V Drozdowska - Seasonal and spatial variability of surface seawater fluorescence - страница 1
Seasonal and spatial variability of surface seawater fluorescence
OCEANOLOGIA, 49 (1), 2007.
© 2007, by Institute of Oceanology PAS.
properties in the Baltic and Nordic Seas: results of lidar experiments
Fluorescent components of seawater Baltic Sea Lidar Fluorescence parameter
Institute of Oceanology, Polish Academy of Sciences,
Powstancow Warszawy 55, PL-81-712 Sopot, Poland; e-mail: firstname.lastname@example.org
Received 8 November 2006, revised 8 January 2007, accepted 18 January 2007.
The paper analyses experimental measurements of laser-induced fluorescence (LIF) spectra in different seawaters. The fluorescence parameters, calculated from LIF spectra as the ratio of the integrals of fluorescence and Raman signal intensities, provide information about the relative changes in the concentrations of fluorescing molecules. Gathered during several cruises in 1994-2004 in the Baltic and Nordic Seas, all the data are presented as scatter plots of the fluorescence parameters of chlorophyll a (Chl a) and coloured dissolved organic matter (CDOM). Satisfactory correlations between these two parameters were found a) for open Nordic Seas waters, b) for the southern Baltic in blooming periods only, and c) for the Gulf of Gdansk in non-blooming periods only.
The structure of the water masses specific to the Baltic Sea is due to its resembling a quasi-enclosed estuary supplied with huge amounts of fresh water from river runoff and sporadic deep inflows of saline Atlantic
water through the Danish Straits. It has therefore been deemed necessary to investigate and monitor the eutrophication of the Baltic (Darecki & Stramski 2004, Wozniak et al. 2000). Since 1993, lidar measurements of in situ seawater fluorescence spectra have been carried out on board r/v 'Oceania' during Baltic cruises as well as during two summer Arctic campaigns (AREX 2001 and 2002) in the Norwegian, Iceland and Greenland Seas. Baltic waters, seriously affected as they are by human activities, have been classified as Case 2 waters; the open waters of the Nordic Seas are Case 1 waters. The difference is due to the quantity and quality of organic matter contained in the upper seawater layer. The present paper analyses the results of the above-mentioned lidar investigations, which were obtained within the framework of a Ph.D. thesis (Drozdowska 2005, unpublished).
Our lidar investigations of seawater have made it possible to obtain continuous fluorescence spectra of seawater in the visible light region, in real-time and without disturbing the aquatic medium. The method can thus be applied to detect the main fluorescent constituents of seawater: phytoplankton pigments, humic-type dissolved organic matter, and oils (Piskozub et al. 1998, Drozdowska & Kowalczuk 1999, Drozdowska et al. 2002, 2004, Drozdowska & Darecki 2005, Drozdowska & Krol 2005, 2006). These spectra provide information on the concentration of chlorophyll a (Chl a) and coloured dissolved organic matter (CDOM). The concentration of Chl a, for which the fluorescence parameter obtained from the lidar-induced fluorescence spectra is a proxy, is an indicator of phytoplankton abundance (Babichenko et al. 1993, Determann et al. 1994, Fadeev 1999, Barbini et al. 2001). The amount of CDOM is also determined from the CDOM fluorescence parameter calculated from the seawater fluorescence spectrum. The position of the CDOM fluorescence spectral band shifts towards the blue or red wavelengths, depending on the dominant fractions of humic substances (HS). High-molecular-weight HS molecules produce a red shift in the fluorescence and absorption spectra and typically, a lower quantum yield of fluorescence; low-molecular-weight HS molecules give rise to a blue shift in these same spectra and a high quantum yield of fluorescence
Analyses of seawater fluorescence spectra obtained by the lidar method yield the fluorescence parameters of Chl a and CDOM, which allow seawater masses to be distinguished according to their individual biophysical and fluorescence properties (Barbini et al. 1998). The classification of seawaters with the aid of lidar-induced fluorescence spectra parameters is based on the correlation coefficient (r2) between the fluorescence parameters of Chl a and CDOM. For Case 1 waters these values are large (close to 1), for Case 2 waters, they are low (close to zero). Moreover, according to
Salyuk and his co-workers (Pavlov et al. 2000, Salyuk et al. 2003), the positive linear regression coefficients (when r2 is close to 1), a and b, respectively supply information about the rate of CDOM formation from phytoplankton communities and the initial content of organic matter in the aquatic environment. Hence, if the correlations between the Chl a and CDOM fluorescence parameters obtained in a given waters are linear, this means that those waters have similar bio-optical properties.
The most important advantages of applying the lidar in marine campaigns is that the results of lidar measurements are obtained in real time without any disturbance to the aquatic environment (Babichenko et al.
1993, Determann et al. 1994, Patsayeva 1995, Barbini et al. 2001). The
lidar light penetrates the seawater, where part of it is absorbed, emitted as fluorescence quanta or transformed into some other kind of energy. The emitted light disperses equally in all directions, but only the fraction reaching the telescope's field of view is recorded by the lidar. So the important parameter of the geometry set-up is the ratio of the solid angle from which the light is collected by the telescope to the full solid angle (47г). It is the ratio of the telescope area to the surface area of the sphere into which the light is dispersed. The radius of this sphere r is equal to the distance between the target that emits the return signal and the telescope. The number of photons reaching the telescope decreases with the square of r and is proportional to the surface area of the telescope.
Intended to create a database of in situ fluorescence spectra of seawater, the lidar experiments were performed on board r/v 'Oceania' with the FLS-12 (LDI, Estonia) lidar system (Babichenko et al. 1989). This consists of an excimer laser (308 nm) used as the pumping source to a tuneable dye-laser, the lidar light source, and the receiving block, which includes the telescope, polychromator and electronic block. The tuneable range of emission is 320-670 nm. Time-gated fluorescence spectra of seawater are recorded in the 400-850 nm range.
The time-gated detection of the return signal permits control of both the optimal distance to the sensing layer (time-gate delay of the receiver) and the thickness of the sensing layer (time-gate duration of the receiver). The return signal is a continuous spectrum and can be divided into separate spectral bands due to Rayleigh scattering (elastic scattering of the laser emission at the water surface and in the water column), Raman scattering (inelastic scattering of the laser emission at water molecules, shifted 3420 cm-1 from the excitation wavelength) and CDOM and Chl a fluorescence (Fig. 2). The total intensity of the recorded fluorescence
Fig. 1. Lidar set-up
______ ll... toot ____________
1 о к u s-
scatte 2 ~700 c
23000 22000 21000 20000 19000 18000 17000 16000 15000 14000 13000 wavenumber k [cm-1]
Fig. 2. Typical lidar-induced seawater fluorescence signal
signal comes from the upper layer. The Raman scattering signal is used to normalise the fluorescence signal in order to obtain the fluorescence parameter describing the relative concentration of the fluorescing molecules (Klyszko & Fadeev 1978, Babichenko 2001, Drozdowska et al. 2002).
3. Results and discussion
Extensive data were gathered on the occurrence and spatial distribution of natural organic matter - phytoplankton and humic substances - in the upper seawater layer.
In the Baltic Sea, the spatial distributions of Chl a during blooms are quite characteristic, which makes it necessary to use different axis scales during blooming as opposed to non-blooming periods (Fig. 3). During blooming periods, Chl a spatial distribution maps resemble complex, patchy, mosaic-like patterns of small areas with different fluorescence properties. In the other periods, Baltic waters are well mixed and characterised by huge areas with similar, homogeneous fluorescence properties.
Similar spatial distribution patterns of the Chl a and CDOM fluorescent parameters were observed in the Nordic (Norwegian, Iceland and Greenland)
55.5o 55.0o 54.5o
fluorescence parameter of Chl a 56°-
14o 15o 16o 17o 18o 19o 20o 21o ЇЇ
8 6 4
18.0o 18.5o 19.0o 19.5o
55 5J April 2003
55 0o 54 5o
55 0o 54 5o
I 55.0o- \
18.0o 18.5o 19.0o 19.5o
56o 55o 54o
18.0o 18.5o 19.0o 19.5o longitude E
14o 15o 16o 17o 18o 19o 20o 21o longitude E
Fig. 3. Spatial distribution of the Chl a fluorescence parameter measured during several Baltic cruises during blooming and non-blooming periods
fluorescence parameter of Chl a [rel. units]
e d tu
78o 77o 76o 75o 74o 73o 72o 71o 70o
04o06o08o 10o 12o 14o 16o 18o 19o longitude E
81o 80o 79o 78o 77o 76o 75o 74o 73o
04o06o08o10o 12o14o 16o18o19o longitude E
Fig. 4. Spatial distribution of the Chl a fluorescence parameter measured during AREX 2001 and 2002
Seas in 2001 and 2002 (Fig. 4). The largest amount of fluorescent organic matter was observed along the shelf of Spitsbergen and along the Scandinavian coast. Additionally, quite significant amounts of fluorescent organics were recorded in central parts of the seas, except for the area around Bear Island, where the fluorescence parameters were low.
Fig. 5 illustrates the relationships between the fluorescence parameters of Chl a and CDOM obtained during marine experiments carried out in different regions of the Baltic Sea at different seasons. To ensure clarity of the plots, the names of the stations have been removed from the figures. The values presented in Fig. 5a (each point is the average of > 50 measurements) were obtained during cruises during blooming (spring) and non-blooming (the other seasons) periods; the stations are marked as circles or triangles, respectively. In general, during algal blooms the fluorescence parameters of Chl a and CDOM obtained in the southern Baltic are subject to considerable variation and represent spatially diverse values. During the non-blooming periods, their values are smaller, and the changes in CDOM fluorescence parameters are much less than those affecting the corresponding parameters of Chl a. Therefore, in the blooming periods a positive correlation (r2 = 0.6)for theChl a and CDOM fluorescence parameters was obtained for all
І 10 ■
ter - autumn (
□ и □ □
y = 1.40+ 1.93x
r2 = 0.67
n = 30
0 12 3