Vegetation indices as a Tool for Mapping Sugarcane Management Zones

Abstract

In precision agriculture, the adoption of management zones (MZs) is one of the most effective strategies for increasing agricultural efficiency. Currently, MZs in sugarcane production areas are classified based on conventional soil sampling, which demands a lot of time, labor and financial resources. Remote sensing (RS) combined with vegetation indices (VIs) is a promising alternative to support the traditional classification method, especially because it does not require physical access to the areas of interest, is cost-effective and less labor-intensive, and allows fast and easy coverage of large areas. The objective of this study was to evaluate the ability of the normalized difference vegetation index (NDVI) and the two-band enhanced vegetation index (EVI2) to classify sugarcane MZs, compared with the conventional method, in the Brazilian Cerrado biome (savannah), where about half of Brazil´s sugarcane production takes place. This study used historical crop production data from 5,500 production fields in three agricultural years (2015 to 2018) and NDVI and EVI2 values of 14 images acquired by the Landsat 8 satellite from 2015 to 2018 in Google Earth Engine (GEE). Although improvements are still necessary and encouraged, a new methodology of classifying MZs according to VIs was proposed in this study. The NDVI was not correlated with MZs classified using the conventional method, whereas EVI2 was more sensitive to biomass variations between MZs and, therefore, could better discriminate between MZs. The EVI2 values measured in crops aged 180 to 240 days in the rainy season proved to be the best strategy for classifying MZs by RS, where MZ A, for example, had EVI2 of 0.37, compared to MZ E, which had an EVI2 of 0.32.