Infiltration as dynamic non-uniform stochastic flow field – repeatable, high-resolution 4D GPR measurements at the plot scale

Infiltration is a central concern in soil physics. The advective and diffusive redistribution of event water depends on various factors such as initial wetting, the establishment of film connectivity, and capillary gradients and hydraulic conductivity. Non-uniform infiltration patterns are prevalent. However, direct infiltration measurements do not account for this reality and tracer experiments require a destruction of the experimental plot. We developed a data acquisition strategy based on time-lapse 3D ground-penetrating radar (GPR) to monitor fast and small-scale subsurface flow processes during irrigation in a non-invasive manner. The technique combines an irrigation pad (1 m2 drip irrigation to simulate moderate, non-erosive rain events) with a GPR measurement platform (manually driven two-channel GPR antenna array with positioning guides). We will present a systematic field experiment consisting of two recurrent irrigations (40 mm/2 h, 1 irrigation per day) and a respective replicate. For evaluation, the GPR measurements were sidelined with soil moisture measurements (TDR profile) and tracer applications (dye and salt). Our data show that the achieved high resolution of less than 5 cm in space and 10 minutes in time makes it possible to monitor and quantify infiltration processes in their spatial and temporal non-uniformity. The experiment supports the hypothesis from earlier experiments at various sites: Non-uniform infiltration patterns and dynamically connected flow-fields are highly heterogeneous but share stochastic features, such as distribution, randomness, and skewness. Our approach opens new options for repeated, spatially resolved infiltration measurements and theory development for soil hydrology and land surface models.