Remote Sensing for the Detection of Old Clandestine Graves

(1)

Remote Sensing for the Detection

of Old Clandestine Graves

Examining the available technologies and their applicability to

drone based searches

by

Marissa Koopman

Student #10689087

A literature thesis presented for the degree of

Master of Forensic Science

Supervised by MSc. Jitteke Struik

Co-assessed by Dr. Ing. Maurice Aalders

Department of Interdisciplinary Sciences

The Netherlands

30 December 2017

(2)

Abstract

The use of remote sensing to detect old clandestine graves brings ad-vantages over traditional techniques. Possible markers for the detection of old clandestine graves are discussed, and several remote sensing tech-niques compatible with drones (TIMS, Raman spectroscopy, LIDAR, and hyperspectral imaging) are examined for their applicability to the detec-tion of buried remains which are five years old or older. The focus on clandestine graves of five years old and older has been chosen due to a marked increase in difficulty in finding these graves, as indicated by the division ”Bijzondere Zoekingen” in the Netherlands.

TIMS and LIDAR are likely unsuitable for the remote detection of old clandestine graves. Although both techniques have a limited body of literature exploring the possibilities, the few studies which exist indi-cate negative results. Raman spectroscopy and hyperspectral imaging are more promising, although as of yet not suitable for implementation in case work due to a lack of burial specific research. There is a general discon-nect between the markers identified for the detection of old clandestine graves in the general forensic literature, and the ability of current remote sensing techniques to utilise these.

1 Introduction

The Dutch police division ”Bijzondere Zoekingen” (Special Searches) works to locate the clandestine graves of missing persons. For some of these searches, there is enough information to narrow down the search area substantially, and it takes relatively little manpower to cover those areas. Occasionally however, little information is available on the missing person’s possible location, and the potential search areas can include many square kilometers. For these searches, the amount of manpower, resources and time needed to cover the search area can be substantial, and as a result methods and technologies which can decrease the amount of manpower, resources and time spent are of great interest to the Special Searches division. In particular, there has been an increased interest in the use of drone-mounted technology to rapidly scan an area for the presence of clandestine graves. The use of drones brings numerous advantages, which will be explored in section 1.2.

Of particular interest to the Special Searches division is the detection of old clandestine graves (five years or older). In their experience, these graves be-come increasingly difficult to locate over time. However, there are still family and friends waiting for news about unsolved cases, due to which there is a need for new technology and methods to locate these graves. The challenges asso-ciated with the detection of old clandestine graves will be discussed in section 1.1.

Before examining the different drone compatible remote sensing techniques available for use, it is important to understand what physical and chemical prop-erties the techniques will be used to searched for. To identify these propprop-erties, it is important to understand what distinguishes buried human remains and the soil surrounding them from the rest of the environment, in order to target these distinguishing properties during a search. Several possible targets for remote sensing techniques will be discussed in section 2.

Various remote sensing techniques will be examined and selected. The con-cepts upon which each technique works, the current forensically relevant lit-erature, and the techniques’ limitations will be discussed in section 3. The techniques discussed will be chosen based on drone compatibility, the availabil-ity of literature, and on how promising they appear with regards to the locating of old clandestine graves.

Found scientific literature will be discussed and critically examined with re-gards to conditions within the Netherlands, and their relevance to clandestine graves located in Dutch soils. As such, the results of this thesis may not be translateable to other geographical regions or soil types.

1.1 The challenges in searching for old clandestine graves

The detection of clandestine graves is of international forensic importance. Graves hold important information which is useful to police investigations. In-ternational efforts exist, through for example the InIn-ternational Commission of Missing Persons, to detect mass graves in active and post war zones. But not

(5)

all clandestine graves are found soon after their creation, and as these graves age, the bodies within them continue to decompose to increasingly advanced stages. Depending on environmental factors such as temperature and burial depth, little more than bone and cartilage may be left to detect in graves more than a year old [Senn and Weems, 2013, Mann et al., 1990], as well as possible items of clothings with which the victim was buried. Many fresher traces will have disappeared or may have been greatly reduced below quantities detectable by traditional methods.

For instance, disturbed soil will have settled so that differences in soil density and vegetation will no longer be easily distinguishable [Pringle et al., 2012, Kil-lam, 2004]. As a result it is not practical to search for newly disturbed soil, or for the air content differences across the top layers of the soil, or for the removal of vegetation. Furthermore it may no longer be plausible to look for the gaseous products of decomposition, known as volatile organic compounds (VOC). There is evidence to suggest that VOCs are no longer released from old graves in suf-ficient amounts, causing cadaver dogs to experience difficulty in detecting them [Forbes et al., 2016]. It may be however, that different VOCs are released later in decomposition compared to the earlier stages, and that cadaver dogs are sim-ply not trained to respond to them. This concept will be explored in section 2.2.

1.2 The advantages and disadvantages of aerial remote

sensing

Remote sensing would allow the detection of clandestine graves without disturb-ing them, and without potentially destroydisturb-ing evidence in dodisturb-ing so. Furthermore, remote sensing would ideally be combined with the use of drones, which can cover a landscape with relative ease and swiftness, while cutting down dramati-cally on the manpower necessary to cover an area. This would be helpful in cases where there is little information on the possible location of the grave(s), and a large area needs to be covered during the search. Furthermore, the use of drones and remote sensing can reduce the risk and exposure of investigative personnel to possibly dangerous environments, such as war zones or difficult to traverse terrain [Kalacska and Bell, 2006]. A drone can also be easily transported and deployed on-site, as opposed to a helicopter or a plane. Aerial remote sensing techniques, which generally rely on the reflectance of light, also make it possible to analyse large areas far more quickly than one would be able to do with most ground based techniques due to swifter travel times of sensors.

However, aerial remote sensing also has its limitations or disadvantages. Some disadvantages are specific to the type of remote sensing chosen, and will be discussed in chapter 3. Other disadvantages are more universal. There are restrictions around the use of drones, especially near airports, which means that it is not always legal to deploy a drone. In the Netherlands it is also required that the pilot of the drone has a clear line of sight to the drone at all times and that the drone remains within a 500 meter radius of the pilot, which can make the drone difficult to use in wooded or urban areas. It is also only allowed to fly

(6)

a drone by daylight, so they cannot be used during night time. The drone may not fly over crowds, which further restricts the forensic use of drones, but this is largely irrelevant for the detection of clandestine graves as these do not tend to occur near crowds. Furthermore, drones carry a risk of crashing if piloted improperly, which could damage and destroy valuable equipment. [Ministerie van Infrastructuur en Milieu, 2016]

2 Targets for remote sensing

Possible targets for remote sensing includes anything found in an old grave which is not usually found beyond it, which will henceforth be referred to as a marker. This could include the mineral cortex of bones, remainders of clothing such as zippers and plastics, or late stage VOCs. Of these markers, specifically human bones would be ideal since these are expected to always be present in a forensically relevant old clandestine grave. As such it is important to understand what sets human bone apart from other materials found within the ground, and even what sets it apart from non-human bone. This will be further discussed in section 2.1.

The presence of a decomposing body affects the vegetation which grows above it through the release of VOCs [Carter et al., 2007]. As the human body decomposes, VOCs are released and seep into the surrounding soil [Putman, 1978]. These are then taken up by the roots of the vegetation growing above the grave. For the detection of old clandestine graves, the main question is how long specific VOCs remain in the soil around a body, and how specific these are to human decomposition. It is important that significantly increased levels of compounds which are characteristic of human decomposition remain in the soil for five years and longer. The presence of VOCs and their ensuing effects on the vegetation growing around a grave will be further discussed in section 2.2.

2.1 Human bone

Human bone consists of a living matrix, and a matrix made mostly of minerals [Whedon and Heaney, 2017]. The living matrix will decompose fairly quickly while the mineral matrix can remain for many years. Due to this resistance to decomposition, and its universal presence in all human graves (unless ac-tively removed), the mineral matrix of bone makes an ideal marker. A concern is whether, if buried bone can be detected at all, human bone can be distin-guished from non-human bone, since skeletal remains originating from animals are expected to be commonly present in soils.

Mineral matrix

The mineral matrix of bone consists of several different minerals, of which the main components are calcium phosphate, hydroxyapatite, and carbonate, which is found in two forms, calcium carbonate and carbonate apatite [Whedon and

(7)

Heaney, 2017]. These carbonates are present in different amounts in different species, thus holding the potential to give limited differentiative power between human and some types of animal bones [Whedon and Heaney, 2017].

Piezoelectric effect

The piezoelectric effect occurs when the atoms of a material are arranged like a crystal, but in a non-symmetrical manner. In these materials there is an equal distribution of negative and positive particles which give an overall neu-tral charge to the material. When stress is applied, however, the positive and negative particles are moved in a non-symmetric manner, causing an electric dipole. As a result of this electric dipole, a current begins to run through the material, which generates an electromagnetic field. The opposite is also true: when a piezoelectric material is subjected a potential difference, it causes the material deform.

Bone is piezoelectric [Lee et al., 2011], and the piezoelectric property of bone is of interest because electromagnetic fields can travel through soil, providing a signal above the surface, which can be detected by sensors. As such it may be possible to measure the electromagnetic fields created by skeletal remains when stress is applied to them. Since the bones are buried, stress already exists in the form of soil pushing down on the bones. However, this constant stress would create an electromagnetic field of constant strength, which would likely be in-distinguishable from other electromagnetic fields in the environment. It would be more interesting to induce stress, and detect any fluctuating electromagnetic fields which result from the application of stress. The potential forensic interest of the weak magnetic fields created through the piezoelectric property of bone has been briefly mentioned by Larson et al. (2011), but no literature could be found expanding on the idea or further investigating it.

An interesting method of inducing this stress would be the use of ultrasound waves, which has been shown to induce the piezoelectric effect in bovine cor-tical bone [Hosokawa, 2017]. The ultrasound waves do need to be capable of penetrating the soil with enough energy to still induce the piezoelectric effect in buried bones. Patents can be found of the design of a ground penetrating sonar which may be suitable for this purpose [Earp, 1998, Earp, 2000].

There are, however, limitations to this hypothetical method. Searching for piezoelectric materials would probably not be plausible in sandy soils with a high quartz content, since the piezoelectric properties of quartz are known to be significantly stronger than that of bone [Gandhi et al., 2014]. Furthermore, the method proposed above would not be human bone specific, but would detect any piezoelectric material within the sensor’s range.

2.2 Soil and vegetation

Soil disturbance

When a grave is dug, the existing vegetation is uprooted and destroyed. After this disturbance, the soil above the grave will need to be repopulated by

(8)

vege-tation which, in fresh graves, can provide an indication as to where the grave is located. Studies show that a succession of species does occur, and could possi-bly be used to locate clandestine graves [Caccianiga et al., 2012]. However, this vegetation recovers to be indistinguishable from its surroundings, and the effect of digging a grave and placing a body in the ground is indistinguishable from other activities which disturb the soil [Caccianiga et al., 2012]. In other words, this change in vegetation is not limited to buried remains.

Despite this, vegetation can still be an interesting target for the search of clandestine graves. Although the disturbance of the soil appears to be solely responsible for the temporary change the species populating the soil above the grave [Caccianiga et al., 2012], clandestine graves also hold a decomposing body. The decomposition process releases numerous VOCs into the surrounding soil, which are taken up by the roots of the vegetation growing above. As a result, it is possible that the elemental composition of vegetation above a grave is signif-icantly different from the surrounding vegetation, and this may be a promising target for remote sensing.

Ground depressions

After soil has been dug up to form a grave, it is placed back on top of the body to fill the grave up. This soil is known as backfilled soil, and is the target of a number of studies in the field of forensic geomorphology [Pringle et al., 2012]. Initially, this backfilled soil tends to be of a lower density compared to the rest of the soil, which has a number of effects. Firstly, it means that there is an increase in oxygen content in the backfilled soil, and secondly, it can cause the backfilled soil to initially form a raised mound [Killam, 2004].

Over time however, this raised mound tends to collapse as the material den-sity of the backfilled soil increases, causing a localised depression in ground elevation. As decomposition progresses, this localised depression can become increasingly pronounced as the chest cavity of the body collapses [Rodriguez and Bass, 1985]. The Netherlands is a relatively flat country where searching for localised depressions with remote sensing may be viable, and as such this may be interesting target for remote sensing. Unfortunately, no literature men-tions actual measurements of the extent of the depressions, or whether these depressions are in the range of millimeters or centimeters. In addition, most mentions of this soil depression are anecdotal, and no studies have been found which fully investigate this phenomenon.

Volatile organic compounds

The changes in the release of VOCs over time, as well as their chemical composi-tions, have been studied with regards to time of death and the training of cadaver dogs [Verheggen et al., 2017, Rosier et al., 2016, Rosier et al., 2015]. As a re-sult, a number of human and pig specific compounds have been identified (ethyl propionate, propyl propionate, propyl butyrate, ethyl pentanoate, diethyl disul-fide, 3-methylthio-1-propanol, methyl(methylthio)ethyl disulfide and pyridine),

(9)

as well as five esters which can distinguish human remains from pig remains (3-methylbutyl pentanoate, 3-methylbutyl 3-methylbutyrate, 3-methylbutyl 2-methylbutyrate, propyl hexanoate and butyl pentanoate) [Rosier et al., 2015]. Of these, 3-methylthio-1-propanol, methyl(methylthio)ethyl disulfide, ethyl pro-pionate, propyl propro-pionate, propyl butyrate, ethyl pentanoate and pyridine re-mained detectable after 12 months of decomposition [Rosier et al., 2015].

Vass et al. (2008) identify numerous VOCs from the top soil of graves which remain detectable even after 16 years of decomposition, which include various benzenes, halogen compounds, and aldehydes. Of note is carbon tetrachloride, which is found in their buried human bone samples, but not in any of the other mammalian samples which included deer, dog, and pig. In addition they ob-serve that clean skeletal bones of various species release compounds which have specific ratios of components for their species, to the extent that the ratios of these compounds could be used for identification purposes [Vass et al., 2008].

These findings show that specific VOCs could make interesting markers for the detection of not only fresh graves with actively decomposing bodies, but also for older graves which contain only skeletal remains. As a result, the previously identified VOCs are potential targets for remote sensing to specifically detect old clandestine graves with human remains. It must be noted however that the hundreds of VOCs which have been identified during the decomposition process in numerous studies are not identified in every study. In fact, there are only very few compounds which are consistently reported in decomposition studies. These inconsistencies could, among others, be caused by different sampling methods, different soil types, and different types of remains [Forbes and Perrault, 2014].

Most studies on VOCs are conducted by analysing soil samples. The com-position of VOCs in the soil, however, is not necessarily the same as the VOC composition in the air above a grave. This should be reflected on when methods of remote sensing which target VOCs are considered. Only one study could be found which compares the VOCs in soil with the VOCs in air. Forbes and Per-rault (2014) examine the differences in VOC composition between top soil and air [Forbes and Perrault, 2014]. Unfortunately their remains were not buried, and rather than comparing the soil above a grave and the air above it, they compared soil below the remains and the air above the remains. Nevertheless their findings provide an indication of the differences in VOC composition be-tween soil and air.

Forbes and Perrault (2014) find that there are indeed compounds which are specific to soil, as well as compounds which are specific to air. Only 23 percent of their identified compounds are found in both air and soil samples. The air specific compounds include mostly sulphides, nitrogen-containing, and aromatic compounds, while the soil specific compounds include mostly esters, hydrocar-bons, ketones, and aromatic compounds. They also report that reduced levels of these compounds were detected in air during skeletonisation, but not a reduced number of them. In soil, however, higher levels of compounds were detected during skeletonisation. This, as well as the increased number of soil specific VOCs, the authors attribute to the activity of microbes in the soil. [Forbes and Perrault, 2014]

(10)

Nitrogen

With the decomposition of the human body, nitrogen is released into the soil and dissolved into ground water [Fancher et al., 2017]. This increases the nitro-gen level in the soil, which can then be taken up by the roots of the vegetation growing above. According to Fancher et al. (2017) this increase in dissolved nitrogen levels persists for at least five years, making it potentially interesting for the detection of old clandestine graves.

Nitrogen levels in plants are highly positively correlated with the amount of leaf pigments such as chlorophyll [Mu˜noz-Huerta et al., 2013]. As such, tech-niques which can quantify the level of nitrogen directly or quantify it indirectly by measuring the concentration of leaf pigments are of interest to the detection of old clandestine graves.

Methane

When a body is buried in wet soil, oxygen diffusion rates are reduced which can lead to a more anaerobic environment [Dalva et al., 2012]. This slows the rate of decomposition, and can cause the formation and release of gases such as methane [Dent et al., 2004]. Dalva et al. (2012) studied old graves (approximately 50 years old) of animals in Quebec and show significant differences between the levels of methane in the air above a grave compared to background air methane levels. They suggest that the levels of methane can be useful in the search for clandestine graves, and they suggest using 2.022ppm - 1.89ppm as a cut-off of the minimum concentration of methane near buried remains [Dalva et al., 2012].

In the Netherlands, a large proportion of the soil is clayey, which tends to facilitate the creation of anaerobic conditions. Clayey soils tend to have a high water content, and the soil pores are often filled with water instead of air, which can make it more difficult for decomposition gases to diffuse to the surface [Alexander et al., 2016]. However, it is exactly this impeded diffusion of decomposition gases which may allow them to be detected at the surface of older graves rather than fresh graves, as water filled soil pores allow gas to diffuse through them up to ten thousand times slower than air soil filled pores [Skvortsova and Utkaeva, 2008].

Despite this, it has been suggested that methane is an unreliable marker, because methane interacts greatly with the soil matrix which can influence its concentration in the air above the grave [Dalva et al., 2015]. This is especially the case in well drained soils [Dalva et al., 2017]. On basis of these findings, methane may be an interesting marker for clayey soils, but not so much for the more sandy soils which are also frequently found in the Netherlands. In addition, methane should not be used as a marker in peat based soils, also common in the Netherlands, where the peat itself contains many sources of methane gas.

(11)

Nitrous oxide

In addition to the analysis of methane levels, the analysis of nitrous oxide levels could be useful for the detection of old clandestine graves. Soil nitrous oxide concentration increases dramatically when the soil is disturbed, where it can be up to ten times higher compared to undisturbed soil [Dalva et al., 2017]. After three years, the difference in nitrous oxide concentration caused by soil dis-turbance disappears, but with buried remains the concentration of soil nitrous oxide remained significantly greater than background nitrous oxide concentra-tions [Dalva et al., 2017]. Based on these results, Dalva et al. (2017) expect that increased soil nitrous oxide will persist over many years, a prediction which is echoed by Stokes et al. (2009) [Stokes et al., 2009].

3 Remote sensing techniques

In this section several promising remote sensing techniques will be examined and discussed with regards to the detection of old clandestine graves. Each of these techniques can be or has been mounted on drones. The literature about the techniques analysed in this chapter were found in fields of forensics, archaeology, geosciences, analytic chemistry, biology, medicine, engineering, and physics.

3.1 Thermal infrared multispectral scanner

How it works

The thermal infrared multispectral scanner (TIMS) is able the determine the temperature of 1 cm2 to an accuracy of 0.1°C [Palluconi and Meeks, 1985]. It does this through detecting the thermal infrared radiation which is emitted from objects, in the range of 8.2µm to 12.2 µm [Lahren et al., 1988]. The technique is used to gain land surface temperatures with an accuracy of 1.5°C [Qian et al., 2016] and has frequently been shown to be suitable for aerial applications such as drones [Palluconi and Meeks, 1985].

Current research

Thermal imaging has been used extensively in archaeology to find prehistoric roads, walls, and other human made structures [McGovern et al., 1995]. These structures, often made of stone, may have a higher specific heat capacity than the surrounding soil. As a result, the soil around these structures warm up and cool down at a lower rate which, if the structure is shallowly buried, is visible on the top layer of soil when analysed with TIMS.

With regards to buried human remains, a shallowly buried body may cre-ate a similar effect as that described above, if the remains have significantly different specific heat capacities compared to the surrounding soil. Bone has a relatively low specific heat capacity (ranging in literature between 0.44 to 0.75 J/gK [Chen and Gundjian, 1976]), whereas most soils have much higher specific

(12)

heat capacities (ranging from 0.80 J/gK for dry and course sand to 2.36 J/gK for fine and sandy clay [Hamdhan and Clarke, 2010]). As such, it may be pos-sible to detect skeletalised buried remains with TIMS.

Little literature could be found on the detection of old buried remains, hu-man or otherwise, with the use of TIMS, in either forensic, archaeologic, or other fields. One article reports that buried remains in advanced decomposi-tion, when only bone, skin, and cartilage are left, are no longer discernible from the surrounding soil through thermal imaging [DesMarais, 2014].

Limitations

Unfortunately with regards to clandestine graves, there are indications that thermal imaging only works with relatively shallow and fresh graves [Larson et al., 2011, DesMarais, 2014], making this technique potentially unsuitable for the detection of old clandestine graves. There is little literature to confirm this however, and it may be worth it to investigate whether TIMS can or cannot be used to locate old clandestine graves. The lack of literature on the use of TIMS in even archaeological burial sites however, where the technique has been frequently and widely applied, could be telling of its incompatibility, with null results potentially remaining unpublished.

3.2 Raman spectroscopy

How it works

Raman spectroscopy is used for the chemical analysis and imaging of materials. When light hits a material, some of the frequencies are absorbed, while others are reflected (scattered). Not all frequencies which are scattered remain unchanged. Small fractions of the scattered light gain a change of energy through interaction with molecular vibrations, which causes their frequency to change. This is known as the Raman scattering process, and this is the principle upon which Raman spectroscopy is based.

Raman spectroscopy is popular in medical practice because, unlike many other chemical imaging techniques, it experiences minimal interference from water molecules [Butler et al., 2016]. This property also makes it suitable for the chemical imaging of soil and vegetation. Due to this property the technique is also used for in vivo studies in the medical field.

Current research

A specific type of Raman spectroscopy, known as spatially offset Raman spec-troscopy, is used to analyse bone beneath skin in in-vivo studies [Mandair and Morris, 2015]. As a result, the spectral graph of human bone is well docu-mented. Unfortunately, the commonly used setup for spatially offset Raman spectroscopy is not capable of searching deep enough to find buried remains, typically detecting several millimetres beneath a surface. This depth could be increased, however, by further separation of the source and the detector.

(13)

Further efforts have led to the development of lightweight Raman spectro-scopes (500g) which would be light enough for drone applications [Meng et al., 2015]. Although these Raman spectroscopes are not capable of analysing deep beneath the surface of the ground, and as such are difficult to use to find skeletal remains directly, they can be useful for the analysis of vegetation and soil.

As discussed in section 2.2, dissolved nitrogen levels are higher in the soil surrounding a buried and decomposing body [Fancher et al., 2017]. As such one would expect the nitrogen level in the vegetation growing on top of a clandes-tine grave to have higher nitrogen levels, which has been observed by Mundorff et al. [Mundorff et al., 2014]. Raman spectroscopy could be an interesting for the detection of these nitrogen levels, where significantly higher levels of nitro-gen compared to the background level of nitronitro-gen could indicate a buried body. However, since Raman Spectroscopy is a highly sensitive technique, a lot of background noise could be expected in the readings of a complex matrix such as a leaf.

Alternatively, Raman spectroscopy could be used to analyse the gaseous compounds close to the ground. As a body decomposes, VOCs are released which may be detected through Raman spectroscopy. As discussed in section 2.2, VOCs have been identified which are present even after 16 years of burial [Vass et al., 2008], which can make for interesting markers. However, little research still exists on the VOC compounds found in air rather than the soil, while evidence exists to suggest that VOC composition in air is not the same as in soil [Forbes and Perrault, 2014]. Dalva et al. (2012) provide a cut-off point of 1.89ppm - 2.022ppm for the concentration of methane above a grave, which may be an interesting marker for Raman gas spectroscopy [Dalva et al., 2012].

Unfortunately no literature could be found on the use of Raman spectroscopy to detect buried remains, human or otherwise. Raman spectroscopy is not un-known to forensic science however, where it has been used to study a vari-ety of body fluids, explosives, and gunshot residues [McLaughlin and Lednev, 2011, Hokr et al., 2014].

Literature can be found about the general analysis of soils and minerals in situ, especially in the field of planetary sciences. Haskin et al. (1997) discuss the best method for the in situ analysis of minerals on planetary surfaces by taking many narrow beam spectra from different locations, which could possi-bly be applied to the search of VOCs in soil [Haskin et al., 1997]. Furthermore, studies in the biomedical field have shown that Raman spectroscopy is capable of analysing particular elements within a complex matrix, such as the presence of specific micro-organisms in soil [St¨ockel et al., 2016].

Limitations

To use Raman spectroscopy to search beneath the earth’s surface, the source and detector need to be separated by some distance depending on the required depth. This could lead to practical problems with regards to drone applications when the required distance becomes too large for one drone to carry both the source and the detector. Furthermore, this method will only allow samples at

(14)

a specific depth to be analysed, and will miss anything falling outside of this range. As such, an estimate will need to be made about the burial depth.

With regards to gas sampling and analysis, atmospheric conditions can have a substantial effect. Wind may interfere with readings, causing the detection of reduced amounts. In addition, when applied to drones, gas spectroscopy will not work if the drone makes use of vertical lift. It is also possible that Raman spectroscopy will prove to be too sensitive for the detection of clandestine graves, and that there will be a significant amount of noise in the readings. This noise to signal ratio may possibly be reduced by the analysis method suggested by Haskin et al. [Haskin et al., 1997].

Raman spectroscopy can be combined with more reflection data with the use of hyperspectral imaging, which will be discussed in section 3.4.

3.3 Light detecting and ranging

How it works

There are two types of light detecting and ranging (LIDAR) systems, called topographic and bathymetric. Topographic LIDAR is used for the taking of precise measurements of the land with a near-infrared laser, and bathymetric LIDAR uses visible green light, which is capable of penetrating water, to make precise measurements of lake, river, sea, and ocean floors. [NOAA, 2017]

The airborne LIDAR emits a pulse of light, either near-infrared or visible green, straight to the surface, where the light is reflected back to the sensor. The LIDAR then records the time of flight of the light pulse, from which the distance to the surface is calculated. Using GPS information and inertial mea-surement units, a point cloud is created, where each point contains three spatial coordinates (height, latitude and longitude) to create a 3D map of the surface. [NOAA, 2017]

Because LIDAR is an active technique which sends out its own light pulses straight down to the surface, rather than a passive technique which relies on reflected sunlight, it has the ability to make measurements in between the gaps of vegetative canopy and map the solid ground beneath the vegetation cover-ing it [Corcoran, 2016]. This makes it preferable to techniques such as aerial photography, which cannot record through dense vegetation.

Current research

LIDAR has been applied extensively in archaeology, where it has been used to locate archaeological structures under dense vegetation cover [Chase et al., 2011]. Since LIDAR is capable of detecting small depressions in the soil even below vegetation cover, it may be useful for the detection of depressions caused by the collapse of backfilled soil in clandestine graves, as discussed in section 2.2. In order to achieve this, LIDAR would need to be sensitive enough to detect the small elevation differences caused by these soil depressions, and it would need to be known how large these depressions tend to be. There are LIDAR like the

(15)

Riegl VQ-480i which has a reported accuracy and precision of 20mm [RIEGL, 2015]. Unfortunately, literature on the measurement of soil depression above a grave is not available, and so little can be said on the feasibility of identifying these depressions over a large area with LIDAR.

Somewhat surprisingly, very little literature could be found on the feasibility of searching for clandestine graves with LIDAR. The only relevant forensic study on LIDAR found was one PhD dissertation by Corcoran (2016) which could not detect distinctive elevation changes at the 19 - 27 months old burial locations [Corcoran, 2016]. Corcoran (2016) also found that there was no significant difference in the elevation loss (after the initial elevation gain from backfill soil) between the control grave, and the graves which contained human remains. Some degree of elevation loss was visible in a simulated mass grave (five human bodies) compared to the surroundings, but the loss of elevation varied greatly over the expanse of the grave [Corcoran, 2016].

Limitations

Searching for ground depressions as a marker for buried remains is a potentially unreliable method, due to the natural mounds and depressions which occur in a landscape, and due to the observations that these depressions do not occur with all buried remains. In addition to this, LIDAR technology can be expensive to use, especially when covering large areas. A technique such as photogrammetry (where multiple pictures of the same landscape taken from different angles are used to make a point cloud 3D model) tends to be less expensive than LIDAR, but suffers from different limitations. It is a passive technique and thus relies on the quality of available light, and cannot penetrate vegetative cover.

Furthermore, although the active nature of LIDAR technology allows it to penetrate through some of the vegetative cover, ground point cloud data re-covered from below denser vegetation can be sparse and insufficient for the proper assessment of elevation changes [Corcoran, 2016]. For this reason, LI-DAR readings can be impeded by simple obstructions such as leaves collecting in any existing depressions caused by buried remains. All of these factors are important to keep in mind when considering the use of LIDAR in the search for clandestine graves, especially since, if improperly used, LIDAR can result in false negative results.

Currently, far too little research exists on how LIDAR can be used for the detection of buried human remains, and too little literature exists which reports measurements of the differences in geomorphology above buried remains.

3.4 Hyperspectral imaging

How it works

All previously discussed techniques can be encompassed within one promising new technique known as hyperspectral imaging. As with all remote sensing techniques, this technique works on the concept that when light hits an object,

(16)

the light is transmitted, reflected, or absorbed. The hyperspectral imager sends out many different wavelengths from the electromagnetic spectrum [Kalacska and Bell, 2006] with different hyperspectral imagers analysing different ranges of wavelengths. Based on the reflected wavelengths, a spectral signature is created, which is used for comparisons and identifications. Hyperspectral imaging is also capable of chemical imaging, which includes techniques such as Raman spectroscopy. The spectral signature of soil is largely a result of its chemical and physical properties, such as organic and inorganic solids, as well as the amount of air and water [Stein et al., 1999].

As a result of its reliance on reflected light, hyperspectral imaging cannot detect below the surface of either the earth or the vegetation canopy. For hyperspectral imaging it is therefore important to focus on markers of buried remains which are present on the very surface of the landscape.

Current research

Kalacska et al. (2009) used airborne hyperspectral imagery to detect a mass grave in Costa Rica. They find that a grave containing mammalian remains can be distinguished from an empty grave using airborne hyperspectral imagery one month after burial. Their study shows that airborne hyperspectral imagery can be used as an initial form of reconnaissance to find potential sites with buried remains. It is important to note however that one month after burial vegeta-tion had not fully regrown, and that the difference in vegetavegeta-tion cover between different graves could have affected their results. [Kalacska et al., 2009]

Kalacska et al. (2006) found that specific bands from the near infrared and shortwave infrared are vital to distinguish the mass grave from controls [Kalac-ska and Bell, 2006]. This is supported by the findings of Silván-Cárdenas et al. (2017), who state that the wavelength intervals which are most useful for distin-guishing buried remains are found between 700nm - 1800nm [Silván-Cárdenas et al., 2017]. They also state that the grave needs to be older than three months in order for the effect of decay to become apparent, which conflicts with reports from Kalacska et al. (2009) who succeeded in identifying buried remains after one month, and also conflicts with reports from Leblanc et al. (2014), who de-tected single graves of both one and four months old. It has been observed that the presence of soil and vegetation can influence the spectral signature [Gibson and Power, 2013]. It is therefore possible that the discrepancies between the studies are caused by the varying time taken for vegetation to regrow, and that bare soil does not allow a distinction between control graves and freshly buried remains to be made.

Leblanc et al. (2014) used an hyperspectral imager with a range of 408nm -2524nm to locate buried pig remains in a blind study. They had a 100% success rate in locating the buried remains, but also identified one extra location as a possible grave which was outside of the boundaries of the testing site [Leblanc et al., 2014]. Unfortunately no information has been provided about the extra location, which may have been a false positive. Leblanc et al. (2014) demon-strated that the detection of single graves is possible with the hyperspectral

(17)

imager, although their study did not include control graves without buried re-mains. As such, it may be that their model differentiates between disturbed and undisturbed soil more so than the presence of buried remains. If this is the case, then it would make their model unsuitable for the detection of old clandestine graves, where the disturbed soil has settled, and cause many false positive iden-tifications in case work. Further research on what their model actually detects would be of added value.

Unfortunately, while the mentioned studies have attempted, and to varying degrees succeeded, to locate buried remains, they did not attempt to charac-terise what signatures are unique to buried remains. Due to this, it is not possible to actually search for burial-specific signatures outside of their experi-mental settings. This is still a large gap in the current literature, and this must be expanded on if hyperspectral imaging is ever to be used in the search for clandestine graves by law enforcement.

Limitations

It is important to note that the spectral signature depends on many variables, such as the degree of bare soil, and that the findings of one location cannot be directly translated to a different location. For instance, spectral signatures developed for buried remains in Dutch clayey soils cannot be directly translated to a spectral signature for buried remains in Dutch sandy soils, or possibly to German clayey soils. This is important to understand to avoid the misuse of hyperspectral imagery and avoid overly pessimistic or optimistic results.

Unfortunately hyperspectral imaging is expensive, and requires a lot of post-processing - results are not analysed and provided in situ or in real time [Silv´ an-C´ardenas et al., 2017]. A last limitation is that the data acquired from a hyper-spectral imaging is complex and requires a large amount of computing power to process. Nevertheless, hyperspectral imaging shows a lot of promise for the wider searches of clandestine graves and to help narrow down suspect locations. Hyperspectral imagers are becoming lighter and less expensive, and they may become more feasible for casework in the future.

4 Further challenges

Although utilising drones for remote sensing is an attractive prospect, the lim-itations of this method of transportation should be kept in mind. Many drones use vertical lift in order to fly, which can greatly disrupt the ground below the drone and displace both gas and objects, disturbing the crime scene. Alterna-tive drones which make use of gliding techniques may be more suitable to avoid these limitations, but drones such as these would not be capable of hovering in one place. As such, both the method of remote sensing as well as the type of drone which transports it need to be customised on a case-to-case basis. In addition to this, local laws and regulations must be followed, as discussed in section 1.2.

(18)

5 Conclusions

Thermal infrared multispectral scanning (TIMS) is a technique which has re-mained largely unexplored with regards to the detection of buried remains. Its main advantage would be that it is a relatively inexpensive technique to use, but while TIMS may hold promise for fresh buried remains in early decomposi-tion, it is less promising for buried remains in advanced decomposition such as skeletalisation. In conclusion, TIMS is not a high priority technique to invest in further with regards to the detection of old buried remains.

Raman spectroscopy is a promising technique which could make use of the large body of forensic literature on the topic of volatile organic compounds (VOCs). Research into the application of Raman spectroscopy to the remote detection of buried remains is encouraged, but due to the lack of current studies in this area, the technique is not ready for use in case work. When research into the applicability of Raman spectroscopy to gaseous VOCs is conducted, care should be taken about the type of drone employed to avoid false negative results due to vertical lift.

Light detecting and ranging (LIDAR) is capable of making highly precise elevation models of the soil, which could be interesting in the search for de-pressions occasionally seen over a grave. However, the technique is expensive, and has limited penetration through vegetation, causing the technique to be relatively easily obstructed by, for example, fallen leaves. In addition there is simply too little literature studying the geomorphology of graves, and no mea-surements of soil depressions can be found on which to base searches. Only one PhD thesis could be found attempting to detect buried human remains with LIDAR, and they report negative results. As such, LIDAR should not be used in casework to look for soil depressions as a marker for old clandestine graves, and it is perhaps also not a promising a technique to focus on developing for this purpose. Photogrammetry, also briefly discussed, would suffer from similar limitations and as such is also not recommended for the search of clandestine graves through soil depressions, although it, as well as LIDAR, has proven to be useful in other forensic domains such as crime scene investigation and docu-mentation [Thali et al., 2000, Thali et al., 2003].

Although hyperspectral imaging is a promising, albeit for now expensive, tool for the search of old clandestine graves, it is not a technique which is ready for use in casework. Too little is as of yet known about burial-specific spectral signatures for hyperspectral imaging to be of proper use outside of experimen-tal setups. Further research into burial specific spectral signatures, as well as the effect of vegetation cover percentage, is needed. Furthermore, since the hy-perspectral imager is especially interesting for widespread searching when little information is available on the grave’s location, it would be a very expensive technique to use both financially and computationally, but future innovations may help reduce these costs to make hyperspectral imaging more viable for

(19)

case-work.

As a general remark, there is still a large disconnect between the discov-ery and identification of markers for buried human remains, and the ability of remote sensing techniques to detect and use these markers. If aerial remote sens-ing is to be used in casework in the future, a significant amount of research into markers specific to the available remote sensing techniques, such as the identi-fication of spectral signatures which specifically describe the probable presence of buried human remains, will need to be conducted. The identification of such markers specific to buried human remains is an important first step to this, but it is not sufficient to be able to use most remote sensing techniques outside of experimental set-ups. further research into the piezoelectric effects of dry bone, and methods to detect fluctuations in weak magnetic fields, may also be worth investing in.

There are very promising aerial remote sensing techniques which can be ap-plied to drones, but as of yet, they do not appear to be ready for use in case work.

References

[Alexander et al., 2016] Alexander, M. B., Hodges, T. K., Wescott, D. J., and Aitkenhead-Peterson, J. A. (2016). The effects of soil texture on the ability of human remains detection dogs to detect buried human remains. Journal of Forensic Sciences, 61(3):649–655.

[Butler et al., 2016] Butler, H. J., Ashton, L., Bird, B., Cinque, G., Curtis, K., Dorney, J., Esmonde-White, K., Fullwood, N. J., Gardner, B., Martin-Hirsch, P. L., Walsh, M. J., McAinsh, M. R., Stone, N., and Martin, F. L. (2016). Using raman spectroscopy to characterize biological materials. Nature Protocols, 11(Supplement C):664–687.

[Caccianiga et al., 2012] Caccianiga, M., Bottacin, S., and Cattaneo, C. (2012). Vegetation dynamics as a tool for detecting clandestine graves. Journal of Forensic Sciences, 57(4):983–988.

[Carter et al., 2007] Carter, D. O., Yellowlees, D., and Tibbett, M. (2007). Ca-daver decomposition in terrestrial ecosystems. Naturwissenschaften, 94(1):12– 24.

[Chase et al., 2011] Chase, A. F., Chase, D. Z., Weishampel, J. F., Drake, J. B., Shrestha, R. L., Slatton, K. C., Awe, J. J., and Carter, W. E. (2011). Airborne lidar, archaeology, and the ancient maya landscape at caracol, belize. Journal of Archaeological Science, 38(2):387 – 398.

[Chen and Gundjian, 1976] Chen, H. L. and Gundjian, A. A. (1976). Specific heat of bone. Medical and biological engineering, 14(5):548–550.

(20)

[Corcoran, 2016] Corcoran, K. A. (2016). A Characterization of Human Burial Signatures using Spectroscopy and LIDAR. PhD thesis, University of Ten-nessee.

[Dalva et al., 2012] Dalva, M., Kal´acska, M., Moore, T. R., and Costopoulos, A. (2012). Detecting graves with methane. Geoderma, 189-190(Supplement C):18 – 27.

[Dalva et al., 2017] Dalva, M., Moore, T., Kalacska, M., and Leblanc, G. (2017). Nitrous oxide, methane and carbon dioxide patterns and dynam-ics from an experimental pig mass grave. Forensic Science International, 277(Supplement C):229 – 240.

[Dalva et al., 2015] Dalva, M., Moore, T., Kalacska, M., Leblanc, G., and Costopoulos, A. (2015). Nitrous oxide, methane and carbon dioxide dynamics from experimental pig graves. Forensic Science International, 247(Supple-ment C):41 – 47.

[Dent et al., 2004] Dent, B. B., Forbes, S. L., and Stuart, B. H. (2004). Review of human decomposition processes in soil. Environmental Geology, 45(4):576– 585.

[DesMarais, 2014] DesMarais, A. M. (2014). Detection of cadaveric remains by thermal imaging cameras. Journal of Forensic Identification, 64(5):489. [Earp, 1998] Earp, R. (1998). Ground penetrating sonar. US Patent 5,719,823. [Earp, 2000] Earp, R. (2000). Mobile sensor for grounds penetrating sonar. US

Patent 6,131,695.

[Fancher et al., 2017] Fancher, J., Aitkenhead-Peterson, J., Farris, T., Mix, K., Schwab, A., Wescott, D., and Hamilton, M. (2017). An evaluation of soil chemistry in human cadaver decomposition islands: Potential for estimating postmortem interval (pmi). Forensic Science International, 279(Supplement C):130 – 139.

[Forbes et al., 2016] Forbes, S., Troobnikoff, A., Ueland, M., Nizio, K., and Perrault, K. (2016). Profiling the decomposition odour at the grave surface before and after probing. Forensic Science International, 259(Supplement C):193 – 199.

[Forbes and Perrault, 2014] Forbes, S. L. and Perrault, K. A. (2014). Decom-position odour profiling in the air and soil surrounding vertebrate carrion. PLOS ONE, 9(4):1–12.

[Gandhi et al., 2014] Gandhi, A. A., Wojtas, M., Lang, S. B., Kholkin, A. L., and Tofail, S. A. M. (2014). Piezoelectricity in poled hydroxyapatite ceramics. Journal of the American Ceramic Society, 97(9):2867–2872.

(21)

[Gibson and Power, 2013] Gibson, P. and Power, W. (2013). Introductory Re-mote Sensing Principles and Concepts. Taylor & Francis.

[Hamdhan and Clarke, 2010] Hamdhan, I. N. and Clarke, B. G. (2010). Deter-mination of thermal conductivity of coarse and fine sand soils. In Proceedings of World Geothermal Congress.

[Haskin et al., 1997] Haskin, L. A., Wang, A., Rockow, K. M., Jolliff, B. L., Korotev, R. L., and Viskupic, K. M. (1997). Raman spectroscopy for mineral identification and quantification for in situ planetary surface analysis: A point count method. Journal of Geophysical Research: Planets, 102(E8):19293– 19306.

[Hokr et al., 2014] Hokr, B. H., Bixler, J. N., Noojin, G. D., Thomas, R. J., Rockwell, B. A., Yakovlev, V. V., and Scully, M. O. (2014). Single-shot stand-off chemical identification of powders using random raman lasing. Proceedings of the National Academy of Sciences, 111(34):12320–12324.

[Hosokawa, 2017] Hosokawa, A. (2017). Investigation of piezoelectric anisotropy of bovine cortical bone at an ultrasound frequency by coupling an experi-ment and a simulation. The Journal of the Acoustical Society of America, 142(2):EL184–EL189.

[Kalacska and Bell, 2006] Kalacska, M. and Bell, L. (2006). Remote sensing as a tool for the detection of clandestine mass graves. Canadian Society of Forensic Science Journal, 39(1):1–13.

[Kalacska et al., 2009] Kalacska, M. E., Bell, L. S., Arturo Sanchez-Azofeifa, G., and Caelli, T. (2009). The application of remote sensing for detecting mass graves: An experimental animal case study from costa rica*. Journal of Forensic Sciences, 54(1):159–166.

[Killam, 2004] Killam, E. (2004). The Detection of Human Remains. Charles C Thomas.

[Lahren et al., 1988] Lahren, M. M., Schweickert, R. A., and Taranik, J. V. (1988). Analysis of the northern sierra accreted terrane, california, with air-borne thermal infrared multispectral scanner data. Geology, 16(6):525. [Larson et al., 2011] Larson, D. O., Vass, A. A., and Wise, M. (2011). Advanced

scientific methods and procedures in the forensic investigation of clandestine graves. Journal of Contemporary Criminal Justice, 27(2):149–182.

[Leblanc et al., 2014] Leblanc, G., Kalacska, M., and Soffer, R. (2014). Detec-tion of single graves by airborne hyperspectral imaging. Forensic Science International, 245(Supplement C):17 – 23.

(22)

[Lee et al., 2011] Lee, K. Y. K., Nyein, M. K., Moore, D. F., Joannopoulos, J., Socrate, S., Imholt, T., Radovitzky, R., and Johnson, S. G. (2011). Blast-induced electromagnetic fields in the brain from bone piezoelectricity. Neu-roImage, 54(Supplement 1):S30 – S36. International Brain Mapping and In-traoperative Surgical Planning Society (IBMISPS)2009 Supplement.

[Mandair and Morris, 2015] Mandair, G. S. and Morris, M. D. (2015). Con-tributions of raman spectroscopy to the understanding of bone strength. BoneKEy Reports, 4.

[Mann et al., 1990] Mann, R. W., Bass, W. M., and Meadows, L. (1990). Time since death and decomposition of the human body: variables and observations in case and experimental field studies. Journal of Forensic Science, 35(1):103– 111.

[McGovern et al., 1995] McGovern, P. E., Sever, T. L., Myers, J. W., Myers, E. E., Bevan, B., Miller, N. F., Bottema, S., Hongo, H., Meadow, R. H., Kuniholm, P. I., Bowman, S. G. E., Leese, M. N., Hedges, R. E. M., Matson, F. R., Freestone, I. C., Vaughan, S. J., Henderson, J., Vandiver, P. B., Tu-mosa, C. S., Beck, C. W., Smith, P., Child, A. M., Pollard, A. M., Thuesen, I., and Sease, C. (1995). Science in archaeology: A review. American Journal of Archaeology, 99(1):79–142.

[McLaughlin and Lednev, 2011] McLaughlin, G. and Lednev, I. K. (2011). Po-tential application of raman spectroscopy for determining burial duration of skeletal remains. Analytical and Bioanalytical Chemistry, 401(8):2511–2518. [Meng et al., 2015] Meng, Z., Petrov, G. I., Cheng, S., Jo, J. A., Lehmann,

K. K., Yakovlev, V. V., and Scully, M. O. (2015). Lightweight raman spec-troscope using time-correlated photon-counting detection. Proceedings of the National Academy of Sciences, 112(40):12315–12320.

[Ministerie van Infrastructuur en Milieu, 2016] Ministerie van Infrastruc-tuur en Milieu, H. (2016). Regelgeving drones - verschillen in regelgeving recreatief-beroeps. Website, Ministerie van Infrastructuur en Milieu.

[Mundorff et al., 2014] Mundorff, A., Figueroa-Soto, C., and Commisso., R. (2014). Assessing the usefulness of plant 15n as a new forensic tool to iden-tify clandestine burials. Proceedings of the American Academy of Forensic Sciences Annual Meeting.

[Mu˜noz-Huerta et al., 2013] Mu˜noz-Huerta, R. F., Guevara-Gonzalez, R. G., Contreras-Medina, L. M., Torres-Pacheco, I., Prado-Olivarez, J., and Ocampo-Velazquez, R. V. (2013). A review of methods for sensing the nitro-gen status in plants: Advantages, disadvantages and recent advances. Sensors, 13(8):10823–10843.

[NOAA, 2017] NOAA (2017). Lidar — light detection and ranging — is a remote sensing method used to examine the surface of the earth. Webpage from https://oceanservice.noaa.gov.

(23)

[Palluconi and Meeks, 1985] Palluconi, F. D. and Meeks, G. R. (1985). Thermal infrared multispectral scanner (tims): an investigator’s guide to tims data. [Pringle et al., 2012] Pringle, J., Ruffell, A., Jervis, J., Donnelly, L., McKinley,

J., Hansen, J., Morgan, R., Pirrie, D., and Harrison, M. (2012). The use of geoscience methods for terrestrial forensic searches. Earth-Science Reviews, 114(1):108 – 123.

[Putman, 1978] Putman, R. J. (1978). Flow of energy and organic matter from a carcase during decomposition: Decomposition of small mammal carrion in temperate systems 2. Oikos, 31(1):58–68.

[Qian et al., 2016] Qian, Y.-G., Wang, N., Ma, L.-L., Liu, Y.-K., Wu, H., Tang, B.-H., Tang, L.-L., and Li, C.-R. (2016). Land surface temperature retrieved from airborne multispectral scanner mid-infrared and thermal-infrared data. Opt. Express, 24(2):A257–A269.

[RIEGL, 2015] RIEGL (2015). Datasheet riegl vq-480i. Datasheet, RIEGL laser measurement systems.

[Rodriguez and Bass, 1985] Rodriguez, W. and Bass, W. (1985). Decomposi-tion of buried bodies and methods that may aid in their locaDecomposi-tion. Journal of Forensic Sciences, 30(3):836 – 852.

[Rosier et al., 2016] Rosier, E., Loix, S., Develter, W., de Voorde, W. V., Tyt-gat, J., and Cuypers, E. (2016). Time-dependent voc-profile of decomposed human and animal remains in laboratory environment. Forensic Science In-ternational, 266(Supplement C):164 – 169.

[Rosier et al., 2015] Rosier, E., Loix, S., Develter, W., Van de Voorde, W., Tytgat, J., and Cuypers, E. (2015). The search for a volatile human specific marker in the decomposition process. PLOS ONE, 10(9):1–15.

[Senn and Weems, 2013] Senn, D. R. and Weems, R. A. (2013). Manual of forensic odontology. CRC Press.

[Silván-Cárdenas et al., 2017] Silván-Cárdenas, J. L., Corona-Romero, N., Madrigal-Gómez, J. M., Saavedra-Guerrero, A., Cortés-Villafranco, T., and Coronado-Juárez, E. (2017). On the Detectability of Buried Remains with Hy-perspectral Measurements, pages 201–212. Springer International Publishing, Cham.

[Skvortsova and Utkaeva, 2008] Skvortsova, E. B. and Utkaeva, V. F. (2008). Soil pore space arrangement as a geometric indicator of soil structure. Eurasian Soil Science, 41(11):1198–1204.

[Stein et al., 1999] Stein, A., van der Meer, F., and Gorte, B. (1999). Phys-ical principles of remote sensing In: Spatial Statistics for Remote Sensing. Springer, Dordrecht.

(24)

[Stokes et al., 2009] Stokes, K. L., Forbes, S. L., Benninger, L. A., Carter, D. O., and Tibbett, M. (2009). Decomposition Studies Using Animal Models in Con-trasting Environments: Evidence from Temporal Changes in Soil Chemistry and Microbial Activity, pages 357–377. Springer Netherlands, Dordrecht. [Stöckel et al., 2016] Stöckel, S., Kirchhoff, J., Neugebauer, U., Rösch, P., and

Popp, J. (2016). The application of raman spectroscopy for the detection and identification of microorganisms. Journal of Raman Spectroscopy, 47(1):89– 109. JRS-15-0205.R1.

[Thali et al., 2003] Thali, M. J., Braun, M., Brueschweiler, W., and Dirnhofer, R. (2003). ‘morphological imprint’: determination of the injury-causing weapon from the wound morphology using forensic 3d/cad-supported pho-togrammetry. Forensic Science International, 132(3):177 – 181.

[Thali et al., 2000] Thali, M. J., Braun, M., Br¨uschweiler, W., and Dirnhofer, R. (2000). Matching tire tracks on the head using forensic photogrammetry. Forensic Science International, 113(1):281 – 287. Selected papers from the 4th International Symposium Advances in.

[Vass et al., 2008] Vass, A. A., Smith, R. R., Thompson, C. V., Burnett, M. N., Dulgerian, N., and Eckenrode, B. A. (2008). Odor analysis of decomposing buried human remains*. Journal of Forensic Sciences, 53(2):384–391. [Verheggen et al., 2017] Verheggen, F., Perrault, K. A., Megido, R. C., Dubois,

L. M., Francis, F., Haubruge, E., Forbes, S. L., Focant, J.-F., and Stefanuto, P.-H. (2017). The odor of death: An overview of current knowledge on char-acterization and applications. BioScience, 67(7):600–613.

[Whedon and Heaney, 2017] Whedon, G. D. and Heaney, R. P. (2017). Bone. Website, Encyclopædia Britannica.

A

Search Strategy

Only articles, books, and web pages written in English or Dutch were used to produce this thesis due to personal language limitations.

A general Google search on the compositions of Dutch soils, the composition of skeletal remains, and the types if remote sensing techniques used in palaeon-tology and archaeology was initially conducted. This was done to gather back-ground information and to gain an overview of the field before analysing the scientific literature.

Next a search was done for the key words ’clandestine graves’ in the Univer-sity database of the UniverUniver-sity of Amsterdam and in Google scholar to gain a first impression on the related challenges and methodologies of detecting clan-destine graves. Further keywords gained from the articles found were used to expand on initial findings and ideas.

(25)

clandestine graves’ and ’remote sensing buried remains’ were searched to find articles about which remote sensing techniques are currently used, and which can be mounted on drones. After this, searches were made about some of the more promising remote sensing techniques found with the initial Google search and gathered from the scientific literature.

The bibliographies of relevant literature found with the search terms were perused for further relevant literature, which expanded the scope of literature read into non-forensic fields.

As the writing progressed, more detailed information was needed and spe-cific search terms were used for topics such as the sensitivities of methods, the properties of specific chemicals, and time spans.

Remote Sensing for the Detection of Old Clandestine Graves