Using surface-NMR measurements to study the effects of compaction measures on the properties of lignite mining dumps

In the last 15 to 20 years, a sudden spike of liquefaction events after groundwater rebound on inner dumps in the Lusatian mining district resulted in around 30,000 hectares of land being closed to public access. One of the common modern compaction methods used is the gentle-blast-compaction (GBC), in which minimal explosive charges are placed in defined depth horizons (below the groundwater table) and detonated one after the other from the bottom upwards. The primary objective is to improve the ground stability by locally collapsing the pore structure of the material. This increases the bulk density of the dump material and reduces the air and waterfilled proportion of the pore space. Usually, direct geotechnical methods like drillings or cone penetration tests (CPT) are used to verify successful compaction. Within the “VerLaUf” project, we investigate the suitability of various airborne and ground-based geophysical methods for the non-invasive evaluation of these compaction measures. In the present study, we focus in particular on the applicability of two electromagnetic methods, transient electromagnetics (TEM) and surface nuclear magnetic resonance (SNMR). The TEM measurements are used to obtain a resistivity model of the subsurface which is needed for the inversion and interpretation of the SNMR data. Due to the direct correlation between SNMR signal amplitude and water content (porosity) as well as SNMR relaxation time and pore size, the SNMR method promises not only qualitative but also quantitative results about the change in the (water-filled) pore space after GBC.

Field campaigns were carried out over the course of three years, where the GBC took place after the first measurement campaign at depths ranging from 7 m to 32 m. The subsequent measurement campaigns were carried out after the GBC, with waiting times of approx. five and 15 months, respectively. The TEM and SNMR measurements consisted of 1D soundings along a 2D profile which was about 450 m long. One reference point, without GBC and about 400 m away from the profile, was measured for comparison and to identify seasonal variations in the data. All measurements were carried out with identical field setups and measurement parameters (loop size, number of averaging measurement repetitions, etc.). The recorded data were processed in an identical manner and a QT-inversion was used to derive a depth resolved partial water content model, i.e., the water content as function of depth and relaxation time. Due to the noisiness of the SNMR data, we used a permeability index, a combination of SNMR signal amplitude and relaxation time, to evaluate the results. By doing so, we reduce the inherent ambiguity (especially in noisy data) between the two parameters. Comparing the results from the first with the last field campaign shows, that a reduction of the permeability index within the GBC targeted layers of about 33 percent is detected, which is indicative for a respective compaction.