A brief overview of modern directions in marine DOC studies Part II—Recent progress in marine DOC studies

Abstract

Progress made in analytical techniques allows the formulation of new concepts in the biogeochemistry of organic carbon. The second part of our review summarizes the latest evolution and introduces new ideas in the biogeochemistry of marine dissolved organic carbon (DOC). Via classification of different fractions and sources of DOC, characterization of its composition, age and availability for bacterial utilization, and fate of DOC, we show the role of DOC in the global carbon cycle and the significance of bulk DOC in the oceans. Special emphasis is placed on the microbial loop in the cycling of DOC and its relation with higher trophic levels (phytoplankton and zooplankton). Significant progress has also been made in the study of the roles of colloidal organic material in metal complexation, ultraviolet radiation in dissolved organic matter photochemical oxidation, and chromophore-containing constituents of DOC as the signature of DOC for satellite observations. The importance of bulk DOC in the global carbon cycle requires the inclusion of this fraction in the regional and global carbon models. We predict that future DOC study in the ocean will focus on the development of sophisticated, almost continuously recording, moored DOC instrument arrays for the monitoring of small-scale DOC horizontal and vertical patchiness; widespread time series stations including estuarine, coastal and open environments; more detailed chemical characterization of different fractions of organic carbon from diverse marine habitats; parameterization of predictive models of DOC cycling on regional and global scales, incorporating the microbial loop; and finally, monitoring of DOC dynamics from satellites on regional and global scales.

---

Evgeny V. Dafner

Peter J. Wangersky

---

1. Introduction

After the recognition of the high-temperature catalytic oxidation (HTC) technique as the standard method for aquatic DOC study, the attention of oceanographers has shifted to the application of this method for marine carbon cycle studies. The period after approximately 1995 can be characterized as a time predominantly of collection of DOC data, rather than the improvement of HTC methodology. As a result, DOC measurement in the ocean has become an obligatory element of any serious oceanographic program, similar to dissolved oxygen, inorganic carbon properties, dimethyl sulfide, or nutrients. The major objective of this research is to understand the role of DOC in the global carbon cycle, and to identify factors controlling the production and fate of the DOC at different spatial and temporal scales.

This objective can be separated into two parts: (i) quantitative or static, i.e., collection of large amounts of DOC data from different locations, depths and seasons of the World Ocean, and (ii) qualitative or kinematic, i.e., estimation of processes influencing DOC cycling in different environments. Having the first type of data, we can draw plots of vertical and lateral DOC variability between different locations, estimate seasonal and daily variability and give explanations in terms of other environmental properties of seawater. The second type of information requires an experimental approach with marine biota. In most cases these experiments are conducted with bacteria, phytoplankton and zooplankton, and provide information on kinematic conversion of DOC (qualitative and quantitative) with time under diverse conditions. In the second part of our review we summarize the most important recent findings in marine DOC studies.

2. Recent progress in marine TOC studies

2.1. Why is it so important to study DOC in seawater?

The oceans are the most important sinks of carbon dioxide in the global carbon cycle, absorbing about half of the carbon of anthropogenic origin released to the atmosphere.¹ DOC in the ocean is recognized as one of the three main reservoirs of organic material on the planet, equal to the carbon stored in terrestrial plants or soil humus.² Despite its importance, the role of dissolved organic material (DOM) in the global carbon cycle is poorly understood. For example, some recent evidence suggests that DOC concentrations are steadily raising in the oligotrophic regions of the world's ocean as a result of climate mediated changes in plankton community structure.³

Azam et al.⁴ noted that we must acquire the ability to predict how the ocean's ecosystem will respond to global change, what role the ocean's biota will play in various global change scenarios, and whether we can safety manipulate the ocean's biota to cause it to absorb additional carbon dioxide. To find answers to these questions we should understand how biological forces determine the fate and spatial–temporal patterns of distribution of organic material (OM) in the ocean. This framework enables one to conceptually integrate the fate of OM from all sources, for example, whether derived from in situ photosynthesis or from river or sewage outfalls.⁴

2.2. Geography of marine DOC studies

During the last decade it has been recognized that for better understanding of the nature, sources and fate of DOC in the ocean and in the global carbon cycle oceanographers should expand the area of research from the open ocean to the coastal, estuarine and riverine waters. It is well known that rivers are very important sources of terrestrial OM, while estuaries play the role of buffers in transformation and partial deposition of the riverine OM (see below). Furthermore, the latest trend in the geography of populations shows increased populations along the coastal zone. For example in the USA, coastal population growth has averaged about 1.1 million people per year and is projected to increase by about 24.7 million by the year 2015.⁵ The anthropogenic influence creates additional stress for aquatic organisms; we must understand this in the near future.

The World Ocean can be characterized as a very dynamic and diverse system, where different water masses form and interact; which is in continuous exchange of properties between the surface and bottom, the northern and southern hemispheres; and which is in continuous exchange of heat and gases with the atmosphere. These processes determine the particularities of life in the ocean, and finally the features of the DOC cycle. Unfortunately, it is very difficult in one review to present all of the marine locations where the DOC pool has been studied even for the last several years. We summarize some results of recent marine DOC observations in Table 1 and illustrate the geography of marine DOC study in Fig. 1.

Fig. 1 Geography of HTC measurements of DOC in the different areas of the World Ocean. In more detail the numbers are given in Table 1 and include area studied, season, depths of sampling and DOC concentrations in surface and deep waters. Note, this Figure as well as Table 1 do not cover all observations conducted to date but give some impressions on the study of bulk DOC in different marine environments.

Table 1 Geography of HTC measurements of DOC in the different areas of the World Ocean by historical data. The numbers characterization include area studied, season, depths of sampling and DOC concentrations in the surface (SW) and deep waters. Locations of these numbers are shown in Fig. 1. When it is not stated, the deep water corresponds to a depth <2000 m. Some of the data were taken from vertical profiles and can contain differences from the referred paper. Note, this Table as well as Fig. 1 do not cover all observations conducted to date but give an impression of the study of bulk DOC in different marine environments

N	Location	Season	Depth	DOC/µM	Reference
1	Central Arctic Ocean	July–August 1994	SW	63–139	Wheeler et al.^6
2	Greenland Sea	April 1995	1000–4000 m	43.6–48.1	Hansell and Carlson^7
3	NE Water Polynya	June–August 1993	SW	110	Skoog et al.^8
			>500 m	<60
4	Trondheimfjord, Norway	Yearlong	SW	<210	Børsheim et al.^9
			400 m	>52
5	Central Arctic Ocean	Summer 1996	SW	43–225	Bussman and Kattner^10
			1000 m	29–112
6	12 Russian river estuaries	June–July 1994, August 1995	SW	230–1006	Lobbes et al.^11
7	Laptev Sea	September 1989,	SW	266–633	Cauwet and Sidorov^12
		September 1991	SW	167–639
8	Hebridian Shelf, West of Scotland	Nov–Dec 1995	SW	58	Álvarez-Salgado and Miller^13
			1400 m	45
9	Northern Atlantic	June–July, 1996	SW	>70	Kähler and Koeve^14
			Deep	40
10	Ría de Vigo estuary	Sept 1994–Sept 1995	SW	>130	Doval et al.^15
			40 m	<60
11	Gulf of Cádiz	February 1997	SW	89	Dafner et al.^16
			450 m	49
12	Azores Front	May 1997, August 1998, April 1999	SW	69	Doval et al.^17
			>500 m	47
13	Gulf of Maine	Summer 1996 and 1997	SW	155	Dai and Benitez-Nelson^18
			25 m	58
14	Georges Bank	April 1993	SW	72–85	Chen et al.^19
			1500 m	54–56
15	Georges Bank	Spring 1993	SW	65–92	Hopkinson et al.^20
		Summer 1994	1500 m	50
16	Chesapeake Bay (estuary)	July 1994	SW	118–215	Guo and Santschi^21
	Galveston Bay (estuary)	July 1995	SW	495–145
17	Bermuda Atlantic Time Series Station	Yearlong	SW	89	Carlson et al.^22
			Deep	50
18	Mississipi River Plume	Summer 1990–1993	SW	359	Benner and Opsahl^23
		Winter 1990–1993	SW	65
19	Equatorial Atlantic	August 1991	SW	97	Thomas et al.^24
		November 1992	400 m	46
20	Ligurian Sea	Annual	SW	92–100	Copin-Montégut and Avril^25
			Deep	50–58
21	Maramara Sea	May, 1988	SW	60–260	Tugrul^26
			750 m	40–50
	Black Sea	June, 1991	SW	<130
			1000 m	105
22	Aegean Sea	September, 1997	SW	52–128	Sempéré et al.^27
			Deep	47–56
23	Gulf of Lions	November 1994	SW	113	Yoro et al.^28
			1000 m	88
24	Gulf of Lions	Years-long, 1989–1993	SW	215	Cauwet et al.^29
			2000 m	53
25	Catalan-Balearic Sea	June 1995	SW	95	Doval et al.^30
			2000 m	51
26	Alboran Sea	November–January 1997	SW	<100	Sempéré and Dafner^31
			1000 m	44–46
27	Strait of Gibraltar	September 1997	SW	139	Dafner et al.^32
		April 1998	800 m	38
28	Arabian Sea, Indian Ocean	Yearlong 1995	SW	80–100	Hansell and Peltzer^33
			Deep	42
29	32°S, 80°E–3°N, 80°E	1995	1000–3000	43	Hansell and Carlson^7
30	Central Indian Ocean	June–September 1997	25–200	52–191	Sardessai and Sousa^34
			500–800	41–57
			3500–5300	41–64
31	North Pacific Ocean		1000–5000	42.3–33.8	Hansell and Carlson^7
32	Columbia River estuary	October 1997	SW	140–180	Klinnhammer et al.^35
33	Tokyo Bay	July 1991	SW	117	Ogawa and Ogura^36
			150 m	73
34	31°N, 159°W, northern Pacific Ocean	June 1987	SW	87	Williams and Druffel^37
			Deep	38
35	Station Aloha, 22°45′N, 158°W	April 1991	SW	82	Benner et al.^38
			765 m	38
			4000 m	41
36	Station Aloha, 22°45′N, 158°W	January 1992	SW	90–115	Tupas et al.^39
			1000 m	42
37	Equatorial Pacific	November 1994	SW	75	Hansell et al.^40
			200 m	50
38	63°S, 23°N, 103°/110°W	April 1994	SW	>80	Hansell and Waterhouse^41
			Deep	∼35
39	Equatorial Pacific, 140°W	Feb, Apr, Aug 1992	SW	80	Peltzer and Hayward^42
			2400 m	36
40	Central equatorial Pacific, 0°, 140°W	March, Apr, Oct 1992	SW	67	Carlson and Ducklow^43
			200 m	46
41	Santa Monica Basin, South Carolina Bight	Apr, July, Aug 1990	SW	103	Hansell et al.^44
			900 m	67
42	Eastern North Pacific	December–January 1996	SW	72	Loh and Bauer^45
			4000	37
43	Eastern North Pacific	June 1995	SW	72	Bauer et al.^46
			4097	37
44	Western South Pacific	January–March 1996	SW	45–73	Doval and Hansell^47
			>1000	43
45	Indian sector of the Southern Ocean	September–November 1995	SW	50	Dafner et al.^48
			4500	37
46	Indian sector of the Southern Ocean	January–March 1994	SW	52–63	Wiebinga and de Baar^49
			Deep	42
47	Atlantic sector of the Southern Ocean	October–November 1993	SW	<55	Kähler et al.^50
			Deep	34–38
48	Pacific sector of the Southern Ocean	June 1995	SW	50	Loh and Bauer^45
			5408	40
49	Pacific sector of the Southern Ocean	Winter 1996	0–700 m	41.7	Hansell and Carlson^7
		1996–1997	1000–4000 m	41.5–41.9
50	Pacific sector of the Southern Ocean	January 1995	SW	45–55	Ogawa et al.^51
			>2000	40–45

---

The DOC distribution in the ocean clearly shows an increase in concentrations in the direction from the open ocean to the coastal zone and from the bottom layer to the surface, following the features of phytoplankton distribution, river discharge and other sources of DOC to the oceans. Lateral increases of DOC concentrations occur from the open ocean areas (65–75 µM C) to the coastal zone (more than 100 µM C) with the highest concentrations of DOC in the estuaries of rivers (120–200 µM C and higher, see for example numbers 6, 10, 16, 23, 24, 32 in Table 1 and Fig. 1). Skoog et al.⁸ have found higher DOC concentrations in the Arctic than in other ocean areas. The vertical DOC distribution in the ocean clearly shows variation in the upper mixed-layer, usually 60–100 µM C, and uniform DOC distribution deeper than 1000 m, about 40 µM C and slightly lower. As we will discuss below, the seasonal cycle in the DOC distribution is connected almost always with the cycle of phytoplankton and with variations in river discharge.

2.3. Different size classes of organic carbon: TOC, POC, DOC, COC

The total organic carbon (TOC) pool consists of two operationally defined phases, particulate organic carbon (POC) and DOC. The traditionally defined DOC used in marine studies includes both colloidal organic carbon (COC) and truly dissolved OM. In turn, more detailed classification of COC suggests that this fraction of DOC can be divided into three different fractions: high-molecular-weight (HMW), low-molecular-weight (LMW), and ultrafiltrate organic carbon (UOC). Fig. 2 shows the size characterization of these fractions. POC and DOC fractions of the organic carbon have been described in the first part of this review.⁵² Here our discussion will focus mostly on the COC.

Fig. 2 Detailed classification of different classes of total organic carbon (TOC) and colloidal organic carbon (COC).

The COC in seawater is one of the largest reservoirs of organic carbon on the planet; it easily outweighs the phytoplankton or bacteria. Colloidal organic particles are also the most abundant particles in seawater.⁵³ The distinction between particulate OM (POM) and DOM is being replaced by the idea of an organic matter continuum—a gel-like matrix of polymers, replete with colloids and crisscrossed by ‘transparent’ polymer strings, sheets and bundles, from a few to 100s of µm—the oceanic ‘dark matter’.⁴

Relatively little is known about the composition and reactivity of COC because of the lack of suitable methods for its isolation from seawater. Until recently the conventional method for isolation has been absorption of acidified DOM onto nonionic XAD resins. This method typically recovers a small fraction (5–15%) of the total DOM in seawater, requires large manipulations of pH during isolation, and is selective for hydrophobic constituents.³⁸ In contrast, tangential cross-flow filtration (CFF) concentrates organic molecules primarily on the basis of size rather than chemical properties. The 1 kDa (kiloDalton) CFF membranes (∼1 nm pore size) are currently used to separate colloids from the truly dissolved fraction.⁵⁴

Several authors have suggested that rates of colloid formation must be rapid, presumably due to a combination of cell exudation and lysis, microbial degradation of POM and excretion by zooplankton.⁵⁵ It has been shown that one of the primary sources of marine colloids is the agglomeration of some fractions of the truly dissolved organic phase.⁵⁶ Chin et al.⁵⁷ have demonstrated that marine polymer gels can assemble from free DOM polymers, transforming DOM into POM with the strong influence of cations such as Ca²⁺ and Mg²⁺.

In the ocean, the vertical distributions of DOC and COC are well correlated. The COC is extensively involved in the biological cycle of DOC, and may represent one of the most reactive fractions of DOC.¹⁸ A recent study by Amon and Benner⁵⁸ has shown that COC supports a substantial fraction of marine heterotrophic bacterial production. But this COC material is not easily accessible to the bacteria and may escape extensive biological degradation by virtue of its particle size characteristics. However, when larger colloids (between 0.2 and 2.0 µm in diameter) are incorporated into aggregates (that are tens of µm to mm across), COC is broken down as the aggregates become bioreactors for organic material.⁵³ In contrast, it is most likely that COC near the sea floor originates, in part, from resuspended sedimentary particles that contain refractory organic matter.¹⁸

CFF technique suggests that more than 60% of the DOC material is of a molecular weight of less than 1 kDa and more than 90% is of less than 30 kDa.⁵⁹ Santschi et al.⁶⁰ and Guo and Santschi²¹ have studied the age and composition of COC by using ¹⁴C and found that COC in the >1 kDa size fraction was composed of a mixture of LMW, refractory compounds ranging in age from 400–4500 years, while COC in the HMW fraction was significantly younger (∼30 years).

3. Sources of DOC in marine systems

The DOC pool in the marine systems can be classified as autochthonous, i.e., produced in situ or allochtonous, i.e., delivered from the land. Most of the DOC in the ocean originates from biological production by marine organisms, while in contrast, fresh water systems are characterized by the terrestrial input of DOC. Reviewing the evidence for the marine origin of DOC, Druffel et al.⁶¹ have included: (i) its marine-like Δ¹³C signatures; (ii) the low abundance or absence of terrestrially derived, lignin-oxidation products within the DOC pool; and (iii) the compositional dissimilarity between riverine-derived humic substances and those isolated from seawater.

The major identified sources of DOM to the oceans are phytoplankton exudation; phytoplankton losses caused by cell damage or lysis; the ‘microbial loop’, including bacteria and microheterotrophs; sloppy feeding and lysis of bacteria by viruses; zooplankton excretion; the terrestrial input largely via rivers and sewage outfalls; sediment and pore waters. Recently it was shown that rain DOC input cannot be ignored in the estimate of the oceanic DOM fluxes.⁶² Dust flux from the biggest deserts is not a significant contributor to deep-sea sedimentation, but it is an important contributor of organic carbon to the ocean.

3.1. Terrestrial input

The oceanic input of organic carbon from continents can be separated into organic carbon transport by rivers and aeolian input. More recent evidence proves that ≤10% of the DOC in the equatorial Pacific is potentially of terrestrial origin because the Δ¹³C signature (relative to the PDB) of DOC indicates that the primary source of DOC is from marine-derived organic carbon.³⁷ It has been estimated that the total amount of organic carbon transported by world rivers ranges from 0.25 Gt C⁶³ to 0.38 Gt C year⁻¹,⁶⁴ with about 60% being transported as inorganic carbon and 40% being transported as organic carbon.⁶⁴ About 21 Gt C year⁻¹ enters the oceans in dissolved form and about 0.17 Gt C year⁻¹ in POC form. Approximately 45% of the organic carbon is discharged from tropical wet regions. The major part of the DOC is discharged into the Atlantic Ocean, while the bulk of the POC is discharged into the Indian and Pacific Oceans.⁶⁴ Global discharge of riverine DOC is sufficient alone to sustain turnover of the entire pool of organic carbon dissolved in seawater. Similarly, the input of terrestrial POC by rivers is adequate to supply all the organic carbon buried in marine sediments.⁶³

Ludwig and Probst⁶⁴ found empirical relationships between the observed organic carbon fluxes and the climatic, biologic and geomorphologic patterns characterizing the river basins. They showed that DOC fluxes are mainly related to drainage intensity, basin slope, and the amount of carbon stored in soil, while POC fluxes are calculated as a function of sediment fluxes, which depend principally upon drainage intensity, rainfall intensity, and basin slope.

Ittekkot⁶⁵ has studied chemical analyses of organic matter associated with suspended material from several major world rivers and found that globally 35% (81 × 10¹² g C year⁻¹) belongs to the labile fraction and may become oxidized in estuaries and in marine environments. The rest (150 × 10¹² g C year⁻¹) appears to be highly degraded and could represent a significant source of organic carbon accumulating in marine sediments. The dominant COC fractions in the riverine and estuarine waters are HMW and LMW colloidal organic carbon, while in contrast, oceanic waters are characterized by the UOC.³⁸ For example, using the smaller 1 kDa CFF in waters of the Amazon River system, Benner and Hedges⁶⁶ found that up to 76% of the DOC was in the HMW and LMW colloidal fractions. About 50–70% of the bulk DOC from estuarine waters of Galveston Bay were also found in these colloidal sizes.²¹ Guo and Santschi have shown that significant fractions of terrestrial COC are rapidly removed in estuaries, and that COC from more marine or internal sources dominate the estuarine distribution of COC. The behavior of DOC during estuarine mixing can be conservative or non-conservative, with removal or extra-source input.^21,67 The global atmospheric flux of terrestrial organic carbon to the surface ocean is difficult to estimate, but may be as large as 0.1 Gt C year⁻¹.⁶³

3.2. Rainwater input

Rain is a significant source of DOC to surface seawater (90 × 10¹² g C year⁻¹), equivalent to the magnitude of river input of DOC to the open ocean and one order of magnitude smaller than the carbon buried in marine sediments per year on a global scale.⁶² DOC in rain is poorly characterized with respect to the chemical compounds present, their sources, and patterns of variations. Recently Willey et al.⁶² presented a report on the DOC input by rain to the ocean surface. They found that DOC is a major component of both marine (23 µM C) and continental (161 µM C) rain. For the coastal zone of North Carolina these authors showed that the largest contribution of rainwater DOC—about 40%—came from organic acids (acetic, formic, lactic, oxalic, pyruvic, malonic, maleic, succinic and methylsulfonic acids, and C6, C9 and C10 oxoacids). Formaldehyde contributes 3%, acetaldehyde contributes 5% and amino acids contribute 2%. The amino acids include glycine, alanine, valine, tyrosine, serine, threonine, aspartate, glutamate, histidine, arginine, asparagines and glutamine. Humic-like substances also contribute DOC.

Rainwater is a significant source of formaldehyde to surface waters and may contribute as much as 30 times the resident amount found in natural waters of southeastern North Carolina during the summer.⁶⁸ Formaldehyde concentrations did not correlate with precipitation volume, suggesting a continuous supply during rain events presumably from direct photochemical production in the aqueous phase. A major fraction of TOC in precipitation in the northeastern US was derived from airborne particulate matter, such as soil dust and plant material. DOC comprised 84% and 80% of these totals, respectively.⁶⁹ The majority of annual deposition occurred in summer (June–September), when both concentrations and rainfall amounts are higher.⁶⁹

The importance of photochemical oxidation on transformation of organic compounds in the rain has also been underlined for the western Pacific where total diacid concentrations amount to about 3% of the TOC.⁷⁰ It has been proposed that shorter chain diacids are in part produced in the marine atmosphere by photochemical oxidation of organic matter through the intermediates such as ketoacids and dicarbonyls.⁷⁰ Recently Kawamura et al.⁷¹ have shown that major portions of the carboxylic acids detected in the rain at west Los Angeles, California, in 1981–1984 were not emitted directly from auto-exhausts, but are most likely produced in the atmosphere by gaseous and/or aqueous phase photo-induced reactions.

3.3. Sediment as a DOC source

One of the most commonly reported parameters for marine sediments is organic content, usually expressed as combustible OM, organic carbon or organic nitrogen. Sediment plays a more important role in the storage of organic carbon than rain. OM preserved in marine sediments provides a molecular record of marine biological processes, accounts for approximately 20% of all carbon burial and plays a key role in balancing the long-term flux of oxygen to the atmosphere.⁷² Berner⁷³ has estimated that about 1.3 × 10¹⁴ g C year⁻¹ is buried in sediment; moreover, more than 80% of all sediment organic matter accumulates on continental margins. It has long been known, for example, that continental slope sediments often contain more OM than do adjacent shelf sediments. But this trend can reflect many possible factors, including high rates of organic production, rapid burial, low rates of degradation or mineralization, and the dominance of fine-grained sediments, which increases the surface area onto which OM can adsorb.⁷⁴

Sorption of OM to mineral surfaces in marine sediments stabilizes the component molecules, slowing remineralization rates by up to five orders of magnitude.⁷² Sorptive protection can therefore account for the enigmatic preservation of intrinsically labile molecules such as amino acids and simple sugars in marine deposits and links the preservation of organic carbon in marine sediments to the deposition of mineral surfaces. Recent studies⁷² have shown that chemically labile, but several hundreds of years old OM can be well preserved on particle surfaces in sediments and can be degraded quickly once it is desorbed. In their experiments with sediments Keil et al.⁷² have found that the youngest desorbed material was mineralized most effectively, showing a 90% decrease in DOC, whereas OM from older sediments was degraded to a lesser extent. They showed that changes in OM were due not to physical losses, but to bacterial utilization. Marine sediments typically contain 1000 times as many bacteria per unit volume as the water column and resuspension should immediately supplement bacterioplankton standing stock.⁷⁵ In experiments with marine resuspended sediments and microheterotrophs (bacteria and protozoa) this author has shown that free bacteria and protozoa, rather than attached bacteria, were the main beneficiaries of the resuspension stimulus, and particles themselves or particle-associated substances were probably not the principal stimulant. Since the suspended organic material is composed of OM fractions of different sizes and varying turnover rates, Guo and Santschi²¹ have emphasized that older, more oxidized and thus more hydrophilic HMW OM might be preferentially released during sediment resuspension. In experiments with the surface sediments resuspension collected from six locations within the Hudson River Estuary and the Inner New York Bight, Komada and Reimers⁷⁶ have shown that the mineral-bound fraction of sedimentary organic carbon was the major source for the excess DOC released into solution, and that across various sedimentary environments, only a small (but fairly constant) fraction of the total sedimentary POC may be poised for rapid transfer to the water column.

Concentrations of pore water DOC are typically an order of magnitude greater than those from the overlying seawater. Therefore, the sediments appear to act as a DOC source to the bottom water.⁷⁷ Benthic fluxes from sediment pore water contain mostly HMW COC.⁷⁸ The relatively low variations in DOC concentrations in pore water and DOC-fluxes from the sediments under widely differing environmental conditions suggest that production and consumption of labile DOC components proceed at similar rates irrespective of what the overall benthic activity is.⁷⁷ Regardless of the mechanisms, suspended OM could be an important source for the DOC pool in coastal and estuarine waters.

Guo and Santschi⁷⁹ performed radiocarbon measurements on COC in the water column and benthic nepheloid layer (BNL) from two continental margin areas (the Middle Atlantic Bight and the Gulf of Mexico) and controlled laboratory experiments to study sources of old DOC in the ocean margin areas. They showed that COC from the BNL was much older than COC from the overlying water column. These results, together with strong concentration gradients of suspended particulate material, POC, PON, and DOC, suggest a sedimentary source for organic carbon species and possibly for old COC as well in BNL waters. They concluded that old COC from continental margin nepheloid layers may thus be a potential source of old DOC to the deep ocean.⁷⁹

3.4. Phytoplankton production

Production of organic carbon by phytoplankton is the most important source of DOC in marine environments. The average amount of carbon fixed photosynthetically in the world's oceans (excluding adjacent seas) has been estimated at between 130 g C (m² year)⁻¹ and 500 g C (m² year)⁻¹ with a steady-state input of DOC of 0.1 × 10¹⁵ g C year⁻¹ (6 × 10¹⁷ g C per 6000 years) or about 0.067–0.25% of the photosynthetically fixed carbon.³⁷ Extant models of organic carbon flux in upper ocean waters identify the dominant DOC flux to heterotrophic bacteria as direct leakage of soluble photosynthate from primary producers; these models generally require 10% or more of primary production to be lost to the phytoplankton cell in this form.⁸⁰

Unfortunately, there is very little information available on the relationship between DOC and the organic exudates released by the phytoplankton. Numerous determinations of uptake kinetics for bulk seawater DOC and algal exudates have yielded turnover times ranging from hours to years. Most of the determinations were based on either short-term radiotracer investigation of simple organic compounds and phytoplankton exudates and detritus, or trace analyses of extremely labile exudates released by algae.⁸¹ However, these determinations have their limitations. The single substrates added might be taken up in a manner different from that of naturally occurring, more complex mixtures of solutes. The uptake rates for the extremely labile compounds do not represent the bacterial uptake rates for bulk DOC in natural seawater, as was demonstrated by Chen and Wangersky.⁸¹ The kinetic studies on the uptake of simple organic substrates are normally conducted by the addition of relatively simple compounds of LMW, such as amino acids and monosaccharides, and are thus likely to apply only to the relatively small labile component of photoassimilated carbon released during the decomposition phase of phytoplankton blooms.

The major portion of DOM released by phytoplankton in the oceans consists of small molecules (LMW) and 20–30% of HMW, carbohydrate rich molecules.⁸² Measurable accumulations of DOM have been observed to take place toward the end of phytoplankton blooms in the oceans. Because these accumulations of OM are clearly linked to seasonal phytoplankton growth events, there is interest in the composition of both the cellular and extracellular release products of phytoplankton. The little that is known about the composition of DOM produced by phytoplankton indicates it consists of all types of biochemical products including carbohydrates (mono, oligo, and polysaccharides), nitrogenous compounds (amino acids, proteins, and polypeptides), lipids (fatty acids), and organic acids (glycollate, tricarboxylic acids, hydroxamate, vitamins). Carbohydrates in the form of exopolymers are released in large amounts by growing phytoplankton in cultures and in the sea. Predominance of carbohydrates in phytoplankton exudates may be one reason why 25–50% of marine HMW DOC is composed of polysaccharides.⁸²

Aluwihare et al.⁸³ have studied the production and accumulation of the oligosaccharides in seawater collected in the Atlantic and eastern tropical Pacific oceans (Mid-Atlantic Bight, between Cape Cod and Cape Hatteras, Hawaii and the eastern North Atlantic) to monitor changes in ultrafiltrate dissolved organic material (UDOM) during decomposition of algal exudates. They found that at the onset of stationary phase growth, DOC concentrations increased from 45 to 430 µM C with the major contribution from polysaccharides (4.3–3.2 ppm), proteins (2.8–1.5 ppm) and minor amounts of lipids (1.5–0.8 ppm). After 19 days of degradation, the amounts of major biochemicals changed considerably; sugars and lipids were relatively more abundant, but a significant protein fraction remains. Finally, after 37 days of incubation the composition of UDOM was similar to that of seawater. The monosaccharide composition of their first incubation sample was dominated by mannose (42%) and galactose (26%), but all other major sugars measured in UDOM were present. After 19 days, the relative abundances of mannose and galactose were unchanged (41 and 27%, respectively). After 37 days, mannose and galactose decreased to respectively 8% and 21% of the total sugar, and the ratio of major neutral sugars approached UDOM values. Oligosaccharide linkage patterns at the end of the experiment were highly specific and nearly identical to UDOM isolated from seawater.⁸³

In contrast, different results for total DOC between experimental and in situ observations have been recently reported by Smith et al.,⁸⁴ who studied the carbon and nitrogen exchanges among the phytoplankton Phaeocystis antarctica as dominant plankton assemblages, bacteria and DOC pools to compare the elemental partitioning in these experimental enclosures with those observed in situ. They used natural seawater assemblages collected in the southern Ross Sea during austral spring 1994 and 1995 and found that DOC levels remained low (less than 50 µM C) while nutrients were present, but increased dramatically (to more than 200 µM C) after nitrate was depleted. By contrast, field studies showed no depletion of nitrate, similar levels of POC and DOC and relatively low levels of bacterial biomass compared to those found during exponential growth in the experiment.⁸⁴

3.5. Planktonic grazers as a source of marine DOC

Another point of view suggests that planktonic grazers are potentially a major source of DOC in marine environments. Strom et al.⁸⁵ reviewing literature about DOC production by grazers summarized that healthy, exponentially growing algae are typified by release rates equivalent to <10% of daily carbon fixation, even when growing at low rates. These low release rates are difficult to reconcile with bacterial DOC demands that can equal 20–40% (or more) of photosynthetic carbon fixation. This discrepancy has led some authors to postulate that grazing supplies much of the DOC supporting bacterial production. Possible modes of grazer-associated DOC release include sloppy feeding, excretion, egestion, and dissolution of fecal material. Further, Strom et al.⁸⁵ have noted that the best studied of these modes is sloppy feeding, or release of dissolved compounds when prey cells are broken by the mouthparts of crustacean zooplankton. Crustacean zooplankton are also thought to release a certain quantity of DOM via direct excretion. Release of DOM by protist grazers such as ciliates and flagellates has also been observed. DOC released by protists may occur via excretion or during egestion, i.e., when food vacuoles fuse with the grazer's cell membrane for the release of waste products. Finally, dissolution of grazer fecal material may contribute to the DOC pool. Theoretical arguments suggest that diffusion can drive the transfer of solutes from crustacean fecal pellets to the surrounding water on a time scale of minutes, and that this diffusion-driven process should be enhanced for the smaller, less coherent waste products of protist grazers. Thus, much grazer-driven detritus may ultimately cycle back into the food web as DOC.⁸⁵

It is difficult to assign relative values to the various mechanisms for the contribution of DOC to the oceans because in our experiments measuring the contribution of any one source, we attempt to eliminate other possible sources, and so obtain a possibly false picture of the overall addition of DOC. Further, the experiments with bacterial utilization of DOC from phytoplankton cultures suggest that the most easily used components may be lost in the normal interval between sampling and sample preservation.⁸¹ We must always remember that the DOC values we measure in oceanic waters are the result of dynamic processes which do not cease when we close the sampling bottles. The true value for total organic matter may well be somewhat higher and more variable than we now believe; we will not know for certain until we devise an in situ analytical procedure.

4. DOC in marine systems

4.1. DOC availability in the ocean

Availability of the DOC pools for bacterial utilization depends upon the age and quality of the DOC and consequently can be classified as highly labile, semi-labile and refractory pools.⁴³ It is accepted in modern marine biogeochemistry that the highly labile pool is biologically available OM that turns over on time scales of hours to days. The semi-labile pool can be determined as DOC turned over on a seasonal time scale. Refractory DOC is representative of deep-water DOC, with an average age of 6000 years.³⁷ Kieber et al.⁸⁶ and Mopper et al.⁸⁷ have suggested that this biologically refractory material is broken down into biologically labile components by photo-oxidation. They have suggested that the biologically refractory material turns over on a time scale as long as deep-ocean mixing, which is less than 6000 years.

All these DOC species are available for bacterial utilization but with different time resolutions. Labile and semi-labile pools play an important role in the DOC cycling in the surface waters, while refractory material is important in the deep and intermediate waters. In productive surface waters, as much as a third to a half of the DOC may be present as labile material, degraded biologically in hours to days.⁸⁸ Studies on bacterial respiration and growth have shown that bacteria are rapidly respiring a very small pool of highly labile, recently produced DOC.⁸⁹ Benner et al.³⁸ have shown that labile and refractory components of DOM coexist in surface water, whereas refractory components dominate in deeper waters. Moreover, HMW polysaccharides made up an important part of the labile DOC in surface water.³⁸

It has been estimated that the global net production of semi-labile DOC is ∼17% of global new production.⁹⁰ In the Sargasso Sea, Hansell et al.⁹¹ have identified two source terms for the semi-labile DOC pool, independent of recently produced photosynthate. First is the photo-oxidation of refractory DOC, of a deep-water source, in the upper few meters of the water column and second, the utilization of DOC produced during the previous spring bloom or some other period of high DOC production, which is semi-refractory to microbial utilization and slowly mineralized. As DOM is cycled through the heterotrophic bacteria the residual DOM becomes more and more recalcitrant and eventually biologically resistant.⁴³

4.2. Age or residence time of marine DOC

The age of DOC and consequently the availability of DOC for bacterial utilization is determined by the origin of the OM. For most of the oceanic waters the origin of the OM depends on the source of water mass formation and its history of circulation. One of the methods for DOC age estimation is the measurement of radiocarbon in the DOC fraction by tandem accelerator mass spectrometry, i.e., the analysis of the deviation in parts per thousand from the ¹⁴C activity of nineteenth-century wood. More recent observations by this technique indicate that the DOC is older in aged and deep-water masses.

Williams and Druffel³⁷ have studied the age of DOC (<1 µm pore diameter) in the oligotrophic gyre of the central North Pacific and found that the apparent mean ages or residence time in the surface water was ∼1310 years and ∼6000 years in the deep waters. In the Sargasso Sea it was found that the age for the deep DOC was ∼2000 years younger than in the Pacific.⁹² This reflects a recycling of about 80% of the deep DOC during each deep ocean mixing cycle, assuming a closed system.⁶¹ The discrepancy in age between two oceans was explained by differences in the Δ¹⁴C of the DOC sources to the deep basins, and by the different deep-water circulation patterns and transit times in the two oceans.⁹² Bauer et al.⁹² have concluded that for each ocean the Δ¹⁴C values of DOC_HTC, DOC_UV and humic substances were remarkably similar, yielding no evidence for a component of DOC that cycled through the system on time scales shorter than several thousands of years. In the discussion that follows we will provide some recent evidence on DOC decay in the deep waters.

4.3. Estimates of DOC exchange between different oceanic basins

For a better understanding of the role of DOC in the global carbon cycle we need to include this carbon pool in the regional models of carbon exchange between different marine basins. Most of the estimates published to date were dedicated to carbon dioxide exchange, while the role of the DOC has been ignored. Although recently several reports on DOM exchange in the marine system have been presented, our knowledge in this field is still limited. One of the best-studied areas covered by this research is the North Atlantic, because it is the most important area for the anthropogenic carbon sink. Meridional CO₂ transport has already been evaluated for the whole North Atlantic and for some of its parts.

These estimates show that the North Atlantic dominates the southward inter-hemispheric inorganic carbon transport. Stoll et al.⁹³ have studied the flux of DOC along the 58^oN section in the North Atlantic and found that about 0.04 × 10⁶ to 0.16 × 10⁶ mol C s⁻¹ have been transported, in contrast with CO₂, in the north direction. In the northeastern Atlantic, the DOC transport has also been evaluated for the Mediterranean water. In the source of this water, in the Strait of Gibraltar, the DOC transport ranged from 0.97 × 10⁴ mol C s⁻¹ to 1.81 × 10⁴ mol C s⁻¹ in September 1997,⁹⁴ and 0.9 × 10⁴ mol C s⁻¹ to 1.0 × 10⁴ mol C s⁻¹ in April 1998.³² In the Gulf of Cádiz, close to the source of this water, the shallow core of this water carries out a significantly lower amount of DOC, about 1.34 to 2.68 × 10² mol C s⁻¹.¹⁶

All these estimates do not take into consideration the decay of DOC due to bacterial remineralization of OM inside deep waters. Although Williams and Druffel³⁷ and Bauer et al.⁹² showed that DOC in the deep ocean is composed of a large fraction of very old material (6000 years) and a smaller fraction of young material, DOC in the deep ocean has long been considered to be uniformly distributed and hence largely refractory to biological degradation. Only recently Hansell and Carlson⁷ analyzed DOC concentrations along the deep path of the global ‘conveyor belt’, including the region of deep-water formation in the North Atlantic, the Antarctic Circumpolar Current, the Indian and Pacific Oceans (see numbers 2, 26, 13, 49 in Table 1 and Fig. 2) and have presented the first evidence of bacterial degradation of DOC in deep waters. They found that DOC concentrations decreased by 14 µM C from the northern North Atlantic to the northern North Pacific, representing a 29% reduction in concentrations. The net rate of microbial utilization of DOC during the ∼80 year transit from the site of NADW (North Atlantic Deep Water) formation to the 32°N sampling site was 0.05 µM C, or half the TOC oxidation (sum of DOC and sinking biogenic carbon oxidation) estimated from apparent oxygen utilization rates in the deep central Atlantic.⁷

4.4. The Mediterranean Sea as a regional conveyor belt model

Going from the global to the regional scale, we can determine in greater detail how local processes affected the DOC cycling during the aging of a water mass. For example, DOC decay in the deep waters of the western Mediterranean Sea has already been hypothesized.³² On the west, this sea connects through the Strait of Gibraltar (depth <600 m) with the northeastern Atlantic and on the east, via the Straits of Bosphorus and Dardanelles with the Black Sea. The semi-enclosed configuration of the Mediterranean Basin, combined with the features of circulation and oligotrophy of this sea, can provide a good model for the study of the recirculation of DOC between different water masses.

The simplest model suggests that the Mediterranean Sea is characterized by the regional conveyor belt. The ‘rotation’ of this belt begins in the Strait of Gibraltar with the inflowing Atlantic Surface Water (ASW). After passing this strait, ASW is modified in the Alboran Sea and spreads further over to the eastern Mediterranean. Along the south coast of France, winter convection causes modified ASW to sink down and form the West Mediterranean Deep Water (WMDW).^95,96 In the Levantine Sea (eastern Mediterranean) surface water, significantly modified due to evaporation, increases in density and sinks, forming the Levantine Intermediate Water (LIW). After recirculation in the eastern basin, both WMDW and LIW spread to the western Mediterranean where they mix in the Strait of Gibraltar, forming the Mediterranean water outflowing to the northeastern Atlantic.

As suggested in the recent literature (Table 1), the DOC concentrations in the surface waters of the Mediterranean Sea are close to, or higher than, 100 µM C, while concentrations of DOC in deep and intermediate waters are significantly lower, showing a trend in decreasing concentrations from ∼45 µM C in the eastern down to 38 µM C in the western Mediterranean (Table 1). The DOC in seawater can be separated into labile, semi-labile and refractory organic carbon. If we accept that the lowest concentrations measured in this sea, i.e., 38 µM C,³² characterizes the refractory DOC pool, we can estimate concentrations of labile and semi-labile DOC over the water column. These estimates suggest that the sum of labile and semi-labile DOC, i.e., the fraction continuously utilized by microplankton and heterotrophic bacteria, ranges from 94 to 222 µM C in the surface waters including coastal waters, and ∼20 µM C and lower in deep waters. These estimates are significantly higher than those presented for other marine systems.

The Mediterranean conveyor belt differs from the global one as described by Broecker⁹⁷ The inflowing ‘young’ ASW (relative to the outflowing Mediterranean waters) does not sink in the area of formation as NADW does but enters the Strait of Gibraltar as the surface current, flows over all of the Mediterranean Sea, changing properties, sinks and goes back to the Atlantic as the intermediate and bottom water masses. It has been estimated that the renewal time scale of water masses forming the Mediterranean outflow ranges from 7–10^98,99 to 100 years^100,101 for the WMDW, and from 10 to 20 years for the LIW.¹⁰² We can apply these age estimates for the residence time of the surface waters in the Mediterranean Sea. The younger age of intermediate and deep-water masses in the Mediterranean basin determines the higher concentration of labile and semi-labile DOC measured in these waters. The importance of oligotrophy as favoring DOC accumulation has been already noted by the elevated concentrations of surface-water DOC in oligotrophic areas of the Arabian Sea³³ and in the northern North Atlantic.¹⁴ Recently it has been estimated that the residence time for DOC in the Mediterranean Sea is something between 103 and 149 years.¹⁰³ As we showed above, the DOC pool can be mineralized only down to ∼38 µM C, which represents the refractory DOC fraction. This DOC amount goes back to the Atlantic Ocean, avoiding remineralization by bacteria.

Although the Mediterranean waters are younger than deep oceanic waters, on leaving the Mediterranean basin the DOC values are close to those measured for the very old deep ocean waters.⁷ This fact illustrates the participation of bacteria in the turnover of the organic material inside the LIW on its pathway to the Strait of Gibraltar from the eastern Mediterranean. Previous investigations for example suggest that metabolic activity, bacterial biomass, and particulate organic material were higher in the LIW of the Western Alboran Sea than farther to the east, and especially higher than in the LIW of the Balearic Sea.¹⁰⁴

4.5. Chemical characterization of marine organic carbon

The composition of marine DOC is still incompletely known. Lee and Henrichs,¹⁰⁵ reviewing different sources of marine DOM, have noted that the sources of DOM should provide clues to the nature of this material. They pointed out that DOM removed from seawater by resins such as XAD-2 contained predominantly aliphatic carbon, in contrast to terrestrial humic substances that contain more aromatic structures.¹⁰⁶ Carboxyl functionality gives a net negative charge to seawater DOM, and adsorbed OM is thus responsible for the negative surface charge of ocean particles.¹⁰⁷

In the central Pacific (number 35 in Fig. 2 and Tab. 1) it has been found that concentrations of total dissolved carbohydrate corresponded to 27, 13 and 7 µM C in surface, oxygen minimum and deep-water samples that were 33, 34 and 17% of the total DOC.³⁸ COC includes a substantial carbohydrate content that can be turned over by biological activity. In the colloidal size fraction, DOM ranged from ∼14 µM C in surface waters to ∼2 µM C in deeper waters. Carbohydrates accounted for 49, 18 and 19% of the total C in surface, oxygen minimum, and deep-water UOC samples.³⁸ HMW components of DOM are reactive and polysaccharides are dominant components of this reactive material, perhaps supporting much of the heterotrophic activity in the surface ocean.³⁸

Chemical characterization of macromolecular DOC at several sites in the Atlantic and eastern tropical Pacific oceans (Mid-Atlantic Bight, between Cape Cod and Cape Hatteras, Hawaii and the eastern North Atlantic) has shown that neutral sugars, acetate and lipids show similar distributions, suggesting that these constituents are linked together in a common macromolecular structure.⁸³ Proton NMR of UDOM yields very similar spectra for all samples, with well-resolved resonance from carbohydrates (5.5–5 ppm (anomeric), 4.3–3.4 ppm (CH), and 1.3 ppm (CH₃); 55% total carbon), acetate (2.0 ppm; 7% total carbon) and LMW lipids (1.3 (CH₂) and 0.9 (CH₃) ppm; 6% total carbon. The abundance of major biochemicals in the UDOM was relatively constant in all samples, and had an average carbohydrate∶acetate∶lipid carbon ratio of 8∶1∶1.⁸³

Compositional similarities of UDOM were also evident in the monomer components of the carbohydrates. Galactose was the most abundant monosaccharide in the samples. Xylose, rhamnose, fucose, glucose and mannose were only slightly abundant. Arabinose, the least abundant sugar, was 35% of the concentration of galactose. This complex distribution of neutral sugars is similar in the Gulf of Mexico, Sargasso Sea and North Pacific.⁸³

As noted by Aluwihare et al.,⁸³ the relatively fixed ratio of major biochemicals and simple oligosaccharide linkage patterns observed in UDOM suggest that a large fraction of macromolecular DOC in surface seawater is not the complex, heterogeneous polymers expected from geopolymerization of simple biomolecules. Rather their results show that most macromolecular OM is a mixture of structurally related acyl-oligosaccharides. Carbohydrate, acetate and lipid portions of UDOM are linked together in a common macromolecular structure, giving rise to the relatively fixed ratio of major biochemicals. The carbohydrate portion of this macromolecule is also characterized by a distinctive and very heterogeneous distribution of seven neutral sugars. They concluded that processes responsible for the formation of these macromolecules must be operative on a global scale and observed variations in monosaccharide distribution and the ratio of carbohydrate, acetate and lipid may result from differences in UDOM sources and removal processes from different sampling sites.

Results from estuarine-, surface-, and deep-water samples show that an important fraction of COM consists of fibrillar material, which is rich in polysaccharides and ‘fresher’ (i.e., has a younger radiocarbon age) than the bulk COM. This result is important because COM makes up 30–70% of oceanic and estuarine nominally ‘dissolved’ organic matter. Other microparticles appear to be quasi-spherical, often attached to the fibrils like pearls.¹⁰⁸

4.6. The fate of organic carbon in marine systems

Analyzing the radiocarbon distribution in the DOM, Williams and Druffel³⁷ have noted that a substantial fraction of DOC, regardless of origin, may be common to all oceanic deep waters and undergo multiple recycling, similar to the recycling of salt in the ocean. They estimated that the apparent mean ages of residence times of about 6000 years BP could result from mixing 51% dead DOC (Δ¹⁴C=−1000‰) with 49% pre-bomb DOC (−40‰), neglecting any recent input of bomb radiocarbon. To arrive at any meaningful predictions of net production by the phytoplankton, it is obviously important to know the rate at which the organic carbon produced by photosynthesis is broken down and respired to CO₂.⁵³

Our views of the ocean's organic matter have changed dramatically and in modern biogeochemistry the dominant conception is that the microbial loop is a major biological force in oceanic carbon flux.⁴ Modern studies have established that bacteria account for a major fraction of the oceanic biomass and particulate carbon pool and use about one-half of the photosynthetically produced organic matter.⁴ Bacterial use of DOM thus mediates large-scale OM fluxes through the pathway: DOM → bacteria → protozoa or viruses (the microbial loop).¹⁰⁹

Indeed, marine bacteria use about one-quarter of the total carbon fixed on the earth. OM flux into bacteria is highly variable. This could affect flux partitioning between major pathways: microbial loop, sinking and grazing food chain. Azam et al.¹¹⁰ have proposed that this variability is due to bacterial interaction with a patchy OM field. Small-scale structure of the organic matter field can create hot-spots of microbial activity. Microbial activity and interaction at such small-scale features have important implications for the fundamental issues of nutrient and metal cycling and food web structure and dynamics.⁴

As has been noted by Azam et al.,⁴ the OM field in eutrophic waters can be envisioned as a denser gel with increased concentrations of hot-spots. The action of bacteria at the hot-spots causes rapid remineralization, particularly of phosphorus; this can confuse the issue of nutrient-ratio-control strategies to remedy eutrophication. Further, ‘uncoupled solubilization’ of OM in the hot-spots could result in the production and ‘storage’ of slow-to-degrade OM in the environment. Finally, since OM hot-spots represent rich and varied environments, they could offer microniches with enhanced bacterial diversity, including bacteria considered maladapted to life in the pelagic ocean, such as human pathogens. Vibrio cholera persists in association with marine copepods.⁴

The rate of the decay of the OM released by phytoplankton depends upon the plankton species present and their physiological state.¹¹¹ In contrast, on the basis of digestion theory and analyzing published empirical evidence Jumars et al.⁸⁰ have suggested that the principal pathway of DOC from phytoplankton to bacteria is through the byproducts of animal ingestion and digestion rather than via excretion of DOC directly from intact phytoplankton. These authors have shown that excess (over ambient) concentrations of solute in fecal pellets of typical size (diameter ∼1 mm) are lost rapidly; 50% of any excess is diffused out of the pellet within 5 min—even in a stagnant water column and without particle sinking. Reasons for rapid loss and its insensitivity to fluid dynamic conditions are the small size of the pelletal reservoir and the sharp concentration gradient between pelletal and ambient concentrations upon pellet release. As a consequence, most solutes initially contained in fecal pellets of zooplankton generally will remain in the 10−100 m thick water layer within which the pellets initially are deposited.⁸⁰

It has been suggested that in estuarine waters the direction of organic matter degradation is from particulate to COC to truly dissolved OM.^21,67 In the ocean, greater than 75% of the net POM loss occurs in the upper 500 m of the water column; because sinking particles contain more viable, metabolically active microorganisms, the process of microbial decomposition is considered to be an important mechanism controlling POM flux.¹¹² This result is consistent with the observed correspondence between POM flux and DOC concentrations and with the reported selective loss of biochemically labile compounds from sinking particles (see reference above). Fig. 3 summarizes the pathway of sinking POM in the ocean as modified by Cho and Azam.¹¹³ As OM degrades from larger to smaller particles and from COM to truly dissolved OM, nitrogen-containing OM in larger sized material is preferentially decomposed, leaving more carbon in the smaller sized COC fraction and at the same time increasing the nitrogen content in the truly DOM fraction. This observation is consistent with observations that bacterial utilization rates and apparent ¹⁴C ages of different size fractions of DOM suggest that the freshness and reactivities decrease from larger size to smaller size organic matter in oceanic environments.⁵⁸

Fig. 3 Schematic diagram for the pathway of sinking POM flux. DOM and DIM represent dissolved organic matter and dissolved inorganic matter. S1–S3 denote the sinking rates of various classes, after Cho and Azam.¹¹³

4.7. Ratio of DOC versus apparent oxygen utilization (AOU)

In the ocean, DOC and dissolved oxygen have quite similar vertical distributions because both are affected by biological and physical processes. Their vertical profiles are characterized by the exponential decrease of concentrations with depth, sometimes with intermediate extremes induced by the lateral spreading from the area of formation of old or young water masses. Since both DOC and oxygen are utilized by bacteria which in turn produce CO₂ through respiration, and because of the similarity of their vertical distribution, it was hypothesized that the oxygen consumption resulted from the DOC regeneration over the water column. As a result, most of the studies have found an inverse relation between molar ratios of DOC and AOU (Table 2) although some of the observations demonstrate weak or insignificant correlation between these parameters.^44,122

Table 2 Comparison of the slopes of DOC versus AOU relationship by historical data³²

Slope	Location	Reference
−0.77	Marine organic matter	Redfield et al.^114
−0.80	NW Pacific	Sugimura and Suzuki^115
−1.00	Surface	Druffel et al.^116
−0.21	Deep water, N Pacific
−0.062	NW Indian Ocean	Kumar et al.^117
−0.24	Sargasso Sea	Bauer et al.^118
−1.58	Surface, N Pacific	Tanoue^119
−1.08
−0.017	Deep water, N Pacific
−0.047
−0.21	N Atlantic Ocean	Kepkay and Wells^120
−0.25
−0.56	N Atlantic	Peltzer et al. (unpublished) cited by Kepkay and Wells^120
−0.43	N Atlantic Ocean	De Baar et al.^121
−0.14	Equatorial Atlantic Ocean	Thomas et al.^24
−0.09
−0.074	Surface, Equatorial Pacific	Peltzer and Hayward^42
0.04	Deep water, Equatorial Pacific
−0.084	Indian sector of the Southern Ocean	Wiebinga and De Baar^49
−0.26	Pacific sector of the Southern Ocean	Doval and Hansell^47
−0.23	Central Indian Ocean
−0.25	The Strait of Gibraltar	Dafner et al.^32

---

Tanoue¹¹⁹ has shown that a single linear regression analysis throughout the entire water column was inappropriate for estimations of the contribution of DOC to the utilisation of oxygen in the water column. In general, correlation was found when subsurface and intermediate levels are considered separately, i.e., variations of both AOU (and probably DOC) were as much due to a mixing as to biogeochemical processes.²⁴ Historical data on the DOC versus AOU relationship represented in Table 2 suggests a large variability of the slope from the different areas of the ocean. One of the problems in comparing the historical data is that prior to Peltzer and Hayward⁴² all authors used a model I regression, resulting in a negative bias in their slopes in addition to the natural variability. Since then, most authors have used the geometric mean regression (a model II regression).

To formalize the regression between DOC and O₂ it is common to use the ‘Redfield ratio’, i.e., C∶O₂ = −0.77.¹¹⁴ Although this ratio has been suggested for fresh marine phytoplankton, it found a wide application for estimation of bacterial remineralization through the entire water column. Over the last decade most of the papers producing information on DOC distribution in the ocean discussed the DOC versus O₂ relationship. Comparing the ‘Redfield ratio’ with the slope of the regression line between DOC and AOU, it is possible to speculate about the oxygen deficit in the surface and deep waters. From vertical profiles of DOC and O₂ collected in the different oceanic locations, it has been estimated that bacterial remineralization of organic matter accounts for approximately one-third of the oxygen deficit. The major difficulty in interpretation of the relationship of DOC versus AOU arises from an estimate of the preformed DOC concentrations when the preformed AOU is at equilibrium with the atmosphere.^123,124

In contrast, another point of view suggests that the similarity in vertical distribution of both parameters is only due to coincidence. Sharp¹²⁵ noted that ‘the DOC–AOU correlation is partially fortuitous since both DOC and dissolved oxygen decrease with depth in the upper few hundred meters of the ocean’. It is a statistical truism that any paired measurement where most of the data is found in two groups, one at each measurement extreme, will display a significant correlation, even if each data group, taken separately, resembles a shotgun pattern. In the surface layer we observe higher concentrations of DOC, mostly the combination of the labile, semi-labile and refractory constituents, and oxygen oversaturation or numbers close to saturation due largely to photosynthesis or ocean/atmosphere exchange. Going deeper, vertical profiles would distinguish older waters which are already depleted in oxygen and degraded in labile and semi-labile DOC.

It seems there is only one option for resolving the problem of the DOC∶AOU ratio. Each water body in the ocean can be characterized by non-conservative properties, from the area of its origin to the area where it mixes with other water masses. One of the best criteria for the definition of different water masses is the density, which is a function of temperature and salinity. To evaluate the contribution of DOC to AOU we have to analyse the molar ratios of both these properties along the surfaces of constant density or isopycnal surfaces. As the water mass spreads laterally along the isopycnal far away from the source of formation, it ages. From a biogeochemical point of view this means that it is characterized by remineralization of labile and semi-labile organic material sinking from the surface. For example, if we study this ratio along the current direction, we need several sections across or to be more precise, a number of sites along this current at different locations with good spatial and temporal resolutions. It is obvious that the best scale for both sections and sites is the mesoscale. Using this approach we can avoid discrepancies with DOC preformed concentrations in the area of water mass formation.

Although the importance of the study of DOC versus AOU ratio along the isopycnal surfaces was formulated a long time ago, the first attempts in this direction have been made only recently. Doval and Hansell,⁴⁷ analyzing this ratio at densities σ_t = 23–27 kg m⁻³ in the South Pacific and in the Indian Oceans at depths <500 m, have shown that the TOC∶AOU molar ratio ranged from –0.15 to –0.34 and from –0.13 to –0.31, respectively. These numbers indicate that TOC oxidation was responsible for 21–47% and 18–43% of oxygen consumption in the upper South Pacific and Indian Oceans, respectively. At greater depths, DOC did not contribute to the development of AOU. They did not find any evidence for significant export of dissolved and suspended organic carbon along the isopycnal surfaces that ventilate the Polar Frontal Zone. Comparing the slopes from Table 2 with estimates from Doval and Hansell,⁴⁷ we can conclude that the numbers from the isopycnal surfaces are close to those found earlier by analysis of vertical profiles of DOC and AOU distributions at different locations in the World Ocean.

4.8. The role of colloidal OM in metal complexation

Colloidal material transports a significant part of DOC (see references above) and trace metals (see references below). Aggregation of colloids from the dissolved to particulate phase represents a major removal process for metals in natural waters.¹²⁶ Colloids are potentially very important in controlling the cycling of many elements that may either be scavenged onto Fe-oxyhydroxides or complexed with large aggregates. Colloidal material is particularly important due to its high specific surface area and, hence, potential adsorptive capacity, which might control the partitioning of an element or compound between dissolved and particulate phases. Colloids can be considered as an additional ligand enriched in different types of functional groups able to complex with trace elements.¹²⁷ One of the most important properties of macromolecular natural organic material is its amphiphilic character, i.e., the fact that it contains both hydrophilic and hydrophobic parts, which can strongly affect its sorption and complexation behavior.¹²⁸

It has been proposed that one of the processes responsible for large variations in scavenging rates could be ‘colloidal pumping’ or ‘Brownian pumping’, i.e., the transfer of dissolved species to large aggregates via colloidal intermediates.¹²⁹ This process involves mainly two steps: (i) the rapid formation of metal–colloid surface site complexes (i.e., adsorption), and (ii) the slow agglomeration and coagulation of these colloids to macroparticles. The ‘colloidal pumping’ process is largely related to the physico-chemical characteristics of the water where the colloids are dispersed. Dai et al.¹²⁷ have noted that the ‘colloidal pumping’ would be more pronounced during estuarine mixing. The colloidal fraction of DOC and trace metals of riverine origin may be removed on the continental shelf or in estuaries. There is some evidence that the behavior of trace elements during estuarine mixing is largely related to their capacity for complexation with organic materials in truly dissolved, colloidal and macroparticulate phases.

COC may be an important mechanism for the stabilization and/or removal of potentially biologically important trace metals, such as iron (e.g., Wells et al.¹³⁰). HMW colloids could be the most reactive component in the cycling of DOC, strongly influencing also the biogeochemistry of many trace elements in estuarine waters.⁶⁷ For example, it has been found that in the Ochlocknee Estuary total Fe and Mn behave non-conservatively; the removal is from the HMW fraction, although Mn is removed at lower salinity than Fe.¹³¹ For Ni, Cu and Cd, the HMW fraction is very important in the river, but these elements are quickly converted from HMW to LMW species with increasing salinity. HMW carbon was strongly correlated with Fe but only weakly correlated with Fe in the smaller size fraction. The two important processes controlling the behavior of metals, carbon and nitrogen in this estuary are colloidal aggregation and desorption or dissociation. Wen et al.¹³² and Wells et al.¹³³ have observed a non-uniform distribution of trace metals within colloidal size fractions and suggested that this was due to specific metal–colloid interaction. COC may help to stabilize trace metals in the upper ocean that are deposited from the atmosphere and/or upwelled from deep waters below.¹⁸ In the Chesapeake Bay and its sub-estuaries Sigleo and Means¹³⁴ have found that organic analyses for amino acids (proteins), carbohydrates and lipids comprised 4–22, 20–60% and less than 1%, respectively, of the COM. The results are significant because amino acids and carbohydrates contain oxygen, nitrogen and sulfur functional groups capable of reacting with trace metals and organic pollutants.

There is also a biological explanation for OM–trace metal complexation as suggested by Azam et al.⁴ Bacteria represent the largest biotic surfaces (>90%) in the ocean; thus surface reactive metals would absorb on bacterial cells. Their ingestion by protozoa would expose the bound metals and radionuclides, since high organic matter density at the hot-spots could absorb metals to high levels. One interesting possibility is that the hydrolytic action of bacteria on the organic ligands releases metals. Small-scale variations in metal distribution could exert growth inhibitory or stimulatory effects on microbial loop organisms as well. Thus, bacteria–OM interaction has implications for control of primary production, OM decomposition and food web structure.⁴

4.9. The role of UV radiation in DOM photochemical oxidation

Recent evidences suggest that stratospheric ozone depletion causes increased incident UV radiation (UVR; 280–400 nm).¹³⁵ This phenomenon is well documented for Antarctic ecosystems.¹³⁶ Increasing incident UV-B radiation (10–20% per decade) has also been reported for temperate latitudes.¹³⁶ Transmission of UVR differs greatly in marine systems and is largely regulated by the concentration and absorptivity of DOM.¹³⁷ The chromophore-containing constituents of DOM (CDOM) in surface waters are important but highly variable components of the total DOC pool. Nelson et al.¹³⁸ have hypothesized that CDOM is produced as a byproduct of the microbial breakdown of DOC and is destroyed due to photo-oxidation. Light absorption by CDOM impacts the aquatic environments through a variety of physical and chemical effects, including the deposition of heat and photochemical consumption of reactive oxygen species, trace gases, and biologically labile organic compounds. Photochemical oxygen consumption has been observed during the oxidation of photochemically produced organic radicals and the photo-oxidation of metal–organic complexes.^139,140 CDOM absorption also affects the UV and visible light field and the passive sensing of water-column components.¹⁴¹

While CDOM compounds can help to shield the water column from the direct effects of UVR, they are themselves subjected to considerable photochemical degradation. Zika¹⁴² has shown that the photochemical transformation of chromophores to an excited state can lead to intramolecular reactions, such as dissociation into radicals, photoisomerization, and intramolecular decomposition, rearrangement, and electron transfer. Other reactions involve photochemically produced free radicals that react with other organic compounds in the water.¹⁴² This degradation is often accompanied by changes in the optical properties of DOC (often referred to as photobleaching). Experimental evidence, mostly obtained with artificial light sources, suggests that photochemical degradation of DOC often reduces DOM absorptivity in the UV range.⁸⁷ Photochemical degradation may proceed to complete photomineralization or may result in production of LMW DOC that in turn may be mineralized through microbial decomposition. The alternative process is when the HMW chemically uncharacterized fractions of DOM may also be modified to more biologically available forms during exposure to natural sunlight.¹⁴³ Photochemical degradation comprises a significant ecological ‘sink’ for UV-absorbing DOC and may potentially affect UV transparency in a number of freshwater ecosystems.¹³⁶ Amon and Benner¹⁴⁴ have shown that photomineralization of biologically refractory riverine DOM appears to be more important than previously believed and could be a major removal mechanism for terrestrially derived DOM in the coastal oceans.

4.10. Monitoring of DOC dynamics from airborne and satellite data on regional and global scales

One of the most obvious prospective directions of marine DOC study lies in the study of bulk DOC by airborne or/and satellite imaging.^141,145,146 It is well known that fluorescence is highly correlated with CDOC absorption at the excitation wavelength despite the presence of multiple sources of CDOM and a changing contribution of CDOM to the total DOC.^147,148 It has been previously shown that samples from the mixed layer exhibit a reasonably well-defined dependence of fluorescence (absorption) on DOC, and this dependence may be developed into an algorithm for estimating DOC.¹⁴¹ Klinkhammer et al.³⁵ have interpreted the data to represent fluorescence from the humic material associated with terrestrial DOM. They were able to convert fluorescent DOM data to quantitative estimates of fluorescent DOC by calibrating the instrument against organic carbon measured by high-temperature combustion. The good agreement between TOC determinations and ZAPS (zero angle photon spectrometer) output is remarkable considering the physico-chemical differences between HTC and fluorescence measurements.³⁵ CDOM can be acquired as absorption in UV–visible range, which is difficult to achieve because of the low sensitivity of the measurements. In earlier works, Hoge et al.,¹⁴⁷ Green and Blough,¹⁴⁹ and Vodacek et al.¹⁴¹ demonstrated that CDOM absorption can be quantitatively retrieved from fluorescence measurements.

For these purposes Vodacek et al.¹⁴¹ used NASA's Airborne Oceanographic Lidar (AOL) with a pulsed frequency tripled Nd∶YAG laser (355 nm) to excite CDOM fluorescence; the upwelled signals are collected and spectrally resolved over a 40 ns gate beginning just prior to the arrival of the laser pulse at the water surface. Thus the pathlength viewed by the sensor is limited to the upper ∼4.5 m of the water column, even if the laser penetrates further. The rapid acquisition of CDOM absorption coefficients from fluorescence measurements is immediately applicable to defining the penetration depths of UV radiation and estimating fluxes of photochemical intermediates and products. With over 3 years of the combined Ocean Color and Temperature Sensor (OCTS) and Sea-Viewing Wide Field-of-View Sensor (SeaWiFS) data available it is now possible to reevaluate the ocean color time series data from these two sensors.¹⁵⁰ The new ocean color sensors are expected to have improved radiometric accuracy and atmospheric correction algorithms compared with the Coastal Zone Color Scanner. Remote sensing application could include fine-scale mapping of surface mixing and development and testing of algorithms for determining CDOM absorption by passive airborne and satellite sensors.

4.11. Other applications of the HTC technique for marine DOC studies

The HTC method has not only been used for DOC study in seawater. The method has been applied to hydrocarbon analysis in waters and soils. The results indicate that the TOC analysis technique, as evaluated, is best suited for samples containing dissolved hydrocarbons in waters and is not a preferred procedure for water samples containing dispersed or floating oil.¹⁵¹ The HTC method has been applied by Pelegri et al.¹⁵² to measure the carbon content of microorganisms. Urban-Rich et al.¹⁵³ employed this technique for analysis of individual copepod fecal pellets. They directly injected individual fecal pellets into the HTC system along with 50 or 100 µl of low carbon seawater and showed that this method gave less uncertainly in the measurement than the standard CHN analysis (cf. 5% and 12–25% respectively).

5. Conclusions

5.1. Future perspectives of marine TOC studies

It is very difficult to predict in which direction the biogeochemistry of marine organic material will go in the future. We can expect that the following research will be focused on the development of new, and the upgrading of existing, analyzers for the wider application of the HTC technique to field research. The most important work will concentrate on the development of almost continuously recording analyzers measuring DOC in real time, as has recently been presented for nutrients. Enormous progress will be made after introduction of the first sophisticated moored DOC arrays allowing detailed observations on small-scale lateral and vertical spatial and temporal resolution (down to the microstructure) of DOC distribution in the ocean. Whenever continuously recording instrumentation has been introduced into the measurement of oceanic variables, the introduction has resulted in the discovery of discontinuities in these variables. There is no reason to believe that the various species of organic carbon will be any different. The continuous analyzer of the future, deployed inside microcosms, will be particularly useful in experiments, for example, with DOM decay during the study of the kinetics of bacterial biodegradation or DOC release by phytoplankton.

Introduction of the moored DOC analyzer will encourage wider distribution of time series stations with continuous DOC monitoring. This type of station already exists. For example, the Bermuda Atlantic Time Series (BATS) and the Hawaiian Ocean Time Series (HOTS) stations provide the oceanographic community with discrete DOC observations. The continuous approach to DOC chemistry opens the possibility for deeper monitoring of the different marine environments including estuarine, coastal and open ocean. Collected data from these stations would help to interpret the dynamics of DOC for better classification of the sources and sinks for bulk DOC.

As has already been noted, to predict the fate of DOC in the ocean and in the global carbon cycle we should identify and distinguish the sources and sinks of this organic carbon pool. This goal can be resolved only by more detailed characterization of the nature of marine organic compounds. We suppose that the research on chemical characterization of different fractions of organic carbon, i.e., TOC, DOC, COC and POC bulk compounds, will be extended through the study of seawater and cultures with phytoplankton and bacteria, to other potential sources of DOC in seawater. We should better understand the roles which natural (river, rain, sediment, aeolian, metal complexation, UV radiation) and anthropogenic sources in marine environments play in the cycling of organic carbon.

The results of the field and experiment investigations would be incorporated in predictive models of organic carbon cycling at the regional and global scales. It is obvious that all kind of models would incorporate a wide range of parameters affecting the nature of the DOC in marine environments. The most important of them are: the microbial loop, phytoplankton release, grazers and victim interaction, metal complexation, UV radiation, and vertical and horizontal circulation. This approach requires increased interdisciplinary collaborations between physical, biological and chemical oceanographers.

Acknowledgements

We thank Edward Urbansky for the opportunity to contribute to this special issue of the Journal of Environmental Monitoring. This paper is a contribution of number 258 from the Center for Marine Sciences, UNCW. The junior author (PJW) acknowledges research support from the Natural Sciences and Engineering Research Council of Canada.

References

1. U. Siegenthaler and J. L. Sarmiento, Nature, 1993, 365, 119.
2. J. I. Hedges, Nature, 1987, 330, 205.
3. M. J. Church, H. W. Ducklow and D. M. Karl, Limnol. Oceanogr., 2002, 47(1), 1.
4. F. Azam, L. B. Fandino, H.-P. Grossart and R. A. Long, Rapp. Comm. Int. Mer. Médit., 1998, 35, 2 , and references cited therein.
5. C. A. Bookman, T. J. Culliton and M. A. Warren, Trends in the U.S. Coastal Regions, 1970–1998. Addendum to the Proceedings, ‘Trends and future challenges for U.S. National and Coastal policy’, August 1999, U.S. Department of Commerce, NOAA, National Ocean Service, Silver Spring, MD, 1999, p. 3. http://state-of-coast.noaa.gov/natdialog/index.html.
6. P. A. Wheeler, M. Gosselin, E. Sherr, D. Thibault, D. L. Kirchman, R. Benner and T. E. Whitledge, Nature, 1996, 380, 697.
7. D. A. Hansell and C. A. Carlson, Nature, 1998, 395, 263.
8. A. Skoog, R. Lara and G. Kattne, Deep-Sea Res. I, 2001, 48, 2613.
9. K. Y. Børsheim, S. M. Myklestad and J.-A. Sneli, Mar. Chem., 1999, 63, 255.
10. I. Bussmann and G. Kattner, J. Mar. Syst., 2000, 27, 209.
11. J. M. Lobbes, H. P. Fitznar and G. Kattner, Geochim. Cosmochim. Acta, 2000, 64, 2973.
12. G. Cauwet and I. Sidorov, Mar. Chem., 1996, 53, 211.
13. X. A. Álvarez-Salgado and A. E. Miller, Mar. Chem., 1998, 62, 325.
14. P. Kähler and W. Koeve, Deep-Sea Res. I, 2001, 48, 49.
15. M. D. Doval, X. A. Álvarez-Salgado and F. F. Pérez, Mar. Ecol.: Prog. Ser., 1997, 157, 21.
16. E. V. Dafner, R. Sempéré, M. Goutx and N. González, Mar. Ecol.: Prog. Ser., 1999, 189, 301.
17. M. D. Doval, X. A. Álvarez-Salgado and F. F. Pérez, J. Mar. Syst., 2001, 30, 33.
18. M. H. Dai and C. R. Benitez-Nelson, Mar. Chem., 2001, 74, 181 , and references cited therein.
19. R. F. Chen, B. Fry, C. S. Hopkinson, D. J. Repeta and E. T. Peltzer, Cont. Shelf Res., 1996, 16, 409.
20. C. S. Hopkinson, B. Fry and A. L. Nolin, Cont. Shelf Res., 1997, 17, 473.
21. L. Guo and P. H. Santschi, Mar. Chem., 1997, 59, 1.
22. C. A. Carlson, H. W. Ducklow and A. F. Michaels, Nature, 1994, 371, 405.
23. R. Benner and S. Opsahl, Org. Geochem., 2001, 32, 597.
24. C. Thomas, G. Cauwet and J. F. Minster, Mar. Chem., 1995, 49, 155.
25. G. Copin-Montégut and B. Avril, Deep-Sea Res. I, 1993, 40, 1963.
26. S. Tugrul, Mar. Chem., 1993, 41, 265.
27. R. Sempéré, C. Panagiotopoulos, R. Lafont, Marroni and F. Van Wambeke, J. Mar. Syst. , in press.
28. S. C. Yoro, R. Sempéré, C. Turley, M. A. Unanue, X. Durrieu De Madron and M. Bianchi, Mar. Ecol.: Prog. Ser., 1997, 161, 255.
29. G. Cauwet, A. Miller, S. Brasse, G. Fengler, R. F. C. Mantoura and A. Spitzy, Deep-Sea Res. II, 1997, 3–4, 769.
30. M. D. Doval, F. F. Pérez and E. Bardalet, Deep-Sea Res. I, 1999, 46, 511.
31. R. Sempéré and E. V. Dafner, 1997, unpublished observations.
32. E. V. Dafner, R. Sempéré and H. Bryden, Mar. Chem., 2001, 73, 223.
33. D. A. Hansell and E. T. Peltzed, Deep-Sea Res. II, 1998, 45, 2171.
34. S. Sardessai and S. N. de Sousa, Deep-Sea Res. II, 2001, 48, 3353.
35. G. P. Klinkhammer, J. McManus, D. Colbert and M. D. Rudnicki, Geochim. Cosmochim. Acta, 2000, 64, 2765.
36. H. Ogawa and N. Ogura, Nature, 1992, 356, 696.
37. P. M. Williams and E. R. M. Druffel, Nature, 1987, 330, 246 , and references cited therein.
38. R. Benner, J. D. Pakulski, M. McCarthy, J. I. Hedges and P. G. Hatcher, Science, 1992, 255, 1561.
39. L. M. Tupas, B. N. Popp and D. M. Karl, Mar. Chem., 1994, 45, 207.
40. D. A. Hansell, C. A. Carlson, N. R. Bates and A. Poisson, Deep-Sea Res. II, 1997, 44, 2115.
41. D. A. Hansell and T. Y. Waterhouse, Deep-Sea Res. I, 1997, 44, 843.
42. E. T. Peltzer and N. A. Hayward, Deep-Sea Res. II, 1996, 43, 1155.
43. C. A. Carlson and H. W. Ducklow, Deep-Sea Res. II, 1995, 42, 639.
44. D. A. Hansell, P. M. Williams and B. B. Ward, Deep-Sea Res. I, 1993, 40, 219.
45. A. Ning Loh and J. E. Bauer, Deep-Sea Res. I, 2000, 47, 2287.
46. J. E. Bauer, E. R. M. Druffel, D. M. Wolgast, S. Griffin and C. A. Masiello, Deep-Sea Res. II, 1998, 45, 689.
47. M. D. Doval and D. Hansell, Mar. Chem., 2000, 68, 249.
48. E. Dafner, R. Sempéré, S. C. Yoro, A. Agatova and G.Cauwet, C. R. Acad. Sci., Ser. IIA: Sci. Terre Planetes, 1999, 329, 345.
49. C. J. Wiebinga and H. J. W. De Baar, Mar. Chem., 1998, 61, 185.
50. P. Kähler, P. K. Bjørnsen, K. Lochte and A. Antia, Deep-Sea Res. II, 1997, 44, 341.
51. H. Ogawa, R. Fukuda and I. Koike, Deep-Sea Res. I, 1999, 46, 1809.
52. E. V. Dafner and P. J. Wangersky, J. Environ. Monit., 2002, 4, 48.
53. P. E. Kepkay, Mar. Ecol.: Prog. Ser., 1994, 109, 292 , and references cited therein.
54. L. Guo and P. H. Santschi, Mar. Chem., 1996, 55, 113.
55. I. Koike, H. Shigemitsu, T. Kazuki and K. Kazuhiro, Nature, 1990, 345, 242.
56. M. L. Wells and E. D. Goldberg, Mar. Chem., 1993, 41, 353.
57. W.-C. Chin, M. V. Orellana and P. Verdugo, Nature, 1998, 391, 569.
58. R. M. W. Amon and R. Benner, Nature, 1996, 387, 166.
59. D. J. Carlson, M. L. Brann, T. H. Mague and L. M. Mayer, Mar. Chem., 1985, 16, 155.
60. P. H. Santschi, L. Guo, M. Baskaran, S. Trumbore, J. Southon, T. S. Bianche, B. Honeyman and B. Cifuentes, Geochim. Cosmochim. Acta, 1995, 59, 625.
61. E. R. M. Druffel, P. M. Williams, J. E. Bauer and J. R. Ertel, J. Geophys. Res., 1992, 7, 15639 , and references cited therein.
62. J. D. Willey, R. J. Keiber, M. S. Eyman and G. B. Avery Jr., Global Biogeochem. Cycl., 2000, 14, 139 , and references cited therein.
63. J. L. Hedges, R. G. Keil and R. Benner, Org. Geochem., 1997, 27, 195.
64. W. Ludwig and J. L. Probst, Global Biogeochem. Cycl., 1996, 10, 23 , and references cited therein.
65. V. Ittekkot, Nature, 1988, 332, 436.
66. R. Benner and J. I. Hedges, Mar. Chem., 1993, 41, 161.
67. L. Guo and P. H. Santschi, Rev. Geophys., 1997, 35, 17.
68. R. J. Kieber, M. F. Rhines, J. D. Willey and G. B. Avery Jr., Atmos. Environ., 1999, 33, 3659.
69. G. E. Likens, E. S. Edgerton and J. N. Galloway, Telus, 1983, 35B, 16.
70. R. Sempéré and K. Kawamura, Atmos. Environ., 1996, 30, 1609.
71. K. Kawamura, S. Steinberg, L. Ng and I. R. Kaplan, Atmos. Environ., 2001, 35, 3917.
72. R. G. Keil, D. B. Montlucon, F. G. Prahl and J. I. Hedges, Nature, 1994, 370, 549 , and references cited therein.
73. R. A. Berner, Am. J. Sci., 1982, 282, 451.
74. J. D. Milliman, Deep-Sea Res. I, 1994, 41, 797 , and references cited therein.
75. S. C. Wainright, Science, 1987, 238, 1710 , and references cited therein.
76. T. Komada and C. E. Reimers, Mar. Chem., 2001, 76, 155.
77. S. Otto and W. Balzer, Prog. Oceanogr., 1998, 42, 127.
78. D. J. Burdige, M. J. Alperin, J. Homstead and C. S. Martens, Geophys. Res. Lett., 1992, 19, 1851.
79. L. Guo and P. H. Santschi, Geochim. Cosmochim. Acta, 2000, 64, 651.
80. P. A. Jumars, L. P. Deboran, J. A. Baross, M. J. Perry and B. W. Frost, Deep-Sea Res., 1989, 36, 483.
81. W. Chen and P. J. Wangersky, J. Plankton Res., 1996, 18, 1521 , and references cited therein.
82. B. Biddanda and R. Benner, Limnol. Oceanogr., 1997, 42, 506 , and references cited therein.
83. L. I. Aluwihare, D. J. Repeta and R. Chen, Nature, 1997, 387, 166.
84. W. O. Smith, C. A. Carlson, H. W. Ducklow and D. A. Hansell, Mar. Ecol.: Prog. Ser., 1998, 168, 229.
85. S. L. Strom, R. Benner, S. Ziegel and M. J. Dagg, Limnol. Oceanogr., 1997, 42, 1364 , and references cited therein.
86. D. J. Kieber, J. McDaniel and K. Mopper, Nature, 1989, 341, 637.
87. K. Mopper, X. L. Zhou, R. J. Kieber, D. J. Kieber, R. J. Sikorski and R. D. Jones, Nature, 1991, 353, 60.
88. P. J. Wangersky, Mar. Chem., 1993, 41, 61.
89. J. A. Fuhrman, Mar. Ecol.: Prog. Ser., 1987, 37, 45.
90. D. A. Hansell and C. A. Carlson, Global Biogeochem. Cycl., 1998, 12, 443.
91. D. A. Hansell, N. R. Bates and K. Gundersen, Mar. Chem., 1995, 51, 201.
92. J. E. Bauer, P. M. Williams and E. R. M. Druffel, Nature, 1992, 357, 667.
93. M. H. C. Stoll, H. M. Van Aken, H. J. De Baar and M. Kraak, Mar. Chem., 1996, 55, 217.
94. E. V. Dafner, M. González-Dávila, J. M. Santana-Casiano and R. Sempéré, Deep-Sea Res. I, 2001, 48, 1217.
95. Medoc Group, Nature, 1970, 227, 1037.
96. N. Stommel, in Studies in Physical Oceanography, ed. A. L. Gordon, Gordon and Breach, New York, 1972, vol. 2.
97. W. S. Broecker, Oceanography, 1991, 4, 79.
98. C. Andrie and L. Merlivat, Deep-Sea Res., 1988, 35, 247.
99. J. P. Christensen, T. T. Packard, F. Q. Dortch, H. J. Minas, J. C. Gascard, C. Richez and P. C. Garfield, Global Biogeochem. Cycles, 1990, 3, 315.
100. H. Lacombe, J. C. Gascard, J. Gonella and J. P. Bethoux, Oceanol. Acta, 1981, 4, 247.
101. J. C. Gascard and C. Richez, Prog. Oceanogr., 1985, 15, 157.
102. K. Stratford and R. G. Williams, J. Geophys. Res., C, 1997, 102C, 12539.
103. R. Sempéré, B. Charrière, F. Van Wambeke and G. Cauwet, Global Biogeochem. Cycles, 2000, 14, 669.
104. T. T. Packard, H. J. Minas, B. Coste, R. Martinez, M. C. Bowin, J. Gostan, P. Garfield, J. Christensen, D. Dortch, M. Minas, G. Copin-Montégut and C. Copin-Montégut, Deep-Sea Res., 1988, 35, 1111.
105. C. Lee and S. M. Henrichs, Mar. Chem., 1993, 41, 105.
106. G. R. Harvey, D. A. Boran, L. A. Chesal and J. M. Tokar, Mar. Chem., 1983, 12, 119.
107. P. P. Newton and P. S. Liss, Deep-Sea Res., 1989, 36, 759.
108. P. H. Santschi, E. Balnis, K. J. Wilkinson, J. Zhang, J. Buffle and L. Guo, Limnol. Oceanogr., 1998, 43, 896.
109. F. Azam, T. Fenchel, J. G. Field, J. S. Gray, L. A. Meyer-Reil and F. Thingstag, Mar. Ecol.: Prog. Ser., 1983, 10, 257.
110. F. Azam, D. C. Smith, G. F. Steward and Å. Hagström, Microb. Ecol., 1993, 28, 167.
111. W. Chen and P. J. Wangersky, Mar. Chem., 1992, 41, 167.
112. D. M. Karl, G. A. Knauer and J. H. Martin, Nature, 1988, 332, 438 , and references cited therein.
113. B. C. Cho and F. Azam, Nature, 1988, 332, 441.
114. A. C. Redfield, B. H. Ketchum and F. A. Richards, in The Sea, ed. M. N. Hill, Interscience, New York, 1963, vol. 2.
115. Y. Sugimura and Y. Suzuki, Mar. Chem., 1988, 24, 105.
116. E. R. M. Druffel, P. M. Williams and Y. Suzuki, Geophys. Res. Lett., 1989, 16, 991.
117. M. D. Kumar, A. Rajendran, K. Somasundar, B. Haake, A. Jenisch, Z. Shuo, V. Ittekkot and B. N. Desai, Mar. Chem., 1990, 31, 299.
118. J. E. Bauer, P. M. Williams, E. R. M. Druffel and Y. Suzuki, Eos Transition, 1990, 71, 154.
119. E. Tanoue, Earth Planet. Sci. Lett., 1992, 111, 201.
120. P. E. Kepkay and M. L. Wells, Mar. Ecol.: Prog. Ser., 1992, 80, 275.
121. H. J. W. De Baar, C. Brussaard, J. Hegeman, J. Schijt and M. N. C. Stoll, Mar. Chem., 1993, 41, 145.
122. J. H. S. Martin and S. E. Fitzwater, Nature, 1992, 356, 699.
123. J. R Toggweiler, in Productivity in the ocean: Past and Present, ed. W. F. Berger, V. S. Smetachek and G. Wefer, Wiley, New York, 1989.
124. J. R. Toggweiler, Nature, 1992, 356, 665.
125. J. H. Sharp, Mar. Chem., 1997, 56, 265.
126. E. R. Sholkovitz, Earth Planet. Sci. Lett., 1978, 41, 77; B. D. Honeyman and P. H. Santschi, J. Mar. Res., 1989, 47, 951.
127. M. H. Dai, J.-M. Martin and G. Cauwet, Mar. Chem., 1995, 51, 159.
128. P. H. Santschi, J. J. Lenhart and B. D. Honeyman, Mar. Chem., 1997, 58, 99.
129. J. T. Honeyman and P. H. Santschi, J. Mar. Res., 1989, 47, 951.
130. M. L. Wells, L. M. Mayer, O. F. X. Donard, M. M. de Souza Sierra and S. G. Ackleson, Nature, 1991, 353, 248.
131. R. T. Powell, W. M. Landing and J. E. Bauer, Mar. Chem., 1996, 55, 165.
132. L.-S. Wen, P. Santschi, G. Gill and C. Paternosto, Mar. Chem., 1999, 63.
133. M. L. Wells, G. J. Smoth and K. W. Bruland, Mar. Chem., 2000, 71, 143.
134. A. C. Sigleo and J. C. Means, Rev. Environ. Contam. Toxicol., 1990, 112, 123.
135. T. Graedel, Nature, 1989, 342, 621.
136. D. P. Morris and B. R. Hargreaves, Limnol. Oceanogr., 1997, 42, 239 , and references cited therein.
137. R. G. Zepp, in Humic Substances and Their Role in the Environments, ed. F. Frimmel and R. Christman, Wiley, New York, 1988.
138. N. Nelson, D. A. Siegel and A. F. Michaels, Deep-Sea Res. I, 1998, 45, 931.
139. C. Langford, M. Wingham and V. Sastri, Environ. Sci. Technol., 1973, 7, 820.
140. O. C. Zafiriou, J. Joussot-Dubien, R. G. Zepp and Zika, Environ. Sci. Technol., 1984, 18, 358A.
141. A. Vodacek, F. E. Hoge, R. N. Swift, J. K. Yungel, E. T. Peltzer and N. V. Blough, Limnol. Oceanogr., 1995, 40, 411 , and references cited therein.
142. R. G. Zika, in Marine Organic Chemistry: Evolution, Composition, Interactions, and Chemistry of Organic Matter in Seawater, ed. E. Duursma and R. Dawson, Elsevier, Amsterdam, 1981, 31, 299.
143. W. L. Miller and M. A. Moran, Limnol. Oceanogr., 1997, 42, 1317.
144. R. M. W. Amon and R. Benner, Geochim. Cosmochim. Acta, 1996, 60, 1783 , and references cited therein.
145. G. M. Ferrari, M. D. Dowell, S. Grossi and C. Targac, Mar. Chem., 1996, 55, 299.
146. F. E. Hoge, C. W. Wright, R. N. Swift, J. K. Yungel, R. E. Berry and R. Mitchell, Deep-Sea Res. II, 1998, 45, 1073.
147. F. E. Hoge, A. Vodacek and N. V. Blough, Limnol. Oceanogr., 1993, 38, 1334.
148. V. I. Esteves, E. B. Santos and A. C. Duarte, J. Environ. Monit., 1999, 1, 251.
149. S. A. Green and N. V. Blough, Limnol. Oceanogr., 1994, 39, 1903.
150. M. Kahru and G. Mitchell, J. Geophys. Res., C, 2001, 106C, 2517.
151. P. Lambert, M. Fingas and M. Goldthorp, J. Hazard. Mater., 2001, 83, 65.
152. S. P. Pelegrí, J. Dolan and F. Rassoulzadegan, Aquat. Microb. Ecol., 1999, 16, 273–280.
153. J. Urban-Rich, D. A. Hansell and M. Roman, Mar. Ecol.: Prog. Ser., 1998, 171, 199.

---

Footnote

† For Part I see ref. 52.

---