Geophysical surveys at Archaeological sites in My Son Sanctuary

I. INTRODUCTION AND SITE HISTORY

1. Mỹ Sơn is situated in Quảng Nam Province about 60 km far from Đà Nẵng City. The group of monumental complexes is scattered over an area of about 10 ha lying some 40 m above the sea level and surrounded by hills rising to a height of 100-300 m.

In 1999, Mỹ Sơn was recognized by UNESCO as a world heritage listed site. The French archaeologists divided the architectural works at Mỹ Sơn into 10 principal groups, A, A', B, C, D, E, F, G, H, K to enable a system of labelling sites by capital letter and number (Fig. 1).

Mỹ Sơn was once a veritable forest of towers, many of them were destroyed by the ravage of time and war. This unique site is now in a state of significant disrepair, urgently requiring conservation efforts. In order to evaluate the evident of the burried remains and underground structures the geophysical investigation was suggested.

From its inception, one of the tasks of the Mỹ Sơn Conservation Project was delimiting the known archaeological areas and locating other possible monument remains still under the ground. The geophysical investigations carried out at Mỹ Sơn have been also aimed to investigate changes in physical properties, in both vertical and horizontal directions, related to the degradation process in the subsurface environment, which affect the stability of the monument foundation structures.

To solve these problems, three methods were used: resistivity survey for both horizontal and vertical exploration, magnetic measurement of the earth’s magnetic field and magnetic susceptibility measurement of the horizontal. Among these, the resistivity and magnetic susceptibility measurements were conducted by a team of geophysicists from the Institute of Geological Sciences, Vietnamese Academy of Sciences and Technology (VAST). The magnetic measurement was carried out by experts from the Lerici Foundation and Milan Polytechnic University (Italy).

2. The Thăng Long Citadel, which is the present Hà Nội centre, was chosen by Lý Thái Tổ, the founding king of the Lý Dynasty, as the capital of Đại Việt Kingdom (present Việt Nam) since Autumn, 1010. Since that, Thăng Long was almost continuously the capital city of Đại Việt Kingdom through different reigns from 1010 to 1789 years. In the periods when it was not the capital, Thăng Long still remained as the central city in the area of North Việt Nam. In 1945, Hà Nội was again chosen by the Government of DRV as the capital.

In the 2004, after more one-year of investigation and archaeological excavation near the Ba Đình Assembly Hall 20.000 m²of land surface were opened. Reaching about 4 m in depth encountered building remains and cultural layers going back to Fifth-Sixth Century very far before the date of Hà Nội foundation (1010) according historical sources (Fig. 2). The event has been attracting the attention of the Vietnamese Government, Hà Nội Municipality, Hà Nội UNESCO Office and many international experts from several countries.

The present relic complex is located in Ba Đình District (Fig. 2), roughly surrounded by Nguyễn Tri Phương, Phan Đình Phùng, Điện Biên Phủ, Trần Phú, Hoàng Văn Thụ and Hùng Vương Streets. Relics in the areas of Forbidden City and Imperial Citadel situated on the central axis of the old citadel are the Cửa Bắc (North Gate), Hậu Lâu, Kính Thiên Palace, Doãn Môn, Cột Cờ (Flag Tower) and the archaeological site at 18 Hoàng Diệu Street. Relics found on the ground remain relatively intact and are well preserved. Underground archaeological artifacts were discovered and studied in many years. Researchers have found out many architectural vestiges and artifacts from the Chinese domination period ago (Đinh, and Early Lê Dynasties) and from the Lý, Trần, Lê and Nguyễn Dynasties.

The Cultural Historical Complex of Thăng Long - Hà Nội can be compared with such World Cultural Heritages as the Complex of Huế Monuments (Central Việt Nam) and the Historic Monuments of Nara Palace (Japan). As compared to Nara Palace, the Thăng Long imperial city relics and other surrounding architectures become the complex of historical, architectural and archaeological relics together with the combination of natural landscape. One interesting point is that around 1,300 years ago, Nara and Thăng Long (then Đại La) were both the politic and cultural centres of Japan and Việt Nam respectively.

The archaeological remains in Hà Nội is destroyed and buried by natural geological hazards and urban development. The old Hà Nội City is situated in the low topographic land of the bank of Red River. Its foundation is unconsolidated soil caused by land subsidence and sedimentation is activated every year by flooding.

The problem how to continue the necessary investigation in the nearby present excavation is important practice. In fact, the results of excavation have been showing that, the different cultural layers with important archaeological and architectural remains are buried and continue to extend in all directions. Because the area is in the important place of the Hà Nội City, it is very difficult for archaeologists to obtain the permission for continuing the excavation.

The main goal of the geophysical investigation is to locate evidences, if possibly, to map and determine the structures concerning to the distribution of buried archaeological remains. Practically, many similar problems have been solved successfully in many countries of the world by using number of geophysical methods, such as electrical tomography, ground penetrating radar, seismic and magnetic surveys. In order to overcome the difficulties in urban conditions for conventional methods, an electromagnetic technique based on non-grounding electrical and magnetic measurements of line dipole source produced by the Russian Company ERA was tested. In this paper an analysis of the obtained results by all methods in the testing areas with known buried archaeological targets is presented and discussed.

II. RESULTS OF GEOPHYSICAL INVESTIGATIONS

1. At Mỹ Sơn Site

Results of Magnetic Prospecting

During the 2004 research season, an intensive magnetic prospecting campaign was carried out on the area of G Group, to detect and locate unknown buried structures. In this case a grid of one-metre side was used to provide a detailed map. The G Group area is a relatively flat area on the top of a hill of nearly rectangular shape (Fig. 1). The long side of the hill stretches in the NW-SE direction. The archaeological complex here comprises five monuments; four of them were almost completely destroyed: the remains take the form of footwalls, with a height of less than 1 m (monuments G2, G3), or are completely erupted (G4, G5).

The magnetic map (Fig. 3) shows many magnetic anomalies: some produced by the existing building, by large-scale dumping in the northeast sector; others were produced by bomb fragment remains and modern features. The anomalies indicated by dotted lines on Fig. 3 were produced by excavation trenches made by Parmentier at the beginning of the last century. Clear groups of anomalies are produced by a laterite enclosing wall, by dumping during the previous archaeological excavation (Parmentier) and by many modern features (remains of shell bombs, etc.), as indicated on the map in Fig. 3. There is no trace of other brick structures.

The investigations were concentrated in the areas of G Group and E7 monument aiming to locate sub-surface inhomogeneities related to archaeological remains; 2D and 3D techniques were applied.

In the area of G Group, 21 profiles were designed. The distance from profile to profile is 2 m, while the distance between the measurement points (minimum distance between electrodes) along each profile is 1 m. The separation (level) of the electrode arrays ranges from 8 to 14 times. This allowed us to investigate on a maximum depth of about 4-7 m. In the area of E7 monument investigations were conducted along three parallel profiles. The central profile passes along the axis of the foundation surface of the monument interior; the two others are parallel one to another, distributed outside the monument. The distance between the profiles is about 3.5-4 m, and the distance between the measurement points along each profile is 1 m.

Data interpretation was carried out by using both 2D and 3D techniques. We used the DC2DSIRT software for 2D, and IP3D software for 3D employing the Simultaneous Iterative Reconstruction Techniques (SIRT) [1]. In the calculation, the initial model was chosen by direct observation of resistivities. The 2D inversion results were presented in the form of geoelectrical cross-sections, through which we were able to understand the resistivity contrast along the vertical plane as well as along each profile. The results of 3D inversion allowed us to calculate the distribution of resistivities on the plane at different depth levels, so as to see the features of the sub-surface environment on the whole investigated area [4, 5].

We can see the more intuitive horizontal distribution of resistivity on the scheme of structure at depth levels Z = -0.64, -1.17 and -1.8 m derived from the 3D inversion. Analysis of results shows that the 3D inversion allowed us not only to reveal a smaller size of resistivity anomalies, but also better to gauge the dimensions and shape of the buried objects (Fig. 4a).

From a depth of more than 2 m below the land surface, the environment shows strong differentiation in resistivity value (Fig. 4b). The higher resistivity anomalies related to construction materials (refined bricks) are separated from the monuments which were scattered and then buried in the surrounding environment. In addition, the higher or lower anomalies, compared with the background value, can be produced by other man-made systems: in some locations, large holes created by bombing and explosions during the war, etc... At levels deeper than 3 m, all the cross-sections show the resistivities becoming much more stable. Experience in those locations suggests that the range of resistivity reflects the presence of soil of high clay content produced from the original sandstone and siltstone by a process of weathering.

2. At Thăng Long Citadel site

In order to evaluate the evidence of the buried remains and structures on the geophysical responses the surveys by conventional methods: resistivity, seismic, magnetometry and ground penetrating radar were carried out. The field observation at the site met some difficulties due to high level of noise produced by urban activities. In addition, surface layer of pavement, recent constructions and underground system of sewers, water supply pipes are obstacles, not only for field works, but either decrease signals from buried archaeological targets. All these effects are the reason of low quality of observed data; therefore the surveyed results not all away satisfied for solving proposed problems.

In combination with conventional methods the new electromagnetic ERA equipment was tested. The obtained results showed that, it allows overcoming difficulties in both field survey and improve some advantages for mapping buried structures at the site.

Resistivity tomography could employed only along single profiles mainly in narrow band of garden or margin of the roads in both Ba Đình excavation and Imperial Citadel.

The results in Resistivity surveys at two excavated sites are presented in Fig. 5: 1) In the Doãn Môn Gate the surveyed profile is distributed nearby the open excavation, where is a buried architectural remains at its bottom (Fig. 2, right); 2) In the Garden of Hậu Lâu residence, the profile is arranged along the axis of excavation filling soil. The high resistivity anomalies correspond to the shape of noticeable excavations. Below these sections the low resistivity zones denote saturated mud and silt. One isolated anomaly with high resistivity distributed under the marks 32-35 m of profile in the Hậu Lâu can be related to buried target of remain? The brick road in the bottom of excavation in Doãn Môn is not identified in the cross-section of resistivity model (Fig. 5b). The no light contrast of resistivity of soil and buried remains caused by noise, that decreases resolution of the spread of array in resistivity tomography.

Magnetic and seismic surveys: The results of theoretical and experimental research show that, magnetic contrast between some archaeological materials (especially burned bricks and its brickbat) and soil is always exist [3]. Due to the main materials of ancient construction here made by burned bricks and flagstones, the exploration was conducted in some limited areas. The high noise level caused by the urban systems became the obstruction to identify the buried remains. Thus, in the areas of Hoàng Thành the magnetic survey is mostly a non-productive one.

In the areas of Hoàng Thành were also carried out the seismic explorations along a number of profiles. The results reflect the subsurface layered geological environment, the underground trench vestige as well as some indications related to the buried remains.

Non grounding electrical and magnetic surveys: The main advantages of the new equipement ERA are sensitive to the ferromagnetic content in soil, so its use allows us to detect more easily a small change in the ground environment. Small size, light weight and non-galvanic contact of measurements are strong points of the equipement, that allow us to easily and quickly carry out the exploration in complicated areas, such as inside the residences, urban locations etc...

For the measurement in Hoàng Thành area, the linear dipole source with length of 150 m and frequency of 625 Hz were installed outside the investigated area. The electrical Ex and magnetic Hz components at the same frequency are measured at distance in every 1 m along profiles distributing parallelly the dipole source.

The collected data are concentrated at sites of Thăng Long Citadel: 1) around the excavation in Doãn Môn Gate, that is related to the buried brick road – a typical remain at the site; 2) area of Imperial (Dragon) Palace (Kính Thiên), where is a known tunnel – facility of underground structure. These examples are chosen because all the mentioned above objects are known and we can use them for evaluate the exploration.

At the Imperial Palace site (Fig. 7a), the Ex anomalies are concentrated in the left of the map. The highest resistivity anomaly is caused by the concrete construction and empty space of the underground tunnel. Next to the right of the map was detected conductive anomaly Ex, which is closely related to the iron materials inside the known underground trenches (note here is high anomaly Hz). Remaining part of the map Ex characterized by moderate values of resistivity, it indicates relatively homogeneity of the soil here.

At the Doãn Môn site (Fig. 7b), evidence of recovered brick road at bottom of the excavation is identified by a high resistivity anomalies Ex in both northern and southern parts of the map. The higher intensity at the nothern part, may be produced at the same time by the brick road and additional effect from the undergronud trench there. The highest resistive anomaly in the northeastern award of the excavation reveals the signal of the known tunnel. The plan map can be divided in two parts with different features: left pattern of the map with relatively high resistivity indicates dry pavement, right one with low resistivity related to the water saturated soil. The two isolated high resistive anomalies distributed in both margins of the open excavation may be explained by a “wall effect” of the EM field. The conductive anomaly in the south-eastern edge of the plan map is related to iron tube going out in the surface.

III. CONCLUSIONS

1. The surveying results by using magnetometry, 2D-3D resistivity, magnetic susceptibility at My Sơn site successfully provide with informations about archaeological remains, in fact the magnetic and magnetic susceptibility anomalies indicate a large quantity of bricks scattered throughout the area, the laterite enclosing wall, remains of bomb fragments and trenches of Parmentier’s excavations. Based on the results obtained from 2D-3D resistivity distribution, we can establish a risk map at different depths (0.64, 1.17 and 1.8 m in G Group) indicating (or describing) features of both the environment and the buried remains under the investigation area.

2. A lower efficiency of identification of buried remains by geophysical investigations in Thăng Long Citadel (Hà Nội) is caused by the complication of urban systems, such as the concrete and asphalt surface, dense construction, high conductivity in the near surface water saturated layer, water supply pipes and electric energy transmition line systems etc.

3. The EM technique with linear dipole source at 625 MHz and non grounding electrical Ex measurements by ERA equipment are possible to assert that: the underground structures of concrete construction as tunnel is revealed by highest anomalies Ex at both Imperial Palace and open excavation near Front Gate (Doãn Môn) sites. The buried brick road as a remain at Doãn Môn site is identified by also high, but smaller than the anomaly Ex caused by the concrete structure. The evidence of underground construction containing iron material were detected by very low anomaly Ex in large area as revealed at Imperial site.

Based on the tested results, the combination of electromagnetic and conventional geophysical methods is more expected for successfully solving proposed problems at archaeological Thăng Long site in future. This result is useful experience for investigating other archaeological and historical sites in Việt Nam.

Acknowledgements: This work is partially supported by International and Vietnamese Agencies: UNESCO, Italian-Việt Nam Government Committee for Scientific Cooperation, Việt Nam National Research Fund of Basic Sciences.

The authors are grateful to the director and managers of Hà Nội ancient wall – Cổ Loa vestiges preserved Centre for the helful cooperation.

REFERENCES

1. Barker R.D., 1992. A simple algorithm for electrical imaging of the subsurface. First break, 10/2 : 58-64.

2. Binda Luigia, Cristina Tedeschi et al., 2006. Characterization of material sampled from some My Sơn Temples. 7th Conf. of ICCE, Teheran, Iran.

3. Cucarzi Mauro, K.M. Ermukhin et al., 2001. Method of volume dipole sources for solving problems of stationary magnetic and electrical prospecting. Prospezioni archeol. - Fondazione Lerici Quaderni, 2 : 183-202.

4. Đoàn Văn Tuyến, Đinh Văn Toàn et al., 2005. Geophysical methods in archaeological investigation and some applied results. Proc. Nat. Conf. on Engin. Geol. and Env., pp. 216-224. Hà Nội.

5. Đoàn Văn Tuyến, Đinh Văn Toàn, Trần Cánh, Trịnh Việt Bắc, Lại Hợp Phòng, 2005. Phương pháp địa vật lý phục vụ công tác khảo cổ và một số kết quả ứng dụng. Tt báo cáo HNKH toàn quốc Địa chất công trình và Môi trường, tr. 216-224. Hà Nội.

6. Parmentier H., 1909. Inventaire descriptif des monuments Chams de l'Annam. Paris.

7. Sapozhnikov B., 2002. Resistivity method without grounding. Ann. Geophys., 14 : 141-162.

8. Won J. et al., 1996. GEM-2, a new multifrequency electromagnetic sensor. J. of Env. and Engin. Geophysics, 6 : 448-451.