Nitrogen, an essential element in chlorophyll and in enzymes needed for photosynthesis, plays an important role in maintaining crop growth and enhancing grain yield [1
]. However, over-fertilization will result in lower nitrogen-use efficiency and environmental pollution. Leaf nitrogen concentration (LNC) can be used to diagnose the nutritional status and guide precise fertilization [2
]. Therefore, many studies have been devoted to accurately monitoring crop LNC [5
]. The close correlation between chlorophyll (Chl) and nitrogen content makes it possible to quantify the crop LNC with empirical methods [8
]. Existing studies have illustrated that reflectance-based parameters/vegetation indices could be used to monitor the LNC [9
], which are based on the absorption characteristics of chemical components. However, until now, these studies have some shortcomings, such as mixed signals from the plants and the soil, lack of specificity of the nitrogen stress, and limitations to the specific ranges of biomass, leaf area, and chlorophyll [12
]. In the past decades, chlorophyll fluorescence (ChlF), the light emitted by chlorophyll has proven to be highly related to crop physiology and sensitive to plant nitrogen status when compared with reflectance signals [13
ChlF, used to probe many aspects of photosynthesis of plants and other photosynthetic organisms, is emitted by Chl, independent of soil interference and biomass. Govindjee [14
] suggested that ChIF could provide abundant information about photosynthetic characteristics, including pigment system composition, de-excitation energy, rates of electron transfer reactions on Photosystem II (PS II), actual photochemical quantum yields and coefficients of photochemical quenching. Since nitrogen is the main element of Chl and enzymes, plant nitrogen content affects the Photosystem I (PS I) and Photosystem II (PS II) functions, and then influences the photosynthetic characteristics by affecting the photosynthetic pigment content and physical changes in pigment-protein complexes. Therefore, plant nitrogen content could affect the photosynthetic function leading to the changes of ChlF emission. Meanwhile, the strong link between LNC and ChlF provides an empirical basis for detecting plant LNC.
ChlF can be measured with active and passive ChlF techniques depending on the type of excitation light source. The active ChlF measurements have been proposed as possible species-specific approach to monitor the LNC and indentify nutrient deficiency of crops by several scientists [15
]. For example, laser-induced fluorescence parameters (F685, F740: fluorescence intensity at 685 nm and 740 nm; F740/F685: ratio of fluorescence intensity at 740 nm and 685 nm) are reported as a potential method for non-destructively monitoring paddy rice LNC [21
]. It seems that much progress has been made to detect LNC using ChlF [22
]. To date, there still exist some limitations to the application of active ChlF. First, due to the artificial light source used to excite the leaf fluorescence emission, the active ChlF is mainly used for individual leaves and small plants. Moreover, it is unrealistic to be applied at the large scale with the limitation of laser pulse energy and background interference [25
]. Additionally, it is difficult to extrapolate the result because the shape and intensity of the active ChlF varies with the excitation light source [14
Sun-induced fluorescence (SIF), also known as passive chlorophyll fluorescence, has been widely used in recent years as a promising approach to probing plant physiology, net photosynthesis, stress status at different scales, i.e., leaf, canopy, region and global [27
]. SIF, a bimodal spectrum ranging from 650 nm to 850 nm, is emitted directly by Chls under the excitation of sun-light. It is composed of two peaks, with the first peak (685 nm) located in the red region, which is mainly attributable to Photosystem II (PS II), and the second peak (740 nm) located in the far-red region, which is attributed to Photosystem I (PS I) and PS II [29
]. Compared with the values of the reflected and transmitted radiation, though the signal of leaf SIF is relatively small (just about 2–5% in the near-infrared), it plays an important role in characterizing the photosynthetic process [29
]. SIF has been employed as an effective means not only for detecting plant photosynthetic capacity [32
], light-use efficiency [33
], stress, and injury [35
], but also for other physiological parameters related to nitrogen fertility conditions. The ChlF peak ratio is known to be an accurate estimator of leaf Chl content [23
], which is an indirect association between ChlF and LNC, mediated by chlorophyll. Tubuxin et al. [37
] has estimated the Chl content using SIF at various growth stages of paprika (Capsicum annuum cv
. ‘Sven’) plants. Moreover, Du et al. [38
] reported a high relationship between SIF at the canopy level and photosynthetically active radiation absorbed by chlorophyll, although it is affected by species-specific, bio-chemical components and canopy structure, particularly at the O2
-B band. The model-based analysis has shown that the slope of gross primary production and SIF tends to be smaller with increasing Chlorophyll a + b content (Cab). The slope is only sensitive when Cab is <20 μg∙cm−2
and is stable when Cab is >20 μg∙cm−2
]. The studies mentioned above provide the experimental basis to probe the LNC utilizing SIF, which is closely linked with the management of nitrogen fertility. However, few studies so far have explored the feasibility and potential of SIF to detect the LNC in agronomic crops.
Concerning a typical bifacial leaf, SIF is emitted from both sides [13
]. Although the upward and downward SIF are generated by the same incident light, there are differences between them due to the internal pigment distribution and structural factors [13
]. Descriptions of the upward and downward SIF characteristics are helpful to interpret the remote sensing signal. Understanding and comparing the contribution of the upward and downward SIF in the total SIF helps in recognizing the change of SIF in the propagation process in a remote sensing manner. Moreover, it should also be noted that two SIF emission peaks are affected by strong internal absorption, which could affect the percentage of the upward and downward SIF and change the ratio between red and far-red peaks for both sides. So far, no studies have investigated the capacity and difference of upward and downward SIF in the LNC detection. Few researchers have compared the ability of the two SIF peaks to estimate LNC, which would be beneficial to understanding the mechanism of monitoring the LNC based on SIF.
The overall goal of this study is to estimate LNC in wheat, a major food crop, using SIF related parameters. To fulfill this goal, four main objectives are pursued: (1) to understand the variation of upward and downward SIF spectra under different LNC levels; (2) to compare the differences of the correlations between the upward and downward SIF spectra and LNC; (3) to construct an empirical model for estimation of LNC based on the upward and downward SIF-related parameters; (4) to evaluate the performance of the LNC models under various Chl content and leaf mass per area (LMA) levels in wheat.