We determined the complex permittivity and clay mineralogy of grain-size fractions in a wet silt soil. We used one clay-size fraction and three silt-size fractions, measured permittivity with low error from [Formula: see text] with time-domain spectroscopy, and estimated mineral weight percentages using X-ray diffraction (XRD). The volumetric water contents were near 30%, and the temperature was [Formula: see text]. For the whole soil, standard fractionation procedures yielded 2.4% clay-size particles by weight, but XRD showed that the phyllosilicate clay minerals kaolinite, illite, and smectite made up 17% and were significantly present in all fractions. Above approximately [Formula: see text], all real parts were similar. Below approximately [Formula: see text], the real and imaginary permittivities increased with decreasing grain size as frequency decreased, and the imaginary parts became dominated by direct-current conduction. Similarly, below approximately [Formula: see text], the measured permittivity of montmorillonite, a common smectite, dominated that of the other clay minerals. Total clay mineral and smectite mass fractions consistently increased with decreasing grain size. Below [Formula: see text], a model with progressively increasing amounts of water and parameters characteristic of montmorillonite matches the data well for all fractions, predicts permittivities characteristic of free water in smectite structural galleries, and shows that the similar real parts above [Formula: see text] are caused by a small suppression of the high-frequency static value of water permittivity by the smectite. We conclude that the clay mineral content, particularly smectite, appears to be responsible for permittivity variations between grain-size fractions. Small model mismatches in real permittivity near the low-frequency end and the greater fractions of kaolinite and illite suggest that the total clay mineral content might have been important for the coarser fractions.

We used [Formula: see text] ground-penetrating-radar (GPR) reflection profiles to determine if high-explosive (HE) depth bombs were possibly detonated within a small New England lake formerly used as a bombing range, as well as the depths and distribution of inert and possibly unexploded ordnance (UXO) within the subbottom sediments. Maximum water depth was [Formula: see text]. The primary munitions were unarmed practice bombs, but terrestrial fragments from HE depth bombs led us to seek stratigraphic obliteration and distortion as evidence of their use. The low water conductivity and quartz-rich sediments provided more than [Formula: see text] of subbottom penetration beneath slope and shelf sediments. The high permittivities of the water and subbottom sediments provided good antenna directivity, excellent 2D migrations, multicentimeter vertical resolution, and a unique pulse waveform signature for metal objects. Our line profiles revealed hundreds of subbottom diffractions, which implied an areal distribution of thou-sands of suspicious objects. Most lay beneath the central basin, but many resided beneath the slopes and shelves as well. Supplementary profiles of a nearby pristine lake and of controlled munitions, a side-scan sonar bottom survey, and time-domain reflectometry dielectric measurements of sediment samples corroborated our findings of natural stratigraphy, a unique metallic diffraction phase signature, the existence of many nonmetallic objects including logs, and subbottom relative permittivities ranging from 28 to 76. The intact natural lacustrine strata, including deltaic foresets and bottom sets, faults, and conformable and unconformable horizons, suggest that live bombs likely did not detonate in the lake. The only stratigraphic disturbances we found were apparent perforations beneath about [Formula: see text] of water, within which diffractions of metallic origin originated. The dielectric results and the prominent slope strata over the deepest horizons led us to conclude that the exceptional subbottom penetration existed under predominantly sandy sediments.

No abstract