Second EARSeL Workshop on Imaging Spectroscopy, Enschede, 2000

ESTIMATION OF CHLOROPHYLL
FLUORESCENCE UNDER
NATURAL ILLUMINATION

FROM HYPERSPECTRAL DATA

Pablo J. ZARCO-TEJADA*, John R.

MILLER**, Gina H. MOHAMMED***,
Thomas L. NOLAND***, and Paul
H. SAMPSON***
*
Centre for Research in Earth and Space
Science (CRESS), York University, Toronto,
Canada pzarco@yorku.ca

**
Dept. of Physics and Astronomy, York
University & Centre for Research in Earth
and Space Technology (CRESTech),
Toronto, Canada
jrmiller@yorku.ca
***
Ontario Forest Research Institute (OFRI),
Ontario Ministry of Natural Resources,
Sault Ste. Marie, Ontario, Canada
gina.mohammed@mnr.gov.on.ca

KEY WORDS: Reflectance, Chlorophyll

Fluorescence, Red Edge, Radiative Transfer,
CASI, Hyperspectral

ABSTRACT

This paper reports a series of laboratory and

field measurements of spectral reflectance
under artificial and natural light conditions
which demonstrate that effects of natural
chlorophyll fluorescence are observable in
the reflectance red edge spectral region.
These are results from the progress made to
link physiologically-based indicators to
optical indices from hyperspectral remote
sensing in the due to CF are measurable under natural
Bioindicators of Forest illumination conditions, although airborne
Sustainability Project. experiments with the CASI hyperspectral
This study is carried out sensor produced promising but less
on twelve sites of Acer convincing results in two diurnal
saccharum M. in the experiments carried out in 1999 and 2000,
Algoma Region, Ontario where small variations of reflectance due to
(Canada), where field the effect of CF were observed.
measurements,
laboratory-simulation
experiments, and
hyperspectral CASI
imagery have been carried
out in 1997, 1998, 1999 1 INTRODUCTION
and 2000 campaigns. Leaf
samples from the study The objective of the Bioindicators of Forest
sites have been used for Sustainability Project (Mohammed et al.,
reflectance and 1997; Sampson et al., 1998) is to develop
transmittance links between physiologically-based bio-
measurements with the indicators from field and laboratory data and
Li-Cor Model 1800 optical indices from hyperspectral remote
integrating sphere sensing data for assessing forest condition.
apparatus coupled to an Previous work (Zarco-Tejada et al.,1999a;
Ocean Optics Model 1999b) showed that optical indices
ST1000 fibre calculated from single leaf reflectance data,
spectrometer in which the infinite reflectance models from optically-
same leaves are thick simulation formulae, and canopy
illuminated alternatively reflectance models, progressively more
with and without closely represented the observed above-
fluorescence-exciting canopy reflectance spectra. Optical indices
radiation. A study of the calculated from modelled canopy
diurnal change in leaf reflectance through infinite and canopy
reflectance spectra, reflectance models were shown to be able to
combined with be used for estimation of pigment content
fluorescence over closed deciduous canopies of Acer
measurements with the saccharum M, and showed high correlations
PAM-2000 Fluorometer with ground truth chlorophyll fluorescence
show that the difference (CF). The strong correlation relationships
spectra are consistent withobtained between selected optical indices
observed diurnal changes calculated in the 690 and 750 nm spectral
in steady-state region from airborne CASI hyperspectral
fluorescence. Small data with ground measured CF were studied
canopies of Acer in detail in Zarco-Tejada et al. (in press, a,
saccharum M. have been b). Experiments were carried out at different
used for laboratory levels of study in the laboratory and through
measurements with the the development of a radiative transfer
CASI hyperspectral model (FRT) that simulates the effect of the
sensor, and under natural fluorescence signal superimposed to the
light conditions with a reflectance spectrum. This paper reports on
fibre spectrometer in further research carried out in this research
diurnal trials, in which the theme, in which experiments with a fibre
variation of measured spectrometer were carried out using small
reflectance is shown canopies of seedlings under natural
experimentally to be illumination conditions, and with CASI-
consistent with a airborne campaigns over Acer saccharum
fluorescence signature M. study sites in 1999 and 2000 with
imposed on the inherent specific bandsets to investigate the ability to
leaf reflectance signature. detect CF from an airborne sensor.
Such reflectance changes
Second EARSeL Workshop on Imaging Spectroscopy, Enschede, 2000
Leaf material used for experimental studies
reported in this paper was sampled from
Acer saccharum M. potted trees grown in
the greenhouse, and from twelve 30x30m
Acer saccharum M. study sites in the
2 LEVELS OF Algoma Region, Canada, where a total of
STUDY 440 single leaf samples were collected per
campaign for biochemical analysis and
The effects of the CF measurement of chlorophyll fluorescence,
signature on vegetation leaf chlorophyll and carotenoid content.
apparent reflectance were Experimental methods to measure leaf and
studied in a series of canopy reflectance were carried out at the
experiments in 1997, different levels of study, and are described
1998, 1999 and 2000 at in the following sections. Chlorophyll
different levels of scale. fluorescence was analyzed with a Pulse
Four different levels of Amplitude Modulation (PAM-2000)
study were carried out, Fluorometer (Heinz Walz GmbH, Effeltrich,
from the leaf level to the Germany), an instrument that has been used
canopy, in order to widely in basic and applied fluorescence
investigate whether the research (Mohammed et al., 1995).
effects of CF on apparent Procedures used for measuring Fv/Fm and
reflectance are effective quantum yield F/Fm were based
measurable in spite of on standard methodologies as documented
increasing complexity: i) in the PAM-2000 manual (Heinz-Walz-
at leaf level, with data GmbH, 1993). For measurement of maximal
collected from Acer fluorescence induction Fv/Fm, leaves were
saccharum M. study sites dark-adapted in bags at room temperature
to develop relationships for at least 30 min.
between leaf reflectance
and CF; diurnal studies 3.1 Leaf-level measurements of
were also carried out at reflectance and transmittance
leaf level to study the
effects of changes in leaf Measurements of reflectance and
apparent reflectance due transmittance at leaf level were acquired on
to diurnal CF patterns; ii) leaf samples using a Li-Cor 1800-12
at canopy simulation level Integrating Sphere apparatus coupled by a
in the laboratory using the 200 m diameter single mode fibre to an
CASI hyperspectral Ocean Optics model ST 1000 spectrometer,
sensor and maple with a 1024 element detector array, 0.5 nm
seedlings; iii) at canopy sampling interval and ~7.3 nm spectral
simulation level with a resolution in the 340-860 nm range. A
fibre spectrometer in a modification was made to the standard Li-
diurnal trial using Acer Cor 1800-12 Integrating Sphere apparatus
saccharum M. seedlings involving a second Li-Cor Lamp/Collimator
under natural illumination housing and the insertion of a Schott RG
conditions; and iv) at 695 coloured long-pass glass filter blocking
above canopy over radiant flux at < 695nm at the exit
selected forest sites with aperture of one of the illuminator units
the airborne CASI (Zarco-Tejada et al., in press, a). These two
hyperspectral sensor in a light sources enabled measurements
series of diurnal alternately of reflectance and transmittance
experiments with differentof a given sample without fluorescence and
CASI modes of operation including the effect of fluorescence. Diurnal
and bandsets. and time-decay studies were carried out
with leaf samples in order to study
variations in the apparent leaf reflectance
3 and transmittance due to normal diurnal
changes of chlorophyll fluorescence, and to
EXPERIMENTAL the effects of fluorescence time-decay on the
METHODS AND DATA measurements of apparent spectral
COLLECTION reflectance, respectively (Zarco-Tejada et
al., in press, a).
3.2 Canopy reflectance Spectralon reflectance panel placed on the
in laboratory and under moveable platform with the plant material
natural illumination enabled reflectance calculation. A filter
holder was custom-designed to permit a
CASI hyperspectral Schott RG695 high pass filter to be placed
canopy reflectance in front of the 1000W halogen light source
measurements in the in order to restrict incident radiant energy on
laboratory were made the scene to > 705 nm. This facilitated the
from a small canopy of collection of canopy reflectance
Acer saccharum M. measurements with CASI in the absence of
seedlings using a Bi- fluorescence generating radiation, similar to
Directional Reflectance measurement protocols at the leaf level.
Factor (BRF) facility. The
CASI sensor was placed A canopy of Acer saccharum M. seedlings
at height of 2.5 m above was used for measurements of canopy
the canopy of plant reflectance under natural illumination
material, and operated in conditions using up/down
a hyperspectral mode at radiance/irradiance optical head coupled by
maximum spectral two 200 m diameter fibre to the same
resolution with 288 Ocean Optics model ST 1000 triple
channels, spectral spacing spectrometer described previously, with 0.5
of 1.8 nm and nominal nm sampling interval, 7.3 nm spectral
bandwidth of 2.5 nm. resolution, and 340-860 nm range. The
Collimated illumination at radiometer optics with the fibre
45 inclination was spectrometer detector was directed at nadir
provided by a regulated view over the seedling canopy using a 1.5m
1000W halogen light tripod, and measurements were made at
source and a collimating 8.20h, 8.42h, 9.02h, 10.47h, 11.34h, 12.52h,
lens. The raw 12 bit CASI and 13.37h to capture variations in apparent
data were calibrated to reflectance due to the effect of diurnal
spectral radiance and a changes in CF.
Second EARSeL Workshop on Imaging Spectroscopy, Enschede, 2000

4 RESULTS

Field-airborne canopy Results of the leaf-level experiments and

reflectance CASI canopy reflectance measurements in
measurements under the laboratory are reported in detail in
natural illumination Zarco-Tejada et al. (in press, a, b). A typical
with CASI pair of reflectance spectra obtained with the
hyperspectral sensor Schott filter measurement protocol is shown
in Figure 1 illustrating the additive effect of
Airborne CASI data were the broad 740 nm fluorescence signal
collected over twelve sites superimposed on the reflectance spectrum
of Acer saccharum M. in due only to the scattering and absorption
1997, 1998, 1999 and effects within the leaf. Results of the time-
2000. Mean reflectance decay experiment showed that changes in
values per plot were CF amplitude subsequent to exposure were
calculated from the also tracked in apparent reflectance spectra.
imagery in each Acer Measurements taken every 2 seconds during
saccharum M. study site five minutes of illumination allowed
of 20x20 m. CASI data comparison between the first and last
were acquired in the spectral reflectance measurement (Figure 2).
hyperspectral reflectance The differences in apparent reflectance are
mode, with 2 m spatial seen at approximately 690 nm and 750 nm
resolution and 72 spectral corresponding to the two chlorophyll
channels (7.5 nm spectral fluorescence emission peaks and, in
bandwidth). A CASI addition, a fluorescence emission of
diurnal mission was unknown origin is observed near 370 nm in
carried out in July 1999 the blue. Reflectance differences in the red
collecting data over two edge were fitted to a double gaussian
study sites at different function as suggested in Subhash and
times of the day, 8.00h, Mohanan (1997) and as used in the FRT
9.30h, 12.20, and 16.12h radiative transfer model for CF simulation
along with ground truth on apparent reflectance (Zarco-Tejada et al.,
CF measurements with in press, a). Parameters showing the centre
PAM-2000. A second maxima and bandwidth for both
diurnal experiment over fluorescence emissions from Acer
two sites was carried out saccharum M. leaf samples are shown in
in June 2000 with a Table 1 and Figure 3. These fits produce
specific CASI mode of high correlation coefficients (R>0.9),
operation in order to generate parameters similar to those of
allow for higher spatial Subhash and Mohanan (1997) and also
resolution data with show bandwidths at 750 nm that are
spectral bands centered at approximately 2 to 3 times larger than near
the PSII photosystem. 690 nm.. The centre peaks show shifts from
CASI data were collected those reported earlier, but these results
in 9 spectral bands at correspond to natural fluorescence from
680.47, 684.26, 688.06, broadband illumination rather than from
691.86, 695.66, 699.46, laser excitation energy in the earlier results.
703.26, 707.06, and
710.87 nm and Measurements of leaf reflectance collected
0.56x1.08m spatial in the diurnal trial showed the diurnal
resolution, re-sampled to variation of fluorescence features Fv/Fm
0.5x0.5 m. Radiometric and Ft110 (Zarco-Tejada et al., in press, a),
calibration and observed during the day compared to the
atmospheric correction of variation of the reflectance difference at 740
the CASI data was nm with and without the blocking filter,
performed as explained in therefore tracking the PSII excitation to
Zarco-Tejada et al. (in visible light
press, b).
0.5
 0.03

r1
r2 0.01
Difference

0.02 0.1
0.1

0.4 0.005 0.4

nce
0.005

0.025

fleRe

0
0.015 0

ctance
0.3 0
0

600
0.02 650
r 700
400
ctance 500 750
600 800
700 0.3
800

Wavelength
Wave le n gth (nm ) (nm)

0.01

With RG695 f ilter No filter Dif

Figure 2.
f erenc e Reflecta
Re fle nce
Figure 1. Single leaf measure
reflectance measurements ments
obtained with the Li-Cor taken at
1800 apparatus and fibre difference t0 (r1)
spectrometer using the
and t1(5
measurement protocol with r

reDiffe
the RG695 filter (thick line) min) (r2)
and with no filter (thin line) in a
from a dark-adapted Acer time-
0.2
saccharum M. leaf sample 0.015 decay
(adapted from Zarco-Tejada experim
et al. (in press, a). ent
0.2 which
demonst
rates the
effect of
fluoresce
0.5 nce
emission bands spectra in the 600-800 nm
on the spectral region due to PSII
reflectance and PSI photosystems.
Second EARSeL Workshop on Imaging Spectroscopy, Enschede, 2000
CASI canopy reflectance measurements in
the laboratory showed that changes in
canopy apparent reflectance from targeted
plant material are observed when the Schott
695nm blocking filter is used. This effect is
superimposed on the evident at 730-750 nm, and is most
reflectance when there is pronounced at 742 nm. When the maple-
no excitation. Results seedling canopy was kept in a fixed position
showed that variations in during 3 minutes of CASI data acquisition
Fv/Fm during the day are in the 72-channel (7.5 nm bandwidth) mode
captured in the leaf of operation, changes in the CASI
reflectance measurements reflectance bands affected by chlorophyll
even when the pigment fluorescence in this time-decay experiment
concentration is constant. were seen (Figure 4)

2
0.36
)2

Table 1. The
reflectance
difference
R=
spectra have A1 exp
been fit with a
double gaussian

curve defined 0.36

above. The best
fit parameters
show spectral
peaks and half- +
widths A2
consistent with exp


A1
fluorescence
emissions.
1.7
Double
Gaussian Fit: 1

1.1

Param
eters
Leaf_Ref
l. A2
Canopy_
Refl.

2.5

(
2
1
2.9
)2

(
 3
Canopy_R
_Dif

2

1

52.9 nm
688.3 nm

81.1 nm
0
682.1 nm
Differe
nce
2.5

-0.5
700
750
800
R
2
1

650
0.94

29.5 nm

0.93 1.5
22.5 nm
Wavelength
(nm)

Figure 3.
%R_ The
1 measured
difference
reflectance
spectra
2 correspondin
_ g to Figures
3.5 2 and 4 have
746.1 nm been fitted
with a
double-
gaussian
752.2 nm
function.
Leaf_R_Dif
Leaf The fit
0.5 parameters
are
presented in
Table 1.

DifferenceR
eflectance
50

Start (0'')
End (3')
Difference
4
70
3

40

60 3.5 Reflecta
4.5 nce (%)
 690)
2.5 calcula
0.64 ted
from
CASI
canop
10
y
0.8 R685 reflect
ance in
0.92
Fv/Fm laborat
0.72
0.5 ory
using
Acer
1.04
sachar
um M.
30
seedlin
0.98 gs.
The
behavi
2 our of
Fv/Fm CF
R685^ during
2/
(R675* the
690) day is
0 1.02
tracke
0.6 d by
the
optical
0.76
0 index
2
derive
d from
0.68 0.9
CASI
600 reflect
20 650
700 ance
750
800
achiev
ing
casi 8:00
10:30 r2=0.9
1.5 13:00
15:30 5.
18:00
0.96
/ 20:30 Maple
(R675*690) 23:00 seedlin
Wavelength (nm)
gs
were
Figure 4. Effect of dark-
the CF variation in a adapte
time-decay d for
experiment in the Time
(h) 15
600-800 nm canopy minute
(%) reflectance. CASI s prior
canopy reflectance to
measurements were readin
made from Acer Figure gs of
saccharum M. 5. Fv/Fm
seedlings in the Diurna (adapt
laboratory taken l ed
after dark adaptation variati from
1 and after 3 minutes ons of Zarco-
of illumination. Fv/Fm Tejada
Differences in and et al.,
reflectances at 680- 1 the in
690 nm and 730-750 optical press,
nm are observed due index a).
to changes in
0.94 R6852/
chlorophyll
(R675
fluorescence.
Second EARSeL Workshop on Imaging Spectroscopy, Enschede, 2000

Results in the diurnal

study using the CASI
sensor in the

indices in the 680-690 nm region track

changes in Fv/Fm;

laboratory with canopy

seedlings showed that
optical

0.01
R680/R630,
R685/R630,
R687/R630
and
R690/R630,
@ 8.20-13.3
0.005

-0.016

-0.021
Difference

achieve correlation coefficients r2=0.93,

r2=0.94, r2=0.92,
 the reflectance curvature in the 675-690 nm
region (Zarco-

-0.026

and r2=0.91, respectively.

Indices sensitive to
changes in

Tejada et al., in press, a, b), also showed

correlation with
Reflectance

670

680
690
700
710

Reflectance

diurnal
changes in Fv/Fm,
i.e.
curvature
index

-0.005

-0.031

R6852/(R675690)
(Figure 5) yields r2=0.95.

4.1 Above-Canopy Data Collection

Experimental
Difference
-0.01

-0.036
0.9-20.8@
Results Under Natural
Illumination Conditions

Effects of CF on apparent
reflectance were found in
canopy reflectance measurements with
-0.015
a fibre

-0.041
 plot), shows that reflectance difference
spectrometer. Results of between 8.20h and
the diurnal experiments
over a

-0.02

-0.046

13.30h is higher than reflectance difference

between 8.20h

rfl diff 8.20-13.37

rfl diff 8.20-9.02

canopy of Acer
Saccharum M. seedlings
(Figure 6, upper

Wavelength (nm)
 of CF is higher in the early morning than in
mid-day,

0.014

and 9.00am, consistent

with previous diurnal
experiments.

when light saturation reduces the CF, and

therefore its

The maximum in the 690-

700nm region due to the
effects
effects
on
the apparent
reflectance.
Dark-adapted CF

pm
0.012

variation in Fv/Fm and Ft during the day,

decreasing in the

0.01

measurements from the

maple seedlings showed
a

afternoon:
Fv/Fm=0.574,
Ft=0.495
at
9:09h;

12.20
8.00 am

with artificial light and laboratory

experiments, showing

Rfl
0.006

Fv/Fm=0.524, Ft=0.418
at 11:22h; and
Fv/Fm=0.516,

0.008

that CF changes can be tracked under

natural illumination

Ft=0.396 at 13:23h. These

results demonstrate
consistency

diff
 670

680
690
700
710

hyperspectral sensor over forest sites in

1999 and 2000

conditions using a fibre

spectrometer over small
canopies.

0.004

Wavelength (nm)

revealed promising but less convincing

results. Analysis

Subsequent experiments
with the airborne CASI
 reflectance at 700nm due to the higher CF
effects in the

results from one of the

healthy forest sites are
shown in
Figure 6. Reflectance
difference obtained in

Figure 6 (lower plot)

where reflectance in the
72-channel
the diurnal experiments
over a canopy of Acer

CASI mode and CF

measurements were
collected at
Saccharum M.
seedlings

(upper
plot, early morning before photosystem
saturation. Ground

different times of the day. truth CF measurements showed a decrease

Reflectance difference in the non-dark-adapted Ft: Ft=0.74 at
between 8.00h, Ft=0.72 at 9.30h, Ft=0.46 at 12.30h,
8.20- 13.30) and dark-adapted Fv/Fm: Fv/Fm:
using Fv/Fm=0.83 at 8.00h, Fv/Fm=0.82 at 9.30h,
a Fv/Fm=0.79 at 12.30h. The shift in the
fiber maximum of the reflectance difference from
spectrometer, 695 nm at leaf level to 700 nm at CASI
canopy level is under investigation. Results
and from the experiments in June 2000 using
higher spatial resolution of 0.56x1.08m and
a CASI bandset of 9 bands centered in the
CASI measurements at 695 nm spectral region did not show clear
8.00h and 12.20h shows and definitive results. The effect of the
a atmosphere with the water vapour and O2
from one of the healthy bands in the 690-700 nm region was an
forest sites (lower issue investigated for both cases in which
different sensors and data correction was
carried out. For the measurements with the
fibre spectrometer, a Spectalon panel
maximum at 700nm,
measurement collected every time
consistent with higher
facilitated the elimination of atmospheric
value of
features in the reflectance data. For the
plot, 8.00- CASI data collected at different times, a
12.20) collected with post-processing refinement after
CASI. atmospheric correction was performed in
order to remove residual correction errors in
the 670-710 nm region by selecting road
spectra from the images. Correction factors
were applied to the data using a flat field
approach, removing residual atmospheric
effects in the difference spectra. Correlation
coefficients were 1999 through SAIL and Kuusk CR models
calculated from the (Zarco-Tejada, in press, a, b). Results show
estimation of Fv/Fm whenthat the indices producing the best
leaf-level relationships estimations of CF were the Curvature index,
between CF and optical derivative indices such as DP22, DPR1,
indices are applied to the DPR2, and DP21 (Zarco-Tejada et al.,
12 sites of Acer 1999a), and indices in the visible such as
saccharum M. collected PRI3 (R570-R539)/(R570+R539), and
with CASI in 1998 and
R440/R690.
Second EARSeL Workshop on Imaging Spectroscopy, Enschede, 2000

ACKNOWLEDGEMENTS

5 Research support for this work is gratefully

acknowledged from GEOIDE, a Canadian
CONCLUSIONS Networked Centre of Excellence, and the
Centre for Research in Earth and Space
Technology.
This paper shows results
obtained at different
levels of study which
demonstrate that CF REFERENCES
signature affects the
reflectance measurements Heinz-Walz-GmbH, 1993. Portable
acquired at leaf and Fluorometer PAM-2000 and data
canopy levels. acquisition software DA-2000, Effeltrich,
Experiments presented Germany.
here focussed on the
study of the effect of CF Mohammed, G. H., Binder, W. D., and
on the canopy reflectance Gillies, S. L., 1995. Chlorophyll
signature under natural fluorescence: A review of its practical
illumination conditions, forestry applications and instrumentation.
as a continuation of Scand. J. For. Res. 10:383-410.
previous work carried out
in the laboratory and Mohammed, G. H., Sampson, P. H.,
through modelling using a Colombo, S. J., Noland, T. L., and Miller, J.
radiative transfer R., 1997. Bioindicators of forest
simulation and the CASI sustainability: Development of a forest
hyperspectral sensor. condition rating system for Ontario project
Diurnal trials using a fibre strategy, Report 137, Ontario Forest
spectrometer over a Research Institute, Sault Ste. Marie, ON.
canopy of seedlings
produced positive results,
showing that canopy Sampson, P. H., Mohammed, G. H.,
reflectance is affected by Colombo, S. J., Noland, T. L., Miller, J. R.,
diurnal variation of CF, Zarco-Tejada, P. J., 1998. Bioindicators of
and that such reflectance forest sustainability progress report, Report
changes are measurable 142, Ontario Forest Research Institute, Sault
under natural illumination Ste. Marie, ON.
conditions. However,
airborne experiments with Subhash, N. and Mohanan, C. N., 1997.
the CASI hyperspectral Curve-fit analysis of chlorophyll
sensor produced fluorescence spectra: Application to nutrient
promising but stress detection in sunflower. Remote Sens.
inconclusive results in Environ. 60:347-356.
two diurnal experiments
carried out in 1999 and Zarco-Tejada, P. J., Miller, J. R.,
2000, in which small Mohammed, G. H., Noland, T. L. and
variations of reflectance Sampson, P. H., 1999a. Canopy Optical
due to the effect of CF Indices from Infinite Reflectance and
were observed. The real Canopy Reflectance Models for Forest
applicability of these Condition Monitoring: Application to
results in vegetation Hyperspectral CASI data, In IEEE 1999
canopies is still under International Geoscience and Remote
investigation due to the Sensing Symposium, IGARSS'99, Hamburg
difficulties associated (Germany).
with the small changes in
reflectance due to CF and Zarco-Tejada, P. J., Miller, J. R.,
the spectral regions where Mohammed, G. H., Noland, T. L. and
they occur, requiring a Sampson, P. H., 1999b. Optical Indices as
very accurate atmospheric Bioindicators of Forest Condition from
correction. Hyperspectral CASI data, In Proceedings
19th Symposium of the
European Association of
Remote Sensing
Laboratories (EARSeL),
Valladolid (Spain)

Zarco-Tejada, P. J.,
Miller, J. R., Mohammed,
G. H., Noland, T. L., (in
press, a). Chlorophyll
Fluorescence Effects on
Vegetation Apparent
Reflectance: I. Leaf-Level
Measurements and Model
Simulation, Remote
Sensing of Environment.

Zarco-Tejada, P. J.,
Miller, J. R., Mohammed,
G. H., Noland, T. L and
Sampson, P.H., (in press,
b). Chlorophyll
Fluorescence Effects on
Vegetation Apparent
Reflectance: II.
Laboratory and Airborne
Canopy-Level
Measurements with
Hyperspectral data,
Remote Sensing of
Environment.