Evaluating soil microbial biomass carbon as an indicator of long-term environmental change

The aim is to assess whether soil microbial biomass carbon (biomass C) could be used as an indicator of environmental change in natural and semi-natural ecosystems. Biomass C was measured by fumigation–extraction in soils from two sites at Rothamsted. One was a plot from the Broadbalk Wheat Experiment, given inorganic fertiliser and chalk, which has been in continuous cultivation for more than 150 yr. The other was a similar sized area, from Geescroft Wilderness, which has been left to revert to woodland since 1885, after being an arable field. Other soil properties (pH, soil organic C and exchangeable cations) were also measured to compare with biomass C. The coefficients of variation (cvs) of the properties measured were calculated for comparison, little difference was found between the cvs for biomass C from each site: cv=26% for Broadbalk and 23% for Geescroft. The cvs for the other, chemical properties, were mostly <10% for Broadbalk and generally >25% for Geescroft, as expected, given the different cultivation histories. Statistical analysis of the variation in biomass C concentration revealed that such measurements would not be valid indicators of environmental change, without processing impossibly large numbers of samples. To decrease the least significant percentage change to less than 5% after three samplings, 320 samples would have to be taken each time. This would be also be true of the other chemical properties in Geescroft Wilderness, where the measured background variation would mask any subtle environmental change. This indicates that, for some properties at least, statistically significant changes will only be detected in the longer term with regular sampling, e.g. 30–40 yr.