Soil heterogeneity in arid shrublands: Biotic and abiotic processes

This dissertation investigates the evolution of soil heterogeneity on arid alluvial fans. Soil formation begins when depositional or erosional processes cease and a geomorphic surface becomes stable. The transformation of sediment to soil (pedogenesis) affects a multitude of hydrologic processes operating across a range of spatial and temporal scales. At the pore scale (mum--mm), soil development results in the formation of structure and secondary porosity. At the field scale, vegetation patches and open soil (or interspace) affect infiltration, partitioning of evapotranspiration, and ultimately vegetation performance and community structure. This feedback between vegetation and soil development is not well-defined in either time or space, and thus predicting their connection to ecosystem production or change is difficult. Ultimately, understanding how small-scale variability impacts larger scale processes is critical to examining whole ecosystem processes. The biotic and abiotic processes that shape arid landscapes operate at different spatial and temporal scales but ultimately control soil moisture variability and ecosystem patterns and processes. This dissertation attempts to expand the canopy interspace binary model of alluvial fans by (1) determining the spatial structure of hydraulic properties from canopy to interspace microsites, (2) quantifying, scaling and defining its spatial extent, (3) investigating the bio-abiotic factors that control the spatiotemporal heterogeneity of arid alluvial soils, and (4) implementing inverse modeling and field characterization to parameterize the vadose properties of these arid soil systems. We use high-spatial resolution measurements of unsaturated hydraulic conductivity, non-dimensional spatial correlation, numerical modeling and multiobjective parameter optimization to address the role of bio-abiotic processes affecting both the spatial and temporal variability of soil moisture in arid shrublands.;Our results support the biotic feedback between soil and hydrologic processes. On young, weakly structured soils with high overall conductivity, this biofeedback may act to retain moisture and not simply increase saturated conductivity underneath vegetation. Spatially, we found a gradient in hydraulic and physical soil properties that extended 1.4 times the canopy radius. Through the use of a chronosequence, much of the soil heterogeneity could be explained spatially but its origin was a coupling of biotic and abiotic processes. Hydraulic properties were weakly correlated in space but 75% of the variance could be attributed to sand content, soil structure grade, mean-particle diameter, and organic matter. Lastly, multiobjective parameter optimization using the AMALGAM algorithm found the effective parameters of the near-surface vadose zone to be highly sensitive to factors often ignored in more mesic applications. In particular, parameters associated with the dry-end of the retention curve were well-defined while saturated conductivity and air-entry lost physical meaning or certainty. The coupled use of water content and potential data as minimization criteria through inverse modeling yielded more realistic parameters than either set alone with no single set of parameters being capable to minimizing both data sets.