Keywords

1 Introduction

Quarrying operations substantially change the landscape as well as adversely alter pre-existing ecosystems. Large territories are destroyed together with archaeological sites which need to be investigated as fully as possible before they disappear forever.

Geophysical methods in recent years have proved to be of great importance in acquiring data for effective archaeological heritage management and, hence, must be applied to determine best targets for excavations (Schmidt et al., 2015; Sarris, 2017).

Geophysical exploration of archaeological sites in mining provinces faces a number of specific challenges. The geology of the region can be a natural obstacle to the application of some geophysical methods (Fassbinder & Bondar, 2013; Bonsall, 2014; Rusch et al., 2020; Bondar et al., 2021a). In addition, the industrial activity noise interferes with the geophysical fields, negatively affecting measurements (Krivanek, 2001; Booth et al., 2010; Bondar et al., 2019; Polin et al., 2020; Schmidt et al., 2020). The weak contrast of physical properties of the soil and the archaeological object against the background of a high-gradient or noisy field makes archaeological objects “invisible” (Krivanek, 2001; Jrad et al., 2014).

The study deals with first archaeological prospection results in the Kremenchuk magnetic anomaly area, where large iron-ore quarries together with a sinter plant had been used since the 1970s. Since the rapid industrialisation of the region, quarrying has become an important threat to the unique archaeological heritage of the region. Alienation of new lands for quarrying of neighbouring fields, storage of dumps and construction of the tailing ponds continues.

In this chapter we evaluate the efficiency of geophysical methods in proving conclusions of visual archaeological inspection and targeting subsequent archaeological excavations to ensure the recording and mitigate the total loss of the idiosyncratic archaeological sites of the area. The article discusses the ability of geophysics to detect subsurface archaeological features against rather non-ideal survey conditions such as unfavorable magnetic geology, superficial deposits, and anthropogenic topsoil pollution. Some successful case studies are presented.

We pay much attention to the geophysical characterisation of burial monuments (kurgans), which provide the main source of information for the study of nomadic archaeological cultures of Ukrainian Steppe. A kurgan is a mound of earth raised over a grave. Large mounds can contain later graves inserted into a primary mound deposit. A few recent papers devoted to geophysical investigations on kurgans from Ukrainian steppe and forest-steppe report the capability of such minimally invasive techniques to detect these monuments even if they are partly or totally truncated by ploughing and there is no trace of them on the surface (Zöllner et al., 2008; Bondar et al., 2019, 2021b; Polin et al., 2020). Showcasing how to distinguish Bronze—Early Iron Age kurgans from natural or modern artificial elevations using geophysics is a particular objective of the chapter.

2 Location of Study Area and Environmental Settings

The region of Kremenchuk Magnetic Anomaly is located on the left bank of the Dnieper River in the forest-steppe zone of the Dnieper lowland, within the Psel and Sukhyi Kobeliachok interfluve (Fig. 1a). This territory is a part of the Ukrainian crystalline shield within the Eastern European platform, thus its landscape is flat and partially crossed by ravines. The average height above sea level is 69 m. The channel of the Dnieper River is heavily indented here, which contributed to the emergence of numerous estuaries and small islands. The climate is moderately continental.

Fig. 1
A map of Kremenchuk Magnetic Anomaly with a zoom out inset map and a radar chart represents the legends of topsoil magnetic susceptibilities, ranging from 15 to 0.42, archeological sites, ore body, chemozem on loess, and sod soil on sand. 2 line graphs of profiles 1 and 2 are represented at the bottom.

Regional map of the Kremenchuk Magnetic Anomaly area showing location of the iron-ore body, quarries, main types of soil, magnetic susceptibility values of the topsoil measured at points marked with red dots, archaeological sites discussed in the article are marked with triangles and labeled (a). Anomalous intensity of the vertical component of the total magnetic field (Za) along transects AB (b) and CD (c). (After Krutihovskaya, 1971)

Full size image

The region comprises both natural and man-made landscapes. The iron ore deposit is located in the floodplain and on the first terrace of the Dnieper. The total area of 140 km2 in Poltava region includes iron-ore processing industrial zone, the residential zone (the city of Gorishni Plavni and villages), transformed areas of relict landscapes and forestry areas. The soil cover is represented by sandy soils on the first fluvial terrace of the Dnieper, meadow-chernozem and chernozem formed on loess loam—on the higher areas of the lowland. Sandy soils are highly permeable and have thin (<20 cm) humic horizon A. Meadow-chernozems and chernozems have horizon A with thickness over 60 cm.

Kremenchuk Magnetic Anomaly stretches along Kryvyj Rig-Kremenchuk fault of the Ukrainian crystalline shield (Dobrokhotov, 1964; Ben’ko et al., 2000). Mining of iron ferruginous quartzites started at the beginning of 1970s at Gorishn’oplavnivskyi quarry. Now the mining complex includes also Yerystovo and Belanovo quarries, which have started being explored recently. The industry is specialised in the production of iron pellets for metallurgical needs.

Kremenchuk Magnetic Anomaly reaches tens of thousands of nanotesla on the ground surface in spots of maximum signal (Fig. 1b, c). The places of three magnetic maxima have been selected for quarrying. Gorishn’oplavnivsky maximum is the strongest one and is characterised by a high horizontal gradient of the total field due to shallow position of the ore body (15–20 m to the surface). At the area of Yerystovo maximum the overburden makes 30–40 m, at Belanovo—more than 100 m. The shape of the magnetic anomaly is controlled by two limbs of the syncline fold at Gorishn’oplavnivskyi area (Profile 1) and by the only eastern limb combined with hinge zone preserved at Belanovo area (Krutihovskaya, 1971).

Quarries and the sinter plant located at the city of Gorishni Plavni are powerful air pollution sources at the region. Wind rose showing the distribution of wind direction (Fig. 1a) evidences that most of the time the winds blow from the north-west, transporting industrial emissions and dust from the dumps and dry beaches of the tailing ponds to the area east from the mining. After being settled on the topsoil, the dust particles change soil properties, in particular, magnetic properties (Evans & Heller, 2003; Fialová et al., 2006). This can significantly influence the results of magnetometer surveys.

3 Archaeology of the Region

The vast majority of archaeological sites in the region are dated to Bronze-Early Iron Age. The Bronze Age (mid-fourth to early first millennium BC) is characterised by the presence of cattle-breeding tribes of the Yamnaya, the Catacomb and the Timber-Grave archaeological cultures. They arranged their burials under kurgans, which became an integral part of the local landscape (Suprunenko et al., 2004; Shylov, 2007; Suprunenko & Sherstiuk, 2009). The final of the Bronze Age – the beginning of the Early Iron Age is represented by settlements of the early stage of the agricultural Belohrudovo-Chornolis culture (Bashkatov et al., 2020). In the Early Iron Age, the interfluve of Psel and Sukhyi Kobeliachok was inhabited by the Scythians (seven to four centuries BC) and Sarmatians (second century BC to third century AD) (Ben’ko et al., 2000; Kulatova, 2011).

Walkover surveys carried out at the area suggested the presence of ~130 possible kurgan groups and settlements of the Bronze—Early Iron Age (Suprunenko, 2014). Each kurgan group can contain from one to twenty kurgans of different size and degree of preservation. Small elevations of the terrain can be also due to geomorphology or associated with rather modern homesteads from the 18th to 20th centuries. On visual inspection, they are easily confused with ploughed kurgans.

4 Materials and Methods

The field measurements were conducted in 2016–2020 at 18 archaeological sites selected from the results of the walkover surveys (Suprunenko, 2014). Six sites were completely excavated by the Dnipro-Psel expedition of the Institute of Archeology of the NAS of Ukraine under the direction of Yu. Bashkatov. A flexible set of geophysical methods was used comprising high-resolution magnetometer survey (14.5 ha covered at 18 sites), electrical resistivity tomography (240 m of profiles recorded at one site) and ground penetrating radar (2650 m of profiles at 14 sites). Geophysical survey areas and individual profiles areas were georeferenced on orthoimages or topographic maps by measuring their coordinates using a GPS.

4.1 Magnetometer Survey

Magnetometer surveys can be a rapid tool for mapping subsurface archaeological structural remains. Their use at kurgan sites can provide information about inner structure and its separate elements like dromos, chambers as well as different objects on the kurgan’s periphery (Smekalova et al., 2005; Parzinger et al., 2016, 2015; Fassbinder et al., 2015; Fassbinder, 2015, 2016; Bondar et al., 2019; Polin et al., 2020; Goldmann et al., 2021). This technique was used at all sites, as a means of quick investigation, although it was not always efficient. The instrument used was a caesium total field magnetometers PKM-1 (Geologorazvedka, Russia), which had a sensitivity of +/−0.01 nT. The instrument records 10 measurements per second, providing a spatial resolution of about 10 cm on the profile by normal walking speed. With traverse spacing of 0.5 m, the total intensity of the geomagnetic field was acquired with a spatial resolution of 50x10 cm.

4.2 Magnetic Susceptibility Measurements

At the vicinity of iron-ore mining and processing area, magnetic enhancement of topsoils could be due to atmospherically deposited magnetic particles of industrial origin (Fialová et al., 2006) and this can affect the results of magnetometer survey.

Weak anomalies from low contrast archaeological features could be hardly detectable against the plough effect of strongly magnetic topsoil. In order to outline such polluted areas, in-situ magnetic susceptibility (k) topsoil measurements were taken using handheld KM-7 Satis Geo kappameter at eight locations. Between 15 and 20 readings were taken from each measured point—a spot of about 4 m2 cleaned from surface vegetation.

4.3 Electrical Resistivity Tomography

ERT can help to characterise the construction features of mounds and their relative stratigraphy (Papadopoulos et al., 2010; Tsourlos et al., 2014; Zhao et al., 2019; Hegyi et al., 2021). Apparent resistivity measurements were acquired using a one-channel device furnished with 64 brass electrodes (Khomenko et al., 2013). All profiles were made using the Wenner-Schlumberger array protocol, the electrodes were placed at every 1 m. Such distribution allowed recognition of electrical resistivity readings to a depth of about 11 m. The Wenner-Schlumberger array was chosen because it is moderately sensitive to both horizontal and vertical structures (Loke, 2009). Measured electrical data were inverted using the interpretation software Res2DINV, employing the robust least-squares optimisation technique (L1-norm) (De Groot-Hedlin & Constable, 1990; Sasaki, 1992; Loke & Barker, 1996). L1-norm tends to produce models that are piecewise constant, which is consistent with the known structure of excavated kurgans. Bad datum points and points with root mean square (RMS) error higher than 90% were removed from the final inversion. The model was accepted after four iterations with a RMS misfit lower than 5%.

4.4 Ground Penetrating Radar

GPR has been used to characterise kurgans and other archaeological mounds. A challenge to survey such sites using this technique can be related to the topographic corrections that are required to a correct interpretation of GPR data collected at relatively well-preserved mounds (Goodman et al., 2007). Other challenges can be derived from other general aspects related to the limitations of the propagation of the electromagnetic energy used in this technique, under specific soil conditions. For example, some clay-rich and highly conductive soil deposits can attenuate the propagation and reflection capacity of the GPR energy and result in a poor depth of penetration or complete failure in the detection of subsurface remains (Schneidhofer et al., 2017; Conyers, 2017; Bondar et al., 2021a). The reason to use GPR to characterise kurgans in the region was the prevalence of highly permeable sandy soils. As GPR is a highly productive technique, so it bridges the gap between magnetometer survey and ERT in cases when the first one is inefficient and the second is extremely time-consuming. The GPR system used was a VIY-3 instrument produced by Transient technologies LLC, Ukraine, equipped with shielded transmitting antennas with nominal centre frequencies of 300 MHz. Data was processed using the Synchro3 software (http://viy.ua/e/software/synchro.htm). The processing steps included zero level setting, dewow operation and wavelet filtering, windowed background removal, time gain and estimation of the average electromagnetic wave velocity by hyperbola fitting. Since the GPR profiled were collected at substantially truncated sites or flat areas around them, there was not need of topographic correction. Obtained processed reflection profiles and annotated reflections of interest were subsequently visualised to correlate them with the results of the other geophysical surveys.

5 Geophysical Results and Archaeological Evidence

Not all the expected sites or known kurgans were confirmed by the geophysical results. Often, only one or two kurgans, of an expected group of five to twenty were defined according to presence of specific geophysical anomalies. In many cases, magnetometer surveys were useful to discriminate between supposed prehistoric kurgans and elevations on the places of homesteads of 18th to 20th centuries. The latter were determined by strong magnetic anomalies from brick buildings and a lot of iron rubbish around and thus excluded from further studies. Below, we describe some case studies showcasing successful results from sites that has been validated via targeted archaeological excavations.

5.1 Detection of Kurgans Integrating Magnetometer Surveys and Complementary Techniques

The large kurgan Bondari is located in an area that soon will be destroyed according to the Belanovo quarry development plan. The maximum height of the kurgan is 4.2 m, and the diameter is about 70 m. There is a big pit at the centre of the mound, which is 3 m deep and has dimensions of 27 × 20 m. For a long time, it was considered to be a “maidan”—a saltpetre fabrication site like those that were common throughout the 17th to 19th centuries in the Poltava region (Sherstiuk, 2013). Saltpetre (or potassium nitrate for the fabrication of gunpowder) was extracted using the soil from ancient mounds. Remains of furnaces, ash dumps and other related infrastructure are usual traces of this activity in the vicinity of remnants of the destroyed kurgans (Zöllner et al., 2008; Bondar et al., 2021b).

An orthoimage of kurgan Bondari was obtained using drone-acquired photographs. A digital terrain model was derived from the orthoimage to record the mound and a large depression visible to the south-east of it (Fig. 2a). The depression is ~40 m wide and 1 m deep. It is well-known that Bronze Age kurgans were formed gradually, due to repeated earthing-up for new burials. The soil for the new mound used to be taken near the primary kurgan, resulting in the formation of a large depression, sometimes ~20–30 m wide (Mozolevs’kyi, 1990; Chernych & Daragan, 2014). As noted by B. Mozolevsky: “kurgans of the Bronze Age always stand, as in a saucer, in a deep and wide depression, the surface around the Scythian kurgan is always flat as a table”. Therefore, kurgan Bondari has been dated to the Bronze Age.

Fig. 2
A heat map of elevation distributions and an orthoimage with grayscale images of surfaces of the study area are represented. At the bottom, a line graph of total field versus distance plots 2 increasing lines of measured values and linear regression.

Bondari kurgan (Bronze Age): (a) Digital terrain model; (b) Results of the magnetometer survey in greyscale overlying an orthoimage of the site area; (c) Total field profile, marked with blue line on the figure (b)

Full size image

The magnetometer survey of the kurgan was performed in a high-gradient field, as the site is in the near vicinity of the Belanovo iron ore deposit (Fig. 1a, b) (Dobrokhotov, 1964; Krutihovskaya, 1971). The total geomagnetic field intensity changes by over 200 nT in the NE-SW direction. However, even against such background, the magnetic anomalies associated with remnants of the mound were visible (Fig. 1b, c). Processing procedure of subtracting profile linear regression values from the measured total field values allows the exclusion of the effect from geological structures. The remnants of the mound caused an anomaly of up to 20 nT. Extremely high values (red) correspond to relics of the geodetic pillar and a fire point of the World War II time. No traces of saltpeter production activity were recognised in the magnetometer survey results.

Excavations proved the conclusions achieved from non-invasive study. The central pit appeared to have formed due to excavation of the central burial performed at the beginning of the twentieth century. Eleven burials of the Yamnaya and the Catacomb cultures of the Bronze Age were inserted into the mound. The individual burials did not cause magnetic anomalies.

5.2 Identification of Kurgans at Magnetically Polluted Areas

The results of the magnetometer surveys carried out at the east of the Gorishn’oplavnivskyi quarry did not detect anomalies potentially associated to kurgans’ mounds. We tend to associate the reason for this with magnetic pollution of the topsoil. The in-situ magnetic susceptibility measurements revealed areas of magnetic enhancement of soil presumably due to the presence of strongly magnetic minerals of anthropogenic origin. Means and standard deviations (SD) of k value are represented on Fig. 1a. In particular, high values are observed to the east from the tailing pool.

The magnetic susceptibility of the topsoil was very high (0.8–1.4 * 10−3 SI) near the kurgan 73 (Fig. 1a) that was excavated at the east of the Gorishn’oplavnivskyi quarry. Targeted trench P excavated near the kurgan exposed two deposits: 0–30 cm—light pale sandy plough horizon; 30–60 cm—brown sandy subsoil (Fig. 3a, b). There was a sharp division between layers at a depth of 30 cm due to ploughing. Below, soil susceptibility decreased to 0.1 * 10−3 SI. Magnetic particles product of quarrying activities and spread in uppermost soil horizon by ploughing seems to be the cause of the enhancement of the uppermost deposit. Microscopic examination of thin sections extracted from the topsoil revealed fragments of ferruginous quartzite dust (a typical by-product from waste rock dumps) as well as magnetite-hematite-glassy spherules (derived from airborne ash emitted by sinter plant) (Fig. 3c, d). Although magnetometer surveys were carried out at this area, the highly magnetic topsoil masked potential anomalies derived from expected buried archaeological features. The ploughing activities carried out at these areas were also visible in the magnetometer results producing a clear stripping effect (Fig. 3a).

Fig. 3
A satellite view of the terrain with some grayscale photos and a highlighted area of the soil excavation site. A photo of an excavated area and 2 micrographs for magnetic susceptibility are depicted on the right. Below, a photograph and a G P R profile of an ancient ground surface under the mound are depicted.

Kurgan 73: (a) Satellite image with georeferenced magnetometer surveys results in greyscale and dynamics of the total magnetic field 50,510 ± 7.0 nT. The kurgan’s mound revealed by excavation is outlined with red dotted line. The location of the soil profile is marked with P (blue dot). The GPR profile 1-2 is shown as a green line; (b) sandy soil profile P with magnetic susceptibility values; (c and d) thin section photographs extracted from the topsoil at P, in reflected light. Fragments of ferruginous quartzite (Q) and magnetite (Mg) are shown in (c). Magnetite-hematite-glassy spherule is shown in (d); (e) GPR profile 1-2 crossing a ploughed area of the kurgan’s mound

Full size image

Kurgan 73 was 4 m high and had a slightly elongated shape because its mound was partially ploughed. Since magnetometer survey could not establish the extent of this site, several areas were explored with GPR profiles. The reflection associated with the mound and its interface with the natural underlying deposits or ancient ground surface were observed at the time of 37–40 ns (depth 1.5 m). The reflections could be attributed to different soil water saturation controlled by different porosity of the undisturbed sandy soil and mixed earth of the mound deposits. Thus, GPR measurements aided to identify the real diameter of the kurgan, which appeared to be 38 m. In the mound, a stone-carved tomb was unearthed as well as 10 more inserted inhumations dated to the Middle Bronze Age. Similar interfaces between the mound deposits-former ground surfaces were observable in the results of other GPR reflection profiles at other neighbouring kurgans.

All the more inspiring were magnetometer survey results obtained on kurgan 54 (Fig. 4), where magnetic susceptibility of the topsoil exceeds 1.0 * 10−3 SI. This is an example of kurgan identified from results of magnetometer survey exclusively. The survey plot occupied the top of a gently elevated area. The circular ditch was interpreted in the place of the +1.6 nT anomaly. Excavation proved a ditch with diameter of 21 m and a depth of 0.4–0.6 m which was partly destroyed by agricultural activity (Fig. 4a, b). Kurgan with a ditch is quite unusual in this region. Three inhumations of Yamnaya culture were excavated within the enclosing circular ditch, no one burial caused magnetic anomaly. Excavations revealed additional mound being constructed for new burials. The total diameter of the kurgan’s mound reaches 40 m. Although the magnetic map is noisy, four anomalies with intensities 1.5–3.2 nT correspond to four excavated burials of Catacomb culture located outside the circular ditch. These burials are chronologically later than ones enclosed with the ditch. They were arranged in underground chambers (catacombs), that were collapsed and filled with dark coloured soil. Figure 4c shows an example of a catacomb.

Fig. 4
A heat map of a study area represents the distribution of magnetic fields. A circular section is marked on it. Two aerial photos of a circular ditch and two circular excavations are represented at the bottom.

Kurgan 54: (a) Results of the archaeological excavation overlying the results of the magnetometer survey in white-blue scale with dynamics of the total magnetic field 50,485 ± 1.5 nT; (b) Aerial photograph of the excavated circular ditch enclosing the Yamnaya culture kurgan mound; (c) Photograph of the Catacomb culture burial 7

Full size image

5.3 Investigation of the Internal Structure of a Large Kurgan Using ERT

The Novoselivska Mohyla kurgan (Fig. 1a) dated to fourth to first millennium BC (Suprunenko, 2014) has a maximum height of 7.5 m and occupies a large area of 64 × 68 m. On the top of this kurgan, there was nineteenth to twentieth century cemetery which is a common aspect at these sites in the region. Since the kurgan was going to be excavated because of the development of a new waste dumping area, all graves were exhumed before the geophysical survey. Given all the open pits left after the exhumation, four ERT profiles were strategically located to investigate the internal structure of kurgan (Fig. 5a).

Fig. 5
A satellite map of a study area highlights the location of Kurgan with labeled arrows, along with a heat map of Kurgan highlighting higher elevations. A 3 D model of kurgan at the bottom depicts the distribution of resistivity.

Novoselivska Mohyla kurgan: (a) orthoimage and digital terrain model with the ERT profiles shown with arrows marked P1-P4; (b) ERT model of the kurgan with the location of potential burials marked as b.1 and b.2

Full size image

The results suggested the location of several structural components of the kurgan characterised by resistivity values varying from 1 to 3100 Ωm (Fig. 5b). The ERT model generally shows lower resistivity zone (<60 Ωm) corresponding to primary mound with a diameter of ~25 m. It is overlayed by secondary mound having higher resistivity values (100–350 Ωm). Overlying this secondary mound, very high resistivity values characterise the uppermost deposits containing the exhumed burials. At the bottom of the kurgan mound deposits, there is a high resistivity zone that may indicate the presence of initial burial (b.1 on Fig. 5b). Another possible burial is interpreted inside the primary mound (b.2 on Fig. 5b).

Although the kurgan has not been completely excavated yet, the ERT results suggested a rather complex stratigraphy of this monument, erected in several stages. This interpretation fits the hypothesis of O. Suprunenko (2014) about this pre-Scythian or Scythian kurgan, when the secondary mound was built over burial(s) inserted into the primary mound of the Eneolithic—Bronze Age.

5.4 Magnetometer Survey Mapping of Bronze Age Settlements

In 2007–2008, during visual observations for new waste dumping areas at the east from Yerystovo quarry, the settlements Yerystivka-1 and Yerystivka-2 were discovered (Suprunenko, 2014). Both settlements stretched along the watershed on a north-south direction and were located and dated by surface finds of the Early Bronze Age (Сatacomb culture, 18th to 17th BC) and Late Bronze Age (Timber-Grave culture, 15th to 14th BC). Suprunenko documented on the surface of Yerystivka-1 seven ash heaps ~0.1–0.3 m high, ranging in size from 8 × 10 to 14 × 16 m. They were located in two almost parallel rows. Yerystivka-2 consists of 20 ploughed ash heaps with heights of 0.2–0.5 m and diameters of 10–21 m. Ash heaps are situated on both sides of the field road. Two of them were excavated in 2008 revealing dwellings of Belohrudovo-Chornolis culture.

A magnetometer survey was carried out on the area of 8 ha (Fig. 6a). The results at Yerystivka-1 shows several small (up to 4 m2) low-intensity (up to +3 nT) anomalies, probably sourced by remains of ancient dwellings and farm structures. However, archaeological research has not been conducted there yet. The results at Yerystivka-2 (Fig. 6b) clearly shows the old field road, which is now ploughed out. A linear anomaly seems to correspond to an accumulation of ferruginous quartzite rubble (blue line in Fig. 6b). Anomaly 1, of oval shape, intensity of 0.7–2.3 nT and significant dimensions (60 × 90 m) was of particular interest. However, the excavation trench that targeted this anomaly did not reveal any archaeological nor natural feature.

Fig. 6
A satellite view of sites Yerystivka 1 and 2 with some grayscale photos. A grayscale scale image highlighting the magnetic field of site Yerystivka 2. 3 satellite views of 3 excavation sites represent anomalies.

Settlements Yerystivka-1 and 2: (a) Satellite image overlaid by the results of the magnetometer survey in greyscale and dynamics of the total magnetic field 49,760 ± 4.0 nT; (b) Results of the magnetometer survey at Yerystivka-2 with anomaly annotations in yellow; (c) Excavation results at the location of anomaly 2 (0.4 m depth); (d) Excavation results at the location of anomaly 3; (e) Excavation results at the location of anomaly 4

Full size image

Anomalies 2,3,4 had intensities of 3–4 nT and a size of 3–10 m (Fig. 6b). These were also targeted with excavation trenches reaching ~0.8–1.4 m depth (Fig. 6c–e) and revealed three pit-houses.

6 Discussion

The efficiency of geophysical techniques at each site within the region is strongly controlled by local conditions. As is shown by the example of Bondari kurgan, large mounds cause magnetic anomalies even on the background of strong geological magnetic effect due to the large amount of earth hosting magnetic minerals (Fig. 2b, c). Destruction of the mound due to looting or saltpeter production could normally be traced by the results of a magnetometer survey. Single inhumations inserted into the mound cause no disturbance in the magnetic field unlike burials in catacombs as illustrated in the case of Kurgan 54. When the catacomb collapses and the overburden is not thick enough, holes are formed on the surface of the mound, gradually refilling with soil. That is what probably happened with the detection of the catacomb burial at Kurgan 54 (Fig. 4a). Magnetic anomalies from catacomb burials as well as from kurgan’s circular ditch are largely originated due to the detrital magnetic remanence of the infill (Bondar et al., 2022). Modern magnetometers are capable of detecting such archaeological features even under a rather thick layer of magnetically polluted soil. However, anthropogenic magnetic enhancement in the topsoil imposes limitations on the use of magnetometry on smaller kurgans. If a small kurgan hosts no ditch or destroyed burial, it stays invisible for magnetometer survey. However, under favourable conditions other geophysical methods could help, as shown on the example of Kurgan 73. Sandy soil facilitated determination of interface between the mound and former ground surface as well as its extent using GPR (Fig. 3e). Valuable archaeological information about internal structure of large kurgans can be obtained using ERT, as shown on example of Novoselivska Mohyla kurgan (Fig. 5b). Many kurgans with highly resistive structures, such as burials in stone-carved tombs, cromlechs, well-preserved catacombs, are often found in the Kremenchuk Magnetic Anomaly area (Suprunenko et al., 2004; Shylov, 2007).

Magnetic prospection of the settlements Yerystivka-1 and Yerystivka-2 has demonstrated good results in searching for Bronze Age dwellings and pit-houses (Fig. 6a, b). However, the excavation results targeting the large anomaly 1 did not provide any archaeological evidence (Fig. 6b). We can attribute it to a “magnetic ghost” case (i.e. an archaeological feature that is invisible during an excavation but detectable in term of their magnetic properties) (Schleifer, 2004; Leckebusch et al., 2000; Breitwieser et al., 2001; Lysenko, 2009; Simon et al., 2012). In this case, this anomaly could be related to a trace of shallow ditch, which did not reach subsoil and was backfilled with organic material with enhanced magnetic properties. It is known that decomposition of organic matter plays the key role in soil magnetic enhancement, the iron minerals stay whereas the organic matter dissolves (Schleifer, 2004).

7 Conclusions

Amongst 18 supposed archaeological sites investigated with the use of geophysical methods at the industrial area of Kremenchuk Magnetic Anomaly, nine had been confirmed to be related to Bronze—Early Iron Age. Relics of the 18th to 20th centuries were recognised under kurgan-like mounds at five sites. Four sites revealed no geophysical traces of archaeology therefore have not been recommended for excavation.

Based on the results we could determine the following geophysical anomalies related to the presence of kurgans: magnetic anomalies caused by the mound of the kurgan; magnetic anomalies caused by circular ditch or catacomb-type burials; reflections from the interface of the mound and former ground surface observable in GPR reflections profiles radargram at sites located in a sandy soil environment.

In this work, we found that cultural heritage sites in many cases can be located and characterised with the aid of geophysics under challenging anthropogenic and natural survey conditions. However, the limitations of a particular method must be considered. The enrichment of magnetic particles in the topsoil together with heavy ploughing add a lot of noise to the magnetometer survey results restraining discrimination of weak anomalies. The use of GPR is limited to recognition of earthwork at sandy soil. ERT as time- and labour-consuming technique is appropriate at large kurgans to establish the stratigraphy.

The mining of ferruginous quartzites at the Kremenchuk Magnetic Anomaly area will continue. More and more new zones will be used to develop activities related to quarrying activities. The need of archaeological investigation of this areas to ensure the recording and safeguard of this heritage from it destruction through modern development will become increasingly important in future. Therefore, the results obtained from our study will facilitate applying geophysical methods for archaeological needs and aid proper understanding of the appearance of archaeo-geophysical anomalies in the region.