Friederike Gnädinger and Urs Schmidhalter
Chair of Plant Nutrition, Technical University of Munich (TUM), Emil-Raman-Str.2, 85354 Freising
Abstract
Yields are primarily dependent on the input of nitrogen. Since fertilization poses serious environmental concerns, better performing maize cultivars with reduced nitrogen inputs are required. The aim of this work is to develop efficient phenotyping procedures to assess differences in nitrogen uptake and nitrogen use efficiency at the reproductive stage of maize cultivars supplied with different nitrogen doses. Both, active and passive sensor systems were tested to assess such traits and the spectral information was related to destructively assessed parameters of the aboveground biomass and nitrogen uptake at flowering, kernel dough stage and at grain harvest. In general, spectral indices obtained from the hyperspectral sensor were better related to the biomass or the nitrogen uptake of maize cultivars than those obtained from active sensors. Optimized hyperspectral indices were closest related to the leaf dry biomass with R2-values of 0.55 and to the Nitrogen Nutrition Index with R2 = 0.77 at the kernel dough stage. PLSR models were calculated for two experimental years, using data from 2015 as calibration data set and from 2014 as validation data set and allowed to predict leaf nitrogen uptake with R2 = 0.70 and RMSE of 10.8 kg N ha-1. The phenotyping platform PhenoTrac 4 developed by TUM allowed to obtain enhanced information of the complex traits nitrogen uptake and nitrogen use efficiency. In combination with improved algorithms using both optimized indices as well as PLSR models promising results were obtained to further narrow the gap existing between genomics and phenomics.