Mining subsidence in a partially flooded abandoned mine : aseismic ground movement and consequences for post-mining risk management

In the Lorraine area of eastern France, decades of iron ore mining from 1850 to 1997 have left vast underground cavities beneath or near urban areas. Several major collapses occurred in the southern part of this iron ore basin in the 1990s, after the mine closure and the flooding of underground mine workings. Following these large scale collapses, the French government initiated a strategy of post mining risk management to prevent and control risks associated with these ground failures. The high risk zones are secured either by reducing the vulnerability while the moderate risk zones are monitored for public safety purposes by using in situ monitoring. This monitoring relies mainly on real time microseismic systems, to detect precursors to a rapid large scale collapse. Data recorded are processing automatically, and may generated alarm in case of abnormal evolution, in terms of number of event as well as in energy. A cell of expertise can be mobilized to analyse the situation and inform the local authorities of the evolutions of the situation. Evacuation can be triggered in case of danger for public safety. After the progressive closing and then flooding of the northern iron basin ending in 2008, subsidence was observed in a town of the Lorrain basin in autumn of 2009. However, this local subsidence, with a low velocity of few centimeters per month, was not clearly detected by the borehole microseismic monitoring station located nearby. Only some microseismic events were recorded, which could not be unambiguously related to the beginning of the subsidence event. To better understand this lack of microseismic precursor a geophysical investigation was launched. A calibration blast experiment was carried out from a remaining old underground access in order to characterise the wave propagation properties in this context. The results of this study show strong anelastic attenuation of the seismic waves though the monitored overburden most likely related to the extensive fault system intersecting the study site. Moreover, robbed pillar extraction and flooding of the site have induced a reduction of the mechanical properties of the overburden. These observations, added to a slow kinetics subsidence mechanism (~ cm/months) with little seismic energy release, may explain the lack of detected microseismicity during the subsidence event. In addition, low frequency microseismic events associated with the very slow subsiding movements might have not been detected by the used high frequency recording instruments, designed initially for rapid collapses (~ cm/hours).