Journal Pre-proof

Building an isoscape based on tooth enamel for

human provenance estimation in Brazil

Ricardo de Oliveira Mascarenhas, João Paulo Sena-

Souza, Stefano M. Bernasconi, Judith A.
McKenzie, Crisógono Vasconcelos, Taís Ribeiro
Muniz, Matheus Pereira Nogueira e Silva, Fábio
Augusto da Silva Salvador, Anelize Manuela
Bahniuk Rumbelsperger

PII: S0379-0738(21)00429-1
DOI: https://doi.org/10.1016/j.forsciint.2021.111109
Reference: FSI111109

To appear in: Forensic Science International

Received date: 5 May 2021
Revised date: 12 November 2021
Accepted date: 15 November 2021
Please cite this article as: Ricardo de Oliveira Mascarenhas, João Paulo Sena-
Souza, Stefano M. Bernasconi, Judith A. McKenzie, Crisógono Vasconcelos,
Taís Ribeiro Muniz, Matheus Pereira Nogueira e Silva, Fábio Augusto da Silva
Salvador and Anelize Manuela Bahniuk Rumbelsperger, Building an isoscape
based on tooth enamel for human provenance estimation in Brazil, Forensic
Science International, (2021) doi:https://doi.org/10.1016/j.forsciint.2021.111109
This is a PDF file of an article that has undergone enhancements after acceptance,
such as the addition of a cover page and metadata, and formatting for readability,
but it is not yet the definitive version of record. This version will undergo
additional copyediting, typesetting and review before it is published in its final
form, but we are providing this version to give early visibility of the article.
Please note that, during the production process, errors may be discovered which
could affect the content, and all legal disclaimers that apply to the journal pertain.
© 2021 Published by Elsevier.
 Building an isoscape based on tooth enamel for human provenance
estimation in Brazil
Ricardo de Oliveira Mascarenhasa,e*, João Paulo Sena-Souzab, Stefano M. Bernasconic,
Judith A. McKenziec, Crisógono Vasconcelosd, Taís Ribeiro Munize, Matheus Pereira
Nogueira e Silvaf, Fábio Augusto da Silva Salvadora, Anelize Manuela Bahniuk
Rumbelspergere

a
Brazilian Federal Police, Technical-Scientific Section, R. Profa. Sandália Monzon 210,
82640-040, Curitiba-PR, Brazil
b
Department of Geosciences, State University of Montes Claros (Unimontes), Campus
Universitário Prof. Darcy Ribeiro, Av. Prof. Rui Braga, s/n - Vila Mauriceia, 39401-089

of
Montes Claros MG, Brazil.
c
Department of Earth Sciences, ETH Zürich, Sonneggstr. 5, 8092 Zürich, Switzerland
d

ro
Center for Technological Development (CEDES), Geological Survey of Brazil, Av. Pasteur
404, 22290-240, Rio de Janeiro-RJ, Brazil
e
Laboratory of Mineral and Rock Analysis (LAMIR), Department of Geology, Federal
f -p
University of Paraná, PO Box 19062, 81531-980 Curitiba-PR, Brazil
Department of Geology, Federal University of Paraná, PO Box 19062, 81531-980 Curitiba-
PR, Brazil
re
*
Corresponding author. Tel.: +55 41 32567631
E-mail adress: mascarenhas.rom@pf.gov.br (R. O. Mascarenhas)
lP

Abstract
In this study, we present a correlation between δ18OC values of carbonate in tooth
a

enamel samples from the modern Brazilian population and the available δ18ODW data for the
rn

meteoric water from the Global Network of Isotopes in Precipitation (GNIP). Tooth enamel
from 119 Brazilian individuals from five different regions of the country were analyzed. The
u

δ18OC isoscape obtained is in good agreement with the isoscape based on regional
Jo

meteoric and drinking water. The regression matrix obtained for the δ18O values of the
carbonate tooth enamel and meteoric water was used to build an isoscape using the
regression-kriging approach. Our data show that Brazil can be divided in two main regions
with respect to the δ18O values of the carbonate tooth enamel: (1.) the most easterly part of
the northeast region, which is characterized by a warm and dry climate and (2.) the
remainder of the country, stretching from the Amazon rain forest to the more southernly
regions. The data herein reported can be used for forensic purposes related to human
identification.

Keywords: Tooth enamel, provenance, stable isotopes, forensic geology, missing person
identification.
1. Introduction
Matching specific missing persons with unidentified deceased individuals during
investigations conducted at forensic medical institutes is a worldwide challenge. This is
especially true for cases in which the bodies are highly decomposed and fingerprint
identification is no longer possible. Also, some countries do not employ fingerprints for civil
identification, thus restricting its use.
Similarly, dental records are not easily accessible and DNA databases in many
countries often lack information regarding missing persons. According to INTERPOL [1],
fingerprints, DNA and dental records are the only way to actually identify an individual and,

of
for this reason, they are called primary identification methods. All other methods (e.g.,
examination of tattoos, clothing used by the victim or personal belongings) are considered

ro
secondary procedures and may be applied to guide the usage of a primary one.
In this last decade, stable isotopic ratios of specific materials have been
-p
successfully used to aid in human identification [2–5]. Because each type of human tissue
has a different turnover rate, isotopic analyses may be used to monitor different periods of
re
an individual´s timeline [6]. For instance, hair and nails may be used to investigate recent
lP

dietary history [7], whereas trabecular (spongy) bones, such as rib bones, have a turnover
of approximately 4-6 years and cortical (compact) bones, such as the femur, have a much
slower turnover of approximately 25-30 years [4,8].
a

Unlike other tissues, tooth enamel preserves its isotopic signature after
rn

mineralization. This happens because the ameloblasts, cells responsible for enamel
formation, expire during this process, leaving a residue of cell-free tissue [9,10]. Although
u

some later mineralization may occur in regions of the teeth experiencing demineralization
Jo

by abrasion or caries [11,12], tooth enamel isotopic signals have been successfully used in
provenance studies, helping to determine the place where an individual lived during
childhood or adolescent years [13].
Tooth enamel has a low organic carbon content. During the maturation phase, in
situ proteases remove almost completely the extracellular matrix, leaving only 1–2 weight %
of the remaining protein and a small amount of water. The result is a dense mineralized
tissue with more than 95% (by weight) of mineral content [14]. This process results in
enamel being very resistant to diagenesis and, for this reason, it is often the preferred
tissue in paleodietary and paleoclimate studies [15–18]. The mineral fraction of tooth
enamel is composed of calcium hydroxyapatite: Ca10(PO4)6(OH)2. Since bio-apatite is, in
general, calcium-deficient, some of the phosphate ions are replaced by carbonate ions in
2
order to maintain electrical balance (B-type substitution). Additionally, part of the hydroxyl
ions may be replaced by carbonate ions in a proportion of 2:1 (A-type substitution),
resulting in the following approximate composition: Ca4.5[(PO4)2.7(HPO4)0.2(CO3)0.3](OH)0.5
[19,20]. Because most of the oxygen in hydroxyapatite is incorporated in the phosphate
ions and because the P-O bonds are stronger and thought to be more resistant to
diagenesis, many isotopic studies focused on the phosphate ions using their oxygen
isotopic ratio (δ18OP) as a proxy [13,16, 21, 22]. However, the analytical procedure is time-
consuming and, therefore, analyzing the oxygen isotopic ratio (δ18OC) of the structural
carbonate in the tooth enamel is more cost effective. Besides, the C-O bonds are more

of
resistant to diagenesis than previously thought [23].
The isotopic ratio of the oxygen atoms contained in the phosphate ions (δ18OP) is

ro
related to the ratio of the oxygen of ingested water (i.e., drinking water and water contained
in food) during the time the tooth is mineralizing [13,24]. Despite the influence of the water
-p
present in food, drinking water represents the dominant portion of the total water ingested
by an individual and, therefore, has a larger influence on δ18OP values of the phosphate
re
ions and similarly on δ18OC values of the carbonate ions [13,25]. Thus, there is a direct
relationship between δ18ODW values of the drinking water and the δ18OC values of the
lP

structural carbonate in tooth enamel [23].

There is still some criticism regarding the application of δ18ODW from drinking water
a

as a direct proxy for provenance determination using the δ18OC of tooth enamel from a
rn

deceased person. Because this requires a rigorous usage of equations, Pollard et al. [26]
proposed that geographical attributions should be based on the direct comparison of
u

measured values of δ18OP rather than on predicted values of δ18ODW. Meier-Augenstein [27]
Jo

also suggested comparing the results obtained using different equations during data
interpretation. Despite this problem, because traceable teeth samples are normally difficult
to obtain, many forensic studies still rely on comparing the results obtained for tooth enamel
with the mapped δ18ODW values for meteoric water, such as the data produced by the GNIP
(Global Network of Isotopes in Precipitation) project compiled by the IAEA (International
Atomic Energy Agency) [2–5,28]. Numerous global scale isoscapes of δ18ODW values for
meteoric water are available [29,30] and can serve as a basis for a spatial prediction of
tooth enamel δ18OC values. The term isoscape (isotopic landscapes) was first coined by
West et al. [31] to represent any spatially-explicit prediction of isotopic values across a
landscape. One of its advantages is that it can provide isotopic values for unmonitored

3
areas, filling sampling voids [32], such as those naturally expected when working with
samples difficult to obtain, for instance human tooth enamel of known provenance.
With this study, we propose to develop a δ18OC isoscape based on the isotopic
composition of tooth enamel for the modern Brazilian population, which may potentially be
used to assist with human identification. The presented spatial prediction represents the
first tooth enamel δ18OC isoscape for human identification and provenance studies in Brazil.

2. Materials and Methods

2.1 - Tooth Samples

of
Dental samples were obtained from the tooth bank of the local Faculty of
Odontology or from private dentists, whereas other samples were voluntarily donated by

ro
patients who gave their informed consent for use in this study. Some of the teeth were
extracted from patients with periodontal disease. Otherwise, the teeth were removed
-p
because they were occluded, badly positioned or for other professional decisions unrelated
to this study.
re
In total, the analyzed teeth belonged to 119 different Brazilian individuals living in
lP

geographic locations throughout Brazil (Figure 1). The provenance of the samples, based
on the patients’ or dentists’ statements, is shown in Table 1.
a

Table 1 – Provenance of the teeth samples according to patients’ or dentists’ statements

rn

Region of Brazil State Number of teeth samples

Amazonas (AM) 2
u

North
Pará (PA) 2
Jo

Bahía (BA) 1
Northwest Ceará (CE) 2
Paraíba (PB) 1
Brasília (DF) 1
Middle-West Mato Grosso (MT) 1
Mato Grosso do Sul (MS) 1
Minas Gerais (MG) 1
Southeast
São Paulo (SP) 7
Santa Catarina (SC) 5
South
Paraná (PR) 95

4
 Regarding biological sex, 30 samples were from males and 59 samples were from
females. The gender identification of 30 samples (2 from Amazonas and 28 from Ponta
Grossa) is unknown. Third-molars corresponded to 80 samples, and 11 samples were
divided into central incisor, lateral incisor, first premolar or second premolars. The dental
elements of 28 samples from Ponta Grossa, Paraná were not identified. Regarding the age
of the individuals, 64 % were between 16 to 39 years old. The geographic location and
available relevant individual data for each studied tooth sample are given in Table 2.
The teeth were first cleaned with a 0.5 % peracetic acid solution for 15 minutes in

of
order to remove organic impurities, then rinsed with mili-Q water and dried at 25 °C to 30 °C
for 36 hours before sampling. Enamel samples were taken using a diamond coated drill bit

ro
attached to a Dremel® 3000 multitool. To avoid contamination, the enamel´s external
surface was discarded. The drill bit was cleaned with 10 % hydrochloric acid between
-p
taking discrete samples. Enamel was sampled in regions of the tooth where it was thicker
because Chenery et al. [23] had previously shown that no significant intra-tooth variation
re
was observed when sampling the distal, mesial, central, upper or lower crown. The enamel
lP

powder obtained as described was analyzed without any further treatment.

a
u rn
Jo

Figure 1 – Place of birth of the individuals who donated the teeth used in this study. The
shaded area contains a map of Brazil showing the distribution of tooth samples within the
country with most samples (95) being sourced from the southern state of Paraná.

2.2 - Oxygen Isotope Analyses

Oxygen isotope analyses were performed on tooth enamel structural carbonate.
Tooth enamel powder (about 500 µg) was weighed into 4.5 ml glass vials (Labco, High
Wycombe, UK, part No. 948 W), capped with Labco butyl rubber septa, following the
procedure described by Breitenbach and Bernasconi [33]. Isotopic analyses were carried
out using a Thermo Fisher Scientific GasBench II, equipped with a CTC autosampler (CTC

5
Analytics AG), and coupled to a ConFlo IV interface and a Delta V Plus mass spectrometer
(both Thermo Fisher Scientific) in the Stable Isotope Laboratory of the Department of Earth
Sciences at the ETH Zürich. The exetainers containing the samples were placed in a hot
block at 72.0 ± 0.2 °C and were flushed for 240 seconds on a model 222XL autosampler
(Gilson™, Middleton, WI, USA), with helium grade 5.0 (99.999% He), at a flow rate of
55ml/min, prior to the addition of 100% H3PO4.The samples were then left to react for 69
minutes with the orthophosphoric acid and the CO2 generated was sampled with a
sampling loop of 100 μl and transported in a helium stream of 2ml/min to the mass
spectrometer. Due to the differences in the relative abundance between the heavier and

of
lighter oxygen isotopes present in the carbonate ion (16O: 99.759 %; 18
O: 0.204 %), the
isotopic ratios are reported as a deviation from a standard value (δ) expressed in parts per

ro
thousand (‰) using the following equation:

( -p
)
[34]
re
where Rsample is the ratio of the heavy to the light isotope measured for the sample and
lP

Rstandard is the ratio for the standard. For carbonate, the standard is VPDB (Vienna PDB),
considered equivalent to PDB (for Pee Dee Belemnite) [35,36]. The results were
a

normalized to the VPBD scale using two carbonate in-house secondary standards, MS2
rn

(δ18O = -1.81 ‰) and Isolab B (δ18O = -18.59 ‰), calibrated with NBS 19 and NBS 18, both
standards supplied by IAEA, Vienna, Austria. The standard deviation for the two secondary
u

standards used was below ±0.09 ‰ (1σ) for δ18O along the three analytical sections.
Jo

Samples from 26 individuals were chosen to be analyzed in duplicate. The standard

uncertainty, calculated using the spreadsheet supplied by Szpak et al. [37] in the Appendix
G of that publication, was ±0.11 ‰. The isotope data are presented in Table 2. For the
samples analyzed in duplicate, an average value is reported.

2.3 – Meteoric Water Isoscape

To test the relationship between the δ18O values of teeth and drinking water, we
assume the existence of a high correlation between drinking water and meteoric water [38].
We are still building a drinking water isoscape for Brazil. However, there are global scale
isoscapes of meteoric water in a grid format adaptable for a regional scale in Brazil [39]. For
this study, we used the regionalized cluster-based water isotope prediction (RCWIP)

6
proposed by Terzer et al. [30]. The authors compiled data from the Global Network of
Isotopes in Precipitation (GNIP) to model a global scale δ18ODW isoscape of precipitation.
Here, we used the regional amount-weighted mean annual δ18ODW of precipitation for South
America, available as GeoTIFF from the International Atomic Energy Agency (IAEA)
website (www.iaea.org/water) in 10 arcmin of spatial resolution.

2.4 – δ18OC Tooth Enamel Isoscape

We organized the database containing the geographical coordinates latitude and
longitude and the δ18OC value of each tooth sample. We aggregated samples with very

of
close distances (<0.01º) and calculated the mean δ18OC of near samples. For each tooth
sample, a value of δ18ODW of meteoric water was extracted from the corresponding pixel of

ro
the RCWIP isoscape to assemble the regression matrix. Since δ18ODW is on the VSMOW
scale (Vienna Standard Mean Ocean Water), whereas δ18OC is on the VPBD scale, the
-p
equation proposed by Coplen [40] was used to convert the isotopic values between the two
scales (from VSMOW to VPBD):
re
δ18OVSMOW = 1.03091 × δ18OVPBD + 30.91
lP

Once the regression matrix was assembled, the procedure chosen to perform the
a

isoscape was the regression-kriging approach (Figure 2).

u rn
Jo

Figure 2 - Methodological workflow for δ18O isoscape modeling. In preliminary estimate

function, a is the regression coefficient and b is the intercept.

First, we tested if there is a correlation between δ18OC and δ18ODW from a global
model isoscape (δ18ORCWIP). Then, we applied a linear regression model using observed
7
δ18OC and δ18ORCWIP. The first regression generated a preliminary function (pi) and
residuals (ei). Then, we applied the preliminary estimate function for each grid cell of the
precipitation isoscape for a preliminary spatial prediction (px). For the next step, we
interpolated the regression residuals to a continuous surface (êx) using a classic variogram
model (γ(h)) and kriging. We chose the regression-kriging method because it allows us to
directly analyze the relationship between the tissue isotope ratio and the spatially explicit
isotopic ratio of precipitation. The regression's residual kriging corrects the equation from
the linear regression, including a spatial model that represents the tissue isotope variability
not explained by the predictor variable.

of
The final model still preserves some patterns of the environment isoscape, as
expected. Several works have used this approach in modeling isoscapes at both regional

ro
[41,42] and global scales [43]. The variogram model describes the variance between pairs
of the data points. We used the automatic parameters detection functions of the R package
-p
gstat [44] to fit an experimental semivariogram, with a cutoff lag distance of 1500 km and a
spherical model. For the final model isoscape, we added the interpolated residuals to the
re
regression surface. To access the final prediction uncertainty, we interpolated the predicted
lP

values and calculated the standard deviation surface from the variance. We applied the
cross-validation to assess the accuracy of ordinary kriging of residuals, with the “leave-one-
out” methodology. We present the accuracy by evaluating the root-mean-square-error
a

(RMSE) and the coefficient of determination (R2) [45]. The RMSE indicates the deviation of
rn

the predicted value by kriging from the linear regression residual:

u
Jo

√ ∑ ̂ ) ))

where ̂ ) is the estimation value and ) is the observed residual.

All procedures were conducted using the R Statistical and Computing Environment
(R Developed Core Tem, 2013). The GeoTIFF archives generated in this study were
transferred to ArcGIS 10.x for layout elaboration.

3. Results and Discussion

The regression analysis showed that the distribution pattern of the observed tooth
enamel δ18OC values was explained in 35 % of the cases by the δ18O of the model of Terzer

8
et al. [30] (R2 = 0.35; p-value < 0.01) (Figure 3). Although the coefficient of determination is
low, the established relationship is statistically significant and allowed us to follow the steps
for the elaboration of tooth enamel δ18OC isoscape. Besides, the relationship between
δ18OC for carbonate teeth enamel and δ18ODW of drinking water is also corroborated by
many previous works [13,23,27]. In this work we used different dental elements, not only
third-molars, but also some central incisors, lateral incisors, first or second premolars to
build the model. Despite the fact that different dental elements mineralize at different ages,
the δ18OC should still depend on the water drunk during that time frame. Therefore, building
models using different dental elements should make models more robust to be used in real

of
cases and should be encouraged.

ro
-p
re
a lP
u rn
Jo

Figure 3 - Scatter plot showing that the observed tooth enamel δ18OC (VPBD) is a function of
δ18ORCWIP (VPBD) extracted from grid format isoscape [30]. R2 = 0.35; p-value < 0.01. The bold
black line is the regression line, the light grey area represents 95% confidence level for the
following model: δ18OC = 0.4006 × δ18ORCWIP (VPBD) + 8.8369.

The regression residuals followed a normal distribution (W = 0.98, p-value = 0.64)

and indicated some spatial patterns. The highest positive residuals are concentrated in the
northeastern region of Brazil. The most negative residuals appeared in the states of
Amazonas, Mato Grosso, and Bahia. In states where there is a higher sample density,
there was also a spatial pattern of residuals, with positive residuals concentrated in the
south of Paraná and Santa Catarina, and negative residuals concentrated in the north of

9
Paraná and São Paulo (Figure 4A). This spatial structure allowed us to perform the ordinary
kriging interpolation (Figure 4B) applying a semivariogram model (Figure 4C).

of
ro
-p
re
lP

Figure 4 - Spatial analysis of the regression residuals. Spatial distribution of residuals (A);
a

Ordinary kriging surface of residuals (B); Experimental semivariogram with following

parameters: Nugget effect (C0) = 0.13177; Sill (C + C0) = 0.70758; Range = 787.43 km (C).
rn

Using only the tooth enamel δ18OC value, it was not possible to distinguish individuals
u

originating from the south-east of Brazil (São Paulo and Minas Gerais), Bahia (North-east),
Jo

South (Paraná and Santa Catarina), Middle-west (Mato Grosso, Mato Grosso do Sul) and
even from the Amazonas (North). This may be explained by the fact that the moisture
generated by the evapotranspiration in the Amazon rainforest is responsible for most of the
meteoric water in these regions, a phenomenon called “flying rivers” [46-48]. In these areas,
lower δ18OC values (from -4.55 ‰ to -6.96 ‰) were observed (Figure 5). Higher values for
δ18OC (up to -2.00 ‰) were observed in the northeast of the country, a region marked by a
dry climate [49]. This gives a maximum amplitude of approximately 5 ‰ for the Brazilian
population. This amplitude is a bit lower than the one predicted for the United States, a
country with similar dimensions to Brazil (about 7.7 ‰) [28].

10
 of
ro
-p
re
a lP
u rn

Figure 5 – Tooth enamel δ18OC isoscape for Brazil.

Jo

The ordinary kriging of the residuals presented a RMSE of 0.53, with R2 = 0.30 (p-
value < 0.001). Although it has a low coefficient of determination, this indicates a robust
model considering the sample design. Most of the observed values are concentrated near
the capital of the state of Paraná. The spatial resolution of the input model causes the
inclusion of a large quantity of samples in the same grid cell. The variation can be explained
by several factors such as people using hydrating sources diverse from tap water, such as
juices, beverages and bottled water [45,46]. Besides, the amount of physical activity is also
known to cause water fractionation among individuals [47,48]. All these factors were not

11
taken into account in this work but may inspire future work regarding the Brazilian
population in order to refine the model here proposed.

The model is reliable in places with sample coverage. The model's uncertainty
increases as the sample density decreases (Figure 6). That is a general phenomenon in
most spatial models [50]. Therefore, future studies that try to improve or apply the tooth
enamel isoscape presented here must increase the amount of sampling throughout the
study area.

of
ro
-p
re
a lP
u rn
Jo

Figure 6 - Standard deviation map for predicted tooth enamel δ18OC representing the spatial
uncertainty of the isoscape.

4. Conclusions
To the best of our knowledge, this is the first attempt to construct an isotopic
database using the isotopic composition of tooth enamel for the modern population in

12
Brazil. We propose that the data provided herein can be used for forensic purposes related
to human identification.
Our δ18OC isoscape compiled from tooth enamel is in good agreement with the
isoscape based on the δ18ODW composition of the regional meteoric and drinking waters.
Despite this positive correlation, a direct comparison between the isotopic value of tooth
enamel should always be preferentially used to provide better results and should be
persued. However, our data offer further validation for the identification approach using the
δ18ODW of meteoric water, whenever tooth enamel data are not available.
In Brazil, there are clearly two regions that can be easily distinguished using δ18ODW

of
data: 1) the most easterly part of the northeast region, which is characterized by a warm
and dry climate and 2) the remainder of the country, stretching from the Amazon rain forest

ro
to the more southernly regions. The latter region covered by this extensive range reinforces
the importance of the moisture provided by the Amazonian forests to the meteoric water
-p
cycle from the northwest to the southeast of Brazil, a vast area which crosses the major
recharge zones for the dams and aquifers used for public water supply.
re
Considering that Brazil is a country with continental dimensions, further isotopic
lP

studies analyzing tooth samples covering regions, beyond those included in this study,
should be pursued to compile a complete nation-wide δ18OC isoscape. Also, to better
understand the overall inter-individual variations throughout Brazil, it is clearly necessary to
a

refine the herein presented δ18OC isoscape.

rn

Table 2 – Available location and relevant individual data for each studied tooth sample with
u

corresponding δ18OC (VPBD) values. Reproducibility of the measurements is better than

Jo

0.17 (1σ). The δ18ORCWIP, converted to the VPBD scale using the equation proposed by
Coplen (1988) [40], was also included for comparison.
Individual Declared City Biological Tooth
Statea Lat.b Long.c Age δ18OC (‰) δ18ORCWIP (‰)
Sample of Birth Sex type
1 Manaus AM 3.12 60.02 - - - -6.57 -34.66
2 Manaus AM 3.12 60.02 - - - -5.93 -34.66
3 Marabá PA 5.38 49.13 28 Female 3rd molar -4.47 -32.95
rd
4 Marabá PA 5.38 49.13 - Female 3 molar -3.64 -32.95
rd
5 Fortaleza CE 3.74 38.53 - Female 3 molar -2.86 -32.14

a
Brazilian state: AM = Amazonas, PA = Pará, CE = Ceará, PB = Paraíba, BA = Bahia, MG = Minas Gerais,
DF = Distrito Federal, MT = Mato Grosso, MS = Mato Grosso do Sul, SP = São Paulo, PR = Paraná, and SC
= Santa Catarina.
b
Lat.: degrees South latitude.
c
Long.: degrees West longitude.

13
Individual Declared City Biological Tooth
Statea Lat.b Long.c Age δ18OC (‰) δ18ORCWIP (‰)
Sample of Birth Sex type
6 Campos Sales CE 7.07 40.37 49 Female 3rd molar -2.40 -32.49
rd
7 Patos PB 7.02 37.28 20 Female 3 molar -2.00 -31.54
8 Irecê BA 11.30 41.86 26 Female 3rd molar -5.73 -33.48
rd
9 Betim MG 19.97 44.2 25 Female 3 molar -5.60 -35.40
Central
10 Brasília DF 15.79 47.88 46 Male -5.12 -35.77
Incisor
São José dos 2nd
11 MT 15.63 58.18 31 Female -6.50
Quatro Marcos premolar -34.96
12 Mundo Novo MS 23.94 54.28 31 Female 3rd molar -5.92 -34.91
rd
13 Avaré SP 23.10 48.92 26 Male 3 molar -6.28 -35.82
rd
14 Jotapiranga SP 24.70 48.01 27 Female 3 molar -5.08 -34.58

of
15 Ribeirão Preto SP 21.17 47.81 30 Male 3rd molar -6.96 -35.14
rd
16 São Paulo SP 22.38 48.45 30 Male 3 molar -4.97 -35.26
rd
17 São Paulo SP 22.38 48.45 24 Female 3 molar -5.66 -35.26

ro
rd
18 São Paulo SP 22.38 48.45 26 Female 3 molar -5.43 -35.26
rd
19 São Paulo SP 22.38 48.45 21 Male 3 molar -6.36 -35.26
20

21
Araucária
Bocaiúva do
Sul
PR

PR
25.45 49.24

25.21 49.11
30

54
-p
Male

Male
rd
3 molar

3rd molar
-5.07

-5.21
-37.09

-37.09
re
Catanduvas do
22 PR 25.24 53.15 20 Female 3rd molar -5.78 -36.30
Sul
rd
23 Curitiba PR 25.45 49.24 42 Female 3 molar -5.47 -37.09
lP

24 Colombo PR 25.29 49.23 19 Female 3rd molar -5.93 -37.10

rd
25 Colombo PR 25.29 49.23 16 Male 3 molar -6.25 -37.10
Central
a

26 Colombo PR 25.29 49.23 69 Male -5.97

Incisor -37.10
27 Colombo PR 25.29 49.23 37 Male 3rd molar -5.35 -37.10
rn

Cornélio rd
28 PR 23.22 50.62 53 Male 3 molar -5.86
Procópio -35.46
rd
u

29 Curitiba PR 25.45 49.24 21 Male 3 molar -5.38 -37.09

rd
30 Curitiba PR 25.45 49.24 19 Female 3 molar -5.35 -37.09
Jo

rd
31 Curitiba PR 25.45 49.24 33 Female 3 molar -6.26 -37.09
rd
32 Curitiba PR 25.45 49.24 20 Male 3 molar -5.68 -37.09
33 Curitiba PR 25.45 49.24 20 Female 3rd molar -5.69 -37.09
rd
34 Curitiba PR 25.45 49.24 23 Female 3 molar -5.60 -37.09
rd
35 Curitiba PR 25.45 49.24 23 Female 3 molar -5.88 -37.09
rd
36 Curitiba PR 25.45 49.24 31 Female 3 molar -5.38 -37.09
rd
37 Curitiba PR 25.45 49.24 31 Female 3 molar -4.96 -37.09
1st
38 Curitiba PR 25.45 49.24 54 Female -5.01 -37.09
premolar
1st
39 Curitiba PR 25.45 49.24 61 Female -4.55 -37.09
premolar
Central
40 Curitiba PR 25.45 49.24 70 Female -5.53
Incisor -37.09
41 Curitiba PR 25.45 49.24 43 Male 3rd molar -5.36 -37.09
rd
42 Curitiba PR 25.45 49.24 40 Male 3 molar -5.33 -37.09

14
Individual Declared City Biological Tooth
Statea Lat.b Long.c Age δ18OC (‰) δ18ORCWIP (‰)
Sample of Birth Sex type
43 Curitiba PR 25.45 49.24 36 Male 3rd molar -5.13 -37.09
rd
44 Curitiba PR 25.45 49.24 33 Female 3 molar -5.97 -37.09
45 Curitiba PR 25.45 49.24 17 Female 3rd molar -5.60 -37.09
rd
46 Curitiba PR 25.45 49.24 41 Female 3 molar -5.52 -37.09
rd
47 Curitiba PR 25.45 49.24 22 Female 3 molar -5.61 -37.09
rd
48 Curitiba PR 25.45 49.24 21 Female 3 molar -5.32 -37.09
rd
49 Curitiba PR 25.45 49.24 25 Female 3 molar -5.35 -37.09
rd
50 Curitiba PR 25.45 49.24 22 Female 3 molar -5.17 -37.09
rd
51 Curitiba PR 25.45 49.24 21 Female 3 molar -5.58 -37.09
rd
52 Curitiba PR 25.45 49.24 23 Female 3 molar -5.17 -37.09
rd

of
53 Curitiba PR 25.45 49.24 24 Female 3 molar -5.96 -37.09
rd
54 Curitiba PR 25.45 49.24 22 Female 3 molar -5.12 -37.09
rd
55 Curitiba PR 25.45 49.24 39 Female 3 molar -5.44 -37.09

ro
rd
56 Curitiba PR 25.45 49.24 30 Male 3 molar -5.42 -37.09
rd
57 Curitiba PR 25.45 49.24 19 Female 3 molar -4.84 -37.09
58
59
Goiorê
Piraquara
PR
PR
24.19 53.03
25.44 49.06
22
27
-p
Female
Female
rd
3 molar
3rd molar
Central
-5.31
-5.73
-35.66
-37.04
re
60 Piraquara PR 25.44 49.06 40 Male -5.74
Incisor -37.04
61 Ivaiporã PR 24.28 51.63 23 Male 3rd molar -6.37 -35.83
lP

rd
62 Lapa PR 25.77 49.87 32 Female 3 molar -6.13 -37.28
rd
63 Lapa PR 25.77 49.87 21 Male 3 molar -5.31 -37.28
64 Londrina PR 23.51 51.11 26 Female 3rd molar -5.43 -35.62
rd
a

65 Paranaguá PR 25.51 48.52 18 Male 3 molar -5.29 -34.64

rd
66 Paranaguá PR 25.51 48.52 19 Male 3 molar -5.61 -34.64
rn

rd
67 Pato Branco PR 26.17 52.66 21 Female 3 molar -5.37 -36.50
rd
68 Pato Branco PR 26.17 52.66 18 Female 3 molar -5.59 -36.50
u

Lateral
69 Pinhais PR 25.44 49.19 59 Female -5.51
Incisor -37.12
Jo

70 Pinhais PR 25.44 49.19 52 Female 3rd molar -5.64 -37.12

71 Pinhais PR 25.44 49.19 23 Male 3rd molar -5.29 -37.12
rd
72 Pinhais PR 25.44 49.19 19 Female 3 molar -5.43 -37.12
rd
73 Pinhais PR 25.44 49.19 19 Female 3 molar -5.51 -37.12
rd
74 Ponta Grossa PR 25.10 50.15 26 Female 3 molar -5.18 -36.98
rd
75 Ponta Grossa PR 25.10 50.15 51 Male 3 molar -6.52 -36.98
76 Ponta Grossa PR 25.10 50.15 - - - -5.96 -36.98
77 Ponta Grossa PR 25.10 50.15 - - - -6.08 -36.98
78 Ponta Grossa PR 25.10 50.15 - - - -5.93 -36.98
79 Ponta Grossa PR 25.10 50.15 - - - -5.22 -36.98
80 Ponta Grossa PR 25.10 50.15 - - - -5.67 -36.98
81 Ponta Grossa PR 25.10 50.15 - - - -6.20 -36.98
82 Ponta Grossa PR 25.10 50.15 - - - -6.23 -36.98
83 Ponta Grossa PR 25.10 50.15 - - - -5.14 -36.98

15
Individual Declared City Biological Tooth
Statea Lat.b Long.c Age δ18OC (‰) δ18ORCWIP (‰)
Sample of Birth Sex type
84 Ponta Grossa PR 25.10 50.15 - - - -6.24 -36.98
85 Ponta Grossa PR 25.10 50.15 - - - -5.74 -36.98
86 Ponta Grossa PR 25.10 50.15 - - - -5.60 -36.98
87 Ponta Grossa PR 25.10 50.15 - - - -5.06 -36.98
88 Ponta Grossa PR 25.10 50.15 - - - -5.27 -36.98
89 Ponta Grossa PR 25.10 50.15 - - - -4.81 -36.98
90 Ponta Grossa PR 25.10 50.15 - - - -5.73 -36.98
91 Ponta Grossa PR 25.10 50.15 - - - -5.35 -36.98
92 Ponta Grossa PR 25.10 50.15 - - - -5.33 -36.98
93 Ponta Grossa PR 25.10 50.15 - - - -5.28 -36.98

of
94 Ponta Grossa PR 25.10 50.15 - - - -5.96 -36.98
95 Ponta Grossa PR 25.10 50.15 - - - -5.81 -36.98
96 Ponta Grossa PR 25.10 50.15 - - - -5.03 -36.98

ro
97 Ponta Grossa PR 25.10 50.15 - - - -5.74 -36.98
98 Ponta Grossa PR 25.10 50.15 - - - -5.26 -36.98
99
100
Ponta Grossa
Ponta Grossa
PR
PR
25.10 50.15
25.10 50.15
-
-
-p-
-
-
-
-5.90
-5.89
-36.98
-36.98
re
101 Ponta Grossa PR 25.10 50.15 - - - -5.53 -36.98
102 Ponta Grossa PR 25.10 50.15 - - - -5.77 -36.98
103 Ponta Grossa PR 25.10 50.15 - - - -5.62 -36.98
lP

Central
104 Renascença PR 26.22 52.95 21 Male -5.25 -36.60
Incisor
Rio Branco do Central
105 PR 25.19 49.31 50 Female -6.59
Sul Incisor -36.94
a

Central
106 Rio Negro PR 26.09 49.68 20 Male -5.74
Incisor -37.04
rn

Santa Isabel Central

107 PR 25.78 53.42 28 Female -5.71
do Oeste Incisor -35.90
Santana do Central
u

108 PR 23.75 49.63 22 Female -6.94

Itararé Incisor -35.54
São José dos Central
Jo

109 PR 25.53 49.20 30 Female -6.14

Pinhais Incisor -37.13
São José dos 1st
110 PR 25.53 49.20 56 Male -5.49 -37.13
Pinhais premolar
São José dos
111 PR 25.53 49.20 42 Male 3rd molar -5.87 -37.13
Pinhais
rd
112 Umuarama PR 23.68 53.41 28 Female 3 molar -5.41 -35.22
União da
113 PR 26.11 51.07 49 Female 3rd molar -5.21
Vitória -37.31
União da
114 PR 26.11 51.07 34 Male 3rd molar -5.90 -37.31
Vitória
Central
115 Itaiópolis SC 26.34 49.91 60 Male -5.35 -37.24
Incisor
116 Joinville SC 26.28 48.84 27 Female 3rd molar -5.52 -34.68
117 Joinville SC 26.28 48.84 27 Female 3rd molar -5.57 -34.68
rd
118 Lages SC 27.81 50.33 30 Female 3 molar -5.51 -37.57
rd
119 Salto Veloso SC 26.91 51.41 35 Female 3 molar -4.73 -37.65

16
5. References
1. INTERPOL. Disaster Victim Identification Guide. (2018).
2. Meier-Augenstein, W. & Fraser, I. Forensic isotope analysis leads to identification of a
mutilated murder victim. Sci. Justice 48, 153–159 (2008).
3. Font, L., Jonker, G., Van Aalderen, P. A., Schiltmans, E. F. & Davies, G. R.
Provenancing of unidentified World War II casualties: Application of strontium and
oxygen isotope analysis in tooth enamel. Sci. Justice 55, 10–17 (2015).
4. Font, L., Van Der Peijl, G., Van Leuwen, C., Van Wetten, I. & Davies, G. R.
Identification of the geographical place of origin of an unidentified individual by multi-
isotope analysis. Sci. Justice 55, 34–42 (2015).
5. Lehn, C., Rossmann, A. & Graw, M. Provenancing of unidentified corpses by stable
isotope techniques - presentation of case studies. Sci. Justice 55, 72–88 (2015).

of
6. Katzenberg, M. A. & Krouse, H. R. Application of Stable Isotope Variation in Human
Tissues to Problems in Identification. Can. Soc. Forensic Sci. J. (1989).
doi:10.1080/00085030.1989.10757414

ro
7. Fraser, I., Meier-Augenstein, W. & Kalin, R. M. The role of stable isotopes in human
identification: A longitudinal study into the variability of isotopic signals in human hair
and nails. Rapid Commun. Mass Spectrom. 20, 1109–1116 (2006).
8. -p
Price, T. D., Manzanilla, L. & Middleton, W. D. Immigration and the ancient city of
Teotihuacan in Mexico: A study using strontium isotope ratios in human bone and
re
teeth. J. Archaeol. Sci. 27, 903–913 (2000).
9. Caruso, S. et al. The process of mineralisation in the development of human tooth.
Eur. J. Paediatr. Dent. 17, 322–326 (2016).
lP

10. Smith, C. E. Cellular and chemical events during enamel maturation. Crit. Rev. Oral
Biol. Med. 9, 128–161 (1998).
11. Selwitz, R. H., Ismail, A. I. & Pitts, N. B. Dental Caries. Lancet 369, 51–59 (2007).
a

12. Li, X., Wang, J., Joiner, A. & Chang, J. The remineralisation of enamel: A review of
the literature. J. Dent. 42, S12–S20 (2014).
rn

13. Daux, V. et al. Oxygen isotope fractionation between human phosphate and water
revisited. J. Hum. Evol. (2008). doi:10.1016/j.jhevol.2008.06.006
u

14. Margolis, H. C., Kwak, S. Y. & Yamazaki, H. Role of mineralization inhibitors in the
regulation of hard tissue biomineralization: Relevance to initial enamel formation and
Jo

maturation. Front. Physiol. 5, 1–10 (2014).

15. Wright, L. E. & Schwarcz, H. P. Stable carbon and oxygen isotopes in human tooth
enamel: Identifying breastfeeding and weaning in prehistory. Am. J. Phys. Anthropol.
(1998). doi:10.1002/(SICI)1096-8644(199805)106:1<1::AID-AJPA1>3.0.CO;2-W
16. Fricke, H. C. & O’Neil, J. R. Inter- and intra-tooth variation in the oxygen isotope
composition of mammalian tooth enamel phosphate: Implications for
palaeoclimatological and palaeobiological research. Palaeogeogr. Palaeoclimatol.
Palaeoecol. (1996). doi:10.1016/S0031-0182(96)00072-7
17. Daniel Bryant, J., Luz, B. & Froelich, P. N. Oxygen isotopic composition of fossil
horse tooth phosphate as a record of continental paleoclimate. Palaeogeogr.
Palaeoclimatol. Palaeoecol. 107, 303–316 (1994).
18. Zazzo, A., Mariotti, A., Lécuyer, C. & Heintz, E. Intra-tooth isotope variations in late
Miocene bovid enamel from Afghanistan: Paleobiological, taphonomic, and climatic
implications. Palaeogeogr. Palaeoclimatol. Palaeoecol. 186, 145–161 (2002).
19. Kohn, M. J., Schoeninger, M. J. & Barker, W. W. Altered states: Effects of diagenesis
on fossil tooth chemistry. Geochim. Cosmochim. Acta 63, 2737–2747 (1999).

17
20. Kubota, T., Nakamura, A., Toyoura, K. & Matsunaga, K. The effect of chemical
potential on the thermodynamic stability of carbonate ions in hydroxyapatite. Acta
Biomater. 10, 3716–3722 (2014).
21. Stuart-Williams, H. L. Q. & Schwarcz, H. P. Oxygen isotopic determination of climatic
variation using phosphate from beaver bone, tooth enamel, and dentine. Geochim.
Cosmochim. Acta 61, 2539–2550 (1997).
22. Luz, B., Kolodny, Y. & Horowitz, M. Fractionation of oxygen isotopes between
mammalian bone-phosphate and environmental drinking water. Geochimica et
Cosmochimica Acta (1984). doi:10.1016/0016-7037(84)90338-7
23. Chenery, C. A., Pashley, V., Lamb, A. L., Sloane, H. J. & Evans, J. A. The oxygen
isotope relationship between the phosphate and structural carbonate fractions of
human bioapatite. Rapid Commun. Mass Spectrom. 26, 309–319 (2012).
24. Longinelli, A. Oxygen isotopes in mammal bone phosphate : A new tool for
paleohydrological and paleoclimatological research ? 48, 385–390 (1984).

of
25. Levinson, A. A., Luz, B. & Kolodny, Y. Variations in oxygen isotopic compositions of
human teeth and urinary stones. Appl. Geochemistry 2, 367–371 (1987).

ro
26. Pollard, A. M., Pellegrini, M. & Lee-Thorp, J. A. Technical note: Some observations
on the conversion of dental enamel δ18Op values to δ18Ow determine human mobility.
Am. J. Phys. Anthropol. 145, 499–504 (2011).
27.
of Stable Isotope Analysis.
-p
Meier-Augenstein, W. Stable Isotope Forensics: Methods and Forensic Applications
(John Wiley & Sons Ltd, 2018).
re
doi:10.1002/9781119080190
28. Ehleringer, J. R. et al. A Framework for the Incorporation of Isotopes and Isoscapes
in Geospatial Forensic Investigations. in Isoscapes (ed. West J., Bowen G., Dawson
lP

T., T. K.) 357–387 (Springer, 2010). doi:https://doi.org/10.1007/978-90-481-3354-

3_17
29. Bowen, G. J. & Revenaugh, J. Interpolating the isotopic composition of modern
a

meteoric precipitation. Water Resour. Res. 39, 1–13 (2003).

30. S. Terzer, L. I. Wassenaar, L. J. Araguás-Araguás, and P. K. A. Global isoscapes for
rn

δ18O and δ2H in precipitation: improved prediction using regionalized climatic

regression models. Hydrol. Earth Syst. Sci. Discuss. 10, 7351–7393 (2013).
31. West, J. B., Sobek, A. & Ehleringer, J. R. A simplified GIS approach to modeling
u

global leaf water isoscapes. PLoS One 3 (2008).

Jo

32. Bowen, G. J. Isoscapes: Spatial pattern in isotopic biogeochemistry. Annu. Rev.

Earth Planet. Sci. 38, 161–187 (2010).
33. Breitenbach, S. F. M. & Bernasconi, S. M. Carbon and oxygen isotope analysis of
small carbonate samples (20 to 100 μg) with a GasBench II preparation device. Rapid
Commun. Mass Spectrom. 25, 1910–1914 (2011).
34. Benson, S., Lennard, C., Maynard, P. & Roux, C. Forensic applications of isotope
ratio mass spectrometry - A review. Forensic Sci. Int. 157, 1–22 (2006).
35. Coplen, T. B. New guidelines for reporting stable hydrogen, carbon, and oxygen
isotope-ratio data. Geochim. Cosmochim. Acta 60, 3359–3360 (1996).
36. Craig, H. Isotopic standards for carbon and oxygen and correction factors for mass-
spectrometric analysis of carbon dioxide. Geochim. Cosmochim. Acta 12, 133–149
(1957).
37. Szpak, P., Metcalfe, J. Z. & Macdonald, R. A. Best practices for calibrating and
reporting stable isotope measurements in archaeology. J. Archaeol. Sci. Reports 13,
609–616 (2017).
38. Wright, L. E. & Schwarcz, H. P. Correspondence Between Stable Carbon, Oxygen

18
 and Nitrogen Isotopes in Human Tooth Enamel and Dentine: Infant Diets at
Kaminaljuyú. J. Archaeol. Sci. 26, 1159–1170 (1999).
39. João Paulo Sena-Souza, F. J. V. C. ; G. B. N. Background and the use of isoscapes
in the Brazilian context: essential tool for isotope data interpretation and natural
resource management. Rev. Ambiente & Água 14, 1–27 (2019).
40. Coplen, T. B. Normalization of oxygen and hydrogen isotope data. Chem. Geol. Isot.
Geosci. Sect. 72, 293–297 (1988).
41. Giustini, F., Brilli, M. & Patera, A. Mapping oxygen stable isotopes of precipitation in
Italy. J. Hydrol. Reg. Stud. 8, 162–181 (2016).
42. Hatvani, I. G., Erdélyi, D., Vreča, P. & Kern, Z. Analysis of the spatial distribution of
stable oxygen and hydrogen isotopes in precipitation across the Iberian Peninsula.
Water (Switzerland) 12, (2020).
43. Bowen, G. J. & Wilkinson, B. Spatial distribution of δ18O in meteoric precipitation.

of
Geology 30, 315–318 (2002).
44. Pebesma, E. J. Multivariable geostatistics in S: the gstat package. 30, 683–691
(2004).

ro
45. Davis, B. M. Uses and abuses of cross-validation in geostatistics. Math. Geol. 19,
241–248 (1987).
46.
47.
48.
-p
Pearce, F. Rivers in the sky. New Sci. 244, 40–43 (2019).
Pearce, F. Weather makers. Science (80-. ). 368, 1302–1305 (2020).
Zemp, D. C. et al. On the importance of cascading moisture recycling in South
re
America. Atmos. Chem. Phys. 14, 13337–13359 (2014).
49. Da Silva, V. D. P. R. On climate variability in Northeast of Brazil. J. Arid Environ. 58,
575–596 (2004).
lP

50. Malone, B. P., Odgers, N. P., Stockmann, U., Minasny, B. & Mcbratney, A. B. Digital
Mapping of Soil Classes and Continuous Soil Properties. in Pedometrics. Progress in
Soil Science (ed. McBratney A., Minasny B., S. U. (eds)) 373– 413 (Springer, 2018).
a

doi:10.1007/978-3-319-63439-5.
rn

Acknowledgments
The authors would like to thank the financial support from the Microbial Project
u

(UFPR/PETROBRAS/ANP n° 2016/00141-1). We are also indebted to Idalina Marly da Luz

Jo

and Ricardo Luiz Vieira, from the tooth bank of the Faculty of Odontology of the Federal
University of Paraná and to the dentist Naiana Mello Cançado for the tooth samples. We would
like also to thank Madalina Jaggi, from ETHZ, for her generous support in the Stable Isotope
Laboratory. Finally, we would like to thank Prof. Dr. Luiz Antonio Martinelli (CENA/USP) and
Prof. Dr. Roberto Ventura Santos (UnB) for their critical reviews of this manuscript.

1.1 Sample CRediT author statement

Ricardo de Oliveira Mascarenhas: Conceptualization, Methodology, Sampling, Analysis,

Writing, Editing;
João Paulo Sena-Souza: Conceptualization, Methodology, Analysis, Writing, Editing;
Stefano M. Bernasconi: Supervision, Sponsoring, Writing and Reviewing;

19
Judith A. McKenzie: Supervision, Writing and Reviewing;
Crisógono Vasconcelos: Supervision, Sponsoring and Reviewing;
Taís Ribeiro Muniz: Sampling, Methodology, Reviewing;
Matheus Pereira Nogueira e Silva: Methodology, Editing, Writing and Reviewing;
Fábio Augusto da Silva Salvador: Supervision, Sponsoring and Reviewing;
Anelize Manuela Bahniuk Rumbelsperger: Supervision, Sponsoring Conceptualization,
Methodology, Reviewing.
*
Corresponding author. Tel.: +55 41 32567631
E-mail adress: mascarenhas.rom@pf.gov.br (R. O. Mascarenhas)

Highlights

of
 Isoscape based on tooth enamel for the Brazilian population
 δ18O isoscape based on tooth enamel for human identification

ro
 Isotopic database based on tooth enamel for the modern population in Brazil
 Isoscape built using the regression-kriging approach

-p
re
a lP
u rn
Jo

20