A leak from an underground fuel pipeline in South Gloucestershire led to serious land contamination and to two families being moved out of their homes, Gerard Madden reports on the clean-up operation and lessons learned

On Tuesday 19 May 1998, at 1.30pm, an underground fuel pipeline was damaged by contractors during planned maintenance work to a valve on the 250mm diameter pipe. The incident occurred on land forming part of an old depot at Patchway, Bristol, within the South Gloucestershire boundary. Significantly the line was not under pumping pressure at the time of the incident; kerosene fuel was, however, released due to hydrostatic head in the line.
Approximately 27,000 litres of oil were recovered from the area of the spill shortly following the incident. Calculations by the company responsible for the pipeline, Serco Gulf Engineering, suggest that only a limited volume was lost to ground, possibly between 500 and 900 litres. The fuel lost was a JP8-type kerosene without additives.
The kerosene spread over an area approximately 40m by 20m around the point of release and was also spread beneath the boundary fence with the adjoining residential property (House A) spreading over parts of the garden including the lawn, patio and garage.

Initial response: Police and fire brigade attended. The fire brigade set up an initial cordon and advised three households to evacuate, SGE initiated a clean-up using tankers and a JCB to dig out visually contaminated soil in the depot and lawn (this continued until Thursday 21 May).
The police and fire brigade left in the evening once the latter was satisfied that all surface liquids (aviation fuel) had been collected by tanker and any risk of ignition had been removed.
The Environment Agency was then called out and it contacted South Gloucestershire Environmental Services. A South Gloucestershire emergency management officer attended the site and an environmental health officer was in telephone contact by 6.00pm on the first day. It was agreed a site visit would be made the next day.

Second day: An EHO visited the site and met the residents of House A and established they had been rehoused overnight in a hotel. evidence of spill was noted. SGE continued to remove visibly contaminated soil and the area reeked of kerosene. The EHO contacted the Environment Agency and Bristol City Council's Scientific Services and established that Wessex Water and Bristol Water were on site the previous day. With the explosive risk dealt with, the next phase was the physical clean-up operation.
Second phase
Residents of a property 16 houses away from the spill site (House B) reported a strong smell of kerosene in their property and a pool of fuel was found under the floorboards. The details were relayed to South Gloucestershire Emergency Service by the fire service at 1.03am on 21 May with a request for an EHO to visit the following day.

Third day onwards: An EHO revisited, viewed House B and met SGE. Health-based advice was sought from Avon Area Health who in turn had access to the Chemical Incident Response Service (Guys' Hospital). More information was required before a definitive medical opinion could be given.
The occupiers of House B were put in a hotel. SGE tried to get ventilation equipment and the clean-up continued. The pathway from the spill to House B was investigated with the service and drainage pipes being looked at and holes being dug in the front and rear gardens of House B.
At a site meeting it was agreed that Scientific Services (of Bristol CC and contracted by South Glos) would undertake sampling and SGE would hire Dames and Moore as environmental consultants, although they could not attend until the following Tuesday.
An EHO visited all properties in the row. At this time we still had no exact figures on the amount of kerosene lost.

Fourth day: On Friday a further site meeting was held and a Second Phase Strategy was agreed between EHOs, doctors from area health and SGE, based on the available information including results from Scientific Services on indoor fume levels. The strategy set down the following:

• all residents to be briefed by doctors
• if health problems, call doctors
• if odours, call the SGE and EHOs.
• if odours/symptoms do occur, SGE to rehouse as a precaution
• occupants of Houses A and B to be rehoused.
• SGE to continue with remediation.

The assistant director of environmental services, emergency call-out EHO and ward councillors were briefed.
The strategy saw us through the first few weeks and was confirmed in writing to all the residents in the row with a series of contact telephone numbers.
Health advice
The first source of information was from Sax's Dangerous Properties of Industrial Materials. It gave an occupational exposure level of - 100mg m3 TWA (time weighted average). In addition, the safety profile indicated it was a suspected carcinogen, a severe skin irritant and could cause somnolence, hallucinations and distorted perceptions.
SGE's product data sheet contained general information with a TWA of 5mg m3 but as an oil mist. It also added exposure to high concentrations of vapour for short periods can result in anaesthesia-like symptoms, and low concentration, if repeated, may cause dizziness, headache and nausea.
Direct contact with the CIRS revealed the need to be very cautious. People can inhale the vapour and be unaware they are affected, which could, for example, lead to a driving accident. In addition, if a room is sealed and concentrations build up, there is potential for acute poisoning and a knock-down effect. It was clearly a health risk.

A definitive medical opinion was not initially available due to the make-up of kerosene. Further information was required, namely: the numbers of people affected; the exact composition of the fluid; and the extent of seepage.
As time went on information became available and after research by the consultants and consideration by Avon Health Authority with the CIRS the following was agreed:
a) For air - the risk assessment methodology was accepted as standard practice* and airborne levels of 10mg m3 for the short term and 1mg m3 for the long term were set.
b) For land - risk based clean-up levels (RBCLs) were agreed (see table on page 48).
Remediation strategy

The objective of the strategy to remediate the contaminated soil was to reduce the contamination to "risk based clean-up levels" and also minimise odours.
The contamination of the ground had occurred through both surface and subsurface flow, so the clean-up needed to address contamination at depth as well as on the surface.
After the emergency services had stepped down, the company began excavating the spill area, taking contaminated material to a licensed landfill. Consultants acting for the company then began assessing the extent of ground contamination through boreholes and soil samples.
As the full extent of the incident and the spread of kerosene to House B became apparent, the remediation plan became more complicated. It was discovered that a number of old land drains ran parallel to Gloucester Road, through the gardens and under the properties. One of these in particular had provided an easy conduit for the kerosene to House B, where it had backed up due to a blockage in the drain. House B was now the most affected property and became the priority for remediation.
The remediation at House B started by basically gutting the property. All household items were removed to storage and carpets, curtains and floorboards were removed. Porous pipes were then sunk across the front living room and front bedroom (the houses are bungalows). Attached to these was a vacuum extraction system.

Also the land drain running under House B was excavated at each side of the property and a vacuum extraction system installed across the drain.
The two systems - the shallow and deeper vacuum extraction systems - were operated alternately and have the ability to push air or pull air through the system. The kerosene vapour extracted was collected through a system of centrifuges - to remove water vapour - and then absorbed onto carbon filters. The quantity of recovered product was then measured so that a graph of accumulated recovered product could be plotted against time. For example: 2.5kg TPH - recovered over 450 hours in the period 12 June 1998 to 24 August 1998.

The system was in place and the extraction continued until the accumulated product graph levelled out, ie when the quantity of product recovered became negligible. A similar system was also installed at House A and for a time extraction was alternated between the two locations. In addition to vacuum extraction, a process involving biodegradation of the kerosene was also used. The area in the vicinity of the spill, the length of the land drain and the ground beneath 141, were inoculated with a solution containing bacteria aimed at degrading the kerosene, mixed with an oxygen enriching compound, which basically keeps the bugs alive as the process is aerobic.

Extensive sampling continued to check both the extent of spread of the kerosene and the progress of the remediation. Final remediation sampling was undertaken once the vacuum extraction system had been removed and after a period of time a steady state was achieved.

Ongoing activities
The monitoring results obtained at each stage dictated the progress of the remediation works on a rolling programme basis. A closing strategy was agreed covering the targets to be achieved and a monitoring protocol extending to one year post-reoccupation and a further year when the company would investigate any unforeseen problems reasonably attributable to the original leak.

Issues raised
The lack of any definitive safe level was a difficulty. Often the chemical is unknown or there is no method readily available, but here we could measure levels but not relate them to "acceptable" levels in homes.
Differing reports of the quantity lost gave concern but it is difficult to calculate the volume of fuel in a pipe as it is not level and has varying quantities of space at the top.
Conflicting advice from two consultants also perturbed residents and ourselves and at one stage, there were two sets of remedial equipment on site. If this had continued, the problems of identifying the person responsible, should the remediation have failed, could have been serious.

Lessons learned
An appreciation of the seriousness of these incidents can only be fully appreciated by meeting those affected on site. There is a feeling of sheer devastation when a home is gutted, residents are forced out and the future is not certain.
The on-site meeting with doctors from Avon Area Health was extremely useful and led to the adoption of a pragmatic approach in the early stages. However a lack of communication between South Glos, Avon Health and the CIRS meant that the offer from the CIRS to attend site was not taken up. A wash-up meeting involving the three organisations has sought to address this point for future incidents.
Having monitoring undertaken independently by Scientific Services (Bristol) was useful to reassure residents, and also to fill in before the arrival of the environmental consultants.
The residents banding together and getting legal representation was very useful. While at some levels it is easier to deal with individuals directly, in this case the number of enquiries greatly diminished once they joined together and, in particular, elected a residents' representative.

A residents' meeting was organised by SGE with invited representatives only. It was an ideal opportunity to address the residents' fears and concerns and, as it was not an open public meeting, it remained organised. The company summarised events and answered questions previously forwarded. In addition, we had an opportunity to explain the position of South Glos. In this instance, it was considered the councils' role was to ensure the properties were safe in terms of health and left free from nuisance, both aspects as covered by the Environmental Protection Act 1990.

The use of a property in the vicinity which was not affected, as a background site, helped in the understanding that kerosene-free may not be practicable. South Glos took a lead in ensuring that written information was given to residents, including a range of contact numbers, events and proposals for action. This was helpful in the absence of any information direct from the company; it reassured residents that the "independent" council was involved and it helped address later allegations that some felt they hadn't been kept informed. It also served to distill current events and have actions agreed in writing.
Advice for others in a similar case would be that if it can't be established that houses are safe, temporarily rehouse occupants immediately. Have a procedure for people to report further problems and actions to be taken. Establish that steps will need to be taken to determine the safety of properties and clarify responsibilities.

Overall, the lasting impression was the all-consuming nature of the case. For the first two weeks, it took 100 per cent of two officers' time and, for the next four weeks, approximately 80 per cent for one and 60 per cent for the other. It is essential for at least two officers to be involved, not just to cover leave but to deal with the mass of information and allegations, and to keep a sense of perspective.

Outcomes
This case has already been used in training for EHOs in-house and in the region. The lessons learned have also been incorporated in the South Glos Inland Pollution Hazards Plan. Roles of relevant organisations, access to health-based advice and lines of communication have been clarified. The company concerned, which has been co-operative throughout, has been advised to revisit its incident planning. There is still disagreement over setting health-based levels of contamination.

Gerard Madden is team leader on pollution at South Glos. These views are his own and may be altered by subsequent findings. He can be contacted on tel: 01454 863569; e-mail: gerard_madden@southglos.gov.vk