Impacto del cultivo de la vid en la variabilidad espacial de las propiedades químicas de un suelo de Denominación de Origen Ribeiro

Abstract

As a result of the complex interaction between different factors and formation processes, soil properties can show significant spatial variability, even over small distances. In addition, in cultivated soils, this spatial variability is strongly influenced by the type of tillage and management, which will be different if the crop is annual or perennial. The objective of this doctoral thesis is to analyze the spatial variability of various soil properties in a vineyard of the D.O. Ribeiro. These include general properties such as percentages of sand, silt, and clay, pH, organic matter, and attributes of the exchange complex; macronutrients (P, K, Ca, Mg;), micronutrients (Fe, Mn, Cu, Zn) and heavy metals (Cd, Cr, Ni, Pb) extracted with nitric acid (US-EPA method), cation exchange resin, DTPA, and Mehlich-3. Its spatial variability was analyzed both along a transect of small dimensions, and in a microplot of the vineyard. All these variables have been studied using statistical and multifractal analyses. In addition, the elements extracted with the Mehlich 3 solution have been studied through a geostatistical analysis. The test area is located in an experimental plot attached to the Galician Viticulture and Oenology Station (EVEGA), which is located in the municipality of Leiro within the Ribeiro Designation of Origin (Galicia, NW Spain). In the transect, 65 samples were taken, at intervals of 0.8 m, for a total length of 52 m. In the microplot, 80 samples were taken, located along five small transects in different positions within the crop distribution: one in the row of the crop, two in the tracks of the tractor used for tillage and two between the tracks of the tractor. Soil texture ranged from loamy to sandy loam. The average pH in the transect and the microplot was 5.30 and 4.95; the average organic matter content was 2.31% and 1.81%; and the average cationic exchange capacity was estimated at 7.56 and 6.65 cmol(+) Kg-1 of soil, respectively. The coefficients of variation of the sand, silt, and clay fractions, pH, organic matter, and cation exchange capacity were less than 9% in the transect and 15% in the microplot. The total Cu content extracted according to the US-EPA method amounted to 100.65 and 93.57 mg Kg-1 in the transect and the microplot, respectively. These values are higher than the background values that can be expected in natural soils of the studied area, of granitic lithology. The concentrations of trace elements extracted according to the US-EPA method were ordered according to the following decreasing sequence: Fe > Mn > Cu > Zn > Pb > Cr > Ni > Cd. Both in the transect and in the microplot, the concentrations of elements extracted with the Mehlich-3 solution were higher than those determined with cation exchange resin and DTPA, and presented different sequences in decreasing order. In the microplot, greater global differences were observed between one side and the other of the row of vines than between the row, entrerrodada and rodada zones. The partition function, q, of all the general properties and elements studied in the transect and the microplot presented a linear relationship with the scale , on a log-log scale, indicating that the variability space in the two sampling schemes presents a fractal nature. The generalized dimension spectra, Dq versus q, and the singularity spectra, f () versus , showed that the spatial variability of all soil properties and elements could be fitted to multifractal or pseudomonofractal models. Scale heterogeneity varied depending on the calculation method used (“box counting” in the transect and “gliding box” in the microplot), the element studied, and the extraction method. In the transect, the highest scale heterogeneity was obtained for Ni and Cd extracted with DTPA. In the microplot, the greatest heterogeneity corresponded to Ca extracted with DTPA and Zn extracted with Mehlich 3. The twelve elements extracted by means of the Mehlich-3 solution that were analyzed using geostatistics presented spatial dependence that was correctly described by means of spherical models, except in the case of Mn, whose best fit was that of an exponential model. The kriging maps showed important differences regarding the spatial distribution of these elements in the areas of the microplot that were sampled and between the different elements studied.