When a road traffic accident swiftly turns into a complex contaminated land investigation, what role does the local environmental health department take? Marc Willimont describes Herefordshire Council's experience

At 5.30pm on 5 September 1997, a road tanker lost control and smashed through a crash barrier on the A49, between Hereford and Leominster, in Herefordshire. It rolled down the hillside, finally coming to rest among trees directly above a bungalow in the small hamlet of Dinmore. The tanker, which was carrying 21,000 litres of formaldehyde, was leaking badly and a huge cloud of formaldehyde vapour swept across the nearby houses. The local environmental health department has been tasked with leading this case in the three years since the accident.

the emergency
The accident was immediately attended by the emergency services and the A49 trunk road was closed for 24 hours. No-one was seriously injured, nevertheless, several residents and emergency personnel had to be admitted to hospital suffering from acute respiratory effects associated with the inhalation of formaldehyde vapour. The county council's emergency planning officer was immediately called out, who then contacted environmental health and the Environment Agency. Assistance was requested from Herefordshire Health Authority's Consultant in Communicable Disease Control (CCDC) and Birmingham University's Chemical Hazard Management & Research Centre (CHMRC).

While the emergency services dealt with the immediate problems, it soon became clear that the incident had occurred at an incredibly problematic site:

the site was adjacent to three residential properties, and the entire hamlet of some 14 properties had to be immediately evacuated;

• the incident occurred beside the border of two district councils, Leominster and South Herefordshire, which were in the middle of becoming a new unitary authority;
• the spillage was on a very steep and wooded site, making both the recovery of the vehicle painfully slow and the later remediation inherently complex;
• the steep slope below the road had a recent history of landslides, sections of which had been back-filled with stone and land drains. This instability would not only limit future remediation options, but the drainage added uncertainty to the behaviour of groundwater migration;
• the spill site was adjacent to the Queenswood Country Park, making public access an immediate issue;
• the site was within 200m of the river Lugg, an important Site of Special Scientific Interest (SSSI) and now designated a Special Area of Conservation (SAC) within Europe. Surface run-off and the groundwater migration had serious implications;
• a railway tunnel passed through the hillside nearby, which complicated matters until the direction of groundwater flow was established;
• council records showed that the spill site was surrounded by private water supplies. There were two boreholes and a shallow well situated downhill from the site. A number of houses were also fed from springs above the spillage, but the routes of their underground supply pipes were unknown; and
• to complicate matters even further, a spring resurging at the bottom of the hill was known to supply a distant hamlet and an organic farm. The owners of this farm were not surprisingly very concerned about their organic status.

After the initial emergency situation passed and the vehicle was removed, the tanker company was advised to contact its insurers and appoint a specialist contractor to investigate the contamination problems and effect its remediation. The case now became a complex contaminated land investigation with massive insurance and liability implications.

formaldehyde
Formaldehyde is a colourless chemical with a very strong odour, that has many industrial, commercial and agricultural applications. The CCDC and Birmingham University's CHMRC advised that as well as the acute respiratory effects already observed, long-term exposure could be carcinogenic. Although formaldehyde is a very well understood chemical, it became obvious that its spillage beyond the industrial environment was unheard of. There was little or no documentation relating to its effect within the natural environment.

the working group
A meeting of all the agencies was essential. Representatives for the tanker company and the chemical manufacturer were also requested to attend. Loss adjusters represented the tanker's insurance company and became a major influence on the financial aspects and hence the outcome of the case. The first meeting was held on the Monday morning after the event. Initially the group met daily, as a reactive group, dealing with the aftermath. Later, it met on a regular basis, seconding experts and consultants according to need. Because of the nature of the group and the sensitive issues involved, the meeting was not open to the public. However, from the very first day all parties recognised the need for the group to be totally transparent and open. Residents were therefore notified, in advance, of the dates and attendees of the meetings.

Herefordshire Council was seen by the residents to be the most appropriate body to co-ordinate this meeting and advise them of its outcome. The council was local and already had a detailed knowledge and understanding of the site, its residents and its problems. It released regular newsletters and information sheets to the residents. Because of the sensitive nature of the case it was agreed that information released to the media would have to be carefully balanced with the wishes of the residents. Legal access to environmental information became an issue.

the aftermath
The householders most affected by the accident, whose bungalow was almost directly below the crash site, had moved in only two days beforehand. They had been in the garden when the accident occurred and had reported seeing "rivers of chemicals" wash over their land. Other residents had reported suffering from the effects of a "vapour cloud". Hospital admissions supported the latter and the CCDC informed all GPs in the area of the event and of the probable symptoms.

Initial investigations could not substantiate that the formaldehyde had left the site, with no obvious contamination of the river or road drains. It had to be presumed that the majority of the spillage had remained in situ below the tanker, even though the pungent odour associated with a formaldehyde spill was far less than expected. It was agreed that the suspected location should be covered with plastic sheeting to minimise vapour release and to reduce the ingress of rain water, which would accelerate migration. A crescent shaped ditch was excavated in the garden below the spill, to inhibit migration in the soil and to act as a collection point.

Passive formaldehyde tubes were rapidly purchased by the council, as well as personal formaldehyde dose meters. Although primarily designed for health and safety monitoring, they proved invaluable as an immediate method of air measurement. All the houses within the hamlet were tested and found to reveal no presence of formaldehyde vapour. A public meeting was held and the council's environmental health team informed the residents of the investigations to date and their conclusions. The residents were offered an air test upon their return.

The residents of the bungalow immediately below the spill site were advised not to return because of its close proximity and the unknown extent of ground contamination. The family never returned and the house remained unoccupied for three years.

water supplies and crop sampling
A database of all private water supplies had been made and people were requested not to use them as bottled water would be provided until further notice, underwritten by the insurers. This had major implications, as a farm and a public house were included. MAFF was advised and a series of crop and fruit samples were taken. Although the results did not reveal contamination, it is understood that some compensation was paid.

Sampling of the private water supplies commenced on a daily basis. The samples were tested for formaldehyde and its natural breakdown products, methanol and formic acid. Initially environmental health and the Environment Agency shared this duty, with the cost being passed to the insurers. After a matter of days it became apparent that formaldehyde had not yet migrated into the water supplies and this remained to be the case.

hydrology
The loss adjusters agreed to appoint specialist hydrologists to work closely with the Environment Agency. Several new boreholes were sunk at the perimeter of the site to assess dispersion and ascertain the direction of groundwater migration. After a number of months, the Environment Agency accepted that the direction of flow was perpendicular to the slope, towards a spring serving a distant hamlet and organic farm. Confirmation was obtained that the other water supplies were not at risk and these were reinstated.

Soil sampling
The loss adjusters also appointed specialist environmental contractors to accurately survey the site using an agreed grid pattern. Ninety-eight surface samples, six trial pits and six window samples were taken at this initial stage. These samples were recorded at various depths and revealed that the contamination was still in situ. A hotspot of contamination as high as 45,000 mg/kg was found with a mean concentration of 24,000 mg/kg within the top 25cm of soil. The survey found that the majority of the contamination lay within 1.5m of the surface. Over the months following, this sampling methodology was repeated on a number of other occasions.

air sampling
Following initial environmental health air sampling, the loss adjusters agreed to carry out their own tests. Initially this constituted passive diffusion tubes which showed that concentrations were high adjacent to the site, but undetectable at the nearest properties. However, this simple technique would not show peaks of high vapour release. Detailed DNPH tube and GC/MS air monitoring was therefore undertaken. This was repeated on a number of occasions, particularly during warm still weather when vapour release would be at its greatest.

clean up targets
Water: as with all contaminated land cases, lengthy discussion and debate on acceptable contamination clean-up levels ensued. There was no maximum admissible concentration (MAC) available for use, nor any reliable data relating to normal background levels. At first formaldehyde was not found, but when detection limits were lowered, fluctuating levels were detected. Sampling of unrelated Herefordshire supplies revealed that this was a normal background occurrence. Consultants proposed 0.9 mg/litre as an acceptable MAC, which was derived from research and relevant papers. Following independent tests and scrutiny this was accepted and could be fed into a risk assessment process.

Air: air quality standards only existed for health and safety, being based on eight-hour time weighted averages. Clearly, this was not appropriate for long-term exposure. World Health Organisation thresholds for formaldehyde detection and irritation were proposed and the irritation limit of 0.08 ppm was eventually accepted. This standard could be used in a risk assessment process and as an objective measure of statutory nuisance.

Soil: because of public access, an acceptable level of formaldehyde within the soil had to be agreed, based on possible direct and indirect ingestion of soil. Again, no relevant standards existed in the ICRCL, Dutch, USEPA or other conventional bodies of guidance. Consultants were therefore asked to use a worst case scenario of a child eating contaminated surface soil. A table of levels proposed for formaldehyde, formic acid and methanol was presented to the working group for different depths. This was eventually accepted and fed into the risk assessment process.

the community
There now followed a period of detailed risk assessment and supporting research, independently commissioned by the insurers. Two of the UK's leading consultants undertook this task, one succeeding the other. A balance between a quick, safe and effective solution was desperately sought against the backdrop of an increasingly unsettled community. Many residents were still using bottled water and the loss adjusters seemed reluctant to resolve this issue until all their investigations had been concluded. Residents were also intermittently complaining of intense odours, especially during the summer months. The adjacent property was still empty and the owners very frustrated with the perceived lack of progress. Both they and a number of other neighbours had sought legal advice and were requesting an immediate settlement.

the legal situation
The local authority considered the legal aspects of the case in depth. Closure of the water supply was possible. However, it was agreed that this was unnecessary because all householders had accepted the supplies as potentially unfit. After the first few months, only the hamlet supplied from the spring was at risk and the insurers were negotiating a new supply.

The vapours were a problem in the summer and an abatement notice under section 80 of the Environmental Protection Act was considered. The insurers responded swiftly, erecting a polythene and scaffold enclosure, vented via activated carbon filters, before service of a notice was necessary. Assessment and measurement concluded that service of the notice was no longer required. A security fence was also erected at this time.

Tempers were getting frayed locally, as remediation had still not been undertaken. From the beginning, the local authority had requested the insurer's consultants to undertake a risk assessment and enact a strategy to abate the statutory nuisance(s) caused by the spillage. The insurers were aware of the duties soon to be imposed under the impending contaminated land regime and they genuinely appeared anxious to effect a speedy solution.

Although the council was under tremendous pressure, it was also conscious that matters were still progressing and it was wary of upsetting a very delicate situation. The Environment Agency had the option of action under section 85 of the Water Resources Act 1985, because of the pollution of groundwater. However, an informal solution was also preferred for the same reasons.

the risk assessment process
A series of detailed site investigations and risk assessments was carried out.

Remediation method 1 - excavation:
The first undertook a mass balance exercise, using soil sample results to approximate the whereabouts of the spilt 21,000 litres. It revealed that by coincidence, the tanker had crashed onto an area that had suffered a landslide and was back-filled with stone. As this was surrounded with clay, the formaldehyde had been encapsulated like a pond. Using the agreed clean up targets for air, water and soil, the case was modelled to find the maximum concentration of formaldehyde that could remain if all these targets were to be met. Worst case scenarios were required and the principles of "source - pathway - receptor" and "suitable for use" deployed.
The first risk assessment concluded that there was a significant risk to both human and ecological receptors. The soil assessment had led to the tightest clean-up standard. This report went on further to recommend:
- removal of the top 1.5m of soil; and
- agreement of a long-term monitoring strategy.
In preparation, the site was cleared of the now dead trees and vegetation and a wide swathe was cut down the hillside for a conveyor system and for strip excavation/back-filling.

During the wet Christmas period of 1997/98, independent sampling by the insurers claimed that 70 per cent of the formaldehyde had leached from the site following heavy rains. It was claimed that remediation was no longer relevant. When these results were scrutinised, it was found that a different laboratory had been used. Reluctantly, the insurers agreed to resample the site, sending one portion of each sample to the original laboratory and the other to the new laboratory. A third, independent, laboratory also analysed the samples and provided the others with samples "spiked" with a formaldehyde concentration known only to them. The results revealed that the first laboratory was by far the most accurate, showing that between 50 and 80 per cent of the formaldehyde still lay in situ. The original remediation proposal therefore still held.

Remediation method 2 - hydrogen peroxide:
By the time this dispute had concluded, the insurers had withdrawn their excavation proposal due to anticipated technical difficulties, but not cost. Instead, they now proposed chemical remediation. A revised strategy was submitted which recommended pumping hydrogen peroxide into the top of the slope to oxidise the formaldehyde, breaking it down into harmless substances. Laboratory tests were already underway to determine the correct concentrations required and to examine any detrimental effects that might be caused. After a few weeks it became apparent that this option was not viable, as the hydrogen peroxide would weaken the soil structure, creating a very real risk of landslide.

Remediation method 3 - soil washing:
By now, almost a year had passed. Another consultancy was brought in and a strategy was submitted for soil washing. This technique involved the injection of water into the top of the site to mobilise the formaldehyde. The resultant leachate would be sucked out at the bottom and then pumped down to a storage tank to await disposal off site. Not surprisingly, this clean-up method raised even more objections because of the increased risk of landslide.

Remediation method 4 - soil washing by containment of natural water:
Finally, the strategy was resubmitted, proposing a unique and site specific solution for this troublesome case. It was called "soil washing by containment of natural water". This entailed the sucking out of natural leachate at the base of the slope, caused by rainfall and groundwater movement. During dry periods, agreed quantities of water would be added at the top of the slope, but only to the extent of the historic measured rainfalls locally recorded. Tests were held in the autumn of 1998 and proved successful.
In keeping with the council's open approach, another public meeting was held and was well received.
The polythene vapour enclosure was removed, 90 well-points were installed and remediation finally commenced in early 1999. This continued for four months and samples were then taken to determine its effectiveness. The results showed that the clean-up targets originally set had been met. The site was fertilised and seeded and replacement vegetation soon established itself.
A comprehensive completion report was requested.

independent testing
Herefordshire Council undertook a series of independent air tests during the warm autumn months, using air pumps and DNPH tubes for analysis by GC/MS. The sampling strategy was drawn up in conjunction with Birmingham University and the CCDC. Twenty-two soil samples were also taken at different depths at known hotspot locations. Water samples continued as before. All these results were considered jointly by the council, the Environment Agency and the CCDC and were found to be in keeping with the conclusions of the completion report. As the CCDC were advised by Birmingham University, the public was reassured that this had been independently reviewed. The completion report was accepted and the Environment Agency agreed a long-term water monitoring programme. All relevant parties were notified and a final newsletter issued in early 2000.

conclusions
- an accident of this nature has to be planned for and emergency procedures and contact lists need to be drawn up;
- in a multi-disciplinary case such as this, the environmental health department is well suited to lead its resolution;
- the establishment of an expert working group will pool resources and facilitate quick decision making;
- regular newsletters can maintain a difficult balance between the need to know and the rights of the residents;
- community involvement, residents' meetings and good public relations can prove to be a useful asset;
- joint working between the local authorities and the Environment Agency is necessary;
- use of an independent consultant, or in this case Birmingham University, should reassure the public of an open and fair-handed approach;
- one accredited laboratory only should be used;
- risk assessments must be site specific and scrutinised by the relevant agencies;
- independent sampling by the regulators on completion may be required in an incident of this scale;
- formal action must be considered, but may not be the best option; and
- the insurers and their loss adjusters are a key stakeholder in this type of incident.

The author would like to acknowledge Bruce Chartres and Andrew Tector, who assisted in the preparation of this article. Readers may be interested in purchasing the WHO Environmental Health Pamphlets on contaminated land, as well as town planning and housing issues, available from the Publications Department for £7.60 each. Tel: 020 7827 5882.