Imaging of hill-slope soil moisture wetting patterns in a semi-arid oak savanna catchment using time-lapse electromagnetic induction

Abstract

► In this study we examine catchment scale soil moisture change. ► We use geophysics to non-invasively investigate wetting patterns. ► We use EC a differencing to eliminate soil property variation effects. ► Results compare well with TDR soil moisture measurements. ► We conclude that EC a differencing improves EC a estimation of catchment soil moisture. Soil moisture ( θ ) is a fundamental hydrological state variable and its spatial pattern is important for understanding hydrological processes. Determination of small catchment-scale soil moisture status and distribution at intermediate scales (0.01–1 km 2 ) is challenging. Primarily because multi-point measurements using sensors are often impractical, while remote sensing resolution is often too coarse. Geophysical methods, e.g. electromagnetic induction (EMI), offer potential for bridging this gap. Our objective was to test the use of time-lapse EMI surveys to separate the influences of ‘static’ soil variables, e.g. texture/mineralogy, from ‘dynamic’, e.g. changes in soil moisture. A novel differencing approach is used for estimating relative changes in soil moisture, subtracting the bulk soil electrical conductivity (EC a ) of the driest seasonal soil map, named the residual EC a , from the EC a collected during subsequent wetting. This approach allows us to remove the effects of spatially distributed mineral electrical surface charge (EC s ) and other factors, and improve estimation of water content. Comparing results with TDR determined soil moisture, we improve the correlation from r 2 = 0.28 to r 2 = 0.48. EC a measurements are observed to be correlated in time ( r 2 > 0.7), but fall into two distinct groups, corresponding to times before and after the onset of stream flow, supporting the concept of preferred soil moisture states. Catchment wetness index predicts areas of convergence resulting in overland flow and stream flow. However, the spatial pattern of soil moisture does not mirror the wetness index, as others have found. We contend that the use of time-lapse EMI imaging provides important insight into the distribution and dynamics of catchment-scale changes in soil moisture, but acknowledge its limitations of requiring moisture dependent contrast of EC a , which will not occur in some soils.