Biological variables as soil quality indicators: Effect of sampling time and ability to classify soils by their suitability

Soil biological variables are considered good soil quality indicators due to their high sensitivity and ability to reflect soil management effects. However, they frequently show high temporal variability. Our objectives were: (a) to analyze temporal stability and seasonal effect on biological variables, (b) to choose between autumn and spring to sample for soil biological variables, and (c) to determine biological variables able to discriminate among selected soil subgroups. Areas with minimal human disturbance were sampled in three soil orders (Mollisol, Vertisol and Alfisol) during two and a half years, each autumn and spring. Microbial biomass C and N (MBC, MBN), basal respiration (Resp), metabolic quotient (qCO2), potential of N mineralization (PNM-AI), soil organic C (TOC) and total soil N (TON) were measured in three composite soil samples collected from homogeneous areas at 0-15 cm depth. For the studied soils, selected soil biological variables presented different levels depending on the time of sampling, spring or autumn. Hence, the importance of pointing out the time of sampling to report results of these variables in this kind of studies is remarked. In general, biological variables presented higher stability when we sampled soils in autumn compared to spring. Because of this, we used autumn soil samples to determine the best soil biological variables to discriminate among selected subgroups of soils. The separation of soil subgroups by means of discriminant analysis using just TOC and TON was scrutinized, considering that these soil variables are routinely measured in soil test laboratories. Nonetheless they were not able to discriminate properly among soil subgroups because they showed high error rates classifying the samples in the correct subgroups. In contrast, the variables PMN-AI, MBC, and MBN adequately discriminated the five soil subgroups. From the biological variables, PMN-AI and MBC were the best ones to characterize (discriminate) among the five soil subgroups. Particularly, PMN-AI was able to separate soils by their suitability for agricultural purposes.