Estimation of irrigated oats yield using spectral indices

Abstract

Spectral indices such as normalized difference vegetation index (NDVI) and Falker chlorophyll index (FCI) are valuable tools for estimating crop yield. However, there remains a need to define the best phenological stages for the evaluation and identification of the indices' critical limits to achieve better agricultural management practices. We aimed to compare the level of accuracy in estimating grain and biomass yield of white oats (Avena sativa L.) as a function of NDVI and FCI at four phenological stages and to define the critical limits of the two indices. Five experimental treatments were performed in four replicates; each treatment received a different irrigation level: 11%, 31%, 60%, 87%, and 100% of crop evapotranspiration volume, respectively. The critical ranges of NDVI and FCI were set when the relative grain yield was between 90% and 110%.The mean NDVI and FCI values in each plot were determined at four phenological stages, which correlated to grain and biomass yield. Active sensors were used and the measured values were later used in regression analysis. The grain and biomass yield of white oats can be estimated with high accuracy. The NDVI presents greater accuracy of grain and biomass yield estimates of white oats than FCI. The critical ranges of the spectral indices for the “Average” class of oat grain yield as a function of the NDVI were 0.71-0.77, 0.66-0.73, and 0.63-0.71 for the phenological stages 8, 10, and 10.5.4, respectively, and of FCI were 58.4–60.2 and 58.2–61.3 for phenological stages 10 and 10.5.4, respectively.