- Email: [email protected]
Tracing water mass mixing in the Baltic–North Sea transition zone using the optical properties of coloured dissolved organic matter
Estuarine, Coastal and Shelf Science 87 (2010) 156–162
Contents lists available at ScienceDirect
Estuarine, Coastal and Shelf Science journal homepage: www.elsevier.com/locate/ecss
Tracing water mass mixing in the Baltic–North Sea transition zone using the optical properties of coloured dissolved organic matter Colin A. Stedmon a, *, Christopher L. Osburn b,1, Theis Kragh c a
Department of Marine Ecology, National Environmental Research Institute, University of Aarhus, Frederiksborgvej 399, Roskilde, 4000 Denmark US Naval Research Laboratory, Marine Biogeochemistry Section, Code 6114, Washington, DC, USA c Freshwater Biological Laboratory, University of Copenhagen, Helsingørsgade 51, DK-3400 Hillerød, Denmark b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 21 July 2009 Accepted 29 December 2009 Available online 18 January 2010
The distribution and characteristics of coloured dissolved organic matter (CDOM) in the Baltic – North Sea transition zone were studied. The aim was to assess the validity of predicting CDOM absorption in the region on the basis of water mass mixing alone and demonstrate the utility of CDOM as an indicator of water mass mixing in coastal seas. A three-end-member mixing model representing the three major allochthonous CDOM sources was sufficient to describe the patterns in CDOM absorption distribution observed. The three-end-member water masses were the: Baltic outflow, German Bight and the central North Sea. Previously, it was thought that water from the German Bight transported northwards in the Jutland coastal current only sporadically influenced mixing between the Baltic and North Sea. The results from this study show that water from the German Bight is detectable at salinities down to 12 in the Kattegat and Belt Sea. On average, 23% of the CDOM in bottom waters of the Kattegat, Great Belt, Belt Sea, Arkona Sea and the Sound originated from the German Bight. Using this conservative mixing model approach, local CDOM inputs were detectable but found to be limited, representing only 0.25% of CDOM in the surface waters of the Kattegat and Belt Sea. The conservative mixing of CDOM makes it possible to predict its distribution and characteristics and offers a powerful tool for tracing water mass mixing in the region. The results also emphasize the need to include the Jutland Coastal current in hydrodynamic models for the region. Ó 2010 Elsevier Ltd. All rights reserved.
Keywords: baltic Sea north Sea german Bight water mass mixing tracers dissolved organic matter
1. Introduction Coloured dissolved organic matter (CDOM) is one of the major light-attenuating components of sea water. It is responsible for much of the ultraviolet light attenuation and, when present in high concentrations, also influences visible light attenuation. The CDOM UV–visible absorption spectrum (300–650 nm) can be modelled using an exponential relationship (Jerlov, 1968; Lundgren, 1976; Bricaud et al., 1981).
aðlÞ ¼ aðl0 Þ eSðl0 lÞ
(1)
where l0 is a reference wavelength l and S is the spectral slope coefficient, characterizing the exponential decline in absorption with increasing wavelength. The absorption at a specific
* Corresponding author. E-mail address: [email protected] (C.A. Stedmon). 1 Present address: Department of Marine, Earth, and Atmospheric Sciences, North Carolina State University, Raleigh, NC, USA. 0272-7714/$ – see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.ecss.2009.12.022
wavelength (e.g. 300 nm (a300)) can be used as an estimate for CDOM concentration. CDOM originates primarily from the degradation of terrestrial and aquatic plant matter. In regions influenced by freshwater run off, terrestrial CDOM often dominates and it is only when this source is considerably diluted that the presence of autochthonous marine CDOM can be detected (Blough et al., 1993; Stedmon et al., 2000; Stedmon and Markager, 2003). When allochthonous supply and mixing rates exceed autochthonous production and degradation rates, the concentration and characteristics of CDOM behave conservatively. There is a considerable allochthonous supply of CDOM in the coastal waters of the North Sea and Baltic Sea and this led early studies to conclude that CDOM represents purely terrestrial material (Højerslev, 1979). Measurements from oceanic regions far from the influence of rivers, however, have since revealed a lower intensity, pelagic, autochthonous CDOM source (Jerlov, 1976; Bricaud et al., 1981; Nelson et al., 1998). In sub-surface open ocean waters, temperature and salinity are used to trace water mass mixing, since both parameters behave conservatively. In coastal or surface waters temperature is not a conservative parameter as it varies seasonally as a result of heat
C.A. Stedmon et al. / Estuarine, Coastal and Shelf Science 87 (2010) 156–162
exchange with the atmosphere. So for these waters alternative parameters are required. In combination with salinity, CDOM absorption at specific wavelengths can be used instead of temperature to trace water mass mixing in coastal waters (Kalle, 1949; Højerslev, 1988; Karabashev et al., 1993). In the Skagerrak and Kattegat, the brackish Baltic Sea water with high CDOM absorption mixes with saline North Sea water with low CDOM absorption (Fig. 1). The Jutland coastal current, flowing north along the western coast of Jutland from the German Bight, also influences the Skagerrak and Kattegat (Højerslev et al., 1996; Warnock et al., 1999; Nielsen, 2000). The current carries water from the rivers flowing into the Southern North Sea, which contain high CDOM concentrations (Laane and Kramer, 1990; Højerslev et al., 1996; Warnock et al., 1999). However, the effect of the German Bight water in the Skagerrak and Kattegat is thought to be limited in comparison to the effect of the large volumes of water transported from the North Sea and Baltic Sea (Aarup et al., 1996; Nielsen, 2000). It has been shown that salinity–CDOM relationships can characterize and quantify the mixing of these three water masses (Malmberg, 1964; Aarup et al., 1996; Højerslev et al., 1996). Despite this seemingly conservative behaviour, Højerslev and Aas (2001) did not find systematic trends in S. Stedmon and Markager (2003) estimated the variability of S during conservative mixing of two end members, facilitating analysis of field data and the identification of nonconservative processes acting on CDOM, such as local autochthonous production (Kowalczuk et al., 2006; Guo et al., 2007). The aims
157
of this work are two fold. Firstly we aim to predict CDOM absorption of water masses in the region based on water mass mixing alone. We assume that mixing is the dominant process controlling both the distribution and character (S) of CDOM. If this is true, a three-end-member mixing model representing the three major allochthonous CDOM sources should be sufficient to describe CDOM absorption. If this assumption is invalid and autochthonous production and removal processes control the distribution and characteristics of CDOM, the model results will indicate the magnitude of these processes. Secondly, we aim to demonstrate the utility of CDOM as an indicator of water mass mixing in coastal seas. By applying simple mixing models water at different depths and locations can be fractionated into the relative proportion of the three dominant end member water masses. 2. Methods 2.1. Sampling and measurement Water samples were collected using the national water quality monitoring cruises on the R/V Gunnar Thorson. Samples were obtained from ten cruises (Table 1) in the transition zone between the North Sea and the Baltic Sea (Fig. 1). Data from February 1999 originates from Stedmon et al. (2000) and is included to expand the data analysis. At each station, samples were collected at 1 m and 5 m intervals from 5 to 30 m. For stations deeper than 30 m, samples were taken at 10 m intervals below 30 m and at 1 m over the sediment. A rosette with a Seabird CTD and 12 5 L Niskin water samplers was used and calibrated according to accredited procedures required by the national monitoring program. CDOM samples were gently filtered through pre-combusted GF/ F filters (approximate pore size 0.7 mm) using a syringe, into acidwashed 100 mL brown glass bottles and stored refrigerated in the dark. Storage time varied from one to two weeks, since the cruise lasted 5 days and all the optical characteristics of the samples were determined during the week directly following each cruise. Stedmon et al. (2000) investigated the effects of storage on CDOM absorption for sample from these waters and found little change for periods up to 27 days. CDOM absorbance was measured on a Shimadzu UV-2401PC spectrophotometer using MilliQ water as a reference, according to Stedmon et al. (2000). A 10 cm quartz cuvette was used for all samples and the absorption coefficient was calculated from absorbance (A) according to Equation (2).
aðlÞ ¼ 2:303AðlÞ =L
(2)
where L is the pathlength in meters (0.1 m). The spectral slope, S, of the absorption curve from 300 to 650 nm was obtained by fitting Equation (1) to the data using the
Table 1 Cruise name, sampling time, number of stations and samples, and location of cruise track. Inner Danish waters include Arkona Sea, Belt Sea, Sound and Kattegat (Fig. 1).
Fig. 1. (a) Map of study site indicating the different surface currents and regional seas. Also shown are the station locations and the coloured contours represent the average salinity at 5 m. (b) A diagram of generalized characteristics of the water column structure in the region. For a map of the bathymetry in the region, see Fig. 1 in Højerslev et al. (1996).
Cruise
period
#Stations
#Samples
Description
GT189
Feb.1999
74
324
GT191
Aug. 1999
72
136
GT198 GT237 GT238 GT239 GT240 GT242 GT243 GT244
Sept. 2000 Aug. 2006 Sept. 2006 Oct. 2006 Feb. 2007 Aug 2007 Sept 2007 Feb 2008
24 24 24 24 22 17 17 17
131 182 135 138 164 131 134 132
Inner Danish waters, Skagerrak and North Sea Inner Danish waters, Skagerrak and North Sea Inner Danish waters Inner Danish waters Inner Danish waters Inner Danish waters Inner Danish waters Inner Danish waters Inner Danish waters Inner Danish waters
158
C.A. Stedmon et al. / Estuarine, Coastal and Shelf Science 87 (2010) 156–162
non-linear regression technique of Stedmon et al. (2000). It is known than S values are dependent on the wavelength range over which they are derived (Del Castillo et al., 1999; Stedmon et al., 2000; Twardowski et al., 2004). Slopes of the absorption spectra in the UV-B region (<300 nm) have been used to provide chemical characteristics of DOM (e.g. Helms et al., 2008). Other studies, focused on bio-optics or remote sensing, have reported slopes derived from longer wavelengths relevant for the penetration of sun light. In this study, the 300–650 nm range is used as one of the aims is to be able to predict CDOM light absorption geographically and temporally in order to aid the development of algorithms for light penetration in both regional hydrodynamic models and remote sensing applications. In this study the absorption coefficient at 300 nm will be used to present the distribution in CDOM concentrations. The choice of wavelength is essentially arbitrary as with knowledge of S the absorption coefficient at one wavelength can be converted to another. However, absorption by CDOM decreases with increasing wavelength and in some of the samples from the North Sea values are below the detection limit for much of the spectrum. 300 nm was chosen as this is the wavelength with the highest signal (furthest from the detection limit) for all samples and it is still within the ecologically relevant wavelength range for studies of underwater light penetration (300–700 nm). 2.2. Water mass mixing analysis The analysis based on Højerslev et al. (1996) assumes conservative behaviour of salinity and CDOM absorption during mixing of three major water masses: North Sea (NS) water originating from the central North Sea and Atlantic, Baltic Sea (BS) water representing the water flowing out of the Baltic, and German Bight (GB) water flowing northwards along the west coast of Jutland from the German Bight to the Skagerrak (Fig. 1). The salinity (Salt) and CDOM absorption (a) in a given water sample from the region can be expected to follow Equations (3)–(5).
Salt ¼ fNS SaltNS þ fBS SaltBS þ fGB SaltGB
(3)
aðlÞ ¼ fBS aBSðlÞ þ fGB aGBðlÞ þ fNS aNSðlÞ
(4)
fBS þ fNS þ fGB ¼ 1
(5)
where f represents the fraction of each water mass in the sample. These three equations with three unknowns can then be solved. After determining suitable end member values for CDOM and salinity (see the Results section), the relative proportion of each water mass can be determined for a given sample from the region. 2.3. Effects of conservative mixing on CDOMs spectral characteristics CDOM absorption during conservative mixing were estimated according to the Stedmon and Markager (2003), expanded to incorporate three end members. Combining the spectral absorption properties of CDOM (Eq. (1)) with the conservative mixing equation from Højerslev et al. (1996) (Eq. (4)), we derive the following equation for CDOM absorption in a sample.
aðlÞ ¼ fBS aBSðl0 Þ eSBS ðl0 lÞ þ fGB aGBðl0 Þ eSGB ðl0 lÞ þ fNS aNSðl0 Þ eSNS ðl0 lÞ
(6)
This absorption spectrum can now be modelled using Equation (1) and a new theoretical, conservative mixing S value can be derived. The comparison of model-derived S values to the measured S values offers an additional test of the conservative CDOM mixing in the region. 3. Results The data indicate three end members and two mixing lines (Fig. 2a, b). The first line represents mixing of German Bight water salinity w 29) with North Sea water (a300 > 5 m1, (a300 w 0.05 m1, salinity > 35). The second mixing line represents mixing of Baltic Sea water (a300 w 5 m1, salinity < 8,) with intermediate water on the German Bight–North Sea mixing line (a300 w 2 m1, salinity w 32–33). The S values generally tend to
Fig. 2. (a) CDOM a300 versus salinity for all data (Table 1), (b) average distribution of a300 at 5 m depth in the region, (c) CDOM spectral slope, S, against salinity.
C.A. Stedmon et al. / Estuarine, Coastal and Shelf Science 87 (2010) 156–162
decrease with increasing salinity in an approximately linear fashion; from 20 mm1 at salinities below 8, to 18 mm1 at salinities above 32 (Fig. 2c). There was a greater variability in S values at high salinities ranging from about 12 to 25 mm1. In order to determine the characteristics of the three end members, averages of samples at each extreme were calculated (Table 2, Fig. 2). For the Baltic Sea end member, data at salinities <8 were averaged; for the North Sea end member, data at salinities >35 and a300 < 0.5 m1 were averaged; and for the German Bight, averages included data where both the salinity >28 and a300 > 5 m1. Averages were used in order to obtain the most representative values for each and to reduce the sensitivity of the mixing calculations to end members input values. These values were then used in Equations (3)–(6) and the relative proportions (f) of each water mass were calculated for each sample. Under ideal conditions with, for example, fixed precise values of end member characteristics, no analytical error, no instrumental noise, nor autochthonous processes, the values for f would only vary between 0 and 1. For the majority of the samples in this data set f varied in this range, which is encouraging (Fig. 3). Values of f greater than 1 and below 0 are largely constrained to samples at the extremes, representing 100% of each of the three end members and this is purely an artefact of the mean values were used (Table 2). We found three mixing regimes (Fig. 3): at salinities below 20, the Baltic Sea CDOM end member dominates, representing over 50% of the total, and the remaining fraction is relatively evenly split between German Bight and North Sea water; at salinities above 32, the influence of the Baltic Sea CDOM is minimal and mixing between North Sea and German Bight water dominates; at intermediate salinities (20–26) the three water masses generally contribute equally. The gradual dilution of the Baltic outflow is clear (Fig. 4a, d) and the mixing of the North Sea waters with the German Bight along the Jutland coast and with Baltic water in the Kattegat is apparent (Fig. 4b, e). Water (and CDOM) from the German Bight can be traced in surface waters as far as the Belt Sea and Sound between Sweden and Denmark (Fig. 4c). In bottom waters German Bight water is present at slightly higher levels and clearly restricted from continuing into the Baltic Sea by the two shallow sills (Darss w 12 E, 54.5 N; Drogden w 12.5 E, 55.5 N) (Fig. 4f). The inner Danish waters (Southern Kattegat, Great Belt, Belt Sea, Arkona Sea, and the Sound) with salinities between 14 and 30 contained on average 23% German Bight CDOM (Fig. 3). Using the results of the water mass mixing calculations and the end member characteristics of the three CDOM pools, estimates for S values can also be calculated. These spectra were derived for each sample and then modelled using Equation (1) to determine conservative mixing values for S. The fit to the a300 values generally was very good; especially at the two salinity extremes (Fig. 5a). However, there was a systematic increase in the unexplained a300 between salinities of 15 and 20 towards a maximum where model estimates were approximately 0.01 m1 lower than measured values (Fig. 5a). This corresponded to the waters of the Kattegat and Belt Sea (Fig. 5b). The deviation decreased approximately linearly as salinities increased further from 24. In contrast, no overall systematic trends for the S value estimates were apparent (Fig. 5c). The conservative model appears to
159
describe much of the linear decrease in S values seen in Fig. 2c. However, the difference between the measured and modelled S values, increases slightly at higher salinities. This scatter corresponded to samples with low CDOM concentrations (a300 < 1 m1). One systematic deviation from the model was apparent. This was for samples from the Skagerrak deep and intermediate waters with salinities above 34.5 (Fig. 5c). Many of these samples had lower S values than predicted by the mixing model. 4. Discussion The major assumptions in this approach are (1) that there are only three dominant CDOM pools in these offshore waters, originating from the North Sea, Baltic Sea and the German Bight, (2) CDOM absorption properties are mixing conservatively, and (3) there are small variations in the properties of CDOM of the end members during the sampling period. The ability of this simple model in describing the large scale gradients in CDOM distribution and spectral characteristics (S) of CDOM in these waters suggests that these assumptions were to a certain extent valid. The presence of three dominant end members is clearly apparent (Fig. 2a and b). For a300 there are two general conservative mixing lines and it is clear that the saline marine end member for the waters of the Baltic Sea consists of a mixture of German Bight water and North Sea water. Although there is a degree of scatter along the Baltic Sea mixing line in Fig. 2a, the overall trend along the mixing lines is linear suggesting that CDOM is mixing on the whole conservatively. The fact that all the data from ten different cruises fall along the same general mixing lines which are similar to that also seen in earlier work (Højerslev et al., 1996; Stedmon et al., 2000) indicates that there are no large systematic changes in the concentrations and characteristics of CDOM in the three identified end members during the study period (1999–2008). Using the S values reported in Table 2 the a300 values can be transformed to approximate a380 values and compared with the end member characteristics reported by Højerslev et al. (1996). Our study’s a380 values for the Baltic Sea and German Bight end members were slightly lower (by 0.15 m1 and 0.30 m1 respectively) and the North Sea end member was higher by 0.05 m1. However, when compared to the degree of scatter in the Højerslev et al. (1996) data (see their Fig. 4) and the different techniques
Table 2 CDOM and salinity characteristics of each end member. Values are means and standard deviations are shown in brackets. End members
n
a300 (m1)
S (mm1)
Salinity
Baltic Sea German Bight North Sea
48 10 28
4.89 (0.19) 5.23 (0.19) 0.60 (0.04)
22.5 (0.8) 18.3 (0.5) 20.2 (1.9)
7.7 (0.20) 29.7 (0.24) 35.1 (0.06)
Fig. 3. Fraction of the three water masses characterized in Table 2 in each sample in the data set derived using the approach described by Højerslev et al. (1996).
160
C.A. Stedmon et al. / Estuarine, Coastal and Shelf Science 87 (2010) 156–162
Fig. 4. The average distribution at 5 m depth (a–c) and 5 m from the bottom (d–f) of the fractions of Baltic Sea water (top panels), North Sea water (middle panels), and German Bight water (lower panels).
(earlier study included CDOM data estimated from in situ beam attenuation measurements) these differences are relatively minor. The modelling results show that the approximately linear decreasing trend in S with increasing salinity (Fig. 2c) can be replicated by the conservative mixing approximations of S (Fig. 5c). This approximate linear behaviour in S is uncommon (Blough and Del Vecchio, 2002) and is here shown to be due to the fact that there are three major CDOM pools with differing concentrations undergoing a complex mixing pattern (Fig. 3). The close match of the conservative mixing estimates with the measured S values indicates that on the whole, CDOMs spectral properties in these waters can be modelled relatively easily. It is likely that some of the scatter in the trends observed is due to local non-conservative processes such as release of DOM from sediments (Skoog et al., 1996), seasonal autochthonous DOM production (Zweifel et al., 1995) and photochemical degradation of CDOM (Stedmon et al., 2007). These processes are important at local scales. For example, at some stations in the data set lower a300
values can be found at 1 m compared to 5 m despite the samples having the same salinity (data not shown). This could be evidence for photobleaching occurring in surface waters. In addition the sample from 1 m above the sediment at times had a slightly higher a300 than the sample above it (data not shown). This was interpreted as a result of release of CDOM from the sediments (Skoog et al., 1996). Autochthonous CDOM production is these waters is expected to be difficult to detect due to the high background concentrations of CDOM from the German Bight and Baltic Sea. In the high salinity waters from the North Sea, which have low background concentrations of CDOM it is, however, detectable, with fresh CDOM having a low S values (Stedmon and Markager, 2001). This can be seen at salinities of 35 in Fig. 5c. However, despite the occurrence of these non-conservative processes their overall impact on the distribution of CDOM and its spectral properties appears to be limited in comparison to the gradients resulting from water mass mixing in the region over timescales of several months. The combination of continual re-supply and high
C.A. Stedmon et al. / Estuarine, Coastal and Shelf Science 87 (2010) 156–162
161
Fig. 5. (a) Difference between the measured and modelled a300 plotted against salinity, (b) the average difference in a300 at 5 m depth in the region, (c) the difference between the measured and modelled S values plotted against salinity.
background concentrations of CDOM are a likely reason for these results. Bowers and Brett (2008) recently demonstrated with a box model that for an estuarine system such as this with relatively long residence time compared to the variability in the source (river water), a time–mean mixing line with salinity would be expected. The large volumes of water involved essentially dampen out the potential seasonal signal in source waters resulting in a constant mixing line. This appears to be what is essentially occurring here. The seasonal changes in the German Bight and Baltic Sea end members are dampened out and a time–mean mixing line exists. Extrapolating this result further and assuming no long term change in the values of the end member characteristics, the results here demonstrate how CDOM distribution and spectral characteristics in the region can be easily predicted (modelled) from knowledge of boundary conditions (end members) alone. This result implies that the regional distribution of CDOM can be easily modelled and implemented in 3D-hydrodynamic models such as COHERENS (Bendtsen et al., 2009), providing estimates on spectral light attenuation for ecosystem models and a novel remote sensing ground-truthing tool. The positive deviations in a300 from the conservative mixing model at salinities 15–20 suggest local freshwater inputs of CDOM into the waters of the Kattegat and Belt Sea (Fig. 5a and b). Without the use of this mixing model approach, these inputs would be difficult to identify. Indeed, earlier studies have considered them negligible due to the relatively low volume of water in comparison to the Baltic outflow (Aarup, 1994; Højerslev et al., 1996). The a300 values measured for these salinities were approximately 4 m1, so the modelled difference of 0.01 m1 represents only 0.25% of the CDOM values and is within the detection limits of the spectrophotometric method (Stedmon and Markager, 2001). However, there is a systematic trend with salinity in the deviations between the modelled and measured a300 values. Maximum deviations are found at a salinity of w20 and there is a clear linear decrease with increasing salinity. This pattern is indicative of conservative mixing of local estuarine inputs of CDOM into the surface waters flowing out of the Baltic. In contrast to the results for a300, the impact of local CDOM inputs at intermediate salinities on S values was not apparent. There are two potential reasons, which are not mutually exclusive. First, local inputs had very low intensities in comparison to the
three major CDOM pools and therefore have little impact on S values (Stedmon and Markager, 2003). Second, CDOM exported from fjords in the region has very similar S values to values found in the Kattegat (Stedmon et al., 2000). The estimates of the relative amounts of CDOM from different sources are intriguing and warrant further study. It is clear that CDOM from the German Bight can be traced into the Kattegat and Belt Sea waters down to a salinity of 12. Despite the sporadic and relatively small inflow from the Jutland coastal current, its impact can still be detected due to its high CDOM concentrations (Table 2). The estimates for the amount of German Bight water in the Kattegat and Belt Sea waters (w23%) agrees well with an earlier radioisotope study, which concluded that Kattegat bottom water contains 28% German Bight water (Dahlgaard et al., 1995). With the modelling approach demonstrated here, the relative amounts of each water mass in any CDOM water sample from these waters can be determined using relatively economical and simple techniques. This can have several benefits for estimating import and export of carbon and nutrients to and from the Baltic Sea. The Jutland Coastal current is controlled by wind, tidal currents and density-driven currents that all contribute to its northwards flow along the North Sea coast of Denmark (Nielsen, 2000). This current is primarily diluted in the southern Skagerrak and joins the Norwegian coastal current in the northern Skagerrak (Fig. 1). It is seldom observed entering the Northern Kattegat. These observations are based primarily on salinity and nutrient analyses, which is not an optimal approach due to the non-conservative nature of nutrient concentrations. By the time the current reaches the Skagerrak–Kattegat boundary, the salinity differences are small and nutrient concentrations, which often are high in the current, do not behave conservatively during the relevant mixing times. The current has little effect on the overall salt transport into the Baltic Sea and therefore is often neglected (Rydberg, 1993) or de-emphasized in hydrodynamic models for the region. The results presented here clearly show that the marine end member for the Baltic Sea (i.e. water entering the northern Kattegat from the North Sea) contains a noteworthy amount of water (and CDOM) from the German Bight and this may have implications for the transport of other constituents such as nutrients and organic matter in general. The patterns seen in CDOM distribution and characteristics cannot be explained without the inclusion of a third water mass with high CDOM concentrations, here found to be water from the German Bight.
162
C.A. Stedmon et al. / Estuarine, Coastal and Shelf Science 87 (2010) 156–162
5. Conclusions The major findings of this study are: 1. The three-end-member water mass mixing model by Højerslev et al. (1996) based on the distribution of CDOM in these waters, has been validated with a more recent, independent and comprehensive data set spanning ten years, and expanded to include spectral data. This approach can explain the trends observed in S values in these waters and paves the way for predicting the spectral light absorption by CDOM using measurements from boundary regions alone. 2. Waters from the German Bight influence the Kattegat and Skagerrak area more than previously perceived. 3. The combination of CDOM measurements and a three-endmember water mass mixing model can be used to resolve the distribution of local freshwater inputs from Germany, Denmark and Sweden into the Kattegat and Belt Sea region. The fact that much of the qualitative and quantitative variability in CDOM absorption properties can be easily modelled is advantageous for phytoplankton remote sensing applications in the region where the signal from CDOM has to be considered (Karabashev, 1992). Additionally, the knowledge gained here can be easily assimilated into ecosystem models in order to improve the estimation of photic depth and the prediction of primary productivity. These results also reemphasize the great potential of using optical properties of CDOM to trace water mass mixing, which to date has not been fully exploited by physical oceanographers. The approach provides a method to resolve the influence of smaller currents and water masses in the supply and transport of carbon and nutrients in coastal seas. Acknowledgements This work was funded by the US. Office of Naval Research (Grant: N00014-06-1-0357), the Danish Ministry of the Environment and the Danish Council for Strategic Research. Peter Kofoed is thanked for his assistance in collecting samples. The crew and laboratory staff of R/V Gunnar Thorson is also thanked for their cooperation. The contour plots were created using Ocean Data View (Schlitzer, R., http://odv.awi.de). References Aarup, T., 1994. Interpretation of satellite ocean color data of the transition zone between the North Sea and the Baltic Sea. Ph.D. thesis, Univ. of Copenhagen. Aarup, T., Holt, N., Højerslev, N.K., 1996. Optical measurements in the North Sea– Baltic Sea transition zone. II. Water mass classification along the Jutland west coast from salinity and spectral irradiance measurements. Continental Shelf Research 16, 1343–1353. Bendtsen, J., Gustafsson, K.E., So¨derkvist, J., Hansen, J.L.S., 2009. Ventilation of bottom water in the North Sea–Baltic Sea transition zone. Journal of Marine Systems 75, 138–149. Blough, N.V., Del Vecchio, R., 2002. Chromophoric DOM in the coastal environment. In: Hansell, D., Carlson, C. (Eds.), Biogeochemistry of Marine Dissolved Organic Matter. Academic Press, pp. 509–546. Blough, N.V., Zafiriou, O.C., Bonilla, J., 1993. Optical absorption spectra of waters from the Orinoco River outflow: terrestrial input of colored organic matter to the Caribbean. Journal of Geophysical Research 98, 2271–2278. Bowers, D., Brett, H.L., 2008. The relationship between CDOM and salinity in estuaries: an analytical and graphical solution. Journal of Marine Systems 73, 1–7.
Bricaud, A., Morel, A., Prieur, L., 1981. Absorption by dissolved organic matter of the sea (yellow substance) in the UV and visible domains. Limnology and Oceanography 26, 43–53. Dahlgaard, H., Herrmann, J., Salomon, J.C., 1995. A tracer study of the transport of coastal water from the English Channel through the German Bight to the Kattegat. Journal of Marine Systems 6, 415–425. Del Castillo, C.E., Coble, P.G., Morrel, J.M., Lopez, J.M., Corredor, J.E., 1999. Analysis of the optical properties of the Orinoco River plume by absorption and fluorescence spectroscopy. Marine Chemistry 66, 35–51. Guo, W., Stedmon, C.A., Han, Y., Wu, F., Yu, X., Hu, M., 2007. The conservative and non-conservative behavior of chromophoric dissolved organic matter in Chinese estuarine waters. Marine Chemistry 107, 357–366. Helms, J., Stubbins, A., Ritchie, J.D., Minor, E., Kieber, D.J., Mopper, K., 2008. Absorption spectral slopes and slope ratios as indicators of molecular weight, source, and photobleaching of chromophoric dissolved organic matter. Limnology and Oceanography 53 (3), 955–969. Højerslev, N.K., 1979. On the origin of yellow substance in the marine environment. In: 17th General Assembly of I.A.P.S.O. (Canberra, 1979), Abstracts, 71. Højerslev, N.K., 1988. Natural occurrences and optical effects of gelbstoff. Report No. 50. Department of Geophysics, University of Copenhagen. Højerslev, N.K., Aas, E., 2001. Spectral light absorption by yellow substance in the Kattegat–Skagerrak area. Oceanologia 43, 39–60. Højerslev, N.K., Holt, N., Aarup, T., 1996. Optical measurements in the North Sea– Baltic Sea transition zone. I. On the origin of deep water in the Kattegat. Continental Shelf Research 16, 1329–1342. Jerlov, N.G., 1968. Optical oceanography. In: Oceanography Series 5. Elsevier. Jerlov, N.G., 1976. Marine Optics. Elsevier. Kalle, K., 1949. Fluoreszenz und Gelbstoff im Bottnischen und Finnischen Meerbusen. Deutsche Hydrografische Zeitschrift 2, 96–105. Karabashev, G.S., 1992. On the influence of dissolved organic matter on remotesensing of chlorophyll in the Straits of Skagerrak and Kattegat. Oceanologica Acta 15, 255–259. Karabashev, G.S., Khanaev, S.A., Kuleshov, A.F., 1993. On the variability of ‘‘yellow substance’’ in the Skagerrak and the Kattegat. Oceanologica Acta 16, 115–125. Kowalczuk, P., Stedmon, C.A., Markager, S., 2006. Modeling absorption by CDOM in the Baltic Sea from season, salinity and chlorophyll. Marine Chemistry 101, 1–11. Laane, R.W.P.M., Kramer, K.J.M., 1990. Natural fluorescence in the North Sea and its major estuaries. Netherlands Journal of Sea Research 26, 1–9. Lundgren B., 1976. Spectral transmittance measurements in the Baltic. Rep. no. 30. Department of Physical Oceanography, University of Copenhagen. Malmberg, S., 1964. Transparency measurements in the Skagerrak. Go¨teborgs Kungliga Vetenskaps och Vitterhets-Samha¨lles Handlingar, Series B 9, 5–18. Nelson, N.B., Siegel, D.A., Michaels, A.F., 1998. Seasonal dynamics of colored dissolved material in the Sargasso Sea. Deep Sea Research 45, 931–957. Nielsen, M.H., 2000. A Dynamic Description of the Hydrographic Classification of the Jutland Coastal Current (in Danish). Danish Meteorological Institute. Scientific Report 00-15. Rydberg, L., 1993. On the Skagerrak circulation and the supply of water from the southern North Sea to the Skagerrak. Department of Oceanography, University of Gothenburg (Sweden), Report No. 53, 34pp. Skoog, A., Hall, P.O.J., Hulth, S., Paxeus, N., Rutgers van der Loeff, M., Westerlund, S., 1996. Early diagenetic production and sediment–water exchange of fluorescent dissolved organic matter in the coastal environment. Geochimica et Cosmochimica Acta 60 (19), 3619–3629. Stedmon, C.A., Markager, S., 2001. The optics of chromophoric dissolved organic matter (CDOM) in the Greenland Sea: an algorithm for differentiation between marine and terrestrially derived organic matter. Limnology and Oceanography 46, 2087–2093. Stedmon, C.A., Markager, S., 2003. Behavior of the optical properties of colored dissolved organic matter under conservative mixing. Estuarine, Coastal and Shelf Science 57, 973–979. Stedmon, C.A., Markager, S., Kaas, H., 2000. Optical properties and signatures of chromophoric dissolved organic matter (CDOM) in Danish coastal waters. Estuarine, Coastal and Shelf Science 51, 267–278. Stedmon, C.A., Markager, S., Tranvik, L., Kronberg, L., Sla¨tis, T., Martinsen, W., 2007. Photochemical production of ammonium and transformation of dissolved organic matter in the Baltic Sea. Marine Chemistry 104, 227–240. Twardowski, M.S., Boss, E., Sullivan, J.M., Donaghay, P.L., 2004. Modeling the spectral shape of absorption by chromophoric dissolved organic matter. Marine Chemistry 89, 69–88. Warnock, R.E., Gieskes, W.W.C., van Laar, S., 1999. Regional and seasonal differences in light absorption by yellow substance in the Southern Bight of the North Sea. Journal of Sea Research 42, 169–178. Zweifel, U.L., Wikner, J., Hagstro¨m, Å, Lundberg, E., Norrman, B., 1995. Dynamics of dissolved organic carbon in a coastal ecosystem. Limnology and Oceanography 40 (2), 299–305.
Contents lists available at ScienceDirect
Estuarine, Coastal and Shelf Science journal homepage: www.elsevier.com/locate/ecss
Tracing water mass mixing in the Baltic–North Sea transition zone using the optical properties of coloured dissolved organic matter Colin A. Stedmon a, *, Christopher L. Osburn b,1, Theis Kragh c a
Department of Marine Ecology, National Environmental Research Institute, University of Aarhus, Frederiksborgvej 399, Roskilde, 4000 Denmark US Naval Research Laboratory, Marine Biogeochemistry Section, Code 6114, Washington, DC, USA c Freshwater Biological Laboratory, University of Copenhagen, Helsingørsgade 51, DK-3400 Hillerød, Denmark b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 21 July 2009 Accepted 29 December 2009 Available online 18 January 2010
The distribution and characteristics of coloured dissolved organic matter (CDOM) in the Baltic – North Sea transition zone were studied. The aim was to assess the validity of predicting CDOM absorption in the region on the basis of water mass mixing alone and demonstrate the utility of CDOM as an indicator of water mass mixing in coastal seas. A three-end-member mixing model representing the three major allochthonous CDOM sources was sufficient to describe the patterns in CDOM absorption distribution observed. The three-end-member water masses were the: Baltic outflow, German Bight and the central North Sea. Previously, it was thought that water from the German Bight transported northwards in the Jutland coastal current only sporadically influenced mixing between the Baltic and North Sea. The results from this study show that water from the German Bight is detectable at salinities down to 12 in the Kattegat and Belt Sea. On average, 23% of the CDOM in bottom waters of the Kattegat, Great Belt, Belt Sea, Arkona Sea and the Sound originated from the German Bight. Using this conservative mixing model approach, local CDOM inputs were detectable but found to be limited, representing only 0.25% of CDOM in the surface waters of the Kattegat and Belt Sea. The conservative mixing of CDOM makes it possible to predict its distribution and characteristics and offers a powerful tool for tracing water mass mixing in the region. The results also emphasize the need to include the Jutland Coastal current in hydrodynamic models for the region. Ó 2010 Elsevier Ltd. All rights reserved.
Keywords: baltic Sea north Sea german Bight water mass mixing tracers dissolved organic matter
1. Introduction Coloured dissolved organic matter (CDOM) is one of the major light-attenuating components of sea water. It is responsible for much of the ultraviolet light attenuation and, when present in high concentrations, also influences visible light attenuation. The CDOM UV–visible absorption spectrum (300–650 nm) can be modelled using an exponential relationship (Jerlov, 1968; Lundgren, 1976; Bricaud et al., 1981).
aðlÞ ¼ aðl0 Þ eSðl0 lÞ
(1)
where l0 is a reference wavelength l and S is the spectral slope coefficient, characterizing the exponential decline in absorption with increasing wavelength. The absorption at a specific
* Corresponding author. E-mail address: [email protected] (C.A. Stedmon). 1 Present address: Department of Marine, Earth, and Atmospheric Sciences, North Carolina State University, Raleigh, NC, USA. 0272-7714/$ – see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.ecss.2009.12.022
wavelength (e.g. 300 nm (a300)) can be used as an estimate for CDOM concentration. CDOM originates primarily from the degradation of terrestrial and aquatic plant matter. In regions influenced by freshwater run off, terrestrial CDOM often dominates and it is only when this source is considerably diluted that the presence of autochthonous marine CDOM can be detected (Blough et al., 1993; Stedmon et al., 2000; Stedmon and Markager, 2003). When allochthonous supply and mixing rates exceed autochthonous production and degradation rates, the concentration and characteristics of CDOM behave conservatively. There is a considerable allochthonous supply of CDOM in the coastal waters of the North Sea and Baltic Sea and this led early studies to conclude that CDOM represents purely terrestrial material (Højerslev, 1979). Measurements from oceanic regions far from the influence of rivers, however, have since revealed a lower intensity, pelagic, autochthonous CDOM source (Jerlov, 1976; Bricaud et al., 1981; Nelson et al., 1998). In sub-surface open ocean waters, temperature and salinity are used to trace water mass mixing, since both parameters behave conservatively. In coastal or surface waters temperature is not a conservative parameter as it varies seasonally as a result of heat
C.A. Stedmon et al. / Estuarine, Coastal and Shelf Science 87 (2010) 156–162
exchange with the atmosphere. So for these waters alternative parameters are required. In combination with salinity, CDOM absorption at specific wavelengths can be used instead of temperature to trace water mass mixing in coastal waters (Kalle, 1949; Højerslev, 1988; Karabashev et al., 1993). In the Skagerrak and Kattegat, the brackish Baltic Sea water with high CDOM absorption mixes with saline North Sea water with low CDOM absorption (Fig. 1). The Jutland coastal current, flowing north along the western coast of Jutland from the German Bight, also influences the Skagerrak and Kattegat (Højerslev et al., 1996; Warnock et al., 1999; Nielsen, 2000). The current carries water from the rivers flowing into the Southern North Sea, which contain high CDOM concentrations (Laane and Kramer, 1990; Højerslev et al., 1996; Warnock et al., 1999). However, the effect of the German Bight water in the Skagerrak and Kattegat is thought to be limited in comparison to the effect of the large volumes of water transported from the North Sea and Baltic Sea (Aarup et al., 1996; Nielsen, 2000). It has been shown that salinity–CDOM relationships can characterize and quantify the mixing of these three water masses (Malmberg, 1964; Aarup et al., 1996; Højerslev et al., 1996). Despite this seemingly conservative behaviour, Højerslev and Aas (2001) did not find systematic trends in S. Stedmon and Markager (2003) estimated the variability of S during conservative mixing of two end members, facilitating analysis of field data and the identification of nonconservative processes acting on CDOM, such as local autochthonous production (Kowalczuk et al., 2006; Guo et al., 2007). The aims
157
of this work are two fold. Firstly we aim to predict CDOM absorption of water masses in the region based on water mass mixing alone. We assume that mixing is the dominant process controlling both the distribution and character (S) of CDOM. If this is true, a three-end-member mixing model representing the three major allochthonous CDOM sources should be sufficient to describe CDOM absorption. If this assumption is invalid and autochthonous production and removal processes control the distribution and characteristics of CDOM, the model results will indicate the magnitude of these processes. Secondly, we aim to demonstrate the utility of CDOM as an indicator of water mass mixing in coastal seas. By applying simple mixing models water at different depths and locations can be fractionated into the relative proportion of the three dominant end member water masses. 2. Methods 2.1. Sampling and measurement Water samples were collected using the national water quality monitoring cruises on the R/V Gunnar Thorson. Samples were obtained from ten cruises (Table 1) in the transition zone between the North Sea and the Baltic Sea (Fig. 1). Data from February 1999 originates from Stedmon et al. (2000) and is included to expand the data analysis. At each station, samples were collected at 1 m and 5 m intervals from 5 to 30 m. For stations deeper than 30 m, samples were taken at 10 m intervals below 30 m and at 1 m over the sediment. A rosette with a Seabird CTD and 12 5 L Niskin water samplers was used and calibrated according to accredited procedures required by the national monitoring program. CDOM samples were gently filtered through pre-combusted GF/ F filters (approximate pore size 0.7 mm) using a syringe, into acidwashed 100 mL brown glass bottles and stored refrigerated in the dark. Storage time varied from one to two weeks, since the cruise lasted 5 days and all the optical characteristics of the samples were determined during the week directly following each cruise. Stedmon et al. (2000) investigated the effects of storage on CDOM absorption for sample from these waters and found little change for periods up to 27 days. CDOM absorbance was measured on a Shimadzu UV-2401PC spectrophotometer using MilliQ water as a reference, according to Stedmon et al. (2000). A 10 cm quartz cuvette was used for all samples and the absorption coefficient was calculated from absorbance (A) according to Equation (2).
aðlÞ ¼ 2:303AðlÞ =L
(2)
where L is the pathlength in meters (0.1 m). The spectral slope, S, of the absorption curve from 300 to 650 nm was obtained by fitting Equation (1) to the data using the
Table 1 Cruise name, sampling time, number of stations and samples, and location of cruise track. Inner Danish waters include Arkona Sea, Belt Sea, Sound and Kattegat (Fig. 1).
Fig. 1. (a) Map of study site indicating the different surface currents and regional seas. Also shown are the station locations and the coloured contours represent the average salinity at 5 m. (b) A diagram of generalized characteristics of the water column structure in the region. For a map of the bathymetry in the region, see Fig. 1 in Højerslev et al. (1996).
Cruise
period
#Stations
#Samples
Description
GT189
Feb.1999
74
324
GT191
Aug. 1999
72
136
GT198 GT237 GT238 GT239 GT240 GT242 GT243 GT244
Sept. 2000 Aug. 2006 Sept. 2006 Oct. 2006 Feb. 2007 Aug 2007 Sept 2007 Feb 2008
24 24 24 24 22 17 17 17
131 182 135 138 164 131 134 132
Inner Danish waters, Skagerrak and North Sea Inner Danish waters, Skagerrak and North Sea Inner Danish waters Inner Danish waters Inner Danish waters Inner Danish waters Inner Danish waters Inner Danish waters Inner Danish waters Inner Danish waters
158
C.A. Stedmon et al. / Estuarine, Coastal and Shelf Science 87 (2010) 156–162
non-linear regression technique of Stedmon et al. (2000). It is known than S values are dependent on the wavelength range over which they are derived (Del Castillo et al., 1999; Stedmon et al., 2000; Twardowski et al., 2004). Slopes of the absorption spectra in the UV-B region (<300 nm) have been used to provide chemical characteristics of DOM (e.g. Helms et al., 2008). Other studies, focused on bio-optics or remote sensing, have reported slopes derived from longer wavelengths relevant for the penetration of sun light. In this study, the 300–650 nm range is used as one of the aims is to be able to predict CDOM light absorption geographically and temporally in order to aid the development of algorithms for light penetration in both regional hydrodynamic models and remote sensing applications. In this study the absorption coefficient at 300 nm will be used to present the distribution in CDOM concentrations. The choice of wavelength is essentially arbitrary as with knowledge of S the absorption coefficient at one wavelength can be converted to another. However, absorption by CDOM decreases with increasing wavelength and in some of the samples from the North Sea values are below the detection limit for much of the spectrum. 300 nm was chosen as this is the wavelength with the highest signal (furthest from the detection limit) for all samples and it is still within the ecologically relevant wavelength range for studies of underwater light penetration (300–700 nm). 2.2. Water mass mixing analysis The analysis based on Højerslev et al. (1996) assumes conservative behaviour of salinity and CDOM absorption during mixing of three major water masses: North Sea (NS) water originating from the central North Sea and Atlantic, Baltic Sea (BS) water representing the water flowing out of the Baltic, and German Bight (GB) water flowing northwards along the west coast of Jutland from the German Bight to the Skagerrak (Fig. 1). The salinity (Salt) and CDOM absorption (a) in a given water sample from the region can be expected to follow Equations (3)–(5).
Salt ¼ fNS SaltNS þ fBS SaltBS þ fGB SaltGB
(3)
aðlÞ ¼ fBS aBSðlÞ þ fGB aGBðlÞ þ fNS aNSðlÞ
(4)
fBS þ fNS þ fGB ¼ 1
(5)
where f represents the fraction of each water mass in the sample. These three equations with three unknowns can then be solved. After determining suitable end member values for CDOM and salinity (see the Results section), the relative proportion of each water mass can be determined for a given sample from the region. 2.3. Effects of conservative mixing on CDOMs spectral characteristics CDOM absorption during conservative mixing were estimated according to the Stedmon and Markager (2003), expanded to incorporate three end members. Combining the spectral absorption properties of CDOM (Eq. (1)) with the conservative mixing equation from Højerslev et al. (1996) (Eq. (4)), we derive the following equation for CDOM absorption in a sample.
aðlÞ ¼ fBS aBSðl0 Þ eSBS ðl0 lÞ þ fGB aGBðl0 Þ eSGB ðl0 lÞ þ fNS aNSðl0 Þ eSNS ðl0 lÞ
(6)
This absorption spectrum can now be modelled using Equation (1) and a new theoretical, conservative mixing S value can be derived. The comparison of model-derived S values to the measured S values offers an additional test of the conservative CDOM mixing in the region. 3. Results The data indicate three end members and two mixing lines (Fig. 2a, b). The first line represents mixing of German Bight water salinity w 29) with North Sea water (a300 > 5 m1, (a300 w 0.05 m1, salinity > 35). The second mixing line represents mixing of Baltic Sea water (a300 w 5 m1, salinity < 8,) with intermediate water on the German Bight–North Sea mixing line (a300 w 2 m1, salinity w 32–33). The S values generally tend to
Fig. 2. (a) CDOM a300 versus salinity for all data (Table 1), (b) average distribution of a300 at 5 m depth in the region, (c) CDOM spectral slope, S, against salinity.
C.A. Stedmon et al. / Estuarine, Coastal and Shelf Science 87 (2010) 156–162
decrease with increasing salinity in an approximately linear fashion; from 20 mm1 at salinities below 8, to 18 mm1 at salinities above 32 (Fig. 2c). There was a greater variability in S values at high salinities ranging from about 12 to 25 mm1. In order to determine the characteristics of the three end members, averages of samples at each extreme were calculated (Table 2, Fig. 2). For the Baltic Sea end member, data at salinities <8 were averaged; for the North Sea end member, data at salinities >35 and a300 < 0.5 m1 were averaged; and for the German Bight, averages included data where both the salinity >28 and a300 > 5 m1. Averages were used in order to obtain the most representative values for each and to reduce the sensitivity of the mixing calculations to end members input values. These values were then used in Equations (3)–(6) and the relative proportions (f) of each water mass were calculated for each sample. Under ideal conditions with, for example, fixed precise values of end member characteristics, no analytical error, no instrumental noise, nor autochthonous processes, the values for f would only vary between 0 and 1. For the majority of the samples in this data set f varied in this range, which is encouraging (Fig. 3). Values of f greater than 1 and below 0 are largely constrained to samples at the extremes, representing 100% of each of the three end members and this is purely an artefact of the mean values were used (Table 2). We found three mixing regimes (Fig. 3): at salinities below 20, the Baltic Sea CDOM end member dominates, representing over 50% of the total, and the remaining fraction is relatively evenly split between German Bight and North Sea water; at salinities above 32, the influence of the Baltic Sea CDOM is minimal and mixing between North Sea and German Bight water dominates; at intermediate salinities (20–26) the three water masses generally contribute equally. The gradual dilution of the Baltic outflow is clear (Fig. 4a, d) and the mixing of the North Sea waters with the German Bight along the Jutland coast and with Baltic water in the Kattegat is apparent (Fig. 4b, e). Water (and CDOM) from the German Bight can be traced in surface waters as far as the Belt Sea and Sound between Sweden and Denmark (Fig. 4c). In bottom waters German Bight water is present at slightly higher levels and clearly restricted from continuing into the Baltic Sea by the two shallow sills (Darss w 12 E, 54.5 N; Drogden w 12.5 E, 55.5 N) (Fig. 4f). The inner Danish waters (Southern Kattegat, Great Belt, Belt Sea, Arkona Sea, and the Sound) with salinities between 14 and 30 contained on average 23% German Bight CDOM (Fig. 3). Using the results of the water mass mixing calculations and the end member characteristics of the three CDOM pools, estimates for S values can also be calculated. These spectra were derived for each sample and then modelled using Equation (1) to determine conservative mixing values for S. The fit to the a300 values generally was very good; especially at the two salinity extremes (Fig. 5a). However, there was a systematic increase in the unexplained a300 between salinities of 15 and 20 towards a maximum where model estimates were approximately 0.01 m1 lower than measured values (Fig. 5a). This corresponded to the waters of the Kattegat and Belt Sea (Fig. 5b). The deviation decreased approximately linearly as salinities increased further from 24. In contrast, no overall systematic trends for the S value estimates were apparent (Fig. 5c). The conservative model appears to
159
describe much of the linear decrease in S values seen in Fig. 2c. However, the difference between the measured and modelled S values, increases slightly at higher salinities. This scatter corresponded to samples with low CDOM concentrations (a300 < 1 m1). One systematic deviation from the model was apparent. This was for samples from the Skagerrak deep and intermediate waters with salinities above 34.5 (Fig. 5c). Many of these samples had lower S values than predicted by the mixing model. 4. Discussion The major assumptions in this approach are (1) that there are only three dominant CDOM pools in these offshore waters, originating from the North Sea, Baltic Sea and the German Bight, (2) CDOM absorption properties are mixing conservatively, and (3) there are small variations in the properties of CDOM of the end members during the sampling period. The ability of this simple model in describing the large scale gradients in CDOM distribution and spectral characteristics (S) of CDOM in these waters suggests that these assumptions were to a certain extent valid. The presence of three dominant end members is clearly apparent (Fig. 2a and b). For a300 there are two general conservative mixing lines and it is clear that the saline marine end member for the waters of the Baltic Sea consists of a mixture of German Bight water and North Sea water. Although there is a degree of scatter along the Baltic Sea mixing line in Fig. 2a, the overall trend along the mixing lines is linear suggesting that CDOM is mixing on the whole conservatively. The fact that all the data from ten different cruises fall along the same general mixing lines which are similar to that also seen in earlier work (Højerslev et al., 1996; Stedmon et al., 2000) indicates that there are no large systematic changes in the concentrations and characteristics of CDOM in the three identified end members during the study period (1999–2008). Using the S values reported in Table 2 the a300 values can be transformed to approximate a380 values and compared with the end member characteristics reported by Højerslev et al. (1996). Our study’s a380 values for the Baltic Sea and German Bight end members were slightly lower (by 0.15 m1 and 0.30 m1 respectively) and the North Sea end member was higher by 0.05 m1. However, when compared to the degree of scatter in the Højerslev et al. (1996) data (see their Fig. 4) and the different techniques
Table 2 CDOM and salinity characteristics of each end member. Values are means and standard deviations are shown in brackets. End members
n
a300 (m1)
S (mm1)
Salinity
Baltic Sea German Bight North Sea
48 10 28
4.89 (0.19) 5.23 (0.19) 0.60 (0.04)
22.5 (0.8) 18.3 (0.5) 20.2 (1.9)
7.7 (0.20) 29.7 (0.24) 35.1 (0.06)
Fig. 3. Fraction of the three water masses characterized in Table 2 in each sample in the data set derived using the approach described by Højerslev et al. (1996).
160
C.A. Stedmon et al. / Estuarine, Coastal and Shelf Science 87 (2010) 156–162
Fig. 4. The average distribution at 5 m depth (a–c) and 5 m from the bottom (d–f) of the fractions of Baltic Sea water (top panels), North Sea water (middle panels), and German Bight water (lower panels).
(earlier study included CDOM data estimated from in situ beam attenuation measurements) these differences are relatively minor. The modelling results show that the approximately linear decreasing trend in S with increasing salinity (Fig. 2c) can be replicated by the conservative mixing approximations of S (Fig. 5c). This approximate linear behaviour in S is uncommon (Blough and Del Vecchio, 2002) and is here shown to be due to the fact that there are three major CDOM pools with differing concentrations undergoing a complex mixing pattern (Fig. 3). The close match of the conservative mixing estimates with the measured S values indicates that on the whole, CDOMs spectral properties in these waters can be modelled relatively easily. It is likely that some of the scatter in the trends observed is due to local non-conservative processes such as release of DOM from sediments (Skoog et al., 1996), seasonal autochthonous DOM production (Zweifel et al., 1995) and photochemical degradation of CDOM (Stedmon et al., 2007). These processes are important at local scales. For example, at some stations in the data set lower a300
values can be found at 1 m compared to 5 m despite the samples having the same salinity (data not shown). This could be evidence for photobleaching occurring in surface waters. In addition the sample from 1 m above the sediment at times had a slightly higher a300 than the sample above it (data not shown). This was interpreted as a result of release of CDOM from the sediments (Skoog et al., 1996). Autochthonous CDOM production is these waters is expected to be difficult to detect due to the high background concentrations of CDOM from the German Bight and Baltic Sea. In the high salinity waters from the North Sea, which have low background concentrations of CDOM it is, however, detectable, with fresh CDOM having a low S values (Stedmon and Markager, 2001). This can be seen at salinities of 35 in Fig. 5c. However, despite the occurrence of these non-conservative processes their overall impact on the distribution of CDOM and its spectral properties appears to be limited in comparison to the gradients resulting from water mass mixing in the region over timescales of several months. The combination of continual re-supply and high
C.A. Stedmon et al. / Estuarine, Coastal and Shelf Science 87 (2010) 156–162
161
Fig. 5. (a) Difference between the measured and modelled a300 plotted against salinity, (b) the average difference in a300 at 5 m depth in the region, (c) the difference between the measured and modelled S values plotted against salinity.
background concentrations of CDOM are a likely reason for these results. Bowers and Brett (2008) recently demonstrated with a box model that for an estuarine system such as this with relatively long residence time compared to the variability in the source (river water), a time–mean mixing line with salinity would be expected. The large volumes of water involved essentially dampen out the potential seasonal signal in source waters resulting in a constant mixing line. This appears to be what is essentially occurring here. The seasonal changes in the German Bight and Baltic Sea end members are dampened out and a time–mean mixing line exists. Extrapolating this result further and assuming no long term change in the values of the end member characteristics, the results here demonstrate how CDOM distribution and spectral characteristics in the region can be easily predicted (modelled) from knowledge of boundary conditions (end members) alone. This result implies that the regional distribution of CDOM can be easily modelled and implemented in 3D-hydrodynamic models such as COHERENS (Bendtsen et al., 2009), providing estimates on spectral light attenuation for ecosystem models and a novel remote sensing ground-truthing tool. The positive deviations in a300 from the conservative mixing model at salinities 15–20 suggest local freshwater inputs of CDOM into the waters of the Kattegat and Belt Sea (Fig. 5a and b). Without the use of this mixing model approach, these inputs would be difficult to identify. Indeed, earlier studies have considered them negligible due to the relatively low volume of water in comparison to the Baltic outflow (Aarup, 1994; Højerslev et al., 1996). The a300 values measured for these salinities were approximately 4 m1, so the modelled difference of 0.01 m1 represents only 0.25% of the CDOM values and is within the detection limits of the spectrophotometric method (Stedmon and Markager, 2001). However, there is a systematic trend with salinity in the deviations between the modelled and measured a300 values. Maximum deviations are found at a salinity of w20 and there is a clear linear decrease with increasing salinity. This pattern is indicative of conservative mixing of local estuarine inputs of CDOM into the surface waters flowing out of the Baltic. In contrast to the results for a300, the impact of local CDOM inputs at intermediate salinities on S values was not apparent. There are two potential reasons, which are not mutually exclusive. First, local inputs had very low intensities in comparison to the
three major CDOM pools and therefore have little impact on S values (Stedmon and Markager, 2003). Second, CDOM exported from fjords in the region has very similar S values to values found in the Kattegat (Stedmon et al., 2000). The estimates of the relative amounts of CDOM from different sources are intriguing and warrant further study. It is clear that CDOM from the German Bight can be traced into the Kattegat and Belt Sea waters down to a salinity of 12. Despite the sporadic and relatively small inflow from the Jutland coastal current, its impact can still be detected due to its high CDOM concentrations (Table 2). The estimates for the amount of German Bight water in the Kattegat and Belt Sea waters (w23%) agrees well with an earlier radioisotope study, which concluded that Kattegat bottom water contains 28% German Bight water (Dahlgaard et al., 1995). With the modelling approach demonstrated here, the relative amounts of each water mass in any CDOM water sample from these waters can be determined using relatively economical and simple techniques. This can have several benefits for estimating import and export of carbon and nutrients to and from the Baltic Sea. The Jutland Coastal current is controlled by wind, tidal currents and density-driven currents that all contribute to its northwards flow along the North Sea coast of Denmark (Nielsen, 2000). This current is primarily diluted in the southern Skagerrak and joins the Norwegian coastal current in the northern Skagerrak (Fig. 1). It is seldom observed entering the Northern Kattegat. These observations are based primarily on salinity and nutrient analyses, which is not an optimal approach due to the non-conservative nature of nutrient concentrations. By the time the current reaches the Skagerrak–Kattegat boundary, the salinity differences are small and nutrient concentrations, which often are high in the current, do not behave conservatively during the relevant mixing times. The current has little effect on the overall salt transport into the Baltic Sea and therefore is often neglected (Rydberg, 1993) or de-emphasized in hydrodynamic models for the region. The results presented here clearly show that the marine end member for the Baltic Sea (i.e. water entering the northern Kattegat from the North Sea) contains a noteworthy amount of water (and CDOM) from the German Bight and this may have implications for the transport of other constituents such as nutrients and organic matter in general. The patterns seen in CDOM distribution and characteristics cannot be explained without the inclusion of a third water mass with high CDOM concentrations, here found to be water from the German Bight.
162
C.A. Stedmon et al. / Estuarine, Coastal and Shelf Science 87 (2010) 156–162
5. Conclusions The major findings of this study are: 1. The three-end-member water mass mixing model by Højerslev et al. (1996) based on the distribution of CDOM in these waters, has been validated with a more recent, independent and comprehensive data set spanning ten years, and expanded to include spectral data. This approach can explain the trends observed in S values in these waters and paves the way for predicting the spectral light absorption by CDOM using measurements from boundary regions alone. 2. Waters from the German Bight influence the Kattegat and Skagerrak area more than previously perceived. 3. The combination of CDOM measurements and a three-endmember water mass mixing model can be used to resolve the distribution of local freshwater inputs from Germany, Denmark and Sweden into the Kattegat and Belt Sea region. The fact that much of the qualitative and quantitative variability in CDOM absorption properties can be easily modelled is advantageous for phytoplankton remote sensing applications in the region where the signal from CDOM has to be considered (Karabashev, 1992). Additionally, the knowledge gained here can be easily assimilated into ecosystem models in order to improve the estimation of photic depth and the prediction of primary productivity. These results also reemphasize the great potential of using optical properties of CDOM to trace water mass mixing, which to date has not been fully exploited by physical oceanographers. The approach provides a method to resolve the influence of smaller currents and water masses in the supply and transport of carbon and nutrients in coastal seas. Acknowledgements This work was funded by the US. Office of Naval Research (Grant: N00014-06-1-0357), the Danish Ministry of the Environment and the Danish Council for Strategic Research. Peter Kofoed is thanked for his assistance in collecting samples. The crew and laboratory staff of R/V Gunnar Thorson is also thanked for their cooperation. The contour plots were created using Ocean Data View (Schlitzer, R., http://odv.awi.de). References Aarup, T., 1994. Interpretation of satellite ocean color data of the transition zone between the North Sea and the Baltic Sea. Ph.D. thesis, Univ. of Copenhagen. Aarup, T., Holt, N., Højerslev, N.K., 1996. Optical measurements in the North Sea– Baltic Sea transition zone. II. Water mass classification along the Jutland west coast from salinity and spectral irradiance measurements. Continental Shelf Research 16, 1343–1353. Bendtsen, J., Gustafsson, K.E., So¨derkvist, J., Hansen, J.L.S., 2009. Ventilation of bottom water in the North Sea–Baltic Sea transition zone. Journal of Marine Systems 75, 138–149. Blough, N.V., Del Vecchio, R., 2002. Chromophoric DOM in the coastal environment. In: Hansell, D., Carlson, C. (Eds.), Biogeochemistry of Marine Dissolved Organic Matter. Academic Press, pp. 509–546. Blough, N.V., Zafiriou, O.C., Bonilla, J., 1993. Optical absorption spectra of waters from the Orinoco River outflow: terrestrial input of colored organic matter to the Caribbean. Journal of Geophysical Research 98, 2271–2278. Bowers, D., Brett, H.L., 2008. The relationship between CDOM and salinity in estuaries: an analytical and graphical solution. Journal of Marine Systems 73, 1–7.
Bricaud, A., Morel, A., Prieur, L., 1981. Absorption by dissolved organic matter of the sea (yellow substance) in the UV and visible domains. Limnology and Oceanography 26, 43–53. Dahlgaard, H., Herrmann, J., Salomon, J.C., 1995. A tracer study of the transport of coastal water from the English Channel through the German Bight to the Kattegat. Journal of Marine Systems 6, 415–425. Del Castillo, C.E., Coble, P.G., Morrel, J.M., Lopez, J.M., Corredor, J.E., 1999. Analysis of the optical properties of the Orinoco River plume by absorption and fluorescence spectroscopy. Marine Chemistry 66, 35–51. Guo, W., Stedmon, C.A., Han, Y., Wu, F., Yu, X., Hu, M., 2007. The conservative and non-conservative behavior of chromophoric dissolved organic matter in Chinese estuarine waters. Marine Chemistry 107, 357–366. Helms, J., Stubbins, A., Ritchie, J.D., Minor, E., Kieber, D.J., Mopper, K., 2008. Absorption spectral slopes and slope ratios as indicators of molecular weight, source, and photobleaching of chromophoric dissolved organic matter. Limnology and Oceanography 53 (3), 955–969. Højerslev, N.K., 1979. On the origin of yellow substance in the marine environment. In: 17th General Assembly of I.A.P.S.O. (Canberra, 1979), Abstracts, 71. Højerslev, N.K., 1988. Natural occurrences and optical effects of gelbstoff. Report No. 50. Department of Geophysics, University of Copenhagen. Højerslev, N.K., Aas, E., 2001. Spectral light absorption by yellow substance in the Kattegat–Skagerrak area. Oceanologia 43, 39–60. Højerslev, N.K., Holt, N., Aarup, T., 1996. Optical measurements in the North Sea– Baltic Sea transition zone. I. On the origin of deep water in the Kattegat. Continental Shelf Research 16, 1329–1342. Jerlov, N.G., 1968. Optical oceanography. In: Oceanography Series 5. Elsevier. Jerlov, N.G., 1976. Marine Optics. Elsevier. Kalle, K., 1949. Fluoreszenz und Gelbstoff im Bottnischen und Finnischen Meerbusen. Deutsche Hydrografische Zeitschrift 2, 96–105. Karabashev, G.S., 1992. On the influence of dissolved organic matter on remotesensing of chlorophyll in the Straits of Skagerrak and Kattegat. Oceanologica Acta 15, 255–259. Karabashev, G.S., Khanaev, S.A., Kuleshov, A.F., 1993. On the variability of ‘‘yellow substance’’ in the Skagerrak and the Kattegat. Oceanologica Acta 16, 115–125. Kowalczuk, P., Stedmon, C.A., Markager, S., 2006. Modeling absorption by CDOM in the Baltic Sea from season, salinity and chlorophyll. Marine Chemistry 101, 1–11. Laane, R.W.P.M., Kramer, K.J.M., 1990. Natural fluorescence in the North Sea and its major estuaries. Netherlands Journal of Sea Research 26, 1–9. Lundgren B., 1976. Spectral transmittance measurements in the Baltic. Rep. no. 30. Department of Physical Oceanography, University of Copenhagen. Malmberg, S., 1964. Transparency measurements in the Skagerrak. Go¨teborgs Kungliga Vetenskaps och Vitterhets-Samha¨lles Handlingar, Series B 9, 5–18. Nelson, N.B., Siegel, D.A., Michaels, A.F., 1998. Seasonal dynamics of colored dissolved material in the Sargasso Sea. Deep Sea Research 45, 931–957. Nielsen, M.H., 2000. A Dynamic Description of the Hydrographic Classification of the Jutland Coastal Current (in Danish). Danish Meteorological Institute. Scientific Report 00-15. Rydberg, L., 1993. On the Skagerrak circulation and the supply of water from the southern North Sea to the Skagerrak. Department of Oceanography, University of Gothenburg (Sweden), Report No. 53, 34pp. Skoog, A., Hall, P.O.J., Hulth, S., Paxeus, N., Rutgers van der Loeff, M., Westerlund, S., 1996. Early diagenetic production and sediment–water exchange of fluorescent dissolved organic matter in the coastal environment. Geochimica et Cosmochimica Acta 60 (19), 3619–3629. Stedmon, C.A., Markager, S., 2001. The optics of chromophoric dissolved organic matter (CDOM) in the Greenland Sea: an algorithm for differentiation between marine and terrestrially derived organic matter. Limnology and Oceanography 46, 2087–2093. Stedmon, C.A., Markager, S., 2003. Behavior of the optical properties of colored dissolved organic matter under conservative mixing. Estuarine, Coastal and Shelf Science 57, 973–979. Stedmon, C.A., Markager, S., Kaas, H., 2000. Optical properties and signatures of chromophoric dissolved organic matter (CDOM) in Danish coastal waters. Estuarine, Coastal and Shelf Science 51, 267–278. Stedmon, C.A., Markager, S., Tranvik, L., Kronberg, L., Sla¨tis, T., Martinsen, W., 2007. Photochemical production of ammonium and transformation of dissolved organic matter in the Baltic Sea. Marine Chemistry 104, 227–240. Twardowski, M.S., Boss, E., Sullivan, J.M., Donaghay, P.L., 2004. Modeling the spectral shape of absorption by chromophoric dissolved organic matter. Marine Chemistry 89, 69–88. Warnock, R.E., Gieskes, W.W.C., van Laar, S., 1999. Regional and seasonal differences in light absorption by yellow substance in the Southern Bight of the North Sea. Journal of Sea Research 42, 169–178. Zweifel, U.L., Wikner, J., Hagstro¨m, Å, Lundberg, E., Norrman, B., 1995. Dynamics of dissolved organic carbon in a coastal ecosystem. Limnology and Oceanography 40 (2), 299–305.