Obtaining a temporal record of heavy metal pollution from reservoirs in the southern Pennines, UK.
LAURA SHOTBOLT
Telford Institute of Environmental Systems, Division of Geography, University of Salford, Greater Manchester M5 4WT, UK.
SIMON M. HUTCHINSON
Telford Institute of Environmental Systems, Division of Environmental Resources, University of Salford, Greater Manchester M6 6PU, UK.
ANDREW D. THOMAS
Telford Institute of Environmental Systems, Division of Geography, University of Salford, Greater Manchester M5 4WT, UK.
ANDREW J. DOUGILL
School of the Environment, University of Leeds, Leeds LS2 9JT, UK
Abstract. The southern Pennine uplands, UK, are surrounded by urban industrial areas and have been subject to a long history of industrially derived atmospheric pollution. The potential for the sedimentary record of reservoirs in this area to provide a temporal record of heavy metal inputs is assessed. Central cores from five reservoirs demonstrate markedly similar magnetic susceptibility (lf) profiles both within and between reservoirs, and therefore undisturbed sediment chronology in these cores. Heavy metal profiles of sediment cores appear largely unaffected by post-depositional metal mobilization with the exception of Fe and Mn. Reservoir sediments are, concluded to provide a temporalrecord of heavy metal inputs and a valuable resource for studying the history of atmospheric pollution in this area.
1
Introduction and aims
The southern Pennine uplands, north-central England, are surrounded by the urban-industrial areas of Manchester, Leeds, Bradford, Stoke and Sheffield (Fig. 1). North-central England is considered the ‘cradle’ of the industrial revolution, and these uplands have been subjected to atmospheric pollution from surrounding cities since the eighteenth century. The high density of reservoirs in these uplands make it an important water supply area for the surrounding population. Both contemporary deposition, and the past accumulation of pollutants (in particular heavy metals) in potentially erodible soils and in reservoir sediments is of concern.
No long-term record of heavy metal deposition is available in this region; however, the use of lake sediments as an historical record of atmospheric pollution is a well established technique. Temporal records of atmospherically derived particulates including trace metals, spherical carbonaceous particles (SCP), magnetic minerals, and polycyclic aromatic hydrocarbons (PAH), have been documented in lake sediment cores across the world (e.g. Rippey, 1990; Hermanson, 1993; Flower et al., 1997).
While there are no lakes in the study area, the application of techniques used in lake sediment analysis to reservoir sediments may provide an alternative historical record. In the UK, reservoir sediments are rarely used as a temporal record: water level drawdown can result in the erosion of marginal sediments, and their redeposition in central areas, disturbing sediment chronology (Anderson et al., 1988). Exceptions include work on heavy metal contamination in reservoirs in the English Midlands (eg. Foster et al., 1991; Charlesworth and Foster, 1999) and a preliminary assessment of the potential of southern Pennine reservoirs for reconstructing water quality and pollution histories (Anderson et al., 1988). Anderson et al. conclude that, while it will be difficult to find sites with undisturbed sediment in the southern Pennines, it should be possible to obtain a temporal record of pollution from reservoir sediments.
To reconstruct a record of heavy metal deposition onto this area from the reservoir sedimentary record it is necessary to ascertain the following: A, the sediment profile has not been physically disturbed; B, there has been no modification of the heavy metal profile by metal solubilisation and mobilization; and, C, the record of atmospheric deposition has not been distorted by a variable input of sediment and associated metals from the drainage basin (either from the weathering and erosion of bedrock, or from metals previously deposited onto the drainage basin).
As part of a wider project investigating the history of heavy metal deposition onto the southern Pennines, this study addresses criteria A and B. It aims to demonstrate that, where undisturbed and unmodified, sediment profiles can represent a record of the net input of metals into the reservoir over its lifespan. This record of inputs can then be used (where variable drainage basin inputs are considered) to reconstruct a history of heavy metal deposition.
Site selection and core collection
Of the 192 reservoirs in the southern Pennines (Fig 1), only a few will be suitable for reconstructing pollution histories (Anderson et al., 1988). A selection procedure (Table 1) was designed to eliminate reservoirs where disturbance to the sediment is likely (Shotbolt et al., in press), leaving those with the highest probability of an undisturbed and retrievable sediment. Agden, Broomhead, Howden, Langsett and Midhope reservoirs were initially selected for analysis (Fig. 1). Table 2 summarizes the characteristics of these reservoirs and their drainage basins.
Between eight and ten 1 m sediment cores were collected from each reservoir with a Mackereth mini-corer (Mackereth, 1969). Cores were taken from central,
1
intermediate and marginal zones. A depth sensor was used to ensure core sites avoided former river channels.
Determination of sediment stratigraphy.
To determine whether sediment profiles have been disturbed (criteria A) volume magnetic susceptibility () was logged at 2 cm intervals down the cores using a Bartington MS2C loop sensor. Magnetic susceptibility (is a measure of the ‘magnetisability’ of a material (Dearing, 1999). The magnetism is derived predominantly from primary and secondary minerals eroded from the catchment, and magnetic minerals deposited from the atmosphere, the majority of which are the product of industrial processes and vehicular emissions (Thompson and Oldfield, 1986). The magnetic profile of an undisturbed sediment core will therefore reflect changes in the amount and type of magnetic inputs through time. Many studies have shown susceptibility profiles of central cores have a distinct ‘signature’ that is repeated across this zone (eg. Thompson et al., 1975) suggesting either sediment has remained undisturbed throughout this deep zone, or any disturbance has affected the whole area. This can then be used to identify any anomalous or disturbed cores (Thompson and Oldfield, 1986) and to identify cores suitable for further analysis. Fig 2 shows cores from Howden reservoir. Central cores are broadly similar and cores 3 and 4 from the deepest zone are almost identical. In contrast, frequently reworked marginal cores show no such consistency.
As cores from the deepest zone of each reservoir were found to be most replicable, three cores were selected from central areas of each reservoir for further analysis. These cores were extruded, divided into 1 cm sections and dried at 40ºC. Water content and bulk density were determined and used to identify the original soil
1
surface where this was retained. Organic matter content was determined, as loss-on-ignition, after heating to 500°C for eight hours (Rowell, 1994). Sections were homogenised and packed into 10 ml pots for single sample mass specific susceptibility measurement. A Bartington MS2B single sample dual frequency sensor was used to measure low (lf) and high frequency (hf) susceptibility. Frequency dependent susceptibility (fd%) was calculated from this (Dearing, 1999). The measurement of susceptibility on a dry mass basis allows a more accurate and detailed analysis of the susceptibility profile. Frequency dependent susceptibility detects ultrafine (<0.03 m) superparamagnetic minerals produced largely by biochemical processes in soil (Dearing, 1999). All samples were measured from one core and alternate samples from the other two selected. Table 3 summarizes the main characteristics of sediment cores from all reservoirs.
Fig. 3 showslf profiles of key cores from all reservoirs. The main features are a peak between 10 and 15 cm from the surface and a general similarity between the profiles. This similarity, both within, and between reservoirs, strongly suggests sediments have not been significantly disturbed. The exception to this is Howden reservoir, which has a zone of lower susceptibility at a depth of 15 to 28 cm. This zone was also evident in deepwater cores collected in a previous study, where 137Cs dating located the whole zone between 1954 and 1963 (Hutchinson, 1995). This low susceptibility zone is thought to be the result of the inwash of material removed from exposed marginal sediments and drainage basin soils by a severe storm following a protracted and exceptionally low drawdown in 1959 (Shotbolt et al. in press). Reservoir and meteorological records show neither drawdown, nor the subsequent storm event were as extreme in the other reservoirs, explaining the absence of any corresponding inwash of low susceptibility material. It appears that only a combination of extreme drawdown and heavy rainfall will disturb sediment in the deepest zone of these reservoirs.
The similarity between susceptibility profiles also implies the same origin of magnetic inputs. Susceptibility will be predominantly derived from atmospheric particulate deposition in this area: previous work has shown magnetically impoverished Millstone Grit bedrock inputs to dilute lf concentrations in the sediment (Hutchinson, 1995).
Radiometric dating is in progress. However, an approximate time scale assuming a constant rate of sedimentation, and using cumulative dry bulk density to remove effects of consolidation (Robbins, 1978), is used to date subsurface peaks to between 1969 and 1981 (Fig 4). While constant rates of sedimentation are unlikely, this approximate dating method is useful, demonstrating peaks are synchronous with peaks of atmospheric pollutants found in studies across the UK. For example, Rose et al. (1995) dated subsurface SCP peaks from 19 lakes from the UK and Ireland to between 1969 and 1982.
The general similarity between susceptibility profiles both within and between reservoirs shows, therefore, that neither variable drainage basin inputs, nor sediment disturbance obscure overall deposition trends. The sedimentary record of these reservoirs appears no less valuable than that of natural lakes. Furthermore, the relatively rapid sedimentation (Table 3) means higher resolution data, and the original soil surface/sediment interface provides a readily obtainable date for the base of the core.
Heavy metal analysis of reservoir profiles.
1
Heavy metal analysis was carried out on the three selected cores from each reservoir using a modified aqua-regia method by Jones and Laslett (1994). Figure 4 shows metal concentration in cores from Agden and Howden Reservoirs. Table 3 summarizes the range and mean metal concentrations in all sediment cores. In all reservoirs concentrations of Zn > Pb > Cr > Cu = Ni. Analytical accuracy was assessed by including two blanks, two duplicates and a certified reference sample in each batch of thirty samples (Table 4).
Mn profiles show very high concentrations at the sediment surface. This corresponds closely to water content profiles (Fig 4) and demonstrates solubilisation and mobilization of Mn through the reduction of Mn oxides and hydroxides in anaerobic sediment (Song & Muller, 1999). Fe profiles also show an increase in the upper 5 cm of both cores largely attributable to redox induced mobilization. Mn and Fe profiles will not therefore represent a temporal record of inputs of these metals to the reservoir. However, there is no evidence of mobilization of any other metals: peaks in Cu and Zn can be attributed to short term cycling of metals between the surface sediments and the water column and are unlikely to affect the long term metal record (Morfett et al., 1988). On the basis of analysis so far, heavy metal profiles, with the exception of Mn and Fe do appear to represent a temporal record of inputs of those metals thus satisfying criteria B.
To conclude, sediments from the deepest part of these reservoirs are physically undisturbed and heavy metal profiles appear largely unaffected by remobilization processes. Sediment profiles therefore represent a record of the net input of metals into the reservoir. Future work aims to evaluate the importance of variable inputs of metal contaminated sediment from the drainage basin, and natural baseline metal levels, allowing interpretation of metal profiles as a record of atmospheric deposition onto this area.
Acknowledgements The authors thank Beth Lethers, Tim Carr, Polly Hardy and Tony Meachen for assistance with sample collection, and Gustav Dobrzynski for cartography. Studies were undertaken in association with a research grant provided by the Manchester Geographical Society.
References
Anderson, N.J., Patrick, S.T., Appleby, P.G., Oldfield, F., Rippey, B., Richardson, N., Darley, J. & Battarbee, R.W. (1988) An assessment of the use of reservoir sediments in the southern Pennines for reconstructing the history and effects of atmospheric pollution. Working Papers no.30, Palaeoecology Research Unit, Dept. Geography, University College London.
Charlesworth, S.M. & Foster, I.D.L. (1999): Sediment budgets and metal fluxes in two contrasting urban lake catchments in Coventry, UK. Applied Geography. 19, 199-210.
Dearing. J.A. (1999) Environmental Magnetic Susceptibility: Using the Bartlington MS2 System. Users Handbook. Second Edition.
Flower, R.J., Politov, S.V., Rippey, B., Rose, N.L., Appleby, P.G., & Stevenson, A.C. (1997) Sedimentary records of the extent and impact of atmospheric contamination from a remote Siberian highland lake. The Holocene.7, 161-173.
Foster, I.D.L., Charlesworth, S. M. & Keen, D. H. (1991) A comparative study of heavy metal contamination and pollution in four reservoirs in the English Midlands. Hydrobiologia.214, 155-162.
Hermanson, M.H. (1993) Historical accumulation of atmospherically derived pollutant trace metals in the Arctic as measured in dated sediment cores. Wat. Sci. Tech. 28, 33-41.
Hutchinson, S.M. (1995) Use of magnetic and radiometric measurements to investigate erosion and sedimentation in a British upland catchment. Earth Surf. Processes and Landforms. 20, 293-332.
Jones, B.R. & Laslett, R.E. (1994) Method for analysis of trace metals in marine and other samples. MAFF.
Mackereth, F.J.H. (1969) A short core sampler for sub-aqueous deposits. Limnol. Oceanogr.14, 145-151.
Morfett, K., Davison,W. & Hamilton-Taylor, J. (1988) Trace metal dynamics in a seasonally anoxic lake. Envir. Geol. Wat. Sci.11, 107-114.
Rippey, B. (1990) Sediment chemistry and atmospheric contamination. Phil. Trans. R. Soc. Lond. B 327, 311-317.
Rose, N.L., Harlock, S., Appleby, P.G. & Battarbee, R.W. (1995) Dating of recent sediments in the United Kingdom and Ireland using spheroidal carbonaceous particle (SCP) concentration profiles. The Holocene.5, 328-335.
Robbins, J.A. (1978) Geochemical and geophysical applications of radioactive lead. In: The biogeochemistry of lead in the environment. (ed. By J. O. Nriagu) Elsevier, Amsterdam. 285-408.
Rowell, D.L. (1994) Soil science: methods and applications. Longman Scientific & Technical, Harlow.
Shotbolt, L., Thomas, A.D., Dougill, A.J. & Hutchinson, S.M. Reconstructing the history of heavy metal pollution in the southern Pennines from the sedimentary record of reservoirs: methods and preliminary results. The North West Geographer (in press)
Song, Y. & Muller, G. (1999) Sediment-water interactions in anoxic freshwater sediments: Mobility of heavy metal and nutrients. Springer–Verlag, Berlin.
Thompson, R. & Oldfield, F. (1986) Environmental Magnetism. Allen & Unwin: London.
Thompson, R., Battarbee, R.W., O’Sullivan, P.E. & Oldfield, F. (1975) Magnetic susceptibility of lake sediments. Limnology and Oceanography. 20, 687-698.
TABLES
Table 4
FIGURE CAPTIONS
Figure 1. The southern Pennines and location of reservoirs.
Figure 2. Volume susceptibility () profiles of cores from Howden Reservoir.
Figure 3. Mass specific susceptibility (lf) profiles from southern Pennine reservoirs.
Figure 4. Heavy metal, loss-on-ignition, and organic matter profiles from Agden and Howden Reservoirs.
1