Effects of liming on soil properties , leaf tissue cation composition and grape yield in a moderately acid vineyard soil . Influence on must and wine quality

1Instituto de Investigación de la Viña y el Vino (IIVV). Universidad de León. Avenida de Portugal, 41, CP 24071, León, España 2RGA Bio-Investigación, SL. Laboratorio 29. Instituto de Recursos Naturales. Avenida de Portugal, 41, CP 24071, León, España 3Instituto Valenciano de Investigaciones Agrarias-IVIA, Centro para el Desarrollo de la Agricultura Sostenible, Crta. Moncada-Náquera km 4.5, CP 46113 Moncada, Valencia, España


Introduction
With an approximate land area affected by soil acidity of 4 billion ha, representing about 30% of the total ice-free land area of the world (Sumner and Noble, 2003), this edaphic characteristic is one of the most yield-limiting factors for crop production as a result of high Al 3+ levels and low availability of basic ions in both surface and subsurface soil.
Liming is one of the main methods used by farmers to enhance the fertility of acid soils because it decreases the contents of exchangeable Al 3+ through replacement with Ca 2+ and Mg 2+ , and also the contents of soluble Al 3+ by precipitation with the hydroxyl anions generated by carbonate hydrolysis in the soil solution.Additionally, liming has beneficial effects on soil structure since it improves soil structural stability (Lanyon, 2001;Hansen and Cass, 2002), and it plays a significant role in the control of certain plant pathogens (Havlin et al., 2014).The effectiveness of liming materials depends on (i) their neutralizing power, which is accounted for by their calcium carbonate equivalent (CCE), and (ii) their fineness (Edmeades and Ridley, 2003;Álvarez et al., 2009).However, soil has a huge buffering capacity that is able to diminish the effects of all kinds of amendments, and the soil acidification process is accelerated by factors such as acid rain and excess application of ammonium-based inorganic nitrogen fertilizers (Zheng, 2010).
Liming is usually performed before vineyard implantation, increasing the soil pH and cation exchange capacity (CEC) values (Joris et al., 2012).Appropriate soil liming after vineyard planting is a high-potential strategy to minimize the toxic effects of Al in established vineyards.However, some common liming materials, such as limestone (CaCO 3 ), used in agriculture to increase yield potential lose their effectiveness in vineyards established on acid soils due to the low solubility and mobility of carbonate in the soil profile, the liming effects being usually limited to the lime application/incorporation sites (Soratto and Crusciol, 2008).Among certain strategies used to ameliorate subsurface soil acidity, it can be cited the replacement of lime by the surface application of more soluble materials (Castro and Crusciol, 2013).Dolomitic limestone and sugar foams are two materials often used for liming.Dolomitic limestone comprises mainly the mineral dolomite, which is made of a calcium and magnesium double carbonate (CaMg(CO 3 ) 2 ).Two important characteristics of this liming material are (i) its high neutralizing capacity, which is higher than that of limestone because of the lower atomic weight of magnesium compared to calcium, and (ii) its low dissolution rate, which is approximately 100-fold lower than that of calcite (Loeppert and Suarez, 1996).Sugar foams are sugar beet-manufacturing residues, which arise from the purification-flocculation of colloid matter in the beet extract by treatment with lime and carbon dioxide (Vidal et al., 2006).This industrial by-product can be used to correct soil acidity and Al phytotoxicity in acid soils because of its high content in active lime.In addition, this liming material contains abundant organic matter and several essential micronutrients (Vidal et al., 1997).
Although the application of sugar foams to the soil is a common practice used worldwide, especially in the last 20 or 30 years due to the improvements observed in some soil properties (Sikora and Azad, 1993), little is known about its effects on the quantity and quality of grapes cultivated on acid soils.Sugar foams may effectively counteract the symptoms of aluminum toxicity in plants, which brings about poor growth and low yields.Specifically, in old vineyards on acid soils, the root system ends abruptly at the depth where the pH value drops and the Al 3+ concentration becomes relatively important (Meyer et al., 1984).In the same way, Cançado et al. (2009) reported that, as in many other plant species, the primary site and target of Al stress in grapevine rootstocks is confined to the actively growing root tip, revealed by a severe inhibition of root growth.Moreover, Kirchhof et al. (1991), in their research on Vitis vinifera cv.Chardonnay, showed how acidic soil conditions, mainly in the subsoil, depressed growth of vine root system.Finally, Wooldridge et al. (2010), on a mixture of cv.Pinot noir and Chardonnay with several rootstock vines cultivated on acid soils, found a more reduced root stress when compared to those cultivated on unlimed soils.
Although extensive research has been carried out on the effects of liming on the properties of acid soils, its effects on nutrient distribution in vine vegetative parts (leaf blades and petioles) and grape berries as well as on crop yield and must and wine quality are still not fully understood.So far the influence of soil liming on Ca, Mg and K contents in grape berries has not been investigated.This is an important knowledge gap because these elements remarkably affect grape juice quality and therefore the vinification process.In the present work, we analyzed the effects of adding two different lime amendments, i.e. dolomitic limestone and sugar foam, on (i) soil acidity, (ii) concentration ranges of Ca, Mg and K in leaf tissues, i.e. blades and petioles, and grape berries, (iii) grape yield and, finally, (iv) must and wine quality characteristics.

Study site
A commercial vineyard located 556 meters above sea level in the municipality of Cacabelos (León; Spain; 42°36'N lat and 6°45'W long) was selected as the study site (Figure 1).From a climatic point of view, the grape growing region would be classified as Region I (≤ 1,390 Celsius degree-days) based on the classification system of Amerine and Winkler (Jackson, 2014).The mean reference evapotranspiration (FAO Penman-Monteith) and mean rainfall were, respectively, 922 and 616 mm yr -1 for 2000-2011(SIAR, 2012)).From a bioclimatic point of view, the site would be classified as upper meso-Mediterranean based on the «thermotypes» classification and subhumid based on the «ombrotypes» classification (IGME, 1995).The soil under study corresponded to an Inceptisol, suborder Xerept, group Haploxerept according to Soil Survey Staff (2010).The parent material of the soils in the study area is made of Tertiary sediments (IGME, 1995).Accordingly, Fe oxyhydroxides are the most common clay minerals in the vineyard soils developed on these Tertiary sediments (Fernández-Calviño et al., 2009), from which calcium minerals are almost completely absent.

Experimental treatments
The research was conducted on a 60-year-old Vitis vinifera L. cv.Mencía vineyard over three consecutive seasons (2009, 2010 and 2011), and the planting density was 4,166 (1.5 m × 1.6 m) plants per hectare.The plants had been grafted on Rupestris du Lot rootstock, which is considered as highly sensitive to soil acidity (Fráguas, 1999).Planting lines displayed a north-south orientation.The trellising system involved head training, with 4-5 arms per plant.Winter pruning left thumb-sized arms with two buds.The vineyard had no irrigation system.No fertilizers or extra amendments other than those used in this research were applied during the study period.The liming factor was applied at three levels (control (C), dolomitic limestone (DL) and sugar foam (SF)), with three replications per level.The study plot was split into nine 450-m 2 subplots.The treatment replications were distributed among the nine subplots in a completely random design.

Characterization of the liming materials and liming dose
The composition of the two liming materials used in this study (Villa, 2005) is shown in Table 1.The dolomitic limestone exhibited a higher Mg content than the sugar foam, whereas this latter presented a higher organic matter content.The high Ca content of the sugar foam is due mainly to the presence of Ca in the form of slaked lime (Ca(OH) 2 ) and, to a lesser extent, as carbonate (CaCO 3 ) (Espejo, 2001).The slaked lime progressively reacts with atmospheric carbon dioxide to produce CaCO 3 .This carbonation occurs at a rate that depends on the aggregate size and porosity, and water content of the sugar foam.
The liming rates were established with the aim of decreasing the aluminum saturation of the effective cation exchange capacity (%Al/CECe) down to 20%.Achieving a %Al/CECe of 20% ensures an adequate degree of base saturation, i.e. 80%, required by most crop plant species (Fageria and Baligar, 2008).Specifically, the lime requirement was calculated using the known Cochrane's formula (Cochrane et al., 1980) and resulted to be 2,720 kg CCE ha -1 , which corresponded to 2,800 kg of dolomitic lime ha -1 and 3,900 kg of sugar foam ha -1 according to the CCE contents of each material (Table 1).The dolomitic limestone was in a powdery state, whereas the sugar foam consisted of aggregates of variable size, which were manually disaggregated before incorporation into the soil.The liming materials were uniformly spread onto the entire surface of the subplots and incorporated with one-pass tillage in November 2008.

Soil sampling and analyses
Before the amendments were added, the following soil properties were evaluated at two soil depths (0-30 and 30-60 cm) using the methods indicated below: texture, soil organic matter (SOM), pH in water (pH W ), pH in KCl 1N (pH KCl ), electrical conductivity (EC), phosphorus, Ca, Mg, K and Al contents, the percentage of base saturation (S) and the micronutrient contents (Fe, Cu, Mn, and Zn) (Table 2).
After the amendments were applied in late November 2008, the effects of liming on the following soil properties were monitored: pH W , S, and exchangeable Ca, Mg, K and Al contents.This monitoring was conducted by sampling the soil at 0-30 and 30-60 cm depths at the veraison stage, which occurs in August in the study area.
The soil samples were collected using an auger, transported to the laboratory in plastic bags and airdried at room temperature.Next, they were (g kg -1 ) (g kg -1 ) (g kg -1 ) (g kg -1 ) (mg kg -1 ) (mg kg -1 ) (mg kg -1 ) (mg kg -1 ) (mg kg -1 ) (g kg -1 ) (g kg  disaggregated to pass through a 2-mm mesh sieve and analyzed.Textural classes according to USDA were determined by the Bouyoucos hydrometer method (Bouyoucos, 1962).Official methods of analysis (MAPA, 1993) were used for the determination of (i) SOM by wet oxidation followed by titration with ferrous ammonium sulfate, (ii) pH in a soil:water (1:2.5)suspension, (iii) EC at 25°C (EC 25 ) in a soil:water (1:2.5)suspension, and (iv) the content of exchangeable cations (Ca, Mg, and K) by successive extraction with ammonium acetate 1M (NH 4 C 2 H 3 O 2 ) and subsequent analysis of the displaced cations by atomic absorption spectrometry (AAS).The exchangeable Al was determined by inductively coupled plasma atomic emission spectroscopy (ICP-AES) using KCl 1M as the extraction solution (Little, 1964).The micronutrients (Fe, Cu, Mn, and Zn) were extracted following the method of Lindsay and Norvell (1978) using a buffer solution of diethylenetriaminepentaacetic acid (DTPA) 0.005 M and calcium chloride (CaCl 2 ) 0.01 M buffered to pH 7.3 and determined by AAS.The P levels were determined by UV-visible spectroscopy after successive extraction with sodium bicarbonate 0.5 M at pH 8.5 following the method proposed by Olsen et al. (1954).CECe was obtained from the arithmetic sum of the concentrations of the exchangeable Ca, Mg, K and Al. S was calculated by dividing the sum of Ca, Mg and K by the CECe.

Leaf sampling and analyses
The Ca, Mg and K contents in blades (Cab, Mgb and Kb, respectively) and petioles (Cap, Mgp, and Kp, respectively) were annually monitored along with the soil properties at the veraison stage.Specifically, 30 basal leaves opposite grape bunches were randomly collected per subplot each year in early August.The leaves were sealed in plastic bags, transported to the laboratory, carefully rinsed with abundant deionized water, and then dried for three days at 70°C (Bavaresco et al., 2010).Then, they were wet digested with an acid mixture of perchloric, sulphuric and nitric acid at 420 °C for 20 minutes (Calleja, 1978), and the cation contents in the extracts determined by ICP-AES.

Grape sampling and analyses
The grapes were sampled at harvest every year on 12th, 20th and 11th September of 2009, 2010 and 2011, respectively.The harvested grapes from each subplot were weighted to determine yield (GY).
Next, 300 grape berries per subplot were randomly chosen to determine the weight of 100 berries (W100) and must quality.In order to calculate W100, 100 of these berries were weighted.The grape must of each subplot was obtained manually from the 300 berries by gently pressing the grapes, using rubber gloves to avoid sample contamination.In the must obtained, the following quality properties were determined: (i) real acidity (pH), (ii) total soluble solids (TSS), which were measured using a refractometer, and (iii) total acidity (TA), which was determined by titration of the grape must with sodium hydroxide (0.1 N) to an endpoint of pH 7, and expressed as the equivalent content of tartaric acid (ta) in g l -1 (OIV, 2014).
The content of each nutrient in grape berries must be investigated in the berry part that acts as its main reservoir.Thus, the concentrations of Ca and Mg were studied in grape seeds (Cag and Mgg, respectively), whereas K content (Kg) was studied in grape skins (Conde et al., 2007;Keller, 2010).The seeds and skins from 100 grapes were manually separated from the flesh and immediately dried at 60°C to constant weight.Ca and Mg in dried seeds and K in dried skins were determined by ICP-AES after wet digestion with an acid mixture of perchloric, sulphuric and nitric acid at 420 °C for 20 minutes (Calleja, 1978).

Winemaking process
Winemaking was performed in three consecutive vintages (2009)(2010)(2011) for all of the biological replicates.The damaged grape clusters (those broken or with visual microbial alterations) were discarded in order to avoid undesirable contamination and degradation compounds.Selected clusters were destemmed, crushed and macerated for 24-48 hours at 10°C (with 0.05 g l -1 of SO 2 ) in 1,000-l stainless steel (AISI 3162B) tanks with cooling jacket (Industrias Céspedes e Hijos, SL, Salvatierra de Miño, Spain).After correcting total acidity (6.5 g l -1 of tartaric acid), the grape extract was inoculated with 0.1 g l -1 of previously hydrated commercial yeast Saccharomyces bayanus P51 and Saccharomyces bayanus Levuline FB (AZ3 Oeno SL, Hernani, Gipuzkoa, Spain).Alcoholic fermentation (AF) was carried out at 25-28°C.The juice was drawn and pumped over the cap daily during the fermentation period.Fermentation was considered completed when both density reached approximately 0.90 g ml -1 (measured using a hydrometer) and glucose and fructose were consumed (measured by enzymatic analysis).
At the end of AF, the wine was transferred to another tank and left on its lees for 24-48 hours.After that, the wine was transferred again to another tank and left on its light lees for 6-8 weeks, removing the wine periodically with nitrogen gas.After that, the wine was transferred into 225-l medium toast, fine grain French oak barrels (for economic reasons, 2nd use in 2009 and 3rd use in 2010 and 2011) where malolactic fermentation took place.The wine was maintained for 12 months in these barrels.

Tannin concentration and wine color analyses
Tannin concentration (Tc, in g l -1 ) was determined by the method proposed by Glories (1988, cited by Ribéreau-Gayon et al., 2006), whereas hue, color index (C.I.), total polyphenol index (TPI) and total anthocyanin content (TAnth, in mg l -1 ) were determined by absorbance measurements (A, in nm) in UV-visible spectroscopy (Ruiz Hernández, 2002).More specifically, hue was evaluated as the quotient of A 420 and A 520 , C.I. as the summation of A 420 , A 520 and A 620 multiplied by the quotient of factor dilution and optical path, TPI as the multiplication between A 280 and the factor dilution and, finally, TAnth as the multiplication between A 520 , the conversion factor 22.76 and the factor dilution.Samples were collected both before and after must AF.
A manifold system (Waters, Barcelona, Spain) was used for SPE.All cartridges were conditioned by rinsing with 3 ml of methanol and 3 ml of Milli-Q water.The wine sample to be extracted (2 ml) was acidified with 1 N H 2 SO 4 (0.5 ml) prior to loading onto the conditioned cartridge.All cartridges were washed with 5 ml of Milli-Q water.The retained phenolic compounds were eluted with 5 ml of ethyl acetate, followed by 5 ml of methanol.The fractions were collected separately to ascertain whether some of the phenolic compounds were eluted during the washing step.
The gradient was linear at a flow rate of 1 ml/min from 0 to 25% solvent B for 15 minutes, from 25 to 70% B for 20 minutes, and from 70 to 100% B for 25 minutes, then isocratic for 5 minutes, and after that from 100 to 0% B for 5 minutes and, finally, isocratic for 20 minutes.
The phenolic compounds analyzed were identified by comparing their retention times and UV-vis spectra with their respective standards.Cyanidin-3-O-lathyroside chloride was used as external standard.
The quantification of the different phenolic compounds was carried out at 520 nm for Malv and 280 nm for the rest of phenolic compounds.The samples were thermostated at 4°C.The quantification of the different phenolic compounds was carried out applying each calibration line constructed using the corresponding standards.The linear calibration of all compounds was satisfactory with squared-R values above 0.99.

Comparisons between the liming treatments
Statistical analyses were performed using R software (R Core Team, 2015).Several analyses of variance (ANOVAs) were carried out to study the effect of liming, with three levels or groups, namely control, dolomitic limestone and sugar foam, on (i) the soil chemical properties (pH W , S, Ca, Mg, K and Al), (ii) the leaf nutrient contents (Cab, Mgb, Kb, Cap, Mgp and Kp), (iii) the grape yield components (GY and W 100 ), (iv) the grape must quality properties (pH, TSS, TA), (v) the grape nutrient contents (Cag, Mgg and Kg), (vi) the tannin concentration (Tc), (vii) the wine color parameters (hue, C.I., TPI and TAnth) and, finally, (viii) the wine phenolic compound content.
In the ANOVAs of the soil chemical properties, the soil depth was also taken into account as a block factor with two levels or groups, i.e. shallow soil (0-30 cm) and deep soil (30-60 cm).In all the ANOVAs the year of sampling with three levels (2009, 2010 and 2011) was also included as a block factor.Therefore, a three-way ANOVA (liming, soil depth and sampling year) was used for each one of the soil properties, whereas a two-way ANOVA (liming and sampling year) was used for the leaf and grape nutrient contents, as well as for grape yield, must quality properties, tannin concentration and wine color measurements.Since the wine ageing conditions changed throughout the investigation period, all phenolic compounds were studied as a function of year and liming directly.
In case the liming factor presented a significant effect on any of the properties, the corresponding "between groups" variance was split into two independent variance contributions or main effects, namely the effect of just liming and the effect of the specific material used to lime.Given the experimental design, these are the only two possible causes leading to a significant effect of the liming factor.To evaluate if either or both effects explain a likely significant ANOVA, two orthogonal contrasts were planned beforehand.Through contrast 1 (C1), liming was compared against no liming, i.e. the dolomitic limestone and sugar foam levels of the liming factor were jointly compared against the control.Through contrast 2 (C2), the liming materials were compared against each other, i.e. sugar foam versus dolomitic limestone.These orthogonal contrasts allowed us to measure the effect sizes in a standardized way.
To carry out an ANOVA the hypotheses of univariate normal distribution and homoscedasticity of the data have to be tested in advance.The univariate normality hypotheses for every variable were tested using the Kolmogorov-Smirnov test.A confidence level of P=0.05 was used to evaluate the significance of the test.All soil, leaf (blades and petioles), yield, must and wine parameters presented univariate normality.Additionally, the ANOVA is fairly robust in terms of the error rate associated to violations of the assumption of homogeneity of variance (homoscedasticity) when sample sizes are equal (Field, 2009) as occurs in the present study.

Soil characterization before liming
The baseline characteristics of the two soil depths before liming are shown in Table 2.The Al content was 1.58 ± 0.89 cmol(+) kg -1 (95% CI in the 0-30 cm layer, and 1.38 ± 0.57 cmol(+) kg -1 (95% CI) in the 30-60 cm layer.These exchangeable Al contents are over the 20% limit in both depth intervals.The 20% limit is considered the highest Al saturation that most plants can tolerate (Fageria and Baligar, 2008).The important exchangeable soil acidity in the plot is further indicated by the larger-than-one differences between pH W and pH KCI at both soil depths.Low exchangeable Ca, K and Mg contents were also found, as well as low P content.Conversely, the micronutrients (Fe, Mn, and Cu) exceeded the levels considered suitable for soil fertility.The high Cu values were likely due to cupric-based fungicides that are frequently applied to the local vineyards.

Soil properties
The time evolution of pH W , Ca, Mg, K, Al and S levels for the treatment and control subplots throughout the three years of monitoring is shown in Figures 2 and 3. Apparently, there are some differences in performance efficiency between both liming materials.More specifically, sugar foam seemed to decrease Al levels more than dolomitic limestone.
An ANOVA was used to investigate if the differences between liming treatments were statistically significant, and furthermore, if they depended on the soil depth, the year of sampling, and the interactions between both.According to the ANOVA, there was a significant effect (P<0.05) of liming in all soil properties (Table 3).By contrast, the effect of soil depth was not significant (P>0.05) in any of the soil properties.Additionally, the effect of liming did not significantly change with the soil depth, which was revealed by the non-significant interaction (P>0.05) between liming and soil depth factors (L x D).There was a significant effect (P<0.05) of year of sampling on exchangeable Ca and K levels.Even though the interactions between liming and year of sampling (L x Y), depth and year of sampling (D x Y), and all three factors (L x D x Y) were non-significant in all soil properties, Ca and K were studied as a function of year and liming, whereas pH W , Mg, Al and S levels were studied as a function of just liming.To know if there were significant differences between liming materials, the "between groups" variance was split into two summands through two contrasts (Table 4).# 0.47 ± 0.16 cmol(+) kg -1 (95% CI), respectively.These increments in both exchangeable cations (Ca and Mg), as well as the decrease in Al saturation caused by liming, were summed up by the increase in S. Specifically, at the end of the investigation, the average S level increase in limed subplots was 33.2 ± 11.9% (95% CI) compared to control subplots.
Similarly, the different effect of the liming material sugar foam in comparison to dolomitic limestone (C2) was observed in pH W and S, which increased by 0.52 ± 0.51 (95% CI) and 17.0 ± 16.4 (95% CI), respectively, and Al, which decreased by 0.57 ± 0.81 (95% CI).The inverse relationship between levels of Ca, Mg and S on the one hand and Al on the other can be observed in Figure 4.

Leaf nutrient contents
Means and standard errors of blade and petiole nutrient contents, arranged by liming material, are shown in Figure 5. ANOVA was used to determine if the slight differences between liming treatments were statistically significant, and furthermore, if they depended on the year of sampling and the interactions between these two factors.According to the ANOVA, there were no significant effects of the liming factor on any of the nutrient contents in both    blades and petioles (Table 5).Interestingly, there was a significant effect (P<0.05) of year of sampling on Ca and Mg content in blades and petioles, as well as on K content in blades.Additionally, the interaction between liming and sampling year (L x Y) was nonsignificant (P>0.05) in all the nutrients evaluated.

Yield, must quality and grape nutrient contents
Means and standard errors of grape nutrient contents (Cag, Mgg and Kg), yield parameters (GY and W100) and must quality properties (pH, TA and TSS) in the treatment and control subplots are shown in Table 6.
According to the ANOVA, liming had no significant effects on any of the nutrient contents.Interestingly, there was a significant effect of year of sampling on Mgg (P<0.05) and Kg (P<0.001).The nonsignificant interaction between liming and sampling year (L x Y) (P<0.05)revealed that the effect of sampling year on grape nutrient contents did not change with liming.Additionally, there were no significant effects of the liming factor on any of the yield parameters and must quality properties except for TA (P<0.01).In the year 2011, both C1 and C2 were significant (P<0.05) in the above parameter.
Interestingly, there was a significant effect of year of sampling on pH (P<0.01) as well as on W100 and  Average values are displayed within bar graphs.Standard errors are shown as bars (± 1 SE mean).The shaded areas in the graphs illustrate the concentration ranges that are optimal for adequate vine growth at veraison according to Fregoni (2005).C: control; DL: dolomitic limestone; SF: sugar foam.
TA (P<0.001).It is also interesting to note the absence of significant interactions (P>0.05) between liming and sampling year (L x Y).Although there are no significant differences with the control subplots in terms of productivity, an increasing trend in grape yields could be observed in limed subplots.Specifically, just liming with either sugar foam or dolomite limestone increased grape yields by nearly 30% at the end of the experiment (2011).These results might be important to farmers in economic terms.

Tannin concentration, color measurements and phenolic compounds in wines
There was a significant effect (P<0.05) of sampling year, as well as liming and sampling year (L x Y), on tannin and all wine color parameters measured both at the beginning and the end of AF.Based on the foregoing, means and standard errors of those parameters arranged by liming material and sampling year are shown in Table 7, and were studied as a function of both factors (Table 8).While many significant differences have appeared because of liming, the results obtained only showed one consistent trend: liming significantly increased hue after AF.No particular trend before and after AF was found with liming treatments compared with control.Interestingly, TAnth levels measured before AF in both control and liming treatments in the 2010 vintage were very different against the other two vintages (2009 and 2011).It was not possible to associate this fact to any particular cause.
Data for phenolic compounds (Gall, Caff, Cou, Cat, Epi, Resv, Malv and Fer), arranged by liming material and sampling year, are shown in Table 9.
Due to the differences in ranges of measurement for each of the phenolic compounds, all the measures were scaled prior to be plotted (Figure 6).
The results obtained by applying one-way ANOVAs revealed that significant differences (P<0.05)existed in most of the quantified phenolic compounds.However, the results showed only two consistent trends: liming significantly (P<0.05)increased Gall and Cat after ageing throughout the experiment.
Similarly to tannin concentration and color measurements, no particular trend before and after   GY (grape yield); W100 (weight of 100 berries); TA (total acidity expressed as the equivalent content of tartaric acid (ta) in g l -1 ); TSS (total soluble solids); Cag, Mgg and Kg (calcium, magnesium and potassium contents in grapes, respectively).

Discussion
In the Haploxerept acid soil studied in this work, liming was effective in decreasing soil acidity, and specifically, sugar foam was more effective than dolomitic limestone.These results agree with those of García Navarro et al. (2009) who, in their assays on Anthrosols and Luvisols (FAO, 2006), reported that the application of sugar foam on dryland crops produced a positive impact on soil fertility through an increase in soil pH W and exchangeable Ca contents.Additionally, although liming materials were incorporated with one-pass tillage at a soil depth of 20-30 cm, no significant differences in soil properties were found between the arable and the underlying soil layer (0-30 and 30-60 cm).
In spite of the higher CCE of dolomitic limestone compared to sugar foam, the latter was more efficient  as liming material than the former.This remarkable higher efficiency of sugar foam over dolomitic limestone could be explained by the higher solubility of Ca(OH) 2 over CaMg(CO 3 ) 2 , which is roughly three orders of magnitude higher.Similarly, the higher exchangeable Mg in soils limed with sugar foam compared to dolomitic limestone subplots could be explained by the higher solubility of Mg(OH) 2 compared to CaMg(CO 3 ) 2 .These results agree with those of Augustin et al. (1997), who investigated the chemical reactions of magnesium fertilizers when they are applied to the soil, and suggested that magnesium fertilizers can be grouped into four classes according to their dissolution rates: MgSO 4 > MgO ≈ Mg(OH) 2 > slag lime ≈ dolomitic limestone ≈ magnesite > basalt.These results differ from the previous findings of Vidal et al. (2006) 9. Effects and effect sizes of liming against control (C1) and liming with sugar foam against liming with dolomitic limestone (C2) on wine phenolic compounds after alcoholic fermentation (AF) and after ageing.
**significant at the P<0.05 level;**significant at the P<0.01 level; ***significant at the P<0.001 level.pH W , pH in water; S, percentage of base saturation; gallic acid (Gall), Caffeic acid (Caff), p-Coumaric acid (Cou), (+)-catechin (Cat), (-)-epicatechin (Epi), trans-resveratrol (Resv), malvidin-3-O-glucoside (Malv) and trans-ferulic acid (Fer); Mean d (mean differences between control and amendments for contrast 1, and between sugar foam and dolomite for contrast 2); df R (degrees of freedom for the residuals of the model); r (effect size).dolomitic limestone showed significantly more exchangeable Mg than those treated with sugar foam.This indicates the need for caution when comparing data derived from laboratory and field studies.
Although chemical soil analysis indicates the potential availability of some nutrients that vine roots may take up, leaf analysis is often the most reliable method of assessing vine nutritional status.
Moreover, because the vine integrates the variability in the soil explored by its roots, leaf analysis is often the basis of the fertilizer recommendation programs for vines.However, in the present study, data from soil testing and leaf analysis have indicated that there is no clear relationship between them.Although blades have higher sensitivity to detect individual macronutrient deficiencies or excesses (   Assuming that liming did not significantly increase yield grape, an increasing trend has been found in limed subplots.The increase sequence in yield data (C DL SF) is consistent with those obtained for Ca and Mg data (C DL SF), as well as for Al data (C DL SF).In view of the foregoing considerations, soil liming may be advisable close to the vine root system.However, in many established vineyards the use of lime could be difficult as a result of the existence of trellis systems.
Although soils that are rich in Ca promote the accumulation of sugar (Jiang and Zhang, 2012), the findings of the current study did not show any significant increase in this parameter.Despite the fact that liming did not cause significant differences in yield quality properties, it is interesting to note that TA decreased more in control than in limed subplots (particularly during the last two years) (Figure 7), and more when sugar foam was used.A possible explanation for these results may be related to the grape K content, which, when excessive, may decrease must TA resulting in an increase in must pH (Conde et al., 2007).The higher K content in leaf tissues showed by limed subplots (although not significant) does not seem to be in agreement with a reduction in K uptake in limed soils.However, this issue could originate from the combined effect of rainfall during the annual vegetative cycle and liming.Thus, Chatonnet (2005) reported that late rainfall close to harvest leads to fast K uptake by vine roots, which drops must TA without necessarily affecting must pH.In 2011, an abnormal rainy August (34.6 mm) and rainfall (8.2 mm) ten days before harvest occurred.Although these conditions are similar to those reported by Chatonnet, no significant differences in Kg were found in this research and several questions remain unanswered.
As reported by Yokotsuka et al. (1999), the addition of limestone to the native soil significantly influenced both tannin concentration and color parameters in Mencía wines in every vintage of the experiment; however, there was a weak consistency in the trends observed.This, together with the results obtained in blades and petioles as well as in grapes and yield, suggests that the specific physiological characteristics of the combination of rootstock and scion, along with marked nutrient storage in the woody parts of the vine, make assessing grapevine responses to amendment application particularly challenging (Jackson, 2014;Olego and Garzón, 2014).However, since soil chemistry is an important factor in terroir, vine nutrition can be instrumental to the specific growth pattern of the vine and thus cause a specific canopy architecture and therefore a ripening pattern (Goode, 2013).
Wine quality is the result of a complex set of interactions (terroir effect), which include soil variables.With respect to the vine's productive system, soil and climate represent the two physical elements fundamental to understand the concept of terroir (van Leeuwen and Seguin, 2006); compared to climate, soil is an even more complex system.Thus, the impact of the pedologic environment on wine quality is difficult to analyze singularly and separately from the rest of the productive system (Seguin, 1986).Accordingly, it is difficult to study the effects of soils that have been artificially modified to improve yield on grape and wine composition, because the soil influences hydro and mineral nutrition and the functioning of the root system.Although phenolic compounds are heavily influenced by the vintage, soil type, mesoclimate and cultural practices (Nadal, 2010), as for the measurements of tannin concentration and color parameters, no definite trends (with consistency before and after wine ageing) in phenolic compounds have been found in the present study as a consequence of liming.Caution must be taken in interpreting the influence of soil amendments on phenolic compounds, as climatic differences between vintages added uncertainty to their interpretation.Additionally, phenolic compounds in wines varied between seasons, regardless of variations in winemaking practices (i.e.different age in the French oak barrels used per vintage) which also could affect wine composition (Cohen and Kennedy, 2010).However, although the composition of wine is heavily dependent on the winemaking technique (Cortell et al., 2008), understanding the impacts of soil types on the phenolic compounds in grapes and wines is necessary for developing an effective viticulture management program (Li et al., 2011).

Conclusion
As expected, liming with either sugar foam or dolomitic limestone decreased soil exchangeable aluminum, improved supply of calcium and magnesium, and increased soil pH and base saturation in acid Haploxerept cultivated with Vitis vinifera L. cv.Mencía.However, sugar foam behaved as a more efficient liming material than dolomitic limestone.This fact may be explained by the higher solubility of sugar foam calcium and magnesium components in the soil.The reduced efficiency of dolomitic limestone might be partly offset by a greater durability.
Although the increasing trend in grape yield with liming, more pronounced with sugar foam than with dolomitic limestone, was not significant, this might have an important economic impact.Finally, no significant trends in properties related to must and wine quality were found.Maybe more evident trends in these must and wine properties would have appeared using higher liming doses or extending the monitoring span to more years.Further research needs to examine more closely the impact of the changes in soil composition on wine quality.

Figure 1 .
Figure 1.Map of the location of the research area (latitude and longitude are showed in Universal Transverse Mercator coordinates).

Figure 4 .
Figure 4. Relationship between Ca, Mg, Al and S levels in soil as a function of just liming.Al in cmol(+) kg-1; S (percentage of base saturation) in %.

Figure 7 .
Figure 7. Monthly rainfall (top) and differences in total acidity (TA) between control and limed subplots (bottom), during the period of investigation.Dashed lines correspond to months of harvest (Sp: September).

Table 3 . Factorial analysis of variance performed on soil properties at veraison stage.
significant at the P<0.05 level; **significant at the P<0.01 level; ***significant at the P<0.001 level; df M , degrees of freedom of the model; df R , degrees of freedom of the residuals of the model.pH W , pH in water; S, percentage of bas saturation; Ca, Mg, K and Al (exchangeable Ca, Mg, K and Al, respectively). *

Table 4 . Effects and effect sizes of liming against control (C1), and liming with sugar foam against liming with dolomitic limestone (C2), on soil properties at veraison stage.
*significant at the P<0.05 level; **significant at the P<0.01 level; ***significant at the P<0.001 level.pHW , pH in water; S, percentage of base saturation; Ca, Mg, K and Al (exchangeable Ca, Mg, K and Al, respectively); Mean d (mean differences between control and amendments for contrast 1, and between sugar foam and dolomite for contrast 2); df R (degrees of freedom for the residuals of the model); r (effect size).F P-value F P-value F P-value F P-value F P-value F P-# ' ' ' ' ' ' ' ' ' ' ' '

Table 5 . Factorial analysis of variance performed on calcium, magnesium and potassium content in blades and petioles at veraison stage.
significant at the P<0.05 level; ***significant at the P<0.001 level; df M , degrees of freedom of the model; df R , degrees of freedom for the residuals of the model; Cab, Mgb and Kb (calcium, magnesium and potassium contents in blades respectively); Cap, Mgp, and Kp (calcium, magnesium and potassium contents in petioles, respectively). *

Table 8 . Effects and effect sizes of liming against control (C1) and liming with sugar foam against liming with dolomitic limestone (C2) on tannin concentration and color measurements before and after alcoholic fermentation (AF) in must and wines respectively.
Mean d (mean differences between control and amendments for contrast 1, and between sugar foam and dolomite for contrast 2); df R (degrees of freedom for the residuals of the model); r (effect size).