Environ. Sci. Technol. 2007, 41, 7273-7279
Characteristics of Dissolved Organic Matter in Baltic Coastal Sea Ice: Allochthonous or Autochthonous Origins? C O L I N A . S T E D M O N , * ,† DAVID N. THOMAS,‡ MATS GRANSKOG,§ H E R M A N N I K A A R T O K A L L I O , |,⊥ STATHYS PAPADIMITRIOU,‡ AND HARRI KUOSA# Department of Marine Ecology, National Environmental Research Institute, University of Aarhus, Frederiksborgvej 399, 4000 Roskilde, Denmark, Ocean Sciences, College of Natural Sciences, University of Wales-Bangor, Menai Bridge, Anglesey, LL59 5AB, United Kingdom, Arctic Centre, University of Lapland, POB 122, 96101 Rovaniemi, Finland, Finnish Institute of Marine Research, POB 2, 00561 Helsinki, Finland, Department of Biological and Environmental Sciences, POB 56, 00014 University of Helsinki, Finland, and Tva¨rminne Zoological Station, J.A. Palme´nin tie 260, 16900 Hanko, Finland
The origin of dissolved organic matter (DOM) within sea ice in coastal waters of the Baltic Sea was investigated using parallel factor (PARAFAC) analysis of DOM fluorescence. Sea ice DOM had distinctly different fluorescence characteristics than that of the underlying humic-rich waters and was dominated by protein-like fluorescence signals. PARAFAC analysis identified five fluorescent components, all of which were present in both sea ice and water. Three humic components were negatively correlated to salinity and concluded to be terrestrially derived material. Baltic Sea ice DOM was found to be a mixture of humic material from the underlying water column incorporated during ice formation and autochthonous material produced by organisms within the ice. Dissolved organic carbon (DOC) and nitrogen (DON) concentrations were correlated to the humic fluorescence, indicating that the majority of the organic carbon and nitrogen in Baltic Sea ice is bound in terrestrial humic material trapped within the ice. This has implications for our understanding of sea ice carbon cycling in regions influenced by riverine input (e.g., Baltic and Arctic coastal waters), as the susceptibility of DOM to degradation and remineralization is largely determined by its source.
Introduction When freshwater freezes, the majority of ions and other dissolved components contained within the water are expelled from the ice, whereas when saline waters freeze, * Corresponding author phone: +45 46301805; fax: +45 46301114; e-mail:
[email protected]. † University of Aarhus. ‡ University of Wales-Bangor. § University of Lapland. | Finnish Institute of Marine Research. ⊥ University of Helsinki. # Tva ¨ rminne Zoological Station. 10.1021/es071210f CCC: $37.00 Published on Web 10/06/2007
2007 American Chemical Society
about 10-40% of the dissolved inorganic and organic matter is retained within the ice matrix (1). Concentrated brines in channels and pores within the ice range in size from micrometers to centimeters in diameter (2, 3). The size and frequency of this interior brine channel system is dependent on the initial salinity of the seawater, growth conditions, as well as ice temperature (1). It is well-known that microbial activity within sea ice can be high, and a large diversity of viruses, bacteria, microalgae, protozoans, and small crustaceans can combine to form active biological assemblages (4-6). In conjunction, high concentrations of dissolved organic matter (DOM) have been reported from sea ice (7), although little is known about its composition, sources, dynamics, and fate. Ice DOM can either be concentrated into brines from the source water during ice formation (8) or be derived from the internal biological community primarily as a result of cell lysis and exudation. Typically, the concentration of dissolved organic carbon (DOC) and dissolved organic nitrogen (DON) in sea ice is higher than predicted from conservative behavior during ice formation, indicating that internal production is a significant source, although there is often no clear relationship between biological standing stock and DOM concentrations (9-11). Sea ice forms an important component of the winter ecosystem of large parts of the northern Baltic Sea. Despite the low salinity of these waters, the characteristics of Baltic Sea ice are more broadly similar to those of sea ice formed in the Polar Oceans than to those formed in freshwater systems (6). The DOM concentrations in Baltic Sea ice are lower than in the underlying waters (12), despite considerable microbial activity within the sea ice (11, 12). Similar trends were found in other sea ice systems where there is a high input of terrestrially derived organic matter (13). To date, there have been few attempts to characterize DOM within sea ice and assess the importance of autochthonous or allochthonous sources. In this study, the characteristics of DOM within and below coastal sea ice in the western Bothnian Sea were examined to determine its source. The overall aim was to investigate whether the composition of Baltic Sea ice DOM is influenced primarily by the underlying DOM in the water column or whether it is dominated by DOM released from biological activity within the ice. We applied fluorescence spectroscopy combined with parallel factor analysis (14, 15) to study the nature of the DOM within the ice and to compare this with the DOM in the underlying and source waters.
Materials and Methods Sampling was conducted on sea ice in the vicinity of the Umeå Marine Sciences Centre (63.578° N, 19.858° E) in the northwestern Gulf of Bothnia (Baltic Sea) on March 18, 2004 (see Supporting Information). Sampling stations started at the outlet of the river O ¨ rea¨lven (station 0) and reached to the outer edge of the land fast sea ice in the O ¨ refja¨rden (station 6) and to ice floes just outside the outer edge of the land fast ice (station 7). Water depth ranged from 3 m at the mouth of the river to about 30 m at station 7. The river plume from O ¨ rea¨lven extended out beneath the ice, which results from the sea ice hindering wind-driven mixing of surface waters, enabling the river water plume to extend further offshore than during ice-free periods (6). Ice cores were taken at stations 1, 3, and 5-7 with a MARK II corer (9 cm i.d.; Kovacs Enterprise). The cores ranged in length from 40 to 60 cm. After retrieval, the cores were immediately divided as follows: two 5 cm sections were cut from the bottom, and the remaining length was divided into VOL. 41, NO. 21, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7273
0.8-5.2 0.5-4.2 1.2-2.9 a
Number in parentheses represents number of samples averaged. Sea ice was only sampled from stations 1, 3, and 5-7, whereas water samples were taken at all stations (0-7).
2.1 1.4 2.0 12.0-21.0 15.1-15.7 18.0-22.0 0.5-3.5 17.0-22.4 3.0-5.8 43 607 318 ice (27) water 0 m (8) water 5 m (7)
0.12 0.16 3.59
0-0.5 0-1.0 3.2-3.9
2-115 522-655 272-380
3.1 13.4 8.6
0-12.4 11.3-15.6 5.5-10.3
32 46 39
4.2-174 40-59 31-56
1.4 21.1 4.14
15.3 15.5 19.2
range mean range mean mean
range
range
mean
range
mean
range
mean
mean
range
Chl (µg L-1)
S (µm-1) a350 (m-1) DOC/DON (mol C mol N-1) DOC (µmol C L-1)
DON (µmol N L-1) 9
salinity
TABLE 1. Average Characteristics of Ice and Seawater Samplesa 7274
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 21, 2007
FIGURE 1. Examples of EEMs from ice sample and water sample. Fluorescence is in Raman units.
FIGURE 2. Spectral properties of five fluorescent components identified by PARAFAC analysis. 10 cm sections, with the exception of the topmost sections, the length of which depended on the total ice thickness. The segments were cut with a stainless steel saw and were put into acid-washed buckets and sealed. Every effort was made to minimize brine drainage, and the elapsed time from the onset of coring until an individual section was segmented was less than 2 min and considerably less for the bottommost sections. The samples were then melted at 4 °C in the dark and were sub-sampled immediately for salinity, chlorophyll a, DOC, DON, and DOM optical measurements. Water samples were collected along the transect at the ice-water
FIGURE 3. DOM in sea ice as compared to water. Averages for fluorescence intensities of each component identified by the parallel factor analysis for ice samples, surface water samples, and water samples from 5 m. Standard deviations are represented by the error bars on the graphs. interface (referred to as 0 m) and 5 m depth using a Ruttner water sampler. Salinity was measured using a WTW LF196 salinometer (salinity is reported here as practical salinity units). Chlorophyll a concentrations were measured in ethanol after 24 h extraction in the dark, using a Jasco FP750 fluorometer, calibrated with pure chlorophyll a (Sigma). The samples for DOC, DON, and DOM were all filtered through syringe filters (Whatman GD/X GMF (borosilicate glass microfiber), pore size 0.45 µm). The DON samples were frozen at -20 °C prior to analysis, whereas the DOC samples were acidified with approximately 10 µL of 85% H3PO4, before storage at -20 °C in pre-combusted (550 °C for 3 h) 4 mL glass vials sealed with Teflon-lined septa. Samples for DOM fluorescence and absorption measurements were stored in acid-washed 100 mL glass bottles at 4 °C. DON concentrations were determined by subtraction of nitrate, nitrite, and ammonium from the total dissolved nitrogen (TDN) analyzed by FIA on a LACHAT autoanalyzer using on-line peroxodisulphate oxidation coupled with UV radiation at pH 9.0 and 100 °C (16), with a precision better than 5%. The DOC concentrations were determined by hightemperature combustion on a MQ 1001 TOC Analyzer (17). Tested daily on the certified reference material of deep (700 m) Florida Strait seawater (D. A. Hansell, University of Miami, Consensus Reference Materials-Deep Seawater Reference, lot 05--05: 46-47 µmol C L-1), the method yielded 46 ( 3 µmol C L-1 (n ) 53), while the precision was better than 10%. One sample from the sea ice was identified repeatedly as an outlier in the subsequent data analyses and was therefore removed from the calculations. Its DOC concentration was above 160 µmol C L-1, whereas the rest of the ice samples had a DOC concentration of less than 120 µmol C L-1. The absorption characteristics of DOM between 240 and 700 nm were measured on a Shimadzu UV-2401PC spectrophotometer using either a 1 cm (water samples) or a 10 cm (ice samples) quartz cuvette and Milli-Q water as a reference and blank. The absorption coefficients (a, m-1) were calculated by multiplying the optical density by 2.303/ L, where L is the cuvette path length in meters. Absorption spectra between 300 and 650 nm were modeled by eq 1, using a nonlinear technique published earlier (18)
A(λ) ) A(λ0)e-S(λ0-λ) + K
(1)
The coefficient S is the spectral slope coefficient and describes the exponential decrease in absorption with increasing wavelength. λ0 is 400 nm, and K is a background constant that allows for any baseline shifts. The absorption at 350 nm (a350) was used to quantify the colored DOM
present, and its spectral slope coefficient (S) was used to characterize the absorption spectrum. Fluorescence measurements were made using a Varian Eclipse fluorescence spectrophotometer. A range of emission spectra spanning 300-600 nm (every 2 nm) was recorded while exciting at wavelengths from 250 to 450 nm (by 5 nm). The excitation and emission slit widths were 5 nm. The fluorescence spectra measured from each sample were combined to create excitation emission matrices (EEMs). EEMs were corrected for instrument specific spectral bias by deriving correction spectra using Rhodamine B (excitation correction) and a ground quartz diffuser (emission correction). Sample inner filter effects were also corrected for using absorption measurements (19-21). The correction procedure is detailed in the Supporting Information. The EEMs were Raman calibrated by normalizing to the integral of the Raman band (excitation 350 nm) from a Milli-Q water sample run the same day as the samples (22). The EEMs were modeled by parallel factor analysis (PARAFAC) in MATLAB software using the N-way toolbox for MATLAB (22-24). In short, PARAFAC is an alternating least-squares regression that splits the EEMs into a series of tri-linear terms and a residual array (eq 2) F
xijk )
∑a b c
if jf kf
+ ijk, i ) 1,..,I; j ) 1,..,J; k ) 1,..,K
f)1
(2)
In this application, xijk represents the fluorescence intensity for sample i at emission wavelength j and excitation wavelength k. aif is directly proportional to the concentration of analyte f in sample i. bjf is proportional to the quantum efficiency of the fluorescence of analyte f at emission wavelength j, and ckf is proportional to the specific absorption coefficient at excitation wavelength k. F is the number of components (factors) in the model, and ijk is the residual matrix representing the unexplained signal. This approach separates the signal from a complex mixture into its underlying components, with no assumptions on the number of components or their spectral shape. A total of 41 EEMs was collected and modeled in this study. The data array consisted of 41 samples by 151 emission wavelengths by 41 excitation wavelengths. The data were split into two halves, and the PARAFAC algorithm was run stepwise on both halves, for three to 10 components. The appropriate number of components was determined by a split-half validation and residual analysis (14, 22, 23). A five component model was found adequate to describe the data (see Supporting Information). VOL. 41, NO. 21, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7275
FIGURE 4. Profiles of (A) chlorophyll a concentrations and (B) component 2 and (C) component 4 fluorescence as examples of humic- and protein-like fluorescence signals, respectively. Only data from stations 3-7 are shown as these were the only ones with complete core profiles. Data plotted at mid-depths of sampled ice segments.
Results General Characteristics of Aquatic and Sea Ice Samples. Noticeable differences were apparent between the sea ice, the surface water, and the water at 5 m depth (Table 1). The surface seawater had a very low salinity (range 0.0-1.0), whereas the water at 5 m depth was more saline (3.5-3.9). For all surface water samples apart from station 7, the salinity was below 1. The overall chlorophyll a concentration range was 0.4-5.2 µg L-1, with no clear differences apparent 7276
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 21, 2007
between the different sample types (Table 1). The DOC and DON concentrations were on average highest in the surface river plume water (607 µmol C L-1 and 13.4 µmol N L-1) and lowest in the sea ice (47 µmol C L-1 and 3.1 µmol N L-1). No trend was apparent in the DOC/DON ratios among the samples; however, the ratios were variable in the sea ice (Table 1). The a350 values of DOM followed the pattern seen for DOC and DON concentrations (Table 1) with the highest values measured in the surface waters influenced by riverine runoff. The S values in the surface waters were remarkably constant at 15.5 µm-1. The S values for the more saline 5 m samples were on average distinctly higher. The sea ice DOM S values were variable but with an average value similar to that of the surface water DOM (Table 1). Fluorescence Characteristics of DOM. Figure 1 illustrates an example of the measured EEMs for an ice sample and a water sample. The fluorescent properties of DOM differed greatly between the sea ice and the water samples. Figure 2 shows the spectral properties of the five individual components identified. The first three components had spectra resembling humic-like material with broad emission at wavelengths greater than 400 m. The two remaining components had spectra that resembled that of the amino acid tryptophan associated with proteins. The spectral properties of the components derived were very similar to those found in earlier studies (see Supporting Information). Components 1-3 are largely thought to be derivatives of terrestrial organic material (15, 25, 26). Components 4 and 5 are similar to components previously identified in coastal and oceanic waters (15, 26) and are thought to be plankton-derived. The contribution of the individual components to the total fluorescence differed notably between the sample types. The sea ice DOM fluorescence was on average dominated by the protein-like fluorescence of component 4, while the humic fluorescence (components 1-3) was very low (Figure 3). In contrast, the surface water DOM fluorescence was dominated by the humic components, while the proteinlike signals (components 4 and 5) were relatively minor. The relative contribution of the dominant humic components differed between the surface and 5 m water samples with the surface samples having a greater proportion of component 1. In general, the protein fluorescence was similar for all water samples (i.e., a low contribution) except for the water samples from 5 m depth at stations 5 and 6, which had very high fluorescence of component 4. These samples are responsible for the large standard deviations (error bars) calculated in Figure 3 for the 5 m samples. Ice Profiles. No consistent vertical patterns across stations were apparent for the majority of the parameters measured in the sea ice. Algal biomass in the sea ice, represented by chlorophyll a concentrations, was variable (Table 1). At stations 5 and 7, the greatest chlorophyll values were measured in the upper layers of the ice column. For stations 3 and 6, however, maximum concentrations were measured at the bottom of the ice column, a few centimeters from the ice-water interface (Figure 4A). The profiles of the fluorescent components can be grouped into two distinct clusters. The humic components had a relatively low fluorescence and showed no vertical trend in the ice column (Figure 4B). The protein-like fluorescence of component 4, however, was greater in the surface and bottom layers of the ice column (Figure 4C). Relationships between Physical, Optical, and Chemical Measurements. Riverine input is often a major source of DOM in coastal waters, and therefore, the data were analyzed for relationships with salinity. In addition, the relationship between fluorescence characteristics with DOC and DON concentrations was investigated. A linear regression analysis was applied using the general linear models (GLM) procedure
FIGURE 5. Relationships between (A) DOC, (B) DON, (C) component 3, and (D) component 5 fluorescence and (E) A350 with salinity. Where significant linear relationships could be derived, regression lines and equations are shown.
in the SAS/STAT software (v. 9.1.3, SAS Institute). Figure 5 shows DOC, humic fluorescence (component 3 shown as an example), and a350 as a function of salinity. There was a significant (P < 0.01) negative relationship with salinity. Significant relationships were also found for the other humic fluorescence signals (components 1 and 2). The sea ice DOM fluorescence and a350 were not significantly (P > 0.05) linearly correlated to salinity. No correlations to salinity were apparent for DON or the protein fluorescence signals (Figure 5). A strong linear relationship was apparent between the humic fluorescence of component 3 and the DOC concentrations for all the samples (Figure 6A). The intercept was not significantly different from zero and was therefore set to zero. Relationships between the other humic components and DOC were significant (P < 0.01) but not appropriately modeled by a linear regression. Components 4 and 5 were not correlated to DOC concentrations. A similar pattern was observed with the correlations between fluorescence and DON concentrations, with component 3 having the strongest significant correlation (P < 0.01, Figure 6B).
Discussion The concentrations of DOC and DON were lower in the ice than in the underlying water, similar to that reported in limnic systems (27). This contrasts with the conditions often reported in polar sea ice where DOC concentrations are an order of magnitude greater than in seawater (10). Experimental studies on the behavior of dissolved compounds during sea ice formation have shown that some of the seawater DOM is incorporated into sea ice (8). Similarly, freshwater studies have shown that a fraction of water DOM remains in the ice (27). These studies suggest that a fraction of sea ice DOM originates from the parent water mass. The fluorescence properties of sea ice and water DOM in this study suggest that, in Baltic Sea ice, the DOM represents a mixture of material from the water column incorporated during ice formation and autochthonous material produced by organisms within the ice. Specifically, the DOM fluorescence signals suggest that it is dominated by autochthonous material represented by protein-like fluorescence signals at waveVOL. 41, NO. 21, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7277
FIGURE 6. Plot of relationships derived between fluorescence of component 3 and te (A) DOC and (B) DON concentrations in all samples. lengths below 400 nm. This is similar to observations in oceanic waters far away from terrestrial sources (28-30). Additionally, the lack of correlation of the protein-like fluorescence signals to salinity in the water samples also supports this, implying that this material is not terrestrially derived but produced within the ice. In contrast, the water samples were characterized by broad fluorescence peaks at wavelengths greater than 400 nm, typical of freshwater and estuarine DOM (30). The PARAFAC analysis indicated that a background pool of terrestrially derived humic material is also present in the ice, most likely retained during ice formation. The relationships between salinity and humic fluorescence signals (components 1-3) confirm its proposed terrestrial origin. Despite the fact that the protein-like fluorescent components (4 and 5) dominated the fluorescence characteristics of the sea ice DOM, the DOC and DON concentrations were not correlated to them. The low background levels of humic fluorescence, possibly representing seawater DOM trapped in or transported into the ice, can explain nearly all the variability in DOC concentrations and a large part of the variability in DON concentrations as illustrated in Figure 6. An approximation of the C/N stoichiometry of this material was obtained by iteratively fitting a combined model for DOC and DON with component 3’s C/N ratio as a model parameter (see Supporting Information). This yielded a C/N ratio of 55 (standard error 5.4 and 95% confidence interval 43-66), which is as would be expected for humic or, generally, terrestrially derived organic matter. This study has demonstrated that the DOM pool in Baltic Sea ice consists of a mixture of material derived from biological activity within the ice and terrestrial material from the source waters from which the ice is formed. This explains why no robust correlations have been found earlier between DOC and DON concentrations and biomass or activity of sea ice communities (11), despite a clear fluorescence signature indicating the presence of autochthonous DOM production. This most likely indicates that the sea ice microbial community is rapidly cycling a labile autochthonous sub-fraction of DOM, which is difficult to detect using chemical measurements due to its rapid uptake. The fact that much of the organic carbon and nitrogen is bound in terrestrial humic material has implications for our understanding of sea ice carbon cycling in the Baltic Sea and other regions influenced by riverine input such as the Arctic (31), as the source of the material largely determines its chemical characteristics and thereby environmental reactivity. The application of fluorescence spectroscopy to studying sea ice DOM appears to be a promising, rapid, and inexpensive approach, indicating the different fractions of organic material present. Future work should focus on how the characteristics of the DOM 7278
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 21, 2007
pool change from ice formation to melt. This, in combination with traditional measurements, will allow us to differentiate between both refractory and labile pools as well as allochthonous and autochthonous contributions. Then, it will be possible to quantify the importance of humic material in sea ice as a nutrient and energy source and evaluate how rapidly the autochthonous material is cycling.
Acknowledgments This work was supported by the Walter and Andre´e de Nottbeck Foundation, Umeå Marine Sciences Centre, and Danish Research Council Grant 274-05-0064. Winnie Martinsen, Matthias Steffens, and Eloni Sonninen are thanked for their assistance with laboratory measurements. The Umeå Marine Sciences Centre is gratefully acknowledged for use of laboratory facilities and nutrient analysis. We also thank four reviewers for their comments on the manuscript.
Supporting Information Available S1, map of sampling location; S2, inner filter correction procedure; S3, examples of the measured, modeled, and residual EEMs; S4, comparison of spectral properties of components to those published in earlier studies; and S5, approach for determination of C/N ratio of component 3 and its error. This material is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited (1) Eicken, H. From the microscopic, to the macroscopic, to the regional scale: Growth, microstructure, and properties of sea ice. In Sea Ice: An Introduction to Its Physics, Chemistry, Biology, and Geology; Thomas, D. N., Dieckmann, G. S., Eds.; Blackwell Science: Oxford, 2003. (2) Eicken, H. The role of sea ice in structuring Antarctic ecosystems. Polar Biol. 1992, 12, 3-13. (3) Weissenberger, J.; Dieckmann, G. S.; Gradinger, R.; Spindler, M. Sea ice: A cast technique to examine and analyze brine pockets and channel structure. Limnol. Oceanogr. 1992, 37, 179183. (4) Thomas, D. N.; Dieckmann, G. S. Antarctic sea icesA habitat for extremophiles. Science (Washington, DC, U.S.) 2002, 295, 641-644. (5) Mock, T.; Thomas, D.N. Sea icesRecent advances in microbial studies. Environ. Microbiol. 2005, 7, 605-619. (6) Granskog, M. A.; Kaartokallio, H.; Kuosa, H.; Thomas, D. N.; Vainio, J. Sea ice in the Baltic SeasA review. Estuarine, Coastal Shelf Sci. 2006, 70, 145-160. (7) Thomas, D. N.; Papadimitriou, S. Biogeochemistry of sea ice. In Sea Ice: An Introduction to Its Physics, Chemistry, Biology, and Geology; Thomas, D. N., Dieckmann, G. S., Eds.; Blackwell Science: Oxford, 2003. (8) Giannelli, V.; Thomas, D. N.; Haas, C.; Kattner, G.; Kennedy, H. A.; Dieckmann, G. S. Behavior of dissolved organic matter and
(9)
(10) (11)
(12)
(13)
(14)
(15)
(16) (17) (18)
(19)
inorganic nutrients during experimental sea ice formation. Ann. Glaciol. 2001, 33, 317-321. Thomas, D. N.; Lara, R. J.; Eicken, H.; Kattner, G.; Skoog, A. Dissolved organic matter in Arctic multi-year sea ice during winter: Major components and relationship to ice characteristics. Polar Biol. 1995, 15, 477-483. Thomas, D. N.; Engbrodt, R.; Gianelli, V.; Kattner, G.; Kennedy, H.; Haas, C.; Dieckmann, G. S. Dissolved organic matter in Antarctic sea ice. Ann. Glaciol. 2001, 33, 297-303. Kaartokallio, H.; Kuosa, H.; Thomas, D. N.; Granskog, M. A.; Kivi, K. Biomass, composition, and activity of organism assemblages along a salinity gradient in sea ice subjected to river discharge in the Baltic Sea. Polar Biol. 2007, 30, 183-197. Granskog, M. A.; Kaartokallio, H.; Thomas, D. N.; Kuosa, H. The influence of freshwater inflow on the inorganic nutrient and dissolved organic matter within coastal sea ice and underlying waters in the Gulf of Finland (Baltic Sea). Estuarine, Coastal Shelf Sci. 2005, 65, 109-122. Krell, A.; Ummenhofer, C.; Kattner, G.; Naumov, A.; Evans, D. W.; Dieckmann, G. S.; Thomas, D. N. The biology and chemistry of land fast-ice in the White Sea, RussiasA comparison of winter and spring conditions. Polar Biol. 2003, 26, 707-719. Stedmon, C. A.; Markager, S. Resolving the variability in dissolved organic matter fluorescence in a temperate estuary and its catchment using PARAFAC analysis. Limnol. Oceanogr. 2005, 50, 686-697. Stedmon, C. A.; Markager, S. Tracing the production and degradation of autochthonous fractions of dissolved organic matter by fluorescence analysis. Limnol. Oceanogr. 2005, 50, 1415-1426. Kroon, H. Determination of nitrogen in water: Comparison of continuous flow method with on-line UV digestion with the original Kjedahl method. Anal. Chim. Acta 1993, 276, 287-293. Qian, J.; Mopper, K. High performance high temperature combustion total organic carbon analyzer. Anal. Chem. 1996, 68, 3090-3097. Stedmon, C. A.; Markager, S.; Kaas, H. Optical properties and signatures of chromophoric dissolved organic matter (CDOM) in Danish coastal waters. Estuarine, Coastal Shelf Sci. 2000, 51, 267-278. Mobed, J. J.; Hemmingsen, S. L.; Autry, J. L.; McGown, L. B. Fluorescence characterization of IHSS humic substances: Total luminescence spectra with absorbance correction. Environ. Sci. Technol. 1996, 30, 3061-3065.
(20) Lakowicz, J. R. Principles of Fluorescence Spectroscopy; Plenum: New York, 1999. (21) McKnight, D. M.; Boyer, E. W.; Westerhoff, P. K.; Doran, P. T.; Kulbe, T.; Andersen, D. T. Spectrofluorometric characterization of dissolved organic matter for indication of precursor organic material and aromaticity. Limnol. Oceanogr. 2001, 46, 38-48. (22) Stedmon, C. A.; Markager, S.; Bro, R. Tracing dissolved organic matter in aquatic environments using a new approach to fluorescence spectroscopy. Mar. Chem. 2003, 82, 239-254. (23) Bro, R. PARAFAC tutorial and applications. Chemom. Intell. Lab. Syst. 1997, 38, 149-171. (24) Andersson, C. A.; Bro, R. The N-way toolbox for MATLAB. Chemom. Intell. Lab. Syst. 2000, 52, 1-4. (25) Cory, R. M.; McKnight, D. M. Fluorescence spectroscopy reveals ubiquitous presence of oxidized and reduced quinones in dissolved organic matter. Environ. Sci. Tecnol. 2005, 39, 81428149. (26) Murphy, K. R.; Ruiz, G. M.; Dunsmuir, W. T. M.; Waite, T. D. Optimized parameters for rapid fluorescence-based verification of ballast water exchange by ships. Environ. Sci. Technol. 2006, 40, 2357-2362. (27) Belzile, C.; Gibson, J. A. E.; Vincent, W. F. Colored dissolved organic matter and dissolved organic carbon exclusion from lake ice: Implications for irradiance transmission and carbon cycling. Limnol. Oceanogr. 2002, 47, 1283-1293. (28) Coble, P. G.; Green, S. A.; Blough, N. V.; Gagosian, R. B. Characterization of dissolved organic matter in the Black Sea by fluorescence spectroscopy. Nature (London, U.K.) 1990, 348, 432-435. (29) Mopper, K.; Schultz, C. A. Fluorescence as a possible tool for studying the nature of water column distribution of DOC components. Mar. Chem. 1993, 41, 229-238. (30) Coble, P. G. Characterization of marine and terrestrial DOM in seawater using excitation-emission matrix spectroscopy. Mar. Chem. 1996, 51, 325-346. (31) Dittmar, T.; Kattner, G. The biogeochemistry of the river and shelf ecosystem of the Arctic Ocean: A review. Mar. Chem. 2003, 83, 103-120.
Received for review May 23, 2007. Revised manuscript received August 13, 2007. Accepted September 6, 2007. ES071210F
VOL. 41, NO. 21, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7279
