Geochemical mapping of England and Wales
Geochemical baseline survey of the environment (G-Base)
The BGS geochemical baseline survey of the environment (G-BASE) project is the national strategic geochemical survey of soils, stream sediments and stream waters in Great Britain that took place over 47 years between 1968 to 2015. This case study relates to English soils surveyed between 1984 and 2014. G-BASE is a big data resource contributing to scientific advancements that support the protection and enhancement of human health from harmful and essential elements in soil. Scientific outputs include geochemical mapping projects and chemical exposure assessments for risk-based land management. The methodology, illustrated in the cartoon below, was developed by BGS and is now used in other national and regional surveys including Nigeria, Eire and the Tellus survey of Northern Ireland, UK. G-BASE has also received commendations from industry and regulators alike. The field procedures manual for the G-BASE project is available for download here.
Soil sampling and analysis
The G-BASE project has carried out annual regional and urban geochemical survey and mapping campaigns to provide baseline data derived by X-ray fluorescence spectroscopy for a large number of major and trace elements in over 44 000 soil samples. Comprehensive quality assurance procedures ensure continuity of data between sampling campaigns.
Systematic sample selection
In rural areas, 1 soil sample is taken per 2 km2 and in urban areas, 4 soil samples are taken per 1 km2 (Figures 2 and 3). The London Earth soil sampling campaigns were carried out from 2005 to 2009, analyses were completed in 2010 and analytical data for over 50 elements in 6600 samples is now available from the BGS under licence. More information on the London Earth project is available here.
Geostatistical sample selection
In order to improve the efficiency of sample collection in the south-west of England, soil sampling was optimized statistically. Figure 4 shows south-west England Survey area. Geostatistical models were used to optimize the distribution of new samples (black symbols) while accounting for existing sample data (red symbols). The sampling and data analysis campaign was conducted over 2 years between 2012 and 2014.
Data workup, analysis and mapping
BGS geochemical maps are currently generated in ArcGIS using the inverse distance weighting (IDW) algorithm to convert the point data to a continuous surface grid. The maps show the distribution of potentially harmful or essential chemical elements in the rural and urban environment. For example, the map of lead in Figure 5 highlights the impact of mineralisation and mining of the Pennine Ore field on soil quality in the Trent Valley.
The map of arsenic in London in Figure 6 shows the influence of London Clay, alluvial parent materials and industrial landuse on its distribution.
Application 1: Derivation of normal background concentrations (NBCs)
Given the unique national scale coverage of soil quality information provided by the G-BASE dataset, BGS was commissioned by the UK Government Department for Environment, Food and Rural Affairs (DEFRA) (Contract reference: SP1008) to derive normal background concentrations (NBCs) for 7 potentially harmful substances (PHS) (Arsenic, Cadmium, Copper, Mercury, Nickel, Lead and Benzo[a]pyrene) in soil in England and Wales (Table 1).
DEFRA commissioned this work to remove ambiguity over the definition of ‘typical’/normal concentrations of PHS in soil used in statutory guidance for Part 2a of the Environmental Protection Act 1990. NBCs are now used by Local Authorities (LAs) and their consultants as lines-of-evidence to help identify contaminant concentrations that do not form part of normal background and therefore might require further investigation and risk assessment under Part 2a. The BGS method defines NBCs as the upper confidence limit of the 95th percentile of the geochemical data for the target geographic domain and provide a conservative estimate of the background concentration over that area, accounting for geogenic and diffuse sources of PHS. Technical reports and peer reviewed publications in high impact journals are available for the project. Figure 7 summarises the methodology.
Figure 8 shows the distributions of arsenic concentrations across England.
Figure 9 shows the geochemical data as well as the NBC for the principal arsenic domain. For more information on the BGS NBCs follow this link.
Application 2: Machine learning for high resolution geochemical mapping
A new application for the G-BASE data includes a novel method for improving the resolution of interpolated geochemical maps (see Figure 5 and 6) using machine learning methods. The following example uses rubidium. Other element maps are available.
Traditional geochemical maps produced by simple univariate interpolation methods such as ordinary kriging (Figure 10a) are limited by the spatial sampling density of the survey, and are blind to other environmental information.
By using machine learning methods instead, supported by remote sensing and geophysical survey data, surface geochemistry can be mapped
with greater resolution, accuracy and interpretability (Figure 10b). The resultant maps and models can be interrogated to provide superior evidence on which to make risk-based decisions about likely effects of potentially harmful or essential element distribution on human health.
Benefit to the economy, society and the environment
“Tellus Northern Ireland cost £6 million to complete and in the year following its release £11 million was spent in the mineral exploration sector”. G Berkenheger, GreenOre Gold PLC, 2015.
“The [G-BASE] data we received is highly satisfactory for the requirements of my role for the Council. The data is used to deal with contaminated land investigations, understanding potential pollution pathways and also dealing with environmental search requests. The data has been used for many years by Trafford Council, [and] primarily assists in understanding risks from contaminated land”. R Pollitt, Trafford Metropolitan Borough Council, 2015.
For further information about this case study contact Dr Darren Beriro ([email protected])