**3. Monitoring archaeological sites using satellite remote sensing and GIS analysis**

In many areas of the world, cultural heritage sites and visible monuments are monitored mostly with on-site observations, including data collection, periodic observations for ar‐ chaeological sites and multi-analysis risk assessments. In this way, on-site observations are time consuming and not cost-effective.

Hadjimitsis et al., (2011) highlighted the beneficial integrated use of satellite remote sens‐ ing with GIS for exploring the natural and anthropogenic hazard risk of the most signifi‐ cant cultural heritage sites in Cyprus. In order to proceed to overall risk and vulnerability assessment of the archaeological sites in Cyprus due to anthropogenic and natural impact, a risk index was attributed to each different factor such as urban activi‐ ty, minimum distance of urban activity in the vicinity of an archaeological site, seismic PGA and air pollution impact. They found that, concerning the seismic risk assessment, that significant monuments are located within the spatial limits of the most seismic prone areas in Cyprus. Additionally, regarding sea erosion, the study proved that 50% of the sites examined in the study, are within a distance of only 500 m away from the coastline making them vulnerable to related coastal hazards such as sea water erosion. The creation of buffer zones in GIS environment around CH sites explored the signifi‐ cant problem of extensive urbanization in the vicinity of cultural heritage sites. Almost 50% of the CH sites are under severe urban pressure and a percentage of 37.5% of the sites are within a radius of 500m from the urban centers. In similar studies, Carlon et al., (2002) and (Alexakis and Sarris, 2010) used both anthropogenic and natural factors to cre‐ ate a risk assessment model concerning archaeological monuments in Venice and West‐ ern Crete respectively. Moreover, Urhus et al (2006) emphasized the human driven agents, such as camping, hunting and woodcutting, for assessing the modern threats to heritage resources and Lanza (2003) addressed the potential threat that is posed at the historical center of Genoa in the case of failure of the urban drainage system.

and applications of several remote sensing techniques for supporting archaeological re‐ search. The section includes detection of subsurface remains at the Thessalian plain based on both satellite and ground spectroradiometric measurements. Moreover, remote sensing and GIS analysis as means for monitoring purposes in the area of Cyprus are also examined. Geophysical surveys from various archaeological sites are also presented as well as the re‐ sults of a study aiming to analyse the impact of atmospheric pollution on archaeological sites. The section ends with discussion of low-altitude airborne systems, as well as 3D laser

**Figure 1.** Film capsule of the CORONA satellite collected from aircrafts. (Photos from Wikipedia and CSNR collection)

**3. Monitoring archaeological sites using satellite remote sensing and GIS**

In many areas of the world, cultural heritage sites and visible monuments are monitored mostly with on-site observations, including data collection, periodic observations for ar‐ chaeological sites and multi-analysis risk assessments. In this way, on-site observations are

Hadjimitsis et al., (2011) highlighted the beneficial integrated use of satellite remote sens‐ ing with GIS for exploring the natural and anthropogenic hazard risk of the most signifi‐ cant cultural heritage sites in Cyprus. In order to proceed to overall risk and vulnerability assessment of the archaeological sites in Cyprus due to anthropogenic and natural impact, a risk index was attributed to each different factor such as urban activi‐ ty, minimum distance of urban activity in the vicinity of an archaeological site, seismic PGA and air pollution impact. They found that, concerning the seismic risk assessment, that significant monuments are located within the spatial limits of the most seismic prone areas in Cyprus. Additionally, regarding sea erosion, the study proved that 50% of the sites examined in the study, are within a distance of only 500 m away from the coastline making them vulnerable to related coastal hazards such as sea water erosion. The creation of buffer zones in GIS environment around CH sites explored the signifi‐

scanner documentation of cultural heritage site.

64 Remote Sensing of Environment: Integrated Approaches

**analysis**

time consuming and not cost-effective.

This section presents the contribution of remote sensing for monitoring the surroundings of archaeological sites in order the managing authorities or governmental related bodies to be able to conduct a risk assessment analysis of cultural heritage sites in Cyprus. Figure 2 presents some of the most indicative threat parameters. Special attention in this section is given to urban expansion during the past 50 years. Anthropogenic factors, such as urban ex‐ pansion and air pollution contribute significantly to the destruction of cultural heritage sites. Remote sensing and GIS provide synoptic views of cultural heritage sites which enable poli‐ cy makers to make appropriate decisions regarding the preservation of cultural heritage sites.

**Figure 2.** Risk assessment analysis for cultural heritage sites (Hadjimitsis et al., 2011)

#### **3.1. Urban expansion and other hazards as a threat to archaeological sites**

In order to study and map urban expansion, a number of significant archaeological sites of Cyprus were examined. These cultural heritage sites are located in the southern coast‐ al part of the island (from west to east): *Tombs of the Kings, Nea Paphos, Palaepaphos (Old Paphos),* and *Amathus.* Urban expansion was monitored with the extensive use of timeseries multispectral and aerial dataset. All images were both geometrically and radiomet‐ ric corrected in ERDAS Imagine 9.3 software. Moreover, atmospheric correction was also performed based on the Darkest Pixel algorithm (see Hadjimitsis et al., 2009, 2002; Aga‐ piou et al., 2011). Post-processing techniques included histogram enhancement, computa‐ tion of vegetation indices, band ratios, principal component analysis and photointerpretation of the results.

in a GIS environment with the use of ground control points (GCP's). The digitization of all the buildings in the broader area of Nea Paphos and Tombs of the Kings was performed for both time periods. Their direct comparison enabled the researchers to map the extent of ur‐ ban development during the last years and revealed the impact of urbanization on the pres‐

Remote Sensing for Archaeological Applications: Management, Documentation and Monitoring

http://dx.doi.org/10.5772/39306

67

**Figure 4.** Urban expansion near the archaeological sites of *Nea Paphos* and *Tombs of the Kings* during the last 50

CORONA satellite images have also indicated the growth of the urban activity around the *Amathus* archaeological site, including the highway that passes 100 m north of the site (see Figure 5) (Hadjimitsis et al., 2010). Several satellite images were used to examine the threat of urban expansion around the *Amathus* archaeological site located just east from the out‐ skirts of the city (Figure 6). The dataset includes Landsat TM/ETM+ images from 1987 until 2009. As shown in Figure 6, urban expansion is clearly observed though interpretation of the

It is very important for researchers to understand the dramatic changes that have occurred due to human activity during the last decades. Figure 7 highlights the potential risk of the archaeological sites due to urban expansion of the city of Limassol. Using archive satellite images, the researchers can map this expansion with great detail and accuracy based on

ervation of archaeological sites (Figure 4).

years (3D view).

images.

classification techniques.

The results showed a dramatic increase in urban expansion of main cities of Cyprus (Limas‐ sol and Paphos) during the last 50 years. For example, in the case of the *Palaepaphos* site (Figure 3), the entire east area of Kouklia village (*Palaepaphos*) is still undeveloped, while at the west area the urban expansion has been increase dramatically (Agapiou et al., 2010a).

**Figure 3.** Palaepaphos archaeological site in 1963 CORONA image (left) and 2004 QuickBird image (right) (Hadjimitsis et al., 2010)

Urban sprawl has been recorded also in the broader area of Paphos during the last decades. Extensive construction and building development has taken place and areas with significant archaeological interest are now affected from urban expansion. Thus, the land use and land cover region of the area was examined to monitor and map the size of urban expansion in the vicinity of the archaeological sites of Tombs of the Kings and Nea Paphos during the last half century. Aerial photos of the study area, acquired in 1963 and 2008 were provided from the Department of Lands and Surveys of Cyprus. Initially, aerial photos were georeferenced in a GIS environment with the use of ground control points (GCP's). The digitization of all the buildings in the broader area of Nea Paphos and Tombs of the Kings was performed for both time periods. Their direct comparison enabled the researchers to map the extent of ur‐ ban development during the last years and revealed the impact of urbanization on the pres‐ ervation of archaeological sites (Figure 4).

al part of the island (from west to east): *Tombs of the Kings, Nea Paphos, Palaepaphos (Old Paphos),* and *Amathus.* Urban expansion was monitored with the extensive use of timeseries multispectral and aerial dataset. All images were both geometrically and radiomet‐ ric corrected in ERDAS Imagine 9.3 software. Moreover, atmospheric correction was also performed based on the Darkest Pixel algorithm (see Hadjimitsis et al., 2009, 2002; Aga‐ piou et al., 2011). Post-processing techniques included histogram enhancement, computa‐ tion of vegetation indices, band ratios, principal component analysis and photo-

The results showed a dramatic increase in urban expansion of main cities of Cyprus (Limas‐ sol and Paphos) during the last 50 years. For example, in the case of the *Palaepaphos* site (Figure 3), the entire east area of Kouklia village (*Palaepaphos*) is still undeveloped, while at the west area the urban expansion has been increase dramatically (Agapiou et al., 2010a).

**Figure 3.** Palaepaphos archaeological site in 1963 CORONA image (left) and 2004 QuickBird image (right) (Hadjimitsis

Urban sprawl has been recorded also in the broader area of Paphos during the last decades. Extensive construction and building development has taken place and areas with significant archaeological interest are now affected from urban expansion. Thus, the land use and land cover region of the area was examined to monitor and map the size of urban expansion in the vicinity of the archaeological sites of Tombs of the Kings and Nea Paphos during the last half century. Aerial photos of the study area, acquired in 1963 and 2008 were provided from the Department of Lands and Surveys of Cyprus. Initially, aerial photos were georeferenced

interpretation of the results.

66 Remote Sensing of Environment: Integrated Approaches

et al., 2010)

**Figure 4.** Urban expansion near the archaeological sites of *Nea Paphos* and *Tombs of the Kings* during the last 50 years (3D view).

CORONA satellite images have also indicated the growth of the urban activity around the *Amathus* archaeological site, including the highway that passes 100 m north of the site (see Figure 5) (Hadjimitsis et al., 2010). Several satellite images were used to examine the threat of urban expansion around the *Amathus* archaeological site located just east from the out‐ skirts of the city (Figure 6). The dataset includes Landsat TM/ETM+ images from 1987 until 2009. As shown in Figure 6, urban expansion is clearly observed though interpretation of the images.

It is very important for researchers to understand the dramatic changes that have occurred due to human activity during the last decades. Figure 7 highlights the potential risk of the archaeological sites due to urban expansion of the city of Limassol. Using archive satellite images, the researchers can map this expansion with great detail and accuracy based on classification techniques.

**Figure 7.** Urban areas of Limassol town in 1987 (red) and in 2009 (pink). The *Amathus* archaeological site is indicated

Remote Sensing for Archaeological Applications: Management, Documentation and Monitoring

http://dx.doi.org/10.5772/39306

69

Vegetation indices are also a key parameters that can be used for monitoring dramatic land use changes over time (e.g. urban activities). The Normalized Difference Vegetation Index (NDVI, with range -1 to +1) was applied to the entire dataset (Figure 8). High values of NDVI (indicated with green in Figure 8) are present vegetated areas while low NDVI values (indicated with yellow) are recorded for areas with no vegetation. Since NDVI values may vary throughout time due to the physical phenological changes of the plants, similar periods

NDVI values were used along with classifications results in order to record NDVI differen‐ ces in urban classified areas. Figure 9 demonstrates the results of the NDVI difference for the period 1987-2009. Although many areas have indicated no dramatic changes, some other areas represented in yellow and red colour (Figure 9) indicate dramatic transformation of the initial landscape. Indeed, such changes have been recorded in a very close proximity of

Further anthropogenic and natural hazards (e.g. landslides; sea erosion; earthquakes etc) can be monitored in a systematic basis using remote sensing data and GIS spatial analysis. Different studies (Hadjimitsis et al., 2010; 2011) have shown the potential of using such

Contemporary technological means such as GIS and satellite remote sensing provide effi‐ cient and detailed maps of the region of CH sites in the island of Cyprus. This specific study revealed the different kinds of natural and anthropogenic hazards that threaten the preser‐

the archaeological site of *Amathus* (see Figure 9 in black square).

methodologies for cultural heritage risk assessment.

in a square.

of Landsat images were examined.

vation of valuable CH sites.

**Figure 5.** Amathus archaeological site in 1963 CORONA image (left) and 2010 Google (right).

**Figure 6.** Landsat images used for mapping the urban expansion of Limassol town during the last 30 years. *Amathus* archaeological site is indicated in a square.

**Figure 7.** Urban areas of Limassol town in 1987 (red) and in 2009 (pink). The *Amathus* archaeological site is indicated in a square.

Vegetation indices are also a key parameters that can be used for monitoring dramatic land use changes over time (e.g. urban activities). The Normalized Difference Vegetation Index (NDVI, with range -1 to +1) was applied to the entire dataset (Figure 8). High values of NDVI (indicated with green in Figure 8) are present vegetated areas while low NDVI values (indicated with yellow) are recorded for areas with no vegetation. Since NDVI values may vary throughout time due to the physical phenological changes of the plants, similar periods of Landsat images were examined.

**Figure 5.** Amathus archaeological site in 1963 CORONA image (left) and 2010 Google (right).

68 Remote Sensing of Environment: Integrated Approaches

**Figure 6.** Landsat images used for mapping the urban expansion of Limassol town during the last 30 years. *Amathus*

archaeological site is indicated in a square.

NDVI values were used along with classifications results in order to record NDVI differen‐ ces in urban classified areas. Figure 9 demonstrates the results of the NDVI difference for the period 1987-2009. Although many areas have indicated no dramatic changes, some other areas represented in yellow and red colour (Figure 9) indicate dramatic transformation of the initial landscape. Indeed, such changes have been recorded in a very close proximity of the archaeological site of *Amathus* (see Figure 9 in black square).

Further anthropogenic and natural hazards (e.g. landslides; sea erosion; earthquakes etc) can be monitored in a systematic basis using remote sensing data and GIS spatial analysis. Different studies (Hadjimitsis et al., 2010; 2011) have shown the potential of using such methodologies for cultural heritage risk assessment.

Contemporary technological means such as GIS and satellite remote sensing provide effi‐ cient and detailed maps of the region of CH sites in the island of Cyprus. This specific study revealed the different kinds of natural and anthropogenic hazards that threaten the preser‐ vation of valuable CH sites.

**3.2. Monitoring air quality in the vicinity of archaeological sites based on satellite and**

der also to cross-validate the AOT values found from MODIS and sun-photometers.

preservation and maintenance of cultural heritage sites.

Spectral variations recorded by satellite sensors are indicators of aerosol particles and, there‐ fore, air pollution. The key parameter for assessing atmospheric pollution in air pollution stud‐ ies is the aerosol optical thickness. Aerosol optical thickness (AOT) is a measure of aerosol loading in the atmosphere (Retalis et al., 2010). High AOT values suggest high concentration of aerosols, and therefore air pollution (Retalis et al, 2010). The use of earth observation is based on the monitoring and determination of AOT either direct or indirect as tool for assessing and measure air pollution. Several studies have shown that satellite data can be used to monitor air pollution and air pollution effects. Tømmervik et al., (1995) compared vegetation cover maps and air pollution emissions data over a 15 year period and found major changes in the environ‐ ment as a result of high air pollution values. Nisantzi et al., (2011) used MODIS satellite data to analyse the relationship between the aerosol optical thickness (AOT) and the PM10 as indica‐ tors of pollution. Satellite remote sensing can be used to assist in air quality monitoring and identify the need to protect cultural heritage in urban areas from air pollution (Hadjimitsis et al., 2002; Kaufman et al, 1990; Retalis, 1998; Retalis et al., 1999). Pollution not only deteriorates cultural heritage sites but may also cause irreversible damage that prevents the proper salva‐ tion of the monument (Skoulikides, 2000). Therefore, improving air quality is critical for the

The study area was the Limassol Castle, located in the center of Limassol, Cyprus. The study utilized a variety of remote sensing tools to measure air pollution. Landsat TM/ETM+ and MODIS satellite images, as well as the GER 1500 spectro-radiometer, were used to directly or indirectly retrieve AOT, as were ground measurements using the Microtops II handheld sun‐ photometer and the Cimel sun-photometer located at the Cyprus University of Technology, which is part of the AERONET program. Air particles' measurements were correlated to the AOT levels to verify the level of pollution. Last, visual observation of the Limassol Castle iden‐ tified the damage caused by air pollution and laser scanning to document and monitor the damage was conducted. Results from satellite remote sensing identified that the centre of Li‐ massol contains high levels of air pollution, with values of AOT higher than other surrounding areas. Determination of AOT measurements using MODIS and Landsat satellite images found that the centre of Limassol, where the Limassol Castle is located, experiences the highest level of AOT values (Figure 10). A PM10 /PM2.5 in situ measurement campaign in the area of the Li‐

Although cultural heritage sites are documented and preserved, there has been limited monitoring and documentation of how cultural heritage sites are affected by air pollution. Themistocleous et al., (2012a) introduced a new approach for monitoring air pollution near cultural heritage sites. By using a variety of tools, including satellite images, sun-photome‐ ters, PM10 monitors, and laser scanners, the level of air pollution and its effect on cultural heritage sites can be determined. The cultural heritage sites were documented, and using GIS tool, any significant areas of air pollution, including urban areas, industrial areas, and roads were determined. The algorithm proposed by Themistocleous (2011) was applied to retrieve the aerosol optical thickness (AOT) from Landsat TM/ETM+ satellite images in or‐

Remote Sensing for Archaeological Applications: Management, Documentation and Monitoring

http://dx.doi.org/10.5772/39306

71

**ground measurements**

**Figure 8.** NDVI maps produced from Landsat dataset.

**Figure 9.** NDVI difference from 1987 until 2009.

#### **3.2. Monitoring air quality in the vicinity of archaeological sites based on satellite and ground measurements**

Although cultural heritage sites are documented and preserved, there has been limited monitoring and documentation of how cultural heritage sites are affected by air pollution. Themistocleous et al., (2012a) introduced a new approach for monitoring air pollution near cultural heritage sites. By using a variety of tools, including satellite images, sun-photome‐ ters, PM10 monitors, and laser scanners, the level of air pollution and its effect on cultural heritage sites can be determined. The cultural heritage sites were documented, and using GIS tool, any significant areas of air pollution, including urban areas, industrial areas, and roads were determined. The algorithm proposed by Themistocleous (2011) was applied to retrieve the aerosol optical thickness (AOT) from Landsat TM/ETM+ satellite images in or‐ der also to cross-validate the AOT values found from MODIS and sun-photometers.

Spectral variations recorded by satellite sensors are indicators of aerosol particles and, there‐ fore, air pollution. The key parameter for assessing atmospheric pollution in air pollution stud‐ ies is the aerosol optical thickness. Aerosol optical thickness (AOT) is a measure of aerosol loading in the atmosphere (Retalis et al., 2010). High AOT values suggest high concentration of aerosols, and therefore air pollution (Retalis et al, 2010). The use of earth observation is based on the monitoring and determination of AOT either direct or indirect as tool for assessing and measure air pollution. Several studies have shown that satellite data can be used to monitor air pollution and air pollution effects. Tømmervik et al., (1995) compared vegetation cover maps and air pollution emissions data over a 15 year period and found major changes in the environ‐ ment as a result of high air pollution values. Nisantzi et al., (2011) used MODIS satellite data to analyse the relationship between the aerosol optical thickness (AOT) and the PM10 as indica‐ tors of pollution. Satellite remote sensing can be used to assist in air quality monitoring and identify the need to protect cultural heritage in urban areas from air pollution (Hadjimitsis et al., 2002; Kaufman et al, 1990; Retalis, 1998; Retalis et al., 1999). Pollution not only deteriorates cultural heritage sites but may also cause irreversible damage that prevents the proper salva‐ tion of the monument (Skoulikides, 2000). Therefore, improving air quality is critical for the preservation and maintenance of cultural heritage sites.

**Figure 8.** NDVI maps produced from Landsat dataset.

70 Remote Sensing of Environment: Integrated Approaches

**Figure 9.** NDVI difference from 1987 until 2009.

The study area was the Limassol Castle, located in the center of Limassol, Cyprus. The study utilized a variety of remote sensing tools to measure air pollution. Landsat TM/ETM+ and MODIS satellite images, as well as the GER 1500 spectro-radiometer, were used to directly or indirectly retrieve AOT, as were ground measurements using the Microtops II handheld sun‐ photometer and the Cimel sun-photometer located at the Cyprus University of Technology, which is part of the AERONET program. Air particles' measurements were correlated to the AOT levels to verify the level of pollution. Last, visual observation of the Limassol Castle iden‐ tified the damage caused by air pollution and laser scanning to document and monitor the damage was conducted. Results from satellite remote sensing identified that the centre of Li‐ massol contains high levels of air pollution, with values of AOT higher than other surrounding areas. Determination of AOT measurements using MODIS and Landsat satellite images found that the centre of Limassol, where the Limassol Castle is located, experiences the highest level of AOT values (Figure 10). A PM10 /PM2.5 in situ measurement campaign in the area of the Li‐ massol Castle found that for the majority of the time periods, the PM10 readings exceeded the limit value (50 μg/m3), indicating a high level of air pollution in the area.

**4. Detection of archaeological sites based on remote sensing techniques**

Remote Sensing for Archaeological Applications: Management, Documentation and Monitoring

http://dx.doi.org/10.5772/39306

73

Several Neolithic settlements ("magoules") are located in the Thessalian plain in central Greece. These sites are typically found as low hills raised up to 5-10 m. Alexakis et al., (2009; 2011) has recently shown that the detection of several unknown sites is possible based on remote sensing and GIS analysis. The study aimed to combine several types of remote sens‐ ing data (e.g. Landsat TM/ETM+, ASTER, Hyperion, IKONOS) and DEM in order to im‐ prove the detection of these subsurface remains (Figure 12). The satellite data were statistically analyzed, together with other environmental parameters, to examine any kind of correlation between environmental, archaeological and satellite data. Moreover, different methods were compared for the detection of Neolithic settlements. The results of the study suggested that the complementary use of different imagery can provide more satisfactory

Further to the Alexakis study, Agapiou et al., (2012a) argued that the detection of the settle‐ ments is possible based on ground spectroradiometric measurements. Several spectroradio‐ metric measurements have indicated that each magoula has its own spectral characteristics related to its own morphological characteristics. The study has found that the highest peak of the magoula tends to give high NDVI and SR values (similar to the flat – healthy regions) while the slope of the magoula has lowest NDVI and SR values (and for the other indices as well). The extraction of each magoula requires further analysis and enhancement techniques in cases where the spatial resolution of the satellite image used is low. Local histogram en‐ hancements can identify magoules as a small difference of NDVI values at the same parcel

**Figure 12.** Magoula *Neraida* using ASTER image (left). Magoula *Melissa 1* using IKONOS image (RGB - 321) (right).

tion); band 3: wetness (interrelationship of soil and canopy moisture).

Similar results were found following the application of the Tasselled Cap algorithm (Figure 14 to a series of Landsat TM/ETM+ multispectral images. The Tasselled Cap transformation is used to enhance spectral information for Landsat images, and it was specially developed for vegetation studies. The first three bands of the Tasseled Cap algorithm result are charac‐ terized as follow: band 1: brightness (measure of soil); band 2: greenness (measure of vegeta‐

results.

(Figure 13).

**Figure 10.** AOT levels in the Limassol area. High AOT levels are noted in the area near the Limassol Castle.

A similar approach was followed for the Paphos town using daily MODIS AOT data. The re‐ sults have shown that 54% of the measurements for air quality was above the threshold of AOT 300 (AOT 0.300) (see Figure 11). This analysis suggest that cultural heritage sites near the Pa‐ phos town (e.g. *Nea Paphos, Tombs of the Kings* etc) are exposed to air pollutants half the time.

**Figure 11.** Paphos AOT values (sample = 109 measurements) in blue. In red circle is the threshold air quality limit of 300 (AOT 0.300). In the y-axis, AOT value is multiplied by 1000 (to match MODIS data) (Themistocleous et al., 2012a).
