United States
Environmental Protection
Agency
Environmental Monitoring
Systems Laboratory
P.O. Box 15027
Las Vegas NV89114
EPA-600/4-79-048
August 1979
Research and Development
c/EPA
Laser
Fluorosensing of
Surface Water
Chlorophyll a
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad categories
were established to facilitate further development and application of environmental
technology. Elimination of traditional grouping was consciously planned to foster
technology transfer and a maximum interface in related fields. The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL MONITORING series.This series
describes research conducted to develop new or improved methods and instrumentation
for the identification and quantification of environmental pollutants at the lowest
conceivably significant concentrations. It also includes studies to determine the ambient
concentrations of pollutants in the environment and/or the variance of pollutants as a
function of time or meteorological factors.
This document is available to the public through the National Technical Information
Service, Springfield, Virginia 22161
-------
EPA-600/4-79-048
August 1979
AIRBORNE LASER FLUOROSENSING OF
SURFACE WATER CHLOROPHYLL a
by
M. Bristow, D. Nielsen and D. Bundy
U.S. Environmental Protection Agency
Environmental Monitoring Systems Laboratory
Las Vegas, Nevada 89114
R. Furtek and J. Baker
Department of Biological Sciences
University of Nevada, Las Vegas
4505 Maryland Parkway
Las Vegas, Nevada 89154
ENVIRONMENTAL MONITORING SYSTEMS LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
LAS VEGAS, NEVADA 89114
-------
DISCLAIMER
This report has been reviewed by the Environmental Monitoring Systems
Laboratory, U.S. Environmental Protection Agency, and approved for
publication. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
11
-------
FOREWORD
Protection of the environment requires effective regulatory actions that
are based on sound technical and scientific information. This information
must include the quantitative description and linking of pollutant sources,
transport mechanisms, interactions, and resulting effects on man and his
environment. Because of the complexities involved, assessment of specific
pollutants in the environment requires a total systems approach that
transcends the media of air, water, and land. The Environmental Monitoring
Systems Laboratory-Las Vegas contributes to the formation and enhancement of a
sound monitoring data base for exposure assessment through programs designed
to:
•develop and optimize systems and strategies for monitoring pollutants
and their impact on the environment
•demonstrate new monitoring systems and technologies by applying them
to fulfill special monitoring needs of the Agency's operating
programs
This report describes the development of an airborne laser fluorosensor
for mapping the distribution of chlorophyll a_ in the surface waters of lakes,
rivers, reservoirs, estuaries and coastal waters. Remote monitoring devices
of this kind will be invaluable for rapidly assessing the concentration of
surface water algae for a given water body and, in particular, for locating
anomalies, gradients and blooms produced by anthropogenic activity. Further
information concerning the development of this system can be obtained from the
Advanced Systems Branch, Advanced Monitoring Systems Division.
.
'Geopg'e M, MorgarjX
Director
Environmental Monitoring Systems Laboratory
Las Vegas
-------
FIGURES
Number Page
1 Principle of operation of airborne'laser fluorosensor 7
2 Schematic illustrating'possible mode of operation of an airborne
laser fluorosensor for mapping surface water chlorophyll a^
distributions 11
3 Optical diagram of airborne laser-fluorosensor for monitoring
surface water fluorescence signal . . 17
4 Dye and water cooling flow diagram for coaxial flash!amp-
pumped dye laser employed in airborne laser-fluorosensor ... 21
5 Gate and voltage divider circuit diagram for RCA C31000A
photomultiplier 23
6 Schematic of airborne laser fluorosensor for measuring
chlorophyll jj fluorescence showing detection, monitoring
and recording systems .' 24
7 Oscillogram showing sequence of airborne laser-fluorosensor
signals obtained over buoy #12 on October 4, 1976, for a
measured chlorophyll a^ concentration of 10.5 yg/1 26
8 Nautical chart of Las Vegas Bay region of Lake Mead, Nevada,
showing location of marker buoy sampling stations 29
9 Corrected fluorescence emission spectra of filtered and
unfiltered Lake Mead surface water sample, excited at 440 nm,
(August 16, 1977) 31
10 Variation of surface water chlorophyll a_ and laser-fluorosensor
signal with distance for surface water transect of'Las Vegas
Bay in Lake Mead, Nevada (flight #4, October 15, 1976) .... 34
11 Variation of surface water chlorophyll a^ and1 laser-fluorosensor
signal with distance for surface water transect of Las Vegas
Bay in Lake Mead, Nevada (flight #8, November 19, 1976). ... 35
VI
-------
CONTENTS
Disclaimer ii
Foreword iii
Summary iv
Figures vi
Tables viii
Abbreviations and Symbols ix
1. Introduction 1
2,
Conclusions 2
3. Recommendations 3
4. Review 5
5. Chlorophyll _a Monitoring with Airborne Laser Fluorosensors ... 12
6. Instrumentation 16
Laser transmitter system 18
Optical receiver system 22
Electronic monitoring and recording system 22
7. Airborne Measurements 28
Field operations 28
Analysis of airborne laser-fluorosensor data 30
Analysis of ground truth water samples . 32
Comparison between airborne and ground truth data .... 33
Factors influencing the relationship between airborne
and ground truth data . 38
Remote monitoring of blue-green algae 46
8. Airborne Measurements of the Influence of the
Attenuation Coefficients » 51
Theoretical considerations ...... 51
Laboratory measurement of fluorescence-to-Raman ratio for
lake water samples . 53
Modifications to laser-fluorosensor needed to measure the
fl uorescence-to-Rainan ratio 58
References . ................ 62
-------
FIGURES
Number Page
1 Principle of operation of airborne laser fluorosensor 7
2 Schematic illustrating possible mode of operation of an airborne
laser fluorosensor for mapping surface water chlorophyll a^
distributions 11
3 Optical diagram of airborne laser-fluorosensor for monitoring
surface water fluorescence signal 17
4 Dye and water cooling flow diagram for coaxial flash!amp-
pumped dye laser employed in airborne laser-fl uorosensor ... 21
5 Gate and voltage divider circuit diagram for RCA C31000A
photomultipl ier 23
6 Schematic of airborne laser fluorosensor for measuring
chlorophyll a_ fluorescence showing detection, monitoring
and recording systems 24
7 Oscillogram showing sequence of airborne laser-fl uorosensor
signals obtained over buoy #12 on October 4, 1976, for a
measured chlorophyll _a concentration of 10.5 yg/1 26
8 Nautical chart of Las Vegas Bay region of Lake Mead, Nevada,
showing location of marker buoy sampling stations 29
9 Corrected fluorescence emission spectra of filtered and
unfiltered Lake Mead surface water sample, excited at 440 nm,
(August 16, 1977). 31
10 Variation of surface water chlorophyll _§_ and laser-fl uorosensor
signal with distance for surface water transect of Las Vegas
Bay in Lake Mead, Nevada (flight #4, October 15, 1976) .... 34
11 Variation of surface water chlorophyll j[ and laser-fluorosensor
signal with distance for surface water transect of Las Vegas
Bay in Lake Mead, Nevada (flight #8, November 19, 1976). ... 35
VI
-------
FIGURES (continued)
Number Pa9e
12 In vivo chlorophyll a^ fluorescence and water transmission
profiles obtained from surface water transect of Las Vegas
Bay in Lake Mead, Nevada (June 8, 1977). Also shown are
beam-attenuation coefficient values calculated for 25 points
in the transmission profile. Transmission data measured at
610 nm 42
13 Variation of beam-attenuation coefficient with in vivo
chlorophyll a_ fluorescence for surface water transect of
Las Vegas Bay in Lake Mead, Nevada (June 8, 1977).
Attenuation data measured at 610 nm 43
14 Variation of laser fluorosensor signal, extractable
chlorophyll _§_, total biomass and algal color group biomass
with distance for surface transect of Las Vegas Bay in
Lake Mead, Nevada (flight #12, August 16, 1977) 48
15 Laboratory simulation of airborne laser-fluorosensor for
monitoring chlorophyll _a fluorescence and water Raman
signals 54
16 Uncorrected fluorescence and Raman emission spectra of Lake Mead
surface water sample, excited at 440 nm. (August 16, 1977). . 56
17 Uncorrected fluorescence and Raman emission spectra of Lake Mead
surface water sample, excited at 438 nm. (March 31, 1978) . . 57
18 Optical diagram of airborne laser-fluorosensor for monitoring
chlorophyll a_ fluorescence and water Raman signals 59
19 Schematic of airborne laser-fluorosensor for measuring
chlorophyll _a fluorescence and water Raman emission showing
detection, monitoring and recording systems ... 60
20 Variation of laser, fluorescence and Raman laser-fluorosensor
signals as a function of time ........... 61
vn
-------
TABLES
Number
1 Airborne Laser Fluorosensor Characteristics
Laser Dyes Evaluated for Use in Coaxial Flash!amp-Pumped Dye
Laser Employed in Airborne Laser-Fluorosensor 20
Correlation Coefficients for Corrected Laser-Fluorosensor
Signal Versus Turbidity and Chlorophyll j3 Data 36
Correlation Coefficient Matrix for Algal Biomass Obtained by
Enumeration Versus Laser-Fluorosensor Signal and Extracted
Chlorophyll a^ for a Group of 13 Samples 49
vm
-------
LIST OF ABBREVIATIONS AND SYMBOLS
ABBREVIATIONS
EMI
f/number
FLD
FWHM
|OH|
ppb
pps
r
RFI
RMS
SBNR
s
yv
X
6
Electromagnetic interference
Ratio of focal length to diameter of optical system aperture
Fraunhofer line discriminator
Full width at half maximum height
Notation for hydroxyl radical, as in water molecule
Parts per billion
Pulses per second
Pearsons product moment linear correlation coefficient
Radio frequency interference
Root mean square value
Signal to background noise ratio
Sample standard deviation
Ultraviolet
Sample mean
Laser beam divergence
SUBSCRIPTS
c
f
F
L
N
PP
P
R
Rec
RMS
Tr
W
X
Backward scatter
Chlorophyll a_
Forward scatter
Fluorescence
Laser
Non-chlorophyllous material
Peak to peak
Peak
Raman
Receiver
Root mean square
Transmitter
Water
Wavelength
-------
LIST OF ABBREVIATIONS AND SYMBOLS (continued)
SYMBOLS
a -- Absorption coefficient
A -- Intercept constant in Equation 9
b -- Scattering coefficient
B -- Slope constant in Equation 9
d -- Terms defined by Equations 16 and 18
D -- Dimensionless radiance distribution factor
H -- Target range
k -- Water diffuse attenuation coefficient
n -- Concentration
P -- Peak power
R -- Reflectivity
T -- Effective area of telescope aperture
v -- Noise signal voltage
V -- Peak signal voltage
x -- Optical path length
X -- Optical attenuation terms, defined by Equations 13 and 14
a -- Water beam attenuation coefficient
3 -- Atmospheric attenuation coefficient
6 -- Constant in Equation 20
A -- Fraction of emission band seen by detector
T? -- System optical efficiency
n -- Refractive index
°" -- Excitation cross section
-------
SECTION 1
INTRODUCTION
In fresh-water environments a high concentration of chlorophyll _a
indicates high levels of planktonic algae. This condition suggests, in turn,
the existence of eutrophic conditions, that situation in which high
concentrations of nutrients create algal bloom conditions. Such conditions
can lead to malodorous and even toxic water conditions. Conversely,
anomalously low chlorophyll a^ levels might indicate the presence or influence
of toxic pollutants. Chlorophyll a_ is also of importance in the marine
environment. Phytoplankton are the principle producers and suppliers of
energy in the marine ecosystem. They are the principal food source for the
herbivores such as zooplankton, which in turn are an important food source of
the carnivores characterized by the fish population. As a consequence,
information relating to phytoplankton productivity is of considerable interest
to the fisheries industries.
Chlorophyll ^determinations are routinely accomplished by laboratory
analyses on grab samples. This approach is both time-consuming and costly in
terms of manpower and facilities. In addition, because of the finite time
required to take grab samples from launches or helicopters, it is not always
possible to obtain a synoptic record for a given water surface due to water
movement and diurnal effects. Also, delays in transporting water samples to
the laboratory for analysis or the practice of freezing algae samples
collected on filters are known to produce large errors due to irreversible
degradation of the chlorophyll a_. In contrast, both airborne and satellite
remote-monitoring techniques are capable of rapidly and cost effectively
providing data for certain water quality parameters from large areas of water
surface without influencing the nature of the sample. The principal
limitation of remote sensing is its inability to provide information from
below the photic zone, commonly defined as that region from the surface to the
depth at which 99% of the surface light has disappeared. For active as well
as passive sensors operating in the optical region, this depth is generally on
the order of 0.5 to 10 meters and is conveniently characterized by the
reciprocal of the attenuation coefficient measured at a given wavelength.
The purpose of the present program is to develop a compact, integrated
airborne system capable of mapping trends, gradients and anomalies in the
distribution of surface water chlorophyll a. present as planktonic algae.
-------
SECTION 2
CONCLUSIONS
The laser-fluorosensor described in this report has been used to produce
flight-line profiles of relative surface water chlorophyll a_ concentrations
that vary over a range from 2 micrograms per liter (yg/1) to 20 pg/1 at
operating elevations from 200 meters (m) to 400 m. In its present form the
laser fluorosensor is estimated to be capable of monitoring chlorophyll a_
concentrations down to 0.4 yg/1 with a minimum signal to background noise
ratio of 3. With the implementation of measures to increase the fluorescence
signal and reduce the background noise, it will be possible to monitor
chlorophyll c[ concentrations down to 0.1 yg/1 with a minimum signal to
background noise ratio of 20 under the same environmental conditions
encountered in the present measurements.
Attempts to monitor changes in the concentration of blue-green algae by
selectively exciting c-phycocyanin, a chlorophyll a-coupled photopigment
unique to this freshwater algal division, were not successful. It is likely
that the reason for this failure is the relatively low fluorescence cross
section for this pigment in relation to that for chlorophyll a^ itself. It is
possible that monitoring specific algal groups can be better accomplished by
employing the multiwavelength excitation approach.
-------
SECTION 3
RECOMMENDATIONS
The fully developed airborne laser fluorosensor will be able to produce
contour maps of surface water chlorophyll d. fluorescence. Due to the problems
encountered in determining absolute values for the chlorophyll a_ fluorescence
cross section and the diffuse water attenuation coefficients, it is not
possible to produce a calibrated surface water chlorophyll a_ concentration map
using airborne data alone. Rather this map must be calibrated in units of
chlorophyll a_ concentration by the provision of ground truth chlorophyll _a
determinations obtained at one or two selected reference sites.
Presently, the degree of correlation between the airborne chlorophyll jj
fluorescence profiles and those for the corresponding ground truth chlorophyll
a^ data is somewhat less than ideal, with linear correlation coefficients lying
in the range between 0.77 and 0.95. Several reasons exist for this
discrepancy between the airborne and ground truth data. Firstly, chlorophyll
a_ determinations made according to presently accepted standards are both
inefficient and of low reproducibility, whereas it is estimated that the
airborne measurements, although prone to systematic errors, are reproducible
to better than ±5%. It is therefore suggested that investigations be
conducted with the purpose of improving both the efficiency and accuracy of
the method used for making chlorophyll a_ determinations at least to the point
where the error is comparable to that for the airborne data. Secondly, the
airborne chlorophyll _a fluorescence signal is known to be dependent on changes
in the surface water diffuse attenuation coefficients in addition to the
chlorophyll ^concentration. It is therefore proposed to obtain a relative
indication of this variation by concurrently monitoring the water Raman
emission signal. Theoretical and experimental considerations indicate that
this approach is feasible. Modifications to the airborne laser-
fluorosensor needed for measuring this Raman signal are straightforward and
can be readily implemented with a minimum of modification to the present
system. Finally, the dependence of the excitation cross section for
chlorophyll _a,
-------
(1) Rapid Installation in a variety of different aircraft will require a
system considerably smaller than the present package which can be quickly and
easily installed and either palletized or constructed of a number of discrete
modular units suitable for rapid transportation and integration in the field.
(2) Provisions must be made for navigation data accurate to within a few
meters in order that the airborne data can be reproduced in the form of a
surface water map showing isopleths of constant chlorophyll ji fluorescence.
This can be accomplished by using an airborne range-positioning system which
employs a microwave transmitter to interrogate two portable ground-based
transponder units.
(3) A degree of real-time capability should be built into future laser
fluorosensors for use as a quality control feature. A contour map of surface
water chlorophyll a_ could then be produced and continuously updated on a
cathode ray tube display for evaluation by the system operator. Current
microprocessor technology is ideally suited to this application.
(4) Finally, the optical receiver for the laser fluorosensor will contain
two extra spectral detector channels, in addition to those for the water Raman
and chlorophyll a_ fluorescence signals. These additional detectors will
monitor spectral intervals close to the water Raman and chlorophyll a^
fluorescence bands, thereby providing estimates of the background fluorescence
due to dissolved organics at these two bands. These additional measurements
are necessary whenever the concentration of dissolved organics is large or
highly variable.
-------
SECTION 4
REVIEW
Two passive remote-sensing methods are presently under development for
monitoring distributions of surface water fluorescent tracer dyes and
chlorophyll a_.
In the first approach, photographic cameras or multichannel radiometric
scanners are used to monitor the backscattered solar radiation in the visible
and near infrared regions. These techniques generally provide imagery of high
resolution with excellent spatial coverage, which can be used to obtain
qualitative information on the location and extent of algae blooms, suspended
sediment concentrations such as effluent plumes, oil spills, pollution-dumping
sites and tracing dye dispersions. However, this approach has achieved only
limited success in predicting the concentration of water quality parameters,
in particular, chlorophyll a^ and suspended sediment. Much effort has been
expended on formulating interpretation procedures for extracting information
on chlorophyll a_, e.g., references 1, 2 and 3, and suspended sediment content
of surface waters, e.g., references 4, 5 and 6. With the aid of ground truth
data, empirical interpretation schemes have been devised which are able to
predict these parameters for waters containing particulates and algal
distributions of a specific color, generally for a limited geographical
region, but are seen to break down when applied to different water conditions.
As yet, no universal model exists which can provide either absolute or
relative chlorophyll ^concentrations from passive radiometric data. These
problems arise in part from the spectral overlap of the radiation
backscattered by algae and particulate matter, in part from the differing
nature of the spectral and scattering properties of different types of algae
and suspended sediments, and in part from the presence of dissolved materials
that impart a characteristic color to the water. In addition, the limited
optical transmission of water, particularly in the near infrared, atmospheric
backscatter, and surface water specular reflection of sky radiation further
complicate the interpretation of the photographic or radiometric data.
Finally, for successful operation of these devices, clear sky conditions in
the absence of solar glitter are generally required. For passive airborne
measurements made under cloud cover, signal-to-noise ratios are considerably
reduced.
In the second approach, two airborne sensing techniques are worth
mentioning with regard to their ability to remotely monitor surface water
chlorophyll £. Both methods are passive but, rather than record changes in
the backscattered solar spectrum, they detect variations in the solar-induced
chlorophyll ^fluorescence emission band at 685 nanometers (nm) that, ideally,
can be directly related to the chlorophyll d_ concentration. Neville & Gower
(7) used a multichannel spectrometer to monitor the ratio of the solar-induced
-------
fluorescence emission from the chlorophyll a^ band at 685 nm to the incident
irradiance. Variations in airborne fluorescence measurements and the
corresponding chlorophyll a_ ground truth data were found to be highly
correlated.
The other passive fluorosensing approach presently under development
employs a Fraunhofer Line Discriminator (FLD) to monitor the solar-induced
target fluorescence observed in the region of the minimum of a solar
Fraunhofer line (8). This device has been flown in a helicopter to map the
surface water concentration of Rhodamine WT fluorescent tracing dye in the
parts-per-billion (ppb) range (8) and more recently has been used to map the
area! extent of marshland contaminated by an oil well blowout (9). When
operated in the scanning mode, spatial coverage can be achieved by making a
single pass over a target with navigation data provided by conveniently
located ground-based transponder units. Surface water chlorophyll monitoring
has not yet been attempted with this device, although it has been used to
monitor the changes in chlorophyll a_ fluorescence of tree vegetation induced
by geochemical soil anomalies (10). As with all passive remote-sensing
devices operating in the optical region, it is dependent on bright daylight
conditions for a usable signal-to-noise ratio. In addition, the specificity
of the FLD technique is reduced both by the limited number and distribution of
usable Fraunhofer lines and by the broad spectral nature of the solar source.
Recently, active remote monitoring systems employing lasers have been used
to excite, detect and record the fluorescence of surface water targets. These
airborne laser-fluorosensor systems operate by exciting fluorescence in a
surface water volume, using a pulsed laser, and by collecting a small fraction
of the multidirectional fluorescence emission, using a large-aperture
telescope sighted on the excitation spot. The backscattered fluorescence
pulse is then passed through a spectral analyzer onto a photodetector, and the
resultant electronic signal is displayed, digitized and recorded. The
schematic shown in Figure 1 illustrates the general optical principle involved
in which the laser spot size might be on the order of 1 m to 2 m in diameter
from a platform elevation of 200 m. Laser beams are highly directional, can
have high average or peak power, and low beam divergence, and can be chosen or
tuned to operate at any desired visible or ultraviolet (UV) excitation
wavelength. Using laser excitation sources, acceptable signal-to-solar
background-noise ratios can be obtained so that laser fluorosensors can be
operated on a 24-hour basis. Three fluorescence parameters can be measured
using this technique. Firstly, either the amplitude of the fluorescence
signal, generally at the wavelength of peak emission, can be related to the
concentration of a particular target of known fluorescence quantum efficiency,
or, alternatively for an opaque or optically thick target, the quantum
efficiency can be measured and used as an identifying feature. Secondly, the
fluorescence emission spectra can be monitored and recorded at a number of
discrete wavelengths. In addition, the fluorescence absorption or excitation
spectrum can, in principle, be determined using a dynamically scanned tunable
laser source, although such systems are not yet commercially available.
Finally, the fluorescence decay time spectrum can be measured as a function of
wavelength. All of these parameters are unique for a given fluorescence
target and in combination constitute an unambiguous signature with which to
-------
Laser
transmitter
/Telescope receiver
Receiver
field of view
Blue or UV
laser beam
Fluorescence emission
from target
Surface water
fluorescent
target
Laser excited
target area
Figure 1. Principle of operation of airborne laser fluorosensor.
7
-------
Identify or characterize the target in relation to the fluorescent background
or to competing fluorescent targets.
Airborne laser fluorosensors have been used to detect surface water oil
spills (11, 12, 13) and in one case to obtain the fluorescence emission
spectra from a number of oil slicks (14). They have been successfully used to
remotely profile surface water chlorophyll a_ present in phytoplankton (15, 16)
and finally they have demonstrated a potential for remotely monitoring the
highly fluorescent lignin materials in pulp and paper mill effluents (12, 13).
The feasibility of remotely monitoring algae and phytoplankton with laser-
induced fluorescence was fi-rst proposed by Hickman and Moore (17). They
demonstrated that an acceptable signal-to-noise ratio could be obtained for
the 635-nm chlorophyll a_ fluorescence emission in relation to the solar
background using existing laser technology. Mumola and Kim (18) were the
first to operate a laser fluorosensor from an airborne platform in which a
flash!amp pumped dye laser operating at 590 nm was used to excite chlorophyll
a_ fluorescence at 685 nn. Comparison of the airborne in vivo chlorophyll a_
Tluorescence measurements with corresponding ground truth chlorophyll a^ data
produced a linear correlation coefficient of 0.43* for a sample size of 19 for
which the average laser-fluorosensor measurement was 11% greater than that for
the ground truth data.
In many situations the ratio of in vivo fluorescence to extractable
chlorophyll ^a has been observed to vary over as much as a tenfold range (19).
The reason for this variability can be explained in terms of the fluorescence
cross section (or excitation coefficient) for a given substance, which is a
measure of the quantity of incident excitation radiation converted to
fluorescence emission. The fluorescence cross section for chlorophyll _a when
measured in vivo is known to be dependent on a number of factors including
water temperature, stress induced by toxic substances or lack of nutrients,
the presence of photopigment degradation products, the intensity of the solar
background, the intensity and wavelength of the excitation radiation, the
duration of the excitation radiation, and the relative concentrations of the
different algae color groups present (16, 19, 20, 21, 22, 23, and 24).
In vivo fluorometry performed on algae has generally employed an
excitation wavelength of about 436 nm located close to the center of the
blue-violet (Soret) absorption band of chlorophyll a_. However, data presented
by Friedman &Hickman (25) and by Mumola et al . (16) showed not only that the
absolute value of the fluorescence cross section varies considerably between
algal divisions or color groups but also that these cross sections exhibit
considerable variability with wavelength. This problem arises because of the
presence of different light-absorbing photosynthetic pigments among the
different algal color groups. For example, blue-green algae or cyanophyta,
often employed as indicators of eutrophic conditions in fresh water, contain,
in addition to significant amounts of chlorophyll a^ and p-carotene, the blue
pigment c-phycocyanin. This substance effectively blocks light absorption by
the various chlorophyll and carotene pigments in the blue-green region and has
*The linear correlation coefficient was calculated by the present authors from
graphical data presented by Mumola&Kim (18).
-------
a strong absorption band centered at 622 nm. This red radiation is then
internally coupled directly into the chlorophyll jj photosynthetic system. It
therefore becomes clear that large changes in relative concentration of the
different algal color groups on either a spatial or a temporal basis can be
expected to have a significant effect on the accuracy of the predicted
chlorophyll ^ concentration when only a single excitation wavelength is used.
This will be true regardless of whether the in vivo measurements are made with
a field fluorometer or an airborne laser-fluorosensor. Attempts to circumvent
this problem have been made by employing an airborne laser-fluorosensor which
sequentially operates at several excitation wavelengths (16,21). The number
of laser wavelengths required is dictated by the number of algal color groups
anticipated to be present in the surface water sample, with the wavelength of
each laser tuned to lie close to the peak absorption wavelength for the given
algal pigment. With a knowledge of the laser power, the water attenuation
coefficients, and the fluorescence cross sections for each algal color group
at each laser wavelength together with the fluorescence power levels at 685 nm
excited by each laser, it is, in principle, possible to solve a series of
laser fluorosensor equations to obtain the equivalent chlorophyll _a
concentration for each algal color group. On a test flight over the tidal
flow region of the James River in Virginia, the 4-wavelength laser-
fluorosensor was used to measure chlorophyll a^ concentrations for each of the
golden, brown, green and blue-green algal color groups (16). Comparison of
the total laser fluorosensor chlorophyll _a_ values with corresponding ground
truth produced a linear correlation coefficient of 0.90* for a sample size of
7. In this case the average laser-fluorosensor chlorophyll _a value was
approximately 28% below that for the ground truth data. A review of this
multiwavelength approach by Browell (20) demonstrated that relatively small
uncertainties in the values for the water attenuation coefficients and the
algal color group fluorescence cross sections undergo considerable
amplification when used to calculate the total in vivo chlorophyll a_ values.
More recently, laboratory studies performed on known mixtures of four
carefully controlled algal monocultures (each from a different color group)
using the 4-wavelength laser fluorosensor, have obtained a linear correlation
coefficient of 0.99 for covariation between the in vivo laser fluorosensor
chlorophyll ^measurements and the corresponding extractable chlorophyll a^
data (21). However, the stable laboratory conditions employed for these
measurements are neither typical nor representative of the true field
environment. A number of unpredictable factors not present in these
measurements are known to have a significant influence on the algal color
group fluorescence cross sections. These are stress induced by nutrient
limitations, the effect of toxic substances, changes in water temperature, age
of the algal communities, the intensity of the solar background radiation and
the wavelength, intensity and duration of the fluorescence excitation light
source (19, 20, 22). In the absence of either in situ or remotely sensed data
relating to the water attenuation coefficient and algal fluorescence cross
sections, Browell (20) has suggested that both single and multi-wavelength
laser fluorosensors are, at best, capable of providing only qualitative or
relative measure of surface water chlorophyll a_ concentration.
*The linear correlation coefficient was calculated by present authors from
graphical data presented by Mumola et al. (16).
-------
In light of these observations and with a view to reducing system
complexity, a program has been initiated to design, build and test a single
excitation wavelength airborne laser-fluorosensor. The remotely sensed
fluorescence signature will then be used to produce surface water maps showing
isopleths of surface water chlorophyll a_ concentration as illustrated in
Figure 2. Whereas field fluorometers almost universally employ a wavelength
of about 440 nm to excite fluorescence in planktonic algae (24), other
studies, based on laboratory investigations performed on water samples and
algal monocultures suggest that wavelengths in the 600-nm region are more
suitable for monitoring the overall chlorophyll a_ level for disparate mixtures
of algae (15, 16, 25). In particular, from the measurements made on specific
algal color groups presented in Reference 16, it is apparent that at 620 nm
the fluorescence excitation cross sections for the green, blue-green, red and
golden brown algal color groups are all approximately equal. With these
observations in mind, a flexible approach was adopted allowing different
excitation wavelengths to be evaluated with a view to optimizing both the
degree of correlation attainable between the airborne and ground truth data
and also system sensitivity. In particular it was planned to employ a
wavelength of 440 nm to facilitate comparison between the airborne data and
the fluorometrically determined ground truth data and also a wavelength of 620
nm for the reasons stated above.
The principal advantage of the fluorometric approach, whether active or
passive, over the reflectance spectroscopy approach is its provision of raw
data directly proportional to chlorophyll a^ concentration. The data can then
be analyzed without the need for empirical interpretation models established
beforehand with ground truth data. In this respect the airborne
laser-fluorosensor is similar in principle and operation to field fluorometers
that are widely used for in situ monitoring of both surface and subsurface in
vivo chlorophyll _a (19, 24, 26). The principal difference between these two
fluorometric techniques is that the in situ system employs a constant
pathlength cell, whereas, with the remote sensing approach, the effective
sample pathlength and resultant fluorescence signal are dependent on the
variable optical attenuation lengths for the laser and fluorescence
wavelengths. A method for compensating for this variable sample length by
concurrently measuring the water Raman band emission is discussed in
Section 8.
10
-------
Portable range
positioning transponders
Shoreline
Isdpleths of surface
water target
fluorescence
Laser excitation
spot
Sampling Points
Figure 2. Schematic illustrating possible mode of operation of an airborne laser fluorosensor for
mapping surface water chlorophyll _§_ distributions.
-------
SECTION 5
CHLOROPHYLL a MONITORING WITH AIRBORNE LASER FLUOROSENSORS
The laser fluorosensor equation for remote monitoring of in vivo
chlorophyll _a at normal incidence has been derived by Browell (20) and is
presented below as Equation 1. The chlorophyll a. concentration, nc, which
is assumed to remain constant with depth, is given as a function of either
known or measurable parameters:
nc =
PF H2]
PL
4-nAp
T ^Tr "Rec
yjexp {H (eL + 0F)f
d-Rw)2
kL + kF
"c
ug/i
(i)
where Pp = Peak detected fluorescence power at fluorescence wavelength F,
watts.
PL
H
T
Ap
''Tp
''Rec
Rw
Peak detected laser output power at laser wavelength L, watts.
Aircraft elevation or range above fluorescent target, m.
Effective area of telescope receiver, m^ .
Fraction of fluorescence band seen by detector.
Laser (transmitter) efficiency.
Telescope (receiver) efficiency.
Specular reflectance for air-water interface at normal incidence
for visible spectrum.
= Refractive index for water over visible spectrum.
= atmospheric beam attenuation coefficient at laser wavelength
PP = atmospheric beam attenuation coefficient at fluorescence
wavel ength F , m~l -
nc = concentration of total in vivo chlorophyll _a from algae in
irradiated sample, pg/1 .
k[_ = diffuse attenuation coefficient for water at laser wavelength L,
12
-------
kp = diffuse attenuation coefficient for water at fluorescence
wavelength F, m"l.
ffc = in vivo cross section for chlorophyll a_ fluorescence over
complete fluorescence band centered at 685 nm when excited at
laser wavelength L, m2/mg.
Factors within the first set of square brackets in Equation 1 are measured
directly from each digital or analog laser fluorosensor record obtained for a
given sampling station. Those within the second set of square brackets are
known or measurable laser fluorosensor system constants, whereas those within
the third set of brackets relate to known environmental factors. However, it
is the difficulty in anticipating the behavior of the environmental factors
within the fourth set of square brackets, specifically
-------
environment by Kiefer and Austin (28) have shown that the transmissometer and
the fluorometer signals remain constant down to 10 meters with a c* 1 m~l.
Similar measurements made in a fresh-water environment by Baker and Baker (29)
have shown that both nc and a remained constant down to about 3 meters for a
> 4m"1. Although no infallible rule exists which guarantees that nc,
aC) k|_ and kp will remain constant to depths characterized by 2(k~1)
or 2(a~1), most coastal, estuarine and inland waters have a values greater
than 1m'1, such that wind-induced mixing should ensure that this surface
layer will remain homogeneous with regard to both particulate and algal
matter.
(5) The assumption is made that the fluorescence lifetime for in vivo
chlorophyll a^ is less than or equal to the laser pulse width. In the present
case, with a laser pulse width of about 200 nsec, and a chlorophyll a^
fluorescence lifetime of less than 1 nanosecond (nsec) (30), this condition is
clearly met. However, in situations where the laser pulse width is less than
or equal to the fluorescence lifetime, the value of Pp would be reduced in
relation to PL due to the effect of pulse spreading induced by the
fluorescence decay phenomenon. This problem can be avoided by either applying
a suitable correction to the peak power measurements or measuring pulse energy
rather than peak power for both the laser and fluorescence signals.
(6) For clear atmospheric conditions over limited ranges of the order of
300 m, the atmospheric attenuation term exp |H(g|_ + 3p)| remains
essentially constant with a value close to unity, so that the effect of the
terms within the third set of square brackets can be considered to be
constant.
(7) Several assumptions are made with regard to the fluorescence emission
cross section
-------
influence
-------
SECTION 6
INSTRUMENTATION
The laser fluorosensor consists of a laser transmitter and telescope
receiver, which are mounted in a lightweight aluminum box in a non-coaxial
configuration, as shown schematically in Figure 3. Airborne operations are
conducted from a Bell UH/1D H (Huey) helicopter, with the system mounted in a
central location over a clear hole cut between two stringers just forward of
the transmission housing.
The system operates in a fixed downward-looking mode which at typical
aircraft ground speeds of 20 m/sec and a laser repetition rate of 1 pulse per
second (pps), gives a ground footprint spacing of 20 m. At an above target
elevation of 200 m, the laser beam divergence of 6 mrad produces a water
surface spot size of 1.2 m diameter. As the ground resolution of the system
is essentially set by the (20-m) spacing between laser excitation spots, it
would be advantageous to utilize a laser spot diameter similar to this
interspot spacing to smooth out the effects of small-scale fluctuations in the
concentration of surface water algae. This advantage can be gained by
increasing the aircraft height H without increasing the laser beam divergence
or the telescope field of view. A limitation to this expedient is set by the
l/Fr drop-off in the fluorescence signal with increasing H. As a result,
the sky and solar background signal and noise within the telescope field of
view eventually compete with and ultimately dominate the fluorescence signal.
In this case the signal-to-background noise-ratio also varies as 1/H^.
Consequently, doubling the aircraft height produces a fourfold reduction in
the signal-to-background noise ratio (31). Another approach to increasing the
laser excitation spot size would be to double the laser beam divergence 9
together with the telescope field of view at a given aircraft altitude. In
this case the signal-to-background noise varies as 1/9, so that doubling the
system field of view halves the signal-to-background noise-ratio (31).
However, this approach is much more difficult to implement than that achieved
by changing the aircraft height due to the problems involved in continuously
and simultaneously adjusting the fields of view for both laser and telescope.
A better approach would be to increase the laser repetition rate, thereby
increasing the sampling frequency so that the effects of algal patchiness can,
if desired, be smoothed out by low pass filtering the measured chlorophyll _a
profile. ~~
Adjustment of the position of the surface water laser spot is achieved
with the laser beam steering mirror shown in Figure 3, which positions the
laser spot at the center of the telescope field of view. This micrometer
adjustment is made during each flight at a prescribed altitude by maximizing
the amplitude of the detected surface water chlorophyll _a fluorescence signal
at a given location.
16
-------
FLUORESCENCE SIGNAL PULSE
DETECTOR HIGH VOLTAGE
GATING PULSE
DIFFUSER
LASER
POWER
MONITOR
FLASHLAMP PUMPED
PULSED DYE LASER
LASER BEAM
STEERING
MIRROR
SHOCK
\
GATED PHOTOMULTIPLIER DETECTOR
FLUORESCENCE BAND FILTER
LASER BLOCKING FILTER
FOCAL PLANE APERTURE
-REFRACTING TELESCOPE
.30 CM DIAMETER f/1.8 FRESNEL
COLLECTOR LENS
X Ai.R CRAFT FLO Q RQQOv X>
Figure 3. Optical diagram of airborne laser fluorosensor for monitoring surface water fluorescence
signal .
-------
LASER TRANSMITTER SYSTEM
Characteristics and typical performance data for the coaxial flash!amp-
pumped dye laser (Phase-R DL-1200) are given in Table 1. Performance data for
this laser using a number of dyes in methanolic solution are given in Table 2.
With the preferred dye Coumarin 120 used for exciting chlorophyll a.
fluorescence at 440 nm, the laser emits pulses of 200 kilowatt peak power and
200 nsec FWHM (full width at half maximum) width at a repetition rate of 1 pps
with a spectral bandwidth of about 0.45 nm. At an energy input of 25 joules
per pulse, the laser output declines to the half-power point after 25,000
joule shots/liter. Spectral tuning is achieved using a 60° prism made from
Schott SF 10 glass with the FWHM spectral bandwidth of the laser emission
varying from about 0.45 nm at 440 nm to 2.1 nm at 639 nm. Laser peak power is
monitored by directing a small fraction of the laser output onto a calibrated
PIN silicon photodiode via a quartz diffuser plate. The laser signal is also
used to provide a signal for triggering the oscilloscope waveform digitizer
and photomultiplier gating electronics. Although the laser can operate at up
to 10 pps, performance tends to deteriorate rapidly above 1 pps due to the
effects of dye degradation and the inability of the flowing dye and cooling
water to remove the flash-induced heat buildup prior to the next laser pulse.
This latter effect produces refractive index inhomogeneities in the flowing
dye which in effect produces a misalignment of the laser cavity. Our
measurements indicated that a dye flow rate of at least 51/min was required
for stable laser performance at a repetition rate of 1 pps. The laser dye and
cooling water flow circuits are illustrated schematically in Figure 4. Both
dye and cooling water temperatures were maintained at 18°C ± 1°C.
Construction materials which do not appear to effect laser performance when
exposed to the laser dye solution are stainless steel, glass, quartz, Teflon,
Delrin, polypropylene, polyethylene and silicone rubber. Considerable effort
was directed towards achieving a satisfactory dye flow rate through the laser.
Magnetically coupled pumps with wetted parts of stainless steel and Teflon are
generally satisfactory because of the inert nature of these materials.
Unfortunately all high-performance pumps produce both cavitation bubbles and a
steady stream of Teflon and stainless steel wear particles, which act as
optical scattering centers within the laser cavity, thereby reducing laser
output power. These contaminants in turn necessitate the use of an inline
filter to trap both bubbles and pump wear products. After much
experimentation, a variable speed magnetically coupled stainless steel and
Teflon centrifugal pump (Micropump Model #10-41-316) was chosen together with
a 142-mm diameter 1-u pore-size cellulose acetate membrane filter (Millipore
Cellotate EA) which together produced a flow rate of 5 liters per minute
(1/min) at 83 kilo pascal(kPa) (12 psi). The possibility of a flashlamp
explosion with subsequent risk of a methanol fire, particularly on board an
aircraft, required careful consideration. A pressure-operated laser cut-off
switch was installed downstream of the flashlamp, as shown in Figure 4. The
switch was arranged to cut off electrical power to both the dye pump and the
laser power supply on loss of dye pressure due to rupture of the flashlamp.
In addition, an inert nitrogen atmosphere is maintained in the laser cavity
during laser operation. Halon fire extinguishers are also made available for
use against a possible methanol fire. With a view to avoiding a flashlamp
explosion due to increasing lamp resistance with age, a 1,000-ohm high voltage
resistor was placed in parallel with the flashlamp. As the value of the lamp
18
-------
TABLE 1. AIRBORNE LASER-FLUOROSENSOR CHARACTERISTICS
LASER TRANSMITTER
Peak power 100KW-300KW
Pulse width (FWHM) 200 nsec - 280 nsec
Pulse energy 40 mj - 70 mj
Beam divergence 6 mrad
Spectral bandwidth 0.45 nm 2.1 nm
Repetition rate 0.1 pps 10pps
Degree of polarization Linear to within 1 part per 100
TELESCOPE RECEIVER
Refractor 30 cm diam., f/1.8, Acrylic Fresnel
Focal plane aperture 4 mm diam.
Chlorophyll a fluorescence filter 685 nm, 13 nm FWHM,
64% Transmission
Laser blocking filter Corning 2-64
19
-------
TABLE 2. LASER DYES EVALUATED FOR USE IN COAXIAL FLASHLAMP-PUMPED DYE LASER EMPLOYED
IN AIRBORNE LASER FLUOROSENSOR
A
Laser dye
Coumarin
120
(HOC)
Coumarin
339
(EOC)
Coumarin
311
(EOC)
Coumarin
314
(EOC)
Coumarin
522F
(EXC)
Rhodamine
110
(EOC)
RhodamineB
Perchlorate
(EOC)
Molecular
Weight
175.19
215.25
203.24
313.35
283.25
366.80
543.02
B
Molecular
Concentration
(M)
5 x 10~4
2 * 10'4
2 x 10"4
1 x 10~4
2 x 10~4
5 x 10~5
5 x 10~5
Peak
Wavelength
(nm)
452
482
485
520
534
570
639
Tuning
Range
(nm)
432-468
440-492
442-492
485-541
502-572
545-590
595-655
C
Peak Power
(KW)
200
210
255
300
240
220
270
Spectral
Bandwidth
FWHM
(nm)
0.45
0.93
0.80
1.25
0.97
2.00
2.10
Pulsewidth
FWHM
(nsec)
200
240
225
250
250
280
275
Dye
Lifetime
Joule-Shots/I
25,000
20,000
50,000
50,000 +
50,000 +
100,000 +
250,000
IV)
o
A) EOC: Eastman Organic Chemicals, Rochester, NY; EXC: Exciton Chemical Company, Dayton, OH.
B) Solvent: Spectral grade methanol.
C) Phase-R DL 1200 Laser: Input energy-25 Joules; Dye temperature-65°F; Dye flow rate- 5 l/min;
Repetition rate-1 pulse/sec.
-------
1.3 litre
insulated
and vented
laser dye
reservoir
Pressure
X) 9auge
Drain
line
142 mm. dia.
membrane filter
holder
V
Variable speed
magnetically
coupled
centrifugal
pump
Emergency
pressure
operated
laser cut off
switch
Air bubble
bypass line
Cooling water pump
Laser coaxial
flashlamp
30 liter thermostatically
controlled constant
temperature water bath
Figure 4. Dye and water cooling flow diagram for coaxial flashlamp-pumped dye laser employed in
airborne laser-fluorosensor.
-------
resistance increases to that of the resistor, a point is reached at which
increased current flow through the resistor will cause the resistor to fail in
preference to the lamp (32). This expedient not only avoids a flashlamp
explosion and fire risk but also acts as a convenient indicator of flashlamp
lifetime.
OPTICAL RECEIVER SYSTEM
The optical part of the receiver system, consisting of a Fresnel lens, a
focal plane aperture, a series of optical filters and a gated photomultiplier,
is shown schematically in Figure 3 with characteristics listed in Table 1.
The principal advantages of using a lightweight acrylic lens (30-centimeter
diameter, f/1.8) are the low cost and minimum requirement for a support
structure because of its negligible weight. Its principal disadvantage, other
than those due to scattering losses from the grooves, originates from its low
f/number refractive nature. Lenses of this type produce large blur circles or
circles of least confusion due to the effects of spherical aberration. For a
plano-convex acrylic lens of refractive index y = 1.49 and focal length of 30
cm the calculated diameter for the circle of least confusion for a point
source at infinity is 6 mm (33), whereas the calculated size of the image of
a laser spot produced by a 6-mrad laser beam at 200 m is 2-mm diameter.
Clearly, if the full fluorescence return signal is to be detected, an aperture
of at least 6 mm diameter should be employed rather than the theoretical 2 mm
value. However, this increase in aperture produces a 9-fold increase in the
solar background signal level that reaches the detector. This change in turn
reduces the signal to noise ratio by a factor of three (31). In this
situation a compromise focal plane aperture of 4-mm diameter was found to be
satisfactory. The chlorophyll a_ fluorescence band at 685 nm is spectrally
isolated with a 5-cm square interference filter centered at 685 nm with a
13-nm width (FWHM) and 64% peak transmission. Additional laser blocking is
provided by a longwave pass filter with a cutoff at 640 nm (Corning 2-64).
Neutral density attenuation filters are employed to ensure that the detected
signal does not exceed the photomultiplier linearity limit. The fluorescence
detector is a 5-cm diameter, 12-stage, red-sensitive (S-20 response) end-on
photomultiplier tube (RCA C31000A). The linearity limit of the detector
output is enhanced by using a capacitatively decoupled linear dynode chain as
shown in Figure 5. Saturation of the detector, caused by continuous exposure
to the backscattered sky and solar background radiation, is avoided by gating
on the supply to dynodes 4, 6 and 8 for the duration of each fluorescence
return signal. A 42-volt pulse applied as shown in Figure 5 provides an
on/off ratio of about 1,000:1. Under these conditions with a gate width of 3
psec, the tube remains linear to 1 volt as measured across a 50-ohm load. In
addition, gating the detector also provides a measurement of the steady (DC)
background current at the instant the fluorescence pulse is measured. A
determination can then be made to ensure that the detector has remained in the
linear operating region during the measurement of the fluorescence pulse.
ELECTRONIC MONITORING AND RECORDING SYSTEM
The circuit for the detector gating, monitoring and recording electronic
system is shown in Figure 6. Events are initiated by the laser power pulse,
which triggers a monitoring oscilloscope (Tektronix R7844). The oscilloscope
22
-------
T
— 2000V
-42V
bate r uise — r\J
in — 1—
0.01
Onni
— 50K<
J
o.ooi 70K<
0.001 100K3
ill 1
II ^
~"^ 0 001
100K<
0.001 5
1 1 1
II "i
•=• 0.001
100K<
0.001 <
1 1
1 1 vj
4
600V zener
• (i
| •
>
:
i
>
L
. BOOK
:^w^p
^
T
> . 500K
:4wC-p
s
• . 500K
:
-------
Delayed gate
generator
_V
e
30 cm dia Fresnel
refracting telescope
Gated
photomultiplier
(RCA C31000 A)
Low pass
filters
Pulse
generator
Dual beam
oscilloscope
(Tektronix R7844)
Pulse
splitters
\
(o ivmzj v x \
\ \r T
. \ \l
is . • to. fTii b ra b •
P.. p> |i«| ^ ;.i| p • ^
f
/
Laser
Flashlamp pumped power t t ^__ Video ^
dye laser monitor 1 | monitor
(Phase-R DL1200)
TV Camera
(silicon t
photodiode) jL
•
i
r
-I
^ ^ t
Video tape Video/digital Interface
recorder time code
generator
Waveform
digitizer
*"(Biomation 8100)
+
t
9 track digital
tape recorder
(Kennedy 9800)
Figure 6
Schematic
detection,
of airborne laser fluorosensor for measuring chlorophyll j[ fluorescence showing
monitoring and recording systems.
-------
time base gate output is then used to trigger a pulse generator which in turn
initiates formation of a delayed detector gate pulse. The delay period for
the detector gate pulse is adjusted so that the detector is activated several
hundred nanoseconds before the arrival of the backscattered fluorescence
signal. This series of events is conveniently illustrated by the oscillogram
shown in Figure 7. In this case, for a height-of 309 m (1,103 ft), the
detector was gated on 1.1 psec after lasing. Also shown are the laser power
pulse and the chlorophyll a^ fluorescence return from the algae, which is
superimposed on the steady 3-psec background signal. This background signal,
consisting of both sky and solar backscatter and sky surface specular
reflection components, does not constitute a serious noise source for the
present laser fluorosensor system. However, the same cannot be said of the
effects of solar glitter, viz., the direct specular surface reflection of the
solar disc seen by the detector. In this case, the solar glitter background
signal not only contributes significantly to the overall signal noise level
but often drives the detector into a region of non-linearity or saturation.
In the light of these observations it was found advisable to avoid making
flights over rough surface water at times of high sun angle when solar glitter
can be seen by a downward-looking sensor. The most favorable time was
generally in the early morning hours, when the sun angle is low and the water
surface is calm. The overall signal waveform such as that shown in Figure 7
is also digitized into 512 channels, each 10 nsec wide with 8-bit resolution
over a 5.12-usec period using a fast waveform digitizer (Biomation 8100).
This digital data is then dumped from an internal buffer memory through an
interface bus onto magnetic tape using a 9-track digital tape recorder
(Kennedy 9800). The effective bandwidth of both digital and analog
electronics is approximately 8 MHz dictated principally by the 50-ohm low-pass
filters used to reduce the amplitude of photon and electromagnetic
interference (EMI) noise. These filters are conveniently made from lossy
50-ohm cable whose low pass cutoff frequency is dependent on the cable
length*. These filters are relatively cheap and easy to install with suitable
(BNC) connectors and transmit signals free of ringing or line reflections.
Navigation, fluorescence target evaluation and track recovery were
facilitated using a TV camera with a wide angle lens mounted directly to the
laser fluorosensor module and boresighted with the laser beam. The helicopter
flight path over a chosen water surface target can then be viewed directly on
a video monitor and the same video image recorded on a video cassette. In
order to use this video recording to assist with interpretation of the laser
fluorosensor data, a video digital time-code generator was used to index each
digital laser fluorosensor.record and each video frame with a time signal to
the nearest 1/100 second. An approximate ground location for each laser
fluorosensor record can then be established.
Despite considerable effort expended on the digital recording equipment,
satisfactory operation of this part of the system was not achieved for several
reasons. The. high ambient daytime temperatures encountered in the southern
Nevada area often produce temperatures inside the helicopter in excess of
*Manufactured by CAPCON, Inc. NY, NY 10001.
25
-------
LASER PULSE
AT 440nm
DETECTOR ON
MEASURE OF
LASER POWER
ALTITUDE =
CHLOROPHYLL a
FLUORESCENCE
AT 685nm
MEASURE OF SURFACE
WATER CHLOROPHYLL a.
CONCENTRATION (EXTRACTED
CHLOROPHYLL a = 10.5 jig/1)
T
MEASURE OF
SOLAR
BACKSCATTERED
RADIATION
Figure 7.
= 309m
Oscillogram showing sequence of airborne laser-fluorosensor signals obtained over buoy
#12 on October 4, 1976, for a measured chlorophyll j[ concentration of 10.5 ug/1.
-------
120° F. This heat buildup invariably induces component failure, particularly
in the digital electronic equipment. This problem has been temporarily solved
by scheduling the test flights during the early hours of the morning when
ambient temperatures are some 30° F to 40° F cooler than at midday. However,
this expedient will not be acceptable when, of necessity, flights must be made
at specific times of the day with a view to monitoring diurnal changes in the
distribution of surface water algae. With these requirements in mind, it is
planned to provide a dedicated cooling system for the digital rack-mounted
instrumentation. Problems that were somewhat more difficult to eradicate
concern the influence of vibrations and electronic noise on the performance of
the digital recording system. Low-frequency vibrations in the region of 3 Hz
to 7 Hz induced by the helicopter rotor blade were found to affect the
precision alignment of the read and write heads of the digital tape recorder
in relation to the tape transport assembly. This problem was eliminated by
the addition of a stiffening member to the tape-drive support frame.
Vibrations may also reduce the integrity of the many mechanical circuit
connections in the digital equipment as exist with printed circuit board
connector pins, wire-wrapped circuits and push-fitted integrated circuit
chips. Although successful isolation from low-frequency vibrations in the
region below 7 Hz is difficult to achieve using available aircraft-rated shock
mount components, efforts are being made to improve the performance of the
shock isolation rack mounting system used to hold the digital recording
system. Finally, both the digital recording system and the monitoring
oscilloscope have demonstrated a sensitivity to noise in the form of radio
frequency interference (RFI) and EMI, particularly in the enclosed aircraft
environment. Efforts are being made to improve the shielding of these
components from external noise sources through the use of improved RFI
screening, elimination of potential ground loops and additional filtering of
the aircraft power supply- Due to the aforementioned problems with the
digital equipment, it was not always possible to produce digital recordings of
the airborne laser fluorosensor data. Consequently, oscillograms obtained at
each sampling buoy location were used as the prime data source for comparison
with the corresponding chlorophyll _a ground truth data.
27
-------
SECTION 7
AIRBORNE MEASUREMENTS
FIELD OPERATIONS
Airborne testing and evaluation of the laser fluorosensor were made over
the Las Vegas Bay region of Lake Mead; a nautical chart of this area is shown
in Figure 8. This bay is an ideal test site for a number of reasons.
Firstly, it is located approximately 10 minutes flying time from McCarran (Las
Vegas International) Airport. Secondly, this region has been extensively
surveyed over a number of years by the Department of Biological Sciences,
University of Nevada, Las Vegas (UNLV), because of the concern regarding the
influence of nutrients on the population level of algae (34). High
concentrations of nutrients and other pollutants enter the western end of Las
Vegas Bay from Las Vegas Wash, which is essentially an open stream carrying
both partially treated and untreated sewage water from the City of Las Vegas.
At certain times of the year this pollution has been observed to support algal
communities to the extent that chlorophyll a_ values will vary from 0.5 yg/1 to
50 yg/1 over a distance of 10 km. Finally, ground truth support for these
airborne measurments was conveniently provided through the facilities and
personnel of the Department of Biological Sciences at UNLV. A series of
sampling station marker buoys positioned by the Department and the National
Park Service are shown in Figure 8. This string of buoys, easily seen from a
height of 200 m, was flown as three separate straight flight lines. The
lengths of these lines from buoys 1 to 7, 7 to 11', and 11' to 12 are
respectively 4452 m, 3544 m, and 1607 m. Not surprisingly, the highest
chlorophyll a_ concentrations generally occurred in the region of buoy 12,
which lies closest to the point where Las Vegas Wash enters Las Vegas Bay,
whereas the lowest concentrations generally occurred at buoy 1. Lower
chlorophyll a_ levels are encountered further out in the center of Lake Mead in
the direction of Sentinel Island.
Airborne laser fluorosensor measurements were made over the three flight
lines defined by buoys 1 to 7, 7 to 11' and 11' to 12 concurrent with
collection of samples close to the buoys. Surface water grab samples were
obtained at three different locations on the circumference of an approximately
25-m diameter circle around each sampling station in order to minimize the
chances of collecting significant amounts of periphyton, which might break
away from the sampling station buoy surface. These three samples were then
mixed in order to minimize the effects of patchiness in the distribution of
surface water algae.
When exciting chlorophyll a_ in the Soret absorption band region, the
wavelength of the laser transmitter was tuned to lie close to 440 nm.
Excitation at 440 nm avoids exciting phaeophytin _a, a degradation product of
28
-------
Figure 8. Nautical chart of Las Vegas Bay region of Lake Mead, Nevada, showing location of marker
buoy sampling stations.
-------
chlorophyll a^ that has an excitation peak close to 420 nm but whose
fluorescence cross section is approximately 10 times less than that of
chlorophyll a_ at 440 nm when using excitation bandwidths smaller than 5 nm
(35, 36).
ANALYSIS OF AIRBORNE LASER-FLUOROSENSOR DATA
As indicated earlier, the airborne data could be collected either
continuously in digital format on magnetic tape or in the form of oscilloscope
photographic records taken at each buoy site as exemplified by the oscillogram
in Figure 7. Flights over desert terrain essentially devoid of vegetation,
located immediately adjacent to Lake Mead, produced no signal at 685 nm other
than from the solar background, indicating that backscattered laser radiation
was not leaking through either the chlorophyll j. fluorescence band
interference filter or the short wavelength cutoff color glass filter used to
block the laser backscatter. However, a small fraction of the pulsed signal .•
at 685 nm comes from the fluorescence of dissolved organic matter in the water
rather than from chlorophyll a.. This fact is illustrated by the fluorescence
spectra shown in Figure 9, obtained in the laboratory from a Lake Mead water
sample with a corrected-spectra spectrofluorometer-(Perkin Elmer MPF4). The
first spectrum in Figure 9 obtained from an untreated subsample shows the
characteristic in vivo chlorophyll a_ fluorescence emission band at 685 nm for
a measured extractable chlorophyll a. level of 8.4 yg/1. The small spikes
visible on this emission band are due to the movement of discrete algal
particles across the instrument field of view. The second spectrum was
obtained on a similar sample after passage through a 0.3 y membrane filter
used to remove the algae. The residual fluorescence background is due to the
presence of dissolved organic matter in the sample and illustrates the fact
that, at least for Lake Mead waters, this is an insignificant source of error
to the present measurement at 685 nm. However for conditions of low
chlorophyll a_ concentration in the presence of a high dissolved organic
content, account will have to be taken of this background signal at 685 nm.
As there are no immediate plans to calibrate the airborne laser-
fluorosensor data directly in units of chlorophyll a_ concentration, the
effects of varying laser intensity and aircraft altitude on the chlorophyll a^
fluorescence signal were eliminated by normalizing the fluorescence signal to
arbitrary reference values of laser power and aircraft altitude. The
elevation H of the helicopter over the water surface target was measured from
each record to an accuracy on the order of ± 3 meters. This value was then
used to normalize the amplitude of the fluorescence signal to a reference
altitude of 200 m using the 1/H^ dependence of the fluorescence signal with
range. Over the period of a specific airborne mission involving about 1,000
laser-fluorosensor sampling sites, the laser output power generally falls to a
value equal to 30% of its original value due to degradation of the Coumarin
120 laser dye. The laser peak power signal was therefore used to normalize
the fluorescence signal to an arbitrary 1-volt laser power signal, assuming a
direct linear dependence of fluoresence emission power on laser power. In
addition, correction factors for the effects of electronic and optical
attenuators were made when necessary. Except for certain pulse-integration
and profile-smoothing procedures applied to the analysis of the digital data,
interpretation of both analog (oscillographic) and digital data was
30
-------
Excitation - 440nm
Bandwidth - 10nm
Emission
bandwidth - 5nm
Chlorophyll ,a_
fluorescence
(concentration
8.4 M9/D
Unfiltered
sample
Fluorescence emission
from dissolved organics
Filtered sample
(Millipore 0.3u
pore size)
550
Figure 9
580
610
640 670 700
WAVELENGTH (nm)
Corrected fluorescence emission spectra of filtered and
unfiltered Lake Mead surface water sample, excited at 440 nm,
August 16, 1977.
31
-------
essentially identical. This involves determination of the amplitude of the
measured laser power and fluorescence signal above the solar background
together with a value for the time between the laser and the fluorescence
pulses. In the case of the digital data, a 5-point least-squares parabolic
curve-smoothing procedure (37) was applied sequentially to the data for each
of the 512 channels using the data from the two channels on either side to
provide the smoothing curve. This process is then applied to the next data
point and so on. This procedure effectively eliminates the short-term random
contributions from digitization, photon and electronic noise while leaving the
long-time constant components of the record unchanged. The elapsed time
between laser and fluorescence pulses, which is used to calculate the aircraft
altitude after correction for photomultiplier transit time, was determined by
calculating the position of the centroids for the laser and fluorescence
pulses.
ANALYSIS OF GROUND TRUTH WATER SAMPLES
Determinations of chlorophyll _a concentration for the surface water
samples were obtained with either the spectroscopic approach of Strickland and
Parsons (38) or the fluorometric approach of Holm-Hansen, Lorenzen, Holmes and
Strickland (39). The procedure employed for extracting the chlorophyll _a from
the algae was the same for both methods. This involves collecting the algae
on a glass fibre filter paper (Whatman GF/C) pretreated with magnesium
carbonate and grinding the filter paper in a prescribed volume of 90% acetone,
which is then allowed to sit for a 3-hour period during which the chlorophyll
a_ is extracted into solution from the algae. The centrifuged supernatant
solution containing the chlorophyll _a is then used undiluted for the
spectrophotometric determination. For the fluorometric determination the
supernatant solution is further diluted in 90% acetone over the range from
25:1 to 100:1 to ensure that self absorption of the fluorescence emission does
not occur. Spectrophotometric determinations were made using a Coleman Model
620 Junior II Spectrophotometer. Fluorometric determinations were made using
a Turner Model 111 filter fluorometer with a standard door, a blue continuous
emission lamp and a Hamamatsu R-136 red-sensitive photomultiplier detector.
For high chlorophyll a^ concentrations, a Corning 5-58 excitation filter was
employed, whereas for low concentrations a Corning 5-60 excitation filter was
used. A Corning 2-64 short-wavelength cutoff filter was used to isolate the
chlorophyll _a emission band at 685 nm. For the analysis of the samples for
flight #12, a UV blocking filter (Corning 3-73) was employed to reduce the
influence of the 404.7-nm mercury line in exciting fluorescence in the
pheophytin a_ fraction (35, 36), while at the same time passing the continuous
blue and discrete 435.8-nm mercury line radiation used for exciting
fluorescence in the chlorophyll _a fraction. Both spectroscopic and
fluorometric determinations were repeated on five subsamples obtained from
each grab sample and the means for these groups of subsamples determined for
comparison with the corresponding airborne data. Calibration of the
fluorometer for each filter door combination was made periodically against a
chromatographically pure chlorophyll a standard*.
*Pure chlorophyll _a extract was obtained from Sigma Chemical Co.
St. Louis, Mo.
-------
An indication of the relative concentration of sample particulates (both
algae and suspended sediment) was obtained using a 90° scattering nephelometer
(Hach Model 2100A).
COMPARISON BETWEEN AIRBORNE AND GROUND TRUTH DATA
Airborne laser fluorosensor data for flights #4 and #8, obtained using a
laser excitation wavelength of 440 nm and taken about 1 month apart, are shown
in Figures 10 and 11 respectively, together with the corresponding chlorophyll
_a ground truth data obtained by the spectroscopic method of Strickland &
Parsons (38).
Elevation above the water surface was about 200 m for both flights. Since
absolute calibration of the airborne fluorescence data in terms of chlorophyll
_a concentration was not provided, it was found convenient to normalize each
airborne data set to the corresponding chlorophyll _a ground truth data set by
minimizing the sum of the squares of the differences between the corresponding
airborne and ground truth values (least squares procedure).
Changes in the 10-kilometer (km) long surface water profiles, for both the
airborne and ground truth measurements for flight #4, shown in Figure 10,
appear to be in general agreement. Comparison between the airborne and ground
truth data for this flight gave a linear correlation coefficient of 0.95. The
data for flight #8 made over the same path on a later date exhibit a somewhat
lower correlation coefficient of 0.77. It is apparent that some of the large
point-to-point fluctuations present in the ground truth profile for this
flight are also present in the airborne fluorescence profile, though in a
somewhat attenuated form. Similar but smaller fluctuations were also present
in the fl uorometrical ly determined chlorophyll jj data for flight #8 which,
when compared to the laser fluorosensor data, gave a correlation coefficient
of 0.85. Linear correlations between airborne and ground truth data for these
and other flights are presented in Table 3 with the values lying in the range
from 0.41 to 0.95. For a sample size of 12 or larger, with r >_ 0.57,
correlations are significant at the 5% level whereas with r >_ 0.70, they are
significant at the 1% level. It is interesting to note that the correlation
(r=0.41) for the data of flight #12 obtained using a laser fluorosensor
excitation wavelength of 622 nm is not significant at the 10% level. One of
the reasons for using this excitation wavelength was to investigate the
correlation between the airborne data and the biomass for specific algal color
groups, and in particular for blue-green algae. Further discussion on the
data for this flight is deferred to Section 7f. Also shown in Table 3 are the
correlation coefficients relating the laser fluorosensor data to that for
water turbidity obtained using a 90° scattering nephelometer. It is
significant that, in all cases, these coefficients are only slightly less than
those for the correlation between the laser fluorosensor data and that for the
spectroscopically determined chlorophyll a_. This suggests, at least from an
optical viewpoint, that the particulate matter in surface waters of Las Vegas
Bay consists principally of algae rather than of suspended sediment.
An estimate of the sensitivity limit for the laser fluorosensor in its
present form can be gauged from fluorescence pulse data provided by the
oscillogram in Figure 7. Assuming that for daytime operation, system
33
-------
18
I 14
o
Ui
_o
.c
o
(Q
g 10
><
LU
8
2 -
O
For linear correlation between laserfluorosensor
and chlorophyll a^ data, r= 0.95 for 12 samples
Airborne laserfluorosensor fluoresence emission
O at 685 nm, normalized to chlorophyll a data
by least squares procedure.
...A... Extractable chlorophyll a
from surface water samples.
_ Sampling station number
1
2
I
3
j
4 5
I I
678
I I I
9 10 11 1V
ll J !
12
i
I
2000
4000
6000
I
8000 10,000
Distance from station 1 (m)
Figure 10. Variation of surface water chlorophyll a^ and laser-fluorosensor signal with distance
for surface water transect of Las Vegas Bay in Lake Mead, Nevada, Flight #4, October
15, 1976.
-------
=r 11 r
O)
3_
10 -
I 9
_o
.C
o
a> 8
.Q
CD
O
CO "7
4-1
X
LLJ
o
For linear correlation between laserfluorosensor
and chlorophyll a data, r = 0.77 for 13 samples.
Airborne laserfluorosensor fluoresence emission
_o at 685 nm, normalized to chlorophyll a data
by least squares procedure.
Extractable chlorophyll a
'.A... r ' —
from surface water samples.
°
,
3C
2
1
0
\s
Y ***•.
™
-
1 2
/ !
...A
3
!
-^
4
i
*
•
5
i
•
'
i
•
«
•
•
'••A/
6 7
I I
: /
• i
* •
• «
'•A
X..A/
8 9 10
I I . I
^x.u
Sampling station
11 1V
I J
number
12
I .
2000
4000
6000 8000 10,000
Distance from station 1 (m)
Figure 11. Variation of surface water chlorophyll a_ and laser- fl uorosensor signal with distance
for surface water transect of Las Vegas Bay in Lake Mead, Nevada. Flight #8, November
19
1976
-------
TABLE 3. CORRELATION COEFFICIENTS FOR CORRECTED LASER-FLUOROSENSOR SIGNAL VERSUS TURBIDITY
AND CHLOROPHYLL a DATA
Flight
Number
3
4
5
8
12
Date
10/04/76
10/15/76
11/04/76
11/19/76
08/16/77
Turbidity
(NTU)
0.801
0.846
0.841
0.708
...
A
Ca
(Spectro.)
(pg/D
...
0.953
0.870
0.770
---
B
Ca
(Fluoro.)
(pg/D
0.815
0.751
...
0.846
0.408
C
Sample
Size
12
12
13
13
13
Laserfluorosensor
Excitation Wavelength
(nm)
440
440
440
440
622
A) Spectrophotometric determination of chlorophyll a using method of Strickland and Parsons (38)
B) Fluorometric determination of chlorophyll a using method of Holm-Hansen et al. (39).
C) Each chlorophyll a laserfluorosensor sample is an average of three measurements.
-------
sensitivity is limited by the backscattered sky solar background radiation
existing within the chlorophyll ^fluorescence band, a peak Signal to RMS
(Root Mean Square) Background Noise Ratio (SBNR) can be calculated using
SBNR = Vp/vRMS
where Vp is the steady peak detector voltage of the fluorescence pulse
above the steady background level and VRM$ ">s the RMS noise voltage of the
background signal. Based on the fact that shot or photoelectron noise is
white noise with an instantaneous Gaussian probability distribution, a
statistical criterion can be adopted whereby the peak to peak fluctuations of
the background lie within a total spread of 5 standard deviations for 99% of
the time. Because one standard deviation is defined as the RMS noise voltage
VR^, we can write the equality
VPP = 5 VRMS
where vpp is the peak-to-peak spread in the noise envelope for 99% of the
time. The expression for the peak signal to RMS background noise ratio then
becomes
SBNR - 5 Vp/vpp
As depicted in Figure 7, Vp = 0.64 volts for a chlorophyll a_ concentration
of 10.5 yg/1 and vpp = 0.04 volts; it follows that for a minimum acceptable
SBNR of 3, the system can monitor chlorophyll a^ levels down to 0.4 yg/1.
This sensitivity limit is subject to a number of qualifiers. First, these
measurements were made under clear skies with the sun near the zenith. With a
water albedo on the order of 0.05, the noise due to the solar backscatter (but
not surface glitter), was at the high end of the anticipated range. However,
albedo values in the range from 0.2 to 0.3, due to high concentrations of
suspended sediment, would substantially increase this background noise.
Second, it is assumed that the contribution to the laser induced fluorescence
signal at 685 nm due to dissolved organics is negligible. For the surface
waters of Lake Mead, the amplitude of the fluorescence signal at 685 nm due to
dissolved organics has an equivalent chlorophyll a_ value of about 0.1 yg/1.
For water conditions where this is not the case, an estimate of this
background signal must be obtained by making an additional laser-fluorosensor
spectral measurement, say in the region of 650 nm. Third, the present
sensitivity limit corresponds to water conditions for which the optical
transmission is dominated by algal particles. For conditions where suspended
sediment is the principal factor limiting transmission, system sensitivity
will be reduced in relation to the values indicated above. This situation can
be expected to exist for suspended sediment concentrations greater than
50 mg/1.
A number of expedients can be employed to improve system sensitivity-
Increasing the laser energy or peak power will directly increase Vp, whereas
reducing the laser beam divergence, so that the receiver collects a smaller
solar background signal, will reduce Vpp. The background noise signal vpp
can also be reduced to near zero by flying at dawn or dusk or during
37
-------
nighttime, in contrast to the present experiments, which were flown close to
midday under clear skies but in the absence of significant solar glitter. In
this situation, system sensitivity is limited by "in-signal" photoelectron
noise. In any case, to achieve a system sensitivity limit down to a
chlorophyll a^ concentration of 0.1 pg/1, with a signal-to-noise ratio of 10
should represent no major technical problems. This level of sensitivity, in
effect, requires a 13-1/3-fold increase in the overall SBNR over the
performance of the present system. With a larger laser flashlamp and
increased laser input energy, a doubling in the laser output power and energy
can be achieved with a corresponding 2-fold increase in the SBNR (31).
Similarly a 1.77-fold-enhancement in the SBNR can be achieved by increasing
the FWHM-bandwidth of the optical filter at 685 nm from its present value of
13 nm to 23 nm, so as to be in close correspondence with the width of the
chlorophyll a_ band. Further, by utilizing a telescope with a reflecting
element rather than a refracting element, the circle of least confusion (the
blur circle) caused by spherical aberration can be reduced from the present
6-mm-diameter to a size much smaller than the 2-mm-diameter image of the
surface water laser spot. A 2-mm-diameter focal plane stop could then be used
rather than the present 4-mm-diameter stop, thereby reducing the magnitude of
the background signal and noise without affecting the magnitude of the
chlorophyll a^ fluorescence signal. This measure produces an additional 2-fold
enhancement in the SBNR (31). Finally, by halving the elevation of the
airborne platform over the water surface target, the chlorophyll ja
fluorescence signal would be enhanced by a factor of 4 based on the
dependence of the fluorescence signal on altitude; this expedient does not
affect the magnitude of the background signal for an extended homogeneous
target. This procedure in turn produces a 4-fold increase in the SBNR (31).
With implementation of all of the aforementioned system modifications, the
SBNR would be increased overall by a factor 27 times the present value,
thereby realizing a SBNR of 20 at a chlorophyll a_ level of 0.1 yg/1 under the
environmental conditions of the present measurements.
It should be noted that reducing the aircraft height also reduces the
laser spot size, thereby increasing the sensitivity of the laser-fluorosensor
signal to small-scale variations in the concentration of surface water algae.
As indicated in Section 4, this can be accommodated for by increasing the
laser repetition rate.
FACTORS INFLUENCING THE RELATIONSHIP BETWEEN AIRBORNE AND GROUND TRUTH DATA
Several explanations exist for the less than perfect covariation between
the laser fluorosensor signal and the corresponding surface water chlorophyll
^determinations. Factors influencing these correlations are either random or
systematic in origin and are discussed separately.
Random errors compounded by a small sample size of 12 or 13 will clearly
reduce the degree of correlation. Differences are known to exist between the
locations at which the grab samples were obtained and those at which the
airborne measurements were made for any one sampling station. Uncertainties
on the order of 50 m combined with significant patchiness in the surface water
algae distribution can be expected to degrade any such correlation. With a
view to minimizing this source of error, samples were collected at three
different locations on the circumference of an approximately 25-m diameter
38
-------
circle around each sampling station buoy and then mixed prior to analysis. At
the present time, it is not clear what residual error is present due to
differences between the airborne and ground truth sampling locations. Direct
grab sampling from the helicopter would eliminate this problem although the
helicopter rotor down-wash might induce considerable horizontal mixing and
water movement at the sampling site. Random errors are also incurred in the
measurement of the extractable chlorophyll a_. The established method for
extracting chlorophyll _a from algae in water samples (40) by filter grinding
and extraction in 90% acetone is known to be both inefficient and subject to
considerable variability due to the production of chlorophyll ^degradation
products (41, 42). For a total of five determinations made on each grab
sample, coefficients of variation (s/x) of up to ±50% have been observed
for chlorophyll a_ levels down to 1 yg/1 . More efficient and reproducible
methods for extracting chlorophyll a^ from algae are presently under
investigation and will be reported elsewhere.
Systematic errors are best discussed in terms of the laser fluorosensor
equation (Equation 1) for remote monitoring of in vivo chlorophyll c[. The
environmental factors within the fourth set of square brackets in Equation 1
are considered to be the primary source of systematic error encountered with
the laser-fluorosensor technique. Specifically, the discussion centers on the
validity of the assumptions that the fluorescence cross section
-------
indicate that the fluorescence excitation cross section at 618 nm for a
blue-green algal monoculture (anacystis marinus) was, on the average, about 5
times larger than the values for the other color groups. Unfortunately, as
mentioned in Section 7 (d), the present airborne measurements, made by using
an excitation wavelength of 622 nm, were not successful (the laser
fluorosensor and chlorophyll a_ ground truth data were poorly correlated, with
r = 0.41). However, as will be discussed in more detail in Section 7 (f), the
reason for this poor correlation is related primarily to faulty ground truth
chlorophyll a^ determinations rather than to an inherent weakness in the laser
fluorosensor method. A final decision on the choice of an optimum excitation
wavelength must therefore await further tests of the airborne laser
fluorosensor. At that time it will then be possible to pass judgement on the
validity of the concept of using a single unique overall value of ffc for
characterizing the fluorescence of in vivo chlorophyll ^ at a given excitation
wavelength.
Inspection of Equation 1 shows that changes in the surface water values
for k|_ and kp, the diffuse attenuation coefficient at the laser and
fluorescence wavelengths respectively, directly influence the laser-
fluorosensor determination of nc, regardless of whether a relative or
absolute value is required. The assumption was made that changes in
concentration of particulate and dissolved matter were sufficiently small over
the extent of a given water surface that k|_ and kp could be considered to
be constant. This premise in turn requires the assumption that optical
effects caused by algal scattering and absorption contribute a negligible
percentage to the total values for kp and k|_. However, situations exist
where the related beam attenuation coefficient a^ is highly correlated with
either extracted chlorophyll a^ or in vivo chlorophyll a_ fluorescence.
Kiefer and Austin (28) obtained linear correlation coefficients of 0.95
for variation of the beam attenuation coefficient a\ versus chlorophyll _a
fluorescence and of 0.95 for ax versus nc, the concentration of extracted
chlorophyll a_. These measurements were made in a marine environment for
chlorophyll _a values ranging from 1 yg/1 to 6 yg/1. The least-squares best-
fit line for the latter correlation is given by
ax = 0.99 + 0.2 nc (2)
where the c^ values have been corrected for the component of attenuation due
to optically pure seawater at wavelength x. Similar measurements by Baker and
Baker (29), made in fresh water environment, realized a correlation
coefficient of 0.83 for uncorrected ax values versus extracted chlorophyll a_
levels up to 100 yg/1. The least-squares best-fit line for this data set is
aA = 3.57 + 0.11 nc (3)
From these measurements, it would appear that, for extracted chlorophyll a_
levels at least greater than 1 yg/l, a linear relationship exists between
extracted chlorophyll _a with a slope factor lying in the range from 0.1
nrVug/l to 0.2 nrVug/1. For surface water samples containing high and
variable concentrations of non-algal matter in addition to that for the
40
-------
chlorophyll-related pigments, this relationship can be expected to break down
due to the unpredictable contribution to the overall attenuation coefficient
from this non-algal matter. This effect is further exacerbated by the fact
that algal population growth can become arrested or even reversed by the
restrictions put on photosynthetic activity by the reduction in available
solar radiation caused by high concentrations of suspended sediment.
Confirmation of a relationship between nc and a.\ for the surface
waters of Las Vegas Bay was obtained by making a concurrent measurement of in
vivo chlorophyll a_ fluorescence and the beam attenuation coefficient at 610
nm. An in situ measurement of the beam attenuation coefficient was obtained
using a 1-meter path length transmissometer (Martek Model XMS). This device,
which measures beam transmittance T with a collimated and filtered light
source, provides values of a\ from the expression
T = exp (-<*x x)
where x (= 1 m) is the optical path length of the beam. The transmissometer
was rigidly mounted to the bow of the survey launch at a depth of about 1
meter and, as such, provided a continuous chart-recorded profile of the
surface water optical transmission. The present measurements were made at a
wavelength of 610 nm with a filter with a half-height bandwidth of 32 nm, so
that this transmission measurement is a close approximation for kp, the
diffuse attenuation coefficient for the chlorophyll _a fluorescence emission
wavelength at 685 nm. The concurrent in vivo chlorophyll a^ fluorescence
profile was made using a filter fluorometer (Turner Model 111) as described in
Section 7 (c), except that a high volume flow-through sample cell was used in
place of the standard cuvette. The entry port for the flow-through sample
tube was located directly adjacent to the optical path of the transmissometer
with a sample pump located downstream of the fluorometer in order to avoid
damage or contamination of the sample by either cavitation bubbles or pump
lubricant. The continuous in vivo fluorescence and optical transmission
profiles obtained for the region in Las Vegas Bay between stations 9 and 12
are shown in Figure 12. The in vivo fluorescence data were calibrated in
terms of equivalent extracted chlorophyll a_ through use of the single
chlorophyll _a determination performed on a sample obtained at station 11, as
indicated by the single reference point in the chlorophyll a_ profile.
The discrete a values, plotted for equal intervals of time, were
calculated directly from the corresponding 1-meter path length transmission
profile data using Equation 4. The variations in the attenuation coefficient
data are in almost complete correspondence with those for the chlorophyll a_
fluorescence data, after corrections are made for the approximately 6-second
time delay in the fluorometric data. This delay represents the finite time
needed to pump a sample from the location of the submerged transmissometer to
the deck location of the fluorometer. The same discrete attenuation
coefficient values are also shown in Figure 13 plotted against the
corresponding in vivo chlorophyll _§_ values obtained from the calibrated
profile in Figure 12; a linear correlation coefficient of 0.96 was obtained
from this group of 25 data points. Extrapolating the curve to a zero
fluorescence value suggests a background attenuation coefficient value of
about 1.7 m"1, due principally to dissolved and non-algal particulate
41
-------
c
.0
'35
Ifl
-------
8
c
• 5
o
05
o
o
14
1-1
re
3
C
05
E
re
05
CD
0
Linear correlation coefficient
r =0.96 for 25 samples.
Attenuation due
to particulate
and dissolved
material
_L
i
0
Figure 13.
10 15 20 25
in vivo chlorophyll a_(|jg/l)
30
35
Variation of beam attenuation coefficient with in vivo
chlorophyll a_ fluorescence for surface water transect of
Las Vegas Bay in Lake Mead, Nevada, June 8, 1977.
Attenuation data measured at 610 nm.
43
-------
matter. The tendency of the data to deviate from a linear relationship at high
chlorophyll _§_ fluorescence levels may be due possibly to self-absorption of
the fluorescence radiation within the 19-mm diameter sample cell or to an
unrelated increase in the concentration of dissolved or non-algal particulate
matter, which occurs in the region of high chlorophyll a. concentration around
sample station 12.
To gauge the the effect of the relationship between ax and nc on the
performance of the laser fl uorosensor, it is necessary to relate the beam
attenuation coefficients a[_ and ap to the corresponding diffuse
attenuation coefficients kp and k|_ as used in Equation 1. According to
Smith & Tyler (27), ax and kx can be written in terms of their constituent
components, such that
<*X = ax + bf,X + bb,X (5)
and kA = Dax + bb,X (6)
where ax is the absorption coefficient due to dissolved, particulate and
viable algal matter, and bf x and bt>?x are the forward and backward
scatter attenuation coefficients respectively, resulting from all particulate
matter. Note that the expression for the diffuse attenuation coefficient
contains no forward scatter loss term because irradiance measurements monitor
both forward scattered and unscattered radiation that has not been absorbed.
A further complication in the process of relating ax to kx stems from the
fact that kx is not a simple spectral property of a given water sample but
is dependent on the radiance distribution. As a first approximation, this
effect is accounted for by the factor D in Equation 6. The lower bound for D
exists for the case of a collirnated illumination source for which D = 1
whereas the upper bound exists for the case of a totally diffuse source for
which D = 2 (27). Fortunately, these two extremes correspond exactly to the
laser excitation (D=l) and the fluorescence emission (D=2) configurations that
occur during the operation of the airborne laser fl uorosensor. The effective
doubling of the influence of absorption on kx in the case of the diffuse
source is due to the increased path length experienced by individual photons
that are multiply scattered on their way to a given vertical penetration
depth.
The expression for kx and ax can be further broken down into the
contributions for pure water (denoted by subscript w) , chlorophyl 1 -related
pigments (denoted by subscript c) and all other dissolved and particulate
non-chlorophyllous substances (denoted by subscript N), such that
+ bb,c,X + bb,N,X (7)
aw>*+ Dac,X + DaN,X+ bb,c,X
+bb,N,,X (8)
44
-------
where the molecular (Rayleigh) scattering contributions from pure water are
considered to be negligible and have been omitted.
Finally a linear expression can be given for ax in terms of nc as
typified by the data given in Figure 13, such that
aA = Ax + BAnc
(9)
where A^ and BA are presumed to be constant for a given locality and
wavelength A. For the laser excitation case where D-l, Equations 7, 8 and 9
can be solved to give k|_ in terms of measurable parameters A|_ and B|_ and
the scattering coefficients bf c L and bf ^ L, where
kL = BLnc + AL - bfjCsL - bf)N>L (10)
Similarly for the fluorescence emission case with D - 2, Equations 7, 8 and 9
provide a similar expression for kp such that
kF = BFnc + AF -
)} j
w,F + ac,F + aN,F
(11)
Substituting the expressions for k|_ and kp, given by equations 10 and 11
respectively, into Equation 1, we have
nr =
"PF H2"
PL
4irAF
T^Tr^Rec
y2exp {H (0L + pF)f~|
(1-RW)2 J
XN
(12)
Xc, which represents the optical loss terms due to the chlorophyll pigments,
is given by
Xr =
nc (BL + BF)
bf CjF
of'n
Although the terms bf c |_ and
are not given explicitly in'terms
dependence such that Xc will also vary
other optical loss terms, is given by
AL + AF
ac,F - (bf,c,L +
within the
,c,F)J
(13)
,
expression for Xc
it is reasonable to assume a linear
as nc. X|\|, which represents all
bf,N,F)
From this expression, it is clear
in which PF varies linearly as nc,
factor Xc is also dependent on nc.
(14)
that we no longer have a simple relationship
but rather one in which the additional
To monitor relative changes in nc,
it now becomes necessary to obtain exact values for all of the other terms in
addition to (PF/P|_)H2, in particular for the chlorophyll a^ fluorescence
cross section
-------
and scattering losses from non-algal sources are dominant, such that
XN/XC»I, the term Xc can be omitted so that the attenuation
coefficients are no longer dependent on nc. A situation then exists in
which nc varies as (Pp/PiJH2; this is the situation which was assumed
to exist in the derivation of Equation 1 by Browell (20). A different
situation exists however when scattering losses arise principally from the
presence of algae such that XC/XN «1. In this case, nc can be
eliminated from both sides of Equation 12 and the laser fluorosensor
measurement represented by the term (Pp/P[jH2 remains constant for all
values of nc. Put in physical terms, the lower the concentration of algae
or equivalent extractable chlorophyll _a, the larger the penetration depth of
the laser beam. In turn this situation means that the laser beam can
ultimately intercept the same total algae count it that would have encountered
in high algae count situations. As a consequence, the received chlorophyll a.
fluorescence power will remain constant, independent of changes in nc. In
reality, the truth will lie somewhere in between these two extremes depending
upon whether optical losses are dominated by algal or non-algal matter.
Because a direct airborne measurement of the absolute values for k|_ and
kp is not possible within the framework of the present program, an
alternative approach must be considered. A possible means for monitoring
changes in k^ and, in particular, its dependence on nc can be provided by
obtaining a relative indication of laser beam penetration through the surface
water layer by concurrently monitoring the water Raman emission signal. As
the Raman emission is a property of the water alone, the intensity of the
observed signal will depend only on k^ and hence only indirectly on nc.
Changes in the water Raman emission signal can therefore be used to provide
information on changes in k^. A more extensive discussion of this method
will be presented in Section 8.
On the basis of the foregoing discussion, the relative contributions of
variations in ac, k[_ and kp to the degradation of the correlation
between the airborne and ground truth data are not clear. With implementation
of the Raman technique for correcting for changes in k[_ and kp, it should
be possible to isolate the residual uncertainty due to
-------
An airborne experiment was conducted employing laser excitation at 622 nm
rather than at 440 nm, with the purpose of selectively exciting fluorescence
in the blue-green algae. Variations in the biomass for each algal division,
extractable chlorophyll a^ and the laser fluorescence signal as measured over
the established Las Vegas Bay flight line, are shown in Figure 14. A further
comparison of these data is presented in Table 4 in the form of a matrix of
linear correlation coefficients relating biomass from each algal division to
either extractable chlorophyll a_ or the laser fluorsensor signal. Biomass
determinations were made by collecting the algae from a given sample volume
onto a Millipore HA 0.45y membrane filter. Enumeration for all species
present in significant numbers was made with a microscope at either 400X power
or 1000X power. At the latter magnification, the filters were soaked in a
refractive index matching fluid to aid in the identification process. Algal
counts were made for 10 Whipple fields for each filter and a volume
established for each algal species. From these data, biomass in mg/1 was
calculated for each algal division or color group assuming a constant algal
density of 1. Finally, for each sample, an estimate of total algal biomass
was obtained by summing the contributions from each algal division.
Figure 14 shows that the laser fluorosensor signal progressively increases
toward station 12 in contrast to the blue-green algal biomass determination,
which shows a slight downward trend. The correlation coefficient for
covariation between these parameters was -0.48, which is significant at the
10% level. Although no straightforward explanation exists for an apparent
inverse dependence of this kind, a number of interfering effects might be
responsible. The laser fluorosensor signal appears to follow more closely the
trends in the total biomass determination for which the correlation
coefficient has a value of +0.82, significant at the 1% level. This
observation suggests that excitation radiation at 622 nm is being absorbed
directly by the chlorophyll a. present in all algae as well as by c-phycocyanin
common only to blue-green algae. This conclusion is supported by data
presented by Mumola, Jarrett and Brown (16). They show that, although the
excitation spectra for blue-green algae peaks close to 618 nm, its absolute
value given in terms of its fluorescence cross section differs only slightly
from the values for the other algal color groups at 618 nm. In this respect
it is interesting to note that more recent measurements made by this same
group (21) indicate that the fluorescence cross section for a single
blue-green algal species (anacystis marinus) was, on the average, 5 times
larger than values obtained from monocultures representing the other three
(red, golden brown and green) algal color groups. Assuming that the different
algal color groups have similar fluorescence cross sections at 622 nm, it is
not difficult to see how the summed effect of all other algal color groups
will be sufficient to degrade any possible correspondence between the
laser-induced fluorescence intensity and the concentration of blue-green algae
even though the blue-greens were the dominant type over much of the flight
line (see Figure 14). Remote monitoring of blue-green algae is further
complicated by the fact that during non-bloom conditions, blue-greens are
known to congregate below the surface layer. In other studies performed on
Lake Mead (43), it has been noted that the concentration of blue-green algae
tends to reach a maximum at depths below the 1-m surface layer. As the
attenuation coefficients for the surface water along the flight line change
from values less than 1 m"1 to values close to 10 m"1 between sampling
47
-------
oo
1.8
-f. 1.6
\
O5
E
CO
E
o
ra
1.4
1 .2
1.0
0.8
0.6
0.4
0.2
\
Figure 14.
T r
Sampling Station Number
T~I
10
11
11'
A
*
O
A
•
Extractable Chlorophyll j) (Fluorometric Determination)
Laserfluorosensor Signal (Excitation - 622nm; Emission - 685 nm)
Total Algal Biomass
Blue Green Algal Biomass
Green Algal Biomass
Dinoflagellate Biomass
Diatom Biomass
2000
4000
6000
12
18
16
14
12.
cn
3.
.G
Q.
O
8 £
JI
O
oj
6 3
o
(0
LLJ
8000 10,000
Distance from Station 1 (m)
Variation of laser-fluorosensor signal, extractable chlorophyll _§_, total biomass
and algal color group biomass with distance for surface transect of Las Vegas Bay in
Lake Mead, Nevada. Flight #12, August 16, 1977.
-------
TABLE 4. CORRELATION COEFFICIENT MATRIX FOR ALGAL BIOMASS OBTAINED BY
ENUMERATION VERSUS LASER-FLUOROSENSOR SIGNAL AND EXTRACTED
CHLOROPHYLL a FOR A GROUP OF 13 SAMPLES
— ^^ Chlorophyll a
_. ;^^^_ Indicator
Biomass by ^~^-^^
Algal Division ^^^-~~^^
(mg/l) ^\^^
Cyanophyta Blue Greens (BG)
Chlorophyta Greens (G)
Bacillariophyta Diatoms (Dia)
Pyrrophyta Dinoflagellates (Dino)
BG + G
BG + G + Dia
BG + G + Dino
Total Count
A
Airborne
Laserfluorosensor
Signal
-0.48
0.88
0.54
-0.08
0.90
0.89
0.81
0.82
B,C
Extracted
Chlorophyll a.
0.57
0.51
0.23
0.36
0.43
0.42
0.29
0.31
A) Chlorophyll a fluorescence emission at 685 nm excited at 622 nm.
B) Fluorometric determination of extracted chlorophyll a^ using method of Holm-Hansen et al (39)
C) Correlation between chlorophyll a and laserfluorosensor signal gave coefficient of 0.408.
49
-------
situations 1 and 12, it is possible that the presence of blue-green algae is
being progressively screened from the laser fluorosensor by increasing
concentrations of the intervening material in the surface layer, in particular
algae from other color groups. This explanation assumes that the depth at
which the blue-green algae congregate remains a constant regardless of ambient
conditions. In reality, however, this situation will be complicated by the
tendency of the algae to adjust their depth according to the available light
level compounded with the effects of other environmental factors, such as
water temperature and availability of nutrients and dissolved oxygen.
It is interesting to note that extractable chlorophyll d_ exhibited a poor
correlation with both total biomass (r = 0.31) and the laser fluorosensor
signal (r = 0.41). The first result is somewhat surprising, as both the algae
counts and the chlorophyll a_ extractions were performed on subsamples of the
same grab sample material. More disturbing is the fact that chlorophyll a_
determinations are employed principally as a substitute or indicator for total
algal biomass. In the light of these correlations between the laser
fluorosensor signal and total algal biomass (r = 0.82), it is suggested that
errors in the chlorophyll at determinations are the principal source of this
discrepancy. As discussed in Section 7 (e), the established procedure for
extracting chlorophyll a_ from algae is both inefficient and unreliable in
terms of reproducibility. The large deviations in the extractable chlorophyll
a_ profile in Figure 14 are considered to be due primarily to this problem.
Improved methods for extracting chlorophyll a_ from algae are presently under
investigation and will be reported elsewhere. In contrast to the
spectrophotometric approach of Strickland and Parsons (38), the fluorometric
method of Holm-Hansen et al . (39), used to make the determinations for this
flight, does not correct for the fluorescence contributions from the
chlorophyll b^ or carotenoid pigments. It is therefore planned to employ the
spectrophotometric technique for establishing extractable chlorophyll a_ values
in all future ground truth surveys. Because of this uncertainty in the
accuracy of the chlorophyll ^ determinations for this flight and its effect on
the correlation between the airborne and ground truth data, it has not been
possible to establish the efficacy of using the 622-nm excitation wavelength
for monitoring chlorophyll a_ in vivo. Future tests will therefore be
conducted at a number of excitation wavelengths, in particular at 622 nm and
470 nm. The latter wavelength, although not responding to the pigments
specific to red and blue-green algae, should provide a better chlorophyll a^
fluorescence response to the chlorophyll JD and carotene pigments present i~n
most algae than the original 440-nm excitation wavelength.
50
-------
SECTION 8
AIRBORNE MEASUREMENTS OF THE INFLUENCE OF THE ATTENUATION COEFFICIENTS
THEORETICAL CONSIDERATIONS
A relative measure of the variation in the diffuse attenuation coefficient
k^ for-surface waters, and consequently its dependence on nc, can be
obtained by monitoring the |OH| stretch water Raman emission at wavelength R
excited at laser wavelength L, concurrently with the measurement of
chlorophyll _a_ fluorescence emission at wavelength F. The wavelength of the
Raman emission is determined by the excitation wavelength and the Raman
frequency shift. For example, with laser excitation at 440 nm and a fixed
JOHJ stretch water Raman frequency shift of 3,418 cm-1, the Raman emission
occurs at 518 nm. The intensity of the backscattered Raman emission signal
will depend on the values of k|_ and k^ and not on nw, the concentration
of water, which remains essentially constant. A laser fl uorosensor equation
for the detected Raman emission power PR can be written in a form analogous
to the expression for the fluorescence power Pp obtained from Equation 1
such that
w
R
H2/ (kR + kL)
(15)
where aw is the Raman emission cross section for the OH stretch
vibrational mode of liquid water.
-------
where
dp =
4irAp
(kF + kL)
(1-Rw)2
(17)
exp
(18)
and is considered to be constant for the same reasons as given for dR. An
expression can now be written for the ratio PF/PR using Equations 15 and
17, so that
dF
nw crw dR
As nw, erw,
-------
it would be advantageous to employ an excitation wavelength which shifts the
water Raman emission band closer to the chlorophyll a_ fluorescence emission
band at 685 nm.
It is therefore proposed to modify the airborne-laser fluorosensor
described in Section 6 so that the water Raman signal can be detected,
monitored and recorded concurrently with the fluorescence signal Pp, which
will still be corrected for variations in P|_ and H. The ratio PP/PR
will then be correlated with chlorophyll a^ ground truth data. An improved
correlation between PF/PR and the chlorophyll a_ ground truth data, rather
than with Pp alone, will justify the assumption that the ratio (k[_ +
kR)/(k[_ + kp) remains essentially constant.
LABORATORY MEASUREMENT OF FLUORESCENCE TO RAMAN RATIO FOR LAKE WATER SAMPLES
Before adding a Raman emission detection channel to the existing laser
fluorosensor, it was considered expedient to conduct a laboratory experiment
to simulate the airborne measurement of the ratio Pp/PR and in particular
to ensure that PR, the Raman emission power, could be measured with the same
degree of sensitivity as the chlorophyll ji fluorescence power, Pp. Liquid
water consists of two spectroscopically distinct forms in thermal equilibrium.
These polymeric and monomeric types influence the |0h| molecular band strength
and ultimately the intensity, spectral shape and polarization state of the JOHJ
stretch Raman emission (45). In particular, the spectral characteristics of
the Raman emission are highly dependent on the state of polarization of the
excitation radiation, the scattering configuration used to view the emission
and the polarization state of the detection system (46). The airborne laser-
fluorosensor, of necessity, operates in a downward-looking mode which
constitutes a 180° scattering angle with a plane-polarized laser source.
Consequently, because of the dependence of the Raman band depolarization ratio
on the scattering configuration, the state of polarization of the laser
output, and the polarization sensitivity of the detection system, it is
essential that the laboratory measurements simulate the airborne measurement
as closely as possible. A plane-polarized laser was therefore employed as the
excitation source, and the Raman radiation viewed in the 0° scattering
configuration. The 0° scattering configuration, which is spectroscopically
equivalent to a 180° format, was found to be more convenient to implement in a
laboratory experiment. Finally, precautions were taken to ensure that the
polarization sensitivity of the grating monochromator did not introduce errors
into the measurement of the PP/PR ratio. This step was accomplished by
depolarizing both the fluorescence and Raman emission prior to transmission
through the monochromator. This procedure was found to be necessary because
spectral discrimination in the airborne system is performed with interference
filters at normal incidence, which are not sensitive to the state of
polarization of the incident radiation.
An optical schematic of this laboratory-mounted system is shown in Figure
15. A nitrogen laser pumped dye laser (AVCO Model 3000A), operating at 440 nm
and, at a repetition rate of 500 Hz, was used to irradiate the lake water
samples. This laser output has a pulse width of approximately 5 nsec with a
spectral bandwidth of about 0.2 nm and is linearly polarized to within 1 part
per 1,000. The stability of the laser output power was monitored with the use
53
-------
1/6.8 scanning
monochromator
(GCA McPherson EU-700)
Laser
blocking
filter (Hoya Y46) 2cm Sample cell
440 nm(Coumarin 120)
dye laser
(AVCO model 3000A)
Red sensitive
PMT(R-928)
Quartz collector'
lens
Laser
beam stop
Quartz wedge
depolarizer
Quartz beam I
splitter
Quartz diffuser
Silicon PIN diode
laser monitor
2Ghz Sampling Tunable
oscilloscope low pass
filters
Chart recorder
Figure 15. Laboratory simulation of airborne laser fluorosensor for monitoring chlorophyll a_
fluorescence and water Raman signals.
-------
of a beam splitter to deflect a small percentage of the beam onto a PIN
silicon photodiode via a quartz diffuser. The fluorescence and Raman
emissions from the 2-cm pathlength cell were collected and focused onto the
entrance slit of an f/6.8 scanning monochromator. A compensated, quartz wedge
depolarizer was used to depolarize the fluorescence and particularly the Raman
emission with a view to avoiding polarization artifacts being introduced into
the ratio PF/PR by the polarization sensitive grating monochromator.
Due to use of the 180° scattering configuration, care had to be taken to
ensure that laser radiation did not reach the detector. In addition to the
spectral discrimination provided by the single monochromator, a short
wavelength cutoff glass filter (Hoya Y46) was employed to block wavelengths
shorter than 460 nm. Further laser blocking was achieved by placing a small
piece of black tape in front of the filter on the laser beam axis as shown in
Figure 15. A fast, high-gain, red-sensitive photomultiplier (Hamamatsu R928),
located close to the monochromator exit slit, was used to detect the
fluorescence and Raman signals. Fluorescence and Raman bands of interest were
obtained by scanning the monochromator over the spectral region of interest
while monitoring the peak value of the emission pulse detected by the
photomultiplier. The amplitude of this pulse was measured using a sampling
oscilloscope (Philips PM 3400) operating in the non-scanning mode. The
calibrated DC output signal from the oscilloscope sample and hold circuit, a
measure of the peak amplitude of the detected pulse, was low-pass filtered to
reduce photon noise effects and then plotted out directly as a function of
wavelength using a strip chart recorder. The monochromator slit width and
scanning rate, the low pass filter cutoff frequency, and the chart recorder
speed were all adjusted so as to ensure that the narrowest spectral feature
could be faithfully resolved. The JOH| stretch water Raman band, which is
located at 518 nm when excited at 440 nm has an established FWHM bandwidth of
11.5 nm at this wavelength and consequently is an ideal source for this
purpose.
The spectra for two Las Vegas Bay surface water samples collected on
August 16, 1977 and March 31, 1978 are shown in Figures 16 and 17
respectively. The respective chlorophyll a^ levels were 8.4 yg/1 and 3.4 yg/1 .
It is clear, even from these rather noisy spectra, that, for the chlorophyll a_
range of interest (about 0.1 yg/1 to 50 yg/1), the Raman band will prov-ide a
large and measurable signal for the airborne laser fluorosensor in relation to
the chlorophyll a_ signal level as represented by the fluorescence pulse shown
in Figure 7 for a chlorophyll _a level of 10.5 yg/1.
Two points are worthy of note with regard to these spectra. First they
are uncorrected for the varying spectral sensitivity of the grating
monochromator and the photomultiplier. Both components have lower
sensitivities in the red region at 685 nm than in the blue region at 518 nm.
The blue-to-red sensitivity ratio for the R928 photomultiplier is 1.71 to 1,
whereas that for the grating monochromator is approximately 1.25 to 1, giving
a total ratio of 2.14 to 1. However, as the C31000A photomultiplier used in
the airborne laser-fluorosensor has a blue-to-red sensitivity ratio of
approximately 2.75 to 1, it is clear that the airborne measurement of the
water Raman signal will be further enhanced in relation to the chlorophyll a_
fluorescence signal. The second point concerns the fluorescence background
55
-------
440 nm Laser
excitation
bandwidth
0.2nm
[OH] Stretch
water Raman
emission
at 518 nm
Emission
monochromator
bandwidth
2 nm
Fluorescence
emission from
dissolved organics
Fluorescence emission
from 'in vivo' chlorophyll a
at 685 nm- measured
concentration 8.4 ug/l
480
540
600
660 720
WAVELENGTH (nm)
Figure 16. Uncorrected fluorescence and Raman emission spectra of Lake Mead surface water sample,
excited at 440 nm, August 16, 1977.
-------
>
LU
LJJ
>
5
LJJ
DC
[OH] Stretch
water Raman
emission
at 515nm
438nm Laser
excitation
bandwidth
0.2nm
Emission
monochromator
bandwidth
2nm
Fluorescence
emission from
dissolved organics
Fluorescence
emission
from 'in vivo'
chlorophyll .a,
— measured
concentration 3.4
ug/l
490
530
570
610 650 690
WAVELENGTH (nm)
Figure 17.
Uncorrected fluorescence and Raman emission spectra of Lake
Mead surface water sample, excited at 438 nm, March 31, 1978,
57
-------
due to dissolved organics, indicated in Figures 16 and 17 by broken curves.
This background signal, which has a fluorescence emission maximum in the
region of 430 nm, constitutes only a small fraction of the total signal at 685
nm for Lake Mead surface water and, as such, its influence on the chlorophyll
a^ fluorescence measurement can be neglected. In addition this background
fluorescence signal has been shown to remain fairly constant over the extent
of Lake Mead at any one time (47). The present plan is therefore to use the
total emission intensity at the Raman band wavelength as a reference signal
rather than just the peak intensity of the Raman band. This expedient is
acceptable provided that the concentration of dissolved organics for a given
water surface remains constant in both the horizontal and vertical dimensions
at any one time as the background fluorescence signal at this wavelength will
be attenuated equally with the Raman signal. This assumption will be
unacceptable for situations where the concentration of surface-water dissolved
organics varies considerably over short horizontal distances at any one time.
A more advanced laser fluorosensor would therefore have to separate the water
Raman band from the background fluorescence. This goal could be achieved by
monitoring the background fluorescence intensity on either side of the Raman
band in addition to the total peak intensity at the Raman band wavelength. By
simple interpolation, the peak intensity of the Raman band can then be
determi ned.
MODIFICATIONS TO LASER FLUOROSENSOR NEEDED TO MEASURE THE FLUORESCENCE-
TO-RAMAN RATIO
The proposed modifications to the airborne laser fluorosensor system are
shown schematically in Figure 18.
The transmitter-receiver unit is identical to the existing laser
fluorosensor shown in Figure 3 except for the addition of the Raman band
photomultiplier detector (RCA 31000A), which is gated in the same manner and
at the same time as the fluorescence detector. The Raman signal is separated
from the fluorescence signal by means of a neutral (50 50) beam splitter. The
modified electronic detection, monitoring and recording system is shown
schematically in Figure 19.
This system is essentially identical to the original one shown in Figure 6
but with the addition of a 50-ohm delay line needed to delay the Raman signal
in relation to the fluorescence signal. This delay allows the fluorescence
and delayed Raman pulses to be combined and recorded sequentially on a single
waveform digitizer input channel. By employing this approach, loss of analog
input bandwidth in the digitizer (Biomation 8100) is avoided. The anticipated
waveform is shown schematically in Figure 20, where the fixed delay period
between the fluorescence and Raman pulses will have a value of 3 ysec.
58
-------
Fluorescence signalj-
Gating pulsej
Gating pulsej
Raman signalj
Raman band filter
Laser blocking filters
Neutral beam splitter
Collimating lens
Focal plane aperture
Diffuser\
Laser
power
monitor^
(PIN diode)
—[Laser pulse
Shock
mounts
Prism tuned
flashlamp pumped
pulsed dye laser
Laser beam
steering
mirror \
, Raman photomultiplier
Fluorescence
photomultiplier
Fluorescence
band filter
Front surface
mirror
Refracting telescope
30cm diameter f/1.8 Fresnel
collector lens
Airc.raft
Y
Figure 18. Optical diagram of airborne laser fluorosensor for monitoring chlorophyll a_ fluorescence
and water Raman signals.
-------
H>
01
o
Fluorescence
gated photo-multiplier
(RCA C31000A)
Delayed gate
generator
30 cm dia.
Fresnel
refracting
telescope
Raman
gated
photomultiplier
(RCA C31000A)
Pulse
generator
Dual beam
oscilloscope
(Tektronix R7844)
V
Prism tuned'
flash lamp pumped
dye laser
(Phase-R DL 1200)
TV camera
Low
pass Pulse
filters splitters
(8 MHe)
power
monitor
(silicon
photodiode)
Waveform
digitizer
(Biomation 8100)
9 Track digital
tape recorder
(Kennedy 9800)
Video tape
recorder
V
Video/digital
time code
generator
Interface
Figure 19. Schematic of airborne laser fluorosensor for measuring chlorophyll a_ fluorescence and
water Raman emission showing detection, monitoring and recording systems.
-------
t
LU
h-
UJ
>
<
LU
(ALTITUDE ~Tc/2)
Fixed
delay
period
Fluorescence
emission
Laser
peak
power
Solar background
Water
Raman
emission
Solar background
h
Fluorescence
detector
gate period
H h
Raman
detector
gate period
H
TIME
Figure 20. Variation of laser, fluorescence and Raman laser-fluorosensor signals as a function
of time.
-------
REFERENCES
1. Johnson, R. W. Quantitative Mapping of Chlorophyll a_ Distributions in
Coastal Zones by Remote Sensing. Presented at 43rd Annual Meeting of
American Society of Photogramrnetry, Washington, D.C. February 27 -
March 5, 1977.
2. Arvesen, J.C., E.C. Weaver and J.P. Mil lard. Rapid Assessment of Water
Pollution by Airborne Measurement of Chlorophyll Content. Presented
at Joint Conference on Sensing of Environmental Pollutants, Palo Alto,
California, November 8-10, 1971, AIAA Paper #71-1097.
3. McNeil, W.R., K.P.B. Thompson and J.Jerome. The Application of Remote
Spectral Measurements to Water Quality Monitoring. Canadian Remote
Sensing Journal 2, 48-58, 1976.
4. Witte, W.G. Evaluation of the Dual Differential Radiometer for Remote
Sensing of Sediment and Chlorophyll in Turbid Waters. Proceedings of
Fourth Annual Remote Sensing of Earth Resources Conference, Tullahoma,
Tennessee, March 24-26, 1975.
5. Johnson, R. W. Application of Aircraft Multispectral Scanners to
Quantitative Analyses and Mapping of Water Quality Parameters in the
James River, Virginia. Presented at XIX Meeting of COSPAR - Latest
Results of Earth Survey. Philadelphia, Pa., June 1976. Published in
COSPAR Space Research Vol. XVII, Pergamon Press, Oxford 1977.
6. Bowker, D.E. and W.G. Witte. The Use of LANDSAT for Monitoring Water
Parameters in the Coastal Zone. Proceedings of AIAA 1977 Joint
Conference on Satellite Applications of Marine Operations, New Orleans,
La., November 15-18, 1977.
7. Neville, R.A., and J.F.R. Gower. Passive Remote Sensing of Phytoplankton
via Chlorophyll _a Fluorescence. J_. Geophys. Res. Q2_, 3487-3493, 1977.
Stoertz, G.E., W.R. Hemphill and D.A. Markle. Airborne Fluorometer
Applicable to Marine and Estuarine Studies. Marine Technology Society
8.
Journal . 3., 11-26, 1969.
9. McFarlane, C., and R.D. Watson. The Detection and Mapping of Oil on a
Marshy Area by a Remote Luminescent Sensor. Proceedings of API 1977
Oil Spill Conference. March 8-10, 1977. New Orleans, La. pp 197-201.
10. Hemphill, W.R., R.D. Watson, R.C. Bigelow and T.D. Hessen,
Measurement of Luminescence of Geochemically Stressed Trees and
Other Materials. Proceedings of First Annual William T. Pecora
Memorial Symposium, October 1975, Sioux Falls, South Dakota, pp 93-112
62
-------
11. Kim, H.H. and G.D. Hickman. An Airborne Laser Fluorosensor for the
Detection of Oil on Water. Proceedings of the Symposium on the Use of
Lasers for Hydrographic Studies, NASA Wallops Island, Virginia,
September 12, 1973, NASA Special Publication, SP375, Washington, D.C.
1975.
12. O'Neil, R.A., A.R. Davis, H.G. Gross and J. Kruus. A Remote Sensing
Laser Fluorometer. Am. Soc. Test. Mater., Special Technical Publication
#573, pp 424-436, 1975.
13. Bristow, M.P.F., Airborne Monitoring of Surface Water Pollutants by
Fluorescence Spectroscopy. Remote Sensing of Environment 7, 105-127,
1978.
14. Fantasia, J.F. and H.C. Ingrao. Development of an Experimental Airborne
Laser Remote Sensing System for the Detection and Classification of Oil
Spills. Proceedings of the 9th International Symposium on Remote Sensing
of the Environment, Environmental Research Institute of Michigan, Ann
Arbor, Michigan, April 15-19, 1974, pp 1711-1745.
15. Kim, H.H. New Algae Mapping Technique by the Use of Airborne Laser-
fluorosensor. Appl . Opt. 12, 1454-1459, 1973.
16. Mumola, P.B., 0. Jarrett and C.A. Brown. Multiwavelength Laser Induced
Fluorescence of Algae in-vivo: A New Remote Sensing Technique.
Proceedings of Second Joint Conference on Sensing of Environmental
Pollutants, Washington, D.C., December 10-12, 1973.
17. Hickman, G.D. and R.B. Moore. Laser Induced Fluorescence in Rhodamine
B and Algae. Proceedings of the 13th Conference on Great Lakes Research,
Buffalo, New York, March 31 - April 3, 1970. pp 1-14.
18. Mumola, P.B. and H.H. Kim. Remote Sensing of Marine Plankton by Dye
Laser Induced Fluorescence. Proceedings of IEEE Conference on
Engineering in the Ocean Environment, Newport, Rhode Island, September
13-15, 1972, pp 204-207.
19. Loftus, M.E. and H.H. Seliger. Some Limitations of the In Vivo
Fluorescence Technique. Chesapeake Science 16, 79-92, 1975.
20. Browell, E.V. Analysis of Laser Fluorosensor Systems for Remote Algae
Detection and Quantification, NASA Tech Note TN D-8447, June 1977.
21. Brown, C.A., F.H. Farmer, 0. Jarrett and W.L. Staton. Laboratory
Studies of In Vivo Fluorescence of Phytoplankton. Proceedings of
Fourth Joint Conference on Sensing of Environmental Pollutants,
New Orleans, La. November 6-11, 1977, pp 782-788.
63
-------
22. Blasco, D. Variations of the Ratio in vivo - Fluorescence/a -
Chlorophyll and its Applications to Oceanography. Effect of Limiting
Different Nutrients, of Night and Day and Dependence on the Species Under
Investigation. NASA Technical Translation TTF - 16,317, 1975 from
original Spanish document published in Investigacion Pesquera 37,
533-556, 1973.
23. Tunzi, M.G., M.Y. Chu and R.C. Bain. In Vivo Fluorescence, Extracted
Fluorescence and Chlorophyll Concentrations in Algal Mass Measurements.
Mater Research 8, 623-635, 1974.
24. Anon., Chlorophyll and Pheophytin. Technical Note 10-577 - C & P, 1977,
Turner Designs, Mountain View, California.
25. Friedman, E. J. and G.D. Hickman. Laser Induced Fluorescence in Algae.
A New Technique for Remote Detection. NASA Contractor Report CR-62090,
1972.
26. Lorenzen, C.J. A Method for the Continuous Measurement of in vivo
Chlorophyll Concentration. Deep Sea Research 13, 223-227, 1966.
27. Smith R.C. and J.E. Tyler. Transmission of Solar Radiation into
Natural Waters. Photochemical and Photobiological Reviews. !_, 117-
155, 1976. Ed by K.C. Smith, Plenum Press, New York.
28. Kiefer, D.A. and R.W. Austin. The Effect of Varying Phytoplankton
Concentration on Submarine Light Transmission in the Gulf of California.
Limnol. Oceanogr. _19_. 55-64, 1974.
29. Baker A.L. and K.K. Baker. Estimation of Planktonic Wind Drift by
Transmissometry. Limnol . Oceanogr. 21, 447-452, 1976.
30. Campillo, A.J., V.H. Kollman and S.L. Shapiro. Intensity Dependence
of Fluorescence Lifetime of in vivo Chlorophyll excited by a Picosecond
Light Pulse. Science, 193, #4249, July 16, 227-229, 1976.
31. Bristow, M.P.F. and N. deVilliers. On the Relative Merits of Using
Either C.W. Helium - Cadimum or Pulsed Nitrogen Lasers as the
Fluorescence Excitation Source in Airborne Laser Fl uorosensing Systems.
Technical Note 73-6. November 1973, Canada Centre for Remote Sensing,
Department of Energy, Mines and Resources, Ottawa, Canada.
32. Grant, W.B. and J.G. Hawley, Prevention of Fire Damage due to Exploding
Dye Laser Flashlamps. Appl. Opt. L4, 1257-1258, 1975.
33. Riede, M.J., The Single Thin Lens as an Objective for IR Imaging
Systems. Electro-Optical Systems Design, 6, #11, 58-65, 1974.
34. Deacon, J.E. Lake Mead Monitoring Program. Final Report. Dept. of
Bilogical Sciences, University of Nevada, Las Vegas, September 1977.
64
-------
35. Loftus, M.E., D.V. Subba Rao and H.H. Seliger. Growth and Dissipation of
Phytoplankton"in Cheasapeake Bay. I. Response to a Large Pulse of
Rainfall. Cheasapeake Science. 13. 282-299, 1972.
36. Baudouin, M.F. and P. Scoppa. Fluorometric Determination of Chlorophyll
ai in the Presence of Pheopigments: Effect of the Half - Value width of
the Excitation Source. Marine Biol. _10, 66-69 (1971).
37. Scheid, F. Chapter 21 in 'Numerical Analysis', Schaums Outline Series,
Schaum, New York, 1968.
38. Strickland, J.D.H. and T.R. Parsons. A Practical Handbook of Seawater
Analysis. Second Edition, Part IV, Section 3.1, Spectroph.otometric
Determination of Chlorophylls and Total Carotenoids. Fisheries Research
Board of Canada, 1972.
. ' !
39. Holm-Hansen, 0., C.J. Lorenzen, R.W. Holmes and J.D.H. Strickland.
Fluorometric Determination of Chlorophyll. J^ Cons, perm, int. Explor.
Mer. 30, 3-15, 1965.
40. Standard Methods for the Examination of Water and Wastewater, 14th
Edition, American Public Health Association, 1975.
41. Shoaf, W.T. and B.W. Li urn. Improved Extraction of Chlorophyll a. and b_
from Algae Using Dimethyl Sulfoxide. Limnol. Oceanogr. 21^, 926-928, 1976.
42. Marker, A.F.H. The Use of Acetone and Methanol in the Estimation of
Chlorophyll in the Presence of Phaeophytin. Freshwat. Biol. 2, 361-385,
1972.
43. Deacon, J.E. and R.W. Tew. Inter-relationships between Chemical,
Physical and Biological Conditions of the Waters of Las Vegas Bay of Lake
Mead. Final Report, Dept. of Biological Sciences, University of Nevada,
Las Vegas, 1972.
44. Walrafen, G.E. Raman Spectral Studies of the Effects of Temperature on
Water Structure, J.. Chem. Phys. 47^ 114-126, 1967.
45. Chang, C.H. and L.A. Young. Seawater Temperature Measurement from Raman
Spectra. AVCO Research Note 960, AVCO Everett Research Laboratory Inc.,
Everett, Mass. January, 1974.
46. Gilson, T.R. and P.J, Hendra. Laser Raman Spectroscopy. Wiley-
Inter-science, New York. 1970.
47. Bristow, M. and D. Nielsen. The Application of Fluorescence
Spectroscopy to Remote Monitoring of Total Organics in Surface Waters.
To be published.
65
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/4-79-048
2.
4. TITLE AND SUBTITLE
AIRBORNE LASER FLUOROSENSING OF SURFACE WATER
CHLOROPHYLL a
6. PERFORMING ORGANIZATION CODE
3. RECIPIENT'S ACCESSI ON-NO.
5. REPORT DATE
August 1979
N/A
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO
M. Bristow, D. Nielsen, D. Bundy, R. Furtek and
J. Baker
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Environmental Monitoring & Support Laboratory
Office of Research & Development
U.S. Environmental Protection Agency
Las Vegas, NV 89114
10. PROGRAM ELEMENT NO.
1HD 620A
11. CONTRACT/GRANT NO.
N/A
12. SPONSORING AGENCY NAME AND ADDRESS
U.S. Environmental Protection Agency—Las Vegas, NV
Office of Research & Development
Environmental Monitoring & Support Laboratory
Las Vegas, NV 89114
13. TYPE OF REPORT AND PERIOD COVERED
Interim 1/76 to 8/78
14. SPONSORING AGENCY CODE
EPA/600/07
15. SUPPLEMENTARY NOTES
16. ABSTRACT
A prototype airborne laser fluorosensor for monitoring surface water
chlorophyll a_ has been tested over Lake Mead, Nevada. Trends in the remotely
sensed data are in close correspondence with ground truth data. It is suggested
that system performance can be improved by concurrently gauging the water optical
attenuation coefficient and by implementing changes to the established procedure
for performing chlorophyll _a analyses on ground truth samples.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. cos AT I Field/Group
Chlorophyll a_
Remote Sensing
Laser Fluorosensor
Fluorescence
Surface Waters
46C
48C
63C
68D
3. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
UNCLASSIFIED
21. NO. OF PAGES
76
20. SECURITY CLASS (This page)
UNCLASSIFIED
22. PRICE
A05
EPA Form 2220-I (9-73)
------- |