Environ. Sci. Technol. 2001, 35, 2682-2689
A Potential Source of Organic Pollutants into the Northeastern Atlantic: The Outflow of the Mediterranean Deep-Lying Waters through the Gibraltar Strait S O L E D A D M A R T IÄ , J O S E P M . B A Y O N A , A N D J O A N A L B A I G EÄ S * Department of Environmental Chemistry, CID-CSIC, Jordi Girona Salgado, 18-26, E-08034-Barcelona, Spain
Small and large-size particles were collected in the water column (50-3000 m) of a Northeastern Atlantic area where deep Mediterranean waters, outflowing through the Strait of Gibraltar, are incorporated at mid-depth. Particles collected by water filtration (0.7 µm pore size) and by vertical hauls of a neuston net (50 µm mesh size) were analyzed for organic pollutants, namely aliphatic and aromatic hydrocarbons, and organochlorine compounds. Smallsize particles represented the largest bulk of particulate organic carbon as well as of hydrophobic organic pollutants. Surface concentrations of n-alkanes (C14-C35), aromatic hydrocarbons (12 parent compounds), PCBs (7 congeners), and DDTs (DDT+DDE) were, respectively, in the range of 50-63 ng/L, 23-68 pg/L, 8-13 pg/L, and 0.05-1.7 pg/L. These concentrations showed a general decrease with depth, particularly significant in the upper 200 m, consistently with the POC contents. Compositional changes with depth were also evident in small-size particles and included the depletion of low molecular weight n-alkanes and low chlorinated PCB congeners as well as a decrease of the fossil to pyrolytic PAHs ratio. Unusual increases of concentrations were observed at mid-depths (900-1100 m), indicating additional particle inputs, either by in-situ formation or by advective transport from the Mediterranean. The latter was recognized because small-size particles within these water veins exhibited distribution patterns out of the vertical sequence and similar to those of deep Mediterranean waters. An input of 8 and 0.5 tons per year of the above PAH and PCB compounds, respectively, has tentatively been calculated as the contribution of these Mediterranean waters to the Northeastern Atlantic.
Introduction North America and Europe are believed to be major source areas for a variety of hydrophobic organic contaminants (HOCs), namely hydrocarbons and polychlorobiphenyls (PCBs), that may find their final repository in the North Atlantic Ocean (1, 2). Hydrocarbons are introduced into the open sea predominantly by accidental or intentional discharges from oil tankers and by atmospheric fall-out, the latter contributing also significantly to the input of organochlorinated compounds, like PCBs and DDTs (3, 4). * Corresponding author phone: +34-93-400 6142; fax: +34-93204 5904; e-mail:
[email protected]. 2682
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 13, 2001
However, the studies on HOCs in the open sea, and particularly in the Atlantic, are scarce mainly because of the poor reliability of analytical techniques. Accurate determinations of organic contaminants in the North Atlantic have recently been performed under the impulse of the Intergovernmental Oceanographic Commission (IOC) (5) and estimations of downward fluxes have also been provided (6, 7). These studies made evident that the ocean is, effectively, an important sink for continentally derived particulated HOCs and that both, autochthonous and allochthonous particulate material, may play an important role in their transport and fate. Moreover, the mixing of different water masses and the scavenging of species from solution onto particulate matter appear to be major controlling factors of the concentrations of HOCs in the open ocean, besides the codistillation concept suggested by Iwata et al. (8). In this respect, the Eastern Atlantic has a permanent exchange of waters with the Mediterranean Sea through the Strait of Gibraltar, and taking into account the particular hydrogeographical conditions of the Mediterranean that have resulted in higher levels of contamination compared with other regional seas (9), a major question has raised concerning the significance of this exchange as a source of organic pollutants for the Northeastern Atlantic. The waters exchange results from the evaporation of the Mediterranean waters, which become more dense than those of the North Atlantic and generate an inflow of fresher and lighter Atlantic upper waters into the Mediterranean and an outflow of more saline and denser deep Mediterranean water to the Atlantic (10). Despite its initially very high density, the Mediterranean water outflow does not reach the bottom of the North Atlantic because it entrains a substantial volume of the overlying Atlantic waters while still in the Gulf of Cadiz. Thus, the resultant mixed Mediterranean waters that finally reach the open North-central Atlantic, in the form of water lenses (eddies), become neutrally buoyant at depths between 900 and 1200 m (11). A field survey was conducted by Price et al. (12) to characterize the physical dynamics of this Mediterranean outflow and, more specifically, to know where and how the outflow is modified by mixing. More recently, and as part of the French JGOFS project, we have undertaken the MEDATLANTE experiment for the biogeochemical characterization of these Mediterranean water veins (13). Based on the lipid composition of the small-size particulate matter we were able to recognize the signature of the Mediterranean water discharges in the mid-depth Northeastern Atlantic. However, it is well-known that the particulate material also plays an important role in the transport of hydrophobic pollutants in the marine environment (14). This paper reports on the measurement of organic pollutants, such as aliphatic and aromatic hydrocarbons and organochlorinated compounds (e.g. PCBs and DDTs), in large (>50 µm) and smallsize (>0.7 µm) particles collected in the Northeastern Atlantic, in the region where the Mediterranean waters are recognized. Large-size particles are responsible for most of the vertical flux of organic pollutants through the water column, whereas small-size particles, basically following the movement of the water masses in which they are incorporated, may contribute to the advective transport of pollutants.
Experimental Section Sample Setting. Samples were collected in January and August 1989, on board the R/V Jean Charcot, at the sites shown in Figure 1. The first cruise studied the formation of the so-called Nicole water lens, resulting from the coalescence 10.1021/es000258p CCC: $20.00
2001 American Chemical Society Published on Web 06/01/2001
TABLE 2. Locations and Characteristics of Samples Collected in the MEDATLANTE Summer Cruisea
FIGURE 1. Area of study showing the Mediterranean vein water at about 1000 m depth and the main lens-like structures studied during the MEDATLANTE I and II cruises. Numbers refer to sampling stations indicated in Tables 1 and 2.
TABLE 1. Locations and Characteristics of Samples Collected in the MEDATLANTE Winter Cruisea ref
date
latitude N
longitude W
water depth (m)
sample typeb
2A 3A 4A 4B 5A 7A 7B 7C 8A 1a 2a 3a 3b 4a 4b 5a 7a 8a 8b
Jan. 11, 1989 Jan. 10, 1989 Jan. 11, 1989 Jan. 11, 1989 Jan. 21, 1989 Jan. 6, 1989 Jan. 6, 1989 Jan. 6, 1989 Jan. 23, 1989 Jan. 1, 1989 Jan. 11, 1989 Jan. 9, 1989 Jan. 10, 1989 Jan. 11, 1989 Jan. 11, 1989 Jan. 21, 1989 Jan. 6, 1989 Jan. 23, 1989 Jan. 23, 1989
35° 53.1′ 35° 51.9′ 36° 31.0′ 36° 31.0′ 36° 55.0′ 36° 25.0′ 36° 25.0′ 36° 25.0′ 36° 20.2′ 36° 41.1′ 35° 53.1′ 35° 51.9′ 35° 51.9′ 36° 31.0′ 36° 31.0′ 36° 55.0′ 36° 25.0′ 36° 20.2′ 36° 20.2′
6° 34.7′ 7° 5.8′ 6° 52.2′ 6° 52.2′ 10° 39.9′ 10° 35.3′ 10° 35.3′ 10° 35.3′ 10° 10.0′ 1° 12.2′ 6° 34.7′ 7° 5.8′ 7° 5.8′ 6° 52.2′ 6° 52.2′ 10° 39.9′ 10° 35.3′ 10° 10.0′ 10° 10.0′
328-480 0-900 0-100 100-350 850-1400 0-80 80-500 500-720 600-1400 1000 450 200 900 50 200 1100 1000 1200 1700
L L L L Lc L L L L s s s s s s sc sc s s
a See the experimental part for the sample type description. b L: large (>50 µm); s: small (>0.7 µm). c Samples corresponding to the Mediterranean water lenses.
of three Mediterranean water veins which, from Gibraltar, followed different paths along the Cadiz slope. The second cruise was focused on the sampling of a pair of more distant and, therefore, more older lenses, called Tristan and Yseult, the latter possibly corresponding to the evolution of Nicole after 6 months. The precise location of the sampling stations are indicated in Tables 1 and 2. Temperature and salinity profiles were recorded at each station with a CTD for plotting the potential temperature-salinity diagrams, which allowed the recognition of the Mediterranean eddies (meddies), at a depth of around 1000 m (15, 16). Stations 1 and 9 were located in the Alboran Sea for sampling, respectively, winter and summer Mediterranean deep waters. Station 2, at the outlet of the Gibraltar Strait, represented Mediterranean waters in their way toward the Atlantic and the vertical profiles of stations 3 and 4 consisted of Atlantic waters, the upper 100 m, in station 3, receiving the influence of the Spanish rivers, Guadiana and Guadalquivir. Stations 8/10 and 11/13 were selected for sampling, respectively, deep and surface Atlantic waters. Finally, stations 5 and 7 and 12-14 included at mid-depths (900-1200 m) the Mediterranean water lenses.
ref
date
latitude N
longitude W
water depth (m)
particle sizeb
12A 13A 13B 13C 9a 10a 11a 12a 13a 13b 13c 14a 14b
Aug. 18, 1989 Aug. 22, 1989 Aug. 23, 1989 Aug. 23, 1989 Aug. 11, 1989 Aug. 20, 1989 Aug. 19, 1989 Aug. 18, 1989 Aug. 22, 1989 Aug. 22, 1989 Aug. 22, 1989 Aug. 31, 1989 Aug. 31, 1989
36° 20.0′ 36° 15.0′ 36° 15.0′ 36° 15.0′ 36° 44.2′ 36° 28.9′ 36° 29.6′ 36° 20.0′ 36° 15.0′ 36° 15.0′ 36° 15.0′ 36° 2.9′ 36° 2.9′
16° 0.6′ 15° 30.1′ 15° 30.1′ 15° 30.1′ 1° 10.0′ 15° 45.3′ 16° 0.3′ 16° 0.6′ 16° 30.1′ 16° 30.1′ 16° 30.1′ 16° 27.8′ 16° 27.8′
920-1012 0-200 800-1200 1500-3000 1000 1200 95 1100 76 700 1000 800 2000
Lc L Lc L s s s sc s s sc sc s
a See the experimental part for the sample type description. b L: large (>50 µm); s: small (>0.7 µm). c Samples corresponding to the Mediterranean water lenses.
Sampling. Two types of samples were collected for the isolation of large (L) and small (s) particles. Large particles were sampled with a neuston net (47 µm mesh size) by vertical hauls through a predetermined transect, with a speed of about 50 cm/s. The water volume sampled was estimated multiplying the inlet surface of the net by the hauling distance. The collected material, that included some living zooplankton, faecal pellets, and detritus, was vacuum sucked for removing the water, and the filter was wrapped in aluminum foil and stored at -20 °C until analysis. Small-size biogenic and detrital particles were the largely predominant material collected with a stainless steel bottle of 220 L of capacity, that was deployed to the selected depth, filled with water, closed automatically, and carried to the surface. Once there, a Teflon impeller pump was connected to the bottle outlet and the water was filtered through a precombusted (500 °C overnight) Whatman (Kent, UK) glass fiber filter of 293 mm of diameter (0.7 µm pore size) in a closed circuit and in a clean laboratory, to avoid external contamination. The filters were folded, wrapped with cleaned aluminum foil, and stored at -20 °C, until they were processed in the laboratory. Chemical Analysis. The filters were spiked with cholestane, octachloronaphthalene and deuterated pyrene (as surrogates) and Soxhlet extracted during 12 h, in the dark, with 50 mL of a mixture of methylene chloride-methanol (2:1). The extracts were vacuum evaporated to almost dryness and fractionated in a 3% water deactivated silica-alumina column, 1:1 (30 cm × 0.5 cm i.d.). The following fractions were recovered: (I) aliphatic hydrocarbons (6 mL of nhexane); (II) organochlorinated compounds and mono- and dicyclic PAHs (6 mL of hexane:dichloromethane, 9:1); and (III) tri- to hexacyclic PAHs (6 mL of the same mixture, 8:2). All fractions were evaporated to dryness and dissolved in 100 µL of isooctane for further analysis. Aliphatic Hydrocarbons. The first fraction was analyzed by gas chromatography (GC) using a Carlo Erba 5300 HRGC, equipped with a FID detector and splitless injector. A fused silica column of 30 m × 0.25 mm i.d. coated with 5% phenylmethyl-polysiloxane DB5 (J&W, Scientific, USA) was used. The operating conditions were as follows: injector and detector temperatures, 300 °C and 330 °C, respectively; column temperature programmed from 60 °C to 100 °C at 15 °C/min, from 100 °C to 300 °C at 4 °C/min, and 15 min at the final temperature. The carrier gas was hydrogen (50 cm/s). n-Alkanes of 14, 22, 32, and 36 carbon atoms were used as external standards for quantification. The Unresolved Complex Mixture (UCM) was quantified using the response factor of the n-alkane eluting in the zone of maximum response (usually C23-C26). VOL. 35, NO. 13, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2683
Polycyclic Aromatic Hydrocarbons. Aliquots of fractions II and III were analyzed by GC coupled to mass spectrometry (GC-MS) (Fisons 800) in the electron impact (EI) mode with selected ion monitoring (SIM) and with full scan for confirmation. Helium was the carrier gas at 30 cm/s. The column, a 30 m × 0.25 mm i.d. CP-Sil 5 CB fused silica (Chrompack, Middelburg, NL), was programmed from 90 °C to 300 °C at a rate of 6 °C/min, keeping the final temperature for 15 min. A standard mixture containing 12 individual PAHs (phenanthrene, anthracene, fluoranthene, pyrene, benz[a]anthracene, chrysene, benzo[b]fluoranthene, benzo[k]fluoranthene, benzo[a]pyrene, indeno-[1,2,3-cd]pyrene, dibenz[ah]anthracene and benzo[ghi]perylene) was used for calibration. Organochorinated Hydrocarbons. The second fraction was analyzed by GC-ECD (Hewlett-Packard 5890) equipped with a HP 7673A data system. The same column as for the aromatic hydrocarbons was used and was programmed from 80 °C to 180 °C at 15 °C/min and from 180 °C to 280 °C at 3 °C/min, keeping the final temperature for 15 min. The carrier gas was helium at a linear flow-rate of 50 cm/s, and the makeup was nitrogen (30 mL/min). The injector and detector temperatures were 280 and 310 °C, respectively. Quantitation was performed using an external standard calibration mixture of selected congeners (IUPAC nos: 28, 52, 101, 118, 153, 138, and 180) supplied by Promochem (Wesel, Germany) as well as the chlorinated compounds p,p′DDT, p,p′-DDE, R-HCH, and HCB. Analytical Performance. This was implemented as customary (17, 18). Procedural blanks and control samples were processed in the same manner as real samples, and they were below 5% of the analyte abundance. Duplicates and standard mixtures were included between batches of four samples. Recoveries calculated from the surrogates were higher than 80%. The GC injections were performed with an automatic injector to improve reproducibility that was better than 14% for the whole procedure. The whole protocols were validated through the participation in intercalibration exercises (UNEP-IOC-IAEA) and the analysis of a certified reference material (SH-4, NRC of Canada). Concentrations of individual PAH and PCB compounds were within the confidence intervals of the reported values.
Results and Discussion General Features of the Particulate Material. Particles in seawater may have different sources and are present in different sizes. Small-size particles collected by water filtration represented by far the largest bulk of particulate organic carbon (POC), as it is the case in other oceanic regions (19). POC associated with these particles (5-50 µg/L) was 2-3 orders of magnitude higher than that associated with larger particles (0.01-0.32 µg/L). Strong spatial and seasonal variations were also observed (13). Thus, surface small particles collected in January accounted for 15-50 µg/L of POC, whereas in August they represented only 5-15 µg/L, reflecting the temporal variability of primary production (20). At the same time, a significant decrease of POC concentrations with depth was observed in both types of particles (13). To characterize the biogeochemical evolution of the outflow of these particles from the Mediterranean toward the NE Atlantic, the lipid composition was determined and reported elsewhere (13). In summary, the larger particles were found to be dominated by zooplankton markers, whereas the small ones showed evidence of a mixed algal composition with, possibly, a bacterial contribution. The vertical profiles of the different lipid classes showed a general decrease with depth, particularly significant in the upper 200 m, consistently with the POC contents. Compositional changes were more evident in small-size particles and included the loss of unsaturated species and the increase of 2684
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 13, 2001
diagenetic and bacterial markers. Unusual increases of concentrations were observed at mid-depths indicating additional particle inputs, either by in-situ formation or by advective transport from the Mediterranean. Microscopic observations of net-collected particles revealed mainly the presence of zooplanktonic organisms (copepods) in the photic zone and increasing amounts of nonmigrating carnivorous microfauna, faecal pellets, and detritus with depth. Below 1000 m water depth the detritus and faecal pellets accounted for more than 70% of the particles (A. Saliot, personal communication). Aliphatic Hydrocarbons. The GC profiles of the surface large-size particles were largely dominated by two resolved peaks, namely pristane and cis-heneicosahexaene (C21:6), which can be attributed, respectively, to several species of copepods and planktonic algae (21-23). Phytadienes were abundant as well, being possibly the result of chlorophyll grazing by zooplankton. Small-size particles contained the same compounds although at lower abundances. The occurrence of these biogenic species was temporally and spatially dependent, particularly in large-size particles. Higher concentrations were found in January and in the photic zone, according to the primary productivity trend (20). In addition to the above components, the series of n-alkanes was also present in all samples, although the distributions exhibited several differences between both types of particles, as shown in Figure 2. Thus, large-size particles from the open Atlantic surface (7A) showed distributions characterized by a predominance of n-C15 and n-C17 and a series of n-C20-C30, without odd-even carbon number predominance. Such distributions may point to inputs from microalgae and cyanobacteria (24). A slight odd-even carbon number predominance in the range of n-C27-C35 was indicative of a certain contribution of higher plant waxes. The profiles of deeper samples (7B) were more uniform, probably reflecting a certain n-alkane degradation during their downward transport through the water column. Matsueda and Handa (25) reported similar distributions of alkanes in sinking particles collected in the Eastern North Pacific open waters and suggested that bacteria grown on these particles was the most probable source. These patterns were qualitatively similar both in winter (shown in Figure 2) and in summer. In turn, small-size particles were strongly depleted in lower n-alkanes, probably as a consequence of their higher exposure to degradation. Most profiles displayed either a modal distribution of higher alkanes with no carbon number preference, and maximizing around n-C20-C24, or a distribution with a significant odd-even carbon number predominance in the n-C27-C33 range (13a). Biogenic sources, respectively microbial and terrestrial, can be invoked for this type of profiles. The distributions in deeper samples reflected a stronger compositional decay (10a). The profiles corresponding to the Mediterranean particulate outflow (2A and 2a) exhibited a marked odd-even carbon number predominance in the n-C23-C33 range, indicative of terrestrial (higher plants) inputs. This was also the main characteristic of the small particles collected in the Mediterranean lenses recognized in the Atlantic at about 1000 m water depth (5a), a feature that was not observed in samples collected, at similar depths, in typical Atlantic waters (10a). On the contrary, large particles (12A) displayed n-alkane patterns consistent with those found in the overlying waters (e.g. 7A), indicating a certain coupling of these particle pools. Furthermore, most large-size particle samples evidenced the concurrence of an unresolved complex mixture (UCM) of aliphatic hydrocarbons, which may be indicative of petrogenic sources. These were confirmed by mass fragmentography with the identification of sterane and triterpane markers. Petrogenic contributions were particularly evident
FIGURE 2. n-Alkane distributions in large (L) and small (s) particles of selected stations. Numbers refer to the n-alkane carbon numbers. in samples influenced by the Mediterranean waters (e.g. stations 2A and 12A). However, they were only measurable in large-size particles, a feature that may reflect a better preservation of petrogenic hydrocarbons by physical packaging within planktonic faecal pellets. n-Alkane concentrations were found mostly in the range of 1-50 ng/L and 0.01-0.45 ng/L in small and large-size particles, respectively, with higher values in January than in August. Concentrations in small particles of Western Mediterranean open surface waters (30 m) were in the range of 6-34 ng/L (18), similar to those found in stations 2-4, and of 0.8-6.8 ng/L in the Alboran deep seawaters (stations 1 and 9). On the other hand, Dachs et al. (18) found concentrations of 48-131 ng/L of unresolved aliphatic hydrocarbons in small particles of the Western Mediterranean surface, consistently with those previously reported in the area (26-28). The concentrations in large particles from surface Atlantic waters (stations 4A, 7A, and 13A) ranged from 2.5 to 3.8 ng/L and decreased to 0.2 ng/L at around 2000 m water depth (station 13C), whereas the Mediterranean influenced waters (stations 2A and 12A) exhibited concentrations of 4.7-5.3 ng/L. The UCM was below the detection level in small particles. Moreover, a low correlation between the UCM and n-alkanes was observed (R2 ) 0.3006) as a result of the predominant biogenic origin of the latter. Polycyclic Aromatic Hydrocarbons (PAHs). PAH distributions were generally dominated by phenanthrene and its methyl derivatives (Figure 3), which revealed a significant contribution of fossil fuel residues (29). It is possible that these patterns reflect a chronic petrogenic pollution, easily accumulated by surface microbiota and originated by accidental oil spills or tanker ballast operations in Atlantic open seawaters. Indeed, large-size particles showed a more significant petrogenic signature, supported by the concurrence of an UCM of aliphatic hydrocarbons in this type of samples. However, it is interesting to notice that the surface distributions exhibited a marked seasonality, with concentrations in winter almost twice that in summer and slightly
depleted in the highly alkylated components (13a). Besides the consequences of the primary productivity on the accumulation of these hydrocarbons, the lower ambient temperature and a higher contribution of pyrogenic sources in winter could also account for these features. Pyrolytic PAHs (e.g. benzofluoranthene and benzopyrene isomers) were present in all samples, being relatively apparent in large particles (e.g. 7A). The fluoranthene/pyrene ratio also showed the signature of combustion inputs, and the predominance of BePy over BaPy was consistent with aged aerosol sources. Finally, the low abundance of perylene, which may be produced diagenetically, may also denote a pyrogenic source. Pyrolytic PAHs are primarily introduced into the open seawaters by atmospheric deposition, bound to soot particles, and once there they may be trapped by marine particles. In this respect, zooplankton faecal pellets have been proposed as a primary mechanism for the rapid transfer of PAHs from the productive surface waters to deep waters and bottom sediments in many areas (30-32). These PAH distributions of mixed origin were similar to those found in particles of different sizes collected in Mediterranean open waters (33, 34) but contrasted with those found in deep sediments, where the pyrolytic higher molecular weight PAHs predominated (35). To explain this apparent inconsistency it has been suggested that a more efficient uptake of fossil fuel than pyrolytic derived PAHs in the photic zone and a faster recycling of the former in the water column and in the sediment-water interface may occur (36, 37). The profiles corresponding to the Mediterranean particulate outflow (2A and 2a) exhibited, as in the case of the n-alkane distributions, some terrestrial signatures, consistent with the hydrogeological characteristics of the basin. Thus, retene was clearly represented in these samples as well as in those of the meddies. This phenathrene derivative is mostly associated to the diagenesis of resin acids of vascular plants (38), although a contribution from conifer combustion cannot be excluded (39). PAH concentrations, including only the main nine parent compounds indicated in Figure 3, were found in the range VOL. 35, NO. 13, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2685
FIGURE 3. Representative distributions of PAHs in large (L) and small (s) particles of selected stations: Ph, phenanthrene; MPh, methylphenanthrenes; DMPh, dimethylphenanthrenes; TMPh, trimethylphenanthrenes; A, anthracene; Fl, fluoranthene; Py, pyrene; BaA, benz[a]anthracene; Ch, chrysene; BFl, benzofluoranthenes; BePy, benzo[e]pyrene; BaPy, benzo[a]pyrene; Per, perylene; IPy, indeno-[1,2,3cd]pyrene; and BPer, benzo[ghi]perylene.
TABLE 3. Concentrations (in pg/L) of Representative Fossil and Pyrolytic PAHs and PCB Congeners, in Small-Size Particles of Open Seawaters Western Med. a 15 m 1500 m phenanthrene fluoranthene pyrene chrysene benzofluoranthenes PBC No. 28 PBC No. 52 PBC No. 101 PBC No. 118 PBC No. 153 PBC No. 138 PBC No. 180 a
References 17 and 34.
45 10 10 2.8 1.2 0.4 0.8 1.0 1.9 2.6 0.4 0.5 b
8 5 7 1.5 1.6 0.06 0.12 0.08 0.18 0.15 0.04 0.06
Alboran Sea (stations 1, 9) 1000 m 0.3-1.6 1-7 0.3-12 1.1-15 1.2-23 0.1-0.2 1.4-1.8 1.2-2.3 1.4-1.9 0.6-0.9 0.9-1.4 0.2-0.3
9
43-54 6-50 27-41 8-13 c 0.01-0.15 0.07-0.5 0.03-1.03 0.06-1.0 0.07-1.2 0.04-1.4 0.05-0.8
7-20 5-10 5-11 c c 0.02 0.02-0.07 0.01-0.08 0.03-0.1 0.04-0.1 0.04-0.09 0.03-0.06
this work
