Heavy Metal Contamination and Mixing Processes in Sediments from the Humber Estuary, Eastern England

Heavy Metal Contamination and Mixing Processes in Sediments from the Humber Estuary, Eastern England

Estuarine, Coastal and Shelf Science (2001) 53, 619–636 doi:10.1006/ecss.2000.0713, available online at http://www.idealibrary.com on Heavy Metal Con...

1MB Sizes 0 Downloads 15 Views

Estuarine, Coastal and Shelf Science (2001) 53, 619–636 doi:10.1006/ecss.2000.0713, available online at http://www.idealibrary.com on

Heavy Metal Contamination and Mixing Processes in Sediments from the Humber Estuary, Eastern England S. V. Leea and A. B. Cundyb,c a

Department of Earth Sciences, Cardiff University, P.O. Box 914, Cardiff CF10 3YE, U.K. School of Chemistry, Physics and Environmental Science, University of Sussex, Brighton BN1 9QJ, U.K.

b

Received 8 October 1999 and accepted in revised form 13 September 2000 The geochemical properties of cores collected from mud flat and salt marsh environments in the Humber Estuary were investigated. A total of 10 cores were collected along a shore-normal transect on the northern bank of the estuary, near Skeffling. Major and trace element concentrations were determined for each core. The vertical distributions of 210Pb and 137 Cs were also examined to provide a measure of the rate of sediment accumulation. Surface intertidal sediments show elevated concentrations of a range of trace and major elements, including Pb, Zn, Cu, Al, Mn and Fe. Concentrations were higher in the upper mud flats and salt marsh where sediment grain size is finer. Dating of salt marsh sediments indicated a local sediment accretion rate of 0·4 cm yr 1. The early-diagenetic remobilization of heavy metals has apparently been limited, and the salt marsh sediments provide a (time-integrated) record of historical pollutant inputs. Heavy metal fluxes have been calculated from the salt marsh sediments and are broadly comparable with other industrialized and semi-industrialized estuaries. Cu, Pb and Zn inputs to the Skeffling area peaked in the mid-20th century, while Ti, Al and Fe, which are discharged into the Humber from two Tioxide-processing facilities, are only slightly enriched in these sediments. On the mud flats, local mixing, resuspension and erosion has resulted in correlatable sedimentary horizons interspersed with mixed sediment layers. Consequently, the vertical distribution of heavy metals in these mud flats is relatively erratic, and the mud flat sediments are unsuitable for studying historical pollution trends.  2001 Academic Press

Keywords: estuarine sedimentation; mud flats; salt marshes; heavy metals; contaminants; Humber Estuary; North Sea

Introduction Estuarine sediments are an important sink for a wide range of contaminants. Heavy metals in particular show high affinities for fine-grained estuarine sediments (e.g. Barr et al., 1990; Middleton & Grant, 1990; Croudace & Cundy, 1995). Once metals have been released into the aquatic environment they may interact with fine-grained suspended sediments, and subsequently be removed from the water column by deposition in mud flat and salt marsh environments. This removal may be only temporary, however, as following deposition and burial, metals become subject to a variety of physical, chemical and biological processes which may mix, remobilize and ultimately rework the metals into the water column. These processes may be natural (salt marsh erosion, earlydiagenesis, bioturbation etc.) and/or anthropogenic (dredging, land reclamation). In addition, the postulated accelerated sea-level rise due to global warming is likely to enhance erosion of intertidal areas c

Corresponding author.

0272–7714/01/050619+18 $35.00/0

210

Pb and

137

Cs dating;

(Baumann et al., 1984; Reed, 1989; Allen & Pye, 1992; Reed, 1995), increasing the reworking and redistribution of older contaminated sediments. Hence, both salt marsh and mud flat environments may continue to be an important source of heavy metals into an estuary even after effluent discharge has ceased. It is therefore important to understand both the spatial distribution of contaminants, to identify those areas which are the largest potential sources of trace metals if subject to erosion/disturbance, and the vertical distribution of contaminants, to assess the degree to which subsurface sediments are contaminated, and the extent to which trace metals have been subject to chemical remobilization or physical/ biological reworking. The vertical distribution of contaminants in recently-accreted sediments has been examined by a number of authors in a range of estuaries (e.g. Heijnis et al., 1987; Sharma et al., 1987; Zwolsman et al., 1993; Milan et al., 1995; Cundy & Croudace, 1996; Cundy et al., 1997; Cundy et al., 1998). Following  2001 Academic Press

620 S. V. Lee and A. B. Cundy

radiometric dating, it is frequently possible to link vertical variations in contaminant concentration with historical pollutant inputs, and examine changes in pollutant loading over time. Salt marshes in particular have been extensively used for the reconstruction of historical pollution trends due to their clay-rich nature and consequent large adsorption capacity for trace metals, and their stabilized, vegetated nature and dense root system which makes them less susceptible to post-depositional disturbance than adjacent mud flat areas. Pollution trends have, however, also been examined in intertidal mud flat settings, although the intense bioturbation and physical reworking of sediments in these areas mean that such studies have met with varying degrees of success (e.g. Cundy & Croudace, 1996). This study examines the spatial and vertical distribution of trace and major elements, including heavy metals, in recent mudflat and saltmarsh sediments from the Humber estuary, eastern England, and uses these sediments to assess the relative scale of recent and historical contaminant inputs to the estuary. In terms of flow and catchment size, the Humber Estuary (Figure 1) is the largest estuary in the United Kingdom with an average mean flow of 250 m s 1 and a catchment area of 27 000 km2. This latter figure is equivalent to 20% of the area of England (National Rivers Authority, 1993). The Humber is a valuable resource for the community, for fisheries and for wildlife and since it is the main U.K. freshwater input to the North Sea, any contamination in the Humber may have a detrimental effect on coastal water quality (NRA, 1993). A considerable amount of research has been undertaken on trace metal transport and metal interaction with suspended particulate matter in the Humber (e.g. Millward & Glegg, 1997 and references therein), showing that the Humber acts as a trap for fine, metal-contaminated sediments. More limited work has been carried out on the distribution of metals in surface sediments (e.g. Grant & Middleton, 1990, 1993) while almost nothing is known about trace metal distribution in subsurface sediments, or the historical pollution burden of the estuary (Millward & Glegg, 1997). The study area: the Humber Estuary The Humber is a macrotidal estuary and experiences tidal ranges at Grimsby in the order of 6 m during spring tides and 3 m during neap tides (Denman, 1979). Current speeds of 2–3 m s 1 are regularly observed (NRA, 1993). The Humber has a characteristic turbid appearance and contains very high concentrations of suspended sediment, typically 1106

tonnes during the summer and 3106 tonnes during a winter spring tide. The estuary has a long history of anthropogenic impacts. The catchment area exhibits a wide variety of land-use types, from large areas of intensively farmed agricultural land in the Vales of York and Trent, urbanized and industrial areas in the Midlands and south and west Yorkshire, to the open moorland of the North York Moors and Pennines (Neal et al., 1996). Though now much reduced, iron smelting in the region dates back to Roman times, and mining activities in the Pennines have provided an additional input of heavy metals to the estuary. A large percentage of the water entering the Humber is still derived from industrial and domestic effluents—the 1993 Quality Status Report of the North Sea (North Sea Task Force, 1993) identified the Humber as a potential source of a range of pollutants. Over the period 1985–1992 up to 420 kg day 1 of copper, 540 kg day 1 of lead and 2670 kg day 1 of zinc were input into the estuary (Millward & Glegg, 1997), while large quantities of Fe, Al and Ti have been introduced into the estuary since the 1950s from two titanium dioxide processing plants located near Grimsby (Figure 1). The input of Fe has been as high as 1·75105 kg day 1 (Newell et al., 1984), an order of magnitude greater than the riverine input of Fe (Millward et al., 1996).

Methods The present investigation was situated on the northern bank of the lower estuary, near Skeffling, where a shore-normal transect was set up (Figure 1). This transect was also utilized during LISP-UK (Littoral Investigation of Sediment Properties). Four stations were situated along the transect (A, on the upper mud flats; B, 600 m; C, 1500 m and D, 2200 m from shore). At station A the mud flat surface is flat, but at station B a ridge-runnel system is developed which increases in size along the transect. Two cores were taken from each station along the transect, and, where appropriate (at stations B, C and D) one core was collected from a ridge, with the other from a runnel. The mud flats are backed by salt marsh, up to 80 m in width. Two cores were also collected from the salt marsh, SM1 and SM2. Cores were collected using two different methods. Salt marsh cores were collected using a vibrocorer (the Atlas Copco Cobra) during summer 1996, while mud flat cores were collected during summer 1997 by inserting a plastic tube into the mud flat which was dug out once filled with sediment. On return to the laboratory, cores were split, with one half kept as an archive and the other

Heavy metals and mixing processes in the Humber Estuary 621 0°30' W

0°20' W

0°10' W

0°00'

0°10' E

53°40' N

53°30' N

F 1. Location of the Humber Estuary including aerial mosaic of the Humber transect (courtesy of Dr Kevin Black). Two cores were collected from each of the four stations on the mudflat (shown by white circles) and, where appropriate (at B, C and D) one taken from a ridge and one from a runnel. Two cores were also collected from the saltmarsh, SM1 and SM2, at the end of the transect. Other labels (e.g. VGU1, EMP2) refer to sampling locations used in the LISP-UK study.

622 S. V. Lee and A. B. Cundy

used for sub-sampling. The cores were then visually described and assigned a colour using a Munsell soil colour chart. 210 Pb in salt marsh cores was determined through the alpha spectrometric measurement of its granddaughter 210Po. The methodology is based on Flynn (1968) using double acid leaching of sediment and autodeposition of the Po in the leachate onto silver discs. Errors were less than 5%. 210Pb in the mud flat cores and 137Cs were determined via gamma spectrometry using a Canberra 30% P-type HPGe detector. Errors were typically in the order of 4% (1), detection limits were 0·5 Bq kg 1. The simple and CRS (constant rate of supply) models were used to calculate 210Pb dates (Goldberg, 1963; Appleby & Oldfield, 1992). The 210Pbexcess (or unsupported 210Pb) activity was estimated by subtraction of the value of constant 210Pb activity at depth, using values from salt marsh core SM1. Two cores were dated from the mud flat (A1 and D Ridge), and one from the salt marsh (SM1). Magnetic susceptibility was used to correlate dated cores with other cores from the transect. Magnetic susceptibility values were determined using the Multi-Sensor Core Logger (MSCL) developed by Geotek Ltd, U.K. The sensor used was a Bartington MS2E Point sensor. Trace element concentrations were measured on powdered, pelletized samples, with major elements determined from fused beads. Analyses were carried out on a Philips PWI400 X-ray spectrometer system at Cardiff University. Precision is 1% RSD, whilst accuracy was determined using a marine standard and was better than 2% RSD. Mineralogy was determined via X-ray diffraction using a Philips PW1729 X-ray generator. Samples were scanned from 5–352 using a copper radiation tube and a voltage of 35 kV and 40 mA. Further analyses were undertaken using a Leo (Cambridge Instruments) Scanning Electron Microscope S360, with secondary and backscatter electron detectors and an Oxford Instruments Energy Dispersive X-Ray A10000 Analyser. Grain size analyses were carried out at cm intervals using a Micromeritics Sedigraph 5100.

Results Core description Grain size analyses of the mud flat cores indicate that the sediments consist mainly of silty clay and that the sand content of the sediments increases towards station D (Figure 2). Laminated sediments are

present in cores from the salt marsh and from all mud flat cores except those collected from station A (see Figure 7). Sand and clay horizons are also present. There are two shelly horizons, possibly representing storm deposits, in the salt marsh cores (at 53·5 cm in core SM1 and 57·5 cm in SM2), although lack of correlation from magnetic susceptibility indicates that these are unlikely to represent the same event. While grain size analyses were not carried out on salt marsh cores (lack of sample prohibited this), both SM1 and SM2 show a visible increase in grain size with depth. There is a distinct absence of burrows within the sediments.

Magnetic susceptibility Magnetic susceptibility in the mud flat cores indicates a number of similar trends and thus a tentative attempt at correlation has been carried out, shown in Figure 3. Magnetic susceptibility is relatively low, generally below 50 SI units, with a maximum of 101 SI units in core D Ridge at a depth of 29·5 cm. Magnetic susceptibility is generally higher in runnels than in ridges and higher in coarser-grained than in fine-grained horizons, a result of size and density segregation. No distinct variation can be seen along the transect. XRD and SEM analyses indicate that the magnetic material is magnetite and ilmenite. Whilst both are present as massive grains and crystals, they can also be found as small spheroidal objects, with a maximum diameter of 40 m (Figure 4). The surface of most spheroids is covered with a secondary precipitation of iron, though there are some spheroids that remain unchanged. These spheroids are smelting products. Iron production in the Humber catchment area dates back to the Roman period and was based on the smelting of clay ironstones from the local coal measures (Jarvie et al., 1997). During the 19th century, iron smelting became concentrated in a few large iron works. Sheffield became an important iron production centre at this time and in the late 1800s, an iron and steel plant was opened in Scunthorpe. This has now closed and iron and steel production in the area has decreased dramatically (Jarvie et al., 1997). Whilst it is possible to state that the smelting spheroids in the sediment are anthropogenic in origin, it is impossible to ascertain whether crystals and grains of magnetite and ilmenite are anthropogenic (i.e. brought into the estuary as the high grade ilmenite ores used in local TiO2 industries) or whether they are derived from the erosion of the glacial sediments in the Holderness cliffs (e.g. Al-Bakri, 1986).

Heavy metals and mixing processes in the Humber Estuary 623 A1

0

B Ridge

C Runnel

D Ridge

5

10

Depth (cm)

15

20

25

30

35

100 50 Grain size (% in fraction) 0

50 100 Grain size (% in fraction) 0

50 100 Grain size (% in fraction) 0

100 50 Grain size (% in fraction) 0

40

F 2. Grain size down core in cores A1, B Ridge, C Runnel and D Ridge. ( ), % sand; ( ), % silt; ( ), % clay.

Geochemistry 210

Pb and

137

Cs dating

Cored sediments were dated via the 210Pb and 137Cs dating methods. Dating using the natural radionuclide 210 Pb is based on determination of the vertical distribution of unsupported 210Pb (210Pbexcess), or 210Pb arising from atmospheric fallout. This allows ages to be ascribed to sedimentary layers based on the known decay rate of 210Pb [see Appleby & Oldfield (1992) for a synthesis of the 210Pb method]. 137Cs is an artificially produced radionuclide, present in the study area due to atmospheric fallout from nuclear weapons testing, waste discharges from nuclear facilities and reactor accidents. Marked maxima in the deposition of 137Cs occurred in the northern hemisphere in 1958, 1963 (from nuclear weapons testing) and 1986 (from the Chernobyl accident). Discharges from nuclear facilities at Sellafield in north-west England have also released significant amounts of 137Cs into UK coastal waters. In favourable conditions, these periods of peak

fallout/discharge provide subsurface activity maxima in accumulating sediments which can be used to derive rates of sediment accumulation (e.g. Ritchie et al., 1990; Cundy & Croudace, 1996). Both 137Cs and 210Pbexcess are present throughout the entire cored depth of sediment (40 cm) in the mud flat cores A1 and D ridge (Figure 5). The distribution of both radionuclides with depth is relatively erratic in these cores. Under conditions of uniform sediment accretion, with a constant supply of 210Pb the vertical distribution of 210Pbexcess should approximate to an exponential decay curve, with activity declining with depth. This is clearly not the case here, indicating that significant mixing or rapid recent deposition of sediment has occurred to produce an irregular vertical profile (see discussion below). Hence, it is not possible to assign accurate dates to the mudflat cores. Based on the presence of 137Cs throughout cored depth, the sediment in these cores has accumulated more recently than 1954 (the date of first widespread dispersion of 137Cs), giving a minimum accretion rate of c. 0·9 cm yr 1. In contrast, in the salt marsh core

Depth (cm)

35

30

25

20

15

10

5

0

0

40

A1

80 120 0

40

A2

0

40

80

120

0

80 120 0

40

80

C Ridge

Magnetic susceptibility (SI)

40

B Runnel

120

0

40

80

C Runnel

0

120

F 3. Magnetic susceptibility in Humber mudflat cores with correlation.

80 120

B Ridge

40

80

D Ridge

120

0

40

80

D Runnel

120

624 S. V. Lee and A. B. Cundy

Heavy metals and mixing processes in the Humber Estuary 625

F 4. SEM image of spheroidal smelt products found in mudflat cores. Spheroid X has relatively little secondary precipitation whilst spheroid Y has a coating of ilmenite. See text for discussion.

SM1, 210Pbexcess shows a steady decline with depth, to insignificant activities at 50 cm. Based on the Constant Rate of Supply (CRS) and ‘ simple ’ models of 210Pb dating (see Appleby & Oldfield, 1992 for discussion of 210Pb dating models), the 210Pb profile gives a sediment accumulation rate of c. 0·4 cm yr 1 (Figure 5), 137Cs shows a broad subsurface maximum between 3 and 11 cm depth, declining to zero at 25 cm. The subsurface maximum in 137Cs activity is likely to have been caused by inputs from atmospheric nuclear weapons testing (peak fallout in 1963) and discharges from Sellafield. 137Cs from the Chernobyl incident is unlikely to have significantly affected the Humber Estuary (e.g. Watt Committee on Energy, 1991). In studies of salt marsh cores from north Norfolk, approximately 90 km south of the Humber estuary, Shimmield (1997) found that the vertical distribution of 137Cs was consistent with derivation from Sellafield discharges, with a temporal offset of 5–8 years (assumed to be the transport time of 137Cs between Sellafield and north Norfolk). Hence, the broad peak in 137Cs observed here is likely to be due to input from Sellafield (in approximately 1980 i.e. peak Sellafield discharges offset by 5 years), with some additional input from weapons testing. It is

not possible to discriminate these sources due to the broad subsurface maximum in activity, hence the observed peak is attributed here to the period 1963– 1980. This age range agrees reasonably well with the age-depth curve derived from 210Pb dating (Figure 5). In addition, the sharp 137Cs increase above 25 cm corresponds to an age of approximately 48 years derived from the 210Pb CRS dating model, consistent with the initial widespread release of 137Cs in the early 1950s. Major and trace elements Normalization of metal concentrations There is usually a close relationship between metal concentration and grain size in estuarine and marine sediments since metals tend to associate with finergrained particles (e.g. Livens & Baxter, 1988; Cundy et al., 1997). In order to compare samples it is necessary to compensate for this by applying a correction. Otherwise any finer-grained samples may show relatively high metal concentrations. Whilst the grain size effect can be compensated for by ratioing metal concentrations to grain size for each sample, this has

626 S. V. Lee and A. B. Cundy (a)

(b) D Ridge 137

5

5

10

10

15 20 25 30 35

Maximum age of base of core = 42 years (i.e. younger than 1954)

0

15 20 Maximum age of base of core = 42 years (i.e. younger than 1954)

25

35 40

210

–1

210

–1

Unsupported Pb (Bq g ) 0 0.01 0.02 0.03 0.04

5

5

10

10

20 25

20 25 30

35

35

40

1954?

30

0

40

10

50

20

210

–1

Unsupported Pb (Bq g ) 0 0.01 0.03 0.05

20

Years before 1996 50 100 150

30 40 50

Simple modelderived sediment accumulation rate = 3.4 mm y–1 (2 σ 2.8–4.5 mm y–1)

30 40 50

40 210

Age-depth curve, core SM1

20

10

15

30

1963– 1980

60

Depth (cm)

15

–1

Cs (Bq g ) 0.01 0.02 0.03

60

Unsupported Pb (Bq g ) 0 0.1 0.2 0.3 0.4

Depth (cm)

Depth (cm)

137

10

30

40

–1

Cs (Bq g ) 0.01 0.02 0.03

0

Depth (cm)

Depth (cm)

0

–1

Cs (Bq g ) 0.15 0.05 0.1

Depth (cm)

137

SM1

Depth (cm)

A1

60 137

F 5. (a) Unsupported Pb and Cs vs. depth for mudflat cores A1 and D Ridge, and saltmarsh core SM1. See text for discussion. (b) Age-depth curve for core SM1, derived using the Constant Rate of supply (CRS) model of 210Pb dating. The open rectangle indicates the age range given by the 137Cs profile. The simple model of 210Pb dating gives an average accretion rate of 3·4 mm y 1 (2 range 2·8–4·5 mm y 1), broadly consistent with dates from the CRS model. The mixing depth for core SM1 is 6 cm.

the disadvantage of being laborious and so many authors have normalized metal concentration to some element found dominantly in the clay-size fraction. Aluminium has been frequently used as a grain size proxy in both marine and estuarine sediments (e.g. Zwolsman et al., 1993; Cundy & Croudace, 1995). To judge whether Al2O3 was suitable for use in this study, grain size data were obtained from a number of samples and plotted against weight % Al2O3 (Figure 6). Whilst regression analysis shows a relatively good degree of correlation (R2 =0·707), Grant and Middleton (1990) suggest that aluminium cannot be used as a grain size proxy in the Humber because of an additional input from the two titanium dioxide smelting plants on the south bank of the estuary located between Immingham and Grimsby. Other authors have suggested caesium (Ackerman, 1980) and lithium (Loring, 1990) for grain size normalization. However, Cs is usually present in very low

concentrations in estuarine sediments and Li cannot be measured using XRF (Grant & Middleton, 1990). Iron and titanium are other potential grain size proxies. However, the Humber is noticeably contaminated with these elements, by up to 50% above natural background (Grant & Middleton, 1990) as a result of acid-iron waste discharges from the two titanium dioxide plants (Barr et al., 1990). In fact, both plants discharge up to 70 000 m3 day 1 of waste material which consists of a variety of elements (such as Ti and Fe, cadmium, copper, chromium, nickel, vanadium and manganese) through two pipelines into the estuary (House et al., 1997). Another possible grain size proxy is rubidium. Whilst there appears to be no anthropogenic input of Rb in the Humber (Middleton & Grant, 1990) it shows a poor correlation with grain size (R2 =0·522, Figure 6). Potassium was eventually chosen. Whilst regression analysis shows a slightly lower level of correlation than for

Heavy metals and mixing processes in the Humber Estuary 627

% Al2O2

preclude using K2O as a grain size indicator as the regression fit is relatively good.

2

15

y = 4.67 + 0.231x R = 0.707

Heavy metal concentrations 10

5

10

15

20 25 30 35 % grain size < 8 µm

40

45

2

0.0085

y = 0.00524 + 6.78e–05x R = 0.522

0.008

% Rb

0.0075 0.007 0.0065 0.006 0.0055 0.005

5

10

15

20 25 30 35 % grain size < 8 µm

40

45

40

45

2

2.4

y = 1.18 + 0.0252x R = 0.7

% K2O

2.2 2 1.8 1.6 1.4 1.2

5

10

15

20 25 30 35 % grain size < 8 µm

F 6. Al2O3, Rb and K2O plotted against % grain size less than 8 µm.

Al2O3 (R2 =0·7), K2O is a more reliable choice since (a) Aluminium concentrations are slightly elevated in the Skeffling sediments, probably due to anthropogenic inputs (see below), and (b) there are no known anthropogenic inputs of K into the estuary. The graph of K2O against % grain size less than 8 m does not pass through the origin, which is what would be expected if all of the K2O was in the fine fraction. Thus, some of the K2O is present in the coarser fraction (Lee, pers. comm.). However, this does not

Bulk (non-normalized) concentrations for Cu, Pb and Zn in mud flat cores range from <3–83 ppm, 28–145 ppm and 88–395 ppm respectively. In the two salt marsh cores, the ranges are <3–60 ppm (Cu), 46–127 ppm (Pb) and 78–344 ppm (Zn). Some of these values are lower than those presented by Middleton and Grant (1990) for the Scrobicularia clay from the nearby Holderness coastline (Cu= 17 ppm, Pb=22 ppm, Zn=84 ppm), which these authors suggest provides a geochemical baseline with which to compare modern day trace element-enriched sediments in the Humber. This discrepancy is likely to be due to differences in grain size between the sediments analysed here and the Scrobicularia clay. To reduce this possible grain size effect, elements were normalized to K2O in the Scrobicularia clay so that they could be compared to similarly normalized values from the salt marsh and mud flats. Normalized elemental concentrations from the Scrobicularia clay and from the cores analysed here are presented in Table 1, along with enrichment factors (see below). Values obtained are for the most part greater in the Humber cores than from the Scrobicularia clay thus indicating that the Scrobicularia clay may still be used as an indicator of pre-industrial elemental concentrations following correction for differences in grain size. The sediments are most heavily enriched in Fe2O3 with a maximum enrichment in core A1 of over 18 times the normalized concentration present in the Scrobicularia clay. The sediments are also heavily enriched in Zn, with a maximum normalized concentration of 169·9 in A2 compared to a value of 33·6 in the Scrobicularia clay. Pb has a maximum normalized concentration of 69·2, compared to 8·8 in the Scrobicularia clay. Other enriched elements include As, MnO, P2O5 and Cr. This is in good agreement with results obtained by Grant and Middleton (1990) who also found that Humber estuary sediments showed elevated concentrations of these elements. Despite the large discharges of A1 and Ti into the Humber, Al2O3 and TiO2 concentrations are only slightly elevated in the sediments analysed here. Pb, Cu and Zn (normalized to K2O) vs. depth are shown in Figure 7. In mud flat cores, the vertical profiles of Cu, Pb and Zn are virtually identical. No clear increase or decrease in heavy metal content with depth is apparent. This is not an effect of varying grain size in the cores as the data are normalized to K. In contrast to the irregular Cu, Pb and Zn profiles seen

Al2O3/K2O TiO2/K2O Fe2O3/K2O P2O5/K2O MnO/K2O V/K2O Cr/K2O Ni/K2O Cu/K2O Zn/K2O As/K2O Sr/K2O Y/K2O Nb/K2O La/K2O Ce/K2O Pb/K2O

Al2O3/K2O TiO2/K2O Fe2O3/K2O P2O5/K2O MnO/K2O V/K2O Cr/K2O Ni/K2O Cu/K2O Zn/K2O As/K2O Sr/K2O Y/K2O Nb/K2O La/K2O Ce/K2O Pb/K2O

5·36 0·28 0·20 0·04 0·02 43·60 39·60 15·20 6·80 33·60 8·80 48·00 10·80 5·60 13·60 29·60 8·80

Scrobicularia clay

5·83 (6·14) 0·47 (0·53) 2·81 (3·21) 0·14 (0·20) 0·05 (0·06) 69·61 (85·73) 56·06 (67·66) 21·85 (24·79) 9·64 (13·43) 100·93 (119·18) 60·76 (80·57) 102·99 (109·67) 12·88 (13·49) 9·77 (14·26) — 29·86 (54·63) 34·55 (44·42)

1

2

1·09 (1·15) 1·66 (1·89) 14·04 (16·05) 3·60 (5·00) 2·52 (3·00) 1·60 (1·97) 1·42 (1·71) 1·44 (1·63) 1·42 (1·98) 3·00 (3·55) 6·90 (9·16) 2·15 (2·28) 1·19 (1·25) 1·74 (2·55) — 1·01 (1·85) 3·93 (5·05)

2

1·17 (1·12) 1·80 (1·96) 16·15 (17·35) 4·50 (5·75) 3·19 (4·00) 1·77 (2·03) 2·01 (5·28) 1·66 (1·92) 2·97 (5·10) 3·82 (4·88) 8·72 (11·83) 1·82 (2·05) 1·28 (1·66) 1·78 (2·17) 1·63 (2·36) 1·60 (2·29) 5·20 (6·81)

C Runnel

6·27 (6·48) 0·50 (0·55) 3·23 (3·47) 0·18 (0·23) 0·06 (0·08) 77·26 (88·59) 79·40 (209·11) 25·22 (29·13) 20·22 (34·70) 128·37 (164·08) 76·77 (104·10) 87·17 (98·52) 13·79 (17·94) 9·98 (12·16) 22·16 (32·13) 47·32 (67·79) 45·75 (59·97)

1

A1

5·41 (6·10) 0·46 (0·54) 2·83 (3·70) 0·14 (0·15) 0·05 (0·06) 68·82 (99·54) 54·49 (82·56) 22·53 (36·81) 7·48 (20·83) 100·35 (166·66) 60·45 (96·61) 114·46 (133·63) 14·04 (21·96) 7·95 (12·35) 9·44 (40·12) 41·46 (86·73) 34·77 (41·64)

1

2

1·01 (1·14) 1·66 (1·93) 14·13 (18·50) 3·44 (3·75) 2·67 (3·00) 1·58 (2·28) 1·38 (2·08) 1·48 (2·42) 1·10 (3·06) 2·99 (4·96) 6·87 (10·98) 2·38 (2·78) 1·30 (2·03) 1·42 (2·21) 0·69 (2·95) 1·40 (2·93) 3·95 (4·73)

2

1·17 (1·20) 1·90 (2·29) 17·05 (18·55) 5·35 (6·50) 3·28 (65·50) 1·860 (2·11) 1·91 (2·61) 1·70 (1·95) 3·46 (5·32) 4·09 (5·06) 10·94 (14·57) 1·84 (2·08) 1·49 (1·80) 1·78 (2·05) 1·74 (2·73) 1·86 (2·42) 5·49 (7·09)

D Ridge

6·28 (6·43) 0·53 (0·64) 3·41 (3·71) 0·21 (0·26) 0·07 (1·31) 80·92 (92·02) 75·74 (103·21) 25·89 (29·60) 23·53 (36·18) 137·33 (169·86) 96·27 (128·24) 88·49 (99·63) 16·10 (19·44) 9·95 (11·48) 23·66 (37·12) 55·19 (71·63) 48·29 (62·36)

1

A2

5·58 (6·30) 0·46 (0·49) 2·78 (3·20) 0·14 (0·16) 0·05 (0·06) 70·88 (87·24) 55·70 (68·77) 20·98 (26·72) 7·36 (12·38) 102·27 (136·93) 64·89 (86·62) 119·95 (127·73) 14·81 (19·58) 8·56 (11·90) 10·46 (58·11) 40·58 (75·59) 34·60 (43·42)

1

2

1·04 (1·18) 1·64 (1·75) 13·90 (16·00) 3·40 (4·00) 2·54 (3·00) 1·63 (2·00) 1·41 (1·74) 1·38 (1·76) 1·08 (1·82) 3·04 (4·08) 7·37 (9·84) 2·50 (2·66) 1·37 (1·81) 1·53 (2·13) 0·77 (4·27) 1·37 (2·55) 3·93 (4·93)

2

1·08 (1·16) 1·63 (1·79) 14·36 (16·00) 4·02 (5·00) 2·53 (3·00) 1·54 (0·37) 1·48 (1·92) 1·51 (1·84) 2·32 (4·13) 3·37 (4·45) 7·63 (11·52) 2·05 (2·28) 1·19 (1·48) 1·44 (1·96) — 1·42 (3·06) 4·47 (5·95)

D Runnel

5·77 (6·20) 0·46 (0·50) 2·87 (3·20) 0·16 (0·20) 0·05 (0·06) 66·95 (15·96) 58·75 (76·20) 22·93 (27·95) 15·77 (28·09) 113·08 (149·57) 67·17 (101·39) 98·27 (109·42) 12·83 (15·96) 8·05 (10·97) — 42·07 (90·44) 39·33 (52·39)

1

B Ridge

5·46 (6·74) 0·42 (0·53) 2·48 (3·28) 0·11 (0·16) 0·04 (0·09) 52·67 (80·91) 52·24 (67·35) 20·13 (25·48) 10·34 (26·42) 97·14 (134·33) 41·55 (75·57) 107·57 (123·06) 14·85 (20·24) 5·84 (12·63) 9·67 (32·36) 37·45 (74·41) 41·88 (52·29)

1

2

1·02 (1·26) 1·52 (1·89) 12·38 (16·40) 2·69 (4·00) 2·20 (4·50) 1·21 (1·86) 1·32 (1·70) 1·32 (1·68) 1·52 (3·89) 2·89 (4·00) 4·72 (8·59) 2·24 (2·56) 1·37 (1·87) 1·04 (2·26) 0·71 (2·37) 1·27 (2·51) 4·76 (5·94)

2

1·02 (1·17) 1·56 (1·75) 13·82 (16·50) 3·85 (5·25) 2·50 (3·00) 1·54 (2·35) 1·53 (2·43) 1·63 (2·49) 2·66 (5·18) 3·69 (5·82) 8·37 (13·18) 2·19 (2·52) 1·25 (1·78) 1·38 (2·18) 1·04 (2·02) 1·56 (2·88) 4·93 (7·87) SM1

5·48 (6·26) 0·44 (0·49) 2·76 (3·30) 0·15 (0·21) 0·05 (0·06) 67·23 (102·55) 60·55 (96·15) 24·76 (37·82) 18·09 (35·25) 124·13 (195·49) 73·64 (73·64) 105·00 (121·14) 13·45 (19·23) 7·74 (12·18) 14·20 (27·48) 46·16 (85·12) 43·35 (69·22)

1

B Runnel

5·70 (6·62) 0·45 (0·52) 2·68 (3·41) 0·12 (0·22) 0·07 (0·41) 60·69 (97·78) 58·33 (83·42) 22·15 (31·46) 12·53 (29·35) 108·16 (168·27) 51·53 (103·70) 105·12 (140·18) 15·36 (22·07) 7·31 (14·36) 18·20 (43·76) 46·13 (92·24) 43·95 (62·12)

1

2

1·06 (1·24) 1·62 (1·86) 13·39 (17·05) 3·05 (5·50) 3·49 (20·50) 1·39 (2·24) 1·47 (2·11) 1·46 (2·07) 1·84 (4·32) 3·22 (5·01) 5·86 (11·78) 2·19 (2·92) 1·42 (2·04) 1·31 (2·56) 1·34 (3·22) 1·56 (3·14) 4·99 (7·06)

2

0·53 (0·56) 1·70 (1·96) 14·23 (16·00) 3·63 (4·25) 2·48 (3·00) 1·64 (2·09) 1·46 (1·80) 1·41 (1·56) 1·43 (1·95) 3·03 (3·75) 6·82 (9·52) 2·02 (2·40) 1·18 (1·67) 1·89 (2·59) 1·35 (2·31) 1·28 (2·14) 3·92 (4·95) SM2

2·85 (3·01) 0·48 (0·55) 2·85 (3·20) 0·15 (0·17) 0·05 (0·06) 71·60 (91·02) 57·86 (71·17) 21·47 (23·72) 9·76 (13·24) 101·64 (125·87) 60·02 (83·76) 97·17 (115·37) 12·72 (17·99) 10·61 (14·52) 18·35 (31·47) 37·93 (63·40) 34·48 (43·57)

1

C Ridge

T 1. Normalized elemental concentrations and enrichment factors in Humber mudflats and salt marsh cores. Column 1 shows the average (normalized) elemental concentration over the entire depth of sediment sampled, while column 2 shows the average enrichment factor over the cored depth of sediment. Normalized elemental concentrations are calculated from the ratio elemental concentration in sample/concentration of K2O in that sample, and enrichment factors are calculated from the ratio normalized elemental concentration in sample/normalized elemental concentration in Scrobicularia clay. See text for full discussion. Maximum normalized elemental concentrations and maximum enrichment factors are shown in brackets

628 S. V. Lee and A. B. Cundy

Heavy metals and mixing processes in the Humber Estuary 629

in the mud flat cores, the salt marsh cores show a broad mid-depth maximum in the concentration of each of these metals, with a general decline towards the sediment surface. Discussion Controls on the spatial distribution of contaminants A simple yet effective way of showing how concentrations vary between sites is by using isocon diagrams (Grant, 1986; Cundy et al., 1997). These plots compare the average concentrations of specific elements from different sites. An average concentration for each element has been calculated from each core. [Figure 8(a)] compares bulk (non-normalized) concentrations between stations, whilst concentrations have been normalized to K2O to take into account grain size variations in [Figure 8(b)]. The isocon diagrams indicate that there is little difference in sediment composition or heavy metal concentrations between the two salt marsh sites, or between ridges and runnels at Station D (despite the higher magnetic susceptibility in runnels compared to ridges). There are, however, more significant differences in element concentrations between the salt marsh core SM1 and station D (lower mud flat), and between station A1 (upper mud flat) and station D. Core SM1 shows higher bulk (non-normalized) concentrations of Cu and Zn than station D. This, however, is likely to be caused by differences in grain size between the two sites, as normalizing to K2O removes this effect [Figure 8(b)]. The upper mud flat core A1 shows higher concentrations than the lower mud flat core D1 for all elements except SiO2 and Zr. Again, this is largely an effect of differences in grainsize between the two sites [Figure 8(b)], although Cu and (to a lesser extent) Ni remain slightly enriched in the upper mud flat after normalizing to K2O. Despite this it seems that, once grain-size effects are considered, heavy metals are relatively uniformly distributed between the different intertidal environments (salt marsh, upper mud flat, lower mud flat) along this transect in the Humber. It should be noted that significant variations in metal concentration may be observed, however, at smaller scales across individual marshes and mud flats (e.g. Cundy & Croudace, 1995). Controls on the vertical distribution of contaminants The similarity between the Pb, Cu and Zn profiles in each mud flat core indicates that either these elements

have undergone similar early-diagenetic remobilization and reprecipitation around redox boundaries, or that all three elements are derived from the same source. In the B Ridge mud flat core Cu, Pb and Zn show a broad concentration maximum at 25 to 30 cm which is coincident with concentration maxima for the redox-sensitive elements S and Fe (Figure 7) and a colour change from orange–brown to dark grey. This possibly represents sulphide formation and corresponding trapping of Cu, Fe, Pb and Zn as insoluble sulphides at depth in the mud flat, a process which has been inferred in other studies (e.g. Cundy & Croudace, 1995). In the other mudflat cores, however, subsurface maxima in Cu, Pb and Zn show little relationship with subsurface Fe, Mn and S maxima. In the salt marsh core SM1 the broad mid-depth maximum in Cu, Pb and Zn is not coincident with concentration maxima in redox-sensitive elements (Figure 7). In SM2, Fe, Mn, Cu, Pb, and Zn show broadly similar distributions with depth, although the anthropogenic input of Fe to the Humber means that this may be due to pollutant input rather than earlydiagenetic remobilization and precipitation. The lack of coincident peaks between Cu, Pb and Zn and redox-sensitive elements, particularly S, in mudflat and salt marsh sediments indicates that earlydiagenetic remobilization has not significantly affected the vertical distribution of these heavy metals. This is consistent with studies carried out elsewhere (e.g. Allen & Rae, 1986; Cundy & Croudace, 1996; Cundy et al., 1997), although early-diagenesis may be an important control on heavy metal distribution in some estuaries (e.g. Zwolsman et al., 1993) Instead, the subsurface enrichments of Pb, Zn and Cu in the cores analysed here are probably due to contaminant input. The vertical distribution of these elements seems to be dominantly controlled by pollutant input to the estuary, and by mixing and resuspension processes. In estuaries, significant mixing and reworking of contaminated sediments prior to burial may lead to a general pulse of contaminated sediments being washed onto salt marshes and upper mud flats over the period of peak industrial activity. In this scenario, discrete peaks in heavy metal concentration are absent, and in-estuary mixing produces broad, coincident subsurface concentration maxima for the most common industrial contaminants. This appears to be the case here, and is seen most clearly in the salt marsh core SM1, where Cu, Pb, Zn and 137Cs show broad subsurface maxima instead of discrete concentration or activity peaks. The lack of discrete activity maxima in the 137Cs profile corresponding to periods of peak input in 1963 (from atmospheric weapons testing) and the late

Depth (cm)

40

35

30

25

20

15

10

5

0

40

35

30

25

20

15

10

5

0

35

40

40

35

30

25

20

15

10

5

0

0

0

0

Fine sand, shell fragments and 0 silt laminae, 10YR 3/2 Discontinuous laminations, clay and silty clay 5Y 3/2 Fine sand (5Y 3/2) Silty clay (5Y 2.5/1) Laminations (as above), clay, silt. Clayey silt (5Y 3/2) Discontinuous laminations, silt and clay (5Y 3/2) Silty clay with silt patches Discontinuous laminations, silty clay (5Y 3/2) and silt (10YR 3/3) Clayey silt, few silt laminations/lenses. 5Y 3/2 Fine sand. 5Y 3/2 with 2.5Y 3/2 Fine laminations. 5Y 3/2 Fine sand. 5Y 3/2 with 2.5Y 3/2 Discontinuous laminations, fine sand and clay. 5Y 3/1 and 2.5Y 3/2 Fine sand with silty clay fingers. 5Y 3/1 and 2.5Y 3/2

Fine sand (7.5 YR 2/0)

Silt lens and patch

30

Clayey silt. 7.5YR 2/0 with 2.5Y 3/2

25

Silt with silty clay and clay patches (7.5YR 2/0 with 5Y 3/1 and 10YR 3/3)

20

Gradational colour

15

Clayey silt. 2.5Y 3/2 and 5Y 3/2

Clayey silt (10YR 3/3) Sharp colour boundary

Sharp colour boundary Clayey silt. 7.5YR 2/0 with 2.5Y 3/2 Gradational colour boundary Wavy, continuous laminations, silty clay with clay (7.5YR 2/0 with 2.5Y 3/2) Fine sand (5Y 3/2)

Clayey silt. 10YR 3/3, blends into 5Y 3/2 and 2.5Y 3/2 with 5Y 3/1 patches Silt and silty clay laminations (4–7 cm) Silt lenses (21–23 cm)

Clayey silt. 7.5 YR 2/0 with 5Y 3/1 and 10YR 3/3

Small clasts of harder clay (7.5YR 2/0)

Sharp colour boundary

Clayey silt (10YR 3/3) Sharp colour boundary Clayey silt. 10YR 3/1 and 5Y Sharp colour boundary Clayey silt. 7.5YR 2/0 with 10YR 4/2 10YR 3/3 Gradational colour boundary Clayey silt. 10Y 3/3 with 10YR 3/1

10

5

0

D Ridge

Depth (cm)

B Runnel

Depth (cm)

B Ridge

Depth (cm)

A1 2

3

2

3

2

3

1

2

3

Fe2O3/K2O

1

Fe2O3/K2O

1

Fe2O3/K2O

1

Fe2O3/K2O

F 7. Part 1.

MnO/K2O wt % sulphur Pb/K2O 4 0 0.075 0.15 0.225 0.3 0 0.25 0.5 0.75 1 0 20 40 60 80 100 0

MnO/K2O wt % sulphur Pb/K2O 4 0 0.075 0.15 0.225 0.3 0 0.25 0.5 0.75 1 0 20 40 60 80 100 0

MnO/K2O wt % sulphur Pb/K2O 4 0 0.075 0.15 0.225 0.3 0 0.25 0.5 0.75 1 0 20 40 60 80 100 0

MnO/K2O wt % sulphur Pb/K2O 4 0 0.075 0.15 0.225 0.3 0 0.25 0.5 0.75 1 0 20 40 60 80 100 0

50

Cu/K2O 100 150 200 0 10 20 30 40 50 0

Zn/K2O

50

Cu/K2O 100 150 200 0 10 20 30 40 50 0

Cu/K2O 100 150 200 0 10 20 30 40 50 0

Zn/K2O

50

Zn/K2O

Cu/K2O 100 150 200 0 10 20 30 40 50 0

Zn/K2O 50

TiO2/K2O 0.2 0.4 0.6 0.8 0

TiO2/K2O 0.2 0.4 0.6 0.8 0

TiO2/K2O 0.2 0.4 0.6 0.8 0

TiO2/K2O 0.2 0.4 0.6 0.8 0 90 120

60

90 120

60

90 120

30

60

90 120

As/K2O

30

As/K2O

30

As/K2O

60

As/K2O 30

630 S. V. Lee and A. B. Cundy

40

35

30

25

20

15

10

5

0

60

50

40

30

20

10

0

60

50

40

30

20

10

0

Orange-brown clay

Orange-brown sand/clay laminations

Mainly clay with patches/laminations of silty clay and sand

Horizon of broken shells

Laminations of clay, silty clay and sand, with larger sandy horizons

Orange-brown clay blending to dark grey at 17 cm

Orange-brown fine sand No material

Mainly clay with patches/laminations of silty clay and sand

Horizon of broken shells

Laminations of clay, silty clay and sand, with larger sandy horizons

Clay with some sand patches (where plant roots stick out)

Large shell

Orange-brown clay blending to dark grey at 15–17 cm

Fine sand (2.5Y 3/2) with thin clay lenses (5Y 3/2)

Waterlogged silt (5Y 3/1) Waterlogged silt (10YR 3/3) Discontinuous laminations, silt, silty clay and clay Clay (5Y 2.5/1) Continuous laminations (wavy = ripples?), of silt and clay down to 16 cm. 5Y 3/1 at top, grading down to 5Y 3/2 and 2.5Y 3/2 Discontinuous laminations of silt and clay down to 20 cm. 5Y 3/1 and 2.5Y 3.2 Clayey silt (5Y 3/1)

0

0

0

2

3

2

3

1

2

3

Fe2O3/K2O

1

Fe2O3/K2O

1

Fe2O3/K2O

F 7. Part 2.

MnO/K2O wt % sulphur Pb/K2O 4 0 0.075 0.15 0.225 0.3 0 0.25 0.5 0.75 1 0 20 40 60 80 100 0

MnO/K2O wt % sulphur Pb/K2O 4 0 0.075 0.15 0.225 0.3 0 0.25 0.5 0.75 1 0 20 40 60 80 100 0

MnO/K2O wt % sulphur Pb/K2O 4 0 0.075 0.15 0.225 0.3 0 0.25 0.5 0.75 1 0 20 40 60 80 100 0

50

Cu/K2O 100 150 200 0 10 20 30 40 50 0

Cu/K2O 100 150 200 0 10 20 30 40 50 0

Zn/K2O

50

Zn/K2O

Cu/K2O 100 150 200 0 10 20 30 40 50 0

Zn/K2O 50

TiO2/K2O 0.2 0.4 0.6 0.8 0

TiO2/K2O 0.2 0.4 0.6 0.8 0

TiO2/K2O 0.2 0.4 0.6 0.8 0 90 120

60

90 120

30

60

90 120

As/K2O

30

As/K2O

60

As/K2O 30

F 7. Fe2O3/K2O, MnO/K2O, weight % S, Pb/K2O, Zn/K2O, Cu/K2O, TiO2/K2O and As/K2O plotted against depth in mudflat and salt marsh cores. Detailed stratigraphy shown alongside.

Depth (cm)

SM2

Depth (cm)

SM1

Depth (cm)

D Runnel

Heavy metals and mixing processes in the Humber Estuary 631

632 S. V. Lee and A. B. Cundy (a) 700

700 SiO2

600

500 D Runnel

D Ridge

500 Zr

400 300 200

SiO2

600

K2O

100

Y

Zr

400 300 200

Zn Cu

100

Nb 0

100 200 300 400 500 600 700 SM1

0

700

100 200 300 400 500 600 700 A1

700

600

600

500

500

400

400

SM2

D Runnel

SiO2

Zr 300

Zr

La

300

200

200

100

100

0

SiO2

100 200 300 400 500 600 700 D Ridge

0

100 200 300 400 500 600 700 SM1

(b) 600

500

TiO2

Cl

300 Ni

400 300

Ca Zr Ni P2O5 Mg Cu

200

100 0

100 100 200 300 400 500 600 SM1

0

Cl

600

TiO2

SiO2

SiO2

400

D Runnel

D Ridge

500

200

Cl

600

100 200 300 400 500 600 A1

600

500

500

TiO2

400 Ca

SiO2

300

400 SM2

D Runnel

TiO2

SiO2 Ca

300

Zr 200

Ni

100 0

200

La

Ni

Zr Cl

100 100 200 300 400 500 600 D Ridge

0

100 200 300 400 500 600 SM1

Heavy metals and mixing processes in the Humber Estuary 633

1970s (corresponding to input of Sellafield-derived 137 Cs) indicates that mixing has produced a timeintegrated record of contaminant input. The laminated nature of the salt marsh sediments indicates that this mixing took place prior to deposition, rather than through bioturbation. Hence, the salt marsh sediments at Skeffling seem to preserve a record of general pollutant loading, but using these sediments to accurately reconstruct temporal trends in metal input is extremely difficult due to in-estuary mixing. This has been documented in the macrotidal Severn Estuary by Allen and Rae (1986), where, rather than individual input events being recognized in the muddy late Flandrian intertidal sediments, estuarine processes have mixed the sediments to produce three zones: chemozone I represents pre-industrial deposits when metal content is low, chemozone II represents industrial sediments with an increase in metal content, and chemozone III (post-industrial) shows a slight decline in metal content with decreasing age up to the present. Cundy et al. (1997) observed a similar phenomenon in salt marsh cores from the mesotidal Hamble Estuary, southern England. In the mud flat cores, the vertical contaminant distribution has been further complicated by local mixing, resuspension and erosion. The deep penetration of 137Cs on the Humber mud flats, combined with the erratic 210Pbexcess profile, suggests that the sediments were deposited very rapidly and/or may be mixed. However, if the mud flat sediments were completely mixed, correlation of core layers by magnetic susceptibility would not be possible. Even though the magnetic susceptibility records are not identical, there are some sections within each record that look extremely similar. It is therefore suggested that whilst some layers have been completely mixed and/or resuspended after deposition, some horizons were preserved since current velocities were not high enough to resuspend them, or their deposition was followed by the deposition of another horizon which acted as a protective cap. If sediment that had been deposited at the same time was preserved in a number of locations, then this would result in correlatable horizons interspersed with mixed layers. This is in agreement with comments made by Andersen et al. (pers. comm.) on cores collected from the Humber

and Kongsmark in Denmark. They suggest that horizons with low unsupported 210Pb concentrations represent older sediment that has been mixed in from elsewhere in the estuary. Thus, of the three mud flat and salt marsh cores which were radiometrically dated here, only the salt marsh core SM1 yielded useful dates. Recent and historical inputs of contaminants to the Humber Estuary As a result of the mixing processes outlined above, it is only possible to reliably reconstruct general trends in pollutant loading in this study using data from the salt marsh core SM1, as local mixing and resuspension/ erosion is apparently less significant here than on the adjacent mud flats (shown by the more uniform accretion illustrated by the 137Cs and 210Pbexcess profiles). A general coincident increase in concentration occurs for a range of heavy metals at 40 cm, corresponding to the beginning of the 20th century. Rather than basing contaminant trends on concentration data, however, fluxes (or rate of metal input in g cm 2 yr 1) were calculated. Using flux, rather than concentration data, allows a more reliable assessment of contamination trends, particularly where sediment accumulation rates have varied. For example, a maximum in heavy metal concentration may simply be due to a slower sediment accumulation rate, rather than an increase in pollutant discharge (e.g. Croudace & Cundy, 1995). This effect is removed by using flux data. Total, rather than excess or anthropogenic, fluxes are presented here. This is due to differences in grain size between the samples analysed and the pre-industrial sediments thought to provide a geochemical baseline for the study area, which would lead to negative excess flux values (see above). Hence, the flux values shown here represent the sum of anthropogenic and inherent detrital inputs for each element. Total Pb fluxes in core SM1 show a general increase through time (i.e. up the core) and then a decrease from the mid-20th Century (i.e. at 20·5 cm) onwards (Figure 9). This is much earlier than the decline in gasoline-derived Pb caused by reduction of alkyllead additives and the introduction of unleaded gasoline in the U.K. (1986, Cundy et al., 1997).

F 8. (a) Isocon diagrams based on Grant (1986). Each point represents an average value of the concentration of that element in the core. Values have been adjusted in order that they plot on the same scale. For clarity, only those points which plot significantly off the line of equal concentration are labelled. LOI10; Wt%-SiO2 10, Al2O3 5, Fe2O3 10, MnO100, MgO20, CaO10, Na2O100, K2O100, TiO2 100, P2O5 1000, S100; ppm–Pb1, Cu1, Ni1, Cr1, Cl100, La10, Ce1, V1, Y10, Zr1, Nb10, As1. (b). Isocon diagrams using normalized (to K2O) concentrations. Adjustments as for un-normalized data except for: Mg100, Ca100, TiO2 1000, S1000, Cu10, Ni10, Cl0.1.

634 S. V. Lee and A. B. Cundy Date (age range –2 –1 Al2O3 flux (mg cm yr ) based on 210Pb CRS model) 0 10 20 30 40 50 60 1996

30

40

Pb flux (µg cm

20

30

30

10

20

30

yr ) 40

0

30

20

30

50 –2

–1

yr )

Cu flux (µg cm

20 40 60 80 100 120

0

10

20

10

40

Zn flux (µg cm 50

–1

yr )

10

20

–1

–2

5

0

40

50 –2

Depth (cm)

c. 1920–1950

Depth (cm)

c. 1955–1970

30

40

10

c. 1970–1985

20

TiO2 flux (mg cm

Depth (cm)

20

Date 50 (age range based on 210Pb CRS model) 0 1996

10

–1

yr )

10

Depth (cm)

c. 1920–1950

10

Depth (cm)

c. 1955–1970

0

–2

5

–2

–1

yr )

10 15 20 25 30

10

Depth (cm)

c. 1970–1985

Fe2O3 flux (mg cm

20

30

40

40

40

50

50

50

F 9. Total Al2O3, Fe2O3, TiO2, Zn, Cu and Pb flux data from salt marsh core SM1. Flux=dry bulk density ` concentration ` sediment accumulation rate.

Instead, the flux decrease observed here may be caused by improved industrial effluent control measures and/or industrial decline. Cu and Zn fluxes show similar trends to Pb (Figure 9). The fluxes observed at Skeffling for these metals are broadly comparable to those reported for other industrialized and semi-industrialized estuaries (Table 2). While total Pb fluxes are slightly higher at Skeffling than those estimated for intertidal areas in other UK estuaries, they are significantly lower than the 120 g cm2 yr 1 total Pb flux calculated for subtidal sediments around an oil refinery discharge pipeline in Southampton Water, southern England (Croudace & Cundy, 1995). Similarly, Cu and Zn fluxes are much lower than the maximum values observed in Southampton Water.

The major discharges of Al2O3, Fe2O3 and TiO2 into the Humber Estuary which began in the 1950s from the two Ti-reprocessing plants have not had a noticeable effect on the fluxes observed at Skeffling—there is no evidence for dramatically increased fluxes to the Skeffling site after 1950. This is possibly because inputs of Al, Fe and Ti from these plants have been low in comparison to, or similar to, detrital inputs at the site. Despite the presence of Fe-smelt products in the sediments, and a clear enrichment in Fe above values found in local Scrobicularia clay, the Fe2O3 flux observed in core SM1 (20–25 mg cm 2 yr 1) is broadly similar to Fe fluxes observed in other estuarine sediments (Table 2). The total TiO2 flux to the sediments examined here is much lower than those observed in

Heavy metals and mixing processes in the Humber Estuary 635 T 2. Total flux data from the Humber and other estuaries. Additional data from Bricker (1993), Croudace and Cundy (1995), Cundy et al. (1997) and Cundy et al. (1998)

Humber (Skeffling marsh) Narragansett Bay, U.S. Augusta Bay, Sicily Southampton Water, U.K. Hythe, U.K. Hamble, U.K.

TiO2 (mg cm 2 yr 1)

Fe2O3 (mg cm 2 yr 1)

Pb (g cm 2 yr 1)

Cu (g cm 2 yr 1)

Zn (g cm 2 yr 1)

2·3–4·5 — 1–40 — — —

13–25 — 5–25 — — 7–10

27–47 19–33 1–20 up to 120 21 19

2–23 32–71 7–25 up to 1520 30–73 3–13

50–118 7–85 17–37 up to 130 20–53 19–24

the Augusta Bay area, Sicily by Cundy et al. (1998). (NB. The high flux values in Augusta Bay are thought to be caused by detrital input of Ti from catchment rocks (Cundy et al., 1998)). The comparatively low Ti flux observed here, in combination with the lack of decline in Al and Fe following documented discharge reductions in 1983, indicates that discharges from the Ti processing plants between Immingham and Grimsby have been considerably diluted before they reach the Skeffling sites. It should be noted, however, that the fluxes reported here apply to the Skeffling marsh only, and much higher values may be expected near to the effluent discharge points, as observed in other estuaries (e.g. Cearreta et al., 2000).

Conclusions Surface and near-surface intertidal sediments from the Humber Estuary show elevated concentrations of a range of metals, including Pb, Zn, Cu, Al, Mn and Fe. Metal concentrations are greatest in salt marsh and upper mud flat sediments primarily as a result of grain size variations: sediments are finer on the salt marsh and upper mud flat and increase in grain size in a seawards direction. The vertical distribution of heavy metals in these sediments is controlled by historical pollution inputs and estuarine mixing and resuspension processes. Early-diagenetic remobilization of heavy metals has apparently been limited. Salt marsh sediments preserve a time-integrated record of historical pollutant inputs to the Skeffling area. Cu, Pb and Zn inputs peaked in the mid 20th century, while Ti, Al and Fe, which are input into the Humber from two Tioxide-processing facilities, are only slightly enriched in these sediments. Heavy metal inputs in this part of the Humber are comparable with other industrialized estuaries. Local mixing, resuspension and erosion on the mud flats have produced a complex vertical distribution of heavy metals, with correlatable sedimentary horizons interspersed with mixed sediment layers.

Hence, the mud flat sediments here are unsuitable for studying historical pollution trends.

Acknowledgements This work was funded by the Commission of the European Communities, Directorate General for Science, Research and Development under contract number MAS3-CT95-0022. The LISP-UK programme (LOIS Special Topic no.122) is acknowledged for additional data. We are indebted to Keith Dyer and Malcolm Christie for organizing the logistics of the fieldwork. Helmar Kunzendorf at Risø National Laboratory, Denmark, carried out additional 210Pb and 137Cs analyses on the cores. Ian Croudace at the Southampton Oceanography Centre is thanked for access to dating facilities and helpful comments on an earlier draft of this paper. The principle author would like to acknowledge the guidance of Adrian Cramp who sadly passed away before this paper was published. The principle author is also grateful to Kevin Black for help with core collection, Colin Lewis for assistance with XRF analyses, Tony Oldroyd with the XRD, Peter Fisher with the SEM and Alice Howe for carrying out the grain size analyses. Geotek Ltd are thanked for the use of a multi-sensor core logger to carry out magnetic susceptibility measurements. An anonymous reviewer is thanked for constructive comments which improved the overall quality of this paper.

References Ackerman, F. 1980 A procedure for correcting the grain size effect in heavy metal analyses of estuarine and coastal sediments. Environmental Technology Letters 1, 518–527. Al-Bakri, D. 1986 Provenance of the sediments of the Humber Estuary and the adjacent coasts, Eastern England. Marine Geology 72, 171–186.

636 S. V. Lee and A. B. Cundy Allen, J. R. L. & Pye, K. 1992 Coastal saltmarshes: their nature and importance. In Morphodynamics, conservation and engineering significance of saltmarshes (Allen, J. R. L. & Pye, K., eds). Cambridge University Press, Reading, pp. 1–18. Allen, J. R. L. & Rae, J. E. 1986 Time sequence of metal pollution, Severn Estuary, southwestern U.K. Marine Pollution Bulletin 17, 427–431. Appleby, P. G. & Oldfield, F. 1992 Application of lead-210 to sedimentation studies. In Uranium-series Disequilibrium Applications to Earth, Marine and Environmental Sciences (Ivanovich, M. & Harmon, R. S., eds). Clarendon Press, Oxford, pp. 731–778. Barr, R., Watson, P. G., Ashcroft, C. R., Barnett, B. E. & Hilton, C. 1990 Humber Estuary—a case study. Hydrobiologia 195, 127–143. Baumann, R. H., Day, J. W. Jr. & Miller, C. A. 1984 Mississippi deltaic wetland survival: sedimentation versus coastal submergence. Science 224, 1093–1095. Bricker, S. B. 1993 The history of Cu, Pb and Zn inputs to Narragansett Bay, Rhode Island as recorded by salt marsh sediments. Estuaries 16, 589–607. Cearreta, A., Irabien, M. J., Leorri, E., Yusta, I., Croudace, I. W. & Cundy, A. B. 2000 Recent anthropogenic impacts on the Bilbao Estuary, N. Spain: geochemical and microfaunal evidence. Estuarine, Coastal and Shelf Science 50, 571–592. Croudace, I. W. & Cundy, A. B. 1995 Heavy metal and hydrocarbon pollution in Recent sediments from Southampton Water, Southern England: a geochemical and isotopic study. Environmental Science and Technology 29, 1288–1296. Cundy, A. B., Collins, P. E., Turner, S., Croudace, I. W. & Horne, D. 1998 100 years of environmental change in a coastal wetland, Augusta Bay, southeast Sicily: evidence from geochemical and palaeoecological studies. In Sedimentary Processes in the Intertidal Zone, 139 (Black, K. S., Paterson, D. M. & Cramp, A., eds). Geological Society of London Special Publication, pp. 243–254. Cundy, A. B. & Croudace, I. W. 1995 Sedimentary and geochemical variations in a salt marsh/mud flat environment from the mesotidal Hamble estuary, southern England. Marine Chemistry 51, 115–132. Cundy, A. B. & Croudace, I. W. 1996 Sediment accretion and recent sea-level rise in the Solent, Southern England: inferences from radiometric and geochemical studies. Estuarine, Coastal and Shelf Science 43, 449–467. Cundy, A. B., Croudace, I. W., Thomson, J. & Lewis, J. T. 1997 Reliability of salt marshes as ‘ Geochemical Recorders ’ of pollution input: A case study from contrasting estuaries in southern England. Environmental Science and Technology 31, 1093–1101. Denman, N. E. 1979 Physical characteristics of the Humber. In The Humber Estuary: a selection of papers on present knowledge of the estuary and its future potential given at two symposia arranged by the Humber Advisory Group and the University of Hull, Number 20. N.E.R.C. Publication Series C, pp. 5–9. Flynn, W. W. 1968 The determination of low levels of 210Po in environmental materials. Analytica Chimica Acta 43, 221–227. Goldberg, E. D. 1963 Geochronology with 210Pb in radioactive dating. IAEA, Vienna, 121–131. Grant, J. A. 1986 The isocon diagram—a simple solution to Gresen’s equation for metasomatic alteration. Economic Geology 81, 1976–1982. Grant, A. & Middleton, R. 1990 An assessment of metal contamination of sediments in the Humber Estuary, U.K. Estuarine, Coastal and Shelf Science 31, 71–85. Grant, A. & Middleton, R. 1993 Trace metals in sediments from the Humber estuary: a statistical analysis of spatial uniformity. Netherlands Journal of Aquatic Ecology 27, 111–120. Heijnis, H., Berger, G. W. & Eisma, D. 1987 Accumulation rates of estuarine sediment in the Dollard area: comparison of 210Pb and pollen influx methods. Netherlands Journal of Sea Research 21, 295–301.

House, W. A., Leach, D., Warwick, M. S., Whitton, B. A., Pattinson, S. N., Ryland, G., Pinder, A., Ingram, J., Lishman, J. P., Smith, S. M., Rigg, E. & Denison, F. H., 1997 Nutrient transport in the Humber rivers. The Science of the Total Environment 194/195, pp. 303–320. Jarvie, H. P., Neal, C. & Robson, A. J. 1997 The geography of the Humber catchment. The Science of the Total Environment 194/195, pp. 87–99. Livens, F. R. & Baxter, M. S. 1988 Particle size and radionuclide levels in some west Cumbrian soils. The Science of the Total Environment 70, 1–17. Loring, D. H. 1990 Lithium: a new approach for the granulometric normalisation of trace metals. Marine Chemistry 29, 155–168. Middleton, R. & Grant, A. 1990 Heavy metals in the Humber Estuary: Scrobicularia clay as a pre-industrial datum. Proceedings of the Yorkshire Geological Society 48, 75–80. Milan, C. S., Swenson, E. M., Turner, R. E. & Lee, J. M. 1995 Assessment of 137Cs method for estimating sediment accumulation rates: Louisiana salt marshes. Journal of Coastal Research 11, 296–307. Millward, G. E., Allen, J. I., Morris, A. W. & Turner, A. 1996 Distribution and fluxes of non-detrital particulate Fe, Mn, Cu and Zn in the Humber coastal zone. Continental Shelf Research 16, 967–993. Millward, G. E. & Glegg, G. A. 1997 Fluxes and retention of trace metals in the Humber Estuary. Estuarine, Coastal and Shelf Science 44 (Supplement A), 97–105. National Rivers Authority 1993 The Quality of the Humber Estuary 1980–1990. Water Quality Series No. 12. Neal, C., Smith, C. J., Jeffrey, H. A., Jarvie, H. P. & Robson, A. J. 1996 Trace element concentrations in the major rivers entering the Humber Estuary, NE England. Journal of Hydrology 182, 37–64. Newell, R. C., Newell, P. F. & Trett, M. W. 1984 Benthic communities in the lower Humber estuary in the vicinity of the Tioxide UK outfall at Grimsby. Tioxide Environmental Services Group, Grimsby, 46 pp. North Sea Task Force 1993 North Sea Quality Status Report 1993. Oslo and Paris Commissions, London, 132 pp. Reed, D. J. 1995 The response of coastal marshes to sea-level rise; survival or submergence? Earth Surface Processes and Landforms 20, 39–48. Reed, D. J. 1989 Patterns of sediment deposition in subsiding coastal salt marshes, Terrebonne Bay, Louisiana: the role of winter storms. Estuaries 12, 222–227. Ritchie, J. C., McHenry, J. R. & Gill, A. C. 1990 Application of radioactive fallout cesium-137 for measuring soil erosion and sediment accumulation rates and patterns: a review. Journal of Environmental Quality 19, 215–233. Sharma, P., Gardener, L. R., Moore, W. S. & Bollinger, M. S. 1987 Sedimentation and bioturbation in a salt marsh as revealed by 210 Pb, 137Cs and 7Be studies. Limnology and Oceanography 32, 313–326. Shimmield, G. 1997 The use of radionuclides in assessing salt marsh accretion rates and relative sea-level change. In (Barrett, R. L., ed.). LOIS-Second Annual General Meeting (Abstract) 18–20 March 1997, University of Hull, NERC, Swindon (LOIS publication 323). Watt Committee on Energy 1991 Five years after Chernobyl: 1986– 1991, a review. Chameleon Press Ltd, London, 58 pp. Zwolsman, J. J. G., Berger, G. W. & van Eck, G. T. M. 1993 Sediment accumulation rates, historical input, postdepositional mobility and retention of major elements and trace metals in salt marsh sediments of the Scheldt estuary, SW Netherlands. Marine Chemistry 44, 73–94.