- Open Access
Eliminating interference by anthocyanin in chlorophyll estimation of sweet potato (Ipomoea batatas L.) leaves
© Huang et al.; licensee Springer. 2014
Received: 24 October 2013
Accepted: 11 November 2013
Published: 30 January 2014
Spectral reflectance was evaluated for its usefulness as a nondestructive estimation of chlorophyll (Chl) content from three cultivars of sweet potato (Ipomoea batatas L.) with green, yellow, and purple leaves grown in a greenhouse for 22 days. While the green and yellow leaves contain variant amount of photosynthetic pigments without or with little level of anthocyanins, the purple leaves, except large amount of photosynthetic pigments, have high quantity of anthocyanins.
For green and yellow leaves, the reciprocal reflectance (R-1) and derived indices incorporating near infrared (NIR) reflectance, [(Rλ)-1 - (RNIR)-1] and [(RNIR/Rλ) - 1], in the green and red edge spectral ranges were shown to be strongly correlated (r2 = 0.8 ~ 0.9) with the chlorophyll content. The root mean square error (RMSE) of the chlorophyll content estimation using these indices was < 50 mg m-2. However, when purple leaves containing high levels of anthocyanins were included in the sample, R-1 in the green spectral range and the above-mentioned indices displayed much weaker correlations with the chlorophyll content. The RMSE of chlorophyll estimation using these indices in the green spectral range sharply increased to > 110 mg m-2 when the sample included purple leaves. The new index, [1 - (Rλ/RNIR)], was therefore inferred and developed to eliminate the distorting effect of anthocyanins on chlorophyll content estimation using reflectance in the green spectral range. For leaves with high levels of anthocyanins, the correlation between [1 - (Rλ/RNIR)] and the chlorophyll content remained strong (r2 = 0.8 ~ 0.9) in the green spectral range, and the RMSE was minimal.
The reflectance index, [1 - (Rλ/RNIR)], therefore represents a new and useful parameter for estimating leaf chlorophyll content in leaves with any level of anthocyanins such as purple rice leaf.
Please see Additional file 1 for translations of the abstract into the six official working languages of the United Nations.
The chlorophylls (Chls), Chl-a and -b, are the most important plant pigments as they are necessary for photosynthesis. Chlorophylls absorb solar radiation and convert light energy to chemical energy. The amount of solar radiation that a leaf absorbs is determined mainly by chlorophylls. Thus, the foliar chlorophyll content directly affects the photosynthetic potential and primary production (Filella et al., 1995). Since chlorophyll contains much of the nitrogen content of a leaf, the chlorophyll content can provide an indirect estimation of a plant’s nutrient status (Moran et al., 2000). The chlorophyll content changes throughout the different stages of plant development, and it universally decreases in situations of plant stress and during senescence (Merzlyak and Gitelson, 1995).
Traditional methods of measuring leaf chlorophyll content are destructive. They require leaf extraction with organic solvents and spectrophotometric or HPLC measurements in solution, and are considered time-consuming and expensive (Gitelson et al., 2006; Blackburn, 2007). However, much attention has been devoted to developing nondestructive leaf optical methods as alternative means to measure leaf chlorophyll contents. These methods include transmittance and reflectance spectroscopy for chlorophyll content measurement, and can be applied across spatial scales because of their simplicity, rapidity, and nondestructive nature (Chappelle et al., 1992; Buschmann and Nagel 1993; Main et al., 2011).
Many theoretical models based on leaf reflectance were developed to predict the leaf chlorophyll content, water content, and other variables associated with the vegetative structure (Inoue et al., 1993; Jacquemoud et al., 1996; Dawson et al., 1998). Relationships between reflectance in the visible region of the spectrum and the chlorophyll content are fundamentally nonlinear (Gitelson and Merzlyak, 1997). However, reciprocal reflectances (R-1) in the green and red edge spectral regions were found to be closely related to the chlorophyll content in leaves with a wide range of pigment compositions and contents (Gitelson and Merzlyak, 1994). Indices using these spectral bands were suggested for chlorophyll estimation in leaves of various plant species (Zou et al., 2011; Dillen et al., 2012; Huang et al., 2012). Gitelson et al. (2003) reported that the index, (Rλ)-1 - (RNIR)-1, was linearly proportional to the chlorophyll content in the wide spectral bands 525 ~ 555 and 695 ~ 725 nm. Moreover, a derived index incorporating near infrared (NIR) reflectance, [(Rλ)-1 - (RNIR)-1] × RNIR, was also used to adjust for differences in leaf structure, and this algorithm gives highly accurate estimation of chlorophyll levels. Recently, a three-band conceptual model of remotely estimating total chlorophyll contents in leaves was developed and produced accurate estimation (Gitelson et al., 2006).
Leaf colors are determined by pigments, and anthocyanins are responsible for the red coloration in plant tissues (Ji et al., 1992; Gould et al., 1995). Therefore, leaves containing anthocyanins tend to be red or purple. Light in the spectral band at around 550 nm is absorbed by anthocyanins. The development of nondestructive methods for chlorophyll estimation using the green spectral band is more difficult for leaves containing anthocyanins, because their main absorption band is also located in the green range (., Merzlyak et al 2003). In anthocyanin-free leaves, nondestructive chlorophyll estimation are possible using either the green band (around 550 nm), red edge band (around 700 nm), or an NIR band of > 750 nm. For leaves containing anthocyanins, use of only the red edge band or an NIR band at > 750 nm is recommended (Gitelson et al., 2006).
In applying the above developed models for nondestructively estimating chlorophyll contents in leaves with or without anthocyanins (Gitelson et al., 2006), big error was found in anthocyanin-containing leaves, suggesting that anthocyanins may affect the nondestructive estimation of chlorophyll content in all purple leaves. Actually, the same phenomenon was also found and bothered the authors for more than fifteen years in biochemical and destructive measurement of chlorophyll in leaves with any level of anthocyanins; in short, the higher content the anthocyanins, the bigger error the chlorophyll measurement (personal data not published). We referred that the anthocyanins strongly absorbing around 500–550 nm may strongly affect the chlorophyll-related compounds, including all chlorophyll biosynthesis and degradation intermediates, strongly absorbing 400–475 nm and 550–700 nm. The more concentration of anthocyanins lead to the more distortion effect on chlorophyll measurement.
Therefore, the objectives of this study were to investigate the properties of the reflectance spectra of sweet potato leaves with and without anthocyanins, and to develop techniques for nondestructive chlorophyll estimation in leaves with a wide range of pigment contents. A technique was developed to eliminate the effects of anthocyanin on reflectance in the green spectral range on chlorophyll estimation, and a new index for estimating chlorophyll contents is finally proposed.
Plant materials and culture practice
Three leafy vegetable sweet potato [Ipomoea batatas (L.) Lam] cultivars of Taoyuan 2, green, yellow, and purple were used in this study. The green leaf cultivar was selected by the Taoyuan District Agricultural Research and Extension Station (TDARES) for food consumption of its leaves. Yellow and purple cultivars are local varieties from Taoyuan, and contain high amounts of carotenoids. Terminals about 40 cm in length were taken from sturdy vines, and cultivated in plastic boxes 60 cm long, 22 cm wide, and 15 cm deep, that contained a medium of sand, vermiculite, and loamy soil in a volume ratio of 2:1:1. Specimens were planted in a greenhouse at the TDARES, and evenly spaced at intervals of 60 cm to encourage similar growth rates and sizes. Plants were treated with the nutrients N:P2O5:K2O (0.1:0.2:0.3 g/box) weekly, watered every other day to maintain optimal irrigation, and allowed to grow for 22 d before the following measurements of spectral reflectance were made.
Measurement of spectral reflectance
Adaxial reflectance spectra of leaves were recorded at a rate of 600 nm/min using an U-3010 spectrophotometer (Hitachi, Tokyo, Japan) equipped with an integrating sphere attachment. The diffuse reflectance of the leaves was measured at values of 400 ~ 800 nm (R400 ~ R800) with a spectral resolution of 1 nm and against barium sulfate as a reference standard. The reflectance was expressed as a ratio of the radiance of the leaf to the radiance of the reference. The spectra were recorded for sections of leaves between the main veins.
Chlorophylls and carotenoids were isolated from leaves by homogenization in liquid nitrogen and subsequent threefold extraction with 80% acetone (v/v). After centrifugation for 5 min at 1500 × g, the absorbance of the supernatant was measured at 663.6, 646.6, and 440.5 nm (Porra et al., 1989). Leaf samples were also extracted with 1% (w/v) HCl in methanol, and the anthocyanin contents were assayed spectrophotometrically. The relative amounts of anthocyanins were expressed by [A530 - 0.333A657] m-2 (Mancinelli et al., 1975). Absorbance was measured with a Hitachi U-2000 UV-visible spectrophotometer.
To develop model datasets, 48 leaves from the three varieties were used. Validation was done using independent datasets derived from 24 leaves, and the chlorophyll content was calculated using reflectance data from the model’s validation datasets. The predicted chlorophyll content was compared to chlorophyll levels physically measured, and the root mean square error (RMSE) of the predictions was calculated. Relationships between the reflectance indices and chlorophyll contents were examined using simple linear regression models. All statistical analyses were conducted using JMP software, vers. 5.01 (SAS Institute, Cary, NC, USA).
The content of chlorophylls (Chl), carotenoids, and anthocyanins in leaves of three varieties of sweet potato
Chl (mg m-2)
Carotenoids (mg m-2)
Anthocyanins (A530-0.333A657) m-2
319 ~ 515
74 ~ 88
99 ~ 182
36 ~ 51
336 ~ 615
73 ~ 88
0.43 ~ 2.90
237 ~ 431
68 ~ 83
104 ~ 145
33 ~ 47
297 ~ 545
74 ~ 89
0.35 ~ 2.71
Correlations of reflectance indices with the chlorophyll content, and the RMSE with chlorophyll estimates
Correlations between the predicted and actual measured chlorophyll contents
It was previously shown that the blue and red spectral ranges, the main absorption bands of chlorophyll, are only sensitive to low chlorophyll levels, but at moderate to high levels, they become insensitive (Gitelson and Merzlyak, 199419961997). In these spectral ranges, reflectance is only sensitive to total chlorophyll levels at 90 mg m-2 (Gitelson et al., 2003). Such low levels of chlorophyll are found in highly stressed or senescent leaves. The three sweet potato varieties used in this study grew well under greenhouse conditions, and the lowest chlorophyll content recorded was 99 mg m-2 in yellow leaves (Table 1). In both the model development and validation datasets, chlorophyll contents were not low. Therefore, the blue and red spectral reflectances for sweet potato leaves should be insensitive to chlorophyll content. For the green and yellow (G + Y) sweet potato leaves, Rλ values in the visible blue (400 ~ 500 nm) and visible red (660 ~ 690) ranges showed the lowest correlations with chlorophyll content, but Rλ values showed maximum sensitivity to the chlorophyll content in the green (500 ~ 600 nm) and red edge (around 700 nm) ranges (Figure 2A). This was due to the strong absorption by chlorophyll in the blue (400 ~ 500 nm) and red (660 ~ 690 nm) spectral ranges, but weak absorption in the green (500 ~ 600 nm) and red edge areas (around 700 nm). Therefore, reflectance in the green and red edge ranges can be used as more-precise measurements for developing leaf chlorophyll content estimation algorithms (Lichtenthaler et al., 1996; Gamon and Surfus, 1999; Gitelson et al., 2002). For green and yellow (G + Y) leaves, three indices derived from the reciprocal reflectance, [(Rλ)-1 - (RNIR)-1], [(RNIR/Rλ) - 1], and [1 - (Rλ/RNIR)], displayed strong correlation with the total chlorophyll content in the green and red edge spectral ranges (Figures 3A, B, 4A).
The relationship between the spectral reflectance and chlorophyll content may be influenced by the presence of other pigments. Curran et al. (1991) found that the red edge was significantly shifted to longer wavelengths by amaranthin, a red pigment found in Amaranthus tricolor leaves. In anthocyanin-containing leaves, both chlorophyll and anthocyanin contribute to absorption and reflectance in the green spectral range. The development of a nondestructive method for estimating chlorophyll content is complicated by the interference of absorption by anthocyanins in leaves and fruits which contain them (Gitelson et al., 2001). In anthocyanin-free leaves and fruits, reflectances at the green (550 nm) and red edges (around 700 nm) were closely correlated with a wide range of chlorophyll contents (Gitelson et al., 2003). In leaves and fruits containing anthocyanins, a strong correlation was observed between the chlorophyll content and R550, while R700 was lost due to anthocyanin absorption at 550 nm causing the reflectance to be significantly lower than at 700 nm. This meant that spectral bands at around 550 nm were seen as unsuitable for nondestructive chlorophyll estimation in anthocyanin-containing leaves (Gitelson et al., 2001). In this study, only purple leaves contained anthocyanins (Table 1); therefore, correlations between the old derived indices of [(Rλ)-1 - (RNIR)-1] and [(RNIR/Rλ) - 1] and chlorophyll content were relatively weak in the green spectral range when the dataset included purple leaves. The RMSE of chlorophyll estimates using R-1 in the green spectral range was dramatically increased by the inclusion of anthocyanin-containing leaves. Independent of the presence of anthocyanins, R-1 indices were found to be highly sensitive to the total chlorophyll content in the red edge spectral range (Figures 2, 3).
When developing indices for nondestructive chlorophyll estimation, spectral bands at around 550 nm are not recommended for leaves which contain anthocyanins (Gitelson et al., 2003). Hence, we attempted to determine whether spectral reflectance in the green range could be used as a sensitive term in developing algorithms for estimating the chlorophyll content of leaves containing anthocyanins. The new index, [1 - (Rλ/RNIR)], was therefore developed to eliminate the effect of anthocyanins on reflectance in the green spectral range. Using this new index to predict the chlorophyll content, no sharp decrease in r2 or increase in RMSE was observed following the inclusion of purple leaves in the dataset (Figures 4, 5). This means that the green range reflectance becomes a usable term for estimating the chlorophyll content in anthocyanin-containing leaves.
In previous studies, spectral bands in green were determined to be unsuitable for chlorophyll nondestructive estimation with leaves containing anthocyanins (Gitelson et al., 2006; Merzlyak et al., 2003). In the present study, we developed a new index to eliminate the effect of anthocyanins on the reflectance in the green range when estimating chlorophyll contents of leaves containing anthocyanins. We concluded that the index, [1 - (Rλ/RNIR)], used with reflectance in the green and red edge ranges, is sufficient for nondestructive chlorophyll estimation in spite of the presence or absence of anthocyanins. Furthermore, the new index is more comprehensively applicable to nondestructively estimate the chlorophyll contents in any kind of leaves and fruits with any level of anthocyanins.
We thank Mr. Chang-Chang Chen for help with the greenhouse work and laboratory experiments.
- Blackburn GA: Hyperspectral remote sensing of plant pigments. J Exp Bot 2007, 58: 855–867.View ArticlePubMedGoogle Scholar
- Buschmann C, Nagel E: In vivo spectroscopy and internal optics of leaves as basis for remote sensing of vegetation. Int J Remote Sens 1993, 14: 711–722. 10.1080/01431169308904370View ArticleGoogle Scholar
- Chappelle EW, Kim MS, McMurtrey JE: Ratio analysis of reflectance spectra (RARS): an algorithm for the remote estimation of the concentrations of chlorophyll a, chlorophyll b, and carotenoids in soybean leaves. Remote Sens Environ 1992, 39: 239–247. 10.1016/0034-4257(92)90089-3View ArticleGoogle Scholar
- Curran PJ, Dungan JL, Macler BA, Plummer SE: The effect of a red leaf pigment on the relationship between red edge and chlorophyll concentration. Remote Sens Environ 1991, 35: 69–76. 10.1016/0034-4257(91)90066-FView ArticleGoogle Scholar
- Dawson TP, Curran PJ, Plummer SE: LIBERTY- modeling the effects of leaf biochemical concentration on reflectance spectra. Remote Sens Environ 1998, 65: 50–60. 10.1016/S0034-4257(98)00007-8View ArticleGoogle Scholar
- Dillen SY, Beeck MO, Hufkens K, Buonanduci M, Phillips NG: Seasonal patterns of foliar reflectance in relation to photosynthetic capacity and color index in two co-occurring tree species, quercus rubra and betula papyrifera. Agric. For. Meteor. 2012, 160: 60–68.View ArticleGoogle Scholar
- Filella I, Serrano I, Serra J, Peňuelas J: Evaluating wheat nitrogen status with canopy reflectance indices and discriminant analysis. Crop Sci 1995, 35: 1400–1450. 10.2135/cropsci1995.0011183X003500050023xView ArticleGoogle Scholar
- Gamon JA, Surfus JS: Assessing leaf pigment content and activity with a reflectometer. New Phytol 1999, 143: 105–117. 10.1046/j.1469-8137.1999.00424.xView ArticleGoogle Scholar
- Gitelson AA, Merzlyak MN: Spectral reflectance changes associated with autumn senescence of asculus hippocastamum and Acer platanoides leaves, Spectral features and relation to chlorophyll estimation. J. Plant Physiol. 1994, 143: 286–292. 10.1016/S0176-1617(11)81633-0View ArticleGoogle Scholar
- Gitelson AA, Merzlyak MN: Signature analysis of leaf reflectance spectral: algorithm development for remote sensing of chlorophyll. J Plant Physiol 1996, 148: 494–500. 10.1016/S0176-1617(96)80284-7View ArticleGoogle Scholar
- Gitelson AA, Merzlyak MN: Remote estimation of chlorophyll content in higher plant leaves. Int J Remote Sens 1997, 18: 291–298.View ArticleGoogle Scholar
- Gitelson AA, Merzlyak MN, Chivkunova OB: Optical properties and nondestructive estimation of anthocyanin content in plant leaves. Photochem Photobiol 2001, 74: 38–45. 10.1562/0031-8655(2001)074<0038:OPANEO>2.0.CO;2View ArticlePubMedGoogle Scholar
- Gitelson AA, Zur Y, Chivkunova OB, Merzlyak MN: Assessing carotenoid content in plant leaves with reflectance spectroscopy. Photochem Photobiol 2002, 75: 272–281. 10.1562/0031-8655(2002)075<0272:ACCIPL>2.0.CO;2View ArticlePubMedGoogle Scholar
- Gitelson AA, Gritz Y, Merzlyak MN: Relationships between leaf chlorophyll content and spectral reflectance and algorithms for non-destructive chlorophyll assessment in higher plant leaves. J Plant Physiol 2003, 160: 271–282. 10.1078/0176-1617-00887View ArticlePubMedGoogle Scholar
- Gitelson AA, Keydan GP, Merzlyak MN: Three-band model for noninvasive estimation of chlorophyll, carotenoids, and anthocyanin contents in higher plant leaves. Geophys Res Lett 2006, 33: L11402. doi:10.1029/2006GL026457 doi:10.1029/2006GL026457View ArticleGoogle Scholar
- Gould KS, Kuhn DN, Lee DW, Oberbauer SF: Why leaves are sometimes red. Nature 1995, 378: 241–242.View ArticleGoogle Scholar
- Huang N, Niu Z, Zhan Y, Xu S, Tappert MC, Wu C, Huang W, Gao S, Hou X, Cai D: Relationships between soil respiration and photosynthesis-related spectral vegetation indices in two cropland ecosystems. Agric. For. Meteor. 2012, 160: 80–89.View ArticleGoogle Scholar
- Inoue Y, Morinaga S, Shibayama M: Non-destructive estimation of water status of intact crop leaves based on spectral reflectance measurements. Jap. J. Crop Sci. 1993, 62: 462–469. 10.1626/jcs.62.462View ArticleGoogle Scholar
- Jacquemoud S, Ustin SL, Verdebout J, Schmuck G, Andreoli G, Hosgood B: Estimating leaf biochemistry using the PROSPECT leaf optical properties model. Remote Sens Environ 1996, 56: 194–202. 10.1016/0034-4257(95)00238-3View ArticleGoogle Scholar
- Ji SB, Yokoi M, Saito N, Mao LS: Distribution of anthocyanins in aceraceae leaves. Biochem. Systematic Ecol. 1992, 20: 771–781. 10.1016/0305-1978(92)90036-DView ArticleGoogle Scholar
- Lichtenthaler HK, Gitelson AA, Lang M: Non-destructive determination of chlorophyll content of leaves of a green and an aurea mutant of tobacco by reflectance measurements. J Plant Physiol 1996, 148: 483–493. 10.1016/S0176-1617(96)80283-5View ArticleGoogle Scholar
- Main R, Cho MA, Mathieu R, O’Kennedy MM, Ramoelo A, Koch S: An investigation into robust spectral indices for leaf chlorophyll estimation. J. Photogrammetry Remote Sens. 2011, 66: 751–761. 10.1016/j.isprsjprs.2011.08.001View ArticleGoogle Scholar
- Mancinelli AL, Huang-Yang CP, Lindquist P, Anderson O, Rabino I: Photocontrol of anthocyanin synthesis, III. The action of streptomycin on the synthesis of chlorophyll and anthocyanin. Plant Physiol 1975, 55: 251–257. 10.1104/pp.55.2.251View ArticlePubMedPubMed CentralGoogle Scholar
- Merzlyak MN, Gitelson AA: Why and what for the leaves are yellow in autumn? On the interpretation of optical spectra of senescing leaves (Acer platanoides L.). J Plant Physiol 1995, 145: 315–320. 10.1016/S0176-1617(11)81896-1View ArticleGoogle Scholar
- Merzlyak MN, Solovchenko AE, Gitelson AA: Reflectance spectral features and non-destructive estimation of chlorophyll, carotenoid and anthocyanin content in apple fruit. Postharvest Biol. Technol. 2003, 27: 197–211. 10.1016/S0925-5214(02)00066-2View ArticleGoogle Scholar
- Moran JA, Mitchell AK, Goodmanson G, Stockburger KA: Differentiation among effects on nitrogen fertilization treatments on conifer seedlings by foliar reflectance: a comparison of methods. Tree Physiol 2000, 20: 1113–1120. 10.1093/treephys/20.16.1113View ArticlePubMedGoogle Scholar
- Porra RJ, Thompson WA, Kriedemann PE: Determination of accurate extraction and simultaneously equation for assaying chlorophyll a and b extracted with different solvents: verification of the concentration of chlorophyll standards by atomic absorption spectroscopy. Biochem. Biophys. Acta 1989, 975: 384–394.Google Scholar
- Zou X, Shi J, Hao L, Zhao J, Mao H, Chen Z, Li Y, Mel H: In vivo noninvasive detection of chlorophyll distribution in cucumber (cucumis sativus) leaves by indices based on hyperspectral imaging. Anal Chim Acta 2011, 706: 105–112. 10.1016/j.aca.2011.08.026View ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited.