Soil Sampling of Contaminated Land

Published: 10th Mar 2009

by Hazel Davidson and Geraint Williams


Sample selection, handling, transportation and storage

Site investigation and risk assessment have evolved into a major industry over the last two decades, with little legislation or guidance in the earlier years.

The errors associated with sampling are now better documented, and any environmental consultant or site contractor needs to be well aware of the large differences which can result from poor site investigation design, and incorrect procedures on site, particularly with respect to sample selection, handling, transportation and storage. The assessment of potential contamination within a site is of critical importance for the future use of the site, predicting the cost of possible remediation, and the re-sale value. One of the main drivers for site investigation is SPOSH – ‘the significant possibility of significant harm’ being caused by the presence of contamination on the site.

The design of a site investigation involves a great deal of research, and is generally referred to as the Desk Study, or a Phase 1 investigation. This will involve gaining information with respect to previous use of the site, locations of buildings (past and present), location of storage depots (particularly fuel tanks), identification of particular chemicals, feedstocks, and final products likely to have been stored on the site, and the possible discharge/spillages/burial of any materials, including waste generated on the site. In addition, the geology, hydrogeology, topography, location of natural water courses, and the presence of aquifers all need to be considered in assessing the potential mobility of any contaminants identified on the site.

The sampling of the soil (natural ground and made ground) will need to reflect all of these considerations with respect to location, depth horizons, frequency, sampling grid pattern, etc.

The errors associated with sampling on site are much greater than the analytical errors measured within the laboratory.

Frequency of sampling

There are now specific standards relating to site investigation and sampling, with one of the most commonly referenced being BS 10175. This is a comprehensive, if rather lengthy document, but is about to undergo a re-write. There are now ISO standards also covering various types of sampling on site (see references). The recommended number of samples to take will obviously depend on all the data gathered in the Phase 1 investigation, but some numbers may help to put this in perspective:

  • Average 1 hectare site = 17 trial pits/bore holes
  • Assume a sampling depth = 3 metres
  • Assume 5 depth horizons for each location between surface and 3 m
  • Assume a soil specific gravity of 2.5
  • Thus we have a volume of soil a hectare square and 3 m deep – the weight of soil in this is 75,000 tonnes
  • Therefore the weight of soil represented with each 1 kg sample submitted is 882 tonnes, or 0.0003 % of the site

Most people would not consider this to be representative…

Once soil samples are received into the laboratory, the situation is even worse, as only 1 – 20 g of soil is used for each test, so the percentage of soil actually tested can be as little as one millionth of a per cent.

It is figures such as these which may assist site personnel to appreciate the significant problems in trying to obtain representative sub samples, and to emphasize the need to take as many samples as possible in order to improve the certainty associated with their interpretation of the risk related to a specific site.


The ISO definition of uncertainty is:

‘An estimate attached to a test result which characterises the range of values within which the true value is asserted to lie.’

The emphasis for the site assessment will be to obtain a ‘reasonable minimum amount of information’ in order to perform a meaningful risk assessment. This is defined as ‘a set of information about the condition of a site and its setting, obtained at each stage of the investigation, that should provide a reasonable basis for assessment’.

There is uncertainty associated with any test result, but this is compounded by the much greater uncertainty associated with the sampling. It is useful to be able to quote the testing uncertainty, as it can affect the classification of the soil, or area from which the soil was taken.

Uncertainty within the laboratory varies with the method – generally, inorganic methods will fall between +/- 5 – 10 %, and organic methods between +/- 15 – 30%. Uncertainty in the on site sampling is commonly +/- 50 – 200%.

Mike Ramsay, with a team working at the University of Sussex (and industry partners), are developing a Decision Support Tool (DST) to assist with site investigations (see references for information on a CLAIRE conference which include a presentation on this). Basically, the DST asks a number of questions about the sampling protocol (e.g. excavation method being used, sampling method, how many sampling points, what analysis is required), and then the DST recommends:

  • How many duplicates should be taken
  • At which sampling point the duplicates should be taken
  • How the samples should be collected
  • How the laboratory analysis should be done (duplicates, etc)

Systematic uncertainty is the inherent uncertainty associated with a method and can be assessed by analysing samples and comparing the data obtained with those of a reference standard analysed by the same method.

Random uncertainty can be estimated by running replicate samples and deriving a value for the precision. This can be applied to sub samples of a sample submitted to a laboratory, or to replicate samples taken on site.

In the laboratory, analytical uncertainty can be reduced by taking replicate samples from the original submitted sample and reporting an average of all the replicates. Unfortunately, in the real world, the cost involved does not usually allow this to happen. On-site variation can be reduced by taking more samples, but again, there will be cost constraints.

Method Advantages Disadvantages
Trial Pits Applicable to environmental and geo-environmental investigations Allows for detailed visual assessment of ground conditions Rapid and inexpensive technique for examination of the upper 3-4m Applicable to a wide range of ground conditions Potential for cross contamination (extra care needed to ensure that the surrounding area is not affected by excavated spoil and the reinstatement does not leave contaminants exposed on the surface) Depth of excavation limited to the size of the machine Limited to working above the water table Installation of monitoring instrumentation not recommended
Window Samplers Applicable to environmental and geo-environmental investigations Generally low sample disturbance Easy re-instatement with only limited spoil for disposal Installation of narrow diameter monitoring instrumentation (gas and groundwater) Low quantity of sample retrieved is not always representative – small disturbed samples only Potential loss of volatiles Unsuitable for rock, with poor recovery and slow progress in coarse, granular materials Installations have limited annular filter pack/response zone
Cable Tool Percussion Drilling (shell and auger) Applicable to environmental and geo-environmental investigation Allows greater depth than trial pits and window samplers (typically drilled depths are between 10m to 30m although greater depths achievable) Allows for the retrieval of undisturbed samples Allows for wide range of in-situ testing for geotechnical purposes (i.e. SPTs) Enables the installation of permanent monitoring instrumentation (gas and groundwater) More expensive than trial pits Unsuitable for rock; large cobbles can inhibit progress Potential for contamination of underlying aquifers and groundwater flow between strata unless properly cased Potential to create migration pathways Less amenable to visual inspection than trial pits
Rotary Drilling Applicable to geotechnical investigation Allows for the greater depth of drilling (in excess of depths reached by cable percussion drilling) Capable of drilling through solid strata (penetration of rock) Permits installation of permanent monitoring instrumentation Use of water flush has potential to change natural groundwater conditions Use of air flush can oxidise samples and affect ground gases by encouraging migration of gases and vapours Not all techniques suitable for soils; core recovery in soft/loose soils is poor or requires more specialist techniques Borehole may need to be progressed to rockhead by cable percussion techniques

Site sampling

Soil samples can be taken by a variety of methods, depending on the location, depth, and type of samples required.

The weight of soil required for laboratory testing is usually 1 – 2 kg, but if the material to be tested is very chunky (> 20 mm particle size), then more material may be required in order for the laboratory to take a representative sub sample for testing.

It is better not to include large chunks of material (bricks, cobbles, etc.) unless these are to be included in the chemical analysis, as the laboratory do not usually remove these, and a representative sub sample will be crushed and analysed as part of the whole sample.

If both inorganic and organic testing are required, then a 250 ml glass jar, plus a 400 g plastic tub will be needed. If leach testing is also required, then a further 400 g plastic tub is also recommended.

Trial pitting

The excavation of trial pits is typically carried out by a mechanical excavator under the direct supervision of an engineer who records the ground conditions, as well as assessing the engineering characteristics of the material encountered. Pits can be extended to trenches (e.g. to search for boundaries).

The excavated material should be placed on a ground sheet, away from other excavated materials. The sample should be taken from the base of the pit rather than bulk excavated materials from differing depths. The sampler should ensure that the samples are representative of the stratum encountered and that any extraneous material from higher strata is removed. Extreme care is required to ensure a representative sample is obtained particularly when sampling below a tarmac or asphalt surface. A small fragment can have a dramatic effect on hydrocarbon results. Samples should be taken with a decontaminated trowel constructed of suitable material.

Window sampling

Window sample holes are driven through soils by a percussive hammer reacting on cylindrical steel windowless or window sample tube. Soil samples can be recovered by using a tube with an inert liner to enable easy removal of the core from the sampler. Window sampling can be used to collect samples at different depths, to rapidly penetrate to the depth at which the sample is to be taken, or to provide a continuous core. This technique allows for the hole to be advanced to between 6m to 10 m below ground level. Dense or stiff ground and other solid obstructions will restrict the depth of advancement.

Cable tool percussion drilling

Cable Tool Percussion, alternatively known as shell and auger drilling, is not suitable for investigating rock; it is most suited to drilling through superficial deposits (soils). Cable percussion boreholes are large diameter (usually 150mm to 250mm). In clay soils a weighted steel tube (clay cutter) is dropped down the borehole and the soil sampled from the contents of the cutter. Chisels aid in the penetration of hard ground obstructions. In granular soils a shell is used however a lubricant, such as water, must be added to the borehole and has potential for cross contamination or facilitating change in natural conditions.

The soil sample should be taken from materials not in direct contact with the drilling tools and transferred into the requisite containers using a decontaminated trowel. Undisturbed samples (U100s) can be collected in cohesive strata by driving a 100mm hallow tube into the ground and withdrawing the subsequent core for analysis. The traditional method of sealing U100s – with paraffin waxes, is inappropriate for samples destined for VOC analysis.

Rotary drilling

Rotary drilling uses a combination of rotary action and downward force and the technique and is applicable to rock, but less so to loose soils. The need for flushing medium makes many rotary-drilling techniques unsuitable for contaminated land investigation per se but may be useful in certain circumstances (e.g. uncontaminated areas upgradient of the site, installation of monitoring points).

Sampling for volatiles

Volatile organic compounds (VOCs) are common determinands requested on soil samples, but site personnel often do not realise the requirements to obtain meaningful volatiles data. It is important that the correct sample vessels are used, and these are usually small, 60 g glass jars for soils, which must only be used for VOC analysis. Some consultants/contractors simply take a bulk sample in a large jar or tub, and then request the laboratory to take a volatile from this sample. UKAS do not approve of this practice, and expect laboratories to remove accreditation from any samples taken in this manner, and to inform their clients that this is not an accepted protocol.

It is extremely important that the soil sample is packed tightly into the jar with minimal gaps or a headspace – volatiles will quickly be lost into these spaces. Pack the soil down, and scrape across the top with a knife or trowel – this will ensure a good seal with the lid. If both VOCs by GCMS and Gasoline Range Organics (GRO) by GCFID are required, then separate samples will be required. In addition, it is best practice to take each sample in duplicate, as this is a one shot analysis – once the laboratory has removed a portion of the soil, a headspace is created, so it is unlikely that a meaningful repeat can be performed, therefore a duplicate sample is advisable.

The samples should be kept cold – a cool box with frozen icepacks is recommended on site, and the samples should be returned to the laboratory as soon as possible.

Sample transportation/storage/holding times

Soil samples should be taken, either in plastic tubs or glass jars, packed carefully into cool boxes with frozen icepacks, and with enough packing so they cannot move, and also so no headspace remains (this will help to maintain a lower temperature). It is rare that chemical preservation is recommended for soil samples, as the complexity of the soil matrix will often cause more issues than the preserving agents will solve. The target temperature is 4oC, but it is accepted that this may be difficult to maintain, and the MCERTS standard for waters, produced by the Environment Agency in 2008, recommends 5oC +/- 3oC for the storage environment.

It is better not to stockpile samples on site, but to return the samples on a daily basis to the laboratory, otherwise holding times may be compromised.

Factors which can affect the stability of the soils include:

  • Changes in water content
  • Biological activity
  • Loss of volatile components
  • Interaction of chemicals/matrix within the soil
  • Reactions with the sample container
  • Light
  • Temperature

The rate of all these reactions will vary with the composition of the matrix of the sample (e.g. clay, sand, or loamy soil), the pH, the moisture content, and the organic content.

For several years, the USEPA guidelines on sample storage have been used in the UK, but a new standard was published in 2008 – ISO 18512, which recommends slightly more stringent holding times for some analytes (at 5 +/- 3oC). Figure 2 compares holding times for some of the common analytes in both guidelines;

Parameter Sample Size (g) Holding time USEPA (days) Holding time ISO 18512 (days)
pH 100 n/a 7
Heavy metals 10 6 months 6 months
Anions – Cl, SO4 20 28 28
Anions – NO3, NO2 20 2 1
Total sulphate 10 28 28
Easily liberated sulphide 100 7 n/a
Total and free cyanide 0 14 n/a
Asbestos 100 n/a n/a
VOCs 2 x 60 14 4
SVOCs 20 7 pre 40 post extraction 7
TOC 10 28 n/a
PAHs 20 7 pre 40 post extraction 4
EPH and mineral oil 20 14 7
Total phenols 20 7 pre 40 post extraction 4
Chlorinated pesticides 20 7 pre 40 post extraction 28 days
Organo N and P pesticides 20 7 pre 40 post extraction 7

Longterm storage of soil samples will usually require freezing the samples.


This article should indicate the many areas which require care and attention when sampling, storing, and analysing soils from contaminated sites. It is important to use a qualified and experienced consultant, and a laboratory which is accredited to ISO 17025 and MCERTS. Currently, the MCERTS standard for soils does not specify accreditation for soil sampling, but the standard for waters does include this. Future versions of the soils standard may go this route, and it would significantly improve the quality of soil sampling. In the interim, site personnel need to be aware of the potential issues, and to implement appropriate procedures to ensure data obtained from the soils is as meaningful as possible.


BS 10175:2001 Code of Practice for the Investigation of Potentially Contaminated Sites

Secondary Model Procedure for the Development of Appropriate Soil Sampling

Strategies for Land Contamination

Environment Agency R&D Technical Report P5-066/TR

ISO 10381-2 Sampling – Guidance on sampling techniques

ISO 18512 Guidance on long and short term storage of soils

Environment Agency: Performance Standard for Organisations Undertaking Sampling and Chemical Testing of Water MCERTS v.1 July 2008

Environment Agency: Performance Standard for Laboratories Undertaking Chemical Testing of Soil MCERTS v.3 March 2006

DST supported by the Technology Strategy Board (sponsored by DIUS)

Project number TP/5/CON/6/I/H0065B

CLAIRE conference proceedings (21/4/09) contentmanagement.servlet.ContentDeliveryServlet/IPM-NET/ 1%253A%2520%2520Knowledge%2520and%2520Skills/Emerging%2520Technologies/EnvMeasurementsProceedings-Issue1.pdf

Copy of DST presentation – acknowledgements to Mike Ramsey and Katy Boon 510a15c9bb85560680e1a0/

Published: 10th Mar 2009 in AWE International

Popular Articles by Hazel Davidson and Geraint Williams