Centre for Ec ology & Hy drology

Disposal of oiled beach sand in c onstaLsJ á Ce ntre for Ec olog y & Hy drology

Disposal of oiled beach sand in coastal soils

R.E. Daniels,' A.F. Harrison,2 G.H. Hall,3 J.S. Garnett,2 A.P. Rowland,2 R. Scott,2 and J. Davies'

' ITE Furzebrook ITE Merlewood 3 1FE Windermere BGS Wallingford

c a Final Report to Maritime and Coastguard Agency ITE is one of four Institutes which form the Centre for Ecology and Hydrology, a part of the UK Natural Environment Research Council. The mission of CEH is to advance the sciences of ecology, environmental microbiology (including virology) and hydrology. It will do this through high-quality and internationally recognised research which will lead to a better understanding and quantifi cation of the physical, chemical and biological processes relating to land and freshwater, and of living organisms within these environments. ITE is responsible for research into all aspects of the terrestrial environment and its resources, and has specialist staff based at six research stations, each with its own particular expertise and regional emphasis.

This report is an official document prepared under contract between the Maritime and Coastguard Agency and the Natural Environment Research Council. Permission to use material contained within the report should be sought from both MCA and NERC.

September 1999 TABLE OF CONT ENTS

EXE CUTIVE SUM MARY

1. INTRODUCTI ON 5 1.1 The incidence of spills 5 1.2 The problem of disposal 5 1.3 Aims of the proj ect 5

2. NATURAL BIO REME DIATIO N 7 2.1 Hydrocar bon degradation 7 2.2 Th e distribution of hydrocarbon degr ader s in the environment 8 2.3 Biodegr adation following spills 8 2.4 The sta rting material 8

3. EXPERIM ENTAL APPROACH 11 3.1 Gener al philosophy 11 3.2 Exper imental sca les 11 3.3 Toxicity trials 12 3.4 Lysimeter experiments 12 3.5 Eskmeals field trials 13 3.6 Extensive study • 15 3.7 First Pendine deposit 16 3.8 Second Pendine deposit 17 3.9 Micr obiological techniques 19 3.10 Decomposition activity 19 3.11 Assessment of degradation 20 3.12 Assessment of leaching 20

4. FACTORS AFFECTING DEGRADATION 21 4.1 Patter n of degr adation 21 4.2 Effects of oil type and concentration 21 4.3 Effects of temper ature 23 4.4 Effects of moisture 23 4.5 Effects of sand type 27 4.6 Effects of micr obial differ ences 29 4.7 Effects of treatment 31 4.8 Inter actions 34

5. DE GRADATIO N PRODUCTS AND THEIR FATE 35 5.1 Introduction 35 5.2 Heavy metals 35 5.3 Changes in hyd rocarbon composition 36 5.4 Vertical migration 36 5.5 Leaching of miner al ions to groundwater 37 5.6 Leaching of hydrocarbons to groundwater 37 6. SITE RECOVERY 39 6.1 Na tu re conser vation value of dunes 39 6.2 Toxicity testing of plants 40 6.3 Recovery of dune sites 40

7. CONCLUSIONS 45

8. IMPLEMENTATION 47 8.1 Site consider ations 47 8.2 Site factors infl uencing degradation 47 8.3 Management factors 48 8.4 Public per ception 49

9. F UTURE RESEARCH NEEDS 51 9.1 Enh ancement of degr adation 51 9.2 Site selection 51

10 ILLUST RATIO NS 53

11. ACKNOWLEDGEMENTS 57

12. REFERENCES 59 EXE CUTIVE SUMMARY

Background Experiments to study the factors identi fi ed in the literature as being important in controlling rates Disposal of oiled sand removed from beaches of decomposition were conducted in the contaminated as a result of oil spills at sea poses a laboratory and in the fi eld. Potential for maj or problem. Regulatory and environmental movement of the introduced hydrocarbons, or constraints are increasingly limiting the use of their breakdown products was measured in landfill sites to dispose of such material . In many laboratory scale experiments and in fi eld trials. remote ar eas the only acceptable disposal method Toxic effects of oil residues on plants were is the long-distance transport of the OBS to an examined in glasshouse tests and the capacity for offi cially licensed disposal sit e. site recovery was measured in fi eld trials. Towards the end of the project, issues of .A research programme to investigate the feasibility selection and management of suitable sites were and environmental acceptability of an alternative explored with representatives of local authorities, method of disposal was instigated by the Maritime statutory 'conservation agencies and the and Coastguard Agency (formerly Marine Pollution Environment Agency. Control Unit of the Coastguard Agency). Its purpose was to assess the practicality and Experimental Work environmental acceptability of using naturally occurring populations of soil micro-organisms to La boratory experiments degrade weathered oil residues incorporated into M icrobiological studies examined the stimulation sandy coastal soils. Especiall y in low grade, non- of respiration of microbial communities when sensitive areas of coastal sand dunes and dune challenged with introduced hydrocarbons andt pastures. If the method proved successful it would the ability of cultured organisms to break facilitate disposal of oil ed beach sand (OBS) in aromatic ring structures. remote areas where disposal is dif fi cult and expensive. Small-scale experiments were used to examine the responses of plants found in coastal dunes and The work was carried out by a consortium led by dune pastures to the presence of different The Institute of Terrestrial Ecology and began in concentrations of weathered and unweathered oil 1994 residues.

M ain conclusions A lysimeter facility (consisting of 30 tubes 2 m deep x 27.5cm diameter) at ITE Merlewood was used to examine the infl uence of a number of • M icrobial action can rapidly reduce the factors on the rate of hydrocarbon degradation concentration of hydrocarbons in oiled sand. and the composition of drainage water in Under favourable conditions, the oriemal replicated experiments. hydrocarbon content can be more than hal ved in less than a year. Field Trials A fi eld trial at Eskmeals, Cumbria was used to • The process has produced no adverse compare degradation rates in replicated plots environmental effects. There was no movement containing buried and landfarmed OBS. of hydrocarbons to groundwater, and plant and Measurements were also made of air and soil animal communities were not affected by the temperatures, and rainfall and water table presence of weathered oil residues in the soil. fl uctuations. Groundwater samples taken from beneath the plots were analysed to determine M ethodology whether hydrocarbons were present. An initial feasibility study showed that the problem of OBS disposal in coastal areas required an A larger scale monitoring trial, following oil ecological rather than a technical approach. residue decomposition, was conducted at Observations were made in the laboratory and in Pendine, where OBS prepared from a stranding of the fi eld to obtain information on the processes tar balls was placed in a dune hollow. The deposit involved in microbial degradation of petroleum was characterised by uneven distribution of hydrocarbons and their ecological consequences. highly weathered oil residues. This site was also used to examine the progress of plant colonisation practical terms to increase the efficiency of natural and sand stabilisation within the dune area. breakdown. However, if the rate limiting factor is related to physical or chemical properties of the A second experiment at Pendine was designed to sand containing the oil residues some form of measure the decomposition of a homogeneous OBS treatment may be possible which is capable of prepared with a high concentration of fuel oil to enhancing the rate and completeness of detect any movement of hydrocarbons from the breakdown. For example, the addition of nutrients deposit to groundwater. may increase the rate of decay but investigations into such additions were outside the scope of this Geographical Variation project. Because areas of the coast differ in climatic conditions and in the type of sand found, an Eff ect of oil concentration extensive trial was carried out, in which small bags Extensive weathering of oil residues reduced the of OBS were buried at fi fteen sites around the British initial decomposition rate, indicating that the most coast from Shetland to Devon. These bags were volatile or soluble components were those which recovered at intervals and analysed for remaining are likely to be most readily metabolised by micro- hydrocarbon content. A collateral trial at a single organisms. Increasing concentration of site (Eskmeals) contained bags of OBS made from hydrocarbons in OBS, up to about 7.5% increased sands collected at all fiftee n sites. the rate of decomposition, but above this concentration there was rapid decrease in Results and observations degradation rate . This suggests there is an optimum starting concentration (between 5% and Main f indings 7.5%), for OBS to obtain maximum breakdown of Trials at all scales showed that micro-organisms hydrocarbons. (principally bacteria) with a capacity for hydrocarbon degradation were present in coastal Eff ect of Temperature sandy soils and that these populations responded Temperature had a clear infl uence on degradation rapidly to the presence of oil residues. In all cases, a rate, as demonstrated in laboratory and lysimeter rapid build-up of the microbial hydrocarbon experiments, where temperature could be degrader populations was accompanied by controlled or measured continuously. Seasonal enhanced soil respiration and a reduction in the temperature dependence is more likely to be concentration of hydrocarbons present in the soil. important than diurnal in terms of site management, and timing of disposal may be Pattern of Decay important in establishing the precise nature of the A common pattern of degradation was found, with decay curve. an initial rapid reduction in hydrocarbon content of OBS , followed by a progressive slowing of the rate Eff ect of moisture of loss with time. This pattern of decomposition can As aerobic decomposition is the predominant best be described by the power function process by which hydrocarbons are degraded, it (Y = A . X"). Actual equations defi ning the decay was considered important to examine factors curves varied depending on the type and infl uencing water content of the OBS. concentration of the oil residue and a number of Waterlogging after periods of rain appeared to be external factors, including temperature, moisture and responsible for reductions in measured microbial nature of the sand into which the oil residues were activity in lysimeter experiments. The infl uence of incorporated. The shape of the decay curve may be waterlogging in reducing hydrocarbon a result of the more readily decomposable decomposition was also seen in the fi eld plots at components of the oil residues being degraded Eskmeals. Those plots at the foot of the sloping quickly, whilst more recalcitrant compounds persist. dune pasture area (where groundwater There may also be some rate limiting factor related to penetration of the OBS deposits was more microbial nutrition or to changes in buffering frequent and prolonged) showed a slower rate of capacity of the soil which means that the bacteria hydrocarbon degradation than those at the top of operate in progressively less favourable nutrient or the slope and in drier dunes. In contrast, within p14 conditions. the unstructured, free draining, sand of the fi rst Pendine experiment, microbial respiration was If the nature of the oil residue s is the main enhanced by periods of rain, suggesting that determinant, then there is little that can be done in water may have been a limiting factor at this site. M icrobial diff erences between sites Future work Differences were found between the initial populations of hydrocarbon-degrading bacteria in This project has demonstrated the feasibility of this beach and dune sands from the same site, and method for disposal of OBS and has covered most between sands taken from different beaches or of the background scientific aspects. Future work dunes. However, initial bacterial counts did not needs to concentrate primarily on the regulatory necessarily reflect the patterns of decomposition and environmental protection issues. Additional found. The results suggest that there may be large research should also be directed at refining the site differences in adaptability of microbial populations selection and management procedures, and at between sites. identifying more prec isely the relationships between particular co ntrolling factors and Comparing burial wills land farming degradation effi ciency. The Eskmeals trials showed that, landfarming with frequent ploughing, was more efficient in reducing the concentration of hydrocarbons present in a given weight of OBS than burial. However, for the same land area the total amount of oil residues degraded was greater in burial plots than in landfarming plots.

Migration of hydrocarbons to ground water Hydrocarbons from OBS deposits only migrated a few centimetres into underlying sand deposits. No evidence was found of the increased presence of hydrocarbons above background concentrations in groundwater.

Ecological impacts The main environmental impacts arise from mechanical disturbance of sites rather than from the presence of hydrocarbons. At all the field sites, colonisation of experimental areas by plants was observed. At those sites where disturbance was Minimal rapid recovery of vegetation was observed . Animal colonisation of the plots at Eskmeals followed development of near-complete plant cover.

Discussion and Consultation

To explain the results and demonstrate the low environmental impact of this alternative method of disposal for OBS a series of meetings, presentations and site visits was arranged. Ori anisations consulted included: English Nature Scottish Natural Heritage, The Countryside Council for.Wales. The Environment Agency, Scottish Environmental Protection Agency and local authorities.

Site selection criteria

A site selection procedure has been developed together with a preliminary set of guidelines on site management designed to maximise the efficiency of degradation of OBS. á 1. INTRODUCTION

1.1 Aims of the project ashore on sandy bearhes on low-energy coasts is likely to stay there and degrade slowly or be The need to fi nd an effective and environmentally- redistributed by successive tides. It then remains a acceptable means for disposal of oiled beach sand physical threat to wildli fe and a problem for human prompted the Marine Pollution Control Unit to use of the beaches. There may also be signifi cant commission a project to investigate the possibility economic impacts where commercial fi sheries or of using a natural bioremediation method. A one- popular tourist beaches are impacted. For which year Feasibility Study produced a report in 1993 ever of these reasons, clean-up and disposal and, on the basis of information obtained, an . become priorities. experimental programme began in January 1994. The project had two major aims: Although clean-up of such beaches is not practically diffi cult, given ease of access, disposal • To test the suitability of burial and landfarming of the material removed does pose an increasingly of weathered oil residues in coastal sandy serious problem. The options for disposal fall into environments, as a practical alternative to three broad categories: disposal in landfi ll sites. • Recovery and re-use • To assess whether such operations would be • Stabilisation and storage environmentally acceptable. • Degradation

R ecovery 1.2 The incidence of spills It may be possible, in some cases to remove oil emulsions from the beach and to recover some of When oil from spills at sea comes ashore on a the hydrocarbons present, following removal of the sandy beach, it poses a number of environmental water component. This is a preferred option, both and practical problems. Considerable publicity economically and environmentally, though the accompanies accidents involving oil tankers at sea appropriate, necessary, facilities may not exist in or iri coastal waters and the consequent release of, the vicinity of a spill site. occasionally, large quantities of crude oil or refined products. Although such incidents may result in the Stabilisation beaching of large volumes of oil emulsion at one Material recovered may be used in civil time, there is also a more continuous problem of engineering works, usually followi ng stabilisation small quantities of weathered oil coming ashore on with a binding agent or quickl ime. The stabil ised many beaches in the vicinity of busy shipping oils may then be used as part of land reclamation lanes. The products of either type of spill will have may be incorporated in tarmac or be used in consequences along the aff ected shoreline, other construction materials. A prime example of involving wildlife, economic and recreational this disposal route was the use of oiled sand, activity. They also pose disposal problems. recovered following the Torrey Canyon spill, in the reconstruction of Brest harbour. Again, the possibil ity of disposal by this means cannot be 1.3 The problems of disposal assured.

The arrival of the products of a spill on the shore Storage inevitably threatens wildli fe conservation on Final disposal of oily waste has frequently been to aff ected coastal sites; though vulnerability will landfi ll sites. Such disposal has involved either depend on shore type (Gundlach & Hayes, 1978; pre-treatment with a stabilising agent or no Gray, 1985). In some instances clean-up may be treatment (especially where oil residues were well- undesirable because, for example, attempts to weathered and contained few low-molecular weight remove oil from salt marshes may be more compounds). Such disposal may have taken place damaging in the long term than allowing the oil to into sites also receiving domestic refuse (as in the remain and degrade naturally. On high energy, case of beached oil from the Eleni V in 1978) or rocky coastlines removal may not be practical and, specially licensed sites. Such sites would normally in any case, wave action may assist the process of -be sealed to reduce the risks of groundwater self-cleaning of rocks. In contrast, oil washed contamination. The major problem with disposal to landfi ll is the development of anaerobic conditions, Proponents of bioremediation suggest the addition which lead to partial decomposition of oil residues of nutrients (principally nitrogen, phosphorus and and the production of a range of intermediate potassium) with or without the addition of cultures compounds, whose environmental effects are of hydrocarbon degrading micro-organisms. poorly known. Where there is potential for Considerable publicity was given to this type of contamination of water resources, the practice bioremediation during clean-up operations becomes increasingly unacceptable. This following the Exxon Valdez spil l (Bragg et al. unacceptability has been emphasised by recent 1994). However, the success of these procedures policy decisions at the British and European level, has been questioned following an initial wave of backed by the introduction of such measures as enthusiasm, and the need for critical examination strict site licensing regulations and a landfi ll tax. of the techniques and their effi cacy has been advocated by a number of authors, including Degradation Mearns (1997). Destruction of oil residues may be achieved by high temperature incineration but effi cient burning Two problems arise in the use of bioremediation requires the provision of appropriate speciali st agents. First, the need for cultured organisms is facilities nearby. On-site burning would be less highly questionable when native microbial fl oras effi cient and create additional problems of air with hydrocarbon—degrading capacity already exist pollution. and, second, the wider environmental effects of introducing nitrates and phosphates into In recent years there has been increased interest in environments which are naturally nutrient poor. bioremediation, predominantly in situ, on shorelines (Venosa et al. 1996; Oudot et al. 1998). Bioremediation involves enhancement of natural rates of hydrocarbon degradation by the addition of micro-organisms, the addition of nutrients, or promoting the transport of oxygen to reaction sites. The oldest method of encouraging natural biodegradation is by ploughing a thin, superfi cial layer of oil y waste into underlying soil . This process (known as landfarming) and means of maximising its effi ciency, have been described by a number of authors including; Knowlton & Rucker (1979), Amaral ( 1987) and Bleckmann et al. (1997).

The two maj or problems associated with landfarming are:

• the large area of land required to accommodate material in a thin layer • the potential for accumulation of heavy metals where a landfarm site is used repeatedly.

M ixing hydrocarbons with some other forms of organic matter is thought to aid the process of degradation and this is the rationale behind landfi ll operations where oily waste is incorporated with domestic refuse. An alternative is to mix the oil with natural sorbents such as straw, to promote so- called " composting." Surface area (and hence access to oxygen) is increased and some nutrient addition may be achieved, depending on the nature of the organic material used as the composting agent. NAT URAL BIOREM EDIATI ON

2.1 Hy drocarbon degradation p ho to-o xid a tion volatiles — bp-a tmos phe re

Hydrocarbons, naturall) occum ng co mpound s containing on ly carbon and in Mogen. are ubiquitous in the environment. Low -molecular w e a the re d o il re s id ues weight hydroc arbons are Eases hut those of higher molecular weight occur as liquids or solids 4 - m ic ro-o rg a n is m s at normal tempe ratures. Bacteria and fungi are 4 - Oxyg e n know n to degrade hydrocarbo ns hut it is BIOLOGICAL 4 - nutrie nt s generally co ns idered that the former assume the DECOMPOSITION 4 - tempe rature maj or role. If protected from microbial degradation. 4 - Mo is ture mixtures of hydrocarbons can accumulate in vast 4 - PH quantities as oil deposits, which may he found • throughout the world. Oils are comple x m ixtures of microbial bioma s s n erme diate s + ca rbo n dioxide many type s of hydrocarbon and the individual /IT compounds are more , or less, susceptible to inco rp o ra tio n into leaching microbial degradation depe nding on the ir precise so il orga nic matter molecular structure. Generally. the saturated alkan es (straight cha in molec ules ) can be Figure 2. 1. Proc ess es, f acto rs influen ci ng th em. degraded most rapid ly but resista nce to microbial and f ates of p rod ucts res ulting •from deg rada tio n activity increases with increasing cha in length, of oi l res idues by micro-orga nisms. and as the ratio of carbon ato ms to hyd rogen atoms increases bec ause of the presence of unsaturated (double) bo nds and bra nched cha ins. Arom atic co mpound s (with ring structured molecules ) tend to be more resistant as cleavage adopted in the present study. As many of ring structures (essential for degradation) is hydrocarbons are hydrophobic, deg rada tion will less easily accom plished. Multip le ring st ructure only oc cur at the interface of o il and aq ueous compOunds and those containing nitroge n or ph ases. sulphur groups arc even more recalcitrant . Such c o m p o u nd s are, therefore, degraded more slowly. In a practica l context, decompositio n w ill de pend The biochem ical mechan isms for metabo lism of on the ava ilability o f oxygen and a su pply of hydroc arbons arc diverse. The degrada tio n o f essential nutrient elements lo r the growing di fferent com ponents of oil may he carried out by microbial populations. For a m icrobia l popu lation diffe rent spec ies or consortia of spec ies with in a to re spond to a large exce ss o f hyd rocarbon, microb ial population. mineral nutrien t requirements must be satisfied . Nitroge n and phosphorus in particular arc The rate at which hydrocarbons are degraded is thought to be lim iting, as oil residues have ve ry subject to several controlling factors. These are low N P conten ts. Concentrations o f th ese summarised in Figure 2. 1. The ge nerally acce pted elements may also be low in some env ironments . theory is th at oxygen is essential for the rapid The natural environment tends to maintain a degrada tio n of hydrocarbons . There is evidence ba lance between di fferent che mica l processes as of some loss o f hydrocarbons in anae robic reactio ns within the maj or hiogeoc hem ica l cycles en vironments and lab oratory trials have counteract one another. The stimulatio n of demonstrated anae robic degradation potential organic carbon decomposition can pe rturb th is using nitrate or su lphate instead of oxygen (e .g. equilibrium and result in major environme nta l Coates et al. 1996 ). Howe ver, it is still not clear if change. For exa mple, the prod uctio n o f ac idity by degradation is complete under anae robic exce ss ive aerobic decomposition may saturate the cond itio ns and the possib le accumulation of na tural buffering capac ity of the medium in w hic h stable, and po tentially toxic, end products should micro-Organisms are grow ing. Microbial activity not be discou nted. may then decrease in response to falling pH . Hyd rocarbon decom positiOn 1n ay. there fore, More complete degradation is to he expected be come self-limiting as a resu lt of th is negative under aerobic conditions and that is the pre mise feed back mec ha nism.

7 2 .2 The distribution of hydrocarbon ran aground on Vancouver Island in 1970. Even degraders in the environment where conditions were found to be unfavourable, e.g. in the Strait of Magellan, there was slow Hydrocarbons, and micro-organisms capable of degradation of oil spilled from the Metula in 1974 degrading them, occur in marine, freshwater and (though this was considered to be the result of terrestrial ecosystems. nutrient limitation rather than low temperature). More recently, even without the extensive use of There has always bee n a discrepancy between bioremediation techniques, there was marked bacterial cell numbers obtained from total counts decomposition of oil spilled from the Ex xon Valdez and those found through the use of culturable in Prince William Sound and of the vast amounts counting methods. Under natural conditions the of oil beached in the Persian Gulf during the Gulf culturable count is only 0.0 1% to 12.5% of the War of 1992. Microbial decomposition thus total count which means that the vast majority of operates over a wide range of climatic conditions. bacteria in the environment cannot be grown under laboratory conditions. This has severely restricted In late 1992, during the Feasibility Study phase of the advance of microbial ecology and only recently this project, a visit was made to the MoD site at has progress been made. Despite this limitation Pendine. Beached oil emulsion from the accident there is good evidence that the distribution of in 1978 involving the Christos Bitas had been hydrocarbon degrading bacteria correlates well buried at this site. No accurate estimates of the with sources of hydrocarbon in ecosystems (Leahy original hydrocarbon content of this material had & Colwell, 1990; Benka-Coker & Ekundayo, been made, though improbable concentrations as 1997)). The proportion of the population able to high as 20% have been quoted. Core samples degrade hydrocarbons appears to be quantitatively revealed a concentration of 0.45% of extractable related to the degree or extent of exposure of that hydrocarbons remaining in the deposit. Even ecosystem to hydrocarbon contaminants. A number allowing for a low (e.g. 5%) starting concentration, of stud ies have shown that areas subj ect to recent this represented a loss of over 90% of the oil or frequent spills have larger populations of residues in 14 years. Core samples showed that hydrocarbon-degraders than similar "clean" sites bacteria capable of mediating the breakdown of (Atlas 1984; Barkay & Pritchard 1988). hydrocarbons were present in the deposit in signifi cantly higher numbers than in the adjacent Bacterial populations are not static and can become sand dunes. adapted to degrade specifi c hydrocarbons when the supply of those hydrocarbons increases. Native There is, thus, evidence that microbial activity can microbial populations have the capacity to act to decompose a large proportion of the metabolise a range of compounds and generally compounds present in weathered oils reaching the contain the appropriate genotypes enabling them to shore. break down most hydrocarbons. In many cases this potential may not be organised into defi nitive enzyme systems but exchange of genetic material 2.4 The starting material within the population, and the induction of appropriate enzyme systems, will allow Some microbial communities are able to survive in, degradation pathways to evolve rapidly. Because of and mediate the breakdown of, crude oil or refi ned this adaptability, signifi cant populations of petroleum products. However, unless a spill occurs hydrocarbon-degrading micro-organisms show a following grounding of a vessel directly on a high degree of variability in both time and space. beach, any oil reaching shore will have undergone physical and chemical transformation at sea. Figure 2.2 shows the major processes involved and 2.3 Biodegradation following spills the time scales over which they operate. Since the more volatile and more "soluble" components will There is evidence from a number of oil spill have been dispersed to the atmosphere and the sea, inc idents that hydrocarbons are degraded as a respectively, they will be largely absent from any result of microbial activity. Changes were reported oil that is beached. The remaining residues can be to have begun within days of beaching, in the case solid (tar balls), liquid or, most commonly, an of the Amoco Cadiz spill in 1978. Almost all the n- emulsion produced by mixing of oil and seawater. alkanes were found to have disappeared within 4 These will lack the most reactive (and hence the years from oil spilled by the Irish Stardust, which most biologically active) components which may HOUR DAY WEEK MONTH YEAR

EVAPORATION

DISPERSION -, BIODEGRADATION

DISSOLUTION -, PHOTO-OXIDATION *

_ 1 0 EMULSIFICATION * - - I - SEDIMENTATION SPREADING *

Figure 2.2. Weathering processes influencing od spill composition and ti me scale over which they operate (Courtesy of ITOPF).

have been present in the original oil. Disposal will therefore involve what might be best referred to as "oil residues" rather than oil. This term will be retained throughout this report. The lon.2er a slick is at sea, the more it will be weathered and the greater will be the change in its nature (both chemical and physical). In addition, any microbial populations present in the weathered oil will begin to degrade the hydrocarbons present. Because of wave action, some of the oil residue on the shore will become mixed with the materials of which the beach is composed. Subsequent removal of the oil residues will also involve removal of part of the surface layer of the beach (large ly sand, but with possible addition of seaweed, shells or pebbles). The resulting mixture is referred to as Oiled Beach Sand (OBS).

9 10 3. EXPERIMENTAL APPROACH

3.1 General Philosophy also examined. The inclusion of appropriate control samples at all experimental scales allowed critical The project specification called for examination of examination of these data. what was considered to be a simple, cheap and environmentally acceptable disposal method. The In order to assess the responses of plants to oil need to adopt this strategy dictated that residues and gain insight into the possible toxic technological input at the field scale was minimal effects of disposal, pot experiments were set up at and that the design of experiments should be such ITE Banchory and performance was recorded. that they would resolve practical questions. A series of experiments used a purpose-built set of That nutrient addition is capable of enhancing lysimeters installed in a trench at Merlewood to degradation has been shown by a number of studies monitor the processes of oil residue decomposition (Atlas 1984; Swannell et al. 1993). Other workers under controlled conditions. LysiMeter have examined the possibilities of selecting experiments offered the capacity to study the particular microbial organisms (Jason et aL 1995) decomposition of buried OBS under natural or of modifying them through molecular genetic climatic conditions in close proximity to laboratory techniques (Padmanabhan et al. 1998). However it facilities. This allowed immediate analysis of has been suggested (Chapelle, 1999) that leachates collected from the bases of the lysimeters introduced micro-organisms might not compete following precipitation events, and the possibility successfully with the, indigenous, populations. The of monitoring (e.g. gas sampling) at more regular objective of this project has not been to design new intervals than was feasible at the remote field bioremediation techniques, so addition of fertilisers locations. From these studies, it was possible to or cultures has played no part in the work derive a much greater understanding of the • programme. processes and mechanisms controlling the oil degradation. In the past there has been a lack of integration between field observation and laboratory-based Small quantities of OBS buried in the ground at a experimentation under highly artificial, and often number of locations were used to relate unrealistic, conditions. In an attempt to remedy this geographical, and attendant climatic and edaphic, mismatch, this project was designed to examine the conditions to oil residue breakdown. These processes involved in hydrocarbon degradation in a experiments again allowed control of starting co-ordinated manner. The approach adopted was conditions and replication, but no control over one in which systematic studies were integrated external variables. A collateral trial at Eskmeals, over a range of experimental scales. using the same procedures, eliminated climate Because of this, it was considered possible to have variables whilst retaining variation attributable to some confidence in any upscaling from sand type . experimental to field trials and to extend this to practical aspects of real disposal operations. Larger field plots were established at the MoD (DERA) site at Eskmeals, Cumbria, and used to compare breakdown rates in landfarming and 3.2 Experimental scales burial plots. Control of initial conditions and replication at a larger scale was still possible in Experiments have been conducted at a range of these experiments. scales, each designed to provide the appropriate level of control over factors thought to influence Two operational scale experiments, one the the process of oil residue degradation. Each scale product of a real spill and one a simulation of a was capable of being linked readily to fi ndings spill, were set up at the MoD (DERA) site at from the next scale up or down, as appropriate. Pendine, Carmarthenshire . Within these experimental sites, data were collected to predict Microbiological studies looked at the stimulation the rates of breakdown from different starting OBS of respiration when microbial communities were type and concentration, and to detect any challenged with hydrocarbons. In addition, the movement of hydrocarbons towards the water ability of cultured organisms to degrade aromatic table. A re-vegetation trial was also established on molecules (a key process in oil degradation) was the real spill site (first Pendine experiment). At this site there was less control of starting conditions in the real spill experiment than in the simulated one and no control oVer environmental variables.

A number of common analytical techniques were appl ied across all scales of exp eriment (see sections 3.9-3.11).

3.3 Toxicit y tr ia ls

Trials set up to assess the toxicity of weathered oil residues used the grass Festuca n ebra (red fesc ue - a species found as a ma jor component of dune pastures) collected from a number of locations around the British coast. For each species, five replicate pots were set up for each of five different OBS types, prepared from each of three different oils (Forties cr ude, Kuwait crude and medium fuel oi0 . The oils were artificially weathered for different lengths of time before being mixed with beach sand to provide growing med ia containing 5% oil residues. Young plants of comparable size were planted into the dra inage bottle s OBS and the proportion of healthy foliage present was assessed at monthly inter vals for six Figure 3.1. Cross section qf a lysimetec showing months . math design Jeatures. A second trial, using only a single oil (Forties) given a single weathering treatment, compared Climatic variables were monitored constantly by an the effect of OB S on the performance of eleven adjacent meteorological station with an integral vascu lar plants and one moss growing in dunes data logger. Hourly values for rainfall and air and dune pastures. Six replicates were used and temperature were recorded together with sand the monitoring methods were the same as in the temperature at four depths (8, 22 (OBS layer), 38 fi rst trial. and 98 cm) in selected lysimeters. Suction samplers were inserted immediately below the OBS layer to The effect of different concentrations of oil collect rainfall as it percolated through, and tipping residues in OBS on germination success was also bucket devices at the base of the lysimeter tested using Festuca rubra and comparisons measured and collected through-flow. Leachates were made between three grasses and three herb from the base of each lysimeter tube were sampled species at a single oil residue concentration . at intervals and analysed using an oil -in-water meter, or forwarded to f i b M onks Wood for PAH analysis. Between collection and ana lysis these 3.4 Lysim eter ex per im ents samples were maintained at 4°C.

Overa ll design F irst experim ent Large lysimeters (cylindrical PVC tubes I 50Cm x The firSt lysimeter experiment was set up in June 30cm internal diameter) were filled from the base 1994 to examine differential breakdown of three oils with gravel (30cm), dune pasture sand (80cm), under three weathering trea tments. Forties crude, 12.5 kg OBS (20cm) or beach sand as a control Kuwait crude and medium fuel oil were given three sample, and f5cm dune pasture topsoil (over- different treatments; untreated; washed in three burden). This left a headspace of 5cm between changes of seawater in a cement mixer ; washed in the surface of the sand and the top of the seawater and then exposed on an artificial beach lysimeter tube. The water table was set at 100cm for 10 days. A 5% OBS was prepared from each of below the sand surface, and designed to allow a the differently weathered oils. The experiment was small amount of fl uctuation. set up as three blocks with 3 oils x 3 weath ering

12 treatments (= 9 combinations) + 1 control in each OBS Preparation block. Overall respiration (as CO, flux from the soil The first stage in OBS preparation was to samples) was measured and leachates were tested make an artificial beach on which the oil was to be for hydrocarbons. Microbial populations and their weathered . A 40m x 20rn area was excavated and activity were assessed when the experiment was lined with a 1.5mm thick polyethylene membrane terminated. overlain by a geotextile fleece. A drainage sump and the necessary pipework were installed, then a Secon d exp erim ent gravel filter (300m m deep) and a 300mm layer of The second lysimeter experiment was designed to coarse sand were placed on top of the fleece. determine how oil residue concentration might Finally, a layer of quarry sand (used for OBS influence decomposition rate and potential for the preparation) was spread over the surface and leaching of hydrocarbons. Only partially sprayed with sea water at a rate of 5 1 m-2. weathered Kuwait crude was used in this experiment, with six nominal concentr ations of OBS was prepared on two occasions; January 1995 OBS being prepared (0%, 2.5%, 5%, 7.5%, 8.5%, for winter burial trials and July 1995 for summer 10%). Again a randomised three block design was burial trials. On each occasion the oil emulsion was used, giving three replicates of each treatment. spread eve nly over the surface of the artificial Carbon dioxide flux was measured and leachate beach using rubber spreaders and allowed to waters were collected for analysis. Samples were remain for two days before being mixed into the also collected at the end of the experiment for surface layer of the artificial beach with a rotovator microbiological examination. to give an oil residue concentration of 5% . This OBS was used for the burial trials. Subsequently, a Th ird exp eriment 10% mix was prepared for incorporation into The third lysimeter experiment was closely linked landfarming trials. into the extensive trial (see 3.6, below) examining differences in site characteristics. Following The exp erimental site preliminary examination of the nature and The pasture burial and landfarming trials were microbial activity of sands from the fifteen established using a randomised block design within experimental sites, five (Askernish, Cresswell, a stock-proof exclosure in an area that had Largo, Pendine, Tain) were selected for inclusion previously been grazed by sheep and cattle. Dune in a lysimeter experiment where they could be burial plots were contained within a separate compared under the same climatic conditions. enclosed area in nearby, disturbed, mature dunes.

Washed and lightly weathered Forties oil was In the summer of 1994, before the setting up of the used to prepare 12.5 kg of OBS with 2-3% plots, soil, vegetation and invertebrate surveys concentrations of oil residues. The experiment were conducted to provide a baseline against which was set up in three randomised blocks, each block subsequent changes could be measured . containing factorial combinations of five beach sands with two treatments (with OBS and without Dun e burial plots OBS). This design allowed comparison of Dune burial plots were set up in winter only, with background activity within beach sands and a no equivalent burial in summer. Four replicated measurement of the stimulation provided by the blocks were established. Within each block, four addition of hydrocarbons to each beach sand holes (each 1m square and 0.7m deep) were type. excavated after removal of the surface turf. Each hole was filled to within 0.2m of the surface with OBS, and backfilled with material previously 3.5 Eskmeals field trials excavated. Turf was replaced over the surface of all plots. In April 1995, transplants of Ammophila Three types of trial were conducted at Eskmeals; arenaria (marram), Hippophae rhamnoides (sea dune burial, dune pasture burial and landfarming. buckthorn) or Hypochoeris radicata (cat's-ear) All used the same starting material, an artificial were made into three plots in each block: the fourth OBS prepared from local quarry sand and a remained as an untreated control. topped Russian oil supplied as an emulsion containing 38% sea water. Quarry sand was used La ndf arming plots because it was not possible to obtain a permit to The landfarming plots were set up on 3 1" January remove sand from nearby beaches. and Th June 1995. A randomised block design was

O 2-weekly plough E l Munro recorder 0 4-weekly plough 0 weather station • no plough

• winter w nter control summer control summer • • • co 0 M • 0 IN • summer control winter control 0 summer winter • • • • 0 • MI • MI Su m m e r summer control winter winter control

Figure 3.2. Layout of landf arndng plots in f ield trial at Eskeneals used with six repl icate sub-plots of each main plots, using a clean rotovator. For winter plots, treatment (w inter and summer land farmed, and ploughing ceased in January 1996 and for the winter and summer control) in each main plot of summer plots on 24th A pril 1996. Revegelation each block. treatments (application of rye-grass seed mixture and control wit h no seeding) were appl ied randomly to Subp lots were paired and each pair was assigned to pairs of plots. one of three ploughi ng treatments: Dun e pasture burial plots • no ploughing; Within the main enclosure, dune pasture burial pl ots • pl oughed at two-weekly intervals; were set up on 24 February and 7th June 1995 in an • ploughed at four-weekly intervals. area adj acent to that used for the landfarming plots. For winter burial plots, two areas (each 7In x 5m) Stakes were used to mark the corners of each 2m x within each block W e r e excavated to a depth of 0.7m. 2m subplot. For the controls, quarry sand waS Six heavy duty plywood frames O m x l m x 0.70 appl ied as a 75mm layer and ploughed into the were put in place within the excavated area to delimit surface to a total depth of 150mm, using a rotovator. the individual sub-plots. Th e area between the For experimental plots, a 75mm layer of OBS was frames was backf il led with the native sand excavated appl ied after all control plots had been established fr om the site. and ploughed in to a depth of 150mm. The schedule of pl oughi ng began on 13th M arch and on each Quarry sand was placed in the control plot frames occasion control plots were ploughed before OBS and prepared OBS in the burial frames, both to

14 ithi n 0.2nt of the sult ace " giving an OBS depth of 3 .6 E xte n sive stu dy 0.5m Tne frames were then remo ed and the plots V1/4ere co% ered h 0.2m ofn at i ve topsoil. The same Followi ng a desk stud) , supported by a procedure w as follow ed or the summer burial subsequent. selective, field survey, 15 sites were plots. Revegetation trials included transplants chosen for burial of small quantities of OBS. as a ryc-grass till ers. in addi tion to the seeding and means of testing variations I n decomposition control treatments used in landfarming plots. attributable to local site conditions. T he choice of sites (Figure 3.3.) was not intended to imply anx Environmental monitoring designation as potential locations for future Changes in water table height, were measured using disposal. On two occasions (autumn and spring ). M unro recorders installed adj acent to the dune at each selected site. OBS was prepared using burial plots, and close to the upper and lower topped Russian oil emulsion (the same as that blocks of the landfarming trial. A Delta-T used for the Eskmeals trials) and local beach automatic weather station was installed close to the sand. A 1:10 w/w mi xture was calculated to give a centre of the dune pasture site to col lect air fi nal oil residue concentration of 6%. The temperature, soil temperature at diff erent depths, prepared OBS was weighed into nylon mesh bags and rainfall data. Comparisons of temperature that were sealed and tagged with wire or string to between plots were made using TinyTalk data aid recovery, and buried 30cm deep in prepared loggers installed in a number of plots across the pits. Three additional bags were taken for site. Piezometer tubes. with bottom end-caps and analysis of initial hydrocarbon concentration. 0.3mm slotted tubing at their lower ends, were installed in and around all plots in both The trials were set up as three replicate blocks at landfarming and burial trials to allow sampling of each site. Within each block six bags were buried groundwater. in two parallel rows in M arch 1996. A t one end of the row a slotted piezometer tube was installed to Carbon dioxide emission was measured using static a depth of at least 2 metres. A djacent to thi s a gas chambers put in place shortl y before a set of bag of sand, wi thout addition of emulsion was measurements was taken. Each chamber consisted buried. Close to the piezometer tube in the central of a 200mm length a I70mm diameter rigid block a pair of TinyTalk automatic temperature plastic pipe, threaded at the top. which was sunk data loggers were buried. one 10cm and one 30cm into the ground, leaving a headspace of 120min deep. A further set of four bags (two per row) between the soil surface and the top oF the tube. A n was buried in September 1996. airt ight screw cap fitted over the top of the pipe. A fter a timed period, bungs fitted into two Hmin diameter holes in the cap were removed and the 4!Spiggie inlet tube to a portable infra-red gas analyser was inserted in one of them. •setr W Bay of Skelt Oldshoremore A peak reading for carbon dioxide concentration was noted, compared with background Tain concentration, and used to calculate the emission Askernishf Camusdarach rate over the ti med period.

Vegetation change, over a two year period, was Cresswell monitored by recording the species present in each Torrs Warren plot and estimating the percentage cover for the most common species found. SelAttetO ztt orsey Pitfall traps placed in and around all plots were used to collect samples of ground-dwelling invertebrates. Their covers were removed at the Pendine beginning of each sampling round and replaced afterwards. Samples were then returned to the Crimping laboratory and examined in order to determine the number of di fferent taxa (at the species or genus level) present in each sample. Figure 3.3. Locations qf extensive study sites

15 At intervals replicate bags from each block were programme was begin involving, initially, the removed and returned for analysis of collection of core samples in a stratified manner hydrocarbon content. Tiny Talk loggers were from random locations within selected squares of removed for downloading of data and new ones a 5in grid set up on the site. The core samples, were buried in the same positions. Water table taken with a 5cm bipartite Ej kelkamp corer, were depth was measured and vegetation was recorded analysed for hydrocarbon content and microbial (species presence and cover value), At the end of activity. At intervals, measurements were made of the experiment the " blank" bags of beach sand CO, fl ux at random points across the site. were removed with the final samples. On 2" April 1996, a collateral trial was established at Eskmeals. Ground water m on itoring Using sand collected from the dispersed set of Because of the problems of ensuring absence of field sites, five replicate bags of OBS, comparable munitions in the vicinity of the site, the to that in the field sites, were prepared. These installation of piezometer tubes to allow were buried i n five blocks, one replicate per block. measurement of water table depth and the collection of groundwater samples for Locations of individual OBS types were hydrocarbon analysis was delayed for six randomised within each block. Adj acent to each months. However, a set of nine holes was dril led OBS bag, a bag of beach sand from the same site in July 1994, using a hydraulically-operated was buried. These plugs of OBS and beach sand percussion drill . were removed at the same time as the last set of samples from the field plots. Lengths of 30mm diameter piezometer tubing (sufficiently long to give at least l m penetration of the ground water table) were installed. The 3.7 Fir st Pendine deposit bottoms of these tubes were capped and the lowest sections had 0.3mm horizontal slots: upper I n itial establishm ent sections were unslotted. Between Christmas 1993 and new Year 1994 a quantity of well-weathered oil was washed up as R evegetation trial tar balls of varying size on the beaches of Because of the problems of wind erosion carry ing Pendine and Laugharne Sands. Following sand from the surface of the deposit, it was agreement between MOD, EA and CCW, the oil decided that a surface vegetation cover should residues were removed, together with the surface be established. Although some marginal layer of beach sand, by Carmarthen District invasion of plants from the adj acent dunes had Council and buried in a dune holl aw adj acent to been observed, this was not rapid enough the the old Christos Bitas deposit. A monitoring stabifi se the sand surface. The opportunity was taken to combine the need for surface stahilisation with a trial of the effectiveness of different vegetation types in achieving this Beach sand stabilisation. In November 1994 the trial was set Piezometer plug up using a series of replicated plots each 4m square, set up within the cells of the sampling Tiny Talk recorder grid establi shed earlier. The following treatments were applied: 1) control 2) surface raki ng Spring plugs 3) surface raking and incorporation of 70g m= of an algi nate soil stabil iser - 4) seeding with a Festuca rubra mix after addition of 70 g nr2 of soil stabiliser 5) seeding with a Festuca rubra mix after addition of 135 g in-' of soil stabiliser Autumn plugs 6) seedi ng with a Festuca j uncifolia mix after addition of 7017 ITI 2 of soil stabil iser 7) planting of marram tillers at 0.5m intervals Figure 3.4. Plan af individual plot ayout tt,ithin 8) planting of Marram tillers at 0.5m intervals an extensive study site. after addition of 70 g m-2 of soil stabiliser

16 A series of samples of groundwater, sand and seawater was taken at intervals between installation A dune hollow of the piezometer tubes and incorporation of the OBS some 13 months later. DE RIEDIEDD Exp eriment establish men t and early dune ridge • 1111111 • dune ridg e 0 0 111DMOM sampling 111• 0 1111DINED Work began on setting up the experiment in March Chrfstos Bfitas 1997. Sand was collected from the mid shore part of deposit the nearby beach as a skim of up to 10cm thickness over an area approximately 100 m x 50 m. An area 0 control • Fes tuca rubra + stabiliser • raked 0 Fes tuca rubra + 2x stabiliser 10m x Sm was levelled and a 300mm high bund was o + sta bilise r • Fes tuca juncifolia + stabilise r built up around its margin. The enclosure produced 0 ma rram • ma rrarn + stabiliser was lined with damp-course quality Nesqueen polythene. the individual sheets of which were Figure 3.5. Layout of experimental p lot and double tucked together to provide a seal capable of revegetation trial in the f irst Pendine experiment. preventing contamination of underlying sand . At the edges of the area, the polythene was raised and Plant species present within the plots were incorporated into the blind to provide lateral recorded , together with measurements of their containment of any oil spilled. The polythe ne lining cover and vegetation height, at yearly intervals. At forming the floor of the enclosure was then covered the same time, measurements of sand accumulation with a 40mm layer of sand. This area was designed were made by recording height differences at I m to receive 12 skips for temporary storage of oil and intervals along a horizontal line stretched between preparation of emulsion. diagonally opposite Marker posts of each grid square. Once the emulsion preparation area had been completed, an area 49m x 33m was similarly levelled, bunded and lined with Visqueen and covered with a 3.8 Second Pendine deposit 200mm layer of beach sand. This formed the artificial beach. The actual width of the mixing area on the Th e site artificial beach was set as 45rn x 29m, giving a total In early 1996, two adjacent hollows were selected area of 1305m2 and leaving a 2m buffer zone around within an area of mature dunes that had been the margin to further reduce the risk of subject to disturbance in the past. The smaller, contamination of the surround ing area. landward, hollow was designated as the control, to be filled with beach sand, and the larger, deeper. 23.000 litres of medium fuel oil, delivered by tanker hollow (the experimental hollow) on the seaward from BP Llandarcy, was maintained in lined skips side o f a low ridge was to contain prepared OBS. within the bunded containment area.

Pre-establishm ent procedures and baseline Emulsion was produced in 2000 litre batches in a sample collection skip dedicated to this function. Equal volumes of oil In February 1996 a series of 12 piezometer tubes and seawater (collected and delivered using a was installed in the hollows and the surrounding bowser) were pumped into the mixing skip and dunes using a percussion drill and tubing similar to combined using a submersible "IlK150 hydraulic that installed at the site of the first Pendine pump. After initial trials, a standard procedure was experiment. adopted which produced a stable, thick, emulsion.

The site was levelled to give relative heights of the Turves were removed from the control hollow, a tube tops and the ground surface. The depths to layer of beach sand (l ni thick) was placed in the water table were measured and eontour maps were hollow and the turves were re placed . Two Them constructed showing the relationship between probes were buried at approximately 0.2m above the surface contours and ground water levels. base of the beach sand layer to allow the moisture status of the sand to be measured. The consulting engineers (Sir William Idalcrow & Partners) produced a contour map of the site, tied Following completion of the control hollow, the into a benchmark in Pendine village. experimental hollow was prepared by removal of

17 turf. A layer of beach sand was placed in the I .3m . A set of samples (four of OBS and four of the hollow and levelled off (0.4m thick over the deepest underlying buffer layer ) was collected from the part of the ho llow). Two Theta probes were buried artificia l beach and sent to the Environment Agency mid-way down this beach sand layer. A thermistor laboratory at Llanelli for analysis. probe, for tempera ture measurement, was laid on the surface close to the middle of the holl6w. The Once placing of the OBM had been com pleted, the connecting wires from all probes were laid back to artificial beach was removed. Sand which had the ridge between the two hollows, which had been formed the buffer layer was transported lo the designated to receive the data logger and rainfall experimental hollow and placed over the OBM to gauge. form what we have termed overburden. This overburden was 0.6m deep over the central part of Emulsion was pumped from the skips into the the hollow, but reduced in thickness towards the bucket of a dumper truck and taken to the artificial margins. beach, where it was dribbled over the lip of the bucket as the dumper made a pass. A single pass of Delta-T automatic data logging equipment and a the dumper across the artific ial beach delivered the solar cell to provide its main power source (plus appropriate quantity of em ulsion to d ye an even backup batteries) were set up on the ridge betwee n spread over the sand surface . This method of the two hollows . Previously installed instruments application prevented contamination of the dumper were connected up. Remaining temperature probes wheels and ensured containment of the emulsion. were installed within the hollows; water table measuring probes were inserted within two Mixing of the emulsion into the upper layer of the piezometer tubes (one in the centre Of the artificial beac h was card ed out using a tractor- experimental hollow and one toward the margin of mounted rotova tor. The OBS was considered to be the control hollow). A tipping bucket rain gauge thoroughly mixed after two complete rounds by the was also set up on the ridge close to the control tractor. After 48 hours of weathering on the hollow. artificial beach, the OBS was transferred to the experimental hollow and spread using a long-arm At regular intervals, samples of groundwater were excavator. Within the deepest part of the hollow taken from the piezometer tubes and core samples the depth of the slightly domed OBM deposit was were also taken from the control and experimental

data logge r E P T

T P 10m

CONTR OL HOLLOW EXPERIMENTAL HOLLOW

ove rburde n be ach s and • : ea ch sa n dune s a nd

wa te r ta ble s ea — 1•

Figure 3.6 Layout of second Peru e experiment

18 hollows. Overburden was sampled using a 7cm The ability of the bacterial community to mediate diameter bipartite corer; the OBS using a 5cm this key reaction was followed using laboratory diameter bipartite corer and the underlying layer of culture techniques. Using ultra-sound, bacteria beach sand (underburden) using a 2.5cm diameter were dislodged from the sand particles after gouge auger. Samples were sent to Llanelli, Monks suspending a known weight of sample in water. Wood, Windermere or Merlewood for appropriate The supernantant was diluted and small volumes analysis. spread onto the surface of nutrient agar plates. The culturable bacteria were allowed to grow for 7 days at 20°C after which the colonies were counted. The 3.9 Microbiological techniques plate was then fl ooded with a I% (w/v) solution of catechol and colonies capable of degrading the The experimental designs used to compare aromatic ring structure turned yellow. These were treatments and sites required a large nuMber then counted separately. It is accepted that there samples. It was important, therefore, that may be other metabolic pathways resulting in the microbiological methods were simple and yet cleavage of the carbon ring structure, and that the provided sufficient information on oil residue culturable count is likely to be only a small decomposition processes and potential within proportion of the total bacterial population active experiments and trials sampled. within the sample. However, despite these The sand dune environment (especially in its less limitations, it was considered that the catechol mature phases) is characterised by low screening method provided useful information on concentrations of organic matter and inorganic the adaptation of microbial populations to nutrients. The addition of oil residues to the sand hydrocarbon decomposition. represented a large input of organic carbon. Measurement of the metabolic activity of the microbial populations provided a representative 3.10 Decomposition activity estimate of their response to the challenge of hydrocarbons. Sand dune soils are well drained so Decomposition activity is defi ned as non- that, given a low water table, only aerobic destructive measurement of the components of the respiration needs to be considered. Replicate degradation process. As in the case of samples of sand or OBS were incubated in sealed measurements of microbial activity, this consisted glass vials and the accumulation of carbon dioxide predominantly of measurements of carbon dioxide in the headspace gases was used as an estimate of fl uxes. respiratory activity. The inclusion of control plots in all experimental designs allowed comparison of In the fi rst lysimeter experiment, we used a volatile activity between oiled and non-oiled samples. organic compound (VOC) monitor to detect the Some respiratory activity should be expected in extent of direct evaporation of hydrocarbons from the control sands as they contain small quantities OBS. Only trace quantities of VOC's were of naturally decomposing organic matter from detectable at the beginning of the experiments, other sources (e.g. plant and animal remains). during periods of hot weather, and thereafter none. Because of low and infrequent emissions, The overall activity of the microbial population measurements were not continued on a regular provided no indication of the range of basis. hydrocarbon compounds being degraded and a method was required, which would demonstrate if In contrast, it was possible to detect diff erences in more stable Compounds, particularly those with fl uxes of CO2 resulting from microbial respiration aromatic ring structures, were being metabolised. in the head-space above the surface of sand in Degradation of aromatic compounds involves capped lysimeters. Initially the head-space was modifi cation of side-groupsenthe carbon ring to sampled via a syringe and injected into an infrared aromatic acids then conversion to catechols. This gas analyser (IRGA). Later, a 24-channel, is followed by oxidative cleavage of the carbon automated, facility was developed for sampling ring. One pathway (the meta-pathway), for this li simeters sequentially and measuring CO2fl ux oxidative step, results in the formation of hydroxy- using an IRGA. muconic semi-aldehyde. This reaction can be observed as the colourless catechol is Protocols were devised to determine the CO2 transformed to the yellow serni-aldehyde concentrations in the 30 lysimeters containing the derivative. OBS and control treatments. Lysimeters were

19 capped with a gas-tight lid and allowed to sulphate. The extract was cleaned on a glass equilibrate for one hour prior to obtaining a single column containing 0.8g of alumina deactivated with value integrated from a 15 second reading at 10:00 5% water and reduced further under a stream of hrs. On each occasion, the instrument was nitrogen. Hexane extracts of OBS were fi ltered calibrated from prepared 0, 1000 and 2000 ppm(vIv) through anhydrous sodium sulphate. In both CO2-C mixtures in nitrogen, and fl ux concentrations cases, Dichlobenil was added as an internal derived from a linear regression. Net fl ux from the standard before extracts were transferred to lysimeter was computed following deduction of chromatography vials. Extracts were analysed ambient atmospheric concentration. using GC-M S. A recovery sample and a blank were included in each analysis and limits of detection A similar principle was used in the fi eld, where were defi ned as the concentration giving a signal static gas chambers with screw caps (see section twice that of the baseline noise. 3.5) were installed in bare sand overlying the experimental deposits. Portable IRGA monitors were OBS samples from the second (1997) Pendine used at the Pendine and Eskmeals fi eld trials, experiment were also sent to the Environment although the data was more di ffi cult to interpret due Agency laboratory at Llanelli, where they were to the greater variability in the plots, and the analysed for total hydrocarbon content and infl uence of other factors such as plant root abundance of individual components using respiration. fl uorescence spectrometry and GC-MS, with Forties and Ekofi sk oils as the calibration standards. 3.11 Assessment of degradation

The main method for estimating the oil residue 3.12 Assessment of leaching content of OBS used a gravimetric procedure to determine changes in the concentration of hexane- Direct evidence of leaching was obtained by extractable hydrocarbon. This procedure was analysis of water samples taken from below the applied to core samples taken from the Eskmeals fi eld trials and from the bases of the lysimeters. landfarming and burial plots, and the two Pendine Water collected from the lysimeters was initially fi eld trials. It was also used to determine the tested using an oil-in-water monitor but because of hydrocarbon content of OBS recovered from its lack of resolution, later samples were tested for lysimeter experiments and the extensive study the presence of hydrocarbons by smell. PAH (including its collateral trial). content was measured at Monks Wood using GC- MS. A I Og sample of pre-dried OBS (dried for 3 hours at I05°C) was refl uxed with 25ml hexane for 30 Fluorescence spectrometry was also used by the minutes. The mixture was fi ltered and the filter EA Llanelli laboratory to estimate the total residue washed with 3 x 25ml volumes of hot concentration of hydrocarbons present in hexane. The fi ltrate and washings were combined groundwater samples taken at the site of the second in a weighed evaporating basin, and the extracted Pendine experiment. The procedure used was that hydrocarbon determined gravimetrically after developed for analysis of oil residue samples evaporation of the hexane. It is possible that some recovered followi ng theSea Empressincident. volatile, low molecular weight, hydrocarbons were Samples of groundwater from the second Pendine lost during sample preparation for hexane trial were analysed for chloride concentration as a extraction. Moreover, asphaltenes and resins are means of tracing the movement of water displaced known to precipitate when oil is dissolved in from the OBS and beach sand. That seawater hexane. Therefore, it was expected that the would also be expected to act as a carrier for measured hydrocarbon content of the sands would "soluble" hydrocarbon fractions from the artificial be lower than the calculated value. OBS. The British Geological Survey, Wallingford, analysed groundwater samples collected from the A nalysis of oil residues for PAH content in selected second Pendine experiment using Inductively samples was undertaken at ITE Monks Wood. Coupled Plasma Optical Emission Spectrometry Hydrocarboff residues were extracted from water (ICP-OES). These measured concentrations of samples using hexane, foll owed by reduction of the inorganic elements providing indications of hexane component at 80°C after separation from the changes in groundwater chemistry resulting from aqueous phase and drying with anhydrous sodium deposition of OBS.

20 4 . FA CTO RS AFFE CTING DE GRADATI ON

4.1 Patter n of degr ada tion As suspected at the time of the beach cleaning operatio n, an un necessarily high prop ortio n of sand All experiments, at all scales, showed a progressive was re moved and the low concentration of oil residue reduction of hydrocarbon content in OBS samples. remaining in the OBS d id not make th is a realistic test There was, with some exceptions in the case of the of the power of the disposal technique . The extensive trial, rapid development of populations of concentrations of oil residue used in other hydrocarbon-degrading micro-organisms, and an experiments approached more closely that expe cted increase in CO, flux in samples containing oil residues. from a normal clea n-up ope ration, whilst the use o f compared with that in controls. less heavily weathered oils produced a starting product more comparable with that from a fresh sp ill. The longest period of sampling was for the first Pendine experime nt, where six sample dates were The initially rap id decline in hydrocarbon content, spread over four and a half y •ars. From these results it followed by a slow ing o f that rate , ag rees with studies has been possible to construct a decay curve showing und ertaken by othe r workers . The fitting of a po wer the decline in concentration of hexane-extractable function to degradatio n curves, as in this proj ect, hydrocarbons present in the OBS . appears to be new. The power funct ion, has the oeneral form: From the position of the points on the graph it is clear that the relationship is curvi-linear rather than linear. Y = A . X' An exponential line cannot be fitted with confidence (R2= 0.23). However, a power curve better fitted the where Y = concentration of hydrocarbons remaining , data and gave an R2 value of 0.87, indicating a X = time since incorporation of the OBS , A = initial statistically significant result. hydrocarbon content of the OBS and n = the rate function . By extrapolation from this curve, the starting concentration of the OBS was estimated to be c. 0.4%. This implies that de composition will de pend on the Even given this low starting concentration, it doeS initial concentration, time and some rate lim iting mean that within less than 5 years some 97.5% of the factor. However, the initial concentration of oil original hexane extrac table hydrocarbon had been residues m ay help determine some of the factors dearaded. In terms of actual quantities, this is contributing to the rate limiting functio n (e .g. ease of equivalent to a loss of 78 tonnes from an original pene tratio n of oxygen or nutrients). The nature of the estimated 80 tonnes present in the OBS. The large loss residues (the ratio of easily decom posed compo unds of hydrocarbon is notable as the original material to those , which are more recalbitrant) may also he consisted of well weathered fuel oil and hence was expected to he lp de termine the rate limiting functio n. expected to contain a relatively low proportion of the Because of this , we may expect tha t the re will be a more easily degradable components. fam ily of curves defining the behaviour of different types of OBS under di fferent environmental conditions . The se differe nt cond itio ns and the 0.14 y = 0.231x-12584 R2 = 0.8747 influe nce they have on the decomposition proc esses a 0,12 have formed the basis of much of the project. •5 0.10

0.08 4.2 Effects of oil typ e and 0.06 con centr ation

.c The first two lysimeter experiments addressed the 2.2 0.02 question of the influence of oil type, including weatherina history (which influences composition 0.00 0 10 20 30 40 50 60 of the initial staring material) and oil concentration time from de position (months) on the degradation process. Results (Table 4.3) showed that overall there was a significant Figure 4.1. Reduction in hydrocath on content of difference between the starting and final OBS with time at the f irst Pendine experiment. A hydrocarbon concentrations (paired student 't' test power curve has been fi tted to the fi eld data. gave t=3.67: p=0.005). However, statistical analysis

2 1 Table 4.1. Starting concentrations of hydrocarbon of variance (A NOVA) also showed significant in OBS p repared f rom diff erent oils and subj ected differences between treatments (F=82.9: p<0.001). to diff erent treatments. With all three oil types the washing and weathering treatment produced a signifi cantly lower hydrocarbon content in the starting OBS than washing alone or no treatment. The amount of reduction depended on oil type. Even without artifi cial weathering, there were clear differences in the fi nal hydrocarbon content of samples taken. This refl ected dif ferential loss through processes such as evaporation during the preparation of the OBS.. Starting concentrations are given in Table 4.1.

Differences in decomposition between residues derived from the three oil types varied depending Table 4.2. Microbial activity (measured as CO, on the initial starting material and the treatment fl ux) in 0 135 p repa red using diff erent oils with given. Washing reduced the hydrocarbon diff erent degrees of weathering. concentration by •a relatively small amount, but exposure on an artifi cial beach produced a much Oil type and CO2 flux (nM CO2 g dry day') larger reduction. Increasing weathering reduced treatment Topsoil OBS Sand the degradation due to microbial activity and that Medium fuel oil degradation varied between oil types, being greatest in Forties and least in fuel oil . unweathered 0.42 0.52 0.09 washed 0.16 0.42 0.03 Weathering treatment infl uenced respiration rate washed & weathered 0.23 0.11 0.04 in the lysimeters during the summer of 1994, with OBS prepared with unweathered oil s producing a Kuwait crude oil greater fl ux of CO2 than that from washed oil . unweathered 0.52 0.31 0.10 Washed and weathered oil residues produced washed 0.38 0.50 0.10 the least amount of CO2. During the winter of washed & weathered 0.29 0.22 0.08 1994/95 there was very little fl ux of CO2 but, during the spring and summer of 1995 rates rose Forties crude oil again. However, although an overall signifi cant unweathered 0.63 0.21 0.08 effect of treatments was detected in washed 0.38 0.23 0.09 measurements made in July 1995 (F=3.2: p<0.02), washed & weathered 0.29 0.19 0.09 the only signifi cant difference between individual treatments was that between untreated Control 0.26 0.05 0.04 Forties, and washed and weathered Forties. Flux rate from untreated Forties oil residues was some fi ve times as high as that from the washed and weathered equivalent (40 mg C m-2 h-' compared with 8 mg C m-2 h4).

Table 4.3. Lysimeter experiment: eff ect of starting hydrocarbon concentration on deg radation aft er eight months and on carbon dioxide fl ux in summer and spring.

Nominal starting Actual concentration (%) Net flux of CO2 concentration (%) Initial Final % degraded August March

2.5 2.77 2.83 0 11.27 7.32 5 5.12 2.83 45 19.21 7.11 7.5 7.17 3.2 55 23.87 11.12 8.5 8.3 7.83 6 4.48 0 10 9.51 6.17 35 7.00 1.01

22 The results suggest that the lighter components of 4.3 Eff ects of tem perature the oil, which are lost during the washing and weathering processes, may be those which are Because microbial activity is normally affected subject to initial and more rapid degradation by micro- strongly by temperature, it is to be expected that organisms . hydrocarbon degradation and carbon dioxide emission would be influenced by this factor. Such Although miciobial activity measured in the a relationship was demonstrated in a microcosm laboratory (samples taken from lysimeters in experiment in which OBS samples taken from the November 1994) showed overall differe nces between dune burial plots at Eskmeals were incubated at treatments, within the OBS prepared with crude oil different temperatures in the laboratory. iesidues, no differences were attributable to Respiratory activity was measured as CO2 flux. weathering treatment. Th is is unlike the fuel oil residues, where weathering significantly reduced the The results showed a more or less linear increase CO2emission rate compared with washing only and in CO2emission with rise in temperature, up to no treatment. An incidental effect of the use of 20°C and a more rapid response above that. different oils was to produce different oil residue Samples of OBS from two depths showed similar, concentrations in the starting OBS. The second elevated, respiration rates compared with controls. lysimeter experiment was designed to investigate The control sample taken from 30-35cm deep more rigorously the effect of differences in starting showed a higher rate of respiration than that from concentration. 60-65cm because of the presence of other organic matter in the surface horizons of the dune soil. After eight months of this experime nt, the effect of different starting concentrations of oil residues in Temperature strongly influenced weekly CO2 OBS was to produce differences in the proportion of emission rates within lysimeters containing OBS hydrocarbon degraded ; from zero (though this may be made with sands from different locations around a spurious result) to 55%. No consistent pattern was the coast, confirming the findings of the found to relate decomposition directly to starting laboratory incubation study (Fig 4.2). The concentration. The results do, however, suggest that relationship betwee n net CO2 flux and temperature there was a progressive rise in decomposition rate within the OBS layer was highly significant. with increasing concentration, followed by a fall, as concentrations rose above 7.5%. A further indication Comparison of the increase in flux between 5 and that this may be the case was shown by the pattern of 15°C (of 3 1 mg C ni-2h-') and between 15 and 25°C emission rates of CO2 from the lysimeters. During (104 mg C m 2. Fr') demonstrated a Qi0 value of August 1996 net flux measurements over a 24 hour approximately 3. That is, a rise in temperature of period, at a time of expected high microbial activity, ten degrees from 5 to 15°C or 15 to 25°C increased increased progressively with increasing oil residue microbial activity by a factor of three (Fig 4.3). concentration up to 7.5% and then decreased to Values of Qmgreater than two are usually taken to values below those for a 2.5% starting concentration. indicate enhanced biological activity. Confirmation The pattern in March was similar although the actual of the effect of temperature over a smaller time values obtained were lower. There was no difference scale was also demonstrated in lysimeters. Carbon between flux rates from OBS with starting dioxide fluxes followed a diurnal pattern over a concentrations of 2.5% and 5% but, above 7.5%, period of 48 hours, refl ecting changes in (where maximum rates were achieved) flux was again temperature recorded in the OBS layer and reduced significantly. tracking variation in air temperature (Fig 4.4).

The most likely explanations are: 4.4 Eff ects of moisture • at high concentrations, much of the pore space is filled by oil which, allied to the increasing Within the first few weeks of the start of the third frequency of flooding on the surface of these lysimeter experiment it became obvious that lysimeters following periods of rainfall, leads to rainfall affected the evolution of carbon dioxide development of anaerobic conditions and into the headspace above the sand surface. reduction in aerobic microbial activity. Recent rainfall events appeared to coincide with much lower flux values, resulting in highly variable • at concentrations greater than 7.5%, the oil between-week measurements. These "quenching" residues show a toxic effect. effects continued to be observed through to the

23 y =484.63e0•0237X R2 = 0.9606 y = 493.82e°.°623" R2 = 0 .9 6 1 3

y = 423.740 .°624 R2= 0.9867

A

• y = 1 7 1 . 5 3 e 0 . 0 6 8 5 X € 1000 Ct R 2 = 0.7429

500

0 5 10 15 20 25 30 temperature (°C)

Figure 4.2. Eff ect of temp erature on carbon dioxide f lux f rom incubated OBS (green) and control sand (blue) samples taken f ront dune burial sites. Samples f ront 30cm are indicated by triangles; those f rom 60cm by squa res.

conelusion of the experiment and affected both oil residues in the OBS at I month and sO allow control and treatment lysimeters. However, easier comparison: because of decreasing microbial activity with in the OBS, the effeets were not as pronounced Winter burial during the latter part of the experimental period. Block I During periods of light rain, presumably when Y= 4.67. X' '2'9 122= 0.96 steady infiltration and through-fl ow were Block 2 suff icient to counteract the tendency for Y= 4.98.X-" R 2= 0.82 waterlogging to occur, the rate of emission of Block 3 carbon dioxide was unaffected. Y= 5.09. X-". R 2= 0.74

The influence of moisture on microbiological Summer burial acti vity in the fi eld has also been demonstrated Block 1 in two trials. Continuous measurements made in Y= 4.59. X-".2" R'= 0.88 the first Pendine experiment showed that, in a Block 2 well drained system with a low concentration of Y= 4.78. X ').2AKI Ra= 0.90 aggregated oil residues, a period of rainfall Block 3 enhanced the rate of carbon dioxide emission Y= 4.93. X-°103 R2= 0.97 (Fig 4.5). I n both winter and summer burial treatments, plots Variation in water table fluctuation within the in block I showed the fastest rate of hydrocarbon Eskmeals trial appeared to infl uence oil residue degradation and those in block 3 the slowest (Fig degradation directly in the dune pasture burial 4.6). Because of the way the experiment was experiment. Power curves fi tted to hydrocarbon arranged. with block 3 on the lower part of a decay data are based on the following sloping site and block 1 on the upper part, equations, but have been adj usted to give a groundwater rose towards the OBS layer more constant starting point of 5% concentration of frequently in block 3. The water table rose and

24 120 - y = 0.0078x3 0.0321x2 - 0.036x + 6.6552 4-- R2 =0.804 =100 -

E 80 1 2 - 97 60 -

40 cri

ON 20 - o o o 5 10 15 20 25 OBS temperature (°C) 0

Figure 4.3. Change in carbon f l ux with rising 800 control temperature in lysimeter experiment using -- 0 1381 4-- diff erent sands. — 0 882 600 — 0 883

? 400

-01180 0 200 8 140 0 0 10 20 30 40 50 60 Le 100 time (hrs) from midnight on 4th July 1994

Figure 4.5, Rainfal l and carbon dioxide fl ux over two days in f irst Pendine experiment. O N 20 10 11 date in October 1997 6i Figure 4.4. Diurnal variation in ai r (black) and — Block 1 — Block 2 OBS (blue) temperature, and as influence on - Block 3 carbon dioxide f lux (red). Temperatures are )', shown as actual values x 10 in order to achi eve a 4 t. common scale. 3 y-t5x-(Lies y ., 5 x- 0.18 1 2 penetrated into both OBS deposits during winter cn .6 E periods of all three years of the experiment (Figure '6 1 E 4.7). However, this penetration was higher and 2 more prolonged in block 3, producing longer oc 0 periods when anaerobic conditions prevailed and € 6 co0 so restricted degradation activity. 2 = Sbi l moisture was also one of the measurements e -, made to characterise sands from different sites in \ " ...„ the extensive field trial. Although overall there was 3 -,‘.',7:.-...... : L ., .-- -- y=5,-0.163 y. 5x-o.2 no relationship between initial moisture content - . _ 2 y=5x-°.239 and amount of hydrocarbon degraded. there appear to be two groups of sites. One had a relatively low i water content and showed a range of capabil ities

for decomposition of oil residues. whilst the other 0 0 20 40 60 80 100 120 had a wider range of soil moisture and a lower time (weeks) overall capacity to degrade oil residues. If the two groups are taken separately, each shows a positive Figure 4.6. Block diff erences between relationship between water content and capacity to hydrocarbon decay cur ves for summer and winter degrade hydrocarbons. This result appears to burial in dune pastime at Eskmeals.

25 9Z

wm autys3 IV m auip ail y a jwi nq fn E :p o p ? (a o w )f a o l g "a ll y 0 3 0 17 sgo Il eo y n i ad 117 S IIO W E IL M If d / q 131 .0 113/4 4 y 7.3 •f i l 8 3 97

amp 96/£0/ 1.0 16/0116Z L6/90/9 Z L6W /9Z 96/0 M 96/90117Z 962 0/ZZ 96/0 1./Z2 96190/ L 96/ZO/ZZ

9• I.—

1 1—

0' L-

9'0—

1 0—

L 0'0 86/£04 0 L6/01./6Z / 6/90/8Z / 62 0/9Z 96/01./9 Z 96/90/I7Z 96/ZOIZZ 96/OLITZ g6/90/ 1.Z 96/TO/ZZ

9 I.—

9 0-

17'0—

T O—

,— 0•0

96/170/90 LS/ Ll /OE L6110193 £6/80/1.Z 96/ 11/14 96/10/0 1. 96/£0/90 96/01./OE 96/90/93 96/30/33

O'C-

97 —

0' 1-

SSO T O— contradict the findings of the lysitheter experiment. The extensive trial was undertaken to deter mine However, in well-drained, coarse-textured systems, whether differences on a larger scale did infl uence such as those found in coastal dunes, moisture the degradation process. Results from this trial availability may be a factor limiting degradation at were assessed in conjunction with those from a times. Where fine material is incorporated in the collateral trial at Eskmeals using sand from all sand, or the sands are well cemented, there may be extensive trial sites. and from a lysimeter experiment a greater tendency for temporary waterlogging in which sands from five of the locations were used and reduced degradation. to prepare OBS to a common starting concentration. These trials, where external 4 .5 Eff ects of sand type conditions were common, revealed differences in the rate of hydrocarbon degradation. The burial trials established at Eskmeals used quarry sand as the bulking medium in preparation Field trial of OBS. The rate of degradation was observed to Because of confounding factors in oil preparation, be lower than that in the first Pendine experiment. site location and site characteristics, it was difficult It is possible that a lack of buffering in the to relate rates of hydrocarbon decomposition to Eskmeals quarry sand, compared with the Pendine specific factors. However, differences were found beach sand contributed to this difference. The (Table 4.4) and these appeared to be related more presence of calcium carbonate in shell fragments to characteristics of the sands used in preparation in the beach sand incorporated in the Pendine trial of OBS than to local climatic factors. This seems could have provided the necessary buffering to somewhat surprising given that temperature is prevent a fall in pH unfavourable to bacterial clearly an important factor controlling hydrocarbon growth. degradation and clear differences were found in temperature regime betwee n sites. Another Variation in the native sand into which OBS was confounding factor was the tendency of some sites introd uced appeared to have little effect at a local to flood, though the extent and duration of such scale. No significant differences were observed flooding is not known. However, at one site between burial plots in dune and dune pasture (Spiggie) final sampling was not possible because soils at Eskmeals. Results show that the driest of the site was inundated. the three blocks in the dune pasture burial trial matched the dune burial plots in terms of the The results of a principal components analysis degradation rate function and the total amount of showed a strong inter-correlation among ten hydrocarbon degraded over a two year period. variables used to describe each site, though no Equations for the two locations were : single variable played a dominant role. Three components produced by the analysis accounted Y=4.67.70 2'9 for 70% of the variation encountered. These were : for the dune pasture burial and • calcium-magnesium-organic matter content of beach sand used ; Y=4.67.X-"" • moisture content of OBS and pH of beach sand; • site rainfall and a possible association with for the dune burial. water table.

Mble 4.4. Variation in chemical and physical characteristics of sands from extensive trial sites used in the third lysimeter experiment.

Loss on Moisture Pa nicle size (% composition) Site pH ignition (%) Ca lcium (%) Nitroge n (%) content (%) 2mm >1m m >0.2mm >0.2m m

Aske mish 8 .3 1.2 14.0 0 .015 22.0 0.0 0.3 94 .0 - 5 .4 Cres swell 6 .9 0.7 2.5 0 .002 1.9 0.0 1.5 98.0 0 .1 La rgo 8 .4 0.9 4 .5 0 .006 6 .1. 0.1 1.6 96.0 2 .6 Pendine 8.2 0.7 3 .6 0 .006 3 .0 0.0 0.5 76.0 24 .0 . Tain 7.9 1.4 0 .1 0 .04 1 20 .0 0.0 4 .8 65.0 30.0

27 Table 4.5.Variation in microbial activity collected from five of the locations used for the (measured as carbon dioxide f lux) in 0 138 extensive field trial. p repared using sands f m ni different extensive trial sites. Values are exp ressed as mg C in 2 Ir' Three of the sands (Askernish, Cresswe ll and Largo) were characterised by relatively large grain size, though Askernish also had a high calcium and organic matter content. Cresswell differed from Largo principally in its lower calcium content and lower pH. Of the two sands with a high proportion of fine particles. Tain had much more organic matter and nitrogen than Pendine, but a lower calcium content.

Measurements of carbon dioxide flux after one week showed that lysi meters containing Askernish OBS gave the highest rates of carbon dioxide production. These lysimeters continued to produce the highest flux values throughout the experiment (Table 4.5). However, OBS prepared using sand from Tain Collateral trial showed a high initial rate of microbial activity but, Within the collateral trial at Eskmeals, effects of on autumn, winter and spring sampling occasions it d ifferences in sand type were revealed . A consistently produced the lowest values of all five significant relationship was found between the sands tested. Sands from Largo and Pendine gave proportion of hydrocarbon degraded and the final consistently similar values approximately two moisture content of the OBS mixture. Soil moisture thirds that of Askernish. Cresswell samples were content is dependent on a number of intrinsic co nsistently somewhat lower than Largo and features of the sand: grain size and organic matter Pendine. In control samples. Askernish and content are two of the more important. The Cresswell had similar, low, rates of microbial influence of water co ntent in this trial re-enforced activity, with Largo and Pendine having slightly the results from earlier lysimeter experiments, and higher activity. The highest rate was found in Tain the burial and land farming field trials at Eskmeals. controls. A high water content inhibited ae robic degradation of oil residues by micro-organisms. A possible explanation is that this sand contained a suitable carbon substrate for microbial Lysim eter experim en t decomposition, though this need not have been Under the standardised conditions of lysimeter based on petroleum hydrocarbons. Tain sand was experiment 3, there was considerable variation in known from analysis to have been finer than the the rate of oil res idue decomposition in sands others sampled and had a higher Organic matter content (as measured by loss on ignition).

Ca rbon dioxide flux has been used to calculate - 700 y = 1071.9C ° 1479 degradation rates without the need for destructive R 2 = 0.8987 sampling. ANOVA showed that, for individual .1 600 samples. variances about the mean in carbon dioxide flux were comparable over time. This meant 2 500 that random errors were consistent between one 7/3 400 set of measurements and another and we could 2 have confidence in using these mean values in the re. 300 • predictive model. A matched pair t-test confirmed 0 50 100 150 200 250 300 350 400 that only Askernish gave a significantly different time (days) (higher) rate of respiration among all tested pairs. Figure 4.8. Ilyarmcarbon degradation in OBS PlotSofchange in estimated oil residue mass with p repared using Askernish sand. Calculated values time have thus been possible assuming a starting are based on measured carbon dioxide ,flux weight of 625g of oil residue (5%) in the OBS values . A power curve has been f itted to the data. (P ig 4.8).

28 present may not be achieved . Hov evcr. this 50 content ion is open to argum ent as almost any population o f bacteria is capable of adapting to 40 degrade lTh drocarbons: and the presence of one active co mmunity does not lead to exclusion of another.

4.6 Effects of microbial differ ences

Based on colimy morpholog). wh ich may not 0 always bc reliable. differe nces were seen in the composition of the bacterial communities from - 10 beach sand and OBS. Some colony types 0 1 2 3 4 5 6 7 8 disappeared after the add ition of oil resid ues . fina l moisture (%) which co uld sugge st differential toxicity. Whilst Fig ure 4.9. relationship between f inal moisture prediction of activity may not ye t he possi ble. our content of ORS, prepared using thjf erent sands, results show that, ge nerally. microbial popu lations and hydrocarbon loss. ada pt rapid ly and become capable of degrading hydrocarbo ns .

After eight months the estimated reduction in oil The rate, and exte nt, of degrada tion in the residues was 43% for Askernish and 28% for extensi ve trial will ha ve de pended on many Factors, Cresswell. Largo, and Pendine. Although Tain OBS not least of which wou ld ha ve been the po llutio n lost a similar percentage, carbon dioxide production history of the beach sand used to prepare the OBS . ceased after 130 days of the experiment. This The inability to detect a capacity for deg rading suggests a very high level of initial activity but that aromatic molecules might suggest the beach sand some factor limiting that activity came into came from a relatively unpolluted location, but operation subsequently. exposure to oil residues still allowed populations with this ability to develop. The results obtained did not precisely match expectations from laboratory incubation of sands from the different sites, which predicted that Thble 4.6. Cultw oble bacterial counts obtai ned Cresswell would show low activity. Pendine and f rom samples taken from the extensive study sites Largo would show intermediate activity, and that in spring 1996. Counts are gi ven in terms of Askernish and Tain would he the most active. Tain number of coloni es x If ' per gram dry weight of samples, in particular, did not behave as expected. material sampled. ns = no sample. Although an initial active population of micro- organisms was present in Tain sand, it would Beach na tive appear that other factors in the field or in lysimeter Site sa nd OBS mix sa nd experiments limit actual degradation. Among the Aske rnish 17.9 124 3.7 more important of these factors may be the Ca musda rach 3.8 1.6 6.5 fineness of the sand itself, where a small pore size, Clirnping 0.8 0.9 69.2 may retain water, and limit the movement of the Cresswe ll 1.1 0,4 21.4 oxygen and nutrients necessary for microbial Eskmea ls 2.6 1.1 1.4 growth. That water content may be important is Horsey 8.9 8.3 12.3 indicated by the results, after two years, from the La rgo 4.6 4 ns collateral trial, where there was a clear relationship Northa m 3.2 3.3 31.7 between hydrocarbon loss and moisture content of 2.6 21.8 9.1 the OBS (Fig 4.9). Oldshoremore Pe ndine 5 6.4 8.7 Bay of Skelt 2.2 0.3 18.1 It has also been suggested (Ellis & Adams 196 1) 8.4 35.8 that the presence of some other organic compound Sa ltfleetby 2.6 14.6 9.4 in soil may provide an alternative, preferred, Spiggie 9.9 substrate which means that the full hydrocarbon- Tain 1.8 8.1 18.1 degrading potential of the microbial populations Torrs Wa rren 3.6 15.7 1.6

29 Microbial respiratory activity was generally greater • Immediate, moderate, stimulation of activity, in dune soils in which the OBS bags were buried in which was maintained (Bay of Skelt, Climping, the extensive trial than in beach sand (Table 4.7). Northam, Pendine, Spiggie) or increased However, the reverse was true for samples from (Eskmeals, Oldshore More, T OR S Warren) over Askernish, Bay of Skelt, Tain and T OR S Warren 3 or 6 months of burial. where high respiratory activity was associated with • Immediate, large, stimulation of activity that incubated samples of beach sand . One possible was maintained for 3 to 6 months (Askernish, explanation is that this may have been related to Largo, and Tain). the recent pollution history of the beaches • Little immediate stimulation, but an increase in co ncerned. As noted in Section 2.2, oil-polluted activity after 3 months (Cresswell, Horsey) or 6 sites tend to have larger populations of months (Camusdarach) exposure to OBS. hydrocarbon-degraders than unpolluted sites. • No stimulation within the first six months However, the beaches in question are not those (Saltfleetby). where there is a high expectation of oil pollution. These fi ndings suggest that there may be large Spring burial differences in the adaptability of microbial With the exception of Saltfleetby, microbial populations between sites. We have not respiration was stimulated after the addition of oil investigated whether this is related to site specifi c residues (indicating metabolism, i.e. degradation of differences in the species, or to the balance those residues). The level of stimulation varied. It between bacteria, fungi, yeasts and algae (all of was high in sam ples from Askernish, Largo and which ta xonomic groups have species capable of Tain, but low in samples from Camusdarach, degrading hydrocarbons). The capacity to maintain Cresswell, Horsey and Saltfleetby. It was moderate any activity which is stimulated is also variable and in all other samples (Table 4.8). this may be a function of the physical and chemical nature of the deposits within which the micro- Although there was considerable variation organisms are found, rather than a property of the between replicate samples removed from each site, microbes themselves. Any feature which reduces the data showed similar trends, with four the availability of nutrients or oxygen, or which observable patterns of microbial response to oil prevents the dispersal of breakdown products and residue addition: so leads to changes in the local environment, may

Table 4.7. Respiratory activity of samples taken Table 4.8. Bacterial respiratory activity in f rom extensive study sites in spring 1996. Values samples taken f rom OBS buried in spring 1996 at gi ven are micromo les CO 2g-' dry weight day ' . diff erent locations. Values are micronw les CO 2g ' ns no sample. dry weight day '

30 Table 4.9. Bacterial respiratory activity in Tabl e 4.10. Fungal distribution in beach sand and samples taken from OBS buried in autumn1996 at OBS samplesfrom 5 sites around Britain. diff erent locations. Values are micromoles CO2( ' Bacterial counts are 106 dry weight of material dry weight day ' sampled Fungal species and colonies ar e total numbers per plate.

those sites with high bacterial counts had low fungal counts.

It is known that, in general, bacteria are favoured by growth media which are circum-neutral, whilst fungi tend to grow in more acidic conditions. However, no correlation was found between pH of initial beach sand and fungal growth. No data inhibit microbial activity. The sands do differ in a exists on the pH of OBS at the time of sampli ng for number of physical and chemical characteristics. fungi because the presence of oil residues makes accurate measurement diffi cult. Autumn burial As in the spring burial, addition of oil residues stimulated microbial activity. The level of 4.7 Effects of treatment stimulation was high in OBS buried at Askernish, Bay of Skelt, Largo and, possibly, Pendine, low at In translating experiments at a small scale to Camusdarach, and intermediate at the other sites. realistic and practical application at the operational There were three notable differences in response of scale, it is necessary to take account of different beach sand populations compared with spring management procedures, which may be adopted, burial. and the effects that these will have on the 1) High activity at Bay of Skelt degradation process. Minimal dif ferences occurred 2) Moderate stimulation at Saltfleetby between the patterns of breakdown of OBS in 3) Low activity at Largo. burial plots, whether these were in dunes or dune pasture. In addition, catechol degraders were found in samples of beach sand from Tain and OBS from The establishment of burial and landfarming trials Climping. side by side at Eskmeals gave the opportunity for comparisons between such treatments. Fungi Examination of samples taken from OBS, which had The two major differences between burial and been buried for six months at each of six sites landfarming plots were: showed different patterns of fungal colonisation. • Whilst OBS in burial plots was not exposed Plates prepared using beach sand had few, if any directly to the atmosphere, landfarming plots fungal colonies, and those using OBS consistently were. showed many more colonies and species. There • OBS in burial plots remained undisturbed once were large differences in actual numbers recorded in place, but in landfarming plots it was mixed at depending on site of origin of the sand. In general, regular intervals.

31 5 dune burial — winte r burial — s ummer buria l

4

2 y 4 . 5 7 X - 0 2 1 4 y = 4.67x4 .219 y = 4.59x4 .239

0 0 20 40 60 80 100 120 time from burial (weeks )

Figure 4 10. Comparison of f itted decay t urves f or dune burial plots, and winter and summer dune pasture bu rial in block 1 within the drier part of the pasture.

However, for a given land area, burial plots were was in the order: 2-weekly plough>4 weekly capable of accepting greater volumes of OBS than plough>no plough. Clearly, frequent ploughing had landfarming plots. Hence, two different a major effect on hydrocarbon breakdown rate, comparisons have been made : between rates of possibly because it increased OBS aeration by degradation and total amount of oil residue bringing buried material to the surface, and also degraded per unit area of land occupied. redistributed, moisture, nutrients and microbial populations. Decay curves for incorporated oil residues in both burial and landfarming plots showed rapid initial Winter burial appeared to be more favourable for falls in hexane-extractable hydrocarbon degradation than summer burial. This may be a real concentration, followed by subsequent slowing of effect, resulting from the creation of more suitable the rate of loss. As in other exper iments, a series of moisture conditions in winter, or an artifact produced power equations best explained the patterns of as a result of incomplete incorporation of the OBS degradation. However, the rate function in the into the upper layer of sand in the summer treatment. equations differed depending on treatment app lied. In summer treatments there was a tendency for the topsoil to dry as large clods. Such drying may have Compar ison between landf arming produced conditions where water became a rate treatm ents limiting factor (cf. first Pendine experiment). The rate of loss of hydrocarbons from OBS in landfartning plots depended on both the time of Based on an initial hydrocarbon concentration of incorporation (winter or summer ) and the frequency 4.8% within the OBS incorporated into the of ploughing. In both winter and summer burial landfarming plots, calculation of the amounts of oil plots the rate of decomposition of the oil residues residues degraded, by weight, have been made.

32 lo

0 20 40 60 80 100 120 Prne (wee ks) from incorpo ration

Figure 4. 14 Progress of 0 8 5 breakdown in lancUam ing plots established in winter and siumner and subjected to different ploughing treatments (mean values f or all three blocks)

These show that ploughing every two weeks plots than in burial plots, though not in unploughed increased by 35% the efficiency of hydrocarbon landfarming plots. Unploughed winter plots showed brea kdown compared with that in the no-plough sim ilar degradation curves to burial plots, but treatment in winter plots. The efficiency of degradation unploughed summer plots appeared to retain high was doubled by ro to vatio n every two weeks in the concentrations of oil residues. Landfarraing with summer incorporation plots (Fig 4.11) frequent ploughing was, thus, more efficient at hydrocarbon removal than burial.

Comparison of landfanning and burial Ca lculation of total quantities of oil residue degraded The degradation rate was consistently higher in showed that burial plots could accept a larger ploughed landfarming plots than in burial plots. quantity of material initially. Burial plots also The process of degradation was more rapid and more degrading a larger absolute amount of hydrocarbon complete over two years in ploughed landfarm ing than landfarming plots of equivalent area.

33 Table 4.11. Compa rison of percentage s umme r: — no ploug hing decomposition of hyd rocarbons under burial and - - burial landtarm ing treatments. - 2-weekly ploughing winte r: — no ploug hing % degraded - • - burial Dis posal method a fter 2 yea rs 2-weekly ploug hing Dune buria l 60.1 Winter burial 18.0 Summe r buria l 20.0

Winter landfarming 2-weekly plough 82.1 4-weekly plough 71.6 20 40 60 80 100 no plough 4 1.0 time (wee ks ) from incorporation/b urial Summe r landfa rm ing Figure 4.12. Fitted decay curves f or breakdown of 2-weekly plough 54.5 OBS in landf ann ing and burial plots at Eskmeals. 4-wee kly plough 44.8 no plough 29.1 4 .8 Interactions Table 4.12. Comparison of total weights (tonnes The results of the field trials and the lysimeter p er hectare) of hydrocarbon degrad ed in burial experiments indicate that there is no single and landf arming p lots at Eskmeals. controlling factor that determines the rate of decomposition of oil residues. This is not an Oil residues inOBS unexpected conclusion, bearing in mind the initial initia l de graded proposition (Figure 2.1) that moisture, oxygen Dis posal method weight aft er 2 ye a rs supply, temperature, micro-organisms, and nutrients were likely to be key factors control h ng Dune buria l 334 20 1 degradation. Fr om a practical po int of view, bearing W inter buria l 1008 182 in mind the starting materials used , temperature and Summer burial 1029 2 10 moisture appear to be the most important variables. Winter la ndfarming The lysimeter experiment's, in particular, illustrated 2-wee kly plough 134 110 the interaction of these two climatic variables and a 4-wee kly plough 134 96 relationship has been derived which links rainfall no plough 134 55 within the previous twelve hours and soil temperature with the flux of carbon dioxide: S umme r la ndfa rming 2-wee kly plough 134 73 Logm(F) = 0.93 + 0.058(T) - 0.042(P) 4-wee kly plough 134 60 no plough 134 39 where F is CO2 flux (mg C.m-2.1r 5 T is temperature within the OBS ("C) P is precipitation over the previous 12 hours ( n m),

This equation would not apply to all situations because the depth of OBS and the initial oil type present would have an influence on degradation potential. However, the degradation pattern would follow the general - power curve- relationship over time in each case.

A fully predictive equation would need to be based on measurements made of all the interacting factors which infl uence the microbial degradation of petroleum hydrocarbons.

34 5. DEGRADATION PRODUCTS AND THE IR FATE

5.1 I ntr oduction monitor the concentrations of hydrocarbons in drainage water from OBS in lysimeters and in Some of the compounds present in the groundwater below fi eld trials. A more restricted unweathered parent oil s are known to be sampling programme was carried out to determine biologically active and to produce toxic eff ects. the presence of PAHs in these waters. Acute eff ects are shown by exposure to the I3TEX group of cyclic compounds (benzene, toluene, ethylbenzene and xylene). Polycyclic 5.2 Heavy metals aromatic hydrocarbons (PAHs), which are ubiquitous in the environment, may be Although crude oil s contain heavy metals, particularly toxic and some are known prinéipally vanadium (V) and nickel (Ni), the carcinogens or mutagens. However, because concentrations found are not normally above a few many of the more volatile, low molecular weight, parts per million. In some oils, e.g. Boscan crude compounds (including the BTEX compounds) from Venezuela, the concentrations of Ni may reach are not present in weathered oil residues, these I 00ppm and V may be as high as 1,000ppm (Stencel residues may not pose a signifi cant threat to the & Jaffe, 1998). Such high concentrations are rare, environment. however. More typical values are shown in Table 5.1 together with the WHO drinking water guideline Hydrocarbons consist almost entirely of carbon value for Ni (though none have yet been set for V and hydrogen atoms and the main product of by WHO). Environment Assessment Levels (EALs) oxidation is carbon dioxide, which poses no set by. the Environment Agency vary depending on pollution threat. Complete degradation of water hardness. For nickel, they range from 50pg hydrocarbons is desirable, but it is possible that in the softest waters (containing less than 50mg 14 partial biological transformation could result in of calcium carbonate) to 200pg 14 for very hard the release of potentially toxic intermediate waters (with more than 200 mg14 of calcium compounds. carbonate). For vanadium, Water wiih less than 200 mg 14 of calcium carbonate, the EAL is 20pg 14 and The production of intermediate compounds for harder waters it is 60pg 14 . In coastal waters, during the course of degradation and the the EAL is 100pg 14 of vanadium. In the context of existence of PAHs in weathered oil s may pose the.current study, heavy metal contamination was threats to microbial, plant and animal not considered to pose a serious environmental communities, if released into the environment at hazard because of the low concentrations involved. concentrations Aer ie known critical thresholds. They are not easily eluted as free ions and pH values are high enough in seawater and coastal In order for a disposal method to be sandy soils to prevent mobil isation in the dune environmentally acceptable, the risks of environment. producing and releasing signifi cant quantities of toxic compounds must be reduced to a minimum. Concentrations must remain below the Table 5.1. Concentrations of nickel and vanadium critical thresholds. As well as release, the in some typical oils and calculated problem of transport of potentially toxic concentrations in 5% OBS p repared using those compounds to groundwater, and hence their oils. ns = no standard. wider dispersal, has had to be addressed. Because of this, the proj ect included a maj or Concentr atb n (mg kg') component designed to determine whether there In on In 5% OBS was any movement of environmentally Oil type NI V NI V unacceptable materials from the OBS to surrounding soil or underlying groundwater. In Kuvaul crude 9.6 31 0.48 1.55 the Feasibility Study phase of the proj ect, Nigerian a ude 5 <0.5 0.25 <0.03 heavy metals, some hydrocarbons and PAHs Libyan a tkle 4.3 1.7 0.21 0.08 were identified as possible sources of concern. Forties (n ide 2 3 0.1 0.15 For reasons given below, heavy metal ion range in 6 fuel oh 1.5 - 66 2.7 - 180 0.07 - 3.3 0.13 - 0.16 concentrations were not measured. Samples WHO guidelines (mg 11 0.02 ns were taken on a regular basis to detect and

35 5.3 Changes in hydrocarbon Within the OBS, some polycyclic aromatic com position hydrocarbons were found. On the basis of concentrations found in the original oil, it is As oil residues degrade, the more readily possible to calculate the concentrations in the OBS decomposed hydrocarbon components, e.g. short deposit at the start of the experiment. chain alkanes, disappear first, leaving progressively more recalcitra nt compounds behind. The current No significant hydrocarbon peaks were present in project did not investigate the changes in the control samples and this pattern was repeated composition of oil residues in detail as the overall in material taken from beneath the OBS deposit process is well recorded (Ke nnicut 1988. after 10 months. In some samples. traces of Blenk insa pp et a l. 1995). Instead, samples were weathered fuel oil were found, but this may be taken at inter vals from a limited number of the attributable to some cross-contamination in experiments and were subj ected to analysis to sampling. determine the overall hydrocarbon composition and , more specifica lly, the concentration of represe ntative PAH's. In particular, sam ples from the second 5.4 Vertical migration Pe ndine trial have been analysed by GC-MS. One objective of the second lysimeter experiment T h e re s u lt of the analyses of samples from the was to examine whether the starting concentration second Pendine trial, over a period of eighteen of oil residue in OBS infl uenced the potential for months, followed the expected course. There was a downward movement of hydrocarbons through the reduction in the co ncentration of hydrocarbons over soil to groundwater. Microbial respiration by a large part of the range and virtual loss, as a result samples of sand from beneath the OBS under of microbial degradation, of short chain compounds laboratory conditions was used to detect the within the upper layer of the OBS deposit. Whilst presence of such hydrocarbons and the potential hydrocarbons were also removed from the lower for their removal. In all lysimeters there was layers of the OBS, larger quantities of short chain greater activity beneath those containing OBS than compounds remained in this part of the deposit. In in the controls and evidence of a trend of both parts of the OBS deposit, molecules with increasing respiratory activity with increasing carbon chain lengths of 20-2 1 produced maximum concentration of oil residues incorporated in the peak heights throughout the period of OBS, despite variation between individual measurement. lysimeters.

Tab le 5.2. PA H con centra tions (pg ,g4) in f ne l oil, em ulsion and OBS p rep a red / Or second Pendine - Immediately be low OBS - 70cm be low OBS

6 0.8 y = 0.0638x + 0.0559 0" R2 -= 0.56 94 2 0.6 E

a 0.4 a) y = 0 .0102x + 0.0769 70 37c R 2 = 0.7422 ° 0.2 a

0 .0 (-) 0 2 4 6 8 10 12 Oil res idue co ncentration

Figure 5. 1. Virriation in microbial resp iration rate in sand f rom two depths beneath 0 135' in Ivsimeters. with di eren t starl ing co ncen tratio ns Qfoil residues.

36 Table 5.2. Changes o v e r t i m e in the concentration of chloride in groundwater sa i np l e s taken fivm piezometer tubes located on the landward side of the experimental area, within the control and experimental hollows and on the seaward side of the experimental lw llow

l i me from incorporation (months)

Figure 5.2. Hydrocarbon content of sand samples taken from beneath OBS deposit at seeond Pendine experiment site. Samples taken from 1-5cm below the base of the OBS.

With increasing depth, the rate of respiration decreased, indicating a combination of lower microbial activity and lower concentrations of substrate hydrocarbons. Although there was some sand and its included seawater had occurred. movement of hydrocarbons from the OBS into the Measurements made of chloride (an easily leached underlying sand, this was accompanied by ion) concentration showed that an initial fl ush of populations of hydrocarbon-degrading micro- inorganic anions had taken place. The rise in organisms. At 70cm below the OBS there was concentration within areas adjacent to the control virtually no respiratory activity or detectable hollow showed a distinct pulse, despite hydrocarbon substrate. The presence of micro- superimposed variations between seasons. organisms in the layer of sand immediately beneath the OBS resulted in breakdown of the migrating Groundwater from beneath the control hollow and hydrocarbons so that, within 70cm, there was its margins showed enhanced levels of chloride, virtually none remaining. but the pattern in concentrations occurring below and around the margins of the OBS deposit I n the second Pendine experiment, core samples suggested that direct percolation of water through showed a clear boundary between OBS and the OBS did not take place. Rather, precipitation underlying beach sand, suggesting there had been was penetrating only a limited distance into the no mass movement of oil residues from the OBS. OBS, and percolating water was shed from within However, analysis of the underburden showed the its upper layers to groundwater around the margins presence of very low concentrations of hydrocarbon of the infilled hollow. This suggests that where the after the first sampling two months after burial of oil residue concentration is high, complete OBS. This suggested a limited amount of initial penetration of water is not possible, and thi s may downward movement and the possibil ity that inhibit the degradation process, as suggested by bacterial populations subsequently degraded the fl ooding found to follow periods of rainfall in hydrocarbons leaching from the OBS. Samples taken the lysimeter experiments containing high at seven and eighteen months were at the limits of concentrations of oil residue in OBS. detection (0.02%).

5.6. Leaching of hydrocarbons to 5.5 Leaching of mineral ions to groundwater groundwater Leachates draining from the lysimeters were Changes in the concentration of mineral ions in odourless, suggesting a minimal hydrocarbon groundwater beneath the second Pendine experiment content. In line with this observation, low confirmed that leaching of these ions from beach concentrations (below the levels considered toxic

37 7able 5.4. PAH concentrations in water draining f rom lysiMeters containing diff erent starting tf - 0.08 = landwa rd - co ntrol concentrations of OBS. under OBS g 0.06 - se award 113 a) E 0.04

-0 0.02

-0 0.00 0 20 4 0 60 80 100 Time from buria l (wee ks ) Figure 5.3. Changes in hydrocarbon concentration in ground water samples taken f rom p iezometer tubes at the second Pendine experiment site.

exceeded in seawater in spring 1997. The rise in concentration of some PAHs in the control hollow following placing of beach sand reflects high concentrations of these same compounds in seawater and appears to be the product of leaching of that seawater from the sand. Analysis of for algae and most invertebrates) of PAHs were samples taken in January, May, July and September also found. 1998 failed to show PAW in water from any of the piezometer tubes. Thus, leaching of these No hydrocarbons were detected in groundwater compounds from OBS deposits in the field appears Samples collected from piezometer tubes around to be, at most, a temporary phenomenon giving rise the trial plots at Eskmeals . Analysis for PAW to no problems of groundwater contamination. sim ilarly failed to reveal the presence of these compounds in groundwater samples taken from the same piezometer tubes.

This lack of movement of hydrocarbons from OBS to groundwater was confirmed in water samples taken at the second Pend ine site. Although concentration of hydrocarbons in groundwater tended to show a slight increase, the amounts present were not significantly above background concentrations. Samples taken after 60 weeks from beneath dunes 2km to the cast of the site (close to the first Pendine experimental site) showed hydrocarbon concentrations ranging trbm 0.007 to 0.0 14 ma ll , whilst seawater and water issuing from the springline on the beaCh gave values of 0.005 and 0.0 11 mg 1.1 respectively.

Samples taken at Pendine before the establishment of the second experiment show baekeround PAH concentrations varying in both time and space. The higher concentrations were not sin ificantly exceeded following the placing of OBS in the experimental hollow. Comparison shows that the concentrations of some compounds found were far

38 6. SITE RECOVE RY

6.1 Nature conser vat ion value of dunes and intervening wet hollows, where the water table may approach. or rise above, the ground Dunes not only form part of natural coastal defence surface. Within these V. et hollows, an systems, but also have high inherent nature additional range of freshwater communities is conservation value. This results from their able to develop. structural features and edaphic characteristics. Dune systems consist of sequential series of ridges Within mature dune systems there may he and hollows showing increasing maturity with some redistribution of wind-blown sand, but increasing distance from the sea. Young dunes large scale disturbance of the surface can lead composed of accumulated wind-blown sand along to excessive erosion. the upper margins of beaches, are well-drained and frequently contain high concentrations of calcium Sand dune systems are, thus, sensitive to but little organic matter. They are, generally. disturbance, which disrupts the delicate sparsely vegetated. As dune systems mature the balance between erosion and deposition, and calcium content is leached away, but there is to nutrient inputs which allow the development surface accumulation of organic matter and, hence, of more mesotrophic (often weedy) vegetation. water-holding capacity of the sand increases. This Despite this, many dune systems have been nutrient poor system supports a specific range of subjected to extensive use in the past and the plant species and the invertebrate fauna dependent majority of the remaining areas of relatively upon them. undamaged dune complexes have some form of nature conservation designation as Sites of The biological diversity of dune systems is further Special Scientific Interest or National Nature increased by the topographic pattern of dry ridges Reserves.

100

90

80

70

60

50

40

30

20 Festuca rubra

10 Plantago lanceolate

0 Holcus lanatus

Fuel oil

Figure 6.1. Survival of f our plant species in OBS prepared using dIfferent oils.

39 In seeking to find suitable areas for disposal of Both oil type and period of weathering affected oil residues, the sensitivity of dune systems has the severity of this effect. After six months, OBS to be taken into account. As part of the process prepared using Forties oil had a greater effect of assessing effects of disposal of OBS in dune than that prepared using medium fuel oil or systems, glasshouse experiments were carried Kuwait crude. out and field observations on vegetation and fauna were made on field trial plots. Plants of the same species (red fescue) collected from different locations around the British coast, or even from different parts of the same 6.2 Toxicity testing of plants population at a single site, showed different responses to the presence of oil residues. This Results from the trial using four different suggests that there is a genetic component to species commonly found in dune systems tolerance of oil residues, which, in practical terms, showed differential survival when planted in may be diffi cult to take into account in the fi eld. 3.6% OBS prepared from different oils. Of the two grasses used, red fesc ue (Festuca rubra) Because of somewhat variable results from these showed good survival after six months in all oil experiments, any overall conclusion on general residue types, but Yorkshire fog (Holcus toxicity is diffi cult to draw. However, experience lana tus) appeared to survive less well, in the extensive field trial suggested that the especially in medium fuel oil OBS (Fig 6.1). presence of oil residues was not a common problem, because roots of dune plants were Ribwort plantain (Plantago lanceolata) showed found to have penetrated many of the bags a considerable reduction in survival in all oil excavated. These roots showed no apparent types and the result for the other broad-leaf adverse effects. spec ies self-heal (Prunella vulgaris) was thought to be affected by drought during one hot, dry, weekend: this species is less tolerant of 6.3 Recovery of dune sites dry conditions tha n the others used. Groundwater chemistry Tillers of red fescue obtained from coastal There is no long term effect on groundwater dunes and planted directly into 5% OBS chemistry resulting from the introduction of a prepared using different weathered oils showed large volume of sand and emulsion, both signifi cant reductions in the proportion of containing seawater. Following an initial fl ush of healthy leaves, compared with control plants. anions and cations from the deposits, the chemistry of groundwater beneath dune systems returned quickly to its initial status. This was Tab le 6.1. Eff ect of diff erent wea thering treatments demonstrated effectively at the site of the second of diff erent oils on growth of red f escue. Results Pendine experiment, where groundwater, exp ressed as p ercentage compared with control characterised by the dominance of calcium and p lants. bicarbonate, passed through a phase in which sodium and chloride became the major ions, before returning to calcium bicarbonate waters once again. The speed of change and its extent depend on the quantity of material incorporated into the dune system, rainfall patterns, infi ltration rate, and the rate of release of ions from the OBS which will, in turn, depend on emulsion composition and concentration of emulsion.

Because the changes observed were rapid and reversible, there is no reason to believe that they would adversely affect the long term survival of dune plant and animal communities though, in the short term, some species may be favoured locally by the temporary introduction of more saline conditions.

40 Plant colonisation Dune secetat ion at the site of the second Clear changes in vegetation were seen in the dune Pendi ne experiment had alread) shown ski ns of burial plots at Eskmeals. A lthough no control plots past disturbance. in particular the presence td had been established. comparable observations tt ere 1/4darden plants in the vicinm of the dune made on the plots themselves, the disturbed arca hollows. though these did not extend into the within the experimental enclosure and in undisturbed hollows themselves. The closed egetation dune vegetation nearby. The results of observatirms present belt de the OBS and control deposits made after t o years. indicated that the undisturbed were put i n place. contrasts with the open dune community was one in which marram vegetation, containinc a large number of (Anunaphil a arenaria l and red fescue had similar, colonising species. which developed high. levels of relati ve abundance. v.ith false oat-grass subsequentl tdlrrhenwherum elahus) and sand sedge (Carex ar enarim as common associates. Disturbance reduced Withi n the dune pasture at Eskineals, signi ficant the abundance of marram and false oat-grass changes also took place in vegetation on the dramatically, whil st common bent Ogrostis capillaris) burial plots. At the begi nning of the trial the increased in abundance. Wi thin the OBS plots, the field consisted of improved grassland with a abundance of marram and fescue was reduced much high stocking rate of grazing sheep. The result further, whilst three grasses, Yorkshire fog, common was a pasture domi nated by perennial ryegrass bent and smooth meadow-grass increased. together (Lol ium perenne). Establishment of the trial with false oat-grass. This indicates a change from dune plots resulted in disturbance to the established vegetation to open, acid grassland. Yorkshire fog, sward and loss of grazing. The vegetation common bent and smooth meadow-grass are typical developing on the plots reflected these early colonists of disturbed sandy soils in general. changes.

Winter burial W inter control

Rumex etosa Rurnex etosa olinm rine

_

, -

11 _7 r 77 r,

Agrostis

-

Summer burial Summer contr ol

nne

Lora n- enne

Agrostis ca

11-4

Fig ure 6.2. Relati ve proportions of diff erent plant species in canny)! and burial p lots at Eskmeals. Only the mare abundant species are named.

4 1 Table 6.2. Relati ve cover-abundance of .sp ecies on No undisturbed vegetation- was monitored. hut 0 135 and canno t plots at Eskmeals. Values vegetation characteristics were assessed on OBS exp ressed as percentage of vegetati on cover and control plots. in p lots. The results show that, although perennial ryegrass W inter Str nmer Winter Summer persisted, common bent was the most abundant species present, as in the case of the disturbed dune burial site. There was no consistent difference in the response of di fferent species to either the presence of OBS in the deposi t or the li min2 o f disturbance. However, winter burial of OBS or control sand produced a more complete vegetation cover after two years. This may have been because winter buri al disturbed any seed present in the seed bank without affecting its capacity to germinate_ whilst summer buri al removed newly germi nated plants and left a dry surface upon which germination was not possible. Nevertheless, by July 1996, summer plots had a vegetation cover in excess of 70%, compared with almost complete cover in the winter plots.

70

60

to 50

ro 15 40

§, 30

° 20

10 2-week plough d ie 4-week plough winter landfarm summer landfarm winter no plough control summer control

Figur e 6.3. Eff ect of landfarnang tr eatments on percentage uf bare ground remaining in plots al Eskineals after three t ears.

42 • •

The etieet of landfarnm g s as different and related T IPP 6.3, .11Tri ii . brute to two factors: d isturbance and retention a l oil pipall Pap with in. and a df. ), di:lit:P( 1p 5 resid ues at the (mrface . Freq uent ploughing V. as qlfield plot at Eshmea l.5 1 9 9 5 . expected to de lay the establishm ent of permane nt ve getatio n and this appears t ( ) heo c bee n the case mea n numbe r in the plots at Eskmeals. Almost co mple te ott side plots inSide plots vegetatio n co ver had established in co ntrol plots Dir ie 168 56 by January 1998 (after three year s). wh ilst in Pa stu re txr iat ploughed plots an average 50 % or the plot was winte r 60.7 47.3 still bare gro und . This overall value has hidde n winter control 150.3 125.7 co nsiderable differences attributable to different sum me r 70.7 23.3 ploughing treatments, with more frequent s umme r cont-ol 29.6 82 ploug hing resulting in less successful Lendfarming colonisation (Fig 6.31. w inte r 60.7 47.3 w inter co ntrol 150.3 125.7 Difference s between blocks were also found but sum me r 20.7 23.3 these we re not co nsiste nt across all treatments. s ummer control 29.6 82 Where no natural du ne or dune pasture soil is returned to the surface, colonisatio n of a d isposal site may need to be managed through seeding or plan ting with appropriate species. Although some :a 450 0 1995 marginal invasion from adjacent co mmunities may 419 400 0 1996 0 occur, this may be slow and irregular, and where 350 To there is a s ignificimt area , may leave a central core 300 e xposed to loss o f sand by wind erosio n. Planting ru 250 with appropriate spec ies will he lp to stabilise bare 200 sand and to enco urage its accumulation rather r? 150 than loss. The results of the planting trial at the 2 loo first Pendine trial site have shown that the ta 50 planting of marram tillers may be used as a 2 s umme r winter s umme r winte r dune success ful method of producin<2 an almost burial buria l control co ntrol buria l complete ve getatio n cover w ithin four years. Sow ing w ith a red fescue seed mix also he lps Figure 6.4. Nitinbets of animals m ight in /99.5 stabilise sand and produce site s for accum ulation. and 1996 in p itfa ll t •aps within trial p lots at Ho we ver the e ffectiveness o f th is sward in Est:weals. trapp ing sand is not as great as the presence of marra m tussoc ks. Dunes appear to support larger populations than A nim al colon isation die dune pasture and the effect of disturbance has At the first Pe ndine experime ntal site, bee n to reduce the number of animals sig nifica ntly. observatio ns suggest that the margins of the large The reduetion in number of animals trapped . fro m a de posit we re colonised by both ants and spide rs sm aller starting base, was less pronounced in dune within the first six mo nths. A ra nge of pa sture burial plots than in landfarming plots. invertebrates was also observed (luring the first summer at the second experime nta l site at Pend ine. As vegetation re-colonised the plots, num bers of invertebrates captured in pitfall traps placed with in Detailed survey s of surface-dwelling invertebra te the plots increased dramatically. The increase was animals have only bee n carried out on the found acros s all types of plot, suggest ing that, at Esk mea ls experimental plots. least in terms of overall number o f animals, the presence of oily resid ues at the sur face in the Resu lts from pitfa ll traps indicates tha t site. land fan n ing p lots was not a factor inhibiting disturbance and a reduction in vegetatio n cover. colonisation . rather than presence of oil residues, ha ve bee n responsible for a reduction in number of anima ls found

43 44 7 CONCLUSIONS

The main conclusions from the project are: 5 Ecological risks are low • Although in experimental situations toxic 1 The disposal method works, because : effects of weathered oils on dune pasture • Micro-organisms present in sandy coastal plants can be demonstrated, such toxicity soils have the capacity to degrade petroleum was not found in fi eld trials. hydrocarbons. • In fi eld trials, disturbance has a greater • There is only limited migration of eff ect in slowing the development of hydrocarbons from OBS to underlying sand. colonising plant communities than the • There is no significant leaching of presence of oil residues. hydrocarbons from OBS deposits to • The number of invertebrate animals is also groundwater. aff ected by disturbance more than the presence of oil residues. 2. The process of degradation proceeds rapidly at fi rst but declines with time . • The pattern of degradation can be expressed most accurately by a power function. • Low molecular weight compounds (with fewer than 10 carbon atoms) are lost from oil residues more quickly than those of higher molecular weight.

3. A number of factors infl uence the rate of degradation. • The type of oil and the degree of weathering infl uence the rate of degradation. • The two main external infl uences on decomposition rate are temperature and soil moisture. • Higher temperature, within the normal climatic range, encourages faster degradation of oil residues. • The presence of a high water table or infi ltration of heavy rainfall inhibits degradation. • Characteristics of the soil also infl uence degradation rate. • Grain size and organic content of soil may affect degradation rate by infl uencing soil water holding capacity • Buffering capacity (observed as calcium carbonate concentration) appear s to be particularly important.

4 Environmental risks are low • Leaching of ions from seawater contained within the OBS is short term and groundwater chemistry quickly returns to its former status. • Hydrocarbon-degrading populations of microbes develop in sand immediately below OBS deposits. • Although polycyclic aromatic hydrocarbons are found in low concentrations within degrading OBS, they do not migrate io groundwater.

45 46 8. IMPLEMENTATION

Although the obj ective of the proj ect was to A rea examine the effi cacy of a natural method for A disposal site needs to be of suffi cient size to degradation of oil residues, the application of the accommodate the OBS to be disposed of and any method to the treatment of the products of real ancillary works associated with such disposal. No spills requires some positive management. This additi onal engineering is envisaged in burial sites should aim to maximise the rate function (other than temporary storage of turves, should controlling decomposition of oil residues their removal and replacement be seen as an contained in OBS. essential part of site rehabil itation).

The technical obj ecti ves may be partly met by Volume appropriate selection of disposal sites and partly Each disposal site, or component of a disposal site, by manipulation of starting conditions. Other should have sufficient available volume to practical considerations, based on factors such accommodate the amount of material to be as location and status of potential sites will also received, without excavation. In practice, for burial need to be taken into account. sites there will need to be, as part of the site specification, a stated limit on the volume of material, which that site, or each component of it 8.1 Site considerations can accept. This would mean that for any site, there might be a range of identified individual areas Habitat typ e an d conservation capable of taking diff erent volumes and so, be designation appropriate for use with diff erent sizes of spill . There would be a presumption against the use of high quality sites for the disposal of OBS. Sites Stability with a nature conservation designation (and Sites where the dunes are accumulating sand will accepting a hierarchy of importance such that provide more acceptable locations than those e.g. sites of international importance would rate where dune erosion is active. In eroding sites, there more highly than those designated for local is a greater possibility of subsequent exposure interest only) would be considered only in before decomposition of hydrocarbons has been extreme circumstances and after other completed. possibil ities have been exhausted. In all cases, a competent agent should carry out an ecological impact assessment. Suitable areas within 8.2 Site factors infl uencing degradation designated sites may be capable of use provided there is no threat to the integrity of those sites The features of sites, which are most likely to from either the burial or associated activity. infl uence degradation, have been identifi ed in the different experiments and trials undertaken in this A ccess proj ect. Any site considered for burial should have suitable access for the type of machinery Water Table Depth required for moving the quantities of material to Soil aeration or the possibil ities of waterlogging be dealt with. Such access should be, for causing anaerobic conditions within any OBS preference, served by a tarmac road or other deposit are key factors in determining site suitable track and should, in all cases avoid suitabil ity. Sites where there is a permanently high direct traversing of active dune systems or other water table, or where fl uctuations in water table areas of high environmental value. If sites of would bring it into close proximity to OBS, are different capacity for burial are defi ned, it may be inherently unsuitable. Sites where the water table is that the access criteria may be modifi ed for less than I m below ground level for all or part of smaller vehicles (e.g. small unattributable spills the year should be rej ected. Sites with a water table may by removed by dumper truck rather than by at I - 2m below OBS level for all or part of the large lorry). This consideration, together with the year should be regarded as probably suitable and possibility of installing temporary trackways those where the water level is below 2m for all or could have consequences for assessment of part of the year should be considered as most accessibility in particularly diffi cult locations. suitable.

47 In most cases of disposal in near-shore locations, oil types and main geoclimatic regions of the water tables will slope towards the sea. However, country in order to optimise degradation rates and where the groundwater slope is inland (as on the end points. landward side of some dune systems) any migration of hydrocarbons under conditions of Thickness of oil layer high water table could present diffi culties for This may be linked to oil concentration. A lower wetland sites. Hence parts of dune systems with concentration of oil in the OBS may permit a included wet slacks should always be avoided. greater thickness of that OBS to be deposited. Options may be related to land confi guration such Climate that a shallow depression may be best suited to a All areas of Britain are likely to be suitable thin layer of more concentrated oil residues (within climatically, but actual rainfall and temperature the limits imposed by penetration of air, water and conditions may need to be taken into account in nutrients) whil st a small , deep hollow may require defi ning management procedures at any particular the deposition of a thick, but less concentrated location. layer of OBS. Once individual locations have been identified, it will be necessary to try to build in OBS type tailored specifi cations. The rate of degradation will depend on the type of OBS to be buried. This will be determined by the Fertiliser Application type of oil and its weathering history, and the grain In naturally nutrient-poor locations (which dunes size and chemistry of sand. A lthough there is no systems are) the additi on of fertiliser is generally to control over this, management prescriptions may be avoided (especially where semi-natural dune be dependent on factors within the OBS. communities are present and their maintenance is to be valued). However, in certain circumstances, Timing of disposal some addition may be recommended for particular Research results indicate that the timing of disposal sand types in particular situations. The effi cacy of infl uences the rate of degradation and that summer nutrient supplements, usually N & P (but possibly and winter incorporation of OBS may lead to including Ca in sands with low Ca concentrations) di fferent patterns of decomposition. It would not has been demonstrated in bioremediation be possible, nor advisable to separate clean-up applications elsewhere. At present there are no operations and disposal (because age of material rules governing the application of fertilisers in can also infl uence breakdown rate). However the dune systems, but minimal applications would be timing of disposal may play an important role in demanded. There are diff erent possibil ities for determining management strategies for the deposit, application techniques (incorporation with OBS, and in infl uencing any proj ections of the rate of the subsequent drilling and fi lling, surface dressing decay process and the time taken to reach any pre- where vegetation is absent) determined fi nal concentrations. Aeration B each poll ution history Under normal circumstances there is not expected A beach with a history of oil pollution is likely to to be any need for aeration. In general sands are have larger populations of micro-organisms well aerated, unless the water table is high. There adapted to hydrocarbon degradation than a beach may only be a need for aeration where high with no such history. Because of this, degradation concentrations of oil are used, or there has been may be expected to begin more quickly at such a some concentration of oil within lower layers of site. Again, there is no control over this, but it may OBS deposits leading to anaerobic conditions. infl uence proj ections of degradation rates. Passive aeration could be achieved by incorporation of porous pipes within OBS or sandwiching layers of OBS and clean sand. Forced 8.3 Management factors aeration requires special engineering solutions and the use of pumps on site. Oil concentration Given that effi cacy of oil degradation is reduced as Revegetation hydrocarbon content rises above 7-8%, it may be The need for specifi c revegetation measures is very necessary to dilute the OBS to an appropriate much dependent on site characteristics and concentration with beach sand before burial. A set disposal method. For burial sites, a standard of optimal concentrations should be set for specifi c recommendation is for removal of surface layers

48 prior to deposition and subsequent return of these reasons of access, security and, public sensitivity, as turves. Any other operations must be seen as certain types of site in public or commercial only "emergency repairs" to be undertaken when ownership may be more suitable than those in problems arise. In the absence of such a turfi ng private ownership or considered as part of the procedure, two diff erent approaches suggest public domain. However, it is hoped that any themselves, depending on site size. For a small problems arising could be resolved by negotiation. site, marginal invasion would accomplish revegetation within a comparati vely short period, Monitoring so alaissez-faire approach is all that is required. Monitoring schemes pose no practical problems, For larger sites, a planting programme would be but do have a cost implication. Any monitoring required, following standard practices for dune scheme needs to obtain baseline data before restoration, or reseedi ng. deposition of OBS. A requirement for any monitoring programme is the setting of clear obj ecti ves. In practice, in the case of burial of 8.4 Public perception OBS, this means establishing pm-defi ned limits to hydrocarbon content which, once met, mean that With all technologies involving materials the site can be decommissioned, as any remaining considered to be potentially capable of damaging materials would no longer be regarded as special the environment or public health, certain waste, and monitoring can cease. precautions need to be taken to ensure that any risk (perceived or actual) is minimised and that the procedures adopted are subj ect to control and scrutiny by appropriate authorities. The main aspects of concern in this context would be:

Publk sensitivity Areas seen as especially sensitive in terms of public opinion would be considered only after examination of the potential of other suitable, but less sensitive, locations in the vicinity. Any operation would need a programme of public education and the public would need to be assured of the low and short term risk element attached to disposal of OBS by burial.

Security Site security has two facets: maintaining any monitoring equipment installed on site, and safety of the public. Sites offering some measure of security for, e.g. water sampli ng wells, would be viewed more favourably than those open to public access. Such sites would also be favoured because of the possible risks of disturbance to the deposit and human contact with OBS (especially early in the life of the deposit). Although such contact is not likely to cause any serious health problems, it should still be considered as a risk to the public. At some locations new fencing may be required. Sites exposed to particular threats or risks of erosion should also be avoided.

Ownership Site ownership would need to be considered carefully, particularly where any special conditions apply to that ownership, or where there are perceived threats to the value of private land. For

49 50 9. FUT URE RESEARCH NEEDS

The current project has established that microbial research should be concentrated on lysimeter degradation of OBS provides a practical answer to experiments, with subsequent fi eld-testing. problems of disposal of oil residues from spills. It has also shown that there are no significant Although we have derived a general function to environmental risks from the release of toxic describe the course of hydrocarbon degradation in compounds during the breakdown process. The buried OBS, the reasons for the decline in main focus of any future research should be breakdown rate with time will be related to some directed at two topics: rate limiting factor. This may related to the substrate itself or to some other, external, limiting • Methods of enhancing the process. This means factor. Once the small molecular weight determining both conditions under which the components of oil residues have been metabolised, initial rate of hydrocarbon break-down can be the process may slow because micro-organisms maximised and investigating possible means by with the capacity for degrading long chains or ring which decomposition rates may be maintained structures are not present in suffi cient numbers in or improved during the latter part of the the soil medium. Alternatively, the decline may process. result from progressive exhaustion of one of the • Investigations to determine availabil ity and major controlling external resources, e.g. oxygen suitability of potential disposal sites around the or nutrients. Research directed at determining British coast. which of the factors may be causing the decline could provide guidance on management of deposits to enhance degradation at an appropriate stage, 9.1 Enhancement of degradation thus ensuring more rapid and more complete decomposition. The second lysimeter experiment indicated that_ there may be an optimal concentration of oil One specific measure which has been suggested as residues in OBS, which would providing a means of enhancing degradation is the mediation of plants — phytoremediation. Some • maximise aerobic degradation plant species are thought to be able to stimulate • minimise the downward movement of hydrocarbon-degrading micro-organisms, hydrocarbons particularly those associated with the rhizosphere, • minimise interference with site hydrology and so promote breakdown of oil residues. Research to follow up this line of enquiry would, The object of future research in this area would be involve pot trials in the fi rst phase and to determine this optimum concentration and any determination of the role of plants in promoting patterns of variation attributable to oil type or hydrocarbon degradation, via stimulation of weathering history. Although initial experiments rhizosphere bacteria and mycorrhizal fungi. Such a might be on a small, lysimeter, scale, they would programme would involve a microbiological also need to be scaled up to fi eld trials to ensure component, quantifying the extent of any such eff ective translation of experimental scale results. stimulation. This would be linked to assessment of differing potential among plants with diff erent As an adjunct to this area of research, it would be morphological or physiological characteristics (e.g. valuable to determine the behaviour of similar grasses, legumes, and non-legume herbs), and starting OBSs in other light-textured soils, e.g. measurement of the eff ectiveness of the plants in sandy loam, sandy clay loam. Th is proposal arises promoting degradation. from suggestions made during the early part of the project that the possibility of including soils other than sands might increase the potential number of 9.2 Site selection sites along parts of the coastline where dunes and dune pasture systems are not present or are Although some indication has been given of the unavailable because of their high conservation questions to be considered in selecting and value. Under this proposal it would be necessary to managing disposal sites, further work is needed to determine both differences in oil residue identify and characterise locations, which might be decomposition rates and possible modifi cation of available to receive oil residues following future oil mobility characteristics. In the fi rst instance such spills. Any work in this fi eld must be carried out in

51 a manner that provides clear guidance on criteria for selection and the rigorous application of guidelines produced. Only wi th a clearly defi ned protocol in place can selection be seen to be based on obj ective assessment.

The site selection exercise needs to be conducted independently of any immediate need arising from a spill in which oil is washed ashore. Only in this way will it be possible to conduct the survey in an independent manner, free from the pressure which a particular incident or situation is bound to provoke.

Choice of potential disposal sites is seen as an exercise which would be a combination of ecological, hydrogeological, environmental conservation and logistic studies, combined with awareness of local social and economic constraints, and the legal requirements associated with disposal of oily waste materials.

52 10. ILL UST RAT IONS

First Pendine ex periment. Bacteria colonies cultured from dune sand, beach sand and OBS .

Pe nd ine

FirS1Pendine experiment. Planting marrarn grass shoots in a trial plot. November 1994.

First Pend ine experiment. Reve getation of trial plots after two years.

53 Eskrneals held trial. Preparing plots withi n dune pasture for receipt al OBS.February 1995.

Eskmeals fi eld tr ial. plough ing oil residues into sand on the artificial beach. February 1995.

Esk meals fi eld trial . Taki ng sequential core samples.

5d Lys imeter fac ility at FEE. Mer lewood.

Extensive study. Removing turves for burial of OB S

Second Pe ndine experiment. M ixin em ulsion, March 1997.

55 56 11. ACKNOWLEDGEMENTS

We gratefully acknowledge the fi nancial support of the Maritime and Coastguard Agency (formerly the Marine Pollution Control Unit of the Coastguard Agency).

In the latter stages of the project the Environment Agency carried out analyses on samples from the second Pendine experiment at their Llanelli laboratory. We are grateful for their assistance and thank Ron West and Anthony Grave11, in particular, for their help.

Without the co-operation of the Ministry of Defence, via DTE0 and its successor DERA , the fi eld trials at Eskmeals and Pendine would not have been possible. We wish to express our grateful thanks to all those at both establishments who so willingly helped us to set up and maintain these experiments.

We also wish to thank all the other landowners and land managers who allowed us to bury bags of OBS for the extensive trial.

Our thanks must also go to County Contractors and Mason Brothers for the care with which they set up the experimental plots at Eskmeals and Pendine respectively. Special thanks go to Bob Denham for his practical good sense and good humour in supervising the OBS preparation at Pendine.

Finally, our thanks to B. Adams, C.J. Barr,N.G Bayfi eld, D.G Benham, H.I.L Black, P.A . Coward, A.J. Dixon, F. Jay, D.K. Lindley, E. Manning, S. McGrorty, M. Misra, J.A. Parkinson, I. Pearce, R.W. Pickup, J.M. Poskitt, W.E. Rispin, J.D. Roberts, S.M.C.R. Robertson, R.J. Rose, M.J. Rossall, D. Seaton, E.C. Waterhouse, D.R. Wilson and J. Wright who were members of the team contributing important planning, fi eld, laboratory and analytical skills.

57 58 12. REFERENCES

Amaral, S.P. (1987). Landfarming oily wastes: Jason, J.D., Gupta, FL, Das, C..M., Nair, S., design and operation. Waier Science Bhar athi, L & Chandram ohan , D. (1995). Technology, 19: 75-86. Candidates for the development of consortia Atlas, 1LM. (1984). Diversity of microbial capable of petroleum hydrocarbon degr adation communities. Advances in Microbial Ecology , in marine environment. Proceedings, 211° 7: 1-48. International Oil Spill Research and Ba rkay, T.M. & Pritchard, H. (1988). Adaptation Development Forum, IM O, London, pp.14-26. of aquatic microbial communities to pollutant Kennicut, M.C.I1 (1988). The effect of stress. Microbiological Sciences, 5: 165- 169. biodegradation on crude oil bulk and molecular Benka-Coker, M.O. & Ekundayo,J.A. (1997). composition. Oil & Chem ical pollution, 4: Applicability of evaluating the ability of 89-112. microbes isolated from an oil spill site to Knowlton, R E. & Rucker, J.E. (1979). degrade oil Environmental Monitoring and Landfarming shows promise for refi nery waste Assessment, 45: 259-272. disposal. Oil and Gas Journal, 77: 108-116. Bleckmann, CA ., Oxley, M E., Wilson, E.J., Hayes, Leahy, J.G. & Colwell, R.R. (1990). Microbial K.W. & Heryck, N.L. (1997). Land treatment of degradation of hydrocarbons in the produced oily sand: field results. Waste environment. Microbiological Reviews, 54: Management & Research, 15: 223-237. 305-315. Blen kinsopp, S, Sergy, G , Wang, Z., Fingas, M.F., Mearns, A.J. (1997). Cleaning oiled shores: Foght, J . & Wes tla ke, D.W.S. (1995). Oil spill putting bioremediation to the test. Spill Science bioremediation agents - Canadian efficacy test & Technology Bulletin, 4: 209-2 17. protocols. Proceedings of the 1995 Oil Spill Oudot, J ., Merlin, F.X. & Pinvidic, P. (1998). Conference: achieving and maintaining Weathering rates of oil components in a preparedness, American Petroleum Institute, bioremediation experiment in estuarine Washington D.C., pp. 91-96. sediments. Marine Environmental Research, Bragg, J.R., Prince, R.C., Harner, E.J. & Atlas, 45: 113-125. R.M. (1994). Effectiveness of bioremediation Padmanabhan, R, S hanker, it & Khanna, P. for the Exxon Valdez oil spill. Nature, 368: 4 13- (1998). A method for extraction of DNA and 4 18. PCR-based detection of polycyclic aromatic Chapelle, EH. (1999). Bioremediation of petroleum hydrocarbon-degrading bacteria in soil hydrocarbon-contaminated ground water: the contaminated with oil and grease World perspectives of history and hydrology. Ground Journal of Microbiology and Biotechnology , Water,37: 122-132. 14: 925-926. Coates, J.D., Anderson, R.T., Woodward, J.C., Stencel, J.R. & Jaffe, P.R. (1998). Trace metal Phillips, E.J.P. & Lovley, D.R. (1996). analysis using ion chromatography in water Anaerobic hydrocarbon degradation in partitioned from crude oils. In: Garcia-Martinez, petroleum-contaminated harbor sediments R. & Brebbia, C.A. (eds.), Oil and hydrocarbon under sulphate-reducing and artificially spills, modelling, analysis and control. imposed iron-reducing conditions. Southampton, WIT Press. pp. 223-229. Environmental Science and Technology, 30: Swan nell, R.P.J , Tookey, D. & McDonagh, M. 2784-2789. (1993). Bioremediation of oil sp ills: a state of Ellis, R.E. & Adams, R.S. (1961). Contamination of the an review. Publication CR 3807, Warre n soils by petroleum hydrocarbons. Advances in Spring Laboratory, Stevenage, UK. Agronomy, 13: 197-216. Venosa, A.D., Suidan, M.T., Wrenn, B.A., Gray, A.J. (1985). Poole Harbour: ecological Strohmeier, K.L., Haines, J.R., Eberhart, B.L., sensitivity analysis of the shoreline. King, D. & Holder, E. (1996). Bioremediation of Huntingdon, Institute of Terrestrial Ecology. an experimental oil spill on the shoreline of Gundlach, E.R. & Hayes, M.O. (1978).. Delaware Bay. Environmental Science and Vulnerability of coastal environments to oil spill Technology, 30: 1764- 1775. impacts. Marine Technology Society Journal, 12: 18-27.

59 - á á