Sunteți pe pagina 1din 16

Remote Sensing of Environment 174 (2016) 181196

Contents lists available at ScienceDirect

Remote Sensing of Environment


journal homepage: www.elsevier.com/locate/rse

Analysis of the effect of chloroplast arrangement on optical properties of


green tobacco leaves
Barbora Barnkov, Duan Lazr, Jan Nau
Department of Biophysics, Centre of the Region Han for Biotechnological and Agricultural Research, Faculty of Science, Palack University, lechtitel 27, 783 71 Olomouc, Czech Republic

a r t i c l e

i n f o

Article history:
Received 15 July 2015
Received in revised form 5 November 2015
Accepted 9 December 2015
Available online xxxx
Keywords:
Diffuse reectance
Diffuse transmittance
Absorptance
Leaf absorbance
Inltration
Chlorophyll uorescence
Collimated light
Diffuse light
Nonlinearity

a b s t r a c t
There are many studies showing the active optical reaction of a green leaf to the changing surroundings based on
chloroplast movement and their rearrangement in plant cells. These studies concentrated mostly on the effect of
one feature (leaf type, leaf side or light type) on the leaf optical spectra. We have measured the diffuse reectance
and transmittance spectra of tobacco green leaves in combination of 4 variants: in normal and water inltrated
leaves, in collimated or diffuse incident light, on both the adaxial and abaxial leaf sides, and for the face or side
chloroplast arrangement. A Simple Explicitly Non-Linear Empirical model for Leaf Optical Properties (SENLELOP
model) is used to theoretically describe, simulate and t the deviations from the LambertBeer's law causing
nonlinearity in the measured spectral changes. It is shown that the incident diffuse light is captured by the leaf
more effectively than the collimated light. The light incident from the adaxial leaf side is more effectively
absorbed than the same light incident from the abaxial leaf side. The air in intercellular spaces of natural leaf
increases about twice the beam path and strongly deepens the non-linearity of the absorption process when
compared with water inltrated leaf. The chloroplast arrangement in the palisade cells is reected in most of
the studied differences. The leaf absorbance changed in our case of tobacco leaves up to 30% when the
chloroplasts moved from the face to the side position. This change depends strongly on the wavelength and
quite slightly on the character of incident light. Further analysis predicts that in practice the effect of chloroplast
rearrangement on the reectance spectra is in dependence on the wavelength of the light about 25% in our case
of fully developed green leaves but can be higher in some cases. Thus it can affect values of some of the indices
used in the remote sensing.
2015 Elsevier Inc. All rights reserved.

1. Introduction
A green leaf of higher plants can, to some extent, react and adapt to
the changing light conditions. This reaction is aimed to optimize the absorption of light quanta both to maximize the energy utilization in photosynthesis and to avoid harmful over-excitation. The process of this
adaptation is rather complex occurring in different time-scales and
may cover a change in the leaf position, leaf structure (e.g., swelling),
chloroplast movement, pigment composition, the state transitions or
energy quenching. We concentrate in this paper on the effect of chloroplast arrangement on the optical properties of leaves.
The ability of leaves to adapt its inner chloroplast arrangement to the
light conditions also inuences the resulting optical properties of leaves,
i.e., the leaf spectra of reectance, transmittance, absorptance and uorescence. The leaf reectance is usually the basis of the image detection
of a canopy by the spectral or multispectral cameras, both in laboratory
and in remote sensing studies. This implies that the detected spectral
image can be also inuenced by the chloroplast arrangement in leaf
Corresponding author.
E-mail address: naus@prfnw.upol.cz (J. Nau).

http://dx.doi.org/10.1016/j.rse.2015.12.011
0034-4257/ 2015 Elsevier Inc. All rights reserved.

cells without changes in leaf structure or chlorophyll content. It can be


expected that even the light environment within the canopy (light spectrum, geometry, intensity) can be inuenced by the chloroplast arrangement in the individual leaves and leaf layers. This indicates that
the knowledge of this phenomenon, including its theoretical description, should be taken into account in the remote sensing studies.
Our work aims to attract the attention of researchers to the changes
in optical spectra of leaves due to changes in chloroplast arrangement in
leaf cells. We suggest a simple method of mathematical description and
evaluation of these changes. Our work is based on previous original experience showing the effect of chloroplast movement in the signal of
chlorophyll uorescence or in the readings of a chlorophyll meter
(Brugnoli & Bjrkman, 1992; Nau, Prokopov, ebek, & pundov,
2010). Our study is restricted to the situation on a level of one leaf.
The upscale transfer to the level of whole plant or canopy is a matter
of further research.
1.1. Optical parameters
By optical parameters we mean here the spectra of leaf reectance,
transmittance, absorptance and uorescence. From this point of view,

182

B. Barnkov et al. / Remote Sensing of Environment 174 (2016) 181196

the most important physical quantity is the leaf absorptance (Gates,


Keegan, Schleter, & Weidner, 1965) showing that part of the incident
light which is really absorbed in the leaf tissue (e.g., Davis, Caylor,
Whippo, & Hangarter, 2011). To get a reliable information about the
leaf absorptance, the method of integration sphere should be used and
both reectance and transmittance spectra should be measured from
both sides of the leaves (see, e.g., Davis et al., 2011; DeLucia, Shenoi,
Naidu, & Day, 1991; Hlavinka, Nau, & pundov, 2013).
The ability to measure the optical parameters from a distance, remote sensing, has received extensive study. Much of this research has
focused on indices such as NDVI (Gamon, Peuelas, & Field, 1992;
Rouse, Haas, Schell, & Deering, 1974; Tucker, 1978), calculated from reectance measurements. Transmittance measurement has been also included (Bergstrsser et al., 2015; Daughtry & Walthall, 1998; Mariotti,
Ercoli, & Masoni, 1996) but it is not so common.
1.2. Natural versus inltrated leaf
Leaves contain in their structure intercellular spaces lled with the
air. These spaces cause most of the complexity of the measured spectra
as the difference in refractive index of cell walls and cytoplasm on one
side and the air on the other side causes light refraction and reection.
Including also light diffraction, scattering (Rayleigh and Mie scattering),

and interference (Steinhardt & Fukshansky, 1987), the complex


phenomenon describes light propagation and is sometimes generally
referred to as scattering.
In order to simplify the optical properties and to obtain some extreme parameters and spectra, the leaves may be vacuum inltrated
with water or suitable solution or oil. The obtained spectra may be a
basis for a discussion on the role of the intercellular spaces in the optical
processes (DeLucia, Nelson, Vogelmann, & Smith, 1996).
1.3. Adaxial versus abaxial leaf side
In many bifacial leaves the leaf structure visualized by microscopic
cross section differs in the parts adjacent to the adaxial (upper) and abaxial (lower) leaf sides (see, e.g., Smith, Vogelmann, DeLucia, Bell, &
Shepherd, 1997). The mesophyll cells at the adaxial leaf side are of palisade character whereas those at the abaxial leaf side are of spongy
structure, see Fig. 1. These differences are reected in different spectra
of diffuse reectance (see, e.g., DeLucia et al., 1991; Hlavinka et al.,
2013). The reectance of a leaf can be divided into the external
one, Re, and internal one, Ri (Hlavinka et al., 2013; McClendon &
Fukshansky, 1990a,b). The adaxial and abaxial leaf sides differ both in
the Re and in Ri. These differences may be attributed to the structural
differences of palisade and spongy parenchyma. It should be noted

Fig. 1. Images of tobacco leaf parts from optical microscope using transmission (C) or uorescence (A, B, D) modes. A and B show side (A) and face (B) chloroplast positions as seen from
adaxial leaf side. C and D show leaf cross section with different architectures in the adaxial (top) and abaxial (bottom) leaf sides. The lighter dots in A, B, and D show uorescing
chloroplasts. The bar indicates 20 m.

B. Barnkov et al. / Remote Sensing of Environment 174 (2016) 181196

that the palisade parenchyma cells enable deeper penetration of the collimated light (Vogelmann & Martin, 1993). On the other hand, the
spongy parenchyma may be more suited for absorption of the diffuse
light (Lee, Bone, Tarsis, & Storch, 1990). This leaf side is directed to the
soil or lower layers of the canopy and the incident light is mostly the diffuse one.
When evaluating the external reection Re, the Fresnel formulas
should be accordingly applied (see, e.g., Gerber et al., 2011; Gorton,
Craig, Brodersen, Williams, & Vogelmann, 2010; see the Supplementary
data). In this case, however, it should be mentioned, that the Fresnel formulas must be used in correct way to express the detected light intensity (see, e.g., Lucarini, Saarinen, Peiponen, & Vartiainen, 2005; Suila
& Nau, 2007).

183

The light incident on the leaf surface may be of different spectral and
geometrical character and different polarization. The geometrical character of the incident light may change from a perpendicular collimated
light (a beam of parallel rays) to a completely diffuse light. The collimated light may make a different angle with the leaf normal, i.e., it may
come from different directions. Only several studies have concentrated
their attention on the comparison of leaf optical properties detected in
collimated and diffuse light (e.g. Brodersen & Vogelmann, 2010; Davis
et al., 2011; Gorton et al., 2010; Williams, Gorton, & Witiak, 2003).
Bird and Riordan (1986) found out that on a clear day light radiation
consists of 85% of direct (collimated) light and 15% of diffuse light. Diffuse light component increases with higher clouds presence and denser
vegetation dispersing the light. Experiments in low-intensity light
showed that in these conditions there is no type of radiation preferred.
On the other hand, photosynthesis was 15% higher in plants grown at
high-intensity direct light when compared to corresponding absorbed
amount of diffuse light (Brodersen & Vogelmann, 2010). These results
indicate that a leaf adapts both anatomically and biochemically to
light intensity during its development and thus it handles different absorbing properties to direct and diffuse light reaching it (Brodersen &
Vogelmann, 2010).

Therefore the main method currently employed for the detection of


chloroplast movement is the measurement of red light transmittance
through a leaf (Wada & Kong, 2011). The light transmittance decreases
when chloroplasts gather to the face position. On the contrary, the light
transmittance increases when chloroplasts move out of the face position
into the side position (Wada, 2013; Wada & Kong, 2011). Difference between the two positions is clearly visible to the naked eye (Davis et al.,
2011; Kong & Wada, 2014) because of the change in light capture. The
position of chloroplasts inuences the light capture decline (light
absorption and scattering) due to changes in size of package effect and
detour effect (see, e.g., Terashima, Fujita, Inoue, Chow, & Oguchi,
2009). Package (grouping) effect is a phenomenon when ray goes beyond chloroplasts and is not absorbed by pigments which results in increase of maximum of spectrum transmittance (Merzlyak, Chivkunova,
Zhigalova, & Naqvi, 2009). The grouping of the chloroplasts causes that
leaf tissue behaves like a sieve for incoming light and therefore it is also
called as the sieve effect (Davis et al., 2011). Detour effect, known as optical path lengthening, is phenomenon when the ray path is lengthening
due to internal reections or scattering (Lee et al., 1990; Davis et al.,
2011; Kirk, 1976; Osborne & Raven, 1986; Terashima et al., 2009). Nevertheless, measuring of the light transmittance does not show actual
chloroplast movement but just the optical result of such movement
(Wada, 2013).
There are numerous studies both on the molecular nature of the
movement and the methods of measurement (for review see,
e.g., Wada, 2013). Also a sensitive method of collimated transmittance has been used to follow some effects of plant stress on chloroplast movement (e.g. Frolec, ebek, Lazr, & Nau, 2010; Nau et al.,
2010). The chloroplast movement can even inuence the reading of
some chlorophyll-meters (Nau et al., 2010) or the chlorophyllmeters can be, on the contrary, used to detect the chloroplast movement (Xiong et al., 2015). This movement can take place in different
directions and in different rates and can strongly vary among plant
species (Kniger, 2014; Kniger & Bolliger, 2012). The variation
may be caused by different size and shape of chloroplasts (and
even of the whole cells) in rst layer of palisade mesophyll cells
(Davis et al., 2011).

1.5. Chloroplast movement

1.6. The effect of chlorophyll uorescence

Every photosynthetic leaf cell contains usually dozens of chloroplasts situated between a central vacuole and a cell wall. Chloroplast arrangement is not stable, though, they respond to illumination reaching
the leaf surface. Chloroplast distribution thus depends on light conditions affecting the plant, i.e., on light intensity, direction, polarization
and spectral composition of incident illumination (Brugnoli &
Bjrkman, 1992; Davis et al., 2011; Kasahara et al., 2002; Nau et al.,
2010; Park, Chow, & Anderson, 1996; Vogelmann & Gorton, 2014;
Wada, 2013; Zurzycki, 1961).
Many papers have been devoted to description and signicance of
the chloroplast movement (e.g. Bana, Aggarwal, abuz, Sztatelman, &
Gabry, 2012; Brugnoli & Bjrkman, 1992; Davis et al., 2011; Kasahara
et al., 2002; Kniger, 2014; Kniger & Bolliger, 2012, Nau et al., 2010;
Park et al., 1996; Wada, 2013). There are two extreme types of the
movement. In case of high light intensity, the chloroplasts undergo the
so called avoidance movement placing the chloroplasts on the sides
(the side position) and let the light penetrate deeper into the leaf tissue,
see Fig. 1. On the other hand, in the case of low light intensity, the chloroplasts move to the periclinal cell walls (the face position) to enlarge
the light absorption (Bana et al., 2012; Gabry, 2004; Wada & Kong,
2011), see Fig. 1. This movement is in most cases in higher plants
evoked by the blue part of the incident light spectrum (Bana et al.,
2012; Gabry, 2004).
As a consequence of chloroplast movement, changes in optical leaf
properties occur. (Brugnoli & Bjrkman, 1992; Davis et al., 2011;
Kasahara et al., 2002; Nau et al., 2010; Park et al., 1996; Wada, 2013).

Chlorophyll uorescence may slightly change the reectance or


transmittance spectra of a leaf in the red and near infrared region
(Lork & Fukshansky, 1985), particularly when white light excitation is
used. This effect has been also modeled by the Monte Carlo method by
Suila and Nau (2007). However, generally this effect is not considered
or it is thought to be negligible (e.g., Gorton et al., 2010).

1.4. Collimated versus diffuse incident light

1.7. Theories of the optical properties


Many mathematical models dealing with plant leaves optical properties have been developed so far, the models being different in used approaches. As it is beyond the goal of this work to describe in details all
the models, only a short summary follows. Probably the rst model for
plant leaf optical properties was the model by Allen and Richardson
(1968) which was based on calculation of forward and backward light
uxes inside a leaf considering the leaf as a single homogeneous slab
(plate) inside which absorption and scattering occur, in analogy with
KubelkaMunk theory (Kubelka & Munk, 1931). The plate model
(Allen, Gausmann, Richardson, & Thomas, 1969) improved the previous
Allen's model by considering also optical effects on the upper surface
of the plate. The generalized plate model (Allen, Gausmann, &
Richardson, 1970) considered the leaf being composed from several
plates (layers). A multilayer concept was then used in several other
models (e.g. Seyfried & Fukshansky, 1983; Terashima & Saeki,
1985) and a concept of specic absorbance, considering one or two
layers of a leaf, was used by Hlavinka et al. (2013). The PROSPECT

184

B. Barnkov et al. / Remote Sensing of Environment 174 (2016) 181196

model (Jacquemoud & Baret, 1990; Jacquemoud et al., 1996;


Jacquemoud, Bacour, Poilv, & Frangi, 1999) is based on the generalized
plate model. It should be noted that the Allen's models and PROSPECT
model are based on the Stokes formulas (Stokes, 1860-1862). Different
approaches were used in RAYTRAN model (Govaerts, Jacquemoud,
Verstraete, & Ustin, 1996; Ustin, Jacquemoud, & Govaerts, 2001), ABM
models (Baranoski, 2006; Baranoski & Rokne, 1997), LEAFMOD model
(Ganapol, Johnson, Hammer, Hlavka, & Peterson, 1998), LIBERTY
model (Barton, 2001; Dawson, Curran, & Plummer, 1998), and SLOP
model (Maier, Ldeker, & Gnther, 1999). In addition to transmittance
and reectance calculation in all the above mentioned models, a model
by Fukshansky and Kazarinova (1980), a Monte-Carlo ray tracing model
by Suila and Nau (2007), and FluorMODleaf model (Pedrs, Goulas,
Jacquemoud, Louis, & Moya, 2010) considered also chlorophyll uorescence emission. The leaf models mentioned above are also incorporated into more general models used in remote sensing (e.g.
Jacquemoud et al., 2009; Malenovsk et al., 2008; van der Tol,
Verhoef, Timmermans, Verhoef, & Su, 2009; Verhoef & Bach, 2007;
Verrelst et al., 2015). Regardless of a very high complexity of most
of the above mentioned models, some of them are intrinsically
based on what we call below (see Section 3.4) as linear approximation. Many models are implicitly non-linear. The model presented
in this work is very simple but it is explicitly non-linear, realizing
thus a distortion from the LambertBeer's law which is observed in
our experimental data.
In this paper we have measured a complex set of optical properties
of a green tobacco leaf under different conditions with the main goal
to detect the inuence of chloroplast positioning. The spectra have
been measured for the face and side chloroplast position. The leaf absorptance was evaluated for both the adaxial and abaxial sides, under
both collimated and diffuse illumination, in natural state and after
water inltration. A quantity of leaf absorbance was calculated as a decimal logarithm of one minus absorptance. A concept of nonlinear deviation from the LambertBeer's law in a quadratic approximation was
used to explain the changes in the leaf absorbance. Both the evaluation
of relative absorbance change (named difference sensitivity) and theoretical simulation supported the acceptability of the non-linear concept.

below). The disks were homogenized in 80% acetone with a low amount
of MgCO3 and then centrifuged (5 min at 3600 g). Absorption spectra of
the supernatant were measured with spectrophotometer Unicam
UV550 (Spectronic, Cambridge, UK). Concentrations of chlorophylls
(a + b) and of carotenoids were determined to be 25.9
3.42 g cm 2 and 6.5 0.70 g cm 2 (mean standard deviation,
n = 10), respectively. In such a way determined concentration of chlorophylls well agreed with measured SPAD-values (see below) as it was
previously calibrated for tobacco leaves by Nau et al. (2010).
2.3. Measurement of leaf optical properties
Perpendicular collimated and diffuse light were used to illuminate
the leaf. The portable spectroradiometer LI-1800 (LI-COR, Lincoln,
Nebraska, USA) with an integrating sphere LI-1800-12 were used for
measurements of diffuse reectance (RD) and transmittance (TD).
White light of illuminator was used for the measurement.
2.4. Collimated perpendicular light illumination
The collimated light was produced by the illuminator. The signal of
the sample was compared to the standard (LI-COR part # 991,804); a
small plate covered by puried barium sulfate which optimally reects
the light in a diffuse manner (diffuse reectance approaches 1) was
used as the standard. If the signal of the sample is Sv() and the standard
signal is Ss(), the nal signal RD() is:
RD

Sv
:
Ss

TD() is calculated using the following formula, where the standard signal is Ss() and the sample signal (the illuminator is behind
the sample) is STv():
T D

STv
Ss

2. Materials and methods

2.5. Diffuse light illumination

2.1. Plant material

The procedure of measurement was in fact similar to that used by


(Gorton et al., 2010).

Plants of Virgin tobacco (Nicotiana tabacum L. cv. Samsun) were


seeded on 26th June, 2014 and planted in plastic pots with gardening
soil. The plants were cultivated in a phytochamber Weiss-Gallenkamp
SGC.170.PFX.J (Weiss-Gallenkamp, Loughborough, England) at a regime
of 10 h of darkness (temperature 18 C, relative air humidity 55%) and
14 h of uorescent light (temperature 25 C, relative humidity 45%) at
150 mol photons m2 s1 of PAR at the upper leaves level. From 6 to
7 month old plants in their vegetative phase having from 18 to 22 leaves
were used for the measurement. Young mature leaves were used for the
measurements. They were usually in the position 47 from apex. The
chlorophyll-meter SPAD-502DL (Konica Minolta Sensing Inc., Osaka,
Japan) was used for measurements of the chlorophyll content. The
leaves detached at the petiole base were placed in a double mask
made of a ltering paper with circular openings of 15 mm in diameter
dening precisely the measured area. The mask ensured that the same
leaf part was measured on both adaxial and abaxial sides. The petiole
of the leaf was immersed in a small tube with water.
2.2. Concentration of pigments
Concentration of pigments was determined spectroscopically according to the method of Lichtenthaler (1987). To do that, disks of
10 mm diameter were cut from the leaves which were previously
used for measurement of relative content of chlorophylls by SPAD (see

2.5.1. Diffuse reectance


The measurement was performed with the integration sphere LI1800-12. The light from the illuminator was incident on the reference
standard white plate so that the sample in parallel position with respect
to the illuminator beam was illuminated by diffuse light. The sample
was replaced by an ideal white diffuse plate (Spekol Rd./0, Carl Zeiss,
Jena, Germany) and a spectrum SRW() was detected. Then the same
measurement was done with no sample in place (the black hole arrangement), the detected spectrum was designated as SRb(). Then a
sample (leaf) was placed in the place and a spectrum SRS() was detected. The spectrum of diffuse reectance RDD() upon diffuse illumination
was calculated according to:
RDD

SRS SRb
:
SRW SRb

2.5.2. Diffuse transmittance


To obtain the incident diffuse light a second integrating sphere was
used. We have adapted the integration sphere of Specol Rd./0 (Carl
Zeiss, Jena, Germany). The light from the illuminator fell on the opposite
white inner surface in a direction parallel to the sample surface. No
direct light from the illuminator could reach the sample. The white

B. Barnkov et al. / Remote Sensing of Environment 174 (2016) 181196

inner reective surface of the sphere distributed the light rays randomly in the sphere. As a consequence, diffuse light was incident
on the sample.
The two spheres were attached together by the openings and the
sample was placed between them. When there was no sample, the detected signal STW() was proportional to the incident diffuse light. After
insertion of a leaf a spectrum STS() was measured. The diffuse transmittance under diffuse illumination, TDD(), was then calculated as:
T DD

STS
:
STW

2.6. Leaf inltration


When the leaf is inltrated with water, the difference of the refractive indexes of cytoplasm/intercellular compartment is signicantly
lowered. Consequently the optical phenomena, such as reection and
refraction on the inside structure of the leaf are limited. The inltrated
leaf was also kept wet on the surface so that there was a partial alignment of the surface irregularities and the surface became more at
with a dened refraction index of water.
The leaf inltration was carried out by an inltration setup; a thickwall Erlenmeyer ask closed with a rubber stopper and attached to a rotational vacuum pump (type RV 1, 5/11, LP, Prague, Czech Republic)
with two valves, for sucking and lling.
The leaf segment was placed in a test tube, immersed completely in
deionized water. The inltration was periodically performed under low
air pressure of about 0.1 of normal pressure (controlled by a manometer). The inltration was observed visually. A well inltrated leaf was almost transparent showing that the intercellular space was lled with
water.
2.7. Induction of chloroplast movement
The avoidance chloroplast movement from the face to the side position in palisade cells was induced by collimated strong blue light
( ~ 400 nm, 563 mol photons m2 s1) generated by KL 2500 LCD
cold light source (Schott Glas, Wiesbaden, Germany) in combination
with BG12 blue glass lter (Schott Glas, Wiesbaden, Germany) and
some lens. The intensity of light was measured by quantum radiometer
LI-189 (LI-COR, USA). The illumination lasted 1 h and the adaxial leaf
side was always illuminated in direction perpendicular to the leaf surface. For the whole time of chloroplast movement we measured also
leaf surface temperature, to check that the leaf is not overheated and
thus damaged. After the light exposition, another SPAD-value measurement took place, this time to control that chloroplasts moved to the side
position. The SPAD-value decreased up to 30% (see also Nau et al.,
2010; Xiong et al., 2015). The reverse chloroplast movement, from the
side to the face position, was induced by 1-h illumination of the leaf
by the blue light of a weak intensity (12.4 mol photons m2 s1)
using the same illuminating system. The movement of chloroplasts to
their face positions was checked at the end of illumination by SPAD
measurement which value came back to the original value. The fact
that the SPAD-value reversibly returned to its original value conrms
that the measured changes were caused by the chloroplast movement
and not by some other processes, e.g., changes in chlorophyll content.
Chloroplast arrangement in both positions could be demonstrated by
microscopy (see below) as shown in Fig. 1. Except the blue light, no
other light was present in laboratory during the chloroplast movement.
2.8. Methodology of measurement
All measurements and light expositions were done at room temperature. In order to enable measurement of all eight variants with
one leaf and thus to ensure comparison of the results, a methodology

185

had to be developed; prior to the measurement, the tobacco plant


was placed into laboratory with a low light intensity (white uorescent light, 58 mol photons m 2 s 1) for 30 min to ensure chloroplasts being in the face position. Then we separated one leaf from the
plant, placed it in a mask and measured SPAD-value. The leaf optical
spectra and parameters of chloroplast movement depend on chlorophyll leaf area content. To have about the same chlorophyll area concentration we have chosen leaves with very similar SPAD-values of
39 4.
The measurement was then divided into four sequences. In the
rst two sequences we measured optical properties of natural leaf
and in the third and fourth sequences we used an inltrated form
of the leaf. In the rst sequence, with chloroplasts in the face position, we measured diffuse light reectance and transmittance on
both adaxial and abaxial leaf sides rst with diffuse incident light
on the sample and then with the collimated perpendicular incident
light. In the end of the sequence we measured SPAD-value again to
make sure that chloroplasts stayed in the same position during
measurements.
After nishing the rst sequence, we placed the leaf in a mask to induce the avoidance movement. When the movement was nished, we
measured the second sequence with the identical procedure as the
rst one described above.
Before measurements of the third sequence, we induced the reverse
chloroplast movement back to their face positions and subsequently we
inltrated the leaf with deionized water. The remaining third and fourth
sequences of measurement were conducted in the identical way as in
the case of the natural leaf described above.
Using the methodology described above, we obtained from one
leaf 24 = 16 pairs of spectra of diffuse light reectance and transmittance, each pair for different state of the sample (natural inltrated), different chloroplast arrangements (face side), different light
incident on the sample (diffuse collimated), and different leaf
side (adaxial abaxial). All the curves presented below in Figs. 3-6
are average curves from 3 to 5 measurements.
2.9. Microscopy
The side and face chloroplast arrangements were induced in the inltrated leaf as described above. The native leaf segment was imbedded
into parafn block and thick leaf cross sections were prepared using a
sledge microtome (Meopta, Czech Republic). Fluorescence microscope
Hund H600 AFL 50/100 (Wetzlar, Germany) with lter block D for incident uorescence light was used and the photographs were taken by
Samsung Galaxy A3 (Suwon, South Korea) mobile telephone camera
(8 Mpx). A mercury discharge lamp was used for uorescence
excitation.
3. Theory
3.1. Absorbance of a leaf
A suitable physical quantity expressing the amount of the incident
light absorbed by the leaf is the diffuse absorptance. This quantity is
dened as:
A 1T D RD ;

where TD() and RD() are diffuse leaf transmittance and diffuse
leaf reectance, respectively. The word diffuse means that an integration sphere is used for the detection of reected or transmitted
light. Under such approach, all the light being reected to the hemisphere in front of leaf and transmitted to the hemisphere behind the
leaf is detected. The only error may come into consideration if some
light leaks through the leaf tissue out of the detection sphere.

186

B. Barnkov et al. / Remote Sensing of Environment 174 (2016) 181196

However, the quantity absorptance A() is not very suitable for


some theoretical considerations. From this reason we use in this paper
a concept of leaf absorbance dened as:
D log1A logT D RD 

3.3. Differential sensitivity


To gain an enlarged difference in the compared spectra, we use the
term sensitivity (see also Carter, 1991) which is equal to the concept
of relative difference. It is dened as:
S

It is important to stress here that employing such a denition, both


rays transmitted through the leaf and the rays returning from the leaf
to the hemisphere of incident light are taken together.
3.2. Spectrophotometry
The below presented description and discussion should substantiate more clearly the concept of leaf absorbance. The above denition of absorbance corresponds to the practically and generally
used LambertBeer's absorption law in spectrophotometry of diluted solutions using collimated (or nearly collimated) light. For the
transmittance T(,x) of a sample the following expression is used
in spectrophotometry:
T ; x

I; x
10D :
I 0; x

where () is the molar absorption coefcient (units:


l mol1 cm1), C is concentration (units: mol l1) and x is the length
of geometrical beam path, i.e., the thickness of the measuring cuvette
(usually 1 cm). The measurements are usually realized in a twocuvette or two-beam systems where the effect of the cuvette reection
and solvent scattering, reection and absorption are thought to be
eliminated.
If we analyze the analogy between the measurements with the
integration sphere and in a spectrophotometer considering the absorbance and absorptance terms, some interesting conclusions can
be drawn; in the spectrophotometric arrangement (e.g., a cuvette
with a diluted pigment solution) the absorbed light can be expressed
also as:
A 1T :

However, this expression is valid only if the inner reection Ri() of


the sample is neglected. To elucidate this statement, we express the
sample reection R() as composed of the surface (external) Re()
and internal Ri() reection. That means:
R Re Ri :

10

Because the external reection of the cuvette Re() (in fact including
also the reection on the second cuvette wall) is mostly corrected by
using the reference cuvette with a solvent, the expression for the absorptance closer to reality should be:
A 1T Ri :

12

where DR() is a reference spectrum and D() the measured (calculated) spectrum after the change in light or leaf structural conditions. The
quantity S() means a change of the spectrum in relation to the original
(reference) one. In fact, there are two possibilities how to express the relation of two absorbances: i) One possibility is to use a known or supposed common reference spectrum obtained under nearly ideal
conditions, DR0(). The spectrum of diluted suspension of broken chloroplasts can serve for this purpose. This approach is suitable for the theoretical simulation of the studied effects. ii) The reference spectrum DR() is
that one of a pair of spectra being more similar to the ideal one. This sensitivity we will designate as the differential sensitivity Sd(). This Sd()
sensitivity is a more practical one, can be easily obtained, and was also
used by us (see Figs. 3-6). As the reference spectra we considered the
spectra obtained for inltrated leaf (Fig. 3), for collimated light (Fig. 4),
for abaxial leaf side (Fig. 5), and for face chloroplast position (Fig. 6).
3.4. Linear approximation

Here the absorbance is expressed by LambertBeer's law as:


D C x;

DDR
D

1;
DR
DR

11

Therefore, the Eq. (9) holds only for samples with negligible inner
reections. Generally, the analogy or differences between the method using integrating sphere and the spectrophotometric method
would desire a deeper analysis. However, this is not the goal of this
work.

Under the linear approximation we mean a validity of the linear expression for the absorbances in LambertBeer's law:
D C x;

13

DR R C R xR :

14

The differential sensitivity is in this case:


Sd

C x
1:
R C R xR

15

If we assume that there is no change in mean leaf pigment concentration (C = CR) and in the basic spectrum (() = R()), the differential sensitivity does not depend on the wavelength and it has a constant value:
Sd

x
1:
xR

16

Ideally, for the case when there is also no change between the length
path for the explored conditions and for the reference, the differential
sensitivity is zero upon the linear approximation.
3.5. Nonlinear approximation
However, the experimental data (Figs. 3-6) show that the differential sensitivity Sd() is not a constant, indicating that the linear approximation of the LambertBeer's law is not valid for leave tissues. The
deviation from the absorption linear law has been already studied and
considered (see, e.g., Lee et al., 1990; McClendon & Fukshansky,
1990a,b and references therein). We have revealed that the Sd() spectra bear inverse features of the reference spectra. This fact can be simulated by some kind of deviation from the linearity of the leaf absorbance.
The deviation can be expressed in the simplest way by introduction of a
weighted quadratic term in the expression of D():
D a C x bC x2 ;

17

where a and b are parameters reecting the linear and nonlinear


contribution, respectively, to D(). In the case of differential sensitivity,

B. Barnkov et al. / Remote Sensing of Environment 174 (2016) 181196

for simplicity we assume that the reference absorbance has the linear
form as expressed by Eq. (14). It should be noted that any other more
complicated function describing well the experimental data could be
used instead of the quadratic function but we have chosen the quadratic
function as the easiest possibility. Moreover, the agreement between
theoretical model and experimental data is a support for our approach.
Let us suppose again that there is no change in mean leaf pigment
concentration (C = CR) and in the basic spectrum (() = R()) and
that the main cause of the spectrum change consists from a change of
the mean path length x. Then the differential sensitivity is:
Sd a

R C R x
R C R x2
b
1
R C R xR
R C R xR

18

and consequently
Sd a

x
R C R x2
b
1 a r b DR r2 1;
xR
xR

19

where
r

x
:
xR

20

We want to point out that the non-linearity of our model is not only
in x as might be concluded from Eq. (19), but explicitly also in () and C
(see Eq. (17)). We call the above mentioned Simple Explicitly NonLinear Empirical model for Leaf Optical Properties as SENLELOP model,
even if it is not a real structure-function-based model but a theoretical
description of experimental data. To our knowledge, the SENLELOP
model is the rst model which explicitly considers non-linearity, opposite to all published models so far (see Section 1.7). As our model well
describes experimental data (Figs. 3-6), the non-linearity should be
considered in more general models used in the remote sensing.
Although in a rather simplied way, we can now describe and discuss the differential sensitivity of selected pairs in terms of three parameters, basic weight of the linear term a, non-linear deviation b and the
relative change in the beam path r. However, we should rst relate
our r and b parameters to the known effect occurring in the leaves;
the sieve and detour effects (see Section 1.5). Even if we have to interpret our r and b parameters with some caution as their meaning is rather complex, in the rst approximation, we suggest to understand the r
parameter as reecting the detour effect (lengthening of the beam
path) and the b parameter as reecting the sieve effect. The suggested
meaning of the r and b parameters makes our non-linear approximation
more realistic.
3.6. Fitting the model parameters to experimental curves
Fitting of the parameters (a, r, b) of the model (Eq. (19); see below)
to experimental Sd() curves was performed using the solver function
of Microsoft Excel by estimating the parameters so that the criterion
function 2 is minimized. 2 is dened as a sum of squares of the difference between experimental Sd() and theoretical Sd() for given lambda. For a simplicity we assumed that a = 1 in all the cases and we tted
only values of b and r parameters.
3.7. Contribution of chlorophyll uorescence

It is generally well known that green leaves illuminated in visible region


emit uorescence. We will discuss here only the effect of chlorophyll uorescence, although other types of luminescence can be present (see,
e.g., Buschmann, 2007). The chlorophyll uorescence spectrum of a leaf
at room temperature is ascribed to chlorophyll a of the inner antenna of
photosystem II and consists of two main bands the maxima of which

187

are approximately situated at 685 nm and 740 nm (see, e.g., Govindjee,


1995; Mader et al., 1982). Both the positions of the band maxima and
mostly the band ratio depend on many factors, both the passive ones
coming from the leaf structure and pigment concentration and active
ones reecting the adjustments of photosynthetic apparatus. The most
important passive effects seems to be the effect of chlorophyll
reabsorptions (see, e.g., Bartokov, Nau, & Vkruta, 1999; Nau,
Klinkovsk, Ilk, & Ciknek, 1994; Suila & Nau, 2007 and references
therein). On the contrary, the active change of the emission spectrum
can be demonstrated by changes in the uorescence band ratio during
the uorescence induction (Buschmann, 2007; Palombi et al., 2011).
It has been already shown in theoretical studies (e.g. Lork &
Fukshansky, 1985; Suila & Nau, 2007) that chlorophyll uorescence
affects to some extent the shape of the R() and T() spectra of leaves
in the red and near infrared region. These theoretical simulations
serve to roughly estimate the effect of uorescence in general. In experiments, real measured uorescence and optical spectra are used to estimate and simulate the extent of the spectral changes caused by
uorescence. The uorescence contribution is demonstrated as an increase in RT() and TT().
The evaluation of the effect of chlorophyll uorescence is quite extensive and we will not realize this evaluation in this work. We just
mention that one visible consequence of the presence of chlorophyll
uorescence is a visible local maximum at about 750 nm in the reectance spectrum (see Fig. 2). Similarly as in other publications
(e.g., Gorton et al., 2010) we will expect the inuence of chlorophyll
uorescence to the conclusions of our results as not substantial.
4. Results and discussion
Two typical pairs of diffuse light reectance (R()) and transmittance (T()) spectra obtained with natural leaf and the same leaf after
inltration, with chloroplasts in their face positions, illuminated with
collimated light on both the adaxial and abaxial leaf sides are shown
in Fig. 2a and b, respectively. Fig. 2c shows absorptances A() calculated
from the reectances and transmittances and Fig. 2d shows absorbances
D() calculated from the absorptances. All the curves are shown as they
were measured (Fig. 2a, b), i.e., from 380 to 800 nm and these curves in
fact represent a case with reliable R() and T() values and low noise.
However, in some cases, there was a high noise of the signal for the
shorter wavelengths and therefore the curves (spectra) shown in the
following gures are presented from 450 nm. Moreover, the reectance
curves (Fig. 2a) show a maximum at about 750 nm which is caused by
chlorophyll uorescence (see Section 3.7) and this effect is also manifested as a decline of absorptance slightly below zero (Fig. 2c) at this
wavelength. Therefore the curves (spectra) shown in the following gures are presented to 700 nm. The effect of chlorophyll uorescence will
be analyzed in another work.
Here, we would like to mention an interesting fact used in further
discussion. In the near infrared region (above 750 nm) the reectance
of natural leaf from the abaxial side is lower than that from the adaxial
side (Fig. 2a, gray and green symbols, respectively). Consequently, the
relation of transmittances in this region (Fig. 2b) is opposite. The situation in the inltrated leaf is in slightly inverse manner (Fig. 2a, blue and
red symbols, respectively).
To make the description of particular measurement variants shorter,
we use four-letter code in the following text. The rst letter describes if
the chloroplasts were in the Side or Face position, second letter describes if aDaxail or aBaxail leaf side was illuminated, the third letter describes if Collimated or diffUse light was used and the fourth letter
describes if Natural or Inltrated state of the leaf was used. Hence, the
measured curves can be described symbolically as, e.g., FDCN and FDCI
for curves (spectra) presented respectively by green and red symbols
in Fig. 2, which means that chloroplasts were in the face position, adaxial leaf side was illuminated, collimated light was used for illumination
and the leaves were either natural or inltrated, respectively. The

188

B. Barnkov et al. / Remote Sensing of Environment 174 (2016) 181196

Fig. 2. Typical spectra of diffuse reectance R (panel a) and diffuse transmittance T (panel b) measured with natural or inltrated Virgin tobacco leaf with its chloroplasts in the face position
upon illumination of the abaxial or adaxial leaf side by collimated light. The spectra then served for calculation of the typical spectra of diffuse absorptance A (panel c) and consequently of
the leaf absorbance D (panel d). The thin vertical line at 750 nm in panels a and c indicates a position of a hump in diffuse reectance R (panel a) and of a hollow in diffuse absorptance A
(panel c) caused by chlorophyll uorescence (see the text), which causes diffuse absorptance slightly below zero value as indicated by a thin horizontal line in panel c.

relative difference (sensitivity) of the absorbance of natural versus inltrated leaf shown in Fig. 2 was calculated using absorbances D[FDCN]
and D[FDCI] by the expression:
Sd FDCN=I FDCN=I DFDCNDFDCI=DFDCI:

In Table 1, for the sake of simplicity, only the symbols of the spectral
pairs are shown. To elucidate the way of interpretation we use, the effect of lengthening r and of non-linearity b on the Sd() spectrum should
be shortly explained. In a strongly simplied approach, an increase of
lengthening parameter r moves the sensitivity curve (understood as a
mean line) up to higher values whereas an increase in absolute values
of b enlarges the difference between the minima and maxima of the

Table 1
Results of the tting of the model parameters (b and r) to experimental Sd spectra. See the text for meaning of denotation of the variants.
Figure

3a

3a

3b

3b

3c

3c

3d

3d

Variant

SDC(N/I)

FDC(N/I)

SBC(N/I)

FBC(N/I)

SDU(N/I)

FDU(N/I)

SBU(N/I)

FBU(N/I)

b=
r=

0.242
1.724

0.199
1.961

0.234
1.400

0.231
1.498

0.191
1.556

0.081
1.552

0.250
1.083

0.147
1.209

Figure

4a

4a

4b

4b

4c

4c

4d

Variant

SD(U/C)N

FD(U/C)N

SB(U/C)N

FB(U/C)N

SD(U/C)I

FD(U/C)I

SB(U/C)I

FB(U/C)I

b=
r=

0.114
1.216

0.069
1.134

0.166
1.150

0.106
1.171

0.155
1.348

0.250
1.404

0.182
1.421

0.210
1.470

Figure

5a

5a

5b

5b

5c

5c

5d

5d

Variant

S(D/B)CN

F(D/B)CN

S(D/B)CI

F(D/B)CI

S(D/B)UN

F(D/B)UN

S(D/B)UI

F(D/B)UI

b=
r=

0.115
1.300

0.107
1.432

0.023
1.006

0.034
1.042

0.096
1.376

0.044
1.298

0.114
1.005

0.004
0.980

Figure

6a

6a

6b

6b

6c

6c

6d

6d

Variant

(S/F)BCN

(S/F)DCN

(S/F)BCI

(S/F)DCI

(S/F)BUN

(S/F)DUN

(S/F)BUI

(S/F)DUI

b=
r=

0.130
0.957

0.118
0.903

0.217
1.065

0.166
1.003

0.181
0.951

0.159
0.988

0.101
0.989

0.066
0.981

4d

B. Barnkov et al. / Remote Sensing of Environment 174 (2016) 181196

189

Fig. 3. Experimental difference sensitivities Sd (symbols) showing the effect of leaf inltration where leaf absorbance of the inltrated leaf was considered as the reference. Panels a and b
show the Sd spectra for the collimated incident light whereas panels c and d show the Sd spectra for the diffuse incident light (see the left side captions). Panels a and c show the Sd spectra
for adaxial leaf side whereas panels b and d show the Sd spectra for the abaxial leaf sides (see the top captions). Red up- and blue down-triangles show the Sd spectra for face and side,
respectively, chloroplasts position whereas the black lines show the best t of the model (Eq. (19)) to the experimental spectra (see the legend inside the graph). Values of the model
parameters are summarized in Table 1. The range of Sd values is the same in all the panels. The error bars indicate standard deviations (SD) from 3 to 5 measurements and for the case
of clarity they are presented only as either +SD or SD.

Fig. 4. Experimental difference sensitivities Sd (symbols) showing the effect of geometrical character of the illuminating light where leaf absorbance upon collimated light was considered
as the reference. Meaning and description of the rest of the gure is in analogy with Fig. 3.

190

B. Barnkov et al. / Remote Sensing of Environment 174 (2016) 181196

Fig. 5. Experimental difference sensitivities Sd (symbols) showing the effect of leaf side where leaf absorbance of the abaxial leaf side was considered as the reference. Meaning and
description of the rest of the gure is in analogy with Fig. 3.

sensitivity curve making the red and blue minima deeper (the curve is
less at). However, the two parameters partly correlate and so the effects of the curve shift and deeper prole are not fully independent.

The correlation coefcient obtained using data of Table 1 was R =


0.426 (see Supplementary data).

Fig. 6. Experimental difference sensitivities Sd (symbols) showing the effect of chloroplast arrangement where leaf absorbance with the chloroplasts in their face positions was considered
as the reference. Meaning and description of the rest of the gure is in analogy with Fig. 3.

B. Barnkov et al. / Remote Sensing of Environment 174 (2016) 181196

It might seem that our presentation (Figs. 3-6) repeats the results
and that the curves (spectra) intuitively thought to be mutually related.
Even if some spectra might look similar, no spectrum is presented twice.
As for mutual relationship of the spectra, we have analyzed this question carefully. The analysis leads to a complex mathematical expression
showing that there is a mutual relation between curves, however it is
very difcult to envisage without the mathematics. For instance a particular Sd() curve can be derived from three other curves in a nonlinear relation.
Because we have explored the effect of chloroplast arrangement for
the natural and inltrated leaf illuminated by the diffuse or collimated
light from adaxial or abaxial leaf sides and because of used denition
of differential sensitivity Sd(), it is easier and more illustrative to describe and discuss rst the effects of leaf inltration (natural versus inltrated leaf), geometrical character of the illuminating light (diffuse
versus collimated illumination), and leaf side (adaxial versus abaxial
leaf side) without discrimination between chloroplast arrangement.
The latter will be described and discussed at the end in a separate part.
4.1. Effect of leaf inltration (natural versus inltrated leaf), the role of
intercellular spaces
The absorbance of the inltrated leaf was taken as the reference
absorbance (see Fig. 3). As can be expected, the inltration of a leaf by
a liquid with the refraction index near to that of cytoplasm would
strongly simplify the optical situation and hence will cause a large difference between D() and DR() consequently leading to a big deviation
of Sd() from a constant value (Fig. 3), resulting in a high absolute values
of the non-linearity (b) and the path lengthening (r). Both the b and r
values (Table 1, row 1) were the largest when comparing the natural
with the inltrated state of the leaf than in other related pairs
(Table 1). The r value was approaching 2 in case of collimated light
and adaxial leaf side (Fig. 3a, Table 1). The non-linearity parameter b
reached also the highest absolute value of about 0.24. If we assume for
a while that the beam path is near to the leaf width of the inltrated
leaf then our results (r approaching 2) are in agreement with those of
McClendon and Fukshansky (1990b) who have found that the beam
path in leaves with a high chlorophyll content is about 2 to 3 times longer that the leaf width. Therefore, this pair may serve for a rough experimental estimation of the beam path length in a natural leaf. We recall
here the meaning of our beam path length; the concept concerns all
rays leaving the leaf (both transmitted and reected).
Interestingly, the lengthening of beam path caused by the air spaces
and represented by the vertical shift of the Sd() curves, is lower when
the abaxial leaf side is illuminated as compared with the adaxial leaf
side both for collimated and diffuse light (compare Fig. 3 right and left
panels and Table 1, row 1). If we take into account the fact that the sensitivity means relative changes in absorbances, the above fact indicates
that some lengthening appears already in the inltrated state being
larger upon illumination of the abaxial leaf side then of the adaxial
leaf side. As it has been shown by Vogelmann and Martin (1993), the
palisade cells enable the beams to penetrate more easily into the leaf
structure. Therefore, it is tempting to expect a shortening of the beam
path when the leaves are illuminated from the adaxial as compared to
the abaxial leaf side. This may be true for the rays of the rst passage
(for explanation see Supplementary data), passing the leaf in transmittance mode and penetrating through the second boundary. However,
our absorbance comprises both rays transmitted by the leaf and those
returning back, which means also the rays of the second and following
passages. As it is shown in Fig. 2 in the near infrared region, the reectance at the adaxial leaf side is in a natural leaf larger than that of abaxial
leaf side; more rays return back at the adaxial leaf side. The fact that
there is a higher chlorophyll concentration at the adaxial side (e.g. Cui,
Vogelmann, & Smith, 1991; Lee et al., 1990) supports the results.
The above comparisons help us to understand the effect of air spaces
in the leaf structure. From the optical point of view, the intercellular air

191

spaces increase about twice the beam path and strongly deepen the
non-linearity of the absorption process.

4.2. Effect of geometrical character of the illuminating light (diffuse versus


collimated illumination)
We have compared the situation under the diffuse illumination to
the perpendicular collimated one. The absorbance in the collimated
light is taken as the reference one (see Fig. 4). As is shown by the homogeneous leaf model presented in the Appendix A, the diffuse illumination should lead to an increase of the mean beam path and thus to an
increase of the absorbance.

4.2.1. Inltrated leaf


The beam path has increased in our case in the inltrated leaf by
quite a high amount of 3547% (values of r were between 1.348 and
1.470, Table 1) in the diffuse light as compared with the collimated illumination. All the Sd() curves are consequently slightly shifted above
zero line (Fig. 4c, d), especially in case of the abaxial leaf side (Fig. 4d).
In discussion of the results obtained with the inltrated leaf, the theoretical model (see Appendix A) should include also the reection in
the second interface between the leaf and air or a multi-reection situation (see Supplementary data). It is predicted that the theoretical increase in the mean beam path after the rst passage may amount up
to a value around 1.2. This is too small to explain the situation in the
inltrated leaves. This indicates that the reection at the second
boundary should be considered and it is different from that of the
rst boundary and the parameter m, relating these two reections
(see expressions A12 and A13) is larger than 1. As an example, if
we assume R1 to be 0.1, to get the lengthening 1.4 in full multireection model, the m value should be around 3. This number is reasonable if we accept that the collimated light changes to the diffuse
one upon passage through the leaf. The mean reection of diffuse
light is several times larger (depending on the refraction index)
than the reection of perpendicular collimated light (see Supplementary data).

4.2.2. Natural leaf


The changes shown in comparison of leaf absorbances in diffuse light
versus collimated illumination in case of natural leaves (Fig. 4a and b)
are quite small. The r value was found to be in the interval 1.134
1.216 and the absolute value of the non-linearity parameter b ranged
from 0.069 to 0.166 (Table 1, row 2). The increase of the beam path
up to about 20% is predicted by the theoretical model (see Appendix
A), if we assume the surface reection governed by the Fresnel formulas
to participate in the rst interface between air and leaf and for the rst
beam passage only. However, in this case the real leaf does not fully correspond to the demanded at surface and homogeneous leaf interior
considered by the model and so the theoretical predictions should be
taken as only approximate. Nevertheless, the theoretical and experimental values were found to be quite near to each other being around
1.2.
There is also a slight increase in the non-linearity (see parameter b,
Table 1, row 2). This result indicates that the intercellular spaces lled
with air cause such scattering of the penetrating light that mostly eliminate the difference between illumination with the collimated and diffuse light rendering all incident light to be mostly of diffuse character
inside the leaf.
Both in the inltrated leaf and the natural leaf the diffuse light entering the interior of the leaf is more effectively absorbed than the collimated one. The absolute value of the non-linearity parameter b increases in
the diffuse light documenting a more effective absorption of the green
light. This effect is stronger in case of inltrated leaf (Table 1), also
due to a slight correlation between b and r parameters.

192

B. Barnkov et al. / Remote Sensing of Environment 174 (2016) 181196

It is of interest to mention that opposite results were found for sunower leaves by Gorton et al. (2010). The difference can be caused by
different architecture of sunower and tobacco leaves.
4.3. Effect of leaf side (adaxial versus abaxial leaf side), the optical polarity
of leaves
Next we have compared the optical properties of different sides of
the bifacial tobacco leaf. A different architecture of the two leaf sides is
shown in Fig. 1. The absorbance of the abaxial leaf side is taken as the
reference one (Fig. 5).
4.3.1. Inltrated leaf
The leaf inltration led to nearly negligible differences between optical parameters of adaxial and abaxial leaf sides (Fig. 5b, d) The b value
is quite near to zero (except the S(D/B)UI value its absolute b value is
0.114) and the r value is near to 1 in collimated light (Table 1, row 3).
This is to be expected as the water inltration and the surface lms of
water strongly enlarge the relative amount of beams transmitted
through the leaf and lowering the surface and internal reection (see
Fig. 2a, blue and red symbols). Then, in terms of the optical law of reversibility of the beam paths, the results obtained from the adaxial
and abaxial leaf side in case of completely homogeneous leaf interior
should not differ. As there remain some differences in the refraction indexes after inltration, a small effect of refraction and scattering may
cause the detected irregularities in the sensitivity curves and even an
asymmetrical passage of rays.
4.3.2. Natural leaf
On the other hand, the natural leaf revealed quite interesting behavior (Fig. 5a, c). It was found that the beam path in the leaf is larger both
for collimated and diffuse illumination (r about 1.2981.432; Table 1)
when the leaf is illuminated from the adaxial leaf side than from the abaxial side. In a possible explanation of this effect, we should take into account again the meaning of the beam path in our concept. As already
mentioned in Section 3.1, our calculation of leaf absorbance covers
both beams transmitted by the leaf and beams returning from the leaf
back to the space of illumination. Although the surface reection Re is
larger from the abaxial leaf side (Hlavinka et al., 2013; McClendon &
Fukshansky, 1990a), we recall here that the near infrared part (no pigment absorption) of the spectra of diffuse reectance and transmittance
show that more light is reected (sum of surface and inner reections)
from the adaxial leaf side than from the abaxial side (see Fig. 2a, green
and gray symbols). Most probably this reection is caused by the
inner reection originating from the interface between the palisade
and spongy parenchyma, the reection being mostly of total reection
character. This effect, together with different leaf structures on its two
sides (Fig. 1) cause that the light reected from the interface must
pass a larger path when reected from the adaxial than from the abaxial
leaf side. There is only a slight increase in the non-linearity when light
comes from the adaxial leaf side (Table 1, row 3).
In conclusion, the light incident from the adaxial leaf side is more effectively absorbed than the same light incident from the abaxial leaf
side.
4.4. Effect of chloroplast arrangement (face versus side
chloroplasts positions)
Until now, we did dot discern between the sensitivities detected for
the face and side arrangements of chloroplasts in cells. These two positions give two curves (spectra) in Figs. 3-5. We will now discuss the differences between these two curves.
4.4.1. Effect of leaf inltration
The effect of chloroplast movement is demonstrated in the relation
of natural versus inltrated leaf (Fig. 3). The fact, that the chloroplast

avoidance movement was induced by blue light illumination of the adaxial leaf side before measurements of related spectra (see Sections 2.7
and 2.8) is reected in a rather different situation of the sensitivity
curves calculated for adaxial and abaxial sides in collimated light
(Fig. 3a, b). Whereas the adaxial leaf side revealed a distinct difference
in the sensitivity for the face and side chloroplast position (Fig. 3a),
the abaxial side showed nearly no difference (Fig. 3b). The side position
in the palisade cells allows the beams in natural leaves to shorten the
paths (Fig. 3a, Table 1) both in reected and transmitted beams when
illuminated from the adaxial leaf side.
Because the beams contributing to the calculated absorbance of the
abaxial leaf side contain a large part of the transmitted rays (see
Fig. 2b, infrared region, blue symbols), these rays change the direction
in the spongy parenchyma and enter the palisades in different directions. The side position of chloroplasts is now not so important and
the sensitivity curves do not differ signicantly.
4.4.2. Effect of geometrical character of the illuminating light
Changes in sensitivities caused by changes in chloroplast positions in
the cells are more pronounced when comparing the effect of the geometrical character of illuminating light in the case of the inltrated leaves
(Fig. 4c, d) than in the natural leaves (Fig. 4a, b) and in the case of adaxial
leaf side (Fig. 4a, c) than in abaxial leaf side (Fig. 4b, d). In the inltrated
leaf representing a more homogeneous space for light beam passage
than in the case of natural leaves, transition of chloroplasts from the
side positions into the face positions distort the ideal situation of the inltrated leaf more signicantly than in the case of the natural leaves.
4.4.3. Effect of leaf side
The effect of chloroplast movement may be demonstrated also when
comparing the sensitivity curves relating the adaxial to the abaxial leaf
side (Fig. 5). In case of the inltrated leaves (Fig. 5b, d), as already mentioned, the effect is very small and may be understood as insignicant
under our experimental conditions. However, in the natural leaves,
some distinct differences are visible. The movement of chloroplasts
from the face to side position lowers the difference in the mean beam
path between the abaxial and adaxial illumination (Fig. 5a). This can
be explained by the fact that the side position enables the perpendicular
collimated light penetrate the palisade cells more directly when illuminated from the adaxial side and shortens the overall light path.
The illumination by diffuse light (Fig. 5c, d) indicates no signicant
changes in the beam path judging from positions of the curves, however
the non-linearity is slightly higher in case of chloroplast side position;
the curves have deeper local minima. We suggest ascribing this nonlinearity increase to the chloroplast grouping effect in the side position.
It should be mentioned that this effect is visible not only for natural leaf
(Fig. 5c) but also in case of inltrated leaf (Fig 5d). This may be understood by a possibility that the diffuse light (Fig. 5d) meets more groups
of chloroplasts than the collimated one (Fig. 5b).
4.4.4. Change in leaf absorbance
For a moment, we can suggest to take the lengthening (parameter r)
as a measure of the increase in the absorption effectiveness. When the
sensitivity Sd() was calculated for comparison of side versus face position of natural leaves in collimated light, the face position being the reference one, the value of about 0.9 (0.903 for illumination from adaxial
leaf side) was obtained (Fig. 6; Table 1, row 4). This means that in this
special case the chloroplast rearrangement from the face to the side position lowers the absorption by about 10%. However, because the nonlinearity parameter b reached absolute values of about 0.12 in this
case, the side arrangement enlarges the non-linearity effect. Table 1,
row 4 shows an important fact that in all cases the side position revealed
a higher nonlinearity than the face position. This supports the notion
that the chloroplast grouping leads also to an increase of the nonlinearity in the light absorption. With some exception in the inltrated

B. Barnkov et al. / Remote Sensing of Environment 174 (2016) 181196

leaves the side position reveals a tendency to shorten the light path in
the leaf (Table 1, row 4, the r parameter is slightly lower than 1).
Generally, our results show (Table 1, row 4) that chloroplast
arrangement mainly effects the non-linearity (mean absolute value
of b parameter being 0.142) whereas the lengthening is almost unaffected (mean value of r parameter being 0.980). Considering our interpretation of the b an r parameters to reect the sieve and detour effects
(see Section 3.5), respectively, then our results agree with those of
Davis et al. (2011) who concluded that chloroplast movement manipulate the sieve but not the detour effect.
4.5. Quantication of the effect of chloroplast arrangement and predictions
for the reectance measurements
Fig. 6, presenting directly the effect of the chloroplast arrangement
on the spectra, shows that maximal absolute values of differential sensitivity, Sd(), are 0.20.3. This means that the leaf absorbance, D(), was
changed by chloroplast arrangement maximally by 2030%. However,
for practical reasons and application in remote sensing measurements,
it is more valuable to relate changes in D() to practically measured
quantity, leaf diffuse reectance RD(). This may also lead to specication of conditions under which the chloroplast rearrangement may be
most reected in the detected leaf diffuse reectance. In the Supplementary data we have estimated that in our case the chloroplast rearrangement can change the leaf reectance by about 2% in the blue and
red regions and by about 5% in the green region of light spectrum. All
the so far mentioned quantications are for our case of fully developed
green leaves.
As it is also shown in the Supplementary data, the highest changes in
the leaf reectance spectrum caused by chloroplast arrangement might
be found in spectral region between 500 and 650 nm and with leaves
having lower content of chlorophylls (young or senescent leaves, leaves
of lower nutrition supply, leaves of plants exposed to long lasting stress,
etc.) and/or leaves with high chloroplasts mobility. The decrease in
chlorophyll content causes a decrease in leaf absorbance, which, according to our derivation in the Supplementary data (equation S63), leads to
more pronounced changes in diffuse reectance. The decrease in chlorophyll content also causes more pronounced changes in leaf absorbance caused by chloroplast movement and thus also more
pronounced changes in diffuse reectance (Eq. S63). In connection to
that, Xiong et al. (2015) reported a more pronounced decrease in
SPAD-values, indicating a more pronounced changes in leaf absorbance,
caused by chloroplast movement for the leaves with a lower chlorophyll
content (caused by a low or no nitrogen supply) than for the leaves with
a higher chlorophyll content. Therefore, more signicant changes in the
leaf reectance caused by chloroplast arrangement than reported in our
case and expected in the other cases are justied. Taking all together, as
reectance signal is used for calculation of many indices in remote sensing, changes in chloroplast arrangement should thus be considered.
5. Conclusion

193

National Program of Sustainability I, Ministry of Education, Youth and


Sports, Czech Republic.
Appendix A
A.1. Mean geometrical length of a beam path in a desk sample
The inltrated leaf may serve as an experimental model for the evaluation of differences between the mean geometrical paths of light rays
in case of collimated and diffuse illumination. We use the term geometrical path to distinguish from the optical path which is usually understood in optics as a product of the geometrical path and the refractive
index. The leaf is modeled as a planparallel homogeneous sample having refractive index n and thickness t.
A.2. Collimated illumination
In case of perpendicular collimated illumination the mean geometrical beam path in the rst beam passage is equal to the leaf thickness t.
All rays are equivalent and enter the leaf equally. This means that they
all are reected according to the LambertBeer's equation:

Rc

n12
n 12

A1

A.3. Diffuse illumination


However, a rather different situation is in the case of diffuse incident
illumination. Now, we have to consider different angles of incidence of
different beams. If the angle of incidence is , then the corresponding
beam path in the leaf is:
n t
t p ;
n2 sin2

A2

where, we have used the law of light refraction:


sin n  sin;

A3

where is the angle of refraction.


Further, we expect the intensities of light coming from the different
direction being equal. Then we should apply the Lambert's law according to which the light intensity incident on the surface is proportional to
cos . The relative amount of beams incident with angle is proportional to the corresponding space angle:
d 2 sin d:

A4

By comparison of changes of optical parameters of leaf pairs differing


in some properties (natural versus inltrated leaf, diffuse versus collimated illumination, adaxial versus abaxial leaf side) we have found
that the chloroplast arrangement is in some cases reected in the studied changes. We have found that the measured changes are non-linear
and can be described well by a non-liner theory, the SENLELOP model,
implying deviations from the LambertBeer's law in leaves. It summary,
chloroplast arrangement and non-linear models should be considered
in future remote sensing studies to obtain more correct results.

Moreover, we have to consider the light reection RF() depending


in the case of nonpolarized light on the angle according to the Fresnel
formulas:

Acknowledgments

It should be noted, that the formula presented above is the correct


formula for the reectance of detectable light intensity, not for the
ratio of the amplitudes of the electrical vectors as presented in Gorton
et al. (2010).

This work was supported by the grant No. LO1204 (Sustainable Development of Research in the Centre of the Region Han) from the

2
p !2
1 4 cos n2 sin2
p
R F
2
cos n2 sin2

p !2 3
n2 cos n2 sin2 5
p
:
n2 cos n2 sin2
A5

194

B. Barnkov et al. / Remote Sensing of Environment 174 (2016) 181196

The summated and weighted light path for rst passage is equal to:

Z2

sin cos
p1R F d
n2 sin2

X 2 t n
0

Z
2 t n

g 1 F d:

However, if only the rst passage is to be considered, it is necessary


to include the Fresnel reections. In this case, the mean path was calculated numerically and for n = 1.36 we have obtained:

A6

X
xd
2 t n
X0

Z2
g 1 F d 1:098:

A14

Using the same theoretical approach in calculation of the mean path


for the case of collimated light, we obtain:

We have designated:
sin cos
g 1F 1 F p 1R F 
n2 sin2

xc 1Rc t:

Let us for a moment ignore the Fresnel reections, then the sum of
rst weighted light path is:

In summary, using the diffuse light instead of perpendicular collimated light should increase the geometrical beam path in the inltrated
leaf by a factor (assuming leaf n = 1.36).

Z2
X 1 2 t n
0

sin cos d
p :
n2 sin2

A7

Calculating the integral we obtain:


h
pi
X 1 2 t n n n2 1 :

A8

The weighted incident light representing all incident beams may be


expressed as:

Z2
I w 2

sin cos d :

A9

The rst mean geometrical path of diffuse light in case of neglecting


the reection and multiple reections gives:
x1


p
X1
2n n n2 1 t:
X0

A10

Here X0 means the weighted beam path of perpendicularly incident


collimated light:
X0 t

A11

For derivation see Supplementary data.


For n = 1.5 we obtain x1 1:146 t, and for n = 1.36 x1 1:19 tand
one would expect the mean beam path around 1.2 t.
Until now, we considered only the rst beam passage through the
leaf structure. However, the beam passes also the second frontier between the sample and air. The beam is partly reected and partly transmitted. A calculation leads to an innite geometric series which sum can
be expressed by the reection coefcient R. Interestingly, in case of the
same reection coefcients at both boundaries, the result is the same as
in the case of neglected reection, i.e., it is expressed by the formula
(A10).
However, as can be shown (Supplementary data), if the reection
coefcients at the two boundaries are different, and related by:
R2 m R1

A12

then the expression (A10) should be multiplied by:


1

R1 m1
1m R21

A13

and the beam path may reach higher values due to m being higher
than 1.

A15

xd
1:124:
xc

A16

The increase in the mean geometrical path should mean an increase


in absorbance. For an ideal homogeneous inltrated leaf this increase
should amount about 12%.
The presented model serves for an approximate estimation of possible increase of absorbance of inltrated leaf and very roughly also for
the natural leaf. Although the situation in the natural leaf is much
more complex, the model predicts an increase of absorbance not exceeding about 20% when going from the collimated incident light to
the diffuse one. This theoretical estimation is in accordance with the
measured data. A change of the collimated light to the diffuse one
leads only to a mild increase in the absorption of the light.
It should be noted that we do not specically separate the external
and internal reections. In principle, it is possible to incorporate this discrimination into the expressions. However, this discrimination leads to
more complicated expressions which would be difcult to understand.
This discrimination would be done in a separate study. The external reection Re amounts about between 4 to 10%. We suppose that its effect
is a marginal one and we do not discuss this issue in this work.
Appendix B. Supplementary data
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.rse.2015.12.011.
References
Allen, W. A., & Richardson, A. J. (1968). Interaction of light with plant canopy. Journal of
the Optical Society of America, 58, 1323-1028.
Allen, W. A., Gausmann, H. W., Richardson, A. J., & Thomas, J. R. (1969). Interaction of isotropic light with a compact plant leaf. Journal of the Optical Society of America, 59,
13761379.
Allen, W. A., Gausmann, H. W., & Richardson, A. J. (1970). Mean effective optical constants
of cotton leaves. Journal of the Optical Society of America, 60, 542547.
Bana, A. K., Aggarwal, C. H., abuz, J., Sztatelman, O., & Gabry, H. (2012). Blue light signalling in chloroplast movements. Journal of Experimental Botany, 63, 15591574.
Baranoski, G. V. G. (2006). Modeling the interaction of infrared radiation (750 to
2500 nm) with bifacial and unifacial plant leaves. Remote Sensing of Environment,
100, 335347.
Baranoski, G. V. G., & Rokne, J. G. (1997). An algorithmic reectance and transmittance
model for plant. Computer Graphics Forum, 16, 141150.
Barton, C. V. M. (2001). A theoretical analysis of the inuence of heterogeneity in chlorophyll distribution on leaf reectance. Tree Physiology, 21, 789795.
Bartokov, H., Nau, J., & Vkruta, M. (1999). The arrangement of chloroplasts in cells inuence the reabsorption of chlorophyll uorescence emission. The effect of desiccation on the chlorophyll uorescence spectra of Rhizomnium punctatum leaves.
Photosynthesis Research, 62, 251260.
Bergstrsser, S., Fanoukaris, D., Schmittgen, S., Cendero-Mateo, M. P., Jansen, M., Scharr, H., &
Rascher, U. (2015). HyperART: non-invasive quantication of leaf traits using
hyperspectral absorptionreectancetransmittance imaging. Plant Methods, 11, 117.
Bird, R. E., & Riordan, C. (1986). Simple solar spectral model for direct and diffuse irradiance on horizontal and tilted planes at the earth's surface for cloudless atmospheres.
Journal of Applied Meteorology, 25, 8797.

B. Barnkov et al. / Remote Sensing of Environment 174 (2016) 181196


Brodersen, C. R., & Vogelmann, T. C. (2010). Do changes in light direction affect absorption
proles in leaves. Functional Plant Biology, 37, 403412.
Brugnoli, E., & Bjrkman, O. (1992). Chloroplast movements in leaves: inuence on
chlorophyll uorescence and measurements of light-induced absorbance changes related to pH and zeaxanthin formation. Photosynthesis Research, 32, 2335.
Buschmann, C. (2007). Variability and application of the chlorophyll uorescence emission ratio red/far-red of leaves. Photosynthesis Research, 92, 261271.
Carter, G. A. (1991). Primary and secondary effects of water content on the spectral reectance of leaves. American Journal of Botany, 78, 916924.
Cui, M., Vogelmann, T. C., & Smith, W. K. (1991). Chlorophyll and light gradients in sun
and shade leaves of Spinacia oleracea. Plant, Cell and Physiology, 14, 493500.
Daughtry, C. S. T., & Walthall, C. L. (1998). Spectral discrimination of Cannabis sativa L.
leaves and canopies. Remote Sensing of Environment, 64, 192201.
Davis, P. A., Caylor, S., Whippo, C. W., & Hangarter, R. P. (2011). Changes in leaf optical
properties associated with light-dependent chloroplast movements. Plant, Cell and
Environment, 34, 20472059.
Dawson, T. P., Curran, P. J., & Plummer, S. E. (1998). LIBERTY-modeling the effects of leaf
biochemical concentration on reectance spectra. Remote Sensing of Environment, 65,
5060.
DeLucia, E. H., Shenoi, H. D., Naidu, S. L., & Day, T. A. (1991). Photosynthetic symmetry of
sun and shade leaves of different orientations. Oecologia, 87, 5157.
DeLucia, E. H., Nelson, K., Vogelmann, T. C., & Smith, W. K. (1996). Contribution of intercellular reectance to photosynthesis in shade leaves. Plant, Cell and Environment,
19, 159170.
Frolec, J., ebek, J., Lazr, D., & Nau, J. (2010). Impact of two different types of heat
stress on chloroplast movement and uorescence signal of tobacco leaves. Plant Cell
Reports, 29, 705714.
Fukshansky, L., & Kazarinova, N. (1980). Extension of the KubelkaMunk theory of light
propagation in intensely scattering materials to uorescent media. Journal of the
Optical Society of America, 70, 11011111.
Gabry, H. (2004). Blue light-induced orientation movements of chloroplasts in higher
plants: recent progress in the study of their mechanisms. Acta Physiologiae
Plantarum, 26, 476478.
Gamon, J. A., Peuelas, J., & Field, C. B. (1992). A narrow-waveband spectral index that
tracks diurnal changes in photosynthetic efciency. Remote Sensing of Environment,
41, 3544.
Ganapol, B. D., Johnson, L. F., Hammer, P. D., Hlavka, C. A., & Peterson, D. L. (1998).
LEAFMOD: a new within-leaf radiative transfer model. Remote Sensing of
Environment, 63, 182193.
Gates, D. M., Keegan, H. J., Schleter, J. C., & Weidner, V. R. (1965). Spectral properties of
plants. Applied Optics, 4, 1120.
Gerber, F., Marion, R., Olioso, A., Jacquemoud, S., Ribeiro da Luz, B., & Fabre, S. (2011).
Modeling directional-hemispherical reectance and transmittance of fresh and dry
leaves from 0.4 m to 5.7 m with the PROSPECT-VISIR model. Remote Sensing of
Environment, 115, 404414.
Gorton, H. L., Craig, R., Brodersen, C. R., Williams, W. E., & Vogelmann, T. C. (2010). Measurement of the optical properties of leaves under diffuse light. Photochemistry and
Photobiology, 86, 10761083.
Govaerts, Y. M., Jacquemoud, S., Verstraete, M. M., & Ustin, S. L. (1996). Three-dimensional
radiation transfer modeling in a dicotyledon leaf. Applied Optics, 35, 65856598.
Govindjee (1995). Sixty-three years since Kautsky: chlorophyll a uorescence. Australian
Journal of Plant Physiology, 22, 131160.
Hlavinka, J., Nau, J., & pundov, M. (2013). Anthocyanin contribution to chlorophyll
meter readings and its correction. Photosynthesis Research, 118, 277295.
Jacquemoud, S., & Baret, F. (1990). PROSPECT: a model of leaf optical properties spectra.
Remote Sensing of Environment, 34, 7591.
Jacquemoud, S., Ustin, S. L., Versebout, J., Schmuck, G., Andreoli, G., & Hosgood, B. (1996).
Estimating leaf biochemistry using the PROSPECT leaf optical properties model.
Remote Sensing of Environment, 56, 194202.
Jacquemoud, S., Bacour, C., Poilv, H., & Frangi, J. -P. (1999). Comparison of four radiative
transfer models to simulate plant canopies reectance: direct and inverse mode.
Remote Sensing of Environment, 74, 471481.
Jacquemoud, S., Verhoef, W., Baret, F., Bacour, C., Zarco-Tejada, P. J., Asner, G. P., ... Ustin, S.
L. (2009). PROSPECT + SAIL models: a review of use for vegetation characterization.
Remote Sensing of Environment, 113, S56S66.
Kasahara, M., Kagawa, T., Oikawa, K., Suetsugu, N., Miyao, M., & Wada, M. (2002).
Chloroplast avoidance movement reduces photodamage in plants. Nature, 420,
829832.
Kirk, J. T. O. (1976). A theoretical analysis of the contribution of algal cells to the attenuation of light within natural waters. New Phytologist, 77, 341358.
Kong, S. -G., & Wada, M. (2014). Recent advances in understanding the molecular mechanism of chloroplast photorelocation movement. Biochimica et Biophysica Acta
Bioenergetics, 1837, 522530.
Kniger, M. (2014). Chloroplast movement in higher plants, ferns and bryophytes: a comparative point of view. In D. T. Hanson, & S. K. Rice (Eds.), Photosynthesis in Bryophytes
and Early Land Plants (pp. 131150). Springer.
Kniger, M., & Bolliger, N. (2012). Chloroplast movement behavior varies widely
among species and does not correlate with high light stress tolerance. Planta,
236, 411426.
Kubelka, P., & Munk, F. (1931). Ein beitragzur optik der farbanstriche. Zeitschrift fr
Technische Physik, 11, 593601.
Lee, D. W., Bone, R. A., Tarsis, S. L., & Storch, D. (1990). Correlates of leaf optical properties
in tropical forest sun and extreme-shade plants. American Journal of Botany, 77,
370380.
Lichtenthaler, H. K. (1987). Chlorophylls and carotenoids: pigments of photosynthetic
biomembranes. Methods in Enzymology, 148, 350382.

195

Lork, W., & Fukshansky, L. (1985). The inuence of chlorophyll uorescence on the light
gradients and the phytochrome state in a green model leaf under natural conditions.
Plant, Cell and Environment, 8, 3339.
Lucarini, V., Saarinen, J. J., Peiponen, K. -E., & Vartiainen, E. M. (2005). KramersKronig relations in optical materials research. Springer.
Mader, M., Nau, J., Makovec, P., Sofrov, D., Vacek, J., Grc, L., & Masojdek, J. (1982). Effect
of ontogeny and additional fertilization with nitrogen on composition and activity of
the spring barley photosynthetic apparatus. Photosynthetica, 16, 161175.
Maier, S. W., Ldeker, W., & Gnther, K. P. (1999). SLOP: a revised version of the stochastic model for leaf optical properties. Remote Sensing of Environment, 68,
273280.
Malenovsk, Z., Martin, E., Homolov, L., Gastellu-Etchegorry, J. -P., Zurita-Milla, R.,
Schaepman, M. E., ... Cudln, P. (2008). Inuence of woody elements of a Norway
spruce canopy on nadir reectance simulated by the DART model at very high spatial
resolution. Remote Sensing of Environment, 112, 118.
Mariotti, M., Ercoli, L., & Masoni, A. (1996). Spectral properties of iron-decient corn and
sunower leaves. Remote Sensing of Environment, 58, 282288.
McClendon, J. H., & Fukshansky, L. (1990a). On the interpretation of absorption spectra of
leaves. I. Introduction and the correction of leaf spectra for surface reection.
Photochemistry and Photobiology, 51, 203210.
McClendon, J. H., & Fukshansky, L. (1990b). On the interpretation of absorption spectra of
leaves. II. The non-absorbed ray of the sieve effect and the mean optical pathlength
in the remainder of the leaf. Photochemistry and Photobiology, 51, 211216.
Merzlyak, M. N., Chivkunova, O. B., Zhigalova, T. V., & Naqvi, K. R. (2009). Light absorption
by isolated chloroplasts and leaves: effects of scattering and packing. Photosynthesis
Research, 102, 3141.
Nau, J., Klinkovsk, T., Ilk, P., & Ciknek, D. (1994). Model studies of chlorophyll uorescence reabsorption at chloroplast level under different exciting conditions.
Photosynthesis Research, 40, 6774.
Nau, J., Prokopov, J., ebek, J., & pundov, M. (2010). SPAD chlorophyll meter reading
can be pronouncedly affected by chloroplast movement. Photosynthesis Research, 105,
265271.
Osborne, B. A., & Raven, J. A. (1986). Light absorption by plants and its implication for
photosynthesis. Biological Reviews, 61, 160.
Palombi, L., Cecchi, G., Lognoli, D., Raimondi, V., Toci, G., & Agati, G. (2011). A retrieval algorithm to evaluate the photosystem I and photosystem II spectral contributions to
leaf chlorophyll uorescence at physiological temperatures. Photosynthesis Research,
108, 225239.
Park, Y., Chow, W. S., & Anderson, J. M. (1996). Chloroplast movement in the shade plant
Tradescantia albiora helps protect photosystem II against light stress. Plant
Physiology, 111, 867875.
Pedrs, R., Goulas, Y., Jacquemoud, S., Louis, J., & Moya, I. (2010). FluorMODleaf: a new leaf
uorescence emission model based on the PROSPECT model. Remote Sensing of
Environment, 114, 155167.
Rouse, J. W., Haas, R. H., Schell, J. A., & Deering, D. W. (1974). Monitoring vegetation systems in the Great Plains with ERTS. Proceedings of the third Earth resources technology satellite-1 Symposium. NASA, SP-351, 309317.
Seyfried, M., & Fukshansky, L. (1983). Light gradients in plant tissue. Applied Optics, 22,
14021408.
Smith, W. K., Vogelmann, T. C., DeLucia, E. H., Bell, D. T., & Shepherd, K. A. (1997). Leaf
form and photosynthesis. Bioscience, 47, 785793.
Steinhardt, A. R., & Fukshansky, L. (1987). Geometrical optics approach to the intensity
distribution in nite cylindrical media. Applied Optics, 26, 37783789.
Stokes, G. G. (1860-1862). On the intensity of the light reected from or transmitted
through a pile of plates. Proceedings of the Royal Society of London, 11, 545556.
Suila, P., & Nau, J. (2007). A Monte Carlo study of the chlorophyll uorescence emission
and its effect on the leaf spectral reectance and transmittance under various conditions. Photochemical and Photobiological Sciences, 6, 894902.
Terashima, I., & Saeki, T. (1985). A new model for leaf photosynthesis incorporating the
gradients of light environment and of photosynthetic properties of chloroplasts within leaves. Annals of Botany, 56, 489499.
Terashima, I., Fujita, T., Inoue, T., Chow, W. S., & Oguchi, R. (2009). Green light drives
leaf photosynthesis more efciently than red light in strong white light:
revisiting the enigmatic question of why leaves are green. Plant Cell Physiology,
50, 684697.
Tucker, C. J. (1978). Red and photographic infrared linear combinations for monitoring
vegetation. NASA Technical Memorandum, 79620.
Ustin, S. L., Jacquemoud, S., & Govaerts, Y. M. (2001). Simulation of photon transport in a
three-dimensional leaf: implications for photosynthesis. Plant, Cell and Environment,
24, 10951103.
van der Tol, C., Verhoef, W., Timmermans, J., Verhoef, A., & Su, Z. (2009). An integrated
model of soil-canopy spectral radiances, photosynthesis, uorescence, temperature
and energy balance. Biogeosciences, 6, 31093129.
Verhoef, W., & Bach, H. (2007). Coupled soil-leaf-canopy and atmosphere radiative transfer modeling to simulate hyperspectral multi-angular surface reectance and TOA radiance data. Remote Sensing of Environment, 109, 166182.
Verrelst, J., Rivera, J. P., van der Tol, C., Magnani, F., Mohammed, G., & Moreno, J.
(2015). Global sensitivity analysis of the SCOPE model: what drives simulated
canopy-leaving sun-induced uorescence? Remote Sensing of Environment, 166,
821.
Vogelmann, T. C., & Gorton, H. L. (2014). Leaf: Light capture in the photosynthetic organ.
In M. F. Hohmann-Marriott (Ed.), The Structural Basis of Biological Energy Generation
(pp. 363377). Springer.
Vogelmann, T. C., & Martin, G. (1993). The functional signicance of palisade tissue: penetration of directional versus diffuse light. Plant, Cell and Environment, 16, 6572.
Wada, M. (2013). Chloroplast movement. Plant Science, 210, 177182.

196

B. Barnkov et al. / Remote Sensing of Environment 174 (2016) 181196

Wada, M., & Kong, S. G. (2011). Analysis of chloroplast movement and relocation in
Arabidopsis. In R. P. Jarvis (Ed.), Chloroplast research in Arabidopsis. Methods and
protocols, 1. (pp. 87102).
Williams, W. E., Gorton, H. L., & Witiak, S. M. (2003). Chloroplast movement in the eld.
Plant, Cell and Environment, 26, 20052014.

Xiong, D., Chen, J., Yu, T., Gao, W., Ling, X., Li, Y., ... Huang, J. (2015). SPAD-based leaf nitrogen estimation is impacted by environmental factors and crop leaf characteristics.
Scientic Reports, 5, 13389.
Zurzycki, J. (1961). The inuence of chloroplast displacement on the optical properties of
leaves. Acta Societatis Botanicorum Poloniae, 30, 503527.

S-ar putea să vă placă și