Desenvolvimento de um sistema de apoio à decisão para definir zonas de manejo em cafeicultura de precisão
Data
2010-03-04
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Federal de Viçosa
Resumo
A aplicação de fertilizantes e corretivos às taxas variáveis, baseada nas propriedades físicas e químicas do solo, exige uma amostragem densa para se determinar a variabilidade espacial no campo. Uma das técnicas para reduzir o número de amostras é definir zonas de manejo. Alguns pesquisadores têm demonstrado a importância desempenhada pelas variáveis elétricas do solo para explicar a variabilidade de suas propriedades. Dessa forma, o presente trabalho teve como objetivo desenvolver e avaliar um sistema de apoio à decisão, para definir zonas de manejo com base na variabilidade espacial da condutividade elétrica aparente do solo em regiões de produção de cafés de montanha. O sistema de apoio à decisão foi estruturado com formulários específicos para ajustes dos semivariogramas, interpolação dos mapas por krigagem e delimitação das zonas de manejo, utilizando-se o algoritmo fuzzy k-means . Para encontrar o número ótimo de classes, determinou-se o valor de dois índices: o Índice de Performance Fuzzy (FPI) e Entropia da Partição Modificada (MPE). A condutividade elétrica aparente do solo foi determinada a 0,20 m (CE20) e 0,40 m (CE40) de profundidade, utilizando-se o medidor portátil ERM-02. Foram criadas seis zonas de manejo definidas a partir do mapa de altitude (ZMA), do mapa de CE20 (ZM20), do mapa de CE40 (ZM40), dos mapas de CE20 e altitude (ZM20A), dos mapas de CE40 e altitude (ZM40A) e dos mapas de altitude, CE20 e CE40 (ZM2040A). Para cada caso, a área foi classificada em duas, três, quatro e cinco classes. Para análise da concordância entre as zonas de manejo e as propriedades do solo, calculou-se o coeficiente Kappa. Os valores médios obtidos para CE20 e CE40 foram de 1,80 mS m -1 e 1,22 mS m -1 , respectivamente. A CE20 e CE40 apresentaram baixa correlação com as propriedades do solo. A correlação mais elevada foi obtida para o fósforo remanescente com valores de 0,427 e 0,465 para a CE20 e CE40, respectivamente. Os mapas de CE20 e CE40 apresentam forte semelhança entre si, com um coeficiente de correlação de 0,969. O mapa de potássio foi o que melhor se correlacionou com o mapa de CE20. Pelos valores obtidos de FPI e MPE, foi observado que a ZMA exige uma análise adicional para se determinar o número ótimo de classes. Para a ZM20 e ZM40, o número ótimo de classes foi de duas. Para a ZM20A, ZM40A e ZM2040A, o número ótimo de classes foi de três. O coeficiente Kappa para ZMA foi significativamente superior na classificação do zinco, saturação por bases, pH, matéria orgânica, acidez potencial e areia fina, em relação às demais zonas de manejo. Para as demais propriedades do solo, os valores mais elevados de coeficiente Kappa foram obtidos utilizando-se a ZM20A e ZM40A. Pela análise dos coeficientes Kappa, concluiu-se que ZM20A, ZM40A e ZM2040A apresentaram melhores resultados na classificação das propriedades do solo. Entretanto, a ZM2040A não apresentou melhores resultados que a ZM20A e ZM40A.
The variable rate application of fertilizers and lime based on chemical and physical properties of soil requires a dense sampling for determining the spatial variability in the field. One technique to reduce the number of samples is defining the management zones. Some researchers have demonstrated the importance played by the soil electrical variables to explain the variability of soil properties. Thus, the objective of this study is to develop and to evaluate a decision support system for defining management zones based on soil apparent electrical conductivity in the mountain coffee production fields. The decision support system was structured to perform semivariograms, to generate maps by kriging and to define management zones using fuzzy k-means algorithm. To find the optimal number of classes, two indices were calculated: the Fuzzy Performance Index (FPI) and Modified Partition Entropy (MPE). The electrical conductivity of soil was determined at 0.20 m (CE20) and 0.40 m (CE40) using a portable meter ERM-02 made by Landviser . The data were grouped into six management zones defined from the map of elevation (ZMA), the map of CE20 (ZM20), the map of CE40 (ZM40), the maps of CE20 and elevation (ZM20A), the maps of CE40 and elevation (ZM40A) and the maps of altitude, CE20 and CE40 (ZM2040A). For each case, the field was classified in two, three, four and five classes. To analyze the correlation between the management zones and soil properties, the Kappa coefficient was calculated. The mean values of CE20 and CE40 were 1.80 mS m -1 and 1.22 mS m -1 , respectively. The CE20 and CE40 showed low correlation with soil properties. The highest correlation was obtained for the remaining Phosphorus with values of 0.427 and 0.465 for CE20 e CE40, respectively. Maps of CE20 and CE40 presented high similarity to each other, with a correlation coefficient of 0.969. The map of Potassium content was more closely correlated with the map of CE20. For values obtained of FPI and MPE, it was observed that the ZMA requires additional analysis to determine the optimal number of classes. For ZM20 and ZM40, the optimal number of classes was two. For ZM20A and ZM40A, the optimal number of classes was three. The Kappa coefficient for ZMA was better when classifying zinc, base saturation, pH, organic matter, potential acidity and fine sand, in relation to other management zones. For the other soil properties, the highest values of Kappa coefficient were obtained using the ZM20A and ZM40A, and they were statistically equal. By analyzing the Kappa coefficient, it was concluded that ZM20A, ZM40A and ZM2040A were the best way to classify soil properties. However, ZM2040A was not better than ZM20A and ZM40A.
The variable rate application of fertilizers and lime based on chemical and physical properties of soil requires a dense sampling for determining the spatial variability in the field. One technique to reduce the number of samples is defining the management zones. Some researchers have demonstrated the importance played by the soil electrical variables to explain the variability of soil properties. Thus, the objective of this study is to develop and to evaluate a decision support system for defining management zones based on soil apparent electrical conductivity in the mountain coffee production fields. The decision support system was structured to perform semivariograms, to generate maps by kriging and to define management zones using fuzzy k-means algorithm. To find the optimal number of classes, two indices were calculated: the Fuzzy Performance Index (FPI) and Modified Partition Entropy (MPE). The electrical conductivity of soil was determined at 0.20 m (CE20) and 0.40 m (CE40) using a portable meter ERM-02 made by Landviser . The data were grouped into six management zones defined from the map of elevation (ZMA), the map of CE20 (ZM20), the map of CE40 (ZM40), the maps of CE20 and elevation (ZM20A), the maps of CE40 and elevation (ZM40A) and the maps of altitude, CE20 and CE40 (ZM2040A). For each case, the field was classified in two, three, four and five classes. To analyze the correlation between the management zones and soil properties, the Kappa coefficient was calculated. The mean values of CE20 and CE40 were 1.80 mS m -1 and 1.22 mS m -1 , respectively. The CE20 and CE40 showed low correlation with soil properties. The highest correlation was obtained for the remaining Phosphorus with values of 0.427 and 0.465 for CE20 e CE40, respectively. Maps of CE20 and CE40 presented high similarity to each other, with a correlation coefficient of 0.969. The map of Potassium content was more closely correlated with the map of CE20. For values obtained of FPI and MPE, it was observed that the ZMA requires additional analysis to determine the optimal number of classes. For ZM20 and ZM40, the optimal number of classes was two. For ZM20A and ZM40A, the optimal number of classes was three. The Kappa coefficient for ZMA was better when classifying zinc, base saturation, pH, organic matter, potential acidity and fine sand, in relation to other management zones. For the other soil properties, the highest values of Kappa coefficient were obtained using the ZM20A and ZM40A, and they were statistically equal. By analyzing the Kappa coefficient, it was concluded that ZM20A, ZM40A and ZM2040A were the best way to classify soil properties. However, ZM2040A was not better than ZM20A and ZM40A.
Descrição
Tese de Doutorado defendida na Universidade Federal de Viçosa.
Palavras-chave
Citação
VALENTE, D. S. M. Desenvolvimento de um sistema de apoio à decisão para definir zonas de manejo em cafeicultura de precisão. 2010. 103 f. Tese (Doutorado em Engenharia Agrícola) - Universidade Federal de Viçosa, Viçosa-MG. 2010.