Spectral response of citrus and their application to nutrient and water constraints diagnosis

Abstract

The early diagnosis of nutrient and water stress is a key issue for the proper management of orchards. Unfortunately, the measurement of indicators in the field to assess those deficiencies is expensive and time consuming. Remote sensing provides the possibility of assessing water and nutrient stress over every single stand on the whole area of interest. Leaf and canopy spectra are affected by every vegetation process or change in leaf pigments. However, the effect of each nutrient deficiency on the canopy spectrum is very subtle and the wavelength ranges affected by the different nutrients are generally overlapping. Then, an analysis in detail is needed in order to properly discriminate between the different stress causes. This chapter presents a review of previous studies dealing with remote sensing of nutrient and water stress, defining the spectral regions affected and the methodologies that have been applied for vegetation stress assessment. Specific multispectral indices related to leaf pigment content and physiological vegetation processes, thermal imagery, and fluorescence studies are presented. Finally, some results obtained from radiative transfer modeling simulations highlight the importance of the spatial resolution for precise agriculture purposes. There is extensive future research needed focusing on dissembling the overlapping effects of different nutrient deficiencies on the electromagnetic spectrum. Generally, vegetation indices used for nutrient assessment are affected by more than one nutrient content; a specific diagnosis is difficult to solve. As an example, pertinent indices can be developed in order to perform a proper stress diagnosis. Additionally, leaf-level radiative transfer models can be improved to account for differences in specific pigments. If leaf reflectance can be modeled for different nutrient levels, tree nutrient status can be assess through modeling inversion. Modeling inversion techniques are especially interesting because they do not rely on site-specific empirical relationships, being applicable to extensive areas.