Cross - borehole slug test analysis in a fractured limestone aquifer

This work proposes new semi-analytical solutions for the interpretation of cross-borehole slug tests in fractured media. Our model is an extension of a previous work by Barker (1988) and Butler and Zhan (2004). It includes inertial effects at both test and observation wells and a fractional flow dimension in the aquifer. The model has 5 fitting parameters: flow dimension n, hydraulic conductivity K, specific storage coefficient Ss, and effective lengths of test well Le and of observation well Leo. The results of a sensitivity analysis show that the most sensitive parameter is the flow dimension n. The model sensitivity to other parameters may be ranked as follows: K > Le ~ Leo > Ss. The sensitivity to aquifer storage remains one or two orders of magnitude lower than that to other parameters. The model has been coupled to an automatic inversion algorithm for facilitating the interpretation of real field data. This inversion algorithm is based on a Gauss-Newton optimization procedure conditioned by re-scaled sensitivities. It has been used to interpret successfully cross-borehole slug test data from the Hydrogeological Experimental Site (HES) of Poitiers, France, consisting of fractured and karstic limestones. HES data provide flow dimension values ranging between 1.6 and 2.5, and hydraulic conductivity values ranging between 4.4x10 and 7.7x10 m.s. These values are consistent with previous interpretations of single-well slug tests. The results of the sensitivity analysis are confirmed by calculations of relative errors on parameter estimates, which show that accuracy on n and K is below 20% and that on Ss is about one order of magnitude. The K-values ha l-0 02 60 61 7, v er si on 1 4 M ar 2 00 8 Author manuscript, published in "Journal of Hydrology 348 (2008) 510-523" DOI : 10.1016/j.jhydrol.2007.10.021 ha l-0 02 60 61 7, v er si on 1 4 M ar 2 00 8