The depth of sampling has recently been highlighted as critical to making accurate measurements of changes in SOC stocks. This paper aimed to determine the effects of land management changes (LMC) on soil organic carbon (SOC) by re-sampling long-term agoecosystem experiments (LTAEs) across Canada using identical sampling and laboratory protocols. The impact of sampling depth on the monitoring of LMC-induced differences in SOC stock in LTAEs in Canada, and the implications on statistical power and sampling design, were assessed. In most cases, four cores would be suitable for detecting a significant difference in SOC stock of 5 Mg ha -1 at 95% confidence for LMCs in western Canada. The impact of eliminating fallow on SOC stocks was typically restricted to the surface 15 cm. The impact of perennial forages on the average cumulative SOC was sufficiently large to be detectable at all sampling depths (to 60 cm). In three of the six LTAEs sampled in western Canada comparing conventional tillage to no-till, there was a significantly greater SOC storage in the 0- to 30-depth than the 0- to 15-cm depth, suggesting that sampling below 15 cm could be necessary. The same comparisons in eastern Canada suggested that sampling often must exceed the 30-cm depth to account for any changes in SOC due to moldboard plow tillage. Nonetheless, there was little evidence to suggest that increasing sampling intensity or sampling deeper would improve the ability to detect a difference in SOC stocks for this LMC.
