Available online at www.sciencedirect.

com

ScienceDirect
Procedia Environmental Sciences 25 (2015) 135 – 141

7th Groundwater Symposium of the

International Association for Hydro-Environment Engineering and Research (IAHR)

Estimating hydraulic conductivity profiles using borehole

resistivity logs
V. K. Kalerisa* and A. I. Ziogasa
a
Hydraulic Engineering Laboratory, University of Patras, 26500 Patras, Greece

Abstract

Measurements of aquifer resistivity are intuitively attractive for estimating aquifer hydraulic conductivity because of
the fundamental relation between hydraulic conductivity and electrical conductivity; both of these properties depend
on porosity, grain size and packing configuration ([1], [2]). In this study porosity and hydraulic conductivity profiles
are estimated in three boreholes in Glafkos aquifer, located near the city of Patras in Western Greece. For this
purpose, resistivity measurements, the law of Archie [1], the Kozeny-Carman model [2] and numerical simulations of
pumping tests performed in the boreholes, have been used. It is shown that resistivity logs and pumping tests
constitute a useful data set for the estimation of hydraulic conductivity profiles. The reliability of the results has been
discussed. The relationships between porosity and hydraulic conductivity resulting for each of the three investigated
boreholes have been compared to each other as well as with literature data. The comparison confirms that the value of
the hydraulic conductivity for a given value of the porosity varies significantly.

© 2015
© 2015The TheAuthors. Published
Authors. by Elsevier
Published B.V. This
by Elsevier B.V.is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the Scientific Committee of the IAHR Groundwater Symposium 2014.
Peer-review under responsibility of the Scientific Committee of the IAHR Groundwater Symposium 2014

Keywords: hydraulic conductivity profiles; resistivity logs; pumping tests, numerical simulations.

1. Introduction

The coastal aquifer of Glafkos River (Fig. 1a) is a very important groundwater reservoir for the water
supply of the city of Patras. The aquifer covers an area of about 25km2. Due to its importance, a network
of seven observation wells shown in Fig. 1b has been constructed in the last years. The depth of the

* Corresponding author. Tel.: +30 2610 996517; fax: +30 2610 996572.
E-mail address: [email protected].

1878-0296 © 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the Scientific Committee of the IAHR Groundwater Symposium 2014
doi:10.1016/j.proenv.2015.04.019
136 V.K. Kaleris and A.I. Ziogas / Procedia Environmental Sciences 25 (2015) 135 – 141

boreholes varies from 80 m to 120 m and their diameter is 12 in. Groundwater level data collected during
the period from 2008 to 2012 have been used for the calibration of a groundwater model used for the
investigation of management issues (Ziogas, [3]). In order to refine the estimation of the hydraulic
properties of the aquifer, data collected during the construction of the boreholes, are used. In this study
we focus on the analysis of the data collected in the boreholes G1, G4 and G5 (Fig. 1b). The analysis of
the remaining wells is still in progress. The data collected in each of the aforementioned boreholes are: (a)
a cuttings log, (b) a gamma ray (GR) log, (c) a spontaneous potential (SP) log, (d) two electrical
resistivity logs, i.e. one short normal and one long normal (see below), and (e) water level measurements
in the borehole during a pumping test of short duration (three to four hours). The logs (b), (c) and (d),
which have been measured before the installation of the casing, have a vertical resolution 0.20 m.
The cuttings logs show that the aquifer consists of mixtures of coarse and fine gravel, sand and clay.
The GR-logs show that the values of the radiation are smaller than 35 API units. According to [4] (see
page 30 therein) such values correspond to clean formations, i.e. to formations with negligible clay
content. Concerning the SP-logs, they add little to the investigation, because, due to the lack of clay layers
in the formation, the clay base line cannot be established. Thus, concerning the structure of the formation
at the location of the boreholes G1, G4 and G5, the logs (a), (b) and (c) provide only qualitative
information.
Concerning the two types of the resistivity measurements, the difference between the short normal and
the long normal log consists in the spacing of the electrodes of the measuring device. In the short normal
device, which is characterized here as 16NR, the separation distance of the electrodes is 16 in (or 406
mm) whereas in the long normal, which is characterized here as 64NR, the separation distance of the
electrodes is 64 in (or 1630 mm). The spacing of the electrodes determines the depth of the current
penetration into the formation for a given borehole diameter ([4], [6]). Which of the aforementioned
resistivity measurements are appropriate for the estimation of the hydraulic conductivity profiles, can be
decided by considering the conditions, under which the law of Archie [1] that is used for the analysis of
the resistivity measurements, is valid.

Fig. 1. The aquifer of Glafkos River: (a) location of the aquifer; (b) the constructed network of observation wells

2. The proposed method

The first step in in the estimation of the hydraulic conductivity profile is to estimate the porosity
profile. This estimation is based on Archie’s law [1], which correlates the formation factor with the
porosity. The formation factor Fa is defined as the ratio of the resistivity of the porous material Ro
saturated with an electrolyte, to the bulk resistivity of the electrolyte Rw:
 V.K. Kaleris and A.I. Ziogas / Procedia Environmental Sciences 25 (2015) 135 – 141 137

(1)

Archie’s law is valid under the condition that the electrical resistivity of the pore fluid saturating the
porous material is very low as to completely dominate the current flow in the formation [5]. Such
conditions prevail in the zone closest around the borehole because the fluid saturating this zone, the so
called mud filtrate, is mainly drilling fluid and exhibits low resistivity (see Fig. 8.18 in [6]). Thus, for the
analysis based on Archie’s law, resistivity measurements, which are representative for the area saturated
with this low resistivity fluid, are required. Since the short normal resistivity device measures the
resistivity in the area closer to the borehole than the long normal device, it is reasonable to assume that
the resistivity values measured by means of the short normal device are representative for the
aforementioned area. Therefore, we perform the estimation of the hydraulic conductivity profiles using
the short normal resistivity logs. It should be noticed that for the analysis, no elimination of clay
contamination effects has been performed (see [7]), since the gamma ray logs show that the clay content
of the formation is not significant (see section 1).
The formation factor [Eq.(1)] is calculated from the resistivity values R16NR obtained from the 16NR-
log, dividing them by the resistivity of the mud filtrate. In our investigations only the resistivity of the
drilling mud has been measured. Concerning the relationship between the resistivity of drilling mud Rm
and that of the mud filtrate Rmf there are contradictory opinions. Sherborne and Newton [8] found
experimentally that the resistivity of the drilling mud in most cases closely approximates the resistivity of
the mud filtrate. Patnode [9] concludes that it is erroneous to substitute drilling mud resistivity for mud
filtrate resistivity and Lamont [10] found experimentally the relationship: Rmf = 0.88Rm. Finally,
Schneider [11] (see Fig. 4.177, page 304 therein) qualitatively indicates that Rmf is slightly smaller than
Rm. Due to the lack of data, we set in this analysis the resistivity of the mud filtrate equal to the measured
resistivity of the drilling mud. It should be noticed that uncertainties in the resistivity of the mud filtrate
significantly influence the resulting porosity and hydraulic conductivity profiles.
In order to calculate the porosity φ from the formation factor Fa= R16NR/Rmf, we use Archie’s law in the
form of Eq.(2), which has been proposed by Schneider [11] for unconsolidated sediments:

(2)

More usually an equation of the form:






is used (see [12]). Since there are no generally valid
rules for the selection of the parameters α and m in this equation, we preferred here to use Eq.(2).
For the estimation of the hydraulic conductivity from the porosity, functional relationships between
porosity and hydraulic conductivity could be used. However, investigations on this issue are either
performed for definite types of sediments (see [13]) or they provide results according to which the
hydraulic conductivity varies over some orders of magnitude in relatively narrow ranges of the porosity
([13], [14]). This is due to the fact that the hydraulic conductivity does not depend only on the porosity
but also on the grain size and the packing configuration [2]. From data presented in the literature (see for
instance [15] and [16]), the relationship between porosity and grain size seems to be more unambiguous.
Therefore, the procedure we follow in this study in order to estimate the hydraulic conductivity profile
consists in the following steps: (a) we estimate the profile of the grain size by using the porosity profile,
which has been estimated from the formation factor [see Eq.(2)], and a relationship between porosity and
grain size (see below). (b) We use then a grain size–porosity-hydraulic conductivity relationship in order
to estimate the hydraulic conductivity. In the literature, a large number of such relationships (see for
instance [2] and [13]) has been proposed. In this study the Kozeny-Carman equation given in [2] has been
used:











(3)










138 V.K. Kaleris and A.I. Ziogas / Procedia Environmental Sciences 25 (2015) 135 – 141

γw is the specific weight of the water, μw is the dynamic viscosity of the water and de is the effective grain
diameter.
The relationship φ= f(de) between porosity and effective grain size de for each of the three boreholes
investigated here has been estimated adaptively, through a calibration procedure based on the pumping
test data. We start for each borehole with a form of the function φ= f(de), which results by approximating
the data presented by Davis [15] and De Marsily [16] (see Fig. 2). From this initial form of the function
φ=f(de) and the previously estimated φ-values [Eq.(2)], we calculate the hydraulic conductivity profile
using Eq.(3).

10
Data from Davis [15]
1

0.1
Grain Size in mm

0.01
Adaptively estimated curve
1E-3

1E-4
Data from de Marsily [16]
1E-5

1E-6
0 10 20 30 40 50 60
Porosity

Fig. 2. Estimation of the porosity – grain size relationship for the observation well G1.

This profile is then used for the numerical simulation of the pumping test performed in the borehole.
Since the vertical discretization of the numerical model is coarser than the vertical resolution of the
hydraulic conductivity profiles obtained from the resistivity logs, the hydraulic conductivity profiles used
for the simulation have been respectively aggregated.
Depending on the deviations between the groundwater level measured during the pumping test in the
borehole and that simulated, the function φ= f(de) is respectively modified and a new hydraulic
conductivity profile is calculated. Since the results of the simulation are more sensitive with respect to the
hydraulic conductivity values at the screened sections of the borehole, the function φ= f(de) is each time
modified over its whole range, however, so that a higher or a lower value for the hydraulic conductivity at
the screened section results, depending on the deviations between the measured and the simulated
drawdown after each simulation. Thus, in this adaptive procedure the variation of the hydraulic
conductivity of all the layers of the profile is not arbitrary but it is conditioned on the resistivity log.
Details concerning the simulations of the pumping tests are given in Ziogas [3]. The MODFLOW-
2000 code [17] has been used. In order to minimize the computational effort, two-dimensional, cross-
sectional aquifer models were used and the technique proposed by Langevin [18] for simulating axially
symmetric (radial) flow in a 2D-model has been applied. The horizontal discretization is fine in the
vicinity of the borehole and becomes logarithmically coarser toward the outer model boundary. The
vertical discretization is fine in the screened parts of the casing and coarser in the rest of the model.
Concerning the discretization in time, an initial time step of the order of 0.1ms has been used for the
drawdown phase whereas for the recovery phase the initial time step was of the order of 1s. In both cases
the time step was increased logarithmically. The borehole-column was simulated as a high conductivity
zone and the pumping was assigned to the bottom cell of the borehole-column. The cells neighboring the
borehole-column have been considered to be inactive, representing the casing of the wells, except of those
corresponding to the screens. A number of columns surrounding the well casing were used to simulate the
filter pack of the well. According to the grain size of the filter material, its hydraulic conductivity is
 V.K. Kaleris and A.I. Ziogas / Procedia Environmental Sciences 25 (2015) 135 – 141 139

considered to vary between 500 m/d and 800 m/d. The columns surrounding the filter pack correspond to
the aquifer. The top and bottom layers of the model are no flow boundaries. The outer model boundary is
a constant head boundary with a value equal to the groundwater level, which was measured prior to the
beginning of the pumping test. The distance of this boundary from the well column was set large enough,
to ensure that the boundary does not affect the head drop around the well. To calculate the hydraulic head
inside the well, observation points were set to all the cells in the screened part of the casing. The heads at
the observation points were averaged using weights, which are related to the thickness of the layer that
contains the corresponding observation point. These averaged heads were compared with the measured
heads. The hydraulic conductivity profile providing the simulation results shown in Fig. 3 is considered to
sufficiently approximate the water level variation measured during the pumping test in borehole G1 and is
accepted as the final hydraulic conductivity profile for this location (see Fig. 4a).

Fig. 3. Comparison between the measured water level during the pumping test in borehole G1 and the water level simulated by
using the hydraulic conductivity profile considered as the final one for this borehole: (a) drawdown period; (b) recovery period.

3. Results and discussion

Figure 4a shows the hydraulic conductivity profiles estimated for the thickness of the saturated zone of
the aquifer penetrated by the boreholes G1, G4 and G5. The depth of the boreholes G1 and G4 is 120m.
The difference in the aquifer thickness penetrated by each of them is due to the fact that the altitude at the
location G1 is significantly larger than that at G4 whereas the groundwater level above the mean sea level
in the two boreholes during the construction period differs only slightly. The depth of the borehole G5 is
80 m and it is located at even lower altitude than G4. The estimated profiles show that the aquifer at the
locations G1 and G4 exhibits significant heterogeneity over the depth. At the location G5 the aquifer is
more homogeneous.
Concerning the reliability of these results there are two indications that the estimated profiles are
reliable. These are: (a) that the profiles simulate the reaction of the aquifer for the short duration pumping
test performed in each borehole and (b) that the profiles are compatible with the resistivity logs. However,
the analysis is based on a series of assumptions, the most important of which concern: (i) the functional
relationship between the formation factor and the porosity [Eq.(2)], (ii) the functional relationship
between porosity and hydraulic conductivity [Eq.(3)] and (iii) that clay contamination effects are not
140 V.K. Kaleris and A.I. Ziogas / Procedia Environmental Sciences 25 (2015) 135 – 141

significant (see section 1). Due to these assumptions, a sensitivity analysis is required for the final
evaluation of the profiles.
Figure 4b shows the relationship between the porosity and the hydraulic conductivity resulting for the
three boreholes investigated here, along with data presented in the literature (see Nelson [13] and Bush
and Luckner [19]), which concern different materials. The hydraulic conductivity and porosity values of
the materials of the Glafkos aquifer are within the wide range of these variables given in [13] but they are
not compatible with the results presented in [19], particularly for the high hydraulic conductivity values.
Further, the hydraulic conductivity-porosity relationships for the boreholes G1 and G5 are similar to each
other but different from the corresponding relationship resulting for G5.

G1 (this study)
0.01
G4 (this study)
G5 (this study)
1E-3

Hydr. Conductivity in m/s

20
Depth with respect to mean sea level im m

uncons. sands [13]

G5 (this study) gravely sand [19]
10 1E-4
G4 (this study) middle-grained sand [19]
0 G1 (this study) sandy gravel [19]
1E-5 sandy loam [19]
-10
loamy sand [19]
-20 1E-6 clay loam [19]
-30 silty clay [19]
1E-7
-40
1E-8
-50
-60 1E-9
0 10 20 30 40 50 60
-70
(b) Porosity
-80

1E-7 1E-6 1E-5 1E-4

(a) Hydr. Conductivity in m/s

Fig. 4. Results for the analysis of the logs in boreholes G1, G2, G3: (a) Hydraulic conductivity profiles; (b) Comparison of the
porosity – hydraulic conductivity relationships with literature data.

These results confirm that assessments of the value of hydraulic conductivity, which are based only on
porosity values, are uncertain. This fact is emphasized also by the field experiments of Morin et al. [5],
which show that there are aquifers, in which the correlation between porosity and hydraulic conductivity
is not significant. Thus, for the estimation of the hydraulic conductivity profiles from the porosity
profiles, additional information is required. Here has been shown that such information can be obtained
from the pumping test performed in each borehole.
It should be noticed that hydraulic conductivity profiles can be estimated also from flowmeter
measurements (Molz et al., [20], Kaleris et al., [21]). The difference between resistivity logs and
flowmeter measurements is that the resistivity logs, which are always constructed before the installation
of the casing, provide information over the whole depth of the aquifer. On the contrary, flowmeter
measurements, which in unconsolidated sedimentary aquifers mostly cannot be performed in uncased
boreholes, provide information over the whole depth of the aquifer only if the borehole is fully screened.
For partially screened boreholes flowmeter measurements provide information mainly for the layers,
which correspond to the screened parts of the borehole. Due to the fact that resistivity logs are low cost
measurements, their combination with simple pumping tests, as described here, represent a cost effective
method for the estimation of hydraulic conductivity profiles.
 V.K. Kaleris and A.I. Ziogas / Procedia Environmental Sciences 25 (2015) 135 – 141 141

Acknowledgements

This investigation has been performed in the frame of a cooperation project between the Hydraulic
Engineering Laboratory of the Department of Civil Engineering of the University of Patras and the
Municipal Enterprise for Water Supply and Sewage of Patras. The project was funded in the frame of the
EU-project INTERREG IIIA Greece-Italy 2000-2006 (Grant I3101034). The contribution of Irene
Karathanasi, from the Municipal Enterprise for Water Supply and Sewage of Patras, who managed the
planning and the construction of the boreholes, is highly appreciated.

References

[1] Archie GE. The electrical resistivity log as an aid in determining some reservoir characteristics. Transactions, Amer. Inst.
Mining, Metallurgy and Petroleum Enginnering 1942; 54-62.
[2] Bear J. Dynamics of Fluids in Porous Media.New York: Dover; 1972.
[3] Ziogas AI. Investigation of coastal aquifer management issues through the use of numerical models. PhD thesis, Hydraulic
Engineering Laboratory, Department of Civil Engineering, University of Patras (in Greek), 2013.
[4] Ellis DV, Singrer JM. Well Logging for Earth Scientists, 2nd Edition. Springer; 2008.
[5] Morin RH, LeBlanc DR, Troutman BM. The influence of Topology on Hydraukic Conductivity in a Sand-and-Gravel Aquifer.
Ground Water 2010; 48, No. 2; 181-190.
[6] Driscoll FG. Groundwater and Wells. 2nd ed. Minnesota: Johnson Division; 1987.
[7] Khalil MA, Ramalho EC, Monteiro Santos FA. Using resistivity logs to estimate hydraulic conductivity of a Nubian sandstone
aquifer in southern Egypt. Near Surface Geophysics 2011; 9, No 4; 349-355.
[8] Sherborne JE, Newton WM. Well Logging Factors Influencing Electrical Resistivity of Drilling Fluids. AIME Transactions
1946.
[9] Patnode WH. Electrical Logging-Relationship of Drilling Mud Resistivity to Mud Filtrate Resistivity. AIME Transactions 1949.
[10] Lamont N. Relation Between the Mud resistivity, Mud Filtrate Resistivity and the Mud Cake Resistivity of Oil Emulsion Mud
Systems. AIME Petroleum Transactions 1957; 210: 387-388.
[11] Schneider H. Die Wassererschließung. 2. Auflage Essen: Vulkan; 1973.
[12] Urish DW. Electrical Resistivity-Hydraulic Conductivity Relationships in Glacial Outwash Aquifer. Water Resour Res 1981;
17, No. 5; 1401-1407.
[13] Nelson PH. Permeability-Porosity Relationships in Sedimentary Rocks. The Log Analyst 1994; May-June; 38-62.
[14] Detmer DM. Permeability, porosity, and grain-size distribution of selected Pliocene and Quaternary sediments in the
Albuquerque Basin. New Mexico Geology 1995; November; 79-87.
[15] Davis SN. Porosity and Permeability of Natural Materials. In: De Wiest RJM, editor. Flow Through Porous Media. New York,
London: Academic Press, 1969, p. 53-89.
[16] De Marsily G. Quantitative Hydrogeology. San Diego: Academic Press Inc., 1986.
[17] Harbaugh AW, Banta ER, Hill MC, McDonald MG. MODFLOW-2000, The U.S. Geological Survey Modular ground-water
model – user guide to modularization concepts and the ground-water flow process, U.S. Geological Survey, Open-File Report
00-92, Reston, Virginia, 2000.
[18] Langevin CD. Modeling axisymmetric flow and transport. Ground Water 2008; 46, No. 4; 579-590,
[19] Bush KF, Luckner L. Geohydraulik. Stuttgart: Ferdinand Enke; 1974.
[20] Molz FJ, Morin RH, Hess JG, Güven O. The impeller meter for measuring permeability variation: Evaluation and comparison
with other tests. Water Resour Res 1989; 25, No. 7; 1677-1683.
[21] Kaleris V, Hadjitheodorou c, Demetracopoulos AC. Numerical simulation of field methods for estimating hydraulic
conductivity and concentration profiles. J. Hydrol. 1995; 171; 319-353.