journa l homepage: www.e lsev ier .com/ locate /e ja
apping suitability for Sangiovese wine by means of ı13C and geophysicalensors in soils with moderate salinity
doardo A.C. Costantinia,∗, Sergio Pellegrinia, Pierluigi Bucelli a, Roberto Barbetti a,tefano Campagnoloa, Paolo Storchib, Simona Maginia, Rita Perriab
CRA–Research Centre for Agrobiology and Pedology, Piazza D’Azeglio 30, 50121 Florence, ItalyCRA–Research Unit for Viticulture, Via Romea 53, 52020 Arezzo, Italy
r t i c l e i n f o
rticle history:eceived 15 September 2009eceived in revised form 31 May 2010ccepted 31 May 2010
A three year experiment was carried out to test the possibility of using the carbon isotope ratio (ı13C)measured in wine, combined with data of proximal and remote soil sensors, to assess viticultural andoenological suitability for Sangiovese. Two specialized vineyards on similar geomorphological conditionswere investigated. Twelve plots were positioned differently along slopes. The soils were similar, exceptfor structure, porosity and related hydrological characteristics, and salinity of the deeper horizons.
Soil electrical conductivity and resistivity were measured at three different depths, as well as cumula-tive soil moisture down to the root-limiting layer. Satellite images were analyzed to obtain the NDVI. Soilwater content in the plots was monitored at different depths. Yield, phenological phases, and chemicalanalysis of grapes were determined. Stem water potential was measured during summer. Grapes of eachplot were collected for wine making in small barrels. The wines obtained were analyzed and submittedto a blind organoleptic testing. The wine was also analyzed for its ı13C isotopic ratio.
Almost all plots had rather high amounts of transpirable water, even during the driest time of the year.However, only yield components of Sangiovese were influenced by water availability. Wine quality wasinstead significantly improved by the moderate salinity of the deeper horizons, which increased plantwater stress during berry ripening and reduced production. The moderate physiological stress affectingvines was reflected in stem water potentials and ı13C values. ı13C was correlated with several viticulturaland oenological parameters, and also with panel test evaluations of wine quality. The threshold betweengood and bad scores corresponded to a ı13C value of −26.7 ± 1.2‰.
Soil salinity affected the geophysical survey and its relationship with the viticultural and oenologicalresult. In particular, the electromagnetic conductivity measured at the beginning of the experiment was
functional in distinguishing the two vineyards, but it was not useful for a more detailed prediction ofSangiovese performance. However, electromagnetic resistivity in the first 0.5 m was not influenced bysalinity of the deep soil horizons, but only by clay content, and permitted a significant estimation of theSangiovese anthocyanins content, colour intensity, and must acidity.
The outcomes of this study recommend the use of ı13C in combination with electromagnetic resistivityr Sansion
to map soil suitability fosoils may encourage diffu
. Introduction
Physiology and production potential of the vine, as well as wine
uality, are significantly determined by summer water stress (Vaneeuwen and Seguin, 1994; Deloire et al., 2004). Plant water stress isrimarily conditioned by total rainfall and crop management, par-icularly irrigation, but also by rooting depth and soil water holding
capacity. This in turn is regulated by soil particle size and structure,as well as by topography, which can convey rain water and sub-surface flows to different areas of the vineyard. Soil salinity canalso lower water potential, limit root uptake, and has significantinfluence on vine performance and wine quality (Lanyon et al.,2004). Recently, in addition to the more commonly used viticul-tural and oenological indicators – namely yield components, leaf
and stem water potential, oenological parameters and sensorialappraisal – a new physiologic marker has been used for a syn-thetic evaluation of the overall vine water uptake conditions duringthe ripening period, i.e., the ratio between the two stable carbonisotopes 13C/12C, called ı13C, measured in the must sugars upon
arvesting or in the alcohol of the wine produced (Van Leeuwen etl., 2001, 2003; Tregoat et al., 2002). Being correlated with impor-ant components of must quality (e.g., sugar content and titratablecidity) ı13C can be used to map terroirs at different scales (Vaneeuwen et al., 2001; Zufferey and Murisier, 2006; Guix-Hébrardt al., 2007). Reported ı13C values range from −21.0 and −28.0‰,nd −25.5 or −26.0‰ were considered threshold values betweenater stress and non-limiting water nutrition (Van Leeuwen et al.,
003; Deschepper et al., 2006). According to the same authors (Vaneeuwen et al., 2003; Tregoat et al., 2002) values of about −23‰hould correspond to a minimum stem water potential close to1.2 MPa, a value that matched an optimal correlation with theerry sugar accumulation rate. However, other authors pointed outlack of correspondence between ı13C and other proxies for soilater conditions, in particular soil electric resistivity, which would
e correlated instead with other components of quality, namelyhenols and anthocyanins (Deschepper et al., 2006). In addition, aenetic variability of ı13C performance has been observed amongultivars, pointing to the need for a calibration according to varietyGaudillère et al., 2002). So far, no information is available aboutarbon isotope discrimination in grapevines cultivated on moder-tely saline soils.
The use of proximal and remote sensors for precision viticul-ure and mapping terroirs has been introduced for some years ands increasing (Tisseyre et al., 2007; Acevedo-Opazo et al., 2008).eophysical sensors, measuring soil electrical conductivity and
esistivity, as well as satellite images, can provide informationbout the spatial variability of edaphic properties influencing cropield, in particular particle size and water content (Corwin andlant, 2005). This information was complemented with vine andine performance indicators, among others ı13C, and used to delin-
ate site-specific management units and terroirs (Deschepper etl., 2006). Nevertheless, there are limitations in using geophysicalensors where others factors affect the measurements, particularlyoil type and salinity (Dabas et al., 1989; Corwin and Lesch, 2005;amouëlian et al., 2005). Managing this variability could consti-ute a significant challenge for vine growers. On the other hand, itas been shown that some parameters, such as yield and canopyigour, can present significant temporal as well as spatial stability.f one could demonstrate the stability of soil differences detectedy means of sensors, then it would be possible to map the soil ofhe vineyard by means of a single survey using a combination ofensors, thus reducing the cost of their application. However, tem-oral stability was not observed in grape quality in non-irrigatedineyards. Therefore, the potential use of maps from previous yearss a decision-making tool for quality management in the years toome was not feasible (Tisseyre et al., 2007).
Sangiovese is one of the most important Italian wine cultivars,specially widespread in central Italy (Caldano and Rossi, 2008)here climatic conditions during the crop season are intermedi-
te between the very hot and dry summers of southern Italy andhe milder climates of northern Italy. Nevertheless, Sangiovese isecoming appreciated in other viticultural areas over the worldor its capacity to express local peculiarities (Paoletti, 1995). Localuitability for Sangiovese might be assessed through using both13C and proximal/remote sensors in the attempt to measure watertress, since this cultivar needs a moderate deficit during summero produce high quality wine (Costantini et al., 1996; Storchi et al.,005; Palliotti et al., 2008). However, the role played by soil salin-
ty should be carefully considered. In fact, vineyards planted withangiovese on moderately saline soils are not infrequent, as it has
een reported in some important viticultural areas such as Montal-ino and Chianti (Brancadoro et al., 2006; Costantini et al., 2006).his could contrast with the assumption that considers salinity aimitation for viticulture (White, 2003; Lanyon et al., 2004). Thus,he use of new technologies for mapping terroirs must be properly
ronomy 33 (2010) 208–217 209
calibrated when the cultivar is Sangiovese and soils are moderatelysaline. The goal of this work was to test the use of ı13C combinedwith multiple proximal and remote sensors in zoning vineyards forthe Sangiovese vine in environmental conditions characterized bymoderate soil water and salinity stresses.
2. Materials and methods
Two experimental vineyards (2 ha each) were selected at Cetona(Chianti area, central Italy, 42◦57′N, 11◦54′E). They had the sameclimate, lithology, and geomorphological setting, but differentsoils. Long term mean air temperature was 12.7 ◦C, annual rainfall644 mm, and Winkler’s index 1800. Both vineyards were planted onsoils covering slopes with similar steepness (from 2 to 13 or 18%)and aspect (E and NE), and formed from fine silty marine sedimentsof the Pliocene era. The vine variety was Sangiovese, plant density3500 per ha, the rootstock 420A (Vitis Berlandieri × Riparia), whichis considered to be resistant to drought and active lime (Fregoniet al., 1978), but not to salinity (Lambert et al., 2008). Both vine-yards were planted in 1991, after slope reshaping by bulldozing anddeep ploughing down to about 0.8–1.0 m. Viticultural husbandrywas similar and the soil surface was periodically cultivated to limitweed growth, interrupt capillarity and reduce evaporation.
The benchmark soil profiles of the two vineyards had a cambichorizon and were strongly calcareous. Profile 1 hosted a perchedwater table during rainy seasons, as a consequence of the mas-sive structure of the Cr horizon, and was then classified StagnicCambisol (Calcaric, Hyposodic, Hyposalic) according to WRB (F.A.O.et al., 2006) and Aquic Haplustept, fine silty, mixed, mesic, active,following Soil Taxonomy (Soil Survey Staff, 1998). Profile 2 was aHaplic Cambisol (Calcaric) and a Typic Haplustept, fine silty, mixed,mesic, superactive.
Agricultural practices, especially land levelling before vineplanting, led to soil scalping, so that the unweathered substratumcould be found at a shallow depth. That was especially the case ofthe vineyard 1 profile (Table 1). Chemical fertility of that soil wasinfluenced by the presence of substratum at shallow depth, whichwas the reason for the relatively higher active lime, lower organiccarbon and micronutrient content. However, the most striking dif-ferences between the two soils were related to salinity, namely thepresence of sodium and magnesium salts. In fact, soil 1 showedrather high electrical conductivity in the lower horizons, as well ashigh exchangeable sodium and magnesium (more than 7% and 50%of CEC, respectively).
Moisture content at −33 kPa and −1500 kPa of disturbed sam-ples of the benchmark profiles were analyzed in laboratory by theceramic-plate system (Kassel and Nielsen, 1986) and bulk densitywith the core method on replicated samples. Soil description androutine analysis of the air-dried <2 mm fraction followed the offi-cial Italian methods (Costantini, 2007; MiPAF, 2000). In particular,soil texture was tested in the laboratory by the sieve and pipettemethod; mean geometric diameter was calculated according to Geeand Or (2002). CaCO3 content was measured gas-volumetrically,by addition of HCl in a Dietrich-Frühling calcimeter; active CaCO3was analyzed with a solution of ammonium acetate. This is themore active fraction of CaCO3, which easily dissolves and precipi-tates. Soil electrical conductivity of the plots was tested every 0.2 m;organic carbon content was determined by using the Walkley-Blackprocedure; pH and electrical conductivity were measured in a 1:2.5(w w−1) water suspension; cation exchange capacity (CEC) wasmeasured by use of 1 M Na-acetate solution at pH 7.0; exchange-
able bases were extracted with 1 M NH4
+ acetate solution at pH7.0 and measured by flame photometry (Na, K and Ca) and atomicabsorption spectrometry (Mg).
Twelve experimental plots, about 300 m2 each, were locatedalong the slope in correspondence with the three basic morpholog-
nd = not determined.a Consistence moist: FR = friable, RE = resistant.b Structure: SB = subangular blocky, AB = angular blocky, MA = massive.c CEC = cationic exchange capacity.
F
itrtss
F
ig. 1. 3D orthophoto of vineyard 1, with experimental plots, aspect, and elevation.
cal positions of summit, backslope, and footslope, so as to consider
he different soil moisture conditions present in the vineyards,anging from the relatively dry to the relatively wet (Cetona 1 Ao F, Cetona 2 A to F) (Figs. 1 and 2). The soil profile of the plots wasampled and analyzed for texture and electrical conductivity (fiveamples every 0.2 m from the surface to 1 m depth).
ig. 2. 3D orthophoto of vineyard 2, with experimental plots, aspect, and elevation.
Many difficulties were encountered in the monitored plots inusing the core method to measure bulk density, because of the highsoil plasticity. Therefore, total porosity was calculated from the fieldmeasured value of moisture when soil was saturated. Then, bulkdensity was estimated assuming a particle density of 2.65 g cm−3
and correcting values according to Faybishenko (1995), to take intoaccount entrapped air (assumed to be 0.05):
�b = 100 × �w × C
(100 × �w × C/�s) + �m
where: �b = calculated soil bulk density at saturation (g cm−3),�w = water density (g cm−3), C = �s/� = ratio between porosityat saturation and total porosity, �s = particle density (g cm−3),�m = gravimetric water content at saturation (%w w−1).
Saturation was empirically assumed after a heavy spring rainthat left ponds on the soil surface. Similarly, moisture content atfield capacity was obtained by averaging sample values recordedabout three days after soil saturation.
On April 4, 2005, a multiple combined geophysical survey wascarried out in collaboration with the Soil Information System (SIS)of John Deere Agri Services at bud break of vines. Electromag-netic conductivity was measured with EM38. A multi-sensor probewas used to evaluate resistivity (resistivimeter), soil moisture(capacitance-based sensor) and resistance (cone penetrometer).The root-limiting layer was assumed to be the first layer offer-ing a resistance higher than 350 psi (2413 kPa). The probe wasinserted into the soil to about 1.5 m depth in 21 random locationsin each vineyard. Spatialization was obtained with Arc GIS and theInverse Distance Weighting method. Thus SIS produced the follow-
ing evaluations: (i) soil electromagnetic conductivity (EC); (ii) soilelectromagnetic resistivity (ER) at three depths (from surface toaround 0.5, 0.9, and 1.3 m), (iii) volumetric soil moisture in the root-ing zone (RL-moist), that is, the cumulative volumetric water fromthe surface to the root-limiting layer.
E.A.C. Costantini et al. / Europ. J. Agronomy 33 (2010) 208–217 211
mpera
oecNfcuoedwetuwate
vdmia(0ubrmtvvvUmc1ebsisdsitrp
Fig. 3. Precipitation and air te
A NDVI analysis was carried out on Spot5 satellite images takenn July 4, 2006 with 1 m resolution. The NDVI is a normalized veg-tation index which uses the different reflection of the vegetationover to the spectral bands of visible red (rR) and near infrared (rIR):DVI = (rIR − rR)/(rIR + rR) (Rouse et al., 1974). NDVI values, ranging
rom −1 and +1, are closely related with vegetation physiologicalonditions and biomass amount. In the present work negative val-es were eliminated and their distribution normalized by meansf the transformed vegetation index (TVI) suggested by Deeringt al. (1975): TVI = sqrt (0.5 + NDVI). The original pixel at 1 m wasegraded to 10 m to compare the TVI map with those producedith the geophysical sensors. Calculation of the average value for
ach plot and vineyard was obtained by using the zonal statisticool of the spatial analyst of Arc GIS. The software Arcscene wastilized to drape the orthophotos on the Triangular Irregular Net-ork and to create the 3D map of the vineyards. In the statistical
nalysis, the data of the proximal and remote sensors were relatedo the average results of the three year viticultural and oenologicalxperiment.
A meteorological station was placed only in vineyard 1, sinceineyard 2 was just a few dozen meters away from it. Hydrope-ological properties and transpirable soil water were identified byeans of field monitoring during the years 2005, 2006, and 2007
n plots A, B, and C of both vineyards, but only in 2007 in plots D, E,nd F. Soil water content was measured by the gravimetric methodthree samplings per position with a hand auger) at 0.1–0.3 m and.4–0.7 m depth. Experimental plots were unrestricted and these of permanent equipment, like neutron probes or capacitance-ased soil moisture sensors, was not possible. Measurements wereeplicated every one/two weeks during the growing season, andonthly in the rest of the year. A daily value of the water con-
ent (total mm in the 0–0.7 m depth) at each position in the twoineyards was obtained by using rainfall amount, then estimatingineyard evapotranspiration and runoff, following the Soil Conser-ation Service Curve Number methodology (SCS-CN USDA, 1969;SDA, 1985), and calibrating the results with the measured soiloisture. Mean daily potential evapotranspiration (ETp) was cal-
ulated with the Priestley–Taylor equation (Priestley and Taylor,972). Cultural coefficients (Kc) were applied to ETp to evaluate realvapotranspiration (ETr) according to the methodology proposedy Allen et al. (1998). The estimated ETr of the vineyard reduced theoil water content according to the logarithmic function reportedn Thornthwaite and Mather (1957). A daily average of transpirableoil water (TSW) from the surface to 0.70 m was computed. Theaily TSW of each plot was the difference between the calculated
oil water content and the absolute minimum value measured dur-ng the three years of testing at the two measurement depths. Theime period of June 10th–September 10th was chosen to be theeference time, because it corresponded to the most sensitive vinehenological phases (from flowering to ripening).
ture during the study period.
During the three years studied, three replicated samplings perplot were conducted on ten vines. The vegetative behaviour of theplants was recorded, in particular the date of phenological phases,yield components (yield per vine, cluster and 100 berry weight,number of clusters per vine), must titratable acidity, and sugar con-tent of grapes (OIV, 2005). Stem water potential (Van Leeuwen etal., 2003; Williams and Araujo, 2002) was measured with a pres-sure chamber, according to the methodology proposed by Choné etal. (2001), on non-transpiring mature leaves that had been baggedbefore measurement. The measurement was carried out during theyears 2005 and 2006, monthly from June to August, to appreciatethe maximum value of stem potential.
One hundred kg of grapes were collected from each plot forwine making in small barrels, using the same oenological techniquefor all samples: crushing and de-stemming, 50 mg L−1 SO2, 8-daymaceration, two pumpings over the cap every day, controlled tem-perature of fermentation (26–28 ◦C), devatting and pressing. Thewines obtained were analyzed for colour intensity, that is the sumof optical density at 420 and 520 nm, and for total anthocyanins andpolyphenols, very important parameters for evaluating colour andstructure of a red wine (Di Stefano et al., 1997). Ten months laterthe wines were submitted to blind organoleptic testing by meansof an unstructured card (Weiss, 1981) with the aim of defining arank of preferences in terms of general harmony. The evaluationcard had a square with a diagonal drawn from the left vertex down(negative evaluation, 0 score) to the right one up (positive evalu-ation, 220 score). The square centre stood for mean quality wine(110 score). The panel was required to position samples along thediagonal on the basis of a comprehensive evaluation.
The isotopic ratio 13C/12C (ı13C) was measured in the wineethanol by Isotope Mass Spectrometry to assess possible waterstress occurring during grape formation and ripening. The ı13Cwas expressed in reference to the international standard V-PDB(Farquhar et al., 1989; Van Leeuwen et al., 2001).
3. Results
3.1. Meteorological conditions during the study years
Meteorological conditions during the experiment were char-acterized by a rather humid and mild 2005, with mean annualair temperature (MAAT) 12.6 ◦C and annual rainfall (AR) 1028 mm(Fig. 3), while both years 2006 and 2007 were rather hot and dry(MAAT 13.9 and 13.6 ◦C; AR 427 and 470 mm, respectively). Choos-ing as reference the time period of June 10th to September 10th –
since late spring and summer rainfall and temperatures are partic-ularly relevant for vine growing – rainfall varied much more thantemperature during testing. In particular, 225.8 mm of rain fell in2005, 9.8 mm in 2006, and 60.0 in 2007, whereas daily mean airtemperatures were 22.4, 23.2, and 22.0 ◦C, respectively. Air tem-
212 E.A.C. Costantini et al. / Europ. J. Agronomy 33 (2010) 208–217
Fi
pyttp
3
3
bbfbv
Cosi5
3v
vgyst
in particular, showed a higher mean value and standard devia-tion in vineyard Cetona 1 (125.4 mS m−1; 24.15) than in Cetona 2(65.5 mS m−1; 7.35). Plots showed similar differences in mean val-ues and standard deviations between vineyards (Table 2). Vineyard
ig. 4. Comparison between averaged mean geometric diameters of soil particlesn the plots of Cetona 1 and 2 vineyards.
erature as a whole can be considered rather high during all threeears, with a relevant number of days with maximum tempera-ure higher than 30 ◦C (37 in 2005, 45 in 2006, and 38 in 2007). Thisemperature level is believed to be the upper threshold for efficienthotosynthesis of Sangiovese (Intrieri et al., 2001).
.2. Soil variability assessment in the two vineyards
.2.1. Soil characteristics in the plotsSoil texture of the plots was similar to that of the corresponding
enchmark profiles of the two vineyards. In particular, as evidencedy the mean geometric diameter (Fig. 4), it was rather fine and uni-orm along the profile in all plots, but there was a slight differenceetween the two vineyards, that is, it ranged from 3 to 9 �m inineyard 1 and from 4 to 10 �m in vineyard 2.
Soil salinity instead clearly differentiated the two vineyards.etona 1 plots were on average 366 �S cm−1, against 242 �S cm−1
f Cetona 2 plots (Fig. 5). Starting from 0.3 m depth, Cetona 1 plotshowed a marked increase in electrical conductivity, particularlyn Cetona 1A, C, and E, where maximum values of 944, 924, and27 �S cm−1 were, respectively, reached.
.2.2. Proximal and remote sensor evaluation of spatialariability in the experimental vineyards and plots
At the beginning of the experiment, both the geophysical sur-
ey performed by SIS at bud bursting of vines and the measuredravimetric water content indicated that the soils of the two vine-ards were close to field capacity. Therefore, the different RL-moisthowed in the two vineyards (Figs. 6 and 7) can be mainly attributedo the different depth of the root-limiting layer (ranging from
Fig. 5. Comparison between soil electrical conductivity (1:2.5, w w−1) in the plotsof Cetona 1 and 2 vineyards.
101 to 130 cm in Cetona 1, and from 131 to 142 cm in Cetona2). In particular, vineyard Cetona 1 had a smaller overall mois-ture (288 mm) and a larger variability (standard deviation 66.6)than vineyard Cetona 2, where the average was 384 mm and stan-dard deviation 34.1. The lower mean moisture and higher localvariability of Cetona 1 was also well represented in the plots(Table 2).
Spatial differences between the two vineyards and plots werealso reported in the maps of EC and ER (Figs. 8–11). The EC maps,
Fig. 6. Map of cumulative soil moisture up to the root-limiting layer (RL-moist) invineyard Cetona 1 at bud bursting of vines.
ig. 7. Map of cumulative soil moisture up to the root-limiting layer (RL-moist) inineyard Cetona 2 at bud bursting of vines.
etona 1 produced average and standard deviation values of ER inhe first 0.5 m similar to Cetona 2, that is, 24.9 ohm-m versus 24.0,nd 2.02 versus 3.88. The measurements in the plots also showedimilar values in Cetona 1 and in Cetona 2 (Table 2). Since the ERrom surface to deeper depths of the plots (around 0.9 and 1.3 m)id not show any significant relationship with all the other soilnd viticultural variables in study, the values were not reported in
able 2. However, the ER mean values from surface to around 0.9 mn vineyards Cetona 1 and Cetona 2 were 23.2 ohm-m and 24.0,espectively, while standard deviations were 6.89 and 3.88. Fromurface to about 1.3 m the mean values of the whole vineyard Cet-
ig. 8. Map of electromagnetic conductivity (EC) in vineyard Cetona 1 at bud burst-ng of vines.
Fig. 9. Map of electromagnetic conductivity (EC) in vineyard Cetona 2 at bud burst-ing of vines.
ona 1 was 22.6 ohm-m, whereas it was 18.2 in Cetona 2. Standarddeviations were 1.41 and 2.71, respectively.
The TVI evaluation obtained through the analysis of the satel-lite images of the two vineyards showed similar normalized meanvalues in Cetona 1 and in Cetona 2 plots, but a larger standarddeviation in the first vineyard. TVI values resulted significantly cor-related with the geophysical sensors. In particular, with ER at 0.5 m(r = −0.471 in Cetona 1 and −0.535 in Cetona 2, n = 340, P < 0.001),
and RL-moist (r = 0.314 in Cetona 1 and 0.427 in Cetona 2, n = 340,P < 0.001), and also with EC, but only in Cetona 2 (r = −0.254, n = 340,P < 0.001). TVI did not show any relationship with viticultural andoenological parameters, resulting correlated only with the mon-
Fig. 10. Map of electromagnetic resistivity (ER) from 0 to about 0.5 m in vineyardCetona 1 at bud bursting of vines.
214 E.A.C. Costantini et al. / Europ. J. Agronomy 33 (2010) 208–217
FC
i(
3
(wiv
scutbstwttbby
adanw
Table 4Linear relationships between score at the panel test and some viticultural and oeno-logical variables.
Table 5Linear relationships between transpirable soil water from 0 to 0.7 m (TSW), soilsalinity (electrical conductivity), and some viticultural and oenological variables.
ig. 11. Map of electromagnetic resistivity (ER) from 0 to about 0.5 m in vineyardetona 2 at bud bursting of vines.
tored soil moisture of the plot at the date of the satellite imager = 0.578, n = 6, P < 0.05).
.3. Vine performance and wine quality
Viticultural parameters ranged notably during the experimentTable 3). The yield per vine was the parameter which varied most,hereas sugar content in the must was the parameter which var-
ed less. The other parameters showed an intermediate range ofariation.
Among the parameters evaluated for wine quality, colour inten-ity showed the widest range of variation. High values of theoefficient of variability (CV) were also found in the sensory eval-ation. During panel evaluation, the mean and median values ofhe scores obtained by the wines were just under the thresholdetween negative and positive judgments. However, the maximumcore indicated the presence of some samples with good quality athe sensory evaluation. The results of the panel test did not correlateith any yield parameter, but correlated positively with sugar con-
ent, colour and phenolic composition, and negatively with mustitratable acidity (Table 4). Vine parameters instead were affectedy both soil water content and salinity (Table 5). TSW increasederry and cluster weights, while soil electrical conductivity loweredield per plant and increased stem water potential.
The carbon isotopic ratio was correlated with several viticultural13
nd oenological parameters (Table 6). In particular, ı C resulted
irectly correlated with soil electric conductivity, i.e., soil salinity,s well as with stem water potential. Furthermore, there was a sig-ificant direct relationship between ı13C values and scores of theine at the panel test.
able 3ummary statistics of the viticultural variables, ı13C, and transpirable soil water (TSW).
Stem water potential −0.764 24 <0.001Electrical conductivity 0.587 30 <0.001Panel test 0.363 30 <0.05
3.4. Relationships between proximal sensors, viticultural andoenological parameters
The results of the geophysical survey were compared with thesoil characteristics and the viticultural and oenological perfor-mance of Sangiovese during the three year experiment. Soil ECin the plots was directly correlated with salinity (r = 0.771, n = 12,P < 0.01) and clay content (r = 0.803, n = 12, P < 0.01), while mean ERup to 0.5 m in the plots was negatively correlated only with clay(r = −0.771, n = 12, P < 0.01). On the other hand, RL-moist of the plotswas neither correlated with clay nor with soil salinity. Moreover,no geophysical variable was correlated with mean TSW of the plotsduring the experiment.
As for the relationships with the viticultural and oenologicalparameters, EC in the plots was directly correlated with the meanweight of berries obtained during the test, as well as with meancluster weight, and ı13C (Table 7). RL-moist was inversely corre-lated with ı13C, while ER in the first 0.5 m was directly associated
ER 0.5 m Wine anthocyanins −0.686 12 <0.05Must total acidity 0.620 12 <0.05Wine col. intensity −0.606 12 <0.05
ith titratable acidity and inversely with colour intensity andnthocyanins of wine. As said before, the values of soil ER fromhe surface to 0.9 and 1.3 m were not related with any viticulturalr oenological parameter.
. Discussion
.1. Climatic and spatial variability during the experiment
Boundary conditions of the experiment can be considered repre-entative of the study area and typical of Mediterranean viticulture.ainfall amounts, more than temperatures, differed most throughhe years. Soil variability was also notable, particularly salinitynd RL-moist. Vineyard Cetona 1 showed a more accentuated soilariability than Cetona 2 and the presence of the substratum at shal-ower depths. As a consequence, mean soil water holding capacity
as lower and local variations higher in vineyard Cetona 1. Alsohe EC maps showed a larger mean value and standard devia-ion in vineyard Cetona 1. The much higher EC found in Cetona 1ontrasted with the lower cumulative moisture, showing that theensor was less influenced by soil water content than by salinityr = 0.806, P < 0.01). ER in the first 0.5 m instead showed a differentrend. Vineyard Cetona 1 produced average and standard deviationalues similar to Cetona 2. Therefore, water content and salinity ofhe surface horizon in the two vineyards at the time of survey didot differ significantly. It is worth note that, in spite of the differ-nces in the results of the three kinds of proximal sensors (RL-moist,C, and ER), as a rule vineyard Cetona 1 produced larger standardeviations, which confirmed its greater local variability.
The TVI obtained through the NDVI analysis of the remote sen-or correlated well with ER at 0.5 m and RL-moist in both vineyards,nd also with EC, but only in Cetona 2. The different relationshipetween TVI and EC in the two vineyards was particularly interest-
ng, because it confirmed that in Cetona 1 soil salinity influencedC, but not vine vigour, at least up to the date of satellite surveyearly July). TVI evaluation showed similar normalized mean val-es in Cetona 1 and in Cetona 2 plots, but a larger standard deviation
n the first vineyard. The outcome confirmed once more the largerariability of vineyard Cetona 1, but also underlined the averageimilar vigour of vines in Cetona 1 and Cetona 2 at the phase oferry formation.
.2. Vine performance and wine quality
The yield per vine was the parameter which varied most. Thisesult underlined the different soil fertility conditions throughouthe vineyards. However, the high maximum values might indi-ate that excessive fertility had in some cases not been adequately
ontrolled by the uniform agricultural husbandry operated by thearmer. On the other hand, sugar content in the must was thearameter which varied less, and indicated a fairly good averageiticultural result. Among the parameters evaluated for wine qual-ty, colour intensity showed the widest range of variation and a
ronomy 33 (2010) 208–217 215
mean value that can be considered only sufficient for a Sangiovesewine. Also, the mean values of total polyphenols and anthocyaninswere rather low, but maximum values reached a very good levelfor Sangiovese. High values of the coefficient of variability werealso found in sensory evaluation. Mean and median scores of paneltesting were just under the threshold between a negative and apositive judgment. On the other hand, in this case also, the maxi-mum score indicated the presence of some samples of good qualityat sensory evaluation. The high variability of the viticultural andoenological variables was only partially reflected in the values ofı13C. Mean, median and minimum absolute values of the carbonisotopic ratio appeared to be rather high if compared with the lit-erature (Van Leeuwen et al., 2001). As the CV was rather low, we caninfer that Sangiovese vines in many plots suffered only from limitedor negligible stress. The results of the panel test did not correlatewith any yield parameter, but correlated positively with sugar con-tent, colour and phenolic composition, and negatively with musttitratable acidity (Table 4). Therefore, we can deduce that the panelevaluation was particularly influenced by wine colour, acidity, andalcoholic degree, which in fact constitute some of the most impor-tant parameters characterizing the quality of a Sangiovese.
4.3. Soil–plant relationships
Vine parameters were affected by both soil water content andsalinity (Table 5). TSW had positive effects on berry and clusterweight, while soil electrical conductivity lowered yield per plantand stem water potential. The carbon isotopic ratio was correlatedwith several viticultural and oenological parameters (Table 6), con-firming its ability to summarize many factors of wine quality. Itis relevant to emphasize that in our experimental conditions ı13Cdid not correlate with TSW. This rather unexpected result can beexplained by considering the influence of salinity on vine stress asexpressed by ı13C. Actually, as shown in Table 6, there was a sig-nificant direct relationship between the two variables. The lack ofcorrelation between TSW, oenological result and stem water poten-tial, confirmed that TSW was in most cases enough to fulfil thevine needs, even during the driest time of the season. On the otherhand, the inverse relationship between ı13C and stem water poten-tial drew attention to a moderate stress that occurred in the vinesof some plots (Table 6). Furthermore, it is important to note thatour results demonstrated a significant direct relationship betweenpanel evaluation of the wine and ı13C (Table 6). If we choose thescore of 110 as a threshold between good and bad in wine evalua-tion, we can infer that this threshold corresponded to a ı13C valueof −26.7 ± 1.2‰. Hence this value could be chosen as a first ref-erence for the delimitation of terroirs for Sangiovese. In our case,a ı13C value of −26.7‰ corresponded to a stem water potential of−0.93 MPa and to a mean electrical conductivity of the soil profile of284 �S cm−1. These values indicated environmental conditions forSangiovese vines ranging from those characterized by the absenceof water stress to those of moderate stress, caused by an increaseof soil salinity in depth.
4.4. Relevance of proximal sensors for Sangiovese
Soil EC in the plots was directly correlated with salin-ity (r = 0.771, n = 12, P < 0.01) and clay content (r = 0.803, n = 12,P < 0.01), while mean ER up to 0.5 m in the plots was only negativelycorrelated with clay (r = −0.771, n = 12, P < 0.01). On the other hand,RL-moist of the plots was neither correlated with clay nor with soil
salinity. Moreover, no geophysical variable was correlated with themean TSW of the plots during the years of study.
Notwithstanding the statistical significance of the correlationbetween ı13C and some geophysical results, an in-depth examina-tion of relationships demonstrated the role played by soil salinity in
216 E.A.C. Costantini et al. / Europ. J. Ag
Fat
dapveTtTcayiptctSiwSssigw1
5
gstowhsmbiwobt
cvbtv
ig. 12. Relationship between electromagnetic conductivity (EC) and ı13C. Squarend triangle symbols represent data of vineyards Cetona 1 and Cetona 2, respec-ively.
ifferentiating the performance of the two vineyards. In fact, usings example the highly significant relationship between EC and ı13C,oints in the diagram showed in Fig. 12 resulted clearly clustered,isibly separating the two vineyards. If the vineyards were consid-red separately, there would not be any significant relationship.herefore, the relation between EC and ı13C only further stressedhe differences in soil salinity and ı13C between the two vineyards.he same held true for the relationships between EC and the yieldomponents, as well as between ı13C and RL-moist. For that reason,lthough EC and RL-moist were able to differentiate the two vine-ards and some soil plot characteristics, they did not give usefulndications about Sangiovese performance. On the other hand, theoints showing the relationships between ER 0.5 m and the viticul-ural parameters were well distributed in the two vineyards, andould be considered representative of the relationship between thewo variables. The ability of ER to predict the differences betweenangiovese plots in terms of must acidity, anthocyanins, and colourntensity was particularly interesting. As ER was neither correlated
ith TSW, nor with soil salinity of the plots, the relationship withangiovese might be based on clay content. In other words, loweroil resistivity values indicated higher clay percentage and at theame time pointed to higher wine anthocyanins content and colourntensity, and lower must acidity, all elements of quality for San-iovese. A linkage between soil clay content and quality of red wineas also found by other authors (e.g., Fregoni, 1997; Tomasi et al.,
999).
. Conclusions
The experiment demonstrated that the performance of the San-iovese vine was influenced by both transpirable soil water andalinity, which varied according to soil type and topographic posi-ion. A moderate salinity of the deeper soil horizons enhanced theenological outcome of Sangiovese, because it caused a limitedater stress during berry ripening in vines cultivated on soils thatad a good water supply even during the driest part of the growingeason. The values of the stem water potential and ı13C showed aoderate stress. For the first time, a significant direct relationship
etween wine panel evaluation and ı13C has been verified. Choos-ng the value of 110 as a threshold in the scores obtained by the
ine at the panel (range 0–220), we could infer that this thresh-ld corresponded to a ı13C value of −26.7 ± 1.2‰. This value mighte chosen as a first reference for the delimitation of Sangioveseerroirs.
Although moderate, salinity influenced significantly the out-
omes of the geophysical survey and their relationship with theiticultural and oenological results. In particular, the correlationetween ı13C and electromagnetic conductivity measured at theime of bud bursting was only useful to distinguish the twoineyards, but could not be used as a proxy for the Sangiovese
ronomy 33 (2010) 208–217
performance. Electromagnetic resistivity in the first 0.5 m insteadallowed a significant estimation of Sangiovese performance interms of anthocyanins content, colour intensity, and must acidity.The NDVI analysis of satellite images was also able to differen-tiate the two vineyards, but did not give significant results as tothe prediction of Sangiovese performance, probably because of theexcessively coarse detail of satellite resolution. However, the NDVIanalysis was very important, since it showed that differences invine vigour were negligible until the phase of berry formation.
In conclusion, the outcomes of this study recommend the useof the carbon isotope ratio in combination with the survey ofsoil electromagnetic resistivity to map terroirs for Sangiovese. Thefavourable performance of Sangiovese in moderately saline soilscould encourage further diffusion of the cultivar outside its tradi-tional areas of cultivation.
Acknowledgements
Authors are deeply in debt with the Aggravi and Sebastianifarms, of the Cantina Sociale Etruria in Cetona, which hosted thetrial, and acknowledge the precious collaboration given by Jean-Pierre Lemoine (Soil Information System, John Deere Agri Services)and Jean-Luc Portalier (SADEF—Suze-la-Rousse, France).
References
Acevedo-Opazo, C., Tisseyre, B., Guillaume, S., Ojeda, H., 2008. The potential of highspatial resolution information to define within-vineyard zones related to vinewater status. Precision Agric. 9, 285–302.
Allen, R.G., Pereira, L.S., Raes, D., Smith, M., 1998. CropEvapotranspiration—Guidelines for Computing Crop Water Requirements. FAO,Irrigation and Drainage, Paper 56, Rome.
Brancadoro, L., Failla, O., Tonnato, L., Scienza, A., 2006. Risultati di indagini di zon-azione a diversa scala effettuate sul territorio toscano. Quad. vitic. enol. univ.Torino 28, 281–291.
Caldano, G., Rossi, A., 2008. Codice Denominazioni di origine dei vini. Unione ItalianaVini (Ed.), Milano.
Choné, X., Van Leeuwen, C., Dubourdieu, D., Gaudillere, J.P., 2001. Stem water poten-tial is a sensitive indicator for grapevine vine status. Ann. Bot. 4, 477–483.
Costantini, E.A.C., 2007. Linee guida dei metodi di rilevamento e informatizzazionedei dati pedologici. In: CRA-ABP, Firenze, Italia, p. 296.
Costantini, E.A.C., Campostrini, F., Arcara, P.G., Cherubini, P., Storchi, P., Pierucci, M.,1996. Soil and climate functional characters for grape ripening and wine qualityof “Vino Nobile di Montepulciano”. Acta Hort., 427 ISHS, 45–55.
Costantini, E.A.C., Barbetti, R., Bucelli, P., L’Abate, G., Lelli, L., Pellegrini, S., Storchi,P., 2006. Land Peculiarities of the vine cultivation areas in the province of Siena(Italy), with indications concerning the viticultural and oenological results ofSangiovese vine. Boll. Soc. Geol. It. 6, 147–159.
Dabas, M., Hesse, A., Jolivet, A., Tabbagh, A., 1989. Intérêt de la cartographie de larésistivité électrique pour la connaissance du sol à grande échelle. Sci. Sol. 27(1), 67–68.
Deering, D.W., Rose Jr., J.W., Mass, R.H., Schell, J.A., 1975. Measuring forage pro-duction of grazing units from LANDSAT MSS data. In: Proceedings of the 10thInternational Symposium on Remote Sensing of Environment, pp. 1169–1178.
Deloire, A., Carbonneau, A., Wang, Z., Ojeda, H., 2004. Vine and water a short review.J. Int. Sci. Vigne Vin 38 (1), 1–13.
Deschepper, G., Cassassolles, X., Dabas, M., Pernet, D., 2006. Complémentarité desmesures de résistivité électrique des sols et du �C13 du moût dans l’étudeet la valorisation des terroirs viticoles. In: VIth International Terroir Congress,Bordeaux, pp. 232–236.
Di Stefano, R., Ummarino, I., Gentilini, N., 1997. Alcuni aspetti del controllo di qualitànel campo enologico. Annali ISEN XXVIII, 105–121.
F.A.O., I.U.S.S., I.S.R.I.C., 2006. World Reference Base for Soil Resource. World SoilResource Report n. 103. F.A.O., Rome.
Farquhar, G., Ehleringer, J., Hibick, K., 1989. Carbon isotope discrimination and pho-tosynthesis. Ann. Rev. Plant Physiol. Plant Mol. Biol. 40, 503–537.
Faybishenko, B.A., 1995. ydraulic behaviour of quasi-saturated soils in the pres-ence of entrapped air: laboratory experiments. Water Resour. Res. 31 (10),
2421–2435.
Fregoni, M., 1997. Effets du sol et de l’eau sur la qualité de la vendange. Proc. Symp.Int. sur la Qualité de la vendange, Cape Town, 151–168.
Fregoni, M., Scienza, A., Miravalle, R., 1978. Évaluation précoce de la résistance desportes-greffes à la sécheresse. In: Génétique et Amélioration de la Vigne, INRAed., pp. 287–297.
audillère, J.-P., Van Leeuwen, C., Ollat, N., 2002. Carbon isotope composition ofsugars in grapevine, an integrated indicator of vineyard water status. J. Exp. Bot.53 (369), 757–763.
ee, G.W., Or, D., 2002. Particle-size analysis. In: Dane, J.H., Topp, G.C. (Eds.), Methodsof Soil Analysis. Part 4. SSSA Book Series No. 5. SSSA, Madison, WI, pp. 255–293.
uix-Hébrard, N., Voltz, M., Trambouze, W., Garnier, F., Gaudillère, J.P., Lagacherie, P.,2007. Influence of watertable depths on the variation of grapevine water statusat the landscape scale. Eur. J. Agron. 27, 187–196.
ntrieri, C., Poni, S., Lia, G., Gomez del Campo, M., 2001. Vine performance and leafphysiology of conventionally- and minimally pruned Sangiovese grapevines.Vitis 40, 123–130.
assel, D.K., Nielsen, D.R., 1986. Field capacity and available water capacity. In: Klute,A. (Ed.), Methods of Soil Analysis, Part 1, 2nd ed. Am. Soc. Agr., Madison, WI, pp.901–926.
ambert, J.J., Anderson, M.M., Wolpert, J.A., 2008. Vineyard nutrient needs vary withrootstocks and soils. California Agric. 62 (4), 202–207.
anyon, D.M., Cass, A., Hansen, D., 2004. The effect of soil properties on vine perfor-mance. CSIRO Land and Water Technical Report No. 34/04.
iPAF: Metodi di analisi chimica del suolo, Franco Angeli, Milan, Italy, 2000.IV, 2005. Recueil des methodes internationales d’analyse des vins et des mouts.
Vol. 1, Paris, France.alliotti, A., Silvestroni, O., Petoumenou, D., Vignaroli, S., Berrios, J.G., 2008. Eval-
uation of low-energy demand adaptive mechanisms in Sangiovese grapevineduring drought. J. Int. Sci. Vigne Vin 42 (1), 41–47.
aoletti, A., 1995. Sangiovese in Tuscany. In: J.M. Rantz (Ed.), Proc. Int. Symposiumon Clonal Selection, Portland, Oregon, Am. Soc. for Oenology and Viticulture,Davis, pp. 60–69.
riestley, C.H.B., Taylor, R.J., 1972. On the assessment of the surface heat flux andevaporation using large-scale parameters. Mon. Weather Rev. 100, 81–92.
ouse, J.W. Jr., Haas, R.H., Deering, D.W, Schell, J.A., Harlan J.C., 1974. Monitoring thevernal advancement and retrogradation (green wave effect) of natural vegeta-tion. NASA/GSFC Type III Final Report, Greenbelt, MD, USA, p. 371.
amouëlian, A., Cousin, I., Tabbagh, A., Bruand, A., Richard, G., 2005. Electrical resis-tivity survey in soil science: a review. Soil Till. Res. 83 (2), 73–193.
CS-CN USDA, 1969. Soil Conservation Service National Engineering Handbook, Sec-tion 4: Hydrology. Washington D.C.
oil Survey Staff, 1998. Keys to Soil Taxonomy—SMSS Technical Monograph n.19,Pocahontas Press, Blacksburg, Virginia, U.S.A., p. 541.
ronomy 33 (2010) 208–217 217
Storchi, P., Costantini, E.A.C., Bucelli, P., 2005. The influence of climate and soil onviticultural and oenological parameters of Sangiovese grapevine under non-irrigated conditions. Acta Horticulturae 689, 333–340.
Thornthwaite, C.W., Mather, J.R., 1957. Instructions and tables for computingpotential evapotranspiration and the water balance. Publicat. Climatol. 10 (3),183–311.
Tisseyre, B., Ojeda, H., Taylor, J., 2007. New technologies and methodologies for site-specific viticulture. J. Int. Sci. Vigne Vin 41 (2), 63–76.
Tomasi, D., Calò, A., Biscaro, S., Vettorello, G., Panero, L., Di Stefano, R., 1999. Influ-ence des caractéristiques physiques du sol sur le développe de la vigne, dans lacomposition polyphénolique et anthocyanique des raisins et la qualité du vin deCabernet sauvignon. Bull. O. I. V. 72, 819–820.
Tregoat, O., Van Leeuwen, C., Choné, X., Gaudillère, J.P., 2002. The assessment ofvine water and nitrogen uptake by means of physiological indicators influ-ence on vine development and berry potential (Vitis vinifera L. cv merlot, 2000,Bordeaux). J. Int. Sci. Vigne Vin 36 (3), 133–142.
USDA, 1985. SCS National Engineering Handbook, Section 4: Hydrology. WashingtonD.C.
Van Leeuwen, C., Seguin, G., 1994. Incidences de l’alimentation en eau de la vigne,appréciée par l’état hydrique du feuillage sur le développement de l’appareilvégetatif et la maturation du raisin. J. Int. Sci. Vigne Vin 2, 81–110.
Van Leeuwen, C., Gaudillere, J.P., Tregoat, O., 2001. Evaluation du régime hydrique dela vigne à partir du rapport isotopique 12C/13C. J. Int. Sci. Vigne Vin 4, 195–205.
Van Leeuwen, C., Tregoat, O., Chone, X., Jaeck, M.E., Rabusseau, S., Gaudillere, J.P.,2003. Le suivi du régime hydrique de la vigne et son incidence sur la maturationdu raisin. Bull. O. I. V. 867, 868, 367–379.
Weiss, J., 1981. Rating scales in the sensory analysis of foodstuffs. II. Paradigmaticapplication of a rating method with unstructured scale. Acta Aliment. 10 (4),395–405.
White, R.E., 2003. Soils for Fine Wines. Oxford Univ. Press, p. 266.Williams, L.E., Araujo, F.J., 2002. Correlations among predawn leaf, midday leaf
and midday stem water potential and their correlations with other measures
of soil and plant water status in Vitis vinifera. J. Am. Soc. Hort. Sci. 127 (3),448–454.
Zufferey, V., Murisier, F., 2006. Appréciation de l’alimentation hydrique de la vignedans divers terroirs viticoles vaudois (Suisse) Estimation of plant hydraulicsof grapevine in various terroirs in the Canton of Vaud (Switzerland). In: VIthInternational Terroir Congress, Nyon.
