Relação dos índices de vegetação com a produtividade de milho em áreas irrigadas no Rio Grande do Sul

Abstract

The supply of food to the world’s population is a largely discussed topic, with the need to increase yield without increasing the field, prioritizing the sustainability of water resources, energy and the environment. Agriculture is highly technological and technified, seeks high productions and reduced costs, thus, the need for information after the crop is sown is of paramount importance for the planning of the crop. The present study objective to identify which vegetation index presents better correlation with corn yield, at different stages and under irrigated conditions in Rio Grande do Sul, using Sentinel-2 images. The study was carried out in five fields located in the cities of Catuípe and Cachoeira do Sul, which respectively, clay and sandy loam soils. The season evaluated were those of 2019/20 e 2020/21, using 25 MSI/Sentinel-2 images. The NDVI, SAVI, ARVI and EVI2 vegetation indices were calculated using Sentinel-2 spectral bands 2, 3 and 8. The image processing was performed in QGIS software. Principal components analysis was performed, the Pearson’s correlation and linear regression. The results showed that for the V6 to V10 stages, the ARVI and the NDVI presented a better determination for estimating corn yield, R² 0.73 and 0.71, respectively. In stage R1/R2 were SAVI with R² 0.67 and ARVI with 0.62 and in R4, NDVI with 0.50. For the R5 stage, the correlation with yield was very low, as the culture start senescence. The different types of soil did not show differences in vegetation index.