 |
| Titel |
The "universal" behavior of the Breakthrough Curve in 3D aquifer transport and the validity of the First-Order solution |
| VerfasserIn |
Igor Jankovic, Mahdi Maghrebi, Aldo Fiori, Antonio Zarlenga, Gedeon Dagan |
| Konferenz |
EGU General Assembly 2017
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 19 (2017) |
| Datensatznummer |
250141595
|
| Publikation (Nr.) |
 EGU/EGU2017-5125.pdf |
|
|
|
|
|
| Zusammenfassung |
| We examine the impact of permeability structures on the Breakthrough Curve (BTC) of
solute, at a distance x from the injection plane, under mean uniform flow of mean
velocity U. The study is carried out through accurate 3D numerical simulations, rather
than the 2D models adopted in most of previous works. All structures share the
same univariate distribution of the logconductivity Y = lnK and autocorrelation
function ρY , but differ in higher order statistics. The main finding is that the BTC of
ergodic plumes for the different examined structures is quite robust, displaying
a seemingly "universal" behavior. The result is in variance with similar analyses
carried out in the past for 2D permeability structures. The basic parameters (i.e. the
geometric mean, the logconductivity variance σY 2 and the horizontal integral scale
I) have to be identified from field data (e.g. core analysis, pumping test or other
methods). However, prediction requires the knowledge of U, and the results suggest that
improvement of the BTC prediction in applications can be achieved by independent
estimates of the mean velocity U, e.g. by pumping tests, rather than attempting to
characterize the permeability structure beyond its second-order characterization. The BTC
prediction made by the Inverse Gaussian (IG) distribution, adopting the macrodispersion
coefficient estimated by the First Order approximation αL = σY 2I, is also quite
robust, providing a simple and effective solution to be employed in applications. The
consequences of the latter result are further explored by modeling the mass distribution that
occurred at the MADE-1 natural gradient experiment, for which we show that most of
the plume features are adequately captured by the simple First Order approach. |
|
|
|
|
|
|