Definição de zonas homogêneas de manejo em área sob cultivo de cana-de-açúcar com histórico de aplicações de vinhaça

Abstract

This study was performed to define the homogeneous zones of management from the sensors information in commercial fields cultivated with sugarcane having a historical of vinasse application. The device Veris 3150 was used, which comprises sensors that read the following wavelength ranges of the electromagnetic spectrum: visible range (VIS), near infrared (NIR) and mid infrared (MID) and also the soil electrical conductivity. The information were acquired within rows spaced of 12,5 meters apart. After collected, the data were filtered and the spectral responses separated in 3 spreadsheets: the first one containing the VIS range, the second one with the MIR and the third one having the MIR range; however the electrical conductivity was analyzed separately. Subsequently, the data were subjected to principal components analysis in exception the soil electrical conductivity, selecting the number of principal components necessary to represent 90% of the data variance. From each principal component an index was generated, representing the data group of the component, being the index selected to the spatialization. With the indexes calculated, the semivariograms were made and the necessary adjustments too. Also the cross validation was made. Identified the spatial dependence of the indexes calculated, for each range of the spectrum, the kriging was applied. The kriging was utilized to interpolate the attributes clay, Ca, Mg, K, P and SOM, being done subsequently the correlation between the soil attributes and the sensors information. This way, the maps of the studied fields were produced and divided in homogeneous zones, using the criterion of the global variance reduction. The soil attributes averages were tested among the homogeneous zones to identify which were the information most efficient and reliable for each attribute. Therefore, it is concluded that there is a possibility of areas separation in management zones using these sensors information, and also distinguish sub-areas within the production unit with different nutrient levels.