Présentations Colloques

Oral Presentation
8.03
Session 8.03: Geometrical structure and hydrogeological properties of Hard-Rock aquifers.
Laurencelle Marc
A workflow for inferring vertical profiles of hydraulic conductivity in regional rock aquifers from specific capacity data affected by drilling and testing biases
Specific capacity (Sc) data from well-performance tests are commonly used to estimate transmissivity (T). Since Sc data are much more abundant than T data from hydraulic tests, they allow an analysis of the spatial variability of T. The objective of our study was to define a representative vertical profile of hydraulic conductivity (K) in a fractured-rock aquifer on the basis of T values derived from Sc data. Although the estimation of T from Sc is apparently simple, wellbore storage significantly affects short duration tests, which is generally overlooked. Also, as poor shallow T preferentially leads to deeper wells, water-well drilling involves an important sampling bias due to non-random well depths. Moreover, T cannot be used as a direct estimator of K variations with depth, as T is the depth-integrated K.**Considering these complications, we developed a methodology to infer regionally representative vertical trends of K from Sc data affected by drilling and testing biases. First, a conceptual model relates large-interval T estimates from Sc data to the actual vertical profile of K. Second, a new approach based on the Papadopulos-Cooper solution is used to estimate T from Sc data considering wellbore storage, and to identify a T threshold below which Sc cannot be used to assess T. Reliability of T(Sc) estimates is assessed through a sensitivity analysis, in order to compute a corrected depth trend for T(Sc). A stochastic modeling workflow is then used to define the actual K trend with depth through simulation of the driller-biased decreasing T(Sc) depth trend. The original Monte Carlo approach developed for this purpose thus provides the best-suited fracture and hydraulic parameters related to the actual aquifer heterogeneity. This methodology was applied to the Montérégie Est study area (~9 000 km2, in southern Quebec, Canada) for an 18 000 Sc data set, providing a rigorous quantitative model of K variation with depth in its regional fractured-rock aquifer.
Canada