Long-term effects of conventional tillage and no-tillage on saturated and near-saturated hydraulic conductivity - Can their prediction be improved by pore metrics obtained with X-ray CT?

Tillage practices have a profound impact on soil structure and soil hydrology, which may affect ecosystem functions like plant productivity. There is an ongoing debate whether a conversion from conventional tillage (CT) to no-till (NT) leads to an increase in (near-)saturated hydraulic conductivity. This is because true effects are often disguised by large spatial and temporal variability, but also by the deficiencies in the measurement technique. In this paper, we measured (near-) saturated hydraulic conductivity (K_S and K-₂) in a long-term tillage trial (26 years) in Germany with three different methods: hood infiltrometer (HI) in the field, tension disc infiltrometer (TI) on undisturbed soil cores and direct simulation (DS) of water flow on X-ray CT images of macropore structure in these soil cores with a Stokes-Brinkmann solver. On average the absolute values varied by two orders of magnitude in the order TI < HI < DS with very low correlation (R² < 0.05) between log₁₀(K_s) measurements. The conversion from CT to NT caused an increase in bulk density, a decrease in air capacity and a small but consistent decrease in grain yields. K_s was increased with NT when measured with HI but decreased in TI and indifferent when measured with DS. This inconsistency is caused by the proportion at which large biopores that are more frequent in NT soil due to higher earthworm abundance contribute to total flow in each measurement technique. Regression analyses between pore space attributes measured with X-ray CT and K_s (and K_-2) showed very strong agreement with DS values, but poorer agreement with HI and TI, suggesting that those values are afflicted with measurement artifacts like poor contact, entrapped air, different average volumes and so on. The pore metric with highest predictive power (> 90%) on simulated K_s in NT soil cores was the critical pore diameter because it represents the bottleneck that restricts a large contribution to flow by elongated biopores. However, in plowed soil (CT) pore metrics that best describe flow through the loose soil matrix, like macroporosity and pore connectivity, have a higher predictive power and the critical pore diameter is rendered meaningless. The relative importance of various pore metrics as good predictors of hydraulic conductivity does not only change in a very small pressure range (K_s vs. K_-2) but also between measurement techniques (HI vs. TI vs. DS). These inconsistencies raise the question if and how existing pedo-transfer functions for estimating (near- ) saturated hydraulic conductivities can be extended by image-derived pore metrics in a meaningful way.