YELLOW SUBSTANCE IN THE NORTH SEA
SPATIAL AND TEMPORAL DISTRIBUTION USING ABSORPTION SPECTROSCOPY
By:
Sandor van Laar Supervisors:
Roddy Warnoek Winfried Gieskes
RU Groningen Groningen, August 1995
"Having problems getting your method right?"
"Yes."
'Well, practical problems, that is science!"
"I think it is quite a frustrating part of science."
A little part of a conversation I had a few months after I started this research. Although I really don't like that 'frustrating part' of research I learned that finally you will get your results, if you keep on trying and work systematically. And then, at the end, if the results you have obtained are good, or maybe, very good, the satisfaction is enormous
This is shortly the history of the research I did from September 1994 until July 1995. Remote sensing, I knew a little about that, but I never heard of yellow substance and I never thought I would use absorption spectroscopy in practical work. The result is that I learned a lot, but more important, I had a lot of fun. I think that is the most important thing in life.
Therefore, I would like to thank Roddy, my supervisor, for having the faith that I would bring this research to a good end, even when I zombied around the Biological Centre after another heavy hockey-weekend. I think the same accounts for Winfried, you looked a little worried sometimes, thanks for your support. And there is always someone you spend your coffee- breaks with, talking about small and big things and computers off course, Patrick thanks for all the company and help. In this report a beautiful figure of the concentrations of yellow substance in the Dutch coastal zone is included. This figure is the result of the work of Koen
Schrader, working at the North Sea Directorate of Rijkswaterstaat, who was willing and able to make some time free for me with this figure as result, Koen, a lot of thanks to you.
Finally I would like to specially thank my mother and father, who knew exactly when to call, the last period was hard but your love and support helped a lot.
SUMMARY
Optical remote sensing is used to study parameters related to water colour. The colour data obtained this way can be converted to spatial and temporal distribution patterns of parameters related to water colour. The result of this conversion are quantitative images which can play an important role in calibration and validation of hydrodynamic and ecological models.
Yellow substance is one of the important constituents in the water column influencing the water colour. It absorbs light increasingly with decreasing wavelength; in this way it interferes with remotely sensed data of chlorophyll a and suspended matter in the visible domain.
Furthermore, yellow substance affects the ecology of water bodies and affects the penetration of UV-light in natural waters.
The collection of dark coloured organic compounds of soil and sedimentary accumulations is called yellow substance. The main components are humic and fulvic acids. Most yellow substance in natural waters is of terrestrial origin, but in open oceans yellow substance may also be derived from decomposition products of marine organisms. The residence time of yellow substance in oceans is estimated at 15 0-900 years; it thus forms a semi-conservative property in the marine environment.
In this research absorption spectroscopy was used to measure the spatial and temporal variation in the concentration of yellow substance in the southern North Sea in samples taken in 1993 and 1994. Spectra obtained in this way are exponential and can be described by two parameters, the absorption coefficient at a reference wavelength, normally 380 nm, -andthe slope of the exponential. The absorption coefficients decreased from on-shore to off-shore stations. The highest coefficients for the absorption by yellow substance were found at the Texel transect, high coefficients were also found at the on-shore stations of the Noordwijk and Walcheren transects and at the Marsdiep station. In the Channel the lowest coefficients were measured, at many off-shore stations low coefficients were also obtained.
The output of rivers was clearly the main source of yellow substance in the North Sea.
Especially at the on-shore stations this influence was very large, but other factors such as origin of water masses may play a role too. The content of yellow substance from the several
such as forests, agricultural land and cities, of these rivers. Yellow substance concentration and salinity had a strong negative correlation. Chlorophyll a and suspended matter, however, did not show any correlation with yellow substance. Although degradation of phytoplankton results in yellow substance, the yellow substance component of the water due to degradation of marine organisms is very small relative to the riverine input.
The values for the slope, calculated for the whole absorption spectrum of yellow substance, were higher than values. This is due to the tendency of the spectra to increase at lower wavelengths. A bump at 275 urn, which is thought to be the result of absorption by dissolved DNA, is another reason for the increase of the slope. If the slope is calculated for segments (360-440 urn and 440-540 urn) of the spectrum, measured values were in the same range as the values in previous research.
CONTENTS
page
PREFACE
SUMMARY ii
1. INTRODUCTION 4
2. YELLOW SUBSTANCE, A REVIEW 5
2.1 Remote Sensing 5
2.2 Yellow substance and its place in the marine environment 6 2.3 Yellow substance, chemical background and optical properties 7
2,3,1 Chemical background 7
2,3.2 Optical properties 9
2.4 Analytical methods 11
2.4.1 Absorption measurements 11
2.4.2 Fluorescence measurements 12
2.4.3 Absorption vs. fluorescence measurements 13
2.5 Yellow substance in the Oceans, the North Sea, the Skagerak, 14 the Kattegat and in lakes
2,5.1 The Oceans 14
2.5.2 The North Sea and adjacent estuaries 15
2.5.3 The Skagerrak and the Kattegat 15
2.5.4 Lakes 17
2.6 Relations between yellow substance and salinity, chlorophyll a and 17 total suspended matter
2.6.1 Yellow substance and salinity 17
2.6.2 Yellow substance and chlorophyll a 18
2.6.3 Yellow substance and total suspended matter 21
3. MATERIALS AND METHODS 22
3.1 Cruises and transects 22
3.2 Investigation area 23
3.3 Samples 26
3.4 Method and measurements 26
3.5 Precautions and problems measuring yellow substance 27
3.6 Data-analysis 30
3.6.1 Statistical analysis 30
3.6,2 Exponential curve fit 30
4. RESULTS 32
4.1 Spatial variation 33
4,1.1 Spatial variation within transects 33
4.1.2 Spatial variation between transects 37
4.2 Temporal variation 39
4.2.1 Temporal variation at transects 39
4.2,2 Temporal variation during cruises 42
4.3 Exponential curve fit 44
4.3.1 Slope 45
4.3.2 Segment analysis 46
4.4 Yellow substance and other water parameters 46
4.4.1 Yellow substance and salinity 46
4.4.2 Yellow substance absorption and chlorophyll a 52
4.4.3 Yellow substance absorption and total suspended matter 54
5. DISCUSSION
5.1 Spatial variation 55
5.2 Temporal variation 57
5.3 Exponential curve fit 58
5.4 Yellow substance and other water parameters 59
Contents
5.4,1 Yellow substance and salinity 59
5.4.2 Yellowsubstance and chlorophyll a 61
5.4.3 Yellow substance and total suspended matter 61
6. CONCLUSION 62
6.1 Spatial variation 62
6.2 Temporal variation 62
6.3 Exponential curve lit 62
6.4 Yellow substancea and other water parameters 62
REFERENCES 64
APPENDIX A 69
APPENDIX B 76
The University of Groningen (RUG), the Netherlands Institute for Sea Research (NIOZ) and the North Sea Directorate (RWS) were participants in the PMNS (Particulate Matter North
Sea) project. The aim of this project is the development of semi-empirical algorithms for the next generation of remote sensing measurements of ocean colour, measurements which will be obtained in the near future with the Sea-Viewing Wide-Field-of-View Sensor (SeaWiFS) (Aiken et a!., 1992). A major goal of the PMNS project is to develop algorithms that will allow discrimination between the principal constituents of the water column like algal pigments, suspended sediment and dissolved material (yellow substance). Using these algorithms it will be possible to quantify phytoplankton, sediment/detritus and yellow substance from multispectral measurements of ocean colour (Wamock eta!., 1995).
The component of the PMNS project covered by the University of Groningen is discrimination between the spectral signatures of these three principal constituents. This involves the measurements of the absorption characteristics of phytoplankton and natural particulates and relating these to the concentration of phytoplankton pigments and yellow substance (Wamock eta!., 1995). The research described in this report is a part of this component of the PMNS project: the yellow substance distribution in the Dutch coastal zone of the North Sea, studied
by absorption spectroscopy. The aim of this study was to get a good picture of the seasonal and spatial distribution of yellow substance. Yellow substance is one of the major constituents in the marine water affecting its colour, The relationship of yellow substance with other important water colour constituents was also analysed. The influence of another important water characteristic, salinity, on yellow substance has also been the subject of my project.
Characteristics of yellow substance (both chemical and optical),
its place in the marine
environment, and results of previous research is reviewed in chapter 2. In chapter 3 the materials and methods used to measure yellow substance absorption are described. After that the results are discussed in chapter 4. The results of the study of the relationship between yellow substance and other important water variables can also be found in this chapter. In the discussion and conclusion, chapter 5, all the results obtained are compared with previous research. The conclusion is used to summarise all the results. Furthermore, two appendices are included which contain summary tables of all results.2. YELLOW SUBSTANCE, A REVIEW
2.1 Remote Sensing
Suspended and dissolved substances in surface waters influence its colour. Clear water is blue, water rich in aquatic humus is yellow to brown and the colour of turbid water, carrying suspended matter, depends on the mixture of the constituents. The water colour can vary from dark blue-green via bright green and brown to red. The colour is determined by the light scattered out of the water and that reflected at the water surface. Light that originates from below the water surface shows characteristic influences of the diverse components in the water. Absorption diminishes light intensities and scattering changes its angular distribution.
The intensity of reflected light increases with the amoimt of scattering and decreases by absorption (Krijgsman, 1994).
Variables related to water colour are used to study environmental processes such as the primary production of algal biomass and the distribution of suspended matter (Krijgsman,
1994). These components in the water interact with the natural light field from the incident light intensity and subsurface upwelling irradiance (Dekker, 1993). Optical remote sensing, both airborne remote sensing and remote sensing by satellites, is used in such studies. This is a
technique to collect colour data by detection of upward radiance at a distance. After
interpretation the colour data can be converted to spatial and temporal distribution patterns of parameters of water quality. Quantitative images obtained in this way can play a role in calibration and validation of two- and three dimensional hydrodynamic and ecological models(Dekker, 1993).
To date, quantitative observations have been made by satellites in a small number of broad wavelength bands. Instruments with higher spectral resolution are flown in aircraft. Line detectors have been available for some decades, imaging instruments for a few years (Dekker,
1993). It is to be expected that these instruments will also be placed in satellites in the near future. These high spectral resolution sensors will require new calibration procedures using
sea-truth data and new algorithms to interpret the spectral colour data (Krijgsman, 1994).
2.2 Yellow substance; its significance in the marine environment
Underwater light is increasingly absorbed with decreasing wavelength. Dissolved and colloidal organic compounds (yellow substance) present in natural waters are partially responsible for
this absorption. There are several important reasons to study the origin and chemical
composition of yellow substance.
Firstly, remote sensing of particulate constituents,
e.g. the phytoplankton or the total
suspended sediment concentration is very difficult when yellow substance concentrations vary simultaneously and also affect the remote signal. It thus represents one of the principal components of seawater that compromise the accuracy of remote sensing algorithms that seek to quantifr the concentration of algal pigments (Tassan, 1988; Davies-Colley and Vant (1987); Coble and Brophy, 1994).Secondly, the absorption of light by yellow substance affects the ecology of water bodies. The ability of aquatic higher plants and many algae to photosynthesise and grow is markedly
affected by the availability of blue light, which in turn is highly dependent on the
concentration of yellow substance in the water (Kirk, 1976; Davies-Colley and Vant, 1987).Thirdly, the study of yellow substance is important to increase the knowledge of dissolved organic matter (DOM). The fourth reason is to examine the advantages and limitations of using DOM as a natural tracer of water movements in the sea (Karabashev et al., 1993).
The fifth important feature of yellow substance absorption is that because its absorption spectrum shows such a strong increase in absorption at shorter wavelengths, yellow substance will greatly affect the penetration of UV in natural waters. This high absorption prevents harmful UV-B from penetrating coastal waters to any significant extent (Kramer, 1987).
The humic substances in the sea water play an important role in the biochemistry of the marine environment (Dujmov et al., 1992), this provides the sixth reason to study yellow substance. These substances are important in processes such as metal complexation, organic pollutant (PAH and pesticides) interactions, lipid solubility, biochemical fates of bio-elements (C, N, P, Ca, Si), transportation and distribution of organic matter in seawater
Yellow substance, a review
(Malcolm, 1985; McCarthy, 1989; Shaw, 1994).
Binding of organic
or inorganic contaminants to humic substances can alter the availability of contaminants for uptake by biota (Dujmov et al., 1992).The last reason is that because of the greater absorption of short compared to long
wavelengths, yellow substance shifts the energy of maximum transmission of the watercolumn to longer wavelengths (ultimately, orange or red). It so changes the colour of the water and its optical properties (Davies-Colley and Vant, 1987).
2.3 Yellow substance; chemical background and optical properties
2,3.1 Chemical background
Yellow substance is the name most used for the largest single category of components of dissolved organic matter (DOM) in natural waters, aquatic humus (Gjessing, 1976). It comprises 40-60% of the total DOM and accounts for 85-100% of the colour ofDOM (Thurman, 1983). Other names have been proposed for aquatic humus; Kirk (1976,1983)
proposed the term "gilvin" (Latin glivus: pale yellow), Kalle (1938) used
the name"Gelbstoff" and Coble and Brophy (1994) used coloured dissolved organic matter(CDOM).
Yellow substance is the most familiar term to workers. Therefore I will use this word in this study.
The term yellow substance does not refer to a specific chemical
compound of fixed composition but rather to the collection of dark coloured organic compounds of soil and other sedimentary accumulations (Nyquist, 1979). The different components of these organic compounds can be divided into three main groups (Christman & Oglesky, 1971; Schnitzer &Khan, 1972; Kalle, 1961). The first group is "humin", which is insoluble in dilute base and acid. The second group is "phenoihumic (humic) acid", which is light to dark brown and soluble in dilute alkaline solutions but is precipitated by acidification of an alkaline extract.
The third group is "carbohydrate humic (fulvic) acid or melanoide", which is light to golden yellow, more stable than humic acid and soluble in both base and acid.
Most humic material may be formed by oxidation and polymerisation directly from the
true that some saprophytic fimgi excrete large amounts of phenolic substances when grown on carbohydrate. These phenolic substances can undergo oxidation and polymerisation to give hurnic-like material (Jackson, 1975; Schnitzer, 1978). Via the air and the rivers the humic material, now called yellow substance, enters the marine environment (Karabashev et al.,
1993).
Marine humic substances found away from continental margins have a distinctly different chemical structure and origin than the yellow substance in coastal regions (Kalle, 1966).
Terrestrial humic substances are more aromatic and have higher C:N and H:C ratios than do marine humic substances. The marine humic material in open oceans is thought to be derived from decomposition products of marine organisms, either directly as unutilized macromolecules (Yentsch and Reichert, 1962; Duursma, 1974; Jackson, 1975), or indirectly by condensation of smaller, non-fluorescent compounds (Kalle, 1966), Most .humic. and fulvic substances have moderate to low molecular weights (Aiken et a!., 1985; Carder et al,, 1989).
Fulvic acids derived from algae have low C :N ratios and low aromaticity. Algal-derived fulvic acids are yellow in colour but absorb light to a lesser extent than fulvic acids from terrestrial sources (McKnight et a!., 1994).
Moran and Hodson (1994) found that in the south-eastern U.S., at the Atlantic coast, huniic substances are released in the bulk dissolved organic carbon (DOC)-pool by bacterial decomposition of vascular plant detritus in coastal wetlands. The primary source of dissolved
humic substances was lignin. The lignin-rich humic substances were used by marine
bacterioplankton, although more slowly and less efficiently than other components of theDOC-pool. Nevertheless, a significant fraction (24%) of these substances was mineralised within 7 weeks. Calculations made by Moran and Hodson (1994) indicate that about half of the dissolved humic substances in this coastal zone was contributed by the coastal salt marshes and half by river export. Højerslev (1988) estimated the residence time of yellow substance in the oceans at 150-900 years. He stated that it consists of a very stable mixture of humic and fulvic acids. This means that yellow substance is a semi-conservative property in the whole marine environment and can be applied to water mass classification (Højerslev, 1988). Most researchers believe that the main part of yellow substance both in coastal and oceanic waters has a long residence time. However, Kieber et a!. (1989) found a potentially important removal process in the ocean for the main fraction of the DOC, the old, biologically refractory material, such as huniic substances. This removal process is the photochemical
Yellow substance, a review
breakdown of high-molecular-weight marine DOC resulting in the production of a compound, pyruvate, that can be used by heterotrophic-bacterial plankton as a substrate. De Haan (1993) found photodegradation of aquatic humic substances by underwater UV-light near the surface to depths where UV-light with wavelengths <320 nm penetrated. The dailyphotodegradation
rate of aquatic humic substances appeared to be of the same order of magnitude as that of the daily pelagic photosynthetic fixation of dissolved inorganic C in typically oligotrophic humic lakes. However, the measured UV degradation of aquatic humic substances on a yearly basis
was small compared to the approximate annual primary
production of temperate lake phytoplankton (8%) and submerged macrophytes (3%) (de Haan, 1993). Other authors (Karabashev, 1987; Laane and Kramer, 1990; Thomas and Lara, 1995) did not find signs of photodegredation of DOC.2.3.2 Optical properties
Hunuc substances in the open oceans and in coastal areas emit blue fluorescence when excited
by near UV light.
Absorption coefficients increase exponentially from visible to UV wavelengths (figure 2.1), which causes the yellowish colour. In coastal areas the absorptioncoefficients of yellow substance decrease in offshore direction (Coble and Brophy, 1994).
Figure 2.2 shows the typical absorption curve of yellow substance, a straight line on a semilog
plot (logarithm of the absorption
coefficient, a, vs. wavelength) (Jerlov, 1968). Theexponential form of this representation was given by Morel and Prieur (1976):
aQ) =
a(X)
exp (S(?vref)) (2.1)The spectral slope factor (S) of yellow substance absorption was initially shown to be independent of water mass type between 375-500 mu. Similar values were found, with a mean of S =-0.014(Bricaud eta!., 1981). Other authors found almost the same mean S-values, S = -0.015 (Jerlov, 1968) and S = -0,013-0.016 for lake waters (Kirk, 1976). However, it has recently been shown by Carder et cii. (1989)that S is both wavelength and water mass type dependent when calculated over two shorter intervals, 370-440 nm and 440-565 mu. Spectral slope increases with decreasing wavelength of measurement and with increasing fulvic acid to humic acid ratio (Carder ci' cii., 1989). Results from the research of Coble and Brophy (1994) support both the results of Carder et a!. (1989) and the results of Jerlov (1968), Kirk (1976) and Bricaud et a!. (1981).
Figure 2.1: Yellow substance spectra (Dekker, 1993).
Figure 2.2: Log-transformed yellow substance spectra from figure 2.1 (Dekker. 1993).
800
A common way to express the yellow substance concentration is to use the absorption coefficients (m') at 375 nm (Bricaud ci cii., 1981) or at 380 nm (Højerslev,1994). Bricaud et
cii. (1981) found values for oceanic waters, or waters which were not influenced by fresh
water, to range from 0.06 to 0.3 m1. In these kind of waters the yellow substance
600 650
waveleu,'tl: nm)
550 600 650
wavelength (nm)
Yellow substance, a review
concentration is presumably related to in situ biological activity. In coastal waters, affected by fluvial discharges, the absorption coefficients at 375 m' lie between 0.36-4.2 m1. The highest concentrations were measured at locations with industrial or domestic effluents (Bricaud et al., 1981). These kind of values for the coefficients are typical for the range of published values from other areas (Coble and Brophy, 1994; Carder et cii., 1989; Ferrari and Tassan,
1992).
Shapiro (1957) and Chanu (1959) observed a peak at 275 urn in the absorption spectrum of extracted organic matter which was attributed to purine and pyrimidine rings by Yentsch and Reichert (1962). Bricaud et al. (1981) observed a weak marked shoulder at 265 nm with a spectral resolution of 10 nm. Lawrence (1980) and McLaughlan (1981) contributed the shoulder at 270 nm to degradation products of tannins and lignins.
Using three-dimensional excitation emission matrix (EEM) spectroscopy, DOM from at least three different sources, soils, rivers and surface seawater, can be distinguished(Coble et cii., 1990). In the waters of the Black Sea at least three fluorophores are present. The largestpeak, with excitation/emission wavelengths of 295 and 345 nm respectively, has fluorescence properties which are similar to those of the indole ring of tryptophan (Coble et cii., 1990).
2.4 Analytical methods
2,4.1 Absorption measurements
Yellow substance absorbs a radiant flux exponentially with wavelength according to a simple exponential law valid for nearly all natural waters. This implies that three different rapid optical measurements for determining yellow substance in tenns of absorbance(concentration) can be listed (Hjerslev, 1988).
1. Hydrocasts, filtration and spectrophotometric measurements on filtered water samples. For measurements in oligotrophic waters the sensitivity of the spectrophotometer ought to be of the order of 0.0005 unit of optical density or less. This method is highly sensitive to contamination of both the sample and the water in the reference cell. The method gives point values and is rather slow in comparison with other methods (Højerslev, 1988).
In this method the samples are filtered on Whatman GF/C® or GF/F® glass-fiber filters
previously rinsed with about 100 ml of distilled or seawater. Without this rinsing
precaution, a significant absorption ( 0,1 m1 at 375 nm) results presumably by discharge of organic matter from the filter, No organic material in the filters, flasks and stoppers should be used. The samples are analysed with a double-beam spectrophotometer in the spectral range 200-700 rim with a path length of 10 cm. A standard water is used in the reference cell in order to subtract absorption by pure, distilled, water (Bricaud et al.,1981).
2. In situ light transmission measurements at two wavelengths - one in the UV part of the spectrum and the second one in the red part of the spectrum. For more details on this method see Joseph (1949), Fløjerslev (1986) and Højerslev (1988).
3. In situ measurements of the depth of the euphotic zone, in addition to the downward and the upward UV-B daylight (310 nm) irradiance Ed and E respectively, combined with an optical model derived by FInjerslev (1982) For more details on this method see Højerslev (1982, 1988).
2.4.2 Fluorescence measurements
Yellow substance is fluorescent, with emission peaks at 420-45 0 nm from excitation at 220- 250 and 320-350 nrn (Coble et al,, 1990; Mopper and Schultz, 1993), The detailed spectral excitation and emission information obtained using high resolution fluorescence spectroscopy provides a more complete view of optical properties and chemical composition of aquatic humic substances than measurements of a single excitation and emission spectrum (Coble et al., 1990; Coble and Brophy, 1994).
Emission spectra are measured with an spectrofluorometer between 250 and 710 nm at forty separate excitation wavelengths between 260 and 455 nm, generating an excitation-emission matrix (EEM). Excitation and emission maxima can be independently detennined from these EEMs. The biggest advantage of the technique is that it is sensitive enough so that no preconcentration or extraction of coloured DOM from seawater is necessary, thus eliminating the potential for chemical alteration or selective retention of certain components which hamper other techniques used for coloured DOM measurement and characterisation (Coble and Brophy, 1994).
Yellow substance, a review
Relative amounts of dissolved fluorescent matter can be estimated by comparison with calibration curves of quinine in sulphuric acid and/or Suwannee River fulvic acid in standard seawater as a standard solution and by scanning the emission spectra at constant excitation wavelengths (Dujmov et cii., 1992). The standardisation of natural fluorescence is usually done by standardising the instrument with the known concentrations of the mentioned standard solutions. Dissolved fluorescent matter is expressed in millifluorescence units after checking the instrument with the named solutions (Dujmov et cii., 1992 and references herein).
2.4.3 Absorption vs. fluorescence measurements
The measurement of yellow substance concentration by absorption spectroscopy and by fluorescence spectroscopy have both important advantages and disadvantages.
Direct measurement of absorption by absorption spectroscopy is affected by important error sources, e.g. scattering by residual particulate material (Bricaud et cii., 1981), which require cumbersome correction procedures (Ferrari and Tassan, 1991). Fluorescence is orders of magnitude more sensitive than absorption since the wavelength of the emitted (detected) light differs from that of the excitation beam. Furthermore it can be used to distinguish between yellow substance from terrestrial and marine sources without preconcentration of water samples (Coble and Brophy, 1994). Other advantages of fluorescence are: it can be performed rapidly in situ and it does not need elaborate filtration (Ferrari and Tassan, 1991). On the other hand, the conversion of induced fluorescence into equivalent absorption may be a source of important errors because the fluorescent properties of yellow substance exhibit significant variability due to composition changes that depend on its origin (Ferrari and Tassan, 1991 and
references therein). However, a high degree of correlation has been found
between fluorescence and absorption coefficient, but errors in retrieval of yellow substance absorption coefficients can be as high as a factor of 2. Since estimates of fluorescence efficiencies of yellow substance vary up to 2.5-fold for a wide range of water types (Donard et cii., 1989), part of the error in estimation of absorption from fluorescence is due to variability in chemical composition rather than to measurement errors (Coble and Brophy, 1994)./
\
/
•
:
' LQ
•
(
—
••—--- ,•
Sc
Figure23: Distribution pattern ol yellow substance concentrations (measured in dissolved fluorescence, mFl) in the North Sea. sampling stations indicated by black dots (Laane and Kramer. 1990).
2.5 Yellow substance in the Oceans, the North Sea, the Skagerak, the Kattegat and in lakes
2.5.1 The Oceans
The highest concentrations of yellow substance occur in regions influenced by land drainage, such as the Baltic Sea, the North Sea and certain coastal waters (Kalle, 1966; Højerslev,
1988; Laane and Kramer, 1990). For oceanic waters unaffected by significant land drainage, there are few direct measurements of the optical influence of yellow substance at visible
',J.
'C
Yellow substance, a review
wavelengths (Carder et cii., 1989). Tn regions of low productivity Bricaud et cii. (1981) have extrapolated measurements from the more absorptive ultraviolet wavelengths into the visible spectrum. Jerlov (1968) measured values for clear oceanic waters at short wavelengths.
Indirect calculations have been made by Morel and Prieur (1976), Smith and Baker (1981) and Carder et cii. (1986).
2,5.2 North Sea and adjacent estuaries
In the North Sea, its major estuaries and the German Bight the concentration of yellow substance is relatively high compared with the concentration of yellow substance in the Atlantic water entering the North Sea (Laane and Kramer, 1990), Using the values found for yellow substance in the North Sea, three different water masses can be distinguished: the Channel and Atlantic water, both with low values, and the coastal zone with higher values.In the coastal zone the values increased towards the coast, see figure 2.3 from Laane and Kramer
(1990).
Figure 2.4 shows the relation between yellow substance and salinity in the Ems, Elbe, Scheldt, Weser and the Rhine. The highest values of yellow substance fluorescence were measured in the water of the river Ems (Eisma et cii., 1982). This river runs through an old peat area (Laane, 1981). Tn the Scheldt estuary the values of yellow substance were between 20-110 mFL (Kramer, 1985). The values in the Rhine estuary (8-49 mFL) (Eisma et al., 1982) and the Wadden Sea (9-23 mFL) (Laane and Kramer, 1990) were lower than the values foundin the Scheldt estuary. But the estimated yearly flux of yellow substance (mFL.dm3 yr1) has been reported to be much higher in the Rhine (258) than in the Scheldt (12) (Laane and Kramer, 1990).
2.5.3' The Skagerak and the Kattegat
The Skagerak and the Kattegat represent an area where the waters of the Atlantic Ocean and the Baltic Sea are mixing and rivers with large catchment zones bring DOM-rich waters into the sea (Fonselius, 1990). The Kattegat region forms a transition zone between the Baltic Sea and the North Sea. Water from the Baltic, which is brackish and high in yellow substance, mixes with water from the North Sea, which is saline and dense with either low orhigh yellow
180 'Ems eElbe
Weser
1604 D Rhine
A Schetdt 14C
k
6 12 if 2C 24 26 32 S
Figure2.4: The relation between yellow substance (measured in mFl) and salinity (S) in the Ems, Elbe, Weser, Rhine and Scheldt estuaries (Laane and Kramer. 1990).
The highest concentrations of yellow substance in late spring were found iii the area from the Oslo ord to the central Skagerak. In autumn the highest concentrations were found in the Kattegat and along the Norwegian coast. The lowest values are very local and associated with the Jutland current (Karabashev et al., 1993). Højerslev (1994) found that a large fraction of the Central North Sea-North Atlantic water intrudes the deeper part of the Norwegian trench to the northern part of the Kattegat. A high fraction of brackish, light Baltic water was found in the surface waters of the Kattegat. The flux of German Bight water into the Kattegatis rather small and has maximal fractions in the intermediate layers (Højerslev, 1994).
Yellow substance, a review
2.5.4 Lakes
Kirk (1976) studied yellow substance concentrations in various inland and coastal waters m Australia Among measurements done in freshwater lakes the concentration of yellow substance was found to vary seven-fold. The absorption of yellow substance ranged from 0.42 m' to 2.90 m at 440 nm. The absorption in coastal waters was much lower than in any of the freshwater lakes (0.0 1-0,08 m' at 440 urn). Because of the high yellow substance in the inland waters 60-80% of the quanta captured was absorbed by yellow substance rather than by water. The inland water with the lowest absorption coefficient were derived from a catchment consisting entirely of forest whereas lakes with higher concentrations of yellow substance were fed from catchrnents that included cleared agricultural land as well as forest (Kirk, 1976).
Dekker (1993) and Krijgsrnan (1994) also studied the absorption of yellow substance in lake waters. In different Dutch lakes the absorption coefficients ranged from 0.78 rn-1 to 3.51 m ( Dekker, 1993) and from 0.91 m'1 to 22.6 rn1 at 440 nrn (Krijgsman, 1994). The differences in the range of the coefficients found in both researches were due to the different catchment areas
of the lakes studied.
The content of yellow substance in the water in a particular water body can vary markedly over a year. It can be concluded that in general coastal waters have substantially lower yellow substance concentrations than inland waters (Kirk, 1976).
2.6 Relations between yellow substance and salinity, chlorophyll a and total
suspended matter -
2.6.1 Yellow substance and salinity
Yellow substance can be used as a semi-conservative property in the marine environmentand
applied to water mass classification and mixing studies
along with traditional t, S(temperature, salinity) analyses. This has certain advantages in shallow coastal regions for which the temperature cannot be considered as being a conservative property (Højerslev,
1988).
Højerslev (1994) used yellow substance combined with salinity to c1assif' the water masses in the Skagerak and the Kattegat region. He based this classification on the characters of the water masses. Baltic water has a low salinity and a high yellow substance content, whereas North Sea waters originate either from the Atlantic having extremely high salinities and low yellow substance content or from the German Bight, which has an intermediate to high salinity and a very high yellow substance content, Hjerslev (1994) found that light absorption by yellow substance in the Baltic Sea was 0,96 m1 with a salinity of 8 %o, in the Central North Sea and Atlantic the light absorption was 0.07 m', at a salinity in this area of 35 %o. The light absorption in the German Bight was 1.50 m', the salinity was 31 %o. Using these values Højerslev (1994) made yellow substance/salinity diagrams for the Skagerak and Kattegat region. Assuming that any Kattegat water mass is formed by mixing of the water types of the German Bight, the Baltic Sea and the Central North Sea-North Atlantic, all measured data pairs should stay within the triangle of the values found for these water masses (figure 2.5).
I-Iøjerslev (1994) also made calculations to obtain mixing ratios for the named water masses and made depth profiles with use of the yellow substance and salinity pairs.
The surface salinity distribution in the North Sea has been described by Colijn et al. (1990).
An inverse relationship between fluorescence and salinity exists (figure 2.4). This indicates that the rivers are the main source of yellow substance in the North Sea (Laane and Kramer,
1990).
2.6.2 Yellow substance and chlorophyll a
One of the most important factors influencing absorption in the blue region of the spectrum is the presence of yellow substance (Kopelevich and Burenkov, 1977). Higher plants and many algae depend highly on this region of the spectrum for photosynthesis (Kirk, 1976). Yellow substance absorbs strongly at 440 nm and the spectral region near the absorption maximum of the Chl a Soret band. Absorption coefficients as low as 0.005
m' are equivalent to the
absorption coefficient due to phytoplankton when Chl a has a concentration of 0.10 mg m3.Thus it is important to understand the role of yellow substance as it affects the remote sensing of ocean colour, even at "negligible" concentrations (Carder et al. 1989).
Yellow substance. a revie\
Figure 2.5: Salinity/yellow substance diagram, three data pairs from different water masses form a triangle in which almost all measured data pairs fall (Hojerslev, 994).
Because of the low values for absorption due to yellow substance in oligotrophic waters the assumption has often been made (Smith and Baker, 1981; Bricaud et al., 1981) that water colour or yellow substance covaries linearly with chlorophyll pigments.
In oceanic areas, or more generally in areas not influenced by
freshwaters, the yellow substance concentration is presumably related to biological activity as a by-product of algal cell degradation (Yentsch and Reichert, 1962). But in certain upwelling areas changes in chlorophyll concentrations of two orders of magnitude, are accompanied by little change in concentrations of yellow substance (Bricaud et a!.,1981). This suggests that huniic
substances have a long half-life in the ocean relative to the algal population that may have produced it.Kopelevich and Burenkov (1977) reported that the correlation between absorption and chlorophyll is explained by the existence of a close relation between chlorophyll and the yellow substance in productive waters. They suggest that yellow substance consists of two parts: "conservative" and "nonconservative". The first is that part of yellow substance whose content is the result of many centuries of mixing of stable organic matter in the ocean. The second part forms as the result of the decomposition of the phytoplankton in the sea water.
Plotting literature values of yellow substance absorption against literature values of primary productivity shows that yellow substance in the open sea can be considered as a by-product of primary productivity (figure 2.6) (Carder et a!., 1989),
It seems also clear that a general relationship exists between yellow substance and primary productivity in the Gulf of Mexico. This may ultimately be refined to the point that residual yellow substance pools may be interpretable as a measure of the primary productivity of a region over the previous 1 or 2 months in open oceans (Carder et a!., 1989).
PERUVIRN
I IJPWELLING
>-
>
J .- 0o
.5W FLA
OUTER OF GUINEA
SUELF --
• _ CLIUOTRO1C
GULF AND CARIBBEAN
LOG LUg (440>]
Figure 2.6: Relationship between annual primary productivity and yellow substance absorption based on literatare values. The rectangles enclose ranges of yellow substance and primary productivity reported for the oligotrophic Gulf of Mexico and Carribean, Mississipi plume, outer Florida shelf, NW African upwellmg (Mauritanian coastal waters), Gulf of Guinea and the Peruvian upwelling (Carder et a!., 1989 and references herein).
Yellow substance, a review
2.6.3 Yellow substance and total suspended matter
As with chlorophyll, yellow substance also interferes with the absorption of suspended sediment (Tassan, 1988). Therefore, it is also important for the estimations made for total suspended sediment derived from ocean colour measurements that the seasonal and spatial changes in yellow substance concentration are known.
The relationship between yellow substance and total suspended sediment or total suspended matter has only been a subject for research a few times. Sørensen and Aas (1994) reported that no correlation was found between yellow substance and the suspended sediment. In the sediments of the Wadden Sea yellow substance can be found, the origin ofthis yellow substance is unclear, It may be produced in situ from non-fluorescent compounds (Laane and Kramer, 1990).
The main sources of suspended matter in the Dutch coastal zone are the Flemish Banks and the Channel. A large variability between different years and seasons was found depending on varying transport through the Strait of Dover. It is possible that strong salinity gradients
influence the sedimentation of suspended matter (Visser et al., 1991).
3.1 Cruises and transects
In the period April 1993 to July 1994 six cruises were carried out in the southern part of the North Sea. These PEGASUS (Pelagic Geographic study of Abundance and composition of
Suspended matter) cruises are part of the PMNS program. The stations where samples were taken were chosen to obtain a maximum of variation in abundance and composition of North Sea particular matter and also yellow substance in the Dutch coastal zone in the course of the seasons. This means that at a total of five transects along the Dutch coastmeasurements were carried out. These transects were Texel (Te), Walcheren (Wa), Noordwijk (No), Channel (Ch) and Frisian Front (Fr), measurements were also done in the Marsdiep (Ma). Figure 3.1 shows these transects. In Table 3.1 the six cruises, the period they were carried out and the transects
measured are listed.
North Sea
I Fr
i,.
I S,1TeMa ad&n
EmsI No ,
I
' Wa
Figure 3.1: Investigation area (Southern Bight) with the transects where samples where taken; Texel (Te), Waicheren (Wa), Noordwijk (No), Channel (Ch), Frisian Front (Fr) and the station at the Marsdiep (Ma).
Materials and methods
Table 3.1: The six cruises and the period in which they were carried out. Per cruise the transectsmeasured, Texel (Te), Waicheren (Wa), Noordwijk (No), Channel (Ch), Frisian Front (Fr) and the station Marsdiep (Ma) are marked (.).
Period Transect
Te Wa No Ch
Fri
MaCruise 1 Cruise 2 Cruise 3 Cruise 4 Cruise 5 Cruise 6
13-16 April 1993 • . .
. —
2 1-25 June 1993 • . . .
29 Sept.-! Oct. 1993 6-9 December 1993 31 Jan,-3 Feb. 1994 25-31 July 1994
.
. .•
.
. .— •___._
• I •—__. • •
The Texel, Noordwijk and Walcheren transects were sampled most thorough: by Texel six times, Noordwijk and Waicheren both five times. The Channel transect was sampled twice, the Frisian Front and the Marsdiep both were sampled only once. Chlorophyll a, total suspended matter and salinity data obtained at the time of sampling are used to study the relationship between yellow substance and these parameters. Two cultures, older than a year, were also analysed. One culture contained Rhodomonas sauna, the other Tetraselmis sueccia.
3.2 Area of investigation
The water in the North Sea circulates according to a fixed pattern. Atlantic water enters the North Sea through the Channel in the south, and along the coast of Scotland in the north.
Water leaves the North Sea along the coast of Norway. The input of oceanic water is due to the predominant western winds and the tidal wave. Ebb-streams do not neutralise the water movement completely. This results in an anti-clock wise turning rest-stream (figure 3 .2a and 3.2b). In a prolonged period of non-western winds the rest-stream may turn temporary. The water in the North Sea is refreshed ito 2 times a year (North Sea Atlas, 1992),
In the Dutch coastal zone four water masses can be distinguished using parameters like salinity, temperature and origin of the water (Bergman et al., 1991).
Figure 3.2a: Rest-streams, watermasses en frontal zones in summer (North Sea Atlas, 1992)
I
I
/ I
\Front
and Baltic
northern North Seawater Scottish
and British coastwater
/, //
II
i Atlantic-oceanic water L...
—
\
central
North Seawater Front -
contuiental
-I
ScottishandBritish
eoastwater / Legenda
— surfacerest-stream
— bottom rest-stream
y/ front
—— boundarywaterniass
/
0 '00
Materials and methods
Scottish and British coastwater
central North Seawater
Norwegian and Baltic coastwater & Skagerrak
Legenda
— surfacerest-stream
—-'- bottom rest-stream 77'7 front
— — boundarywatermass
Figure 3.2h: Rest-streams, vatermasses and frontal zones in inter (North Sea Atlas. 1992)
\ northern
\
N N
\
Atlantic-oceanic water
I-
/ /
N
I.
/
/N
/
/ /
contmental/
coastwater\ /
/ Scottish and British / coastwater
,1v
//1
//
'00
• Channel water: Clear and nutrient poor water with high salinity. As a result of the origin, surface layers of the Atlantic Ocean, it is low in
organism and organic material content,
• Central North Sea water: This watermass is qualitative related to the Channel water, it has the same origin. But the input is mainly from the north. In winter it is mixed, in summer stratification occurs. The surface
layer is 20 to 30 meters deep and low in nutrient content, the deep layer is low in temperature and has high nutrient contents.
• English Coast water: The water follows the British coast from northern Scotland to East-Anglia and gets rising dissolved and particulate matter from rivers on its route. This water also contains erosion products from the South-England coast.
• Continental Coast water: A mixture of Channel, Scheldt, Maas and Rhinewater.
Especially the Rhine output has great impact on this
watermass. An important factor of high river input is the big pollution of nutrients and contaminants.
3.3 Samples
At each cruise bottle samples were collected from depths of 0 and 5 meters at each station.
The samples were filtered through Whatman GF/F® glass-fibre filters, The filtrates of these samples were kept for the yellow substance measurements. The bottles with the filtrates were kept in the dark, using brown bottles or tinfoil, in a coldroom (4°C). The storage period ranged from 8 months (cruise 6) to 19 months (cruise 1). During the storage period and the measurements the use of organic material for filters, stoppers and tubes was avoided.
3.4 Method and measurements
The spectral analysis of the samples was carried out with a
Cary® 3E double-beam spectrophotometer in a range of 250-800 nm with absorption measurements every 1 nm.The pathlength of the cuvette was 10 cm. With this pathlength we were able to measure even very low concentrations of yellow substance. The transmittance for the 10 cm pathlengths were converted into coefficients m1 using Napierian logarithms;Materials and methods
a(X) = 10[-ln(T(A))] (3.1)
In this equation the absorption coefficient at a certain wavelength is a(X), the measured transmittance at a certain wavelength is T(7.).
The accuracy of the measurements was estimated as 0.0005 unit of optical density ( 0.0115 m'). Before the measurements the samples were filtered on 0.2 m Nucleopore® filters. The filtrateswere measured for the content ofyellow substance.
A standard water (distilled Milli-Q® water) was used in the reference cell, in order to subtract
absorption by pure water. Eventual differences between the
cells were eliminated by memorising the baseline, obtained when both reference and sample cells were filled with standard water.3.5 Precautions and problems measuring yellow substance
Because the concentrations of yellow substance are often very low, the chance that some kind of contamination interferes with the measurements is conceivable. Therefore precautions were taken to minimise the chance of contamination in the samples, the filtrates and the water in the
reference cell.
Before every measurement-session all glass-ware was put in a NaOH solution for at least 16 hours. After that the glass-ware was cleaned using HC1 and rinsed with Milli-Q® distilled water.
This cleaning procedure wasn't used in the first few months of this research, The glass-ware was only rinsed with Milli-Q® distilled water. The results of the measurements in this period showed a big bump at 275 nm (figure 3.3), which was notrepeatable. This meaned that something in the method was influencing the results. Using a yellow substance-like matter, namely tea, every step of the protocol was checked.
Q
-C,
C,
C,
P..
C,z -c
7
6
3
2
0
S
250 300 350 400 450 500 550 600 650
wavelength (nm)
Figure 3 3: Two spectra of yellow substance absorption of the same sample. once measured with the glassfilter- holder ( ) which shows the bump and once measured with the iron filter-holder ( —-) whichshows no bump.
4
3
2
0
650
wavelength(nm)
Figure 3.4: Two spectra of absorption by tea, once measured without filtration (— ) and once measured with filtration ( ---).Usingfiltration the peak of the spectrum of tea becomes higher than without the filtration.
250 300 350 400 450 500 550 600
Materials and methods
In this check-experiment the following was tested and measured;
1, Milli-Q water against air to make sure that no bump occurs due to the water.
2. Solutions of NaOH and HC1 to see if any chemical is left on the glass-ware after cleaning.
3. The position of the cuvetttes in the spectrophotometer.
4. Tea-measurements, without filtration.
Tea-measurements, with filtration. Comparison of 4 and 5 to see if the filtration is the source of the bump.
Test 1,
2, 3 and 4 showed that the Milli-Q® water, the chemicals, the position in the
spectrophotometer and the machine itself had no effect on the spectra. Test 5, however, showed that sometimes the peak of the tea spectrum (figure 3,4) was higher than other times.This result suggests that the filtration has an effect on the spectra. After this test the holder of the Nucleopore® filter, which was composed of sintered glass, was changed for a stainless steel one. The glass-holder is a matrix of glass wherein contamination seems to accumulate and comes out this holder during filtration once in a while. This problem cannot occur with the iron-holder because this is an one-layer holder with small holes in it. After this change
sometimes a small bump occurred at 275 nm, but this bump was repeatable which means that it is due to a component in the sample.
With the iron-holder another problem arose. Because of the big pressure difference on the filter, bubbles of air occurred on the filter which slowed the filtration speed down. This problem was solved by bubbling Helium gas through the samples before the filtration, whereby air in the samples is replaced by Helium.
The absorption spectra of yellow substance show a dip around 740 nm. The origin of this dip was presumably absorption by water (Pegau and Zaneveld, 1993). The reason for this absorption was thought to be a temperature difference between the reference cuvette and the sample cuvette in the spectrophotometer. By warming the samples in a warm water-bath we tried to remove this temperature difference. However, after getting both cuvettes on the same temperature, the dip was still the same. it seems that the difference in temperature between the
reference and sample cell was not the main reason for the observed dip at 740 nin.
3.6 Data-analysis
3.6.1 Statistical analysis
The results of this research are statistically tested using analysis of variance (ANOVA) and analysis of covariance (ANCOVA). The variance between transects is tested, as well as differences within transects and between stations Using data-pairs of absorption coefficients at 380 mu and salinity the influence of salinity on the absorption by yellow substance is tested with ANCOVA.
3,6.2 Exponential curve fit
An exponential curve fit model is used to describe the exponentials of the spectra. Three equations, derived from equation 2.1, are important for the analysis.
Y(X) =a(A.f)*exp(_S*(A._Xf)) (3.2)
In this equation, Y is the absorption at a certain wavelength ?, a(Xref) is the absorption at the reference wavelength, usually 380 mu, S is the slope of the exponential and X1- is the reference wavelength depending on the segment which is analysed (see below).
This exponential relationship was fitted to the raw absorption data by non-linear regression based upon the Marquardt algorithm. The exponential relationship was fitted to the data from three broad bands:
- 250-650 nm (whole spectrum) ?¼.ref380nm
- 360-440 nm (Bricaud et al., 1981)
k= 400 nm
- 440-540 nm (Carder et al., 1989) 490urn
The exponential relationship was also fitted to the data from five narrow wavebands:
- 250-280 mu
Xf=265nm
- 280-320nm
?f=30Onm
- 320-360nm X,ef=34Onrn
- 360-400nm
f=38Onm
- 400-440 mu
refr42Omu
Materials and methods
The ratio of the absorption of hurnic substances at 465 nm and 665 nm were used by Power and Langford (1988) for the characterization of humic substances. In this research the ratios of segments of the spectrum were used to see if humic substances can be charatensed this way.
Below 650 nm our data became very noisy and irratic, because yellow substance absorption is extremely low in the 650-800 urn region. Therefore, these data are not discussed m theresults.
This region of the spectrum, however, was used to correct the data for negativevalues which would make the above mentioned exponential curve fit impossible. To achieve this the lowest value in the 650-800 nm region was subtracted of the values of the whole spectrum. This correction is based on the same kind of correction Bricaud et al. (1981) carried out to correct their data for scattering.
At all stations samples were taken at the surface and at 5 metres depth. Analysis of variance (ANOVA) is executed on the mean absorption coefficients (m1) at 380 nm, calculated from the whole spectrum (250-650 nm), of the cruises, the transects and of all samples together.
This analysis showed that no significant difference (p<O.O5) excisted between shallow and deep samples measured at different cruises (Table 4.1). Comparing the seasonal differences between shallow and deep samples at the transects didnot show significant (p<zO.05) differences (Table 4.2). The differences between shallow and deep samples, when all samples were taken into account, were also not significant (p<0.05), see Table 4,3, These results
meaned that the shallow and deep samples measured at each station could be used at
replicates.
Table 4.1: Statistical analysis (ANOVA) of the differences between shallow and deep samples at different cruises (sh means shallow, d means deep), p<O.05 (ns =notsiificant, s =siiificant).
April 1993 cruise Sum of Squares df Mean Square F p-level Mean (± sd)
Effect 0.000008 1 0.000008 0.0003 0.9854 sh 0.4990 (± 0.1445)
error 0.9813 42 0.0234 ns d 0.4499(±0.1608)
June 1993 cruise Sum of Squares df Mean Square F p-level Mean (± sd)
Effect 0.0011 1 0.0011 0.0188 0.8914 sh 0.5031 (± 0.2429)
error 3.3259 56 0.0594 ns d 0.5119(±0.2444)
Sept.-Oct. 1993 cruise Sum of Squares df Mean Square F p-level Mean (± sd)
Effect 0.00011 1 0.00011 0.0045 0.9469 sh 0.5282(±0.1657)
error 0.9282 37 0.0251 ns d 0.5248(±0.1503)
December 1993 cruise Sum of Squares df Mean Square F p-level Mean (± sd)
Effect 0.0019 1 0.0019 0.0534 0.8189 sh 0,8099(±0.1925)
error 0.9805 28 0.0350 ns d 0.8257 (± 0.1816)
Jan-Feb. 1994 cruise Sum of Squares df Mean Square F 'p-level Mean (± sd)
Effect 0.0428 1 0.0428 0.2807 0.5990 sh 0.7578 (± 0.4054)
error 6.4011 42 0.1524 ns d 0.6954(±0.3748)
July 1994 cruise Sum of Square df Mean Square F p-level Mean (± sd)
Effect 0.0194 1 0.0194 0.5462 0.4622 sh 0.4755(±0.1862)
error 2.6270 74 0.0355 ns d 0.4435 (± 0.1908)
Results
Table 4.2: Statistical analysis (ANOVA) of the differences between shallow and deep samples at different transects (sh means shallow, d means deep), p<O.O5 (ns not significant, s significant).
Channel Sum of Squares df Mean Square F p-level Mean (± sd)
Effect 0.0437 1 0.0044 0.3326 0.5683 sh 0.3467 (± 0.1265)
error 0.4072 31 0.0131 ns d 0.3237(±0.1004)
Noordwijk Sum of Squares df Mean Square F p-level Mean (± sd)
Effect 0.00002 1 0.00002 0.0002 0.9885 sh 0.5642 (± 0.3134)
error 6.1463 71 0.0866 ns d 0.5632 (± 0.273 1)
Texel Sum of Squares df Mean Square F p-level Mean (± sd)
Effect 0.0254 1 0.0254 0.2984 0.5864 sh 0.6121 (± 0.2909)
error 6.7194 79 0.851 ns d 0.5767(±0.2924)
Walcheren Sum of Squares df Mean Square F p-level Mean (± sd)
Effect 0.0022 1 0.0022 0.0673 0.7961 sh 0.5489 (± 0.1648)
error 2.3989 73 0.0329 ns d 0.5598 (± 0.1967)
Table 4.3: Statistical analysis (ANOVA) of the differences between shallow and deep samples, when all samples were taken into accoimt (sh means shallow, d means deep), p<O.O5 (ns not significant, s
significant).
4.1 Spatial variation
The spatial variation of yellow substance could be tested by comparing the absorption coefficients at different stations within the transects and between transects.
4.1.1 Spatial variation within transects
The Channel transect has been measured two times, once in June 1993 and once in July 1994.
Measurements were done from the coast of England at Dover to the coast of France near
obtained at this transect. The Waicheren and Noordwijk transects were both sampled 5 times.
Measurements were carried out from on-shore to off-shore stations. Figure 4.2 and Figure 4.3 show the absorption spectra obtained from the measurements at these transects. TheTexel transect has been measured most frequently, samples were collected on all six cruises. Spectra obtained from measurements at this transect are shown in Figure 4.4, The Fnsian Frontand the Marsdiep have both been sampled once, in July 1994 and December 1993 respectively.
The spectra obtained during those cruises are shown in Figure 4.5 and Figure 4.6.
0.6
0.5
04
U
U 0.3
I I I I I I
0 1 2 3 4 5 6
station no.
Figure 4.1: Absorption coefficients (380nm) from the stations in the Channel. Both cruises, June 1993 and July 1994 are shown. Station 1 is at the English side of the Channel. station 8 is at the French side (° are deep samples, are shallow samples).
At all transects all spectra, obtained from measurements of samples from different cruises show a decrease of absorption by yellow substance from the on-shore stations to the off-shore stations. Because the Channel transect was sampled from the English to the French coast it didnot really show a decrease of yellow substance from on-shore to off-shore, but from the
French coast to the English coast. The Marsdiep samples were all taken very near to
eachother, therefore the Marsdiep shouldnot be used as a transect but as a station with replicates. In the 275 nm region at all spectra a bump occurs, this is most clearly seen at off- shore stations with lower absorption coefficients.June 1993
June 1993 July 1994 July 1994
7 8 9
