Conference Paper

1

CHAPTER 1

INTRODUCTION 1. 1 Role of magnetic methods in near-surface investigation Magnetic methods have a prominent place in near surface geophysics for a number of reasons. First, the sources of interest often have strong magnetic signature; sometimes the only measurable geophysical of the target is its magnetic field. On the other hand, magnetic measurements are comparatively simple, rapid, and completely noninvasive. Although the resolution required for near-surface applications generally makes airborne data acquisition impractical, even this is possible in some cases. Finally, in near-surface investigations magnetic data are often easy to interpret; in some cases, visual inspection of essentially raw data is sufficient. For all these reasons, magnetic methods are widely used in almost all areas of near-surface geophysics. In some subfields, such as buried ordnance detection, magnetic methods are particularly important.

1. 2 Basic concept of magnetic survey The aim of a magnetic survey is to investigate subsurface geology on the basis of anomalies in the Earth’s magnetic field resulting from the magnetic properties of the underlying rocks. Although most rock-forming minerals are effectively non-magnetic, certain rock types contain sufficient magnetic minerals to produce significant magnetic anomalies. Similarly, man-made ferrous objects also generate magnetic anomalies. Magnetic surveying thus has a broad range of applications, from small-scale engineering or archaeological surveys to detect buried metallic objects, to large-scale surveys carried out to investigate regional geological structure for explore of oil, gas and coal. Within the vicinity of a bar magnet a magnetic flux is developed which flows from one end of the magnet to the other (Source: magnetic flux theory). This flux can be mapped from the directions assumed by a small compass needle suspended within the magnet where the flux converges are known as the poles of the magnet. A freely suspended bar magnet similarly aligns in the flux of the Earth’s magnetic field. The pole of the magnet which

2

tends to point in the direction of the Earth’s north pole is called the north seeking or positive pole, and this is balanced by a south seeking or negative pole of identical strength at the opposite end of the magnet.

Figure 1.Magnetic fields on the Earth: The giant bar magnet

Magnetic fields can be visualized as magnetograms, which are used to make observations of the Sun. Magnetograms are visual representations of the polarity and the strength of magnetic fields that point directly away or towards the observer. Black regions have the strongest southward field (the field points away from the observer and into the Sun), and white regions have the strongest northward field (the field points towards the observer and away from the Sun). Gray areas have little or no magnetic field. Using the concept of fields that we explored in the previous section, the magnetic force exerted on a given particle can be expressed as follows in SI units: Fmag = qvBsin

Where, q is the amount of charge, v the velocity of the particle, B the magnetic field, and the angle formed between the velocity and magnetic field directions. The direction of this magnetic force is determined by convention using what is called the right hand rule (source: A text book of modern physics). According to this rule, if we were to curl our fingers starting in the direction of the velocity to the direction of the magnetic field, our thumb would point in the direction of the magnetic force. In other words, the force is perpendicular to the velocity and the magnetic field. If we are familiar with vectors and vector math, in particular the cross product, the magnetic force equation can be expressed in a more elegant form in SI units as: Fmag = qv x B

3

In a word, Magnetic surveys are one of the tools used by the exploration geophysicist in the search for mineral-bearing ore bodies or even oil-bearing sedimentary structures. The essential feature magnetic survey is the measurement of the magnetic-field intensity and sometimes the magnetic inclination, or dip, and declination (departure from geographic north) at several stations. In physics, a magnetic field is that part of the electromagnetic field that exerts a force on a moving charge. A magnetic field can be caused either by another moving charge (i.e., by an electric current) or by a changing electric field. The magnetic field is a vector quantity, and has SI units of tesla, 1 T = 1 kg·s-1·C-1. Magnetic field is usually denoted by the symbol . Historically, was called the magnetic flux density or magnetic induction. A distinct quantity was called the magnetic field (strength), and this terminology is still often used to distinguish the two in the context of magnetic materials (non-trivial permeability µ). Otherwise, however, this distinction is often ignored, and both quantities are frequently referred to as "the magnetic field." (Some authors call the auxiliary field, instead.) In linear materials, such as air or free space, the two quantities are linearly related:

Here,

is the magnetic permeability of the medium, measured in henries per meter. In

SI units, and are measured in teslas (T) and amperes per meter (A/m), respectively; or, in cgs units, in gauss (G) and oersteds, respectively. Two parallel wires carrying an electric current in the same direction will generate a magnetic field that will cause a force of attraction between them. This fact is used to define the value of an ampere of electric current. While like charges repel and unlike ones attract the opposite holds for currents: if the current in one of the two parallel wires is reversed, the two will repel. Now we shall discuss about the geomagnetic field, which mainly help us during the magnetic survey. We know that, Geomagnetic field is a vector quantity. It has strength and a direction. In figure 2, F- the field vector, Z- the vertical component, X- the geographic north, H- the magnetic north, Y- the geographic east.

4

Figure 2.The relationship between the components of geomagnetic field

1.3 The concept of magnetic gradient The word gradient (from grade) means the inclination of a surface along a given direction. A magnetic gradient is defined as the amount and direction of the linear rate of change of the magnetic field in space. So, the vertical magnetic gradient means the linear rate of change of the magnetic field vertically. For a dipole, the vertical gradient is expressed by taking the derivative of the simplified expression of the dipole (Source: Application manual for portable magnetometers by S. Breirer). The gradiometer anomaly from a dipole varies as 1/z4 which explains why the gradiometer automatically removes non local anomalies, such as the regional gradient. In other words, the gradient varies much more with distance than the total field, or expressed in other terms the difference in intensity between two nearby sensors from distant sources is so small that it is negligible compared to the difference in intensity from nearby sources. The expression 3T/z is a convenient for rapidly estimating the gradient from a dipole given only the total field anomaly and the distance to the source. For example, the earth’s field itself thus has a vertical gradient of 0.004 and 0.008 gammas per foot at the equator and the poles respectively (Source: Application manual for portable magnetometers by S. Breirer).The following two figures are the example of the vertical magnetic gradient.

5

Total field

Vertical gradient Figure 3.Gradiometer as a filter for removal of regional gradient

Total field

Vertical gradient

Figure 4.Gradiometer for resolving local anomalies

1.4 Objective of the study Magnetic survey techniques can be used for a variety of purposes and are fast, more effective, inexpensive, & accurate. It is one of the best methods for reconnaissance survey. It is used to detect subsurface features such as rock fragment, manmade foreign materials, buried walls, structures, Klins, bricks, roof tiles, fire pits, buried pathways, tombs, buried entrances and monuments (AM Abbas, TF Abdallatif & Mancheol Suh, 2005). The main objective of our study is to trace and find out the near-surface archaeological features, particularly in the suchon area by the help of magnetic method.

6

CHAPTER 2

HISTORICAL BACKGROUND & GEOLOGY 2.1 Gongju history Gongju (Gongju-si) is a city in South Chungcheong province, South Korea. It is located at 36°27′N 127°7′E. Gongju was formerly named Ungjin and was the capital of Baekje from AD 475 to 538. Suchon area belongs mainly to the Baekje period and expected to contain some of the representative relics of the Baekje Kingdom. Baekje is one of three competing kingdoms emerged during the Han dynasty in Korea. After the 1st century, three states developed in the Korean peninsula and Manchuria: Goguryo, Shilla, and Baekje. People of the Baekje state built step pyramids and used the key hold-shaped tombs called Kofun. Baekje, occupying the southwestern part of the Korean Peninsula, is traditionally said to be founded in 18 BC in the Seoul area by a legendary leader named Onjo. The Baekje Kingdom was actively engaged in both cultural and trading relations with foreign countries from the early stages. Baekje capital was moved in the ancient times (475A.D.) to Ungjin (present Gongju), and large-scaled castles and graves coexist in Gongju due to the living of many people in the castles. Baekje culture is represented by there periods: the Hansong Period (foundation-475A.D.), the Ungjin Period (475A.D.538A.D.), and the Sabi Period (538A.D.-660A.D.). The stone built Kongsan Mountain Castle in Kongju City, is a typical example of a castle of the Ungjin Period, constructed to straddle the valley floor. Excavations at the site yielded evidence for structures, ponds, and others architectural details, indicating that the site is likely to be used as a royal palace at some time. Tombs of the Baekje Kingdom are divided into two main types; the first, stone-mound tombs, and the second, earth mound tombs. The earth mound tombs may be further classified into those with wooden coffins, stone or brick-chambers, and jar coffins. During the Ungjin Period (475A.D.-538A.D.), stone chamber tombs became predominant between the ruling classes, although this period also referred to the introduction of brick tombs. King Muryong’s tomb, the representative tomb of the Ungjin Period, is a good example of brick tombs. Pottery technology of the Baekje Kingdom appears both to incorporate, and be influenced by, the technology of Lolang, Goguryo, and Chinese six dynasties' potteries. Features unique to the roof-tiles of the Baekje

7

Kingdom only began to appear after the capital of the Baekje Kingdom moved from Hansong to Ungjin (present Kongju) in 475A.D. Baekje people harbored a high regard for protocol and formality, which is manifested in numerous faces of social life ranging from hairdo to marriage and funeral rites. There were strict guidelines concerning methods of greetings and street etiquette, and the coming-of-age ceremony was particularly demanding on its participants. The sexagenarian cycle system, adopted early in the Baekje period, was similar to the customs of today. Also, cattle, horses, chicken as well as hawks were raised and used for farming and hunting. As for the house of the ordinary people, the records of inhabitation described in the historical books indicate that people's houses in the Baekje period were dugout huts with Ondol heating system. The installation of Ondol system is a remarkable improvement over buildings of the previous era. The dugout huts were built on the mountains (hilly districts).The agricultural industry of Baekje had wide plains that are suitable for growing rice. In particular, the agricultural iron-tools (e.g. scythe, plow…etc) were increased the production of grain. The production and use of farming iron-tools were under direct control of the ruling class. Farming irontools are found with weapons at Mongchon earthen-castle and Guui-dong relics. It is to be suggested that farming tools were government properties like weapons rather than personal possessions. The dominance of agricultural industry reduces the fishing industry in Baekje era. Also, hunting with hawks was restricted and popular among the noble class of Baekje (Source: WikipediA, the free Encyclopedia).

2.2 Geology Geology of Korea: Physiographically, Korea is a mountainous peninsula extending south-southeast from the northeastern part of the China mainland. The north-northwest, south-southeast trend forms the Taebaeksan Range, which is close to the east coast. The east coast is of an uplifted topography, showing a relatively straight shoreline, whereas the west coast shows the features of a submerging shoreline. The mountains are not high, rarely exceeding 1,200 meters, but they are found almost everywhere. As a consequence, the terrain is rugged and steep. Only near the west and southwest coasts are there extensive flat alluvial or deluvial plains and more subdued rolling hilly lands. Being a mountainous peninsula, Korea is of a diverse geologic make-up. It is composed largely of Precambrian rocks, such as granite gneisses and other metamorphic rocks. Two separate blocks of Paleozoic Strata are found in South and North Korea. The one in the South covers the Taebaeksan Range, and the one in the North is near Pyongyang. Mesozoic Strata are found in the southeastern part of the peninsula and Cenozoic Strata are limited to some

8

small areas scattered around the peninsula. Jurassic and Cretaceous granites intrude through the older rocks in a northeastward-southwestward direction in some places, but show no specific trend in others.

Figure 5.Geological Map showing the study area in Suchon.

Geology of Suchon: Geologically, the basement rocks around the study area are composed of Precambrian banded gneiss and mica schist. Along the Jungan Stream, west of the study area, the Quaternary alluvial deposits are developed (Tareq Fahmy Abdallatif, 2004). The study area was a steep slope and fully covered by small vegetation. It was gently plunging to the WWS and surrounded by low paddy field. The eastern side of the area is gradually elevated and extended up to several kilometers. Suddenly discontinue of highland and presence of low plain paddy field in the western side of our study area indicates the presence of faulting.

9

CHAPTER 3

MAGNETIC INSTRUMENTATION & FIELD PROCEDURE 3.1 FM (256) Fluxgate Gradiometer A magnetic surveying instrument used to measure the vertical gradient of the geomagnetic field. In a word it’s a magnetometer using two sensors in a light, selfcontained instrument, which measure vertical earth’s magnetic field, and finally we get the vertical gradient of the earth magnetic field. It measures magnetic fields, technically flux density up to several times the strength of the Earth magnetic field. It has a resolution of 0.1, 1, 10 nT and a range of +/-204.7 nT.

Figure 6.Fluxgate gate Gradiometer (FM256) with casing

The above FM256 Fluxgate gradiometer system is designed as a one man rapid location, mapping and identification system for a wide range of targets, which can be

10

archaeological, environmental, utility services, geological or military in origin. Archaeological targets include fired structures such as kilns, furnaces, hearths and ovens, and structures with an enhanced magnetic susceptibility such as pits, ditches, enclosures, field system, barrows etc. Other target includes environmental waste, oil drums, pipelines, cables, unexploded ordnances and geological formations. The FM256 can be operated as a single stand alone gradiometer or in dual gradiometer mode. The dual mode uses two instruments carried together in a CF6 carrying frame (fig.6) to double the survey speed or , by using interleaving, to provide increase survey density(double or quad). The FM256 instrument can be used in either scanning mode to search rapidly for disturbed areas, or in logging mode, to collect detailed data in parallel or zig-zag traverse. The datalogging facilities, with integral sample trigger, provide powerful functions for fast and efficient surveying, keeping track of survey position, and giving both audible and visual indication of current survey position. Data can be collected at up to 16 samples/m and stored in a 256000 reading memory.

Figure 7.Fluxgate gate Gradiometer (FM256) with stand

The dual gradiometer system uses two instrument carried together, 1m apart, either to double the speed at which a survey can be made or to increase the sampling density of a survey. Basing the system on two individual gradiometers gives optimum flexibility since they can also be used separately at different sites when required. The FM256 comes

11

complete with robust padded carry case for transportation, instruction ,manual, data dump lead, universal charger and adapter pins sets, balance alignment tools, screwdriver and battery holder fir alkaline batteries. The carrying case cut-out has compartments designed for the standard items provided and also compartments for other accessories.

3.2 Preparing the Fluxgate Gradiometer First the fluxgate gradiometer was taken out into the field and switched on to allow the electronics to stabilize with the field environment. Stabilization time of at least 15 minutes, and preferably longer, is required before a detailed survey is started. The fluxgate gradiometer (FM256), with 0.5 m sensor spacing, was then set to a sensitivity of 0.1 nT, required grid length, width, sample interval 0.5 m and traverse interval also 0.5 m to survey those two site. A zero reference point was used for balancing and zeroing the instrument to match individual grids and to ensure high quality data. The instrument was positioned vertically over the selected zero reference point, normally of a stable magnetic field, and aligned in the N-S direction. Two additional steps were then followed: The first step is the adjustment of the balance control keys to get nearly the same reading at the upper and lower positions. The second step is the adjustment of the two sensor keys to get nearly the same reading through the two main geographic directions (N-S and E-W). After that, we zeroed the instrument over the same point and in the same direction of the planned traverse lines. This step was repeated after completing each grid.

Figure 8.Programm setting of the FM256 fluxgate gradiometer

12

3.3 Procedure of the field survey Rapid assessments of large areas may be made using scanning techniques. Scanning requires an experienced operator who can assess the background level response and compare this with any anomalous responses. Traverses are made by walking forwards and backwards over the Site, as shown in fig.10. In a detailed area survey the site is partitioned into a number of grids, typically 20m or 30m.But in our case we used grid (20mX10m) for the first site and (10mX10m) for the second site, which are in turn subdivided into a mesh of smaller squares, typically 1m square. One or more instrument readings are taken within each 1m square, giving a detailed and systematic coverage of a site. Figure11, sketch map shows a site subdivided into (20mx10m) & (10mx10m) squares and that are used during our survey. Readings are taken by walking along a traverse. The separation between readings along a traverse is known as the sample interval. The separation between traverses is known as the traverse interval (fig.9).The length of the traverse specifics the grid length. The number of traverses and the traverse interval define the grid width. All these measurements are specified in terms of meters. For surveying, first we located the scanning area and than we placed markers at the ends of the traverses to act as visual guides whilst walking. We removed all magnetic items from our body and from the site. Than the gradiometer was taken a start position about 1m outside the grid, in line with the first traverse direction, than we pressed enable log so the logging display is shown. At the start position, we orientated the gradiometer appropriately and started walking in the traverse direction (fig.10).In figure10, traverse direction is along the X-axis and the dimension of the grid in the Y direction is known as the grid width. We set off walking at a constant pace and as the sensor tube passes over the edge of the grid we pressed the start/stop switch. We took great care not to tilt the instrument or change the body posture when pressed the switch. Thirdly, we continued to walk at a constant pace, and ensured the 1m marker ‘beeps’ coincide with the 1m marks on the perpendicular guide line. Fourthly, for the next traverse, again we took position the gradiometer at a start position which should be about 1m outside the grid. We repeated the previous process until the whole grid was surveyed. We surveyed the site using the same procedure, checking and readjusting alignment, balancing and zeroing periodically. More than one survey session may be required to cover a site; depending on the size of site and memory limitations.Than we kept the gradiometer powered up between two sessions to optimize stability. The first study area and the second study area are shown in the following fig.10. The two study area is situated side by side and also we can see the start point of the gradiometer survey in that same figure. We started our survey form the start point and than proceeded zigzag way along Y-axis up to the last ends of the site.

13

Site Star

•

Alignment & Balance station & zero reference Finish station

Grid

Grid Length

Grid

Grid Width

Sample Interval Traverse Interval

Traverse Direction Figure 9.Diagram to illustrate the relationship between Site & Grid (Source: FM256, instruction manual)

Figure 10.Photograph of the study area showing the axes of the fluxgate gradiometer survey

14

3.4 Acquisition of magnetic data: The study area was surveyed using fluxgate gradiometer (FM256 of Geoscan Research, 2004). We conducted a gradiometer survey in Suchon area, Gongju. The corners of the whole area were marked with wooden sticks. These were well hammered in the ground for repetition of the measurements. For carrying out the measurements of the vertical gradient of the geomagnetic field, the total studied area of 40mX20m (800m²) & 20mX20m (400 m²) was divided into four grids respectively. Distances were measured using a non-magnetic tape and traverses were outlined with a non-magnetic rope marked at 1m interval, than right angles were taken using a small optical square. After that the fluxgate gradiometer was prepared and taken out into the field for surveying. Finally, we downloaded the data into a computer, examined the results graphically and decided if the survey area needs extending or parts need resurveying. Thus we completed the survey successfully. For downloading the raw magnetic data we used the Geoplot software. The ‘new grid input template’ of Geoplot software was used for data acquisition. In the acquisition section we enter site name: Suchon; Map reference: metal road; Dir. 1st traverse: NE; Grid length: 20m; Sample interval: 0.5m; Grid width: 10m and Traverse mode: zigzag. We also filled up the instrumentation menu bar as Survey type: Gradiometer; Instrument: FM256; Units: nT; Range: auto; Log zero drift: Off; Baud Rate: 19200; Data Format: Hex D+R and Computer Buffer size: 32767.After writing the above necessary information we should save it correctly. Geoplot has very specific requirements for the data format. So, it is extremely important to define and use a survey strategy before downloaded the data from the Geoscan instrument. For each site, always we should use grids that have the same size, survey pattern and orientation. The initial traverse must be in a clockwise direction from the first reading, with respect to a plan view of a grid. We should choose the traverse direction so that it will be horizontal with respect to the computer screen when grids are finally assembled using a master grid. Always we should complete unsurveyed part of a grid with dummy readings using the instrument Dummy. Whilst logging data into the memory of Geoscan research instrumentation we must not change grid size, sample or traverse interval, log zero drift status etc. It is extremely important that the input details are correct. If they are inaccurate this may result in invalid data which may be uncorrectable at a later data, except by tedious manual re-entry. In some circumstances even this may not be possible. Once we are satisfied the input details, we should press the download data key to download the data into PC. Simultaneously we should connect the data lead between the instrument and PC, switch the instrument on and wait 1 second. Data will be downloaded within a short time. Finally, the distribution of the grids and the raw gradiometer data are presented in fig.11.

15 20m

N G1 20m

G2 40m

G3

G1

G2

G3

G4

20m

G4

nT

Figure 11.Sketch map showing the distribution of the field grids & Magnetic images of the two study area showing the raw gradiometer data

16

CHAPTER 4

MAGNETIC DATA PROCESSING Since the acquisition of the magnetic data has been achieved through the fluxgate gradiometer (vertical magnetic gradient), the enhancement of the obtained data is the next step and the main target of this chapter in order to facilitate the interpretation process. However, the processing and the treatment steps will concentrate on the obtained data of the vertical magnetic gradient of the geomagnetic field. The processing of the magnetic data of the vertical magnetic gradient have been achieved by using Geoplot software (Geoscane Research, 1994), and than extracting to Surfer software to be more representative for interpretation. In the present processing techniques we detected several field errors like noises, grid edge discontinuity and stripe effects. We want to remove the errors and enhance the presentation of the resultant magnetic images and profiles. So, we used several processing functions to remove our field errors.

4.1 Applied processing function The obtained raw data of the study area (fig.12) have been shown the presence of edge discontinuities between the surveyed grids, stripping effects between the marked traverses and displacement in the location of some features. Moreover, the presence of scattered iron fillings is expected even after cleaning the top surface of the study area. The field errors concerning grids and traverses are normally occurred with the gradiometer survey due to the sensitivity of the instruments to any kind of movements (e.g. windy days), measuring the last limit of grid two times, irregularities of the ground surface and instability of the operator. The process menu and toolbar of Geoplot software provides a comprehensive range of functions for manipulation of all data types, together with specific routines to correct for data collection artifacts such as edge matching and drift correction. Some functions are specifically designed for Geoscan research instrumentation but all may be equally applied to other instrumentation data sets. Mathematically, any real bipolar or monopole two dimensional data array may be processed. The necessary process functions which we are applied too process our raw magnetic data is mention below.

17

Clip: The Clip function can be used to clip, or limit, data to specified maximum values. This can improve graphical presentation and also forms a useful pre-process procedure for many other functions. Zero Mean grids (ZMG): The Zero Mean Grid function scans the data set; grid by grid .It sets the background means of each grid to zero. It is useful for removing grid edge discontinuities often found in gradiometer or similar bipolar data. Zero Mean Traverse (ZMT): The Zero Mean Traverse function sets the background mean of each traverse within a grid to zero. It is useful for removing striping effects in the traverse direction which often occur in Fluxgate gradiometer data. Despike: The Despike function can be used to automatically locate and remove random, spurious readings often present in resistance data and locate and remove random “ iron spikes” often present in gradiometer and magnetometer data. It operates over the whole of the data set. Low Pass filter (LPF): The Low Pass filter function may be used to remove high frequency, small scale spatial detail. It is useful for smoothing data or for the enhancement of larger weak features. It can operate over the whole of the data set, or any inclusive or exclusive block. High Pass Filter: the high pass filter function may be used to remove low frequency, large scale spatial detail. A typical use is the removal of slowly changing geological background response commonly found in resistance survey. It can operate over the whole of the data set, or any inclusive or exclusive block. High pass filter scans the data set with a Gaussian or uniformly weighted window, which may be square or rectangular. It calculates the weighted average within the window and subtracts this from the central reading in the window. All other readings remain unchanged and dummy readings are ignored. Interpolation: function may be used to increase or degrease the number of data points in a survey. Increasing the number of data points can be used to create a smoother appearance to the data. The Interpolate function scans the data set in the X or Y direction. Sometimes we also use the following functions if necessary:

18

Median Filter: This filter function may be used to remove random spurious readings present in survey data and smoothes the data at the same time. It is most useful for high sample density data. It can operate over the whole of the data set, or any inclusive or exclusive block. Median Filter scans the data with a uniformly weighted window, which may be square or rectangular. It sorts the date in the window into ascending order and picks the middle or median value as the new value. All other readings remain unchanged and dummy readings are ignored. Periodic Filter: The Periodic Filter function may be used to remove or reduce the amplitude of regular, periodic features. There are also other Geoplot processing functions, such as Power, Randomize, Search and Replace, Spectrum, etc. Search and replace: This function looks for numbers in a specified band and replaces them by another specified number. It is a general purpose numeric tool with a wide variety of application when used in conjunction with other process functions. For example, regions which are strongly perturbed by nearby iron fences, pipelines etc can be converted to dummy regions, allowing other statistical functions to perform correctly. It can operate over the whole of the data set, or any inclusive or exclusive block. Standard Deviation/Variance Map: The standard deviation/variance Map function replaces the data by either the local variance or standard deviation, which parameter is chosen. A graphics plot of this new data set indicates areas of statistically different activity. It only operates over the whole of the data set.

4.2 Sequence of image processing The obtained raw data have been presented through the Geoplot program as a shade plot (grey 99.ptt).The gradient of the vertical component of the geomagnetic field ranges from -90.85 to +85.43 nT for the first image and -149.75 to +153.00nT for the second image (fig.11).The resultant magnetic image looks unclear and disturbed with many noises that make it difficult to interpret easily and simply. To remove the field errors and enhance the obtained data, the following processing sequences have been done: (I) (II) (III) (IV)

Presentation of raw data with effects Clipping the data sets Removing of grids discontinuities and traverses strips Removing of random iron spikes

19

(V) (VI)

Smoothing the acquired data using Low Pass Filter & High Pass Filter Finally interpolate the data set to create a smoother appearance

N

Stripe effects within grids

Noise Spikes

Grid edge discontinuity

Figure 12.Different effects showing in the raw gradiometer data

(grey 99.ptt)

(I)Presentation of raw data with effects: Neutralization of the effect of major geological and ferrous responses involves replacement with dummy readings, where necessary, though in many cases the process functions are able to handle these situations without replacement. The above raw magnetic image appears blurred and disturbed by noise & traverse stripe effects. The defects of the images includes: (a) Stripe effects within grids, (b) Noise Spikes and (c) Grid edge discontinuity. So, we are going to remove the above image effects gradually. But the order of the image processing can be very important for some functions. For example, we should Despike before applying a low pass filter, to avoid the spike energy smearing out.

20

(II) Clipping the data sets: But before applying these function we should clipped the

data. The clip function is used to clip, or limit, data to specified maximum and minimum values. This can improve graphical presentation and also forms a useful pre-process procedure for many other functions. It can operate over the whole of the dataset, or any inclusive or exclusive block. In our case we used different clipping values (-2, +2; -10, +10 & -20, +20), to compare the difference between the original dataset and the clipping dataset. But finally, we choose the clipping value (Min= -2, Max= +2).

Raw gradiometer data

Clipping, Min=-2, Max=+2

Figure 13.Difference between



the original raw data and the clipping dataset (grey 99.ptt)

In above figure, we can see the difference among the clipping dataset. The Clip function was used to clip the dataset in three different ways. Such as Clipping, Min= -2, Max= +2; Clipping, Min= -10, Max= +10: Clipping, Min= -20, Max= +20. (III)Removing of grids discontinuities and traverses strips: After Clipped, we used the Zero Mean Grid (ZMG) & Zero Mean Traverse (ZMT) function. The grid edge discontinuities arise due to improper choice The stripes and joints between traverses and grids, which are obvious in most of the survey area, and due to the tilting of the instrument during the zig-zag survey process, have been treated by the

21

application of Zero Mean Grid (ZMG), and Zero Mean Traverse (ZMT).The Zero Mean Grid function scans the data set, grid by grid. It evaluates the mean of a grid, using only those values below a specified threshold. The calculated mean is then subtracted from that grid, resulting in the background with zero mean. The process is repeated for each grid. Dummy readings are ignored and remain unchanged. The Zero Mean Grid parameters group requires one entry, a Threshold value. The Threshold can be tabbed through predefined values: 0.25, 0.5, 1.0, 1.5, 2.0 or 2.5 standard deviations. On the other hand, The Zero Mean Traverse function can be used to remove stripe effects within grids. It can also correct for any slope or drift along a traverse if the Least Mean Square fit parameter is set to on. The Zero Mean Traverse function scans the data sets, grid by grid, traverse by traverse. It evaluates the mean of a traverse in a grid, using only those values below an internal threshold (an iterative process is used to define the threshold, depending on the data set itself). The calculated mean is then subtracted from that traverse, resulting in the background with zero mean. The process is repeated for each traverse and each grid. Optionally, a least mean square straight line fit and removal can also be applied to each traverse to further reduce striping effect caused by drift within a traverse. Dummy readings are ignored and remain unchanged. The present work has been used a Threshold value equal 2.5 Standard Deviation to remove the grid discontinuities, and Zero Mean Traverse (ZMT), Grid=All, LMS=on, thresholds not applied, reduced striping along the traverses. (fig.14 & fig.15).First we removed the grid edge discontinuity effect form the raw data set. So we applied ZMG (zero mean grids) function on the raw data. But unfortunately, the ZMG function couldn’t remove the grid edge discontinuity effect from the data set, which is shown in figure14. Secondly, we applied the ZMG function (zero mean traverses) and ZMG function could remove the grid edge discontinuity effect from the dataset successfully (fig.15). The Zero Mean Traverse parameters group requires one entry, the Least Mean Square Fit status. This may be tabbed to either On or Off. The present work has been used the Least Mean Square Fit setup to “On”. The required parameters for applying the Zero Mean Grid and Zero Mean Traverse functions have been shown in the Table(1). And the resultant magnetic image after the application of ZMG and ZMT has been presented in (Fig.15).

22

A

B

Figure 14.Data set after the application of Zero Mean Grid function to remove Grid Edge discontinuity effect. [Threshold (STD Dev=1)]Image A, raw data & Image B, after the application of the ZMG function

C

D

Grid edge discontinuity effect present

Grid edge discontinuity effect removed

Figure 15.Data set after the application of zero mean traverse function to remove grid edge discontinuity and stripe effects.

23

The applied ZMT function was, lest mean square fit =off; grid num=all; positive threshold= +5, negative threshold= -5.The grid edge discontinuity effect which couldn’t remove by the ZMG function, was removed by the ZMT function successfully. Image C, before the application of ZMT function & Image D, after the application of the function. The stripe effects within grids also removed by the ZMT function. So, we think the ZMT function is for effective than ZMG function to process our data.

Process (Advanced) File Function Choice S Sitename

Suchon01

G Filename

Suc-1

------General ------

----Numeric----

T Cut+ Combine J Absolute

F File Format

Composite

O De Slope

V Save

Off

K De Spike G De Stagger

L Reload

Off

E Edge Match H High Pass I

Parameters

L Low Pass

C Functions Choice

Z

-- (Zero Mean Grid) - ---

0.25

N Threshold (SD) C

Y

C Clip B Compress M Multiply N Power #

Randomise

Q Search+ Repl Y Zero Mean Grid

----Analysis-----

Z

S Statistics

Zero Mean Traverse

F Spectrum V Variability

Functions Choice

(Zero Mean Traverse)--

Interpolate

A Add

Off

N Least Mean Square. Fit

Table1.The formal presentation of application of Zero Mean Grid and Zero Mean Traverse functions in Geoplot software (Geoscan Research, 1994).

24

(IV)Removing of random iron spikes:Now we want to remove the noise spikes from our data set. So, we used the Despike function.The Despike function can be used automatically to locate and remove random iron spikes, which is present in our gradiometer data. It operates over the whole of the data set. As the archaeological sites are generally contains some modern surface and near surface iron objects distributed randomly, so the study area should be clean of their effect, which can clutter the survey results and make difficulty in the interpretation. Using Despike function (K) can minimize the effect of these objects and enhance the data for other processing functions. The Despike parameters group requires four entries: X and Y radii, Threshold and spike replacement. The X and Y radii determine the size of window used for the mean and standard deviation determination. They are entered as positive integer values, which may be in the range 0 to 10 (readings). The Threshold can be tabbed through predefined values: 0.5 to 3.0 standard deviations. The spike displacement can be tabbed between Mean or Standard Deviation. In the present study, we used different Despiking parameters one after another to see the change exactly. Despiking parameter was, (Xrad:1,Yrad:1); (Xras:2,Yrad:2);(Xrad:3,Yrad:3).But if we scrutinized, we can see, Despiking parameter (Xrad: 1, Yrad: 1)can remove the noise spikes in a best way.(fig.16). The required parameters for applying the DeSpike function have been shown in the Table (2). Process (Advanced) File

Function Choice

S Site name

Suchon01

G Filename

Suc-2

------General ------

-------Numeric----

T Cut+ Combine

F File Format Composite

J

Absolute

O De Slope A Add

V Save

L Reload

Off Off

K De Spike

C Clip

G De Stagger

B Compress

E Edge Match

M Multiply

H High Pass

N Power

25

I

Interpolate

#

Randomise

L Low Pass

Q Search+ Repl

D Period. Dect

----Analysis-----

Y Zero Mean Grd

S Statistics

Z Zero Mean Trv

F Spectrum V Variability

Parameters C Functions Choice

K

----- (De Spike) ----X X-Radius

1

Y Y-Radius

1

N Threshold (SD)

2.0

R Spike

Mean

Replacement

Table2. The formal presentation of application of DeSpike functions in Geoplot software (Geoscan Research, 1994).

Zero Mean Traverse applied data

Despiking parameter: Xrad: 1; Yrad: 1; Thres: 1; SpRep=Mean

E

F

Noise spikes present

Noise spikes removed

Despk.parameter: Xrad: 2; Yrad: 2; Thre: 2; SRep=Meadian

G

Noise spikes partially removed

Despk.parameter: Xrad: 3; Yrad: 3; Thre: 3; SRep=Std Dev

H


Figure 16.Difference between the ZMT applied data & Despiking data. Image E before Despiking And Image F, G, H after Despiking (grey99.ptt).

26

Noise spikes present


Noise spikes totally removed

Magnetic anomaly

Figure 17.Trace plot before and after Despiking which shows anomaly of the subsurface features

(V)Smoothing the acquired data using Low Pass Filter & High Pass Filter:

In the 5th step, we used the Low Pass Filter (LPF), X=1, Y=1, Wt=Gaussian, to remove high frequency, smooth the data and enhance the larger weak features.(fig.18).It can operate over the whole of the data set, or any inclusive or exclusive block. Low pass filter scans the data set with a Gaussian or uniformly weighted window, which may be square or rectangular. It calculates the weighted average within the window and replaces the central reading in the window with this value. All other readings remain unchanged and dummy readings are ignored. In our case we used different window to smooth the data. But the window=2, weighting=Gaussian, parameter was best to remove high frequency, small scale spatial detail and smoothing the data. In our processing techniques, Low pass filter is one of the most important functions to process the raw data, because it can suppress higher frequency components such as noises in the data whilst at the same time preserving low frequency, large scale spatial detail and improve the visibility of larger, weak archaeological features. It can also be used to improve the appearance of relief or artificial sunshine plots, especially if data has been sampled at 0.5m intervals or better. Also we used High pass Filter to remove low frequency in some cases (fig.19).

27

Low pass Filter X window=1 Y window=1 Weighting=Gaussian



Low pass Filter X window=3 Y window=3 Weighting=Uniform

Figure 18.Comparison of the process magnetic image of the study area after application of LPF using different window & Weighting function (grey99.ptt). Process (Advanced) File S Site name

Function Choice Suchon01 Suc-3

G Filename F File Format V Save L Reload

------General -----

------Numeric-----

T Cut+ Combine Off Off

J

Absolute

O De Slope K De Spike G De Stagger E Edge Match H High Pass I

Interpolate

L Low Pass D Period. Dect Y Zero Mean Grid Z Zero Mean Traverse

A Add C Clip B Compress M Multiply N Power #

Randomise

Q Search+ Repl ----Analysis----S Statistics

28

C Functions Choice

L

F Spectrum

---- (Low Pass) ---X X-Radius Y Y-Radius W Weighting

2 2 Gaussian

Table3. The formal presentation of application of Low Pass Filter function in Geoplot software (Geoscan Research, 1994).

High pass Filter X windo=10 Y windo=10 Weighting=Uniform



High pass Filter X windo=10 Y windo=10 Weighting= Gaussian

Figure 19.Comparison of the process magnetic image of the study area after application of HPF using different window & Weighting function (grey99.ptt).

The following figure20 is the colored magnetic image after application of LPF. We used the color image to increase the visualization of subsurface features.

29

Figure 20.Colored process magnetic image of the study area after application of LPF to see the contrast of the magnetic anomaly with surroundings

(VI)Finally interpolate the data set to create a smoother appearance:

Finally, we interpolate Y, Expand –SinX/X, x2 & Expand-Linear, x2 than Interpolate X, Expand –SinX/X, x2 & Expand-Linear was applied to create a further smoother appearance to the data (fig.21).

Figure 21.Magnatic image after application of interpolation

30

Raw data

AFT E R CLIPPING

AFT E R ZM G, ZMT

AFT E R DESPIKINNG

AFTER INTERPOLATION

AFTER LPF

Figure 22.Complete Process sequence of the first magnetic image

In above we can see the complete process sequence at a glance. The sequence of processing of the first magnetic data was initiated as just above. The above process sequence is not unique for other gradiometer data. But in case of our magnetic data the process sequence was perfect as we can see in the above figure. For the Processing of our second magnetic image we used the same procedure and functions as the previous one. The same processing sequence is shown in as follows:--

31

N

Stripe effects within grids Grid edge discontinuity

Noise Spikes

Figure 23.Different effects showing in the second raw gradiometer data

The

above raw magnetic image is for the second study area. The second image looks also blurred and disturbed by noise & traverse stripe effects. So we have to process the second magnetic image using the same sequence of Geoplot functions as we used for the first image. The defects of the images includes:(a) Stripe effects within grids,(b) Noise Spikes and (c) Grid edge discontinuity, that is same defects as the first image. So we are going to apply the following sequence of processing to process the second magnetic image. Clipping

ZMG & ZMT T


Despiking

L PF


Interpolation

Process data


Figure 24.Different clipping function applied in the original raw data (grey99.ptt)

Figure 24 is showing the clipping dataset. The Clip function was used to clip the data in three different ways. Such as Clipping, Min= -2, Max= +2; Clipping, Min= -10, Max=

32

+10 & Clipping, Min= -20, Max= +20.After using different clipping function, we can see the difference among the dataset. Secondly, we used the ZMG and ZMT function. We used these two functions to remove the grid edge discontinuity and stripe effect form the raw dataset. After removing these effects successfully (fig.25), we want to remove noise effects from the dataset. Grid edge discontinuity present

Grid edge discontinuity effect has removed

Figure 25.Data set after the application of ZMG & ZMT function to remove Grid edge Discontinuity & Stripe effects, [Lest Mean Square fit =Off; Grid num=all; positive threshold= +5,negative threshold= -5]

Zero Mean Traverse applied data Noise spikes present

Despiking parameter: Xrad: 1; Yrad: 1; Thres: 1; SpRep=Mean Noise spikes removed

Despk.parameter: Xrad: 2; Yrad: 2; Thre: 2; SRep=Meadian Noise spikes removed

Despk.parameter: Xrad: 3; Yrad: 3; Thre: 3; SRep=Std Dev Noise spikes present

Figure 26.Difference between the ZMT applied data & Despiking data

So, we used the Despike function. The Despike function can be used automatically to locate and remove random iron spikes, which is present in our gradiometer data. It operates over the whole of the data set. The Despiking function employing a threshold value of 3, replace=mean, X=1, Y=1 removed the effects of random iron spikes (fig.26). Figure.26, showing the difference between the Zero mean traverse applied data and Despiking data.

33



High pass Filter X window=2 Y window=2 Weighting=Uniform

High pass Filter X window=10 Y window=10 Weighting=Uniform

Figure 27.Application of LPF & HPF using different window & Weighting function (grey99.ppt)


Figure 28.Colored LPF magnetic image of the study area

Here we used LPF function to remove high frequency, smoothing the data and enhance the larger weak features, simultaneously we used HPF function to remove low frequency, larger scale spatial detail (fig.27).Low pass filter & High pass filter function scans the data set with a Gaussian or Uniformly weighted window, which may be square or rectangular. But the X window=2, Y window=2, weighting=Gaussian, parameter was best for both filtering techniques to remove high frequency and low frequency respectively. Removal of such local variability in the data can improve the visibility of sub-surface weak archaeological and surface features too. We used also the colored LPF magnetic image to observe the subsurface features correctly (fig.28).

34

Raw data

Af t e r Despiking

Af t e r Clipping

Af t e r ZM G, ZMT

Af t e r LPF After Interpolation n

Figure 29.Complete Process sequence of the second magnetic image

Complete process sequence of the second magnetic image. The sequence of processing of the second magnetic images was initiated as above (fig29).After applying all the necessary processing techniques for the both magnetic images, the resulted final Processed magnetogram(fig.30) was very impressive to see the buried subsurface archaeological features of the study area.

35

N

nT

Figure 30.Complete Process magnetic image of the study area after application of necessary process functions

36

4.3 Data & image analysis Image analysis is used to extract information contained in images. Image analysis aims at detecting, diagnosing and describing the texture and geometry of the images. The information content and, consequently, the meaning of an image are function of the processes used for its generation and acquisition. The features through which an image is identified and categorized are of spectral and geometric (or morph metric) nature. All of the basic image manipulation functions are rotation, brightness, contrast, hue, saturation, gamma adjustments, inverting and resizing (Source: Surfer-8 Help menu). Usually we can analyze magnetic data by different types of images like Vector map, Contour map, Histogram, Relief map, Wire frame map, 3D surface map etc. On the other hand, data analysis is essential to apprehend the dataset correctly. The following dataset are collected and sorted from the original Geoplot dataset of the survey area. The data are sorted and arranged in the following table 4 because of there exceptional effect to create anomaly. The dataset are as follows: Profile Name

Raw data Maximum

Minimum

Processed data Maximum

Nature of anomaly

Minimum

AB

202.1

-49.5

57.4

-26.7

Monopole

CD

203.7

-18.3

50.7

-13.7

Dipole-Dipole

EF

203.4

-125.9

65.0

-21.3

Dipole-Dipole

GH

198.0

-15.5

2.1

-8.2

Monopole

Table 4.Data table is prepared based on original Geoplot data to show the difference of raw data & processed data and to find out the nature of anomaly

We can see in the above table, that the raw data posses the largest value than the processed data. In case of raw data the highest value is, 203.7 and the lowest value is,125.9.For the processed data, the maximum value is, 65.0 and minimum value is,-26.7. The difference of the maximum value, (203.7-65.0) =138.7 & the difference of the minimum value is, (-125.9+26.7) = -99.2. So, we can say apparently that, the magnetic field intensity of the processed data has reduced 138.7 nT. Thus the image was smoother to see the subsurface archaeological features clearly. Three dimensional modeling of magnetic anomalies is complex. Probably the most convenient methods are to approximate the causative body by a cluster of right rectangular prism or by a series of horizontal slices of polygonal outline. Based on the dipolar nature of magnetic anomaly we named the profile AB, CD, EF, GH as monopole, dipole-dipole, dipole-dipole and monopole. But the anomaly shape is not closely related to the geometry of the causative body. So we can’t determine the shape of the subsurface body by the help of magnetic anomaly.

37

Grid Vector Map & Contour Map: direction and magnitude of grid vector map can be derived from one grid. The arrow symbol points in the downhill direction and the length of the arrow depends on the magnitude, or steepness of the slope. A vector is drawn at each grid node unless some nodes are skipped by changing the frequency. The data of first site and second site represents the following grid vector map (fig.31). A vector map is different than a grid map. Instead of dividing the map area into equally sized cells, a vector map is made up of objects, or obstacles. In the first vector image we can see three round obstacles at the top left most corners. These three obstacles or disturbance indicate the presence of subsurface features. The second vector map represents only one disturbance of the magnetic domain. The disturbance of magnetic domain means presents of unknown subsurface features. The following contour map (fig.32) also supports the grid vector map for its different contour values. The lowest contour value is ‘-20’ shown in fig.32.

Figure 31.Grid vector map of the study area

38

Figure 32.Contour map of the Study area

Shaded Relief Map: is raster maps based on grid files. These maps (fig.33) use colors to indicate the local orientation of the surface relative to a user defined light source direction. The light source can be thought of as the sun shining on a topographic surface. Portion of the surface that face away from the light source reflects less light toward the viewer, and thus appear darker. The colors on a shaded relief map are based on the reflectance value of zero means that no light is reflected toward the viewer.A reflectance value of one means that all incidents light are reflected toward the viewer. Shaded relief is a method for representing topography on maps in a natural, aesthetic, and intuitive manner. The following shaded relief map is an overview of the elevation of our study area. In the map, we can see six elevated picks as anomaly. These all picks are the basic characteristics of shaded relief map. The elevated picks are surrounded by subduction zone, which means the negative magnetic anomaly.

39

Figure 33.Shaded relief map of the study area

Wire frame Map: The following Wire frame maps are three dimensional representations image of the survey area. The Wire frames map is created by connecting Z values along lines of constant X and Y. At each XY intersection the height of the wire frame is proportional to the Z value assigned to that node. The number of columns and rows in the grid file determines the number of X and Y lines drawn on the wire frame. Wire frames can display any combination of X lines, Y lines, or Z lines. On the Wireframe, X lines correspond to the columns in the grid file and Y lines correspond to rows in the grid file. Figure34 represents the wire frame Map of the study area before data processing and Fig. 35 represents the Wire frame Map of the study area after data processing. There is some basic difference between these two ware frame maps. In fig.34 the surface of the 3-D wire frame looks rough unsmooth than the wire frame maps of fig.35.In Fig.34 the first image is showing high Z values and several positive magnetic anomalies and the second image is showing low Z values (negative magnetic anomaly).

40

In the first wire frame map the spikes and the four grids can be differentiate easily. But after data processing the wire fame maps look smoother because of the absence of iron spikes. In figure35 the surface of the wire frame map are more undulated and six positive anomalies are more developed than the previous one.

Figure 34.Wire frame Map of the study area before data processing

41

Figure 35.Wire frame Map of the study area after data processing

3D Surface Map: are sometimes referred to as surface charts or surface maps. 3D surface map are trivariate plots in which a comparison of 3 measures is presented. 3D surface map includes grid lines in the X and Y direction as well as contour lines. Fig.36 and Fig.37 is 3D surface map of the study area before and after data processing respectively. There are some clear differences between the Fig.36 and Fig.37.We can see 14 spikes or positive magnetic anomaly circled by ‘I’ (fig.36) before data processing but after processing the data, six positive anomalies circled by ‘K’ (fig.37) are developed and

42

others are removed. These positive anomalies indicate significant subsurface features. In Figure 37, first image, a depressed zone centered by the positive anomaly represents the negative anomaly, which may be subsurface structure too. The other negative anomaly marked by ‘L’, (fig.37) may be the subsurface ditch or gap.

I

J

Figure 36.3D surface map of the study area before data processing

43

K

L

Figure 37.3D surface map of the study area after data processing

44

CHAPTER 5

INTERPRETATION OF MAGNETIC DATA Before the interpretation of magnetic data we shall interpret the magnetic anomaly. Because, magnetic anomaly is essentially related to the interpretation of magnetic data. The interpretation of magnetic anomalies is similar in its procedures and limitations to gravity interpretation as both techniques utilize natural potential fields based on inverse square laws of attraction. There are several differences, however, which increase the complexity of magnetic interpretation. The magnetic anomaly of a finite body invariably contains positive and negative elements arising from the dipolar nature of magnetism. The intensity of magnetization is a vector, and the direction of magnetization in a body closely controls the shape of its magnetic anomaly. Thus bodies of identical shape can give rise to vary different magnetic anomalies. For this reasons magnetic anomalies are often much less closely related to the shape of the causative body than are gravity anomalies. (Source: An introduction to Geophysical Exploration, by P. Kearry) Negative pole

+

Total field magnetic anomaly

Combined effects of both poles Positive pole

Magnetic North Depth

+

B

Figure 38.Standard magnetic anomaly of a subsurface feature approximated by a dipole

45

The above fig.38 is the total field magnetic anomaly of an elongated body approximated by a dipole. Indirect interpretation of magnetic anomalies is similar to gravity interpretation in that an attempt is made to match the observed anomaly with that calculated for a model by iterative adjustments to the model. Simple magnetic anomalies may be simulated by a single dipole. Such an approximation to the magnetization of a real geological body is often valid for highly magnetic ore bodies whose direction of magnetization tends to align with their long dimension. In such cases the anomaly is calculated by summing the effects of both poles at the observation points, however, require a different approach. On the other hand the magnetic anomaly of most regularly shaped bodies can be calculated by building up the bodies from a series of dipoles parallel to the magnetization direction. We have sketched several dipole and monopole anomalies and some other anomalies for geological bodies and archaeological features at various orientations and different inclinations for the field. (fig.39, fig.40, fig.41).The main objective of showing the following twelve standard anomalies are to correlate with our four original anomalies showing in fig.42. In fig.39 Dipole and Monopole exhibit different shape and sizes. On the other hand fig.40 exhibits various anomalies for various geological bodies such as Dipping dike, Sphere, Vertical cylinder, Anticline or Ridge etc. But, figure 41, represents four anomalies of archaeological features.

Monopole Dipole

F

F Monopole Dipole

Figure 39.Free Hand Sketch of Dipole and Monopole for various inclinations

46

Sphere Dipping Dike

Vertical Cylinder

Anticline or ridge

Figure 40.Anomalies for geological bodies at various orientations and different inclinations for the field

Kiln Baked Brick wall

Shallow Tomb

Shallow grave

Fire Pit

Figure 41.Typical magnetic anomalies of common archaeological features

47

A

B

C E

D

F

G

nT

H

nT

250

300


200 150

200

100 100

50 0

0

A

-50

.

B

C

D

-100

0

4

8

12

16

PProfile-1(AB) rofile AB

20

M

0

4

+ -

--

8

12

Profile CD

16

M

--

Monopole

+

+ ++

-

-

-

-

Dipole

nT

nT

300

200


200 100

H

G

0 -200

E

0

F

-400 -600

-100 -800

-200 -1000

0

4

+ ---

16 M

8 12 Profile-3(EF) Profile EF

Dipole

-

+

+

+

-

+

0

5

10

15

20

- - --

-

-

-

M

25

Profile GH

Monopole

Dipole

Figure 42.The four magnetic profiles over our study area based on raw magnetic data

48

The above profiles (AB; CD; EF; GH) indicates total field magnetic anomaly of our study area. AB & GH profile are monopole, in other hand, CD & EF profiles are Dipole-Dipole. In above fig.42, there are four magnetic anomalies or profiles (AB; CD; EF; GH) over our study area based on raw magnetic data. If we correlate our four original anomalies with the above standard anomalies, we can’t see much similarity among those anomalies. But the anomalies of the processed data (fig.43) have some similar shape, size and orientation with the standard anomalies. Because of the dipolar nature of magnetic anomalies, trial and error methods of indirect interpretation are difficult to perform manually since anomaly shape is not closely related to the geometry of the causative body. The magnetic anomaly of a body of regular shape is calculated by determining the pole distribution over the surface of the body using the specific equation. Each small element of the surface is then considered and its vertical and horizontal component anomalies are calculated at each observation point. In vice versa, we can determine only the polar characteristics of the archaeological features from our anomalies. So we determined only the monopole and dipole characteristics of our archaeological features and presented in the above figure. nT

nT 60 Total field magnetic anomaly

50 40 30 20 10 0 - 10 - 20

Profile CD

Profile AB

nT

nT 4

Total field 2 magnetic 0 anomaly -2 -4 -6 -8 -10

Profile EF

Profile GH

Figure 43.Four anomalies/profile of the study area based on the process data

The anomalies formed after data processing is some how different than previous one. Profile AB and profile GH slightly change than previous one but still represents

49

monopole.Though new profile CD is dipolar, yet size is different than previous one. Further, Profile EF was dipolar, but after processing it changed into monopole. Quartzite rocks (Positive anomaly)

Porcelain Potteries

(Positive anomaly)

Blank room structure (Negative anomaly)

-

Quartzite wall (Negative anomaly)

Other potteries Stone mound tomb Surrounded by quartzite rock Small tomb

Figure44. Northern excavated site of Suchon and hand trace diagram of our potential survey area showing the correlation between magnetic anomalies and archaeological features for the authenticity of our study.

Above hand trace diagram is correlated with the previous small discoveries which excavated in the northern side of the Suchon area. Based on the correlation we can specify several archaeological features in our latest study area. Specially, the archaeological features are marked in the top leftmost corner and middle of the sites. The specified archaeological features may be of ancient Gongju contemporary to the Baekje period. The above figure44 shows us more accurate position of the subsurface

50

archaeological features to identify or explore. The first site is more important than the second site because we can see more anomalies in that site. The above 3D views and those anomalies also support the presence of subsurface archaeological features. The features in the left-hand side of the above image (fig.44), which also shows 5 picks of positive anomaly (fig.37), are accumulation of potteries surrounded by Quartzite wall; others are stone mound tomb surrounded by quartzite rock, potteries and porcelains. The second study are which represents a large negative anomaly partly surrounded by positive anomaly may be a blank room structure surrounded by quartzite rocks. There is a possibility to find out more historical relics and artifacts in the Suchon area. These all features are the destruction of the Baekje inheriting King Muryeong (501–523) style features.

Max

Depth: 0.5m Diameter: 1.07m

M N

/

Max 2

Depth: 1m Diameter: 2m

O

X½


Half-Width

Z F

P


Q


R Depth: 1.78m Diameter: 3.5m

Figure 45.Figure showing the estimated depths of archaeological features based on the half-width rules

51

Depth & Diameter: The half width rules are derived from the formula “Dipole and Monopole signatures in vertical and horizontal fields” The half-width is the horizontal distance between the principal maximum (or minimum) of the anomaly and the point where the value is exactly one-half the maximum value (fig.45). This rule is only valid for simple shaped forms such as sphere (dipole), vertical cylinder (monopole), and the edge of a narrow, nearly vertical dike (line of monopoles) in the Polar Regions. We have got about six excellent subsurface archaeological features. We calculated the depth and diameter of the six archaeological features based on half-width method. The archaeological features are marked as M, N, O, P, Q, and R. The depth and diameter of feature M is 0.5m, 1.07m.The depth and diameter of feature N is 1m, 2m; for feature O is 1.07m,2.14m and the depth and diameter of features P,Q,R are (1.25m,2.5m);( 0.89m,1.78m) &(1.78m,3.5m) respectively. In this method first we measured the diameter of the six archaeological features by the centimeter scale than we converted the diameter into field scale. In half-width method, we know the depth of archaeological features equal to the half of the real diameter. So, we divided the measured diameter by two and thus we got the depth of those features.

52

CHAPTER 6

CONCLUSIONS Detailed geophysical studies using the fluxgate gradiometer (FM 256) were carried out at the site of Suchon, Gongju. The aim of this study was to discover the subsurface archaeological features. The survey was carried out over two sites and produced some important points. The raw magnetogram of Suchon showed some field errors. The application of Zero Mean Grid was performed to remove the difference in grid to grid mean value and the Zero Mean Traverse function to remove the stripping effect caused along the traverse direction. The application of Despike function was essential for eliminating the presence of some spike readings that interfered with the archaeological features, especially in the central part of the study area. The final magnetogram was then displayed for each area, which is considered to be the best magnetogram for interpretation. The application of a High Pass Filter (HPF) to the gradiometer data is not favorable, but it was applied just to examine the effect of high pass filtering on certain archaeological features to represent them in higher resolution. The Low Pass Filter (LPF) was applied for smoothening and enhancing the final image in order to be helpful in tracing the archaeological features. The interpretation was performed using the final magnetograms obtained from the two study area. The LPF magnetogram was of great help in delineating the presence of some deep archaeological features. While the HPF magnetograms was of great help in mapping the exact shape and dimensions of shallow depth features. The survey confirmed the presence of several subsurface archaeological features within a shallow depth. Our investigation helped identify the locations of interesting subsurface archaeological features like ancient tombs, room structure, Quartzite rocks, burial mounds, quartzite wall, porcelain potteries, and some other Baekje house hold utensils etc. The process magnetic image and the 3D subsurface map also proved the excellent subsurface archaeological features. Our study has also shown the several practical

53

advantages of the fluxgate gradiometer (FM256).It takes less time and effort, it works well with shallowly buried features, It can accurately detects some important parameters such as vertical gradient of the geomagnetic field, the positions, the horizontal dimensions, and the polarity of the near surface features. Actually, The most of the area of Gongju maintains its interest for all geophysicists working in the field of archaeological prospection to explore the huge hidden cultural heritage buried in that area. Finally, I like to say that the above area is really an interesting site for archaeological study and to find out more historical relics and artifacts.

CHAPTER 7

REFERENCES Albert Hesse, Luis Bwrba, Karl Link and Agustin Ortiz. A Mapping and Electrical Study of Archaeological Structures at loma Alta, Michoacan, Mexico. Archaeological prospection, Vol.4, 53-67(1997). Guy Desvignes, Alain Tabbagh and Christophe Benech. The Determination of the Depth of magnetic Anomaly Sources. Archaeol. Prospect. 6, 85-105 (1999). Salah El- Deen A. Mousa, Tareq F. Abdallatif, Ahmed G. Hussain, Ahmed A. Al Bassiony. Application of Geophysical tools for mapping the Archaeological features at Saqqara, Giza, Egypt. Proceedings of the nineteenth annual meeting, April 22-23, 2001. Tareq F. Abdallatif, Jung Mo Lee. Shallow magnetic survey of the Younghae basin area, South Korea: Evaluation of structural setting. Geosciences Journal, Vol.5, No.4, p.327338, December (2001). E.V. Aidona, A. Sarris, D. Kondopoulou and I. Sanakis. Appliction of Magnetic and Spectrometry methods in the Detection of the Human Activity in Soils: A case study at the Archaeological Site of Kitrod( Northern Greece).Archaeol. Prospect.8, 187-198(2001) W. Neubauer, A. Eder-hinterleitner, S.seren and P. Melichar. Georadar in the Roman Civil town Carnutum, Austria: An Approach for Archaeological Interpretation of GPR Data. Archaeol. Prospect.9, 135-156 (2002). T.F.Abdallatif, S.E. Mousa and A. Elbassiony. Geophysical Investigation for Mapping

54

the Archaeological Features at Qantir, Sharqyia, Egypt.Archaeol. Prospect. 10, 27-42 (2003). A. Vsfidis, M. Manakou, G. Kritikakis, D. Voganatsis, A. Sarris and Th. Kalpaxis. Mapping the Ancient Port at the Archaeological Site of Itanos(Greece) Using Shallow seismic Methods. Archaeol. Prospect. 10, 163-173 (2003). H. Ghazala, A.S.El- Mahmoudi and T.F. Abduallatif. Archaeogeophysical study on the Site of tells Toukh El-Qaramous, Sharkia Governorate, East Nile Delta, and Egypt. Archaeol. Prospect. 10, 43-55 (2003). J. M. Maillol, D.L. Ciobotaru and I .Moravetz. Electrical and Magnetic Response of Archaeological Features at the Early Neolithic Site of Movila lui Deciov, Western Romania. Archaeol. Prospect. 11, 213-226 (2004). Tareq fahmy Abdallatif, Esmat Mohamed Abd-All, Mancheol Suh, Reem Mostafa Mohamad, and Ibrahim Aly El-Hemaly. Magnetic Tracing at Abu Sir (Land of Forgotten Pyramids), Northern Egypt. An International Journal, Vol.20, No.5, 483503(2005). M. Suh, T.F. Abdallatif, J. Han, S. Choi, J. Oh. Delineating the northern part of the Socotra basin, offshore Korea, using marine magnetics. Geo-Mar Lett(2005). Abbas Mohamed Abbas, Tareq fahmy Abdallatif, Fathy A. Shaaban, Ahmed Salem, Mancheol Suh. Archaeological Investigation of the Eastern Extensions of the Karnak temple Using Ground-Penetrating Radar and magnetic Tools. Geoarchaeology: An International Journal, Vol, 20, No.5, 537-554 (2005). V.paoletti, M.Secomandi, M. piromallo,F. Giordano,M.Fedi and A . Rapolla. Magnetic Survey at the Submerged Archaeological site of Baia, Naples, Southern Italy. Archaeol.prospect.12, 51-59(2005). Henrique Lorenzo and Pedro Arias. A Methodology for Rapid Archaeological site Documentation Using Ground Penetration Radar: An International journal, Vol.20, No. 5, 521-535 (2005). Rene Efrain Chavez, Maria Encarnacion Camara, Rocio Ponce, and Denisse Argote. Use of Geophysicsl Methods in Urban Archaeological Prospection: The basilica de Nuestra Senora de la Salud, Patzcuaro, Mexico. Geoarchaeology: An International Journal, vol.20, No.5, 505-519 (2005). Atchuta Rao, D. Ram Babu H.V, & Sanker Narayan P.V, (1981). Interpretation of magnetic anomalies due to dikes: The complex gradient method. Geophysics, 46, 15721578. Blakely, R.J., & Simpson, R.W. (1986).Approximation edges of source bodies from magnetic or gravity anomalies (short note). Geophysics, 51, 1494-1498.

55

David, A(1994).The role of geophysical surveys in early medieval archaeology. AngloSaxon Studies in Archaeology and History, 7, 1-26. Debeglia, N., & Corpel, J (1997).Automatic 3-D interpretation of potential field data using analytical signal derivatives. Geophysics, 62, 87-96. Geoscan Research (1994).Instruction manual1.01 (Geoplot2.01).Bradford, west Yorkshire, England: Geoscan Research. Hounslow, M.W., & Chroston, P.N (2002).Structural layout of the suburbs of Roman Butrint, southern Albania: Results from a gradiometer and receptivity survey. Archaeological prospect ion, 9, 229-242. Hsu, S.K., Sibuet, J.C., & Shyu, C.T. (1996).High –resolution detection of geological boundaries from potential anomalies: An enhanced analytical signal technique.Geophysics, 61,373-386. Ellwood BB, Harrold FB, Msrks AE. 1994. Site identification and correlation using geoarchaeological methods at the Cabal O do Porto Marinho(CMP) locality, Rio Maior, Portugal. Journal of Archaeological Science21:779-784. Gibson TH. 1986. Magnetic prospection on prehistoric sites in Western Canada. Geophysics 51(3):553-560. Weymouth JW. 1986. b. Geophysical methods of archaeological Methods and Theory 9:311-395. Hussain AG. 1983. Magnetic prospecting for archaeology in Kom Oshim and Kiman Faris, Fayoum, Egypt. ZES 110: 36-51. Scollar I, Tabbagh A, Hesse A, Herzog I. 1990. Archaeological Prospecting and Remote Sensing, Topics in Remote Sensing. Cambridge University Press: Cambridge:375-516. Wemouth JW. 1986a. Archaeological site surveying program at the University of Nebraska. Geophysics51:538-552. El-Behiry MG.2000.Nondestructive geophysical sueveys for delineating buried tombs and identifying their environmental status. Proceeding of Symposium on Application of Geophysics to Engineering and Environmental problems (SAGEEP), Virginia; 399-408 Clark AJ. 1990. Seeing Beneath the soil: Prospection methods in Archeology. B.T. Batsford: London. Baker GS, Steeples DW, Schmeissner C, Pavlovic M, Plumb R. 2001. Near-surface imaging using coincident seismic and GPR data. Geophysicical research letters 28:627630.