Field Analysis:Field Sampling

SASSA Home Page &#8658; Field Analysis Home Page &#8658; Field Sampling

Sampling of soils and sediments associated with archaeological sites
SASSA is designed to support the use of analytical techniques to aid the interpretation of archaeological sites. Clearly such techniques can only be applied to a very limited amount of material collected from or near archaeological sites, thus sampling is inevitable and essential. As in all field sciences careful thought needs to be given in geoarchaeological investigations to the most appropriate sampling methods. The well-quoted statement that results are only as good as the samples is especially applicable to geoarchaeology. It is also interesting to reflect that sampling costs are often miniscule compared to subsequent analytical costs, an unfortunate balance since much time and effort should be spent on designing and executing a sampling strategy.

The whole aim of sampling is to be able to characterize a deposit based on analysis of a limited amount of material. This immediately raises two fundamental points about sampling: (1) there needs to be from the beginning of an investigation a clear understanding of WHY sampling and subsequent analysis are potentially important for answering specific interpretational questions – in other words the approach much be problem driven. Samples collected from sites without a research plan are unlikely to be of much value and are usually ultimately binned. (2) Sampling contexts need to be carefully defined so that generalisations from the results can be made. The use of the field tool in SASSA can play a key role in this regard.

The following general principles apply to the sampling of soils on archaeological sites, and are in part based on Quine (1995):

1. The design should incorporate sampling defined contexts in both vertical and lateral sequences since ultimate interpretation of results is dependent upon evaluation of variation in analytical results.

2. A consistent sampling approach should be followed whereby the same volume of soil should be extracted at each sampling point.

3. Interpretation of single point analytical values is problematic. Priority should also be given to sampling other contexts including the underlying soil parent material as well as soil in the wider locality.

Besides on-site sampling, geoarchaeologists may well need to sample soils on a wider landscape basis, as is the case for geochemistry. In studies of this type the aim is to identify spatial patterns in elemental loadings which can be interpreted in terms of former land management practices. Much work has been done on spatial variability in phosphorus which is often taken as an indicator of former anthropogenic activity. Sampling at the landscape scale raises a different range of challenges to on-site work. If differences in elemental loadings are to be investigated, then statistical analyses including ANOVA will be essential. Thus sampling will need to incorporate the requirements for the inferential testing of differences and a random component to sampling is essential. Readers seeking guidance on sampling design at the landscape scale should consult any of the standard statistical texts, all of which outline different approaches, for example, random, stratified random and systematic. Besides the issue of the spatial sampling pattern, consideration has also to be given to sampling depth, for example, fixed depth or based on a particular soil horizon. As always, the question which is being tackled should govern details decisions on sampling strategy.