Geoderma 334 (2019) 99–112

Contents lists available at ScienceDirect

Geoderma
journal homepage: www.elsevier.com/locate/geoderma

Scale effect of terroir under three contrasting vintages in the Chianti Classico T
area (Tuscany, Italy)
⁎
Simone Prioria, , Sergio Pellegrinia, Rita Perriab, Sergio Puccionib, Paolo Storchib,
Giuseppe Valboaa, Edoardo A.C. Costantinia
a
CREA, Research Centre for Agriculture and Environment, Firenze, Italy
b
CREA, Research Centre for Viticulture and Enology, Arezzo, Italy

A R T I C LE I N FO A B S T R A C T

Handling editor: A.B. McBratney In viticulture, terroir is a concept used to explain the specific combination and interaction of natural and human
Keywords: factors that provides distinctive characteristics to the wine. The role of soil and geology on wine characteristics is
Viticulture debated and sometimes considered less important than either climate or the human component.
Soil The present study, performed on one of the largest farms of the “Chianti Classico” wine district (Tuscany,
Cru Italy), focused on the effect of terroir on wine characteristics using two different zoning scales. At a broader
Proximal soil sensing scale, called macro-terroir (MT), the experimental vineyards were selected based on lithology, soilscape, mor-
Hydrology phology, and mesoclimate. Each vineyard was then subdivided at a detailed scale into two homogeneous zones
for soil features, the Basic Terroir Units or Unité Terroir de Base (UTB). The study was conducted during three
different vintages (2012, '13 and '14), in vineyards located in four different MT, which are representative of large
parts of the Chianti Classico wine district. The vineyards were surveyed by proximal sensors, namely electro-
magnetic induction sensor (EMI) and gamma-ray spectroscopy to characterize soil spatial variability and to
define two homogeneous areas (UTB) of about 2 ha in each MT. The UTB differed for some soil features, mainly
texture, gravel content, soil depth, available water capacity, and internal drainage. The weather for the three
vintages was very different e during the growing season, which was very dry and hot in 2012, moderately wet
and warm in 2013 and chilly and very wet in 2014. Grape harvest, wine-making and six-month ageing were
carried out separately for the different UTB, using the same methodology. Mixed-design analysis of the variance
of several must and wine features demonstrated that MT played the major role on must pH, as well as total
acidity, glycerine content and colour intensity of the wine. The climate of the vintage played a stronger role than
MT on the content of must malic acid, as well as polyphenols, anthocyanins and dry extract of the wine. Blind
wine sensory analysis performed for all vintages showed significant differences between wines from the different
UTB, in particular for colour intensity and wine aroma, but the differences between UTB within each MT were
not stable over the three contrasting vintages, being less pronounced in the most humid vintage (summer 2014).
This study demonstrates that characteristics of pedo-geological landscapes can be used for a wine district
zoning, while a more detailed soil mapping, leading to UTB identification, is needed for differentiating particular
wine characteristics.

1. Introduction identity and a tight connection with its place of origin (Salette et al.,
1998; Bucelli et al., 2011; Costantini et al., 2016).
The concept of “terroir” has long been used in viticulture to describe To date, the terroir concept is not easily understood and remains
the relationships between the sensory attributes of wine and the geo- one of the most debated issues in the world of wine, because of the large
graphical territory from which it is derived (Vaudour, 2002; Deloire variety of interacting natural and human factors, on which there is not
et al., 2005). The terroir distinction has progressively gained relevance always agreement. These factors include climate, soil, topography,
in wine marketing as a tool to endorse the quality of wines and improve grapevine cultivar, viticultural and oenological practices, which to-
their competitiveness and profitability on the international markets. gether create unique and distinctive characteristics in the wine from a
Probably more than many other foods and beverages, wine has a strong given place that is perceived and recognizable by consumers and

⁎
Corresponding author.
E-mail address: [email protected] (S. Priori).

https://doi.org/10.1016/j.geoderma.2018.07.048
Received 26 January 2018; Received in revised form 23 July 2018; Accepted 28 July 2018
Available online 04 August 2018
0016-7061/ © 2018 Elsevier B.V. All rights reserved.
S. Priori et al. Geoderma 334 (2019) 99–112

experts (Fischer et al., 2016; Costantini et al., 2016; Barham, 2003; (Martínez-Casasnovas and Concepción Ramos, 2009) or land prepara-
Vaudour et al., 2015; Wilson, 1998). According to the definition tion activities for vineyard planting, including levelling, bedrock
adopted by the International Organization of Vine and Wine (OIV), the crashing, and deep ploughing. These activities can reduce soil depth,
viti-vinicultural terroir refers to “an area in which collective knowledge disturb the natural soil profile, and can increase the short-range spatial
of the interactions between the identifiable physical and biological variability of the soil across a vineyard (Costantini et al., 2015).
environment and applied viti-vinicultural practices develops, providing Therefore, at the farm or “within-vineyard” scale, soil characteristics
distinctive characteristics for the products originating from this area” are credited as major terroir components (Bramley et al., 2011a, 2011b;
(OIV, 2010). Tardaguila et al., 2011; Priori et al., 2013a, 2013b). Soil physical
Beside the human factor, which plays the most important role properties, such as texture, structure, internal drainage, and soil depth,
through viticultural and oenological practices, the natural factors that influence soil temperature, soil/water relationships, and root develop-
are most important in the expression of terroir may vary depending on ment, which subsequently influences water and chemical nutrition of
the spatial scale. At a “regional scale”, macroclimate in interaction with the vine (Morlat and Bodin, 2006). Chemical nutrition is critical for
the grapevine cultivar is likely to be most important (Jones et al., grapevine development and berry production, but there is often a weak
2005). At a “within-region” and “wine district” scale, the interaction or no relationship between the soil nutrient status and wine quality, due
between mesoclimate, topography and geology might be the dominant to factors regulating plant nutrient uptake, including soil nutrient dy-
factors driving grapevine performance and grape peculiarities (Nicholas namics and availability, soil water content, vine rooting patterns and
et al., 2011; Priori et al., 2014b; Ramos et al., 2015). Topography antagonism between nutrients (Garcia et al., 2001; Mackenzie and
greatly affects mesoclimate by altitude, proximity to large water bodies, Christ, 2005). More attention is given to soil water status and water
aspect, and slope. It is well known that during grape ripening, the uptake conditions, which are confirmed as key factors of terroir
spatial variability of day and night temperatures plays a very important (Costantini et al., 2013; Marciniak et al., 2013; Bonfante et al., 2015).
role in separating wine producing areas characterized by different High grape quality, especially for red wine, is often associated with
grape maturation, aroma, and colouration (Tonietto and Carbonneau, mild water deficit, which in rainfed vineyards is related to a complex
2004). The role of geology on wine peculiarities is much more debated. interaction between climate (rainfall, evapotranspiration), soil hy-
While many authors (Vaudour, 2002; Van Leeuwen and Seguin, 2006; drology (water holding capacity, internal drainage) and the density and
Costantini et al., 2012; Bonfante et al., 2015) assert that vineyard distribution of vine roots (Bonfante et al., 2011; Costantini et al., 2013;
geology contributes significantly to wine peculiarities, other authors Deloire et al., 2005; Dry, 2016; Marciniak et al., 2013; Brillante et al.,
consider the effects of soil and bedrock on grape and wine negligible 2016).
(Matthews, 2016). A recent approach to investigate and manage soil spatial variability
Although vineyard geology is widely hypothesized to contribute in vineyards consists of mapping homogeneous management zones
significantly to wine typicity, there is still little scientifically-based using proximal and remote sensing methods, which provide increased
knowledge on how this connection is established and which specific resolution and accuracy of soil spatial characterization, while reducing
geological parameters are involved. Some significant examples are the the sampling costs, and improving management of wine quality in re-
relationship of Chablis wines with Kimmeridgian limestone, or that of lation to soil features (Acevedo-Opazo et al., 2008; Taylor et al., 2009;
Beaujolais wines with granite, which along with many others are taken Bramley et al., 2011a, 2011b; Bonfante et al., 2015; Vaudour et al.,
as a crucial for the expression of wine typicality (Van Leeuwen and 2017; Tardaguila et al., 2017). Some authors refer to homogeneous
Seguin, 2006). management zones as the “Basic Terroir Units” or “Unité Terroir de
Matthews (2016) asserts that the word “terroir”, interpreted as Base” (UTB), to underline the concept that each of them represent the
geology and soil associated, is often abused and not preceded by sci- smallest useful area for vineyard management, in which the natural
entific discoveries of soil- and rocks-derived flavours or characteristics factors (soil, geology, climate) are homogeneous and have uniform ef-
of the wine. He also writes that “grapevines have next to no interaction fect on vine biology and wine quality (Deloire et al., 2005). Key ques-
with rocks” (Matthews, 2016) which supports Maltman (2008), who tions in this approach, also reported by Bramley (2016a, 2016b), in-
wrote that a direct rock geochemical influence on wine is undemon- clude: does variation in soil properties have a functional impact on
strated and scientifically impossible. The author disapproves of the use grape and wine composition? At what scale are these effects expressed?
of direct connection between wine flavours and rocks, like “minerality, How stable are these effects across vintages characterized by con-
rocky flavour”, “quartz taste”, and “smell of graphite”, the latter for trasting climatic conditions?
wines produced in schists with graphite in Priorat (Maltman, 2008). On The present study was conducted in one of the largest and most
the other hand, he also affirms that rock and soil features can indirectly renowned wineries in the “Chianti Classico” district (Tuscany, Italy),
influence the bio-chemical pathway of elements during grape growing the Barone Ricasoli farm. The objective of the research is to evaluate
and vinification, and then wine peculiarities. Characteristic isotopic the effect of terroir on wine quality at two different zoning scales: i) the
speciation and transfer can been followed from bedrock to soil, vine, “macro-terroir” (MT) level, as defined according to geology, soilscapes,
grape, and wine, thus allowing wine origin to be traced at a very de- morphology and climate, and ii) the UTB level, based on the division of
tailed scale (Braschi et al., 2018). each MT into homogeneous sub-zones according to soil proximal sen-
Certainly, the role of geology might be expressed indirectly. The sing survey and soil physical-hydrological properties (texture, gravel
bedrock geology determines the relief and the landforms of an area, and content, depth, available water capacity).
is a key factor in soil genesis. The nature of the bedrock, along with its
physical status (colour, hardness, compaction, presence of planes of 2. Materials and methods
weakness) and degree of weathering, greatly influence soil physical and
hydrological properties, which influence root development and water 2.1. Study variety and area
uptake. Moreover, the bedrock geochemistry affects soil pH, nutrient
supply, and balance, which are crucial for vine growth and grape The grapevine cultivar studied was Sangiovese, the most important
composition (Kodur, 2011; Retallack and Burns, 2016). for “Chianti Classico” and other high quality wines of Central Italy, such
However, the relationship between soil properties and the under- as “Brunello di Montalcino”, “Vino Nobile di Montepulciano” and
lying bedrock may not always be so clear. For instance, soils can de- “Morellino di Scansano”. The Sangiovese cv. can express a wide variety
velop from allochthonous parent material, such as aeolian sediments, of wine peculiarities, due to its high responsiveness to the environ-
colluvial depositions, or human transported materials (Dazzi et al., mental factors (Bucelli et al., 2004; Dalla Marta et al., 2010; Ducci,
2009). In other cases, the relationship can be broken by erosion 2013; Mattii et al., 2005; Scalabrelli et al., 2001). Moreover, it is very

100
S. Priori et al. Geoderma 334 (2019) 99–112

Fig. 1. Main soilscapes of the Chianti Classico DOCG area, as reported by Pollini et al. (2014) and Amato and Valletta (2017, modified). SAND: hills and low
mountains on feldspathic sandstone (Macigno formation); CALC: hills on clayey-calcareous flysch (Monte Morello formation); MS: hills on flysches mainly made by
marls, shales and calcarenites; MAR_s: low hills on marine sandy deposits; MAR_c: low hills and plains on marine clayey and silty deposits.

sensitive to water stress and it is identified as an anisohydric cultivar composition and different period, from Pliocene (ancient terraces) to
(Poni et al., 2007). Anisohydric and near-anisohydric cultivars, like Holocene. In this area, the Ligurian Units and the upper part of Tuscan
Sangiovese, Sirah, etc., continue to transpire even when soil water unit are generally on the top of the hills (400–600 m a.s.l.) or along
content diminishes because of poorer stomatal adjustment capacity slope at high-medium elevation (300–500 m a.s.l.), whereas marine
than isohydric and near-isohydric cultivars like Montepulciano, Ca- deposits and fluvial terraces are situated beneath 300–350 m a.s.l. From
bernet Sauvignon, and Grenache (Schultz, 2003; Palliotti et al., 2014). a morphological point of view, the slopes of Ligurian Units and Tuscan
This strategy makes the anisohydric plants less water-efficient and more Units are usually steeper than the others, because of the nature of hard
affected by soil water shortage (Poni et al., 2007). bedrock (limestone, sandstone and marls).
A 3-year project was carried out on the “Barone Ricasoli” estate, one For this work, we defined as “macro-terroir” (MT) a large area,
of the widest and oldest farms in the Chianti Classico wine district, characterized by similar lithology, morphology, and climate, which
where the Chianti wine “formula” was formalized in the year 1872 delimits a soilscape. A soilscape is defined as an area which groups soils
(Simone et al., 2015). The 1200 ha farm is in the northern part of the having functional relationships and similar pedogenesis and that can be
province of Siena with 230 ha dedicated to viticulture. The vineyards, represented at 1:250,000 scale (Finke et al., 1998). We defined as
planted at elevations that span from 180 to 490 m a.s.l., with different “Basic Terroir Units” (or Unité Terroir de Base, UTB; Morlat, 2001) sub-
slopes and aspects, display characteristic features of the typical terroir areas within MT of about 2 ha in size, characterized by homogenous soil
of the Chianti Classico district (Fig. 1). The farm includes the four main features (texture, stoniness, soil depth, available water capacity, etc.).
geological units of the Chianti Classico district, which are: i) the Li- The UTB size was based upon the capacity of the cellar tanks (9 tons),
gurian Unit: a succession of clayey‑carbonate sedimentary rocks of which is a suitable amount to produce a commercial wine brand
Cretaceous-Eocene period; ii) the upper part of Tuscan unit: a succes- (around 10.000 bottles).
sion of clay-calcareous-marls, covered by a thick layer of feldspathic A total of seven vineyards were selected (Fig. 2), across 4 MT
sandstone (Macigno Formation) of Oligocene period; iii) the marine showing the following characteristics:
deposits: silty-clay and sandy-gravelly marine deposits of Pliocene - Calcareous flysch (CALC): MT developed on clayey-calcareous
period; iv) the fluvial terraces: fluvial deposits of different textural rocks (Monte Morello Formation), situated on regular slopes between

101
S. Priori et al. Geoderma 334 (2019) 99–112

Fig. 2. Experimental vineyards plotted on geological map 1:10.000. Legend: all-recent alluvial deposits; AT-alluvial terraces; PLIs: marine sands with conglomerate
lenses; SIL: shales of Ligurian Unit; MLL: clay-calcareous flysch of Ligurian Unit; MAC: feldspathic sandstone; STO: limestones, marls and shales of Tuscan Unit.

400 and 450 m a.s.l. According to the maps of soilscapes reported by - Fluvial terraces (FLUV): MT situated at lower elevation than the
Pollini et al., 2014, this formation covers around 35% of the total area others (250–320 m a.s.l.) and characterized by soils developed on an-
of Chianti Classico DOCG. The regional soil map (scale 1:250,000, cient fluvial deposits (Pliocene-Quaternary). This MT is not reported in
Gardin and Vinci, 2006) reports a soilscape formed by the association of Fig. 1, since the fluvial terraces are discontinuous in the study area.
shallow and very gravelly soils, with loam, silty clay loam or clay loam Although this MT is not as widespread as the previous, several ancient
texture, usually very rich in calcium carbonates, classified as Calcaric fluvial terraces are present in the Chianti area. In the ancient fluvial
Cambisols, Calcaric Leptosols and Calcaric Regosols. terraces of central Tuscany, the regional soil map shows soils with
- Sandstone (SAND): MT characterized by poorly weathered soils different weathering and pedogenetic development, usually calcareous
developed on feldspathic sandstone (Macigno del Chianti Formation), and with loamy and loamy clay texture, varying from Calcaric Fluvisols,
situated on regular slope or high-plains of the hills between 430 and Calcaric Cambisols, Eutric Cambisols, and Cutanic Luvisols (Gardin and
470 m a.s.l. This formation covers around 25% of the total area of Vinci, 2006).
Chianti Classico DOCG (Priori et al., 2013a, 2013b). The soils are Only one important MT of the Chianti Classico DOCG, which covers
characterized by sandy or loamy-sandy texture, usually high stoniness around 15% of the total area of the district (MS, Fig. 1) and is char-
and very low content or absence of calcium carbonate (< 1%). The soils acterized by soils developed on sequences (flysches) of marls, shales
of this macro-terroir are classified as Eutric Regosols, Eutric Cambisols and calcarenites, was not investigated during this work.
and Cutanic Luvisols (Gardin and Vinci, 2006). The studied MT are not only present in the Chianti area, but also in
- Marine deposits (MAR): MT developed on marine sands and other territories of the Tuscany region. In the Province of Siena, where
gravelly-sands of early Pliocene period, situated on slopes around other wines are produced from the same Sangiovese variety, two MT
300–350 m a.s.l. Such MT is very common in central Tuscany, where reported in this paper, namely CALC and SAND, correspond to the
deep sandy deposits accumulated in a marine basin during the Pliocene “Natural Terroir Units” n.6: loamy soils with frequent stoniness, high
period (4.8–4.2 Ma ago) (Martini et al., 2011). These deposits cover calcium carbonate developed on calcareous flysches; and n.9: sandy-
around 18% of the total area of the Chianti Classico DOCG (Priori et al., loamy soils with frequent stoniness, low calcium carbonate, developed
2013a, 2013b). The soils of this MT can have different degree of ped- on sedimentary rocks rich in sand (arenites) (Priori et al., 2014b).
ogenesis, due to the balance between erosion and stability during the Climatic data were acquired in each MT by weather stations during
Quaternary period. The soil types span from Calcaric Regosols, Calcaric the three years of the project (2012–2014). Mean daily temperature,
Arenosols, Calcaric Cambisols, and in some cases, Cutanic Luvisols minimum and maximum daily temperature, as well as daily precipita-
(Gardin and Vinci, 2006). The texture is variable between sandy-loam tion were collected. A modified Winkler index (WImod) was calculated
to clay-loam and the calcium carbonate content is usually medium following the method of Amerine and Winkler (1944), changing the end
(around 15%). of the time range from 31st October to 30th September. The

102
S. Priori et al. Geoderma 334 (2019) 99–112

modification was necessary because Sangiovese grape harvest in this grapes from each UTB, which was the volume of grapes needed by the
area always occurs between the second half of September and the be- winery to fill in their steel tanks. Each MT was then subdivided into two
ginning of October. CALC and SAND were comparable for average UTB, for a total of 8 UTB. Small areas, characterized by deep and fertile
temperature and WImod, whereas MAR and FLUV showed slightly lower soils due to the downslope accumulation, waterlogging, or exceptional
mean temperatures. Night temperatures were lower in the area closer to soil erosion, were excluded from the delineation of UTB. The significant
the valley bottom (FLUV and MAR) than at higher altitude (CALC and differences of the soil proximal sensing data between the couples of
SAND). The “Cool night index” (CI, Tonietto and Carbonneau, 2004) UTB within each MT were investigated following the method of Taylor
showed temperate nights in CALC and SAND (14.1–14.7 °C), and cool et al. (2007). The authors proposed that two classes have sufficient
nights in MAR (12.9–13.8 °C) and FLUV (11.9–12.6 °C). All the ex- variation if:
perimental vineyards had comparable summer precipitation (P7/8),
2
calculated from 1st of July to 31th of August. YClass1 − YClass2 ≥ ( σkrig × 1.96) × 2 (1)
All the experimental vineyards were cultivated with the Sangiovese
where YClass1 and YClass2 is the mean of the selected soil variable in
cultivar. The age of the vineyards was similar (12–16 years) and the
UTB1 and 2.
trellis system was the simple spurred cordon with vertical shoot posi-
tioning. The vigour was managed with moderate shoot topping and the
soil was tilled during early summer. Grapevine density was similar, 2.3. Soil analysis and monitoring
ranging between 2 × 0.75 m (CALC and SAND) and 2.1 × 0.8 m (MAR
and FLUV), which means 6200–6600 vines/ha. The experimental Within each UTB, three representative soil profiles of about 1.5 m
blocks received the same viticultural treatments during the growing were dug. The location of digging followed the variability captured by
season. None of the vineyards were irrigated, like most of DOCG vi- the soil survey with proximal sensors (see below). Soil profile descrip-
neyards from Chianti Classico and other wine districts from central tion followed the national and international guidelines for soil de-
Italy. scription (Costantini, 2007; Jahn et al., 2006) and classification ac-
cording to the Word Reference Base for Soil Resources (WRB; IUSS
2.2. Mapping Basic Terroir Units (UTB) Working Group WRB, 2014). Soil stoniness was estimated in the field
using a reference frame of 30 × 30 cm.
The experimental vineyards were surveyed by soil proximal sensing Soil samples were collected from the several genetic horizons of
to obtain high-detail maps of soil spatial variability and to delimit Basic each soil profile. The samples were air-dried, sieved to 2.0 mm and
Terroir Units (or Unitè Terroir de Base, UTB) within each MT. analysed for physical and chemical properties. Soil texture was de-
The proximal sensors used were: i) EM38-Mk2 electromagnetic in- termined by the X-ray/sedimentation method, using a Micromeritics
duction sensor (Geonics Ltd., Ontario, Canada) and ii) “The Mole”, Sedigraph III analyser (Andrenelli et al., 2013).
gamma-ray spectroradiometer (Soil Company, The Netherlands). Total organic C (TOC) and total N (Ntot) were measured by dry
EM38-Mk2 measures the soil apparent electrical conductivity (ECa) combustion with a ThermoFlash 2000 CN soil analyser, after removal of
across two depth ranges of 0–75 and 0–150 cm, approximately carbonates by HCl 10%. The total equivalent CaCO3 content was cal-
(McNeill, 1990). The ECa measurements are influenced by several soil culated from the difference between the total C measured by dry
properties, including clay content (Sudduth et al., 2005; Morari et al., combustion in the untreated soil (mineral C + organic C) and in the
2009), gravel content (Morari et al., 2009; Priori et al., 2013a), soil HCl-treated soil (organic C) (Sequi and De Nobili, 2000). The active
moisture and water availability (Cousin et al., 2009; Zhu et al., 2010), lime was determined according to the Drouineau method, based on a
bulk density (Taylor et al., 2009), salinity (Triantafilis et al., 2000), and 2 h treatment with 0.1 M ammonium oxalate and following back-titra-
soil depth (Priori et al., 2013b). tion with 0.1 M KMnO4 (Loeppert and Suarez, 1996). Soil pH was
“The Mole” spectroradiometer continuously measures the gamma- measured potentiometrically in a 1:2.5 soil–water suspension. Electrical
ray natural emission coming from the first 30–40 cm of the soil and conductivity was measured in a 1:2 soil–water filtered extract after 2 h
rocks, through a Cesium Iodide scintillator crystal (van Egmond et al., shaking and overnight standing. Soil cation exchange capacity (CEC)
2008). Proximal gamma-radiometrics has been used to survey topsoil and exchange bases were analysed with the BaCl2-triethanolamine
features, such as texture (Pikki et al., 2013; Priori et al., 2014a), gravel (pH 8.2) method. The amounts of Ca, Mg, K and Na in the extracts were
content (Priori et al., 2014a), potassium (Castrignanò et al., 2012), and quantified by flame atomic absorption spectrometry, using an Agilent
organic carbon (Dierke and Werban, 2013, Priori et al., 2016). For the SpectrAA 220FS spectrometer (Gessa and Ciavatta, 2000). The soil
survey, the gamma-ray spectroradiometer was mounted on the back of water retention was determined using a pressure plate apparatus (Klute,
a tractor, whereas EM38-Mk2 was inserted in a non-metallic chariot, 1986). Water retention at field capacity (FC) and wilting point (WP)
which was pulled by the same tractor through a 2 m long shaft. The (−33 and −1500 kPa matric potential, respectively) were measured on
latter is needed to avoid the influence of metallic mass of the tractor on the < 2 mm soil fraction. The values of moisture content at FC and WP
ECa measurements. The sensors were supplied with GPS and rugged PC were corrected for gravel content according to Gardner (1986).
for data-logging. A proximal soil survey was carried out in April 2012,
when the soil moisture was between 10 and 20% in volume. Soil sen- 2.4. Wine characteristics
sing was performed continuously around every 8 vine rows, and the
data were interpolated across the whole vineyard areas using ordinary The grapes from each UTB were carried to the farm winery, where
kriging (OK). The parameters of OK, namely lag size, number of lags, they were separately vinified in stainless steel tanks, using the same
and maximum range were selected in order to minimize the estimation oenological techniques, as follows: i) crushing and destemming; ii)
error (mean kriging variance, σkrig). Two clusters in each MT were addition of SO2 (15 mg·L−1) and dry selected yeast (200 mg·L−1); iii)
delimited according to the k-means clustering (Fig. 3). The variables 15-day maceration at controlled temperature punching the cap down
used for clustering were: apparent electrical conductivity ECa1 and six times a day until 3 Babo units; iv) the wines were poured after a soft
ECa2, total count of gamma-ray (TC), as well as slope and aspect, ob- pressing into 5 hectolitres oak barrel (tonneaux) for a 6-months ageing.
tained by a digital elevation model with a detail of 10 m. Because of the All these operations were repeated every vintage of the trial (2012-'13-
manual grape-harvest, the cluster areas were simplified to delineate the '14). The grape musts were analysed for sugar content, pH and malic
UTB based upon a compromise between the results of clustering and the acid.
farm needs (Fig. 2). Seven months after grape-harvest, the wines were analysed to assess
A minimum size of 1.5–2 ha was adopted to obtain around 9 tons of the alcohol content, total polyphenols and anthocyanins, total acidity,

103
S. Priori et al. Geoderma 334 (2019) 99–112

Fig. 3. Example of UTB mapping in the vineyard of SAND MT. In the first row, the maps obtained by the proximal sensing: gamma-ray total counts (TC), apparent
electrical conductivity of 0–75 cm and 0–150 cm. In the second row, the maps obtained by DEM: slope and aspect, and the two clusters obtained by k-means
clustering using the previous maps after value standardization. The polygons SAND1 and SAND2 showed the UTB used for the grape harvest.

dry extract, glycerine, and colour intensity. Malic acid, titratable design models were adopted for each must and wine variable, using
acidity, polyphenols, anthocyanins, net dry extract and glycerine were vintage as random factor, MT as fixed factor and UTB nested in MT
determined by FTIR (Fourier Transform Infrared Spectroscopy), using a (hierarchical nested ANOVA). The same methods were used to analyse
wine analyser WineScan FT (FOSS, Denmark) (Bevin et al., 2006). the scores of wine tasting.
Colour intensity (CI) and hue (Hue) were measured according to Glories A multivariate analysis of the oenological data was also carried out
(1984) using an AGILENT (USA) 8453 DAD spectrophotometer. using principal component analysis (PCA), considering as active vari-
Moreover, the wines were evaluated by a panel of 10 wine tasters ables alcohol content, total wine acidity, total polyphenols, total an-
through a “blind tasting”. Sensory analysis was performed to assess thocyanins, dry extract, glycerine, and colour intensity index.
differences among the wines, therefore the evaluation method was Pedological and climatic variables were also plotted in the factor
mainly comparative. The wine tasters gave a score, ranging from 1 to 10 loadings graph (Fig. 5, left) as supplementary variables. These are not
to several wine parameters, giving score 10 to the wine which best included in the calculation of PCA, but they are plotted in the factor
expressed the parameter. The sensory parameters were: colour in- loadings graph according to the correlation with the PCA factors. The
tensity, flavour intensity and balance, structure, acidity, astringency, factor score graph (Fig. 5, right) demonstrates the statistical separation
and persistence. A score was also given to the overall evaluation of the among MT using a multivariate approach.
wine. In addition, the tasters indicated their feeling about the wine in
terms of aroma typology (fruity, floral, spicy, herbaceous), using 0-
3. Results
absent, 1-scarce, 2-medium, and 3-strong.
3.1. Basic Terroir Unit (UTB)
2.5. Statistical analysis
The apparent electrical conductivity (ECa) measured at 0–75 cm
Several statistical approaches were used to investigate the effects of (ECa1) and 0–150 cm (ECa2) depths, ranged between 5 and 45 mS·m−1,
vintage, MT and UTB on oenological parameters and wine tasting generally showing higher values in ECa2 (Table 1). On average, the
scores. Effect of the spatial variability of mesoclimate throughout MT lowest ECa values (around 5 mS·m−1) were observed in SAND2
was investigated by non-parametric correlation (Spearman's ranks) (ECa1 = 5.2 mS·m−1) and SAND1 (ECa1 = 8.7 mS·m−1), both char-
between WImod, CI, P7/8 and wine features. acterized by sandy soils and SAND2 by high stoniness (> 35%,
Significant differences among groups have been tested by one-way Table 2). On average, the ECa of CALC, MAR and FLUV were similar
ANOVA and Fisher's LSD test, using the vintage, the MT, and the UTB as and vary between a mean of about 16 mS·m−1 (FLUV2) and about
grouping factor. Fisher's LSD test determines the significant differences 33 mS·m−1 (MAR1). According to Eq. (1), ECa1 showed significant
between group means in an analysis of variance. Parametric Student's t- differences between the clusters within every MT (Table 1), whereas
test and non-parametric Mann-Whitney U test were run to verify sta- ECa2 showed no significant differences between the clusters of SAND.
tistical difference between the couples of UTB within a MT. To verify Gamma-ray spectroscopy provided very different total counts (TC)
the interactions between the effects vintage, MT, and UTB, mixed according to the MT (Table 1). The mean TC value of soils on calcareous

104
S. Priori et al. Geoderma 334 (2019) 99–112

Table 1 The 8 UTB represented areas very suitable for high quality grape
Mean values of the variables used for clustering and within each UTB, as sim- and characterized by a certain internal homogeneity in terms of soil
plified for grape harvest. In bold, the significant differences, following the features, hydrology and microclimate. The geometrical simplification of
method of mean kriging variance (σkrig) (Taylor et al., 2007). ⁎: aspect is re- the cluster areas in UTB did not show any significant loss of statistical
ported in radiant (rad) and main direction (S: South, SE: South East, SW: South
differences between groups, with the only exception of CALC (Table 1).
West).
The soil profiles of CALC1 and CALC2 were all classified as Skeletic
MT Cluster/UTB Area ECa1 ECa2 TC Slope Aspect⁎ Calcaric Cambisol (Loamic) (IUSS Working Group WRB, 2014) and
characterized by clayey loamy texture, high stoniness and calcium
ha mS·m−1 Bq·kg−1 % rad (direction)
carbonate, as well as moderately low water permeability (Ksat). CALC1
CALC σkrig – 1.6 1.2 17 n.d. n.d. profiles showed significant lower sand content, total calcium carbonate
Cluster1 2.1 17.7 24.0 297 12.0 4.1 (SW) and slightly higher TOC and exchangeable potassium (K) than CALC2
Cluster2 2.4 24.1 29.2 343 8.9 3.5 (SW) (Table 2).
CALC1 2.2 20.5 28.2 316 13.6 3.8 (SW)
The soil profiles of SAND1 and SAND2 were classified as Eutric
CALC2 2.3 22.3 26.5 333 9.3 3.9 (SW)
SAND σkrig – 0.6 0.6 17 n.d. n.d. Cambisol (Arenic) and Skeletic Eutric Cambisol (Arenic), respectively.
Cluster1 2.6 8.7 13.4 671 15.6 3.8 (SW) These soils showed sandy loam texture, high stoniness and perme-
Cluster2 1.3 5.2 11.6 657 6.9 3.9 (SW) ability, as well as very low organic matter and nitrogen content. SAND1
SAND1 2.0 8.2 12.4 669 14.0 3.7 (SW)
showed significant lower stoniness and K content, but higher available
SAND2 1.9 5.7 12.2 666 8.8 3.6 (SW)
MAR σkrig – 2.4 1.7 12 n.d. n.d.
water capacity (AWC) than SAND2 (Table 2).
Cluster1 2.5 32.7 30.8 408 11.1 3.0 (S) MAR1 and MAR2 showed the best differentiation between the UTB
Cluster2 2.6 21.7 22.3 361 18.4 2.2 (SE) of the MT, and their soils were classified as Calcaric Cambisols (Loamic,
MAR1 2.3 32.9 31.0 410 12.8 2.9 (S) Colluvic) and Calcaric Cambisol (Arenic), respectively (IUSS Working
MAR2 2.8 23.4 23.4 358 17.6 2.2 (SE)
Group WRB, 2014). MAR1 was a reddish-brown, deep, well preserved
FLUV σkrig – 0.7 0.7 12 n.d. n.d.
Cluster1 1.9 28.0 32.8 485 10.9 2.9 (S) and strongly structured soil. The texture was variable between loam
Cluster2 1.9 15.6 22.8 379 16.7 3.6 (SW) and clay-loam, with variable content of rounded cobbles and pebbles of
FLUV1 2.0 22.5 28.5 443 11.3 3.1 (S) heterogeneous lithology. Calcium carbonate was low (< 5%), TOC and
FLUV2 1.8 19.0 25.6 404 19.5 3.6 (SW) Ntot were medium and pH was neutral or sub-alkaline (7.5–8.3). MAR2
was the soil situated in the areas of the slopes more sensitive to erosion,
therefore the soil was shallower, less preserved and structured than
flysch (CALC) was 317 ± 40 Bq·kg−1, and the differences between
MAR1. The colour was pale brown or yellowish and the texture was
CALC1 and CALC2 were not significant. The TC value of soils on
sandy or sandy-loam, with a general high content of cobbles and peb-
feldspathic sandstone (SAND) was on average 667 ± 21 Bq·kg−1, with
bles. Calcium carbonate was moderate (10–20%), whereas pH was sub-
no significant differences between SAND1 and SAND2. The only sig-
alkaline. This soil was also characterized by a low content of organic
nificant difference of TC values between the couple of clusters was
matter and nitrogen.
calculated in MAR, and varied between 358 ± 59 Bq·kg−1 (MAR2) and
The UTB FLUV1 and FLUV2 were situated on two different land-
410 ± 35 Bq·kg−1 (MAR1). In FLUV, no significant difference was
forms: the first on gentle slope at the top of the hill (ancient terrace
observed, although TC varied between 404 ± 62 Bq·kg−1 (FLUV2) and
surface), the second one along the slope connecting the ancient to the
443 ± 65 Bq·kg−1 (FLUV1).
recent fluvial terraces. Both soils showed loamy or clay-loamy texture,
Radionuclide concentration (data not reported) followed the same
poor soil structure, moderate calcium carbonate (8–19%) and very low
trend of TC, although the ratio between 40K/238U and 40K/232Th was
content of organic matter and nitrogen. FLUV1 was deeper and showed
slightly higher in the soils of feldspathic sandstone. The high gamma-
significant lower sand, calcium carbonate, stoniness, TOC, nitrogen and
ray emission and 40K radionuclide concentration in these soils was
water permeability than FLUV2. Both the UTB soils showed sub-alka-
probably due to the high percentage of potassium rich minerals, like
line pH (8.1–8.4) and were classified as Calcaric Cambisol (Loamic)
muscovite and feldspars, which are very frequent in such kind of
(IUSS Working Group WRB, 2014).
sandstone (Macigno del Chianti formation). Within of a same parent
material, the spatial variability of gamma-ray TC and radionuclides
concentration is mainly influenced by the topsoil texture and surface 3.2. Weather of the three vintages
stoniness. The relationships between gamma-ray spectroscopy and soil
features in these vineyards are reported in detail by a previous work The weather was variable over the three years of study, as reported
(Priori et al., 2014a). in Fig. 4, with the following trends during the grapevine growing season
(1st April–30th September):

Table 2
Means and standard deviations of soil features analysed in each UTB (3 profiles and, only for clay, sand, and stoniness, 6 augerings), calculated as average values on
soil horizons (0–90 cm) with the only exceptions of TOC and Ntot, measured in Ap horizons (about 0–30 cm). 1: total calcium carbonate; 2: total organic carbon; 3:
total nitrogen; 4: exchangeable potassium, 5: available water capacity; 6: saturated hydraulic conductivity. In bold, significant differences between the UTB couples of
MT, calculated by Student's t-test (p < 0.05).
UTB Clay Sand CaCO31 Stoniness TOC2 Ntot3 K4 AWC5 Ksat6

(g·100 g−1) (m2·m−2) (g·kg−1) (mg·kg−1) (mm·m−1) (mm·h−1)

CALC1 36.0 ± 3.9 16.5 ± 2.7 38.7 ± 9.1 30.8 ± 10.9 6.6 ± 0.1 0.7 ± 0.1 132 ± 12 114 ± 27 2.1 ± 0.4
CALC2 33.6 ± 5.8 21.6 ± 1.2 57.0 ± 1.7 36.2 ± 10.5 5.4 ± 0.1 0.8 ± 0.1 107 ± 7 94 ± 10 2.3 ± 1.0
SAND1 9.8 ± 3.5 59.8 ± 5.5 1.7 ± 0.6 19.2 ± 2.7 4.4 ± 0.6 0.4 ± 0.1 74 ± 4 113 ± 27 19.7 ± 5.9
SAND2 10.0 ± 2.7 61.0 ± 5.2 2.6 ± 0.6 36.5 ± 6.6 3.9 ± 1.2 0.3 ± 0.1 170 ± 23 64 ± 11 22.8 ± 14.7
MAR1 35.3 ± 5.9 33.5 ± 6.8 13.3 ± 3.2 8.1 ± 3.7 8.4 ± 1.1 0.9 ± 0.1 149 ± 67 139 ± 3 7.5 ± 7.7
MAR2 19.2 ± 4.3 51.3 ± 5.8 26.7 ± 2.1 21.8 ± 10.6 5.8 ± 2.1 0.6 ± 0.2 86 ± 30 103 ± 21 11.6 ± 2.6
FLUV1 31.5 ± 3.8 31.4 ± 5.1 12.3 ± 6.4 7.0 ± 3.4 3.3 ± 1.1 0.4 ± 0.1 126 ± 28 79 ± 8 2.5 ± 0.6
FLUV2 25.5 ± 7.9 39.8 ± 6.0 28.3 ± 8.4 27.2 ± 10.2 6.1 ± 1.9 0.7 ± 0.2 100 ± 36 67 ± 19 9.1 ± 1.3

105
S. Priori et al. Geoderma 334 (2019) 99–112

Fig. 4. Monthly mean temperature (a) and precipitation (b) during the grapevine growing season (1st of April–30th of September) of the three vintages. Black line
showed the long term data (20 years) of the area.

Table 3
Mean elevation of the four MT and the climatic data of the three years. 1: WImod, Winkler index modified (calculated according Amerine and Winkler, 1944, using the
time range from 1st April to 30th September, instead than 31st of October); 2: CI: Cool night Index (Tonietto and Carbonneau, 2004); P7/8, Summer precipitation
during grape veraison period, measured from 1st of July to 31th August.
MT h 2012 2013 2014

1 2 3 1 2 3
WImod CNI P7/8 WImod CNI P7/8 WImod1 CNI2 P7/83

(m a.s.l.) Σ°C °C (mm) Σ°C °C (mm) Σ°C °C (mm)

CALC 435 1897 14.3 58 1790 14.3 52 1523 14.7 136

SAND 480 1878 14.3 58 1799 14.1 52 1514 14.4 136
MAR 325 1761 13.8 42 1567 12.9 70 1511 13.4 127
FLUV 315 1677 12.6 40 1474 11.9 67 1456 12.6 127

- 2012: mean temperatures were close to long-term average in spring, following dates: 23–25 September 2012, 1–3 October 2013, 29
but much higher in June, July and August. Mean daily temperature September–3 October 2014.
was higher than 22 °C from the middle of June to the beginning of The mild and humid summer of 2014 provided grape musts with
September. The modified Winkler Index (WImod) resulted about higher acidity (pH 3.19 on average) and malic acid, while vintage 2012,
1800 °C on average (Table 3). Summer was extremely dry, with no characterized by high temperature and scarce precipitation during
precipitation between 12th of June and 30th of August. The 31st of summer, produced grape musts with low malic acid and wines with the
August and 1st of September were characterized by two intense lowest content of polyphenols and anthocyanins. Summer 2013, which
rainstorms (around 50 mm each). can be considered representative of a good vintage for Sangiovese
- 2013: April was warm, but with a sudden temperature decrease in cultivar, provided, on average, the highest values of polyphenols, an-
the second half of May. Summer was warm and long, with the only thocyanins, and dry extract in the wine.
exception of a short period of temperature decrease at the end of Spatial variation of WImod, CI, and P7/8 did not show any statistical
June. WImod was about 150 °C lower in 2013 than 2012. correlations with the must and wine parameters in 2012. In 2013 CI
Precipitation was higher than the average in July, with 6 rainy was correlated with must sugar and wine anthocyanins (rs 0.71 and
events (> 3 mm) well distributed between June to August. 0.73), whereas P7/8 showed significant inverse correlation with must
September was slightly warmer and drier than the average. sugar and pH (rs −0.83 and −0.85), as well as wine anthocyanins and
- 2014: Mean temperature was slightly lower than average during glycerine (rs −0.79 and −0.73). In 2014, only P7/8 showed significant
most of the summer, with the only exception being a short warmer correlation with wine dry extract (rs 0.77).
period at the beginning of June. WImod was around 1500 °C. Discriminant analysis on vintage effect on wine distinguished the
Precipitation frequency was very high and above the average for vintages for all oenological variables, except alcohol and glycerine
summer, with 11 rainy days between July and August and 6 in content (results not reported). The same outcome was obtained by the
September. Fisher's LSD test (Table 4).
Principal component analysis (PCA), carried out on the wine fea-
The MT CALC and SAND, which were located near the top of the tures (alcohol, total acidity, polyphenols, anthocyanins, dry extract,
hills with south aspect, showed the highest heat sum (WImod, Table 3), glycerine, and colour intensity), showed clusters associated with MT
whereas the MT FLUV showed the lowest WImod and CI, because of the (Fig. 5) more evidently than those associated with vintages. SAND and
thermic inversion induced by the river valley. Precipitation in the dif- MAR wines, especially, showed strong differentiation from CALC and
ferent MT were comparable (Table 3). FLUV for Factor 1 (more related to polyphenols, anthocyanins, alcohol,
dry extract and glycerine), whereas SAND and MAR were better dif-
ferentiated between them by Factor 2 (related to colour intensity and
3.3. Must and wine analytical features
total acidity).
The best differentiation among the wines occurred in vintage 2012,
The contrasting climate conditions between the experimental years
the driest and warmest, whereas in the other years only CALC wines
resulted in significant vintage-to-vintage differences in the grape and
were well discriminated (Fig. 5).
wine quality. The harvest date was when the grape in each vintage
Analysing the three vintages pooled together, the discriminant
reached average values of 220 g∙L−1 of sugars, corresponding to the

106
S. Priori et al. Geoderma 334 (2019) 99–112

Table 4
Results of Fisher's LSD tests after one-way analysis of variance, using as grouping factor vintage, MT and UTB, respectively. Letters showed the groups significantly
different for p < 0.05.
Must Final wine

after 6 months ageing

Sugar pH Malic acid Alcohol Tot. acid. Polyphen. Anthocyan. Dry extr. Glycer. Colour int.

−1 −1 −1 −1 −1 −1 −1
g∙L g∙L %vol g∙L mg∙L mg∙L g∙L g∙L

Vintage
2012 220 3.27 (b) 0.7 (c) 13.2 6.0 (b) 1548 (c) 191 (b) 27.2 (b) 6.3 8.4
2013 225 3.28 (b) 1.5 (b) 13.5 6.6 (a) 2040 (a) 245 (a) 28.7 (a) 6.6 8.1
2014 223 3.19 (a) 2.3 (a) 13.4 6.8 (a) 1695 (b) 222 (a) 26.8 (b) 6.6 8.3

MT
CALC 226 (a) 3.25 (b) 1.2 (c) 13.6 (a) 6.8 (a) 1760 (b) 245 (a) 28.2 (a) 6.5 (a) 9.7 (a)
SAND 223 (ab) 3.34 (c) 2.0 (a) 13.4 (ab) 5.7 (b) 1684 (b) 209 (b) 27.5 (ab) 6.6 (a) 6.6 (c)
MAR 216 (b) 3.14 (b) 1.2 (c) 12.9 (b) 6.8 (a) 1653 (b) 203 (b) 26.8 (b) 5.9 (b) 8.0 (b)
FLUV 225 (ab) 3.25 (b) 1.6 (b) 13.5 (ab) 6.7 (a) 1948 (a) 221 (ab) 28.0 (a) 6.9 (a) 8.8 (ab)

UTB
CALC1 223 (ab) 3.25 (abc) 1.36 (cd) 13.4 (ab) 6.6 (ab) 1674 (bc) 243 (a) 28.1 (ab) 6.6 (abc) 9.4 (ab)
CALC2 230 (a) 3.26 (abc) 1.05 (d) 13.8 (a) 6.9 (a) 1846 (ab) 247 (a) 28.2 (a) 6.5 (abc) 9.9 (a)
SAND1 221 (ab) 3.31 (bc) 1.91 (ab) 13.3 (ab) 6.0 (bc) 1623 (c) 200 (b) 27.3 (ab) 6.6 (ab) 6.2 (d)
SAND2 226 (a) 3.38 (c) 2.13 (a) 13.6 (a) 5.5 (c) 1745 (bc) 217 (ab) 27.7 (ab) 6.5 (abc) 6.9 (cd)
MAR1 211 (b) 3.13 (c) 1.24 (cd) 12.7 (b) 6.8 (a) 1676 (bc) 209 (ab) 26.5 (b) 5.8 (c) 8.6 (abc)
MAR2 220 (ab) 3.16 (c) 1.24 (cd) 13.2 (ab) 6.7 (ab) 1639 (bc) 196 (b) 27.1 (ab) 6.1 (bc) 7.5 (bcd)
FLUV1 227 (a) 3.28 (ab) 1.57 (bc) 13.6 (a) 6.3 (ab) 1957 (a) 222 (ab) 27.8 (ab) 7.1 (a) 9.5 (a)
FLUV2 223 (ab) 3.22 (bc) 1.54 (bc) 13.4 (ab) 7.0 (a) 1939 (a) 220 (ab) 28.1 (a) 6.7 (ab) 8.0 (abcd)

analysis demonstrated that wines produced in the four MT were sig- calcareous flysch (CALC) and ancient fluvial terraces (FLUV). The wines
nificantly different (p < 0.05) for colour intensity, glycerine and total made from grapes produced on the ancient fluvial terraces had sig-
acidity. Squared Mahalanobis distances showed the highest differences nificantly higher polyphenols content, whereas grapes produced in the
between CALC and SAND (13.3, p < 0.01) and between MAR and calcareous flysch provided significant higher anthocyanins and colour
SAND (11.3, p < 0.01), whereas there were not significant differences intensity.
between CALC and MAR, and CALC and FLUV. When considering UTB as grouping variable instead of MT, the
Fisher's LSD test (Table 4) showed significantly lower wine total discrimination between groups in general did not increase. Fisher's LSD
acidity and colour intensity in the wines from grapes produced on test showed several significant differences among the wines of the
sandstone (SAND), and lower dry extract and glycerine content in the different UTB. On average, during the three experimental vintages,
wines made from grapes on marine deposits (MAR). The latter also CALC2 wines provided the highest alcohol and colour intensity, SAND1
showed lower wine colour intensity as compared to wines from wines the lowest colour index, SAND2 the lowest total acidity, MAR1

Fig. 5. PCA results. On the left, factor loadings of the active and supplemental (Suppl.) variables of PCA. On the right, factor scores of the PCA. The dashed polygons
separate the four MT. The supplemental variables are pedological and climatic features not included in the calculation of PCA, but plotted in the graphs, according to
the correlation with the PCA factors. Active variables: Alcohol, TAc-total acidity, Pol-polyphenols, Ant-anthocyanins, DExtr-dry extract, Gly-glycerine, Col-colour
intensity/Supplemental: TC: gamma-ray total count, ECa1 and ECa2: apparent electrical conductivity 0–75 cm and 0–150 cm, Sk-stoniness, AWC: Available water
capacity, Ca: total calcium carbonate, Clay, Sand, h: elevation, WImod: Winkler index modified (1 April–30 September), CI: Cool night index, P7/8: total precipitation
in July and August. On the right, factor scores of the PCA.

107
S. Priori et al. Geoderma 334 (2019) 99–112

Table 5
Results of mixed-design analysis of variance of the main must and wine features, using MT as fixed factor, vintage as random factor, and UTB as nested design in MT.
In bold, significant values (p < 0.05).
Variable % variability due to: F

MT Vintage MT × vintage UTB (MT) Error MT Vintage MT × vintage UTB (MT)

Effect Fixed Random Random Fixed Fixed Random Random Fixed

df 3 2 6 4 8 3 2 6 4

Must Must sugar 24.2 4.6 31.5 14.9 24.9 1.5 0.4 1.7 1.2
Must pH 48.4 15.9 19.1 6.2 10.5 5.1 2.5 2.4 1.2
Must malic acid 17.8 75.8 3.7 1.5 1.1 9.6 61.3 4.3 2.7
Wine Total acidity 40.5 25.1 15.8 9.1 9.5 5.1 4.8 2.2 1.9
Polyphenols 20.0 64.8 4.7 4.5 5.9 8.4 41.0 1.1 1.5
Anthocyanins 23.4 45.4 20.7 2.6 7.9 2.3 6.6 3.5 0.7
Dry extract 19.0 44.8 25.7 3.0 7.5 1.5 5.2 4.6 0.8
Glycerine 43.5 4.3 26.6 6.4 19.2 3.3 0.5 1.9 0.7
Colour intensity 56.0 1.1 30.0 11.2 1.7 3.7 0.1 23.4 13.1

the lowest alcohol content, and FLUV2 wines the highest total acidity. The results of the wine tasting are summarized in Table 6. This table
On the other hand, Student's t-test and Mann-Whitney's U test did reports the results of the Fisher's LSD test after one-way ANOVA for
not show significant differences between the couples of UTB within a each vintage, using UTB as the grouping factor and tasters as replicates.
MT, with the only exception of significant higher total acidity in FLUV2 The results show that the differentiation of the wines, especially the
(6.9 g·l−1) than in FLUV1 (6.3 g·l−1) during the three years. differentiation between the couples of UTB within a same MT, de-
The results of mixed-design analysis of variance (Table 5), using MT creased from 2012, the driest and warmest vintage, to 2014, the wettest
as fixed effect, vintage as random effect, and UTB nested in MT, showed and coldest year. The only exceptions were FLUV1 and 2, that tended to
that: increase the differentiation in 2013 and 2014 for colour intensity in
2013 and for fruity notes in 2014 (Fig. 6).
- The variability of must pH, wine total acidity, glycerin and colour SAND2 and SAND1 generally obtained the highest score among all
intensity were better explained by MT than climate variability of the wines, in all the vintages, for flavour and fruity notes, but SAND2
different vintages. constantly outperformed SAND1 (Fig. 6). In 2012, also MAR1 and 2
- The content of must malic acid, wine polyphenols, anthocyanins and were very different in terms of flavour, acidity and general score,
dry extract were more influenced by the vintage, although MT whereas they became similar in 2013 and, in 2014. On the other hand,
variability played an important role for anthocyanins and dry ex- CALC1 and CALC2 never showed any significant differences in tasting.
tract.
- UTB nested in MT explained very low variance, comparable with the 4. Discussion
variance of the error, with the only exception of colour intensity.
The results of this work demonstrated that, although climate of the
The F-test demonstrated that the wine produced in the different MT vintage was extremely important for determining wine peculiarities,
were significantly different (p < 0.05) for must pH and malic acid, the role of terroir, at both scales (MT and UTB) was fundamental and
wine total acidity, and polyphenols. Vintage showed a strong significant stable over the years for several wine peculiarities.
effect for must malic acid and polyphenols, and lower but still sig- According to the analysis of variance for must and wine features
nificant effect on anthocyanins and dry extract. Differentiation of wine (Table 5), the effect of MT was stronger than that of vintage climate for
colour intensity was mainly due to the interaction between MT and must pH, wine acidity, glycerine and colour intensity. Macro-terroir
vintage, though UTB within each MT also showed a significant influ- played a strong role also in the aroma of the wine, as demonstrated by
ence. the wine tasting analysis (Table 7).
The climate conditions of the vintage were particularly important
3.4. Wine sensory analysis for must malic acid, polyphenols, anthocyanins and dry extract. On the
other hand, the interaction between vintage and MT significantly af-
The blind sensory analysis of the wines produced in the 8 UTB fected the same wine variables (Table 4). Spatial variability of meso-
provided results with very high standard deviation. This is also ob- climate within a single vintage did not seem to effect the wine, since
servable in the results of mixed-design analysis of variance of wine only few significant correlations between climatic indices (WImod, CI,
tasting parameters, where most of the variance percentage is explained and P7/8) and wine features were observed in 2013 (anthocyanins and
by the error (Table 6). glycerine) and in 2014 (dry extract).
This outcome can be due to the general high quality of the studied The subdivision of two different UTB within a same MT, mainly
wine, however, the standardization and reliability of wine sensory based on homogeneous soil physical and hydrological features, seemed
analysis are very complicated issues and still subject to discussion to play an important role only during dry summers, like in 2012 and, to
(Rodríguez Donate et al., 2017; Cicchetti, 2017). Nevertheless, some a lesser extent, in 2013. The geometrical simplification of the cluster
remarkable wine peculiarities due to the terroir effect were recogniz- areas in UTB, needed to facilitate the grape harvest and to satisfy the
able and stable throughout the three vintages. The interaction between minimal amount of grape for winery fermentation tanks, did not sig-
MT and vintage showed significant (p < 0.05) differentiation among nificantly decrease the variance between the groups of the variables
wines for colour, flavour intensity, floral aroma, body and general used for clustering, with the only exception of CALC1 and 2 (Table 1).
evaluation. The effect of UTB enhanced the discrimination of the wines The CALC vineyards exhibited short-range soil spatial variability,
for colour, flavour intensity and fruity notes. On the other hand, her- mainly due to the stoniness percentage. For this reason, it was not
baceous and spicy notes showed higher relationship with the vintage possible to simplify the areas of the CALC clusters without loss of be-
than with the MT or UTB. tween-groups variability. In this case, the separated grape-harvest

108
S. Priori et al. Geoderma 334 (2019) 99–112

Table 6
Results of mixed-design analysis of variance of the wine tasting indices, using MT as fixed factor, vintage as random factor, and UTB as nested design in MT. In bold,
significant values (p < 0.05).
Tasting parameter % variability due to: F

MT Vintage MT × vintage UTB (MT) Error MT Vintage MT × vintage UTB (MT)

Effect Fixed Random Random Fixed Fixed Random Random Fixed

df 3 2 6 4 232 3 2 6 4

Colour 15.0 4.1 8.7 4.6 67.5 3.4 1.4 5.0 4.0
Flavour 0.2 4.3 5.3 4.8 85.3 0.1 2.4 2.4 3.3
Fruity 2.5 1.5 3.6 5.2 87.2 1.4 1.3 1.6 3.5
Floral 0.3 1.1 5.8 2.0 90.8 0.1 0.6 2.5 1.3
Herbaceous 0.1 8.0 1.2 3.3 87.5 0.2 20.3 0.5 2.2
Spicy 0.6 5.9 1.2 1.2 91.2 0.9 14.7 0.5 0.7
Body 1.6 3.8 11.6 0.8 82.2 0.3 1.0 5.5 0.6
Acidity 4.3 4.9 4.4 1.2 85.3 1.9 3.3 2.0 0.8
General eval. 0.6 0.5 17.6 1.8 79.4 0.1 0.1 8.6 1.3

should be done in small and scattered areas to highlight the UTB effect the total acidity of wines was around −15% of the average, and
on wine (Bramley et al., 2011a, 2011b). This could be possible only lower colour intensity. Ricci Alunni (2004) described the wines
through selective harvest managed by GPS and digital maps. produced on sandy non calcareous soils as elegant, scented, but
From the oenological and wine tasting results, the wine produced in scarce in colour. The inverse relationship between soil pH and wine
the different MT and UTB showed the following features: pH has been also found by other authors (Retallack and Burns,
2016). Lower acidity of the wines produced in SAND terroir, could
- Clayey-calcareous soils developed on Cretaceous calcareous flysch be due to two main causes: high content of potassium and very low
(CALC), one of the most representative MT of Chianti Classico wine content of calcium carbonate in the soil. Several authors (Morris
district (Fig. 1), provided wines with general higher alcohol, total et al., 1983; Mpelasoka et al., 2003) reported that high potassium
anthocyanins, dry extract and colour intensity than average. A si- content in soil, due to natural causes or to fertilization, tends to
milar study (Priori et al., 2013a, 2013b) was carried out in other increase must and wine pH. The same authors reported a negative
vineyards of the farm on clayey-calcareous flysch, during the vin- effect of high potassium on wine colour, which is lighter. High
tage 2010. The wines showed similar high colour intensity (from 9.1 concentration of potassium in grape juice tends to decrease the
to 11.6), alcohol (13.6–14.2%vol) and polyphenols concentration of free acids, such as tartaric acid (Kodur, 2011).
(1823–2004 mg·L−1). In a previous study, also Scalabrelli et al. During wine fermentation, potassium tends to bind with tartaric
(2001) reported higher anthocyanins, total acidity and dry extract acid, with a consequent precipitation of potassium bitartrate (Kodur,
than the average on wines produced from grapes cultivated on 2011). This process causes tartaric acid of the wine to decrease, and
calcareous flysch. Wine tasting confirmed that such wines have hence the increase of wine pH. The direct relationship between
generally higher colour intensity and general high fruity flavour and content of potassium in soil and in grape must has been reported in
medium-high acidity and body. These characteristics corroborate many other studies (e.g., Freeman and Kliewer, 1983; Chan and
the review of Italian terroir and wine features of Ricci Alunni Fahey, 2011), however it was not tested for this study.
(2004), who reported high alcohol and colour intensity as main
features of the wines produced on calcareous-clayey soils. The dif- The wine tasting showed higher flavour intensity and fruity and
ferences between the wines produced in CALC1 and CALC2 were not floral notes in SAND2 than in SAND1 in all the vintages. The causes of
significant. Indeed, the geometrical simplification of CALC UTB the positive effect of the SAND2 soil features on wine flavour should be
areas caused the loss of the very detailed soil spatial variability analysed in more detail to understand which parameters are involved.
within the vineyard. Similar to the findings of González-Barreiro et al. (2015), it is likely that
- Loamy-sand soils, developed on feldspathic sandstone (SAND), higher drainage and soil porosity of SAND2, as well as lower grapevine
which are very common at the higher altitude of the Chianti hills, vigour contributed to the grape aroma precursors. These UTB were
characterized the wine with light colour intensity and low acidity. mainly differentiated by the slope and for the ECa1, which was sig-
The grapes in our trial had higher must pH and malic acid, whereas nificantly lower in SAND2 and indicated a significant higher soil

Fig. 6. Spider graphs reporting the means of the most significant wine tasting parameters (colour, flavour intensity, and fruity notes) for each UTB in the three
vintages. The axis scale was standardized 0 to 10 for all the parameters, setting the maximum value of each parameter for that vintage to 10.

109
S. Priori et al. Geoderma 334 (2019) 99–112

Table 7
Results of Fisher's LSD tests after one-way ANOVA for each vintage, using UTB as grouping factor and tasters as replicates. Letters showed the groups significantly
different for p < 0.05. In bold, significant differences (p < 0.05) between the UTB couples of each MT.
Vintage UTB Flavour Taste Gen. score

Colour Intensity Fruity Floral Vegetal Spicy Body Acidity

2012 CALC1 9.0 (ab) 8.7 (ab) 2.1 (abc) 2.1 (ab) 1.5 (a) 1.9 (a) 8.5 (a) 8.0 (abc) 8.3 (abc)
CALC2 9.6 (a) 8.3 (ab) 2.1 (abc) 1.9 (ab) 1.3 (a) 1.5 (a) 8.4 (a) 6.9 (cd) 7.2 (cd)
SAND1 6.7 (e) 8.1 (ab) 1.9 (bc) 2.0 (ab) 1.5 (a) 1.6 (a) 7.9 (a) 7.2 (bc) 8.6 (ab)
SAND2 7.1 (de) 8.9 (a) 2.6 (a) 2.4 (a) 1.5 (a) 1.9 (a) 8.3 (a) 6.0 (d) 8.2 (abc)
MAR1 7.6 (cde) 7.2 (b) 1.7 (c) 1.8 (b) 1.8 (a) 1.6 (a) 7.3 (a) 7.8 (bc) 6.7 (d)
MAR2 8.0 (cd) 9.0 (a) 2.5 (ab) 2.2 (ab) 1.2 (a) 1.6 (a) 8.6 (a) 8.7 (a) 8.9 (a)
FLUV1 8.5 (bc) 7.9 (ab) 1.8 (c) 2.2 (ab) 1.9 (a) 1.6 (a) 7.2 (a) 8.5 (ab) 7.9 (abcd)
FLUV2 8.8 (ab) 8.4 (ab) 2.1 (abc) 2.1 (ab) 1.5 (a) 1.8 (a) 7.2 (a) 8.4 (abc) 7.5 (bcd)
2013 CALC1 8.6 (ab) 7.7 (ab) 2.8 (a) 2.7 (a) 1.7 (a) 1.9 (ab) 8.5 (ab) 7.3 (a) 8.6 (ab)
CALC2 8.2 (ab) 8.6 (a) 2.8 (a) 2.5 (ab) 1.7 (a) 2.5 (ab) 9.1 (a) 7.0 (a) 9.7 (a)
SAND1 5.8 (d) 6.1 (c) 1.5 (b) 1.5 (c) 2.1 (a) 1.8 (ab) 6.3 (d) 7.4 (a) 5.5 (d)
SAND2 6.1 (d) 7.1 (ab) 2.4 (a) 2.2 (ab) 1.7 (a) 2.2 (ab) 6.8 (cd) 6.9 (a) 6.6 (cd)
MAR1 7.5 (b) 7.8 (ab) 2.1 (a) 1.9 (bc) 2.3 (a) 2.2 (ab) 7.9 (abc) 7.6 (a) 8.2 (b)
MAR2 6.4 (cd) 7.3 (ab) 2.4 (a) 2.5 (ab) 1.4 (a) 2.6 (a) 6.6 (d) 7.5 (a) 7.8 (bc)
FLUV1 9.8 (a) 7.6 (ab) 2.3 (a) 1.9 (bc) 1.6 (a) 2.4 (ab) 7.7 (b) 8.0 (a) 8.2 (ab)
FLUV2 6.7 (cd) 8.0 (a) 2.4 (a) 2.5 (ab) 1.9 (a) 1.8 (b) 6.6 (cd) 8.0 (a) 6.5 (cd)
2014 CALC1 7.4 (bc) 7.6 (b) 2.1 (abc) 1.8 (ab) 1.4 (a) 2.0 (a) 7.6 (ab) 8.2 (ab) 7.0 (b)
CALC2 7.8 (abc) 8.3 (ab) 2.4 (a) 2.0 (ab) 1.2 (a) 1.6 (a) 7.3 (b) 8.6 (ab) 7.2 (b)
SAND1 7.0 (c) 7.8 (b) 1.8 (bc) 2.2 (ab) 1.4 (a) 1.4 (a) 8.4 (a) 8.4 (ab) 8.1 (ab)
SAND2 8.3 (ab) 9.0 (a) 2.2 (abc) 2.4 (ab) 1.0 (a) 1.7 (a) 8.5 (a) 7.7 (b) 8.8 (a)
MAR1 9.0 (a) 8.2 (ab) 2.3 (ab) 2.2 (ab) 1.0 (a) 1.7 (a) 8.5 (a) 8.8 (a) 8.0 (ab)
MAR2 7.9 (bc) 8.4 (ab) 2.2 (abc) 1.9 (ab) 0.7 (a) 1.5 (a) 7.9 (ab) 8.1 (ab) 7.7 (ab)
FLUV1 8.5 (ab) 9.3 (a) 2.5 (a) 1.6 (b) 0.9 (a) 1.3 (a) 8.8 (a) 8.2 (ab) 8.0 (ab)
FLUV2 8.3 (abc) 8.3 (b) 1.7 (c) 2.0 (ab) 1.1 (a) 1.4 (a) 8.7 (a) 8.6 (ab) 8.3 (ab)

stoniness. west-southward. This province-scale study demonstrated that this

type of terroir provides wines generally higher in alcohol and
- Loamy and loamy-sandy soils developed on marine sands (MAR), polyphenols than the average of the province.
common in a wide area of the southern-central Tuscany, such as the
“Nobile di Montepulciano DOCG” (Priori et al., 2014b, and
Costantini et al., 2012) and “Chianti DOCG” districts (Pollini et al., 5. Conclusions
2014), provide wines with low (MAR1) or medium (MAR2) alcohol,
as well as low glycerine and dry extract. Very similar values of must The four MT studied in this work, which represent soil and litho-
sugar content and total polyphenols were reported in Sangiovese logical features typical of many vineyards of the Chianti Classico wine
wines in the southern area of Chianti DOCG by Costantini et al. district, produced wines with different peculiarities through vintages
(2009 and 2010). Scalabrelli et al. (2001) reported that wines made with contrasting climate. The parameters that better showed MT effect
by Sangiovese cv. grapes, cultivated on marine sandy deposits in were must pH, wine total acidity, glycerine and colour intensity.
Chianti Classico DOCG area, were characterized by lower antho- Climate of the vintage instead played a major role on the content of
cyanins and dry extract, as well as higher total acidity, than the must malic acid, polyphenols, anthocyanins, and dry extract.
average Chianti wines. Water availability seems to play a key role in In general, the vineyards on clay-calcareous flysch (CALC) produced
this MT since MAR needs a moderate water deficit to produce wines wines with the highest colour intensity, alcohol, anthocyanins and dry
recognizable by high fruity-floral aroma. The two UTB of this MT extract, whereas those on feldspathic sandstone (SAND) produced
showed significant differences of ECa1, ECa2 and gamma-ray TC, wines with the weakest colour and the lowest total acidity.
which showed higher values in MAR1. MAR2, characterized by The marine deposits (MAR) wines showed lower dry extract and
sandy-loamy soils, very low fertility, rich in gravel, and with high glycerine than the average, and intermediate colour intensity, whereas
internal drainage, produced wines characterized by general high the ancient fluvial terraces (FLUV) wines showed higher polyphenols,
flavour intensity and higher fruity and floral notes than MAR1. The dry extract, and glycerine content. Wine tasting confirmed the results of
differences between the two UTB disappeared during vintage 2014, analytical data in that CALC provided wines with the strongest colour,
the most humid year, therefore, a separated grape-harvest is useless whereas SAND wines with the lowest colour and acidity.
in these conditions. Differences between the UTB within a same MT were not stable over
- Loamy and clay-loamy soils, calcareous, developed on fluvial ter- the three contrasting vintages. The rainiest summer of 2014 showed the
races (FLUV), made wine richer in alcohol, polyphenols, and gly- weakest tasting differences of the wines produced in the two UTB of a
cerine, as well as high colour intensity. FLUV1 showed significantly same MT. This was mainly due to the primary features discriminating
higher ECa1 and ECa2, mainly due to lower sand content and sto- two UTB within a same MT, which were stoniness, soil depth, water
niness than FLUV2. According to wine tasting, the wines of FLUV1 holding capacity, and water permeability (Ksat). These soil variables are
had higher colour intensity in 2013 and in 2014 higher fruity notes strictly related to plant water nutrition, so they have a major influence
than FLUV2. Very few studies reported the Sangiovese cv. wine fea- on grape quality in dry summer. On the contrary, wet summers like
tures in this type of MT because it is not widespread in this area. The 2014 smooth the differences in wine peculiarities between UTB in a
studies of Bucelli et al. (2010) and Priori et al. (2014b), that ad- same MT, making the costs of a separated grape-harvest unnecessary.
dressed viticultural zoning in the province of Siena, reported some In conclusion, geology, soilscape and climate features, which char-
results of Sangiovese cv. cultivated on calcareous soils of fluvial acterize MT, drive some major wine peculiarities over time, while soil
terraces in Brunello di Montalcino DOCG, approximately 50 km physical-hydrological features, which typify UTB within the same MT,
play a key role on wine distinctiveness, mainly during dry vintages. The

110
S. Priori et al. Geoderma 334 (2019) 99–112

use of a more robust delineation of homogeneous areas, performed Biotechnology. 689. pp. 333–340.
using soil maps obtained by proximal sensors, is fundamental to dis- Bucelli, P., Costantini, E.A.C., Storchi, P., 2010. It is possible to predict Sangiovese wine
quality through a limited number of variables measured on the vines. J. Int. Sci.
criminate UTB within vineyard. On the other hand, selective harvest of Vigne Vin 44 (4), 207–218.
small and scattered UTB within a vineyard is possible only if digital Bucelli, P., Costantini, E.A.C., Barbetti, R., Franchini, E., 2011. Soil water availability in
maps and GPS are used, both for the manual and the mechanical har- rainfed cultivation affects more than cultivar some nutraceutical components and the
sensory profile of virgin olive oil. J. Agric. Food Chem. 59, 8304–8313.
vest. Castrignanò, A., Wong, M.T.F., Stelluti, M., De Benedetto, D., Sollitto, D., 2012. Use of
The outcomes of this study are of particular interest since it was this EMI, gamma-ray emission and GPS height as multi-sensor data for soil character-
variability of natural conditions under which the original “formula” of isation. Geoderma 175, 78–89.
Chan, K.Y., Fahey, D.J., 2011. Effect of composted mulch application on soil and wine
the Chianti wine was conceived. grape potassium status. Soil. Res. 49 (5), 455–461.
Cicchetti, D.V., 2017. Evaluating the value of replicate tastings of a given wine: bio-
Acknowledgements statistical considerations. J. Wine Res. 28 (2), 135–143.
Costantini, E.A.C., 2007. Linee Guida dei Metodi di Rilevamento e Informatizzazione dei
Dati Pedologici. CRA-ABP, Firenze, Italia. 296 p. (In Italian, with English summary)
This work was supported by the “Barone Ricasoli” s.p.a. farm [On-line at. http://abp.entecra.it/soilmaps/ita/downloads.html.
(VignaCRU project). The authors wish to thank the agricultural man- Costantini, E.A.C., Pellegrini, S., Bucelli, P., Storchi, P., Vignozzi, N., Barbetti, R.,
ager of the farm Massimiliano Biagi and co-workers (Fabio Cascella, Campagnolo, S., 2009. Influence of hydropedology on viticulture and oenology of
Sangiovese vine in the Chianti area (Central Italy). Hydrol. Earth Syst. Sci. Discuss. 6,
Marco Cerqua, and Claudio Carapelli) for their support. Many other 1197–1231.
people contributed to VignaCRU project and to this work. In particular, Costantini, E.A.C., Pellegrini, S., Bucelli, P., Barbetti, R., Campagnolo, S., Storchi, P.,
we want to thank Pierluigi Bucelli, Nadia Vignozzi, Stefano Mocali, Magini, S., Perria, R., 2010. Mapping suitability for Sangiovese wine by means of
δ13C and geophysical sensors in soils with moderate salinity. Eur. J. Agron. 33,
Arturo Fabiani, Romina Lorenzetti, Maria Fantappiè, Giovanni L'Abate, 208–217.
Roberto Barbetti, and Nadia Bianconi (CREA-AA, Agriculture and Costantini, E.A.C., Bucelli, P., Priori, S., 2012. Quaternary landscape history determines
Environment, Firenze), Alessandra Zombardo and Marco Leprini the soil functional characters of terroir. Quat. Int. 265, 63–73.
Costantini, E.A.C., Agnelli, A., Bucelli, P., Ciambotti, A., Dell'Oro, V., Natarelli, L.,
(CREA-VE, Viticulture and Enology, Arezzo); Aldo Ciambotti and Pellegrini, S., Perria, R., Priori, S., Storchi, P., Tsolakis, C., Vignozzi, N., 2013.
Valentina Dell'Oro (CREA-VE, Viticulture and Enology, Asti). Moreover, Unexpected relationships between deltaC13 and wine grape performance in organic
the authors want to thank Prof. Mark Sperow, of West Virginia farming. J. Int. Sci. Vigne Vin 47 (4), 269–285.
Costantini, E.A.C., Agnelli, A.E., Fabiani, A., Gagnarli, E., Mocali, S., Priori, S., Simoni, S.,
University, for their suggestions on paper editing. Valboa, G., 2015. Short-term recovery of soil physical, chemical, micro-and meso-
biological functions in a new vineyard under organic farming. Soil 1 (1), 443.
References Costantini, E.A.C., Lorenzetti, R., Malorgio, G., 2016. A multivariate approach for the
study of environmental drivers of wine economic structure. Land Use Policy 57,
53–63.
Acevedo-Opazo, C., Tisseyre, B., Guillaume, S., Ojeda, H., 2008. The potential of high Cousin, I., Besson, A., Bourennane, H., Pasquier, C., Nicoullaud, B., King, D., Richard, G.,
spatial resolution information to define within-vineyard zones related to vine water 2009. From spatial-continuous electrical resistivity measurements to the soil hy-
status. Precis. Agric. 9 (5), 285–302. draulic functioning at the field scale. C. R. Geosci. 341, 859–867.
Amato, V., Valletta, M., 2017. Wine landscapes of Italy. In: Soldati, M., Marchetti, M. Dalla Marta, A., Grifoni, D., Mancini, M., Storchi, P., Zipoli, G., Orlandini, S., 2010.
(Eds.), Lanscapes and Landforms of Italy. World Geomorphological Landscapes, Analysis of the relationships between climate variability and grapevine phenology in
Springer. https://doi.org/10.1007/978/3-319-26194-2. the Nobile di Montepulciano wine production area. J. Agric. Sci. 148 (6), 657–666.
Amerine, M., Winkler, A., 1944. Composition and quality of musts and wines of California Dazzi, C., Papa, G.L., Palermo, V., 2009. Proposal for a new diagnostic horizon for WRB
grapes. Calif. Agric. 15 (6), 493–675. Anthrosols. Geoderma 151 (1–2), 16–21.
Andrenelli, M.C., Fiori, V., Pellegrini, S., 2013. Soil particle-size analysis up to 250 μm by Deloire, A., Vaudour, E., Carey, V.A., Bonnardot, V., Van Leeuwen, C., 2005. Grapevine
X-ray granulometer: device set-up and regressions for data conversion into pipette- response to terroir: a global approach. J. Int. Sci. Vigne Vin 39–4, 149–162.
equivalent values. Geoderma 192, 380–393. Dierke, C., Werban, U., 2013. Relationships between gamma-ray data and soil properties
Barham, E., 2003. Translating terroir: the global challenge of French AOC labeling. J. at an agricultural test site. Geoderma 199, 90–98.
Rural. Stud. 19 (1), 127–138. Dry, P.R., 2016. Understanding the components of terroir. In: Beames, K.S., Robinson,
Bevin, C.J., Janik, L.J., Cozzolino, D., 2006. Development of a rapid “fingerprinting” E.M.C., Dry, P.R., Johnson, D.L. (Eds.), Proceedings of the 16th Australian Wine
system for wine authenticity by mid-infrared spectroscopy. J. Agric. Food Chem. 54 Industry Technical Conference: Adelaide, South Australia, 24–28 July. 2017. The
(26), 9713–9718. Australian Wine Industry Technical Conference Inc., Glen Osmond, S.A., pp. 39–44.
Bonfante, A., Basile, A., Langella, G., Manna, P., Terribile, F., 2011. A physically oriented Ducci, E., 2013. The Qualitative Characterization of ‘Sangiovese’ Grapevine According to
approach to analysis and mapping of terroirs. Geoderma 167–168, 103–117. the Area and Cultivation Conditions (PhD thesis). Department of Agriculture, Food
Bonfante, A., Agrillo, A., Albrizio, R., Basile, A., Buonomo, R., De Mascellis, R., ... Manna, and Environment University of Pisa, Italy. https://etd.adm.unipi.it/theses/
P., 2015. Functional homogeneous zones (fHZs) in viticultural zoning procedure: an available/etd-10242013-104405/.
Italian case study on Aglianico vine. Soil 1 (1), 427. Finke, P.A., Hartwich, R., Dudal, R., Ibáez, J., Jamagne, M., King, D., Montanarella, L.,
Bramley, R.G.V., 2016a. Vineyard variability and terroir – making sense of a sense of Yassoglou, N., 1998. Georeferenced Soil Database for Europe; Manual Of Procedures
place. In: Beames, K.S., Robinson, E.M.C., Dry, P.R., Johnson, D.L. (Eds.), Proceedings Version 1.0. European Communities.
of the 16th Australian Wine Industry Technical Conference: Adelaide, South Fischer, U., Bauer, A., Koschinski, S., Schmarr, H.G., 2016. Terroir in the old and new
Australia, 24–28 July. 2017. The Australian Wine Industry Technical Conference world – what sensory is telling us. In: Beames, K.S., Robinson, E.M.C., Dry, P.R.,
Inc., Glen Osmond, S.A., pp. 39–44. Johnson, D.L. (Eds.), Proceedings of the 16th Australian Wine Industry Technical
Bramley, R.G.V., 2016b. Vineyard variability and terroir – making sense of a sense of Conference: Adelaide, South Australia, 24–28 July. 2017. The Australian Wine
place. In: Beames, K.S., Robinson, E.M.C., Dry, P.R., Johnson, D.L. (Eds.), Proceedings Industry Technical Conference Inc., Glen Osmond, S.A., pp. 30–35.
of the 16th Australian Wine Industry Technical Conference: Adelaide, South Freeman, B.M., Kliewer, W.M., 1983. Effect of irrigation, crop level and potassium fer-
Australia, 24–28 July. 2017. The Australian Wine Industry Technical Conference tilization on Carignane vines. II. Grape and wine quality. Am. J. Enol. Vitic. 34 (3),
Inc., Glen Osmond, S.A., pp. 45–51. 197–207.
Bramley, R.G.V., Ouzman, J., Boss, P.K., 2011a. Variation in vine vigour, grape yield and Garcia, M., Gallego, P., Daverede, C., Ibrahim, H., 2001. Effect of three rootstocks on
vineyard soils and topography as indicators of variation in the chemical composition grapevine (Vitis vinifera L.) CV. Négrette, grown hydroponically. I. Potassium, cal-
of grapes, wine and wine sensory attributes. Aust. J. Grape Wine Res. 17 (2), cium and magnesium nutrition. S. Afr. J. Enol. Vitic. 22 (2), 101–103.
217–229. Gardin, L., Vinci, A., 2006. Carta dei suoli della Regione Toscana in scala 1: 250.000.
Bramley, R.G.V., Ouzman, J., Thornton, C., 2011b. Selective harvesting is a feasible and Available on line at. http://sit.lamma.rete.toscana.it/websuoli/ (Last accessed on 4/
profitable strategy even when grape and wine production is geared towards large 2018).
fermentation volumes. Aust. J. Grape Wine Res. 17 (3), 298–305. Gardner, W.H., 1986. Water content. In: Klute, A. (Ed.), Methods of Soils Analysis. Am.
Braschi, E., Marchionni, S., Priori, S., Casalini, M., Tommasini, S., Natarelli, L., Buccianti, Soc. Agron., Madison, Wisc., pp. 493–544 (Part 1).
A., Bucelli, P., Costantini, E.A.C., Conticelli, S., 2018. Tracing the 87 Sr/86 Sr from Gessa, C., Ciavatta, C., 2000. Complesso di scambio. In: Angeli, F. (Ed.), Metodi di Analisi
rocks and soils to vine and wine: an experimental study on geologic and pedologic Chimica del Suolo. Ministero per le Politiche Agricole e Forestali, Osservatorio
characterisation of vineyards using radiogenic isotope of heavy elements. Sci. Total Nazionale Pedologico e per la Qualità del Suolo, XIII, pp. 1–31.
Environ. 628, 1317–1327. Glories, Y., 1984. La couler des vins rouges. II Partie – mesure, origine et interpretation.
Brillante, L., Bois, B., Léveque, J., Mathieu, O., 2016. Variations in soil-water use by Connaiss. la Vigne du Vin 18, 253–271.
grapevine according to plant water status and soil physical-chemical character- González-Barreiro, C., Rial-Otero, R., Cancho-Grande, B., Simal-Gándara, J., 2015. Wine
istics—a 3D spatio-temporal analysis. Eur. J. Agron. 77, 122–135. aroma compounds in grapes: a critical review. Crit. Rev. Food Sci. Nutr. 55 (2),
Bucelli, P., Storchi, P., Costantini, E.A.C., 2004. The influence of climate and soil on 202–218.
viticultural and enological parameters of Sangiovese grapevines under non-irrigated IUSS Working Group WRB, 2014. World reference base for soil resource 2014. In: World
conditions. In: VII International Symposium on Grapevine Physiology and Soil Resources Reports n. 103. FAO, Rome (Italy).

111
S. Priori et al. Geoderma 334 (2019) 99–112

Jahn, R., Blume, H.P., Asio, V.B., Spaargaren, O., Schad, P., 2006. Guidelines for Soil learning methods. Geoderma 226–227, 354–364.
Description. FAO, Rome (Italy). Priori, S., Barbetti, R., L'Abate, G., Bucelli, P., Storchi, P., Costantini, E.A.C., 2014b.
Jones, G.V., White, M.A., Cooper, O.R., Storchmann, K., 2005. Climate change and global Natural terroir units, Siena Province, Tuscany. J. Maps 10–3, 466–477.
wine quality. Clim. Chang. 73 (3), 319–343. Priori, S., Fantappiè, M., Bianconi, N., Ferrigno, G., Pellegrini, S., Costantini, E.A., 2016.
Klute, A., 1986. Water retention: laboratory methods. In: Klute, A. (Ed.), Methods of Soil Field-scale mapping of soil carbon stock with limited sampling by coupling gamma-
Analysis, 2nd edn. ASA and SSSA, Madison, WI, pp. 635–662 (Part 1). ray and vis-NIR spectroscopy. Soil Sci. Soc. Am. J. 80 (4), 954–964.
Kodur, S., 2011. Effects of juice pH and potassium on juice and wine quality, and reg- Ramos, M.C., Jones, G.V., Yuste, J., 2015. Phenology and grape ripening characteristics of
ulation of potassium in grapevines through rootstocks (Vitis): a short review. Vitis cv Tempranillo within the Ribera del Duero designation of origin (Spain): influence of
50–1, 1–6. soil and plot characteristics. Eur. J. Agron. 70, 57–70.
Loeppert, R.H., Suarez, D.L., 1996. Carbonate and gypsum. In: Sparks, D. (Ed.), Methods Retallack, G.J., Burns, S.F., 2016. The effects of soil on the taste of wine. GSA Today
of Soil Analysis, Part 3: Chemical Methods. SSSA and ASA, Madison, WI, pp. 26 (5).
437–474. Ricci Alunni, G., 2004. Uno degli autori della qualità del vino: il terreno. Proceedings of
Mackenzie, D.E., Christ, A.G., 2005. The role of soil chemistry in wine grape quality and Italian conference “I Paesaggi del vino”, Perugia, 6–8 February 2004.
sustainable soil management in vineyards. Water Sci. Technol. 51, 27–37. Rodríguez Donate, M.C., Cano Fernández, V.J., Guirao Pérez, G., 2017. Comparative
Maltman, A., 2008. The role of vineyard geology in wine typicity. J. Wine Res. 19 (1), evaluation of malvasia wines: concordance and reliability of judgments. J. Wine Res.
1–17. 28 (2), 144–158.
Marciniak, M., Reynolds, A.G., Brown, R., 2013. Influence of water status on sensory Salette, J., Asselin, C., Morlat, R., 1998. Le lien du terroir au produit: analyse du système
profiles of Ontario Riesling wines. Foodserv. Res. Int. 54 (1), 881–891. terroir-vigne-vin; possibilité d'applications à d'autres produits. Sci. Aliment. 18 (3),
Martínez-Casasnovas, J.A., Concepción Ramos, M., 2009. Soil alteration due to erosion, 251–265.
ploughing and levelling of vineyards in north east Spain. Soil Use Manag. 25, Scalabrelli, G., D'Onofrio, C., Ducci, E., Bertuccioli, M., 2001. Grapevine performances in
183–192. five areas of Chianti Classico. Rev. S. Vitic. Arboric. Hortic. 33, 253–260.
Martini, I., Aldinucci, M., Foresi, L.M., Mazzei, R., Sandrelli, F., 2011. Geological map of Schultz, H.R., 2003. Differences in hydraulic architecture account for near-isohydric and
the Pliocene succession of the Northern Siena Basin (Tuscany, Italy). J. Maps 7 (1), anisohydric behaviour of two field-grown Vitis vinifera L. cultivars during drought.
193–205. Plant Cell Environ. 26 (8), 1393–1405.
Matthews, M.A., 2016. Terroir and other myths of winegrowing. Univ of California Press Sequi, P., De Nobili, M., 2000. Carbonio organico. In: Angeli, F. (Ed.), Metodi di Analisi
(328 pp.). Chimica del Suolo. Ministero per le Politiche Agricole e Forestali, Osservatorio
Mattii, G.B., Storchi, P., Ferrini, F., 2005. Effects of soil management on physiological, Nazionale Pedologico e per la Qualità del Suolo, VII.1, pp. 1–13.
vegetative and reproductive characteristics of Sangiovese grapevine. Adv. Hortic. Sci. Simone, C., Barondini, M.E., Calabrese, M., 2015. Firm and territory: in searching for a
19 (4), 198–205. sustainable relation. Four cases study from Italian secular firms. Int. J. Environ.
McNeill, J.D., 1990. Geonics EM38 Ground Conductivity Meter: EM38 Operating Manual. Health 7 (4), 329–358.
Geonics Limited, Ontario, Canada. Sudduth, K.A., Kitchen, N.R., Wiebold, W.J., Batchelor, W.D., Bollero, G.A., Bullock, D.G.,
Morari, F., Castrignanò, A., Pagliarin, C., 2009. Application of multivariate geostatistics Clay, D.E., Palm, H.L., Pierce, F.J., Schuler, R.T., Thelen, K.D., 2005. Relating ap-
in delineating management zones within a gravelly vineyard using geo-electrical parent electrical conductivity to soil properties across the north-central U.S.A.
sensors. Comput. Electron. Agric. 68, 97–107. Comput. Electron. Agric. 46, 263–283.
Morlat, R., 2001. Terroirs viticoles : étude et valorisation. Chaintré : Oenoplurimédia Tardaguila, J., Baluja, J., Arpon, L., Balda, P., Oliveira, M., 2011. Variations of soil
(2001)b. pp. 118. properties affect the vegetative growth and yield components of “Tempranillo”
Morlat, R., Bodin, F., 2006. Characterization of viticultural terroirs using a simple field grapevines. Precis. Agric. 12, 762–773.
model based on soil depth–II. Validation of the grape yield and berry quality in the Tardaguila, J., Diago, M.P., Priori, S., Oliveira, M., 2017. Mapping and managing vine-
Anjou vineyard (France). Plant Soil 281 (1–2), 55–69. yard homogeneous zones through proximal geoelectrical sensing. Arch. Agron. Soil
Morris, J.R., Sims, C.A., Cawthon, D.L., 1983. Effects of excessive potassium levels on pH, Sci. 64 (3), 409–418.
acidity and color of fresh and stored grape juice. Am. J. Enol. Vitic. 34 (1), 35–39. Taylor, J.A., McBratney, A.B., Whelan, B.M., 2007. Establishing management classes for
Mpelasoka, B.S., Schachtman, D.P., Treeby, M.T., Thomas, M.R., 2003. A review of po- broadacre agricultural production. Agron. J. 99, 1366–1376.
tassium nutrition in grapevines with special emphasis on berry accumulation. Aust. J. Taylor, J.A., Coulouma, G., Lagacherie, P., Tisseyre, B., 2009. Mapping soil units within a
Grape Wine Res. 9 (3), 154–168. vineyard using statistics associated with high-resolution apparent soil electrical
Nicholas, K.A., Matthews, M.A., Lobell, D.B., Willits, N.H., Field, C.B., 2011. Effect of conductivity data and factorial discriminant analysis. Geoderma 153 (1), 278–284.
vineyard scale climate variability on Pinot noir phenolic composition. Agric. For. Tonietto, J., Carbonneau, A., 2004. A multicriteria climatic classification system for
Meteorol. 151 (12), 1556–1567. grape-growing regions worldwide. Agric. For. Meteorol. 124 (1/2), 81–97.
O.I.V, 2010. Definition of Viticultural “terroir”. Resolution OIV/Viti 333/2010. Triantafilis, J., Laslett, G.M., McBratney, A.B., 2000. Calibrating an electromagnetic in-
Palliotti, A., Tombesi, S., Frioni, T., Famiani, F., Silvestroni, O., Zamboni, M., Poni, S., duction instrument to measure salinity in soil under irrigated cotton. Soil Sci. Soc.
2014. Morpho-structural and physiological response of container-grown Sangiovese Am. J. 64, 1009–1017.
and Montepulciano cvv. (Vitis vinifera) to re-watering after a pre-veraison limiting Van Egmond, F.M., Loonstra, E.H., Limburg, J., 2008. Gamma-ray sensor for topsoil
water deficit. Funct. Plant Biol. 41 (6), 634–647. mapping: the mole. In: 1st Global workshop on High Resolution Digital Soil Sensing &
Pikki, K., Söderström, M., Stenberg, B., 2013. Sensor data fusion for topsoil clay mapping. Mapping, 5–8 February 2008, Sydney, Australia.
Geoderma 199, 106–116. Van Leeuwen, C., Seguin, G., 2006. The concept of terroir in viticulture. J. Wine Res. 17
Pollini, L., Bucelli, P., Calò, A., Costantini, E., L'Abate, G., Lisanti, M.T., Lorenzetti, R., (1), 1–10.
Malorgio, G., Moio, L., Pomarici, E., Storchi, P., Tomasi, D., 2014. Atlante dei territori Vaudour, E., 2002. The quality of grapes and wine in relation to geography: notions of
del vino italiano. Pacini Editore, Firenze (864 pp.). terroir at various scales. J. Wine Res. 13, 117–141.
Poni, S., Bernizzoni, F., Civardi, S., 2007. Response of “Sangiovese” grapevines to partial Vaudour, E., Costantini, E.A.C., Jones, G.V., Mocali, S., 2015. An overview of the recent
root-zone drying: Gas-exchange, growth and grape composition. Sci. Hortic. 114 (2), approaches to terroir functional modelling, footprinting and zoning. Soil 1, 287–312.
96–103. Vaudour, E., Leclercq, L., Gilliot, J.M., Chaignon, B., 2017. Retrospective 70 y-spatial
Priori, S., Martini, E., Andrenelli, M.C., Magini, S., Agnelli, A.E., Bucelli, P., Biagi, M., analysis of repeated vine mortality patterns using ancient aerial time series, Pléiades
Pellegrini, S., Costantini, E.A.C., 2013a. Improving wine quality through harvest images and multi-source spatial and field data. Int. J. Appl. Earth Obs. Geoinf. 58,
zoning and combined use of remote and soil proximal sensing. Soil Sci. Soc. Am. J. 77 234–248.
(4), 1338–1348. Wilson, J.E., 1998. Terroir: The Role of Geology, Climate and Culture in the Making of
Priori, S., Fantappiè, M., Magini, S., Costantini, E.A.C., 2013b. Using the ARP-03 for high- French. University of California Press, Wines (ISBN: 9780520219366).
resolution mapping of calcic horizons. Int. Agrophys. 27 (3), 313–321. Zhu, Q., Lin, H., Doolittle, J., 2010. Repeated electromagnetic induction surveys for de-
Priori, S., Bianconi, N., Costantini, E.A.C., 2014a. Can γ-radiometrics predict soil textural termining subsurface hydrologic dynamics in an agricultural landscape. Soil Sci. Soc.
data and stoniness in different parent materials? A comparison of two machine- Am. J. 74, 1750–1762.

112