ABSTRACT Conventional tillage and intensive machinery traffic are the major causes of physical soil degradation in sugarcane fields. This study evaluates the impact of adopting conservation management practices during sugarcane planting on soil physical properties and stalk yield of sugarcane in the municipality of Ibitinga, state of São Paulo, Brazil. The experimental design (split-block) included four cover crops and three soil tillage systems, with three repetitions. For comparison purposes, a control treatment was also included (without cover crop and under conventional tillage). Sampling for soil physical analysis was performed in three layers that coincide with soil horizons A (0.00-0.20 m), AB (0.20-0.30 m), and Bt (0.30-0.70 m), during cane-plant and first sugarcane ratoon cycles. The results showed that cultivation of sunn hemp associated with deep subsoiling induced high stalk yield of sugarcane in both production cycles, cane plant (116 Mg ha -1 ) and first ratoon (114 Mg ha -1 ), with a net gain of 11 and 9 Mg ha -1 compared with the control treatment, respectively. However, these results were not sufficient to induce significant differences in sugarcane yield. Nonetheless, the use of sunn hemp and millet, associated with subsoiling (at 0.40 or 0.70 m depth) during sugarcane planting, are promising management strategies to sustain better soil’s physical quality when compared to traditional management, conventional soil tillage without cover crops and/or cash crop, as peanuts, that increase the risks of soil compaction and physical degradation.

ABSTRACT: The Random Forest algorithm is a data mining technique used for classifying attributes in order of importance to explain the variation in an attribute-target, as soil CO2 flux. This study aimed to identify prediction of soil CO2 flux variables in management systems of sugarcane through the machine-learning algorithm called Random Forest. Two different management areas of sugarcane in the state of São Paulo, Brazil, were selected: burned and green. In each area, we assembled a sampling grid with 81 georeferenced points to assess soil CO2 flux through automated portable soil gas chamber with measuring spectroscopy in the infrared during the dry season of 2011 and the rainy season of 2012. In addition, we sampled the soil to evaluate physical, chemical, and microbiological attributes. For data interpretation, we used the Random Forest algorithm, based on the combination of predicted decision trees (machine learning algorithms) in which every tree depends on the values of a random vector sampled independently with the same distribution to all the trees of the forest. The results indicated that clay content in the soil was the most important attribute to explain the CO2 flux in the areas studied during the evaluated period. The use of the Random Forest algorithm originated a model with a good fit (R2 = 0.80) for predicted and observed values.

ABSTRACT: The use of data mining is a promising alternative to predict soil respiration from correlated variables. Our objective was to build a model using variable selection and decision tree induction to predict different levels of soil respiration, taking into account physical, chemical and microbiological variables of soil as well as precipitation in renewal of sugarcane areas. The original dataset was composed of 19 variables (18 independent variables and one dependent (or response) variable). The variable-target refers to soil respiration as the target classification. Due to a large number of variables, a procedure for variable selection was conducted to remove those with low correlation with the variable-target. For that purpose, four approaches of variable selection were evaluated: no variable selection, correlation-based feature selection (CFS), chisquare method (χ2) and Wrapper. To classify soil respiration, we used the decision tree induction technique available in the Weka software package. Our results showed that data mining techniques allow the development of a model for soil respiration classification with accuracy of 81 %, resulting in a knowledge base composed of 27 rules for prediction of soil respiration. In particular, the wrapper method for variable selection identified a subset of only five variables out of 18 available in the original dataset, and they had the following order of influence in determining soil respiration: soil temperature > precipitation > macroporosity > soil moisture > potential acidity.

RESUMO O cultivo de eucalipto tem se expandido consideravelmente no Brasil, sobretudo em regiões em que os solos apresentam baixa fertilidade, como nos Cerrados brasileiros. Para alcançar maiores produtividades, é fundamental saber a necessidade e o momento adequado para correção do solo. Para auxiliar esse processo de tomada de decisão, modelos matemáticos e computacionais têm sido utilizados e são uma alternativa promissora. O objetivo deste trabalho foi modelar a influência dos atributos físico-químicos do solo e climáticos no desenvolvimento do Eucalyptus urograndis em Neossolo Quartzarênico, por meio da técnica de indução da árvore de decisão. Para isso foram utilizados 30 atributos, sendo 29 preditivos e um atributo-meta ou variável resposta, a saber, a altura do eucalipto. Foram avaliadas quatro abordagens para seleção de atributos: sem seleção, seleção de atributos baseado em correlação (CFS), método do Qui-quadrado (χ2) e Wrapper. Para classificar os dados foi utilizada a técnica de indução de árvore de decisão por meio do software Weka 3.6. As técnicas de mineração de dados através da indução de árvore de decisão permitem o desenvolvimento de um modelo de classificação do desenvolvimento inicial de eucalipto eficiente para previsão de novos casos em diferentes classes de produção, onde o volume individual de madeira (VOL) e o diâmetro altura do peito (DAP) são atributos determinantes para previsão do crescimento do Eucalyptus urograndis, além de atributos químicos do solo como: Magnésio (Mg+2), Fósforo (P), Alumínio(Al+3), potássio (K+), acidez potencial (H+Al) e potencial hidrogeniônico (pH), atributo físico como a resistência do solo à penetração e relacionado ao clima como a temperatura mínima.

ABSTRACT Eucalyptus plantation has expanded considerably in Brazil, especially in regions where soils have low fertility, such as in Brazilian Cerrados. To achieve greater productivity, it is essential to know the needs of the soil and the right moment to correct it. Mathematical and computational models have been used as a promising alternative to help in this decision-making process. The aim of this study was to model the influence of climate and physico-chemical attributes in the development of Eucalyptus urograndis in Entisol quartzipsamment soil using the decision tree induction technique. To do so, we used 30 attributes, 29 of them are predictive and one is the target-attribute or response variable regarding the height of the eucalyptus. We defined four approaches to select these features: no selection, Correlation-based Feature Selection (CFS), Chi-square test (χ2) and Wrapper. To classify the data, we used the decision tree induction technique available in the Weka software 3.6. This data mining technique allowed us to create a classification model for the initial development of eucalyptus. From this model, one can predict new cases in different production classes, in which the individual wood volume (IWV) and the diameter at breast height (DBH) are crucial features to predict the growth of Eucalyptus urograndis, in addition to the presence of chemical soil components such as: magnesium (Mg+2), phosphorus (P), aluminum (Al+3), potassium (K+), potential acidity (H + Al), hydrogen potential (pH), and physical attributes such as soil resistance to penetration and related to climate, such as minimum temperature.

A consorciação com espécies hibernais vem sendo utilizada como alternativa para aumentar a produtividade de entressafra (outono/inverno) das pastagens irrigadas. Este trabalho foi conduzido no período de maio a outubro de 2009, no município de Icaraíma, Estado do Paraná, com o objetivo de avaliar a produtividade e a composição morfológica do capim Tifton 85 com sobressemeadura de forrageiras de inverno em sistema irrigado. O delineamento utilizado foi de blocos casualisados com quatro repetições e com medidas repetidas no tempo. As parcelas experimentais foram implantadas por meio de cinco tratamentos, sendo três na forma de sobressemeadura de aveia, utilizando os genótipos Iapar 61, IPR 126, e FMS2 em pastagem de capim Tifton 85 e, um na forma de sobressemeadura de aveia em consórcio com centeio (IPR 126 + IPR 89), também em capim Tifton 85 e, por fim, um tratamento testemunha com apenas capim Tifton 85 sem sobressemeadura. A sobressemeadura com aveia IPR 126 obteve a maior produtividade acumulada, igual 4.102 kg MS ha-1, com razão folha/colmo superior ao capim Tifton 85 exclusivo, iguais 1,77 e 1,08, respectivamente. No entanto, as sobressemeaduras de forrageiras de inverno não aumentaram a produtividade total da pastagem.

Intercropping with cool season species has been used as an alternative to increase productivity off-season (fall/winter) of irrigated pastures. This study was conducted from May to October 2009 in Icaraíma, Paraná State, to evaluate the productivity and morphological composition of the Tifton 85 grass overseeded with winter forage in irrigated system. The experiment was a randomized block design with four replicates and repeated measures over time. The experimental plots consisted of five treatments, three as oats overseeding using the genotypes IAPAR 61, IPR 126, and FMS2 on Tifton 85 pasture and one as overseeding of oats combined with rye (IPR 126 + IPR 89) also on Tifton 85 grass and, finally, a control with only Tifton 85 without overseeding. The overseeding of IPR 126 oats achieved the highest cumulative productivity, 4102 kg DM ha-1, with leaf/stem ratio higher than that of exclusive Tifton 85 exclusive, 1.77 and 1.08, respectively. However, overseeding of winter forages did not increase the overall productivity of the pasture.