Assessment of spatial variability of soil fertility index in a Dystrofic Yellow Latosol

The lack of knowledge about the variability of the soil fertility index often leads to the use of inadequate amounts of fertilizers for certain areas within the crop. Aiming to analyze a set of variables from the soil chemical analysis, the multivariate analysis was applied in order to condense these variables into a smaller group of factors without important information loss to create a Soil Fertility Index. Soil sampling was carried out in a total area of approximately 740 ha, with a sampling grid of 5 ha, making a total of 148 sampling points. Four factors: (1) exchangeable aluminum (Al), aluminum saturation (m%), base saturation (V%), calcium, pH, Mg and P. Al and m% (2) the CTC and OM variables (3) Fe and Mn variables and (4) B, S and Zn were extracted from the factorial analysis explaining 74.79% of the total variance of the data, which is satisfactory by the criterion of the variance percentage. The analysis showed that in the sample dataset, an intersection was occurred in the mean dataset forming three centroids which coincides with the number of properties where the set of samples was collected. The use of multivariate analysis proved to be efficient for the proposed study, since the analysis of variance could not show efficiency due to the interrelations between the variables causing bias in the results. Based on the universe analyzes studied here, approximately 97% of the sampled area presented satisfactory or high soil fertility levels, which leads to the use of reduced amounts of fertilizers in most of the growing area.