The suitability of using domestic pigs (Sus spp.) as human proxies in the geophysical detection of clandestine graves

Research in many forensic science fields commonly uses domestic pigs (Sus spp.) as proxies for human remains, due to their physiological and anatomical similarities, as well as being more readily available. Unfortunately, previous research, especially that which compares the decompositional process, has shown that pigs are not appropriate proxies for humans. To date, there has not been any published research that specifically addresses whether domestic pigs are adequate human proxies for the geophysical detection of clandestine graves. As such, the aim of this paper was to compare the geophysical responses of pig cadavers and human donor graves, in order to determine if pigs can indeed be used as adequate human proxies. To accomplish this, ground penetrating radar (GPR) and electrical resistivity tomography (ERT) responses on single and multiple pig cadaver graves were compared to single and multiple human donor graves, all of which are in known locations within the same geological environment. The results showed that under field conditions, both GPR and ERT were successful at observing human and pig burials, with no obvious differences between the detected geophysical responses. The results also showed that there were no differences in the geophysical responses of those who were clothed and unclothed. The similarity of the responses may reflect that the geophysical techniques can detect graves despite what their contents are. The study implications suggest that experimental studies in other soil and climate conditions can be easily replicated, benefiting law enforcement with missing persons cases.


Highlights
• The suitability of pigs as human proxies is contested in various areas of forensic science research.
• GPR and ERT can successfully detect both human and pig burials.
• Pigs can be used as adequate human proxies in the geophysical detection of clandestine graves.
• Experimental pig studies can be replicated worldwide without the barrier of access to human donors.

| INTRODUC TI ON
Finding clandestine graves (defined as hidden/illegal burials containing human remains [1]) is a difficult, time-consuming, and laborintensive task.The importance of locating these graves, however, cannot be overstated.Not only does locating these graves facilitate the judicial process [2,3], but it can also provide the family of the missing person with answers about their loved one's death [4,5].
Research in the field of forensic science, most notably in forensic anthropology, entomology, and taphonomy (defined as the study of the phenomena that affect human remains at the time of and after death [32]) fields, often uses domestic pig (Sus spp.) cadavers as a proxy for human remains (see Matuszewki et al. [33] for a comprehensive list of forensic entomology and taphonomy studies that have used pig and other animal cadavers as human analogs).Not only are pig cadavers physiologically and chemically similar to humans, with a similar size, tissue-to-body fat ratio, and skin and hair types [34,35], but they are also more readily available for research purposes [36].
Fortunately, countries such as Australia, Canada, the USA, and The Netherlands (see Pecsi et al. [39] for a list of the 12 existing facilities that undertake research on donated human cadavers) have research facilities dedicated to human cadaver research (gaining bequeathal via body donation programs), that have allowed researchers to better understand the post-mortem interval (defined as the time elapsed between death and discovery [40]).Despite this, access to this type of facility is not widely available and there may not be enough human cadavers available for all the planned studies, therefore the use of animal cadavers remains valuable [38].
Although near-surface geophysical techniques have been used in forensic cases (see [25-27, 41, 42]), most of the literature focuses on experimental studies that test ground penetrating radar (GPR) and resistivity (fixed-probe resistivity, and electrical resistivity tomography-ERT) methods in their ability to locate pig burials, as a simulation of human burials [35,[43][44][45][46][47][48].Despite this, there has been no published research confirming the suitability of domestic pigs as human proxies in the geophysical detection of clandestine graves.
As such, the aim of this paper is to determine if pig cadavers are adequate human proxies in this context.The objective is therefore to compare the geophysical responses of identically conducted GPR and ERT surveys of domestic pig burials (20 months post-burial) and human donor burials (56 and 78 months post-burial).If successful, geophysical techniques can continually be tested using pig cadavers in different soil and climate conditions, which is less problematic with the ethical constraints around human donors.

| Study sites and graves
The two study sites are located at the Australian Facility for Taphonomic Experimental Research (AFTER), with the pig cadaver and human donor graves at the animal and human sites, respectively.
The two sites are separated by a road, with approximately 100 m between them.The sites are 19 m above sea level and located in Western Sydney, New South Wales (see Figure 1A).The bedrock geology is Hawkesbury sandstone with minor shale and laminite lenses [49] and the soil on the site is a poorly developed sandy Tenosol [50] lacking sedimentary or pedogenic structures.
The burial types included in this research are one single human grave (HS1), one mass human grave with six individuals (HM1), two single pig graves (PS1 and PS2), and one mass grave with three pig cadavers (PM1), the specifics of which can be found in Table 1.With the exception of PS2 (see Table 1 and Figure 1), the depths of each grave were specifically chosen to mimic what could be found in homicide cases where the victim(s) are placed in covert burials, which is generally 0.5 m deep [45].PS2 was intentionally dug deeper to 1.8 m mimic typical cemetery burials.
The 50-80 kg domestic pigs (Sus spp.) were purchased from a local abattoir, and to abide by the ethical conditions set out by AFTER and the University of Newcastle's Animal Care and Ethics F I G U R E 1 (A) General area of the AFTER animal and human sites.The site is located in Western Sydney, New South Wales, Australia.For confidentiality reasons, the exact location is not included.(B) Schematic of the pig cadaver graves at the animal site.The line and arrow across the graves denote the survey line which was 16 m long.PS1 is 2 m × 0.75 m × 0.5 m, PM1 is 2 m × 3 m × 1 m, and PS2 is 2 m × 1 m × 1.8 m.There is 2.75 m between PS1 and PM1 and 1.5 m between PM1 and PS2.The pig icons show how each pig was placed in the graves.(C) Schematic of the human donor graves at the AFTER facility.The line and arrow across each grave denote the survey line which was 16 m long over each grave.HM1 is 5 m × 2 m × 1.4 m and HS1 is 2 m × 0.3 m × 0.3 m.In the single grave (HS1), the circle denotes where the donor's head was placed in the single grave.In the mass grave (HM1), the dotted rectangle denotes where the donors were placed.According to Blau et al. (53), the individuals in this grave were placed in a haphazard mound and grouped at the NW end of the grave).Schematics not to scale.
Committee, the pigs were euthanized by a captive head bolt and delivered to the site the same day (the research was conducted in compliance with the relevant guidelines for animal experiments).The graves were dug parallel to each other with approximately 2.75 m between PS1 and PM1, and 1.5 m between PM1 and PS2.The pigs were placed on their side within each grave.See Figure 1B for a schematic of the pig graves at the animal site.The three pig cadaver graves were surveyed 20 months post-burial.
The human graves (schematic in Figure 1C), which are all part of existing research projects at AFTER, contain human donors from the University of Technology Sydney's body donation program (approved by the UTS Human Research Ethics Committee Program Approval, HREC ETH15-0029 and HREC ETH18-2999-all procedures were performed in compliance with the relevant laws and institutional guidelines; consent was previously obtained from the human donors when consent was provided to bequeath their remains).The single human donor grave (HS1) was surveyed 78 months post-burial and the mass human donor grave (HM1) was surveyed 56 months post-burial.
The soil profile at the animal site is lacking in both pedogenic or sedimentary structures, with no differences in texture with depth.
Grain size, loss on ignition, and magnetic susceptibility analyses demonstrated that the soil consists mostly of fine and very finegrained sand, with small amounts of clay and silt, and little organic material.Refer to Table 2 for the climate conditions during the survey period (Apr 2021-Feb 2023), which includes the total rainfall for each month, the average monthly temperature (calculated using the daily 3 pm temperature), and the average shallow (0-1 m) and deep (1-6 m) soil moisture levels.Figure 2B shows the animal site on the day of the 20-month GPR and ERT surveys, highlighting a lack of vegetation directly above the graves, but sparse vegetation around them.At the human site, no soil samples were taken, as they were already interred when the research project began, however, other research projects on the same site have published that the soil is classified as a sandy clay loam or gravelly sandy clay [51][52][53] and is composed of mostly sand (38%-76%), silt (22%-48%), and clay (1%-14%) [54].Although no photographs were taken of the vegetation at the human site, Figure 2A is an analog for the conditions at both geophysical survey sites.

| RE SULTS AND D ISCUSS I ON
The results of the GPR and ERT surveys of PS1, PM1, and PS2 are shown in Figure 3 and Table 3.Although both Dipole-Dipole (B) and Wenner (C) arrays have small, higher resistivity anomalies within the grave when compared to background values, the grave would not be considered highly observable.
The results of the GPR and ERT surveys of HS1 are shown in Figure 4 and Table 3.Although the grave is observable by a small hyperbola in the GPR (A) data, and a low resistivity anomaly in the Dipole-Dipole (B) and Wenner (C) data, the respective anomalies are too small and likely would not be selected as an area of interest in a forensic search.
The results of the GPR and ERT surveys of HM1 are shown in Figure 5 and Table 3.The grave is successfully observable by multiple reflector disturbances in the GPR data (A).Based on the position of the human donors (see Table 1), the two small hyperbolae near the left grave wall could be the electromagnetic waves reflecting directly off the human donors.The grave is also observable in the Dipole-Dipole (B) and Wenner (C) data as a low resistivity anomaly encompassing the grave.In both ERT arrays, the grave wall of the deepest part of the grave (left grave wall at 6.75 m along the survey line) is indicated by a small high resistivity anomaly.Although there is another small high resistivity anomaly at the other end of the grave (right grave wall at 11.75 m along the survey line), it is likely to highlight a natural disturbance.This is because this grave was not dug as a rectangle, but rather a triangle with the deepest part at the West end (left grave wall) and gradually becoming more shallow towards the East end (right grave wall).This is visually depicted by Blau et al.
[53] with a photograph of the donors in the grave (figure 7 of Blau et al. [53]) as well as a diagram of the graves vertical profile (figure 4 in Blau et al. [53]).
Based on the results of this study, both human donor and pig cadaver burials were observable and showed a similar response with both GPR and ERT.This is an encouraging finding because it means that experimental studies that explore the detectability of covert graves can be conducted without being impeded by not having access to human donors.Being able to replicate experimental studies is important because the literature strongly suggests that the success of a geophysical survey is contingent on the soil and climate conditions [58], meaning that they can successfully locate graves in some conditions, but not in others.This means that law enforcement agencies can have a better idea of when/where geophysical techniques will be most useful, as they can have area-specific geophysical data.
In the geophysical responses of the human burials, the mass grave (HM1) was the most observable.This indicates that the size TA B L E 3 Summary of observations of the GPR and ERT responses of the pig cadaver and human donor burials. of the grave (and thus its volume) is a large determining factor in its resulting observability.Additionally, the grave was more clearly observable in the GPR data than in the ERT data, highlighting the importance of utilizing a multi-technique survey design.The GPR profile demonstrates that the electromagnetic waves were reflecting off both the grave contents (small hyperbolae at the left grave wall-this is where the donors were placed in the grave), as well as the soil disturbances associated with creating the grave (reflector disturbances at the right grave wall-no human donors were placed here, only backfill).As the grave was surveyed 56 months post-burial and the human donors were likely skeletonized, or at least mummified, the fact that the GPR was able to detect the grave based on skeletal remains under field conditions is rather rare.There is only one other published instance where GPR was successfully able to detect skeletal remains [59].

Description of observations PS1 (1 in
As for the single grave, although HS1 was observable with GPR, the grave-related anomalies are small and not unique, as the rest of the profile is highly disturbed with reflector disturbances and hyperbolae.The ERT data did show that the grave was encompassed by a low resistivity anomaly, potentially picking up on the disturbed soil of the grave, however, similar low resistivity anomalies are seen in the surrounding soil.This issue is exacerbated by the shallow depth (0.3 m), as the resistivity profiles are particularly heterogeneous at this depth.Despite observing the grave, the lack of uniqueness in the GPR and ERT surveys compared to the rest of the profile, and the small size could mean that they may not be selected for further investigation if this were a forensic investigation.
Like the mass human grave, the mass pig grave (PM1) is the most observable, again due to the increased dimensions and volume.
Additionally, the GPR response to PM1 is similar to that of HM1, as the electromagnetic waves look to be reflecting off the intact chest cavities of the pig cadavers throughout the grave.As the pigs were only in the ground for 20 months, they are (likely) still in active decay with tissue attached.The single graves are both observable in the GPR data, however, only PS1 is observable in the Wenner data.This finding further supports the importance of using multiple geophysical techniques, as well as using multiple electrode arrays.The lack of observability of PS2 could indicate that an increased depth does not increase observability.This is further supported by Hansen et al. [60] who stated that deeper, cemetery-type burials were more difficult to detect with GPR and resistivity methods.
Another interesting finding was that although most of the human donors were buried with clothes and the pigs were buried naked, this is not observable in the geophysical data.More specifically, there is no difference in the GPR and ERT responses of the clothed and unclothed donor/cadaver graves.This is a unique finding because other published research, such as Molina et al. [61], has shown that burials with clothed victims are imaged better than those who are buried naked because the clothing acts as a target surface for the electromagnetic waves.
To further assess the changes in the geophysical responses in older burials, the authors conducted an additional survey of HM1 13 months later (69 months post-burial).The GPR responses show similar reflector disturbances between the 56-(Figure 5) and 69month (Figure 6) post-burial surveys, including the two small hyperbolae by the left grave wall.In the Dipole-Dipole data, the low resistivity anomaly encompassing the grave and the small high resistivity anomaly indicating the left grave wall are similar between the two surveys.In the Wenner data, the low resistivity anomaly indicating the grave is similar between the two surveys, however, the anomalies at both grave walls are different (larger) in the later survey.The similarities between the geophysical responses may indicate that the time since burial, specifically with older forensic burials (6+ years as per the graves in this project), does not affect its detectability, which is important for cases where the individual has been deceased and buried for at least 6 years.encountered in standard Australian field conditions, especially the field conditions at the AFTER sites, because of the reduced moisture (and thus higher contact resistance).Despite the higher RMS values, the ERT profiles that were generated are still valuable because they correctly identified anomalies that coincide with the known location of the human and pig burials.
It is important to note that the anomalies observable in the geophysical data can be both relevant and irrelevant to the investigation at hand.For example, although the observable anomalies in Figures 3-6 were because of intentionally dug graves (and thus intentional and relevant disturbances), natural disturbances (e.g., from animal burrows) and unnatural disturbances (e.g., buried utility pipe systems) can also be detected as anomalies.In Figure 5A specifically, there is an anomaly between 12 and 14 m that may be selected for excavation, however, considering there are no graves there, it is likely from a natural disturbance.As such, the only way to confirm the location of a covert grave is with a physical excavation.

| CON CLUS ION
In conclusion, the similarities in the GPR and ERT responses of the pig cadaver and human donor graves demonstrate that, under field conditions, pigs may be an adequate human proxy in the geophysical detection of clandestine graves.Other important findings include that the size/volume of the grave is a factor in its resulting observability (increased volume may increase observability), that GPR was successfully able to detect a grave with skeletal (or mummified) remains, and that both GPR and ERT are useful techniques to detect both more recent (less than 2 years) and older burials (more than 6 years).Finally, the inconsistency in the observability of the graves with GPR and ERT highlights the benefit of using multiple geophysical techniques and multiple electrode arrays.
Although the results of this study were positive, there were limitations that need to be addressed.First, the time elapsed from burial to the geophysical survey is different for the pig cadavers and human donors, which means that any direct comparisons need to take that into account.Second, the survey design consisted of only a single GPR and ERT line, which is not conventional, especially when the location of the grave is not known (i.e., in forensic searches).The pig graves were created as part of a larger project that only required one survey line and to facilitate a meaningful comparison, the human burials were also only surveyed using a single line.Despite this, however, the presented method still successfully located the graves with GPR and/or ERT.Therefore, if similar studies are conducted in other soil and climate conditions, the pig cadavers and human donors should be buried contemporaneously, and the geophysical surveys should be conducted in a grid.
In other experimental studies that explore the use of geophysical techniques for clandestine grave detection, pigs are more commonly used; especially in the United Kingdom, where the use of human donors is prohibited [47].However, by successfully demonstrating that pigs are acceptable human proxies for this type of research, it means that the results can be directly used to further inform law enforcement as to which soil and climate conditions are the most beneficial when conducting a GPR and/or ERT survey.This provides them with a viable forensic search technique for missing persons investigations that involve a clandestine grave.It also means that future experimental studies can be done in other soil and climate conditions without needing access to human donors, providing law enforcement with soil-/ climate-specific intelligence on the effectiveness of using geophysical techniques for forensic active and cold case missing person searches.

A 16 m 20 A
measuring tape was laid over the pig graves, with the resulting GPR and ERT surveys being 15.6 m and 15.75 m long, respectively.To facilitate a meaningful comparison, a 16 m line was also placed over each of the human graves, with the resulting GPR survey of the single and mass graves being 17.2 m and 16.3 m long, respectively, and the resulting ERT surveys being 15.75 m.In each survey, except for HM1, the survey line was placed across the midline of the grave of the shorter axis (Figure 1A, B).For HM1, the survey line was placed across the midline of the longer axis (Figure 1B).TA B L E 1 Information for the graves included in this research project.The information on the human burials was obtained from Blau et al. [53].× 1.4 "Placed in a mounded haphazard manner and grouped at the deepest end of the grave" (Blau et al, pg.326 [53]) MALA X3M GPR system (MALA GPR Australia, Brookvale, NSW, Australia) with a 500 MHz central frequency antenna was used to survey the burials, with specific survey parameters including a 17,659-sampling frequency, 4 stacks, 56.3 ns time window, 1024 samples, and a 2 cm trace increment.The raw profiles were imported into ReflexW (Sandmeier geophysical research, Karlsruhe, Germany), TA B L E 2 Climate data sourced from the Bureau of Meteorology.

F
I G U R E 2 (A) Photograph of the site prior to the graves being created in March 2021.Due to confidentiality, photographs of the human site are not published, however, this is also typical of the type and height of vegetation at the AFTER (human) facility.(B) Site photograph of the animal site at the AFTER facility when GPR and ERT geophysical surveys were undertaken 20 months post-burial.There is a lack of vegetation directly above the grave, however, there is vegetation at the south end of the site, between PM1 and PS2. and then underwent a series of sequential processing steps, including, (1) subtracting the mean from traces via a "de-wow," (2) applying an energy decay gain, (3) applying a bandpass Butterworth filter, (4) applying a 2D running average filter, and (5) moving the start time via a static correction.For the profiles of the human burials, a sixth step was used which was a whole line background removal filter.The final 2D profiles were then imported into InkScape software v.1.1.2(b8e25be833, 2022-02-05; Free Software Foundation, Inc. Boston, MA, USA) and compiled for a visual comparison.A ZZ FlashRes64 system (v.19.2; ZZ Resistivity Imaging Pty. Ltd., Adelaide, South Australia, Australia) was used to collect the ERT data, with specific survey parameters including 64 electrodes, Wenner (K = 20) and Dipole-Dipole (K = 15, L = 5) arrays, 120 voltage, and a 1.2/0.2on/off time.An electrode spacing of 0.25 m was chosen to maximize depth penetration and the resulting resolution.The raw data was extracted using ZZ-RdatacheckU96 software (v.4.0; ZZ Resistivity Imaging Pty. Ltd., delaide, South Australia, Australia), and then transformed and filtered in R Software using a custom R script [55, 56].Each profile was then individually processed (exterminating bad data points) and inverted using an L2 robust inversion (following a least-squares inversion approach [57]) in Res2DInv (Seequent, The Bentley Subsurface Company).The final 2D ERT profiles were then imported into InkScape software v. 1.1.2(b8e25be833, 2022-02-05; Free Software Foundation, Inc., Boston, MA, USA) and compiled for a visual comparison.
Firstly, PS1 is successfully observable in the GPR (A-reflector disturbances) and Wenner (C-low F I G U R E 3 GPR (A) and ERT (B, Dipole-Dipole; C, Wenner) responses to three pig burials. 1, PS1; single pig burial, with one pig cadaver that is 2 m × 1 m × 0.5 m. 2, PM1; mass pig burial, with three pig cadavers that are 3 m × 1 m × 1 m. 3, PS2; single pig burial, with one pig cadaver that is 2 m × 1 m × 1.8 m.PS1 (1) is observable in the GPR data by reflector disturbances, and by a low resistivity anomaly in the Wenner data, however, it is not observable in the Dipole-Dipole data.PM1 (2) is observable in the GPR data by reflector disturbances and by a low resistivity anomaly in the Dipole-Dipole and Wenner data.PS2 (3) is observable in the GPR data by reflector disturbances and is not highly observable in the Dipole-Dipole and Wenner data.resistivity anomaly not seen in the immediately surrounding soil) data.As for the Dipole-Dipole (B) data, there is a small low resistivity anomaly on the right grave wall when compared to the background values, however, it is not a highly observable feature.Next, PM1 is successfully observable in the GPR (A-reflector disturbances), Dipole-Dipole (B-low resistivity anomaly throughout grave with lower resistivity anomaly at lower left base of grave), and Wenner (C-low resistivity anomaly not seen in the immediately surrounding soil) data.Finally, PS2 is successfully observable in the GPR (A-stratigraphic breaks near the surface) data.
Small hyperbola; not highly observable Reflector disturbances within grave with two small hyperbolae indicating the human donors near the left grave wall ERT: Dipole-Dipole (B) No clear grave anomaly Low resistivity anomaly throughout grave High resistivity anomaly close to surface; not highly observable Low resistivity anomaly throughout grave; not highly observable Low resistivity anomaly within grave; left grave wall indicated by small high resistivity anomaly ERT: Wenner (C) Low resistivity anomaly throughout the grave Low resistivity anomaly throughout the grave High resistivity anomaly close to surface; not highly observable Low resistivity anomaly throughout grave; not highly observable Low resistivity anomaly throughout grave; left grave wall indicated by small high resistivity anomaly

F I G U R E 4
GPR (A) and ERT (B, Dipole-Dipole; C, Wenner) responses of HS1 78 months post-burial.The grave contained one individual and measured 2 m × 0.3 m × 0.3 m.The grave, denoted by a green rectangle, is observable by a small hyperbola (white arrow) in the GPR data, and a low resistivity anomaly throughout the grave in the Dipole-Dipole and Wenner data.Due to the small size of this anomaly, it is not considered highly observable.

F I G U R E 5
GPR (A) and ERT (B, Dipole-Dipole; C, Wenner) responses of HM1 56 months post-burial.The grave contained six individuals and measured 5 m × 2 m × 1.4 m.The grave, denoted by a green rectangle, is observable by reflector disturbances within the grave in the GPR data, and a low resistivity anomaly in the Dipole-Dipole and Wenner data.

F I G U R E 6
GPR (A) and ERT (B, Dipole-Dipole; C, Wenner) responses to HM1 69 months post-burial.The green rectangle denotes the mass human grave, with six human donors, that measures 5 m × 2 m × 1.4 m.The GPR and ERT responses are similar to those in Figure 5, despite the added 13 months in the ground.

Month/year Total rainfall (mm) Average temperature (°C)* Average soil moisture (0-1 m, percentile rank) Average soil moisture (1-6 m, percentile rank)
*Average temperatures were calculated using the daily temperatures taken at 3 pm.