Brivaldo Gomes de Almeida
◽
Bruno Campos Mantovanelli
◽
Thiago Rodrigo Schossler
◽
Fernando José Freire
◽
Edivan Rodrigues de Souza
◽
...
<p>Geostatistical and multivariate techniques have been widely used to identify and characterize the soil spatial variability, as well as to detect possible relationships between soil properties and management. Besides that, these techniques provide information regarding the spatial and temporal structural changes of soils to support better decision-making processes and management practices. Although the Zona da Mata region is a reference for sugarcane production in the northeast of Brazil, only a few studies have been carried out to clarify the effects of different management on soil physical attributes by using geostatistical and multivariate techniques. Thus, the objectives of this study were: (I) to characterize the spatial distribution of soils physical attributes under rainfed and irrigated sugarcane cultivations; (II) to identify the minimum sampling for the determination of soil physical attributes; (III) to detect the effects of the different management on soil physical attributes based on the principal component analysis (PCA). The study was carried out in the agricultural area of the Carpina Sugarcane Experimental Station of the Federal Rural University of Pernambuco, 7&#186;51&#8217;13&#8221;S, 35&#186;14&#8217;10&#8221;W, characterized by a Typic Hapludult with sandy clay loam soil texture. The investigated plot, cultivated with sugarcane, included a rainfed and an irrigated treatment in which a sprinkler system was installed according to a 12x12m grid. The interval between consecutive watering was fixed in two days, whereas irrigation depth was calculated to replace crop evapotranspiration (ETc) and accounting for the effective precipitation of the period. Daily ETc was estimated based on crop coefficient and reference evapotranspiration (ETo) indirectly obtained through a class A evaporation pan. In both treatments, the soil spatial variability was determined according to a 56x32m grid, on 32 soil samples collected in the 0.0-0.1m soil layer, spaced 7x8m, and georeferenced with a global position system. The soil was physically characterized according to the following attributes: bulk density (BD), soil penetration resistance (SPR), macroporosity (Macro), mesoporosity (Meso), microporosity (Micro), total porosity (TP), saturated hydraulic conductivity (Ksat), gravimetric soil water content (SWCg), geometric mean diameter (GMD) and mean weight diameter (MWD). The results of the descriptive statistics showed that among the studied attributes, Ksat, SPR, and Macro presented higher CV values, equal to 63 and 69%, 35 and 40%, and 32 and 44%, under rainfed and irrigated conditions, respectively. The minimum sampling, adequate to characterize the different soil attributes, resulted in general smaller in the rainfed area, characterized by higher homogeneity. Thus, the GMD, SWCg (both with 2 points ha<sup>-1</sup>), and SPR (with 6 points ha<sup>-1</sup>) were identified as the soil physical attributes requiring the lowest sample density; on the other hand, MWD and Ksat, with 14 and 15 points ha<sup>-1</sup>, respectively, required the highest number of samples. Pearson&#8217;s correlation analysis evidenced that soil BD was the most influential physical attribute in the studied areas, with a significant and inverse effect in most of the investigated attributes. The geostatistical approach associated with the multivariate PCA provided to understand the relationships between the spatial distribution patterns associated with irrigated and rainfed management and soil physical properties.</p>
RFJ Spatial variability in fertigated coffee yields and plant nutrients in soil saturation extracts