reconstruction of microalgal suspension absorption spectra ...backscattering (tassan and ferrari,...

© 2012 Science From Israel / LPPLtd., Jerusalem

IsraelJournalofPlantSciences Vol.60 2012 pp.231–242DOI: 10.1560/IJPS.60.1-2.231

*Author to whom correspondence should be addressed.E-mail: [email protected]‡Deceased.

Reconstruction of microalgal suspension absorption spectra from reflectance spectra of the cells deposited on GF/F filters

Alexei Solovchenko,a,b,* olgA chivkunovA,a Wen JuAn Sun,a And MArk MerzlyAka, ‡

aBiologicalFaculty,M.V.LomonosovMoscowStateUniversity,1/12LeninskieGori,GSP-1,119991Moscow,Russia

bK.A.TimiryazevInstituteofPlantPhysiology,RussianAcademyofSciences,Moscow,Russia

(Recieved30June2011andinrevisedform9October2011)

HonoringAnatolyGitelsonontheoccasionofhis70thbirthday

ABSTRACT

In this communication we present an approach for reconstruction of suspension absorption spectra of microalgae as measured in standard 1-cm spectrophotometric cuvettes from reciprocal reflectance, R(λ)–1, spectra of cells deposited on glass-fiber filters. Reciprocal reflectance exhibited the highest dynamic range (eight times wider in comparison with the functions –log(T) and –log(R)) and showed almost linear relationships with suspension absorbance and chlorophyll content from 0.7 to 9.9 μg/ml. The relationships between suspension absorbance and the –log(T) and log[(1–R)/T] functions were non-linear and were fitted by second-order polynomials with high contribution of the quadratic term. The use of R(λ)–1 allowed the recon-struction of suspension absorbance of six species from key groups of planktonic microalgae in the range from 0.001 to 0.250, with RMSE < 0.02. The advantages of the suggested approach over traditional techniques employing –logT(λ) of the filters are discussed, including the suitability for plankton studies in the field and for the monitoring of cultivated microalgae.

Keywords: Microalgae optics, absorbance spectra reconstruction, reflectance, trans-mittance, quantitative filter technique

INTRODUCTION

Spectral information obtained in aquatic environments may be successfully used for the estimation of phy-toplankton concentration, its spatial distribution and taxonomic composition (Mayo et al., 1995; Yacobi et al., 1995; Schalles et al., 1998; Morel and Maritorena, 2001). A special challenge is posed when the density of microalgae in water is extremely low (<1 mg m–3 of total chlorophyll, Chl), and requires the concentration of relatively large water samples, mostly on glass-fiber

filters, followed by spectrophotometry and reconstruc-tion of the absorbance spectra of the suspended particles (Mitchell et al., 2003). The measurements of light ab-sorption by filters carrying microalgal samples (known as quantitative filter technique, QFT) are routinely used for laboratory measurements (Mitchell et al., 2003; Tas-san and Ferrari, 2002). Measurement of spectra of cells collected onto filters may be handy in studies of lab and (outdoor) mass cultures of microalgae, especially for characterization of microalgal species possessing cells that tend to clot or rapidly fall on the cuvette floor during

Abbreviations: Chl—(total) chlorophyll(s), CLSM—confocal laser scanning microscope, QFT—quantitative filter tech-nique

IsraelJournalofPlantSciences 60 2012

232

spectrum scanning (Solovchenko et al., 2010, 2011).The reconstruction of absorbance spectra of microal-

gal cell suspensions from spectra of the cells deposited on filters requires an understanding of the relationships between optical properties of the microalgal cells as measured in standard spectrophotometric cuvettes and on filters. Differing values have been reported for the correlation between absorption by a microalgal suspen-sion sample and attenuation of light passing through the sample collected onto a filter. These studies revealed a wide variation of bio-optical properties of phytoplank-ton (Dall’Olmo and Gitelson, 2005). This variation is caused by changes in taxonomic composition of the phytoplankton and the environmental conditions where the phytoplankton thrives (Bricaud et al., 2004). However, it should be remembered that, in addition to the mentioned factors, bio-optical variation is also con-tributed by factors inherent to the spectrophotometry of algal samples deposited on glass-fiber filters, such as strong light scattering. According to Mitchell (1990, 2003), the optical density of microalgae in cuvette and on Whatman glass-fiber filters exhibits non-linear re-lationships, which could be empirically described by a second order polynomial. A“transmittance–reflectance” method was suggested to account for light losses due to backscattering (Tassan and Ferrari, 1995, 2002). How-ever, this method, though more precise, appears to be more laborious since it requires at least one additional (reflectance) spectrum scan.

Reflectance-based algorithms are widely used in remote sensing of phytoplankton in water bodies (Gi-telson et al., 1993; Mayo et al., 1995; Schalles et al., 1998). Gitelson and colleagues successfully applied a reflectance-based technique for remote estimation of cell mass and Chl content in Spirulina (Gitelson et al., 1995) and ultra-high Nannochloropsis culture (Gitelson et al., 2000). Reciprocal reflectance, R(λ)–1, spectra were successfully employed for non-destructive assay of pigment content in whole higher plant leaves and fruits (Merzlyak et al., 2003; Merzlyak et al., 2005; Gitelson et al., 2009). That experience inspired the at-tempt to reconstruct the suspension absorption spectra of microalgae from their reflectance spectra measured after deposition on a filter, which is the subject of the current report. The hypothesis that pigment content, absorbance, and reflectance of microalgal cells loaded onto a filter maintain close relationships is seemingly in line with the theory of diffuse reflectance developed in previous years (Atherton, 1955; Aldeeson et al., 1961). To the best of our knowledge, no attempts have been made so far to obtain the spectral characteristic of microalgal samples solely via reciprocal reflectance spectra measurements. Therefore, we investigated the

relationships between absorbance of microalgal cell suspensions measured in a standard spectrophotometric cuvette and optical properties (reflectance and absor-bance) of the same cells deposited on glass-fiber filters. Finally, we compared the efficiency of different spectral functions in reconstruction of the microalgal suspension absorbance spectra.

METHODS

Microalgal cultures

Cultures of microalgae representing key groups of phy-toplankton were used in this work: the cyanobacteria AnabaenavariabilisKütz. ATCC 29413 and Anacystisnidulans (Richt.) Drouet & Daily (Synechococcuselon-gatusNäg.) IPPAS B-267 (Cyanobacteria) cultivated on BG-11 medium (Stanier et al., 1971), the chlorophytes Chlorella pyrenoidosa Chick and Scendesmus quad-ricauda (Turp.) (Chlorophyta) cultivated on Tamiya medium (Tamiya, 1966), and the diatoms Thalassiosiraweissflogii (Grunow) G. Fryxell & Hasle andSkeletone-macostatum (Grev.) cultivated on f/2 medium (Guillard and Ryther, 1962). The cyanobacterial cultures were obtained from the collection of the Department of Bio-engineering of M.V. Lomonosov MSU, the cultures of the chlorophytes and the diatoms were kindly supplied by Profs. D.N. Matorin and S.I. Pogosyan (Department of Biophysics, Biological Faculty of M.V. Lomonosov MSU) who originally isolated the strains. The cultures were maintained at 25 °C under the irradiance of 100 μmol PAR photons m–2 s–1. The irradiance was measured with LI-850 quantometer (LiCor, USA)

Sample preparation

A suspension aliquot was pelleted by centrifugation and re-suspended in fresh medium to obtain stock sus-pension of ca. 0.15–0.45 OD at 678 nm. Then, seven dilutions were made stepping 1/8 (from 7/8 to 1/8 of the stock suspension) using corresponding cultivation medium for each species. Absorbance spectra of the stock and each of the diluted samples were recorded. Then aliquots of 0.75 mL were taken from each sample. The aliquots were made up to 15 mL with correspond-ing cultivation medium and applied on glass-fiber GF/F (Whatman) filters, 21 mm diameter, using a Millipore # xx10 025 03 filter holder with a glass funnel and a vacuum aspirator. All measurements we performed in five replicates.

Total chlorophyll assay

Cells were pelleted by centrifugation, transferred to a glass–glass homogenizer with a chloroform–methanol (10 mL, 2:1, v/v) mixture and extracted to remove all

Solovchenkoetal./Reconstructionofmicroalgalspectra

233

pigment. The lipid fraction including Chl was separated according to Folch et al. (1957). The chloroform phase was used for further pigment analysis. Total Chl was quantified using absorption coefficients for chloroform (Wellburn, 1994).

Spectral measurements

The spectra of the suspensions were recorded with Hita-chi 150–20 spectrophotometer equipped with 150 mm integrating sphere fitted with a cuvette compartment (Fig. 1). In 1 cm glass cuvette, the spectra of cell suspen-sions were taken at two distances: as close as possible to, D(λ;g

0) (Fig. 1a), and 1 cm apart from the detector

entry window, D(λ;gI) (Fig. 1b), for additional detail see

(Merzlyak et al., 2008; Merzlyak and Naqvi, 2000). In all cases, a cuvette with corresponding cultivation me-dium was used as a reference. The scattering-corrected optical density spectra, Dc (λ), were then calculated as

(1)where and —light attenuation at 800 nm

where pigments do not possess detectable absorption, measured as close as possible and 1 cm apart from the detector window, respectively; for further details, see Merzlyak et al., 2008, and Merzlyak and Naqvi, 2000.

Routinely, the transmittance and reflectance spectra of the same samples were recorded after deposition on a glass-fiber filter (GF/F, 25 mm diameter, Whatman) as described above against an empty filter soaked in the corresponding medium (Fig. 1c,d). This approach allowed us to obtain a flat baseline in the spectral range studied (350–800 nm, see the spectra in Figs. 7–12).

In order to characterize optical properties of the filters, in certain experiments transmittance and reflec-tance spectra were recorded against a BaSO4 reflectance standard (Hitachi, Japan). In this case measured sample spectra were offset by subtracting the spectrum of a blank filter from the measured spectra (see, e.g., Fig. 4). Spectra calculated using both approaches were nearly identical.

For transmittance measurements, the wet filters were mounted with the deposited cells facing the interior of the IS) on the entry ports of the integrating sphere of the spectrophotometer (Fig. 1c) or, for reflectance measure-

Fig. 1. Schematic representation (top view) of (a, b) suspension absorbance, (c) filter transmittance, and (d) reflectance measure-ment setup. 1—integrating sphere; 2—cuvette compartment; 3, 4—sample and reference entry ports, respectively; 5, 6—BaSO4 high reflectance plates covering sample and reference exit ports, respectively; 7—detector; 8—reference cuvette; 9—sample cuvette; 10—blanc cuvette; 11—sample filter (loaded with microalgal cells which face the interior of the integrating sphere); 12—blanc filter; S, R—sample and reference measuring beams. For explanation, see text.


234

ments, at the exit ports of the IS (Fig. 1d). Optical den-sity of the filters, Df (λ), was expressed as –log T, where T is transmission. Reciprocal reflectance was calculated as Rf (λ)–1, where Rf is reflectance of the filter.

In a separate series of experiments, wet filters with the cells deposited onto them were dried over silica gel under vacuum at room temperature and then saturated with Immersol TM 518N immersion oil (Carl Zeiss, Germany) after transmittance and reflectance measure-ments, and then the measurements were repeated.

Optical microscopy

The integrity of the cells of the studied algal species upon deposition on GF/F filters was controlled visually on the microphotographs made using an Axioskop 2 light photomicroscope fitted with an Axiocam MRc digital camera (Carl Zeiss, Germany).

Analysis of the algal cell distribution within GF/F filters

The distribution of cells within GF/F filters was in-vestigated using confocal laser scanning microscope (CLSM) LSM 510 Meta (Carl Zeiss). Fluorescence of Chl present in the algal cells and used as an intrinsic fluorescent probe was excited with the laser line at 633 nm and its emission in the band of 650–710 nm was registered. The recorded z-stacks were processed using the bundled software (Carl Zeiss) to obtain color-coded depth maps representing the distribution of algal cells within the GF/F filter.

RESULTS

Our light microscopy examination showed that the cells of the studied algal species suffered no damage upon deposition onto the GF/F glass-fiber filter and were more or less evenly distributed in the plane of the filter surface (Fig. 2a,b). At the same time, the CLSM analysis showed that cell size and the irregular landscape of the filter exerted a considerable impact on the pattern of the cell distribution within the filter depth (Fig. 2b). As apparent from Fig. 2c, virtually all cells were captured within the first 100 μm of GF/F filter depth, while the entire filter is ca. 400 μm thick.

Absorbance spectra of the stepwise diluted microal-gal cell suspension samples were recorded in standard spectrophotometric cuvettes and compensated for scat-tering, as described above in “Methods”, yielding scat-tering-corrected spectra Dc (λ) similar to those shown in Fig. 3. For all of the microalgal species studied, the amplitude of the main absorption bands in the red and Soret band of the scattering-compensated suspension absorbance spectra was linearly related with sample

cell density and Chl content in the whole range studied (r2 > 0.99). The Dc(λ) spectra did not show a detectable absorption in the NIR region, and after normalization to the red maximum coincided (not shown), suggesting that distortion of the shape of spectra due to scatter-ing-related effects and photometric inaccuracy did not occur.

The same suspension samples were deposited on GF/F filters and subsequently transmittance, Tf (λ) (Fig. 4A), and reflectance, Rf(λ) (Fig. 4B), spectra of the filters were measured. The Tf (λ) of a wet filter in the NIR was ca. 37% monotonously decreasing to ca. 25% at 350 nm; Rf(λ) increased from 66% in the NIR to 81% at 350 nm (see curves labeled “0” in Fig. 4).

The measured Tf (λ) and Rf (λ) spectra were then used for calculation of absorbance Df (λ) (Fig. 4A), and reciprocal reflectance, Rf (λ)–1, spectra (Fig. 4B). The Df (λ) of the wet blank filters displayed featureless absorption increasing towards longer wavelengths; the Rf (λ) amplitude of the blank filters decreased and that of accordingly increased towards longer wavelengths (not shown). After subtraction of blank filter spectrum, the Tf(λ), Df (λ), and Rf(λ) spectra became almost flat in the NIR (Fig. 4).

The spectra of the microalgal cells deposited on the GF/F filters possessed essentially the same spectral fea-tures as the Dc(λ) spectra and retained the positions of main maxima in the blue and red regions (Figs. 3, 4). In all cases the difference between normalized Dc (λ) and Df (λ) or Rf(λ) spectra did not exceed 2% with the ex-ception of the cyanobacterium A.variabilis in which the difference reached 6% in the Soret band (not shown). The amplitude of the bands in Df(λ) spectra values was, on an average, three times higher (Figs. 3, 4) and the amplitude of Rf(λ)–1 maxima was 25 times higher than that of Dc(λ) (Figs. 3, 4B). Accordingly, the variation in the magnitude [Rf(λ)]–1 spectra of diluted suspension samples deposited on GF/F filters was more than one order of magnitude higher than that in the corresponding Df(λ) spectra.

Saturation of GF/F filters, both blank and loaded with cells, with immersion oil brought about a dramatic reduction in their scattering, e.g., the attenuation in NIR decreased from 0.4 (Fig. 5A) to ca. 0.01 (Fig. 5). Visually the oil-saturated filters became completely transparent resembling a solid piece of glass. As a result, the Df (λ) spectra of oil-saturated filters became very similar to Dc(λ) of the same samples (Fig. 5).

The optical properties of suspended cells, Dc (λ), and after deposition of the cells onto GF/F filters (Df (λ),–log[Rf(λ)], and Rf(λ)–1) were compared in the whole spectral range used in this study (Fig. 6). In all cases these relationships could be fitted by a second-order polynomial


235

Dc(λ) = ax+bx2 (2)

where x— Df (λ),–log[Rf(λ)], or Rf(λ)–1).The comparison revealed that the non-linearity, as

it appeared from the contribution of the quadratic term (coefficient b) in the polynomial fit equation was con-siderably higher in the cases of the relationships “Dc (λ) vs. Df (λ)” (Fig. 6A) and “Dc (λ) vs. –log[Rf (λ)]”

(Fig. 6B) than for “Dc (λ) vs. Rf (λ)–1” (Fig. 6C). It should be noted that the latter relationship was charac-terized by relatively low contribution of the quadratic term (b, see Fig. 6) in a wide range of cell densities and regardless of the species studied. As could be seen from Fig. 6C, these relationships departed from linearity at Rf (λ)–1 values higher than 2–2.5.

The empirical equations (eq 2) obtained as a result

Fig. 2. Microalgal cells (the diatom Thalassiosira weissflogii) deposited on GF/F glass-fiber under optical microscope (a) and and pattern of their distribution as revealed by chlorophyll fluorescence registered with a confocal laser scanning microscope (b, c). In (b), a typical depth map is shown for T. weissflogii (the cell depth is color-coded; see the scale at the bottom). In (c), quantitative data on chlorophyll distribution within the volume the filter are presented for the four microalgal species studied.


236

Fig. 3. Typical Dc(λ) spectra of dilution series (the corresponding dilution factor is indicated below each curve). Total chloro-phyll concentration is shown for each sample.

Fig. 4. Representative spectra of GF/F filters carrying different amounts of C.pyrenoidosa cells (the same series as presented in Fig. 3). (a) Tf(λ) spectra (black curves) and calculated on their basis Df(λ) spectra (magenta curves). (b) Rf(λ) spectra (black curves) and calculated on their basis Rf(λ)–1 spectra (magenta curves). Dilutions are indicated in each panel in corresponding color; ‘0’—the spectra of blank (carrying no algal cells) GF/F filters.


237

(3)

where Dc(λi) and Dc¢(λi) are the measured and the cor-responding reconstructed Dc(λ) spectra. Results of the reconstruction for all investigated species (excepting the deviating datasets for Chlorella and Anabaena in the cases of Df(λ) and Rf(λ)–1, see Fig. 6) are shown in Figs. 7–12. All three spectral functions provided a reasonable precision of Dc (λ) reconstruction over whole cell density and spectral range studied (Figs. 7–12). It should be noted, however, that including the whole dataset brought about a decline in the precision of Dc (λ) reconstruction viaDf(λ) values in comparison that achieved using Rf(λ)–1 values (not shown). Higher deviation of the reconstructed Dc¢(λ) from the measured Dc(λ) spectra was observed at higher cell densities dur-ing reconstruction with the use of Df(λ). Reconstructed Dc¢(λi) spectra retained a tight linear (r2 > 0.98) relation-ship with cell density and Chl content of the microalgal suspension sample studied characteristic of the mea-sured Dc (λ) spectra regardless of the reconstruction algorithm employed.

DISCUSSION

As it could be seen from Fig. 2, the deposition of mi-croalgal cells on glass-fiber filter results in formation of an optically complex system comprised by small pigmented particles (Chl-containing cells) distributed

Fig. 5. Dc(λ) spectrum of C.pyrenoidosa cell suspension (blue symbols), Df (λ) of the same sample before (magenta), and after (green) saturation of the filter with immersion oil.

Fig. 6. Relationships between OD (a), antilogarithm of reflectance (b) and reciprocal reflectance (c) of the GF/F filters carrying deposited microalgal cells and scattering-corrected absorption of corresponding cell suspension samples. The approximation (solid line) and prediction limits (dashed lines) were calculated for the entire visible range and for all the dilutions studied.

of comparative analysis of the relationships between optical properties of the microalgal cells measured in cuvettes and on GF/F filters were tentatively employed for reconstruction of Dc(λ) spectra from spectra of the samples deposited on the filters. At the same time root mean square error (RMSE) of the reconstruction was calculated in the whole spectral range and for all dilu-tion studied as


238

within strongly scattering medium (Fig. 4A). Though the spectra of the cells deposited on filters contained the same maxima as could be found in the spectra of suspended cells, the amplitude of the former is consider-ably higher than that of the latter (Figs. 3, 4), obviously

due to effects related with amplification of light path due to multiple scattering in the filter (Roesler, 1998; Lohrenz, 2000). A number of approaches have been proposed for the correction of the microalgal spectra measured using QFT most of which were designed for

Fig. 7. Reconstruction of A.variabilis Dc(λ) spectra using (a) Df (λ), –log[Rf(λ)], and (b) Rf(λ)–1 spectra. Thick solid lines represent Dc(λ), lines with symbols—Dc¢(λ); corresponding spectra of RMSE of Dc¢(λ) is shown below.

Fig. 8. Reconstruction of A.nidulans Dc(λ) spectra using (a) Df (λ), –log[Rf(λ)], and (b) Rf(λ)–1 spectra. For designation of the curves see Fig. 7.


239

Fig. 9. Reconstruction of C.pyrenoidosa Dc(λ) spectra using (a) Df (λ), –log[Rf(λ)], and (b) Rf(λ)–1 spectra. For designation of the curves see Fig. 7.

Fig. 10. Reconstruction of S.obliqus Dc(λ) spectra using (a) Df (λ), –log[Rf(λ)], and (b) Rf(λ)–1 spectra. For designation of the curves see Fig. 7.

transmittance-based measurements (Mitchell, 1990; Roesler, 1998) or transmittance measurements using re-flectance measurements for a backscattering correction (Tassan and Ferrari, 1995).

It should be noted, however, that the amount of light

reflected by the filter was two to three times higher than the amount of light transmitted (cf. Figs. 3A,B). The bal-ance of light transmitted and reflected by the filter, taking into account the inaccuracy of the measurements, shows that losses of light due to attenuation by the filter per se


240

Fig. 11. Reconstruction of T. weissflogii Dc(λ) spectra using (a) Df (λ), –log[Rf(λ)], and (b) Rf(λ)–1 spectra. For designation of the curves see Fig. 7.

Fig. 12. Reconstruction of Sk.costatum Dc(λ) spectra using (a) Df (λ), –log[Rf(λ)], and (b) Rf(λ)–1 spectra. For designation of the curves see Fig. 7.

and an increase in reflectance towards blue-violet which was observed during the filter measurements (curves 0 in Fig. 4). These phenomena could, at least in part, be ascribed to the increase of scattering coefficient with the decrease of wavelength (Butler and Norris, 1960).

were relatively small. It appears that the increase in light attenuation observed in samples deposited on filters ap-parently results from strong backscattering (reflection) of light. This is in line with high reflectance values (ca. 80% in the NIR), a monotonous decrease of transmittance,


241

A considerable increase in absorption of light by microalgal cells deposited on filters in comparison with suspended cells (Figs. 2, 3) stems from amplification of the effective optical path due to multiple scattering, it could be illustrated by the reduction of amplitude of Df (λ) to the level characteristic of Dc(λ) by saturating the filter loaded with cells in the immersion oil (pos-sessing a refraction index close to that of glass of which the filers are made) thereby removing the glass–air interfaces (Fig. 5).

In this work we tried to employ light reflected by microalgal cells deposited on filters since it appears to constitute a more ample signal in comparison with transmitted light (Fig. 4) for reconstruction of the spec-tra of corresponding cell suspensions. The relationships between Dc (λ) and spectral function obtained from on-filter measurements appeared to be essentially non-linear (Figs. 6A,B). In line with previous findings for Dunaliella tertiolecta, Chlorella sp., Nannochloropsissp., and Thalassiosira fluviatilis (Mitchell, 1990; Mitch-ell et al., 2003) as well as a large number of samples from natural water bodies (Lohrenz, 2000; Roesler, 1998; Tassan and Ferrari, 1998), these relationships were successfully fitted with second-order polynomials with significant contribution of quadratic term: the ratio b/a (see eq 3) was equal to 1.63 and 0.99 for Df (λ) and –log[Rf (λ)], respectively (eq 3, Figs. 6A,B). By con-trast, the relationships “Dc(λ) vs. Rf(λ)–1” were consid-erably more linear (b/a = 0.06; Fig. 6C). These findings are in accordance with the theory of diffuse reflectance developed by Atherton (1955) widely employed for the description of absorption of dyes applied to textiles or paper. According to this theory, the following is true for monochromatic beam in a pigment-containing strongly scattering medium:

R(λ)–1–R0(λ)–1 = c1e1(λ) + c2e2(λ) + ... + cnen(λ), (4)

where R0(λ)—reflectance of the medium (blank GF/F filter in this case), cn—concentration and en—absorption coefficients at wavelength λ for the pigments present in the system (Atherton, 1955; Aldeeson et al., 1961). In-deed, blank corrected Rf (λ)–1 was linearly related to Chl content in the range studied (r2 = 0.99). The departure from linearity was observed at higher Rf (λ)–1 values, probably due to general non-linearity of the relationships between pigment content and reflectance signal in strong-ly scattering media at high content of the pigment.

An essential point is that the relationships “Dc(λ) vs. Rf (λ)–1” obeyed a single law for all six species studied (Fig. 6C). This fact suggests that reciprocal reflectance-based approach could be applied for analysis of natural samples with diverse taxonomic composition without

compromising the precision of reconstruction. Our pilot experiments with samples containing different propor-tions of cyanobacteria, green microalgae and diatoms confirmed this suggestion. Furthermore, the reciprocal reflectance-based technique is particularly suited for studying plankton in the field since it could be easily implemented with portable fiber-optic spectrometers and reflectance probes. Another series of pilot experi-ments demonstrated the feasibility of reliable Rf (λ)–1 measurements even without the use of a conventional integrating sphere.

Collectively, the results obtained in the present work show that the reconstruction of Dc(λ) solely via reflec-tance measurements is feasible and is at least as efficient as that performed using transmittance measurements. Furthermore, due to the wider dynamic range, the higher linearity of “Rf (λ)–1 vs. Dc(λ)” the sensitivity of Dc(λ) reconstruction via Rf (λ)–1 increases. As a result, the use of reflectance measurements of cells deposited on filters, and subsequent reciprocation of the measured values is advantageous for quantitative characterization of plankton in natural ecosystems and (mass) cultivated microalgae.

NOMENCLATUREDc(λ)—scattering-corrected OD spectrum of algal cell

suspension recorded in standard spectrophotometric cuvette.

Dc¢(λ)—reconstructed Dc(λ) spectrum.Df(λ)—OD spectrum of algal cells deposited onto GF/F

filter.Rf (λ)—reflectance spectrum of algal cells deposited

onto GF/F filter.Rf(λ)–1—reciprocal reflectance spectrum of algal cells

deposited onto GF/F filter.

ACKNOWLEDGMENTS

The authors are grateful to Prof. D.N. Matorin and G.A. Dolnikova for generous donation of microalgal cultures, and to Dr. D.B. Kiselevsky for kind assistance with CLSM. Financial support of the Russian Foundation for Basic Research (Grants #09-04-00419, #12-04-00743) and Ministry of Science and Education of Russian Federation (Project #16.512.11.2182) is greatly ap-preciated. The authors are also would like to thank two anonymous reviewers for their thorough analysis of and insightful comments on the manuscript.

REFERENCES

Aldeeson, J., Atherton, E., Derbyshire, A. 1961. Modern physical techniques in colour formulation. J. Soc. Dyers Colour. 77: 657–669.


242

Atherton, E. 1955. The relation of the reflectance of dyed fab-rics to dye concentration and the instrumental approach to colour matching. J. Soc. Dyers Colour. 71: 389–398.

Bricaud, A., Claustre, H., Ras, J., Oubelkheir, K. 2004. Natu-ral variability of phytoplanktonic absorption in oceanic waters: influence of the size structure of algal populations. J. Geophys. Res. 109.

Butler, W., Norris, K. 1960. The spectrophotometry of dense light-scattering material. Arch. Biochem. Biophys. 87: 31.

Dall’Olmo, G., Gitelson, A. 2005. Effect of bio-optical param-eter variability on the remote estimation of chlorophyll-a concentration in turbid productive waters: experimental results. Appl. Optics 44: 412–422.

Folch, J., Lees, M., Sloane-Stanley, G. 1957. A simple method for the isolation and purification of total lipids from animal tissues. J. Biol. Chem. 226: 497–509.

Gitelson, A., Garbuzov, G., Szilagyi, F., Mittenzwey, K., Karnieli, A., Kaiser, A. 1993. Quantitative remote sensing methods for real-time monitoring of inland waters quality. Int. J. Rem. Sens. 14: 1269–1295.

Gitelson, A., Laorawat, S., Keydan, G., Vonshak, A. 1995. Op-tical properties of dense algal cultures outdoors and their application to remote estimation of biomass and pigment concentration in Spirulina platensis (Cyanobacteria). J. Phycol. 31: 828–834.

Gitelson, A., Grits, Y., Etzion, D., Ning, Z., Richmond, A. 2000. Optical properties of Nannochloropsissp. and their application to remote estimation of cell mass. Biotechnol. Bioeng. 69: 516–525.

Gitelson, A., Chivkunova, O., Merzlyak, M. 2009. Nonde-structive estimation of anthocyanins and chlorophylls in anthocyanic leaves. Am. J. Bot. 96: 1861.

Guillard, R., Ryther, J. 1962. Studies of marine planktonic diatoms: i. Cyclotellanana Hustedt, and Detonulaconfer-vacea (Cleve) Gran. Can. J. Microbiol. 8: 229–239.

Lohrenz, S.E. 2000. A novel theoretical approach to correct for pathlength amplification and variable sampling load-ing in measurements of particulate spectral absorption by the quantitative filter technique. J. Plankton Res. 22: 639–657.

Mayo, M., Gitelson, A., Yacobi, Y., Ben-Avraham, Z. 1995. Chlorophyll distribution in lake Kinneret determined from Landsat Thematic Mapper data. Int. J. Rem. Sens. 16: 175–182.

Merzlyak, M.N., Naqvi, K.R. 2000. On recording the true ab-sorption spectrum and the scattering spectrum of a turbid sample: application to cell suspensions of the cyanobac-terium Anabaenavariabilis. J. Photochem. Photobiol. B: Biol. 58: 123–9.

Merzlyak, M., Solovchenko, A., Gitelson, A. 2003. Reflec-tance spectral features and non-destructive estimation of chlorophyll, carotenoid and anthocyanin content in apple fruit. Postharvest Biol. Technol. 27: 197–212.

Merzlyak, M.N., Solovchenko, A.E., Smagin, A.I., Gitelson, A.A. 2005. Apple flavonols during fruit adaptation to solar radiation: spectral features and technique for non-destruc-tive assessment. J. Plant Physiol. 162: 151–60.

Merzlyak, M., Chivkunova, O., Maslova, I., Naqvi, K., So-

lovchenko, A., Klyachko-Gurvich, G. 2008. Light absorp-tion and scattering by cell suspensions of some cyanobac-teria and microalgae. Russ. J. Plant Physiol. 55: 420–425.

Mitchell, B. 1990. Algorithms for determining the absorption coefficient for aquatic particulates using the quantita-tive filter technique. Proc. SPIE, Ocean Optics, X 1302: 137–148.

Mitchell, B., Bricaud, A., Carder, K., Cleveland, J., Ferrari, G., Gould, R., Kahru, M., Kishino, M., Maske, H., Moisan, T. 2003. Determination of spectral absorption coefficients of particles, dissolved material and phytoplankton for discrete water samples. NASA Tech. Memo 209966: 125–153.

Morel, A., Maritorena, S. 2001. Bio-optical properties of oceanic waters: a reappraisal. J. Geophys. Res. 106: 7163–7180.

Roesler, C.S. 1998. Theoretical and experimental approaches to improve the accuracy of particulate absorption coef-ficients derived from the quantitative filter technique. Limnol. Oceanogr.: 1649–1660.

Schalles, J., Gitelson, A., Yacobi, Y., Kroenke, A. 1998. Esti-mation of chlorophyll a from time series measurements of high spectral resolution reflectance in an eutrophic lake. J. Phycol. 34: 383–390.

Solovchenko, A., Merzlyak, M., Khozin-Goldberg, I., Cohen, Z., Boussiba, S. 2010. Coordinated carotenoid and lipid syntheses induced in Parietochloris incisa (Chlorophyta, Trebouxiophyceae) mutant deficient in Δ5 desaturase by nitrogen starvation and high light. J. Phycol. 46: 763–772.

Solovchenko, A., Khozin-Goldberg, I., Recht, L., Boussiba, S. 2011. Stress-induced changes in optical properties, pigment and fatty acid content of Nannochloropsis sp.: implications for non-destructive assay of total fatty acids. Mar. Biotechnol. 13: 527–535.

Stanier, R., Kunisawa, R., Mandel, M., Cohen-Bazire, G. 1971. Purification and properties of unicellular blue-green algae (order Chroococcales). Microbiol. Mol. Biol. Rev. 35: 171–205.

Tamiya, H. 1966. Synchronous Cultures of Algae. Annual review of plant physiology 17: 1–27.

Tassan, S., Ferrari, G. 1995. An alternative approach to ab-sorption measurements of aquatic particles retained on filters. Limnol. Oceanogr. 40: 1358–1368.

Tassan, S., Ferrari, G.M. 1998. Measurement of light absorp-tion by aquatic particles retained on filters: determination of the optical pathlength amplification by the ‘transmit-tance-reflectance’method. J. Plankton Res. 20: 1699.

Tassan, S., Ferrari, G. 2002. A sensitivity analysis of the’Transmittance-Reflectance’method for measuring light absorption by aquatic particles. J. Plankton Res. 24: 757–774.

Wellburn, A. 1994. The spectral determination of chlorophyll a and chlorophyll b, as well as total carotenoids, using vari-ous solvents with spectrophotometers of different resolu-tion. J. Plant Physiol. 144: 307–313.

Yacobi, Y., Gitelson, A., Mayo, M. 1995. Remote sensing of chlorophyll in Lake Kinneret using highspectral-resolution radiometer and Landsat TM: spectral features of reflectance and algorithm development. J. Plankton Res. 17: 2155.

reconstruction of microalgal suspension absorption spectra ...backscattering (tassan and ferrari,...

Documents