Environ. Sci. Technol. 2005, 39, 9440-9445
Spatially Complex Distribution of Dissolved Manganese in a Fjord as Revealed by High-Resolution in Situ Sensing Using the Autonomous Underwater Vehicle Autosub P. C. J. N. J. P.
J . S T A T H A M , * ,† D . P . C O N N E L L Y , † R. GERMAN,† T. BRAND,‡ O. OVERNELL,‡ E. BULUKIN,† MILLARD,† S. MCPHAIL,† M. PEBODY,† PERRETT,† M. SQUIRE,† STEVENSON,† AND A. WEBB†
National Oceanography Centre, Southampton, European Way, Southampton SO14 3ZH, U.K., and Scottish Association for Marine Science, Dunstaffnage Marine Laboratory, Oban, Scotland
Loch Etive is a fjordic system on the west coast of Scotland. The deep waters of the upper basin are periodically isolated, and during these periods oxygen is lost through benthic respiration and concentrations of dissolved manganese increase. In April 2000 the autonomous underwater vehicle (AUV) Autosub was fitted with an in situ dissolved manganese analyzer and was used to study the spatial variability of this element together with oxygen, salinity, and temperature throughout the basin. Six along-loch transects were completed at either constant height above the seafloor or at constant depth below the surface. The ca. 4000 in situ 10-s-average dissolved Mn (Mnd) data points obtained provide a new quasi-synoptic and highly detailed view of the distribution of manganese in this fjordic environment not possible using conventional (water bottle) sampling. There is substantial variability in concentrations (600 nM) and distributions of Mnd. Surface waters are characteristically low in Mnd reflecting mixing of riverine and marine end-member waters, both of which are low in Mnd. The deeper waters are enriched in Mnd, and as the water column always contains some oxygen, this must reflect primarily benthic inputs of reduced dissolved Mn. However, this enrichment of Mnd is spatially very variable, presumably as a result of variability in release of Mn coupled with mixing of water in the loch and removal processes. This work demonstrates how AUVs coupled with chemical sensors can reveal substantial small-scale variability of distributions of chemical species in coastal environments that would not be resolved by conventional sampling approaches. Such information is essential if we are to improve our understanding of the nature and significance of the underlying processes leading to this variability.
* Corresponding author phone: +44 (0)23 8059 2679; fax: +44 (0)23 8059 3059; e-mail:
[email protected]. † National Oceanography Centre. ‡ Scottish Association for Marine Science. 9440
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 24, 2005
Introduction Most of the processes influencing the biogeochemistry of near shore marine systems are characterized by short time and space scales (1). However, our understanding of such systems has been largely based on sparse data from onedimensional data sets in which the lateral variability is poorly defined or unknown. If significant horizontal mixing and concentration gradients exist, then the vertical variability in concentrations at one point will give a poor representation of the broader distribution of the material of interest and the wider applicability of the conclusions drawn from onedimensional data will be uncertain. Where modern highresolution techniques have been applied (2), new insights into the biogeochemistry of the system studied have resulted. Thus there is clearly a need for more detailed survey data to better define variability of chemical species and to allow a better understanding of the physical, biological, and chemical processes in operation. Previous systems that have provided high-frequency chemical data on marine systems have used moored buoys, towed bodies with shipboard analysis, or manned submersibles (3). Relative to these techniques, autonomous underwater vehicles (AUVs) have many advantages as instrument platforms to study spatial variations. These include the ability to reach inaccessible areas, e.g., under ice (4), no risk to operator, the capability of carrying out complex three-dimensional missions over increasingly long ranges, and the potential to carry a range of instrument payloads. The potential of AUVs coupled to a range of sensors has already been noted by the marine science community and they have been used in, for example, optical and biological studies (5, 6) and are planned to be incorporated into operation of future sea floor observatories (e.g., ORION). Autosub is an example of AUV technology that has been developed at the National Oceanography Centre, Southampton, and used as a platform for a wide variety of instruments in science missions (7); here we describe its novel use in a loch when coupled to an in situ chemical analyzer to investigate the biogeochemistry of Mn.
Experimental Section Study Site. Loch Etive is a narrow sea loch on the west coast of Scotland (Figure 1), which has two main basins with classic fjordic features. The upper basin is separated from the lower basin by a sill at 13 m depth, and this feature allows isolation of waters in the deepest part (maximum depth 145 m) behind the sill for periods of, on average, 16 months (8). A recent side-scan sonar and bathymetric survey of the upper basin (9) shows a range of morphological features common to such fjord systems ranging from granitic outcrops and glacial erratics to zones of fine-grained sediments in the deeper basins. In a complementary paper the biogeochemistry of Mn in the upper basin of Loch Etive is discussed by Overnell et al. (10), based on data from a single station that was reoccupied on ca. monthly intervals between June 1999 and December 2000. In the isolated deep waters behind the sill oxygen decreases due to primarily benthic respiration, and dissolved manganese increases to nearly micromolar concentrations, as compared to the low nanomolar concentrations typical of coastal waters. While it was possible to follow the dynamic removal of oxygen and to discuss the underlying Mn geochemistry, the work drew on data from vertical profiles at a single station. As a result there was no information on variations in manganese distributions spatially through the upper basin of the Loch, and so it was not possible to assess 10.1021/es050980t CCC: $30.25
2005 American Chemical Society Published on Web 10/21/2005
FIGURE 1. Loch Etive. The darker zone indicates the extent of the deep basins behind the sill in the southwestern part of the upper loch. how representative the data from the deep basin were, and what other processes might be influencing the geochemistry of Mn. In the present paper we describe the use of Autosub, equipped with an in situ dissolved Mn sensor, to investigate the quasi-synoptic 3D spatial variability of Mn and O2 in the upper basin of Loch Etive, and consider the implications of this new data. Methods. Autosub is a 7-m long autonomous underwater vehicle that is capable of carrying a wide range of instruments. In the version used in this study, depth capability was 500 m and duration was >1 day. The major features and operation of Autosub have been described in detail in Millard et al. (7). A description of the in situ dissolved Mn analyzer package and its interfacing with Autosub is given in Statham et al. (11). In overview, the analyzer was based on the spectrophotometric detection of the 1-(2-pyridylazo)-2-naphthol (PAN)-Mn complex at 560 nm in a continuous flow system (12). Reactive Mn is determined by this method, and this fraction is expected to be primarily dissolved Mn(II) although labile colloidal Mn may also be measured; the data are reported as dissolved Mn (Mnd). Dissolved iron can interfere at high concentrations, and as the Loch Etive system would be expected to have elevated Fe under low oxygen conditions, an iron-specific chelating reagent (desferrioxamine B) was also added to the reagent stream. The effectiveness of this approach is demonstrated by the ability of the method to
work in hydrothermal plumes (13) where Fe concentrations are highly elevated. The light transmission of the seawaterMn complex was measured in a 50-mm path-length cell using a green LED as light source (Toshiba TLPGA183P), and a silicon photodiode device that converted light intensity to frequency (TAOS TSL235) as detector. The system operated continuously, with a time delay of 10 s between sample being taken in and the analytical measurement for this sample being recorded. Data were logged on the onboard data acquisition system of Autosub at ca. 1 Hz. Solutions were propelled through the manifold using a peristaltic pump. To allow in situ calibration of the system, either a blank seawater solution or a 1000 nM Mn standard could be switched into the chemistry manifold instead of the sample solution for a short period on a ca. 48 min cycle, using a series of electrically operated valves. During the missions reported here an electromechanical problem arose with the switching valves and calibration was done using off-line measurements of Mn in samples collected during the mission (see below). Reagents, standards, and blanks were held in previously acidcleaned plastic 500-mL blood transfusion bags. The manifold, pump, and valves were housed in a pressure-balanced acrylic tube filled with silicone oil, while the intake to the analyzer was positioned on the top of Autosub, and was fitted with a coarse filter to prevent large material from entering and blocking the manifold system. The detection limit estimate (2σ) for the analyzer, based on the precision of sequential low-concentration measurements in a homogeneous part of the loch, was 25 nM. An estimate of the precision of the method was given by averaging a set of data points in a relatively homogeneous part of the loch at 70 m, which resulted in a mean concentration of 271 nM with a standard deviation of 8.4%; this value may also reflect some natural variations in the system. An Aquamonitor water sampling system (WS Oceans) was deployed in Autosub at the same time as the in situ analyzer to allow water collection for subsequent metal analyses and further calibration of the instrument. Every 4 min a 200-mL sample was collected into an acid-cleaned blood bag. Samples were filtered in the ship’s laboratory immediately after collection and acidified for preservation. Details of analysis for suspended particulate matter (SPM) concentration, and dissolved and leachable (1 M HCl) Mn are given in Overnell et al. (10). The Mnd data presented in this paper are based on calibration of the sensor using analysis of Aquamonitor samples collected from zones of constant Mnd concentration in the loch. The use of an independent solvent extraction/ atomic absorption method validated with a certified reference material demonstrates the accuracy of the in situ measurements. Additionally there is good correspondence of distributions and concentrations of Mn in the Loch with separately collected and analyzed samples (10). Salinity, temperature, and dissolved oxygen were measured using a Seabird 911 system mounted on Autosub; the oxygen sensor calibration was checked onboard ship using air-saturated water of known salinity and temperature for the high value and sulfitecontaining water as the zero. The analyzer system was deployed with Autosub in the loch on April 6 and 7, 2000 from the ship Terschelling. Autosub was deployed on several occasions, and in these missions the main modes of use were bottom hugging (i.e., constant depth from bottom) and constant depth below surface, along the axis of the loch. In one mission (230) a square wave pattern transect along loch was used to give information on lateral as well as axial variations in the wider, upper part of the loch. These missions are listed in Table 1, and mission tracks are shown in the later distribution plots. Autosub typically traveled at 1 m/s during these missions. Data from the Mnd analyzer was averaged over 10-s periods, giving a spatial resolution of about 10 m. The time of Mnd measurement was VOL. 39, NO. 24, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
9441
TABLE 1. Missions Completed during the Study Autosub sampling date mission April 6, 2000
226
April 6, 2000
227
April 6, 2000 April 7, 2000
228 230
April 7, 2000
231
description bottom hugging (between 10 and 20 m above seafloor) constant depth from surface (100 and then 80 m) constant depth from surface (70 m) constant depth from surface (80 m) with square wave lateral pattern; near to shore Autosub came closer to surface to avoid the bottom constant depth from surface (20 m)
offset from the remaining data by 10 s to allow for color development and mixing in the instrument, thus making the Mnd data contemporaneous with the T, S, and oxygen data. The combined transects provide a thorough three-dimensional coverage of the upper basin of the Loch.
Results The water mass structure of Loch Etive was typical of fjord like systems with the following: (1) the surface layer being a mixture of river waters and saline waters from outside the loch, (2) waters below sill depth being isolated between overturn events, and (3) an intermediate mixing layer reflecting a gradual decrease in deep water salinity through primarily tidally driven mixing with low-salinity surface waters (8, 10). The Loch was visited with Autosub at the end of a period of deep-water isolation (partial and then full deepwater renewal occurred later in April, and then early May, respectively (10)). During this isolation there had developed significant oxygen depletion in the deep waters and concurrent increases in dissolved manganese. Strong halo- and thermo-clines were present, with the main changes in properties between 30 and 80 m. Salinity and temperature in the deep basin (>90 m) were remarkably constant, averaging 27.20 ( 0.12 (1σ), and 12.11 ( 0.07 (1σ) °C respectively, and the lack of variability infers a well-mixed water column after the deepwater turnover event which led to the formation of this water mass. The upper water column (down to about 30 m) was well oxygenated (>300 µM), while deep waters (>80 m) had low concentrations of oxygen of about 50 µM. Slight differences in the low oxygen concentrations were evident between different parts of the loch in these deep waters. Between the deep and surface layers was a mixing zone with intermediate values of oxygen, with the most pronounced concentration gradients between 30 and 45 m. Some oxygen was present in all waters sampled in the deeper part of the loch. Data with respect to time into the mission from the bottom hugging transect (M226; varying between 10 and 20 m off bottom) along the length of the Loch are shown in Figure 2. Note that during this transect Autosub moved off the center, deepest, line of the loch and moved to shallower waters, and to maintain the distance above the seabed moved up in the water column. Thus the transect line between about 1000 and 2000 seconds does not follow the deepest line through the main basin. The major time gap in the record corresponds to Autosub returning to the surface for routine checks, and minor gaps correspond to programmed in situ calibration periods where no data on surrounding waters were collected. Major features in the record are the clear increase in Mnd in the deeper basins. However, there is not a simple correlation of Mnd with depth, thus demonstrating variability in Mnd within the deep basins. There are also increased Mnd concentrations (>100 nM) toward the head of the loch in some shallow water areas, with associated reductions in dissolved oxygen as compared to surface water concentra9442
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 24, 2005
tions of e25 nM at the southwestern, seaward, end of the loch. The square wave pattern track of Autosub across and parallel to the main axis of the loch at 80 m depth (M230) also shows lateral variation in Mnd (Figure 3), with highest concentrations adjacent to the southeastern side of the loch and freshwater inputs. On this square wave transect Autosub moved to shallower waters at either side of the loch to maintain a safe distance above bottom. The overall vertical distribution of Mnd along the main axis of the upper basin of Loch Etive is shown in Figure 4. This figure is constructed using the full set of Mnd data points (ca. 4000 values, each averaged over 10 s) from the Autosub transects in Table 1. While a few features in data poor zones may reflect the contouring package, the remaining information gives a detailed and striking picture of smallscale variability superimposed on the general feature of high Mnd in the deep basins. The large number of data points and their variations in concentrations over short space scales indicate real changes that are not reflections of geometric anisotropy introduced by the contouring program. Elevated concentrations of Mnd in shallower waters, particularly toward the head of the loch, are evident in this figure as well as in Figure 2.
Discussion A detailed interpretation of the complex distributions shown in this data set is beyond the scope of the present paper. However, a preliminary discussion is provided below to demonstrate that the information is consistent with our current understanding of Mn cycling in such systems, thus supporting the validity of the data, as well as providing the basis for new insights into the geochemistry of this element. Distributions of Dissolved Manganese. In the surface lower salinity waters of the loch Mnd concentrations were generally less than 25 nM, as one would anticipate in such loch systems with mixing of low Mn concentration marine and river waters (14). The elevated Mnd distributions in the deeper main water-body of the loch are clearly nonhomogeneous both horizontally and vertically, and these features must reflect the net effect of input, removal, and transport processes. In the interpretation of the cycle of Mn in Etive based on changes at one vertical profile with time, Overnell et al. (10) demonstrated the importance of sedimentary sources of dissolved reduced Mn(II) and subsequent water column oxidation to solid phase Mn(IV) oxides. As all waters in the loch contain some dissolved oxygen, the only obvious mechanism for in situ water-column reduction would be through the presence of micro-reducing zones, and while this may occur, organic matter in suspension appears low and sedimentary sources of Mnd are expected to dominate in this system. The oxidation of Mnd is directly dependent on the concentrations of dissolved and particulate Mn, oxygen, and pH (15), and is auto-catalyzed by particulate Mn. Thus Overnell et al. (10) predict short removal times for Mnd (order of 7-10 days) in the deeper part of their profile, which, although having low oxygen concentrations, has the highest Mnp and Mnd. Once converted to the particulate form, Mnp can aggregate to a size at which it will sediment out to the seafloor, and where it may remain or be rereduced. In the study of Overnell et al. (10) the total Mn inventory in the deep water column was not conservative, and as overall there is a net increase in dissolved Mnd through the period of isolation, inputs must exceed removal. The most southwesterly zone of the Loch (sampling depth about 90 m) had slightly higher oxygen and lower Mnd than the central zone (sampling depth about 135 m); this behavior presumably reflects greater respiration and intensity of reduction of Mn to dissolved forms in the deepest basin. An example of relatively high concentrations of Mnd and decreases in oxygen in shallow (20-30 m) waters is evident
FIGURE 2. Data from the along-Loch transect, M226. Autosub followed the seafloor along the long axis of the loch, and therefore the data reflect Mnd and dissolved oxygen concentrations at varying depths in this system. Data are plotted against time into the mission, and the transect extends from the southwest to the northeast part of the Loch.
FIGURE 3. Two-dimensional lateral distribution of Mnd at 80 m depth (transect M230). Arrows indicate the input point of the named rivers to the loch. toward the northeast limits of the M226 transect. These features infer reducing sediments and release of Mnd well above the deeper waters of the main basin. For Mission 226 as a whole, which provides the best range of data and depths of all the missions, and is the only transect to reach the deepest part of the Loch, the data give only a scattered inverse relationship between Mnd and oxygen. However, when the data are broken down into separate zones more consistent variations are evident, and these data are plotted in Figure 5. In the deepest basin Mnd shows a wide range of values for only a small range in dissolved oxygen values. Variability in release of Mnd with location, coupled with differing rates of oxidation of the reduced Mnd that is dependent on Mnp, oxygen, and pH, will give changes in concentration of Mnd. The low oxygen values observed are expected to be driven by benthic respiration and oxidation of reduced chemical species. An additional factor that may enhance variability in dissolved and particulate Mn is disturbance of sediment at the side of the water body through, e.g., internal waves and advection of released material along isopycnals toward the
center of the loch. Such lateral advection has been demonstrated in lake systems (16), and the slow oxidation rate of Mnd could allow its use as a tracer of physical processes in such systems. In the shallow higher-oxygen waters at the head of the loch (M226) the elevated Mnd and reduced oxygen concentrations appears to reflect shallow benthic respiration and Mnd release plus mixing in these near surface waters. Sedimentary Sources of Mn. The variability in the Mnd data presented here indicates changes in Mn input to the water column across the loch. A sedimentary source of Mnd will be largely driven by the redox conditions present, which in turn is dependent on the bacterially driven oxidation of organic carbon within the sediments. Once oxygen and nitrate are exhausted then oxidized manganese will be reduced to release Mnd (17, 18). While additional factors such as bioturbation (10, 19) can enhance release of Mnd, the major factor influencing Mnd release will be the amount of labile carbon in the sediment.The two significant sources of organic carbon to sediments in this pristine system are phytoplankton and riverine inputs. However, in these west-coast loch systems the work of Overnell and Young (20) for the adjacent Loch Linnhe clearly shows that the seasonal carbon inputs from primary production are much less important than riverine particulate organic carbon (POC) inputs. They estimated that 83% of the total carbon flux to the sediments was from terrigenous sources, while only 17% was of phytoplankton origin. These estimates are consistent with Abril et al. (21) who report that the predominance of terrestrially derived POC over algal POC is a common feature of estuarine environments not influenced by anthropogenic contamination. The Loch Etive system has a very substantial watershed of 1400 km2 (8), and in the upper basin a series of rivers (Noe, Kinglass, and Liver) and streams feed into the loch. The River Awe that enters midloch (southern end of the upper basin investigated here) is the major river input to Loch Etive, and freshwater is pushed into the upper basin during the flood tide. Each freshwater source will have a loading of POC derived from the pristine watershed, and also particles produced on flocculation of dissolved organic matter (DOM) VOL. 39, NO. 24, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
9443
FIGURE 5. Dissolved Mn and oxygen from different sections of upper Loch Etive during mission M226.
FIGURE 4. Vertical section of (a) Mnd and (b) oxygen along the axis of the upper basin of Loch Etive. The 10-second average data points used in this plot are shown in black on the figure. Note that with the square wave lateral pattern track at 80 m, Autosub came closer to the surface near-shore to avoid the bottom and therefore the track covered a range of depths. as river water enters the saline regime. Much of this particulate material will be expected to deposit close to the mouth of the rivers and through slumping, gravity flows, re-suspension, and other mechanisms (22, 23) much eventually will be transported to the deep basins in the loch. Thus these riverine inputs are expected to leave a significant imprint of carbon-containing sediments in the loch, and such sedimentary features are evident in sidescan data from the loch (9). Overnell et al. (24) studied benthic respiration and organic matter concentrations in four Scottish sea lochs and within Loch Etive they found elevated respiration at the site nearest the head of the loch, which was ascribed to the degradation of riverine particulate carbon. When the location of significant riverine inputs to Etive are plotted in a plan view there is a striking correspondence between high Mnd and zones influenced by these river inputs (Figure 3 and at head of loch, Figure 2). These data thus suggest that particulate carbon introduced via rivers may be an important carbon source driving benthic respiration, sub-oxia, and Mnd release. This situation in the pristine Loch Etive is in contrast to eutrophic coastal systems such as the northern Adriatic Sea, where anthropogenic releases of nutrients lead to large amounts of highly labile phytoplankton detritus being formed 9444
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 24, 2005
that exceed riverine carbon sources, and its decay leads to redox changes and release of dissolved Mn and Fe from sediments (25). In the present study the impact of river derived carbon on redox conditions is suggested even in the shallower parts of the loch (right section of Figure 2) where elevated Mnd is coincident with small reductions in dissolved oxygen. These findings are consistent with those of Hall (26) for surface waters in the nearby Loch Linnhe where occasional high Mnd concentrations of up to 200 nM were observed near river sources and edges of the Loch, and were ascribed to benthic inputs from reducing sediments at the head and around the edge of the loch. Thus processes leading to release of Mnd appear to be important in both deep basin and surface waters, but in surface waters the signals are expected to be more ephemeral because of the enhanced mixing and transport relative to the deeper waters. These observations provide a new dimension to our knowledge of spatial distributions of Mn in this system, and raise important questions over carbon inputs and Mn cycling in such coastal areas. The overall highly complex distribution of Mnd revealed by the AUV surveys is the net result of variable sources, cycling, and mixing processes, and further detailed process studies are now needed to fully elucidate the mechanisms leading to the observed complex distributions. There are important implications from the work reported here for sampling strategies in dynamic marine waters, where previously limited data collection procedures have been used. Without a detailed knowledge of the spatial variability our understanding of processes controlling these distributions and elemental budgets will remain limited, at best. This study clearly demonstrates the power of combining AUVs with in situ measurement systems for defining the spatial variability of reactive elements in coastal waters. To the authors’ knowledge, this is the first example of use of an in situ chemical analyzer on an AUV, and indicates the great potential of this approach for environments extending from tropical estuaries to polar regions at the ice-land interface.
Acknowledgments The Captain and crew of Terschelling were a delight to work with and their positive and supportive attitude was greatly appreciated. Colleagues in DML are thanked for their help and support, including Dave Meldrum. Constructive comments from three anonymous reviewers improved the
manuscript and we are grateful to them. The work was supported by NERC grant GST/02/2145.
Literature Cited (1) Tokar, J. M.; Dickey, T. D. Chemical sensor technology: current and future applications. In Chemical Sensors in Oceanography; Varney, M. S., Ed.; Gordon and Breach: Amsterdam, 2000; pp 303-329. (2) Klinkhammer, G. P.; McManus, J. Dissolved manganese in the Columbia River estuary: Production in the water column. Geochim. Cosmochim. Acta 2001, 65 (17), 2835-2841. (3) Varney, M. S., Ed. Chemical Sensors in Oceanography; Gordon and Breach: Amsterdam, 2000. (4) Brierley, A. S.; Fernandes, P. G.; Brandon, M. A.; Armstrong, F.; Millard, N. W.; McPhail, S. D.; Stevenson, P.; Pebody, M.; Perrett, J.; Squires, M.; Bone, D. G.; Griffiths, G. Antarctic krill under sea ice: Elevated abundance in a narrow band just south of ice edge. Science 2002, 295 (5561), 1890-1892. (5) Brown, C. A.; Huot, Y.; Purcell, M. J.; Cullen, J. J.; Lewis, M. R. Mapping coastal optical and biogeochemical variability using an autonomous underwater vehicle and a new bio-optical inversion algorithm. Limnol. Oceanogr. Methods 2004, 2, 262281. (6) Yu, X. R.; Dickey, T.; Bellingham, J.; Manov, D.; Streitlien, K. The application of autonomous underwater vehicles for interdisciplinary measurements in Massachusetts and Cape Cod Bays. Cont. Shelf Res. 2002, 22 (15), 2225-2245. (7) Millard, N. W.; Griffiths, G.; Finnegan, G.; Mcphail, S. D.; Meldrum, D. T.; Pebody, M.; Perrett, J. R.; Stevenson, P.; Webb, A. T. Versatile autonomous submersibles - the realising and testing of a practical vehicle. Underwater Technol. 1998, 23 (1), 7-17. (8) Edwards, A.; Edelsten, D. J. Deep water renewal of Loch Etive. Estuarine, Coastal Mar. Sci. 1977, 5, 575-595. (9) Howe, J. A.; Overnell, J.; Inall, M. E.; Wilby, A. D. A side-scan sonar image of a glacially overdeepened sea loch, upper Loch Etive, Argyll. Scott. J. Geol. 2001, 37, 3-10. (10) Overnell, J.; Brand, T.; Bourgeois, W.; Statham, P. J. Manganese dynamics in the water column of the upper basin of Loch Etive, a Scottish fjord. Estuarine, Coastal Shelf Sci. 2002, 55, 481-492. (11) Statham, P. J.; Connelly, D. P.; German, E. B.; Millard, N.; McPhail, S.; Pebody, M.; Perrett, J.; Squires, M.; Stevenson, P.; Webb, A. Mapping the 3D spatial distribution of dissolved manganese in coastal waters using an in situ analyzer and the autonomous underwater vehicle Autosub. Underwater Technol. 2003, 25 (3), 129-134. (12) Chin, C. S.; Johnson, K. S.; Coale, K. H. Spectrophotometric Determination of Dissolved Manganese in Natural-Waters with 1-(2-Pyridylazo)-2-Naphthol - Application to Analysis in Situ in Hydrothermal Plumes. Mar. Chem. 1992, 37 (1-2), 65-82. (13) Chin, C. S.; Coale, K. H.; Elrod, V. A.; Johnson, K. S.; Massoth, G. J.; Baker, E. T. In Situ Observations of Dissolved Iron and Manganese in Hydrothermal Vent Plumes, Juan-De-Fuca Ridge. J. Geophys. Res.-Solid Earth 1994, 99 (B3), 4969-4984.
(14) Hall, I. R.; Hydes, D. J.; Statham, P. J.; Overnell, J. Dissolved and particulate trace metals in a Scottish sea loch: An example of a pristine environment? Mar. Pollut. Bull. 1996, 32 (12), 846854. (15) Yeats, P. A.; Strain, P. M. The oxidation of manganese in seawater: rate constants based on field data. Estuarine, Coastal Shelf Sci. 1990, 31, 11-24. (16) MacIntyre, S.; Flynn, K. M.; Jellison, R.; Romero, J. R. Boundary mixing and nutrient fluxes in Mono Lake, California. Limnol. Oceanogr. 1999, 44 (3), 512-529. (17) Slomp, C. P.; Malschaert, J. F. P.; Lohse, L.; VanRaaphorst, W. Iron and manganese cycling in different sedimentary environments on the North Sea continental margin. Cont. Shelf Res. 1997, 17 (9), 1083-1117. (18) Van Cappellen, P.; Viollier, E.; Roychoudhury, A.; Clark, L.; Ingall, E.; Lowe, K.; Dichristina, T. Biogeochemical cycles of manganese and iron at the oxic-anoxic transition of a stratified marine basin (Orca Basin, Gulf of Mexico). Environ. Sci. Technol. 1998, 32 (19), 2931-2939. (19) Aller, R. C. Bioturbation and Manganese Cycling in Hemipelagic Sediments. Philos. Trans. R. Soc. London Ser. A 1990, 331 (1616), 51-68. (20) Overnell, J.; Young, S. Sedimentation and Carbon Flux in a Scottish Sea Loch, Loch Linnhe. Estuarine, Coastal Shelf Sci. 1995, 41 (3), 361-376. (21) Abril, G.; Nogueira, M.; Etcheber, H.; Cabec¸ adas, G.; Lemaire, E.; Brogueira, M. J. Behaviour of Organic Carbon in Nine Contrasting European Estuaries. Estuarine, Coastal Shelf Sci. 2002, 54 (2), 241-262. (22) McCave, I. N. Erosion, transport and deposition of fine-grained marine sediments. In Fine-Grained Sediments: Deep-Water Processes and Facies; Piper, D. J. W., Stow, D. A. V., Eds.; Blackwell: Oxford, 1984; pp 35-69. (23) Pierce, J. W. Suspended sediment transport at the shelf break and over the outer margin. In Marine Sediment Transport and Environmental Management; Stanley, D. J., Swift, D. J. P., Eds.; Wiley: New York, 1976; pp 437-458. (24) Overnell, J.; Harvey, S. M.; Parkes, R. J. A biogeochemical comparison of sea loch sediments. Manganese and iron contents, sulphate reduction and oxygen uptake rates. Oceanol. Acta 1996, 19 (1), 41-55. (25) Tankere, S. P. C.; Statham, P. J.; Price, N. B. Biogeochemical cycling of Mn and Fe in an area affected by eutrophication: The Adriatic Sea. Estuarine, Coastal Shelf Sci. 2000, 51 (4), 491506. (26) Hall, I. R. Cycling of trace metals in coastal waters: biogeochemical processes involving suspended particles. Ph.D. Thesis, Department of Oceanography, University of Southampton, 1993; 329 pp.
Received for review May 24, 2005. Revised manuscript received September 16, 2005. Accepted September 22, 2005. ES050980T
VOL. 39, NO. 24, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
9445
