Please use this identifier to cite or link to this item: http://repositorio.ufla.br/jspui/handle/1/43583
Title: Soil survey as support to precision coffee crop and Winter wines development in Southeast Brazil
Other Titles: Levantamento de solos como suporte à cafeicultura de precisão e desenvolvimento dos vinhos de inverno no Sudeste do Brasil
Authors: Menezes, Michele Duarte de
Curi, Nilton
Mota, Renata Vieira da
Ferraz, Gabriel Araújo e Silva
Pozza, Adélia Aziz Alexandre
Silva, Sérgio Henrique Godinho da
Keywords: Soil survey
Coffee crop
Wine quality
Hierarchical cluster on principal components
Fuzzy logic
Levantamento de solo
Colheita de café
Qualidade do vinho
Lógica Fuzzy
Issue Date: 28-Oct-2020
Publisher: Universidade Federal de Lavras
Citation: GONÇALVES, M. G. M. Soil survey as support to precision coffee crop and Winter wines development in Southeast Brazil. 2020. 118 p. Tese (Doutorado em Ciência do Solo)-Universidade Federal de Lavras, Lavras, 2020.
Abstract: Soil surveys provide subsidies for several applications, including decision making on the management of several crops. Thus, this work had as objectives: i) to define a method by means of grouping analysis for the management zones outline based on data from soils and coffee plantations, collected in areas with defined land parcels and adapt this method in a management of culture already implanted, ii) to characterize the soils, the climate, as well as to verify their relation with the composition of the Winter Wines produced in seven commercial vineyards of the Syrah cultivar, iii) search for areas similar to the soil mapping units that include vineyards of the Southern region of Minas Gerais and verify the relationship between the wines and grapes produced in two reference vineyards. In the first part of this study, a series of tests were carried out involving selection of variables by Random Forest, reduction of dimensionality by principal component analysis (PCA) and factor analysis for mixed data (FAMD), ending with the generation of clusters by hierarchical cluster analysis on principal components (HCPC). The most important variables to explain coffee yield and thus compose the management zones outline, classified by Random Forest, were crop age, crop density, silt fraction and soil organic matter content. The PCA explained total variance of 76.1% in the first two dimensions. Three clusters with a statistically significant difference in coffee production were outlined by the HCPC. In general, the following sequence of cluster generated (123) was found, increasing the crop age, and the content of the silt fraction, and decreasing in the yield and crop density. In the second part of this study, climatic and soil characterization was carried out in seven commercial vineyards, including soil classification, chemical, physical and mineralogical analyses, as well as the identification of the parent material of the vineyard soils. The qualitative profile of wines from the Syrah vine was also characterized. Four groups of vineyards were formed from their similarities in terms of edaphoclimatic characteristics: a) soils with high levels of sand on the surface, in places with high rainfall, originated wines with lower pH; b) Soils with homogeneous clay contents along the profile, in vineyards with high thermal amplitude, presented intermediate values for most wine compounds; c) shallow and young soils with high sand content, in a vineyard with low precipitation and high temperature, produced wine with the highest flavanol content. This wine also has high levels of most other evaluated compounds; d) deep soils, with basalt as the parent material, is related to wines with the highest levels of most compounds, however, this is due to the late harvest carried out in this vineyard. The edaphoclimatic conditions were important for the characterization of the typicity of the wines, and such conditions associated with the handling of double pruning allowed the production of quality wines, compared to wines from world traditional wine-growing regions. The third part of this work involves the extraction of climatic, geological and terrain information from soil mapping units that contain commercial vineyards in two municipalities in Minas Gerais. Such information was applied, using fuzzy logic and similarity vectors in an area of interest (provenance area), in order to verify areas with greater similarity in relation to the conditions extracted from the mapping units (soil map). Most of the variables of the two mapping units, mainly the mean temperature, rainfall and evapotranspiration are very similar. The mapping units differ mainly in terms of the higher altitude in Três Corações and by the different parent material of the soil. Keywords: yield,
URI: http://repositorio.ufla.br/jspui/handle/1/56426583
Appears in Collections:Ciência do Solo - Doutorado (Teses)



Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.