Plot-scale variability of organic carbon in temperate agricultural soils—Implications for soil monitoring#

Abstract

Background: Detecting changes in soil organic carbon (SOC) stock requires systematic and random sampling errors to be kept to a minimum. Especially in soil monitoring schemes based on soil profiles pits, it is important to understand if a minimum spatial shift of that profile pit during resampling could render resampling errors caused by spatial variability negligible. Aims: We aimed at (1) quantifying the random SOC stock error caused by a minimum shift in sampling location of one profile and (2) assessing whether an increase in the number of profile pits to three could significantly decrease the resampling error caused by spatial variability of the relevant parameters. Methods: Eight croplands and grasslands in northeast Germany were sampled. Three sampling designs were compared: one profile resampled (1) by one, (2) by three profiles or (3) three profiles resampled by three. In addition, 16 soil cores were taken per site to characterise overall plot-scale heterogeneity and assess general patterns of spatial dependence of relevant parameters. Results: Spatial dependence of all assessed parameters was weak. Accordingly, the resampling of one profile by one induced a high mean absolute error of 5.1 and 7.6 Mg C ha–1 at a 0–30 cm depth for croplands and grasslands (7.5% and 8.5%). This error was reduced by approximately 50% when three profiles were resampled by three profiles. Conclusions: Even with the smallest spatial shifts possible, monitoring of SOC stocks relies on replicated resampling to detect management or climate change-induced trends in reasonable and relevant timescales.