In sandy soils as they have been investigated in this study, GPR is a valuable technique for mapping soil water content at an intermediate scale in between point (TDR) and area measurements (remote sensing). The observed strong correlation between the change of the reflection time of the electromagnetic wave and the change in soil water content as determined from gravimetrical measurements (r2 = 0.82) confirmed the accuracy of the GPR technique. The popular Topp equation proved suitable for the registration of the seasonal water status development in the top soil. However, the analysis of the velocity of the electromagnetic wave alone does not provide depth-dependent information about the soil water regime. A more detailed inspection of the radargram, especially of the reflection amplitudes, is necessary when information about a change of soil structure or soil water content with depth is needed. In addition, the statistical analysis of the reflection amplitudes or other parameters obtainable from the GPR measurements might be a way to reveal the heterogeneity of the soil water content and the related water flux field.

First results presented in this study indicated that the assumption of a heterogeneous distribution of soil hydraulic properties coincides better with the geophysical data than a generated homogeneous case (homogeneous soil properties within a single soil layer). It can be concluded that even in homogeneous sand profiles, local variabilities dominate water flow which confirms previous findings from other studies in which, for instance, the dye tracer technique has been used to identify flow patterns (Flury et al., 1994). The variances of the soil hydraulic and geophysical parameters differed, although in general the heterogeneous case performed better than the homogeneous case.

The next step essential for evaluating the potential of the method would be using the analyzed variance of GPR amplitudes for each depth interval directly in the numerical models for simulating water flow. Comparisons of water discharge rates derived from numerical models based on soil hydraulic parameter distributions determined from the GPR analysis with real measured data will show if a better agreement of model and experiment can be reached. We expect a better description of the water flux variation by using the information about soil heterogeneity derived from GPR measurements.

35 cm

125 cm

0 0

Post a comment