Detecting change in forest floor carbon

Changes over time in forest soils are important to global C balance and to local ecosystem function. Detecting change in C storage in the forest floor is hampered by high variability and the use of study designs that are not adequate to statistically detect change. Using estimates of variability from previous forest floor studies, mostly conducted in the northern USA and Canada, we conducted statistical power analyses to assess the ability of such studies to detect various magnitudes of change in forest floor C. The studies we surveyed were unable to detect statistically significant changes in forest floor C or mass smaller than 15 to 20%. Studies that remeasure plots or sites (i.e., paired designs) have greater statistical power to detect changes than those in which experimental units are independently located for the two sampling dates. The causal mechanisms of forest floor change influence the magnitude of the change, and accordingly our ability to detect such changes. The direct effects of climate change may be too small to be detectable by current designs, but larger changes in forest floor mass resulting from forest management, changes in tree species, changes in fire regime, or the introduction of earthworms are more likely to be detectable. With paired resampling and more efficient allocation of sampling effort, it should be possible for future studies to detect smaller changes.