| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dpto. Producción Vegetal, Univ. de Almería, La Cañada, 04120 Almería, Spain
b Research Station of the Cajamar Foundation, Autovía del Mediterráneo, km. 416.7, 04710 El Ejido, Almería, Spain
c Dpto. Ciencias del Agua y Medio Ambiente, Instituto Tecnológico de Sonora, 5 de Febrero 818 Sur, Cd. Obregón, Sonora, México
* Corresponding author (rodney{at}ual.es).
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: CS, capacitance sensor • EC, electrical conductivity • ECe, electrical conductivity of saturated soil extract • ECns, electrical conductivity of nutrient solution • ECsw, electrical conductivity of soil solution • ECw, electrical conductivity of water • ETc, crop evapotranspiration • PVC, polyvinyl chloride • +SAL, treatment of standard nutrient solution with added salts • SAL-N1.9, treatment of 5 dS m–1 nutrient solution where capacitance sensors were normalized in a solution of 1.9 dS m–1 • SAL-N5.2, treatment of 5 dS m–1 nutrient solution where capacitance sensors were normalized in a solution of 5.2 dS m–1 • SMP, soil matric potential • SWC, volumetric soil water content
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The EnviroSCAN system consists of individual cylindrical ring sensors mounted on a vertically installed probe at different depths (Paltineanu and Starr, 1997) within a polyvinyl chloride (PVC) access tube. Before installation, each individual sensor is "normalized," whereby sensor-specific readings of oscillation frequency are determined in water (Fw) and air (Fa); these are subsequently related to the frequency measured in the soil (Fs) to calculate the scaled frequency (SF) as
 | [1] |
Intensive vegetable production systems are well suited for the use of soil moisture sensors for irrigation scheduling. Being high-value systems on relatively small fields, they are better able to afford sensor systems and to monitor soil moisture representatively. Commonly, these systems are located in relatively dry climatic regions, such as the Mediterranean coast, where fresh water resources are limited and subject to increasing competition from other economic sectors. One such system is the greenhouse-based vegetable production system located on the southeastern coast of Spain, in which approximately 80% of cropping occurs in soil (Castilla and Hernández, 2005; Thompson et al., 2007c).
Salinity in the root zone can be influenced by various sources. There is appreciable geographic variation in the salinity of irrigation water. In drier climatic regions, there is variation in naturally occurring soil salinity. During irrigated cropping, there is a tendency for soil salinity to increase. Even where salt accumulation during irrigated cropping is well managed, there may be fluctuations in soil salinity during an irrigated crop. When fertigation is used to apply nutrient solutions, the salinity of the applied nutrient solution may be varied for crop management purposes, such as to promote fruit set and to increase fruit sugar content. For these reasons, intensive vegetable production occurs within a range of soil salinity conditions, and soil salinity may change appreciably during an individual crop.
There are reports that SWC measured by the EnviroSCAN sensor, hereafter referred to as the capacitance sensor (CS), can be positively affected by increasing salinity (Mead et al., 1995; Baumhardt et al., 2000; Kelleners et al., 2004b). Working with solutions, Kelleners et al. (2004b) demonstrated that CS readings were affected by solution electrical conductivity (EC). In a soil column study, Baumhardt et al. (2000) observed a large and rapid overestimation in SWC following the addition of a highly saline 11.3 dS m–1 solution. Working with soil columns, Mead et al. (1995) reported, in moist soil, a 10% overestimation for a soil with EC of the saturated extract (ECe) of 3.4 dS m–1 (after applying a moderately saline 5 dS m–1 solution), and that the relative overestimation increased appreciably at higher ECe. In the latter study, normalizing sensors with the applied saline solutions, rather than tap water, either eliminated or appreciably reduced the salinity-induced overestimation. There are very few published studies of salinity effects on SWC measurements made with the CS system in field studies under crop production conditions. There are no published studies examining whether the use of saline solutions in the normalization procedure reduced salinity effects under field conditions.
This work was conducted to evaluate two objectives. The first was to assess the effects of changes in soil salinity on SWC measurements made with the CS in field studies under vegetable production conditions. The second was to assess whether normalizing individual sensors with more saline water reduced salinity effects on SWC measurement.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Theoretical Considerations of Salinity Effects on Soil Water Content Measurement
General descriptions of the theoretical basis of the measurement of volumetric SWC using capacitance sensors are given by Paltineanu and Starr (1997), Gardener et al. (2001), Starr and Paltineanu (2002) and Fares and Polyakov (2006). A detailed theoretical description is given by Kelleners et al. (2004b). Essentially, (i) SWC is a function of the oscillation frequency (F) of an electrical circuit formed by the sensor within the soil medium, (ii) the oscillation frequency is determined by the capacitance of the circuit, and (iii) the capacitance of the circuit (in which soil acts as a capacitor) is a function of the relative dielectric permittivity (
r) of the soil being measured.
To account for the complexity of its behavior in soil, measured r is considered as being the complex relative permittivity (r*), which consists of a "real" part (r') describing energy storage, and an "imaginary" part (r'') describing energy losses (Gardener et al., 2001; Kelleners et al., 2004b):
 | [2] |
(–1), and r'' is the sum of a conductivity term and a relaxation term:
 | [3] |
is ionic conductivity (in soil, the bulk soil electrical conductivity), 
is the angular frequency ( = 2
F), 0 is the permittivity in vacuum, and r,rel'' are losses due to relaxation.
Capacitance and other dielectric sensors are based on measurement of r*, which is considered to be r' based on the assumption that r'' << r'. Evett and Parkin (2005) suggested using the term apparent complex permittivity (a*) rather than the term complex relative permittivity (r*) for the measured parameter, to emphasize that this parameter is indirectly measured and that energy losses (r'') may not be negligible.
It has been established that at operating frequencies of <50 MHz, soil electrical conductivity can influence SWC measurement with capacitance sensors (Gardener et al., 2001). It has been reported that at operating frequencies of >50 MHz, soil electrical conductivity does not affect capacitance sensor measurement (Gardener et al., 2001). The present CS operates at frequencies of >100 MHz (Paltineanu and Starr, 1997; Kelleners et al., 2004b). Initial descriptions of the present CS assumed that salinity effects, on the r'' term, would be minimal at these operating frequencies (Paltineanu and Starr, 1997; Starr and Paltineanu, 2002).
Location and Cropping Details
The three experiments were conducted in two greenhouses at the field research station of the Cajamar Foundation, in El Ejido, Almeria province, in southeastern Spain (2°43' W, 36°48' N, and 151-m elevation).
Greenhouses and Soils
The two greenhouses used were constructed of plastic film placed over wire mesh supported by a metal frame; they were unheated and passively ventilated. Experiments 1 and 2 were conducted in Greenhouse 1, which measured 28 m long by 22.5 m wide, and Exp. 3 in Greenhouse 2, which measured 24 m long by 20 m wide. Both greenhouses had an artificial layered soil, as is commonly used in greenhouses in the region, formed by placing a layer of soil, imported from a quarry, over the original stony, loam soil and a 10-cm layer of coarse river sand over the imported soil as a mulch (Castilla and Hernández, 2005; Thompson et al., 2007c). The <2-mm fraction of the original soil had 46% sand, 32% silt, and 22% clay, and a bulk density of 1.6 Mg m–3. In Greenhouse 1, the layer of imported soil was 20 cm deep, had a sandy loam texture (55% sand, 28% silt, and 17% clay), bulk density of 1.5 Mg m–3, and 11 g kg–1 of organic C. In Greenhouse 2, the layer of imported soil had a depth of 30 cm, a clay texture (15% sand, 37% silt, and 48% clay), bulk density of 1.5 Mg m–3, and 7 g kg–1 of organic C. The dominant clay mineral in the three soils is illite.
In both greenhouses, drip irrigation tape was placed on the surface of the sand mulch, with 1.5-m spacing between drip lines and 0.5-m spacing between emitters within drip lines; the emitters had a discharge rate of 2.5 L h–1. All crops were grown with single plants adjacent to and 8 cm from individual emitters, such that emitters and plants formed paired lines. The irrigation water had an electrical conductivity (ECw) of 0.4 dS m–1.
Experiments 1 and 2 were conducted in plots in Greenhouse 1. Individual plots on the southern side of the greenhouse measured 10.5 by 4.5 m, and those on the northern side measured 8.5 by 4.5 m. There were three lines of emitters, each with an adjacent line of plants in each plot. Adjacent plots were partially hydraulically separated by vertically placing plastic sheeting to a depth of 30 cm from the surface of the imported soil. The experimental design was a randomized block design with four blocks (two in the northern side and two in the southern side of the greenhouse), with the two treatments randomly allocated within each block. Single additional plots in each block were not used in these experiments.
Experiment 3 was conducted in plots measuring 2.0 by 1.5 m; each plot consisted of five consecutive emitters and the corresponding adjacent plants. One or two such plots were located within each drip line. There were three replicate plots for each treatment.
Crops
In Exp. 1, a spring-cycle tomato (Lycopersicon esculentum Mill. cv. Boludo) was grown from 23 Jan. to 6 July 2004. In Exp. 2, melon (Cucumis melo L. cv. Sirio) was grown from 21 Feb. to 25 June 2003. In Exp. 3, a winter-cycle tomato (cv. Raf) was grown from 5 Sept. 2005 to 16 Mar. 2006 and a spring-cycle tomato (cv. Raf) from 17 Mar. to 4 July 2006. Unless otherwise indicated, crop management followed local practices.
Treatments and Experimental Design
Experiment 1
Experiment 1 compared the effect of two nutrient solutions with different salinities on CS measurement (Fig. 1
). In the control treatment, a standard complete nutrient solution with an electrical conductivity (ECns) of 2.4 dS m–1 was continuously applied (Table 1
). In the +SAL treatment, a complete nutrient solution with a ECns of 6.5 dS m–1 was applied between 22 Mar. and 24 June 2004 (Fig. 1). Before and after this period, a 2.4 dS m–1 nutrient solution was applied. The +SAL nutrient solution was prepared by adding equal amounts of CaCl2 and NaCl to the same nutrient solution as used in the control treatment. Both treatments were irrigated identically, receiving 100% of estimated crop evapotranspiration (ETc), using relationships derived for vegetable crops in this horticultural system, described in Fernández et al. (2000) and Orgaz et al. (2005). Manual tensiometers (Model ISR-300, Irrometer Co., Riverside, CA) were used to determine irrigation frequency (threshold: –30 kPa).
|
|
|
|
The three replicate plots of the control treatment were located within two drip lines of emitters and associated plants that received the 1.9 dS m–1 nutrient solution. The six plots for the two salinity treatments were located, in a randomized block design, within three lines of emitters that received the 5 dS m–1 nutrient solution. The lines of emitters receiving 5 and 1.9 dS m–1 nutrient solutions were alternated. The experimental area was located within tomato crops of the same cultivar and planting dates.
Measurements
Capacitance Sensors
In the three experiments, SWC was measured with EnviroSCAN CSs supported on probes positioned vertically within PVC access tubes. The SWC measurements were made and recorded every 30 min on a datalogger (Model RT6, Sentek Sensor Technologies) for the 0- to 10- and 10- to 20-cm soil depths in Exp. 1 and 2, and from the 5- to 15- and 15- to 25-cm soil depths in Exp. 3. In Exp. 1 and 2, average SWC values for 0 to 20 cm were calculated from 0- to 10- and 10- to 20-cm values. In Exp. 3, SWC data for the 5- to 15- and 15- to 25-cm depths were considered separately. All soil depths given are relative to the surface of the layer of imported soil; the overlying layer of coarse sand was considered solely as a mulch. The PVC access tubes (56.7-mm o.d.) were installed by augering a hole of slightly larger diameter (approximately 65-mm diameter), backfilling the hole with a thick slurry made from the same imported soil as used in the greenhouse, and pushing the access tube, sealed at the lower end, into the hole. The opening at the soil surface around the perimeter of the access tube was sealed with a 2-cm-wide ring of thick soil paste. Access tubes were installed at least 4 mo before treatments were first applied.
Before installation, each CS was normalized, as previously described, by determining Fw and Fa in a PVC access tube in a water bath and air, respectively, at room temperature. In Exp. 1 and 2, Fw was determined in water (ECw of 0.4 dS m–1). In Exp. 3, for the control and SAL-N1.9 treatments, Fw was determined in a 1.9 dS m–1 nutrient solution, as applied to the control treatment; for the SAL-N5.2 treatment, Fw was determined in a nutrient solution with an electrical conductivity very similar to the 5 dS m–1 nutrient solution applied to the SAL-N1.9 and SAL-N5.2 treatments.
The CSs were calibrated in situ for each of the soils used. For the 0- to 20-cm layer of the sandy loam soil used in Exp. 1 and 2, in Greenhouse 1, the CS was calibrated against SWC measurements made with a time-domain reflectometry (TDR) system (Model Trase 6005X1, Soil Moisture Corp., Santa Barbara, CA). A previous study had demonstrated the accuracy of this TDR system in the same soil (Fernández et al., 2004). The calibration equation for the CS in this soil was SWC = 1.1807SF – 0.6972 (r2 = 0.882, n = 20); SF was defined in Eq. [1]. The calibration for the 5- to 25-cm depth of clay soil used in Exp. 3, which was determined gravimetrically, was SWC = 0.92SWCuncorrected – 0.0154 (r2 = 0.600, n = 13), where SWCuncorrected was measured with the manufacturer's calibration of SWCuncorrected = (0.792SF – 0.023)2.475. This calibration corresponded to a soil matric potential (SMP) range of 0 to –80 kPa.
Soil Matric Potential
In Exp. 1 and 2, SMP was measured with tensiometers equipped with pressure transducers (Model SKT 600/IE, Skye Instruments, Llandrindod Wells, Wales, UK); measurements were made every 5 min, which were then averaged and recorded every 30 min on a datalogger (Model Data Hog 2, Skye Instruments). Each tensiometer had been individually calibrated by the manufacturer. The midpoint of the ceramic cup was installed at 10-cm depth from the surface of the imported sandy loam soil layer. Four replicated tensiometers were installed, each in a different plot. For Exp. 1, 0600-h data from each day were selected for presentation.
In Exp. 3, manual tensiometers (Model ISR-300, Irrometer Co., Riverside, CA) were used. Measurements were made visually at 0900 h on working days, before any irrigation. Tensiometers were installed so that the midpoint of the ceramic capsules was at 10 or 20 cm. One replicate tensiometer was installed at each depth in each plot.
Salinity Measurements
Electrical conductivity was measured with a conductivity meter (Model Basic 30, Crisol Instruments S.A., Alella, Spain). Samples of applied nutrient solutions were collected at drip irrigation emitters using plastic trays. Samples of soil solution were obtained using soil solution samplers (Model SSAT-30, Irrometer Co., Riverside, CA). The depths of installation and numbers of replicates were as described for tensiometers. The sampling frequency in Exp. 1 was every 7 to 14 d. In Exp. 2, in the first irrigation treatment period, it was 7 to 11 d until 1 May 2004, after which it was every 1 to 2 d; in the second irrigation treatment, it was every 4 to 7 d. In Exp. 3, between 29 Dec. 2005 and 12 Apr. 2006, ECsw was measured every 2 to 7 d. Between 13 Apr. and 29 June 2006, measurements were generally every 3 to 9 d, with the exception of two periods of 17 and 18 d due to difficulties of maintaining sampler vacuum. The soil solution suction samplers were sampled by applying a vacuum of –80 kPa for 6 to 24 h, which was applied 9 to 24 h after irrigation. Additionally, in Exp. 3, ECe (Rhoades and Miyamoto, 1990) from the 5- to 15- and the 15- to 25-cm soil depths was measured on 17 Feb., 26 Apr., 22 June, and 30 June 2006.
Sensor Location
In Exp. 1 and 2, all sensors and samplers were installed along the central line of the three lines of emitters in each experimental plot. In Exp. 3, the CS sensors were located next to the central line of the five emitters that formed the experimental plots; the two tensiometers and two soil solution samplers were located on either side of the immediately adjacent emitters. In Exp. 1, 2, and 3, and associated calibration studies, all sensors were located 6 cm from the line of emitters (perpendicular to the line of emitters) and 8 cm from a paired emitter and plant in the direction parallel to the line of emitters.
Statistical Analysis
Statistical analyses were conducted with Statgraphics Plus Version 4.1 (Manugistic Co., Rockville, MD).
|
RESULTS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
Soil matric potential (10-cm depth) was very similar in both treatments throughout the treatment period (Fig. 5a ). The relationship between SMP measured at 0600 h in the +SAL and control treatments was described by the linear regression y = 1.0511x – 0.3768, with an r2 value of 0.92 (Fig. 5a). There were no statistically significant differences (P < 0.01) between the slope and intercept of this equation and those of the 1:1 line, indicating that SMP, and therefore the actual soil water content, were very similar in both treatments throughout the experimental period.
|
Experiment 2
There were clear differences between ECsw (10-cm depth) of the well-irrigated treatment during the first and second irrigation treatment periods (Fig. 6a
and 6b). During the first irrigation treatment period, following commencement of the addition of the more saline solution on 10 Apr. 2003 (Fig. 2), there was a progressive increase in ECsw from an initial value of 1.8 dS m–1, reaching a maximum value of 7.2 dS m–1 on 30 April (Fig. 6a). The subsequent salt leaching water-only irrigations applied during 2 to 7 May (Fig. 2) caused a rapid decline in ECsw in early May (Fig. 6a). During the second irrigation treatment period (Fig. 2), the ECsw of the well-irrigated treatment remained relatively constant, with an average value of 2.3 dS m–1 (Fig. 6b).
|
In the first irrigation treatment period, measured SWC (0–20 cm) in the well-irrigated treatment increased progressively from 14 April to 1 May; the within-day range increased from a range of 0.27 to 0.30 m3 m–3 to a range of 0.33 to 0.36 m3 m–3 (Fig. 6e). During this period, the within-day range of corresponding SMP values was relatively constant (Fig. 6c). There was a rapid decline in measured SWC (Fig. 6e) during 2 to 7 May when the large water-only irrigations (EC of 0. 4 dS m–1) were applied. During this period, SMP data (10-cm depth) indicated that the soil was consistently wetter than during the preceding 3-wk period (Fig. 6c). The sequential increase and subsequent rapid decrease in measured SWC in the well-irrigated treatment during the first irrigation treatment period (Fig. 6c) paralleled similar changes in the ECsw (Fig. 6a). For the deficit-irrigated treatment in the first irrigation treatment period, there was a similar discrepancy between the SWC and SMP data. Whereas SWC remained relatively constant (Fig. 6e), SMP showed a reduction (Fig. 6c) that was consistent with being irrigated with 25% of ETc.
For the equivalent treatments in the second irrigation treatment period, the SWC data were very different (Fig. 6f) from those measured in the first irrigation treatment period (Fig. 6e). Throughout the second treatment period, the overall range of SWC in the well-irrigated treatment remained constant at 0.27 to 0.31 m3 m–3 (Fig. 6f). This range was very similar to that measured in the well-irrigated treatment of the first treatment period before increasing the salinity of the applied nutrient solution (Fig. 6e). In the second treatment period, when ECsw remained constant (Fig. 6b), SWC and SMP data were in agreement for each irrigation treatment (Fig. 6 d,f), and were consistent with the corresponding irrigation management.
Experiment 3
Immediately before commencing application of the 5 dS m–1 nutrient solution to the two salinity treatments (SAL-N1.9 and SAL-N5.2) on 11 Jan. 2006 (Fig. 3), the ECsw (10-cm depth) was very similar for the three treatments, with values of 2.1 to 2.3 dS m–1 (Fig. 7a
). After commencing application of the more saline nutrient solution to treatments SAL-N1.9 and SAL-N5.2, an increase in ECsw was apparent 1 wk later in both treatments (Fig. 7a). Subsequently, the ECsw increased rapidly to 4.6 dS m–1 by 31 January. Thereafter, until 14 February, there was a much slower increase to 4.8 to 4.9 dS m–1. During most of the first salinity treatment period (11 January to 16 February; Fig. 3), the ECsw (10-cm depth) of the two salinity treatments (SAL-N1.9 and SAL-N5.2) was clearly much higher than that of the control treatment (Fig. 7a). The series of large water-only irrigations between 17 February and 16 March caused a large and rapid reduction in ECsw of the two salinity treatments to 1.2 to 1.8 dS m–1. Throughout the period from 29 Dec. 2005 to 24 Feb. 2006, the ECsw of the SAL-N1.9 and SAL-N5.2 treatments fluctuated similarly and had very similar values (Fig. 7a). There was some variation with time of the ECsw of the control treatment, firstly increasing from 2.1 to 2.9 dS m–1 by 31 January, and then decreasing to 1.0 dS m–1 by 15 March because of the large water-only irrigations, which were applied to all treatments.
|
Soil matric potential (10-cm depth) was very similar in the three treatments from 23 Dec. 2005 to 11 Apr. 2006 (Fig. 7e), indicating that there were no real differences in soil moisture between the three treatments during this period. From 17 February to 15 March, there were large and rapid reductions in measured SWC in the SAL-N1.9 and SAL-N5.2 treatments (Fig. 7c). During this period, SMP for the three treatments was constant and consistently higher than –10 kPa (Fig. 7e). These SMP data indicate that, between 17 February and 15 March, soil in the SAL-N1.9 and SAL-N5.2 treatments was constantly very moist and not drying as suggested by the SWC data.
Saturated soil extract ECe values (10-cm depth) for the control and saline treatments were 1.8 and 3.2 dS m–1, respectively, on 17 February before applying the salt leaching irrigations, and 1.4 and 1.5 dS m–1, respectively, on 26 April after applying the salt leaching irrigations (Table 2 ).
|
The SWC data for the two salinity treatments for 2 to 17 May 2006 were lost due to technical problems. After 18 May, SWC data (5–15-cm depth) showed much higher values in the two salinity treatments than in the control (Fig. 7d), despite receiving identical irrigation volumes. Rapid reductions in measured SWC in the two salinity treatments followed the water-only irrigations that commenced on 15 June. The SMP data (10-cm depth) indicated that SMP was consistently very similar in the three treatments (Fig. 7f). After 15 June, SMP in the salinity treatments did not change (Fig. 7f) when large and rapid reductions were observed in SWC (Fig. 7d). For the control treatment during the second salinity treatment period, there was a progressive increase in the ECsw at 10-cm depth (Fig. 7b) and a similar small progressive increase in SWC at 5 to 15 cm (Fig. 7d). There was a larger progressive increase in SWC of the control at 15 to 25 cm (data not presented), which paralleled a similar increase in the ECsw at 20-cm depth (data not presented).
For the entire study period of Exp. 3, from 24 Dec. 2005 to 30 June 2006, the data for ECsw, SWC, and SMP for the 20-cm depth (data not presented) were generally similar to the corresponding measurements for the 10-cm depth (Fig. 7). For the ECsw and SWC, the same general tendencies and relative differences between treatments were observed in each of the two depths; the differences at 20 cm were generally somewhat slower and smaller than those at 10 cm.
The effect of normalization of CS with solutions with different ECns values was assessed by linear regression analysis of the relationship between SWC measured at 0600 h in the SAL-N5.2 and SAL-N1.9 treatments, for the period 24 Dec. 2005 to 11 Apr. 2006, at the 5- to15- (Fig. 8a ) and 15- to 25-cm depths (Fig. 8b). For both soil depths, the fitted linear regression had a slope that was not significantly different from that of the 1:1 line (P < 0.01), indicating that the EC of the solution used for normalization had no effect on the overestimation of SWC caused by salinity.
|
|
|
DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
In the three experiments reported here, positive effects of salinity on CS measurement occurred when the ECsw increased. The rapidity of the response and the parallel nature of the changes, in SWC measurement and in the ECsw, indicated that the CS was sensitive to increases from background levels of ECsw in these studies. Background ECsw before addition of the more saline nutrient solutions was 1.8 to 2.6 dS m–1. Given that these soils were at or close to field capacity, ECe can be estimated as 0.5ECsw (Rhoades and Miyamoto, 1990), suggesting approximate background ECe values of 0.9 to 1.3 dS m–1. Using crop sensitivity to soil salinity as a guide (Rhoades and Miyamoto, 1990), the salinity effects on CS measurement reported here can be considered as occurring under conditions of moderately low to moderate soil salinity.
The effects of salinity on CS measurement of the SWC in the present field studies are consistent with previously reported results of laboratory studies. Mead et al. (1995) and Baumhardt et al. (2000) observed that the CS overestimated SWC following the addition of saline solutions to soil columns. Kelleners et al. (2004b) reported that increasing ionic conductivity of the measured medium reduced the oscillation frequency measured by the CS, resulting in higher SWC values.
Estimates of the relative error per unit of ECsw in the work reported here are approximate because of the dynamic nature of the data collected under field production conditions. In Exp. 1, in a sandy loam soil, the relative increase in SWC was 7.5% for each 1 dS m–1 increase in ECsw. In Exp. 3, in a clay soil, the relative increase in SWC was 4% for every 1 dS m–1 in ECsw. In a soil column study, Mead et al. (1995) reported a relative increase of 3.5% for every 1 dS m–1 of ECe in moist soil; assuming ECe 
0.5ECsw (Rhoades and Miyamoto, 1990), these values would correspond to approximately 7% for every 1 dS m–1 in ECsw, which is similar to estimates in the current work. In a study with sand columns, relative increases in SWC were approximately 2 to 5% for each 1 dS m–1 increase in ECw, where ECw was the EC of applied water flushed through the columns, the higher values being recorded with moister soil (P. Cepuder, personal communication, 2007). A summary of available experimental data, with the same CS, suggests that responses to salinity are 2 to 8% relative increases in SWC for every 1 dS m–1 increase in the ECsw.
The regression analysis conducted in Exp. 3 suggested a linear increase in the overestimation of SWC with increasing salinity. Working across a much larger range of soil salinity, Mead et al. (1995) reported that the overestimation increased disproportionately with increasing soil salinity. A similar curvilinear response to an EC range similar to that of the present work was observed in another column study (P. Cepuder, personal communication, 2007). There are insufficient data available to characterize the nature of the response of SWC measurement to salinity; however, there are suggestions of a disproportionate curvilinear response.
The sensitivity of SWC measurement to changes in the ECsw suggests that commonly occurring changes in soil salinity during cropping, from periodic fertilizer addition or from salt accumulation during irrigated crop production, can influence the SWC measurement with this CS. Therefore consideration of salinity effects is required whenever changes in soil salinity occur during crop production. Such consideration is required for quantitative SWC measurement, and also where trend analysis is used for on-farm irrigation management (e.g., Thompson et al., 2007b). Where salinity changes are likely, measurement of soil salinity is required. For quantitative SWC measurement where salinity changes occur, in situ calibration of the CS may require a secondary calibration for salinity effects, done either in soil columns or in situ, or at least the use of mathematical corrections. In horticultural crops where appreciable increases in soil salinity are used for crop management purposes, considerable care should be taken for research applications and when used for on-farm irrigation management.
Developments in permittivity mixing models and calibration equations (Gardener et al., 2001) may assist in reducing the effects of salinity on SWC measurement with the CS. Kelleners et al. (2004b) developed a calibration procedure for this CS based on electric circuit theory that corrected permittivity losses due to ionic conductivity and relaxation. Kelleners et al. (2004a) combined this and another calibration model to calibrate the CS for SWC measurement in saline to very saline soils (ECe of 6–35 dS m–1) with generally satisfactory results.
The sensitivity to salinity of SWC measurement indicates that the assumption of r'' << r' (see above) does not apply to this CS under conditions that are generally representative of agricultural mineral soils. Ionic conductivity (), which in soil is the bulk soil electrical conductivity, can appreciably influence r'' (Eq. [3]) and consequently the measured r* value (Eq. [2]). Working with several different soil moisture sensors that respond to electromagnetic properties of soil (not including this CS), in a laboratory study with solutions, Blonquist et al. (2005) observed that those operating at lower frequencies (50–200 MHz) were generally sensitive to . The results of the present field studies are consistent with that observation. The implications of bulk soil electrical conductivity influencing r'' and therefore r* were discussed by Evett and Parkin (2005).
In Exp. 3, normalizing the CS in a 5.2 dS m–1 nutrient solution compared with a 1.9 dS m–1 nutrient solution did not reduce the effect of salinity on SWC measurement. This contradicts the soil column results of Mead et al. (1995), where normalizing the CS with saline solution rather than water completely removed the salinity-induced error at moderate soil salinity levels (ECe of 3.4 dS m–1) and appreciably reduced it at higher salinity levels (ECe of 7.7 and 15.4 dS m–1). For the normalization procedure to account for salinity effects, it is necessary that the numerator and denominator terms in Eq. [1] remain proportional. To do so, the salinity-induced reduction in the terms Fs and Fw in Eq. [1] must be very similar. Under dynamic field conditions where soil salinity is constantly changing, salinity effects on Fs will be constantly changing, whereas the effect on Fw is fixed. Normalizing with more saline solutions will be more effective in column studies (e.g., Mead et al., 1995) where soil salinity will be more stable and more similar to that of the solutions used for normalization.
Further information is required to characterize the behavior of the CS to changing soil salinity and to higher salinity conditions. For example, more information is required to: (i) characterize and quantify the nature of the response (linear, exponential type curve, etc.); and (ii) understand and quantify interactions with other factors such as soil moisture content, temperature, texture, soil bulk density, and mineralogy.
|
CONCLUSIONS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Received for publication September 5, 2006.
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Vadose Zone Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
