Connecting soil structure and hydraulic properties under different land use on a European climate gradient

The infiltration of water into the soil, especially during extreme rainfall events, is controlled by soil hydraulic properties such as saturated hydraulic conductivity. Usually, the saturated hydraulic conductivity of the soil at larger scales is estimated by pedotransfer functions that use easily available soil properties such as soil texture, bulk density, and soil carbon content. Unfortunately, it has been already shown that those predictors do not contain the information for precise prediction of the saturated hydraulic conductivity. Moreover, it is widely accepted that the soil structure caused by aggregation, which defines the soil pore network, are important characteristics towards correctly estimating the saturated hydraulic conductivity.

To analyze and quantify the impact of aggregation on the saturated hydraulic conductivity we combined analyses of soil structure based on drop-shatter tests and aggregate size fractionation, with analyses of infiltration pathways via dye tracer application and in-field infiltration measurements. As soil structure is strongly influenced by soil management and climate, we sampled croplands, grasslands, and forests along a European climate gradient.

Our results indicate that both soil structure parameters and the classical predictors used in pedotransfer functions (soil texture, bulk density, and soil carbon content) had a significant influence on the saturated hydraulic conductivity. Regression models using soil structure parameters had a very similar Aikaike Information Criterion (AIC) as regression models without taking soil aggregation into account. This was different for the near-saturated hydraulic conductivity (K_{-2 cm}), where the regression models based on soil texture, bulk density and soil organic carbon content  performed better than the model using soil structure parameters. Additionally, it was found that landuse and plant type had a large influence on soil structural parameters. We found less stained areas (0- 30 cm depth) in forests than in croplands and grasslands, which indicates more occurrences of preferential flow, and this was also  negatively correlated with the initial soil moisture at the time of measurement. In addition, higher levels of aggregation, indicated by a higher mean weight diameter of the soil aggregates, was associated with higher preferential flow as indicated by the dye tracer Both, stained area and peds and clods were influenced by the plant type as well, since the sites with vegetation having predominantly fibrous root systems responded differently than the sites with tap-rooted plants, trees, or heathland vegetation. The enhanced information on soil structure can therefore help us better understand landuse and land cover effects on saturated hydraulic conductivity and soil water infiltration.