Tracing the
87
Sr/
86
Sr from rocks and soils to vine and wine: An
experimental study on geologic and pedologic characterisation of
vineyards using radiogenic isotope of heavy elements
Eleonora Braschi
a,b,1, Sara Marchionni
c,d,1, Simone Priori
b, Martina Casalini
c, Simone Tommasini
c,
Laura Natarelli
b, Antonella Buccianti
c, Pierluigi Bucelli
b, Edoardo A.C. Costantini
b,⁎
, Sandro Conticelli
a,c,⁎⁎
a
C.N.R., Istituto di Geoscienze e Georisorse, U.O.S. di Firenze, via Giorgio La Pira 4, I-50121 Firenze, Italy
b
CREA, Centro di Ricerca Agricoltura e Ambiente, via di Lanciola 12a, Cascine del Riccio, I-50125 Firenze, Italy
c
Dipartimento di Scienze della Terra, Università degli Studi di Firenze, via Giorgio La Pira 4, I-50121 Firenze, Italy
dDipartimento di Scienze, Università degli Studi di Roma III, Largo San Gesualdo Murialdo, 1, I-00146 Roma, Italy
H I G H L I G H T S
• Geologic and pedologic traceability of wines
• Sr-isotope in wines from micro-vinification, grapevine sap, soil, and rocks
• Decoupling of radiogenic Sr-isotope in soil and parental rock
• Selective uptake of less radiogenic-Sr isotope component from soils in the bio-available fraction
• Conservation of radiogenic-Sr composi-tion through bio-vegetative life of the vine G R A P H I C A L A B S T R A C T
a b s t r a c t
a r t i c l e i n f o
Article history: Received 23 October 2017Received in revised form 6 February 2018 Accepted 6 February 2018
Available online xxxx
Editor: Yolanda Picó
In this paper we report an experimental study to assess the process of Sr-isotope uptake from the soil and its transfer to the grapevine and then to the wine made through micro-vinification. The experimental work has been carried out with a deep control of the boundary conditions (i.e., type of soil, geologic substratum, ground water supply, etc.) on 11 selected vine-plant sites over a period of four harvest years. Sr-isotopes have been de-termined on grape-bunches, grapevine sap, on the bioavailable fraction of the soil, on bulk soil, and on the rocks of the substratum. No significant Sr-isotope variability has been observed among micro-vinifications from differ-ent harvest years. A slight but significant Sr-isotope variability occurred among wines from rows embedded on different soil type. The Sr-isotope data on micro-vinifications well match those of grapevine sap and bioavailable fraction of soils, all of them falling well within the whole geological range of the bedrock, despite an evident decoupling between bioavailable fraction, whole soils and bedrocks does exist. This decoupling has been ascribed to differential geochemical behaviour of minerals in response to pedogenetic processes. Thefindings of our
Keywords: Tuscany Chianti Classico Sr-isotopes
⁎ Corresponding author.
⁎⁎ Correspondence to: S. Conticelli, Dipartimento di Scienze della Terra, Università degli Studi di Firenze, via Giorgio La Pira 4, I-50121 Firenze, Italy. E-mail addresses:edoardo.costantini@crea.gov.it(E.A.C. Costantini),sandro.conticelli@unifi.it(S. Conticelli).
1
Eleonora Braschi and Sara Marchionni contributed equally to the present paper.
https://doi.org/10.1016/j.scitotenv.2018.02.069
0048-9697/© 2018 Elsevier B.V. All rights reserved.
Contents lists available atScienceDirect
Science of the Total Environment
j o u r n a l h o m e p a g e :w w w . e l s e v i e r . c o m / l o c a t e / s c i t o t e n vexperiments indicate that the biological activity of the vine is not able to change the original87Sr/86Sr composi-tion up-taken from the bio-available fraccomposi-tion of the soil. Thus, the87Sr/86Sr of the wine is an unadulterated feature of the terroir.
© 2018 Elsevier B.V. All rights reserved.
Wines Micro-vinification Geologic traceability
1. Introduction
In order to guarantee the authenticity and the origin of wines and to provide a quality warranty for the consumers, the major wine produc-ing countries have developed severe regulations to control the use of re-gional names for wines. Control authorities deputed to verify wine adulteration and provenance fraud are searching for precise and reliable scientific methods for tracing geographic provenance of grapes and hence of produced wine.
Geographic limits of designated wine production areas often cor-respond to specific geologic and geomorphologic boundaries. In-deed, beside cultivar variety and oenological practice, the soil type, geomorphology, and climate are thought to give distinctive attri-butes to wines (Van Leeuwen and Seguin, 2006). The recognition for wines of their“terroir” of provenance is a further certification of quality of the wine (Costantini and Bucelli, 2014).
Scientists are thus challenged to look for and to establish unambigu-ous parameters to identify the“terroir” that clearly defines the geo-graphic area of wines production and provenance (Vaudour et al., 2015). The determination of the stable isotope composition for light el-ements, such asδ2H,δ13C, andδ18O, in wines is presently used to
iden-tify wine adulteration. In addition, the isotopic composition of stable isotope of light elements provide indication of the latitude of grapes grown, although discrepancies from expected values are sometimes ob-served on the basis of seasonal rain variability (e.g.,Roßmann, 2001;
Christoph et al., 2003, 2004;Hölzl et al., 2004). Similarly, NMR analysis, coupled with multivariate statistical analysis has been used for deter-mining possible adulteration in wines and also to track geographical provenance of grapes and wines (e.g.,Ogrinc et al., 2003; Nilsson et al., 2004;Viggiani and Castiglione Morelli, 2008;Caruso et al., 2012). Chemical and isotopic parameters of wine that might have direct re-lationship with the geochemistry of the substratum of vineyards have also been used to recognise the geographical provenance. Metal distri-bution in grapes and their wines has been widely used, but elemental fractionation in the biologic-cycle of the vine from the root to the grape occurs, thus producing elemental speciation from soil to wine (e.g.,Baxter et al., 1997;Thiel and Danzer, 1997;Castiñeira et al., 2001;Stafilov and Karadjova, 2006;Censi et al., 2014). To encompass the relationships between wines and geological substratum of prove-nance, multivariate statistical analysis has been used with controversial results (e.g.,Marengo and Aceto, 2003;Thiel et al., 2004;Kment et al., 2005;Aceto et al., 2013).
Radiogenic isotope compositions of heavy elements of geological interest (e.g.,87Sr/86Sr,143Nd/144Nd,206Pb/204Pb, 207Pb/204Pb, and
208Pb/204Pb) have been widely used in food science to define the
geo-graphical provenance with some successes (e.g.,Horn et al., 1993;
Kelly et al., 2005;Mihaljevič et al., 2006;Di Paola-Naranjo et al., 2011;
Braschi et al., 2013). Unfortunately most of the studies on the heavy el-ement isotope compositions of wines have barren correlation with the geological substrata, due to the poor analytical reproducibility on foods, which is not comparable to that used in geology, geochemistry, and cosmochemistry for dating and tracing the origin of rocks and min-erals (e.g.,Thirlwall, 1991). Recently,Marchionni et al. (2013)reported a robust analytical method for the determination of87Sr/86Sr in wines
and organic matter at comparable uncertainty level of geological sam-ples, also showing a straightforward Sr-isotopic correlation between young volcanic geological substrata of the vineyards and their wines. A less pronounced correlation, however, was found byMarchionni et al. (2013, 2016)when substrata of vineyards were made up by old
siliciclastic sedimentary rocks, as it was also confirmed by several other studies (Durante et al., 2013, 2015;Petrini et al., 2015;Tescione et al., 2015;Vinciguerra et al., 2016).
In this study we present87Sr/86Sr composition in wines from
micro-vinification over a period of four consecutive harvest years. In addition we determine the87Sr/86Sr of the vine soil from which the grapes
were harvested. Comparison among87Sr/86Sr of wines from
micro-vinification, grapevine sap, soil, and bedrocks was performed to explore the possible components that might change the Sr-isotopic values dur-ing the bio-vegetative cycle of the vine. The use of wines from micro-vinification winemaking technique ensures that during the experimen-tal work no external factors in winemaking affected the Sr-isotopes. The experimental work has been performed in an historical Chianti Classico vineyard, in the area of Castello di Brolio, own by the“Barone Ricasoli” winery. The geochemical study has been supported by a detailed geo-logical and pedogeo-logical survey of the vineyards substrata.
1.1. Geological and pedological settings
The area between Florence and Siena (Fig. 1a) is characterised by a strong variability of the geological substratum, made of sedimentary rocks ranging from siliciclastic to scaly marls and limestones belonging to the Meso-Cenozoic sedimentary sequences, and in particular to i) the Tuscan nappe, and ii) the Apennine oceanic sequences, someplace cov-ered by Pliocene marine deposits (e.g.,Braschi et al., 2012;Marroni et al., 2015, and references therein).
The Chianti Classico wine area (Fig. 1a) is made of sedimentary rocks mainly from the Ligurian oceanic domains, ranging from limestones, to marly limestones and marlstones, but the westernmost vineyards of the wine area lie on sedimentary rocks belonging to the Tuscan nappe, ranging from feldspathic sandstones to siltstones, shales, and marly shales (Fig. 1a) (http://www502.regione.toscana.it/geoscopio/geologia. html).
The Barone Ricasoli vineyards are sited within the Chianti Classico wine area, in the region between Florence and Siena, in Central Tuscany (Fig. 1a). They are located in the countryside near the Castello di Brolio, at south-east of the Gaiole in Chianti village, covering about 1200 ha (Fig. 1b). The studied area, near the Torricella locality, is geologically sited at the edge of the Ligurian domain, where the Pliocenic marine sediments unconformably cover it. Three different bedrocks make the substratum of the vineyards: i) Upper Paleocene to Middle Eocene marly-calcareousflysches (autoc. Monte Morello formation – MLL), be-longing to the External Ligurides sequence, ii) Early Pliocene marine poligenetic conglomerates (PLIb), and iii) Early Pliocene marine sands and sandstones (PLIs) (http://www502.regione.toscana.it/geoscopio/ geologia.html).
Recent integrated geological and pedological surveys of the studied vineyards (Martini et al., 2013;Priori et al., 2013, 2014) have revealed that the bedrock distribution in this area is more complicated than pre-viously reported. Indeed in the area south-eastward the Torricella local-ity Pliocene (PLI) marine to Pleistocene (PLE) continental clays, sands and conglomerate lenses occur above the MLL formation (Martini et al., 2013;Priori et al., 2013, 2014). Strict relationships between soil parent material and the different recognised soils are observed to con-firm the above geological distribution (Priori and Costantini, 2013). The different soils were discriminated on the basis of pedogenetic pro-cesses and chemical/physical factors, such as texture, total carbonates, pH, soil depth, drainage and stoniness, according to the WRB classi fica-tion system (IUSS Working Group WRB, 2014). These vary from Skeletic
Calcaric Cambisol (Torricella) to Abruptic Eutric Luvisol (Leccio-1), Calcaric Cambisol (Leccio-2), Cambic Calcisol and Stagnic Calcaric Cambisol (Miniera), Chromic Cambisol (Nebbiano), and Stagnic Calcaric Cambisol (Santa Lucia) (Priori and Costantini, 2013) (Fig. 1c and Data in Brief,Braschi et al., 2018).
Few isotope data are available in literature for the sedimentary rocks from Tuscany (Conticelli, 1998;Conticelli et al., 2001, 2015a, 2015b;
Melluso et al., 2003). Among these data, limestone samples show the lowest87Sr/86Sr measured values (0.707551
–0.707669) and shales to siltstones and sandstones among the highest87Sr/86Sr measured values
(0.721555–0.738229). Marlstones, which are a mixture of two compo-nents, limestone and shale to siltstone, have intermediate87Sr/86Sr measured values (0.708797–0.711234) on the basis of the amount of each component present in the rock.
2. Materials and methods
The study is focused on wines made through micro-vinification of “Sangiovese” monocultivar and on grapevine saps from single vine-rows in the Torricella area of the Brolio vineyards, own by the“Barone Ricasoli” winery, together with the bedrocks and soils making the sub-strata of the vineyard.
For the present study 11 (eleven) sampling points (single vine plants), rooted on different soils and bedrocks, were selected for the har-vesting of grapes for micro-vinification (Fig. 1c). Among themfive vines were selected for sampling the grapevine saps. For each of the selected sampling points (Fig. 1c) pedologic profiles were described (Priori and Costantini, 2013) and soils were sampled. Micro-vinification was per-formed over four different harvest years (2008–2011). Twelve (12) Fig. 1. a) Geological sketch map of the Chianti Classico wine area with boundary of the wine area defined by the “disciplinare” (http://www.chianticlassico.com/en/territory/ characteristics/), the red inset is the area of the geological map blow-up in“b”. Legend: OLO = Olocenic deposits; VIL = continental deposits from “Villafranchiano” period; PLIs = Marine Sands and Sandstones (Early Pliocene age); FAA = Blue Shales formation; PLIb = Marine Poligenetic Conglomerates (Early Pliocene age); MES = Messinian post-evaporitic lacustrine deposits. External Ligurian domain: OMT =“Ottone-Monterverdi” flysch formation; MLL = “Monte Morello” formation (made by Upper Paleocene to Middle Eocene marly-calcareous rocks); SIL =“Sillano” formation; PTF = “Pietraforte” formation; AVR = “Varicolori” shales formation. Tuscan domain: MAC = “Macigno” formation; STO = “Scaglia Toscana” formation. The red square represent the location of the study area enlarged in b and c. b) Geological sketch map of the area under study. MLL = Monte Morello formation, which is made by Upper Paleocene to Middle Eocene marly-calcareous rocks belonging to the External Ligurid sequence; PLIb = Marine Poligenetic Conglomerates with Early Pliocene age; PLIs = Marine Sands and Sandstones with Early Pliocene age. c) Geo-pedological sketch map of the La Torricella– Nebbiano vineyard of the “Barone Ricasoli” vineyard. For each local soil type recognised (Priori and Costantini, 2013) the main pedological classification is as follow, on the basis of pedogenetic processes and chemico/physical factors (i.e., composition, porosity, density, consistency, drainage and stoniness;IUSS, 2014): Torricella soil = Skeletic Calcaric Cambisol; Nebbiano soil = Chromic Cambisol; Leccio1 = Abruptic Eutric Luvisol; Miniera = Stagnic Calcaric Cambisol; Santa Lucia = Stagnic Calcaric Cambisol; Leccio2 = Calcaric Cambisol. For detailed descriptions see Data in Brief (Braschi et al., 2018). (For interpretation of the references to colour in thisfigure legend, the reader is referred to the web version of this article.)
rock samples of the geological substratum of the vineyards were also sampled (Table 1 of Data in Brief,Braschi et al., 2018).
2.1. Wines from micro-vinification and grapevine saps
For each harvest point, grapes from three different adjacent vines were harvested, collecting about 1.5 kg of Sangiovese bunches. The se-lected vines were the head plants of three adjacent vine rows with the central vine labelled with“b”, the left one labelled with “a”, and the right one labelled with“c”. Each bunch of grape, collected for each single sampling point, was carefully cleaned and separately processed under laboratory controlled conditions through micro-vinification to obtain three experimental wines for isotope analyses in each different harvest year (Table 1).
Grapevine saps were sampled from six (6) selected vine plants dur-ing the sprdur-ing season, when the bio-vegetative period corresponds to the maximum grapevine sap production time. The sampling was made by the trunk of the vine plant, or one of the main brunches, which was carefully carved with a cutter to collect the secreted sap in a clean vessel.
Wine samples (some 5 ml) and grapevine saps (some 2 ml) were digested before elemental purification to completely oxidise the organic matter and thus to efficiently isolate the Sr from their inorganic fraction. Sample digestion and purification was performed for all type of samples in a clean lab (“Class 1000” environment) at the Dipartimento di Scienze della Terra of the Università degli Studi di Firenze. Digestion was
performed under horizontal HEPA-filtered laminar flow workstations, which ensures a low-blank working area.
The chemical procedure employed for sample digestion and Sr puri-fication is the same reported inMarchionni et al. (2013)and experi-mentally set up for wines. The sapfluid samples were processed following the same method used for the wines considering their similar fluid organic matrix (Data in Brief,Braschi et al., 2018).
2.2. Rock, soil and bio-available fraction
Along with grapes, 3–5 kg of soil was collected in correspondence of each sampling point from the horizons richest in grapevine roots (gen-erally between 10 and 30 cm). This depth represents the level from which the adsorption of vine roots preferentially takes place. The soil samples were analysed both as bulk material and as leachate (i.e., the bioavailable soil solution fraction, hereafter labile fraction) to assess the different Sr isotope composition, if any, between the soil and the soil solutions (i.e., liquid water in a moist soil). Bulk rocks have also been sampled from the three major bedrock formations and treated ac-cordingly to standard cleaning, crushing and powdering methods. A fraction of the powdered samples of both bulk rock and whole soil was analysed for major (wt%) and trace element (ppm) contents at the Activation Laboratories Ltd. (Ontario, Canada), using the 4Lithores schedule (Table 2 of Data in Brief,Braschi et al., 2018). Detection limits and errors are given athttp://www.actlabs.com/page.aspx?page= 517&app=226&cat1=549&tp=12&lk=no&menu=64.
Table 1
Description and Sr isotope results for wine samples of each single micro-vinification from the studied harvest points. Harvest point Sample vine row Local soil name WRB classification (2014) Geologic formation 2008 2009 2010 2011 87 Sr/86 Sr 2 s.e. 87 Sr/86 Sr 2 s.e. 87 Sr/86 Sr 2 s.e. 87 Sr/86 Sr 2 s.e. BRO1 a Torricella Skeletic Calcaric Cambisol MLL 0.708144 ±0.000007 0.708199 ±0.000006 0.708295 ±0.000007 0.708267 ±0.000006 BRO1 b Torricella Skeletic Calcaric Cambisol MLL 0.708148 ±0.000005 0.708135 ±0.000006 0.708277 ±0.000005 0.708237 ±0.000006 BRO1 c Torricella Skeletic Calcaric Cambisol MLL 0.708249 ±0.000006 0.708192 ±0.000006 0.708285 ±0.000008 0.708313 ±0.000005 BRO2 a Torricella Skeletic Calcaric Cambisol MLL 0.708195 ±0.000006 0.708215 ±0.000007 0.708236 ±0.000006 n.d. n.d. BRO2 b Torricella Skeletic Calcaric Cambisol MLL 0.708195 ±0.000006 0.708197 ±0.000006 0.708314 ±0.000006 n.d. n.d. BRO2 c Torricella Skeletic Calcaric Cambisol MLL 0.708238 ±0.000006 0.708197 ±0.000006 0.708258 ±0.000006 n.d. n.d. BRO4 a Nebbiano Chromic Cambisol PLE 0.708459 ±0.000006 0.708427 ±0.000006 0.708470 ±0.000007 0.708565 ±0.000005 BRO4 b Nebbiano Chromic Cambisol PLE 0.708425 ±0.000006 0.708463 ±0.000006 0.708330 ±0.000005 0.708521 ±0.000005 BRO4 c Nebbiano Chromic Cambisol PLE 0.708499 ±0.000006 0.708384 ±0.000006 0.708458 ±0.000005 0.708535 ±0.000006 BRO5 a Miniera Stagnic Calcaric Cambisol FAA 0.708749 ±0.000006 0.708783 ±0.000006 0.708753 ±0.000006 0.708736 ±0.000005 BRO5 b Miniera Stagnic Calcaric Cambisol FAA 0.708750 ±0.000006 0.708775 ±0.000006 0.708725 ±0.000005 0.708774 ±0.000006 BRO5 c Miniera Stagnic Calcaric Cambisol FAA 0.708734 ±0.000005 0.708756 ±0.000006 0.708771 ±0.000005 0.708777 ±0.000005 BRO6 a Nebbiano Chromic Cambisol PLE 0.708494 ±0.000006 0.708306 ±0.000006 0.708422 ±0.000008 n.d. n.d. BRO6 b Nebbiano Chromic Cambisol PLE 0.708476 ±0.000006 0.708327 ±0.000005 0.708613 ±0.000009 n.d. n.d. BRO6 c Nebbiano Chromic Cambisol PLE 0.708357 ±0.000006 0.708383 ±0.000006 0.708515 ±0.000005 n.d. n.d. BRO8 a S.·Lucia Stagnic Calcaric Cambisol PLE 0.708578 ±0.000006 0.708560 ±0.000006 0.708427 ±0.000006 n.d. n.d. BRO8 b S. Lucia Stagnic Calcaric Cambisol PLE 0.708531 ±0.000005 0.708586 ±0.000006 0.708513 ±0.000008 n.d. n.d. BRO8 c S.·Lucia Stagnic Calcaric Cambisol PLE 0.708517 ±0.000006 0.708512 ±0.000005 0.708545 ±0.000006 n.d. n.d. BRO9 a Leccio1 Abruptic Eutric Luvisol PLI 0.708947 ±0.000006 0.709086 ±0.000005 0.708914 ±0.000007 n.d. n.d. BRO9 b Leccio1 Abruptic Eutric Luvisol PLI 0.709089 ±0.000006 0.709080 ±0.000006 n.d. n.d. n.d. n.d. BRO9 c Leccio1 Abruptic Eutric Luvisol PLI 0.708980 ±0.000006 0.708957 ±0.000006 0.708859 ±0.000009 n.d. n.d. BRO10 a Leccio1 Abruptic Eutric Luvisol PLI n.d. n.d. 0.708590 ±0.000006 0.708516 ±0.000006 n.d. n.d. BRO10 b Leccio1 Abruptic Eutric Luvisol PLI n.d. n.d. 0.708737 ±0.000006 0.708599 ±0.000009 n.d. n.d. BRO10 c Leccio1 Abruptic Eutric Luvisol PLI n.d. n.d. 0.708622 ±0.000005 0.708538 ±0.000007 n.d. n.d. BRO11 a Leccio2 Calcaric Cambisol PLI 0.708333 ±0.000006 0.708376 ±0.000006 0.708433 ±0.000005 0.708422 ±0.000006 BRO11 b Leccio2 Calcaric Cambisol PLI 0.708365 ±0.000006 0.708399 ±0.000006 0.708411 ±0.000008 0.708442 ±0.000006 BRO11 c Leccio2 Calcaric Cambisol PLI 0.708357 ±0.000006 0.708410 ±0.000006 0.708397 ±0.000006 0.708397 ±0.000005 BRO12 a Leccio2 Calcaric Cambisol PLI 0.708647 ±0.000006 0.708720 ±0.000006 0.708756 ±0.000005 n.d. n.d. BRO12 b Leccio2 Calcaric Cambisol PLI 0.708645 ±0.000005 0.708677 ±0.000006 n.d. n.d. n.d. n.d. BRO12 c Leccio2 Calcaric Cambisol PLI 0.708615 ±0.000006 0.708652 ±0.000006 0.708683 ±0.000005 n.d. n.d. BRO13 a Miniera Cambic Calcisol FAA 0.708719 ±0.000005 0.708637 ±0.000005 0.708667 ±0.000005 0.708664 ±0.000005 BRO13 b Miniera Cambic Calcisol FAA 0.708616 ±0.000005 0.708658 ±0.000005 n.d. n.d. 0.708674 ±0.000005 BRO13 c Miniera Cambic Calcisol FAA n.d. n.d. 0.708646 ±0.000006 0.708694 ±0.000005 0.708711 ±0.000006 Footnotes: Local soil names are fromPriori and Costantini (2013); Bedrocks names: MLL = Monte Morello Formation, (Upper Paleocene to Middle Eocene); PLI = PLIb (Pliocene Polygenic Marine Conglomerate) and PLIs (Pliocene Marine Yellow Sandstone); PLE = Plio-Pleistocenefluviolacustrine deposits; FAA = Blue Shale Formation. Legend: 2 s.e. = two standard error of the mean at 95% confidence level; n.d. = not determined. Each isotopic value is given by the average calculated on 120 cycles of measurements, using the peak-jumping method; each cycle value is given by the average of three measurements with different cup configuration. Labels a, b, and c indicate the vine row sampled (see text for explanation).
In order to define the type of sampled soil, laboratory analyses were also carried out on each soil horizon up to the rooting depth, for each sampling site. Available water capacity (AWC) was estimated as the dif-ference between soil water content atfield capacity (FC) and wilting point (WP), as determined at the CREA laboratories in Florence, using the Richards pressure plate extractor. FC, WP, and AWC (mm m−1) were measured on thefine earth, whereas available water until the root limiting layer was reduced by taking into account the skeleton con-tent. Layers without measured bulk density were beyond the rooting depth. Bulk density was obtained with the core method, averaging two or three replicated samples. Routine analyses were carried out fol-lowing the Italian official methods (MiPAF, 2000) (Data in Brief,Braschi et al., 2018).
For isotope determination both rock and whole soil samples were analysed following the sample preparation, digestion and Sr purification method for whole rock reported inAvanzinelli et al. (2005).
In order to obtain a representative bioavailable soil solution fraction (labile fraction) we used specific UNIBEST® capsules made up of ion ex-change resins designed to reproduce the uptake of the plant roots from the soil (Skogley and Dobermann, 1996). The capsule is left in a mixture of soil and Milli-Q® water to allow the bioavailable elemental up-take. The extraction and further Sr purification procedure is detailed in
Marchionni et al. (2016). 2.3. Sr isotope measurements
Some 100–150 ng of sample were loaded on single Re filament as ni-trate form along with a Ta activator and H3PO4, aiming to a properly stable
beam intensity during the measurements. Sr isotope ratios were mea-sured with a Thermal Ionisation Mass Spectrometer (TIMS) Triton TI® using a multi-dynamic mode, applying the triple jump procedure (Thirlwall, 1991) described in detail byAvanzinelli et al. (2005). Each re-ported isotope ratio was the result of 120 sets of cycles (with each cycle representing the average of three measurements performed during triple-jumping), taken in 6 blocks, each consisting of 20 cycles with 8 s in-tegration time. An idle time of 3 s was set before the start of the collection after each jump, to eliminate possible memory effect due to the decay of the signal in the faraday cups (Avanzinelli et al., 2005).
The instrumental mass bias was corrected off-line using the88Sr/86Sr
ratio measured on the main configuration (jump 2). The measured and the natural88Sr/86Sr (88Sr/86Sr
N= 8.375209) were used both to
calcu-late the mass discrimination factor (ε) and to subsequently apply the correction through the exponential fractionation law.
87Sr/86Sr
tripleaverage value for NIST SRM987 international reference
standard was 0.710248 ± 0.000016 (2σ, n = 121). The Sr procedural blank measured during the course of these analyses was between 120 and 300 pg, which is in the blank range of our lab (Avanzinelli et al., 2005).
The reproducibility of the analytical method we used in this study is reported inMarchionni et al. (2013), where 31 different aliquots of the same sample of wine were processed and measured for87Sr/86Sr com-position, yielding a 2σ = ±0.000017 (i.e., ±23 ppm), which is well con-sistent with that of the international reference standard.
3. Sr-isotope results and discussion
3.1. Isotopic variability in wines and grapevine saps
The results of the 107 samples micro-vinifications are reported in
Table 1. These are the results of 11 different harvest points from which grape bunches were sampled over four consecutive vintage years (2008–2011) and used for micro-vinification transformation. Ac-tually three different micro-vinification were performed for each har-vest point (labelled a, b, and c inTable 1) coming from three adjacent vine rows. Overall, the87Sr/86Sr values on micro-vinification samples
vary from 0.708135 ± 0.000006 to 0.709089 ± 0.000006. It is worthy
to note that in most cases the three micro-vinifications, from the same harvest point of the same vintage year, generally showed87Sr/86Sr
values within the analytical uncertainty. Thus to evaluate the87Sr/86Sr
variations through the different vintage years under consideration we merged the data of the three adjacent micro-vinifications (a–c in
Table 1) of each harvest point (Table 3 of Data in Brief,Braschi et al., 2018).
The87Sr/86Sr values of micro-vinification of each harvest point
(BRO1-BRO13) were plotted inFig. 2to evaluate the possible discrepan-cies of the isotopic value through time. The covariation plots showed in
Fig. 2a–c were chosen to consecutively compare the results obtained by the Sr-isotope analyses on micro-vinifications year by year. All data for each paired years of harvest fall on, or very close to, the 1:1 correlation line, indicating a preservation of the isotopic signature during two con-secutive harvest seasons. InFig. 2d data are fully compared during the four harvesting years, using a bar chart diagram that display an overall good reproducibility of the micro-vinification isotopic values over the period under study.
The general conservation of the87Sr/86Sr values of wines from the
same harvest point over the four different vintage years (2008–2011) enabled us to statistically treat the overall data obtained on each harvest points (BRO1–BRO13). The corresponding averaged values represent the large 87Sr/86Sr average of each micro-vinification, hereafter
(87Sr/86Sr)
M. The calculated descriptive statistics is reported inTable 2.
The collected grapevine sap samples show87Sr/86Sr values ranging from 0.708306 ± 0.000006 to 0.708582 ± 0.000006 (Table 3). These Sr-isotopic values correlate with the (87Sr/86Sr)
Mof the corresponding
micro-vinification from each harvest point (Fig. 3), providing further support to thefinding ofMarchionni et al. (2016)who have shown that in the case of winemaking process the87Sr/86Sr does not change
from soil to grape, must, and wine.
In summary,87Sr/86Sr isotopic values in micro-vini
fication remained constant during the different years of sampling, providing further evi-dence that it may be used as a tool for characterising the traceability of wines not only in a whole production area (e.g.,Marchionni et al., 2013;Durante et al., 2013;Petrini et al., 2015) but also at a more local scale within the same vineyard. In addition, the correspondence be-tween the87Sr/86Sr of grapevine saps (Table 3) and the (87Sr/86Sr)Mof
micro-vinification from the same harvest point argue for the absence of any metabolic factor that may affect the87Sr/86Sr during the
biologi-cal life of the vine from the Sr uptake by the roots to the fruit maturation. 3.2. Sr-isotopes in rocks and soils
In order to verify experimentally the existing relationships, if any, between the wine produced by single micro-vinification experiments and the substrata of the correspondent harvested vine plant, we sam-pled and analysed the soil profile and the bedrock underling each har-vest point. The87Sr/86Sr values obtained on each sample of bedrock,
whole soil and labile fraction are reported inTable 3.
The bedrocks characterising the substrata of the vineyards chosen for this experimental works have different lithological features reflected by their whole rock chemistry (Table 2 of Data in Brief,Braschi et al., 2018) and87Sr/86Sr values (Table 3). Indeed the marly calcareous
rocks (MLL) have quite homogeneous chemical compositions, both for major (wt%) and trace (ppm) elements, with also homogeneous
87Sr/86Sr values, found to be in the range between 0.707974 ±
0.000006 and 0.707997 ± 0.000006 (Table 3andFig. 4a). Contrarily, early Pliocene marine sands and polygenic conglomerate (PLIs and PLIb, respectively) show a larger chemical compositional variability, with also higher and lower87Sr/86Sr values than MLL. The87Sr/86Sr values indeed vary from 0.707791 ± 0.000007, of the calcareous conglomerate, to 0.716998 ± 0.000012, of a bluish shale found interlayered within PLIb (Table 3andFig. 4a), although most of them never exceeding the87Sr/86Sr value of 0.709535 ± 0.000008
and PLIb terranes are part of the same lithological unit, being the PLIs making the matrix of PLIb conglomerate, we thus henceforth refer to a unique PLI lithological unit for these two formations. Moreover, several authors have recognised in the studied area a further discon-tinuous, thin (50–100 cm; Table 1 of Data in Brief,Braschi et al., 2018) lithological unit made by Pleistocenicfluvio-lacustrine con-glomerate and sands (PLE;Martini et al., 2013;Priori et al., 2013, 2014) that, however, have a similar lithological composition with
respect to PLI and thus possibly similar geochemical and isotopic composition.
On these different lithologies six different soil types were found (Fig. 1c;Table 3). One of them developed exclusively on the MLL (i.e., the Skeletic Calcaric Cambisol, autoc. Torricella of BRO1 and BRO2), whilst another one (i.e., Abruptic Eutric Luvisol, autoc. Leccio1 of BRO9 and BRO10) developed on PLE with MLL rocks at shallow depth (Priori and Costantini, 2013). The Calcaric Cambisol (i.e., autoc. Fig. 2. Variation of the87
Sr/86
Sr values of wines during the four harvest years under study. a-c) Selected covariation diagrams showing the87
Sr/86
Sr between two consecutive years of harvest for each sampling point. Each value is the average of the three different micro-vinifications obtained by adjacent vine plants (a, b, c inTable 1, see text for explanation). Error bars represent the standard deviation (s.d.) for each mean value (seeTable 3of Data in Brief,Braschi et al., 2018). Different symbols discriminate among sampling (harvest) points (i.e., BRO1–13); empty and full symbols are used to discriminate between two different sampling points belonging to the same row; colours indicate different soil type; d) Bar chart di-agram showing the overall variation of87Sr/86Sr of micro-vinification, averaged by years, during time. Error bar represents the uncertainty (s.d., see Table 3 of Data in Brief,Braschi et al.,
2018). Colourfields under bars indicate the relative soil type for each sampling point (i.e., BRO1–13). (For interpretation of the references to colour in this figure legend, the reader is re-ferred to the web version of this article.)
Table 2 Overall87
Sr/86
Sr data used to calculate the large87
Sr/86Sr average of each wines from micro-vinification over the four vintage years for each harvest point, i.e. (87
Sr/86
Sr)M.
(87
Sr/86
Sr)M BRO1 BRO2 BRO4 BRO5 BRO6 BRO8 BRO9 BRO10 BRO11 BRO12 BRO13
Mean 0.708230 0.708228 0.708463 0.708759 0.708436 0.708533 0.708992 0.708604 0.708399 0.708678 0.708671 Median 0.708233 0.708223 0.708467 0.708761 0.708421 0.708538 0.708989 0.708597 0.708399 0.708674 0.708669 Standard deviation (s.d.) ±0.000074 ±0.000056 ±0.000082 ±0.000043 ±0.000111 ±0.000063 ±0.000095 ±0.000088 ±0.000052 ±0.000057 ±0.000049 Sample variance 5.43E−09 3.19E−09 6.65E−09 1.87E−09 1.24E−08 4.03E−09 8.93E−09 7.73E−09 2.75E−09 3.26E−09 2.39E−09 Min 0.708040 0.708069 0.708083 0.708630 0.708006 0.708284 0.708693 0.708272 0.708233 0.708512 0.708528 Max 0.708516 0.708392 0.708898 0.708901 0.708859 0.708746 0.709218 0.708831 0.708567 0.708861 0.708831
Counts 1429 1080 1402 1440 1075 1080 959 700 1440 943 1200
Leccio2 of BRO11 and BRO12) developed on the PLI formation, whereas the Chromic Cambisol and a Stagnic Calcaric Cambisol (i.e., autoc. Nebbiano of BRO4 and autoc. Santa Lucia of BRO8, respectively) devel-oped indifferently on PLE (Table 3). We also recognised that in the case of the Stagnic Calcaric Cambisol and Cambic Calcisol (i.e., autoc. Miniera of BRO5 and BRO13 respectively) the parent material was characterised at shallow depth by a level of blue shale, attributed to the FAA formation (Table 1 of Data in Brief,Braschi et al., 2018).
The87Sr/86Sr on whole soils ranges from 0.709370 ± 0.000013 to
0.714969 ± 0.000006 (Table 3) characterising the Skeletic Calcaric Cambisol (BRO 1, developed on MLL rocks) and the Abruptic Eutric Luvisol (BRO 9, developed on PLE/MLL rocks), respectively (Table 3,
Fig. 4). InFig. 4a, the87Sr/86Sr of the analysed whole soils and bedrocks
is compared with their Sr content (ppm). From this plot is clear the decoupling between the87Sr/86Sr of whole soils and that of their
sup-posed geological sources (i.e., MLL, PLE, and PLI bedrocks). Most of the bedrock samples forming the substrata of the studied vineyard show in-deed largely variable Sr contents (ranging from 200 to 800 ppm) but
87Sr/86Sr values clustered between 0.707791 ± 0.000007 and
0.709535 ± 0.000008 (with the sole exception of the bluish shale show-ing an outlier value of 0.716998 ± 0.000012 with less than 100 ppm of Sr) whereas the whole soil samples show a wide range of Sr contents and 87Sr/86Sr values in between the two extreme bedrock end-members (i.e., the calcareous lithologies and the bluish shale).
It is worth to note that soils form from the chemical and physical weathering of the parental material thus representing a mixture of the different lithologies occurring in the bedrock substrata. In this light the samples of whole soil show variable87Sr/86Sr values following a
hy-perbolic mixing trend between the two extreme end-member composi-tions defined above (Fig. 4a). Although the bluish shale found within PLI is a thin layer (5–10 cm thick) unlikely representing the only, real end-member able to enrich the whole soil in radiogenic Sr, we may argue that its isotopic composition might be representative of generic radiogenic-Sr shale present in the area. Despite the lack of correspon-dence between the87Sr/86Sr values of the soils and those of their
bed-rock sources is undoubted, a rough negative correlation is observed between87Sr/86Sr and total CaCO
3content of bulk soils (Fig. 4b). This
correlation testifies the progressive increase of a more Sr radiogenic fraction (possibly a clay-rich, strongly radiogenic source) in the bulk soil with decreasing of the carbonate component, as generally expected. Anyway, despite this rough correlation, in many cases the amount of Table 3
87Sr/86Sr values for grapevine saps, whole soils, labile fraction of soils and bedrocks.
Harvest point Local soil name WRB classification (2014) Geologic formation Sap Whole soil Labile fraction
87 Sr/86 Sr 2 s.e. 87 Sr/86 Sr 2 s.e. 87 Sr/86 Sr 2 s.e. BRO1 Torricella Skeletic Calcaric Cambisol MLL 0.708306 ±0.000006 0.709370 ±0.000013 0.708045 ±0.000006 BRO4 Nebbiano Chromic Cambisol PLE 0.708332 ±0.000007 0.712818 ±0.000012 0.708160 ±0.000006 BRO5 Miniera Stagnic Calcaric Cambisol FAA 0.708332 ±0.000007 0.709990 ±0.000006 0.708762 ±0.000006 BRO8 Santa Lucia Stagnic Calcaric Cambisol PLE 0.708582 ±0.000006 n.d. n.d. n.d. n.d. BRO9 Leccio1 Abruptic Eutric Luvisol PLI/MLL n.d. n.d. 0.714969 ±0.000006 0.709033 ±0.000006 BRO10 Leccio1 Abruptic Eutric Luvisol PLI/MLL 0.708549 ±0.000006 0.713977 ±0.000015 0.708596 ±0.000005 BRO11 Leccio2 Calcaric Cambisol PLI 0.708400 ±0.000006 0.710116 ±0.000006 0.708321 ±0.000006 BRO12 Leccio2 Calcaric Cambisol PLI n.d. n.d. 0.713433 ±0.000006 0.708622 ±0.000006 BRO13 Miniera Cambic Calcisol FAA 0.708559 ±0.000006 0.711405 ±0.000006 0.708429 ±0.000006
Bedrock sample Lithology Geologic formation
87 Sr/86 Sr 2 s.e. MML1 Limestone MLL 0.707974 ±0.000006 MML2 Marly-limestone MLL 0.707996 ±0.000009 MML3 Marl MLL 0.707997 ±0.000006
PLIs Sand PLIs 0.709535 ±0.000008
PLIb0 Sandy matrix PLIb 0.709192 ±0.000007
PLIb1 Pebble PLIb 0.708050 ±0.000006
PLIb2 Pebble PLIb 0.707852 ±0.000007
PLIb3 Pebble PLIb 0.708352 ±0.000008
PLIb4 Pebble PLIb 0.707791 ±0.000007
PLIb5 Pebble PLIb 0.709064 ±0.000011
PLIb6 Pebble PLIb 0.708489 ±0.000010
PLIb7 Bluish shale PLIb 0.716998 ±0.000012
Footnotes: Local soil names are fromPriori and Costantini (2013); Bedrock names: MLL = Monte Morello Formation (Upper Paleocene to Middle Eocene); PLI = PLIb (Pliocene Polygenic Marine Conglomerate Pliocene) and PLIs (Pliocene Marine Yellow Sandstone Pliocene); FAA = Blue Shale Formation; PLE = Plio-Pleistocenefluvio-lacustrine deposit. Legend: 2 s.e. = two standard error of the mean at 95% confidence level; s.d. = standard deviation; n.d. = not determined. In italic is reported the value for either BRO4 or BRO5 vine sap. This sap sample cannot be correctly attributed to none of the two sampling points (both belonging to the same vine row) due to an uncertainty in the sampling procedure. As a consequence the possible corre-lation with the relative BRO4 or BRO5 micro-vinification remains ambiguous and not straightforward. The result is thus not reported inFig. 3, and not discussed in the text but is reported in the table for completeness.
Fig. 3. Covariation diagram of the87Sr/86Sr of the wines (87Sr/86Sr
M) vs. the87Sr/86Sr of the
grapevine saps. The87Sr/86Sr of the wine is the great mean calculated through the
descriptive statistics reported inTable 3using the overall wine produced through micro-vinification for each harvest point. Error bars represent the standard deviation (s.d.) for each mean value (Table 3of Data in Brief,Braschi et al., 2018).
CaCO3in soils is not directly related to the composition of the
corre-sponding bedrocks (e.g., the Abruptic Eutric Luvisol of BRO 9–10). In summary,87Sr/86Sr values obtained on whole soils from the differ-ent harvest points indicate a decoupling with respect to the87Sr/86Sr
values of the parent material of the soil, suggesting that possibly most of them might undergo further mixing of material from the nearby vineyards, due also to erosional to runoff processes along the gently slope of some vineyards.
3.3. Sr-isotopes from soils to wine
It was shown that the biological cycle of the vine affects the total content of micro-elements, inducing differential enrichments among roots, brunches, leafs, and fruits (Garcia et al., 2001;Censi et al., 2014). On the other hand,Marchionni et al. (2016)andDurante et al. (2016)
showed that no87Sr/86Sr fractionation occurred during winemaking procedure, from grape to wines, and that the isotope composition is pre-served from the labile fraction uptake by the roots from the soil. Our ex-periments confirmed that87Sr/86Sr fractionation does not occur from
the grapevine sap to wine from micro-vinification of the same harvest-point.
InFig. 5the box plots for the (87Sr/86Sr)
Mof the wines from
micro-vinification are shown, with the samples ordered according to the soil type and relative substrata. In most cases, micro-vinifications from the same soil type overlap completely, but significantly differ from others. This is the case of wines from the harvest points BRO1 and BRO2 with vines grown on Skeletic Calcaric Cambisol developed on MLL rocks (autoc. Torricella soil;Priori and Costantini, 2013), and those from the harvest points BRO4 and BRO6 with vines grown on Chromic Cambisol (autoc. Nebbiano soil; Priori and Costantini, 2013) developed on Pleistocenic continental deposits (Fig. 5). Similarly but less evidently, the harvest points BRO5 and BRO13, with substrata made by Stagnic Calcaric Cambisol and Cambic Calcisol soil type (autoc. Miniera soil;
Priori and Costantini, 2013) and FAA geological formations, show simi-lar values within the analytical error. On the other hand, an evident decoupling does occur for some wines obtained by micro-vinification of vines growing on the same soil type and bedrocks (Fig. 5). This is the case of wines from the BRO 9 and BRO 10 harvest points, both characterised by vines growing on Abruptic Eutric Luvisol (autoc. Leccio 1 soil;Priori and Costantini, 2013) soil type and developed on bedrock attributed to PLI overlying the MLL (Fig. 5). Here the (87Sr/86Sr)
Mof
BRO 9 wine is consistently higher than that obtained from BRO10 micro-vinification. Similarly, BRO 11 and BRO 12, embedded in Calcaric Cambisol soil type (autoc. Leccio 2 soil;Priori and Costantini, 2013), which developed on PLI, are considerably different. BRO12 wine is in-deed more radiogenic in Sr than BRO 11 wine, matching perfectly the value obtained from wines by micro-vinifications of BRO 13 vine
plant. The latter harvest point is characterised by Cambic Calcisol soil type (i.e., autoc. Miniera soil,Priori and Costantini, 2013), developed on a thin PLI bedrock, overlying the FAA at shallow depth. For this spe-cific case the only difference observed with respect to the other harvest points is the closeness with a local fault that divide the MLL from PLI for-mations and that in the case of BRO9 soil it might have mixed it with ad-jacent ones, developed on the different rock types beyond the fault (Fig. 1c).
In spite of these connections, when the (87Sr/86Sr)
Mof wines from
micro-vinification and that of whole soils are plotted one vs. the other one to verify the possible relationships, no correspondence on a 1:1 cor-relation line is observed (Fig. 6), with87Sr/86Sr of whole soils plotting at
higher values than those of wines. Nonetheless, a rough positive trend can be observed, indicating that in general the wines from micro-vinification inherited the same behaviour of soils (Fig. 6). To encompass the mismatch, the87Sr/86Sr of the labile fraction of the soil, considered to be the bioavailable fraction adsorbed through the vine-root uptake, was determined in the sampled soils (Table 3). The lowest87Sr/86Sr
value (0.708045 ± 0.000006) was found for the labile fraction of the Fig. 4. a)87Sr/86Sr vs. Sr (ppm) of whole soils and bedrock samples. The carbonatic pebbles and bluish shale samples are assumed as possible end-member compositions from which the
soil of the studied vineyard derived. The theoretical mixing curve between these two end-members is shown; labels (in per cent) represent the increasing shale content in the mixture. Error bars (2 s.e.) are within the symbols. For MLL, PLIs and PLIb meaning see caption ofFig. 1. b)87
Sr/86
Sr vs. CaCO3% of whole soil. Error bars (2 s.e.) are within the symbols.
Fig. 5. Box plot for the (87
Sr/86
Sr)Mof each harvest point. The overall measured values of
wines from the same harvest points have been used to calculate the corresponding box: the box highness is defined by the value of q1 and q3 (first and third quartile), the solid line and the star symbol represent the median and the mean, respectively; the external bars indicate the minimum and the maximum values. Boxes for harvest points on the same soil type are paired one beside the other one. The abscissa corresponds to the roughly North-South distribution of the different harvest points. Shaded area reports the
87
Sr/86
Skeletic Calcaric Cambisol (autoc. Torricella soil;Priori and Costantini, 2013) of the harvest point BRO1, developed on MLL rocks, whilst the highest87Sr/86Sr value was found for the labile fraction of the Abruptic
Eutric Luvisol (Ruptic Hypereutirc; autoc. Leccio 1 soil; Priori and Costantini, 2013), of the harvest point BRO9, developed on PLI rocks (Table 3).
InFig. 7a the (87Sr/86Sr)
Mof wines are plotted vs. the87Sr/86Sr values
of the labile fraction of the soil sampled beneath each correspondent harvest point. A positive correlation is displayed and all the Sr isotope values fall close to the 1:1 correlation line, with the samples at lower (87Sr/86Sr)
Mthat plot offset (e.g., BRO1, BRO4 and BRO13). However,
when a statistical regression is applied, all the samples fall well within the 95% confidential level at 95% prediction limit, with at high R = 0.949 (Fig. 7b), suggesting the strict relationship between the87Sr/86Sr
of wine and that derived by the labile fraction of the soil. The observed shift of the87Sr/86Sr toward higher values for wines with respect to the
expected labile fraction values is likely explained by the role played by the CaCO3fraction of the soils where the single vine plants are
embed-ded. InFig. 4we already showed that the carbonatic component is less radiogenic in Sr with respect to the shale one, being also more prone to be mobilised by circulating water and weathering processes. As a consequence the higher is the CaCO3content, the more diluted is the
ra-diogenic Sr composition of the labile fraction (Fig. 4b). This holds true for all the samples offset from the 1:1 correlation line with the only ex-ception of BRO4. Considering the strong heterogeneity of the soil parent material (i.e., bedrocks) the87Sr/86Sr may depend upon the possibility
that the vine roots might sample a volume of soil larger than that we collected. If this is the case a slightly discrepancies in the correlations be-tween wines and labile fractions might occur.
This effect, however, does not deny the existing relationships be-tween the87Sr/86Sr offinal product (micro-vinification) and that of
the geo-pedologic substrata of the harvest points. Thus, the labile frac-tion of the soil still represents the component that strongly controls the supply of Sr at the vine plant via its roots. As discussed in the above paragraph the overall labile fractions have87Sr/86Sr significantly lower than87Sr/86Sr of whole soil from which they were extracted,
ap-proaching to the87Sr/86Sr values of the geological substrata (Table 3).
This argues for a strong control of the geology of the area on the release of Sr and thus on the87Sr/86Sr signature offinal product. This reinforces
the potential use of87Sr/86Sr for tracing the geographic provenance
(i.e., terroir) of the vineyard, suggesting that geology plays an important role in the isotopic characterisation of the agricultural products and their derivate goods.
4. Conclusion
This experimental study showed excellent reproducibility of the
87Sr/86Sr of each micro-vinification during the four harvesting years,
suggesting that the Sr uptake process from the grapevine roots to its final product is time-independent. Thus the87
Sr/86Sr of wines can be safely considered a meaningful parameter to define the characteristics for the terroir typicity of the wine even at a very small scale (single vine).
87Sr/86Sr of wines strongly depends upon the87Sr/86Sr values of the
Sr uptaken by the roots of the vine through the labile (bioavailable) frac-tion of the soil. The lack of 1:1 proporfrac-tion between wine and either the soil or the bedrock underlying the vineyards, confirms that the process of release of the Sr micro-elements is more complex than originally Fig. 6. Plot of the87
Sr/86 Sr of whole soils vs.87 Sr/86 Sr of wines (87 Sr/86 Sr)Mfrom
micro-vinification, the latters are calculated on a great mean using all the data collected for each harvest point through the a–c adjacent sampling points and the four vintage years. Shaded areas refer to different bedrocks characterising the geological substratum of the experimental area. Error bars represent the standard deviation for each mean value (Table 3 of Data in Brief,Braschi et al., 2018) and are within the symbols where not shown. For MLL and PLI meaning see caption ofFig. 1.
Fig. 7. Plot of the87
Sr/86
Sr of labile fraction (bio-available) of soils vs. (87
Sr/86
Sr)Mof wines
from micro-vinification, the latters are calculated on a large mean using all the data collected for each harvest point through the a-c adjacent sampling points and the four vintage years. Plot a) reports the 1:1 correlation line; b) Linear regression model for
87
Sr/86
Sr in labile fraction and wines (87
Sr/86
Sr)M. Internal dashed curves are the
confidence bands defining the area that has a 95% chance of containing the true regression line. External dashed curves represent the prediction band, the area in which 95% of all data points are expected to fall. Error bars represent the standard deviation for each mean value (Table 3 of Data in Brief,Braschi et al., 2018) and are within the symbols where not shown. Normality of distribution for x and y was checked by using normal probability plots, whilst the influence of outliers (not present) was verified by performing also a robust linearfitting. The linear correlation has been calculated using the robust regression method of the add-in program for Microsoft Excel Isoplot 4 (Ludwig, 2008).
thought, possibly depending upon the different nature and mineralogi-cal compositions of bedrocks (i.e.,Marchionni et al., 2013, 2016). The good correlation observed between the87Sr/86Sr of the whole soil and
the labile fraction suggests, anyway, that87Sr/86Sr represents a good
“geologic” tracer for geographic traceability of agricultural products, such as wine, independently by the extension of the area of provenance, and they well relate to geo-pedological substrata of the vineyards.
The87Sr/86Sr of the bulk soil is widely different from that of the
cor-respondent labile (bio-available) fraction. The latter shows87Sr/86Sr close to the values of the dominant carbonatic bedrock from which the soils developed. With respect to them, some whole soils appear to have higher87Sr/86Sr values likely controlled by a higher proportion of the clay component, which being rich in87Rb produces higher amounts
of radiogenic87Sr through time.
In conclusion, the87Sr/86Sr of wine, not depending by the
bio-vegetative life of the vine, represents a perfect geochemical tracer to guarantee the geographic origin and the production terroir of wines and possibly of other agriculture products. In addition, the small-scale
87Sr/86Sr variation can allow to further characterise, if properly
com-bined with other geographic parameters (such as best exposure, moder-ate wmoder-ater stress etc.), an area suitable for cultivation of specific vineyards of recognised quality, usually defined in the oenological liter-ature as cru, to yield wines of the best quality.
Acknowledgement
This experimental study represents a pilot project developed in co-operation with“Barone Ricasoli” winery, one of the more ancient and traditional winemakers of Tuscany, who allowed access to their vineyards allowed thefine check of agricultural procedures and the sampling of grapes from the same vineyards producing the DOC and DOCG wines known worldwide. Financial support was provided by CRA (now CREA) internal fundings (ISSUOVINO project) through a fel-lowship granted to thefirst author (E.B.), and by the Dipartimento di Scienze della Terra of the Università degli Studi di Firenze forfinancial support for the isotopic analyses. We would like to thank Massimiliano Biagi for providing continuous support during thefield work, Maurizio Ulivi for the constant and sincere help during measurement and for assisting with the laboratory management, Sergio Pellegrini and Giuseppe Valboa of CREA for the soil laboratory analysis, Paolo Storchi and Rita Perria of CREA for the collaboration in sap and soil sampling, Riccardo Avanzinelli and Lorella Francalanci for suggestions and critical reviews of original draft of the manuscript. Eventually, thefirst and the senior authors greatly appreciated Dr. Elena Boari for pioneering the set up of Sr-purification and Sr-isotope composition determination in wines.
References
Aceto, M., Robotti, E., Oddone, M., Baldizzone, M., Bonifacino, G., Bezzo, G., Di Stefano, R., Gosetti, F., Mazzucco, E., Marengo, M., 2013.A traceability study on the Moscato wine chain. Food Chem. 138, 1914–1922.
Avanzinelli, R., Boari, E., Conticelli, S., Francalanci, L., Guarnieri, L., Perini, G., Petrone, C.M., Tommasini, S., Ulivi, M., 2005.High precision Sr, Nd, and Pb isotopic analyses using the new generation Thermal Ionisation Mass Spectrometer ThermoFinnigan Triton-Ti®. Periodico di Mineralogia 74, 147–166.
Baxter, M.J., Crews, H.M., Dennis, M.J., Goodall, I., Anderson, D., 1997.The determination of the authenticity of wine from its trace element composition. Food Chem. 60, 443–450.
Braschi, E., Conticelli, S., Manetti, P., Pandeli, E., 2012.Geodiversità litologica della Toscana (Lithological geodiversity of Tuscany). In: Pratesi, G. (Ed.), Il Museo di Storia Naturale dell'Università degli Studi di Firenze– Le collezioni mineralogiche e litologiche. 4. Fi-renze University Press, pp. 200–227 (2012).
Braschi, E., Marchionni, S., Bucelli, P., Priori, S., Costantini, E.A.C., Bollati, A., Mattei, M., Conticelli, S., Tommasini, S. (2013). 3.4. Analisi Isotopiche. In: Oltre la zonazione (Be-yond zonation) - tre anni di studio al Castello di Brolio E.A.C. Costantini (edt.), 102–116. Edizioni Polistampa, Firenze.
Braschi, E., Marchionni, S., Priori, S., Casalini, M., Tommasini, S., Natarelli, L., Buccianti, A., Bucelli, P., Costantini, E.A.C., Conticelli, S., 2018.From vine to wine: data on87
Sr/86
Sr from rocks and soils as a geologic and pedologic characterisation of vineyards. Data in Brief 18 xx-xxx.
Caruso, M., Galgano, F., Castiglione Morelli, M.A., Viggiani, L., Lencioni, L., Giussani, G., Favati, F., 2012.Chemical profile of white wines produced from ‘Greco bianco’ grape variety in different Italian areas by nuclear magnetic resonance (NMR) and conventional physicochemical analyses. J. Agric. Food Chem. 60, 7–15.
Castiñeira, M.M., Brandt, R., Von Bohlen, A., Jakubowski, N., 2001.Development of a pro-cedure for the multi-element determination of trace elements in wine by ICP–MS. Fresenius J. Anal. Chem. 370, 553–558.
Censi, P., Saiano, F., Pisciotta, A., Tuzzolino, N., 2014.Geochemical behaviour of rare earths in Vitis vinifera grafted onto different rootstocks and growing on several soils. Sci. Total Environ. 473, 597–608.
Christoph, N., Roßmann, A., Vörkelius, S., 2003.Possibilities and limitations of wine au-thentication using stable isotope and meteorological data, data banks and statistical tests. Part 1: wines from Franconia and Lake Constance 1992 to 2001. Mitteilungen Klosterneuberg 53, 23–40.
Christoph, N., Baratossy, G., Kubanovic, V., Kozina, B., Roßmann, A., Schlicht, C., 2004. Pos-sibilities and limitations of wine authentication using stable isotope ratio analysis and traceability. Part 2: wines from Hungary, Croatia and other European countries. Mitteilungen Klosterneuberg 54, 155–169.
Conticelli, S., 1998.Effects of crustal contamination on ultrapotassic magmas with lamproitic affinity: mineralogical, geochemical and isotope data from the Torre Alfina Lavas and Xenoliths, Central Italy. Chem. Geol. 149, 51–81.
Conticelli, S., Bortolotti, V., Principi, G., Laurenzi, M.A., Vaggelli, G., D'Antonio, M., 2001. Pe-trology, mineralogy and geochemistry of a mafic dyke from Monte Castello, Elba Is-land, Italy. Ofioliti 26, 249–262.
Conticelli, S., Avanzinelli, R., Ammannati, E., Casalini, M., 2015a.The role of carbon from recycled sediments in the origin of ultrapotassic igneous rocks in the Central Mediter-ranean. Lithos 232, 174–196.
Conticelli, S., Boari, E., Burlamacchi, L., Cifelli, F., Moscardi, F., Laurenzi, M.A., Ferrari Pedraglio, L., Francalanci, L., Benvenuti, M.G., Braschi, E., Manetti, P., 2015b. Geochem-istry and Sr-Nd-Pb isotopes of Monte Amiata Volcano, Central Italy: evidence for magma mixing between high-K calc-alkaline and leucititic mantle-derived magmas. Ital. J. Geosci. 134, 266–290.
Costantini, E.A.C., Bucelli, P., 2014.Soil and terroir. In: Kapur, Selim, Erşahin, Sabit (Eds.), Soil Security for Ecosystem Management - Mediterranean Soil Ecosystems 1 Springer Briefs in Environment, Security, Development and Peace. Mediterra-nean Studies No. 3 vol. 8. Springer-Verlag, Heidelberg– New York – Dordrecht – London, pp. 97–133.
Di Paola-Naranjo, R.D., Baroni, V., Podio, N.S., Rubinstein, H.R., Fabani, M.P., Badini, R.G., Inga, M., Ostera, H.A., Cagnoni, M., Gallegos, E., Gautier, E., Peral-García, P., Hoogewerff, J., Wunderlin, D.A., 2011.Fingerprints for main varieties of Argentinean wines: terroir differentiation by inorganic, organic, and stable isotopic analyses coupled to chemometrics. J. Agric. Food Chem. 59, 7854–7865.
Durante, C., Baschieri, C., Bertacchini, L., Cocchi, M., Sighinolfi, S., Silvestri, M., Marchetti, A., 2013.Geographical traceability based on87Sr/86Sr indicator: afirst approach for
PDO Lambrusco wine from Modena. Food Chem. 141, 20779–22787.
Durante, C., Baschieri, C., Bertacchini, L., Bertelli, D., Cocchi, M., Marchetti, A., Manzini, D., Papotti, G., Sighinolfi, S., 2015.An analytical approach to Sr isotope ratio determination in Lambrusco wines for geographical traceability purposes. Food Chem. 173, 557–563.
Durante, C., Bertacchini, L., Bontempo, L., Camin, F., Manzini, D., Lambertini, P., Marchetti, A., Paolini, M., 2016.From soil to grape and wine: variation of light and heavy ele-ments isotope ratios. Food Chem. 210, 648–659.
Garcia, M., Gallego, P., Daverede, C., Ibrahim, H., 2001.Effect of three rootstocks on grape-vine (Vitis vinifera L.) cv. Negrette, grown hydroponically. I. Potassium, calcium and magnesium nutrition. S. Afr. J. Enol. Vitic. 22, 101–103.
Hölzl, S., Horn, P., Roßmann, A., Rummel, S., 2004.Isotope-abundance ratios of light (bio) and heavy (geo) elements in biogenic tissues: methods and applications. Anal. Bioanal. Chem. 378, 270–272.
Horn, P., Schaaf, P., Holbach, B., Hölzl, S., Eschnauer, H., 1993.87
Sr/86
Sr from rock and soil into vine and wine. Zeitschrift für Lebensmittel Untersuchung und Forschung 196, 407–409.
IUSS, Working Group WRB, 2014.World reference base for soil resources 2014 interna-tional soil classification system for naming soils and creating legends for soil maps. World Soils Resources Reports. 106. FAO, Rome.
Kelly, S., Heaton, K., Hoogewerff, J., 2005.Tracing the geographical origin of food: the ap-plication of multi-element and multi-isotope analyses. Trends Sci. Technol. 16, 555–567.
Kment, P., Mihaljevič, M., Ettler, V., Šebek, O., Strnad, L., Rohlová, L., 2005.Differentiation of Czech wines using multielement composition–a comparison with vineyard soil. Food Chem. 91, 157–165.
Ludwig, K.R., 2008.Isoplot/Ex Version 3.7. A Geochronological Toolkit for Microsoft Excel. Berkeley Geochronological Centre (Special publication).
Marchionni, S., Braschi, E., Tommasini, S., Bollati, A., Cifelli, F., Mulinacci, N., Mattei, M., Conticelli, S., 2013.High precision87
Sr/86
Sr analyses in wines and their use as geological fingerprint for tracing geographic provenance. J. Agric. Food Chem. 61, 6822–6831.
Marchionni, S., Buccianti, A., Bollati, A., Braschi, E., Cifelli, F., Molin, P., Parrotto, M., Mattei, M., Tommasini, S., Conticelli, S., 2016.Conservation of87
Sr/86
Sr isotopic ratios during the winemaking processes of‘Red’ wines to validate their use as geographic tracer. Food Chem. 190, 777–785.
Marengo, E., Aceto, M., 2003.Statistical investigation of the differences in the distribution of metals in Nebbiolo-based wines. Food Chem. 81, 621–630.
Marroni, M., Moratti, G., Costantini, A., Conticelli, S., Benvenuti, M.G., Pandolfi, L., Bonini, M., Cornamusini, G., Laurenzi, M.A., 2015.Geology of the Monte Amiata region, south-ern Tuscany, central Italy. Ital. J. Geosci. 134, 171–199.
Martini, E., Comina, C., Priori, S., Costantini, E.A.C., 2013.A combined geophysical-pedological approach for precision viticulture in the Chianti hills. Boll. Geofis. Teor. Appl. 54.
Melluso, L., Conticelli, S., D'Antonio, M., Mirco, N., Saccani, E., 2003.Petrology and miner-alogy of wollastonite-melilite-bearing pyrometamorphic rocks from Colle Fabbri and Ricetto, Central Apennines, Italy. Am. Mineral. 88, 1287–1299.
Mihaljevič, M., Ettler, V., Šebek, O., Strnad, L., Chrastný, V., 2006.Lead isotopic signatures of wine and vineyard soils-tracers of lead origin. J. Geochem. Explor. 88, 130–133.
MiPAF, 2000.Metodi di analisi chimica de suolo. Collana di metodi analitici per l'agricoltura. Franco Angeli Editore. Italy Fasciculated, Milano.
Nilsson, M., Duarte, I.F., Almeida, C., Delgadillo, I., Goodfellow, B.J., Gil, A.M., Morris, G.A., 2004.High-resolution NMR and diffusion-ordered spectroscopy of port wine. J. Agric. Food Chem. 52, 3736–3743.
Ogrinc, N.K.I.J., Košir, I.J., Spangenberg, J.E., Kidrič, J., 2003.The application of NMR and MS methods for detection of adulteration of wine, fruit juices, and olive oil. A review. Anal. Bioanal. Chem. 376, 424–430.
Petrini, R., Sansone, L., Slejko, F.F., Buccianti, A., Marcuzzo, P., Tomasi, D., 2015.The
87
Sr/86
Sr strontium isotopic systematics applied to Glera vineyards: a tracer for the geographical origin of the Prosecco. Food Chem. 170, 138–144.
Priori, S., Costantini E.A.C. (2013). Soil mapping at the winery scale, and thematic maps. In: Oltre la zonazione (Beyond zonation) - tre anni di studio al Castello di Brolio E.A.C. Costantini (edt.), 102–116. Edizioni Polistampa, Firenze.
Priori, S., Martini, E., Andrenelli, M.C., Magini, S., Agnelli, A.E., Bucelli, P., Biagi, M., Pellegrini, S., Costantini, E.A.C., 2013.Improving wine quality through harvest zoning and combined use of remote and soil proximal sensing. Soil Sci. Soc. Am. J. 77, 1338–1348.
Priori, S., Bianconi, N., Costantini, E.A., 2014.Canγ-radiometrics predict soil textural data and stoniness in different parent materials? A comparison of two machine-learning methods. Geoderma 226, 354–364.
Roßmann, A., 2001.Determination of stable isotope ratios in food analysis. Food Rev. Int. 17, 347–381.
Skogley, E.O., Dobermann, A., 1996.Synthetic ion-exchange resins: soil and environmen-tal studies. J. Environ. Qual. 25, 13–24.
Stafilov, T., Karadjova, I., 2006.Methods for determination and speciation of trace ele-ments in wine. Int. J. Pure Appl. Chem. 1, 273–305.
Tescione, I., Marchionni, S., Mattei, M., Tassi, F., Romano, C., Conticelli, S., 2015.A compar-ative87
Sr/86
Sr study in Red and White wines to validate its use as geochemical tracer for the geographical origin of wine. Procedia Earth Planet. Sci. 13, 169–172.
Thiel, G., Danzer, K., 1997.Direct analysis of mineral components in wine by inductively coupled plasma optical emission spectrometry (ICP-OES). Fresenius J. Anal. Chem. 357, 553–557.
Thiel, G., Geisler, G., Blechschmidt, I., Danzer, K., 2004.Determination of trace elements in wines and classification according to their provenance. Anal. Bioanal. Chem. 378, 1630–1636.
Thirlwall, M.F., 1991.Long-term reproducibility of multicollector Sr and Nd isotope ratio analysis. Chem. Geol. 94, 85–104.
Van Leeuwen, C., Seguin, G., 2006.The concept of terroir in viticulture. J. Wine Res. 17, 1–10.
Vaudour, E., Costantini, E.A.C., Jones, G.V., Mocali, S., 2015.An overview of the recent ap-proaches to terroir functional modelling, footprinting and zoning. Soil 1, 287–312.
Viggiani, L., Castiglione Morelli, M.A., 2008.Characterization of wines by nuclear magnetic resonance: a work study on wines from the Basilicata region in Italy. J. Agric. Food Chem. 56, 8273–8279.
Vinciguerra, V., Stevenson, R., Pedneault, K., Poirier, A., Hélie, J.F., Widory, D., 2016. Stron-tium isotope characterization of wines from Quebec, Canada. Food Chem. 210, 121–128.
