Nutrients (Nitrogen, Potassium, and Phosphorus)
Among the nutrients, nitrogen (N) is an essential element in plant growth and productivity and thus crucial to numerous ecological processes. Several absorption features are listed in the literature along the SWIR region (Curran 1989). Spectral indices derived from handheld, airborne, and spaceborne spectrometers are used for assessing N content. The majority of them are based on indirect indicators, mostly chlorophyll content, which was proven to be physiologically linked to N content. Thus, the previously listed chlorophyll indices, for example, CARI, MCARI, TCARI, and TCARI/OSAVI, have also been frequently used for assessing nitrogen.
On the other hand, since the 1510 nm is directly related to nitrogen content, SWIR-based nitrogen indices were developed and implemented. The normalized difference nitrogen index (NDNI), proposed by Serrano et al. (2002), is a log 10 transformed reflectance nitrogen index based on the absorption feature of nitrogen at 1510 nm and a reference band at 1680 nm:
Herrmann et al. (2010), who explored the performances of different indices with respect to N content, found that the firm advantage of SWIR-based indices lies in their ability to predict, and in their sensitivity to, this constituent. The best index, named the normalized ratio index (NRI1510), utilized the 1510 and 660 nm bands:
The two other constituents, potassium (K) and phosphorus (P), are also important macronutrients required by plants after N. However, VIs have shown very limited success for assessing K and P. Pimstein et al. (2011) conducted an experiment in order to assess P and K in wheat plants using indices and a partial least squares-regression (PLS-R) (Section 10.4.1). The correlation coefficient (r) value found using a VI based on a two-wavelength index (1645 and 1715 nm) was 0.73, and the value found using a PLS-R analysis of K content was 0.88. K appears in plants as an ion K+; therefore, it is not expected to be spectrally detected by its bonds. Pimstein et al. (2011) and Thulin et al. (2014) have shown that wavelengths that are highly related to K content mainly occur in the NIR and the SWIR since the K status is closely related to leaf structure and water regime, respectively. Studies exploring the spectral assessment of K content in in vivo plants have achieved less satisfactory results than studies examining the spectral assessment of K stress. P appears in plants in adenosine di- and triphosphates (ADP and ATP, respectively) and, therefore, is expected to be spectrally detected as a result of bonds with other elements. Since P affects plant development and conditions (Homolova et al. 2013), it is assumed that the material itself is not identified. Mutanga and Kumar (2007) applied a neural network (Section 10.4.1) and concluded that integrating the RE and SWIR depth of absorption features is important for P content assessment in the African savanna. Homolova et al. (2013) presented studies analyzing spectral data but did not mention studies exploring P indices; they did mention, though, that there are no pronounced P absorption features. Therefore, the exploration of early nutrient stress identification by spectral means with a high spatial resolution is suggested, along with searching for alternative methods of nutrient concentration assessment.