Spatial interpolation techniques for site-specific irrigation management in a mango orchard

Abstract

Interpolation techniques can be a suitable approach to apply site-specific management for irrigation in fruit fields. However, which interpolation method will produce more accurate maps? In order to respond this question, the aim of this study was to test the performance of spatial interpolation techniques for mapping soil physical properties for site-specific irrigation management purpose in a mango orchard in the Brazilian semi-arid region. The experimental site was structured in a grid of 60 georeferenced points, which correspond to the number of mango trees spaced about 8 × 10-m in a fruit field in Petrolina county, Pernambuco. Soil texture, bulk density, volumetric water content at -0.01; -0.033; -0.06; -0.1; and -1.5 MPa, and soil water available were determined. Stochastic and deterministic interpolation methods were tested. Soil physical properties did not show spatial dependence preventing the interpolation by stochastic method (Kriging). However, it was possible to interpolate using deterministic methods such as inverse distance weight (IDW), local polynomial interpolation (LPI), and Radial basis functions (RBF). IDW and RBF showed the best results of map quality for physical properties, however, all interpolation method showed relative small errors and could be all used for delineating zones for site-specific irrigation management.