Comparative soil sampling in the Dornach site (Switzerland) for soil three-dimensional pollution description.

Fifteen institutions from 13 European countries and Switzerland participated in a comparative test sampling at the Dornach site, near Basel in Switzerland. They received the site description and were asked to develop their own sampling plans, based on their national guidelines for a three-dimensional description of the Pb, Zn, Cu and Cd pollution, with a maximum of 15 samples. The comparative sampling test took place in late August 1997. The sampling plans and records, provided by the participants, were compared and evaluated in terms of sampling strategy throughout the site, the sampling strategy inside the sampling unit, strategy at the sampling point and last minute alteration of the sampling plans in the field. The object of this was to define the similarities and the differences in the sampling plans and identify the crucial steps inducing error or uncertainty which should be harmonized throughout Europe. The number of the total composite samples varied from 4 to 16, while the number of sampling points varied from 4 to 224 and the sampling density varied from 0.25 to 16 single samples per 100 m2. The number of 10 x 10 m2 quadrats from which samples were taken varied from 3 to 56, out of the total 61. The number of sampling units varied between 3 and 12. The criteria in descending order of frequency were: both land use and pedology, land use, pedology. The majority of the national participants divided the sampling units into two or three sub-units, while seven participants collected replicate composite samples from the same sampling unit. Only one sampled from the existing profiles, eight national representatives ignored the plough depth. The sampling pattern inside the sampling unit was in descending order of frequency: random, point sampling, regular, 'W' and 'X' shaped sampling. The number of single samples in the composite varied from 1 to 20. Most of the national participants sampled to pre-defined (ad-hoc) sampling depths, while others sampled soil horizons and others both. The maximum soil depth sampled also varied from 20 to 117 cm. Most participants mixed in the top organic matter horizon while others sampled it separately. Most sampled separately the top Ah horizon in the forest while the rest mixed it with the underlying soil horizon. The last-minute sampling plan alterations were in: the sampling depth (nine), the movement of sampling points (five), the reduction of sub-samples in the composite sample (four), changes in the number of sampling units (three), ignoring the bottom sample (three), not sampling the L/H layer (three) although this was planned, unplanned sampling of the L/H layer (two) and changing the number of planned samples (two). This study has highlighted the need for soil sampling harmonization throughout Europe for pollution and soil quality determination.