No abstract

Soil water repellency (WR) is a widespread phenomenon caused by aggregated organic matter (OM) and layers of hydrophobic organic substances coating the surface of soil particles. These substances have a very low surface free energy, reducing a soil's water attraction.There is focus on WR due to its effects on germination, root growth, liquid-vapour dynamics, surface erosion and leaching of chemicals through fingered flow paths. However, common techniques for measuring WR are time-consuming and expensive. Meanwhile, it is well established that visible near infrared (vis-NIR) spectroscopy is a reliable method for determining soil OM. Potentially it could therefore provide fast measurements of WR through autocorrelation with OM. The aim of this study was to test the feasibility of vis-NIR spectroscopy for estimating the WR of soils with a small gradient in soil organic carbon (SOC) and texture, and to evaluate the effect of soil pretreatment on the predictive ability of WR models. A total of 87 soil samples from an agricultural coarse sandy field in Denmark were analysed for SOC, particle size fractions, water content and WR. Soil samples were scanned with a vis-NIR sensor (350-2500 nm) after air-and oven-drying at 60°C and 105°C. WR, expressed as liquid surface tension (mN m −1 ), was determined using the molarity of ethanol droplet test. Partial least squares regression models of SOC, texture and water content showed no predictive ability (r 2 values between 0.10 and 0.51). However, successful models (r 2 = 0.85) were generated for WR. The majority of bands important in the vis-NIR region of WR models were related to different components of OM indicating that, across the investigated field, WR was related to specific hydrophobic components of soil OM rather than to the total amount of carbon. A lower prediction error of the WR model for soils dried at 105°C (1.93 mN m −1 ) than at 60°C (2.52 mN m −1 ) can be explained by a lower range of WR values for the soils dried at 105°C. Moreover, a higher temperature reduced the number of absorption bands related to OM, indicating a degradation of hydrocarbon groups and a more hydrophobic character of the soil.

The gas diffusion coefficient, air permeability, and their interrelations with air-filled porosity are essential for characterization of diffusive and convective transport of gases in soils. Variations in soil bulk density can affect water retention, air-filled pore space, and pore-network connectivity and tortuosity and, thereby, control gas diffusion and air permeability. Considering 86 undisturbed core samples with variable bulk densities that were extracted on a 15 by 15 m grid from the top layer of a sandy field, the effects of soil bulk density on gas transport parameters and the soil water characteristic were investigated. Interactions with soil organic matter, sand, and clay fractions were also examined. To evaluate bulk density effects, two constitutive parameters were derived from each of the three measured relationships. The Campbell pore-size distribution index (b) and the air-entry matric potential (y ae ) were derived from the soil water characteristic; the diffusive percolation threshold (e DPT ), the air-filled porosity where gas diffusivity ceases to almost zero because of interconnected water films creating isolated-inactive air content, and a pore-network connectivity index (A 2 ) were derived from the gas diffusivity curve, and the analogous parameters convective percolation threshold (e CPT ) and convective pore-network connectivity index (B 2 ) from the air permeability curve. All six parameters showed significant negative correlations with bulk density. To further account for the effects of both bulk density and macroporosity in parametric gas transport models, a diffusive-analog macroporosity-dependent model (DAMP) for gas diffusivity and a generalized Kawamoto et al. model (GK) for air permeability, which yielded improved predictive capabilities when compared with previous models, were developed. Both new models apply a reference point of prediction at −100 cm H 2 O matric potential (macroporosity drained), corresponding to the point where analysis of pore-network tortuosity (T) and equivalent pore diameter for gas transport (d g ) showed diminishing effects of water blockage on gas transport in the sandy soil.