Estimation of maize properties and differentiating moisture and nitrogen deficiency stress via ground – Based remotely sensed data

•Accurate estimations of maize yield and properties via remotely sensed data.•Spectral distinguishing between moisture and nitrogen induced stress.•Models of PLDA and PCA for distinguishing sources of stress.•Further research is required to apply this technique on other crops and environments.•Statistical approaches such as PLSR models may enhance predictions of crop traits. Moisture and nitrogen deficiency are major determinant factors for cereal production in arid and semi arid environments. The ability to detect stress in crops at an early stage is crucially important if significant reductions in yield are to be averted. In this context, remotely sensed data has the possibility of providing a rapid and accurate tool for site specific management in cereal crop production. This research examined the potential of hyperspectral and broad band remote sensing for predicting maize properties under nitrogen and moisture induced stress. Spectra were collected from drip irrigated maize subjected to various rates of irrigation regimes and nitrogen fertilization. 60 spectral vegetation indices were derived and examined to predict maize yield and other properties. Highly significant correlations between maize crop properties and various vegetation indices were noticed. RVI and NDVI were found to be sensitive to maize grain yield in both tested seasons. Cred edge demonstrated the strongest significant correlations with maize yield. The correlations with grain yield were found to be strongest at the flowering stage. Penalized linear discriminant analysis (PLDA) showed the possibility to distinguish moisture and nitrogen deficiency stress spectrally. The implications of this work for the use of satellite based remote sensing in arid zone precision agriculture are discussed.