The methods for remote sensing of phytoplankton described above make use of the fact that algal cells absorb light in a certain region of the spectrum. However, algal cells also emit light: in the laboratory, typically about 1% of the light a photosynthesizing cell absorbs is re-emitted as fluorescence, with a peak at about - 685 nm. Fluorescence quantum yield (ff) in the sea is calculated on the basis of measurements of the spectral distribution of downwelling irradiance and upwelling radiance underwater, together with measurements of absorption spectra of the particu-late and soluble fractions. In the tropical Pacific Ocean, Maritorena et al. (2000), for near-surface waters, found an average fluorescence quantum yield of 0.84% in oligotrophic waters of the central Pacific and 1.53% for the productive waters of the Peruvian upwelling. ff increased with depth, to ~2% in the Prochlorococcus-dominated oligotrophic waters, but to 5 to 6% in the diatom-dominated productive waters. Maximal values were found near 70 m. In the North Sea, off the northeastern coast of the United Kingdom, Morrison (2003) also found ff to increase with depth, typically from ^0.5 to 1% at the surface to 4 to 6% at depths ranging from 5 to 20 m. In the turbid waters of the North China Sea, Xing et al. (2008) found ff near the surface to have an average value of 1.7%. The maximum was 6%, but most values were in the range from 0.1 to 2%. Phytoplankton fluorescence is too feeble to be detected in the downwelling light stream within the water but can show up as a distinct peak in the spectral distribution of the upwelling stream or in the curve of apparent reflectance against wavelength (see Fig. 6.7).956,1385
This peak can also be detected in the spectral distribution of the emergent flux. Calculations indicate that the increased fluorescence associated with an increase in phytoplankton chlorophyll of 1 mg m-3 in the water would lead to an additional upward radiance of 0.03 Wm-2sr-1 mm-1 above the water.396,397 Neville and Gower (1977) showed that in the radiance reflectance (upward radiance/downward irradiance) spectra of productive British Columbian coastal waters obtained at the Brewster angle from a low-flying aircraft, a distinct peak at 685 nm was present, the height of which was proportional to chlorophyll concentration in the upper few metres of water. In a coastal inlet the height of the fluorescence peak above the baseline (measured above the water, but from a boat) was linearly correlated (r2 = 0.85) with a weighted average (allowing for vertical distribution) chlorophyll a concentration over the range 1 to 20mgm-3.501 On the basis of the observed height of the peak it seems unlikely that the results would be of acceptable accuracy below about 1 mg chlorophyll a m-3.
The MERIS spaceborne sensor has a waveband, Band 8, centred on 681.25 nm, on the short-wavelength side of the chlorophyll emission maximum to avoid the atmospheric oxygen absorption band at 687 nm. Gower and King (2007) extracted the fluorescence signal (as fluorescence line height, FLH) from the top-of-atmosphere radiance values using the radiance excess in Band 8 over a linear baseline computed from the radiances in Bands 7 and 9 (665 and 709 nm). By comparing the fluorescence line height from MERIS radiances with in situ measurements of chlorophyll on the continental shelf off Vancouver Island and Washington State they were able to express FLH as a function of chlorophyll concentration (chl, mg m~3) and solar zenith angle (00)
The purpose of the 0.2 chl term in the denominator is to take account of the re-absorption of fluorescence by chlorophyll: as might be expected this becomes more significant at higher levels of phytoplankton. This equation can readily be inverted for the calculation of chlorophyll concentration from measured fluorescence.
The MODIS sensor on the Terra and Aqua satellites is also equipped to measure solar-stimulated fluorescence of phytoplankton, using Bands 13 (665.1 nm), 14 (767.7 nm) and 15 (746.3 nm), all with 10 nm bandwidth. To obtain the FLH, a baseline is formed between the radiances for Bands 13 and 15, and then subtracted from the Band 14 radiance. Hu et al. (2005) used MODIS FLH to monitor the development and movement of a 'red tide' of the toxic dinoflagellate, Karenia brevis, in coastal waters off southwest Florida. Using the algorithm
developed with in situ chlorophyll samples collected within a few hours of the satellite overpass, they were able to achieve satisfactory mapping of the red tide. The standard chlorophyll algorithm using reflectance ratios, by contrast, did not work well for this system, consistently overestimating chlorophyll several-fold.
A problem with using Sun-induced chlorophyll fluorescence to measure chlorophyll is that the relationship between the two is not at all constant. For British Columbia coastal waters, Gower et al. (1999) found the slope of the approximately linear relationship between FLH and chl concentration to vary through a factor of ^2.5 for measurements made in the period 1975 to 1981. The height of the fluorescence peak is markedly affected by the concentrations of CDOM and suspended particles in the water.502,889 Another problem with the method is that even in the chlorophyll fluorescence waveband, most of the light in the emergent flux arises from scattering rather than fluorescence. To address this problem, Gilerson et al. (2006) have made use of the fact that the elastically scattered component is partially polarized, while the fluorescence signal is unpolarized. With a rotating polarizer in front of the optical probe it was possible, both in algal cultures and two inshore coastal sites, to greatly reduce the proportion of scattered light in the measured radiance. The method does not work well in the presence of very high concentrations of suspended mineral particles, which cause depolarization of the scattered light, but should be generally applicable to open ocean and coastal waters. On the basis of their data the authors conclude that the traditional method of extracting fluorescence height using the baseline method can give significant errors, particularly for coastal waters where it strongly overestimates the fluorescence values.
Phytoplankton chlorophyll can be estimated using the fluorescence excited by light from an airborne laser rather than by sunlight, and the method has certain advantages. A key problem in the use of remotely sensed chlorophyll fluorescence as an indication of phytoplankton biomass is that the proportion of the fluorescent light that succeeds in passing up to the sensor, as well as the proportion of the exciting light (solar or laser) that succeeds in reaching the algae, depends on the optical properties of the water. Two different water bodies with the same phyto-plankton content, but different attenuation properties, could give quite different fluorescence signals. To correct for the effects of attenuation by the aquatic medium, Hoge and Swift (1981) and Bristow et al. (1981) have made use of the laser-induced Raman emission of water.
When water molecules scatter light, most of the scattered light undergoes no change in wavelength. A small proportion of the scattered photons, however, when they interact with the scattering molecule, lose or gain a small amount of energy corresponding to a vibrational or rotational energy transition within the molecule, and so after scattering are shifted in wavelength. These appear in the scattered light as emission bands at wavelengths other than that of the exciting light, and are referred to as Raman emission lines, after the Indian physicist who discovered this phenomenon. A particularly strong Raman emission in the case of liquid water arises from the O-H vibrational stretching mode: this shows up as an emission band roughly 100 nm on the long-wavelength side of the exciting wavelength (Fig. 7.12). Since the water content of the aquatic medium is essentially constant, the intensity of this Raman emission, when remotely sensed from above the water is, for a given exciting light
source, determined entirely by the light-attenuating properties of the water, low Raman signals indicating high attenuation, and vice versa.
Thus to correct for variation in the optical character of the water, the measured chlorophyll fluorescence intensity at each station is divided by the corresponding water Raman emission intensity. Chlorophyll fluorescence values measured from the air, and normalized in this way, have been found to correlate closely with chlorophyll contents determined on water samples.166
Bristow et al. (1981) used a pumped dye laser, emitting at 470 nm, and exciting Raman emission at 560 nm, operated from a helicopter at 300 m above the water (Fig. 7.13a). The light returning from the water was collected by a telescope with a 30 cm diameter Fresnel lens; a beam splitter, interference filters and separate photomultipliers (Fig. 7.13b) being used to separately detect the chlorophyll fluorescence and Raman emission. Hoge and Swift (1981) used a neodymium-doped yttrium aluminium garnet (Nd:YAG) laser emitting at 532 nm, and exciting a water Raman emission at 645 nm: the equipment was flown in a P-3A aircraft at 150 m above the water surface. With a frequency-tripled Nd:YAG laser emitting at 355 nm there is a Raman emission peak at 402 nm, which can be used together with the CDOM fluorescence at 450 nm to estimate absorption due to yellow substances (see later). Hoge et al. (2005) report
that combining these data with the chlorophyll 683 nm fluorescence and 645 nm Raman data (from 532 nm laser excitation) leads to improved estimates of phytoplankton chlorophyll content. The 355 nm laser on its own also excites chlorophyll fluorescence at ^680 nm, and the fluorescence to Raman (404 nm) ratio (which they find to be proportional to chlorophyll) has been used by Barbini et al. (2004) with a shipborne instrument to continuously measure surface chlorophyll content in transects from Italy to New Zealand and back again.
In waters in which the phytoplankton population includes a significant proportion of cyanophyte and/or cryptophyte algae, the laser-induced fluorescence spectrum includes a substantial peak at ^580 nm due to emission from the biliprotein photosynthetic pigment, phycoery-thrin,372,586 and thus can provide some information about the types of algae present.
Was this article helpful?