Quantitative comparison of impeller flowmeter and particle-size dsitribution techniques for the characterization of hydraulic conductivity variability

Abstract

Basic univariate statistics and key geostatistical parameters of estimates of hydraulic conductivity obtained at the decimeter scale by two different methods are presented and compared. The two estimates are based on (1) the empirical Kozeny-Carman formulation, and (2) impeller flowmeter tests. The former provides values of conductivity, KGS, based on particle size distributions. Impeller flowmeter techniques allow inferring conductivities, KFM, from measurements of vertical flows within a borehole. Data obtained during an extensive monitoring campaign at an experimental site located near the city of Tübingen, Germany, are considered. Statistics of the natural logarithm of KGS and KFM at the site are similar in terms of mean values (with averages of ln KGS being slightly smaller than those of ln KFM) and differ in terms of variogram ranges and sample variances. The correlation between the two sets of estimates is virtually absent. Additional data from two different sites already presented in the literature allow comparing conductivity estimates from flowmeter and grain-size distributions (or permeameter measurements) taken at adjacent wells and support the finding that KGS and KFM lack correlation. The analysis highlights the difficulty in obtaining meaningful quantitatively comparable hydraulic conductivity data at the decimetric scale.