Anais do XXXV Congresso Brasileiro de Ciência do Solo

SPATIAL PREDICTION OF ZN, PH, P AND CLAY FROM STATISTICAL AND GEOSTATISTICAL METHODS.

MARIANA RUGER SACCO⁽¹⁾; MICHELE DUARTE DE MENEZES⁽¹⁾; ANA PAULA BRANCO CORGUINHA⁽¹⁾; GUILHERME AMARAL DE SOUZA⁽¹⁾; LUIZ ROBERTO GUIMARÃES GUILHERME⁽¹⁾;
1 - UNIVERSIDADE FEDERAL DE LAVRAS;

The spatial prediction of Zn, pH, P and clay at non-sampled areas in a continuous way can be source of valuable information for agriculture and land management. Thus, this work aimed to assess the spatial prediction of Zn, pH, P and clay from ordinary kriging (OK), inverse distance weighted (IDW) and multilevel B-spline. A validation data set was used to assess the best prediction method, and two indices were calculated from the observed and predicted values: mean prediction error and root mean square of prediction error. IDW was the best to describe most of the studied soil properties, followed by multilevel B-spline. It was noticed that interpolation of soil properties requires testing different methods once the accuracy can be highly variable depending on the nature of the method and data variability.

[baixar arquivo pdf]