Zj. Kabala, ANALYTICAL SOLUTIONS FOR THE COEFFICIENT OF VARIATION OF THE VOLUME-AVERAGED SOLUTE CONCENTRATION IN HETEROGENEOUS AQUIFERS, Stochastic hydrology and hydraulics, 11(4), 1997, pp. 331-348
Citations number
37
Categorie Soggetti
Mathematical Method, Physical Science","Water Resources","Environmental Sciences","Statistic & Probability
Under the assumption that local solute dispersion is negligible, a new
general formula (in the form of a convolution integral) is found for
the arbitrary k-point ensemble moment of the local concentration of a
solute convected in arbitrary m spatial dimensions with general sure i
nitial conditions. From this general formula new closed-form solutions
in m = 2 spatial dimensions are derived for 2-point ensemble moments
of the local solute concentration for the impulse (Dirac delta) and Ga
ussian initial conditions. When integrated over an averaging window, t
hese solutions lead to new closed-form expressions for the first two e
nsemble moments of the volume-averaged solute concentration and to the
corresponding concentration coefficients of variation (CV). Also, for
the impulse (Dirac delta) solute concentration initial condition, the
second ensemble moment of the solute point concentration in two spati
al dimensions and the corresponding CV are demonstrated to be unbound.
For impulse initial conditions the CVs for volume-averaged concentrat
ions axe compared with each other for a tracer from the Borden aquifer
experiment. The point-concentration CV is unacceptably large in the w
hole domain, implying that the ensemble mean concentration is inapprop
riate for predicting the actual concentration values. The volume-avera
ged concentration CV decreases significantly with an increasing averag
ing volume. Since local dispersion is neglected, the new solutions sho
uld be interpreted as upper limits for the yet to be derived solutions
that account for local dispersion; and so should the presented CVs fo
r Borden tracers. The new analytical solutions may be used to test the
accuracy of Monte Carlo simulations or other numerical algorithms tha
t deal with the stochastic solute transport. They may also be used to
determine the size of the averaging volume needed to make a quasi-sure
statement about the solute mass contained in it.