The recommendations and declarations of UNESCO, along with the doctrines and practices promoted by ICOMOS, constitute the core regulatory reference for the conservation of cultural heritage. In particular, the 1972 UNESCO Convention on the Protection of the World Cultural and Natural Heritage establishes the obligation of State Parties to identify, protect, conserve, and transmit to future generations the properties recognized as World Heritage.

ICOMOS, as an advisory body to UNESCO, has produced a series of fundamental documents, including the Venice Charter (1964), which represents a milestone in defining the principles of conservation and restoration. More recently, the "Guidelines for Risk Management for Cultural Heritage" (I.C.O.M.O.S., 2019) offers a crucial resource for understanding how to integrate risk assessments into decision-making processes for conservation.

These guidelines emphasise the importance of proactive and integrated planning that accounts for the environmental, cultural, and social specificities of the context. They recommend adopting methodologies capable of assessing the exposure and vulnerability of cultural properties to natural and anthropogenic risks, including climate change. They encourage the implementation of permanent monitoring systems to promptly identify critical issues.

In 2023, UNESCO established the Climate Action Working Group (CAWG), an initiative aimed at integrating climate action into the management of cultural and natural heritage. This working group focused on formulating concrete strategies to strengthen the resilience of UNESCO sites to climate change, with particular emphasis on developing monitoring systems and timely responses to climate threats (U.N.E.S.C.O., 2023).

World Heritage Action Goal 3, included in the "Policy Document on Climate Action for World Heritage" (U.N.E.S.C.O., 2021), promotes the adoption of mitigation and adaptation measures based on scientific evidence. Among the most relevant articles, Article 3.2 encourages the development of decision-making tools that integrate climate data into conservation plans. This goal specifically suggests the implementation of integrated systems capable of correlating climate data with the cultural and landscape values of the site, to formulate more appropriate adaptation responses. This objective is a cornerstone of this research for the conservation of mosaics in situ.

The 2023 IPCC Report introduces the concept of "Value-Led Risk Assessment" (VLRA), a methodology that proposes integrating the cultural and social value of a property into risk assessment models. This approach is particularly relevant in contexts such as Butrint, where the cultural significance of the mosaic goes far beyond its material value, encompassing aspects of collective identity, historical landscape, and ecosystem interrelations.

According to the Report, this approach allows for a fairer and more sustainable evaluation of conservation priorities. Using participatory techniques, communities are involved in defining the values of the site and clarifying their relationship with cultural heritage (I.P.C.C., 2023). The Value-Led methodology also involves integrating quantitative and qualitative data and applying predictive models based on climate scenarios to estimate the probability and impact of risks.

The Burra Charter, adopted by ICOMOS Australia in 1979 andrevised in 2013, is one of the main theoretical references for the values-based approach to heritage conservation. It states that every conservation decision must begin with an understanding of the cultural values of the asset, including not only material aspects but also social, spiritual, and symbolic ones (ICOMOS, 2013).

In the context of Butrint, the approach advocated by the Burra Charter suggests that adaptive conservation strategies should be developed in dialogue with local stakeholders and based on a shared analysis of the site's values. This approach aligns perfectly with the principle of "informed participation" also promoted by UNESCO and ICOMOS, which ensures community involvement in defining intervention priorities, as well as with the VLRA methodology.

The 2015 Paris Agreement, a global treaty on climate change, establishes in Article 7 the need to "enhance adaptive capacity, strengthen resilience and reduce vulnerability to climate change" (United Nations Framework Convention on Climate Change (UNFCCC, 2015). When applied to cultural heritage, this principle translates into the need to develop adaptation strategies that are both environmentally sustainable and capable of safeguarding the identity and cultural values of heritage properties.

Adaptation, in this sense, must be regarded not as an extraordinary action but as a systemic component of heritage management. Conservation plans must, therefore, include measures for climate risk mitigation and resilience enhancement, adopting an integrated and flexible approach capable of responding to ongoing environmental transformations.

The site has been characterized by continuous topographical variations with relevant expansion or retraction of its use. The morphology of the territory is a consequence of the deposition of alluvial sediments carried to the large estuary. The site became a relevant site because of its position along the coast, favourable for access to a hinterland rich in natural resources, the connections with the cities located further north, and considering the control of routes and commerce towards Greece and the Adriatic and Ionian seas. For all these reasons, the area of the city was first occupied during the late Bronze Age and continuously used until the fall of the Ottoman Empire. The alternate phases of use and reuse have given rise to a highly stratified settlement.

The first traces of occupation date back to the 10th-8th century BC, while the walls and gates date back to the Illyrian period (4th century BC). Among the reasons for choosing the site for the first settlements was the presence of hot springs believed to be beneficial. The Temple of Asclepius, dating back to the Hellenistic period, was built near a thermal spring at the end of the 4th century and expanded after the Roman conquest in the early 3rd century BC (Hodges, 2010). After Augustus's victory at Actium, the city of Butrint prospered as a Roman colony, taking advantage of its location near the Via Egnatia. At that time, Butrint played a fundamental role in the epic founding of Rome, celebrated in Virgil's epic poem, the Aeneid. In the Aeneid (III, 291-293), Virgil attributes the founding of ancient Buthrotum to fleeing Trojans; this information is confirmed by the Greek historian Dionysius of Halicarnassus, his contemporary, in Roman Antiquities (I, 42). During the Roman phase, the urban fabric expanded over the area of alluvial deposits with richly decorated buildings. A monumental aqueduct reached Butrint via a bridge and supplied the city and its numerous baths. After the Roman phase, Butrint became an episcopal centre until the 7th century. During the early Byzantine period, several Christian monuments were erected. Partially renewed in the Medieval era, much later, Butrint was controlled by the Venetians and Ottomans (Raynaud & Islami, 2018).

The catastrophic earthquake dating back to the 3rd-4th century AD caused profound damage to architectural structures, with subsidence due to tectonic movement and consequent flooding and rising water level (Pavlides et al., 2001). Some structural damage caused by this event is still visible in the Hellenistic-Roman Theatre, where the deformed composition of the stone structures and the pattern of diagonal cracks on the western side of the auditorium can be seen. The cracks crossing the foundation of the Triconch Palace Portico, documented in the past archaeological excavations, and the impressive subsidence shown in Figure 1 are further evidence of this event. The diagrams published in The Past (Hodges & Prince, 2022) show the ground level before and after the catastrophe, when part of the forum dropped by almost a meter, with subsequent serious flooding. Inundation by water in the lower city from AD 400 became a regular nuisance. Afterwards, sediment deposition saw ground level rise again. Another seismic event affecting Byzantine monuments is also probable (1153).

However, the disastrous seismic events did not discourage the inhabitants, who repaired their houses and returned to occupy them, making the struggle against periodic flooding a leitmotif of life in Butrint. Currently, the Roman Theatre orchestra is permanently waterlogged by a water depth of 0.5 m. The Triconch Palace, built at the beginning of the 5th century and then abandoned, presumably due to flooding, is also mostly completely submerged, as shown in Figure 2. The archaeological study and on-field research “demonstrated beyond doubt that the environmental circumstances at Butrint have never been static, and in different ways over nearly three millennia the people here have adapted accordingly” (Hodges, 2010).

The latest study on continuous subsidence of part of the ancient city, and the documentation of the presence of evaporate diapirs, confirms that the seismic activity in Butrint is still active, and this condition is an aggravation of the condition of rising waters due to climate change.

A document of fundamental importance for the present study is the publication by Marie-Patricia Raynaud and Agron Islami (Raynaud & Islami, 2018), who are the authors of a vast research project on the entire body of mosaics in Albania. The first outcome is Volume 1, dedicated to the Butrint intramuros mosaics. In volume 25, mosaics are analyzed according to the following areas of investigation:

The construction of mosaics in Butrint covers three main chronological periods: the Hellenistic Period (Figure 3), the Roman Period, including first and second phases (Figure 4), and the Protobyzantine Period (Figure 5).

The surviving part of the Hellenistic mosaic is on display in the museum. Dating back to the end of the 3rd or beginning of the 2nd century BC, is a fragmentary mosaic panel that decorated the interior space of the Asklepios temple and depicts a coiled snake, an animal considered sacred to Asklepios’s cult. The mosaic shows a mixed technique of opus signinum dotted with white cubes for the background and opus tessellatum for the serpent, with thin lead strips to outline its contours. Red, blue, and yellow tesserae model the snake by neatly laid rows, with shades of pink, white, and yellowish-green tesserae appearing on its underside. The execution is refined, and the panel, once removed from its original setting, is now exhibited in the Butrint Museum. The original Hellenistic mosaic has been subsequently overlapped by another mosaic during the Roman period.

Most of the mosaics in situ belong to the Roman period and date from the early 3rd to early 5th centuries. An exemplification is the numerous mosaics in the Triconch Palace Complex. It was built in several stages over several centuries and, in its final phase, extended over two insulae. The mosaics in the building date back to various periods and display great originality of geometric design. Some include figural elements, such as the Mosaic of the Masks, finely crafted in opus vermiculatum. The decorative motifs of the Roman mosaics mainly use black, white, grey, and pink tesserae.

The most important mosaic from the Protobyzantine period in Butrint is the complex of the Baptistery. The wellpreserved and iconic mosaic that adorns the Baptistery was completed at the end of the 6th century. The remarkable size of the Baptistery, which exceeds that of many other Eastern baptisteries, as well as the fact that this religious architecture is the last early Christian example in the western Mediterranean region, make this archaeological monument and its almost completely preserved mosaic unique in terms of value and significance. The Baptistery's mosaic pavement displays at least seven circular friezes, with intertwined pink circles filled with a multitude of animals symbolizing God's creation in the air, land, and sea, and includes floral motifs. Similar to the other Protobyzantine mosaics in the city, the tesserae materials include assorted hues of glass in addition to stone, marble, and terracotta tesserae.

The study by Marie-Patricia Raynaud and Agron Islami (Raynaud & Islami, 2018) has made it possible to define the basic knowledge framework for the in-situ conservation project for the mosaics. For each mosaic, information is available on the materials components of the substrate, generally deduced from visual observation of the layers that emerged from the gaps and cracks before restoration work began. It should be noted that there is a lack of data on petrographic characterization of both the bedding mortars and individual mosaic tesserae. The only laboratory analyses carried out concern a few vitreous tesserae, the results of which, relating to the chemical components identified, have been published.

Observation of the composition of the substrate confirms that it was constructed in accordance with Vitruvius' treatise (De Architectura), namely with three layers: Nucleus, Rudus, Statumen. Mortars rich in dust or brick fragments are often present, a detail that informs us of the presence of mortars with more or less hydraulic characteristics depending on the size or smallness of the brick fragments.

Regarding the characteristics of the mosaic’s tesserae, unfortunately, nothing else is known except that they are made of local limestone in various colours ranging from white to grey and black or from light pink to dark pink. Red tesserae are usually made from squared fragments of brick. Based on a study of the geological map of Albania and the area around Butrint, the soils outcropping in this area generally date from the Lower Jurassic to the Burdigalian (Lower Miocene) periods, so it can be said that they are indeed limestones. Their characteristics are believed to be similar to those of the coeval limestones found in the Adriatic region on Italian territory; based on these assumed similarities, the low porosity of this type of limestone is a positive factor for its in-situ conservation.

The above Figure 1 shows a synthesis of the main information on the most important mosaics in the Triconch Palace Complex, including information on value: data supports a general strategy for mosaics conservation.

The accurate descriptions provided by Raynaud and Islami inform us about the extent of the mosaics based on the investigations carried out, the state of conservation observed at the time of their protective layer removal for the purpose of the study, and the safety measures performed. In addition, the publication provides information on the level of exposure of the mosaics to flooding, defined as periodic/seasonal or permanent. Based on the information provided, the state of conservation of the mosaics has been defined as limited to the following parameters:

Table 2 shows a synthesis of other key information collected on a selection of the most important mosaics, including data on altitude.

Overall, although the information gathered indicates that the mosaics have a good level of integrity, it is equally clear that their conservation is seriously at risk because of fluctuations/variations in the sea level. The continuous water level variation is affecting the conservation of the archaeological structures and the on-site mosaics. In fact, rising damp on the supporting structures, with consequent evaporation and efflorescence salt formation, has a destructive effect on the archaeological remains. This phenomenon is emphasized in the areas affected by seawater and brackish water flooding.

It should be noted that the mosaics are currently neither inspectable nor visible as they are covered by a protective layer composed of Bidim® (non-woven fabric), sand, and gravel. Given that it is impossible to inspect mosaics, their state of conservation has been assessed based on the following data collected from various sources, as indicated below:

The monitoring data analysed are related to the period from September 2012 to November 2022. Measurements of the water level of the Vivari Channel were taken twice a day, at 8:30 am and 15:30 pm, corresponding to the low tide and high tide periods, respectively (see Figure 6).

The following observations can be drawn from the analysis of the 10 years of monitoring of the canal water level measured near the site entrance:

During the long on-site mission in Butrint held in August 2024 for survey and investigation, 5 samples of the water flooding the mosaics were collected. Table 3 shows the laboratory analysis results.

The results of ion chromatography analysis reveal that the water flooding the Triconch Complex’s mosaic is actually seawater, while all other sites have chloride salt content that makes them brackish water. It is surprising that this result also applies to sampling site S1, which is the highest point on the slope (see Figure 7).

The high chloride content in the water flooding the mosaics, combined with fluctuations in the water level, is a key factor affecting the conservation of the mosaics. The diagrams below (Figure 8) show the fluctuations in the water level at six relevant sites in the Butrint area, based on monitoring data collected over a period of approximately three months.

The monitoring period at the six sites was very limited in time, covering only seven months, corresponding to the rainiest months. Nevertheless, the data collected clearly show a direct correlation between rainfall events and the rise in water level, which peaks after the end of the weather event. The same observation was made by the author during the repeated on-site visits after the reining events. Furthermore, it is interesting to note that the highest sites near the mountain slope maintain a water level of about 10 cm, while the Nymphaeum area tends to dry out again, and finally, the Gymnasium and Triconch sites tend to dry out with the arrival of summer. Considering the data collected and the impossibility of carrying out a visual assessment of the conditions on site due to the protective layer, two questions arise: 1) Does the protective layer really perform a protective function, or is it rather a factor amplifying the degradation? 2) How far has the degradation progressed since the last inspection?