676 AJCS 11(06):676-682 (2017) ISSN:1835-2707 doi: 10.21475/ajcs.17.11.06.p382 Precision agriculture applied to soybean crop: Part II – Temporal stability of management zones Eduardo Leonel Bottega 1 *, Daniel Marçal de Queiroz 2 , Francisco de Assis de Carvalho Pinto 2 , Domingos Sárvio Magalhães Valente 2 , Cristiano Márcio Alves de Souza 3 1 Federal University of Santa Maria, Cachoeira do Sul city, Rio Grande do Sul State, Brazil 2 Federal University of Viçosa, Agricultural Engineering Department, Viçosa city, Minas Gerais State, Brazil 3 Federal University of Grande Dourados, Faculty of Agricultural Sciences, Dourados City, Mato Grosso do Sul State, Brazil *Corresponding author: bottega.elb@gmail.com Abstract Precision farming techniques have potential applications in the Brazilian Savannah owing to the spatial variability in crop yield. The objective of this study was to evaluate the temporal stability of management zones based on the measurement of apparent soil electrical conductivity (ECa). The study was performed in a field located in the Brazilian Savannah. During the 2011/2012 crop season, ECa was measured at 160 sampling points in a 47-ha area. ECa values were then grouped into classes using the fuzzy k- means algorithm and fertilizers were applied to the soybean crop (BMX Potencia RR) in the 2012/2013 and 2013/2014 crop seasons according to soil analysis for each class. ECa was also measured in the 2012/2013 and 2013/2014 crop seasons and the values were grouped into classes, and the maps of apparent soil electrical conductivity classes were compared using kappa coefficients. Apparent soil electrical conductivity was significant and positively correlated with each pair of crop seasons for all three seasons analyzed. The cluster analysis indicated the presence of two management zones in 2012. These management zones were adopted for fertility management in the 2012/2013 and 2013/2014 crop seasons. A comparison of the 2012 and 2013 management zone maps, based on the kappa coefficient, showed that they remained stable. The use of kappa coefficients was found to be a promising tool in the analysis of the temporal stability of management classes. Keywords: Site-specific management, No-tillage, Soil electrical conductivity, Yield, Oxisol. Introduction In Brazil, precision agriculture is mainly used for managing fertilizer and lime application regimens at variable rates. These fertilizer and lime recommendations are usually defined on the basis of the analysis of soil samples obtained by grid sampling. Therefore, a large number of samples are required for precise detection of the spatial variability in soil attributes, which is a limiting factor for the use of precision agriculture (Souza et al., 2014). The spatial and temporal variability of soil attributes can affect growth, grain development, crop quality, and final yield (Guastaferro et al., 2010). The management practices that aim to increase the profitability of agricultural production with minimal environmental impact have to be adapted to variability in field conditions, and this is the main objective of using precision agriculture techniques. However, the large number of samples required to accurately map variability also increases the cost of implementing management practices based on this variability. Dividing the field into sub-regions called management zones that have similar factors that limit yield can help guide fertilizer input. This is a potentially important tool for managing spatial variability in crops and soils (Dalchiavon et al., 2012; Rodrigues Junior et al., 2011; Moral et al., 2011), apart from being a significant part of precision farming techniques (Xiaohu et al., 2016). For generating management zones, it is first necessary to collect information closely related to crop yield, and the variables chosen to define these management zones should be quick and inexpensive to measure. For these reasons, apparent soil electrical conductivity (ECa) is an excellent potential variable for the delineation of management zones (Lück et al., 2009). Numerous studies have demonstrated a correlation between soil ECa and soil chemical and physical attributes (Morari et al., 2009; Moral et al., 2010; Rodríguez-Pérez et al., 2011). The measurement of ECa is both fast and inexpensive compared to traditional soil sampling and laboratory analysis. Given its convenience and reliability, ECa has been used as an indirect estimator of soil variability in agricultural fields (Corwin and Lesch, 2003). The establishment of management zones based on ECa is well documented, but the information on their temporal stability is required. In this context, the objective of this study was to evaluate the temporal stability of management zones based on the mapping of apparent soil electrical conductivity. Results and Discussion ECa variability The mean values for apparent soil electrical conductivity (ECa) ranged from 1.42 mS m –1 in 2013 to 12.24 mS m –1 in 2014, and from 0.82 mS m –1 to 19.31 mS m –1 for the years AUSTRALIAN JOURNAL OF CROP SCIENCE | SUBMITTED: 24-OCTOBER 2016 | REVISED: 14-FEBRUARY-2017 | ACCEPTED: 06-APRIL-2017