Soil science expertise was used to help solve a double murder by identifying the similarities between soil and clay assemblages on a shovel and from a quarry. The soil material and clay mineralogy had a common provenance and revealed the location of two buried bodies. This successful case led to the formation of the Centre for Australian Forensic Soil Science (CAFSS) in 2003. CAFSS was formed in recognition that soil expertise is not part of the forensic scientist’s repertoire. Since then, CAFSS has advised on more than 100 criminal and environmental forensic investigations (Fitzpatrick, 2008, 2009, 2012a,b; Fitzpatrick et al., 2007, 2009, 2012). However, the shift from traditional soil science and pedology to forensic soil science is not straightforward and requires a wide understanding of crime scene protocols, the evidential requirements of forensic workers, and the nature of legal constraints within which forensic work takes place.
With a focus on the author’s first and most public case, the double murder, this paper has the following three principal objectives:
To summarize established concepts and standard terminologies used in forensic soil science, but especially pedology and mineralogy that have practical relevance to forensic science;
To illustrate the use of pedological and mineralogical methods (e.g., XRD analyses) in forensic science by detailing the comparison of a soil adhering to a shovel found in the trunk (boot) of the suspect’s vehicle with a soil of human origin that, in this case, originated in a quarry; and
To show the educational value of forensic soil science through the development of an interactive exhibit, which succinctly explains how pedologists and mineralogists successfully used soil morphology, soil survey, soil chemistry, and clay mineralogy expertise to help solve this double murder case.
What Is Forensic Soil Science?
Soil forensics is the science or study of soil that uses of a wide range of soil information to answer legal questions, problems, or hypotheses (Fitzpatrick, 2009). The transfer of soil trace evidence is governed by what has become known as the Locard Exchange Principle (Chisum and Turvey, 2000), which states that when two surfaces come into physical contact there is a mutual exchange of trace evidence between them. Natural and human-made soil materials are being recognized and used in forensic investigations to associate a soil sample taken from an item, such as a victim’s clothing (questioned soil), with a soil from a specific known location such as the crime scene (control soil) (Fig. 1). Very often, trace soil evidence can link an object or suspect to the scene of a crime, as well as rule out a suspect or support an alibi and has been used as evidence in courts of law (e.g., Fitzpatrick, 2008, 2009, 2012a,b; Fitzpatrick et al., 2007, 2009, 2012; Murray, 2011; Murray and Tedrow, 1992; Pye, 2007; Ritz et al., 2008; Ruffell and McKinley, 2008). Consequently, it is important to understand and know the different kinds of natural and human-made soils and how they form. It is also especially important to know how to carefully sample and analyze soils so that robust forensic comparisons can be made. A wide diversity of natural soils exists, and each has its own characteristics (e.g., morphology, mineralogy and organic matter composition). For example, according to the United States Department of Agriculture, which collects soil data at many different scales, there were more than 21,000 soil series identified in the United States alone (Guo et al., 2003)! Soil properties such as parent material, climate, soil organisms, and the amount of time it takes for interaction among soil properties to occur will vary worldwide. Human-made soils, called Anthroposols in the Australian Classification (Isbell, 1996), Technosols in the World Reference Base (IUSS Working Group WRB, 2007), and recently proposed as human-altered and human-transported (HAHT) soils by Galbraith (2012) in U.S. Soil Taxonomy (Soil Survey Staff, 1999), are characterized by diversity, heterogeneity, and complexity, which enables forensic soil examiners to distinguish between soils.
First and foremost, soil evidence must be recognized on questioned items and subsequently at known proposed crime scenes and alibi localities (Fig. 1a). Second, evidence must be well documented. Finally, the meticulous collection and preservation of soil samples must be maintained so as to ensure the integrity of the soil evidence (Fig. 1a)—followed by soil characterization (e.g., Fitzpatrick and Raven, 2012). Soil forensic investigations may involve the collection of one or more specific types of soil forensic samples, which are categorized as: (i) questioned soil samples whose origin is unknown or disputed (often from suspect or victim), (ii) control samples whose origin is known and usually from specific sites directly related to the investigation such as the known or proposed crime scene, and (iii) alibi samples whose origin is known and that provide a measure of the uniqueness of the questioned and control samples (Fig. 1b) (e.g., Fitzpatrick, 2009; Murray, 2011; Ruffell and McKinley, 2008).
The Value of Soil in Criminal Investigations
The aim of forensic soil analysis is to associate a questioned soil sample taken from an item, such as a shovel, vehicle, shoes, or clothing, with a control soil sampled from a specific known location (Fig. 1b, Fig. 2). Soil materials were identified by Fitzpatrick (2009) as powerful, perhaps ideal, pieces of contact trace evidence for the following six reasons:
Soil is highly individualistic in that there are an almost infinite number of different soil types, and soils may change rapidly over very short distances both horizontally and vertically, enabling forensic examiners to distinguish between soil samples.
Soil materials are easily described and characterized by color and by using various analytical methods such as XRD (mineralogy) and spectroscopy (chemistry).
Soil has a strong capacity to transfer and stick, especially the fine clay- and silt-size fractions.
Unlike the more obvious bright transfer colors of blood, lipstick smears, and paint, soil is nearly invisible. Fine soil materials, especially when they impregnate vehicle carpeting, shoes or clothing, are often not visible to the naked eye; a suspect will often make little effort to remove them.
Soil materials are easily located and collected using hand lenses or light microscopes when inspecting crime scenes or examining items of physical evidence.
National and international computerized databases of soil profile data and maps can be readily accessed by police or soil scientists through the Internet. for example, the Australian Soil Resources Information System (ASRIS), where a soil map can be produced by downloading information directly from the internet (www.asris.gov.au; Johnston et al., 2003).
Methods for Forensic Soil Science
All sampling methods, sample storage security, soil characterization, and analytical methods must be subject to the most rigorous procedures of sample and data handling available so they can be discoverable and questioned in court. The forensic investigation of soil usually involves the following two main activities:
Soil collection and sampling of one or more locations
Soil characterization and evaluation
CAFSS staff members are regularly subpoenaed to testify in court. To meet these responsibilities CAFSS has developed a publication entitled “Guidelines for Conducting Criminal and Environmental Soil Forensic Investigations” (Fitzpatrick and Raven, 2012). The guidelines provide a systematic approach and use of appropriate standard methods for sampling, characterizing, and examining soils for forensic comparisons. They also assist CAFSS in its mission by ensuring efficiency and accountability in the proper handling, storage, and tracking of soil evidence, which is essential to evidence collection and ultimate prosecution. Several case studies, which have successfully used soil evidence to help solve legal cases, are likely attributed to the systematic methodology of crime scene sampling, processing, and soil characterization and interpretation (Fitzpatrick and Raven, 2012).
Soil Collection and Sampling
Knowing how many questioned, control, or alibi samples to collect is difficult. The ideal number, size, and type of samples to be taken are strongly dependent on the nature of the environment being investigated, especially the type of soil (e.g., wet or dry soil) and nature of activity that may have taken place at the sampling location (e.g., suspected transfer of soil from the soil surface only or from a depth in the case of a buried object or body—or both).
Soil Characterization and Evaluation
Soil characterization requires a multidisciplinary approach, which combines descriptive, analytical, and spatial information (e.g., mapping). Methodology for characterizing soils for a forensic comparison broadly involves subdividing methods into four stages, each comprising several steps and involving a combination of techniques:
Stage 1: Initial Characterization for Screening of Samples
This stage involves morphological characterization of bulk soil samples. Soil morphological interpretation provides a visual, quick, and nondestructive approach to screen and discriminate among the various types of samples (Table 1). Morphological soil descriptors are arguably the most common and simple; it is for this reason that all bulk samples are characterized first using morphological descriptors using standard Australian (McDonald and Isbell, 2009) and international (Schoeneberger et al., 2002) soil morphological methods. The eight main soil morphological descriptors of: (i) matrix color (moist and dry using Munsell Soil Color Charts, (Munsell, 2000), (ii) mottles (retained from geologic sources), (iii) redoximorphic features (providing an indicator of drainage or redox status because soil color relates to soil aeration of weakly reducing conditions; e.g., Bigham et al., 2002), (iv) concentrations (nonredox; e.g., carbonates, nodules or inherited brick fragments), (v) texture (e.g., sands, loams or clays), (vi) structure (e.g., massive or platy), (vii) effervescence class (reaction to 6N HCl, which indicates the presence of carbonates), and (vii) water repellence class are the most useful properties for visual soil characterization and assessing soil conditions (e.g., Fitzpatrick et al., 1999, 2003). Other useful soil morphological descriptors are quartz grain shape (if sandy) and rock or other fragments (if easily observable on questioned items, such as on a shovel).
|Type of forensic soil sample Collected or referenced||Stages for sequential examination of questioned (Q), control referenced (R), control collected (C), and alibi collected (A) samples†|
|Questioned samples (Q)||Stage 1Q. Initial characterization for screening of sample|
|Collected samples of unknown or disputed origin (e.g., shovel, shoes, clothing, etc.)||Stage 2Q. Semi-detailed characterization|
|Stage 3Q. Detailed characterization|
|Stage 4Q. Comparisons between various collected questioned samples and construction of generalized soil models (e.g. using soil classification, small-scale soil maps, landscape, etc.)|
|Control samples (R)||Stage 4 R. Comparisons between questioned samples and known reference control samples/soil maps and construction of generalized soil-landscape models (e.g. using soil classification, small-scale soil maps, etc.)|
|Referenced samples from known maps, collections, archives, museums|
|Control samples (C)||Stages 1C to 3C: As for Stages 1Q to 3Q.|
|Collected samples from a known and undisputed origin (e.g., proposed crime scene)||Stage 4C. Comparisons between questioned samples and collected known control samples and construction of detailed soil–landscape models (e.g. using soil classification, large-scale soil maps, etc.)|
|Alibi samples (A)||Stage 1A to Stage 3: As for Stages 1Q to 3Q.|
|Collected samples from a known locality to provide a measure of the uniqueness of the questioned and control samples||Stage 4A. Comparisons between questioned samples and collected known alibi samples and construction of more refined or detailed soil-landscape models to illustrate uniqueness of the Q and C samples (e.g. using soil classification, large-scale soil maps, etc.)|
Stage 2: Semi-detailed Characterization
This stage involves identification, semi-detailed characterization, and semi-quantification of minerals and organic matter in bulk samples and on individual soil particles, often following sample selection and size fractionation (e.g., <50 μm or <0.4 mm). X-ray diffraction methods are arguably the most significant for both qualitative and quantitative analyses of solid materials in forensic soil science (Fitzpatrick, 2009; Kugler, 2003). Extremely small sample quantities (e.g., few to a few tens of milligrams) as well as large quantities can be successfully analyzed using XRD. The critical advantage of XRD methods in forensic soil science is based on the unique character of the diffraction patterns of crystalline and even poorly crystalline soil minerals. Elements and their oxides, polymorphic forms, and mixed crystals can be distinguished by nondestructive examinations. Part of the comparison involves identification of as many of the crystalline components as possible, either by reference to the International Centre for Diffraction Data (ICDD) Powder Diffraction File (Kugler, 2003) or to a local collection of standard reference diffraction patterns (e.g., Rendle, 2004), coupled with expert interpretation.
Stage 3: Detailed Characterization
This stage involves detailed characterization and quantification of minerals and organic matter in bulk samples and on individual soil particles using additional analytical techniques and/or methods of sample preparation, separation, or concentration (e.g., size or magnetic or heavy mineral fractionation).
Stage 4: Integration and Extrapolation of Soil Information from One Scale to the Next
This stage involves building coherent soil models of soil–landscape information from microscopic observations to the landscape scal e, which may involve soil classification and use of soil, geological and vegetation maps, terrain analysis, remote sensing, and geophysics. Classifying soils for a particular purpose involves the ordering of soils into groups with similar properties and for potential end uses such as for police (e.g., special purpose soil classification systems; Fitzpatrick, 2012a,b). In general, soil classification systems currently used in most countries involve the use of three “General-purpose broad soil classifications” (Fitzpatrick, 2004), which communicate soil information at international [U.S. Soil Taxonomy (Soil Survey Staff 1999) and IUSS Working Group WRB (2007)] and national scales [e.g., Australian (Isbell, 1996)]. Soil classification is extremely useful in the development of a labeling system for soil landscape mapping units. For example, soil types have been broadly mapped for the entire Australian continent and have been published as soil maps (Northcote et al., 1960–1968; Australian Soil Resources Information System). The construction of a conceptual or coherent soil model for questioned samples uses integrated and extrapolated soil information (Table 1) to ensure better-informed location, sampling, and characterization of control samples (e.g., from the crime scene) based on how “similar” the soil is to the questioned sample (e.g., from the suspect’s shoe).
Decision making in forensic soil science is sometimes guided by the development of successive mechanistic process models describing processes with physical, chemical or biological mechanisms at a range of scales (Table 1; Murray and Tedrow, 1992; Ruffell and McKinley, 2008; Fitzpatrick et al., 2009). Some models use multiple data layers as spatial input. The progression of a soil examination through each of the four stages will depend on a number of factors, such as the purpose of the investigation, amount of sample available, and the results from the early stages of examination. Not all stages may be required for all investigations. However, in some investigations it may be necessary to repeat all four stages during the course of a soil investigation to examine sequentially various questioned, control referenced (maps or archived), control collected, or alibi samples (Table 1).
When carrying out soil comparisons it is important to first define the word “comparable” because no two physical objects can ever, in a theoretical sense, be the same (Murray and Tedrow, 1992). Similarly, a sample of soil or any other earth material cannot be said, in the absolute sense, to have come from the same single place. However, according to Murray and Tedrow (1992), it is possible to establish “with a high degree of probability that a sample was or was not derived from a given place.” For example, a portion of the soil (or other earth material) could have been removed to another location during human activity. CAFSS has developed a terminology scheme, which uses “Categories of Comparability” with defined “Examples of Type of Evidence” for soil or geological evidence interpretation (Table 2). The scheme has been specifically developed with no statistical significance of the ranks implied and is used as guidance for both the interpretation of results and to meet admissibility requirements in courts.
|Categories of comparability||Examples of type of evidence|
|None||Different in virtually all aspects|
|Limited||Some general comparison in terms of soil morphology (color, texture, and/or relatively common particle types present)|
|Moderate||General comparison in terms of soil morphology, especially in having a similar assemblage of relatively common particle types in common, some of which may have distinctive textural or chemical features|
|Moderately strong||Fairly high degree of comparison in terms of soil morphology as well as chemical, mineralogical, and/or biological properties; including relatively unusual particle types in common|
|Strong to||High degree of comparison in terms of soil morphology as well as chemical, mineralogical, and/or biological properties; including several relatively unusual particle types present|
|Extremely strong to||Physical fit (rocks) and very high degree of comparison in terms of soil morphology as well as chemical, mineralogical, and/or biological properties; including one or more very unusual particle types present.|
When comparing soil forensic samples a professional judgment should be made to establish the “comparability category” that the soil materials originate from a similar locality. Consequently a judgment should be made as to whether a questioned sample (e.g., from a shovel or shoe) and soils at, for example, the crime scene site (control) “are comparable” or “are not comparable.” If the samples are comparable, a further judgment must be made as to whether the samples have limited, moderate, moderately strong, strong, very strong, extremely strong, or conclusive “degree of comparability” of being from a single location.
Example Case Study: Double Murder Case
Between 1905 and 1990, soil information was used extensively by police. However, by the 1990s, soil analysis was becoming too specialized and expensive for in-house use in most forensic laboratories worldwide. As a consequence, critical soil forensic evidence was often missed or ignored completely, hidden among trace evidence, and insufficiently analyzed. Consequently, problems in recognizing the potential of forensic soil science can be overcome by sharing successful case examples to make soil scientists and crime scene personnel aware of the value of soil evidence. For example, Dr. Raymond Murray in his book entitled Evidence from the Earth (Murray, 2011) presented several high-profile legal cases in which geological materials and methods contributed significantly to solving cases from around the world. To draw attention to this somewhat underutilized forensic tool, some important ways by which soil evidence was proven useful in helping to solve a double murder case in 2000 are presented in the following case study.
On the 18th of September 2000 neighbors reported a disturbance at the home of two women living in the Adelaide suburb of Oakbank within the Adelaide Hills, South Australia (Fig. 2) to police. That same evening, the husband of one of the women called the police after returning from work to his home to find that both his wife and mother-in-law were missing (Porter, 2007; Barrett, 2003). He found blood stains and broken glass on the lounge room floor, while towels were missing, along with the quilt and pillow from his son’s bedroom. His wife’s car and his 22-yr-old son Matthew Holding were also missing.
The following day, the wife’s car was found more than 200 km away by police near Moonta on Yorke Peninsula (Fig. 2). Inside the trunk, crime scene investigators found a pine post and a faintly bloodstained shovel, both sides of which were caked with pinkish powdery soil (Fig. 3). In the vehicle, investigators also located a green jade bracelet and boots, which were thinly coated with fine pinkish soil (Fig. 3), as well as a blood-stained knife, blood-stained towels, bedding, and a pile of sticks, the kind suitable for starting a fire. Shortly afterward, police arrested 22-yr-old Matthew Holding in Moonta (Fig. 2), where he had been attempting to get help for his broken-down car. Later in the day he was charged with the murder of his mother and grandmother. The forensic and circumstantial evidence against Matthew Holding was compelling, but he was telling police nothing.
Collection and Sampling of Questioned Samples
The nature of the soil on the shovel suggested that the suspect might have buried his victims. But where? His vehicle was located near Moonta (Fig. 2), and a reported sighting of his car had led the police to a search of the Moonta cemetery. Consequently, the police spent 3 days searching various localities on Yorke Peninsula (Fig. 2), in the vicinity where they had found and arrested him. Time was passing, and the police needed scientific reasons for focusing their search in one place and ruling out others. According to Porter (2007), the police needed to find his victims’ bodies so the Crown pathologist could confirm their cause of death—“for example, by matching their wounds to the blood-stained knife and corroborate any admissions from the accused.”
Four days had passed since the father had reported the two women missing; it was possible they were alive and injured, but Matthew Holding continued to refuse to answer questions (Barrett, 2003). The location and fate of the missing women remained unknown. Needing to consult soil specialists, crime scene investigators called in soil scientists from the Commonwealth Scientific and Industrial Research Organization (CSIRO) on Thursday, 21 September to examine and characterize the deposited layers of soil adhering to the shovel, jade bracelet, and boots (Fig. 3) and to commence the systematic and sequential soil forensic approach outlined in Table 1.
Morphological Characterization of Bulk Questioned Samples: Stage 1Q
The bulk questioned soil samples on the shovel (back and front), bracelet, and boots were examined, photographed (Fig. 3 and 4), and characterized. From visual observations using Munsell color notation, it was established that the fine coatings (e.g., <50 μm) on the bracelet and boots had a similar matrix pink color to the soil adhered to the shovel (Munsell, 2000). These similar morphological properties indicated they were likely to have originated from the same source or locality. Consequently, because the shovel (back and front) contained larger amounts of caked soil with coarse particles (>50 μm) compared to the bracelet and boots, soil samples adhering strongly to both the back and front of the shovel were recovered with the aid of a stainless-steel spatula for further characterization. Samples were visually described using both hand-held samples and through a Wild Leitz M420 stereo microscope using Schott LED light sources and photographed in a glass petri dish (Fig. 4). In this paper, the soil morphological descriptors outlined in the footnote to Table 3 were used and are in accordance with standard methods (e.g., Soil Survey Staff, 1993; Schoeneberger et al., 2002; USDA, 1996). The following morphological properties (Table 3) indicate that the samples from the back and front of the shovel samples have similarities. The pink matrix color of coatings on surfaces of the grains indicates likely origin from subsoil/saprolite materials. Combinations of concentrations and depletions in varying degrees of abundance within weakly cemented massive lumps of soil samples removed from the shovel revealed morphology similar to that of a poorly drained soil.
|Sample type and (number)||Matrix color||Redoximorphic features†||Texture‡||Gravel >2 mm||Quartz§||Structure||Conc.¶ clay fragments||Efferv. Class#||Roots||WR††||pH||EC|
|Back of shovel Questioned soil sample(UJU 12.2)||7.5 YR 8/4 Pink||7.5 YR 7/4 Pink||f1d, 7.5 YR 6/8 Reddish yellow||f1d, 10YR 7/2 Light gray||CS K||30||A, CF||Massive parts to Single grain||c2p||NE||None||N||4.7||0.15|
|Front of shovel Questioned soil sample(UJU 12.3)||7.5 YR 8/4 Pink||7.5 YR 7/4 Pink||f1d, 7.5 YR 6/8 Reddish yellow||f1d, 10YR 7/2 Light gray||CS K||25||A, CF||Massive parts to Single grain||f1p||NE||None||N||4.9||0.15|
|Quarry sump Control soil sample (Q1)||7.5 YR 8/4 Pink||7.5 YR 7/4 Pink||f1p, 7.5 YR 6/8 Reddish yellow||f2p, 10YR 7/2 Light gray||CS K||55||A, CF||Massive parts to Single grain||m3p||NE||None||N||5.1||0.16|
The clayey sand (CS), coarse sandy texture (McDonald and Isbell, 2009), and the large amounts (25–30%) of quartz gravel particles with angular shapes and conchoidal surface fractures of these questioned soil materials are typically encountered in deeply weathered saprolite at depth. Similarly, the morphology of the prominent, hard irregular clay fragments (<2 mm; 2 to <5 mm) with a generally pale yellow (2.5Y 8/3) color has the likely appearance of originating from clay lenses in strongly weathered saprolite zones or regolith at depth. The massive structure, which parts when dry to a single grain structure (loose to powdery) with no observable roots or detectable carbonate, also suggests an origin from deep layers (e.g., saprolite).
Chemical and Mineralogical Characterization of Bulk Questioned Samples: Stage 2Q
The dried questioned bulk soil samples from both the back and front of the shovel were dry sieved through a 2- and 0.4-mm sieve. Air-dried <2-mm sieved samples were analyzed for pH and electrical conductivity (EC) by 1:5 soil/water suspension (Rayment and Higginson, 1992). The pH of 4.7 and 4.9 (Table 3) indicates very strongly to medium acid conditions (Soil Survey Staff, 1993). The EC of 0.16 dS/m suggests low electrolyte content with low levels of mineral salts (i.e., nonsaline environment).
The <0.4-mm fractions were thoroughly ground in an agate mortar and pestle. The resulting fine powders were lightly pressed into aluminum sample holders for XRD analysis. The XRD patterns were recorded with a Philips PW1800 microprocessor-controlled diffractometer using Co Kα radiation, variable divergence slit, and graphite monochromator.
The XRD patterns were recorded in steps of 0.05° 2θ with a 3.0-s counting time per step, and logged to data files for analysis using in-house developed XPLOT software and HighScore Plus, a commercial software (PANalytical, Almelo, The Netherlands). Mineralogical phase identification were made by comparing the measured XRD patterns with the ICDD database of standard diffraction patterns using computer aided search/match algorithms.
Mineralogy by XRD of the <0.4-mm fraction soil samples from the back of the shovel is shown in Table 4, and the diffraction pattern is displayed in Fig. 5. The mineralogical composition indicates dominant quartz, subdominant well crystalline kaolinite, minor muscovite, with traces of feldspar and talc (front of shovel only). The nature of this mineralogical composition suggests the soil materials originate from subsurface kaolinite-rich weathered zones, such as a mining area or quarry.
|Back of shovel (bulk) (<0.4-mm fraction) UJU 12.2||SD (c)||D||M||T||ND|
|Back of shovel (small white fragments)||D (wc)||T||T||T||ND|
|Front of shovel (bulk) (<0.4-mm fraction) (UJU 12.3)||SD (c)||D||M||T||T|
|Quarry sample (bulk) (<0.4-mm fraction) Q1||SD (c)||D||M||T||ND|
|Quarry sample (small white fragments) Q3||D (wc)||T||T||T||ND|
|Quarry sample (large white fragment) Q4||D (wc)||T||T||T||T|
Mineralogical Characterization of White Fragments in Questioned Samples: Stage 3Q
The pale yellow (2.5Y 8/3), irregular clay fragments (1–5 mm diameter) were hand-picked from both the back and front of the shovel and thoroughly ground in an agate mortar and pestle. The resulting fine powders were lightly pressed into aluminum sample holders for XRD analyses (Fig. 6). The mineralogical composition of these white fragments extracted from the back of the shovel was comprised dominantly of well crystalline kaolinite with traces of quartz, muscovite, and feldspar (Table 4).
Comparisons between Various Questioned Samples and Likely Origin of Samples: Stage 4Q
Sufficient descriptive (Table 3), mineralogical (Table 4), and chemical (Table 3) data were acquired on the questioned soil materials from the back and front of the shovel to compare properties, classify the soil materials, and attempt to build a conceptual soil model of their likely origin and mode of formation on the shovel.
Soil morphological (Table 3) and mineralogical data (Table 4, Fig. 5 and 6), indicate that the soil samples on the back and front of the shovel consist of diverse “physical mixtures” of angular quartz and irregular hard bright yellow colored fragments comprised of well crystalline kaolinite with no evidence of organic matter, suggesting that the materials were not “natural soils” and likely classify as “human-made soil materials” comprising dominantly of mixed and/or transported weathered saprolite (regolith). Consequently, based on the composition of these materials they may be classified as likely originating from the following equivalents of human-made soils from different classification systems: Spolic Technosol (Arenic, Skeletic) (IUSS Working Group; WRB, 2007), Scalpic and Cumulic Anthroposols (Isbell, 1996), and Anthroposols (i.e., human-altered and human-transported, HAHT, soil materials; Galbraith, 2012). The development of well-expressed redoximorphic features in both questioned samples (Table 3) suggests that these human-made soils comprising mainly transported/mixed regolith material have also been subjected to periodic anoxic conditions.
The remarkably clean working edges and smeared soil structure on both the back and front of the shovel shown in Fig. 4 strongly points to use in a wet soil. The cemented soil stuck to the rear sections at both the back and front of the shovel blade (Fig. 4) displayed well-expressed redoximorphic features, which likely equates to the original soil having undergone periodic conditions of wetness. The soil appearance on the front of the shovel blade showed that it had been wet and sloppy (Fig. 4). Finally, soil was compacted in the handle housing on the back of the shovel, indicating that it had been used to flatten or pat down moist soil (Fig. 4).
Comparisons between Questioned and Reference Control Samples/Soil Maps—Construction of Generalized Soil–Landscape Model: Stage 4R
The ASRIS database has compiled the best publicly available soil information available across Australian agencies into a national database of soil profile data, digital soil and land resources maps, and climate, terrain, and lithology datasets. Most datasets are thematic grids that cover the intensively used land-use zones in Australia (Johnston et al., 2003). Hence, the next stage (i.e., Stage 4R) was to compare the comprehensive data obtained from the questioned samples (Tables 3, 4, and 5) against published data recorded in available ASRIS databases for the region (Fig. 7). Although the geographic coverage of sampling points is not evenly spread across the region, it is informative to examine the distribution of soil types after producing a soil map of the region between Oakbank and Moonta based on the ASRIS database (Fig. 7). It is critical to first consult such existing/available soil maps of the region of interest (Fig. 7) in conjunction with or with help from experienced pedologists and geologists. The areas of broadly similar soil types or properties to the questioned sample can then be identified as high priority areas for further investigations and possible sampling of control samples followed by comparative analyses using morphological and analytical information.
|Samples||The World Reference Base for soil resources (IUSS Working Group WRB 2007)||Australian Soil Classification (Isbell, 1996)||U.S. Soil Taxonomy modifications (Galbraith 2012)|
|Shovel†||Spolic Technosol (Arenic, Skeletic)||Scalpic and Cumulic Anthroposols (Confidence level 3)||HAHT‡ soil material|
|Quarry||Spolic Technosol (Arenic, Skeletic)||Scalpic and Cumulic Anthroposols(Confidence level 3)||HAHT soil material|
The low pH (acid; pH 4.5), low electrolyte concentrations (low salt levels; EC of 0.15 dS/m; Table 3) and kaolinitic-rich mineralogical composition indicated that the samples were likely to originate from a deeply weathered landscape specifically consisting of Acid Chromosols (Palexeralfs, Haploxeralfs) and Kurosols (Rhodoxeralfs) in a high rainfall zone (e.g., >600 mm) as shown in the Adelaide Hills region near Oakbank (Fig. 7). This combination of soil properties and rainfall conditions specifically ruled out the Yorke Peninsula, where police had been searching initially, as soil and regolith materials from this region are predominantly alkaline (pH >7), very frequently saline (EC >2 dS/m) and comprise the following predominantly calcareous soils: Calcarosols (Calcixerepts), Alkaline/Neutral Chromosols (Aridisols, Xeralfs, Calcic Rhodoxralfs) and Sodosols (Natrixeralfs). This evidence suggested that the investigation must focus on landscapes in the Adelaide Hills, where soils and underlying weathered saprolite are acidic and have low levels of salts (Fig. 7).
The focus on the Eastern Mt Lofty Ranges (Adelaide Hills) coincided with several major ongoing soil–regolith research programs across the region, investigating land degradation caused by surface waterlogging and rising groundwater tables (Fitzpatrick et al., 1996), mineral exploration (Skwarnecki and Fitzpatrick, 2003), and aggregate, sand, and clay resources (Pain and Keeling, 1990), which provided invaluable firsthand soil–regolith data of the region. The underlying regolith geology of the region is Cambrian with generally strongly weathered metasediments of the Kanmantoo Group, which consists of interbedded, vertically dipping micaceous sandstones and schists (Fitzpatrick et al., 1996). There are also numerous sulfide-rich lenses or bands, some of which have been mined for lead and zinc.
The questioned soil sample from the front of the shovel contained the following five out of five similar mineral components as several archival weathered saprolite samples taken by CSIRO staff at depth (3.4 m) from a nearby catchment (Fitzpatrick et al., 1996): very crystalline kaolinite of industrial extraction grade, quartz, muscovite, feldspar and traces of the uncommon soil mineral “talc” (Table 4). The identification of the highly crystalline nature of these minerals combined with the observations of the: (i) “physical mixtures” of angular quartz grains and irregular hard clay fragments and (ii) absence of plant organic materials was compelling evidence to suggest the soil likely originated from subsurface kaolinite-rich weathered zones such as a mining area or quarry (i.e., quartzite rubble or gravel quarry) rather than from a natural topsoil. The Adelaide Hills area has many quarries, including current workings associated with aggregate, sand, and clay resources (Pain and Keeling, 1990) and several associated with gold mining activities in the past.
In summary, based on the characteristics of the soil on the shovel and on a knowledge of local soils (Fig. 7) and geology from ongoing projects and geological maps from South Australian Resources Information Geoserver (SARIG: https://sarig.pir.sa.gov.au/Map), CSIRO pedologists determined that the soil–regolith material on the shovel likely came from one of a small group of industrial gravel/soil quarries near Oakbank that were known to contain talc.
In the absence of obvious map features, it was decided to conduct a systematic field soil–regolith survey and sampling of a range of quarries near Oakbank (i.e., attempt to visually observe soil and regolith patterns or features in quarries, especially in possible waterlogged areas within quarries, which may visually compare (e.g., color and texture) with the soil–regolith on the shovel).
Collection and Sampling of Control Samples from Quarry
The accumulated soil–regolith information, together with other physical evidence, led police and soil scientists to undertake field investigations in several quarries near Oakbank. The father was kept informed throughout the investigation and was told of the forensic soil scientists’ observations that the soil on the shovel was likely to have originated from a mine site or quarry in the Adelaide Hills. He directed investigators to one particular quarry at Oakbank. He had taken Matthew Holding there when searching for appropriate places to dispose of old car tires. The quarry was also near the location where Matthew Holding’s car had been seen the evening the women disappeared (Barrett, 2003). There were also diffuse tire tracks with a pattern similar to Matthew Holding’s car in and around the quarry but the wheelbase and tire type were very common, and this latter evidence was discounted. As a consequence, police and search teams searched the quarry for several days but could not find any signs of a recent burial and were unable to locate the bodies. The possibility of finding any surface features was negated by the high frequency of footprints and the hoof prints of cattle that presumably had entered the quarry to drink from the pool that had accumulated at the quarry’s lowest point or sump with recent rain. An unsuccessful search was also performed using a “sniffer” dog, although it was only at an early stage of training and was probably already compromised by the heavy human traffic in the quarry.
However, the pedologists were convinced that their conceptual soil–regolith model was correct in suggesting the questioned soil likely originated from human-made soil–regolith material, which in turn was derived from deep kaolinite-rich weathered zones such as a mining area or quarry (i.e., quartzite rubble or gravel quarry) rather than from a pristine natural topsoil or subsoil B horizon. Consequently, after extensive field investigations of the quarry, CSIRO pedologists noticed that the color and texture of the materials being extracted from the quarry, especially in the wet and ponded area in the sump of the quarry, bore a strong visual resemblance to the materials found on the shovel. For this reason, two control soil samples were collected from wet areas immediately adjacent to the large pool of water in the center of the quarry (Fig. 8).
A control clay fragment sample consisting of a very large (5 by 15cm) white to yellowish clay fragment or lump (sample Q4 in Fig. 9) was sampled from the quarry face (see Fig. 10 for the indicated position of sampling in the side of the quarry face). The main purpose for this sampling was to gain detailed mineralogical information on properties of an in situ large pure clay fragment from the quarry face, which was likely to be source of the smaller clay fragments following quarry excavation and via colluvial processes becoming inherited in the human-made soils in the quarry floor (Fig. 8).
Morphological Characterization of Bulk Control Samples: Stage 1C
The two saturated bulk control soil samples taken near the ponded area in the quarry were dried (in air) and visually described using hand-held samples and through a Wild Leitz M420 stereo microscope (Table 3) and photographed in a glass petri dish (Fig. 9). The two bulk control samples revealed the following highly distinctive morphological properties (Table 3): (i) pink matrix color, (ii) redoximorphic features comprising a combination of concentrations and depletions indicating a poorly drained soil, (iii) abundant quartz particles with angular shapes and conchoidal surface fractures, (iv) prominent hard irregular clay fragments (<2 mm; 2 to <5 mm) with a pale yellow (2.5Y 8/3) color, (v) a massive structure, which parts when dry to a single grain structure (loose to powdery) with no observable roots or detectable carbonate, and (vi) clayey sand and coarse sandy texture with large amounts of gravel (55%). The control clay fragment sample observed in Fig. 9 (sample Q4) has a pale yellow (2.5Y 8/3) Munsell color and is essentially a hard lump or fragment of clay with a distinctive waxy feel.
Chemical and Mineralogical Characterization of Bulk Control Samples: Stage 2C
The dried bulk control soil samples were dry sieved through a 2-mm sieve and analyzed for pH and EC. The pH (5.1) and EC (0.16 dS/m) values (Table 3) indicated very strongly to medium acid conditions and low levels of mineral salts (i.e., nonsaline environment).
The two dried bulk control samples were dry sieved through a 0.4-mm sieve and ground in an agate mortar and pestle. The resulting fine powders were lightly pressed into aluminum sample holders for XRD analysis. The XRD pattern for the fine powdered control sample (Q1) is shown in Fig. 5 (red line). The two control soil samples from the quarry contained the following similar mineral components: well crystalline kaolinite, quartz; muscovite, feldspar, and talc (Table 4). Also, the detail in the XRD patterns from these samples was similar as shown in Fig. 5 and 6.
Mineralogical Characterization of White Fragments in Control Samples: Stage 3C
The pale yellow (2.5Y 8/3), irregular clay fragments (1–10 mm diameter) were hand-picked from the bulk control samples (Q1 and Q3) and ground in an agate mortar and pestle. A subsample from the large clay fragment (Q4) was also ground in an agate mortar and pestle. The resulting fine powders were lightly pressed into aluminum sample holders for XRD analyses (Fig. 6, Table 4). The mineralogical composition of both the white clay fragments extracted from Q3 (Fig. 9 and 6) and the large clay fragment (Q4) were similar with dominant well crystalline kaolinite and traces of quartz, muscovite, and feldspar (Table 4). However, the large clay fragment (Q4) also contained detectable minor talc, anatase, and rutile.
Comparisons between Questioned and Control Samples—Construction of Soil–Landscape Model: Stage 4C
The two questioned soil samples from the shovel had the following 10 out of 10 similar soil morphological features and pH and EC values (Table 3) as the two control samples collected from the wet sump of the quarry:
Pink matrix Munsell soil colors
Similar redoximorphic features
Clayey sand and coarse sandy texture with large amounts of gravel (25–55%)
Angular quartz particles with conchoidal fractures
Prominent hard irregular clay fragments (1–5 mm) with a pale yellow (2.5Y 8/3) color
Non effervescence class (NE), which indicates that no carbonates are present
A massive structure which parted when dry to a single grain structure (loose to powdery) with no observable roots
Non water repellent class (N)
Very strongly to medium acid conditions (pH 4.7–5.1)
Low levels of mineral salts (i.e., nonsaline environment) (EC 0.14–0.16 dS/m)
The questioned bulk soil samples from the shovel (back and front) contained the following five out of five similar mineral components as the bulk control samples collected from the sump of the quarry: well crystalline kaolinite (subdominant), quartz (dominant); muscovite (minor); and trace of feldspar and talc (Table 4). The fine details expressed in the XRD patterns from these two bulk soil samples were remarkably similar as shown in Fig. 5 (i.e., can be likened to “finger print pattern comparisons” showing subtle similarities and differences). Similarly, the hand-picked pale yellow fragments extracted from the questioned sample (back of shovel) and the control samples (Q3) and (Q4), had similar mineral composition: well crystalline kaolinite (dominant) with traces of quartz; muscovite, and feldspar (Table 4). X-ray diffraction patterns can also be likened to “finger print comparisons,” with the XRD patterns for the white fragment sample extracted from the back of the shovel (UJU 12.2) and the white fragment samples extracted from sample Q3 being similar (e.g., subtle similarities and differences as shown in Fig. 6).
In summary, the soil morphological descriptors of the bulk soil samples and mineralogical data indicated that both groups of samples had remarkably similar properties, which is reflected in their similar soil classification, as shown in Table 5. Based on the soil forensic evidence the questioned soil samples from the shovel (back and front) and the control samples collected from the Oakbank quarry had a “conclusive” degree of comparability (Table 2) indicating that the samples were virtually “indistinguishable.” To conclude, the soil evidence was unequivocal in terms of all comparison criteria used, thus revealing a specific area in the quarry comprising complex human-made soil–regolith materials with overlying wet and/or pooled water as a very likely location of the two buried bodies.
Despite there being no indication of soil disturbance at the quarry, based on the soil forensic evidence the CSIRO pedologists remained convinced that the bodies of the two women would be located in a wet area of the quarry. Consequently, police performed daily inspections of the quarry. Some 3 weeks after the murder the bodies of the women were found in the wet area of the quarry 15 m from where the control samples were collected when one of the bodies was uncovered by foxes. One day later, foxes uncovered the other body.
The case against Matthew Holding was overwhelming. Holding pleaded guilty in the Supreme Court to the murders of his mother and grandmother. According to Barrett (2003), in his confession Holding said he had woken that morning and asked his grandmother for a cigarette. When she refused and started lecturing him about smoking, he “snapped” and hit her on the head with a bottle. He then slit her throat with a kitchen knife. When his mother intervened, Holding also cut her throat. As he loaded the bodies in the boot of the car, he realized his mother was still alive and struck her with a permapine post, inflicting a fatal blow. He had partially cleaned up the scene of the crime and took a shovel from the house to bury the bodies in a locality (quarry) comprising complex human-made soil with overlying ponded water where they would be unlikely to be found immediately.
Psychiatric reports were ordered before sentencing, and Holding was found to be delusional and psychotic (Barrett, 2003). He had been schizophrenic for many years, exacerbated by his use of illegal drugs. The judge accepted that Holding knew what he was doing at the time of the killings. The judge took Holding’s mental illness into account in sentencing and, after the mandatory term of life imprisonment, imposed an 18-year non-parole period.
Soil Forensic Exhibit
Today we need to work harder to attract students to study soil science, especially pedology, because they have a somewhat dry and dusty, or a muddy and smelly, image of soil science. This stereotype can be far from reality. Forensics in general is attractive to students, just look at the popular television shows CSI, X-files, and numerous others. There has also been an emergence of university courses geared toward producing forensic scientists. As an educational aid for soil scientists, forensic scientists, and police, an interactive soil forensic exhibit was developed based on the case study reported in this paper (Fig. 10). The exhibit includes a replica shovel with quarry soil adhered in clear epoxy resin taken from the original burial site. Four small clear epoxy resin blocks incorporating various components of the quarry soil are placed on the exhibit’s illustrative laminated poster using fabric hook-and-loop fastener tape. This enables the epoxy resin blocks to be easily removed and closely observed with the naked eye or magnifying glass supplied to view individual particles and fragments of angular quartz particles and small white fragments of clay (Fig. 10). The exhibit also includes photographs of the back and front of the original shovel used by the suspect to bury the two victims in a quarry comprising “human-made soils” (marked E in Fig. 10 and 11). The photograph of the replica shovel has arrows pointing to three removable clear epoxy resin blocks (5 by 2 by 1 cm), which incorporate original quarry soil features such as the small white clay fragments (see Fig. 10b). The text is presented in the style of a newspaper article using old style typewriter fonts to report the major points under the following four headings on how the double murder case was solved using soil evidence:
The Crime and Evidence—wife, mother and son missing; with locality map.
Enter The Scientists—soil clues on the shovel found in the suspect’s car.
Closing in on the Site—evidence from soil/geological maps and field survey narrowed the search down to a small group of gravel quarries.
Case Solved—samples from quarries were collected and one located in the sump or wet area of the quarry produced identical mineralogy to the soil–regolith on the shovel, which helped identify the burial site of the two bodies. The suspect was convicted and received a life sentence.
Several copies of the interactive exhibit have been made and are on permanent display in various national and international institutes, police departments, universities and forensic laboratories. The sidebar briefly expands the text, selected photographs, and diagrams included on the exhibit to highlight the significance and educational value of forensic soil science, which succinctly explains how pedologists and mineralogists successfully used soil morphology, soil chemistry, soil survey, and clay mineralogy (using X-ray diffraction) expertise to help solve this double murder case.
Summary and Conclusions
In the first part of the paper, we briefly reviewed the history of forensic soil science, the reasons why soil materials are considered to be powerful, perhaps ideal, pieces of contact trace evidence, and the key methods used to systematically conduct soil forensic investigations.
In the second part of the paper, a case study was detailed, which demonstrates how standard pedological methods and application of map, field, and laboratory approaches involving mainly X-ray diffraction are critical in developing the following three sequential predictive soil–regolith models from: (i) microscopic scale involving the questioned soil features on the shovel, to (ii) landscape scale involving soil/geological maps, and (iii) field investigations involving the critical sampling of control samples in a quarry to solve a complex double murder case at a range of scales. When examining soil evidence, there are a range of stages involving screening soil tests that help provide pieces of a puzzle and more detailed tests that provide definitive answers. If one has enough puzzle pieces a picture starts to form. In this double murder investigation, by combining pedological information on the nature of the questioned soil on the shovel with known soil–regolith data from the region, which included soil map information together with associated evidence, pedologists and detectives were led to investigate a quarry in the Adelaide Hills. The specific sampling of control soils in the quarry followed by definitive soil characterization using X-ray diffraction added crucial assistance in solving the puzzle to locate and pinpoint within 15 m the two bodies in a “subaqueous human-made soil” where they would be unlikely to be found immediately.
In the third part of this paper, we presented an educational aid for soil scientists, forensic scientists, and police in the form of an interactive soil forensic exhibit based on the complex double murder case study. The exhibit includes a replica shovel with quarry soil taken from the original burial site adhered in clear epoxy resin. Four small clear epoxy resin blocks incorporating various components of the quarry soil are also placed on the exhibit’s illustrative laminated poster using fabric hook-and-loop fastener tape.
The successful use of CSIRO soil expertise in helping solve a double murder led to the formation of the Centre for Australian Forensic Soil Science (CAFSS) in 2003 involving six core partners across Australia. Since then, the center has advised on more than 100 criminal investigations and run national and international workshops.