A qualidade do café dentro de um talhão pode variar temporal e espacialmente devido a variações dos atributos do solo, do "status" nutricional , das condições climáticas e da forma como as operações agrícolas são conduzidas. O conhecimento da variabilidade da qualidade pode auxiliar na tomada de decisão em sistemas de produção de café, uma vez que o preço do produto é influenciado pela sua qualidade. Este trabalho foi desenvolvido com o objetivo de analisar a variabilidade espacial e temporal da qualidade de bebida do café de montanha, colhido no estádio cereja, e gerar mapas temáticos dos níveis de qualidade obtidos selecionando as áreas com características interessantes à produção de cafés especiais. Pode-se concluir que, na safra 2006/2007, 95% dos talhões que obtiveram notas de qualidade de bebida superior a 80 pertencem às classes com valores acima de 20 graus brix. Os valores de grau brix e as notas de qualidade de bebida do início e fim da safra em 2006/2007 foram significativamente diferentes pelo teste t para dados pareados a 5%. Além da variabilidade espacial existe uma variabilidade temporal da qualidade de bebida, sendo que, 86% dos talhões com qualidade média maior ou igual a 80 estavam nas áreas de coeficiente de variação menor que 4%, indicando que os talhões com qualidade especial tiveram menor variabilidade temporal. Para o índice de dependência espacial (IDE), as variáveis de índices Dris, com exceção para zinco e cálcio com IDEs respectivos de 1,00 e 0,84 (fortes), todas as demais variáveis de "status" nutricional apresentaram dependência espacial moderada. A percentagem de frutos verdes apresentou dependência espacial forte, enquanto que a percentagem de frutos maduros teve dependência fraca com IDE de 0,22. As percentagens de frutos verdolengos e passas apresentaram dependência moderada, com IDEs respectivos de 0,66 e 0,70; para o teor de sólidos solúveis, brix dos frutos, houve uma moderada dependência espacial (0,33), enquanto que, para a qualidade de bebida do café ocorreu uma distribuição aleatória caracterizada pelo efeito pepita puro.
The quality of agricultural products within a plot can vary temporally and spatially due to variation of soil attributes, weather conditions and the way the agricultural operations are conducted. The knowledge about the variability of the quality in the field can help with the decision making process in coffee production systems, since the price of the product is primarily determined by its quality. In this context, precision agriculture based on the management of production factors in the field can be an important tool in the quest for excellence in coffee production. This work was developed to analyze the spatial and temporal variability of mountain coffee quality harvested at the cherry stage, and to generate thematic maps displaying the quality levels obtained by selecting the areas with optimal features for gourmet coffee production. The research was conducted at the Agricultural Engineering Department of the Federal University of Viçosa and at two coffee (Coffea arabica L.) producing properties, and was divided into three stages. The first stage was based on the study of the spatial and temporal variability of soluble solids content (Brix) in the mature fruits (cherries) and their relation with beverage quality. In the second stage, thematic maps were generated using the quality scores of the beverage obtained during five consecutive seasons at the same property (2003/2004 to 2007/2008). From these maps, the autocorrelation study, or quality spatial arrangement, was conducted using Moran's Spatial Autocorrelation Index. In the third stage, using geostatistical and cluster analysis, the spatial beverage quality, brix grades of the fruits, nutritional status (DRIS and SPAD indexes) such as productivity and maturation, were each identified and characterized in a coffee plantation under pivot. It could be concluded that in the 2006/2007 season, 95% of the plots which obtained beverage quality scores above 80 belonged to the classes with brix grades above 20. The brix grade values and the beverage quality scores from the season from the beginning and end of the 2006/2007 season were significantly different by the t-test for paired data at 5%. In addition to spatial variability, there is also temporal variability of the coffee's quality since 86% of the plots with average beverage quality greater or equal to 80 were in the areas with a coefficient of variation less than 4%, indicating that the plots with special quality had lower temporal variability. For the Spatial Dependence Index (SDI), the DRIS Index Variables, with the exception of zinc and calcium with SDI's of 1.00 and 0.84 (both strong), respectively, all other nutritional "status" variables showed moderate spatial dependence. The percentage of immature fruits presented a strong spatial dependence, while the percentage of mature fruits presented a weak spatial dependence with an SDI of 0.22. The percentage of under and over mature fruits presented moderate dependence, with SDI's of 0.66 and 0.70, respectively; for the soluble solids content and fruits' brix, there was a moderate spatial dependence (0.33), while for the beverage quality there was a random distribution characterized by the pure nugget effect.
