Environmental Engineering Reference
In-Depth Information
to classical error analysis (Lerner
et al
.,
1990
).
Approaches for such an analysis for water-
budget methods are described in
Chapter 2
.
Concerns about inaccuracy in recharge esti-
mates should not deter application of any rea-
sonable method for estimating recharge. Simple
techniques, applied with careful consideration
of conceptual models, can be not only useful,
but enduring. Theis (
1937
) used a simple appli-
cation of the Darcy equation to estimate natural
recharge rates of between 3 and 7 mm/yr for the
southern High Plains aquifer, values that fall
midway in the range of estimates generated in
subsequent years with more sophisticated tech-
niques (Gurdak and Roe,
2009
). White (
1930
)
used a salt tracer to estimate subsurface flow
into a portion of the Mimbres River watershed
in southwestern New Mexico; the groundwater-
flow model of Hanson
et al
. (
1994
) corroborated
that estimate.
Because much of the error associated with
a recharge estimate is not quantifiable, it is
wise to apply multiple methods for estimat-
ing recharge in any study (Lerner
et al
.,
1990
;
Simmers,
1997
; Scanlon
et al
.,
2002b
). Estimates
from multiple methods may not quantitatively
reduce uncertainty; consistency in results,
while desirable, may not be a reliable indicator
of accuracy. Application of different approaches
may have qualitative benefits, however; incon-
sistencies in estimates may provide insight into
measurement errors or the validity of assump-
tions underlying a method and, thus, may
provide direction for revising the conceptual
model.
1.5 Challenges in estimating
recharge
1.5.1 Uncertainty in recharge estimates
Accurate estimates of recharge are always
desired; yet it is beyond our current capabilities
to determine, with any degree of confidence,
the uncertainty associated with any recharge
estimate, let alone claim that an estimate is
accurate. Actual recharge rates are unknown;
therefore, there are no standards that can be
used to evaluate the accuracy of recharge esti-
mates. The most serious errors are those asso-
ciated with an incorrect conceptual model. An
incorrect conceptual model can lead to the appli-
cation of inappropriate estimation techniques
and meaningless recharge estimates. Any esti-
mates based on an incorrect conceptual model
are inherently unreliable. Additional sources of
error arise from improper application of meth-
ods and measurement errors.
Improper application of a method can result
from lack of understanding of the method or
from failure to adequately account for spatial
and temporal variability. Errors related to the
latter arise from an inability to measure at
enough locations or failure to make measure-
ments for a sufficient length of time. Spatial
and temporal variabilities of recharge cannot be
determined exactly, but they can be examined
in some detail with numerical models. A map of
the annual recharge for Texas (
Figure 1.3
) was
developed by Keese
et al
. (
2005
) by combining
a one-dimensional variably saturated water-
flow model with spatially variable soil, vegeta-
tion, and climate properties. Techniques for the
upscaling of point estimates of recharge to large
areas are discussed more fully in
Chapter 3
.
Measurement errors relate to inaccuracies
in data collection. Complications arise because
the magnitude of a recharge flux is generally
quite small and often cannot be measured dir-
ectly. Most methods presented in this text are
indirect methods that rely on more readily
measured parameters, such as changes in water
storage or tracer concentrations, to make infer-
ences on recharge rates. Measurement errors
are the only type of errors that are conducive
1.5.2 Spatial and temporal scales
of recharge estimates
The concept of spatial scale is important
in terms of selecting appropriate methods.
Different methods provide estimates that are
integrated over various spatial scales. Some
methods provide essentially point estimates;
these methods are useful for evaluating aqui-
fer vulnerability to contamination in support of
land-use decisions, for example; however, appli-
cation at many points may be required to deter-
mine an areally averaged recharge rate. Other
