Earl Pumping Station ground investigation, risk assessment and remediation strategy, South East London by Mott MacDonald
Shortlisted for Brownfield Awards Category 1 - Best Project Preparatory Work
London relies on a 150-year-old sewer system built for a population less than half its current size. As a result, millions of tonnes of raw sewage spills, untreated, into the River Thames each year. The Thames Tideway Tunnel scheme is a landmark 25km ‘Super Sewer’ that runs largely beneath the River Thames in London. It is the largest project ever undertaken by the UK water industry and has been the subject of close scrutiny and huge public interest since its inception.
The scheme will intercept combined sewer overflows along the River Thames to clean up the river for future generations. There are a number of shaft sites which will transfer sewerage from the existing network into the tunnel. Earl Pumping Station (EARPS) in South East London is one of these sites.
Figure 1: Location of site in relation to Thames Tideway Tunnel scheme
The EARPS site is a small site (measuring approximately 70m by 70m in plan area) and one of the smallest of the Thames Tideway Tunnel sites. It is located partially within the existing Thames Water sewage pumping station and partially within recently cleared commercial premises. It is in an area that has undergone rapid regeneration from derelict former industry to residential land use since the EIA process commenced in 2010 and is now surrounded on four sides by housing. As well as the challenges associated with the constraints of a small site with existing below ground infrastructure (e.g. access for boreholes, excavated materials management, remediation equipment) the increasing number of residents and other construction activities in the areas has meant ever increasing interest and scrutiny.
Figure 2: EARPS site and surroundings
The diminutive size of the site contrasts with the complex and considerable ground engineering that is required as part of the scheme. This includes a 51m deep, 17m internal diameter drop shaft connecting to the Greenwich Connection Tunnel which is planned to pass through the site in 2021. The project also requires the construction of a large interception chamber and connection culvert between the new shaft and the existing pumping station extending to approximately 12m below ground level.
Figure 3: EARPS proposed structures
Site setting and contaminant legacy
The underlying geology comprises a cover of made ground overlying soft clayey alluvium with pockets of peat in turn overlying Kempton Park Gravel Formation, Thanet Formation, and Chalk. The London Clay Formation and cohesive Lambeth Group that provides a barrier to contaminant migration throughout much of London is absent at the site.
Figure 4: Ground Model
The southern part of the site and adjacent area to the south have a legacy of contamination from a more extensive tar/pitch/naphtha works in the late 19th and early 20th Century as well as more recent light industrial and commercial use. The Earl Sewer crosses the site and the existing sewerage pumping station on the northern part of the site dates back to the 1940s.
The principal receptors for the site were considered to be the underlying Principal (Thanet Formation and Chalk), the surrounding residential properties and workers in the TWUL compound.
Ground investigations comprised 50 boreholes by Tideway in 2012 and 2014, which first indicated the existence of significant contamination at the site. This was supplemented by additional investigation of a further 10 locations by CVB to refine the conceptual site model (CSM) and for geotechnical and hydrogeological assessment to inform design of structures and dewatering.
A programme of monitoring was devised and the datasets from all investigations brought together for the first time to provide a detailed understanding of contaminant distribution across the site. This information was used to inform the risk assessment and remediation strategy for the site as required by the Development Consent Order for the project.
Figure 5: Contaminant distribution
The investigation showed that the principal contamination comprised petroleum hydrocarbons, with a significant proportion of polycyclic aromatic hydrocarbons typical of a light distillate tar. This contamination was extensive in the highly permeable Kempton Park Gravel and the Thanet Formation, but at much lower concentrations in the top of the underlying Chalk. The deeper Chalk was largely uncontaminated.
Hydrocarbons were largely present as dissolved phase with intermittent observations of light non-aqueous phase liquids (LNAPL). Globules of neutrally buoyant non-aqueous phase liquids (NNAPL) were also observed intermittently during sampling. More significantly an apparently isolated area of dense non aqueous phase liquids (DNAPL) contamination was found in the Thanet Formation in the north western corner of the site away from the main identified NAPL sources.
Although there is no London Clay beneath the site, particle size distribution analysis shows that there is a finer grained component at the base of the Thanet Formation that had restricted the migration of dissolved phase contamination and prevented the migration of DNAPL into the Chalk.
The reason for the isolated DNAPL away from known sources is postulated as being associated with historical construction activities in this area of the site.
Detailed Quantitative Risk Assessment
Following generic quantitative risk assessments, more detailed understanding of risk was needed. Key receptors comprised groundwater in the Chalk, surrounding housing, construction workers and Thames Water personnel.
Observations of groundwater data indicated that flow directions were found to be consistently in a westerly/north westerly direction in the Kempton Park Gravel, a north westerly direction in the Thanet Formation and north/north-westerly direction in the Chalk.
Groundwater data indicated that upgradient dissolved phase concentrations were higher than those downgradient with the exception of the area associated with the DNAPL in the Thanet Formation. Detailed quantitative risk assessment (DQRA) using Environment Agency Remedial targets methodology was undertaken to model potential downgradient impacts for groundwater assuming an infinite, constant source at the site.
The DQRA demonstrated that the main contaminants of concern included ammoniacal nitrogen, benzene and naphthalene in the Kempton Park Gravel. We also established that measured downgradient concentrations are low, indicating rapid biodegradation and significant attenuation leading to best case results with no significant risks at the chosen compliance point.
A numerical groundwater flow model was also generated to demonstrate that the proposed construction would not pose an unacceptable risk at the site and surrounding residential areas. A MODFLOW model, with particle tracking using MODPATH, was generated to assess whether the shaft would act as a significant barrier to groundwater flow and thus raise groundwater levels upgradient and significantly divert contaminant flow paths sufficiently to alter impacts to off-site receptors. The modelling demonstrated that the impacts of the shaft and associated sewer diversion on groundwater levels and flow directions will be small and would not significantly affect impacts to off-site receptors.
Figure 6: MODFLOW outputs
In view of the housing located upgradient of the site and the presence of volatile contaminants in groundwater, a DQRA for vapours originating from groundwater was undertaken. Site specific assessment criteria (SSAC) were derived using the CLEA model for benzene, naphthalene, ethylbenzene and acenaphthene. Sensitivity analyses were run for differing ground conditions to account for uncertainties in the ground model. Overall, the observed shallow groundwater concentrations downgradient were significantly lower than the SSACs. A programme of construction groundwater monitoring continues to monitor this potential pollutant linkage as the project progresses.
Embedded mitigation and risk communication
With a significant scheme that has been through a detailed scrutiny as part of a DCO and EIA process, there was an opportunity to tie the narrative in the EIA and principles of ‘embedded mitigation’ into the Remediation Strategy. These are good or best available practice measures that aim to reduce risks to the identified sensitive receptors. These measures in the design phase comprised:
a review of shaft construction methodology and use of in ground-treatment (geomix wall) to reduce risks to the Chalk (effectively using the principles of ‘clean drilling’ boreholes but on a shaft scale);
review of shaft and pump station design and assessment of potential ingress of contaminated groundwater through cracks and subsequent volatilisation of contaminants using Henry’s law and comparisons with workplace exposure limits; and
numerical modelling to determine groundwater flow paths caused by the presence of the shaft and temporary works.
In addition, the importance of risk communication for construction workers, a series of activity specific RAG assessment tables and figures were produced to aid communication of the presence and severity of contamination associated with various construction activities.
A number of construction phase mitigation measures were also considered as part of embedded mitigation and used in the assessment of risk. One example of this was a comprehensive odour and air quality monitoring plan to protect local residents. The plan included regular liaison with residents together with a feedback and complaints procedure. This has been successfully implemented as the works have progressed.
Figure 7: RAG assessments
Remediation options appraisal and strategy
Although remediation of impacted groundwater on site was theoretically possible, the combination of on-going sources upgradient, restrictions on site access and the combination of active remediation techniques that would be required make this unfeasible. However, monitoring data from on-site and down-gradient off-site boreholes indicated that natural attenuation of hydrocarbons was occurring, with lower concentrations in down-gradient boreholes and geochemically reducing conditions.
We therefore successfully put a case forward that active remediation of the dissolved phase contamination in the Kempton Park Gravel would be disproportionate in relation to the likely long-term improvement of groundwater quality. Therefore, proposed remediation should be restricted to placement of a geomix wall around the CSO drop shaft to restrict any contaminant migration to the chalk during excavation, the management of the identified DNAPL in the Thanet Formation and contingency monitoring and watching brief for the remainder of the site. Ongoing monitoring provided evidence that natural attenuation of contaminants continued in accordance with a wider MNA scheme for the surrounding 3rd party land.
Stakeholder engagement and ongoing assessments
One of the key challenges associated with this project was to gain regulatory buy-in for the complex engineering, construction programme and detailed risk assessments. Key stakeholders included the ultimate client Thames Water Utilities Limited; the local planning authorities, London Borough of Lewisham and London Borough of Southwark; and the Environment Agency.
Engagement was undertaken through a series of collaborative stakeholder workshops over sever4al months while additional groundwater monitoring continued. Contamination data and work-in-progress assessments, together with the details of the ground engineering, discussed between all stakeholders and MML and CVB. This included information on the ground model, contamination distribution and risk assessment approach. Findings of initial DQRA as well as outline remediation options and strategy were presented to stakeholders as part of these workshops.
This collaborative approach meant that there were no surprises when the combined risk assessment, remediation options appraisal and strategy was submitted as part of the DCO consent process. The report was accepted by the LB Lewisham in April 2018 and provided the necessary data and assessment to allow for start on site. However, owing to the complexity of the project and the need to finalise some design elements, some of the embedded mitigation needed to be confirmed through the provision of additional assessments which are ongoing.
The key additional assessments required as part of the remediation strategy relate to the submission of a foundation works risk assessment (FWRA) and a remediation method statement (RMS). The FWRA is currently being updated to reflect the revised ground engineering proposals associated with the interception chamber. The RMS was submitted to regulators in June 2019 ahead of the shaft construction which commenced in June 2019. The excavation of shaft was completed in March 2020.
Site specific materials re-use/import criteria were also developed using CONSIM. A robust set of RAMS (including DNAPL control) were also developed to maintain risks to as low as practicable during below ground works. As part of embedded mitigation, a groundwater monitoring also continues in line with the agreed construction groundwater monitoring plan.
The site is heavily constrained by both its size and significant below ground infrastructure as well as by being surrounded on four sides by housing. However, through comprehensive site investigation, robust and detailed risk assessment, risk communication and collaborative engagement with multiple stakeholders, we were able to find sustainable remediation solutions and appropriate embedded mitigation measures. This has enabled risks to be managed appropriately and allowed construction works to commence safely on programme.