Next Article in Journal
What Can Neuromarketing Tell Us about Food Packaging?
Next Article in Special Issue
n-Alkanes and n-Alkenes in Virgin Olive Oil from Calabria (South Italy): The Effects of Cultivar and Harvest Date
Previous Article in Journal
Sulfonic Acid-Functionalized SBA-15 as Strong Cation-Exchange Sorbent for Solid-Phase Extraction of Atropine and Scopolamine in Gluten-Free Grains and Flours
Previous Article in Special Issue
1H NMR Spectroscopy to Characterize Italian Extra Virgin Olive Oil Blends, Using Statistical Models and Databases Based on Monocultivar Reference Oils
Open AccessArticle

Combination of Stable Isotope Analysis and Chemometrics to Discriminate Geoclimatically and Temporally the Virgin Olive Oils from Three Mediterranean Countries

1
MED—Mediterranean Institute for Agriculture, Environment and Development, Universidade de Évora, Pólo da Mitra, Ap. 94, 7006-554 Évora, Portugal
2
HERCULES, Universidade de Évora, Palácio do Vimioso, 7000-089 Évora, Portugal
3
Departamento de Fitotecnia, Escola de Ciências e Tecnologia, Universidade de Évora, Núcleo da Mitra, Ap. 94, 7006-554 Évora, Portugal
4
Departamento de Química, Escola de Ciências e Tecnologia, Universidade de Évora, Rua Romão Ramalho, 59, 7000-671 Évora, Portugal
*
Author to whom correspondence should be addressed.
Foods 2020, 9(12), 1855; https://doi.org/10.3390/foods9121855
Received: 26 October 2020 / Revised: 4 December 2020 / Accepted: 9 December 2020 / Published: 12 December 2020

Abstract

The knowledge of the isotopic composition of virgin olive oil (VOO) allows the evaluation of authenticity and geographical origin, being an important tool against fraud. This study aimed to assess if VOOs produced in three Mediterranean regions could be discriminated on the basis of multivariate statistical analysis of geoclimatic and isotopic data. A total of 138 geo-referenced VOO samples from Portugal, France and Turkey from two different cultivation years were collected. The isotopic composition (δ13C, δ2H and δ18O) of VOOs was obtained using an elemental analyzer coupled to an isotope ratio mass spectrometer (EA-IRMS). One-way analysis of variance for δ13C, δ2H and δ18O showed some significant differences either between crop years or geoclimatic conditions. Based on multiple regression analyses using meteorological and geographical parameters, a meteoric water line for olive oil from Portugal, France and Turkey, in two harvest years, were created to assess the impact of climate change on their δ2H and δ18O values. Principal component analysis and Linear Discriminant Analysis, used to sort samples according to geoclimatic origin, performed best for French and Portuguese olive oils. In light of the results, multivariate isotopic analysis of VOO samples may discriminate not only between geoclimatic regions but also among cultivation years.
Keywords: olive oil; meteoric water line; EA-IRMS; stable isotopes olive oil; meteoric water line; EA-IRMS; stable isotopes

1. Introduction

Traceability of olive oils is still a popular research topic, addressed in many different ways, including exploring the benefits of different analytical tools due to the increased demand for an effective analytical methodology that can verify the geographic origin of olive oils. As claim of origin is becoming more important, so is the need for new tools to establish it. Traditional chromatography techniques (gas or liquid) coupled to mass spectrometry [1,2,3] or UV detection [4], or the “more environmentally friendly” analytical methods based on spectroscopy [5,6] have been used for this purpose.
However, olive oil stable isotope ratio analysis (SIRA), either in bulk, using Isotope Ratio Mass Spectrometry (IRMS) coupled to an elemental analyzer (EA), or in a class of compounds using IRMS coupled with a gas chromatographer (GC), is nowadays used more often to achieve separation among olive oils from different geographical origins [7,8,9,10,11,12]. The SIRA is based on the fact that the stable isotope content of bio-elements reflects the geo-climatic characteristics of the area of production [13] and was initially performed in bulk analysis. In the case of olive oil, bulk δ13C, δ18O and δ2H determinations allowed the differentiation among different geographic origins, at least in some cases. With innovations in this analytical technique that now allow the study of stale isotopes in specific compounds, like fatty acids, or volatile compounds, more detailed information can be retrieved allowing improvement in geographical discrimination.
Oxygen and hydrogen isotope ratios (δ18O and δ2H), as a natural tracer, have been effectively used in diagnosing water sources [14], as well as a marker of authenticity and origin of food or beverages. Particularly, these isotopes have been employed to trace the authenticity and geographical origin of European extra-virgin olive oils [12,15]. The stable isotope compositions of oxygen and hydrogen in soil and ground waters are controlled by local precipitation in the recharge and catchment areas, depending on various geoclimatic factors (e.g., temperature, distance from the sea, amount of atmospheric precipitation, etc.) [16,17]. Therefore, the variations in δ2H and δ18O values in plants and their products, through their relationship with humidity-precipitation and temperature conditions during plant growth, may allow the tracing of the geographic origin of vegetable oils [18,19,20,21].
Bearing this in mind, our study aimed to assess if VOOs from three different Mediterranean countries (Portugal, France and Turkey) could be discriminated on the basis of multivariate statistical analysis of geoclimatic and isotopic data (δ13C, δ2H and δ18O) obtained from bulk analysis of 138 varietal VOOs.

2. Materials and Methods

2.1. Samples and Experimental Design

A total of 138 varietal VOO (Olea europaea L.) samples from three different Mediterranean countries, Portugal (27 in 2016 and 39 in 2017), France (25 in 2016 and 25 in 2017) and Turkey (10 in 2016 and 11 in 2017), were collected from November to December in 2016 and in 2017. For the majority of the selected VOO samples, approximately 5 kg of each different olive variety was processed separately in an Abencor® system (Seville, Spain) within 24 h after harvesting. Fruits were crushed with a hammer mill and the olive paste was malaxed at 25 °C for 30 min in an olive paste mixer. Finally, the olive oil was separated by centrifugation. Other samples were taken directly from olive oil mills to avoid possible undeclared mixtures with olive oils from other cultivars and geographical origins, before bottling. They were stored in dark-brown glass bottles at 20 °C in the dark. The VOO samples were perfectly geo-referenced, obtaining data for latitude (UTM), longitude (UTM), altitude (m.a.s.l.), sea distance (km), mean annual temperature (°C) and mean annual rainfall (mm) (Table S1).

2.2. Stable Isotope Analysis

The carbon, hydrogen and oxygen isotope composition ratios (13C/12C, 2H/1H and 18O/16O, respectively) of VOO samples were determined by elemental analysis/isotope ratio mass spectrometry (EA/IRMS) following the procedure described by Jiménez-Morillo et al. [12]. Briefly, bulk isotopic signature of light elements (δ13C, δ2H and δ18O) was analysed using a Flash 2000 HT with two reactors (combustion and pyrolysis, respectively). The elemental analyzer is coupled by a ConFlo IV interface unit to a continuous flow Delta V Advantage isotope ratio mass spectrometer (IRMS) (Thermo Scientific, Bremen, Germany). Isotopic ratios “δ” are reported in milli Urey (mUr) from appropriate standards recognized by the International Atomic Energy Agency (IAEA). The standard deviation of bulk δ13C, δ2H and δ18O were ± 0.1, 1.0 and 0.5 mUr, respectively. Each sample was measured in duplicate (n = 2).

2.3. Statistical Analysis

Multivariate statistical analysis of data was carried out with Statgraphics Centurion XV software (Statgraphics Technologies, Inc, The Plains, VA, USA). Principal component analysis (PCA) was used for simultaneous ordination of different geographic and climatic variables and the δ13C, δ2H and δ18O values, illustrating their mutual relationships. Multiple linear regression (MLR) was applied to generate an isotopic forecasting model by harvest year, using geographic and climatic factors as independent variables. The validation of the model was confirmed by the production of spurious models computed from the fully randomized dependent variables (isotopic composition of hydrogen, oxygen and carbon) from the 138 VOO samples from three different Mediterranean countries in two harvest years. The alternative models displayed a poor correlation (p > 0.05) with geoclimatic variables (data not shown). Therefore, the well-correlated models generated with experimental data were not overfitted [12,22]. Linear discriminant analysis was also carried out to evaluate whether olive oil samples from the three countries could be easily separated based on isotopic data. The statistical significance of each discriminant function was evaluated by Wilk’s lambda [23] and the predictive ability of the LDA model was evaluated by leave-one-out cross validation.

3. Results and Discussion

3.1. Stable Isotope Analysis of VOO Samples

Table S1 collects the carbon, hydrogen and oxygen isotope values of VOO samples from production areas in France, Portugal and Turkey in two harvest years. Concerning δ13C value, the Portuguese VOO samples ranged between −31.8 and −27.9 mUr for 2016, and between −31.3 and −28.1 mUr for 2017. In the case of French samples, the δ13C composition ranged between −31.5 and −27.6 mUr for 2016, and between −31.6 and −27.1 mUr for 2017. Finally, the VOO samples from Turkey showed δ13C value ranging between −30.6 and −27.8 mUr, and −30.6 and −27.7 mUr for 2016 and 2017, respectively. These isotopic values are in accordance with those reported by Gumus et al. [9] for olive oils from western Turkey. The δ13C values of the VOO samples from the three Mediterranean countries were within the typical value of C3 photosystem plants [24]. However, as observed, there were significant differences between location and harvest years. One-way analysis of variance (one-way ANOVA) using δ13C composition (Figure 1a) as independent variable (IV) indicated that Portuguese and Turkish VOOs are significantly (p < 0.05) different between themselves, but not between respective cultivation years. The French samples were significantly different (p < 0.05) between cultivation years. Traditionally, it has been accepted that δ13C composition of food-stuff depends on the localization of the production area, i.e., on the geomorphological and climatic conditions, such as temperature, rainfall, sea distance, longitude and latitude [12]. These factors influence leaf stomata opening and hence the photosynthesis yield [25], as dry and warm conditions may produce a restriction in the opening time of stomas, limiting the admission of atmospheric CO2 to the leaf, causing an increase in δ13C value. The existence of no significant differences between harvest year in Portugal and Turkey may be related to the similitude in mean annual temperature between 2016 and 2017 (Table S2). However, in French regions a significant difference between harvest year was observed, which was also displayed in the temperature values (Table S2).
On the other hand, hydrogen and oxygen isotope composition of VOO samples (Table S1) showed particular trends for each one. In the case of δ18O, the values of EVOOs from Portugal ranged between 21.2 and 25.9 mUr for 2016 and between 23.4 and 28.6 mUr for 2017, while from France the value ranged between 19.9 and 26.4 mUr for 2016 and between 20.1 and 23.3 mUr for 2017. For Turkish VOO samples, the δ18O composition was between 22.1 and 27.2 mUr for 2016, and between 21.5 and 28.1 mUr for 2017. The statistical analysis of δ18O values (Figure 1b) displayed significant (p < 0.05) differences amongst cultivation years for Portuguese and French samples. In addition, the VOOs cultivated in 2017 were considerably different between the three analyzed Mediterranean countries, while for 2016 only Turkish VOOs were significantly different. With respect to δ2H composition in 2016 and 2017 (Table S1), the values ranged: (i) between −146.6 and −127.1 mUr, and between −155.6 and −122.0 mUr for Portugal; (ii) between −186.7 and −141.8 mUr, and between −178.2 and −127.1 mUr for France; and (iii) between −159.0 and −125.6 mUr, and between −178.1 and −135.1 mUr for Turkey. One-way ANOVA analysis of δ2H values (Figure 1c) showed that only the samples from Turkey were significantly (p < 0.05) different according to the cultivation year. In contrast, the samples from Portugal and France are significantly (p < 0.05) different from each other, which may be directly related to a quite different climate, mainly temperature and precipitation (Table S2). With respect to Turkey, the significant differences between harvest years may be due to specific climatological events that could have altered the isotopic composition of hydrogen. This hypothesis may be supported by the fact that the δ2H and δ18O composition of foodstuffs are significantly correlated to (1) the local climatic conditions (temperature and precipitation) [12,16,26]; (2) soil and plant evapotranspiration [27,28]; (3) air humidity; and (4) fractionation associated with the biosynthesis of plant tissues [29,30]. Therefore, the differences observed between countries and harvest years may be related to the above mentioned factors. However, the existence of irrigation activities may also modify the isotopic composition of hydrogen and oxygen [31,32].

3.2. Meteoric Water Line

Figure 2 shows the correlation between δ2H and δ18O of VOO samples from the three Mediterranean countries in two harvest years. In agreement with other researchers, the H and O isotope values of VOO samples are directly correlated to the meteoric water uptake by olive plants, so it has been observed that the H isotope value of olive oil is about 100 times lower than the value of rainwater, while the O isotope value is about 27 times higher than that of rainwater [30,33]. Therefore, in this work, these criteria were used to generate the local meteoric water line (LMWL) for each location and harvest year (Figure 2). In addition, the correlation lines were compared with the global meteoric water line (GMWL). The GMWL was generated using the δ2H and δ18O values obtained from IAEA WISER (Water Isotope System for data analysis visualization and Electronic Retrieval database) [34]. The slope variation of each LMWL with the GMWL allows assessment of the alteration of climatic conditions in a geographical location and year [35,36]. In general, the LMWL for the Mediterranean countries studied displayed a depletion of slope in comparison with the GMWL, which is related to rise in temperature and reduction of precipitation [35]. The current climatic change and global warming scenario is dominated by the rising of mean temperature in the Mediterranean basin together with the reduction of the number of rainfall episodes and the amount of water [37,38]. A more detailed analysis of δ2H and δ18O by country and year shows that in 2016 all studied countries showed the same depletion slope, which may be correlated to a rise in drought conditions. However, in 2017, the trends were different. In the case of French samples in 2017, a slope parallel to the GMWL is displayed, while in 2016 its slope displays a conspicuous depletion. The average value of temperature and precipitation (Table S2) in France showed a significant (p < 0.05) depletion of 39% in rainfall in 2017. A possible explanation may be the existence of an irrigation process, since, in this year, the cultivation area suffered a significant reduction in rainfall (over 30%, Table S2). Irrigation may alter the natural δ2H and δ18O value [39,40]. Therefore, the non-alteration of the 2017 slope in France may be due to the anthropogenic factor. Displacement may be seen on the y-axis (δ18O composition), which may be related to the isotope irrigation water value. In the case of Portugal, slopes in different harvest years display the same value. However, in 2017, the LMWL showed a shift in δ18O composition, i.e., in the y-intercept. This may be supported by a significant reduction in rainfall of 31% (Table S2). On the other hand, Turkey shows a slight change in slope in 2017. The analysis of precipitation data (Table S2) showed a rise in rainfall of 21%, but this is not significant (p > 0.05). The good correlation of LMWL (correlation between treated δ2H and δ18O of VOOs), as well as their connection to the GMWL, conditioned by inherent geoclimatic factors (temperature, latitude, sea distance, rainfall, etc.), confirm the potential use of isotope composition of VOOs for their geoclimatic characterization. Some recent works have explored this possibility in VOO samples produced from the same olive variety and harvest year in different countries, for example, in Italy [19] and Spain [20].

3.3. Chemometric Analysis

Principal component analysis (PCA) was employed on the one hand to identify the possible correlation between independent (stable isotopes of VOO samples) and dependent (geoclimatic parameters of produced areas) variables (Figure 3a), and on the other hand to sort the VOO samples by geographical region and harvest year (Figure 3b). The analysis of the three first components (component 1: 29.35%, component 2: 20.95%, and component 3: 17.37%) can explain up to 67% of the total variance (Table S3). The scatterplot of the loadings of PC-1 vs. PC-2 vs. PC-3 (Figure 3a) showed that δ2H values were directly correlated with temperature, sea distance and altitude, and negatively with latitude, longitude and rainfall. In contrast, δ18O values were correlated with longitude, altitude and temperature, while longitude and rainfall were negatively correlated with oxygen isotope composition. These correlations were previously observed in other articles by the authors. For example, Jiménez-Morillo et al., [11], using for the first time the analytical pyrolysis–compound specific isotope analysis (Py-CSIA), found that temperature and sea distance were directly correlated with the hydrogen isotope composition of specific compounds of EVOO samples from five Mediterranean countries. However, other authors showed inversely correlation between δ18O and sea distance [12]. With respect to carbon isotope (δ13C), this displays direct correlation with longitude and altitude but correlates inversely with rainfall. Camin et al. [41] and Jiménez-Morillo et al. [12] observed the same correlations between carbon isotope composition of olive oil samples and the geoclimatic factors. More rain will cause plant’s stomata to stay open longer, causing more isotopically light CO2 molecules to pass inside the leaf and the δ13C value to drop, relative to plants growing in a more arid environment. The existing correlations observed in Figure 3a were corroborated by a multi linear regression (MLR) approach. The MLR, using the geoclimatic parameters of the VOO producer locals (n = 138) from the studied Mediterranean countries, led to significant (p < 0.05) forecasting models for the carbon, hydrogen and oxygen isotope composition for the harvest year.
Figure 4 plots the observed vs. predicted values of each models for carbon, hydrogen and oxygen in 2016 and in 2017. The model validation was confirmed by a strict criterion based on comparing the MLR cross-validation tests (observed vs. predicted) with the alternative model computed from the fully randomized isotope values of VOO samples studied in this work (data no shown).
This cross-validation method has been previously used in several works [11,12,42]. The MLR analysis allowed us to obtain the Equations (1)–(6) of the prediction model, using automatic backwards variable selection:
δ13C (2016) = −42.24 + 0.24 × Latitude + 0.02 × Longitude − 0.01 × Rainfall + 0.22 × Temperature + 0.01 × Oceanic distance
δ13C (2017) = −39.91 + 0.20 × Latitude + 0.02 × Longitude − 0.01 × Rainfall + 0.10 Temperature + 0.01 × Oceanic distance
δ18O (2016) = 27.61 − 0.68 × Latitude + 0.08 × Longitude + 0.01 × Rainfall − 0.14 × Temperature + 0.02 × Oceanic distance
δ18O (2017) = 55.68 − 0.67 × Latitude − 0.02 × Longitude − 0.01 × Rainfall − 0.28 × Temperature + 0.01 × Oceanic distance
δ2H (2016) = −83.92 − 2.83 × Latitude + 0.03 × Longitude + 0.03 × Altitude + 0.05 × Rainfall+ 1.01 × Temperature + 0.02 × Oceanic distance
δ2H (2017) = −234.01 + 0.70 × Latitude − 0.34 × Longitude + 0.03 × Altitude + 0.03 × Rainfall + 2.15 × Temperature + 0.04 × Oceanic distance
In this way, the new predicting isotope value models (Equations (1)–(6)) will allow forecasting of changes that foodstuffs will undergo in different climate change scenarios, and adapting the criteria for the protection of designations of origin, safeguarding society from possible food fraud. In addition, they could be used to assess the isotopic composition of virgin olive oils under different climatic change scenarios (from the mildest to the most severe). However, it is known that the isotopic forecasting models developed should be corrected every year owing to intrinsic fluctuations in isotopic composition. Therefore, the next step will be the validation and correction of the possible isotopic alterations, generating correction factors.
On the other hand, the 3D scores in the PCA plot of VOO samples (Figure 3b), generated using their unique isotopic composition, displayed three clusters, one for each country. In the case of the French and Turkish cluster, their separation is dominated by the difference in the stable isotope value of hydrogen and oxygen. The French samples displayed higher isotope values than the VOO samples from Turkey (Figure 2b,c). The Portuguese cluster showed an intermediate location in the 3D PCA plot (Figure 3b). These samples show no significant difference with respect to the other two regional VOO samples (Figure 2b,c). Although the samples are well-grouped into three clusters depending on the country, a more detailed study shows that the samples can be separated slightly depending on the harvest year. The most prominent case is the samples from Portugal (Figure 3b). This difference may be directly related to the variation in the δ18O (Figure 2b), as well as the significant difference in rainfall (Table S2). With respect to the other Mediterranean VOO samples, this difference between cultivation years was not so clear.
In order to verify if a better classification of olive oils can be achieved, a Linear Discriminant Analysis (LDA) was also performed. Figure 5 shows the scatter plot for olive oils from the three countries in each year. Two discriminant functions were constructed based on Wilks’ Lambda values, which explained 100% of the variance, 81.2% of the total variance explained by function 1 and 18.8% explained by function 2, for 2016 olive oils; and 95.3% of the total variance explained by function 1 and 4.7% explained by function 2, for 2017 olive oils (Table S5). The Wilks’ Lambda values were, for function 1 and 2, 0.377 and 0.795 with a p-value of 0.000 and 0.001, for 2016 olive oils; and 0.230 and 0.878 with a p value of 0.000 and 0.010, for 2017 olive oils, which implies that there are significant differences among the three regions and the first function has higher discrimination power, since Wilk’s Lambda is lower and p-value is equal to 0. Discriminant functions can be shown as:
Function 1 (2016) = 20.082 − 0.074 δ13C − 0.174 δ18O + 0.125 δ2H
Function 2 (2016) = −0.196 + 0.434 δ13C + 0.593 δ18O − 0.006 δ2H
Function 1 (2017) = −29.207 − 0.535 δ13C +0.733 δ18O + 0.028 δ2H
Function 2 (2017) = 9.266 − 0.497 δ13C − 0.392 δ18O + 0.099 δ2H
Table S5 summarizes the performance of the discriminant models regarding the classification of olive oil samples, and shows that the application of cross-validation (CV) exposes some misclassifications, mainly regarding samples from Turkey. For 2016 samples, 79.0% of the olive oil was correctly classified and this value diminished to 74.2% after cross validation; for 2017 samples, results were better, with 92.9% of the cases correctly classified before and after CV. Olive oils from Portugal performed better than the others with the highest percentage of correctly classified olive oils, according to country of origin.

4. Conclusions

In light of the results, the combination of multivariate isotopic analysis of VOO samples and chemometrics allows discrimination between geoclimatic regions but also among cultivation years. Therefore, it presents a powerful tool to fight against food fraud in the VOO sector. In addition, the design of this work has allowed the generation of isotopic forecasting models using the geoclimatic parameters as predictive factors. Therefore, the interpolation of geoclimatic data (mainly temperature and rainfall) under different climatic change scenarios (from benevolent to drastic) will allow us to assess the isotopic alteration of virgin olive oil in future scenarios.
This displays a great potential for establishing new correction factors to determine the geographical origin (Designation of Origin) of olive oils.

Supplementary Materials

The following are available online at https://www.mdpi.com/2304-8158/9/12/1855/s1, Table S1: Varieties, origin, region, geo-climatic information and stable isotope composition of VOO samples; Table S2: Average and Standard deviation of Temperature and Rainfall in France, Portugal and Turkey in two years (2016 and 2017). Alteration of rainfall amount (%); Table S3: Principal component loading scores; Table S4: Variable standardized canonical discriminant function coefficients along with discriminant functions classification parameters; Table S5: Overall classification performance (count and %) of the LDA model, with and without cross validation (CV), for the three countries differentiation study.

Author Contributions

N.T.J.-M.—Formal analysis, data treatment, writing-original draft preparation; V.P., investigation; R.G.—writing-review and editing; C.B.D.—writing-review and editing; M.J.C.—supervision, funding, statistical analysis, writing—original draft and review. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by National Funds through Foundation for Science and Technology (FCT) through Project “MedOOmics—Mediterranean Extra Virgin Olive Oil Omics: profiling and fingerprinting—Arimnet2/0001/2015”, Project “Por3O—Portuguese Olive Oil Omics for traceability and authenticity—PTDC/AGRPRO/2003/2014 and Project UIDB/05183/2020”. Project “Por3O—PTDC/AGRPRO/2003/2014” was also funding by European Regional Development Fund (FEDER).

Acknowledgments

Pedro N. Jiménez-Morillo is acknowledged for statistical assistance.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Zunin, P.; Boggia, R.; Salvadeo, P.; Evangelisti, F. Geographical traceability of West Liguria extravirgin olive oils by the analysis of volatile terpenoid hydrocarbons. J. Chromatogr. A 2005, 1089, 243–249. [Google Scholar] [CrossRef] [PubMed]
  2. Gil-Solsona, R.; Raro, M.; Sales, C.; Lacalle, L.; Díaz, R.; Ibánez, M.; Beltran, J.; Sancho, J.V.; Hernández, F.J. Metabolomic approach for extra virgin olive oil origin discrimination making use of ultra-high performance liquid chromatography—Quadrupole time-of-flight mass spectrometry. Food Control. 2016, 70, 350–359. [Google Scholar] [CrossRef]
  3. Şişik Oğraş, Ş.; Kaban, G.; Kaya, M. Volatile compounds of olive oils from different geographic regions in Turkey. Int. J. Food Prop. 2018, 22, 1833–1843. [Google Scholar] [CrossRef]
  4. Carranco, N.; Farrés-Cebrián, M.; Saurina, J.; Núñez, O. Authentication and Quantitation of Fraud in Extra Virgin Olive Oils Based on HPLC-UV Fingerprinting and Multivariate Calibration. Foods 2018, 7, 44. [Google Scholar] [CrossRef]
  5. Valli, E.; Bendini, A.; Berardinelli, A.; Ragni, L.; Riccò, B.; Grossi, M.; Toschi, T.G. Rapid and innovative instrumental approaches for quality and authenticity of olive oils. Eur. J. Lipid Sci. Technol. 2016, 118, 1601–1619. [Google Scholar] [CrossRef]
  6. Alves, F.C.G.B.S.; Coqueiro, A.; Março, P.H.; Valderrama, P. Evaluation of olive oils from the Mediterranean region by UV–Vis spectroscopy and Independent Component Analysis. Food Chem. 2019, 273, 124–129. [Google Scholar] [CrossRef]
  7. Camin, F.; Pavone, A.; Bontempo, L.; Wehrens, R.; Paolini, M.; Faberi, A.; Marianella, R.M.; Capitani, D.; Vista, S.; Mannina, L. The use of IRMS, 1H NMR and chemical analysis to characterise Italian and imported Tunisian olive oils. Food Chem. 2016, 196, 98–105. [Google Scholar] [CrossRef]
  8. Paolini, M.; Bontempo, L.; Camin, F. Compound-specific δ13C and δ2H analysis of olive oil fatty acids. Talanta 2017, 174, 38–43. [Google Scholar] [CrossRef]
  9. Gumus, Z.P.; Celenk, V.U.; Tekin, S.; Yurdakul, O.; Ertas, H. Determination of trace elements and stable carbon isotope ratios in virgin olive oils from Western Turkey to authenticate geographical origin with a chemometric approach. Eur. Food Res. Technol. 2017, 243, 1719–1727. [Google Scholar] [CrossRef]
  10. Bontempo, L.; Paolini, M.; Franceschi, P.; Ziller, L.; García-González, D.L.; Camin, F. Characterisation and attempted differentiation of European and extra-European olive oils using stable isotope ratio analysis. Food Chem. 2019, 276, 782–789. [Google Scholar] [CrossRef]
  11. Jiménez-Morillo, N.T.; Cabrita, M.J.; Barrocas-Dias, C.; González-Vila, F.J.; González-Pérez, J.A. Pyrolysis-compound-specific hydrogen isotope analysis (δ2H Py-CSIA) of Mediterranean olive oils. Food Control. 2020, 110, 107023. [Google Scholar] [CrossRef]
  12. Jiménez-Morillo, N.T.; Palma, V.; Garcia, R.; Pereira, J.A.; Barrocas-Dias, C.; Cabrita, M.J. Multivariate geostatistical analysis of stable isotopes in Portuguese varietal extra virgin olive oils. Microchem. J. 2020, 157, 105044. [Google Scholar] [CrossRef]
  13. Laursen, K.H.; Bontempo, L.; Camin, F.; Roßmann, A. Advances in isotopic analysis for food authenticity testing. In Advances in Food Authenticity Testing; Downey, G., Ed.; Woodhead Publishing: Cambridge, UK, 2016; pp. 227–252. [Google Scholar]
  14. Gat, J.R.; Gonfiantini, R. Stable Isotope Hydrology: Deuterium and Oxygen-18 in the Water Cycle; IAEA Technical Report Series 210; IAEA: Vienna, Austria, 1981; p. 337. [Google Scholar]
  15. Chiocchini, F.; Portarena, S.; Ciolfi, M.; Brugnoli, E.; Lauteri, M. Isoscapes of carbon and oxygen stable isotope compositions in tracing authenticity and geographical origin of Italian extra-virgin olive oils. Food Chem. 2016, 202, 291–301. [Google Scholar] [CrossRef] [PubMed]
  16. Clark, I.; Fritz, P. Environmental Isotopes in Hydrogeology; Lewis Publishers: Boca Raton, FL, USA; New York, NY, USA, 1997; p. 328. [Google Scholar]
  17. Rozanski, K.; Araguás-Araguás, L.; Gonfiantini, R. Isotopic patterns in modern global precipitation. GMS 1993, 78, 1–36. [Google Scholar]
  18. Angerosa, F.; Breas, O.; Contento, S.; Guillou, C.; Reniero, F.; Sada, E. Application of stable isotope ratio analysis to the characterization of the geographical origin of olive oils. J. Agric. Food Chem. 1999, 47, 1013–1017. [Google Scholar] [CrossRef] [PubMed]
  19. Iacumin, P.; Bernini, L.; Boschetti, I. Climatic factors influencing the isotope composition of Italian olive oils and geographic characterisation. Rapid Commun. Mass Spectrom. 2009, 23, 448–454. [Google Scholar] [CrossRef] [PubMed]
  20. Aramendía, M.A.; Marinas, A.; Marinas, J.M.; Sánchez, E.; Urbano, F.J.; Guillou, C.; Moreno Rojas, J.M.; Moalem, M.; Rallo, L. A nuclear magnetic resonance (1H and 13C) and isotope ratio mass spectrometry (δ13C, δ2H and δ18O) study of Andalusian olive oils. Rapid Commun. Mass Spectrom. 2010, 24, 1457–1466. [Google Scholar] [CrossRef]
  21. Richter, E.K.; Spangenberg, J.E.; Kreuzer, M.; Leiber, F. Characterization of rapeseed (Brassica napus) oils by bulk C, O, H, and fatty acid C stable isotope analyses. J. Agric. Food Chem. 2010, 58, 8048. [Google Scholar] [CrossRef]
  22. Jiménez-González, M.A.; Álvarez, A.M.; Carral, P.; Almendros, G. Chemometric assessment of soil organic matter storage and quality from humic acid infrared spectra. Sci. Total Environ. 2019, 685, 1160–1168. [Google Scholar] [CrossRef]
  23. Zhao, H.Y.; Guo, B.L.; Wei, Y.M.; Zhang, B.; Sun, S.M.; Zhang, L.; Yan, J.H. Determining the geographic origin of wheat using multielement analysis and multivariate statistics. J. Agric. Food Chem. 2011, 59, 4397–4402. [Google Scholar] [CrossRef]
  24. Hobbie, E.A.; Werner, R.A. Intramolecular, compound-specific, and bulk carbon isotope patterns in C3 and C4 plants: A review and synthesis. New Phytolgist 2004, 161, 371–385. [Google Scholar] [CrossRef]
  25. Ferrio, J.P.; Voltas, J.; Araus, J.L. Use of carbon isotope composition in monitoring environmental changes. Manag. Environ. Qual. 2003, 14, 82–98. [Google Scholar] [CrossRef]
  26. Spangenber, J.E. Bulk C, H, O, and fatty acid C stable isotope analyses for purity assessment of vegetable oils from the southern and northern hemispheres. Rapid Commun. Mass Spectrom. 2016, 30, 2447–2461. [Google Scholar] [CrossRef] [PubMed]
  27. Dawson, T.E. Fog in the California redwood forest: Ecosystem inputs and use by plants. Oecologia 1998, 117, 476–485. [Google Scholar] [CrossRef]
  28. Roden, J.S.; Lin, G.G.; Ehleringer, J.R.A. A mechanistic model for interpretation of hydrogen and oxygen isotope ratios in tree-ring cellulose. Geochim. Cosmochim. Acta 2000, 64, 21–35. [Google Scholar] [CrossRef]
  29. Yakir, S. Variations in the natural abundance of Oxygen-18 and deuterium in plant carbohydrates. Plant. Cell. Environ. 1992, 15, 1005–1020. [Google Scholar] [CrossRef]
  30. Schmidt, H.T.; Werner, R.A.; Rossmann, A. 18O pattern and biosynthesis of natural plant products. Phytochemistry 2001, 58, 9–32. [Google Scholar] [CrossRef]
  31. Gómez-Alonso, S.; García-Romero, E. Effect of irrigation and variety on oxygen (δ18O) and carbon (δ13C) stable isotope composition of grapes cultivated in a warm climate. Aust. J. Grape Wine Res. 2010, 16, 283–289. [Google Scholar] [CrossRef]
  32. Ellsworth, P.Z.; Ellsworth, P.V.; Cousins, A.B. Relationship of leaf oxygen and carbon isotopic composition with transpiration efficiency in the C4 grasses Setaria viridis and Setaria italica. J. Exp. Bot. 2017, 68, 3513–3528. [Google Scholar] [CrossRef]
  33. Schmidt, H.L.; Werner, R.A.; Eisenreich, W. Systematics of 2H patterns in natural compounds and its importance for the elucidation of biosynthetic pathways. Phytocheistry Rev. 2003, 2, 61–85. [Google Scholar] [CrossRef]
  34. International Atomic Energy Agency. Available online: https://nucleus.iaea.org/wiser/index.aspx (accessed on 7 April 2020).
  35. Gat, J.R.; Airey, P.L. Stable water isotopes in the atmosphere/biosphere/lithosphere interface: Scaling-up from the local to continental scale, under humid and dry conditions. Glob. Planet. Chang. 2006, 51, 25–33. [Google Scholar] [CrossRef]
  36. Wang, S.; Zhang, M.; Che, Y.; Zhu, X.; Liu, X. Influence of below-cloud evaporation on deuterium excess in precipitation of arid central Asia and its meteorological controls. J. Hydrometeorol. 2016, 17, 1973–1984. [Google Scholar] [CrossRef]
  37. Miranda, J.D.; Armas, C.; Padilla, F.M.; Pugnaire, F.I. Climatic change and rainfall patterns: Effects on semi-arid plant communities of the Iberian Southeast. J. Arid Environ. 2011, 75, 1302–1309. [Google Scholar] [CrossRef]
  38. Fraga, H.; Pinto, J.G.; Viola, F.; Santos, J.A. Climate change projections for olive yields in the Mediterranean Basin. Int. J. Climatol. 2020, 40, 769–781. [Google Scholar] [CrossRef]
  39. Bigwood, T.; Sharman, M.; Aldus, A.; Dennis, M.J. Global variations in the deuterium/hydrogen isotope ratios of wine. J. Wine Res. 1998, 9, 155–166. [Google Scholar] [CrossRef]
  40. Simpson, H.J.; Herczeg, A.L.; Meyer, W.S. Stable isotope ratios in irrigation water can estimate rice crop evaporation. Geophys. Res. Lett. 1992, 19, 377–380. [Google Scholar] [CrossRef]
  41. Camin, F.; Larcher, R.; Nicolini, G.; Bontempo, L.; Bertoldi, D.; Perini, M.; Schlicht, C.; Schellenberg, A.; Thomas, F.; Heinrich, K.; et al. Isotopic and elemental data for tracing the origin of European olive oils. J. Agric. Food Chem. 2010, 58, 570–577. [Google Scholar] [CrossRef]
  42. De la Rosa, J.M.; Jiménez-González, M.A.; Jiménez-Morillo, N.T.; Knicker, H.; Almendros, G. Quantitative forecasting black (pyrogenic) carbon in soils by chemometric analysis of infrared spectra. J. Environ. Manag. 2019, 251, 109567. [Google Scholar] [CrossRef]
Figure 1. Boxplots of the bulk carbon (δ13C values, a), oxygen (δ18O values, b) and hydrogen (δ2H values, c) isotope composition of olive oils from three Mediterranean countries (France, Portugal and Turkey) in two harvest years (2016 and 2017). Boxplots display the ranges, lower and upper quartiles (Q1, Q3), and the median (Q2). Box plots with different letters indicate significant difference (ANOVA; Means compared using Tukey test, p = 0.05).
Figure 1. Boxplots of the bulk carbon (δ13C values, a), oxygen (δ18O values, b) and hydrogen (δ2H values, c) isotope composition of olive oils from three Mediterranean countries (France, Portugal and Turkey) in two harvest years (2016 and 2017). Boxplots display the ranges, lower and upper quartiles (Q1, Q3), and the median (Q2). Box plots with different letters indicate significant difference (ANOVA; Means compared using Tukey test, p = 0.05).
Foods 09 01855 g001
Figure 2. Local Meteoric Water Lines (LMWL) generated from the bulk δ18O and δ2H of virgin olive oil (VOO) samples. In all cases, the correlation value was highly significate (p < 0.05). The global meteoric water line (GMWL) is shown for reference.
Figure 2. Local Meteoric Water Lines (LMWL) generated from the bulk δ18O and δ2H of virgin olive oil (VOO) samples. In all cases, the correlation value was highly significate (p < 0.05). The global meteoric water line (GMWL) is shown for reference.
Foods 09 01855 g002
Figure 3. Principal components analysis (PCA). (a) 3D Plot of component weights for the geographical variables and the carbon, oxygen and hydrogen (δ13C, δ18O and δ2H) isotope composition of VOO samples; (b) 3D Score plot of VOO samples from France, Portugal and Turkey in 2016 and 2017.
Figure 3. Principal components analysis (PCA). (a) 3D Plot of component weights for the geographical variables and the carbon, oxygen and hydrogen (δ13C, δ18O and δ2H) isotope composition of VOO samples; (b) 3D Score plot of VOO samples from France, Portugal and Turkey in 2016 and 2017.
Foods 09 01855 g003
Figure 4. Observed vs. predicted values for carbon, oxygen and hydrogen isotopic composition in two harvest years (2016 and 2017) calculated by multiple linear regression (MLR) using geoclimatic parameters of VOO producer areas as predictors. (a,b) plots of δ13C from 2016 and 2017; (c,d) plots of δ18O from 2016 and 2017; (e,f) plots of δ2H from 2016 and 2017.
Figure 4. Observed vs. predicted values for carbon, oxygen and hydrogen isotopic composition in two harvest years (2016 and 2017) calculated by multiple linear regression (MLR) using geoclimatic parameters of VOO producer areas as predictors. (a,b) plots of δ13C from 2016 and 2017; (c,d) plots of δ18O from 2016 and 2017; (e,f) plots of δ2H from 2016 and 2017.
Foods 09 01855 g004
Figure 5. Linear discriminant analysis performed on isotope composition of EVOO samples (δ13C, δ18O and δ2H) from 2016 (a) and 2017 (b).
Figure 5. Linear discriminant analysis performed on isotope composition of EVOO samples (δ13C, δ18O and δ2H) from 2016 (a) and 2017 (b).
Foods 09 01855 g005
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Back to TopTop