Linking Soil Structure and Soil Habitat at Scales Appropriate to Microinvertebrates and Ecological Function.

Our aim was to develop scale-appropriate, three-dimensional estimations of hierarchical soil structure for use in spatially explicit, basic ecological studies of soil microfaunal population dynamics, habitat partitioning and their effect on nutrient cycling. Soil pore habitat heterogeneity relates directly to the biological diversity in soils. Soil pores are hierarchically structured and heterogeneous across spatial scales ranging from microns to meters. Biological interactions alter pore structure, and interactions with moisture alter these habitats on a continuum, ranging from hours to seasons. Few models of soil structure exist at the spatial resolution required to examine soil microfauna habitat and contributions of microfauna to N mineralization and compensatory microbial growth that results from grazing. This project specifically addresses macro- and micro-porous regions and spatial scales from microns, centimeters and meters. At meter scales, 208 geographically referenced determinations of soil aggregates suggest spatial autocorrelation of aggregates 2-cm and greater. Plant roots appear to account for this correlation. At micrometer scales, there are 30 soil thin sections with five unique fields of view. Applying an existing method of modeling soil structure using conditional probabilities to predict pixel values resulted in an inaccurate image. This result was determined to be related to both resolution and the high proportion of pore volume in our study soil. A subsequent strategy used binned pixel classes to constrain the pixel model. That resulted in accurate two and three dimensional simulations. Conditional probabilities were collected using 105 images. Void area averaged 37.6% ± 9.5 in these images. Resulting simulations also converged on 37.6% void volume.  Void volume could be increased to 47% by increasing the probabilities of transitional classes in the binned model. Visual inspection indicates that the model produces numerous blind pores connected to the macroporous network. Work is proceeding to extract pore geometry and topology to determine unsaturated water dynamics.