This article was recently published in the December 2022 edition of the Environmental Scientist quarterly by the Institute of Environmental Science. The article’s authors were Conor Armstrong (Director at Avada), and Adam Bamford (Senior Environmental Scientist at Avada)
Assessing risk following a domestic oil spill is often problematic. There are various receptors in a typical property that are affected in different ways by hydrocarbon contamination. Therefore, it is crucial that an appropriate sampling methodology is implemented from the outset to ensure an accurate contaminant plume is identified before remediation can proceed. However, there are often limitations to sampling in certain locations, compounded by a lack of relevant assessment criteria for many of the affected receptors and the desire by stakeholders (often a home insurer) to minimise costs by limiting the extent of testing where possible.
This case study outlines the remediation of a property close to Enniskillen in County Fermanagh, Northern Ireland (see Figures 1 and 2).
The property sits on an elevated site surrounded by agricultural land, with a watercourse at the bottom of the slope. An on-site septic tank is used for wastewater, and storm and rainwater are piped to the nearby watercourse.
In the centre of the house, an oil-fired AGA cooker (see Figure 3) was used to provide domestic heat and hot water, fed from an oil storage tank to the rear of the property that contained approximately 1,300 litres of kerosene. A braided metal flexible hose immediately behind the range failed causing a substantial volume of kerosene to escape into the property.
When assessing such a situation, the following receptors are considered as part of a conceptual site model (CSM):
• Human health;
• Buildings and other structures; • Services;
• Controlled waters; and
• Third-party impacts.
To assess the risk to each receptor, various testing methodologies are utilised.
An initial on-site survey of the internal air quality was tested in the first instance using a handheld photoionisation detector (PID) that measures volatile organic compounds (VOC) originating from kerosene. However, this device can be sensitive to other compounds commonly found in a domestic setting such as air fresheners, perfumes, detergents and polishes. VOCs deriving from these can occasionally present false positive readings. The wide spectrum of compounds to which the PID is sensitive can make it difficult to obtain a single screening value above which an unacceptable risk to occupants exists. Assessors will often set different thresholds, leading to conflicting advice over whether immediate intervention is necessary.
A more robust air quality analysis can be undertaken in a laboratory. When the contaminants of concern are hydrocarbons, analysis that provides speciated – or grouped based on boiling points – aliphatic and aromatic bandings, total VOCs, and benzene, toluene, ethylbenzene and xylene (BTEX) internal air concentrations should be obtained. However, this analysis can take time, so the initial PID readings are used as an interim measure.
Using guidance on suitability for use from Land Quality Management and the Chartered Institute of Environmental Health – known as LQM/CIEH S4ULs – in-house standards are derived for each of the speciated bands and BTEX compounds reported on by the laboratory. This allows for greater certainty in the risk assessment. Of note, however, is the additive effect of the different hydrocarbon bands and the use of hazard quotients and a total hazard index to assess risk. While this is mentioned in the LQM/CIEH S4ULs and expressed in more detail in guidance issued by the Environment Agency, the additive effect of individual bands or compounds is often overlooked.
Assessing the risks stemming from soil contamination can be fraught with difficulties. Stakeholders often strive for the most economical testing method toevaluate the risk posed by hydrocarbon contamination. The use of probe holes is often encouraged and can be undertaken by a handheld drill fitted with a 1m bit. After digging a hole, the PID nozzle is inserted and a reading obtained. The limitations with this approach are that it is impossible to determine which soil horizon gave rise to a positive reading. It is likely that contamination closer to the surface will cross contaminate any clean vapours from deeper within the probe hole or vice versa. In addition, the various soil type characteristics can give differing readings. Finally, as previously noted, the PID will respond to VOCs other than kerosene making it difficult to identify the source. Furthermore, any shallow groundwater present in the probe hole can suppress the on-site VOC concentration.
A more robust approach is to use boreholes or trial pits. These enable the retrieval of physical samples for a visual assessment and accurate soil characterisation and determination. Samples can then be sent to a laboratory for contamination analysis using the Total Petroleum Hydrocarbons Criteria Working Group analysis, which is a method of dividing a blended hydrocarbon into fractions based on their boiling point by means of gas chromatography. (While the BTEX analysis is similar, it is limited to the four discrete compounds it is designed to measure.) Different levels of toxicity attach to the various fractions. This enables a comparison with generic assessment criteria such as the LQM/CIEH S4ULs. It is important that this analysis considers any exceedances in individual bands as well as takes into account the potential additive effects and relevant pathways, and the LQM/CIEH S4ULs offers guidance on this. Relevant pathways are important, as the risks arising from contamination 1 m below ground are substantially different to those on the surface, even from the same contaminant.
The property was surrounded by hardstanding with drainage gullies and roof downpipes feeding into a culvert that exited downhill at a nearby river. Land to the rear of the property was agricultural and uphill. Precipitation and groundwater flowed under the property and discharged into the watercourse below.
When groundwater under a property is contaminated with hydrocarbons, that groundwater will release VOCs, which could then contaminate a property’s indoor air. The concentration of contamination in the groundwater and the depth to that groundwater are key factors in assessing this risk. Contamination was being carried by the groundwater and storm system drainage network into the nearby watercourse. Emergency containment measures were installed (see Figure 4), and the Northern Ireland Environment Agency was notified of the incident.
The assessment of risk to structures is not straightforward because the effects of contamination are not homogenous across the various building features. The main elements considered are masonry and concrete, insulation, membranes and damp courses, which can have an indirect effect on other receptors such as human health, utilities and controlled waters. There is scant information available regarding the impact of hydrocarbon contamination on concrete, and the LQM/CIEH S4ULs are not appropriate since these only apply to soils.
Most damp-proof membranes in domestic properties are made from polyethylene (also known as polythene) and act as a barrier for water coming up from below ground. This material has an extremely poor chemical resistance to gasoline-range organics (GRO) and diesel-range organics (DRO). Any contact with kerosene will likely compromise the membrane, causing it to fail. Similarly, damp-proof courses – which are built into walls to block rising damp – are also often also made of polyethylene. While they tend to fare slightly better than membranes because they are thicker, they will ultimately fail if there is direct contact with the contaminant.
Insulation can come in many types. Polystyrene sheets have exceptionally poor resistance to GRO and DRO, including kerosene. Bead insulation is generally made from polystyrene and also reacts poorly. High-density insulation is more resistant but is difficult to sample, as the frictional heat of any drilling or coring equipment can cause the release of VOCs, potentially leading to false positive readings even if there is no contamination. Expanding foam (polyurethane) is, however, quite resistant to kerosene. Yet, even resistant insulation can be problematic since it can absorb kerosene and act as a continuing source of hydrocarbon vapours.
Most drainage pipes are made from polyvinyl chloride and offer excellent chemical resistance. However, the seals used in joints are generally made from ethylene propylene diene monomer rubber, which has a poor resistance to GROs and DROs leading to failure when there is direct contact with contaminants. Potable water pipes are typically made from polyethylene. Similar to membranes made from the same material, they offer poor resistance to DROs and GROs such as kerosene. It is often the case that a water main passing through a contaminant plume will need to be replaced. In addition, the plumbing pipework in a property can be tainted, leaving an oily taste in tap water.
THE REMEDIATION PROCESS AT THE PROPERTY
The programme of remedial works at a domestic property largely adheres to the approach set down in the UK Government’s land contamination risk management guidance. In addition, investigations should follow the relevant standards such as the code of practice for ground investigations (BS 5930:2015+A1:2020); the investigation of potentially contaminated sites (BS 10175:2011+A2:2017); and taking soil samples to determine VOC presence (BS 10176:2020).
An initial study was undertaken at the property followed by a detailed site investigation to adequately determine the extent of the contamination, with the findings used to create a site CSM. To undertake an effective assessment of a property, a good understanding is required of both the land contamination and the nature of the various structure types and how they are assembled.
Once the extent of the problem was understood, a schedule of remedial works was designed that factored in all the available data together with the economic requirements and needs of the homeowner (e.g. costs and a desire to return home as soon as possible). While there are many ways to remediate a property, a balance should be struck that provides the most sustainable approach. For example, the severance of a pathway may be preferable to the removal of a contamination source; both methods remove the pollutant linkage and are equally valid remediation approaches. However, homeowners may be resistant to anything other than source removal, so it is important to guide them through the process and the validity of the proposed approach.
Ultimately, the remedial works required were extensive. Internally, a chimney breast was removed and a large section of the internal flooring was excavated (see Figures 5 and 6). The removal of significant quantities of contaminated material inside the property down to foundation level was necessary. Accumulated groundwater in the excavations was found to have much free product (i.e. kerosene in its pure form) (see Figure 7) and iridescence was present on the surface (see Figure 8). This indicated that the solubility limits for at least some of the contaminants had been exceeded. An oil–water separator was used on site until all the free product was removed.
The base of the excavation was treated with oxidisers, and the exposed rising wall blockwork was scrubbed.
The contamination present within the wall cavity was flushed out. Saturated concrete was replaced, and insulation, membranes and damp-proof courses were removed and replaced if they had come into direct contact with kerosene. Where removal was impractical and a particular building element could act as a VOC source, it was either encapsulated to lock in the contamination or treated to degrade the contamination to lower levels. Where no good assessment criteria existed, professional judgement was used. Externally, contaminated soil in contact with the walls of the property was excavated and disposed of. This was necessary to prevent re-contamination of the property and to lower the risk
Following this, the property was restored to its original layout. As part of the reinstatement process a hydrocarbon vapour membrane was installed to sever the pollutant pathway and assist in mitigating the risk of vapours entering the property from any residual low-level contamination that could still be present at depth (see Figure 9).
The entire remediation project took nearly nine months. Some initial delays arose over negotiation of the nature of the work and the cost, but it was important that a mutual agreement was reached to ensure final sign-off to everyone’s satisfaction. Finally, a validation survey was undertaken to ensure that all previously identified unacceptable risks to receptors had been addressed and the amenity of the property restored. By undertaking a high-quality investigation and thorough understanding of the risks and uncertainties present in the assessment, stakeholders were satisfied that the remedial measures employed were sufficient and appropriate.