Time-lapse electrical resistivity tomography of a water infiltration test on Johannishus

Esker, Sweden

Ulusoy, Inan; Dahlin, Torleif; Bergman, Bo

Published in:
Hydrogeology Journal

DOI:
10.1007/s10040-014-1221-2

Published: 2015-01-01

Link to publication

Citation for published version (APA):

Ulusoy, I., Dahlin, T., & Bergman, B. (2015). Time-lapse electrical resistivity tomography of a water infiltration
test on Johannishus Esker, Sweden. Hydrogeology Journal, 23(3), 551-566. DOI: 10.1007/s10040-014-1221-2

General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors
and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the
legal requirements associated with these rights.
LUNDUNI
VERSI
TY
• Users may download and print one copy of any publication from the public portal for the purpose of private
study or research.
• You may not further distribute the material or use it for any profit-making activity or commercial gain PO Box117
• You may freely distribute the URL identifying the publication in the public portal ? 22100L und
+4646-2220000
Hydrogeology Journal (2015) 23: 551–566
DOI 10.1007/s10040-014-1221-2

Time-lapse electrical resistivity tomography of a water infiltration test

on Johannishus Esker, Sweden

İnan Ulusoy & Torleif Dahlin & Bo Bergman

Abstract Managed aquifer recharge (MAR) is an the geological, hydrogeological and structural character-
efficient way to remove organic matter from raw water istics of the site.
and, at the same time, reduce temperature variation.
Two MAR sites were constructed by Karlskrona Keywords Artificial recharge . Groundwater
municipality on Johannishus Esker in Sweden. One of management . Sweden . Conductivity . Resistivity change
these sites, Vång, was monitored for electrical conduc-
tivity and electrical resistivity (using electrical resistiv-
ity tomography - ERT) during a 9-week tracer Introduction
infiltration test. The aim of the monitoring was to map
the pathways of the infiltrated water, with the overall Organic pollution of surface water in areas with high forest
goal to increase the efficiency of the MAR. ERT proved coverage and temperate climates is a common problem.
useful in determining both the nature of the esker Organic matter in drinking water can cause aesthetic and
formation and the water migration pathways. In Vång, odoriferous problems, as well as increase bacterial growth
the esker ridge follows a tectonically controlled paleo- (Lindroos et al. 2002). If the organic matter content is high,
valley. The fault/fracture zone in the bedrock along this chlorine disinfection can result in the formation of mutagenic
paleo-valley was mapped. During the tracer test, the or carcinogenic compounds (Lindroos et al. 2002). Organic
infiltrated water was detected in the area close to the matter can also interact with pollutants in the soil and
infiltration ponds, whereas far-situated observation wells groundwater (Thurman 1985; Lindroos et al. 2002), and
were less affected. For sequential infiltration and form compounds and complexes with Fe (Liang et al. 1993;
recharge periods in MAR, the timing of the well Frycklund and Jacks 1997) and other metals (Nilsson and
pumping is another important factor. Natural ground- Bergkvist 1983; Lundström 1993; Lindroos et al. 1995).
water flow direction was a determinant in the infiltration An efficient way to remove organic matter from the raw
process, as expected. ERT measurements provide water is managed aquifer recharge (MAR). Using basin
supplementary data for site selection, for monitoring infiltration, MAR has been used extensively for decades (e.g.
the functionality of the MAR sites, and for revealing Helmisaari et al. 1998; Frycklund 1998; Lindroos et al.
2002; Greskowiak et al. 2005; Katko et al. 2006; Nöjd et al.
2009). The main idea of MAR is to infiltrate organically
Received: 6 May 2014 / Accepted: 5 December 2014 polluted water into a structurally well-defined limited area,
Published online: 24 December 2014 using the subsurface as a natural filter, and then to produce
* Springer-Verlag Berlin Heidelberg 2014 the water from production wells situated around the
infiltration site. The infiltration is made through infiltration
basins or ponds, or by sprinkling surface water directly onto
Present Address: the forest floor via a network of pipes (Lindroos et al. 2002).
İ. Ulusoy Organic matter is removed from the infiltration water by
Department of Geological Engineering, Hacettepe University,
06800, Beytepe, Ankara, Turkey physical, chemical and microbiological processes in the soil
(Lindroos et al. 2002).
İ. Ulusoy ()) Eskers are long, narrow, sinuous, steep-sided ridges
Department of Geology, Lund University, Sölvegatan 12,
22362, Lund, Sweden composed of irregularly stratified sand and gravel depos-
e-mail: [email protected] ited by subglacial or englacial streams flowing between
Tel.: +90 312 297 77 00 ice walls or in an ice tunnel of a stagnant or retreating
T. Dahlin glacier and left behind when the ice melted (Flint 1971;
Engineering Geology, Lund University, Box 118, 221 00, Neuendorf et al. 2005). Deposition of an esker happens
Lund, Sweden within or in front of the tunnel mouth (Banerjee and
B. Bergman McDonald 1975; Mäkinen 2003; Boulton et al. 2009),
SWECO Environment, Hans Michelsensgatan 2, Box 286, within the subglacial (Syverson et al. 1994; Warren and
201 22, Malmö, Sweden Ashley 1994; Brennand 2000; Campbell 2003; Boulton
552

Fig. 1 Map showing the study area. Digital elevation model superimposed on a shaded relief map. Infiltration site, resistivity profiles, and
production and observation wells are indicated. Coordinate system: UTM (m), Datum WGS84

et al. 2009) or englacial (Syverson et al. 1994; Campbell Geer 1968; Parsons 1970; Artimo et al. 2003; Bolduc
2003; Burke et al. 2009) tunnels, or even in the supra- et al. 2006; Okkonen et al. 2010), and they are effective
glacial environment (Russell et al. 2001). Eskers have sites for MAR (e.g. Lindroos et al. 2002; Katko et al.
long been known as a natural source of groundwater (De 2006; Nöjd et al. 2009).
Hydrogeology Journal (2015) 23: 551–566 DOI 10.1007/s10040-014-1221-2
 553

Fig. 2 Geological and structural context of the study area. Coordinate system: UTM (m), Datum WGS84. a Geological map (SGU 2012), in which the
area between the red lines shows the fracture zone revealed by resistivity measurements. Equipotential lines (a.s.l.; Hägg et al. 2010) around the infiltration
site interpolated from the data of observation/production wells are shown. b Magnetic anomaly map. The map is superimposed on a shaded relief map
and shows measured variations in the magnetic total field after subtraction of the geomagnetic reference field (DGRF 1965.0), and is based on airborne
measurements performed at an altitude of 30–60 m (SGU 2012). Structural features (see legend) and the Johannishus esker (white line) are shown. c
Simplified geological section of the valley perpendicular to the esker ridge (with changes from Hebrand and Björkman 2005), showing the natural state
and the excavated area for infiltration basin construction
Hydrogeology Journal (2015) 23: 551–566 DOI 10.1007/s10040-014-1221-2
554

Fig. 3 Schedule showing the

timeline of the reference and
repeated resistivity
measurements

Karlskrona Municipality built two MAR facilities on changes from Hebrand and Jeppson (1999) and Hebrand
Johannishus Esker (Sweden) for treatment of organically and Björkman (2005) are given in Fig. 2a,c. An
polluted water: the Hillerslätt and Vång infiltration sites. equipotential contour map (m a.s.l., Hägg et al. 2010)
Tracer tests were performed at these MAR sites in 2008. of the infiltration site is projected on Fig. 2a. Johan-
Tests were conducted with salt as a tracer; dosed salt was nishus Esker is a NNE trending ridge over 15 km in
added to the water and injected through the infiltration length (Fig. 2a,b). Near Vång, the crest of the esker
ponds. For the Hillerslätt infiltration site in the northern reaches 55 m a.s.l. and 12 m above the surrounding flat
part of the esker, hydraulic connection was shown by the topography. The width of the esker reaches 130 m in the
salt recovered in the wells; however, at the southern site, study area (Fig. 1). The geology is defined by glacial and
Vång, the rate of recovered tracer in the observation wells post-glacial Quaternary sedimentation over mesoproter-
remains low when compared to the rate infiltrated. For ozoic acidic/intermediate granitic bedrock (SGU 2012,
further investigation of the transport pathways of the Fig. 2a,c) which is tectonised mainly along N–S fault
infiltrated salty water in the Vång test area, additional lines (Berglund et al. 2005; Fig. 2b). Lundqvist and
tracer experiments were conducted in 2009. New tests Wohlfarth (2001) state that the coastal area of Blekinge
were coupled with measurement of electrical conductivity and southeastern Småland up to the highest shoreline
in the observation wells and time-lapse electrical resistiv- became deglaciated ∼14,500 years ago. During the
ity tomography measurements along six profiles which are deglaciation, Johannishus Esker was deposited in the
detailed and discussed in this report. subaquatic conditions at the margin of the retreating ice-
Electrical resistivity tomography (ERT) is an efficient sheet and at the mouth of an ice-tunnel (Malmberg
technique (Loke and Barker 1996; Barker and Moore Persson and Persson 2009). As a result, the bedrock is
1998) for a wide variety of earth science applications overlaid by sandy moraine and esker deposits, and after
especially in hydrogeology to visualize and monitor the the glacial retreat, by post-glacial sand, clay and organic
groundwater. Besides aquifer mapping and characteriza- deposits (Fig. 2a,c). Overburden on the bedrock can be
tion (Barker and Moore 1998; Dahlin 2001; Corriols up to 30 m within the study area.
and Dahlin 2008; Mendoza et al. 2006), tracer tests can The esker deposit can be divided into three main
also be monitored with ERT (White 1988, 1994; Morris units: the continuous body of the esker deposit and two
et al. 1996; Singha and Gorelick 2005; Cassiani et al. mantling units above the body (Fig. 2c, Hebrand and
2006). Björkman 2005). The body, referred to as coarse-grained
esker in Fig. 2c, is the core of the esker deposition, lies
on the bedrock and forms the Johannishus ridge. The
Geology body consists of well-sorted blocks, gravel and sand,
saturated with water and is the most permeable part of
Vång MAR is built on Johannishus Esker near Vång the esker (Hebrand and Björkman 2005). Units mantling
village 4.5 km north of Johannishus and 15 km the esker can be grouped as fine-grained and coarse-
northwest of Karlskrona (Fig. 1) in Blekinge province. grained esker mantle (Fig. 2c, Hebrand and Björkman
The study area encompasses a NNE-trending valley with 2005). Coarse-grained mantle mainly consists of gravel
a flat, southward sloping morphology and with bound- and sand, while fine-grained mantle is mostly stratified
ing heights at two edges (western and eastern side) up fine sand and silt.
to 20 m above the valley floor (Fig. 1). Edge to edge,
the width of the valley varies from 300 to 900 m. In the
middle of the valley, Johannishus Esker is an elongate Fig. 4 Resistivity profile P1. The reference resistivity profile and the„
topographic ridge parallel to the valley direction percentage change in resistivity relative to the reference are shown for
the repeated measurements. Bedrock level (BRL) is indicated with a
(Fig. 1). dashed line and the relative positions of the closest observation wells
The geology map of Geological Survey of Sweden are indicated with black lines. Areas marked with A and B represent the
(SGU 2012) and a generalized geological section with resistivity decrease and are discussed in the text
Hydrogeology Journal (2015) 23: 551–566 DOI 10.1007/s10040-014-1221-2
 555

Hydrogeology Journal (2015) 23: 551–566 DOI 10.1007/s10040-014-1221-2

556

Water infiltration and monitoring methodology 160 m, while profiles P3 and P4 are 200 m and profile P6
is 320 m long. Take-out distances of the electrodes are
Water infiltration site for MAR equal (2 m) in all measurements, giving a total electrode
Part of the esker in Vång was previously used as a gravel spread of 160 m at a single layout. A roll-along procedure
quarry (Ringberg 1969). It was excavated to depths of (Dahlin 1996) was used to get longer lines with a 160 m
41 m a.s.l. (Fig. 2c) and in the spring of 1979, the spread of four electrode cables. When measurements have
excavated area, the hummocky terrain and the remnant been completed on the initial spread, the first cable is
esker were levelled out (Hebrand and Björkman 2005). shifted to the end of the layout and the new spread is
This area was evaluated as an infiltration site by measured; thus, for sections longer than 160 m, the
Karlskrona Waterworks (Hebrand 2001); the infiltration initially installed electrode number is more than 80
ponds were built in the middle portion of the esker crest, electrodes (100 electrodes for 200 m and 160 electrodes
close to the Liatorp road (Fig. 1). Two main ponds were for 320 m length). However for the profiles traversing the
constructed: one to the north and one to the south. The 8-m-wide road passing by the esker ridge, no electrodes
southern dam is larger than the northern dam. The topsoil were installed on the road; two to three electrodes were
(including the organic horizon and other genetic soil skipped (Fig. 1; profiles P1, P2, P3, P4, and P6).
horizons) and fine-grained soil were removed (Fig. 2c), LaBrecque and Daily (2008) showed that iron and steel
and replaced with coarse-grained pebbles. The constructed (and rebar) electrodes produce the smallest errors and
infiltration basins have horizontal floors and are 1 m deep create very low noise in resistance and chargeability
with respect to the outer terrain. Infiltration is achieved by measurements. Rebar electrodes (∼50 cm length) which
sprinkling water over the basin through a suite of parallel were installed before the first measurement and retained
pipelines. throughout all the repeated measurements were used.
A tracer experiment was carried out from September to Resistivity measurements were carried out with an ABEM
October 2009 when the natural groundwater level is often Terrameter SAS4000 resistivity meter integrated with a
at its lowest. Salt infiltration tests were carried out over Lund Imaging System, which is a versatile, multi-channel
6 days by the addition of a total of 15 tonnes of NaCl, resistivity system for roll-along profiling (Dahlin 1996,
which gave a chloride content of 150 mg/L for the water 2001).
infiltrated. The natural chloride content of the groundwater Measurements were made using pole-dipole array,
was 19.1 mg/L (Persson 2008); chloride content at the end including forward and reverse measurements. The pole
of the experiment was 26 mg/L. at infinity for profiles P1 and P2 was located at 625 m
Following the infiltration of dosed water, the electrical north of the infiltration area (Fig. 1, site i1) and for the rest
conductivity of the water in the infiltration dams and in the of the profiles, it was located at 630 m south of the
observation wells (Fig. 1) was logged. Conductivity measure- infiltration area (Fig. 1, site i2).
ments were carried out by Karlskrona municipality starting Data processing, interpretation and initial evaluation
just after the infiltration process; data were recorded for were done continuously during the experiment. Each
9 weeks. repeated resistivity measurement was inverted using
time-lapse inversion, and then the percentage difference
relative to the inverted reference resistivity model was
ERT methodology calculated. Inversion was carried out with Res2DInv using
In order to trace the infiltration pathways spatially and robust (L1-norm type) data and model constraint which is
temporally, repeated resistivity measurements were made better at handling large contrasts in resistivity and noise in
along six profiles with fixed electrodes following an initial data compared to least-squares (L2-norm) inversion (Loke
set of reference resistivity measurements. Reference et al. 2003). The time-lapse difference (%) calculations
measurements were carried out in order to get a resistivity and plotting of the inverted sections were made by
image of the area under undisturbed conditions for each Erigraph software. Time-lapse datasets were inverted
profile. The reference data sets were measured when the simultaneously using a robust (Claerbout and Muir
soils at the surface were very dry, creating some problems 1973) smoothness constrain which minimises the absolute
in establishing good electrode contact, which to some changes in the model resistivity values (Loke 1999).
extent affected the quality of these data. Repeating
measurements started in the first week following the
initiation of the infiltration test, and during these measure- Results
ment rounds, electrode contact conditions were more
favourable. The first-week measurements were made on After the reference measurements of all the profiles, and
the closest profiles, Profiles P2 and P3. Repeated measure- after the start of the infiltration, time-lapse resistivity
ments on the remaining lines were completed in the measurements were conducted. Repetition times for all
second, third, fourth, sixth and eighth weeks.
Resistivity tests were carried out on six ∼E–W oriented Fig. 5 Resistivity profile P2. The reference resistivity profile and„
the percentage change in resistivity relative to the reference are
profiles (Fig. 1); two profiles are located at the northern shown for the repeated measurements. Features are defined as in the
side and four of them are at the southern side of the caption for Fig. 4. Areas marked withA, B, C and D represent the
infiltration area. The lengths of profiles P1, P2 and P5 are resistivity decrease and are discussed in the text
Hydrogeology Journal (2015) 23: 551–566 DOI 10.1007/s10040-014-1221-2
 557

Hydrogeology Journal (2015) 23: 551–566 DOI 10.1007/s10040-014-1221-2

558

profiles are summarized in Fig. 3. The initial measure- western edge were also decreasing in resistivity, while the
ments within the first week were conducted on the two percentage change in resistivity in the fourth week is lower
profiles closest to the infiltration area, profiles P2 and P3 compared to the third week (Fig. 4).
(Fig. 1), while the rest of the measurements were made In the sixth and eighth weeks, decrease in resistivity is
within the following second, third, fourth, sixth, and observed below the esker ridge and in the western part of
eighth weeks (Fig. 3). A total of 31 measurements were the esker. The change in the resistivity remained at more
conducted including the reference measurements. For each than –20 % between depths 30–40 m a.s.l., whereas the
repeat measurement, the resistivity change (in percentage) decrease at shallow (0–7 m) depths went below −40 %
was calculated. The profiles showing the reference (Fig. 4). In the fourth, sixth and eighth weeks, a very small
resistivity and the percentage change in resistivity are decline in resistivity (less than −20 %) is indicated at the
given in Figs. 4, 5, 6, 7 and 8. Bedrock level was observed eastern edge of the profile between 10 and 30 m a.s.l.
in 24 of 42 core logs of observation and production wells (Fig. 4, area B).
(between 18.4 and 39.2 m a.s.l.; Hebrand and Björkman
2005; Hägg et al. 2010) and interpreted with resistivity
sections. Inferred bedrock levels are plotted in the Profile P2
resistivity sections (Figs. 4, 5, 6, 7 and 8, dashed line, Profile P2 is one of the two profiles closest to the infiltration
BRL). Relative projections of the closest observation site (Fig. 1). Esker sediments stand out with higher
wells are also plotted in the reference resistivity sections resistivities (>1,600 ohm.m) at levels above ∼35 m a.s.l
(Figs. 4, 5, 6, 7 and 8). (Fig. 5). Separated with a seemingly vertical transition
around 40 m on the profile, the western edge of the profile
presents higher resistivities (>630 ohm.m) when compared
Resistivity signature of Johannishus Esker to the eastern edge (<630 ohm.m) below 30 m a.s.l.
and surrounding geology At all the repeated measurements, a negative change
Inverted model resistivities are in the range from <100 (down to below −40 %) in the resistivity was observed
ohm.m to >10,000 ohm.m. The esker ridge is represented below the two sides (Fig. 5, areas A, C) of the esker
by high resistivities over ∼1,600 ohm.m, which is mostly ridge. Another negative zone ∼5 m below the esker
because of the coarse-grained composition and dry state of ridge disappears in the sixth week (Fig. 5, area B).
the esker. The depth of the high-resistivity signature of the Starting from the third week, the area indicated with
esker extends to 15–25 m (Figs. 4, 5, 6, 7 and 8). Deeper ‘D’ in Fig. 5 presents negative change. This change is
areas in the sections present two distinct resistivity in excess of −40 % in the fourth week. As the intensity
characteristics—first, those with higher resistivities of the resistivity change decreases, the negatively
(>1,000 ohm.m), denoting most likely the bedrock at the affected area increases in size throughout time in this
eastern and western edges of all the profiles; and, second, profile (Fig. 5). In the shallowest few meters, while the
the low-resistivity areas (<630 ohm.m). These low- resistivities increase during the first time steps, decrease
resistivity areas were observed at (1) the middle portion can be observed following the third week. In the first
of the bedrock just beneath the esker ridge (Figs. 4, 5, 6, 7 time steps, increasing resistivities are seen at larger
and 8), representing most likely the fractured, water- depths in parts of the sections.
saturated bedrock (Fig. 2c); and (2) above the bedrock
along the profile (Figs. 6, 7 and 8), probably representing
the younger sedimentation related to fluvioglacial pro-
cesses over the bedrock.
Profile P3
Profile P3 is the closest profile to the southernmost
infiltration pond (Fig. 1). In the reference resistivity
section, the esker ridge appears again with a highly
Profile P1
resistive signature (Fig. 6, >1,600 ohm.m). A layer with
Profile P1 is the northernmost profile within the resistivity
160–630 ohm.m resistivities lies along the profile
sections (Fig. 1). Esker sediments are observed with high
between ∼15 and 33 m (a.s.l.), whereas below 20 m
resistivities (>1,600 ohm.m) at shallow depths
(a.s.l.), the western and eastern edges of the profile
(>38 m a.s.l.; Fig. 4). The eastern and western edges of
present high resistivities above 630 ohm.m, and a low-
the profile are represented with resistivities between 630
resistivity zone near the eastern side separates them
and 1,600 ohm.m. The middle part of the section between
(Fig. 6).
∼70 and 120 m horizontally presents low resistivities
Within the first week of the experiment, a rapid
below 630 ohm.m.
decrease in resistivity is indicated below 20 m a.s.l. at
The largest change in resistivity was observed from the
the western side of profile P3, (Fig. 6, area A), while at the
western edge to the 80th meter along the profile (Fig. 4, area
A). In the third and fourth weeks, a decline in resistivity
down to between −20 and −40 % was observed mainly under Fig. 6 Resistivity profile P3. The reference resistivity profile and„
the percentage change in resistivity relative to the reference are
the esker ridge and at the western edge of the profile. The shown for the repeated measurements. Features are defined in the
depth of this negative change extends to 27 m a.s.l. In the caption for Fig. 4. Areas marked with A, B, C, D and E represent the
third week, shallow depths (∼2–3 m from the surface) at the resistivity decrease and are discussed in the text
Hydrogeology Journal (2015) 23: 551–566 DOI 10.1007/s10040-014-1221-2
 559

Hydrogeology Journal (2015) 23: 551–566 DOI 10.1007/s10040-014-1221-2

560

eastern side, a relatively small decrease is observed steps. Following the third week, a decrease in the
(Fig. 6, area C). The area marked with ‘A’ in Fig. 6 has resistivity is observed in the shallow parts of the section.
high resistivity in the reference section. The change in
resistivity in this area reaches more than −60 % on the
third day. Following the third week, however, there is no Profiles P4 and P5
more decrease in this area. Rather, the layer marked with Profiles P4 and P5 are located on the southern edge of the
‘B’ shows a decrease in resistivity (this is the layer esker ridge (Fig. 1). Reference resistivity sections of
represented by 160–630 ohm.m resistivity range in the profiles P4 and P5 are analogous (Fig. 7) and they also
reference section which overlies the bedrock, Fig. 6). have similarities with profile P3 (Fig. 6). The esker ridge
After the third week, relatively lower negative changes appears as a high resistivity zone (>1,600 ohm.m); at
occur at the areas indicated by ‘D’ and ‘E’. greater depth, the eastern and western edges show
In the shallowest few meters, resistivity increase is resistivities higher than 630 ohm.m and the area in
obvious and most pronounced in the first couple of time between these edges is represented by low resistivities

Fig. 7 Resistivity profiles P4 and P5. Reference resistivity profiles and the percentage change in resistivity relative to the reference are
shown for the repeated measurements. Features are defined as in the caption for Fig. 4. Area marked with A represents the resistivity
decrease and is discussed in the text
Hydrogeology Journal (2015) 23: 551–566 DOI 10.1007/s10040-014-1221-2
 561

(<630 ohm.m). In profile P5, two zones defined by high steps after the reference, whereas a decrease is seen at the
resistivity (630–4,000 ohm.m) are also evident at depth end of the monitoring.
(Fig. 7).
The results for profiles P4 and P5 are clearer than the
previous resistivity profiles. In the second week, a Profile P6
decrease in resistivity below −60 and −40 % for profiles Profile P6 is located just outside of the esker ridge, on the
P4 and P5, respectively, is evident around the area plain surface (Fig. 1). High resistivities corresponding to
indicated with ‘A’ (Fig. 7). This decrease in resistivity esker sediments are visible in superficial depths (Figs. 8
spreads through the area under the esker in the third, and 2). Deeper in the section, high resistivities at the
fourth and sixth weeks, and for the profile P5, deeper eastern and western edges of the profile suggest bedrock.
infiltration can be interpreted in the fourth and sixth Similar to profile P3, a layer with lower resistivity lies
weeks. As with the previous, increasing resistivity is along the western part of the profile between 10 and
visible in the shallowest part in the first couple of time 35 m a.s.l. (Fig. 8).

Fig. 8 Resistivity profile P6. Reference resistivity profile and the percentage change in resistivity relative to the reference for the repeated
measurements. Features are defined as in the caption for Fig. 4. Areas marked with A and B represent the resistivity decrease and are
discussed in the text
Hydrogeology Journal (2015) 23: 551–566 DOI 10.1007/s10040-014-1221-2
562

A decrease in the resistivity is barely visible starting through the retreating ice. Even in englacial formation
from the fourth week and gradually spreading spatially at of eskers, the importance of structural control is noted
the end of the eighth week at the western edge of the by Burke et al. (2009).
profile (Fig. 8, area A). The intensity of the change is Blekinge is a lowland area exposed to a shallow sea,
more than −20 %. Another relatively smaller decrease is but with a characteristic relief caused by the N–S-trending
observed around 25 m a.s.l. (Fig. 8, area B). Zones of bedrock ridges (Berglund et al. 2005). Between these
increasing resistivity are scattered throughout the relative- ridges there are valleys with fine-grained sediments
ly shallow parts of the section. (Ringberg 1976; Berglund et al. 2005), and glaciofluvial
deposits such as eskers and deltas are common in these
N–S-trending tectonic valleys (Yu 2003).
Discussion Faults and inferred faults coincide with the relative
magnetic lows on the magnetic anomaly map of the
Structural control on esker formation region (Fig. 2b). The NNE–SSW directed Johannishus
In Finland, Hyyppä (1954) and Härme (1961) have Esker is elongated along an inferred fault and along a
related esker trends to fault lines. They proposed that NNE–SSW-directed negative magnetic anomaly
reactivation of old faults, perhaps accompanying iso- bounded by positive anomalies at eastern and western
static rebound, resulted in lines of weakness in the sides (SGU 2012; Fig. 2b). ERT results show that a
glacier that then become loci for meltwater flow corridor of low resistivity elongates through the esker
(Banerjee and McDonald 1975). The majority of the below 30 m a.s.l. (Figs. 4, 5, 6, 7, 8 and 2a, see the
eskers are isolated small ridges that overlie other area between the red lines). Cored drilling data indicate
deposits and follow the underlying topography that this low-resistivity area is the bedrock itself, while
(Campbell 2003). The majority of the meltwater the edges of the ERT profiles gave higher resistivities
channels follow existing bedrock/fault-controlled for the bedrock. The reason for the resistivity contrast
valleys and landscape (Banerjee and McDonald 1975; within the bedrock is most likely because this low-
Shreve 1985; Hebrand and Åmark 1989; Campbell resistivity area is highly fractured, saturated with water
2003) that form subglacial channels draining meltwater and possibly clay weathered, while another possibility

Fig. 9 Change in the ratio of the superficial (top 10 m) negatively affected area with respect to the total area (right vertical scale)
throughout the infiltration test with respect to precipitation (left vertical scale). Precipitation data are from Karlshamn station (∼36 km from
the MAR site) of the Swedish Meteorological and Hydrological Institute
Hydrogeology Journal (2015) 23: 551–566 DOI 10.1007/s10040-014-1221-2
 563

Fig. 10 Change in the affect-

ed area and change of electri-
cal conductivity in the
observation wells throughout
the infiltration test. a Area
marked with resistivity de-
crease showing the negatively
affected area to total area ratio,
b Electrical conductivity mea-
sured in the northern observa-
tion wells, c Electrical
conductivity measured in the
southern observation wells

could be clay infill in the lower parts of the valley. This indicating structural control on the formation of the
fracture zone itself is the fault zone, most probably esker. The paleo-valley controlling the subglacial
Hydrogeology Journal (2015) 23: 551–566 DOI 10.1007/s10040-014-1221-2
564

channel that is draining meltwater southward and, thus, sides of the infiltration site are given in Fig. 10b,c
forming the esker, was following this fractured zone of respectively. Two general trends were observed in the
the bedrock (Fig. 2c). conductivity measurements. The first is marked with an
Eskers may have been formed on tectonically con- almost linearly increasing trend with low slope angle. This
trolled valleys/zones; thus, it is important to state that for is the characteristic trend for the observation wells situated
an MAR site, escape of water through the vertical far from the infiltration site (Ox0309, Ox0308, Ox0321,
structural features is highly possible. In an unsaturated Ox0302, Ox0325, Ox0323; Figs. 1 and 10b,c). Conduc-
environment with similar structural conditions, extreme tivity observed in the northernmost observation well,
care must be taken to prevent the undesirable loss of Ox0321, was almost constant throughout the experiment.
water. The second trend is a rapid increase in conductivity until
the fourth and fifth weeks and then a decrease for the rest
of the experiment (Fig. 10b,c). This trend was observed in
Change in resistivity and conductivity during the the observation wells close to the infiltration site (Ox0311,
infiltration test Ox0324, Ox0326, Ox0303; Fig. 10b,c) and shows some
To visualise and quantify the change in the resistivity, the similarities with the negatively affected area in the
area showing a decrease in resistivity (negatively affected resistivity profiles near to them. After a rise in conductiv-
area) was calculated for each profile. This negatively ity until the fifth week, conductivity starts to decrease in
affected area was calculated both for shallow parts of the Ox0311 (Fig. 10b); a similar decrease in the negatively
sections (top 10 m; Fig. 9) to quantify the superficial affected area was observed in the closest resistivity profile
change in the resistivity and for the whole sections (profile P1) 1 week later (Fig. 10a). The same relation is
(Fig. 10a), which implies the infiltration pattern of the observed between the conductivities measured in obser-
salty water. Coloured lines in Figs. 9 and 10a present the vation wells Ox0303, Ox0324, Ox0326 (Fig. 10c) and the
‘ratio of the negatively affected area with respect to total/ negatively affected area in profiles P4 and P5 (Fig. 10a),
superficial area’ versus ‘time’ for each profile. both represented by a decrease starting with the fourth
The increase in resistivity in the shallow ground is week.
evident in many of the profiles in the early steps of the At the western edge of profiles P2, P3 and P6, negative
monitoring period (Figs. 4, 5, 6, 7 and 8). It is almost change in the resistivity is first observed in depths
certainly due to drying out of the shallow ground as it associated with the bedrock which is characterized by
coincided with a very dry period following the summer. high resistivity (Figs. 4, 6 and 8). This could be
The superficial decrease in resistivity that is generally interpreted as infiltration to well-developed fracture/joint
observed in all profiles in the latter weeks similarly was systems of the highly resistive bedrock. Another possible
preceded by rain (note the precipitation in Fig. 9). explanation could be 3D effects, i.e. the decrease that
A decrease in the resistivity was not observed directly appears to take place at depth is in fact taking place at
at the surface of the esker ridge, but mostly beneath the shallower depths beside the line. Infiltrated salty water
ridge and in the deeper areas. The elevation of the main rapidly draining into the fracture/joint systems of the rock
infiltration pond (southern dam) is below the ridge may reduce the resistivity of the rock. In profiles P4 and
(Fig. 1); therefore, the infiltration occurs mostly within P5, the initial decrease in the resistivity also started in the
the base levels of the esker. resistive areas (Fig. 7, area A); in profile P4, decrease was
When one focuses on the resistivity change in the expanded laterally and in profile P5, both laterally and
whole area of the sections, the largest negatively affected through the deeper areas in the section within the latter
area is observed on the profile P3, which is one of the two weeks (Fig. 7).
profiles closest to the infiltration site (Figs. 10a and 1). On the profiles traversing the road, a bipolar lenticular
The farthest profiles, profile P1 and P6, are the least shaped anomaly was observed beneath the road. The low
affected ones. Although profile P6 is almost 200 m away resistive zones might be associated with application of salt
from the infiltration site relative to profile P1, it is affected on the road for de-icing purposes; there may also be
at least as much as profile P1 (Fig. 10a). A similar artefacts due to the skipped electrodes (Figs. 4, 5, 6, 7 and
phenomenon is observed between profile P2 and profiles 8). The western edge of profile P5 is located very close to
P3, P4 and P5. Although the distances of profiles P3 and the road, which probably caused the abnormal increase of
P2 to the infiltration site are equal (88 m), decrease in resistivity in the repeated sections (Figs. 1 and 7).
resistivity was observed in a larger area in profile P3. In
the fourth week, the area showing a decrease in resistivity
is larger in profiles P4 and P5 with respect to profile P2. Conclusion
When the distance to the infiltration site is considered,
profiles lying south of the infiltration site are affected As inferred from the fault/fracture system mapped by
more than the ones at the northern side (Fig. 10a). This is magnetic and ERT measurements, it is evident that the
most likely because the natural groundwater flow direction structural control is an important factor in the formation of
is from north to south in the valley. Johannishus Esker. Pathways of the water during the
Results of the electrical conductivity measurements for infiltration test were mapped using ERT. The infiltrated
the observation wells situated at the northern and southern water appears to be concentrated in zones that are the most
Hydrogeology Journal (2015) 23: 551–566 DOI 10.1007/s10040-014-1221-2
 565

rapid pathways for groundwater transport. There is no Barker RD, Moore J (1998) The application of time-lapse electrical
indication that the infiltrated water escaped through tomography in groundwater studies. Lead Edge 17:1454–1458
Berglund BE, Sandgren P, Barnekow L, Hannon G, Jiang H, Skog
unexpected channels. Infiltrated water largely remained G, Yu S-Y (2005) Early Holocene history of the Baltic Sea, as
in a limited area relatively close to the infiltration ponds reflected in coastal sediments in Blekinge, southeastern Sweden.
and more distant observation wells were less affected by Quat Int 130:111–139
the infiltrated salty water. ERT and electrical conductivity Bolduc A, Paradis SJ, Riverin M-N, Lefebvre R, Michaud Y (2006)
measurements showed that the infiltrated water affected A 3D Esker geomodel for groundwater research: the case of the
Saint-Mathieu-Berry esker, Abitibi, Québec, Canada. In: Russell
the down-gradient area more than the up-gradient area HAJ, Berg RC, Thorleifson HL (eds) 3D geologic mapping for
relative to the location of the ponds. Natural groundwater groundwater applications: workshop extended abstracts. Geol
flow direction seems to be a determinant factor as Surv of Can Open File 5048
expected. Boulton GS, Hagdorn M, Maillot PB, Zatsepin S (2009) Drainage
beneath ice sheets: groundwater–channel coupling, and the
For efficient production of the filtrated water, location origin of esker systems from former ice sheets. Quat Sci Rev
of the recharging wells must be selected carefully in 28:621–638
relation to the preferential pathways for groundwater Brennand TA (2000) Deglacial meltwater drainage and glaciody-
transport. For example, profiles P3, P4 and P5 reveal namics: inference from Laurentide eskers, Canada. Geomorphol
locations that may be optimal for production wells from 32:263–293
Burke MJ, Woodward J, Russell AJ, Fleisher PJ (2009) Structural
that point of view. Transport time is also important for controls on englacial esker sedimentation: Skeiðarárjökull,
effective recharge, on one hand to be able to recharge Iceland. Ann Glaciol 50(51):85–92
sufficient amounts of water, but on the other hand to have Campbell JE (2003) Quaternary investigations in the Patterson
sufficient time in transit to achieve the desired cleaning Island area (part of NTS 64E-10 and -15), Reindeer Lake,
effect. Conductivity monitoring and water infiltrated areas Eastern Peter Lake Domain. In: Summary of Investigations
2003, vol 2. Misc. Rep. 2003–4.2, Paper A-7, Saskatchewan
mapped with ERT show that the highest concentration Geological Survey, Sask. Industry Resources, Regina, SK, 16
time was around 4–5 weeks with the tested production rate pp
(Fig. 10). This result was supported both by the Cassiani G, Bruno V, Villa A, Fusi N, Binley AM (2006) A saline
conductivity and ERT measurements (Fig. 10). trace test monitored via time-lapse surface electrical resistivity
The efficiency of the MAR sites may vary according to tomography. J Appl Geophys 59(3):244–259
Claerbout JF, Muir F (1973) Robust modeling with erratic data.
the geological, structural and hydrogeological conditions Geophys 38:826–844
of the site. Efficiency in terms of the rate of infiltrated and Corriols M, Dahlin T (2008) Geophysical characterization of the
recharged water does not promote good quality of the León-Chinandega aquifer, Nicaragua. Hydrogeol J 16:349–362
filtrated water. Water quality is also another important Dahlin T (1996) 2D resistivity surveying for environmental and
engineering applications. First Break 14(7):275–283
factor in determining the location of the production wells. Dahlin T (2001) The development of DC resistivity imaging
ERT measurements before the site selection and ERT techniques. Comput Geosci 27:1019–1029
monitoring during the initial stage of the functionality of de Geer J (1968) Some hydrogeological aspects on aquifers,
the MAR sites may reveal important information for especially eskers. In: Eriksson E, Gustafsson Y, Nilsson K
(eds) Groundwater problems. Pergamon, New York
further planning for efficient production. Flint RF (1971) Glacial and Quaternary geology. Wiley, New York
Frycklund C (1998) Long-term sustainability in artificial ground-
water recharge. In: Peters JH et al (eds) Artificial recharge of
Acknowledgements This work was funded by Karlskrona Munic- groundwater. Balkema, Rotterdam
ipality, Sweden, whom we thank for permission to publish the Frycklund C, Jacks G (1997) Iron and artificial recharge of
results. The field test and evaluation of the results, from a design groundwater. Boreal Environ Res 2:171–81
and production point of view, was carried out by Sweco VIAK, Greskowiak J, Prommer H, Massmann G, Johnston CD,
Malmö, Sweden. İnan Ulusoy was guest reseacher at Lund Nützmann G, Pekdeger A (2005) The impact of variably
University thanks to a grant from the Swedish Institute, which is saturated conditions on hydrogeochemical changes during
gratefully acknowledged. The authors are grateful to P. Comida for artificial recharge of groundwater. Appl Geochem 20:1409–
supporting the field activity. We are very grateful to Dr. Alfredo 1426
Mendoza and Dr. Mats Hebrand for their astute comments and Hägg M, Bergman B, Lindberg J, Hebrand M (2010) Karlskrona
constructive remarks on the earlier draft of the manuscript, and Jan Kommun, Karlskrona vatten: fullskaleförsök med saltinfiltration
Lindberg for help with documentation from the field test. We are och geofysiska mätningar [Karlskrona municipality, Karlskrona
thankful to Dr. Richard Owen for his constructive remarks and for water: full scale tests with salt infiltration and geophysical
improving the English of the manuscript. Furthermore, we wish to surveys]. SWECO VBB VIAK 1240121000, Sweco, Malmö,
thank Jacopo Boaga and two anonymous reviewers for constructive Sweden, 56 pp
criticism on the manuscript. Härme M (1961) On the fault lines in Finland. Bull Comm Geol
Finlande 196:437–444
Hebrand M (2001) Karlskrona municipality, Karlskrona water:
technical description of water activity at Johannishus ridge (in
Swedish). VBB VIAK 1240121, Malmö, Sweden, 48 pp
References Hebrand M, Åmark M (1989) Esker formation and glacier dynamics
in eastern Skåne and adjacent areas, southern Sweden. Boreas
Artimo A, Mäkinen J, Berg RC, Abert CC, Salonen V-P (2003) 18:67–81
Three-dimensional geologic modeling and visualization of the Hebrand M, Björkman F (2005) Karlskrona vattenförsörjning:
Virttaankangas aquifer, southwestern Finland. Hydrogeol J undersökning av infiltrationsområden vid Johannishusåsen
11:378–386 [Karlskrona water supply: investigation of infiltration areas at
Banerjee I, McDonald BC (1975) Nature of esker sedimentation. Johannishus Ridge]. SWECO VIAK 1240121, Sweco, Malmö,
Glaciofluvial Glaciolacustrine Sedimentol 23:132–154 Sweden, 59 pp
Hydrogeology Journal (2015) 23: 551–566 DOI 10.1007/s10040-014-1221-2
566

Hebrand M, Jeppson H (1999) Karlskrona municipality, Karlskrona Neuendorf KKE, Mehl JP Jr, Jackson JA (2005) Glossary of
water: hydrogeological investigations at Johannishus Ridge Geology, 5th edn. American Geological Institute, Alexandria,
1997–1999 (in Swedish). SWECO VBB VIAK 1240121, VA
Malmö, Sweco, Sweden, 130 pp Nilsson SI, Bergkvist B (1983) Aluminium chemistry and acidifi-
Helmisaari H-S, Kitunen V, Lindroos A-J, Lumme I, Monni S, Nöjd cation processes in a shallow podzol on the Swedish west coast.
P, Paavolainen L, Pesonen E, Salemaa M, Smolander A, Water Air Soil Pollut 20:311–330
Derome J (1998) Sprinkling infiltration in Finland: effects on Nöjd P, Lindroos A-J, Smolander A, Derome J, Lumme I,
forest soil, percolation water and vegetation. In: Peters JH et al Helmisaari H-S (2009) Artificial recharge of groundwater
(eds) Artificial recharge of groundwater. Balkema, Rotterdam through sprinkling infiltration: impacts on forest soil and the
Hyyppä E (1954) Äsarnas uppkomst [Genesis of ridges]. Geolog nutrient status and growth of Scots pine. Sci Total Environ
6:45 407:3365–3371
Katko TS, Lipponen MA, Rönkä EKT (2006) Groundwater use and Okkonen J, Jyrkama M, Kløve B (2010) A conceptual approach for
policy in community water supply in Finland. Hydrogeol J assessing the impact of climate change on groundwater and
14:69–78 related surface waters in cold regions (Finland). Hydrogeol J
LaBrecque D, Daily W (2008) Assessment of measurement errors 18:429–439
for galvanic-resistivity electrodes of different composition. Parsons ML (1970) Groundwater movement in a glacial complex,
Geophysics 73(2):F55–F64 Cochrane district, Ontario. Can J Earth Sci 7(3):869–883
Liang L, McCarthy JF, Jolley LW, McNabb JA, Mehlhorn TL Persson K (2008) Karlskrona vattenförsörjning, Utvärdering av
(1993) Iron dynamics: transformation of Fe(II)/Fe(III) during spårämnesförsök i Johannishusåsen, etapp Västra Vång [Eval-
injection of natural organic matter in a sandy aquifer. Geochim uation of tracer tests in Johannishus Ridge, West Vång stage].
Cosmochim Acta 57:1987–1999 SWECO 1240121000, Sweco, Malmö, Sweden, 8 pp
Lindroos A-J, Derome J, Niska K (1995) The relationship between Ringberg B (1969) Inventory of gravel deposits in the county.
dissolved organic matter and percolation water chemistry in County Administrative Board of Blekinge, Karlskrona, Sweden,
northern Finland. Water Air Soil Pollut 79:191–200 314 pp
Lindroos A-J, Kitunen V, Derome J, Helmisaari H-S (2002) Ringberg B (1976) Glacialgeologi och isavsmältning i östra
Changes in dissolved organic carbon during artificial recharge Blekinge [Glacial geology and deglaciation in the eastern
of groundwater in a forested esker in southern Finland. Water Blekinge]. SGU C661, Swedish Geological Survey, Uppsala,
Res 36:4951–4958 Sweden, 174 pp
Loke MH (1999) Time-lapse resistivity imaging inversion. Proceed- Russell AJ, Knudsen O, Fay H, Marren PM, Heinz J, Tronicke J
ings of the 5th Meeting of the Environmental and Engineering (2001) Morphology and sedimentology of a giant supraglacial,
Geophysical Society European Section, Em 1, Budapest, ice-walled, jökulhlaup channel, Skeidarárjökull, Iceland: impli-
September 1999 cations for esker genesis. Glob Planet Chang 28:193–216
Loke MH, Barker RD (1996) Rapid least squares inversion of SGU (2012) Digital geological and geophysical maps. Geol. Surv.
apparent resistivity pseudosections by a quasi-Newton method. of Sweden, Uppsala, Sweden
Geophys Prospect 44:131–152 Shreve RL (1985) Esker characteristics in terms of glacier physics,
Loke MH, Acworth I, Dahlin T (2003) A comparison of smooth and Katahdin esker system, Maine. Geol Soc Am Bull 96:639–646
blocky inversion methods in 2-D electrical imaging surveys. Singha K, Gorelick SM (2005) Saline tracer visualized with three-
Explor Geophys 34(3):182–187 dimensional electrical resistivity tomography: field-scale spatial
Lundqvist J, Wohlfarth B (2001) Timing and east–west correlation moment analysis. Water Resour Res 41(W05023):17
of south Swedish ice marginal lines during the Late Weichse- Syverson KM, Gaffield SJ, Mickelson DM (1994) Comparison of
lian. Quat Sci Rev 20:1127–1148 esker morphology and sedimentology with former ice-surface
Lundström US (1993) The role of organic acids in the soil solution topography, Burroughs Glacier, Alaska. Geol Soc Am Bull
chemistry of a podzolized soil. J Soil Sci 44:121–133 106:1130–1142
Mäkinen J (2003) Time-transgressive deposits of repeated deposi- Thurman EM (1985) Humic substances in groundwater. In: Aiken
tional sequences within interlobate glaciofluvial (esker) sedi- GR, McKnight DM, Wershaw RL, MacCarthy P (eds) Humic
ments in Köyliö, SW Finland. Sedimentology 50:327–360 substances in soil, sediment and water. Wiley, New York
Malmberg Persson K, Persson M (2009) Beskrivning till jordart- Warren WP, Ashley GM (1994) Origins of the ice-contact stratified
skartan 3 F Karlskrona NV & SV [Description of soil map 3 F ridges (eskers) of Ireland. J Sediment Res A64:433–449
Karlskrona NW & SW]. Sveriges Geologiska Undersökning K White PA (1988) Measurement of ground-water parameters using
155, SGU, Uppsala, Sweden, 11 pp salt-water injection and surface resistivity. Groundwater
Mendoza JA, Dahlin T, Barmen G (2006) Hydrogeological and 26:179–186
hydrochemical features of an area polluted by heavy metals in White PA (1994) Electrode arrays for measuring groundwater flow
central Nicaragua. Hydrogeol J 14:1052–1059 direction and velocity. Geophys 59:192–201
Morris M, Rønning JS, Lile OB (1996) Geoelectric monitoring of a Yu S-Y (2003) The Littorina transgression in southeastern Sweden
tracer injection experiment: modeling and interpretation. Eur J and its relation to mid-Holocene climate variability. PhD Thesis,
Environ Eng Geophys 1:15–34 Lundqua Thesis 51, Lund University, Sweden, 28 pp

Hydrogeology Journal (2015) 23: 551–566 DOI 10.1007/s10040-014-1221-2