Exploratory study of winter wheat reflectance during vegetative growth using three‐mode component analysis

Remote sensing is a potentially important source of data for site‐specific crop management, providing both spatial and temporal information. Remote‐sensed data have been used in nitrogen application strategies to achieve optimal yield potential or grain quality. Our objective was to investigate and describe the use of repeated multispectral measurements and three‐mode component analysis in order to interpret measurements on winter wheat grown at different plant densities and nitrogen application strategies. A multi‐band spectroradiometer with eight different wavelengths was used three times during vegetative growth. The Tucker3 model was used for the three‐mode component analysis (3MPCA). A comparative study of single reflectance wavelengths and selected calculated vegetation indices was performed. Results showed that the 3MPCA technique gave better separation of samples from different treatments than a single principal components analysis (PCA) using a single measurement day. Furthermore, vegetation indices seemed to perform better than single wavebands. The 3MPCA model is capable of extracting important structural information about measurement conditions and cropping history. The temporal repeated measurements provided additional information about average growth rate and possibly also irradiation conditions at the time of measurement.