Why it pays to do some ground work

The increasing emphasis on statistically assessing site data on ground investigations can be time-consuming and costly if you're not careful, explains Cathy Swords

Changes in national and international guidance are shaping the way site investigation and contaminated land assessments are carried out in the UK. Increased regulatory emphasis on data collection methods and statistical confidence is leading to better designed investigations with more consideration of sampling strategies and how this may affect interpretation.

The geotechnical design code Eurocode 7 aims to implement these changes for geotechnical investigations to increase confidence in the ground model and in the assessment of geotechnical design data. This shift towards statistical methods is not simply to introduce more complications. Rather it provides a mechanism to manage uncertainty in datasets and interpretation, although it must be recognised that the foundation for statistical assessments is an accurate and well-defined conceptual site model.

Increased costs associated with site investigation in accordance with Eurocode are likely to be inevitable. However, on more complex sites the increased knowledge of the ground may result in added value to geo-environmental aspects of a project, particularly if it can provide more confidence in pollutant fate and reduce remediation costs.

Recent developments in contaminated land are not only related to data acquisition.

Earlier this year the Environment Agency issued a revised CLEA model and an updated UK framework for the assessment of chronic human health risks from soil contamination. As part of the model update, the agency has withdrawn the soil guideline values (SGVs) generated using the old methodology and is currently in the process of issuing new SGVs in line with the updated model.

To date, SGVs have been issued for four metals (arsenic, nickel, selenium and mercury) and four volatile (BTEX) compounds. Further SGVs are promised – however, until all of the planned SGVs have been published, the contaminated land community must generate generic assessment criteria (GACs) to plug the gap.

Current SGVs limited in their use
Furthermore, the new SGVs are not quite what the contaminated land community expected. SGVs are now only available for three land use scenarios, namely commercial, allotments and residential; the latter assuming homegrown produce in a private garden. The withdrawal of the land use scenario ‘residential without plant uptake’ means that practitioners must now generate assessment criteria for residential scenarios where vegetable uptake is not a concern – for example, communal soft landscaping.

It is also important to note that SGVs are no longer the minimum screening values they once were. Based on the assumption of a sandy loam soil with 6% soil organic matter, they have effectively been generated for topsoil. Site conditions varying from this assumption, such as low organic content made ground, will require the calculation of an appropriate, possibly lower, assessment criterion which can be used in a generic risk assessment.

The small number of published SGVs and the need to generate GACs to deal with the land use gap and soil type issues with the new SGVs is a complex process requiring significant resources and in the absence of centrally issued guidance many consultants are turning to a few specialist suppliers of GACs. However, practitioners using commercially available GACs must be aware of the assumptions behind the values as in many cases they may not be wholly applicable or representative of the sites to which they are applied.

While generating our CARD in-house GACs, it has become apparent that some of the commercially available GACs include highly conservative, often lower bound, assumptions. Such assumptions include how plant uptake is modelled, choice of toxicological inputs, choice of exposure pathways and the application of correction factors to account for the conservatism in the vapour ingress model.

For example, a GAC of 0.06mg/kg was generated for Benzene earlier this year. More recently, the Environment Agency published a SGV for benzene of 0.33mg/kg for the same land use scenario – a five-fold increase from the initial GAC. The difference in the values is due primarily to the application of a sub-surface soil to indoor air correction factor to account for over-prediction in empirical studies relating to vapour ingress. The factor applied in the generation of the SGV was 10 compared to the value of 1 used to generate the GAC.

This example highlights the conservatism applied to commercial GACs so that they are applicable to a wider range of site conditions, with the trade-off being that risk may be overestimated if these numbers are not applied correctly. Our experience suggests that commercially purchased GACs are often used to inform judgements on the need for remediation, although they are provided as initial screening values not necessarily as remediation triggers.

The application of these highly conservative lower bound values in this manner will increase project costs and lead to costly and non-sustainable solutions. Instead, the GACs and indeed the SGVs are not intended as remediation trigger values without further refinement for site-specific conditions. In the current economic climate, overly conservative assessments could lead to many brownfield projects being abandoned.

Conversely, a thorough understanding of the underlying science and a detailed site-specific assessment more often than not will yield savings on remediation costs.

Given the situation, pragmatic and defensible approaches are more important than ever, and are still possible through carefully designed investigations and a thorough understanding of the underlying science behind contamination and risk assessment.

Cathy Swords is regional manager at Card Geotechnics

Action inspires action. Stay ahead of the curve with sustainability and energy newsletters from edie