干旱区生态环境知识资源中心

Factors determining the distribution and structure of soil and litter macrofaunal assemblages remain still poorly understood, despite the overriding importance of the spatio-temporal mosaic of biotic and abiotic conditions as main drivers of soil biota and processes. Analysis of the effects of different factors on soil communities have been usually restricted to responses to litter, despite the fact that litter and mineral soil layers are connected. Therefore, whether organisms using the litter layer respond to the same biotic and abiotic factors as organisms using the mineral soil still remains poorly known. We hypothesize that the role of biotic and abiotic factors as determinants of the distribution of faunal components of soil communities differ between litter and mineral soil assemblages in arid systems and that both levels are connected by animals moving across both levels. During two years, macroinvertebrates were sampled in litter and soil at an arid region of SE Spain, and different biotic and abiotic factors were measured. We performed structural equation model analysis to uncover the factors related to macrofaunal distribution. Our results show that abiotic factors, litter production and litter and root quality, as well as relationships among different trophic groups were key factors affecting faunal densities in our system. While abundance variations in litter assemblages were principally related to temperature and moisture, belowground faunal densities responded to resource factors. Despite differences in structuring factors at both levels, faunal interactions link both assemblages across the litter-belowground interface. The results highlight three important issues to understand soil communities and food web structure. First, abiotic factors structure soil macrofaunal food webs directly and indirectly, because of the effect of litter as habitat, and not only as food. Second, overlooking the differences found between above and belowground regulation may cause problems in the interpretation of food web structure and dynamics. Third, our models also suggest that both litter and belowground assemblages are dynamically connected. (C) 2009 Elsevier Ltd. All rights reserved.