Middle to Late Pleistocene faulting history of the Heerlerheide fault, Roer Valley Rift System, influenced by glacio-isostasy and mining-induced displacement

Faults of the Roer Valley Rift System (RVRS) are characterized by seismicity, scarps and displaced fluvial terraces, showing that they are active. The Heerlerheide fault is part of the southern boundary fault system of the RVRS, the Feldbiss fault zone (FFZ). During the late 19th and first half of the 20th century coal was mined in the subsurface south of the FFZ. As a result, general subsidence, sinkholes, and fault scarplets appeared on the surface. One of the induced fault scarps coincides with the location of the Heerlerheide fault, indicating that the upper part of the fault was reactivated by the mining. A trench was opened across the fault to study the mining-induced fault reactivation as well as the tectonic fault displacement history. The Heerlerheide fault offsets a 340 ka Meuse river terrace overlain by loess and loess-like deposits of Eemian to Late Pleniglacial age, including the Rocourt soil, the Eben/Patina discordance and the Nagelbeek soil complex. At least three tectonic faulting events were reconstructed, which are most likely surface rupturing earthquakes. However, the two oldest displacements could represent multiple faulting events. The age of the youngest event is well constrained between 17 and 15 ka. The duration of the time interval between the penultimate and youngest events is at least 8 ky (the inter-event time is an important parameter for assessing seismic hazard). The vertical coseismic displacement of the youngest event was around similar to 0.25 m; the estimated moment magnitude is around 6.2. This event is more or less synchronous with the age of events found in other fault trench studies of the FFZ. The timing is roughly contemporaneous with surface rupturing earthquakes along the northern boundary fault zone of the RVRS, the Peel Boundary fault zone (PBFZ), and a phase of volcanism in the nearby Eifel area, suggesting a common mechanism. The timing also corresponds to the start of the glacio-isostatic forebulge collapse, which has been invoked before to explain earthquake events in the RVRS, and in northern Germany and Denmark. Previous studies provided evidence for a Holocene surface rupturing earthquake with an offset of about 1 m at the Geleen fault in the Meuse valley on Belgian territory, taking place between 2.5 +/- 0.3 and 3.1 +/- 0.3 ka. The Geleen and Feldbiss faults on Dutch territory experienced faulting during the Late Glacial - Holocene. The incision and deposition history of a brook crossing these faults suggests an age around 7.5 ka for this event. However, in contrast, our results for the Heerlerheide fault show no evidence for Late Glacial - Holocene tectonic fault activity. We also observed 0.34 m of vertical displacement of the base of the plough layer, which corresponds to the amount to what was observed at the surface during the mining in 1936, indicating that no fault motions have occurred afterwards, despite re-flooding of the mines and consequent surface rebound. However, the fact that the fault was reactivated by subsurface mining shows that it is a weakness zone in the subsurface, and therefore fault reactivation might occur due to still ongoing rebound. In contrast to the tectonic fault displacement, the mining-induced offset was accompanied by downslope movement of the upper part of the hanging wall, resulting in splaying of the fault tip and crack formation. (C) 2021 The Authors. Published by Elsevier Ltd.