A comparison of GPR early-time signal approaches for mapping changes in shallow soil water content

We employ early-time signal (ETS) amplitude analysis of ground-penetrating radar (GPR) data collected during a controlled irrigation study in a sandy soil agricultural field. We carried out two irrigation experiments, one with 250 MHz, 400 MHz, and 1000 MHz GPR antennae, and a second with 1000 MHz antennae only. Supporting data on the effects of irrigation on soil water content were provided by direct measurement and the extent of soil wetting by electrical resistivity tomography (ERT). We compare two statistics used to quantify the ETS: (1) average envelope amplitude (AEA), and (2) carrier frequency amplitude (CFA), which has recently been proposed as a more robust alternative to AEA analysis. We also investigate the dependence of these statistics on the time window used to calculate them. In the first experiment, both statistics successfully relate the GPR ETS to increasing water content for the 250 MHz and 400 MHz antennae, although the CFA was sensitive to changes across a wider range of moisture contents. The 1000 MHz data did not show the expected relationships, especially in the CFA data. This is likely due to signal associated with shallow reflections. In the second experiment, the effect of this signal was removed by using a shorter, earlier time window when calculating ETS statistics, rather than using the first positive half cycle (FPHC). Our work suggests the cautious use of time window for ETS analysis, as the FPHC may not be an acceptable choice when shallow reflectors are present. At our field site, the AEA was less affected than time window selection while CFA appears more robust over a wider range of soil moisture content values.