Modulo Computacional para cálculos topográficos e microserviços para seleção de variáveis aplicados ao delineamento de Zonas de Manejo

Abstract

Management Zones (MZs) are subareas in a field with similar characteristics that can be treated as homogeneous areas from the point of view of sampling and management and allow the application of Precision Agriculture (PA) on rural properties. The delineation of MZs is usually done by clustering algorithms and can be performed using computational tools and other technologies. The selection of variables for the delineation of MZs, one of the steps in the delineation of MZs, is a complex task, and some methods can be applied for this purpose: Spatial correlation analysis, Principal Component Analysis (PCA) and Multivariate spatial analysis based on the Moran index and in PCA (MPCA). This work presents the development of resources within the AgDataBox (ADB) platform for the ADB-Map application, which is a module for calculating topographic variables (ADB-Map-MCT) such as slope, aspect, topographic position index, and curvature; implementation of functionality for building histograms, development of a user manual to improve usability, and the implementation in the ADB-Map application of two variable selection methods proposed in the literature, which are PCA-SC (PCA + Spatial Correlation) and MPCA-SC (MULTISPATI-PCA + Spatial Correlation). The computational module was tested for calculating topographic variables in three areas, and its results were compared to those of the QGIS tool using statistical metrics. In most results, the correlation was moderate to high, indicating that the tool was valid and useful. ZMs were outlined in three experimental areas to validate the selection methods. The results showed that the methods implemented for ADB-Map, PCA-SC, and MPCA-SC could delineate high-quality ZMs, as confirmed by indexes such as MGQI. Regarding the impact of the abovementioned topographic variables on productivity, it was observed in the results that these were selected in most of the clusters created, demonstrating their influence on soybean and corn crops.