Content uploaded by James J. Butler Jr.
Author content
All content in this area was uploaded by James J. Butler Jr.
Content may be subject to copyright.
Calibration and Reliability in Groundwater Modelling: A Few Steps Closer to Reality
(Proceedings
of
ModclCARE'2002. Prague, Czech Republic. June 2002). IAHS Publ. no. 277. 2002.
419
High-resolution characterization of chemical
heterogeneity in an alluvial aquifer
M. K. SCHULMEISTER, J. M. HEALEY, J. J. BUTLER, JR
Kansas Geological Survey, The University of Kansas, 1930 Constant Av., Lawrence,
Kansas 66047. USA
mkschul(S),ku.edu
G. W. McCALL
Geoprohe Systems Inc., 601 North Broadway Street, Sa/ina, Kansas 66620, USA
S. BIRK
Centre for Applied Geoscienees (ZAG), University of Tubingen, Sigwartslrasse 10,
D-72076 Tubingen, Germany
Abstract The high-resolution capabilities of direct push technology were
exploited to develop new insights into the hydrochemistry at the margin of an
alluvial aquifer. Hydrostratigraphic controls on groundwater flow and contam-
inant loading were revealed through the combined use of direct push electrical
conductivity (EC) logging and geochemical profiling. Vertical and lateral
variations in groundwater chemistry were consistent with sedimentary features
indicated by EC logs, and were supported by a conceptual model of recharge
along the flood plain margin.
Key words alluvial aquifer; direct push; electrical conductivity; flood plain margin;
geochemical profiling; heterogeneity
INTRODUCTION
Prediction of the fate of reactive constituents in groundwater requires accurate
assessment of hydrochemical conditions in the subsurface. In alluvial aquifers, the
presence of complex hydrostratigraphic features can lead to chemical zonation over
short distances. Conventional drilling-based approaches typically result in data points
(wells) that are too far apart to allow for the identification of such variations. In the last
decade, direct push technology has become a widely used alternative to drilling based
methods for investigations of organic contaminants in groundwater. We recently
developed a direct push geochemical profiling approach that allows detailed
characterization of inorganic chemical parameters (Schulmeister et al, 2001). When
used in combination with direct push electrical conductivity (EC) logging, hydro-
stratigraphic controls on the inorganic hydrochemistry of an aquifer can be identified.
An example application of this new high-resolution approach for the geochemical
characterization of an aquifer is presented here.
The investigation took place at the Geohydrologic Experimental and Monitoring
Site (GEMS), a research site of the Kansas Geological Survey at which a great deal of
previous work has been done (e.g. Butler et al, 1999, 2002; Schulmeister, 2000).
GEMS is located near the margin of the flood plain of the Kansas River, a tributary to
420 M. K. Schulmeister et al.
Fig. 1 Location of the study site (GEMS).
one of the major river systems in the USA (Fig. 1). The site is underlain by 22 m of
alluvial sediments above a silty sandstone bedrock. The alluvium is composed of
11
m
of sand and gravel that is overlain by 11 m of clay, silt and sand (Fig. 2(a)). The sand
and gravel interval, which acts as a semiconfmed aquifer and is the primary focus of
this study, consists of a fining-upward sequence with interbedded clay lenses (Fig.
2(b)).
Data collected previously from monitoring wells at the site demonstrated vertical
variations in
NO3
concentrations that were difficult to interpret based on assumptions
of regional nonpoint-source loading (Fig. 2(c)).
NO3
was present in both the silt-and-
sand interval and the lower portion of the aquifer, but not in the upper portion of the
aquifer. The stratified nature of the alluvium and the site's proximity to the edge of the
flood plain led to an investigation of the potential importance of the flood plain margin
in controlling the
NO3
distribution and other aspects of the hydrochemistry of the site.
Direct push geochemical profiling and EC logging were conducted using a track-
mounted rig with a hydraulic slide and a high frequency percussion hammer. The
geochemical profiling tool consisted of a short (10 cm) screened section that was
attached to a string of steel drive rods. Samples were obtained in a single probehole as
the tool was progressively advanced to depths of interest. Field measurement of
specific conductance and dissolved oxygen (DO) was performed with a flow through
cell and multiparameter sonde, and samples were collected for laboratory analysis of
NO3,
Fe, and Mn (Schulmeister et al, 2001). EC logging was conducted using a
Wenner-array probe (Christy et al, 1994). Previous work at this site has shown that it
is possible to use electrical conductivity measurements to represent the distribution of
clay materials in the alluvium (Butler et al, 1999; Schulmeister et al, 2003). A series
of EC logs were therefore used to construct a conceptual hydrostratigraphic model of
the flood plain margin. This model guided subsequent geochemical profiling activities.
High-resolution characterization of chemical heterogeneity in an alluvial aquifer All
0 100 200 0 2 4 6 8 10
Hydraulic Conductivity (m day1) N03-N (mg I1)
(a) (b) (c)
Fig. 2 Geology, hydraulic conductivity, and N03 at GEMS: (a) geological description
of core materials; (b) hydraulic conductivity profiles in the lower 11 m of alluvium
from multi-level slug tests at two wells (Butler et al, 2002); and (c) N03 profiles from
conventional monitoring wells with screened intervals of 0.7 m (Schulmeister, 2000).
ELECTRICAL CONDUCTIVITY MODEL OF HYDROSTRATIGRAPHY
Direct push electrical conductivity logs were obtained along a transect oriented
perpendicular to the bedrock valley wall (Fig. 3). This transect was believed to be
roughly aligned along the direction of flow in the upper alluvium. The EC cross-
section revealed the coarsening of sediments in the upper part of the alluvium in the
direction of the flood plain margin, and the disappearance of the fine-grained silt and
clay layers that extend laterally across a large portion of the floodplain. As described
below, the horizontal continuity of interbedded fine and coarse sediments provides a
possible explanation for the chemical stratification observed at the site (e.g. Fig. 2(c)).
HYDROCHEMICAL FACIES AT THE FLOOD PLAIN MARGIN
Geochemical profiling was conducted at three locations in clay-free zones indicated by
the EC logs (Fig. 3). The data from the direct push profiles were combined with those
from the monitoring wells to delineate lateral and vertical variations in groundwater
specific conductance. Relatively low specific conductance values were observed in the
overlying less-permeable material, in the upper part of the aquifer, and near the flood
plain margin. The deep portion of the aquifer generally contained groundwater with
relatively high specific conductance. This contrast in fluid chemistry suggests different
sources for the groundwater. Groundwater in the upper alluvium and near the edge of
the flood plain appears to have originated as recharge along the more permeable flood
422 M. K. Schulmeister et al.
plain margin. In contrast, groundwater in the deeper portions of the alluvium appears
to have originated as lateral inflow from the more central parts of the floodplain. The
intermediate specific conductance values in the middle profile (profile B, Fig. 3)
appear to indicate a region in which these waters are mixed.
Variations in concentrations of additional constituents were consistent with those
observed for specific conductance (Fig. 4). In direct push profile A, a steep chemical
gradient occurred at an elevation of 238 m with an increase in DO and
NO3,
and the
disappearance of Fe and Mn. The
NO3
profile was in agreement with the monitoring
well data (Fig. 2(c)). Higher
N03,
specific conductance, and hydraulic conductivity in
the deeper portion of the aquifer suggest that the
NO3
in that interval may have entered
the site through deep lateral flow from the central reaches of the floodplain. Ground-
water chemistry in this portion of the aquifer was similar to that observed in wells
located in the aquifer at greater distances from the flood plain margin.
Consistent with the specific conductance data at profile C, a groundwater chem-
istry was observed that was different elsewhere at the site. The low levels of DO and
Fe were similar to those observed in the upper part of the aquifer in the other profiles.
However, the higher levels of
NO3
and Mn indicate a different groundwater source and
possible contamination from agricultural activities in the overlying soils. The vertical
movement of
NO3
through the coarse sediments that occur near the flood plain margin,
coupled with transport of this contamination along the sandier strata indicated on the
EC cross-section, provides a possible explanation for the presence of
NO3
in shallow
wells elsewhere at the site (Fig. 2(c)).
Wells
250
H
CD
>
_CD
CO
S
245
CD
>
O
_o
CD
"f
240
c
0
Jj 235
LU
230
•560--T'-
• 660
:
11^620
•700
'710
•708»
•702-720
'702
-•-700
-'699
-695> - " ~-
• 693- - -_
•510
• 555_
•61 ~f~
•303 <
1342 _.
:$ëjdi;c3çk:
Electrical Conductivity
(mS m-1)
0 50 100
Distance
(m)
Fig. 3 Electrical conductivity cross-section X-X' (see Fig. 1). Intervals sampled by
monitoring wells are indicated by black bars. A, B, and C indicate direct push
geochemical profile locations. Groundwater specific conductance measurements
(uS cm"1) are listed next to each sampling point. Direct push samples were taken
approximately every 0.61 m when clays did not block the screen of the sampler.
High-resolution characterization of chemical heterogeneity in an alluvial aquifer 423
Fe,
Win (mgH) Fe Mn (mg \-n Fe, Mn (mgr1)
012301230123
241 H 1 1 1 1 1 1 H 1 ' 1 1 1 1 H 1 1 1 1 > 1
Fig. 4 Direct push chemical profiles from three locations along the EC traverse.
In profile B, the increase in DO and absence of Fe and Mn near the top of the
aquifer indicates that the redox conditions observed in the first profile may extend
laterally across the site. However, very low levels of
NO3
in the high-DO zone suggest
that
NO3
either did not reach this part of the aquifer or was diluted by another source.
The presence of Fe in the high-DO zone supports the hypothesis of mixing of ground-
water from different sources within the vicinity of profile B. The absence of
NO3
in the
upper portion of the aquifer in profiles A and B may also be explained by
NO3
reduction. This is consistent with nitrogen-isotope data collected in previous work
(Schulmeister, 2000).
CONCLUSIONS
The unprecedented level of detail afforded by direct push technology allowed
important information regarding site-specific controls on groundwater flow and solute
transport to be obtained. The conceptual model generated as a result of the direct push
investigation provided insights that would have been difficult to recognize using
conventional methods. The direct push profiling identified recharge at the flood plain
margin as a possible contributor to the stratified groundwater chemistry and the nitrate
contamination observed in the upper portion of the alluvium. These insights should
prove to be extremely valuable for directing and constraining future modelling efforts.
Acknowledgements This research was supported by funding provided by the Kansas
Water Resources Research Institute under grant no. HQ96GR02671
Modif.
008
(subaward
SO
1044).
The multi-parameter sonde and flow cell used to monitor field
424 M. K. Schulmeister et al.
chemical parameters were provided through a collaborative exchange with YSI
Incorporated. Steffen Birk was one of the participants in the 2001 Applied
Geohydrology Summer Research Assistantship program of the Kansas Geological
Survey. This programme is open to students at any university with an interest in
learning more about recent developments in hydrogeological field methods.
REFERENCES
Butler, J. J., Jr, Healey,
.1.
M., Zheng, L., McCall, G. W. & Schulmeister, M. K. (1999) Hydrostratigraphic characterization
of unconsolidated alluvial deposits with direct push sensor technology. Kansas Geol. Survey Open-File Report 99-40
(also available at www.kgs.ukans.edu/Hvdro/Publications/OFR99 40/index.htmO.
Butler, J. .!., Jr, Healey, J. H., McCall, G. W., Garnett, E. J. & Loheide, S. P., II (2002) Hydraulic tests with direct push
equipment. Groundwater 40(1), 25-36.
Christy, C. D., Christy, T. M. & Wittig, V. (1994) A percussion probing tool for the direct sensing of soil conductivity, ln:
National Outdoor Action (Proc. 8th
Conf,
National Ground Water Association, Las Vegas, Nevada, USA, May
1994),
381-394.
Schulmeister, M. K. (2000) Hydrology and geochemistry of an alluvial aquifer near a flood plain margin. PhD Thesis,
University of Kansas, Lawrence, Kansas, USA.
Schulmeister, M. K., Butler, J. J., Jr, Whittemore, D. O., Birk, S., Healey, J. M„ McCall, G. W., Sellwood, S. M. &
Townsend, M. A. (2001) A new direct push-based approach for the chemical investigation of stream-aquifer
interactions. Geol. Society of America Abstracts with Programs 33(6), 426.
Schulmeister, M. K., Butler, J. J., Jr., Healey, J. M., Zheng, L., Wysocki, D. A. & McCall, G. W. (2003) Direct push
electrical conductivity logging for high-resolution hydrostratigraphic characterization. Ground Water Monitoring
Remed.
{in press).