Representativeness of mixed samples for evaluation of organic carbon in soil

The aim to solve the methodological problem of assessing variability of soil organic carbon content (C org ) in layer 0–20 cm based on 10-fold independent mixed sampling from the same field during one day. The soil sampling was carried out on 4 fields ranging from 6 to 16 hectares (Tver region). The soil cover pattern is represented by patches of agrosod-podzolic gleyic sandy loam and loam soils (Albic Retisols (Aric, Loamic)) developed from binary deposits and agrosod-podzols (Albic Podzols (Aric, Arenic)) developed from sands underlying loamy moraine from the depth 80–150 cm. It has been established that the variation in C org content in mixed samples reflects the spatial component that is sometimes comparable to, but more often significantly higher than analytical reproducibility of laboratory methods for determination C org (Turin method and dry combustion method). To compare single measurements of C org in mixed samples obtained at different time periods or from different fields, the use of even the full confidence range based on the interlaboratory reproducibility of the determination method turns out to be an unreasonably low estimate. To conduct monitoring of C org content, it is necessary to assess the spatial variation of the indicator, even when selecting mixed samples.