Peter Brimblecombe, School of Environmental Sciences, University of East Anglia, Norwich UK


Our works of monumental heritage are designed to last. They use durable materials to survive, and often their architects see their preservation on the millennial scale. After the early 18th century restoration of Westminster Abbey, Nicholas Hawksmoor wrote to the Dean: "I am of the opinion, if violence does not happen, this fabrick will stand 1000 years".

Despite this assurance by Hawksmoor materials are damaged by their environment and despite careful choice of building materials degradation and weathering is almost inevitable. The environmental pressures outside are distinguished most notably in terms of liquid water, direct solar irradiation and higher wind speeds compared with the indoor environment. Outside there is always a risk through exposure to air pollutants, but these can be prevalent indoors as well especially where there is indoor heating or pollutants volatilise from indoor construction materials. Understanding the historic weathering and damage to materials gives not only a sense of their long term vulnerability, but also how society responded to the damage.

Historically the environmental forcing factors have been (i) climate change, (ii) a shift in the nature of the urban atmosphere and (iii) a change in the rural environment often driven by innovations in agriculture.

(i) The climate has recovered since the period since Europe’s Little Ice Age in the middle of the last millennium to be somewhat warmer and less prone to freezing in many areas, so the ravages of frost weathering have become less apparent. Across the period 1750-1950 many aspects of the heritage climate have seemed relatively stable, although the 1780’s and the last half of the 20th century have been periods of considerable change

(ii) Although the impact of smoke on building facades and their interiors has been a problem since classical times it was not until the 17th century that there was an awareness of the corrosive nature of the pollutants from fossil fuels. London architects such as Sir Christopher Wren witnessed a notable increase in recession rate and attributed “fretting quality” to “smoaks of the sea-coal”. In the late 19th and early 20th century the sulfation of building stone reached catastrophic proportions and much detail was lost from stone mouldings, while architectural metals suffered badly. Iron corroded rapidly and copper formed less coherent patina’s that failed to protect the underlying metal. The problem of primary urban pollution was addressed by the changing use of fuels in the last part of the 20th century to leave rather more subtle problems such as that which has accompanied the rise of diesel vehicles, particularly  the deposition of diesel smoke and associated organic materials on urban surfaces. Increased nitrate concentrations on stone facades has encouraged biological growth

(iii) Agriculture and land management has caused huge changes to the environment. The water table and surface flooding now respond very differently to precipitation than in the past. The changes in crops, flora and fauna in the countryside has also imposed damage in rural areas. New insect pests are found damaging heritage materials under a warmer climate, while there are fears (not entirely justified) that emissions of ammonia from modern fertilizer use have damaged frescos and other decorations within buildings in the countryside.

Often the changes imposed by the environment show effects relatively slowly, but this makes them difficult to manage as they are seemingly invisible and it is only after a long period that the damaging effects are obvious.  This makes long term historic exposures worthy of study and they can also set a base-line against which future changes can be measured and give observations against which models of future change can be calibrated.

Historical studies also reveal past practices and attitudes towards accumulating damage, some of which were climatically driven. We can see shifting priorities and allocation of resources towards the maintenance of buildings. Buildings can also reveal the way in which interventions illustrate choices about what was worthy of preservation and the kind of aesthetics adopted. We also find changes in the away visitors used properties that reflect a shifting climate and society over recent centuries.



Stefan Michalski, Senior Conservation Scientist, Canadian Conservation Institute


This paper describes an attempt to go beyond just listing the risks facing archaeological sites, towards a quantified assessment of these risks, both the probabilistic events one usually associates with the word risk, as well as the cumulative decay processes that drain sites of their value. The first goal is the assessment of the old risks that we've known about since long before climate change, which will continue into the future. The second goal is the assessment of the new risks brought by climate change, and their comparison to the old risks. The third goal is the creation of risk reduction options, and their evaluation in terms of effectiveness and cost-effectiveness. The final goal will be to prioritize practical risk reduction options for the site manager.

Our first step is to rank order the risks at a global scale. Nineteen risks were analyzed. The "tornado" graphs below (figure 1) present them in rank order.

Figure 1. Graph of risks to global archaeological heritage assets. After each risk name is a letter: H: Human cause (ignoring climate change); CC: Climate change cause. N: natural cause (ignoring climate change). In this context, "item" refers to whole classes of material, as listed in Table 1. The left hand graph shows Magnitude of Risk on  a logarithmic scale. Each unit is a factor of 10. The coloured bars are the "expected" or average risk, subdivided on the left graph by the three components of the analysis: the Frequency or Rate (red); the Loss to each "item" affected (yellow), and the Amount of items affected (blue).  The right hand graph shows the total risk on a linear scale, to emphasize the differences in the top 10. The small black and white bars on the left graph are the range in estimates between a Low estimate and a High estimate. The break from black to white is the Most probable estimate.

The highest score for magnitude of risk, 15, is set at total loss of everything in 1 year. Each of the four largest risks at magnitude ~11 represents an expected rate of loss of 10% of everything per century. The lowest risks at magnitude ~9  are  100 times slower, so 0.1% loss per century. Note that the three highest risks are "old" risks. Two are Human (urban development, and theft from sites "unknown" to archaeology.) The third risk is an ironically overlooked natural risk: the continued natural decay of unexcavated historic sites, known and unknown. Each 100 years that passes adds significantly to their relative decay, simply because they are relatively young, and they represent a large proportion of all archaeological "stock."  While one may object to the consideration of such "inevitable" risks, it is useful to compare risks driven by our activity to those driven by our inactivity.

Figure 2. Magnitude of Risk totals for groupings of risks shown in figure 1: (from top to bottom) Human (except climate change);  Natural (except climate change) and Climate Change. Note, this graph and figure 1 are not a complete assessment of climate change risks. Additions as well as revisions are expected to emerge during discussion.

In summary, considering figures 1 and 2, we can say that although climate change is bringing new risks, and some of these will clearly become the biggest risk locally, at a global scale they will not displace the old risks to archaeological sites. Managers must continue to manage the same old risks as before. On a philosophical note, however, we must confront the fact that the biggest of the old risks – urban development and theft from (officially) unknown sites – lead back to the same cause as climate change itself, i.e., the unsustainable growth of humanity's footprint on the planet.

How was the assessment done?

I used the CCI-ICCROM-RCE risk assessment method and its database tool.. Although designed for individual heritage assets, the method can be applied to global assets as a single entity. Sixteen of the risks were identified based on our experience looking at risk in historic house museums, and our experience with six case studies of archaeological sites during the 2011 ICCROM-CCI-RCE course on the method. Three risks were added that are always mentioned in discussions of climate change and heritage. (See figure 1 or the Appendix for the risks analyzed). I hope to add more as a consequence of the roundtable discussions. Analysis of each risk relied, as always, on available data, common sense, theorising, and imagination (as in "Fermi problems"). Individual analyses are in the attached Appendix, for the curious, in a format generated by the database tool, sorted by magnitude of risk.

Risk analysis of all kinds is an exercise in stretching calculations, so an explicit treatment of uncertainty is important for transparency. This is done in this method in terms of a "Low" and a "High" estimate, alongside the "most probable" estimate (known as a triangle distribution). These enter the graph as the narrow black and white bars at the tip of each risk bar. Currently these uncertainty ranges are much larger than the differences between adjacent risks. Whereas the differences between the top third, middle third, and bottom third of the risks are significant, rank order within each third is not.

Archaeological sites were divided into categories that would aid analysis in terms of generalizations, shown in the table below. I included unknown sites, since the assessment presumes to assess risks in terms of all archaeological material, discovered or not.

The table shows the "relative quantity of material" or if one prefers "relative value" assigned to each category for purposes of this assessment. The database tool allows one to quickly change these ratios, so as to explore the sensitivity of risk ranking to different opinions about these quantities. Except for the frozen sites, unknown material in each category was estimated to be of similar quantity as the known material. This was based on an analysis of the number/size distribution of sites, which range from the ~300 city-state complexes/temples large enough to make a tourist guide list of "the world's best archaeology sites" through the 1500 "notable sites list" in Wikipedia,  finally to an estimate of ~30M domicile scale sites (assembled from various arguments). Assuming a simple inverse relation between the number of sites in a given scale range and the value of each site in that range, the total value of each scale range becomes the same. Whereas it is inevitable that almost all the very large sites are already known, I estimated only 1-10% of an estimated ~30M small sites are known. On average, known site material was estimated as ~1/2 of all possible material, i.e., the unknown ~ known.

Each of the 16 site categories was in turn subdivided into 4 subcategories of site material, based on Inorganic vs Organic, and High value vs Average value. These four subcategories were assigned equal "quantity of material" in terms of value, i.e., an average site may have 1k high value items and 100k average value items, but in terms of a prioritization exercise like risk management, the 1k high value items might carry as much weight as the 100k average items. Perhaps more. Perhaps less. As already noted, it is not that these value judgements must be determined a priori in the method, rather they are just a starting point for the database, after which the sensitivity of the risk assessments to changes in these judgements (or lack of sensitivity for some risks) can be studied quickly and transparently.

In total, 64 separate categories of "site materials" were available for analysis of each risk. For example, analysis of weathering of outdoor stone could be limited to inorganic material at known dry sites, all ages, and all values. Analysis of storage risks such as fire and rot could focus on organic material of average value, on the assumption that most high value material was on display or in special locations.  Theft by professionals favours high value material. Etc.

During the round table, it is hoped to hear about risks that have been overlooked. Some of the sensitivity analysis will be presented. A list of the estimation details that give rise to the most significant uncertainties will be assembled, in order to elicit information where it will be most useful for improving the assessment.

For the site manager, I will present some examples of how local risk assessment and option analysis may differ from the global assessment presented here.


Erasmo Buonomo, Hadley Centre, Met Office, Exeter, UK


This contribution includes i) a short review of the main issues associated with climate model projections on a centennial timescale and ii) an overview of results from climate model projections in the Mediterranean region. These two topics will help to understand the issues  in estimating climate change variables at scales relevant for impact studies and in estimating the range of possible outcomes from model projections, which are discussed in the final section.


Climate Models and Inter-comparison Projects


Climate scenarios and climate impact studies rely on the climate projections generated  by Atmosphere-Ocean General Circulation Models (AOGCMs). These models include a detailed description of the atmosphere, the ocean and the other components of the Earth climate system which are relevant for its evolution up to  centennial time scales. Many components have been added to AOGCMs in twenty years of development; atmospheric chemistry and carbon cycle are among the most recent modules which have been included in many AOGCMs. Input data for climate projections derive from the best available information about natural external factors (e.g. solar input and volcanic emissions) and emissions of greenhouse gases, derived from plausible assumptions about possible evolutions of the world  for this century.


The increased model complexity has contributed to make them computationally very expensive. For this reason, global model have a quite coarse horizontal resolution, which limits their ability to resolve processes at local scales. Since it has also been recognised that coarse horizontal and vertical resolution severely limit the possibility of reproducing the main features of atmospheric and oceanic circulation, current AOGCMs have also a larger number of vertical levels (~ 30 levels) and an increased horizontal resolution (~ 100km for the atmospheric component). While these improvements have produced  more detailed and physically plausible descriptions of climate processes, they are still not sufficient to produce useful data for scales commonly used in impact studies (usually of the order of few km). This problem also  exists for temporal scales, most short-lived events which are important at local scales, e.g. extreme weather events, happen on spatial scales not explicitly resolved by AOGCMs. The mismatch in resolution between AOGCM output and impact models requires the additional step of downscaling AOGCM output to the resolution needed for impact studies. One of the methods which have been used to solve this problem is based on the use of Regional Climate Models (RCMs), a limited area version of AOGCMs (usually without ocean component), run at higher resolution (typically 12-25km) from  large scale flow and oceanic conditions derived from AOGCM simulations. This strategy provides a rather inexpensive way to resolve faster processes of smaller extents (e.g. mesoscale systems, 200-2000km) while still providing model output for the most commonly used climate variables which is both physically and spatially coherent.


The possibility of generating very detailed maps of changes, however, doesn't solve the problem of climate projections. In addition to the impossibility of knowing future emissions, there is also considerable uncertainty in climate models due to the assumptions needed  to build them. The largest part  of model uncertainty originates from processes which are not explicitly resolved at the resolution used for climate projections, some of which are modelled by using statistical or empirical relationships that are dependent on parameters which cannot be determined neither by theory nor by observation. As a consequence of these different assumptions, AOGCMs produce climate projections which could be significantly different at regional scales.


At smaller scales and for rare weather events, the uncertainty on climate projections is also affected by large contributions from climate variability, generated by the chaotic nature of the system with contributions from all spatial and temporal scales which might become predominant.


Model deficiencies and uncertainty in climate projections are fundamental topics of research in climate science. For this reason, a series of multi-model set of AOGCM integrations performed under the Coupled Model Intercomparison Projects (CMIP)  have been distributed in association with Intergovernmental Panel of Climate Change (IPCC) assessments. The main purpose of these sets of climate integrations is the analysis of the skill of AOGCMs in reproducing the Earth climate and to understand the physical reasons for differences in climate change responses generated by climate models with different formulation. These sets have also been extensively used for impact studies and constitute the best available resources for this purpose.  As I noted above, AOGCMs are not able to describe local processes: for this reason, downscaling strategies have been applied to IPCC AOGCM sets, with the aim of producing climate scenarios at local scales and to generate climate data for impact studies.  The idea of model intercomparison as a way to assess model biases and the robustness of climate projections has also been extended to regional climate models. Areas of the world which have been investigated by a range of RCMs, driven usually by a subset of CMIP AOGCMs projections, include Europe (PRUDENCE and ENSEMBLES), South America (CLARIS) and North America (NARCCAP). Regional climate model data from these projects have been used extensively in impact and adaptation studies, sometimes directly as a input for impact models.


Based on these successful experiences, the World Climate Research Program (WCRP) is supporting a new project,  Coordinated Regional climate Downscaling Experiment (CORDEX), which is extending this approach to all the whole world. More in details, this project aims to downscale the CMPI5 multi-model set on 11 regional domains covering all the continents (including the polar regions) using all the RCMs available to the research community. In the interest of maximising the participation to this project, in particular from research institutes from developing countries, RCMs will be using a resolution of 50km for these integrations. Climate model data from this project will be available before the end of 2012 at the Danish Meteorological Institute (DMI) data server.


Assessing Climate Change: the Mediterranean Basin


The Mediterranean region provides a good example of common issues in assessing the vulnerability to climate change of a particular area. From a climatological perspective, the choice of a region is determined by the homogeneity of its main climatic features and in the changes expected in the future climate. While the size of the region might be too big to be relevant for most impact studies and vulnerability assessments, the spatial homogeneity of its climatic features provides additional information on the geographical robustness of the changes. In other words, it will help to understand whether a particular change in hazards has been obtained by chance (i.e. purely from climate variability) or by some underlying physical process, which would necessarily act on a spatial scale much larger to those relevant for impact studies. Another good reason for the choice of the Mediterranean basin is the availability of observed data and regional climate model simulations, mostly as a result of EU funded projects such as PRUDENCE, ENSEMBLES and CIRCE, which makes it one of the best studied regions of the world.


The Mediterranean region has been already identified as one of the most vulnerable to climatic change. Climate models agree in the description of a hotter, drier mean climate by the end of this century (Christensen et al, 2007; Giorgi and Lionello, 2008), with a reduction on large scale storminess, both in winter and summer. The reduction of precipitation (20% annually under a medium-high emission scenario, Mariotti et al, 2008) is the aspect of the changing climate which is raising concerns, since many areas in this region are already under water stress in the present climate. Local changes in sea level are more uncertain as reported by a recent multi-model study based on the IPCC AR4 set of Atmospere-Ocean General Circulation Models (AOGCMs) (Marcos and Tsimplis, 2008), while the contribution from the global ocean, including possible melting of ice sheets, is strongly dependent on the dynamics of exchange at the Gibraltar Strait (Artale et al, 2006) which is not modelled with sufficient accuracy in climate models.


Changes in extremes could have larger impacts than changes in the mean climate and it is not uncommon that extremes might change in the opposite direction with respect to the mean climate.  An example of this behaviour comes from projected changes in precipitation extremes from the PRUDENCE  RCM integrations, showing a significant increase on the European coasts of the Mediterranean basin in winter (Christensen and Christensen, 2005, Frei et al, 2006). Analyses of changes in circulation also show that, while there is a reduction in storminess, the number of intense cyclones is not expected to decrease (Lionello et al, 2008). At the extreme end, the Mediterranean basin can also generate very intense cyclones which can have a very high impact due to their strong winds and intense rainfall (Homar et al, 2007) . Recent studies have focused on the predictability of the explosive cyclones (“bombs”), which are not too infrequent in this area, from a weather forecast perspective (Genoves et al, 2006). However, there are not too many studies on the effect of climate change on these weather phenomena.


Among these extreme cyclones, there are also systems whose structures looks quite similar to tropical cyclones as seen from satellite images. Their meteorological features are different from tropical storms, since they are mid-latitude systems; however they are also strongly dependent on the temperature of the sea surface, in particular their energy depends on the heat extracted from the sea (e.g. Emanuel, 2005). These phenomena are quite rare, only 15 events have been classified as tropical-like hurricanes in the last 20 years for the whole Mediterranean basin (Fita et al, 2007). As in the case of tropical hurricanes, it is very difficult to model Mediterranean hurricanes at the model resolution of AOGCMs. Regional climate models seem to be able to reproduce the main features of these systems: a multi-model study based on the PRUDENCE set of RCM climate projections (Gaertner et al, 2007) has shown that all the models are able to reproduce the basic features of these systems. The same study has also shown an increase in intensity for the future climate for some RCMs in the PRUDENCE set of integrations. Given the rarity of these events, the set of PRUDENCE RCM integrations would not be sufficient not make any quantitative estimate, not even for the whole basin. However, since these events could be substantially more important in the future climate, an assessment of hazards in the Mediterranean region would not be complete without an analysis of these weather systems. One common aspect in the investigation of climatic extremes is the use of appropriate post-processing tools, usually state-of-the art statistical and numerical methods, which allow an efficient evaluation of the signal from the available climate data and an estimate of the usually large statistical noise associated with extremes.


In the discussion above, I gave examples of changes in hazards which could be studies directly from the output of regional climate models.  For many other phenomena, meteorological events are just one of the contributions, sometimes not even the most important,  to the increase in risks.  These phenomena usually require to understand how these non-climatic factors might be changing as well. An example is given by the quantification of sea level changes including a non-climatic contribution from land motion (Woppelmann and Marcos, 2012). Other examples include storm surges, coastal erosion, river floods and flash floods, for which quantitative estimates can only be given by taking into account all the contributing non-climatic factors. In some cases,  the complex interaction of these factors with the changed climate can only be studies by modelling the impact under investigation. Such impact models could be quite complex from a computational point of view, they might require good quality observations for their setup and could also add large contributions to the uncertainty of predictions from non-climatic factors. Examples of work done in the Mediterranean region can be found in the work done under the Research Theme 6 (RT6) of the ENSEMBLES project.


Using Climate Model Data


The assessment of climate change projections for the Mediterranean basin shows some characteristics which are common to all the climate change assessment at regional scales. More in details, there will be some features reproduced by all climate change projections and some others for which there is not even an agreement on the sign of change. In the Mediterranean areas, warming in all seasons and drying in summer are robustly identified by climate models. However, even in this case, there might be considerable differences in the size of responses from different projections. Model uncertainty doesn't stop with AOGCMs: the PRUDENCE set of RCM integrations has been used to estimate of the effects due to small scales (Deque et al, 2006): changes in seasonal temperature and precipitation show a strong component due to regional models, which are mainly responsible for describing the scales not resolved by AOGCMs (< 400km). This component could be basically identified as the climate signal due to local processes not resolved by AOGCMs.  Quantitative estimates of hazards would have to take into account the most comprehensive range of responses, e.g. by including measures of uncertainty based on the different model outcomes and by including a sample of small scale signals due to different downscaling methods.


The assessment of meteorological changes in hazards related to rare events is even more complex for the strong dependence on local features and on the way these features interact with the meteorological systems in the changed climate. In this case, it is necessary to use downscaled climate scenarios. In addition, there could be an increased weight of variability at small scales, as it has been reported for intense precipitation (Kendon et al, 2008), which can be minimised either by considering larger regions or by sampling the variability by increasing the number of model simulations with different initial conditions.


Any attempt to quantify the range of possible outcomes for a meteorological variable would have to start from a meaningful sample of climate model projections, which would allow  a reasonable estimate of all the contributions to the spread of climate signals. As it has been discussed above, impacts and adaptation assessments need additional input data which could be even more uncertain than climate model projections, thereby adding another contribution to the range of projections.


Available multi-model sets from the international research projects mentioned above are good starting points to evaluate climate signals and to estimate the range of outcomes due to different model formulations and different emission scenarios. However, any attempt to estimate climate change signals at local scales and for extremes would be affected by the limited availability of downscaled climate scenarios and from the lack of ensembles of climate projections designed to sample climate variability at small scales. Current effort to tackle these two problems include the WRCP project CORDEX, which will soon produce AOGCM-RCM multi-model sets of downscaled simulations for all the continents and by research projects such as weatherathome, designed to produce an ensembles of GCM and RCM integrations with millions of members, giving for the first time the possibility of assessing changes at small scales for localised weather events.


David Orrell


Climate change is expected to affect many aspects of our lives and environment in coming decades. Cultural monuments which have survived for millennia will not be immune, and may find themselves subjected to increased temperature, pollution, and even a rise in sea levels. We would clearly like to predict these effects, but are accurate forecasts possible? This paper presents a brief history of prediction, from the oracle at Delphi to modern climate models. It shows that our predictive models are based on a reductionist paradigm which can be traced back through scientists such as Newton to the ancient Greeks. While mathematical models have proved successful in areas such as physics and chemistry, they are demonstrably poor at predicting complex organic systems such as the Earth’s climate system. Such systems are dominated by emergent features and nested feedback loops which resist the reductionist approach. This lack of predictive ability has not reduced our reliance on sophisticated models. Instead, they have taken on a cultural status similar to that of the ancient oracle. To predict the effect of climate change on cultural monuments, this paper suggests an approach which eschews complicated models for a method which combines scenario forecasting and simple models.

1. Introduction Climate change has been a contentious issue for many years. Scientists have disagreed over the historical record, the role of human emissions in driving climate change, and our ability to accurately model and predict the climate. There is now accumulating evidence both that the climate is changing, and that it is driven at least in part by carbon emissions. But many scientists – and not just those in the pay of oil companies – remain skeptical about our ability to predict the evolution of the climate system. This paper looks at the historical context of climate models, and argues that they have become cultural monuments in their own right.

Section 2 begins with a condensed history of forecasting, from the time of the ancient Greeks until the present day. It shows that our predictive models are based on a reductionist paradigm, which attempts to predict and control a system by breaking a system down into its components, and applying mathematical laws to see how it will evolve. Sections 3 to 5 show how this reductionist approach has been applied, with limited success, to weather and climate modelling, and discusses the causes of forecast error. Section 6 argues that climate models are maintained less for their forecast accuracy, than for the fact that they embody strong cultural values, such as a belief in order and rationality. In other words, they have shifted from being solely a scientific institution to being also a kind of cultural monument. Section 7 proposes an alternative approach, which combines scenario forecasting and simple models, that can be applied to the problem of assessing the effect of climate change on cultural monuments. Finally, Section 8 summarises the argument and suggests ideas for how to handle climate uncertainty.


Dr Ehud Galili, Israel antiquities Authority and Zinman Institute of Archaeology, University of  Haifa

Sea level rise and coastal erosion- the time perspective: Most civilizations and urban centers in the world evolved in the last 4000 years during relatively stable sea level condition. During the 20th century a sea level rise of about 0.2 meter was recorded all over the globe. Predicted sea level rise in the 21th century is up to about one meter. Such rise will have crucial impact on coastal regions. While urban centers and living human societies can adapt, change and move, coastal archaeological sites can not. After four millennia of relative stability, we are facing a rapid global sea level rise. Records from the last million years indicate that sea level can reach as high as + 7 meters above present sea level even without human intervention.
Human intervention and coastal development: The Mediterranean region is considered to be the cradle of many civilizations, religions and cultures. The cultural and the natural heritage of the Mediterranean countries were and are strongly modified by climatic change and human intervention.
In recent decades marine and coastal environments were intensively disturbed by human activities. Massive construction works took place along the Mediterranean coasts. Archaeological evidences for coastal-marine, including coastal and underwater archaeological sites recording an evolution lasting millennia, are rapidly eroded and destroyed. There is an extremely narrow time span to salvage, protect and preserve the coastal and marine cultural resources of the Mediterranean.

The Israeli perspective: The coastal archaeological heritage of Israel reflects important chapters and events in the history of humanity, including the Neolithic revolution and the appearance of the first Empires. It contains the physical evidences for the foundation of the major monotheistic religions and other major historical events. Human activity in the coastal region (particularly sand quarrying and the construction of marine structures) resulted in an extreme shortage of sand along the coast, rapid erosion and destructive effects on coastal and underwater sites. The problem has been exacerbated by the global rise in sea level over the last century. Ancient sea walls, structures and installations are collapsing (fig. 1). The coastal settlements at risk (fig. 2) can be classified into three categories: Fortified coastal cities with a sea wall, founded on a rocky platform; stratified coastal tells and submerged Neolithic settlement. Erosion rate in several locations is more than one meter per year. If this process continues, a significant portion of the coastal and marine cultural heritage of Israel will disappear and archaeological, tourist and economic assets of great value will be lost.

Actions taken by Israel to rescue the coastal and underwater cultural heritage: Israel conducts a series of actions including: monitoring, risk assessment studies, preservation and protections plans and established special legislation for the protection of the coastal environment. Underwater surveys aimed at rescuing and documenting underwater sites are being carried. A comprehensive GIS data base was established for the coastal and underwater sites. Damage to the ancient coastal sites is constantly monitored visually, photographically and by field surveys. During 2009 a policy document and risk assessment study of the coast and the ancient coastal heritage was prepared by the Prime Minister's Office jointly with other organizations. A detailed master plan for protecting and preserving the sea fronts of the ancient coastal settlements of Israel was established by the Israel Antiquities Authority. Pilot projects for preserving and protecting selected sections of the sea fronts of Ashkelon, Ashkelon north, Ashdod Yam, Apollonia, Caesarea and Akko were conducted.

Coastal settlements at risk in Ashkelon region
Tell Ashkelon:  The site is a national park. The fortified city was founded during the Bronze Age. Numerous granite and marble architectural elements are indicative of the magnificence of the city during the Hellenistic and Roman periods. In the Early Islamic and the Crusader periods the city was fortified by a heavy wall and the seawall was reinforced by granite columns taken from ancient buildings. Principal risks: collapse and massive destruction of the seawalls, erosion and run-off accelerate the destruction of the coastal cliff. There is a significant narrowing of the sandy beach due to the construction of harbor installations south of the site. As a result waves are directly damaging the archaeological remains. There is a dire safety problem stemming from landslides, deterioration and collapse of sections of the coastal cliff, parts of buildings and installations. In the absence of protective beach sands, the tell strata are being washed away after the collapse of the sea walls. Shipwrecks and cargoes are exposed on the seabed and are threatened by treasure hunters. Measures required for salvaging the site: archaeological salvage excavations of the buildings and the installations on the tell seafront which are in immediate danger of destruction, dismantling dangerous buildings that cannot be stabilized. Conservation works of  buildings and installations at the seafront: filling empty overhanging spaces, pointing up joints, stabilizing slopes by means of terraces, vegetation and nets, arranging drainage on the slope and above it, protecting the seafront by reconstructing Crusaders seawalls (figs. 3, 4), or building a protective seawall of boulders and/or depositing sand. Yearly monitoring: panoramic photography and locating new risks when they occur; underwater surveys to be carried out year round to locate, document and salvage the remains of shipwrecks and cargos newly exposed on the sea bottom.
Ashkelon Mayumas (Ashkelon North): Settlement remains from the Byzantine period and a fortified compound are located on the coastal cliff and are currently being destroyed by the sea (figs. 5, 6). Erosion rate has reached more than one meter a year (Fig 7). A temporary protection of the cliff foot using geo-technique plastic sleeves filled with sand was constructed several years ago (fig 8). It was heavily damaged by the sea and failed. Shipwrecks and cargoes are often being exposed on the sea bottom. Archaeological salvage excavations of buildings and installations on the seafront, which are in immediate danger of destruction, are required as well as underwater surveys year round.
Ashdod-Yam: The site, an Early Islamic and Crusader fortress, is surrounded by a wall. Parts of the seawalls have undergone conservation and were restored. The fortress and the seawalls are damaged by the waves. Conservation and restoration measures are required, as well as protection of the sea front.

Conclusions: Sea level rise and human activities, mainly building and quarrying, are causing massive coastal erosion and the rapid destruction of unique coastal and underwater sites. In recent years, significant damage has been caused to the ancient coastal settlements of Israel and valuable archaeological assets have been lost. Further developing works in the coastal region and rising sea levels during the 21st century will cause a severe damage to the ancient coastal settlements. Without the implementation of protective and conservation measures, substantial parts of the ancient coastal settlements of Israel will be lost within several decades. The detailed master plan for protecting and preserving the sea fronts of the ancient coastal settlements of Israel should be applied.


Christos S. Zerefos, Academy of Athens

The Mediterranean is universally recognised as particularly vulnerable to the impacts of anthropogenic climate change. The climate change in Greece concerning air temperatures are projected to rise significantly by the end of the 21st century, while precipitation is projected to decrease. Available observations indicate that precipitation during the 20th century decreased by around 20% in Western Greece and by 10% in Eastern Greece. This is attributed in part to the positive trend in the North Atlantic Oscillation (NAO) index, which will continue through the 21st century. Future climate projections indicate that the next decades will see a significant increase in the frequency of extreme temperature and precipitation events.

According to the simulations of anthropogenic climate interference under the two extreme scenarios (B2 and A2), by 2100 the air temperature is expected to increase by 3.0ºC and 4.5ºC respectively. The temperature increase will be greater in the mainland than in the island areas, and more pronounced in the summer and autumn than in the winter and spring. Meanwhile, the decrease in rainfall, countrywide, is expected to range between 5% (Scenario B2) and roughly 19% (Scenario A2). The climate simulations estimate that relative humidity, countrywide, will decrease between 1% (Scenario B2) and 4.5% (Scenario A2). Relative humidity levels are projected to drop more markedly in the mainland regions, especially during the summer, but are expected to remain unchanged in the island areas.

The simulations also point to a decrease in cloud cover in Greece in the coming decades, compared with the baseline period 1961-1990, by between 8% (Scenario B2) and 16% (Scenario A2). Solar radiation on a national scale is expected to increase by between 2.3 W/m² (Scenario B2) and 4.5 W/m² (ScenarioA2), while mean annual wind speed is not expected to change significantly, apart from the Etesian winds the intensity of which is expected to increase by as much as 10%.

Based on the available simulations, even under the intermediate Scenario A1B, the Greek mainland in 2071-2100 will experience some 35-40 more days with a maximum daily temperature exceeding 35ºC, while even greater will be the increase (by around 50 at the national level) in the number of tropical nights (when minimum temperatures do not fall below 20ºC). On the other hand, the number of frost nights is expected to drop significantly, especially in Northern Greece (by as many as 40). Moreover, the rise in average temperature will prolong the vegetation period by 15-35 days.

Changes are also projected in precipitation extremes. In Eastern Central Greece and North-Western Macedonia, the maximum amount of precipitation occurring within 3-day periods is expected to increase by as much as 30%, whereas in Western Greece it is expected to decrease by as much as 20%. By contrast, the greatest increases in drought periods are projected for the eastern part of the mainland and for Northern Crete, where 20 more drought days are expected per year in 2021-2050 and up to 40 more drought days are expected in 2071-2100. As a result of climate pattern changes, the number of days with a very high risk of fire is expected to increase significantly by 40 in 2071-2100 across Eastern Greece (from Thrace down to the Peloponnese), while smaller increases are expected in Western Greece. The number of days with a humidex value of >38 ºC will be prolonged by as many as 40 in the coastal areas along the Ionian Sea and in the Dodecannese, and by somewhat less (roughly 25 additional days) in the low-lying areas of continental Greece and Crete, as projected for the period 2071-2100.

Based on estimations, sea level is projected to rise between 0.2m and 2m by the year 2100. Of course, any assessment of an area’s vulnerability to a rise in average sea level (coastal risk assessment) inevitably involves considerable uncertainty, as such risk is determined not only by the rate and extent of the sea level rise, but also by other local factors, such as tectonics, sediment supply (from inland) and coastal geomorphology/lithology.

Typical examples are the coastal areas of the Northern Peloponnese (projected to gain 0.3 to 1.5 mm/year in elevation), Crete (0.7 to 4.0 mm/year) and Rhodes (1.2 to 1.9 mm/year). Thus, for instance, an average rise in sea level of about 4.3 mm/year could be reduced to 3.5 mm/year as a result of a compensatory, tectonically-induced mean elevation uplift of about 0.8 mm/year. Changes in sediment-laden inflows in the deltaic estuaries of large rivers can potentially offset a rise in sea level, if the sediment deposits increase, causing the delta in question to advance. Conversely, a decrease in fluvial sediment discharge could result in a greater marine inundation of the area in the event of a rise in sea level. Finally, coastal morphology, especially in terms of coastal slope inclination and lithologic composition, is an important factor, directly related to the rate of erosion.

Calculations of shoreline length showed that of Greece’s total 16,300 kms in coastline, some 6% (or 960 km) correspond to coastal deltaic areas of high vulnerability; 15% (or 2,400 km) to newly-formed soft sediment layers of moderate vulnerability, while the remaining 79% (or 12,900 km) correspond to rocky coastal areas of low vulnerability. Therefore, the total length of shoreline with a moderate-to-high vulnerability to a rise in sea level comes to about 3,360 km, i.e. 21% of Greece’s total shoreline.

Assuming that there are no tectonically-induced and geodynamic corrections, a rise in sea level by 0.5m to 1m would result in a shoreline retreat of between 30m and 2,750m in the high risk deltaic areas, such as the Axios-Aliakmon or the Alfeios deltas, while a rise in sea level by 1m would result in a shoreline retreat of between 400 m and 6,500 m.



Peter Brimblecombe, School of Environmental Sciences, University of East Anglia, Norwich UK


The future is uncertain. Although there are excellent models for climate change which  extend across the current century they are often poorly tuned to understanding the weathering of our heritage. This problem required the NOAHs ARK project to first establish key meteorological parameters affecting cultural heritage and cartographical representations of potential damage to materials in the form of an atlas. It brought a requirement to define Heritage Climatology. Classical climatological maps, such as those of Köppen, can be applied to heritage, but often miss some of the environmental pressures that affect monuments, buildings and sites. The heritage climate needs to be projected into the future to allow strategic management of heritage through the 21st century. However, the way we express climate change impacts on heritage and the reliability of model outputs and predictions of damage remain difficult.

Only subset of meteorological parameters seems be relevant to heritage. For example, the affect of a few degrees change in temperature on the deterioration process might be seen as relatively slight, because stone or metal in themselves are insensitive to temperature, although at risk from intense solar radiation. However, there are ways small changes can be amplified. Warmer temperatures give longer frost free periods, decreases in snow cover and a lengthening of growing season that lead to a broad range of phenological impacts. These can be most noticeable in spring: e.g. earlier breeding or singing of birds, flowering of plants or spawning of amphibians. Freezing and thawing is common in climates with winter temperatures close to zero. Increasing winter temperatures may make frost damage less frequent in future in mild climates such as that of Britain but potentially more frequent in colder climates.

Freezing represents a phase change for water. Such changes are important in causing damage to materials, such that when water freezes or salts crystallise the volume changes can impose mechanical stress on materials. Phase changes are sensitive to climate as they occur at a discrete values of temperature or relative humidity. This means even slight changes in climate can allow phase boundaries to be crossed more or less frequently. Thus the frequency of freeze-thaw events is likely to decrease substantially in many culturally important sites in temperate Europe, even though the winter temperature change is only a few degrees.

Other critical factors that were readily identified in NOAH’s ARK. It was especially clear that the water relation of heritage materials was especially important. The presence of liquid water is linked to temperature, which frequently increases in colder climates. This often comes about through prolonged times of wetness, which for metals and leads to higher rates of corrosion or higher deposition rates of pollutants and more favourable conditions for microbiological activities, greater salt mobilisation. In the vapour phase water is also responsible for deterioration. This is usually described in terms of relative humidity. When this increases most materials show enhanced rates of deterioration rate. Changes in RH lead to crystallisation and dissolution processes within porous stone and the pressure exerted can be high enough to disrupt the stone in a process known as salt weathering. 

Increased precipitation can increase the damage caused by wet deposition by dissolution of surface layers of materials. Changes in the chemical composition, especially pH, can affect the deterioration rate. Wind can increase eddies and turbulent flows around historical buildings and alter the deposition rates of both gaseous and particulate pollutants. It can and strengthen the effect of driving rain and abrasive windblown sand. A very serious effect may be the increased transport of sea salt inland.

Air pollutants, while often under control through regulation can increase in a warmer sunnier world. Most notably ozone, which is formed through the action of sunlight and volatile organic compounds could well increase. Future climates could lead to greater evaporation of volatile organic compounds and an enhanced photochemistry.

The NOAH’s ARK project modelled changes in cumulative processes of deterioration for Europe across the period through to the end of the century. This work reveals patterns of change that are dependent on geography and climate. The Mediterranean is quite different to the rest of Europe.  A particularly relevant example of the role of climate and geography is in salt weathering across an area of the continent stretching from Britain, France and northern Spain through to central Europe. This is especially important given the dominance, in this region, of vulnerable and highly detailed gothic architecture where sculptural detail, such as gargoyles, is carved from soft porous stone. Although climate change does not seek to threaten particular architectural forms, but conversely architecture styles are naturally influenced by climate. Gothic architecture focuses on the pointed arch and verticality with finely moulded pinnacles.  It originated outside the Mediterranean and emphasised the requirement of Abbot Suger (~1081 – 1151) that churches should be airy and bright. This required large glass windows in northern latitudes. Thus Gothic architecture initially occupied a particular climate regime, so it should not be especially surprising that this is also an area where climate change has a particular character. In this case the changes are potentially negative for the architecture.

Such observations and predictions stress the relevance of geography and climate to the cumulative pressures on heritage.




Rohit Jigyasu, UNESCO Chair Professor, Research Center for Disaster Mitigation of Urban Cultural Heritage, Ritsumeikan University, Kyoto, Japan


Based on available empirical data from various sources, the paper will trace link between climate change and increasing frequency and intensity of related momentary hydro-meteorological hazards such as rainfall. However there are two main challenges in drawing clear conclusions based on this analysis. Firstly, there is high degree of uncertainty as one is unsure which impacts can directly be attributed to climate change, since many local factors related to vulnerability and exposure of cultural heritage can be the root cause of disasters rather than hazards themselves. Secondly, time horizon for climate change is too long to ascertain the probable rate of change. While highlighting these issues, the author takes the standpoint that hazards such as earthquakes and floods are not disasters in themselves, rather these are events that may trigger catastrophic situations that we call disasters.

Cultural heritage located in certain geographical regions such as coastal and mountain areas are increasingly exposed to disasters resulting from climate change. Primary hazards such heavy rainfall, GLOF, cyclones and drought also often result in secondary hazards, which can be momentary such as landslides and fires as well as slow and progressive such as soil erosion and weathering. Their combined impact exacerbates the scale of disasters. This would be illustrated through various examples.

However climate change has not only increased scale of events such as rainfall but is impacting their intensity for example there might be fewer days of rainfall but the amount of rainfall on those few days is much higher. As a result there are increasing instances of cloud burst and flash flooding in and around many heritage sites. Therefore average figures of rainfall may not help to understand the nature of disasters due to climate change. This would be substantiated through examples.

It is also important to consider what kind of heritage sites are more vulnerable to disasters resulting from climate change. This depends not only on the nature of hazards but also on the characteristics of heritage site in terms of material, construction systems, and most importantly, their values. Disasters resulting from climate change can cause irreversible damage to authenticity and integrity of cultural heritage. However more severely, delicate balance of local ecological systems of cultural landscapes can be damaged, thereby affecting their long term sustainability. Climate change related hazards have also significantly increased vulnerability of historic urban areas to flooding, especially due to transformations in historic built fabric resulting from rapid urbanization and failing infrastructure that is unable to cope with increased demand. Floods in Mumbai, India (2005) and recent ones in Thailand (2011) illustrate this issue. We cannot discount increased economic vulnerablilty of heritage sites due to potential impact on livelihoods, especially related to tourism, and institutional vulnerability due to weak management systems.

In order to meet the challenge posted by potential increase in disasters due to climate change and resulting loss of lives as well as economic impacts, it would be best that governments as well as heritage management institutions invest heavily on enhancing warning systems as well as monitoring of potential impacts on the heritage. This would require much closer cooperation between heritage related institutions, line departments in planning, services and infrastructure and those responsible for disaster mitigation and response. This would essentially need a conceptual shift from reactive post disaster response and recovery to proactive disaster risk reduction. While on one end, disaster risk reduction needs to be part of heritage management, heritage concerns also need to be mainstreamed into larger development, disaster risk reduction and climate change agendas as these are interlinked in more than one way.

Last but not the least, local capacities and knowledge systems for disaster mitigation that have evolved through series of trials and errors in response to past experience of disasters, need to be updated for facing new challenges emerging due to changing nature and intensity of disasters because of climate change. ‘Living with risk’ approach has defined response to disasters in many traditional societies. Probably traditional management systems need to be evolved to identify ways of adaptation rather than resistance to the inevitable phenomena of climate change.




Sujeong Lee: National Research Institute of Cultural Heritage of Korea


This presentation examines the effect of climate change from management perspective of conserving a heritage site.  The most problematic obstacle to deal with the issue, is the uncertainty of the effect of climate change and the impact on cultural heritage, and therefore, it makes site managers and policy makers difficult to decide whether climate change should be recognized as an influential factor to change the present management plans or not. Such context also has made managerial response to the issue be slower than that of academic field.

However it is important to remind that there is one certain fact that changing climate is affecting and will affect to the physical state of heritage.  Considering that the nature of conservation is to manage the physical change, it is necessary for heritage professionals to incorporate the issue to their decision-making process.

Under the present circumstances, it will be more realistic and reasonable for this presentation to aim at providing ideas on what practical initiatives can be taken in decision-making at management level at present stage.

In order to provide logical and constructive ideas, the presentation first identifies which part of a whole process of conservation can be closely related to the issue of climate change. The first section will briefly explore conservation as a social process to understand the nature of decision-making and to locate the climate change issue in the process. The section will point out that scientific research for conservation treatment and establishing management guidelines are deeply connected with the issue of climate change. Also, these two activities are reciprocal.

The second and the third sections will take each activity to examine the reason why and the way how climate change issue can (or cannot) be working as an influential factor in decision-making of each activity.  The sections aim at providing as set of areas to be tackled by conservators, scientist, and site managers, and policy-makers to respond to the effect of climate change at management level.

The presentation recognizes that government authority plays an important role in executing suggested practical initiatives by providing necessary resources, such as funds, national policy, human resources, and etc. So the final section will examine what is government role and what are their tasks to tackle the issue in collaboration with related professionals and fields.

Climate change and conservation

Conserving heritage is a social process to understand values, to preserve attributed values in ‘authentic’ state, to decide a way of use for the benefit of present generations, and to deliver what is valued to future generations.  Two main practical activities of the social process, which relate to the effect of climate change, are: 1. conducting scientific research and intervening to material fabrics; and 2. setting out legal frameworks and establishing management guidelines. The latter guide decision-makers to a rational decision to take an appropriate respond to meet the purpose of conservation.

Scientific research and conservation treatment

The purpose of scientific research and conservation treatment is to diagnose the cause of physical damage and identify other factors to devalue heritage values, and to intervene to the fabric and its surroundings to restore it into or preserve in sound state. The way how effect of climate change can affect to decision-making is much dependent on the notion of ‘damage’.  Some conservators define damage as ‘any kind of material change’ whereas others as ‘negative state to cause the loss of heritage’. Contemporary approach toward the notion of damage takes the latter in decision-making.  It means that it is important to construct reliable scientific knowledge on whether the effect of climate change will be an additional risk factor to the existing ones.  If so, it also needs to evaluate the degree of the effect to decide whether the effect is serious enough to consider in decision-making.  (This section will be more investigated, revised, and added based on the information from presentations of our round table session A, B, and C.)

Setting out legal frameworks and management guidelines

Although there are holes and gaps in understanding of the effect of climate change on cultural heritage, it is important to include the possible scenario of climate change into legal frameworks because it is not possible to return to the earlier state once heritage is lost or damaged. It means that it is not ethical approach for conservators to put heritage at any possible risk. Therefore, this section will examine what possible options for policy makers to incorporate climate change issues in legal frameworks and management guidelines under the current situation with limited knowledge on the effect. One possible option is to set out and provide procedural guidelines for site managers who set out management plan.  The guidelines provide what site managers should consider in relation to the effect of climate change at each step of setting out management plan, such as analysis of risk factors, monitoring such factors, and etc. (This section will be more explored with the discussions and insight from the Session E to make this presentation to be useful bridge to connect Session A – C and Session E)

Role of government in improving managerial response to climate change.

This presentation will finish with a set of recommendation that government should involve in and cooperate with conservators, site managers, and scientist. As a resource provider, a government needs to overview the climate change issue from holistic perspective to fill the gaps and holes in taking an action at management level. (this section aims at clarifying a state role in dealing with climate change issue.  It recognises the importance of a state role as a leading institute to fill the gap between scientific progress of academia on the issue and managerial respond to the issue and practical application of scientific knowledge at management level )



John A. Fidler, RIBA, IHBC, Intl. Assoc. AIA, FRICS, FSA, FIIC, FAPT., Principal, John Fidler Preservation Technology Inc., Marina Del Rey, California, USA

Local responses to climate change impacts on archaeological sites must be based on informed conservation, foresight and preparedness. Technical and other forms of evaluation based on detailed information gathering should be expedited through a management plan process. Such processes seek to understand the values and significance of the site (not all monuments or parts of monuments have equal value); make an assessment of its physical layout and condition; evaluate current and potential future hazards and risks in relation to the physio-chemical sensitivities of the site; and form prioritized policies, procedures, actions to prepare for, manage and mitigate disaster, damage and loss within resource constraints.

Management plans need to be regularly tested and improved through peer review, scenario game planning and ultimately post disaster evaluations. Management plans are not static tools: they are dynamic and must respond to local changes of circumstance. They need not be expensive or overly sophisticated but, above all, they should be realistic, and based on local resources in the context of national and international frameworks.

Climate change impacts on archaeological sites will vary dependent upon their geographical location and climate zone. So, for example, Britain is likely to expect increased incidences of more exaggerated weather and as a consequence some or all of the following hazards: flooding (flash floods, ocean storm surge, sea-level rise and river inundation, rising water tables); wind damage (i.e., increased frequency of strong gusts and gales); higher general levels of precipitation; and increased days of ground and air frost. Archaeological sites in other parts of the world may suffer greater frequencies of lightning strikes, fire damage and mud slides (e.g., Mesa Verde National Park, Colorado, USA); hurricane damage and so on. Consequential impacts may also involve flora and fauna migration (including arrival of pest species); desertification and site burial.

Some of the natural forces involved are so physically overwhelming and uncontrollable that their impacts cannot be physically mitigated except at extraordinary expense. For example, by the installation of permanent flood barriers; roofing over sensitive parts of sites; or dismantling and reassembling monuments on higher ground (e.g., Abu Simbel). So an increasingly politically popular and pragmatic alternative approach has been to record the monument for posterity’s interest and then to allow inundation or loss through ‘managed retreat.’

However, between these poles there are a variety of ways to prepare for, manage and cope with the aftermath of disasters caused by climate change impacts because the changes and their impacts are currently part of a relatively slow-moving process.

The speaker will address these issues and offer palliative, ephemeral, temporary and permanent solutions for improved conservation management of archaeological sites.



Evangelos Kyriakidis, Senior Lecturer, University of Kent, Canterbury; Director, Initiative for Heritage Conservancy

Values assigned to archaeological sites by the various stakeholders represent the contribution of the existence of these sites to modern day life. Regardless of how one assigns values to an archaeological site, these values are crucial in understanding the threats this site is facing: a disappearance or an alteration of these values is equal to the disappearance or alteration of the entire site. Some values may have a primacy while others come as secondary. Some values reflect some wishful thinking on the part of the heritage manager, while others can be readily observed. Some values are associated with the tangible aspects of this site, while others are strongly connected to intangible qualities or associations to a site.

This paper aims at discussing how climate change increases the risk of losing or degrading values of sites and consequently their interpretation. The risk of losing main values of a site is equally important to any other major threat. This value-based thinking (as opposed to a monetary value thinking – which is the monetary part of the values of a site) helps us conceive a number of otherwise unpredictable threats. Of course this adds a layer of subjectivity to the already quite subjective risk assessment of a site or an item.

For example much of the historic value or the use value of a cooking pot is lost by the very fact that it is removed from its context when moved to a museum display. Much of the historic value of an heirloom is lost if the story about it being passed from one generation to another is forgotten. But also if the characteristics of a cooking pot erode away, and it is not recognisable readily as such, many of the associated values also disappear. The values may be abstract, but they are directly linked to the physical of the monuments.

Many archaeological sites do not stand alone. They belong to a specific natural surroundings and are very much affected by it. A great example of such a correlation is the Minoan Peak sanctuary of Philioremos Gonies in Crete. Peak sanctuaries are Minoan sites (2nd millennium BC) that are attested in peaks close to zones of human habitation. These peaks are chosen as sanctuaries because they have a fabulous visibility over the surrounding landscape, they have a strategic position and value as they control an entire area, they unite lands that cannot see each other otherwise, they are perfect negotiation points, they are unique lookout points, and they have a visual and audio connection between them and the surrounding habitation areas. The essence of these sites is not in what is in them, their archaeological finds, but in what they ‘see’ and control. The landscape is more important for the values of the sites than all the archaeological finds put together. Any change in that landscape will affect the values of the site itself, and any past change of that landscape changed the values of the site through time and the way these were interpreted.

(Photo: The Village of Gonies, courtesy of Andreas Smaragdis)

Climate change, we can extrapolate, did change the landscape during the times that the peak sanctuary was in use. This means a change in economic activity in the area, and therefore a different human (not only natural) landscape. Different types of economic activity were possible in different times (and climatic conditions). This needs to be accounted for in the management of the content of the site (e.g. the research, the narrative, the signage, the experience that the visitor has). Climate change is going to change the landscape again (which also changes because of other human activities – not only climate). The change of landscape for this and other peak sanctuaries does not mean the end of the site, except if the human environment around it is not sustainable any more, if no traditional human economic activity is possible (pastoralism / agriculture).

(Photo: outside the village of Gonies, courtesy of Andreas Smaragdis)

The site of Zominthos for example 300 meters higher in altitude than this site, was originally considered by the excavators as a seasonal site. Seasonality has been a particularly important value of this site. This is because the area around the site, during the first excavations, was not habitable during the winter. All activity in the site as well as its visitation, takes place nowadays during the summer. Seasonality, moreover, affected (at least in the beginning) the narrative around the site. Recent climate change (and the worsening of the Cretan winter in the last centuries) has been crucial in the formulation of ideas around the site (including the bias that we as researchers have), the values of this site, and therefore the management of this site. Yet new climatic data show that during the use of the site winters were milder than today, and it would have been possible to use the site throughout the year. 

In other words, understanding climate change is doubly important for the management of archaeological sites. Ancient climate change is crucial in the understanding of the values of archaeological sites in the present. Yet it is also important for the management (and risk management) of these sites. Modern climate change threatens these values, and affects seriously the management of our sites.




Thomas D. Andrews, Territorial Archaeologist, Manager, NWT Cultural Places Program, Prince of Wales Northern Heritage Centre, Yellowknife, NT  Canada


Hunting caribou/wild reindeer (Rangifer tarandus) on alpine ice patches (above 1500 m a.sl.) during summer months was a part of a broader-spectrum subsistence economy for many northern indigenous peoples.  Recent archaeological and paleoecological research in northwestern North America and in Scandinavia has demonstrated that alpine ice patches have been relatively stable for much of the Holocene, serving as preservation environments for biological remains and fragile hunting tools abandoned during hunting episodes.  Preservation of organic components (wood, bone, antler, hide, sinew, feathers, and mastics), as well as structural and decorative elements of complex tools rarely seen in circumpolar archaeological contexts, has led to significant advances in our understanding of the manufacture and use of hunting technologies.  In the Northwest Territories, analysis of biological remains preserved in alpine ice patches (largely caribou dung and bone) has demonstrated long-term stability in caribou food habits and population genetics, and that alpine environments have remained largely unchanged for at least 5000 years. In recent decades, however, alpine ice patches in the Northwest Territories have been melting at alarming rates, exposing fragile hunting implements and biological remains to taphonomic conditions unfavourable to their long-term preservation. As the ice patches melt, caribou dung is exposed in increasing amounts leading to a tipping point where the heat gained from summer insolation surpasses the cooling effect of the ice, causing ice patches to melt rapidly and disappear in a just a few years. 




Patrick E. McSharry, Smith School of Enterprise and the Environment, University of Oxford, Oxford OX1 2BQ, UK.


It may be necessary to take action now to ensure that cultural heritage sites are resilient to climate change over the coming decades. Policymakers, however, are currently struggling to make sense of the forecasts and scenarios arising from of a growing number of climate models. The situation is further complicated by the need to pick from several assumptions about the future emissions of greenhouse gases. There is a substantial challenge to relate these forecasts to particular geographical regions where knowledge of the interactions between environmental variables and socio-economic systems is required. Practical advice is needed to allow policymakers to formulate adequate management plans to adapt to climate change and reduce risks. Accurate utilization of climate models for informing policy is complicated by the inherent uncertainty in the modeling process and diversity of assumptions that can be made about the future rates of anthropogenic forcing (McSharry, 2011). Even a single climate model running an individual assumption will lead to an ensemble of scenarios for the future environmental variables. The state-of-the art approach is to combine multi-models ensembles into a single probabilistic density forecast for each variable. This approach relies on the construction of a meaningful probabilistic forecast, which is statistically reliable (Pinson et al., 2010).

 Damage to cultural heritage sites may be caused by a number of mechanisms such as corrosion, bio-degradation and soiling (Watt et al, 2009). Over the last decade, changes in both the sources and amounts of emissions of air pollution have altered the rate and extent of damage. The threats posed by climate change to natural and cultural sites on UNESCO's World Heritage List have been recently outlined (UNESCO, 2007). In order to understand and evaluate the impacts of climate change, it is convenient to use the approach adopted when constructing catastrophe models for the insurance industry, which is to divide the modeling process into four modules: (1) hazard; (2) exposure; (3) vulnerability; and (4) economic costs (see Figure 1). This technique can be viewed as a systems approach to risk forecasting and offers a means of propagating uncertainty (Orrell and McSharry, 2009). In the following, we discuss each of these modules in turn.

 (1) Hazard: The climate model provides an assessment of the hazard through probabilistic forecasts of relevant environmental variables such as temperature and precipitation at various horizons in the future. Forecasts of air pollution would then be derived from the environmental variables simulated by the climate models.

(2) Exposure: This module relates to the actual spatial locations of the

particular portfolio of heritage sites that are being considered. For example, this could contain the latitudes and longitudes of a number of key monuments in a particular country.

(3) Vulnerability: This module assesses how changes in the environmental variables (measured in the hazard module and downscaled to specific sites in the exposure module) will affect the value of the heritage portfolio. One example is identifying the relationship between environmental variables such as air pollution and the degradation of monuments from erosion and blackening, loss of tourism revenues and the subsequent cost of restoration.

(4) Economic Costs: This module attempts to measure the impacts as

reflected through economic costs such as loss of tourism revenues and

restoration. This is a key module in order to assess the risk, value the costs for future potential impacts, measure their net present value and undertake an evidence-based cost-benefit analysis.


Figure 1: Schematic diagram showing how the hazard, exposure and vulnerability are combined to estimate the risk, which is measured by the economics costs.

Using these four modules, it would be possible to construct a heritage climate index (HCI) for policymakers. The HCI would provide a means of modelling the relationship between the hazard and the potential economic impacts using a parsimonious approach, which aims to deliver an appropriate hybrid technique combining scientific knowledge and existing empirical observations. For example, in a related application, the tourism climate index (TCI) has been used effectively to forecast economic impacts of climate change on tourism (Simpson and McSharry, 2010). This was facilitated by the simplification of encapsulating the complicated nonlinear interactions in the TCI, which is faithful to psychological studies and supported by empirical evidence (Mieczkowski, 1985).


The advantage of having a HCI would be to assess future impacts and to rank and prioritise cultural heritage sites. In addition, the HCI could be deployed to construct spatial geographic maps indicating where the economic impacts of climate change would be most detrimental. The HCI index would have considerable practical applications and would facilitate evidence-based decision-making. In short, the HCI could be used to simplify the overwhelming body of scientific information into a single number.