. & VAJDA, M., 1982, “Sedimentological Influences of Detached Breakwaters”, Coastal Engineering Conference, (p 1930-1949), (fig 8).
For single detached breakwaters:
Tombolo a/D = 0.33 L/D (a is sand-free distance along inner side of BW).

DALRYMPLE, R., 1976, “Study of equilibrium beach profiles”, Coastal Engineering Conference, (p 1277-1296).
“Dean Nb”: H0/Vf T (H0, T of waves, Vf is fall velocity of sediment).