Environmental Engineering Reference
In-Depth Information
into the calculation of groundwater ages. Cook
and Solomon (
1995
) determined that the time
lag associated with diffusion of CFCs through
a coarse-textured unsaturated zone is 1 to 2
years for a 10 m water-table depth and 8 to 15
years for a 30 m water-table depth. In practice,
recharge dates for systems with thick unsatu-
rated zones may need to be adjusted to account
for this lag time. The tritium/helium-3 (
3
H/
3
He)
method, which does not require information
on atmospheric concentrations, and the use of
radionuclide tracers with long half lives, such
as
14
C, to date very old waters are not affected
by this lag time.
Any groundwater sample represents a dis-
tribution of different ages, particularly in het-
erogeneous media (Maloszewski and Zuber,
1982
; Weissmann
et al
.,
2002
). Mixing of water
of different ages occurs naturally in ground-
water systems due to molecular diffusion and
hydrodynamic dispersion. Mixing can also be
an artifact of sampling procedure. For pur-
poses of estimating recharge, minimal mixing
is desired. Sampling locations and sampling
procedures need to be selected with this con-
cept in mind. In areas of recharge, deeper
water within an aquifer is generally older, and
therefore more mixed, than shallower waters.
Sampling depths near the water table should
minimize natural mixing. To avoid artificial
mixing while sampling, small screened inter-
vals are desirable so that only a small volume
of the aquifer is sampled, approaching as near
as possible a point sample. Delin
et al
. (
2000
)
used screen lengths of 30 mm. Screen inter-
vals of a meter or more provide an integrated
(as opposed to a point) sample, with waters of
multiple ages being possibly mixed together.
Tracer concentrations determined on samples
collected from large screened intervals may be
useful for determining proportions of old and
young waters and identifying groundwater flow
paths, but water samples with mixed ages are
not ideal for estimating recharge rates. Analysis
of multiple gas tracers can be used to assess the
degree of mixing that a sample may have expe-
rienced (Aeschbach-Hertig
et al
.,
1999
).
The collection of groundwater samples
for tracer analysis often involves detailed
procedures and special collection vessels to
avoid exposing the sample to contamination
from the atmosphere. Specific procedures may
be required for each tracer. The CFC laboratory
of the US Geological Survey provides sampling
instructions for all of the tracers mentioned
above and several other tracers (http://water.
usgs.gov/lab; accessed April 10, 2009). It is best
to contact the laboratory that will be doing the
analysis for sampling instructions prior to ini-
tiating the sampling. Improper sampling is a
major cause of sample rejection.
Estimation of recharge from age dating of
groundwater relies on the simple assumption
of piston flow (Cook and Solomon,
1997
); young
water moves downward displacing older water
with little mixing or dispersion. Recharge
rates must be relatively high to use dissolved
gas concentrations for age dating; Cook and
Solomon (
1997
) suggested a minimum of about
30 mm/yr, and Böhlke (
2002
) suggested a min-
imum of about 100 mm/yr. At lower rates, dis-
persion dominates advection and the validity of
the piston-flow assumption is questionable.
Determining a value for
v
v
in Equation (
7.16
)
requires an assumption on the way in which
ages are distributed with depth in an aquifer.
Cook and Böhlke (
2000
) provided equations
for estimating velocity and recharge for piston
flow in generic aquifer types. For a hypotheti-
cal unconfined aquifer of constant thickness,
Z
,
underlain by an impermeable layer, groundwa-
ter age contours are horizontal when recharge
is uniform (
F ig u r e 7.9
). In this highly idealized
flow system, age increases with logarithm of
depth from zero at the water table to infinity
at the base of the aquifer (where vertical veloc-
ity equals zero), and recharge can be calculated
from an age measurement,
t
, at a given depth
below the water table,
z
, as (Böhlke,
2002
):
R
=
ϕ
ln(
Z
ZZ z t
/(
−
)) /
(7.17)
where ln is natural logarithm. The dependence
of age on logarithm of depth has been found to
be a reasonable assumption in a number of stud-
ies (
F ig u r e 7.10
), but other assumptions are pos-
sible as well. Solomon
et al
. (
1993
) determined
that ages increased linearly with depth (i.e. flow
was essentially vertical) within the top 15 m of
