TY - JOUR
AU - Quarta,, Tatiana
AB - Abstract Non-invasive geophysical investigations are usually the only way to gain information on subsurface properties that can affect the stability of historical structures and accelerate degradation processes. A combined multi-frequency ground-penetrating radar (GPR) geoelectrical and induced polarization (IP) survey was performed in the cloister of ‘Palazzo dei Celestini’, Lecce, southern Italy, in order to investigate possible subsurface causes of deterioration. The historical palace was originally a convent connected to the Basilica of ‘Santa Croce’ and is now the head office of the Province of Lecce Administration and the Prefecture. Built in Pietra Leccese, a fine-grained calcarenite, Santa Croce and Palazzo dei Celestini is the most famous baroque architectural complex of the historical centre of Lecce. The high capillarity of the building material causes deterioration problems especially at some altars of the church and in the lower portion of the walls and pillars of the monumental building. The integrated geophysical survey yielded a detailed description of the shallow stratigraphical and hydro-geological setting of the area and an accurate location of ancient and modern drainage systems. The geophysical information was essential for identifying natural or anthropogenic causes of the local increase in subsoil moisture that could accelerate the degradation process and for developing effective remediation activities. cultural heritage, urban and engineering geophysics, GPR, ERT, IP 1 Introduction Geophysical methods are becoming increasingly popular in civil engineering and cultural heritage monitoring as a cost-effective, non-invasive way of obtaining information on the shallow subsurface and the interior of monuments and architectural structures. Ground-penetrating radar (GPR) and electrical resistivity tomography (ERT) are frequently used in combination as complementary techniques to gain information on the electrical properties of the earth and building materials at high (hundreds of MHz to GHz) and low (<1000 Hz) frequencies of the electromagnetic spectrum. GPR is the preferred prospecting method in resistive non-magnetic environments as under these conditions it can provide a resolution by far greater than other geophysical methods (Davis and Annan 1989, Basile et al2000, Conyers 2004). In addition, through the use of advanced processing and visualization methods (Nuzzo et al2002, 2009), GPR can provide effective means of communicating the relevant information on subsoil and building properties to the people involved in environmental and architectural management (archaeologists, architects, administrative and political bodies). Numerous studies have proven the usefulness of this technique to evidence geological, hydro-geological and structural properties, moisture content and contaminant extent as well as the presence of faults and defects in masonry structures (Leucci et al2002, Nuzzo et al2007, Masini et al2007). The integration of ERT can help in overcoming the limited penetration of GPR signals in moderately conductive soils and high levels of clutter in strongly heterogeneous terrains, such as urban and archaeological environments (Gaffney et al2004, Cardarelli et al2008, Soupios et al2008). Non-destructive geophysical methods are incomparable where direct investigations (coring) are impossible, such as inside standing historical buildings. In these situations GPR is usually the easiest and fastest methodology (Nuzzo 2004, Cataldo et al2005, 2009), although occasionally low-invasive ERT techniques, using base-plate electrodes or thin nails, have also been attempted (Tsokas et al2008, Nuzzo and Quarta 2005, 2009). Recent advances in multielectrode instrumentation and inversion software (Oldenburg and Li 1994, Loke and Barker 1996) make combined (dc) resistivity and induced polarization (IP) imaging feasible for improved subsurface characterization. IP measurements are sensitive to the low-frequency capacitive properties of rocks and soils (Olhoeft 1985), which are controlled by different diffusion polarization mechanisms operating at the grain–fluid interface. Since the interpretation of IP data is more difficult than resistivity, the IP method is not common in complex urban contexts. Recent studies on environmental applications have shown that the complementary information provided by recovered conduction and polarization properties helps to discriminate between the effect of lithology and fluid conductivity. In the present study we describe a combined GPR, ERT and IP survey performed in June–July 2004 in an urban setting to achieve a comprehensive characterization of the shallow subsurface for cultural heritage monitoring and conservation. Our results demonstrate that, despite the logistical problems that need to be overcome in such environments, this integrated strategy has strong potential for discrimination between internal (subsurface) and external (atmospheric) degradation sources and helps to identify the main processes of decay, which are essential for the development of effective remediation actions. 2 Geological and historical background 2.1 Geological and hydro-geological setting of Salento The study area is located in Lecce, the chief town of Salento, namely the peninsular part of the Apulia region, southern Italy (the inset in figure 1). The geological setting of Salento comprises a Mesozoic carbonate sequence overlain by thin deposits of Paleogene, Neogene and Quaternary age. The persistence of a shallow water environment during the Jurassic and Cretaceous periods caused the deposition of a thick carbonate sequence (up to 6000 m), which constitutes the bedrock of the region. The Paleogene was characterized by compressive tectonic phases which deformed these deposits creating a horst-and-graben structure. During the Miocene a marine transgression allowed the deposition of biomicritic sediments, composed mainly of planktonic and secondarily glauconitic Foraminifera, known as the Pietra Leccese formation. Subsequent sedimentation cycles determined the formation of Plio-Pleistocenic (bio)calcarenite, sandstone, clay, silt and sand deposits. Miocene limestone and Miocene fine calcarenites outcrop in the study area. The Salento peninsula is a karstic landscape marked by a wide, deep aquifer within the Mesozoic limestone, which lies on seawater intruded from the nearby coastal area. Additional shallow groundwater bodies can be found locally in the most recent deposits (Margiotta 1994, Margiotta and Negri 2005, Nuzzo et al2007). Figure 1 Open in new tabDownload slide (a) Aerial photograph of the historical centre of Lecce, Apulia, Southern Italy (inset). (b) Facade of ‘Basilica di Santa Croce’. (c) Baroque façade of ‘Palazzo dei Celestini’ along Via Umberto I. Figure 1 Open in new tabDownload slide (a) Aerial photograph of the historical centre of Lecce, Apulia, Southern Italy (inset). (b) Facade of ‘Basilica di Santa Croce’. (c) Baroque façade of ‘Palazzo dei Celestini’ along Via Umberto I. 2.2 History of the complex ‘Santa Croce e Palazzo dei Celestini’ (Lecce, Italy) The Basilica of ‘Santa Croce’ with the annexed former Celestines' Convent ‘Palazzo dei Celestini’ is the most extraordinary example of the architectural Baroque style known as ‘Barocco Leccese’ (Colangeli 1978). The convent was commissioned by Walter VI of Brienne, Count of Lecce and Duke of Athens, and its construction began in 1549. The internal part of the building was completed in 1659. The first order of the monumental façade along ‘Via Umberto I’ was built between 1659 and 1688 by the famous local architect Giuseppe Zimbalo, whereas the second order was built by his disciple Giuseppe Cino during 1688–1695 (figure 1(c)). More recent is the façade in front of Via XXV Luglio built during 1817–1842 by architect Maiola. To the Renaissance taste belongs the central square cloister (figure 1(a)) designed by the architect Gabriele Riccardi. He also built the interior of the Basilica, whose simplicity and equilibrium counterbalances the decorative richness of the façade designed by Giuseppe Zimbalo (figure 1(b)). The first level of its façade up to the magnificent balustrade was completed in 1582 by Riccardi (apart from the three portals built in 1606 by Francesco Antonio Zimbalo), whereas the second and third levels were completed in 1646 by the architect and sculptor Giuseppe Zimbalo, grandson of the former, including the finely decorated rose window realized by Cesare Penna (figure 1(b)). Due to the different building periods, while a more sober 16th-century style characterizes the lower part of the façade, a more sumptuous baroque style characterizes its upper part. The church dome was built in 1590. The luxurious richness of the Barocco Leccese style is attributable to the specific properties of the Pietra Leccese calcarenite, which make the stone very easy to cut and carve. Unfortunately, its porosity makes it prone to degradation. 2.3 Deterioration processes affecting Pietra Leccese Pietra Leccese is a calcarenite—calcilutite mainly consisting of a very fine calcium carbonate matrix with phosphates and glauconitic micro-grains whose relative abundance determines the variation in colour from white-yellowish to light brown or greenish. Sound rocks have a degree of compactness between 0.54 and 0.69, porosity of 0.30–0.45 and an imbibition capacity of 11–21% (referred to the weight). Pietra Leccese has scarce resistance to atmospheric weathering. Degradation manifests itself in reduced compactness and increased porosity and imbibition capacity (Margiotta 1994). Three types of deterioration processes may affect this stone: physical, chemical and biological. To the first type belongs the action of water and wind. Pietra Leccese has strong aptitude towards absorbing and retaining water, either in the form of rainwater, atmospheric moisture and condensed vapour or as capillary water rising from the subsoil, especially along micro-fractures and zones of structural weakness (figures 1 and 2). This high hygroscopic power can remarkably reduce its resistance to compression and induce crushing and exfoliation. The abrasive action of the particles transported by the wind causes the weak exterior zones of the stone to crumble generating an unpleasant corroded aspect (figure 2(e)). As a result of the joint action of these factors, the stone loses the lithoid mechanical properties and assumes the consistency of weakly cemented calcareous sand. Also indoor environments, like the interior of a church, can have particular microclimatic conditions favourable for the development of deterioration phenomena (Turnone 2002, Carrozzo et al2003). Figure 2 Open in new tabDownload slide Example of deterioration problems affecting structures in Pietra Leccese. Figure 2 Open in new tabDownload slide Example of deterioration problems affecting structures in Pietra Leccese. Atmospheric pollution, especially in urban areas, can contribute to the degradation of the stone via the chemical aggression of the impurities contained in the air, such as nitrates, sulfur and chlorine (ICOM 1983). For example, sulfur dioxide (SO2), which is a product of petroleum combustion, reacts with atmospheric water and oxygen to produce sulfuric acid (H2SO4), a component of acid rain, responsible for lowering the pH of soil and freshwater bodies as well as causing chemical weathering of statues and structures. The aggression of the sulfuric acid can occur even in the absence of rain: after sunset the stone irradiates and cools more rapidly than the air nearby attracting a flux of humid warm air towards its surface where the water condenses; if SO2 is present in the air it reacts with the condensed water, giving H2SO4; the latter reacts with the stone (mainly constituted by CaCO3, insoluble) producing gypsum (an easily soluble sulfate) and carbon dioxide, according to the reaction The salts dissolved within the capillary water rising from the subsoil cause the development of saline crusts and efflorescence (figures 2(a), (c)). This degradation phenomenon is particularly noticeable in the lower part of the façade of Palazzo dei Celestini and Santa Croce as well as on some of the church lateral altars. Previous investigations have demonstrated the presence of anions of chlorides, nitrates and sulfates and cations of calcium, sodium, potassium and magnesium (CNR-ISCOM 2000). The different salt distributions in the basilica's left aisle (mainly nitrates and sulfates of calcium, sodium and potassium) and right aisle (mainly sodium chloride and in part nitrates) have suggested two different sources. The prevalence of sodium chloride in the right aisle could be due to localized exceptional provision of underground brackish water, whereas the presence of nitrates on both aisles could be ascribed to the rising of capillary water rich in substances deriving from the decomposition of organic material. Sulfates could derive from localized reaction phenomena due to candle lighting. More recent researches using energy-dispersive x-ray fluorescence (EDXRF) analysis (Turnone 2002, Carrozzo et al2003) assessed the presence of sulfur and chlorine on the altars of both aisles. The EDXRF results showed a slight decrease in sulfur concentration with distance from the front doors in the left aisle (from 0.5 to 0.4%), where the door is open all year long, and a sharper reduction in the right aisle (from 0.5 to <0.1%), where the door is open only during summer and was closed at the time of the measurements. The good correlation between the observed high sulfur concentration and high temperature gradients near the left entrance door supported the hypothesis of a strong flux of incoming polluted air from outside where the sulfur concentration was about 3%. The higher chlorine concentration on the altars of the right aisle, the most damaged ones, than that on those of the left aisle confirmed the previous results and was interpreted as different stone quarry provenance or local increase of near-surface water content. Biological attack from bacteria, some algae, fungi and especially lichens (figure 2(d)) can accelerate the weathering of stone material both via the mechanical stresses induced by the ramification of the thallus in the stone substrate and the alternation of expansions and contractions induced by its retention and loss of rainwater and via the chemical attack by the lichenic acids. In the case of Pietra Leccese, the most common types are crustose lichens (growing as patinas especially on the surfaces exposed to the northern winds) and squamulose lichens (developing mainly inside the masonry joints). The deterioration effects induced by lichens (crumbling) are generally very slow and some researchers suggest not removing them as their patinas can even have a protective and hardening function (Cassiano and Cazzato 1997). Because of the tenderness of Pietra Leccese, past restoration experiences have shown the inadequacy of common cleaning and consolidation treatments (such as synthetic resins and water repellents) traditionally used for other kinds of monumental stones (Margiotta 1994). The difficulty in finding adequate techniques to reduce the degradation effects enhances the importance of studies to understand their causes and possibly mitigate them. 3 Geophysical data acquisition and processing 3.1 Ground penetrating radar Figure 3 shows the layout of the geophysical survey superimposed on the map of the (36 by 36) m2 cloister of Palazzo dei Celestini. Additionally a topographical survey using a total station was performed in order to position all the manhole covers, anomalous paving slabs and stone grids presumably related to the presence of modern utility networks (telephone, gas, etc) mainly located in the portico and modern or ancient water drainage systems located both inside the portico and the cloister central area (small boxes in figure 3). A GSSI Sir System2 equipped with 100, 200 and 500 MHz antennas was used for an extensive GPR survey inside the cloister (figure 3). The low penetration depth achieved along test profiles even using lower-frequency antennas (35 MHz) suggested the presence of shallow conductive layers causing heavy absorption of the electromagnetic energy. The survey grid consists of fifteen 5 m spaced 100 MHz profiles along two orthogonal directions (approximately W-E and N-S), 34 longitudinal and 7 transversal 1 m spaced 200 MHz profiles and five longitudinal 5 m spaced 500 MHz profiles (figure 3). Data were acquired in continuous mode within decreasing time windows (200, 80 and 60 ns for the 100, 200 and 500 MHz, respectively) and using a manual time-varying gain function. Figure 3 Open in new tabDownload slide Map of Palazzo dei Celestini and Basilica di Santa Croce with the layout of GPR, ERT and IP survey. The position of the most damaged altars (A, O and F) and that of a post-acquisition excavation (E) are also shown. Figure 3 Open in new tabDownload slide Map of Palazzo dei Celestini and Basilica di Santa Croce with the layout of GPR, ERT and IP survey. The position of the most damaged altars (A, O and F) and that of a post-acquisition excavation (E) are also shown. The data were subsequently processed using standard 2D techniques by means of the ReflexW software (Sandmeier 2008) according to the following processing sequence: horizontal scaling, by data interpolation to a 2.5 cm sampling interval along the inline direction; header editing for inserting the geometrical information; gain adjustment to enhance the visibility of deeper anomalies (time-varying but laterally constant); spectral analysis and band-pass frequency filtering; adapted background removal to attenuate the horizontal ringing and enhance localized anomalies without destroying the continuity of nearly horizontal stratigraphical interfaces; velocity analysis by hyperbola fitting; Kirchhoff migration, using a constant average velocity value of 0.072 m ns-1. After application of the Hilbert transform to extract the signal envelope, the migrated data were merged together into 3D volumes and visualized in various ways in order to enhance the spatial correlations of interesting anomalies. 3.2 Electrical resistivity tomography and induced polarization A 48-channel geo-resistivity meter Syscal R1 (Iris Instruments) was used to acquire the ERT and IP data along two profiles coinciding with the GPR lines P4 and P6 (figure 3). A low-invasive acquisition technique was adopted using as electrodes 47 small iron nails infixed in the joints between the limestone slabs at 0.75 m intervals (figure 4) and tap water to reduce coupling problems. The dipole–dipole configuration was chosen for the ERT and IP acquisition to enhance shallow lateral variations related to anthropogenic features (draining network and possible archaeological remains), whereas the Wenner–Schlumberger configuration was used to image deeper hydro-geological structures. Figure 4 Open in new tabDownload slide Photograph of the ERT line P4 inside Palazzo dei Celestini (a) and close-up of the geoelectrical instrumentation and low-invasive electrodes (b). Figure 4 Open in new tabDownload slide Photograph of the ERT line P4 inside Palazzo dei Celestini (a) and close-up of the geoelectrical instrumentation and low-invasive electrodes (b). The data were processed using a 2D inversion algorithm by means of the RES2DINV software (Loke 2004). In this method, the subsurface is divided into a number of rectangular cells of constant resistivity. The resistivity of each cell is then evaluated by minimizing the difference between observed and calculated data in an iterative manner. In the smoothness-constrained approach, the sum of the squares of the data misfit is minimized while the sum of the absolute values of the data misfit is minimized in the robust approach, which is less sensitive to outliers (Loke and Barker 1996, Loke et al2003). In our case the robust inversion was used to obtain the resistivity and chargeability 2D model sections, although good results were also obtained with the smoothness-constrained method after filtering bad data due to poor electrode contact. 4 Data interpretation 4.1 Ground penetrating radar The 100 MHz antennas were used for a reconnaissance survey as their low resolution and only slightly deeper penetration suggested that these antennas were not suitable for the aims of the study. Therefore, only the results from the 1 m grid 200 MHz survey are presented in this paper. A comparison of 200 and 500 MHz profiles along the same lines surveyed with the electrical methods (P4 and P6) is shown in figure 5. Figure 5 Open in new tabDownload slide Comparison of un-migrated GPR sections acquired with two antennas along two parallel profiles inside Palazzo dei Celestini. The location of pipes (arrows), a trench (T) and possible archaeological remains (A1 and A2) superimposed over hydro-stratigraphical interfaces (white lines) is shown. Figure 5 Open in new tabDownload slide Comparison of un-migrated GPR sections acquired with two antennas along two parallel profiles inside Palazzo dei Celestini. The location of pipes (arrows), a trench (T) and possible archaeological remains (A1 and A2) superimposed over hydro-stratigraphical interfaces (white lines) is shown. Although the 500 MHz provided higher resolution of the shallowest layers, it was more heavily affected by clutter and attenuation, suggesting that the 200 MHz was the best compromise for obtaining good penetration depth and reflector continuity. Numerous high-amplitude hyperbolas (arrows in figure 5) characterize both profiles and are interpreted as pipes of the modern or ancient drainage networks of rainwater. One of them, located at a distance of 16 m and a depth of 1.5 m on profile P6, seems to be located at the base of a trench-like structure (T in figure 5) whose walls are at a 0.8 m depth. Other horizontal or vertical features (such as the flat reflection A1 at a 25–27 m distance and a 1 m depth on profile P4, and the vertical structure A2 at a 21 m distance and a 0.4–1.4 m depth on profile P6) could be related to archaeological remains. This hypothesis is supported by the finding of walls and channel-like features, probably related to a Roman or Medieval oil mill, recorded at comparable depths in an archaeological excavation less than 100 m from the study site. Although marked by discontinuous, low-amplitude reflection surfaces (white lines in figure 5), one can recognize the presence of a resistive, highly scattering shallow layer less than 0.5 m deep followed by a relatively more attenuating intermediate layer down to about 1 m depth and another resistive layer down to 1.5–2.5 m depth. A presumably highly conductive layer prevents further penetration of the electromagnetic energy. These layers are likely different hydro-geologic units. Signal penetration does not exceed 2.5 m in both profiles. The eastern part of profile P6 shows better signal propagation than the western part, whereas P4 shows greater lateral uniformity. The analysis of consecutive 2D radar sections allowed a preliminary estimation of the orientation and depths of the buried pipes. While in some cases their depth was nearly constant, in other cases they showed a noticeable inclination towards the north. However, only the generation of time slices and data volumes allowed the reconstruction of their 3D spatial development. Figure 6 shows some representative (migrated) slices obtained after interpolating onto a 0.10 by 0.25 m grid the absolute data amplitude averaged within partially overlapping time windows. The first three slices of figure 6 were generated using a constant thickness of 15 ns, corresponding to 0.54 m or 1.5 wavelengths. Four SN and two SE-NW shallow pipes can be easily recognized in the first slice as well as a very strong anomaly close to the southern side of the cloister probably related to an EW pipe or pipes junction. Other remarkable features are a central circular anomaly interpreted as the buried remains of a water well (W), as seen in other cloisters, two nearly parallel features interpreted as a buried trench (T) probably related to the well and a localized anomaly of possible archaeological meaning (A). Deeper pipes and further possible archaeological structures (A1 and A2) are visible in the second and third slices. It is worth noting that, due to its slope and considerable depth extent, only the initial part of the westernmost pipe is visible in the second slice while its final part is visible in the third slice. In the third slice, a pipe is visible whose initial part is housed inside the gap of the trench T of the first slice, while close to the well it turns at a straight angle to join the final part of the main sloping pipe. The time slice representation also shows a globally higher reflectivity (or lower attenuation) in the East half of the cloister. A fourth slice has been created using a thicker time window in order to image all the features at a depth greater than 0.5 m. Despite the reduced amplitude caused by the average process within a longer window, all the deep features are visible including the main sloping pipe in its whole length. Figure 6 Open in new tabDownload slide Selected GPR time slices with location of shallow and deep drainage pipes (dashed lines), a buried trench (T), well remains (W) and other possible archaeological structures (A, A1 and A2). Figure 6 Open in new tabDownload slide Selected GPR time slices with location of shallow and deep drainage pipes (dashed lines), a buried trench (T), well remains (W) and other possible archaeological structures (A, A1 and A2). Although time slices are very useful for imaging the main anomalies in map views while reducing the clutter through time averaging, thus allowing the recognition of their planimetric connection to surface evidences such as manholes and stone covers (small boxes in figure 6), only 3D volumes allow us to appreciate the 3D spatial development of the structures. For the case at hand, an even better way was as 4D representations, i.e. as animated cubes. Snapshots of the animated volumes (provided as supplementary mpg and wmv files) are shown in figure 7. By means of animation one can visually appreciate the inclination of gently or steeply sloping pipes as reflectivity anomalies ‘flowing’ from south to north (or at the right angle) with progressing depth (figure 7(a)). Other effective animations include iso-surface anomaly volumes (figures 7(b) and (c)) that can be ‘stripped’ of low-medium energy clutter by raising the threshold to highlight only the most energetic features (highly-reflecting pipes, water well and presumable archaeological remains) or rotated around a user-defined axis in order to identify their mutual spatial relationships. Figure 7 Open in new tabDownload slide Snapshots of 3D animated volumes showing the progressing of high-amplitude anomalies at increasing depth (a) or the iso-surface visualization of the most reflecting bodies (pipes and archaeological structures) using different thresholds (b) and view angles (c). Attached files type and size: MPEG (2.68 MB) and WMV (0.36 and 0.37 MB). Figure 7 Open in new tabDownload slide Snapshots of 3D animated volumes showing the progressing of high-amplitude anomalies at increasing depth (a) or the iso-surface visualization of the most reflecting bodies (pipes and archaeological structures) using different thresholds (b) and view angles (c). Attached files type and size: MPEG (2.68 MB) and WMV (0.36 and 0.37 MB). 4.2 Electrical resistivity tomography and induced polarization Figure 8 shows the resistivity and IP model sections obtained through the robust inversion of the experimental pseudo-sections along profiles P4 and P6 for two different electrode configurations. The dipole–dipole configuration is mainly sensitive to lateral changes and allows a more detailed characterization of the shallow subsurface, whereas the Wenner–Schlumberger configuration allows the reconstruction of stratigraphical or hydro-geological variations at greater depths. Since the experimental system allows the simultaneous acquisition of resistivity and IP data using the same multi-electrode setup with a relatively small increase of acquisition times, IP data were also acquired for the dipole–dipole configuration for a more comprehensive electrical characterization of the near-surface layers in order to assess potential underground sources of problems to the monumental buildings. Figure 8 Open in new tabDownload slide Electrical resistivity (top) and induced polarization (bottom row) model sections for profiles P4 and P6 showing alternation of resistive and conductive layers (bounded by black lines) and a high-chargeability anomaly (white) in the IP model. The outline of the IP anomaly has been superimposed onto the resistivity models to evidence the correlation with a low-resistivity zone. Figure 8 Open in new tabDownload slide Electrical resistivity (top) and induced polarization (bottom row) model sections for profiles P4 and P6 showing alternation of resistive and conductive layers (bounded by black lines) and a high-chargeability anomaly (white) in the IP model. The outline of the IP anomaly has been superimposed onto the resistivity models to evidence the correlation with a low-resistivity zone. The dipole–dipole model sections (figure 8) show a relatively well-defined electro-stratigraphical sequence for both profiles, although the predominant horizontal layering is more evident for profile P6 than P4. In both cases, the top 1 m depth is characterized by a very shallow heterogeneous layer with more resistive inclusions towards the limestone-paved surface. This layer is followed by a fairly resistive (150–300 Ω m) layer up to about 1.6 m depth, and then a more conductive layer up to about 2.5 m depth, and finally a high-resistivity basement (150–1000 Ω m). The shallow layers, and in particular the conductive one, show an increase of the average conductivity in the left part, which is more pronounced for profile P6 and accounts for the stronger attenuation in the initial part of the corresponding GPR section and the globally lower reflectivity in the western part of the time slices (figures 5 and 6). A localized high-resistivity anomaly near the abscissa 21 m on profile P6 explains the vertical feature (A2) interpreted as probable archaeological remains in the GPR section (figure 5). The P4 resistivity model (figure 8) shows a less sharp interface between the top resistive layer and the intermediate conductive unit. At about 19 m, both profiles show a marked increase in conductivity (resistivity < 50 Ω m) which is mirrored in the corresponding IP model sections by a high chargeability anomaly (20–25 mV V-1) that stands out from a globally non-polarizable background. While the interpretation of resistivity data is relatively straightforward, the interpretation of IP data is much more difficult because of various concurrent physical phenomena that can determine a chargeability anomaly. For instance, disseminated clays and metallic minerals have significant IP signatures and, without other information, cannot be distinguished from other sources (e.g. contaminants). However, in our case the IP anomaly (white ovals in figure 8) coincides very well with the high-conductivity anomaly in the corresponding resistivity models and with the location of the shallow central pipe evidenced by the GPR survey (figures 5–7). The high reflectivity of the pipe suggests that it could be made of a metallic material. Therefore, although the presence of a local accumulation of clayey sediments cannot be excluded, the most plausible interpretation for the resistivity and chargeability anomaly is as a gravity-induced vertical leakage from the presumably rusty central pipe. This would cause a local increase in water saturation as well as in the concentration of polarizable minerals. In any case, the presence of a similar anomaly in both profiles clearly outlines its correlation to the pipe. The general shallow electro-stratigraphy is confirmed also in the Wenner–Schlumberger models (figure 8), although due to the different cell discretization one should not expect a perfect correspondence in the shape of the electrical interfaces. The deepest resistive layer outlined in the dipole–dipole models extend to an average depth of 5–6 m according to the Wenner–Schlumberger results and is followed by a conductive layer presumably related to the phreatic aquifer, as confirmed by the level of the water table in nearby wells. An interesting observation is that the shallower and deeper conductive units seem to be interconnected in correspondence to the central resistivity and chargeability anomaly. An explanation is that the leakage is not confined within the shallow conductive unit but penetrates through the resistive basement to reach the deeper aquifer. Although the pipe content is presumably water, the interconnection of the two hydro-geological units outlines the potential risk of local groundwater contamination. 4.3 Synthesis of the results Figure 9 shows an overview of the main results of the combined geophysical survey including the pipe network outlined by GPR (light blue lines) and the shallow hydro-geological units (ERT models) superimposed on to the map of Palazzo dei Celestini, as well as photographs of a post-survey excavation performed inside the cloister close to the church northern wall (E). The accurate topographical survey of surface evidences (green boxes in figure 9) assisted the interpretation of the GPR survey and helped distinguish between possible pipes and natural heterogeneities. Figure 9 Open in new tabDownload slide Map of Palazzo dei Celestini with a synthesis of the main geophysical results, including the pipe network outlined by GPR (light blue and black-dashed lines) and the shallow hydro-geological units (ERT models). A, F and O mark the location of the most damaged altars and the photographs show the results of a post-survey excavation located at E. Figure 9 Open in new tabDownload slide Map of Palazzo dei Celestini with a synthesis of the main geophysical results, including the pipe network outlined by GPR (light blue and black-dashed lines) and the shallow hydro-geological units (ERT models). A, F and O mark the location of the most damaged altars and the photographs show the results of a post-survey excavation located at E. The central shallow pipe (dashed line) correlates very well with the position of the high-conductivity and high-chargeability anomaly outlined by the electrical survey (figures 8 and 9) and is deemed responsible for localized infiltration of fluids containing metallic minerals in the shallow aquifers. The main stratigraphical interfaces deduced from the GPR survey (figure 5) agree well with the boundaries of the electro-stratigraphical layers in the ERT models (figures 8 and 9). The higher GPR attenuation in the western part of profile P6 and the higher conductivity in the corresponding ERT model correlate well with the higher degree of stone decay and efflorescence phenomena in the altars located in the western part of the Basilica. Unfortunately, the geophysical profile could not be extended further westward where the most damaged altars (A, F and O) are located, but the trend observed from P4 to P6 suggests a further increase in conductivity towards SW and supports an underground origin as the main cause of stone degradation via the rising of capillary water. However, other anthropogenic causes can contribute to worsen the situation. In the occasion of heavy rains, probably because of obstructed drainage pipes, stagnation of water in the SW corner of the cloister has been observed, which gradually infiltrates in to the subsoil and is subsequently released as capillary water. Moreover, during the spring following the geophysical survey, an inexplicable spillage of water from the ground in correspondence to the position marked with E was observed. The subsequent excavation showed that it was caused by the presence of a hole (pictures on the left side of figure 9) accidentally made in a drain hidden in the church wall during conservation works to insert resin materials (row of darker dots) aimed at stopping the rising of capillary water (Cirielli, personal communication). This paradoxical circumstance highlights the importance of non-destructive testing, involving also geophysical methods, before any conservation action on cultural heritage in order to prevent the risk of worsening its state of preservation. During the same reconnaissance excavation, under a nearby bas-relief decorated with dolphins, symbol of water, a void was found leading to an underground cistern containing water at a depth consistent with that of the main conductive unit of the geoelectrical survey, which was probably in use during the period of activity of the convent (pictures on the right side of figure 9). 5 Conclusions The paper reported the main results of an integrated geophysical campaign performed inside the cloister of a former convent in Lecce, southern Italy, in order to investigate the possible subsurface causes of deterioration affecting its pillars and walls and, more importantly, some altars of the adjacent Basilica. The deterioration of ancient buildings can be due to various causes: external, such as pollution, biological degradation and adverse climatic or microclimatic conditions, and internal, such as a particular geological or hydro-geological setting or a combination of both. Therefore, being able to discriminate between the different sources and to identify the main process of decay becomes essential for the development of effective remediation actions. These results demonstrate the crucial role of near-surface geophysics in the framework of cultural heritage diagnostics to gain information on subsurface properties that can affect the stability of historical structures and accelerate degradation processes. The geophysical survey aimed at evidencing the general stratigraphical and hydro-geological setting of the area, and possible natural or anthropogenic causes of local increase in subsoil moisture that could promote the rising of moisture through the building structure. The use of different GPR antenna frequencies (100, 200 and 500 MHz) allowed a detailed stratigraphical mapping of the shallow subsurface at slightly different penetration depths and to identify different-sizes ancient and modern drainage systems. Time-slice maps, 3D iso-surface representations and 3D animations yielded effective visualizations of the pipe networks and archaeological anomalies and allowed the analysis of their spatial interrelation. A conductive layer between 1.3 and 2.5 m depth, especially in the western portion of the ERT profile closest to the Basilica of Santa Croce, could indicate a local increase in the water content, which could be responsible for the higher degree of stone decay and efflorescence phenomena in the altars located in the western part of the Basilica. A narrow high-conductive feature, evidencing also anomalous IP values, was interpreted as a presumable leak of fluids enriched with metallic minerals from a shallow pipe. The example shown in this study demonstrates that the combination of different geophysical methodologies, complemented by other physical, geological and biological techniques, is crucial for obtaining an exhaustive knowledge of the status of historical monuments and for identifying the possible causes of damages in order to improve the cultural heritage management. Acknowledgments The authors are grateful to Professor MT Carrozzo from the University of Salento for helpful discussions and to the student U Guidotti and the technicians M Luggeri and G Fortuzzi for their contribution to the acquisition and processing of the geophysical and topographical data and the elaboration of the CAD drawings. They also want to thank F Panico from the Province of Lecce Administration for the permission to perform the geophysical survey inside the cloister of Palazzo dei Celestini. They express a special gratitude to R Cirielli for the precious information and photographs on the maintenance works performed inside the monumental building. References Basile V , Carrozzo M T , Negri S , Nuzzo L , Quarta T , Villani A V . , 2000 A ground-penetrating radar survey for archaeological investigations in an urban area (Lecce, Italy) , J. Appl. Geophys. , vol. 44 (pg. 15 - 32 )http://dx.doi.org/10.1016/S0926-9851(99)00070-109269851 Google Scholar Crossref Search ADS WorldCat Cardarelli E , Fischanger F , Piro S . , 2008 Integrated geophysical survey to detect buried structures for archaeological prospecting. A case-history at Sabine Necropolis (Rome, Italy) , Near Surf. Geophys. , vol. 6 (pg. 15 - 20 ) OpenURL Placeholder Text WorldCat Carrozzo M T , Cataldo R , De Nunzio G , Leucci G , Marzo L , Nuzzo L , Villani A V . , 2003 Microclimatic and geophysical surveys on cultural heritage (Basilica of Santa Croce—Lecce, Italy). Data integration and visualization (GIS) , Proc. SAGEEP , vol. 16 (pg. 1103 - 20 ) OpenURL Placeholder Text WorldCat Cassiano A , Cazzato V . , 1997 , Santa Croce a Lecce: Storia e restauri Galatina Congedo Editore Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Cataldo R , De Donno A , De Nunzio G , Leucci G , Nuzzo L , Siviero S . , 2005 Integrated methods for analysis of deterioration of cultural heritage: the Crypt of ‘Cattedrale di Otranto’ , J. Cultural Heritage , vol. 6 (pg. 29 - 38 )http://dx.doi.org/10.1016/j.culher.2004.05.00412962074 Google Scholar Crossref Search ADS WorldCat Cataldo R , Leucci G , Siviero S , Pagiotti R , Angelini P . , 2009 Diagnostic of the conservation state in the crypt of the Abbey of Montecorona: biological, microclimatic and geophysical evaluations , J. Geophys. Eng. , vol. 6 (pg. 205 - 20 )http://dx.doi.org/10.1088/1742-2132/6/3/0011742213217422140 Google Scholar Crossref Search ADS WorldCat CNR-ISCOM . , 2000 Basilica di Santa Croce in Lecce. Indagini sugli altari e sulle navate laterali Report of the Istituto Conservazione Opere Monumentali C.N.R., Lecce Colangeli O . , 1978 , S. Croce. L'ex-convento dei Celestini Galatina Editrice Salentina Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Conyers L B . , 2004 , Ground-Penetrating Radar for Archaeology New York AltaMira Press, Rowman & Littlefield Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Davis J L , Annan A P . , 1989 Ground-penetrating radar for high-resolution mapping of soil and rock stratigraphy , Geophys. Prospect. , vol. 37 (pg. 531 - 51 ) Google Scholar Crossref Search ADS WorldCat Gaffney V , Patterson H , Piro S , Goodman D , Nishimura Y . , 2004 Multimethodological approach to study and characterize Forum Novum (Vescovio, central Italy) , Archaeological Prospection , vol. 11 (pg. 201 - 12 )http://dx.doi.org/10.1002/arp.2351075219610990763 Google Scholar Crossref Search ADS WorldCat ICOM . , 1983 La pietra: interventi, conservazione, restauro Atti del Convegno internazionale del comitato nazionale italiano dell'ICOM Lecce, 6–8 November 1981 Galatina Congedo Editore Leucci G , Negri S , Carrozzo M T , Nuzzo L . , 2002 Use of ground penetrating radar to map subsurface moisture variations in an urban area , J. Environ. Eng. Geophys. , vol. 7 (pg. 69 - 77 )http://dx.doi.org/10.4133/JEEG7.2.691083136319432658 Google Scholar Crossref Search ADS WorldCat Loke M H . , 2004 Tutorial: 2-D and 3-D electrical imaging surveys Copyright (1996–2004). Geotomo Software, Malaysia Loke M H , Acworth I , Dahlin T . , 2003 A comparison of smooth and blocky inversion methods in 2D electrical imaging surveys , Explor. Geophys. , vol. 34 (pg. 182 - 7 )http://dx.doi.org/10.1071/EG0318208123985 Google Scholar Crossref Search ADS WorldCat Loke M H , Barker R D . , 1996 Rapid least-squares inversion of apparent resistivity pseudosections by a quasi-Newton method , Geophys. Prospect. , vol. 44 (pg. 131 - 52 )http://dx.doi.org/10.1111/j.1365-2478.1996.tb00142.x0016802513652478 Google Scholar Crossref Search ADS WorldCat Margiotta B . , 1994 Monumenti a vita breve: il condizionamento della pietra leccese sul barocco salentino: l'esempio del monastero degli Olivetani Report of the Centro studi di geotecnica e di ingegneria, Lecce Margiotta S , Negri S . , 2005 , Alla Ricerca dell'acqua perduta Galatina Congedo Editore pg. p 208 Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Masini N , Nuzzo L , Rizzo E . , 2007 GPR investigations for the study and the restoration of the rose window of Troia Cathedral (southern Italy) , Near Surf. Geophys. , vol. 5 (pg. 287 - 300 ) OpenURL Placeholder Text WorldCat Nuzzo L . , 2004 The contribution of GPR to investigate damages in the Crypt of the Cathedral of Otranto (Apulia, Italy) Proc. 10th Int. Conf. on Ground Penetrating Radar Delft, The Netherlands, 21–24 June 2004 Nuzzo L , Leucci G , Negri S . , 2007 GPR survey in the karstic area ‘Spedicaturo’ near Nociglia (Lecce, Italy) , Near Surf. Geophys. , vol. 5 (pg. 67 - 76 ) OpenURL Placeholder Text WorldCat Nuzzo L , Leucci G , Negri S . , 2009 GPR, ERT and magnetic investigations inside the Martyrium of St Philip, Hierapolis, Turkey , Archaeological Prospection , vol. 16 (pg. 177 - 92 )http://dx.doi.org/10.1002/arp.3641075219610990763 Google Scholar Crossref Search ADS WorldCat Nuzzo L , Leucci G , Negri S , Carrozzo M T , Quarta T . , 2002 Application of 3D visualization techniques in the analysis of GPR data for archaeology , Ann. Geophys. , vol. 45 (pg. 321 - 37 ) OpenURL Placeholder Text WorldCat Nuzzo L , Quarta T . , 2005 Preliminary results of the geophysical survey inside ‘Palazzo dei Celestini’, Lecce, Italy Proc. 11th Eur. Meeting of Environmental and Engineering Geophysics—Near Surface 2005 Palermo, Italy, 4–7 September 2005 (pg. 1 - 4 ) CD-Rom, P074.pdf Nuzzo L , Quarta T . , 2009 Near-surface geophysical investigations inside the cloister of an historical palace in Lecce, Italy EGU General Assembly 2009 EGU2009-11770 Vienna, Austria, 19–24 April 2009 , vol. vol 11 Geophysical Research Abstracts Oldenburg D W , Li Y . , 1994 Inversion of induced polarization data , Geophysics , vol. 59 (pg. 1327 - 41 )http://dx.doi.org/10.1190/1.14436921070485X Google Scholar Crossref Search ADS WorldCat Olhoeft G R . , 1985 Low-frequency electrical properties , Geophysics , vol. 50 (pg. 2492 - 503 )http://dx.doi.org/10.1190/1.14418801070485X Google Scholar Crossref Search ADS WorldCat Sandmeier K J . , 2008 ReflexW Version 4.5.5. Sandmeier Software, Zipser Strasse 1, D-76227 Karlsruhe, Germany Soupios P M , Loupasakis C , Vallianatos F . , 2008 Reconstructing former urban environments by combining geophysical electrical methods and geotechnical investigations—an example from Chania, Greece , J. Geophys. Eng. , vol. 5 (pg. 186 - 94 )http://dx.doi.org/10.1088/1742-2132/5/2/0051742213217422140 Google Scholar Crossref Search ADS WorldCat Tsokas G N , Tsourlos P I , Vargemezis G , Novack M . , 2008 Non-destructive electrical resistivity tomography for indoor investigation: the case of Kapnikarea Church in Athens , Archaeological Prospection , vol. 15 (pg. 47 - 61 )http://dx.doi.org/10.1002/arp.3211075219610990763 Google Scholar Crossref Search ADS WorldCat Turnone A . , 2002 Tecniche per indagini fisiche volte alla conservazione dei beni monumentali: la Basilica di Santa Croce in Lecce , MSc Thesis University of Lecce © 2010 Nanjing Geophysical Research Institute
TI - Near-surface geophysical investigations inside the cloister of the historical palace ‘Palazzo dei Celestini’ in Lecce, Italy
JO - Journal of Geophysics and Engineering
DO - 10.1088/1742-2132/7/2/S06
DA - 2010-06-01
UR - https://www.deepdyve.com/lp/oxford-university-press/near-surface-geophysical-investigations-inside-the-cloister-of-the-0uQesXAozz
SP - 200
VL - 7
IS - 2
DP - DeepDyve
ER -