Regional organic carbon stock variability: A comparison between depth increments and soil horizons

Soil organic carbon (OC) stocks are highly variable in space. This study aimed at analysing regional soil OC stock variability with direct measurements of OC stocks at 89 sampling points distributed across a forested area (160 km^{- 2}) of the Hainich region in Thuringia/Germany. We sampled complete soil profiles by horizons and depth increments. The major goal of this study was the analysis of regional OC stocks. Soils of our study area were classified as Cambisols, Luvisols and Stagnosols. The averaged total OC stocks of the solum ranged from 7.4 kg m^{- 2} in Luvisols to 10.3 kg m^{- 2} in Cambisols, but the analysis of variance revealed no significant differences among the soil groups. In contrast, significant differences in A and B horizon OC stocks of soil groups could be detected. This was not the case, when depth increments instead of soil horizons were studied. Our results implied that sampling by depth increments is favourable for the detection of regional OC stocks and that sampling by soil horizons should be used for the study of pedogenic processes which control OC stocks in soil. Differences in OC stocks among soil groups are more pronounced if horizons are considered, which shows that mixing of soil horizons by fixed depth increments destroy important pedogenetical information. A stepwise multiple regression exposed horizon thickness as the most important factor affecting OC stocks in soil horizons. This has major implications for detection of changes in carbon stock, since the variability of regional OC stocks is less controlled by OC concentration and bulk density rather than by soil horizon thickness. To detect differences smaller than 10% of regional OC stocks one would need more than 300 A horizon samples. If 0-10 cm increments were sampled, less than 100 samples would be sufficient to detect the same difference.