| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep of Physical Chemistry of the Soil, Soil and Water Inst., ARO, Volcani Center, Bet Dagan, 50-250, Israel
b Lab. for Pest Management Research, Agricultural Engineering Institute, ARO, Bet Dagan, 50-250, Israel
c Hevel Maon Enterprises, M. P. Hanegev, Israel
vwnad{at}volcani.agri.gov.il
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMaterials and methodsResults and discussionSummary and conclusionsREFERENCES |
|---|
), for conventional probes, only when the soil electrical conductivity (EC) is low. The exact limiting level, above which the reflection is unnoticeable, depends on soil texture, salinity, , and probe geometry. Recently, the validity of TDR, measured in saline soils, was questioned and overestimation was attributed to it. This study tested such a possible bias by simultaneously measuring the soil-water content by the TDR and neutron scattering techniques, in two field experiments. The experiments lasted 150 d, and were irrigated with saline waters. In the Zeelim sandy site, peanut (Arachis hypogaea L.) was sprinkler irrigated at a rate of 6500 to 7500 m3 ha-1 with 1.9 dS m-1 irrigation waters. In the Nirim loamy site, cotton (Gossypium hirsutum L.) was drip-irrigated at a rate of 3010 to 4580 m3 ha-1 with waters of EC ranging 3.6 to 4.0 dS m-1. The objective of this study was to observe if salinity increased TDR relative to neutron. Measured neutron values were found to be up to 0.02 or 0.08 m3 m-3 above and below TDR for the sandy or loamy tested soils, respectively, with no significant effect of the salinity. Similarly, both, under- and over estimation were related to the effect of salinity on TDR in over 20 previously published studies comparing TDR to gravimetric under wide ranges of clay, organic matter, and salt contents. The EC effect, expressed as bulk soil EC (
a), may depend also on texture and geometric parameters like particle aspect ratio. At field capacity, for the two soils tested, no detectable effect of salinity was found.
Abbreviations: BD, bulk density • EC, electrical conductivity • EM, electromagnetic • TDR, time domain reflectometry
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and methodsResults and discussionSummary and conclusionsREFERENCES |
|---|
from air dry to water saturation with an error estimate of 0.02. It was accepted that for most soils an empirical relationship between apparent dielectric constant (
) and the volumetric water content () is independent of soil type, density, temperature, and soluble salts content. The potential sources of error that have received attention are (i) clay and organic soils which caused a sharper-than-average curvature of - relations, (ii) a different behavior attributed to the first few water molecules added to the soil because they were constrained by the electrical field of the soil particles showing =ice
3 (later called also "`bound water"), and (iii) a temperature effect on which was found to be small (Topp et al., 1980) and was actually smaller than the theoretical error, under most conditions (Pepin et al., 1995).
Over the past 15 yr, the TDR technique has become a successful method under most field and laboratory conditions. On the basis of TDR measurements in soils wetted with solutions of a given salt concentration, it was suggested initially that measurement of was independent of salt, among other factors. Presently, most users apply Topp's "global" equation to calculate TDR from . Only under exceptional conditions are site-specific calibrations needed to minimize measurement error.
There are conflicting reports on the effect of salinity on the apparent dielectric permittivity, , and on . Some studies suggest that elevated salinity of the soil solution (w) can cause overestimation of , resulting in an overestimation of (Dalton, 1992; Noborio et al., 1994; Wyseure et al., 1997), while others show no effect (Topp et al., 1980; Dalton and van Genuchten, 1986; Nadler et al., 1991). Salinity affects TDR functionality in measuring by increasing the attenuation of the TDR signal, reducing its accuracy and eventually leading to its disappearance.
From 21 studies (Table 1)
reporting the accuracy of measured by TDR in aqueous solutions and in several soil types under different moisture and salinity levels and comparing TDR with gravimetrically determined values (gravimetric), the following can be concluded: (i) of the 22 examples, four TDR values showed no bias, three underestimation, six overestimation, and nine showed both under- and overestimation, relative to gravimetric; and (ii) clays are significant contributors to bulk soil EC (a) and are therefore expected to affect similarly to w.
|
-gravimetric Relations for four frequency bands, representing the TDR frequencies spectrum, in three soils were given by Brisco et al. (1992). For two frequency bands (0.45 and 1.45 GHz) the Rideau soil (53% clay) and the Rubicon soil (sandy loam, 9%) cross each other, and in two other bands (5.3 and 9.3 GHz), the clay curve lies below or above it, respectively. The authors concluded that there is no clear differentiation of the response as a function of soil texture, and texture effects are minimal at all frequencies. – – BD relations for soils, Perdok et al. (1996) compared –BD experimental relations at eight moisture levels for a sandy and clay soils (4 and 36% clay, Perdok et al., 1996, Fig. 3a and 3b, respectively). The clay experimental values fitted least accurately, especially for extreme conditions, and if one cuts across the curves from low to high level, estimation is shifted from under- to overestimation. TDR, calculated by Topp's global calibration, to underestimate by 0.08 (Bridge et al., 1996, Fig. 1.).
Extensive, relevant, and contradicting data regarding salinity effect on was reported by Dalton (1987, 1992) and Dalton and van Genuchten (1986). Dalton (1992) showed an aggregate data for TDR giving a good, one-to-one correlation with gravimetric for saline-solutions saturated soil. However, an overestimation effect of EC on TDR was demonstrated by segregating the very same data according to w, indicating 8 dS m-1 as the value of w above which overestimation of TDR is seen to occur.
These observations by Dalton deserve seven comments.
TDR-gravimetric curve fitting for the aggregate data has a coefficient of r = 0.976 (Dalton, 1992). TDR-gravimetric data are based, each, only on 3 or 4 data points. TDR values at the dry side is ignored. TDR values seem to "feel" the EC effect at salinity lower than 8 dS m-1. TDR was measured in six levels wetted by waters ranging in salinity from 1 to 10 dS m-1. The overestimation bias seems to start at salinity as low as w = 2, 3, or 4 dS m-1. In some cases, EC effect on TDR between 2 to 6 dS m-1 is as effective as the 6 to 10 dS m-1 range (see Dalton 1987, Table 1, = 10, 15 and 20%). w on . Ten columns were wetted to an equal with water of differing w which was later also verified by EC of columns' extracts. From composite results (Dalton and van Genuchten, 1986, Table 1), it is seen that, whereas w ranged 0.8 dS m-1 to 11.1 and a = 0.3 to 1.37 for a = 0.34 ± 0.01, was stable, 19.5 ± 0.5, within the experimental error, indicating that w = 11.1 had no measurable effect on TDR - relations. on salinity
was observed (Shen et al., 1985). However, because the medium is quite different from a characteristic soil, such evidence may not be applicable.
Recently, Wyseure et al. (1997) defined the effect of EC on TDR signal travel time as significant because a played an important role in the loss of energy and slowed down the propagating EM pulse. They set the upper limit of the EC of the soil solution (w) to
for saturated soils.
To enhance the use of TDR for monitoring the soil salinity in drip-irrigated fields, the objective of the present study was to evaluate the practical aspects of salinity effect on in field plots for a sandy and loamy soils.
|
Materials and methods |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and methodsResults and discussionSummary and conclusionsREFERENCES |
|---|
Zeelim Site
The soil is sandy (Xeric or Typic Torripsamment), 3.3% clay, 4.3% silt 92.4% sand, CEC = 3.3 mmol/100g, 2.5% CaCO3, and field capacity 4.8 (w/w). During the last 10 yr, the field was irrigated annually at a rate of 4300 to 7800 m3 ha-1, commercially growing mainly peanut, potato (Solanum tuberosum L.), wheat (Triticum aestivum L.), and sunflower (Heliathus annuus L.). Before seeding the soil was chisel-ploughed, disked, and rototilled.
Peanut (cv Oded) was seeded 7 May in flat raised beds, 1.93 m apart (center to center), in three rows per bed. Seed was inoculated with nitrogen fixing bacteria and therefore no N-based fertilizers were added during growing season except for iron chelates. Sprinkler irrigation (12 by 18 m) was scheduled at 5- to 7-d intervals, during 50 to 165 d after seeding. Salinity of irrigation water was 1.9 ± 0.15 dS m-1, achieved by mixing 1:7 treated sewage (1.5 dS m-1, recycled sewage from Dan Region activated sludge plant) with saline water from a local well (4.65 dS m-1) except during Days 30 to 50 when, because of an increased demand in neighboring fields, fresh, National Carrier waters (1 dS m-1)—rain water which drains into the Sea of Galillee and is pumped to the coastal area—were applied. Experimental treatments included (Table 2) (i) soil mulching with 3-µm-thick black latex sprayed over the field after seeding and instrument installation (Plots 2, 4, 6); (ii) plots without mulch (Plots 1, 3, 5); and (iii) additional 1000 m3 ha-1 water during 1 mo, starting 40 d after seeding (Treatments 5 and 6, Table 2). Plots without plants were established by removing seedlings right after emergence. Three measurement stations were placed in each treatment.
|
by manually analyzing the TDR trace for improved accuracy, and the Topp et al. (1980) equation for calculating .
A neutron probe (Ronly Industries, Rishon Lezion, Israel), located about 1 m away from the TDR probes, measured weekly the soil water content at the 0.2-, 0.35-, 0.50-, 0.65-, 0.80-, 0.95-, and 1.10-m depth (one aluminum access tube per station). Site specific calibrations (separate for the shallowest depth and the rest) were obtained. The accuracy of the neutron was 0.015 and was calculated from the replicates and was in agreement with average accuracy according to Evett and Steiner (1995), Chanasyk and Naeth (1996), and Ould et al. (1997). Soil samples were augered to 0.9 m (0.15-m increments, one profile per station) three times during the growing season for gravimetric water content (oven dried at 105°C) and salinity. The EC of 1:1 aqueous extracts were adjusted to at time of sampling by the ratio 1:1/original and are referred to as w1:1.
Nirim Site
The soil is a loam (Calcic Palexeralf), 14% clay, 38% silt, 48% sand, CEC = 11.9 mmol/100 g, 11% CaCO3, and field capacity of 14.6% (w/w). During the last 10 yr, the studied field was annually irrigated with 4920 m3 ha-1 (salinity levels= 1.4 to 4.4 dS m-1), and cotton or wheat were grown commercially. Detailed description of experimental conditions are given in Nadler and Heuer (1997).
Instrumentation
Operational procedures of TDR and the neutron probe were the same as in Zeelim with a few exceptions. Twenty-four TDR probes in six plots were horizontally installed at the 0.3- and 0.6-m depths in duplicates. The probes were installed perpendicular to the dripper lines, each 0.25 m away from a separate dripper. This is in contrast to the conventional installation approach in which probes are placed under the drippers to minimize variability effects. Weekly measurements were carried out just before the start of the next irrigation event (namely, when conditions were driest). Soil samples were augered to 0.9 m (at 0.15-m increments, one profile per plot) three times during the growing season (Days 22, 97, and 152) for gravimetric water content (oven dried at 105°C) and salinity.
|
Results and discussion |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and methodsResults and discussionSummary and conclusionsREFERENCES |
|---|
w Values were calculated in two ways: (i) applying measured a, TDR, temperature, and clay content to a resistivity-based procedure (Nadler et al., 1984); and (ii) 1:1 aqueous extracts corrected for at sampling time (w1:1). w1:1 at the beginning, middle, and end (Days 1, 75, and 145, respectively) of growing season (Fig. 1
—sand and Table 3
—loam). In the sand, because of irrigation with 1.9 dS m-1 waters, a scattered increase in salinity during the growth season was observed for all six treatments. Mid-season w1:1 values approach 8 dS m-1 and exceed this value at the end of the season (Fig. 1). In the loam (Table 2), averages of w1:1 values for the three treatments are close to 8 dS m-1 at beginning (Day 22) of season, above is (8–14 dS m-1) from mid season (Day 97), and ranging 15 to 27 dS m-1 at the end (Day 152).
|
|
Levels by TDR and Neutron ProbeTDR and neutron could be found during growing season while the salinity was gradually increasing in all six treatments. Although w values exceeded 8 dS m-1 (the critical value of -EC relations), no bias towards overestimation was observed. We do not have a convincing explanation why TDR is consistently below (depth = 0.2 m) or above (depth = 0.5 m) neutron. A priori, a high degree of agreement between these two techniques is not expected under variable conditions because of the order of magnitude of difference in sampled, and therefore averaged, soil volume. It should be noted that the full range of , by the two techniques, throughout the season, at both depths, was 0.04 to 0.08.
|
range was wider, this phenomenon is not observed. When comparing TDR with neutron in the loam, absolute differences were larger relative to the sand but inconsistently so, and did not relate to salinity levels. These differences decreased at the growing season end, when salinity is maximal. A multifactor analysis of variance was conducted to identify the parameters most significantly affecting TDR, including time, depth, w1:1, measuring technique, latex cover, and plant presence, for Zeelim, and time, depth, w1:1, measuring technique, irrigation, and plot, for Nirim. No statistical correlation between TDR and w1:1 was found at either site (Table 4) . The low r2adj for Nirim indicates that other parameters may better explain the model, but are irrelevant to the present study.
|
|
TDR and neutron values. As already mentioned, temperature change was found to be only half the theoretically expected value. Fifty days after seeding, temperature changes in soil are minimal because of surface shading by plants canopy. Furthermore, was measured at the same time of the day and in the 0.3- to 0.6-m depth, temperature changes are suppressed.
These seemingly inconsistent observations on the EC effect on may be explained by the following mechanism. The exchangeable ions residing on the particles' surfaces (e.g., on clays, charged colloids), being a high ion concentration confined to a thin layer, form a two-dimensional surface conductive system. Their physical mobility is large along the surface but small relative to the surface (O'Konski, 1995). This two-dimensional surface conductance (s) is interacting with the electrical component of the EM signal. This interaction is a function of signal frequency, signal direction, and soil properties like s, size distribution, and salinity.
O'Konski (1995) characterized this interaction in the clear physical picture of the high/low EM frequencies interacting with the low/high surface conductance. The distribution of particle sizes will affect the interaction because the relaxation time (
) is proportional to the particle size
, where R is the particle radius. For nonspherical particles, there will be more than one depending upon the direction of the polarization field with respect to the particle axes.
Inconsistency in findings because of the variation in clay content, water content, salinity, and its composition is related by Sen (1984) to the anisotropy in the grain shape of rocks and soils, and he attributed this to rock texture and porosity (expressed as the distribution of spherical and platy grains). Also according to Kenyon (1984), the limiting value of does not depend on w, but rather the frequency at which this value is approached depends upon w, BD, porosity, and the ratio of spherical/platy particles.
Tyc et al. (1988) compared two calculation models, one based on randomly dispersed spherical and platy grains with a single aspect ratio, and the other based on a model in which the platy grains were confined to a thin layer on the pore grain interface. Calculations based on the second model, including a distribution of aspect ratios, fitted better the experimental results. At low frequencies, the platy grain layer reduces the effective area of the pore throats, thereby lowering the conductivity. At high frequencies, the platy particles do not block the flow of current and the entire throat area contributes to the conductivity. High aspect ratio particles must be situated in regions of space where the electric field strength in the direction of the platy grain principal axis, is large.
Another indirect indication for this reasoning may be found in Roth's et al. (1990) functional approach in which is derived from dielectric mixing models. They relate the composite dielectric number (c) of a multiphase mixture to the dielectric numbers (i) and volume fractions (
i) of its constituents, assuming geometrical arrangements of the constituents. For example, a layered two phase medium gives
where 
can have values from -1 to +1. The relative location in packing of the soil components (sand grains, clay platelets, ions, and H2O molecules) was critical and the same compositions, resulting in different geometrical patterns, will result in different c.
|
Summary and conclusions |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and methodsResults and discussionSummary and conclusionsREFERENCES |
|---|
overestimation. Since the total conductivity has to be considered, the proper factor to quantify EC effect is a, which takes into account also texture and water content, rather than w. The clay fraction, similarly to soil salinity, is not monotonously affecting , but depend on particles aspect ratio. Unfortunately, this may serve only as a general indicator of the upper value for a safe use of TDR. The effectiveness of the integrated interaction between the electrical fields of the pulse and the ions depends on the specific parameters of the soil components.
According to the present study data (supported also by studies cited in Table 1), under field capacity, for sandy and loamy soils, TDR can be safely used up to a 2 dS m-1 (which is roughly equivalent to w 16 to 26 dS m-1). A higher or clay content may mean that this upper limit will be reached under lower w. Unfortunately, EC is just one factor out of several and its effectiveness is not constant and therefore a separate confirmation is encouraged.Evett Steiner 1995; O'Konski 1955; Ould Mohamed Bertuzzi Bruand Raison Buckler 1997; Zegelin 1992
|
ACKNOWLEDGMENTS |
|---|
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and methodsResults and discussionSummary and conclusionsREFERENCES |
|---|
Received for publication May 1, 1998.
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and methodsResults and discussionSummary and conclusionsREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  C. Barton, D. Marx, D. Adriano, B. J. Koo, L. Newman, S. Czapka, and J. Blake Phytostabilization of a landfill containing coal combustion waste Environmental Geosciences, December 1, 2005; 12(4): 251 - 265. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Logsdon Time Domain Reflectometry Range of Accuracy for High Surface Area Soils Vadose Zone J., November 11, 2005; 4(4): 1011 - 1019. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. D. Logsdon Soil Dielectric Spectra from Vector Network Analyzer Data Soil Sci. Soc. Am. J., June 2, 2005; 69(4): 983 - 989. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Persson, D. Bendz, and P. Flyhammar Time-Domain Reflectometry Probe for Water Content and Electrical Conductivity Measurements in Saline Porous Media Vadose Zone J., November 1, 2004; 3(4): 1146 - 1151. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Vadose Zone Journal | Journal of Plant Registrations | ||||
| Journal of Natural Resources and Life Sciences Education |
Journal of Environmental Quality |
||||
) and a neutron probe (
) for two depths in six plots as a function of time, in Zeelim site. SD of error is presented by the dotted (for TDR) and dashed (for the neutron probe) lines