Interpolation methods for thematic maps of soybean yield and soil chemical attributes

Nelson Miguel Betzek, Eduardo Godoy de Souza, Claudio Leones Bazzi, Ricardo Sobjak, Vanderlei Artur Bier, Erivelto Mercante

Abstract


The application of precision agriculture considers the values of non-sampled places by the interpolation of sample data. The accuracy with which the maps of spatial distribution of yield and the soil attributes are produced in the interpolation process influences their application and utilization. This paper aimed to compare three interpolation methods (inverse of the distance, inverse of the square distance, and ordinary kriging) in the construction of thematic maps of soybean yield and soil chemical attributes. A set of data referred to 55 sampling units for the construction maps of soybean yield and of eight soil chemical attributes, by different interpolation methods. The comparison was made based on the error matrix, by calculating the Kappa and Tau indices, beyond the relative deviation coefficient (RDC). It was noticed that the inverse of the square distance was the interpolator that less influenced the data behavior, and the best interpolation method dependent of the variability of the studied attribute. The kriging and the inverse of the square distance were considered the methods that presented the best results in the interpolation of data.

Keywords


Geostatistics; Precision agriculture; Spatial variability.

Full Text:

PDF


DOI: http://dx.doi.org/10.5433/1679-0359.2017v38n2p1059

Semina: Ciênc. Agrár.
Londrina - PR
E-ISSN 1679-0359
DOI: 10.5433 / 1679-0359
E-mail:  semina.agrarias@uel.br
Este obra está licenciado com uma Licença  Creative Commons Atribuição-NãoComercial 4.0 Internacional