Application of spectral reflectance to differentiate between leaf diseases and abiotic stresses in wheat
Andrea Ficke: NIBIO
<div>The transition from the uniform application of fertilizers and fungicides to site-specific management requires accurate measurement and differentiation of the stresses involved in predicted yield reduction. Our study was designed to show how automated non-contact sensor technology, based on spectral reflectance, can be implemented to differentiate between diseases, nitrogen and water stress. We designed field trials with three nitrogen levels, three irrigation conditions and two fungicide treatments, to obtain a range of different biotic and abiotic stresses important for yield. We then obtained subsamples of leaves at GS 60, with and without diseases (septoria nodorum blotch, yellow rust and powdery mildew) in combination with different water and nitrogen conditions. The ASD FieldSpec 3 spectrophotoradiometer was used to take images under controlled light conditions to determine the energy spectra emitted from each leaf surface. Utilizing multivariate statistics, we related the variance in reflectance to the determining factors involved. Disease severity was assessed at GS 70-75 in the field and yield determined. The Principal Component Analysis (PCA) model developed from the analyzed images described over 90% of the total variance in the five first components, with each of the different stresses being clearly separated. Analysis of the yield determining factors for all three years showed that year, nitrogen, water, disease and year X disease were all significant.</div>
View Presentation |
