Lower Thames Crossing, Kent, Essex - Arcadis
Shortlisted for Brownfield Awards Category 2 - Best Scientific/Technical/Digital Advance
Development of the Conceptual Site Model
A geo-environmental CSM was required to understand and communicate the potential contamination risks associated with the project and develop a robust ground investigation program. The need to integrate vast amounts of data from numerous sources, including geology, hydrogeology, environmental datasets and historical and current land uses; whilst maintaining flexibility to adapt to an evolving design and the need to communicate findings to internal and external, technical and non-technical stakeholders, were the key challenges faced.
To overcome the challenge, on the behalf of Highways England LTC-Cascade (a joint venture comprising Arcadis, COWI and Jacobs) produced a fully georeferenced digital CSM in a Geographical Information System (GIS). This integrated numerous datasets from third parties, digital collection of field data, and published sources. Thousands of environmental data points covering an area of more than 70 km2 were collected and interrogated to identify potential pollutant linkages that could credibly impact the project. Over 160 potentially contaminated sites were identified and categorised for potential contaminants. Sensitive land uses and other receptors such as aquifers and surface waters were digitised. Geospatial analysis was then used to produce a robust CSM, delivering a consistent approach across areas of highly variable land histories and environmental settings, and enabling additional areas to be efficiently assessed as the project developed. The digital CSM linked with the GIS-based GI design and scheduling of appropriate chemical testing of soils and waters.
The Lower Thames Crossing (LTC) is a Nationally Significant Infrastructure Project under the Planning Act 2008, and the UK’s largest road project since the construction of the M25 motorway. The project includes 14.3 miles of new road and the largest twin-bored road tunnels (over 16 metres in diameter and 2.6 miles length) ever constructed in the UK (top 5 in the world) which will pass underneath the River Thames between Gravesend and East Tilbury. Over 160 mapped, potentially contaminative historical land uses including, but not limited to, major landfills, filling stations and garages and a WWII airfield have been identified.
The multidisciplinary Ground Engineering team for LTC brings together Ground Investigation, Hydrogeology, Geo-environmental, Geology, Geotechnics, and GIS, CAD and BIM specialists. The team has responsibility for all ground engineering-related aspects of the project’s preliminary design. The nature of the project has required a unique integrated, innovative and truly collaborative approach, with the development and utilisation of technical and digital enhancements for the preparation and refinement of the Conceptual Site Model (CSM); the collection of geo-environmental data; assessment of data and presentation of its interpretation to a wide technical and non-technical audience.
The team designed and supported the execution of a multi-million pound major phase of ground investigation works. This included advanced geo-environmental and hydrogeological assessment, data collection, monitoring and testing and innovative surface geophysics surveys. Advanced ground risk management and communication tools were developed, including a sophisticated project-wide 3D ground model using the LeapFrog software and GIS-based geo-environmental CSM and story boarding.
The above tools and techniques have been integrated to manage data, support decision making and ensure robust and appropriate management, and effective communication, of Brownfield risks, whilst this very large and complex project continually evolves.
Refining the Conceptual Site Model
Construction of the LTC across the Tilbury Marshes (the most significant brownfield site along the route) poses a range of engineering challenges. The area has complex geology; comprising a thick sequence of clays, sand, peat bands, overlying sands and gravels, and Chalk bedrock. Historical landfilling has created complex Made Ground deposits.
The project has a large database of borehole stratigraphy, geotechnical and geo-environmental testing data arising from archive borehole data and the drilling of new boreholes, managed in HoleBase and ESdat databases. Other than traditional 2D visualisations, LTC have developed the ground model further, by the commissioning a high-resolution LiDAR survey to give an accurate digital elevation model of the route corridor, and map landforms, including the landfills.
Geospatial data have been processed in ArcGIS to map areas of adverse ground conditions. The desk study suggested that up to five peat layers could be expected within the alluvium. However, visualisation of borehole logs in ArcScene showed a more complex pattern, with beds rarely being traceable over more than 100m and little evidence for concentration of peat at the depths suggested by the desk study. To map the spatial peat patterns, the total peat thickness in each borehole was extracted from HoleBase and results interpolated in ArcGIS. The resulting ‘heat maps’ show total peat thickness in the alluvium and the proportion of alluvium that is peat. The data show that most the area of the Tilbury Marshes crossed by the LTC has less than 1m of peat in total, but there are several small areas, mostly near the river margin, where up to 3m has been encountered. The data indicate that peat comprises up to 5% of alluvium in the north of the Tilbury Marshes, increasing to 10 to 20% at the river margin. This detailed understanding of alluvium and peat deposits enhanced the understanding and refinement of the potential ground gas risk from these sources. This fed directly into the on-site monitoring scope, and project design and mitigation measures following the initial sporadic high methane concentrations recorded.
A TROMINO® geophysical survey comprised point data indicating the depth to a geophysical boundary interpreted as the top of the chalk. The data showed a good correlation in most areas, with notable anomalies in the data east of the LTC alignment, correlating to known areas of former landfill. Results suggest that TROMINO® could be used to infer the top of the chalk where borehole data are absent, and indicate where atypical ground conditions, like Made Ground, are present.
Traditional borehole datasets and novel remote sensing datasets were therefore combined to develop a preliminary site ground model to characterise the ground and to help define areas of potential contamination. The models will continue to be defined as further ground investigation and geo-environmental data is obtained. These will directly feed into the on-going refinement of the CSM and the understanding and communication of potential Brownfield risks to the LTC scheme.
Roboduck - Innovative Collection of Data
Roboduck was designed and built by LTC as a remotely operated floating drone equipped with water quality sensors, a GPS tracker and a video camera. This craft was deployed within various watercourses and at different states of the tide, and collected continuous water quality data for assessment. The small size of many of the watercourses precluded commercial waterborne UAVS with enough payload capacity for the sensors required. A bait boat, used to drop bait to attract fish, was selected due to its small size, high payload capacity and electric waterjet drive to reduce weed entanglement. The standard lead acid batteries were upgraded to lithium to decrease weight and increase range.
A commercially available multiparameter probe, commonly used for groundwater monitoring applications or stream sampling of point locations was clamped to the base of the craft, allowing easy removal for calibration, or for point measurements in locations where even this small craft could not access due to vegetation obstructions. The data was transmitted via Bluetooth® technology and stored on a mobile device, allowing for live viewing of the data in most circumstances.
The two data streams, GPS position and Multiparameter readings, are automatically combined via the timestamps of two data sources for data processing to allow synchronising of the multiparameter probe recording device with GPS time. Less than 1 second of drift occurred between the time sources over the 3-day survey.
Roboduck produced over 1000 water quality data sets, at sampled distances of up to 5 metres, often shorter, as controlled by the operator for fine detail. This provided far more samples than could be collected safely using traditional grab-sampling techniques. The resulting water quality mapping gave excellent resolution of subtle variations of water quality, identifying even small changes caused by tributary ditches. The use of Roboduck data has enabled LTC to map the variation of salinity, temperature, electrical conductivity, pH, ORP, identifying the fine detail of fresh and brackish water transition and the possibility of anthropogenic discharges to surface water. Roboduck offers excellent repeatability and better comparability with historic data when compared to a traditional grab sample approach.
Continuous Monitoring Data
The proposed tunnels North Portal is located within an area of historical landfill, overlying organic Alluvium containing significant thicknesses of peat albeit of variable extents/presence. Due to the complex ground conditions, various landfilling phases and a tidal groundwater regime, specialised Continuous Gas and Groundwater Monitoring Instruments (CGGMI) were employed to assess ground gas risks, in conjunction with “standard” spot monitoring.
The data is collected by permanently installed GasfluxTM and dissolved gas probes, deployed in boreholes drilled by LTC within the landfill, Alluvium, river terrace deposits and Chalk. The CGGMI are fully automated and powered by solar panels, reducing time and cost spent on gathering data on site. The data is sent remotely via wireless telemetry to an online portal allowing the LTC Ground Engineering Geoenvironmental team real-time monitoring of the incoming data without the need for conventional site visits.
A separate on-site weather station ensures that high-quality localised information is available for atmospheric pressure and other parameters important for understanding the gas risk.
Data collected includes:
bulk gas concentrations (methane, carbon dioxide, oxygen), trace gases (carbon monoxide, hydrogen sulphide), volatile organic compounds (VOC);
pressure, gas flow, water levels;
dissolved methane and groundwater temperature.
Extensive monitoring, producing a robust dataset on which gas risks can be assessed and communicated without the need for site visits, especially during the coronavirus pandemic, has been pivotal in refining the CSM related to ground gas risk and refining the hydrogeological model at the North Portal.
Original Numerical Groundwater Modelling
The potential impacts from LTC on hydrogeology include drawdown and contaminant transport from nearby landfills. The BGS 3D lithostratigraphic model and the LTC ground investigations data have been merged within a block model, created using the USGS FloPy, a set of Python scripts enabling one to build, run and post-process MODFLOW, MT3D and SEAWAT models. Visualisation and MODPATH simulations have also been completed in Groundwater Vistas 7.
Using Python script the model geometry and parameters can be easily manipulated. This enables rapid testing of how changes in design may affect the results, such as the depth of groundwater cut-off walls. The calibration includes a Monte Carlo analysis that tested thousands of parameter combinations. Each simulation included a steady state and time-variant calibration assessment.
The advantages of Python coding and block modelling include:
perfect, automated parameterisation by scripted manipulation of the arrays making up the model, giving rapid convergence resulting in short run-times using MF2005;
automated construction of the model even for different cell sizes, extents or layer thickness;
automated generation of infrastructure boundary conditions and zone budget files from mapping;
automated generation of results, such as grids of head and drawdown and calibration residuals; and
automation of sequential runs and Monte Carlo assessments.
Communication and Engagement
The collection of robust data sets without effective communication of the findings would be wasteful. Therefore, the following communication tools were employed:
Digital storyboarding utilising ArcGIS StoryMaps with instant hyperlinked access to detailed information. The StoryMaps approach has been integral to the communication of potential Brownfield risks to the scheme, both face to face and virtually, via the use of Microsoft Teams.
Visualisation of the ground model with LeapFrog.
Presentation of large datasets using PowerBI in an easily digestible format allowing real-time interpretation of findings.
Continuous monitoring data is assessed via Ambisense's Ambilytics platform to expedite trend analysis and allow real-time assessment.
The use of innovative technology and digital tools to produce and continually refine the CSM has allowed a large and varied pool of information to be quickly and effectively assessed to provide clarity on potential contamination issues along the proposed route. This also enabled seamless and real-time visual communication between multidisciplinary teams, supporting decision making and communicates risk to stakeholders from an early stage in the project.