Evaluation of cliff recession in the Valle dei Templi in Agrigento (Sicily)

• Water flow through fractures leads to the progressive collapse of the foundation soils.
• Sand collapsible behaviour can be responsible of instabilities in the Valle dei Templi.
• Additional discontinuities occur because of collapse phenomena.
• The modelling process confirmed the in situ survey measurements.
• The soil parameters that need to be monitoring can be then identified.

The sacral complex of the Valle dei Templi in Agrigento, added in the UNESCO Heritage Site List, stands over the crest of a rigid calcarenite cuesta, which overlies a layer of partially saturated (S = 6–12%) carbonate sand. In turn, the sand stratum lies on a thick stratum of clays. The environment is highly prone to landslides as highlighted by several previous studies that identify the undermining of slopes and the discontinuity pattern, occurred in the late Neotectonic phase, as causes of cliff failure. For wider and less exposed areas, where the undermining local instabilities can be present because of the specific morphology of the site, a recent research has proposed a new interpretation of the failure general mechanism. The new hypothesis here proposed would ascribe to the collapsible behaviour of the sand; both the failure mechanisms of the cliffs and the development of the discontinuities in the rock mass. A unitary framework is then proposed. A series of direct shear and oedometer tests was performed on the collapsible sand samples and values of collapse potential were estimated. As typical for collapsible soils when flooded, vertical strains occurred rapidly causing an increase of shear stresses in the upper calcarenite, characterised by low values of yielding and low values of tensile strength. As a consequence, additional discontinuities occur, volumes of rock fall and cliffs move back. Both by means of in situ surveys and the analyses of involved rock deformability and strength characteristics, a soil model was considered and a modelling process developed. The model results were compared with measurements of displacements as well as dimensions of fallen blocks and consequently the comprehension of the effect of sand collapse was then possible. The identification of the new mechanism key factors allows the definition of the soil parameters that need to be monitored. In order to prevent the cliff recession for the archaeological site, the work proposes a plan of monitoring methods. These methods will be able to estimate the cliff recession, to localize “sensitive areas” and to quantify the parameters that could activate the instabilities.