III. Determination of Soil Salinity from Soil Paste or Bulk Soil Electrical Conductivity
Tải bản đầy đủ  0trang
MEASURING AND MAPPING SOIL SALINITY
213
air
..
..
A
Figure 3. Schematic representation and model of electrical conductivity in soil. (A) The
three paths that current can take in unsaturated soil. (B) Simplified soil model consisting of
the three conductance elements (ac) in parallel. (After Rhoades eta/., 1989a).
seriescoupled and continuous pathway volumetric contents of soil water;
0, is the total volumetric content of soil water; 0, is the total volumetric
content of soil particles; and EC, is the average specific electrical conductivity of the soil particles. The soil water in the continuous pathway, dWc
(=0,  Ow), is envisioned as the “mobile” water phase. It can be different
in electrolyte composition (i.e., EC,) than that in the “immobile” water
phase (i.e., EC,), which is associated with the fine and intraped pore water
(i.e., the immobile water, 0,). At equilibrium, EC, and EC, would be
the same, but during transientstate periods, such as immediately after
irrigation or rainfall, they would likely be different. This model assumes
that EC, is independent of 0, and EC,, which appears to be the case for
most practical purposes (Shainberg et al., 1981; Bottraud and Rhoades,
1985; Rhoades et al., 1990b).
For conditions of EC, greqter than about 2 dS/m and for soils with
typical values of EC, (I 1.5 dS/m), the product 0,EC, is so much larger
than the product B,EC, that the latter can be neglected. Equation ( 5 ) then
simplifies to
[
1
EC, I(0,  O,)EC,
0,
For such cases, the relation between EC, and EC, in Eq. (6) is linear for
EC,
=
(”
+

J. D. RHOADES
214
any value of O,(Ow  0,) beyond some threshold level and the y intercept
depends on EC,, O,, and 0,. Because the ratio [(O, O,)z/O,]
is typically
close to the value 1, the intercept of Eq. (6) is approximately equal to EC,.
The earlier EC, model of Rhoades et al. (1976) is analogous to this limiting
case version of Eq. (9,as discussed elsewhere (Rhoades et al., 1989a).
At low levels of EC,, the relation between EC, and EC, is curvilinear,
as described in Eq. (5). The first term of the equation determines the shape
of the EC,  EC, curve. Over the remainder of the EC, range, EC, and
EC, are linear, with Ow  0, representing the slope, as described above.
So, although Eq. ( 5 ) describes the full relation between EC, and EC,, Eq.
(6) can be used for conditions of EC, 2 24 dS/m (which corresponds
approximately to EC, 1 1 2 dS/m).
A typical set of data illustrating the appropriateness of the abovedescribed model and of generalizationsusing the model is shown in Fig. 4
for Waukena loam soil. The solid line is that described by Eq. ( 5 ) and the
+

.c
0
h
c
4
L    _    _    _ _   _ _  _ :
9,
Waukena loam
12
16
20
Electrical Conductivity of Soil Water, ECw,dS/m
Figure 4. The electrical conductivity of Waukena loam soil as a function of the electrical
conductivity and volumetric content of soil water. The measured data points (0)are shown
and the solid line is the “fit” of these combined data by Eq. (5). (After Rhoades et al., 1989a.)
MEASURING AND MAPPING SOIL SALINITY
3
a
I .o
1
I

0.6

0.4

I
 0.0209
EC,= 0.023%C
0.8
I
215


0.2
I
1
I
I
Clay Content, %
Figure 5. Correlations between EC, and clay percentage for a number of soils from the
San Joaquin Valley of California. (After Rhoades et af.,1989a.)
circles represent experimental data. EC, represents the EC of the equilibrating water or the water expressed from the soil by pressure filtration.
The soil had been extensively leached with waters of different salinities
(EC,), therefore EC, was essentially equal to EC, and to EC, under the
conditions of this experiment. The data and model relations also show that
the ability to determine accurately EC, (or EC,) from EC, decreases as 0,
decreases. This is so because the required accuracy of measurement of EC,
becomes limiting as the EC, =f(EC,) relation flattens at low values of 0,.
At very low values of 0, (ZO.l), it is not possible to determine EC, (or
EC,) from EC, at all (see Rhoades et al., 1976).
To use Eq. ( 5 ) or (6) to assess soil salinity (EC, or EC,) from EC,, the
values of EC, , Ow, and 0, must be known. EC, and 0, can be estimated
using Figs. 5 and 6, respectively. The means of obtaining these relations are
described elsewhere (Rhoades et al., 1989a). The value 0, can be measured
in the field using time domain reflectometry (TDR) methods or it can be
adequately estimated, for many practical purposes, by “feel.” The TDR
method is described later. The value 0, can be estimated from bulk density
(pB) as 0, = &/2.65, where 2.65 is a reasonable estimate of the average
particle density of most mineral soils.
Equation (5) may be solved for EC,, with the assumption that EC, =
EC,, by arranging it in the form of a quadratic equation and solving for its
J. D. RHOADES
216
Volumetric Content of Soil Water, 8,
Figure 6. The volumetric content of soil water in a series path as a function of the total
water content for various soils. (After Rhoades el a/., 1989a.)
positive root:
EC,
=
b+dFXG
+
where a = (O,)(Ow Om), b = (0, Om)z(EC,)
(O,EC,), and c = O,EC,EC,.
If EC, is desired, it can be obtained from
(ECwOw)= (EC,O,
(7)
2a
+ EC,O,)
+ (Ow  O,)(O,EC,)
= EC,(SP/lOO)p,

(8)
where SP is the gravimetric water content of the saturation paste expressed
as a percentage and p~ is the bulk density of the soil (Rhoades et al.,
1989a).
The relations described above give the opportunity to distinguish between the “mobile” and “immobile” contents of soil salinity. Because the
soil water that plants draw on to meet transpiration requirements is primarily 0, and because the salinity effect is primarily one of reducing water
availability to the plant, the salt in the mobile water phase, EC,, ideally
should be used as a basis for relating crop responses to soil salinity. In as
much as EC, is the dominant contributor to EC, in saline soils, it may be
MEASURING AND MAPPING SOIL SALINITY
217
determined from Eq. (6) for a given value of EC, using an intercept term
estimated from the relations of Figs. 5 and 6. This procedure is applicable
2 dS/m. For conditions of lower salinity, the intercept term
where EC,
of Eq. [ 5 ] must be used. The appropriate y intercept value, and thus EC,,
may be determined using successive approximation techniques.
Because of the linear relationship [Eq. ( 6 ) ] that exists at a fixed water
content (such as field capacity) between EC, and EC, at significant values
of EC, (or EC,), soil salinity (either EC, or EC,) may be empirically
related to EC, by an expression of the following type:
EC, (or EC,) = m(EC,  EC:)
(9)
where EC: is [(O, 8,)2/8s](EC,), or EC,, and is predicted as previously
described (i.e., using Figs. 5 and 6), and rn is the slope of the EC, (or
EC,) =f(EC,) relation at field capacity. A typical linear relationship
between EC, and EC, of this type is shown in Fig. 7. Analogous relations
+
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
Electrical Conductivity of Soil, ECa (dS/m)
Figure 7. The relation between bulk soil electrical conductivity and the electrical conductivity of the saturation paste extract (soil salinity) for Dateland soil at field capacity
content. (After Rhoades, 1981.)
218
J. D. RHOADES
have been developed between an EC, and EC, (Rhoades, 1980). With such
calibrations, one can predict EC, (or EC,) from EC, for field soils of
various types, provided they are at or near field capacity water content at
the time of EC, measurement. Simple calibration procedures have been
developed in order to obtain the calibration relations of Eq. (9) appropriate
to field soils with their natural structures, pore size distributions,and water
retention properties (Rhoades, 1976, 1980, 1981; Rhoades and Ingvalson,
1971;Rhoades and van Schilfgaarde, 1976; Rhoades et al., 1977). Numerous satisfactory field calibrations (as in Fig. 7) have been obtained for
many soils around the world and they have been found to be very similar
for soils of similar textures (Rhoades and Ingvalson, 1971;Halvorson and
Rhoades, 1974; Rhoades, 1976, 1979, 1980, 1981; Halvorson et a/., 1977;
Rhoades et al., 1977;Yadau et al., 1979;Loveday, 1980; van Hoorn, 1980;
Nadler, 1981; Bohn et al., 1982). It has been shown that the calibrations
[slopes and intercepts of Eq. (9)] can be predicted from soil properties such
as texture (or percentage clay content) and saturation percentage (or field
capacity) water content (Rhoades, 1981; Rhoades et al., 1989a). For more
information on methods for calibrating EC, =f(EC,), see Rhoades (1976)
and Rhoades et al. (1977).
The effect that varying the soil water content (i.e., departure from Calibration water content) has on salinity determined from Eq. (9) depends on
whether or not salt loss from the soil occurs with the change of water
content (Rhoades et al., 1981). Immediately following an irrigation, salt
loss occurs as the water drains to field capacity; hence, EC, is very sensitive
to changes in 0, during such times. Afler the rapid drainage ceases and the
soil is at field capacity, further major losses of soil water in cropped soils
occur mainly through evapotranspiration. Almost all of the salt in the
water taken up by the plant root system is excluded from entering by the
root membranes; the salt is left behind in the remaining water. Likewise,
no salt is lost through evaporation. Hence, the salt concentration (or
electrical conductivity) of the remaining soil water is increased proportionately as 0, is reduced by evapotranspiration. Because of this inverse proportional relationship between EC, and Ow, the product EC,0, found
within a given soil volume at field capacity will not change appreciably as
0, is reduced below field capacity; thus, the product EC,0, may be considered (ignoring salt precipitation) as approximately constant at water
contents of field capacity or less. However, changes in 0, do affect EC,
through its influence on the partitioning of 0, and Om, as shown in Fig. 6.
As 0, decreases below field capacity due to evapotranspiration, EC, will
show an approximately linear decrease according to the relationship
where (Y
A EC, = aA0,K
(10)
is a factor related to the relation between 0,, and ,
8 and K =
MEASURING AND MAPPING SOIL SALINITY
219
EC, 0, = a constant. For typical soils the error in EC, caused by A 0, is not
large, with reasonable deviation in 0, from field capacity water content.
Experimental data supporting these conclusions about the relatively insignificant effects of A 0, on AEC, and AEC, during and following an irrigation are given by Rhoades et al. (198 1). The appropriateness of using 0, as
a reference for water content in establishing EC, =f(EC,) calibrations is
supported by the results of Bottraud and Rhoades (1985).
It is apparent from the above discussionsthat EC, is primarily a measure
of the total dissolved salt in a soil on a volumetric basis. Hence, given the
value of (EC, 0,) at field capacity, an estimate of EC, at any lower water
content can be calculated from Eq. (8). Hence, the limits of EC, (or
osmotic potential) over an irrigation cycle can be ascertained (estimated)
from EC, (and hence from EC,). For many such practical applicationsthis
procedure can be used in place of Eqs. (5) and (6) to estimate EC, from
EC,.
An equation analogous to Eq. (5) established for bulk soil electrical
conductivity exists for saturated soil pastes, as follows:
where EC, is as defined previously, EC, is the electrical conductivity of the
saturated paste, 0, and 0, are the volume fractions of total water and solids
in the paste, respectively, ,8 is the volume fraction of water in the paste
that is coupled with the solid phase to provide a seriescoupled electrical
pathway through the paste, EC, is the average specific electrical conductivity of the solid particles, and the difference e,  e, is e,, which is the
volume fraction of water in the paste that provides a continuous pathway
for electrical current flow through the paste (a parallel pathway to 0,).
Assuming the average particle density ( p , ) of mineral soils to be 2.65 g/cm3
and the density of saturation soil paste extracts (p,) to be 1.00, 6, and 0,
for saturated pastes can be directly determined from SP as follows:
and
The saturation percentage of many mineral soils can be adequately estimated in the field for purposes of salinity appraisal from the weight of a
pastefilled cup of known volume (Rhoades et al., 1989b). Figure 8 may be
used for this purpose; for details of the relations inherent in this figure, see
Wilcox (195 1).
220
J. D. RHOADES
CUP
GRAMS PASTE
Figure 8. Theoretical relation between saturation percentage (SP) and weight (in grams)
of 50 cm3of saturated paste, assuming a particle density of 2.65 g/cm3.(After Rhoades et al.,
1989b.)
B. DETERMINING
SOILSALINITY
FROM SATURATED
PASTE
ELECTRICAL
CONDUCTMTY
EC, can be determined from measurements of EC, and SP [using Eqs.
(1 1)  ( 13), if values ofp,, Ow, and EC, are known. These parameters can be
adequately and simply estimated, as demonstrated by Rhoades et al.
(1989b,c). For typical arid land soils of the southwestern United States, ps
may be assumed to be 2.65 g/cm3; EC, may be estimated from SP as
EC, = 0.019SP  0.434, and the difference 0,  Om may be estimated
from SP as Ow ,8 = 0.0237(SP).0M57
The measurement of EC, and SP
MEASURING AND MAPPING SOIL SALINITY
22 1
can be easily made using an EC cup of known geometry and volume. The
method is suitable for both laboratory and field applications, especially the
latter, because the apparatus is inexpensive, simple, and rugged and because the determination of EC, can be made much more quickly than that
of EC, .
In this method a saturated soil paste is made as described previously and
is then placed in a conductivity cup of known volume. From the weight of
the paste, SP is determined; from the conductance, EC, is determined.
Then EC, is obtained from Fig. 9 given EC, and SP, using the curve
corresponding to the SP value, or else it is calculated using the following
equation:
EC,
=
b+
2a
+
+
 O,EC,,
where a = O,(Ow  OA, b = (0, 0,,,J2EC, (Ow  O,)O,EC,
and c =  O,EC,EC,. The values of EC,, O,, Ow, and 0, are estimated
from SP using Eqs. ( 12) and ( 13) and the relations given above.
Sensitivity analyses and tests have shown that the estimates used in this
method are generally adequate for salinity appraisal purposes of typical
mineral aridland soils of the southwestern United States (Rhoades et al.,
1989~).For organic soils or soils of very different mineralogy or magnetic
properties, these estimates may be inappropriate. For such soils, appropriate values for p,, EC,, and 0, will need to be determined using techniques
analogous to those of Rhoades et al. (1989b). The accuracy requirements
of these estimates may be evaluated using the relations given by Rhoades et
al. (1 989c).
It should be noted that EC,O, is not equivalent to ECJ, because different amounts of soil are involved in the two measurements. The relation
between these two products is
Ecw~w/Pb= EC,~,/p,
(15)
Data to support this are given by Rhoades (1981) and Rhoades et al.
(1990b). The ratio OJp, is equivalent to SP/lOO (see Rhoades et al.,
1989a,b).
C. DETERMINING
SOILSALINITY
FROM BULKSOILELECTRICAL
CONDUCTIVITY
Soil salinity can be determined from bulk soil electrical conductivity by
essentially one of three ways. Here we discuss these alternative methods of
salinity appraisal, but first the various instrumental means of measuring
EC, will be briefly reviewed.
J. D. RHOADES
222
€
10
80
\
I
I
I
v)
U
90 I
' 100 I
I
I
6
aJ
0
w
I
6
I
1 4
I
I
I
I
2
c
I
I
I
I
w
1 2
I
*0
X
C
0
c
I
I
I
I
I
I
2a
c
Q
v)
v
0
>I
zw
,
c
0
a
U
C
0
0
Q
0
I
L
c
0
Q,
iii
0
2
4
6 8
10 12 14 16 18 20
Electrical Conductivity of Saturation
Paste, EC,,,
dS/m
Figure 9. Relations between electrical conductivity of saturated soil paste (EC,), electrical conductivity of a saturation extract (EC,),and saturation percentage(SP),for representative andland soils. (After Rhoades et al., 1989b.)
MEASURING AND MAPPING SOIL SALINITY
22 3
1. Sensors for Measuring Bulk Soil Electrical Conductivity
Three types of soil conductivity sensors presently exist that are capable
of measuring bulk soil electrical conductivity. Two are fieldproven, portable sensors that are now commercially available: (1) a fourelectrode sensor
and (2) an electromagnetic induction sensor. A third sensor, based on time
domain reflectometry technology, has shown good promise and utility in
certain experimental applications. Each method has its own advantages
and limitations.
a. FourElectrode Units
Bulk soil electrical conductivity can be measured using four electrodes
inserted into the soil, a combination electric current source/resistance
meter, and connecting wire. Such a surface array of electrodes and a
generator/meter unit are shown in Fig. 10. The current sourcemeter unit
may be either a handcranked or a batterypowered type. Units designed
for geophysical purposes generally read in ohms and, if used for general soil
salinity appraisal, should measure from 0.1 to 1000 ZZ. A commercially
Figure 10. A “fixedamy” fourelectrode apparatus and commercial generator/meter.
(After Rhoades, 1992a.)