Evidence for the Preservation of Technogenic Tritiated Organic

Article pubs.acs.org/est

Evidence for the Preservation of Technogenic Tritiated Organic Compounds in an Estuarine Sedimentary Environment Ian W. Croudace,* Phillip E. Warwick, and Jenny E. Morris Ocean and Earth Science, University of Southampton, National Oceanography Centre, Southampton, SO14 3ZH, United Kingdom S Supporting Information *

ABSTRACT: The macrotidal Severn Estuary (southwestern UK) has received a broad range of industrial discharges since the beginning of the Industrial Revolution. A more recent anthropogenic input to the estuary has been technogenic tritium (specifically organically bound tritium, OBT). This was derived from a specialized industrial laboratory producing custom radiolabeled compounds for life science research and diagnostic testing from 1980 until 2008. While it was generally acknowledged that the radiological impact of the tritium discharges into the Estuary was small, public concern motivated the company and regulatory agencies to commission several research studies from 1998 to 2005 to better understand their environmental impact. This study examined OBT interaction with estuarine sediment by acquiring a broad range of geochemical and sedimentological data from a suite of sediment cores collected from the northern side of the Estuary. Two important observations are that the OBT compounds are strongly bound to the clay/ silt fraction of sediment and that the down-core OBT profiles in intertidal and subtidal sediments are broadly similar to the discharge record. Geochemical and chronometric methods (Cu, Pb and Zn elemental profiles, 210Pb, 137Cs) provide important corroboration of the OBT record. A key additional piece of evidence that firmly authenticated the established chronology was the discovery of a previously unreported sedimentary marker layer that was generated by a major storm surge that occurred on December 13, 1981. Although this study has provided clear evidence of systematic accumulation of OBT in sedimentary sinks of the region, an estimation of its depositional inventory shows it represents only a small fraction of the total discharge. This modest retention in the principal sedimentary sinks of the Severn Estuary system reflects the particular dynamics of this highly macrotidal sediment starved estuary.

■

INTRODUCTION The 150 km long Severn Estuary with its classic funnel-shape and its high and variable tidal range (14.5 m max) is categorized as an extremely macrotidal vertically mixed estuary. The very large tidal range makes the estuary rare in world terms and explains why it has been regularly proposed as a location for major tidal energy production. It has an overall area of 557 km2 along with an intertidal area of 100km2 but when its seaward extension is included (Bristol Channel) it covers 2000 km2. The estuary has a larger proportion of exposed bedrock (mostly of Carboniferous age) on its base than any other estuary in northwest Europe1,2 indicating the relative importance of erosion over deposition. The most significant accretionary areas for muddy sediment occur close to the mouth of the River Usk (Newport Deep) and Bridgwater Bay that actively accumulate over an area of 5 km2 and 10 km2 respectively. Fringing salt marshes are also well developed along the margins of the Severn estuary. The Estuary has a history of industrialization dating to the Industrial Revolution and has been a major center for coal mines, smelters, incinerators, paper mills, chemical, coal and steel plants, a radiopharmaceutical synthesis center, and three nuclear power stations. Additionally, sewage (treated and untreated) from major urban centers (Bristol, Cardiff, Newport, Gloucester) has added to the estuarine pollutant load. A © 2012 American Chemical Society

substantial list of chemical pollutants have entered the estuary over the last century including heavy metals (Cd, As, Cr, Hg, Cu, Pb, Zn, and Ni), organometallic compounds, phosphates, hydrocarbons, PCBs, pesticides, and radionuclides from civil nuclear and industrial sources. Concentrations of many elements and/or compounds are above sediment quality guidelines and occasionally some exceeded probable effect levels. Hydrocarbons, PAHs, and PCBs are present locally and have been shown to have a high particle affinity and microbial refractivity leading to their burial with fine sediment. Some of these organic compounds are also reported to be sorbed on the ubiquitous finely divided coal particles3−5 derived from the widespread coal mining operations of the 19th and 20th centuries. This study was primarily concerned with determining the fate of a relatively rare industrial source of radioactivity that was discharged to the estuary. This technogenic tritium (specifically organically bound tritium, OBT) entered the Severn Estuary from permitted discharges over some 25 years (1980−2009) from the Maynard Centre at Cardiff (Figure 1). This specialized Received: Revised: Accepted: Published: 5704

December 3, 2011 April 8, 2012 May 7, 2012 May 7, 2012 dx.doi.org/10.1021/es204247f | Environ. Sci. Technol. 2012, 46, 5704−5712

Environmental Science & Technology

Article

Figure 1. The study area showing estuarine sediment types and the location of civil nuclear reactors (O, Oldbury; B, Berkeley, and H, Hinkley Point). Also marked are Bridgwater Bay, BW and Lydney, L.

industrial laboratory was a global center for the production of tritium-labeled compounds used in the biosciences.6 The Centre has variously been the Radiochemical Centre (1971− 1981), Amersham International Ltd. (1981−1997), Nycomed Amersham plc (1997−2001), Amersham plc (2001−2004), and GE-Healthcare Life Sciences (2004−present). From 1981 to 1998 the site discharged designated radioactive wastes and byproducts (including tritiated water and a complex mixture of tritiated organic compounds) into a sewer under strict regulations. After 1998, changes in the proportions of selected compound groups occurred as well as significant reductions in the amount of OBT discharged. From 2003 onward further significant reductions occurred due to changes in the pretreatment of the discharged waste as a result of the opening of a major wastewater treatment works for the region. In 2008 GE-Healthcare announced its plan to cease production of all radiolabeled compounds and to decommission and delicense the site. When the site was first established an understanding of the behavior of the discharged materials (OBT) was fragmentary and it was assumed that biodegradation and vigorous estuarine mixing would act to minimize any environmental impact. Several hydrophilic organic compound groups were expected to have a low impact as they were watersoluble (e.g., alcohols and sugars) and should rapidly disperse and/or biodegrade. There was greater uncertainty over the fate of the more-hydrophobic organic compounds known to exist in the OBT fraction (hydrocarbons, amino acids, proteins, nucleotides, fatty acids, lipids and purine/pyrimidines; J. Williams, personal communication). In the 1990s the environmental impact began to be recognized when bottom-feeding fish (flounder), caught in the general vicinity of the outfall, were reported with tritium concentrations several thousand times

enriched over seawater values. In 1998, the UK Environment Agency called for significant reductions in the discharges of specific compound classes containing tritium from the site and they advised that research programs be commissioned to increase understanding of the accumulation mechanisms and fate of OBT in biota and in sediment. These research studies were eventually commissioned to consider the accumulation in biota and sediment and also to re-evaluate the radiological implications of OBT to potentially affected human populations.7−11 Revised discharge authorizations in 1998 led to a significant reduction in OBT effluents and total tritium concentrations in biota declined significantly.12 The introduction of a new regional wastewater treatment works in 2002 (Cardiff East Waste Water Treatment Works) further reduced the impact of the OBT discharges. Tritium is a radioactive isotope of hydrogen with both cosmogenic and anthropogenic components that emits low energy beta particles (average 5.7 keV, maximum 18.6 keV) during its decay (half-life of 12.3 years) to stable helium-3. Most environmental tritium exists as tritiated water HTO although another class termed organically bound tritium OBT may also be present. Organically bound tritium represents the case of a tritium atom bound to a carbon atom in an organic molecule which is normally considered to be nonexchangeable OBT. There may also be situations where tritium is bonded to N, O, or S in organic compounds (e.g., amines, alcohols, or thiols) but these are not generally classed as OBT. Conversion of HTO to OBT in the environment may occur through biochemical reactions in organisms but this is generally inefficient13 compared with the case where there is direct incorporation of technogenic OBT.9,10 Technogenic OBT uptake by biota in aqueous environments will be affected by 5705

dx.doi.org/10.1021/es204247f | Environ. Sci. Technol. 2012, 46, 5704−5712

Environmental Science & Technology

Article

of the water. Total tritium (3Htot) in the fresh wet sediment sample was extracted using a multitube extraction furnace19 (Pyrolyser-6 system, Raddec Ltd., UK). This measurement approach allowed an evaluation of whether 3H existed as tritiated water (HTO) or organically bound tritium (OBT). For practical purposes OBT was considered to be any tritium that was not water extractable. Tritium was measured using Quantulus ultra low-level liquid scintillation counters (PerkinElmer) with a method detection limit of 0.02 Bq/g. Additional insights into the way that tritium existed in the sediment samples were gained by carrying out Soxhlet, ultrasonic, and microwave extractions of the fresh wet sediment using a series of polar and nonpolar solvents (water, acetic acid, methanol, dichloromethane, and toluene).20 Elemental Analysis. General elemental geochemistry on sediment core material was carried out using a Panalytical Magix-PRO WD-XRF spectrometer to quantitatively analyze subsamples taken at 1 cm increments. Cores were also nondestructively analyzed using an Itrax XRF core scanner.21 This permitted the rapid and automatic recording of a radiographic image and a range of heavy metal variations at 200 μm intervals down the cores. Radiochronological Determinations. Cesium-137 was measured using well-type HPGe gamma spectrometry systems (Canberra UK Ltd., Didcot). Total 210Pb was determined via its grand-daughter 210Po which was leached from 3 g sediment subsamples with aqua regia after addition of 50 μL 209Po spike (0.2 Bq/mL).22 The leachates were evaporated to dryness, redissolved in 1 M HCl and autodeposited on silver discs before counting for 2 days using an Octete alpha spectrometry system (Ortec, Oak Ridge, TN).

OBT interactions and associations with potential nutrients (e.g., sewage) and also the extent to which their dispersal is controlled by their hydrophobicity (e.g., hydrocarbons, lipids) or hydrophilicity (e.g., alcohols, sugars). The global inventory of cosmogenically produced tritium (∼1018 Bq) was swamped by “bomb-tritium” (∼2.4 × 1020 Bq) during the main period of atmospheric nuclear-weapon testing from 1952 to 196314 although this latter contribution has now decreased by approximately 90% through radioactive decay. Considerable quantities of tritium (>50 × 1015 Bq) have also been released into the marine environment of NW Europe from nuclear reprocessing (La Hague15 and Sellafield16) and nuclear power stations ( refs 12,17 and Figures 1 and 2).

Figure 2. Sources and magnitude of tritium discharges to the Severn estuary and NW European seas from various sites (Cardiff plant, the three nuclear power stations at Hinkley, Berkeley, and Oldbury) and the Sellafield and La Hague nuclear fuel reprocessing plants.16,17.

■

RESULTS AND DISCUSSION HTO and OBT Interactions with Sediment and Biota. Discharges of tritium (as HTO) into the Severn Estuary have occurred from three nuclear reactor sites (Hinkley, Oldbury, and Berkeley from 1962 to 2008 (Figures 1 and 2). To evaluate the magnitude of any possible conversion of HTO to particlereactive OBT, sediment and biota samples were collected close to the Sellafield nuclear reprocessing site in Cumbria that discharges significant HTO (Figure 2). Samples of muddy surface sediment all showed low 3H concentrations that were near detection limit23 while macrofauna (Inner Solway flounder, a benthic flatfish) showed low tritium concentrations of 11 Bq/kg. These studies indicated that any conversion from HTO to OBT was small. By comparison Severn Estuary flounder caught near the Maynard Centre discharge pipeline in 2000 showed tritium up to 54 000 Bq/kg.7−10,23 Investigations into how strongly the tritium was sorbed to Severn Estuary sediments collected near Cardiff was determined by leaching samples with seawater and deionised water for 24 h. Results showed only very minor extraction of tritium implying that HTO and other hydrophilic tritium species were not present in any significant quantity. Further investigations using a range of other liquid media (ethanoic acid, methanol, dichloromethane, and toluene) also showed limited extraction20 indicating that the OBT present was strongly bound to the sediment. Most of the OBT was found to be associated with the clay and silt-sized fraction of the sediment (>70% was present in the 60 years old (an age corresponding to approximately three half-lives of 210Pb). The age estimates are cautionary minimum values that are practically limited by measurement uncertainties. The very flat 210 Pb profile seen for the lower part of core NPD7 makes it more likely that all the sediment below 27 cm is >110 years old. The distinctive step profiles seen for all NPD cores cannot be explained as continuous sedimentary sequences but must represent discontinuous sequences and the inflections are best explained through a sedimentary hiatus.

Figure 1) show good similarities with the organically bound tritium discharge record reported by GE-Healthcare when the two profiles are decay-corrected to the core sampling date (Figure 3). The proportion of OBT relative to total tritium discharged was approximately 20% OBT up to 1998 after which the proportion changed to approximately 70−80% (Figure 3). By 2002 the opening of a major sewage treatment works led to significant reductions of what were considered to be highimpact OBT compounds into the estuary. The various features seen in the discharge profile are readily identifiable in the saltmarsh profiles including the marked rise in OBT from zero values in 1982 up to the 1986 maximum in tritium discharges, followed by a slow decline and later rise up to a second maximum corresponding to 1998. The declining concentration after 1998 are also seen which relate to the changes in the composition and volumes of waste following agreements between the site and the regulators. The Peterstone PET2 core data do not show the start of discharges because insufficient core was recovered but it shows the other distinctive features of the discharge record. The tritium profiles for the Newport Deep subtidal cores (NPD10, Figures 3 and 4) also have a general similarity to the OBT discharge record. The finding that cores taken from fine-sediment sinks (saltmarsh and subtidal accretionary environments) have retained a record of the OBT discharges is hitherto unreported and significant. It implies that the OBT compounds (at least certain classes of the more hydrophobic compounds) are systematically immobilized by components in the fine sediment. It was experimentally determined that the mineralogical location of the OBT compounds was in the clay-silt fraction and it is likely that the ubiquitous finely dispersed coal particles (1−2 wt %; 5) found in this fraction could be a likely substrate. These fine coal particles, also termed carbonaceous geosorbents, have been shown elsewhere to be responsible for trapping organic components such as PAHs, PCBs, etc.24 Adsorption of organic compounds to these carbonaceous geosorbents is shown to be nonlinear and generally exceeds the absorption capacity of amorphous organic matter (AOM) by a 5707

dx.doi.org/10.1021/es204247f | Environ. Sci. Technol. 2012, 46, 5704−5712

Environmental Science & Technology

Article

Figure 4. Geochemical and radionuclide profiles for subtidal cores from the Newport Deep. Zinc data are shown from the WD-XRF and the Itrax core scanner.21 The shaded areas correspond to the position of the sedimentary hiatus.

Cesium-137. The 137Cs sediment dating method normally relies on identifying its first appearance in sediment cores and/ or clear attributable spikes in core profiles. In NW Europe the clearest signal is usually that due to the 1963 global “bomb” pulse and occasionally a smaller spike corresponding to fallout from the 1986 Chernobyl accident. In particular coastal sediments of the western UK it is also possible to identify signs of known industrial/nuclear marine discharges from the Sellafield nuclear plant that produced a broad maximum during the period 1970−1980 (refs 13 and 27 and Figure 5). Any

contributions from southward transport of Sellafield discharges (or via Sellafield-labeled sediment)29 are generally considered to be small since the main proportion of this reprocessingderived 137Cs was transported northwards or is retained in the Eastern Irish Sea mudpatch. However, two studies30,31 suggest that an average flux of 1% of the Sellafield discharge is transferred southwards and reaches the English Channel and at least such a quantity might be expected to have been introduced into the Severn Estuary. Other nuclear discharges that would have an influence in the Severn estuary would be 5708

dx.doi.org/10.1021/es204247f | Environ. Sci. Technol. 2012, 46, 5704−5712

Environmental Science & Technology

Article

absence of a 1963 peak for the Newport Deep cores is significant and requires an explanation (see later). Annual Regulatory Monitoring. Additional and independent chronological information is obtained by comparing 137Cs trends in the cores with those reported from annual surface sediment samples collected as part of long running UK regulatory environmental monitoring programs.13,32 Sampling of these surface sediments around the various nuclear sites was carried out from 1977 onward. The time series data reported from four sampling sites (Hinkley Point, Berkeley/Oldbury, Lydney and Cardiff; Figure 5) show a distinctive peak for 1977 for all of the sites except Cardiff where sampling had not started until 1979. This 1977 peak can arguably be identified in the Peterstone 1 profile although it is not displayed in the Newport Deep cores. Such an absence can be explained by erosional loss of the record. The origin of the 1977 sharp peak is uncertain. It has been postulated that a small proportion of the Sellafield discharges travel southwards30,31 and if transported into the Severn Estuary it would be expected to show a broader 137Cs peak than is evident. The narrowness of the upward inflection seen for 1977 implies a single impulse event (Figure 5), possibly arising from an intraestuary pollution pulse(s). Heavy Metal Variations. The long and varied history of anthropogenic (metallogenic) inputs to the Severn Estuary would have produced distinctive temporal variations in the environment and previous research has shown33−36 that heavy metal profiles show features reflecting preindustrial, industrial and postindustrial periods.37,38 In the current study, particularly for the Newport Deep cores, a broad range of heavy metal profiles have been acquired (e.g., Zn, Cu, Pb, Cr, and Sn) that show significant anthropogenic enrichments. There is no evidence that any of these enrichments can be related to early diagenetic elemental redistributions. For conciseness only Zn data are used to illustrate the case (see Supporting Information Table S1). These zinc profiles show abrupt downward steps in four of the five Newport Deep cores and are also coincident with similar sharp changes seen in the radionuclides 137Cs, 210Pb and 3H. Such coincident changes are not to be expected since the various elements and radionuclides have independent source terms. In NPD7 for example Zn shows a 3-fold change from a typical geogenic value for clay-rich sediment37 of 120 ppm to 380 ppm at the step in the core. The extreme sharpness of this change cannot be taken to mark the onset of industrial inputs because of its abruptness and also since the upper part of the step is known to correspond to approximately the early 1980s based on the radionuclide evidence. A more credible explanation is that evidence for the onset of Zn pollution input has been expunged from the sedimentary record, most likely through erosion. Timing, Origin and Scale of the Sedimentary Hiatus. The various geochemical observations seen in the Newport Deep cores strongly imply that the step profiles originate from a significant and singular erosional event. Accurately dating this event would provide an important constraint for the OBT record because it lies at the boundary of the first appearance of OBT in the Newport Deep cores. The possibility that the hiatus resulted from a man-made process such as dredging has been discounted as enquiries suggested that no such engineering operations had occurred in the area (Associated British Ports, Newport, personal communication). Using the radiochronological data to determine sediment accumulation rates it is clear that the downward step corresponds approximately to

Figure 5. 137Cs profiles for annual regulatory surface sampling of sediment36 along with the profile for Peterstone-1 saltmarsh core with a calculated sediment accumulation rate (SAR) of 1.25 mm/y (based on OBT profile). The Sellafield discharge record is shown to evaluate whether it contributes to 137Cs in the Severn.16

from the three nuclear power stations located on the southern side of the estuary at Berkeley, Hinkley Point, and Oldbury; these started operating in 1962, 1965, and 1967, respectively (Figure 5). It is significant that none of the Newport Deep or Peterstone profiles shows a peak that could be assigned to the classical 1963 bomb pulse. The Peterstone cores do not show any increases from zero data for 137Cs, and this coupled with the absence of a 1963 peak is explained by the cores being too short to see the earlier record. The situation for the longer Newport Deep cores is different and they do show zero values at depth. Overall, the 137Cs profiles for Newport Deep cores display a broad elevation that could relate to discharges from Sellafield and/or the three Severn Estuary nuclear reactor discharges. The 5709

dx.doi.org/10.1021/es204247f | Environ. Sci. Technol. 2012, 46, 5704−5712

Environmental Science & Technology

Article

Figure 6. Loge Pb-210 profiles of Newport Deep subtidal cores with calculated estimates of the sediment accumulation rate (SAR) using the Simple Model.26 The triangular shaded areas below the storm-induced hiatus represent projections of each regression line to the supported Pb-210 level. They indicate the minimum amount of lost sediment only.

event readily accounts for the geochemical discontinuities and sedimentological observations. Interestingly and instructively, more recent investigations by the authors further afield (Bridgwater Bay) have found additional evidence for the erosive nature of this storm surge in the sub-tidal sediment record. Estimating the impact of the storm surge on the sedimentary record can be inferred from the Newport Deep core data using 210 Pb dating. With the Simple Method a linear negative trend is expected for natural log activity data versus depth down to the 210 Pbsupported values of ∼10 Bq/kg (Figure 6). A steady negative decline is expected to be able to continuously track a trend over three to five half-lives of 210Pb (limited by counting statistics). In this study, estimating the minimum amount of lost sediment is made using a simple graphical technique. At least 25 cm for cores NPD7 and NPD10, 10 cm for core NPD8 and 35 cm for core NPD3 are minimum estimates of eroded sediment but these are likely to be conservative estimates. For NPD7, for example, the combination of unsupported 210Pb and geogenic concentrations of Zn (∼120 ppm) implies that the subhiatus sediment is preindustrial and was possibly deposited more than 100 years ago. Cores NPD3 and NPD4 were both collected from the deeper part of the Newport Deep (approximately 5 m water depth below the highest astronomical tide; see Supporting Information Table S2) The Preservation of OBT in Severn Estuary Sediment. This study has demonstrated that some of the classes of tritiated organic compounds (more hydrophobic compounds) discharged from the Cardiff radiopharmaceuticals laboratory became strongly sorbed to muddy estuarine sediment. The OBT-labeled sediment must have mixed with unlabeled

the early 1980s. A compelling and key observation that allowed the layer to be precisely dated was found when a tightly organized layer of centimeter-scale pebble debris was discovered in Newport Deep core NPD10. In this core the sediment consists of regularly laminated mud but at −27 to −30 cm a pebbly layer comprising cobble-sized clasts of various shapes and compositions (granite, sandstone, slate, and oxidized coal (see Supporting Information Figure 1) was identified. The brown muddy sediments below the pebbly layer also had a subtly different color from the sediment above which seems to reflect the lower abundance of finely divided coal particles. High resolution CT radiographic images of the pebble layer showed coarse lithic fragments mantled by finer clastic debris (see Supporting Information Figure 1) and such an arrangement was consistent with accretion of the finer grains around the larger spherical grain during energetic rolling of the larger rounded clasts. An energetic and erosive process that could have led to significant sediment loss along a broad estuarine tract would be a storm surge. A suitable candidate event for a powerful storm was identified from a literature search that described a storm surge occurring on December 13, 1981. This arose from a secondary depression that crossed southern Britain and led to the highest water levels in the Severn Estuary seen during the 20th Century.38 The severe winter storm reportedly caused major flooding along the southern Severn Estuary coast and a surge developed with a peak height of approximately 2 m (near Newport) driven through a combination of high spring tides and westerly wind speeds up to 95 mph. It is notable that this study is the first to report the effects of this powerful storm on the sedimentary record. The severity and timing of the 1981 5710

dx.doi.org/10.1021/es204247f | Environ. Sci. Technol. 2012, 46, 5704−5712

Environmental Science & Technology

Article

pebble layer. This material is available free of charge via the Internet at http://pubs.acs.org.

sediment and would have become distributed in the estuary according to the operational sedimentary regime. Portions of the labeled sediment accumulated in the various fine sediment sinks such as the subtidal Newport Deep and the fringing saltmarshes. These accumulations must have occurred in a systematic manner in order to preserve OBT profiles that broadly matched the OBT discharge pattern (decay corrected). This systematic sediment accumulation behavior is supported by the other observations that show a chronology can be obtained using a variety of proxies (heavy metals, 210Pb, 137Cs and sediment marker layer). The preservation of a sedimentary OBT record shows that only very limited biodegradation or remobilization could have occurred over the two decades or so of discharge. The absence of biodegradation results from the nonhydrolyzable and insoluble nature of the sorbed OBT compounds and their sorption to clay-sized particles (clays and/or carbonaceous materials). The accumulation of such materials would be associated with the low sediment porosity that serves to inhibit bioavailability, the spread of bacterial enzymes and microbial degradation.39 The recognition that a portion of the discharged Amersham OBT became effectively isolated in the fine-grained sediment of the Severn Estuary shows the chronometric potential of the likely compounds in the OBT (hydrocarbons, proteins, lipids, etc.). It remains to be established that the buried OBT was protected from subsequent remobilisation and biodegradation through sorption to carbonaceous geosorbents that are known elsewhere to effectively trap a variety of organic compounds (e.g., PCB, DDT, PAH, etc.).24,25 The total inventory of OBT deposited in the Newport Deep sediment (decay corrected to 2000) can be approximated from the data determined in this study. By taking the area of Newport Deep (5km2) and taking an average OBT activity concentration in the sediment of 0.1 Bq/g to a depth of 30 cm and the sediment density to be 1.6 g/cm3, the total activity of 3 H in the Newport Deep is estimated to be 240 GBq. Similarly, total OBT in the fringing saltmarsh sediments (e.g., Peterstone marsh) is estimated to be 28−192 GBq. On the basis of these approximations the proportion of deposited OBT in the main estuarine sinks represents