Geophysical instruments show great potential for the detailed quantification of soil stratigraphy. In this study, two electromagnetic induction sensors were evaluated on their capacity to map small‐scale variations of the depth to the interface (zin) in a two‐layered soil. On a 2‐ha study site, zin between the silty topsoil and the contrasting clayey subsoil was modeled first by relating the two apparent electrical conductivity (ECa) measurements of the EM38DD sensor to observations of zin obtained by augering. A substantial number of these calibration observations was needed, however, to account for the modeling parameters. To avoid this step, an entirely noninvasive procedure was proposed based on one survey with the DUALEM‐21S sensor. This sensor simultaneously records four ECa values with different coil configurations. These measurements correspond to four different depth response functions that allow modeling zin without calibration observations. The only assumption was a two‐layered soil profile. The zin predictions were validated with 24 independent depth observations. Both procedures resulted in equal correlation coefficients (0.85) between predicted and measured zin and average estimation errors (0.26 m). This indicated that both sensors allowed the accurate mapping of the depth to a contrasting textural layer. With the EM38DD, calibration observations are needed, whereas the four different coil conurations of the DUALEM‐21S sensor provided sufficient information to predict the interface depth without augering.

No abstract

The multiple coil configurations of two electromagnetic induction sensors were tested on a field with strong electrical and magnetic contrasts. The first sensor, EM38DD, measures either the apparent electrical conductivity (ECa or s a ) or the apparent magnetic susceptibility (MSa or x a ) of the soil at two coilorientations.The secondsensor,DUALEM-21S, measuresboth ECa and MSa at two coilorientations and two coil separations. The goal was to test if measuring with the multiple coils resulted in a better detection of near-surface artefacts and the natural soil variability.The ECa of all coil configurations was closely related to the depth of a clay substrate beneath the topsoil sandy loam, which was verified by soil augering. Configurations with a shallower theoretical depth of exploration were less influenced by the clay substrate. Combining two coil configurations revealed important ECa anomalies, not visible on individual measurements, associated with a brick wall foundation and a formerditch.The MSa maps showed totally different anomaly patterns, related to anthropogenic disturbances in the soil, such as the filling-in with brick rubble of a former pond. Depending on the depth and thickness of the disturbance and the relative response of the sensor configurations, the MSa anomalies were entirely positive for one configuration but other configurations also had negative anomalies. It was concluded that multiple coil configurations provide a better insight into the buildup of the soil profile and are better able to detect anomalies than single measurements.

No abstract

Understanding palaeotopographical variability forms the basis for understanding prehistoric societies.Alluvial and lacustrine environments, in particular, are key areas with both a high archaeological and palaeoecological potential. However, the often deep stratification of these sites, the high water table and the complex sedimentological variations can hamper a detailed reconstruction of the spatial relationship between prehistoric settlement and their environment. Combining different remote and proximal sensing techniques and coring data, can offer detailed insight into such landscapes. More specifically, the integration of mobile geophysical methods allows the collection of unprecedented continuous information on large-scale palaeolandscape variability. In this study we present a combined approach in order to map and model prehistoric landscapes and river systems in and around a Late Glacial palaeolake in north-western Belgium. Based on filtered and unfiltered digital elevation models, a survey area of 60 ha was selected, in which detailed mobile multi-receiver electromagnetic induction survey was conducted. The results allowed for the delineation of palaeochannels in the area and enabled modelling the depth of these features in the survey area, providing insight into their flow characteristics.14 C sampling enabled the dating of the evolving river system to the transition between the Late Glacial and the Early Holocene. Through additional coring, this river system could be traced further through the palaeolake area. Based on these results a detailed reconstruction was made of the palaeotopography that harboured the Final Palaeolithic andEarly Mesolithic occupation of the study site.