Temporal and spatial variability in soil food web structure.

Heterogeneity is a prominent feature of most ecosystems. As a result of environmental heterogeneity the distribution of many soil organisms shows a temporal as well as horizontal and vertical spatial patterning. In spite of this, food webs are usually portrayed as static networks with highly aggregated trophic groups over broader scales of time and space. The variability in food web structure and its consequences have seldom been examined. Using data from a Scots pine forest soil in the Netherlands, we explored (1) the temporal and spatial variability of a detrital food web and its components, (2) the effect of taxonomic resolution on the perception of variability over time and across space, and (3) the importance of organic matter quality as an explanatory factor for variability in food web composition. Compositional variability, expressed using the Bray-Curtis similarity index, was measured over 2.5 years using a stratified litterbag design with three organic horizons per litterbag set. Variability in community composition and organic matter degradation increased over time in the litter horizon only. Seasonal variation in community composition was larger than variation between samples from the same season in different years. Horizontal spatial variability in community composition and organic matter degradation was relatively low, with no increase in variability with increasing distance between samples. Vertically, communities and organic matter degradation was more different between the non-adjacent litter and humus horizons than between adjacent layers. These findings imply that soil food webs, at least in temperate forest plantations, are more variable than is currently appreciated in experiments and model studies, and that organic matter turnover might be an important factor explaining variability in community composition. Species composition was more variable than functional group composition, which implies that aggregated food webs will seem less sensitive to local temporal and spatial changes than they in fact are. © Oikos.