| Article ID | Journal | Published Year | Pages | File Type |
|---|---|---|---|---|
| 1043559 | Quaternary International | 2010 | 10 Pages |
Sediment–paleosol-sequences are natural archives for a variety of proxies. Sampling of a profile is thereby a primary part in the analysis. The main goal of this study is to uncover the influence of the sampling resolution on the resulting classification of a profile. Therefore a new profile was investigated nearby the famous paleosol-sequences of “Stillfried B”. After the field interpretation of this profile a 180 cm part was sampled with 1 cm resolution. The sampling was taken out of a monolithic block that is located orographically left within the new profile. The following laboratory analysis was done without referring to the field interpretation, only by an arithmetic distance. From each sample, five particle-sizes and additionally the carbonate and dithionite iron content were measured. For all seven variables, 5 cm subsamples centred on the group mean depth with resolutions of 5, 10, 15 and 20 cm were calculated. For each resolution and variable, kriging was used to interpolate on the 1 cm resolution. The five data sets, which were constructed from the same profile, were then used in a cluster analysis. According to the results of the classical pedological analysis, this profile had to be classified into 5 groups. After optimising each grouping using discriminant analysis, the resulting classifications for the lower resolutions were compared with the one based on the 1 cm measurements. Surprisingly, the differences in the group boundaries also for 20 cm resolution are acceptable. With one exception, the resulting boundaries were in agreement with the pedological classification also. Taking the structural differences of the variables into account, for the profile under investigation, 10 and/or 15 cm would be an optimal sampling distance. However, the 20 cm samples also include enough structural information about the profile.
