Exploring hyperspectral reflectance indices for the estimation of water and nitrogen status of spinach

Optimizing irrigation and nitrogen (N) fertilizer management in vegetable production is key to meeting economic and ecological demands. Inadequate application increases the risk of quality and yield losses as well as environmental pollution. Accurate determination of both crops' water and N supply status is essential for an effective management strategy that couples N and water supply, thus accounting for the water-N interaction. In a field experiment with spinach as model crop, hyperspectral reflectance was measured and related to the relative water content (RWC), nitrogen and chlorophyll (Chl) content of leaves indicating plants' supply status. Three statistical methods were applied to estimate plants' water and N status by reflectance spectroscopy: i) Published Vegetation Indices (VI) for N and water status, ii) Competitive Adaptive Reweighted Sampling -Partial Least Squares Regression (CARS-PLSR) to identify key wavelengths and to build a simple robust model based on full spectra, and iii) inflection points (IP) based on the first and second derivatives. The best model fit for Chl was obtained for VI REIP and IP1. For N, best results were found for mrNDVI and IP1. The estimation of RWC was not significant. Higher Chl content was found in stressed plants due to smaller and thicker leaves. At full water supply, REIP and IP1 as well as CARS-PLSR differentiated N status levels. In water-deficient plants, VIs and IPs successfully detected water stress levels. The basis for a differentiating nitrogen and water management using spectral data has been established. (c) 2021 The Authors. Published by Elsevier Ltd on behalf of IAgrE. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

Automated summary

Optimizing irrigation and nitrogen management is crucial for vegetable production. This study used hyperspectral reflectance to determine the water and nitrogen supply status of crops. Statistical methods were applied to estimate the plant's status, and the basis for differentiating nitrogen and water management using spectral data has been established.