Communications in Soil Science and Plant Analysis

ISSN: 0010-3624 (Print) 1532-2416 (Online) Journal homepage: http://www.tandfonline.com/loi/lcss20

Establishing soil electrical conductivity ‐ Depth

relations from electromagnetic induction
measurements

D. L. Corwin & J. D. Rhoades

To cite this article: D. L. Corwin & J. D. Rhoades (1990) Establishing soil electrical conductivity
‐ Depth relations from electromagnetic induction measurements, Communications in Soil
Science and Plant Analysis, 21:11-12, 861-901, DOI: 10.1080/00103629009368275

To link to this article: http://dx.doi.org/10.1080/00103629009368275

Published online: 11 Nov 2008.

Submit your article to this journal

Article views: 12

View related articles

Citing articles: 11 View citing articles

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=lcss20

Download by: [University of Florida] Date: 05 June 2016, At: 09:06

 COMMUN. IN SOIL SCI. PLANT ANAL., 21(11&12), 861-901 (1990)

ESTABLISHING SOIL ELECTRICAL CONDUCTIVITY - DEPTH RELATIONS FROM

ELECTROMAGNETIC INDUCTION MEASUREMENTS

D. L. Corwin and J. D. Rhoades

Downloaded by [University of Florida] at 09:06 05 June 2016

USDA-ARS
U.S. S a l i n i t y Laboratory
4500 Glenwood Drive
R i v e r s i d e , CA. 92501

ABSTRACT;

P r i n c i p l e s of electromagnetic i n d u c t i o n (EM) and f i e l d

c a l i b r a t i o n approaches a r e discussed a s they p e r t a i n t o t h e
a p p l i c a t i o n of EM t o s o i l systems f o r t h e purpose of d e r i v i n g
soil electrical conductivity - depth relations. Evidence i s
provided to support the u t i l i t y of EM-derived estimates of
ECa-depth relations. Limitations of using electromagnetic
induction to determine ECa for discrete depth intervals through
the soil are discussed. Current research designed to increase
the accuracy of ECa-depth determinations by dealing with the
spatial variability problem associated vith salinity in soil and
by mitigating some of the inherent limitations of the calibration
approaches i s described.

861

Copyright © 1990 by Marcel Dekker, Inc.

 862 CORWIN AND RHOADES

INTRODUCTION;

I± S o i l s a l i n i t y can be determined i n t h e f i e l d from

measurements of bulk s o i l e l e c t r i c a l c o n d u c t i v i t y , ECa ( 1 , 2 ) . A
depth-weighted value of s o i l e l e c t r i c a l c o n d u c t i v i t y through t h e
s o i l p r o f i l e , ECg, can be determined from above-ground
electromagnetic (EM) measurements ( 1 1 ; 1; 2 ) . The p e n e t r a t i o n
depth of t h e EM measurement depends on t h e c o i l o r i e n t a t i o n , t h e
spacing between t r a n s m i t t e r and r e c e i v e r c o i l s , and t h e
Downloaded by [University of Florida] at 09:06 05 June 2016

e l e c t r i c a l frequency of t h e instrument.
The p o t e n t i a l f o r t h e use of electromagnetic i n d u c t i o n (EM)
techniques a s a means of performing reconnaissance surveys of
soil salinity i s self-evident. Williams and Baker (19) measured
ECg t o depths ranging from 7.5 t o 60 m f o r an area of 10,000 km2
using EM t e c h n i q u e s . From t h e s e measurements a r e a s of apparent
high s a l i n i t y were i n f e r r e d . Others have demonstrated t h a t s o i l
bodies of widely d i f f e r i n g s a l i n i t y can be d e l i n e a t e d by EM
techniques ( 4 ) .
Since plant root activity occurs primarily within the top
0.9 m of soil, the electrical conductivity of this portion i s
extremely important in assessing soil salinity from the
standpoint of agricultural productivity. Once areas of potential
soil salinity hazard are delineated, i t i s desirable to be able
to survey these areas more intensively within relatively shallow
soil depths (i.e., 0-0.3, 0.3-0.6, 0.6-0.9 etc. m). In situ
methods of measuring soil electrical conductivity within such
depths are available (9; 10), but remote measurements would offer
advantages of reduced labor, time and cost. Rhoades and Corwin
(11) demonstrated that bulk soil electrical conductivity, ECa,
could be determined for discrete depth intervals of the soil
profile from EM readings made at a succession of heights above
the ground (e.g., 0, 30, 60, 90 and 120 cm) by solving a set of
empirical equations containing depth-specific coefficients. The
values of the calibration coefficients utilized in these
equations were determined by multiple linear regression analysis
 SOIL ELECTRICAL CONDUCTIVITY-DEPTH RELATIONS 863

of successive EM readings measured with a Geonics EM-381»2/

device, and of corresponding ground t r u t h ECa values for the
soil-depth i n t e r v a l s of 0-30, 30-60, 60-90 and 90-120 cm measured
using a Martek'1'3/ SCT, four-electrode probe (10).
A simpler and almost a s accurate method was l a t e r developed
to determine the d i s t r i b u t i o n of ECa within the s o i l p r o f i l e
using only two EM measurements taken at the s o i l surface with the
long-axis of the EM-38 device's electromagnet oriented p a r a l l e l
(EMfj) and then perpendicular (EMy) t o the s o i l surface (1).
Downloaded by [University of Florida] at 09:06 05 June 2016

Equations containing depth-specific c o e f f i c i e n t s were developed

r e l a t i n g ECa within s p e c i f i c soil-depth i n t e r v a l s t o EM^ and EMy.
Corwin and Rhoades (2) l a t e r found t h a t d i f f e r e n t
c o e f f i c i e n t s were needed i n the equations r e l a t i n g ECa of the
s o i l p r o f i l e t o EM() and EMy for s o i l s i n which ECa decreased with
depth, i . e . , inverted s a l i n i t y p r o f i l e s , compared t o s o i l s i n
which ECa increased with depth, i . e . , normal s a l i n i t y p r o f i l e s .
The data s e t s used t o forumulate the c o e f f i c i e n t s for both
inverted and normal s a l i n i t y p r o f i l e s were q u i t e limited (<20
profiles). A s u b s t a n t i a l l y l a r g e r data s e t (about 900 p r o f i l e s )
was subsequently used, along with more rigorous s t a t i s t i c a l
techniques t o obtain new c a l i b r a t i o n c o e f f i c i e n t s for the
empirical r e l a t i o n s of Corwin and Rhoades ( 1 , 2); as a r e s u l t ,
more consistently accurate predictions were obtained (15).
This paper reviews the c a l i b r a t i o n methods and the
p r a c t i c a l f a c t o r s required for applying electromagnetic induction
to s o i l systems for the purpose of developing ECa-depth
relations. In addition, the l i m i t a t i o n s t o the methodology of
converting EM measurements t o ECa for d i s c r e t e depth i n t e r v a l s

1/ Product i d e n t i f i c a t i o n i s provided solely f o r the benefit of

the reader and does not imply the endorsement of the USDA.

2/ Geonics Limited, 1745 Meyerside Dr., Mississauga, Ontario,

Canada.

3/ Martek Instruments, Inc., 17302 Daimler St., I r v i n e , CA.

 864 CORWIN AND RHOADES

a r e d i s c u s s e d a s well a s c u r r e n t r e s e a r c h intended t o m i t i g a t e
the influence of these limitations upon the practical application
of EM as a soil salinity survey tool.

Principles of Electromagnetic Induction and Description of

Equipment :

II. Current flow i s induced in a conductor (such as a moist,

saline soil) that i s within a primary electromagnetic (EM)
Downloaded by [University of Florida] at 09:06 05 June 2016

field. This induced current flow i s proportional to the

electrical conductivity of the conducting body (ECa) for a given
strength of EM field. The current flow, in turn, creates a
secondary electromagnetic field, the strength of which i s
proportional to the current flow, and hence, to ECa. Thus, ECa
may be inferred from the magnitude of the induced secondary EM
field generated upon imposition of a primary EM field on the
conductor (6). This permits a depth-weighted ECa, ECa, to be
measured without soil contact using inductive electromagnetic
techniques.
An electrodeless electrical conductivity meter, operating
on the above described principle, was constructed by Geonics
Limited^/ to determine EC| within relatively shallow soil
depths. The unit i s designated by Geonics as the EM-38 soil
electrical conductivity meter or more simply as the EM-38 (see
Figure 1).
The EM-38 device i s a reduced-scale modification of another
Geonics Limited device, the EM-31, which i s used for geophysical
surveying purposes. The EM-31 unit i s not as suitable for soil
salinity appraisal purposes because i t responds to the EC of
earth material below the root zone depth; about 50% of i t s
reading i s associated with the material below 2.75 m. Figure 1
shows the EM-38 instrument lying in the "horizontal" position
(coil axis parallel to the soil surface) and in the "vertical"
position (coil axis perpendicular to the soil surface). The
EM-38 i s a totally portable device. I t has an intercoil spacing
 SOIL ELECTRICAL CONDUCTIVITY-DEPTH RELATIONS 865
Downloaded by [University of Florida] at 09:06 05 June 2016

Pig. 1. Geonics EM-38 prototype electromagnetic s o i l

conductivity meter (top) lying i n the horizontal
position with i t s c o i l s p a r a l l e l t o the s o i l surface,
and (bottom) lying i n the v e r t i c a l position with i t s
c o i l s perpendicular t o the s o i l surface.

of 1 m, operates a t a frequency of 13.2 kHz, i s powered by a 9V

t r a n s i s t o r b a t t e r y , and reads EC^ d i r e c t l y i n three ranges
(0-100, 0-300, and 0-1,000 mmho/m). The c o i l configuration and
i n t e r c o i l spacing, s (s = 1 meter for the EM-38), were chosen t o
permit measurement t o e f f e c t i v e depths of approximately 1 or 2 m
when the instrument was placed a t ground level with c o i l s
p a r a l l e l p,r perpendicular t o the s o i l surface, r e s p e c t i v e l y . The
device contains the appropriate c i r c u i t r y t o minimize instrument
response t o the magnetic s u s c e p t i b i l i t y of the s o i l and t o
maximize response t o ECa.
 866 CORWIN AND RHOADES

The b a s i c p r i n c i p l e of o p e r a t i o n of t h e EM-38 s o i l
e l e c t r i c a l c o n d u c t i v i t y meter i s shown s c h e m a t i c a l l y i n Figure
2. A t r a n s m i t t e r c o i l (T) l o c a t e d a t one end of t h e instrument
induces c i r c u l a r eddy c u r r e n t loops i n t h e s o i l . The magnitude
of t h e s e loops i s d i r e c t l y p r o p o r t i o n a l t o ECa i n t h e v i c i n i t y of
t h a t loop. The c u r r e n t loops generate a secondary
electromagnetic field that i s proportional to the value of the
current flowing within the loop. A fraction of the induced
electromagnetic field from the loops i s intercepted by the
Downloaded by [University of Florida] at 09:06 05 June 2016

receiver coil (R), and the signal i s amplified and transformed

into an output voltage that i s a measure of ECa.
The cumulative relative contribution, R(Z), to the
instrument reading of a l l soil below Z (where Z i s the depth
divided by the intercoil spacing, s) i s shown in Figure 3 for a
homogeneously conductive material 4 /. This relation was
established by the instrument's developer. The relation shows
that 75% of the device's response comes from depths of <90 cm
when the instrument i s placed in the horizontal position. The
corresponding depth when the device i s read in the vertical
position i s 190 cm. Thus, measurements of ECa appropriate to the
whole root zone and even deeper may be measured in single
integrated readings. From the EM-38's ECa measurement, ECa for
discrete depth intervals can be determined as shown by Rhoades
and Corwin (11), Corwin and Rhoades (1), Corwin and Rhoades (2)
and Rhoades et al (15).

4/ The relative contributions to the secondary magnetic field

(or ECa) from a l l material below a depth Z for vertical are
calculated from
RV(Z) = 1/(4Z2 + 1)½,
2
and R H ( Z ) = (4Z + l)½ - 2Z,

for vertical and horizontal dipoles, respectively (McNeill,

J. D., 1980, Electromagnetic terrain conductivity measurement
at low induction numbers, Technical Note TN-6, Geonics
Limited, Mississauga, Ontario, Canada)
 SOIL ELECTRICAL CONDUCTIVITY-DEPTH RELATIONS 867

t:
CURRENT LOOPS
Downloaded by [University of Florida] at 09:06 05 June 2016

T - TRANSMITTER COIL
R-RECEIVER COIL

INDUCED CURRENT FLOW IN GROUND

Fig. 2. Principle of operation of the electromagnetic soil

conductivity meter.

Development of Empirical Calibrations;

I I I . I n i t i a l efforts to determine ECa for discrete depth

intervals from above-ground EM measurements (i.e., from ECa) were
based upon consideration of the variable nature of the depth
response relation (Figure 3). I t was hypothesized that since
varying the height of the EM-38 above the soil surface changes
the relative strength of the EM field with depth in the soil, an
ECa-depth distribution could be determined from a succession of
EM readings taken at various heights above the soil. Attempts to
use the relation described i n Figure 3 and a series of
simultaneous equations corresponding to each measurement height
 868 CORWIN AND RHOADES
Downloaded by [University of Florida] at 09:06 05 June 2016

ce
0.4

0.3

0.2

O.I
Horizontal

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 24 2.6 2.8 3.0
DEPTH Z(meters)

Fig. 3. Cumulative relative contribution of a l l soil electrical

conductivity, R(Z), below various depths for the EM-38
reading when the device i s held in a horizontal
(parallel) and vertical (perpendicular) position.

failed to give acceptable results. Therefore, s t a t i s t i c a l

relationships (or calibrations) between the EM-38 readings a t
various heights above ground and the soil profile ECa
distribution were established using data obtained at nine
different locations near Lakeview, California. The EM-38
readings were found to be highly correlated with the soil
 SOIL ELECTRICAL CONDUCTIVITY-DEPTH RELATIONS 869

depth-ECa d i s t r i b u t i o n s f o r t h i s study area. The c a l i b r a t i o n

c o e f f i c i e n t s obtained were as follows:

EM0 = 0.4418 EC, + 0.2049 EC2 + 0.1045 EC3 + 0.0604 EC4

+ 0.1940 EC, - 0.0005; r 2 = 0.999 [la]

EM, = 0.0829 EC, + 0.2526 EC2 - 0.0507 EC3 + 0.0736 EC4

+ 0.1658 EC5 + 0.0143; r 2 = 0.988 [lb]
Downloaded by [University of Florida] at 09:06 05 June 2016

EM2 = 0.6662 EC, - 0.0231 EC2 + 0.0412 EC3 + 0.0307 EC4

+ 0.0324 ECS + 0.1188; r 2 = 0.994 tic]

EM3 = 0.5512 EC, - 0.0479 EC2 + 0.0339 EC3 + 0.0195 EC4

+ 0.0122 ECS + 0.1181; r 2 = 0.986 [Id]

EM4 = 0.4352 EC, - 0.0468 EC2 + 0.0287 EC3 + 0.0093 EC4

+ 0.0073 ECS + 0.0961; r 2 = 0.983 [le]

where the subscripts o, i , 2, 3, 4, s represent 0 and 0 to 30 cm,

30 and 30 to 60 cm, 60 and 60 to 90 cm, etc., respectively, for
the height above ground a t which the EM-38 unit was held in the
horizontal position and the soil-depth interval in which the ECa
was measured with a four-electrode s a l i n i t y probe.
When calibration coefficients are established for an area,
values of EC,, EC2, EC3, e t c . can be calculated from the
succession of EM readings by various numerical techniques, such
as LU decomposition (5). The ECa-depth distributions calculated
(predicted) in t h i s manner, have been shown to correspond well
with measured values for s o i l s with similar ECa distribution
patterns (11).
Due t o the complexity of solving for EC,, EC2, EC3, EC4 and
ECs in Equations [ l a - e ] , a more practical approach was
developed. The ECa values of each s o i l depth interval were
correlated with the succession of EM readings as
 870 CORWIN AND RHOADES

ECo-0.3 = ßoEMo + ß,EM, + B2EM2 + B3EM3 + ß4EM4;

ECo.3-0.6 = 7oEMo + 7,EM, + 7aEM2 + 73EM3 + 74EM4; e t c . ,

where EC0-0.3, EC0i3_0.6, etc., represent the ECa as measured

over the soil depth intervals of 0-0.3 m, 0.3-0.6 m, etc. Using
multiple linear regression the following relations were
determined in this manner for the Lakeview study area:
Downloaded by [University of Florida] at 09:06 05 June 2016

ECo_o.3 = -0.1285 EM0 + 0.1446 EM, + 5.3878 EM2

- 17.4476 EM3 + 15.0549 EM« - 0.1309; r 2 = 0.991 [2a]

ECo.3-0.6 = -1.3259 EM0 - 4.8938 EM, + 55.8250 EM2

- 94.0405 EM3 + 47.4196 EM4 - 0.9169; r s = 0.975 [2b]

ECo.6_o.9 = 9.1705 EM0 - 8.4116 EM, - 18.3090 EM2

+ 50.6298 EM3 - 42.5033 EM4 - 0.1224; r 2 = 0.999 [2c]

EC0.9-1.2 = 1.1090 EM0 + 0.2352 EM, - 23.3536 EM2

+ 221.0100 EM3 - 266.8789 EM4 + 3.5012; r 2 = 0.986 [2d]

The ECa-depth distributions calculated using this simple

technique corresponded very well with the measured values found
in this study area. This calculation procedure can be carried
out in the field with a programmable hand-held calculator. The
method i s practical, accurate, and permits the determination of
soil profile ECa distributions from a succession of EM readings
obtained at various heights above ground. However, different
geographical areas have been found to give different
coefficients. Apparently, appropriate calibrations must be
established for different geographical areas. The reason for
this variation in calibration coefficients from one geographical
area to the next i s probably the consequence of variation in the
general ECa distributions between the areas, since this approach
 SOIL ELECTRICAL CONDUCTIVITY-DEPTH RELATIONS 871

i s sensitive to the general ECa distribution pattern upon which

i t i s generated: normal, uniform or inverted.
Appropriate calibrations needed for different geographical
areas and ECa distribution patterns can be obtained relatively
quickly using the four-electrode salinity probe and the
techniques described above. Once the calibrations are made, the
EM-38 device can be used to determine ECa-depth distributions
more rapidly than with four-electrode techniques and with
sufficient accuracy to appraise salinity. I t s use saves the time
Downloaded by [University of Florida] at 09:06 05 June 2016

required to make the access hole necessary for the insertion of

the salinity probe, and i t can be used on dry, rocky and compact
soils where i t i s difficult to insert the salinity probe and to
make good electrode-soil contact.
The preceeding calibration method requires the solution of
a system of equations (one for each soil depth increment), the
acquisition of five EM readings taken at as many heights above
ground (0, 0.3, 0.6, 0.9 and 1.2 m) and ECa measurements taken
from within the four soil depth increments (0-0.3, 0.3-0.6,
0.6-0.9, 0.9-1.2 m). The subsequent use of these calibration
equations to predict ECa for discrete depth intervals requires
five EM measurements at the point of interest.
A simpler method for obtaining calibrations and making
ECa-depth predictions uses only two EM measurements taken on the
soil surface (1). In this approach the EM readings are related
to ECa by a series of simpler equations which are based, in part,
on the theoretical response functions for the EM-38 established
for homogeneous media.
Assuming that the EM-38 response curves hold for
measurements taken over non-homogeneous media, i t i s possible to
derive a series of equations which relate ECa within a soil-depth
interval to the horizontal (EM(j) and vertical (EMy) EM
measurements (which give depth-weighted ECa values to about 1 and
2 m depths, respectively). For the 0 to 0.3-m increment of soil,
 872 CORWIN AND RHOADES

the equations derived from the theoretical, homogeneous response

functions (Figure 3) are:

EMOtv = 0.15 EC0_o#3)V + 0.85 EC>0>3tV and [3]

EM0,H - 0.435 EC0_0.3,H + 0.565 EC>0.3,H 1*1

where EM0>v and EM0)y are the electromagnetic apparent

conductivities measured a t the soil surface in the vertical and
Downloaded by [University of Florida] at 09:06 05 June 2016

horizontal positions, respectively; and EC0_0,3>y, EC>0#3>y,

EC
o-o.3,H» a n ( lEC>o.3,H are tne actual bulk soil electrical
conductivities (ECa values) for the 0 to 0.3-m and > 0.3-m soil
depth intervals, respectively. The subscript a in ECa i s dropped
in these equations and some of the others that follow in this
section in order to minimize the clutter in the subscripts.
Since the volume of soil measured within the 0 to 0.3-m
increment i s very similar for the vertical and horizontal
orientations, i t i s reasonable to assume that EC0_0>3jy =
EC
o-o.3,H» However, in the case of the > 0.3-m increment, the
volumes of measurement are quite different; consequently, i t i s
unlikely that EC>0.3iy will equal EC>0,3^\\. This fact presents a
problem because in order to arrive at a relationship between
EC
o-o.3 and EMo.V and ^o.H u s i n s Equations [3] and [4], i t i s
necessary to equate EM>0,3)y and EC >OI3) H^ This problem was
overcome when i t was found that EMO .¡| could be adjusted using
empirical relationships so that EC>o>3y calculated from Equation
[3] would equal EC>o.3)n calculated from Equation [4] for s i t e s
where ECO_O#3> EM0>y, and EM0)n were measured.
Values of EC>0#3)y were calculated with Equation [3] using
ECo-o.3 as measured with the four-electrode probe, and using
EMoy as measured with the EM-38 device. I t was assumed that
EC
>o.3,V is a better estimate of actual EC>0.3 than EC>0<3in
because i t contributes 85% of the EM0)y reading, whereas EC>0.3,H
only contributes 56.5% of the EM0)n reading. An adjusted EM0>n
 SOIL ELECTRICAL CONDUCTIVITY-DEPTH RELATIONS 873

Adjusted Electromagnetic Conductivity Curves for Composite Increments

9
' 0-06m

o ~
E s

Adjusted - 0.9502 Measured + 0.1521 Adjusted« 1.0645Measured -0.0017

r2 • 0.995 ÍÍ - 0.977

" 0 1 2 3 4 5 6 7 "0 1 2 3 4 5 6 7 8 9 10
Downloaded by [University of Florida] at 09:06 05 June 2016

MEASURED ELECTROMAGNETIC CONDUCTIVITY MEASURED ELECTROMAGNETIC CONDUCTIVITY

EM 0H,(dS/m) EMOH,(dS/m)

Adjusted - 1.4355 Measured - 0.3298 Adjusted - 1.7476 Measured - 0.4S02

r2 . 0.988

0 1 2 3 4 5 6 7 8 9 1 0 0 I 2 3 4 5 6 7 8 9 IO II

MEASURED ELECTROMAGNETIC CONDUCTIVITY MEASURED ELECTROMAGNETIC CONDUCTIVITY

EM0H,(dS/m) EM0H,(dS/m)

Fig. 4. Relationship between electromagnetic soil conductivity as

measured with the EM-38 in the horizontal position at the
soil surface, EM0 R(measured) a n d adjusted electro-
magnetic soil conductivity, EM
V0 H(adjusted)
composite depths.

(relevant to the 0 to 0.3-m depth increment) was then calculated

from Equation [A] using the measured values of EC0_0#3 and the
values of EC >Oi3 y calculated from Equation 13]. A plot of
measured and adjusted EM0>n values for each depth increment of
a l l 16 sample s i t e s revealed a set of linear relations (see
Figure 4). Assuming that these relations can be used,
 874 CORWIN AND RHOADES

i r r e s p e c t i v e of t h e s i t e of measurement, measured values of EM0,H

f o r a s p e c i f i e d depth increment (0-h meters) can be a d j u s t e d s o
t h a t EC>(,ty = EC>( )J H» a s demonstrated i n Equations [5J and [ 6 ]
f o r t h e 0 t o 0.3-m increment:

EM0)v = 0.15 ECo-o.3 + 0.85 EC>o.3tv, and [5]

EMo,H(adjusted,0-0.3 m) = °- 4 3 5 EC0_0#3 + 0.565 EC>0.3>y. [6]

Downloaded by [University of Florida] at 09:06 05 June 2016

Equations [5] and [6] can now be reduced by substitution to a

single equation:

ECa,o-o.3 = 2.982 EMOjH(adjusted,0-0.3 m) - 1-982 EMo,y, [7]

where EM0,H(adjusted 0-0.3 m)i s a n empirically obtained linear

expression of the form as shown i n Figure 4.

EMo,H(adjusted 0-0.3 m> = k i EMo.K + k2» [8]

Equations [7] and [8] can be reduced by substitution to a

single equation. Following the same rationale, an analogous set
of equations can be obtained which provide predictions of ECa for
other s o i l depth intervals from EM^ and EMy. These equations are
of the form

, X1-X2 EMH - kV EMV + k, [9]

where k^, ky and k are empirically determined coefficients for

the depth interval xi-xa. The value k should ideally be zero,
but often i s not, due to experimental error or "noise" in the
data.
In the f i r s t application of t h i s technique, calibrations
were obtained i n three different areas of California: Lakeview,
San Joaquín Valley near Lost Hills and Imperial Valley. The
 SOIL ELECTRICAL CONDUCTIVITY-DEPTH RELATIONS 875

s i t e s were selected to provide a wide range of ECa-depth

relations on different s o i l types. Electromagnetic conductivity
measurements were taken with the EM-38 device positioned i n
horizontal and v e r t i c a l positions at the s o i l surface. Direct
measurements of ECa were then taken a t 0.3-m increments through
the s o i l p r o f i l e using the four-electrode s a l i n i t y probe.
Simple, untransformed linear regressions were performed on
the ECa values derived from the EM(j and EMy measurements, and the
values for corresponding e l e c t r i c a l conductivities were measured
Downloaded by [University of Florida] at 09:06 05 June 2016

with the four-electrode probe. The r e s u l t s obtained (see Figure

5) were quite s a t i s f a c t o r y for survey and diagnostic purposes
though somewhat i n f e r i o r to the multiple regression technique of
(11). Though l e s s accurate, the "established c o e f f i c i e n t "
approach requires only two EM readings as opposed t o five for the
"multiple regression c o e f f i c i e n t " method. Furthermore, the
apparent s i t e s p e c i f i c i t y observed i n the "multiple regression"
approach was not found for the "established c o e f f i c i e n t " approach
for the three s i t e s evaluated. Considering the reduction i n the
number of electromagnetic measurements required, the ease of
calculation, and the p o s s i b i l i t y of broader a p p l i c a b i l i t y , the
s l i g h t reduction i n accuracy of the established coefficient
approach seemed t o be an acceptable tradeoff. Hence, the authors
advocated the use of t h i s l a t t e r technique for p r a c t i c a l s a l i n i t y
appraisal a t that time.
In evaluating the use of the "established c o e f f i c i e n t "
approach on a wider variety of s i t e s , i t was found that the
predicted bulk s o i l conductivities for inverted conductivity
p r o f i l e s ( i . e . , p r o f i l e s where ECa decreases rapidly with
increased depth) consistently deviated from the corresponding
"ground t r u t h " conductivities as measured with the four-electrode
probe. This fact pointed out an obvious insufficiency i n the
newly developed method and an a l t e r n a t i v e approach for the
measurement of inverted conductivity p r o f i l e s using
electromagnetic induction was developed (2).
 876 CORWIN AND RHOADES

ELECTRICAL CONDUCTIVITY (dS/m)

O 5 10 15 20
Lakeview Area X multiple regression
Site 1 coefficient approach
0.15 2
3 O Established coefficient'
approach-successive
increments

• Measured with four

Downloaded by [University of Florida] at 09:06 05 June 2016

probe

045 A Established coefficient'

X
approach- successive
I- increments from
Q.
HI
Q

0.75 •

I.05

Fig. 5. Graph of representative ECa-depth profiles for three

Lakeview sample sites showing measured and three
different calculated profiles for each s i t e .

As discussed previously, when an electromagnetic induction

conductivity reading i s taken with the EM-38 meter placed on the
surface of a conductive medium, the resultant value of EC|
reflects the cumulative relative contributions of the
conductivities within that medium above some depth. Furthermore,
relative contributions of conductivity are dependent on the
orientation of the receiver/transmitter coils with respect to the
 SOIL ELECTRICAL CONDUCTIVITY-DEPTH RELATIONS 877
Downloaded by [University of Florida] at 09:06 05 June 2016

COMPOSITE DEPTH
INCREMENTS (meters)

Fig. 6. Ratio of v e r t i c a l and horizontal weighted responses a t

various composite depth increments ( i . e . , 0-0.15, 0-0.3,
0-0.45, 0-0.6 m, e t c . ) .

s o i l s u r f a c e . An a n a l y s i s of the r a t i o of response of the

v e r t i c a l and horizontal c o i l configurations f o r various composite
depths as determined from Figure 3 for a homogeneous medium
( i . e . , for 0-0.15 m, R H ( Z ) / R V ( Z ) = (1.0-0.818)/(1.0-0.978) =
8.3, f o r 0-0.3 m, R H ( Z ) / R V ( Z ) = (1.0-0.670)/(1.0-0.925) = 4.4,
e t c . ) i s given i n Figure 6. The f i g u r e reveals a p o t e n t i a l
reason for the discrepancy between the bulk s o i l e l e c t r i c a l
c o n d u c t i v i t i e s predicted from EM readings and the "ground t r u t h "
measured ECa's for inverted conductivity p r o f i l e s . As seen i n
Figure 6, the g r e a t e s t r e l a t i v e response discrepancy between
measurements taken with the c o i l s oriented v e r t i c a l l y and those
 878 CORWIN AND RHOADES

oriented horizontally to the soil surface i s for the top 0.3 m.

As a result, the influence of a substantial portion ( i . e . ,
approximately 0.15 m) of the top 0.3 m of soil upon the vertical
measurement i s quite negligible compared to i t s influence on the
horizontal measurement. If there i s a sizeable change in
conductivity within the top 0 to 0.3 m of s o i l , especially
between the 0 to 0.15 m and 0.15 m to 0.30 m increments, as may
occur in an inverted soil salinity profile, then the predicted
conductivities from the EM reading using the simplified, general
Downloaded by [University of Florida] at 09:06 05 June 2016

method of Corwin and Rhoades (1) will be in error. The error i s

due to the fact that the vertical coil configuration scarcely
detects the top 0.15 m of s o i l . With the vertical dipole
orientation, 0.022 fraction of the EM-38 conductivity reading i s
contributed by the top 0.1 m of soil and 0.978 i s contributed by
the soil below 0.1 m.
In order to deal adequately with inverted profiles,
therefore, some other method must be developed which compensates
for this inherent weakness of the electromagnetic inductive
measurement of inverted conductivity profiles. The equations
derived by Corwin and Rhoades (1) must be either modified to
reflect the lack of response of the vertical EM measurement to
the top 0.3 m and/or new adjustment curves based solely on
inverted profile data must be developed to shift the emphasis on
adjustment to the surface.
Two approaches were evaluated (2) for using the EM-38
instrument for measuring ECa-depth relations of inverted
profiles. The f i r s t approach (referred to as Modified-Approach
no. 1) compensates for the discrepancy in relative response
between horizontal and vertical measurements over the 0 to 0.3-m
depth by the establishment of EM0>n adjustment curves as
described in Corwin and Rhoades (1) but, in this case, using data
only from inverted conductivity profiles. The second approach
(referred to as Modified-Approach no. 2) eliminates the influence
of the top 0 to 0.3-m depth increment from the vertical
 SOIL ELECTRICAL CONDUCTIVITY-DEPTH RELATIONS 879

TABLE 1.

Modified-Approach No. 2 Equations Used t o Calculate E l e c t r i c a l

Conductivities f o r Soil Increments From Electromagnetic
Conductivity Measurements.

Depth, m Equations f o r e l e c t r i c a l conductivity

Composite Depths
Downloaded by [University of Florida] at 09:06 05 June 2016

0-0.3 ECa,0-o.3 = 2.185 EM0>n _ 1.299 EM0,V + 0.350

0-0.6 ECa,o-o.6 = 1.828 EM0>H _ 0 .717 EMo,V " ° - 0 0 3

0-0.9 ECa,o_o.9 - 2.399 EM0,H _ 0 .671 EM0,V " °-551

0-1.2 ECa(0-,.2 = 2.943 EM0,H - 0.684 EMO,V - 0.809

Successive Depths

0-0.3 ECat0_0.3 = 2.185 EM0)H - 1.299 EMO,V + 0.350

0.3-0.6 EC a)0#3 _ 0#6 = 1.537 EMOjH - 0.135 EM0>v - 0.345

0.6-0.9 ECa>o.6-o.9 = 3.541 EM0,H - 0.579 EMOtV - 1.647

0.9-1.2 ECai0.9_,.2 = 4.576 EMOfH - 0.723 EM0,V -
1.581

measurement by the removal of the weighted response of that layer

and i t s addition t o the adjacent 0.3 t o 0.6-m layer i n the
equations used by Corwin and Rhoades (1) t o r e l a t e the surface
v e r t i c a l electromagnetic measurement, EMOïy, t o bulk e l e c t r i c a l
conductivity, ECa. This resulted i n the s e t of composite and
successive increment bulk e l e c t r i c a l conductivity equations shown
in Table 1, which are analogous t o and obtained following the
procedure previously outlined by Corwin and Rhoades (1) f o r
non-inverted s o i l ECa p r o f i l e s .
 880 CORWIN AND RHOADES

These two approaches were evaluated using data obtained at

eleven sites scattered throughout California. The sites were
selected in order to provide a variety of soil types and
conditions. Since the horizontal EM conductivity reading for an
inverted conductivity profile i s characteristically greater than
the vertical reading, the identification and location of an
inverted conductivity profile i s simplified. Measurements were
made with a Geonics Limited EM-38 meter at the soil surface of
each s i t e , with the coils in both the vertical and horizontal
Downloaded by [University of Florida] at 09:06 05 June 2016

configurations. Corresponding ground truth electrical

conductivity readings were taken at 15-cm increments through the
soil profile to a depth of 0.90 m with a four-electrode salinity
probe at the exact same location. Since the measurement of the
top 0.3 m was extremely critical for this particular study, from
6 to 10 four-electrode salinity probe ECa measurements were taken
and averaged to obtain an accurate value for this depth
increment. Up to three four-electrode salinity probe readings
were taken at each remaining depth increment and an average was
then determined from the accumulated data for each 0.3-m
composite increment down to 0.9 m: 0 to 0.3 m, 0 to 0.6 m, and 0
to 0.9 m.
Linear regression techniques were used as a means of
evaluating the correspondence of the predicted and measured bulk
soil electrical conductivities. Results of this evaluation (see
Table 2; and (2) for more detail) showed that both modified
approaches provided a good one-to-one correspondence between
measured and calculated ECa's. Approach no. 1, however, yielded
a slightly better correspondence with less scatter. The
unmodified approach, showed sharp discrepancies between measured
and calculated ECa's with increased increment depth in the soil.
This deviation became even more pronounced as the magnitude of
the conductivities increased. A comparison of the significant
linear regression s t a t i s t i c s (see Table 2) for the three methods
confirms that the unmodified approach was not as consistently
 SOIL ELECTRICAL CONDUCTIVITY-DEPTH RELATIONS 881

TABLE 2.

Linear Regression Analysis S t a t i s t i c s f o r Various Methods Used t o

Calculate Bulk E l e c t r i c a l Conductivities of Inverted Conductivity
P r o f i l e s From EM Induction Conductivity Measurements ( 2 ) .

Std. e r r o r
Method Slope Y-intercept R2 of e s t .

Unmodified-Corwin &
Downloaded by [University of Florida] at 09:06 05 June 2016

Rhoades (1982) 1 .51 -0,.37 0 .78 2,.65

Modified-approach no. 1 0,.98 0,.06 0 .97 0,.46

Modified-approach no. 2 0 .96 0,.12 0 .92

0.73

accurate a s the two modified approaches i n the c a l c u l a t i o n of ECa

f o r d i s c r e t e s o i l depth i n t e r v a l s from EM readings.
The data obtained i n t h i s study provide some i n s i g h t as t o
the reason f o r the i n s u f f i c i e n c y of the unmodified approach.
Since the adjustment curves c o r r e c t the EM0 ) \\ readings f o r
i n e q u a l i t i e s i n volume of measurement between v e r t i c a l and
horizontal c o i l configurations ( 1 ) , the adjustments f o r inverted
p r o f i l e s would be expected t o be the opposite of those f o r
e l e c t r i c a l conductivity p r o f i l e s which are found t o increase with
depth. The difference i n the two e l e c t r i c a l conductivity
p r o f i l e s manifests i t s e l f i n d i f f e r e n t slopes of the adjustment
curves; consequently, the slope should decrease sharply with
depth up t o some point. The r e s u l t s of the modified approaches
show t h i s t o be the case.
Based on the above, the need f o r a means of handling
inverted conductivity p r o f i l e s i s evident. Separate c a l i b r a t i o n s
(values of c o e f f i c i e n t s i n Equation 9) should be established
using only data f o r inverted conductivity p r o f i l e s .
 882 CORWIN AND RHOADES

Since i t i s possible to determine whether or not an

inverted conductivity profile has been encountered by the fact
that the EMOJH reading i s greater than the EMoy reading, i t i s
an easy matter to program a hand-held programmable calculator of
sufficient memory capacity to convert horizontal and vertical EM
readings to an electrical conductivity profile.
The accuracy of the ECa profile determined using these
calibration approaches depends on how closely the actual
conductivity profile conforms to the general shape of an
Downloaded by [University of Florida] at 09:06 05 June 2016

increasing or inverted electrical conductivity profile. For

example, ECa-depth relations that fluctuate abruptly are not as
closely predicted as profiles that show a steady increase (normal
salinity profile) or decrease (inverted salinity profile) i n
electrical conductivity. This places s t r i c t limitations on the
accuracy and application of this measurement technique, since
rapid fluctuations of conductivity with depth are smoothed out.
More s t a t i s t i c a l l y rigorous EM - ECa depth calibrations
have recently been developed (Rhoades et a l , 1989c). EMH and EMy
were measured using the Geonics EM-38 device at 900 sites within
a thirty-nine square-kilometer area of the South Kings River
Watershed of the San Joaquín Valley of California (representing
at least 10 different soil types varying in texture from loamy
sand to clay). At each s i t e , corresponding values of ECa were
measured using the four-electrode salinity probe for the
intervals of 0-30 and 30-60 cm; at every third s i t e ECa was also
measured for the interval 60-90 cm. ECa was also measured at
each s i t e using an array of electrodes inserted into the soil
surface in the Wenner configuration (9). Inner-electrode
spacings of both 30 and 60 cm were used in these measurements.
ECa values for the 0-30, 0-60 and 30-60 cm intervals were
estimated from the Wenner array measurements as described
elsewhere (9; 10).
Exploratory data analysis techniques (18) were used to
determine the frequency distributions of the measured EM and ECa
 SOIL ELECTRICAL CONDUCTIVITY-DEPTH RELATIONS 883

data. The data were transformed i n order t o obtain normal

d i s t r i b u t i o n s so that v a l i d s t a t i s t i c a l analyses could be
performed. This was necessary because the d i s t r i b u t i o n s of each
parameter were highly skewed with a preponderance of low values
and a s u b s t a n t i a l l y smaller number of high values. Three
transformations were t r i e d : natural log, square root and
fourth-root. As revealed by the Shapiro-Wilk t e s t s t a t i s t i c
(17), the fourth-root transformation normalized a l l of the
v a r i a b l e s ; consequently, a l l subsequent s t a t i s t i c s were applied
Downloaded by [University of Florida] at 09:06 05 June 2016

to fourth-root transformed data. Hereafter, the symbol ~ w i l l be

used t o denote parameters which have been raised t o the
one-fourth power.
The data was s p l i t according t o ECa-depth p r o f i l e type
(inverted o r normal) as indicated by the r a t i o of EM^/EMy.
Inverted p r o f i l e s were those where EMn/EMy > 1.05; normal
p r o f i l e s were those where EM^/EMy < 1.05.
Regression a n a l y s i s was used t o solve Equation [9] f o r k\\
and ky, f o r the data obtained using the four-electrode s a l i n i t y
probe, (16). Six regressions were obtained using the whole data
set. Each corresponded t o the appropriate s o i l depth and p r o f i l e
condition. These r e s u l t s are given i n Table 3. The r e l a t i v e l y
2
high r values show the good correspondence obtained between ECa
and EM|j and EMy. The Wenner data were also "processed" i n the
same way. Results are given i n Table 4.
Because of suspected c o l l i n e a r i t y between EMH and EMy,
evaluation of c o e f f i c i e n t s t a b i l i t y was made by randomly
s p l i t t i n g the data i n t o two s e t s and comparing the general
a p p l i c a b i l i t y of the c o e f f i c i e n t s of the equations i n Table 3.
In order to test for significant differences in the coefficient
estimates between the different s p l i t s , a categorical variable
(DS) was introduced into these equations to identify the subset.
The analogous relation to Equation [9] then became

ECa,xi-X2 = Bo + B, SASEM,_xa + B2 (DS). [10]

 884 CORWIN AND RHOADES

TABLE 3,
Relations Found Between Soil Electrical Conductivity in the
Different Soil Depth Increments and the Electromagnetic
Measurements Made With the EM-38 Device, where ECa was Measured
with the Martek SCT Insertion Probe.
Downloaded by [University of Florida] at 09:06 05 June 2016

depth, cm Equations for Electrical Conductivity^/ n

for EMH

0-30 EC¡ = 3.023 EMH " I - 9 8 2 ^V 673

°'731
0-60 EC¡ = 2.757 EMH " !-539 ^V ' ° ' 0 9 7 639
°'835
0-90 EC¡ = 2.028 EMH ~ ° - 8 8 7 ^V 198
°-852
30-60 EC¡ = 2.585 EMH ~ i - 2 1 3 ^V " ° - 2 0 4 647
°'782

60-90 EC¡ = .958 EM^ + 0.323 EMy - 0.142 195 0.736

for EMH ^ EMu

0-30 ECa = 1.690 EMH " 0.591 EMy 117 0.866

0-60 EC¡ = 1.209 EMH ~ ° - 0 8 9 147
°'917
0-90 EC¡ = 1.107 EMH 54 0.903
30-60 ECa = .554 EMH+ -595 ^V) 113
°-840
60-90 ECa" = -0.126 EM^ + 1.283 EM^ - 0.097 53 0.812

if ECa, EMH and EMy are t h e fourth roots of ECa, EMHa n d

 SOIL ELECTRICAL CONDUCTIVITY-DEPTH RELATIONS 885

TABLE 4.
The Relations Between Soil Electrical Conductivity in the (ECa)
as Measured by the Wenner Array and Electromagnetic (EM)
Measurements Made with the EM-38 Device.

depth, cm Equations for Electrical Conductivity^/

for EMH S -EMy

0-30 ECa = 2.5395 EMH " 1.4130 EMy - 0.0683 759 0.810
Downloaded by [University of Florida] at 09:06 05 June 2016

0-60 ECa = 2.0918 EM(¡ - 0.8109 EMy - 0.1791 761 0.895

30-60 EC¡ = 1.8941 EM^j - 0.4075 EMy - 0.2925 758 0.840

for EMH ^ EMy

0-30 = 1.1643 EMH ~ 0.0785 EMV 165 0.922
0-60 = 0.6403 EMH + 0.5684 EMy - 0.1141 163 0.969
30-60 = 1.3671 EMy - 0.2088 162 0.919
37 ECa, EMH and ^V are the
fourth roots of ECa, EMH and

If a difference existed between the two data s p l i t s , the B2

parameter would be significantly different from zero. Since the
values of Ba were not significantly different from zero in any
case, i t was concluded that coefficient "stability" was
achieved. Subsequently, the remainder of the analyses were
performed using the total data.
The relations of Tables 3 and 4 were then used to predict
ECa values from given EMH an<* ^V values. The correspondence
between predicted values of ECa and measured values was good.
Linear regression analysis of measured and predicted ECa yielded
r 2 values near 1, slopes near 1 and intercepts close to 1 (see
Table 5). This close correspondence demonstrates the general
 TABLE 5. 00
CO

Results of Linear Regression Between Predicted and Measured Values o

of ECa by Ssoil Depth I n t e r v a l , P r o f i l e Condition and Method of
Downloaded by [University of Florida] at 09:06 05 June 2016

Measurement of ECa (SCT vs Kenner).

SCT Method Wenner Method
Depth, cm n S loiIS. Intercept ra ti Sloi je Intercept r2

for EMH i EMy

0-30 698 1 .088 +0.011 0.702 784 1.065 -0.007 0.812

0-60 663 .965 +0.102** 0.854 786 1.042 0.016 0.875

0-90 206 1 .031 -0.013 0.854

30 - 60 671 0 .922 +0.220** 0.801 782 1.040 0.022 0.807

60 - 90 216 .953 0.168 0.751

for EMH * EMv

0-30 88 1.149 -0.218 0.857 101 1 .179 -0. 138** 0. 924

0-60 85 1.019 +0.060 0.899 99 1 .101 -0. 048** 0. 968

n
o
0-90 30 1.057 +0.066 0.830 -.__ -—
M
S3
30 - 60 85 1.099 -0.048 0.851 99 1 .099 -0. 049 0. 926
>
60 - 90 32 1.275 -0.112 0.816 -— -— o
JO

g
O
M
NOTE: ** indicates intercept i s significant at the 0.01 alpha level.
 SOIL ELECTRICAL CONDUCTIVITY-DEPTH RELATIONS 887

TABLE 6.
Comparison of Measured 1ECa Values With Those Predicted by New
and Previous Relations '

Relation'/ Profile Depth, cm n Slope Int ra

new EMU * EMy 0-30 88 1.15 -0.22 0.86

tt
previous ft II
1.41 -0.45 0.86
II
new 30-60 85 1.10 -0.05 0.85
Downloaded by [University of Florida] at 09:06 05 June 2016

II
previous tl tt
2.77 -1.32 0.67
It
new 60-90 32 1.28 -0.11 0.82
II
previous II II
1.51 -0.10 0.71
II
new 0-60 85 1.02 0.06 0.90

previous II tt tl
1.76 -0.54 0.84
new tl
0-90 30 1.06 0.07 0.83
H
previous II It
1.80 -0.47 0.79
new EMH S EMy 0-30 698 1.09 0.01 0.70

previous t) II ft
1.29 -0.13 0.61
if
new 30-60 671 0.92 0.22 0.80
ft
previous tt ft
1.39 0.45 0.80
H
new 60-90 216 0.95 0.17 0.75
M tl II
previous 0.57 1.24 0.72
11
new 0-60 663 0.96 0.10 0.85
It II It
previous 1.44 0.07 0.85
II
new 0-90 206 1.03 -0.01 0.85
tl n II
previous 0.89 0.62 0.86
EMH i. EMv & 0-30, 1790 0.99 0.11 0.80
EMH S EMy 30-60 &
60-90

previous '¿ '¿ '¿ 0.98 0.56 0.61

}J Measured ECa = (slope) predicted ECa i intercept; n = number of samples;
r 2 = coefficient of determination.
^J ECa values measured with four-electrode insertion probe (10) and values
predicted using relations developed herein (Table 3) and those of Corwin
and Rhoades (1,2).
 00
CD
00
Downloaded by [University of Florida] at 09:06 05 June 2016

TABLE 7.
Measured Values of EMfl, EMy and ECa and Predicted Values of ECa f o r
a Small, Well-Sampled Data Set

Measured, dS/m
EM EC* Predicted ECa, dS/m
S i t e No. H 0-30»/ 30-60 60-90 0-30" 30-60 60-90

1 4.30Í.08) 7.1 (.3) 4.5 (.3) 4.5 (.5) 7.0 6.7 4.7
2 .73(.01) •83(.00) • 72(.03) .49Í.07) 1.2K.05) 0.66 1.1 1.22
3 .80(.00) .78Í.00) .94(.O4) .94(.02) 1.13(.02) 0.98 1.5 1.30
4 2.40(.02) 2.18(.O1) 3.5 (.3) 2.9 (.2) 2.2 (.1) 3.7 4.0 2.9
5 .36(.00) •42(.00) .39(.O3) .26(,02) .5O(.O5) 0.31 0.46 0.55 S
6 1.2K.01) 1.04(.01) 1.29(.O6) .68(.05) .78(.06) 1.9 1.9 1.29
7 .42(.00) .59(.00) 0.23 0.44 4/
pa
3.4 (.1) 1.3 (.1) .44(.O4) re
8 1.77Í.O4) 1.48Í.03) 2.8 2.8 1.9 o
o
9 .4K.00) .16(.00) 1.25(.O8) *l •/ 0.92 4/ 4/ M
to
 SOIL ELECTRICAL CONDUCTIVITY-DEPTH RELATIONS 889

CM -3- in
ri n en oo m -3-
cn 00 CO
—' CM f—* —' m o — r-H CN

cn m -3-
r^ m
en
cn CO O 00
o
CM CO m o CM CM m

CM -3-
m o en CM S o
Downloaded by [University of Florida] at 09:06 05 June 2016

CM
*—« in O o m O CM r—t -3-
(90
(80

(C0
04)

07)

CO
O
o
.24<
.31

.91
.69

.67
.09

.00

in
>
ai

m 00
o O
o
CM O ~* o *-• I
4->
CM
.37
.65

CU
.02

.82

in i—t h~ CN in
_H -3- r-c —
i
•H
O
-3- in ri 00 co •H
m
o O O ri O o CN
o U •H
CM vD m in vD X!
ri Í—t
m en m Cl
-< in m CN ri O

-3- o i—t r^. o «—1 -3-

O O o O O O O O O 0>

o
»—t ri f- en CM in t—1 -x> ¡M
«—i 00 h- CN in 00
e CU
»—i CN «—i «—1 H
O
to

u
in t—« 00 CM CN in 2
u
<v
O O O O O O O
o o •H
I
'S
en co en o ri O CO oo
I—1
ri m m 4-1 -0 10 O
o o CU *J C 4->
-H ^H CM
<U M -H
T3 CO
u vO -H
X!
o e

- 1 " I
 890 CORWIN AND RHOADES

applicability of the equations and coefficients given in Tables 3

and 4 for the prediction of ECa from EM^ and EMy, at least for
this part of the San Joaquín Valley of California.
The r 2 values closer to 1 obtained with the Wenner method
of measuring ECa compared to the Martek SCT method are most
likely due to differences in the two instrument's sampling
volumes. The sample volumes for the Wenner and EM methods are
similar, while the volume sampled by the SCT salinity probe i s
much smaller.
Downloaded by [University of Florida] at 09:06 05 June 2016

To compare the new relations with those previously obtained

by Corwin and Rhoades (1, 2), linear regression analyses were
performed between the measured and predicted values of ECa in the
0-30 and 30-60 cm depths. The new relations consistently give
slopes and intercepts closer to 1.0 and 0.0, respectively, and
frequently r 2 values closer to 1.0 (see Table 6).
A more intensively sampled data set (consisting of eighteen
s i t e s ) was used to further evaluate the accuracy of the ECa
values predicted from EMj) and EMy using the values of k(j and ky
established with the extensive data s e t . These new data were
acquired from within the same study area as previously sampled,
but this time multiple measurements of EM and ECa were made at
each of the sampled s i t e s in order to obtain more representative
readings of EM and ECa within the soil volume measured by the
EM-38 device. Eight readings of EM^ and EMy were made at each
s i t e ; the EM-38 device was read in both positions (H and V) when
"pointed" at each of the eight cardinal compass directions. ECa
was measured at eight locations within the 0 to 30, 30 to 60 and
60 to 90 cm soil-depth intervals; these locations were 25 cm out
from the site-center along each compass heading. The data and
results are given in Tables 7 and 8. Values of r 2 (relating
predicted and measured values of ECa) closer to 1 were obtained
in this more accurate data s e t , especially in the 0 to 30 cm
soil-depth interval, compared to the major data s e t . The better
relationship i s attributed to the multiple Martek SCT probe
 SOIL ELECTRICAL CONDUCTIVITY-DEPTH RELATIONS 891

TABLE 8.

Results of Linear Regression Between Predicted and Measured

Values of ECa f o r a Small, Well-Sampled Data S e t j /

Soil Depth, cm n r2 Slope Intercept

Downloaded by [University of Florida] at 09:06 05 June 2016

for EM|1 < EM V

0-30 9 0.92 0.9K. 10 W 0.0K.09)

30-60 8 0.82 0.96(. 17) -O.2K.24)

60-90 8 0.79 0.82(. 17) 0.1K.25)

for EMH *• E
0-30 9 0.92 1.0K.11) -0.07Í.47)

30-60 8 0.74 0.93(.22) -0.70Í.93)

60-90 8 0.84 1.03(.18) -O.7K.51)

for both p r o f i l e types

0-30 18 0.96 1.0K.05) -0.06(.15)

30-60 17 0.84 0.8K.09) -O.13Í.27)

60-90 16 0.82 0.84(.ll) -0.06(.24)

¿/ Measured ECa = (slope) predicted ECa ± intercept; n = number of

samples; r 2 = coefficients of determination.
2
I Values within ( ) are standard errors.
 892 CORWIN AND RHOADES

measurements t h a t were made i n t h e a c q u i s i t i o n of these d a t a ,

which gives a c l o s e r approximation of t h e mean ECa value of t h e
r e l a t i v e l y l a r g e s o i l volume measured by t h e Geonics EM-38.
P r e d i c t e d and measured ECa values o f t e n d i f f e r e d
considerably from t h e EM^ and EMy values p e r s e (see Table 7 ) .
The c l o s e r correspondence e x i s t i n g between measured and p r e d i c t e d
ECa values than between measured ECa and E c | ( i . e . , EMH and EMy
values) c l e a r l y demonstrates t h e advantage of using t h e
p r e d i c t i o n s of ECa from EM^ and EMy i n l i e u of t h e values of EM^j
Downloaded by [University of Florida] at 09:06 05 June 2016

and EMy p e r s e f o r t h e purpose of s a l i n i t y a p p r a i s a l . Even

though t h e p r e d i c t i o n s a r e not a s a c c u r a t e a s d e s i r e d , they a r e
s t i l l reasonable e s t i m a t e s t h a t provide more meaningful
information with which to interpret soil salinity within the
plant root zone than the EM values themselves.

Limitations:

IV. Each calibration approach described for the EM-38

presupposes that the ECa profiles used in the calibration for a
given geographic location are of the same general shape. In
otherwords, the salinity profiles f i t one of three general
distributions (see Figure 7): normal, inverted or uniform.
Therefore, the determination of ECa-depth distributions for
profiles which are variable in shape over an area must be
calibrated for each given shape. This i s particularly
significant in areas where there are substantial textural
discontinuities. Because ECa i s a function of texture (12),
profiles which are horizontally layered with sands will be
particularly difficult to characterize with the EM-38 since the
ECa of the sand layers will tend to be low, and therefore,
"ignored" by the EM measurement. As an example, Figure 8a
represents the ground truth salinity of a site stratified with a
sand layer. Layer 2 represents the sand layer and i t s '
associated ECa. The ECa-depth relation which will be determined
 SOIL ELECTRICAL CONDUCTIVITY-DEPTH RELATIONS 893

01
i-, •H s
;*.
4-1 •rl
r—(

o o
<w T3 a •rl <M
Ol
a OVi
Downloaded by [University of Florida] at 09:06 05 June 2016

tfl •H
s
fi •H *—t

u O * n) a
•rl
4->
g
0)
(fl >>
4-<
9 4-> •rl
ja O) • • •a
ai C
•rl •73 4-> •rl
m
t—• *-H
u B u
ai (Ö
(fl ai •#> > (fl
•H ai es c
•o X! •H
g
>, ai O
4-1
•H
> ^r XI
V • <M
•rl
Ci
•H
.c CO e
••H 9
Cd w
4-1
(fl o •
ai •rl
o
73 rH
•rl
•rl
a x: 9
4-1

O •rl 4-> X> • o

(0 <M •rl
tM U cni
Vi
ta 4-1
ai
O o •rl (fl B
o (fl •H O
(fl S •o •rl
01 4->
s (fl 01 9
•H
o
i* •rl 01 PH X)
O 4-> •H •rl
•rl
<W Vi *
ta «J •Ö
ai
4-> Xt
9 O 4-1 B
4-J S-, (fl
•rl (fl O.-H •H
o
m i-I
• o 4->
o ni (fl >>
*-< c 4-> ai XI
« O •H •rt
•rl S •H U
*rl
o ai t—1 o (fl
ai Sí
•rl s-, n) •rl
O 3 a (fl Q.t3
 894 CORWIN AND RHOADES

Electrical
Salinity conductivity

< i

Depth
"o,
Q
/
4
Downloaded by [University of Florida] at 09:06 05 June 2016

Fig. 8. Comparison showing how an ECa-depth relation calculated

from an EM-38 measurement represents a salinity
distribution profile which i s highly variable due to the
presence of a low conductivity sand layer (layer 2):
(a.) schematic representation of the ground truth
salinity distribution and (b.) schematic represen-
tation of how the ECa-depth relation compresses the
conductivity distribution by "ignoring" layer 2.

by the EM-38 i s represented by Figure 8b where the EC profile i s

compressed where the sand layer occurs. Thus, a limitation of
the EM-38 calibration technique i s that i s requires prior
knowledge of the general ECa distribution ( i . e . , normal,
inverted, uniform) and/or the textural layering.
Theoretically, a calibration i s necessary for every profile
shape, but since there are an infinite number of potential shapes
this i s completely impractical. Though this limitation would
appear to render the EM-38 calibration techniques as
inapplicable, there are assumptions which can be made which
mitigate the severity of this problem. Where the soil i s uniform
in texture there are seven general salinity profile shapes which
likely occur in irrigated s o i l s . These salinity distributions
are presented in Figure 9A-G. By determining calibration
coefficients for each profile shape i t i s then possible to
 SOIL ELECTRICAL CONDUCTIVITY-DEPTH RELATIONS 895

determine the ECa d i s t r i b u t i o n from EM measurements provided that

there i s p r i o r knowledge of the general p a t t e r n of the s a l i n i t y
p r o f i l e so that the appropriate c a l i b r a t i o n c o e f f i c i e n t s can be
applied.
Of the seven s a l i n i t y d i s t r i b u t i o n s (see Figure 9 ) , one
p r o f i l e ( P r o f i l e D) i s unlikely t o be encountered and a second
p r o f i l e ( P r o f i l e C) i s a specialized type of inverted p r o f i l e .
P r o f i l e A occurs under normal arid-zone i r r i g a t i o n management
s i t u a t i o n s and w i l l vary i n i t s magnitude depending on leaching
Downloaded by [University of Florida] at 09:06 05 June 2016

f r a c t i o n , evapotranspiration and i r r i g a t i o n water q u a l i t y . The

net movement of s a l t i n P r o f i l e A i s downward and the processes
of leaching and evapotranspiration r e s u l t i n an increase i n s a l t
concentration down through the s o i l p r o f i l e . P r o f i l e B, an
inverted conductivity p r o f i l e , w i l l occur i n areas where the net
movement of s a l t s i s upward due t o evaporation and/or
t r a n s p i r a t i o n processes. This net upward movement of s a l t s
r e s u l t s i n the accumulation of s a l t a t the s o i l surface. Profile
C i s an inverted p r o f i l e where the groundwater t a b l e i s very
shallow. At the c a p i l l a r y fringe the conductivity w i l l increase
s l i g h t l y with depth due t o increases i n water content and thereby
g r e a t e r t o t a l dissolved s a l t s . P r o f i l e D resembles P r o f i l e A i n
the upper portion of the p r o f i l e , but decreases with depth i n the
lower portion of the p r o f i l e . This s a l i n i t y p r o f i l e can occur i n
s i t u a t i o n s where there i s a shallow water t a b l e consisting of a
higher q u a l i t y water than the s o i l water e x i t i n g the root zone.
Instances where t h i s p r o f i l e would occur include areas where
l a t e r a l fluxes of good q u a l i t y water come i n t o the system from
upslope locations. Normally t h i s p r o f i l e would not occur with
great frequency. P r o f i l e E i s a uniform conductivity p r o f i l e and
i s t y p i c a l of a highly leached s o i l with a deep groundwater
table. P r o f i l e s F and G are specialized forms of P r o f i l e E.
P r o f i l e F i s a uniformly leached p r o f i l e with a shallow water
table. The water e x i t i n g the root zone i s lower i n s a l i n i t y than
t h a t i n the water t a b l e because of l a t e r a l contributions from
 896 CORWIN AND RHOADES

upslope s o u r c e s . As was t h e case f o r P r o f i l e D, P r o f i l e F i s not

l i k e l y t o occur with g r e a t frequency. P r o f i l e G i s a leached
s a l i n i t y profile where a pulse of saline soil solution i s moving
through the s o i l profile.

Current Research:

V. Because s a l i n i t y , and thereby soil electrical conductivity,

i s a highly spatially variable property of the s o i l , i t can be
Downloaded by [University of Florida] at 09:06 05 June 2016

expected that calibrations of the EM-38 require more than just

single-point measurements of ECa upon which to serve as ground
truth measures of conductivity. Due to the spatial variability
problem, the volume of measurement of ground truth must be nearly
the same as the measurement volume of the EM-38 for accurate
calibrations to be made. In previous work by Rhoades and Corwin
(11), Corwin and Rhoades (1), and Corwin and Rhoades (2), only
single-point measurements of ECa were made with the
four-electrode probe at successive depths to determine the ground
truth conductivity distribution. As a result, there i s some
question as to the accuracy of these measurements in actually
describing what i s being measured by the EM-38 since the volume
of measurement of the EM-38 i s substantially larger than the
accumulated volumes measured by the four-electrode probe. As
shown by Rhoades et a l . (15), the correspondence between ECa
values predicted from EM and measured ECa values i s improved when
increased numbers of measurements of ECa are made with the
four-electrode salinity probe within the larger volume sensed by
the EM-38 unit or when Wenner measurements of ECa (which more
closely corresponds to the EM-38 volume) are used for ground
truth.
In addition, the depth of penetration of the EM-38 i s
substantially deeper than was sampled in any of the calibrations
made in previous work. This fact has been minimized in the past
by assuming that below about 150 cm the conductivity was the same
 SOIL ELECTRICAL CONDUCTIVITY-DEPTH RELATIONS 897

as t h a t measured a t 150 cm. In s i t u a t i o n s where the conductivity

below 150 cm i s low (e.g.. P r o f i l e B), t h i s assumption w i l l not
drastically effect calibration coefficient determination since so
l i t t l e of the total response of the EM-38 i s below 150 cm.
However, if the conductivity i s high, as can occur in Profile A,
then this assumption does not necessarily hold true. For this
reason, ground truth measurements of ECa below 150 cm should not
be overlooked especially for profiles with high conductivity at
the deeper depths (below 150 cm).
Downloaded by [University of Florida] at 09:06 05 June 2016

Obtaining a more accurate ground truth measurement of the

sphere measured by the EM-38 i s essential for a reliable
calibration. Future calibrations of the EM-38 are forthcoming
which will involve an intensive spatial sampling consisting of
nine samples on a 30-cm grid space repeated at 25 cm depth
intervals down to 3 meters. In the top 25-cm, measurements are
taken over the same grid, but at 5 cm depth intervals. Each
profile shape presented in Figure 9 will be calibrated using
salinity profiles of the same general shape, but differing in
magnitude to obtain a low, medium and high conductivity profile
of that particular shape. With this database of ground truth
conductivity distributions generalized calibrations will be
prepared for both procedures of measurement (i.e., [1] taking EM
measurements in the horizontal and vertical positions at
successive heights above ground and [2] taking EM measurements in
the horizontal and vertical positions solely at the soil
surface). Calibration coefficients for each profile shape will
be developed from 5-10 different ground truth conductivity
profiles each varying slightly in i t s magnitude.
The relative response function of the EM-38 device as
affected by conductivity distribution with depth and horizontally
by depth will also be determined and used in the developments of
the calibration coefficients. Because the output of the EM-38
instrument i s not fully linearly related to soil electrical
conductivity as the latter increases to high values, future
 898 CORWIN AND RHOADES

Salinity Salinity Salinity Salinity

Downloaded by [University of Florida] at 09:06 05 June 2016

Fig. 9. Types of s a l i n i t y distributions potentially encountered

in irrigated, arid-zone soils which are texturally
homogeneous: (A.) normal s a l i n i t y distribution, (B.)
inverted s a l i n i t y distribution, (C.) inverted s a l i n i t y
distribution with a shallow water table, (D.) normal
salinity distribution with a shallow water table of good
quality water from an upslope l a t e r a l flow source, (E.)
uniform s a l i n i t y distribution, (F.) uniformly leached
salinity profile with a shallow water table of poor
quality water and (G.) leached salinity profile where a
saline pulse of soil water i s moving through the soil
profile.

calibrations will also include adjustment for t h i s phenomenon

using the data given in Table 9. EMH readings exhibit linearity
to greater values of ECa than do EMy readings. While these
corrections for non-linearity only apply s t r i c t l y to homogeneous
profile conditions, i t may be possible to further adjust them for
non-uniform conditions if the response depth i s known.
 SOIL ELECTRICAL CONDUCTIVITY-DEPTH RELATIONS 899

TABLE 9.

Relation Between Actual and Indicated Soil E l e c t r i2 c a l Conductivity

for EM-38 Assuming Homogeneous Conditions, mS/m'» '
Actual Indicated ECS
ECg EMH EMJL

5 4.955 4.910
10 9.872 9.744
15 14.765 14.529
Downloaded by [University of Florida] at 09:06 05 June 2016

20 19.638 19.276
30 29.335 28.670
40 38.976 37.953
60 58.119 56.240
80 77.105 74.213
100 95.954 91.915
150 142.570 135.158
200 188.566 177.167
300 279.010 258.117
600 540.766 482.062
800 708.938 618.947
1000 872.923 747.691
37 P e r s o n a l communication from J . D. M c N e i l l , G e o n i c s L i m i t e d .
_f/ For conversion to EC units of dS/m divide the values in the
table by 100.

Summary:

VI. Even though the use of electromagnetic induction to

determine soil electrical conductivity at discrete depth
intervals through the soil profile requires an i n i t i a l investment
of time and labor to obtain the necessary calibration
coefficients, there i s a substantial return in the ease and
accuracy of the interpretations of salinity over a broad
landscape. I t i s particularly accurate in the prediction of ECa
(and hence salinity) within the top 0-30 cm and 0-60 cm of soil
which are the most relevant depths to the production of field
 900 CORWIN AND RHOADES

crops. The v a l u e s of ECa w i t h i n t h e s e depths p r e d i c t e d from EM

a r e reasonable e s t i m a t e s t h a t provide more meaningful information
with which t o i n t e r p r e t s o i l s a l i n i t y within the plant root zone
than the EM values themselves. The i n t e r p r e t a t i o n of s a l i n i t y
from ECa i s an equally important aspect of the s a l i n i t y
appraisal. This topic i s reviewed i n the accompanying paper by
Rhoades and Corwin (12).
Downloaded by [University of Florida] at 09:06 05 June 2016

REFERENCES

1. Corwin, D.L. and J.D. Rhoades. 1982. An improved

technique f o r determining s o i l e l e c t r i c a l
c o n d u c t i v i t y — d e p t h r e l a t i o n s from above ground
electromagnetic measurements. Soil Sci. Soc. Am. J.
46:517-520.

2. Corwin, D.L. and J.D. Rhoades. 1984. Measurement of

inverted e l e c t r i c a l conductivity p r o f i l e s using
electromagnetic induction. Soil Sci. Soc. Am. J.
48:288-291.

3. Davis, John C. 1986. Principal Components Analysis, pp.

527-546. In: S t a t i s t i c a l and Data Analysis i n
Geology. 2nd Edition. Kansas Geological Survey.

4. DeJong, E., A.K. Ballantyne, D.R. Cameron and D.W.L. Read.

1979. Measurement of apparent e l e c t r i c a l conductivity of
s o i l s by an electromagnetic induction probe t o aid
s a l i n i t y surveys. Soil Sci. Soc. Am. J. 43:810-812.

5. Gerald, C.F. 1970. Applied Numerical Analysis,

Addison-Wesley Publishing Co., Inc., Reading, MA.

6. Keller, G.J. and F.C. Frischknecht. 1966. E l e c t r i c a l

Methods i n Geophysical Prospecting. Pergaman Press, New
York, NY.

7. McNeill, J.D. 1980. Electromagnetic t e r r a i n conductivity

measurement a t low induction numbers. Technical Note
TN-6. Geonics Limited, Ontario, Canada.

8. Myers, Raymond H. 1986. Classical and modern regression

with a p p l i c a t i o n s . Duxberg Press, Boston Mass.
Copyright 1986 by PWS Publisher. Chapter 4 Press
S t a t i s t i c s , pp. 100-136 and Chapter 7, Detecting and
Combating M u l t i c o l l i n e a r i t y , pp. 243-263.
 SOIL ELECTRICAL CONDUCTIVITY-DEPTH RELATIONS 901

9. Rhoades, J.D. 1976. Measuring, mapping, and monitoring

f i e l d s a l i n i t y and water t a b l e depths with s o i l
r e s i s t a n c e measurements. FAO Soils Bull. 31:159-186.

10. Rhoades, J.D. and Jan van Schilfgaarde. 1976. An

e l e c t r i c a l conductivity probe f o r determining s o i l
s a l i n i t y . Soil S c i . Soc. Am. J. 40:647-651.

11. Rhoades, J.D. and D.L. Corwin. 1981. Determining s o i l

e l e c t r i c a l conductivity - depths r e l a t i o n using an
inductive electromagnetic s o i l conductivity meter. Soil
Sci. Soc. Am. J. 45:255-260.
Downloaded by [University of Florida] at 09:06 05 June 2016

12. Rhoades, J.D. and D.L. Corwin. 1989. Soil e l e c t r i c a l

conductivity: Effects of s o i l p r o p e r t i e s and application
to s o i l s a l i n i t y a p p r a i s a l . Comm. i n Soil Science and
Plant Analysis. (Submitted).

13. Rhoades, J.D., N.A. Manteghi, P.J. Shouse and W.J. Alves.
1989. Soil e l e c t r i c a l conductivity and s o i l s a l i n i t y :
New formulations and c a l i b r a t i o n s . Soil S c i . Soc. Am.
J. 53:433-439.

14. Rhoades, J.D., P.J. Shouse, W.J. Alves, N.A. Manteghi and
S. M. Lesch. 1989. Determining s o i l s a l i n i t y from s o i l
e l e c t r i c a l conductivity using d i f f e r e n t models and
estimates. Soil S c i . Soc. Am. J. ( I n P r e s s ) .

15. Rhoades, J.D., S.M. Lesch, P.J. Shouse and W.J. Alves.
1989. New c a l i b r a t i o n s f o r determining s o i l e l e c t r i c a l
conductivity - Depth r e l a t i o n s from electromagnetic
measurements. Soil Sci. Soc. Am. J. 53(1):74-79.

16. SAS/STAT/Proc. Guide f o r Personal Computers, pp. 269-360.

6th Edition. 1985. SAS I n s t i t u t e Inc., Gary, NC.

17. Shapiro, S.S. and M.B. Wilk. 1965. An a n a l y s i s of variance

t e s t f o r normality (complete samples). Biometrika.
52:591-611.

18. Tukey, J.W. 1977. Exploratory data a n a l y s i s , pp. 1-16.

Stem and Leaf analyses. Addison-Wesley, Reading, MA.

19. Williams, Baden G. and G.C. Baker. 1982. An

electromagnetic induction technique f o r reconnaissance
surveys of s o i l s a l i n i t y hazards. Aust. J. Soil Res.
20:107-118.