1 Introduction

Nitrogen (N) is an essential element with high demand by plants. It acts in the synthesis of amino acids, enzymes, chlorophyll, nucleic acids, and other structural components (Krapp 2015; Marschner 2012; Taiz et al. 2014). Due to its high importance, farmers tend to add significant N fertilizer amounts to ensure good plant growth and high yield. However, high inputs of N above the plant demand may cause environmental damage and economic losses.

The bell pepper (Capsicum annuum L. var. annuum) is one of the most important cultivated crops in Brazil and around the world. It requires massive amounts of fertilizers, especially N, which in the right amount results in optimal vegetative and reproductive development (Aminifard and Bayat 2018; Molla et al. 2019). Also, the adjustment of N application contributes to minimizing the risks of environmental impacts. Fortunately, most of the bell pepper production is irrigated using a drip-irrigated system, which optimizes water use and facilitates the application of fertilizers by fertigation. In this system, the growers usually divide the fertigation into sessions over the crop cycle to provide the nutrient as the plant needed.

The growth of bell pepper in plastic bags filled with substrate may be an alternative to reduce the negative environmental impacts in crop systems. These plastic bags are commonly referred to as slabs that have been widely used in Europe and have gained space in Brazil (Moura 2018; Pivoto 2015). The use of slabs has several advantages, including reducing labor costs, especially those related to soil preparation and weed control, optimization of water and fertilizer use, and as a potential resource for urban agriculture. The bell pepper cultivation in slab requires N fertilization management throughout the cycle in a non-destructive and straightforward way, which is indirectly obtained using digital chlorophyll meter instruments (e.g., SPAD, Dualex).

The SPAD is a chlorophyll sensor used to measure the greenness of the leaf. This index correlates positively with the chlorophyll content of the leaf and with the N status of the plant (Fontes 2016). SPAD readings are calculated based on the transmission of red light at 650 nm and the transmission of infrared light at 940 nm. The ratio of each transmitted and emitted wavelength is used as an indicator of the amount of chlorophyll in the leaf per unit of leaf area (Markwell et al. 1995). Dualex combines chlorophyll (CHL) readings with a simultaneous determination of flavonoid (FLV) concentration, based on the difference in light transmittance at two wavelengths, both in the near-infrared region. It provides the nitrogen balance index (NBI) that improves plant nutrition evaluation, as chlorophyll and flavonoid contents are linked to the crop N status (Cartelat et al. 2005; Cerovic et al. 2005; Fontes 2016).

Several studies have been carried out evaluating the efficiency of chlorophyll sensors, especially SPAD, in determining the N status of peppers (Godoy et al. 2003; Madeira et al. 2003; Madeira and Varennes 2005; Costa et al. 2018; De Souza et al. 2019). These measurements are usually made during some vegetative periods of the crop, and few studies have been considering the possible alterations of the index over the cycle. Similarly, the Dualex has been used to determine the N status in different crops such as maize (Tremblay et al. 2007), wheat (Cartelat et al. 2005), broccoli (Tremblay et al. 2009), and potato (Milagres et al. 2018). However, literature reports on the use of SPAD and Dualex in bell pepper grown in slab are not found so far.

Therefore, the objectives of this study were the following: (1) to evaluate the effects of N doses on the indices measured with SPAD and Dualex through the bell pepper crop cycle, (2) to evaluate the diagnostic effect of SPAD and Dualex on N nutrition, and (3) to establish the accuracy of the chlorophyll indices measured through the bell pepper cycle as yield estimators.

2 Material and Methods

2.1 Plant Material

Bell pepper, cultivar Arcade F1, was grown in a protected environment in the Federal University of Viçosa, from October 2018 to March 2019. The plants grew up in slabs (1.0 × 0.4 m) filled with 40 dm3 of fertilized substrate. The substrate consisted of a mixture of soil, sand, and manure, in the proportion of 2:1:1, respectively. Each slab was fertilized with 160 g of simple superphosphate (18% P2O5), 16.89 g magnesium sulfate (9% Mg), 1.0 g of borax (11% B), 200 mg zinc sulfate (22% Zn), 320 mg copper sulfate (24% Cu), 20 mg of ammonium molybdate (54% Mo), and 0.8 g of potassium chloride (60% of K2O), which corresponded to 5% of the total dose. Additional K was applied via drip fertigation over the crop cycle totaling 15.27 g slab−1 of potassium chloride. The slabs were placed in a suspended bench at 0.1 m wide, in double rows. Single rows were spaced 0.5 m apart, and double rows were 1.05 m apart. Each experimental unit consisted of one slab with 4 plants spaced 0.2 m apart.

2.2 Treatments

The treatments consisted of six N doses (0, 1.5, 3.0, 4.5, 6.0, and 7.5 g plant−1) divided into 10 bi-weekly applications via drip fertigation, according to Fontes et al. (2005). The amount of 5% of each dose was applied to the substrate mixture (before planting). The N source used for the first 30 days after transplanting (DAT) was ammonium sulfate (20% N) and for the remainder of the cycle was calcium nitrate (15% N and 19% Ca). The experiment was arranged in a randomized block design with four replications.

2.3 Nitrogen Index Assessment and Relative Sufficiency Indices

The SPAD index was determined using the chlorophyll sensor SPAD-502® (Konica Minolta Inc., Japan) and the indices CHL, NBI, and FLV were obtained using the Dualex® Scientific (Force-A, Orsay, France). Twelve readings were taken from the uppermost mature leaves of the 4 plants at 30, 60, 90, 120, and 150 DAT on each plot. Each reading treatment was calculated as the average of the 12 single readings. All the measurements were carried out between 8:00 and 10:00 am. The relative nitrogen sufficiency index (NSI) was calculated by dividing the average of measured values from each treatment (N rate) by the average of values from the reference slab (well-fertilized with 7.5 g plant−1). We considered 0.95 as the critical level for comparison of the results.

2.4 Nitrogen Content Determination and Nitrogen Diagnostic

After reading with chlorophyll sensors, the uppermost mature leaves were collected, placed in paper bags, and dried at 65 °C to constant weight. After drying, the samples were processed in a Wiley mill (20-mesh sieve). The N content was determined by the Kjeldahl method described by Bremmer (1965). The relationships between leaf N contents and the chlorophyll sensor readings were evaluated using correlation analysis. All readings from every 30-day intervals were analyzed by correlation analysis with the N content previously determined in the uppermost leaves. The significance of the correlation was used to verify the efficiency of the sensors to be used as N status indicators.

2.5 Yield Prognostic Index

The determination of the indices for the prognosis of bell pepper yield was based on the significance of the correlation between the reading obtained with the chlorophyll sensors at 30, 60, 90, 120, and 150 DAT and final commercial yield. The commercial yield was determined by weighing the fresh fruits harvested during the crop cycle when they turned bright green. We considered the rate of 5.51 g plant−1 as the optimal rate of N (Da Silva et al. 2020).

2.6 Statistical Analysis

Statistical analysis was carried out with the R 3.5.1 software (R Core Team 2019). The significance of differences between means of the variables evaluated at different DAT was assessed by the F test (P ≤ 0.05) in ANOVA followed by the Tukey’s honest significant difference (HSD) post hoc test (P ≤ 0.05). The relationship between optical sensor indices and the N dose was performed by regression techniques followed by the lack of fit test. Pearson correlation analysis (r) was used to assess either the relationship between optical sensor indices and N content in the leaf and optical sensor indices and the commercial yield.

3 Results

3.1 Nitrogen Indices

There was a significant difference regarding the N doses and DAT in relation to the evaluated indices. However, the interaction effect between N and DAT were significant for SPAD and CHL indices only (Table 1). Since the interaction was significant, we proceeded to study the N dose inside each DAT using regression analysis (Fig. 1a and Fig. 1c) and the DAT factor inside each dose of N applied (Fig. 1b and Fig. 1d) using mean comparison test. The results show that the SPAD and CHL indices were significantly modified by N during the evaluation period. The highest values for the SPAD index were obtained at 30 and 60 DAT. In the same period, increasing N dose did not alter the SPAD and CHL indices. As the crop cycle advanced, the low N supply drastically decreased SPAD and CHL values (Fig. 1).

Table 1 Summary of analysis of variance (ANOVA), showing the mean square for the indices SPAD, chlorophyll (CHL), flavonoids (FLV), and nitrogen balance index (NBI) in response to days after transplanting (DAT) and doses of nitrogen (N). When the interaction was found significant, the factors were pooled together. If not, they were evaluated separately. N, quantitative factor; DAT, qualitative factor
Full size table
Fig. 1
figure 1

Changes in SPAD and chlorophyll (CHL) indices in response to nitrogen (N) doses in fertigation evaluated at 30, 60, 90, 120, and 150 days after transplanting (DAT). a, c Regression models adjusted for each DAT, considering significant interaction between the main factors for SPAD and CHL, respectively. b, d Box-Whisker plots showing the median (horizontal line), the interquartile range (box), the whiskers (5–95% percentile), and the mean (“+” symbol) of DAT inside each N dose for the SPAD and CHL, respectively. Boxes that did not overlap are significantly different (P ≤ 0.05). Data represent the mean of 4 replications * and ns: significant and non-significant according to F test (P ≤ 0.05)

Full size image

At 90 DAT, the N supply increased the SPAD and CHL indices linearly. However, at 120 and 150 DAT, this trend was represented by a quadratic model, indicating a possible saturation effect (Fig. 1a and c). We observed a decrease of the indices at 90, 120, and 150 DAT when there is a low amount of N application, which may be related to the high demand of the plant during the reproductive stage (Fig. 1b and d).

The FLV and NBI indices were significantly affected by the N supply and DAT. However, no significant interaction was found between N and DAT, and consequently, these factors were evaluated separately. Nitrogen significantly decreased FLV and increased NBI values. Both indices were represented by a significant linear model (Table 1, Fig. 2). Since there was no interaction, the regression curves for the 5 different evaluation periods run parallelly, but they differed in their intercepts (data not shown). The FLV index was low at the beginning of the cycle (30 and 60 DAT), reached maximum at 90 DAT, and decreased at 120 and 150 DAT. The highest values were observed when there was not N application (control) (Fig. 2a). The NBI index shows an opposite pattern in relation to FLV. This index was higher at the beginning of the cycle (30 and 60 DAT), decreased to lower values at 90 DAT, and increased again at 120 and 150 DAT. The highest values of NBI were obtained with the higher N doses applied (Fig. 2b).

Fig. 2
figure 2

Changes in flavonoid (FLV) and nitrogen balance index (NBI) indices in response to nitrogen (N) doses in fertigation evaluated at 30, 60, 90, 120, and 150 days after transplanting (DAT). a, b The regression model adjusted for the mean of DAT (line), considering no significant interaction between the main factors for FLV and NBI, respectively, and marginal means of FLV (a) and NBI (b) at the DAT levels compared by the Tukey test. Means followed by different letters are significantly different at P ≤ 0.05. Data represent the mean of 4 replications * significant according to F test (P ≤ 0.05)

Full size image

3.2 Relative N Sufficiency Index

The relative NSI calculated with the readings from chlorophyll sensors showed an interesting relationship between N application and the ontogeny of the bell pepper cultivated in slab. For the SPAD values, the N application superior to 4.5 g plant−1 was efficient in keeping the NSI above the critical level during the entire cycle. However, it falls below the critical level within low N applications after 90 DAT. Interestingly, at 120 DAT, the treatment with 3 g plant−1 revert the NSI to the optimal point, but it falls apart by the end of the cycle (Fig. 3a). The NSI calculated from CHL values showed a similar pattern with the ones from SPAD. However, at the beginning of the cycle (30 DAT), the relative NSI was slightly reduced for all treatments (Fig. 3b). The evaluation of NSI from NBI readings did not show a regular trend during the crop cycle, although it confirms the sharp reduction of NSI with low N application (below 4.5 g plant−1) (Fig. 3c). The FLV readings, which increase with N deficiency, showed an opposite pattern in the relative NSI. With the growing cycle, the absence of N resulted in a rapid increase of NSI, which indicates N stress (Fig. 3d).

Fig. 3
figure 3

Relative nitrogen sufficiency index (NSI) calculated based on the ratio of the average chlorophyll readings of the treatments and the average chlorophyll readings of the reference slab (treatment with highest N rate) for SPAD (a), CHL (b), NBI (c), and FLV (d). CL, critical level considering ± 5% of the reference. DAT, days after transplanting

Full size image

3.3 Correlations between SPAD, CHL, NBI, FLV, and N Content in Leaves

A significant correlation was found between the indices evaluated with SPAD and Dualex and the N content in the bell pepper leaf at 30, 60, 90, 120, and 150 DAT, except for the CHL, NBI at 30 DAT, and SPAD and FLV at 30 and 60 DAT (Fig. 4). The CHL and NBI indices positively and significantly correlated with the N content in pepper leaves at 60 DAT (Fig. 4b and Fig. 4c). A similar pattern was observed between SPAD and CHL (Fig. 4a and Fig. 4b), although the range of CHL values was lower than the SPAD. Although the correlation between FLV and N content was not significant at 30 DAT, it was possible to observe that the lowest values of FLV were obtained at this period (Fig. 4d). Regardless of the evaluation period, the leaf N content was represented by the polynomial regression model in the N supply function. However, the N content was significantly higher at 30 and 60 DAT and was lowest at 90 DAT due to the dilution effect (Fig. 5).

Fig. 4
figure 4

Relationship between leaf N content and a SPAD, b chlorophyll (CHL), c nitrogen balance index (NBI), and d flavonoids (FLV) at 30, 60, 90, 120, and 150 days after transplanting (DAT)

Full size image
Fig. 5
figure 5

Changes in the leaf N content in response of nitrogen (N) doses evaluated at 30, 60, 90, 120, and 150 days after transplanting (DAT). The quadratic model represents the marginal means of leaf N content at different N doses, and the letters compare the marginal means of each DAT (Tukey test). Means followed by different letters are significantly different at P ≤ 0.05

Full size image

3.4 Yield Prognostic Index

The CHL and SPAD indices showed the highest correlation coefficients with yield throughout the cycle, and the results confirm their efficiency in predicting the yield of bell pepper grown in slab (Fig. 6a and Fig. 6b). However, there was no significant correlation between the indices evaluated and the commercial fruit yield at 30 DAT. The NBI index also showed high correlation coefficients with the final yield from 60 DAT (Fig. 6c). However, the FLV index did not significantly correlate with the final yield (Fig. 6d). Even though the SPAD, CHL, and NBI were significantly correlated with the final yield in several evaluation dates, these indices were not very efficient in discriminating the variation of the yield up to 60 DAT (Fig. 6). The results showed little variation in the indices in comparison to the increase in the yield. However, after 90 DAT, the increase in the indices was correlated with the rise in the commercial yield.

Fig. 6
figure 6

Relationship between a SPAD, b chlorophyll (CHL), c nitrogen balance index (NBI), and d flavonoids (FLV) indices evaluated at 30, 60, 90, 120, and 150 days after transplanting (DAT) and the final yield of bell pepper grown in slab. ns, non-significant; *significant at P < 0.05; **significant at P < 0.01; ***significant at P < 0.001

Full size image

4 Discussion

Nitrogen significantly affected the SPAD index through the bell pepper cycle. With the growth of the bell pepper crop, SPAD tended to reduce under conditions of sub-optimal application of N. However, this reduction was low until 60 DAT even in the low doses (Fig. 1a), which may be explained by the mineralization of N present in the manure and the soil that composed the substrate. The mineralization process transforms organic N to inorganic forms, which are more available to plants. Godoy et al. (2003) reported that growing bell pepper in a vase containing 50 dm3 of substrate (manure and soil) was sufficient to meet the N requirements of the crop up to 63 DAT. From 60 DAT onwards, a marked decrease was observed in the SPAD index for the lowest N doses. This reduction was due to the greater pepper demand for N during the reproductive phase (Grainfenberg et al. 1985).

There may be a stabilization of the SPAD values at a higher N dose since the instrument does not detect the so-called luxury consumption (Blackmer and Schepers 1994). Therefore, even with increases in the N dose, plants tend not to invest in chlorophyll formation. Plants reach the point of photosynthetic maturity (Costa et al. 2001) and start to accumulate N as nitrate (Larcher 2000), which does not translate into a higher degree of greenness of the leaf.

The CHL index had a similar pattern with the SPAD, but the CHL values were lower than SPAD values. The CHL index evaluation showed that this index tends to drop dramatically without N application (Fig. 1b). SPAD and CHL indices are similar because each instrument indirectly estimates the chlorophyll content in the leaf (Padilla et al. 2018; Parry et al. 2014; Perez-Patricio et al. 2018), and there is a positive relationship between SPAD and the CHL index (Padilla et al. 2018).

The FLV index shows an opposite pattern if compared with the CHL and SPAD indices, and namely, N supply tends to reduce FLV values (Fig. 2a). However, with the advance of the cycle, this index tends to increase, especially when there is a deficiency of N. In the reproductive phase, bell pepper starts to demand high amounts of N (Miller et al. 1979), which justifies the increased stress and consequently the FLV values at low N doses. The increase in the concentration of flavonoids in leaves under resource-limiting conditions, such as N, can be explained by the reduction in vegetative growth being more significant than the restriction of photosynthesis. This reduction leads to a greater accumulation of carbohydrates, which in turn will be allocated to the polyphenol formation route (Gershenzon 1984; Herms and Mattson 1992).

The NBI index, which is the ratio of FLV and CHL, was efficient in showing the trend that occurs with the N balance throughout the cycle. At the beginning of the cycle, the ratio is high and positive and declines when there is no N supply to the crop (Fig. 2b). This drop is probably due to the reduction in CHL levels and the increase in the concentration of flavonoids, which is a response of the plants to N stress. The NBI is one of the advantages of Dualex over other chlorophyll sensors (Cartelat et al. 2005). Using NBI, the user can automatically check the current N status taking into account the relationship between the CHL and FLV indices.

Relative NSI is a common procedure to evaluate the readings obtained with chlorophyll sensors (Padilla et al. 2020). It can be applied in the decision-making of applying N to the crop. This study shows that this index could be satisfactorily used to help interpret the readings from SPAD and CHL. The results confirm that the plants did not have N deficiency until 60 DAT. The low application of N drastically reduced the indices at 90 DAT (below the critical value), which is associated with the high demand of the bell pepper during the reproductive phase. Although the relative NSI calculated from NBI and FLV readings show a similar pattern during the growing cycle, care should be taken to interpret them. The NBI is already a ratio of CHL and FLV. Therefore, both factors should be considered to analyze the NSI values. The FLV can also be confused if the user does not consider the inverse values obtained with this index regarding the N deficiency.

The utilization of reference plots to calculate the NSI reduces the influence of different factors capable of changing the greenness of the leaf and impacting the optical measurements (Padilla et al. 2020). According to them, abiotic stress, diseases, and different cultivars would affect at the same level both reference and measured area. Therefore, it allows the adjustments of the fertigation and helps to identify possible factors of stress, other than N deficiency. One potential limitation is the necessity of using additional fertigation line for the reference plot. However, with some adaptations in the fertigation system and considering the uniformity of the substrate, the growers could use a few slabs as the reference for many others. Godoy et al. (2003) also proposed the use of NSI as an indicator of the application timing of N fertilizer in bell pepper to reduce the total N applied to the crop. This technique has also been used in other vegetable crops such as muskmelon (Gianquinto et al. 2010), cabbage, carrots, and onions (Westerveld et al. 2004).

The highest correlation coefficients between the indices and N content in leaves were obtained at 90 DAT, indicating that it is the time that could be used to determine the N status and, if necessary, proceed to further N application. There was still a reduction in the N content in the leaf at that time, indicating resource allocation to fruit production. The absence of a significant correlation between these indices and the N content in the leaf up to 60 DAT can be explained by the relatively high N content in leaves of all treatments. At this time, the N concentration was above 45 g kg−1 (Fig. 4), which is within the sufficiency range for the culture (Maynard and Hochmuth 2006).

The occurrence of a plateau at 30 and 60 DAT, especially for SPAD and CHL (Fig. 1), may be related to the saturation effect commonly reported in vegetable crops at relatively high N contents (Goffart et al. 2008). However, the CHL and NBI indices correlated positively and significantly with the N content in pepper leaves at 60 DAT (Fig. 4). These indices presented high precision at high N values and can be an alternative for N status diagnosis in saturated conditions. The indices obtained with the chlorophyll sensors have been widely used to predict the yield of many crops such as tomatoes (Padilla et al. 2015), potatoes (Milagres et al. 2018), wheat (Islam et al. 2014), rice (Liu et al. 2017), cassava (Haripriya Anand and Byju 2008), and corn (Gabriel et al. 2019). In the present study, there was a significant correlation between the indices obtained with the chlorophyll sensors and the commercial fruit yield. The high correlation coefficients observed along the cycle with the final commercial yield indicates that SPAD and CHL indices could be used to predict the final yield at different times of evaluation. The results also showed that the singly FLV assessment does not correlate with the final yield. However, FLV provides valuable information when used to calculate the NBI index. As mentioned above, the NBI is an advantage of the Dualex chlorophyll sensor because it is a ratio of CHL to FLV and indicates the N status of the crop. The correlation between the indices and commercial yield did not expressively differ over the cycle. Therefore, it suggests that the bell pepper yield is directly associated with the N fertilization and can be monitored using both SPAD and Dualex chlorophyll sensors. The highest coefficients were obtained at the end of the cycle, as reported by Gabriel et al. (2019) in corn. According to them, these results could support the recommendations of applying N to correct deficiencies at late growth stages in cropping systems such as the bell pepper crop, which allows a late N application.

Monitoring plant N status is essential to optimize nitrogen fertilization. In this study, the results showed that the evaluated indices change over the cycle, as observed by De Souza et al. (2019) using SPAD. According to Varinderpal-Singh et al. (2020), it was possible to save about 30 kg N ha−1 and reduce N2O emission by 29.7%, using optical sensing-based methods, e.g., SPAD, compared with soil test-based N application. These results confirm the importance of using sensors such as SPAD and Dualex in monitoring throughout the bell pepper cycle to reduce costs and possible environmental contamination.

5 Conclusions

The SPAD and Dualex chlorophyll sensors were able to evaluate the nitrogen status of the bell pepper grown in slab throughout the cycle. The N rates necessary to increase one chlorophyll reading unity depended on the time of evaluation. Thus, at 30 and 60 DAT, it would not be possible to change the SPAD and CHL reading with the increase in N rate. At 90, 120, and 150 DAT, would be necessary for SPAD, 0.546, 0.242, and 0.199 g of N plant−1, and for CHL, 0.549, 0.261, and 0.224 g of N plant−1, respectively. Both sensors could predict the nitrogen status of the leaves and the commercial yield improving the correlation as the crop development progressed. However, the utilization of these indices as yield predictors was not satisfactory up to 60 days after transplanting. The utilization of Dualex in the bell pepper crop instantly calculates the ratio of chlorophyll to flavonoids (nitrogen balance index) that proved to be highly correlated with the commercial yield. Both sensors can be effectively used to monitor the nitrogen sufficiency index in order to improve the calendar of fertigation by applying the nitrogen as a plant needed. The utilization of reference slabs can help growers adjust the nitrogen application according to their cultivar, substrate, and management.