Abstract :
[en] A better understanding of the water balance of a landfill is crucial for its management, as the waste water content is the main factor influencing the biodegradation process of organic waste. In order to investigate the ability of long electrical resistivity tomography (ERT) profiles to detect zones of high infiltration in a landfill cover layer, low resolution time lapse data were acquired during a rainfall event. Working at low resolution allows to cover large field areas but with the drawback of limiting quantitative interpretation.
In this contribution, we use synthetic modeling to quantify the effect of the following issues commonly encountered when dealing with field scale ERT data: (i) the effect of low resolution on electrical resistivity changes interpretation, (ii) the effect of the original heterogeneous resistivity distribution on the observed relative resistivity changes, (iii) the need for temperature and pore fluid conductivity data in order to compute water content and absolute changes of water content, and (iv) the interpretation error commonly made while neglecting the dilution effect during fresh water infiltration.
Firstly, due to the lack of spatial resolution, the regularized inversion process yields a smoothed distribution of resistivity changes that fail to detect small infiltration zones and yields an overestimation of the infiltration depth and an underestimation of the infiltrated volume in large infiltration areas. Secondly, the analysis of relative changes, as commonly used in literature, is not adequate when the background water content is highly heterogeneous. In such a case, relative changes reflect both the initial water content distribution and the observed changes. Thirdly, the computation of absolute water content changes better reflects the infiltration pattern, but requires spatially distributed temperature and pore fluid conductivity input data. Lastly, the dilution effect, if not considered, leads to an underestimation of the infiltrated volume.
Taking into account these elements, we extracted the maximum amount of information from our field data without over-interpreting the results. This allowed the detection of larger infiltration areas possibly responsible for a large part of the annual water infiltration and landfill gas loss.
