SSSAJ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 1 May 2008
Published in Soil Sci Soc Am J 72:758-766 (2008)
DOI: 10.2136/sssaj2007.0170
© 2008 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Kallestad, J. C.
Right arrow Articles by Mexal, J. G.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Kallestad, J. C.
Right arrow Articles by Mexal, J. G.
Agricola
Right arrow Articles by Kallestad, J. C.
Right arrow Articles by Mexal, J. G.
Related Collections
Right arrow Soil Methods/Instrumentation
Right arrow Plant and Soil Interactions
Right arrow Irrigation

SOIL PHYSICS

Comparison of Galvanic and Chemi-Luminescent Sensors for Detecting Soil Air Oxygen in Flood-Irrigated Pecans

J. C. Kallestad, T. W. Sammis* and J. G. Mexal

Department of Plant and Environ. Sci., MSC-3Q, New Mexico State Univ., Las Cruces, NM 88003

* Corresponding author (tsammis{at}nmsu.edu).


   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Low soil O2 levels have been shown to limit growth in pecan [Carya illinoinensis (Wangenh.) K. Koch] seedlings and may limit yield in mature trees. To assess changes in the gas-phase O2 concentration in a pecan orchard soil in response to flood irrigations throughout a growing season, two types of O2 sensor were field tested: a galvanic O2 sensor and a spectrometer-coupled chemical sensor (FOXY sensor). Galvanic sensors, housed in diffusion chambers, were buried at four depths and a datalogger recorded continuous voltage output. The FOXY O2 sensor was utilized as part of a mobile O2 detection system to field analyze gas samples withdrawn periodically from buried diffusion chambers. The FOXY sensor was found to be unstable, however, and difficult to calibrate under conditions of changing temperature and humidity. Laboratory experiments simulating submersion of the galvanic sensor indicated that voltage outputs were comparable to the range observed in the field, but the absence of diurnal concentration fluctuations, typically found in soil measurements, provided a way to discriminate between normal and aberrant output. The responsiveness of the galvanic sensor and its capability to continuously gather hourly data makes it superior to methods dependent on manual sample collection. Galvanic sensors were adequately suited for long-term in situ use in agricultural soil when housed in appropriate diffusion chambers. Higher costs, limited access to diffusion chambers during flood periods, and high variability associated with manually collected data make the FOXY mobile O2 detection system comparatively less optimal for use in agricultural settings.

Abbreviations: FEP, fluorinated ethylene–propylene • FOXY, fiber optic oxygen sensor system • LPSRC, Leyendecker Plant Science Research Center • PVC, polyvinyl chloride


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
The O2 status of soil can be measured by several approaches. Amperometric measurement of the rate of O2 reduction at a polarized Pt electrode in the soil was proposed to be an effective in situ simulation of the O2 diffusion rate through the surface of plant roots covered with water films (Lemon and Erickson, 1952; Kirstensen and Lemon, 1962; Lemon and Wiegand, 1962). Although Pt electrodes have been deployed in soil for an entire growing season using automated instruments (Phene et al., 1976), a number of researchers have identified inherent limitations with this method that might result in inaccuracies and high variability should the electrodes be left in place and used continuously for extended periods (Stolzy and Letey, 1964; Van Doren and Erickson, 1966; Rickman et al., 1968; McIntyre, 1970; Devitt et al., 1989). The most important limitation of this method is that the electrode must be continuously wetted.

Polarographic (Clark-type) O2 electrodes, which can be used in liquid or gas phases, feature the merging of the Pt electrode and reference electrode into a single compact probe. These sensors have been placed directly in buried diffusion wells (Willey and Tanner, 1963; Meek et al., 1980) and have been used as a portable O2 detector for analysis of gas withdrawn from buried diffusion chambers (Patrick, 1977; Carter et al., 1984; Faulkner et al., 1989); however, the probe must be removed and the electrolyte changed periodically.

Utilizing an O2 fuel cell and galvanic cell technologies, a strictly gas-phase O2 sensor was developed by the Japan Storage Battery Company in 1985. The GS-type galvanic O2 sensor (Fig. 1 ), named for the company founder Genzo Shimadzu, features a Pb anode and porous Au cathode in a weak acid electrolyte. Oxygen diffuses through the fluorinated ethylene–propylene (FEP) membrane at a slow but proportional rate to the partial pressure and is reduced at the cathode. The resulting current is converted to voltage output using a temperature-compensating thermistor and an adjusting resistor. The sensor occupies a small volume and consumes little O2. This type of sensor has measured the O2 diffusion rate through soil columns in laboratory settings (Dziejowski et al., 1997; MacKay et al., 1998; Jones et al., 2003), O2 consumption by legume root systems (Witty and Minchin 1998), in situ O2 concentration in termite mounds (Turner, 2001), diffusivity of a geomembrane (Trefry and Patterson, 2001), and subsurface gas in vapor extraction wells during air-injection remediation efforts (Hall et al., 2000). Li and Lundegard (1996) found that field-deployed GS-type sensor readings agreed closely with laboratory soil gas analysis after 1 yr of continuous operation.


Figure 1
View larger version (41K):
[in this window]
[in a new window]

  		Fig. 1. Details of the GS-type (KE50) galvanic O2 sensor and electrochemical reactions.

 
Recently, advances in organic luminescent materials have led to the development of several O2-sensitive light-emitting metal complexes that are relatively stable and can be immobilized in polymers or on silica gel particles in thin coatings (Bacon and Demas, 1987; Hartmann et al., 1995). Placed on microprobes and used in concert with an optical excitation source and spectrometer, systems have been devised to detect fluorescence quenching in proportion to O2 partial pressure in gas and liquid phases. Miniaturization of spectrometers and portable computers have permitted a variety of novel in situ environmental sensing applications for these sensors. Microscale measurements of O2 have been made within woody plants (Gansert et al., 2001; del Hierro et al., 2002; Spicer and Holbrook, 2005), porous media (Holst et al., 1997), and soil (Johnson et al., 2005). These sensors do not consume any O2 and can sample small volumes.

The objectives of the research were: (i) to determine if a low-cost GS-type sensor for continuous monitoring of soil O2 in buried diffusion chambers in a pecan orchard resulted in comparable measurements to a more sophisticated spectrometer-coupled chemical sensor deployed as a field-based detection system; and (ii) determine if soil O2, measured with either system, fell to levels known to cause stress in potted pecan seedlings or mature trees following flood irrigation events.


   MATERIALS AND METHODS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
The FOXY Oxygen Sensor System Construction, Settings, Field Calibration, and Measurement Protocol
A portable O2 detection system, comprised of a spectrometer-coupled chemical sensor system (FOXY, Ocean Optics, Dunedin, FL), an end-to-end syringe device for gas sampling and bladder filling as described by Kanwar et al. (1989), and a laptop computer with FOXY application software was assembled as shown in Fig. 2 . The spectrometer-coupled sensor assembly consisted of a USB2000 spectrometer with attached USB-LS-450 light-emitting diode (LED) light source, a 600-µm-diameter bifurcated optical fiber assembly, a FOXY–R probe, a USB-LS-450-TP16 temperature probe, and an FIA-Z flow cell. The volume of the inner chamber of the flow cell was 70 to 90 µL. Soil air drawn through the flow cell passed the temperature probe and the O2 sensor before entering the end-to-end syringe assembly. The gas sampling tubing was low-diffusion 0.318-cm o.d. by 0.159-cm i.d. FEP Teflon tubing (Supelco, Bellefonte PA). The end-to-end syringe sampling device was made with two 30-mL plastic syringes. The spectrometer and LED excitation light source were powered by the portable laptop computer batteries (Latitude D800, Dell).


Figure 2
View larger version (30K):
[in this window]
[in a new window]

  		Fig. 2. Diagram of the FOXY O2 detection system components.

 

Figure 3
View larger version (21K):
[in this window]
[in a new window]

  		Fig. 3. Details of diffusion chambers housing the galvanic O2 sensor.

 
Soil air samples measured with the FOXY mobile detection system were collected from Type A flow-through-type chambers (Fig. 2), featuring a 0.86- by 0.86-mm-opening stainless steel screen cylinder, 2.68-cm diameter by 14 cm long, with the lengthwise wire edges annealed (Hillside Wire Cloth, Belleville NJ). The chamber also housed a plastic polyethylene teraphthalate balloon bladder (no. 2500600ZAA, Advanced Polymers, Salem NH), which during the withdrawal of soil air was filled with an equal volume of water via the end-to end syringe. A 2.67-cm-diameter polyvinyl chloride (PVC) access pipe, capped at each end with bored rubber stoppers, allowed FEP Teflon tubing and vinyl bladder-fill tubing to pass from the diffusion chamber to the surface. All stopper edges and holes were sealed with silicone sealant. A three-way stopcock was placed on the top end of the FEP tubing to facilitate checking for free water in the diffusion chamber using a third syringe. A two-way stopcock capped the top end of the bladder-fill tubing. To minimize preferential flow of water along the outside of the access pipe, silicone sealant was applied around the pipe in a ring at 15-cm intervals.

The spectrometer was set to integrate intensity counts at the 600-nm analyte wavelength during a 128-ms period with a bandwidth of ±25 pixels. Every four integrations were averaged with a boxcar smoothing value of 8 according to the manufacturer's recommendations. A user-defined timed data acquisition program was used to send averaged intensity values to a log file every 5 s for a 1-min scan period. The spectrometer and probe were calibrated in the field before each use and after 6 or 12 chamber measurements using ambient air, 6.02% O2 in N2, and the 99.999% N2 gas standards. To minimize the difference in temperature between the flow cell and the soil, measurements were made in the early morning when air and soil temperature typically differed no more than 3°C. To further stabilize the flow cell temperature during measurement, gel polar-pack bags (Midland Chemical Co., Midland, MI), kept at 22°C, were placed on top of the flow cell and probe.

In a typical measurement sequence, the estimated gas volume within the access tubing was drawn into the syringe before the scan to ensure that the first measured volume of air within the flow cell was sampled immediately from the chamber. Soil air was then drawn stepwise through the flow cell in 2- to 3-mL increments every 10 to 15 s to allow the temperature of the sampled gas and the flow cell to equilibrate. Typically, 25 to 28 mL of soil air was pulled through the flow cell during 1-min scan periods and then injected back into the chamber. Oxygen concentrations were derived by averaging 18 of the 24 measured intensity values across both scan periods, excluding the first four values of each scan to account for any lag in initial sensor response time. The signal intensity to O2 concentration conversion function was derived from a second-order polynomial regression of standard concentration vs. intensity using Microsoft Excel software.

Galvanic Sensor Calibration, Flood Simulation, and Diffusion Chamber Construction
Ten GS-type galvanic O2 sensors were obtained through Figaro USA (Model KE50, Glenview IL), and a single galvanic O2 sensor was purchased from Apogee Instruments (Model O2S, Logan, UT). The two types of sensors were essentially the same except that the O2S sensor had a smaller millivolt output range and was housed in an aluminum casing. A turned aluminum flow-cell attached to the O2S was removed. The O2S sensor assembly was larger in diameter but diffusion inlet diameter and configuration were the same as the KE50. All sensor leads were soldered to 7.6 m of 22-gauge shielded wire cable to normalize wire-related resistance. Linear calibration functions to convert voltage output to O2 concentration were derived by two-point regression of voltage output vs. O2 concentration in ambient air (assumed to be 20.94%), and in a 99.999% N2 environment (Valley Gas, Las Cruces, NM). Measurements were also made with a subset of sensors using 6.02% O2 in N2 standard (Scott Specialty Gases, Plumsteadville, PA) to verify that the output was linear. The 0.00 and 6.02% O2 calibration gases were directed across the sensor opening at a rate of 95 to 100 mL min–1. Sensor voltage output was measured in the laboratory and at the field sites using a datalogger (Model CR10X, Campbell Scientific, Logan, UT), sampled at 10-s intervals and averaged across a 1-h period.

To assess the KE50 galvanic sensor's range of zero-offset drift in a humidified atmosphere, 99.999% N2 was bubbled through deionized water at a rate of 95 to 100 mL min–1 and directed into a 6-cm-diameter by 10-cm-long plastic cylinder containing the sensor.

Laboratory simulations were conducted to test the performance of the KE50 sensor (or O2S sensor) in four scenarios during a possible submergence event in which soil around the diffusion chamber might become temporarily saturated: (i) air trapped in a head-space diffusion chamber (Fig. 3, Type C diffusion chamber); (ii) air trapped in the sensor's outer lid opening (Fig. 1); (iii) water condensate sealed one to three of the four inner lid holes while air became trapped in the outer lid opening; and (iv) water condensate sealed all of the inner lid holes or the gas-permeable membrane became saturated.

To simulate these scenarios the following experiments were conducted:

  1. The opening of the Type C head-space diffusion chamber with enclosed sensor was submerged 1 to 1.5 cm below water or 1 to 1.5 cm below the surface of water in a beaker of submerged soil with the remaining exposed water surface sealed with wax.
  2. The outer lid opening of the sensor was submerged 0.5 cm below water or 0.5 cm below the surface of water in a beaker of submerged soil with the remaining exposed water surface sealed with wax.
  3. Five microliters of water was placed in one, two, or three of the inner lid openings but not touching the membrane. The sensor was then placed over water with the outer lid opening submerged 0.5 cm below the surface.
  4. Water was placed in all of the holes of the inner lid and the sensor then placed with the outer lid opening submerged 0.5 cm below water, or a strip of wetted filter paper was placed next to the inner lid openings, held in place with water-conducting sponge material, then the sensor was placed over a beaker of water with the opening submerged 0.5 cm below the surface.

Diffusion Chamber Design and Installation
Three configurations of buried chambers were used to house the galvanic O2 sensors (Fig. 3) with dimensions listed in Table 1 . The Type B and C chambers housed the KE50 sensor and the larger Type D chamber housed the O2S sensor because of the increased sensor assembly diameter.


View this table:
[in this window]
[in a new window]

  		Table 1. Diffusion chamber details and the effect of chamber geometry on the rate of concentration change.

 
Flow-through chambers (Types B and D) featured a 1.18-mm-opening stainless steel screen shaped into a cylinder. The head-space chamber (Type C) was similarly constructed except a 6-cm length of 4.26-cm-diameter PVC pipe was glued to the threaded coupler instead of screen, creating a sealed inverted cup. The Type C chambers were installed at shallow depths in anticipation of the possibility of free water percolating into the augured and backfilled hole surrounding the chamber. During such an event, measured O2 concentrations represent the O2 trapped in the chamber at the time of submergence and any O2 that diffuses from the water into the open end of the chamber. The sensor opening was positioned 2.6 cm above the bottom edge of the cup. The Type D chamber also featured the FEP Teflon access tube with a two-way stopcock so that O2 levels could also be measured with the FOXY system. The balloon bladder was not included in the Type D chamber due to size restrictions.

All of the chambers were placed in 6.8 ± 0.2-cm-diameter auger holes, except the Type A diffusion chamber, which was installed in a 4.2 ± 0.1-cm-diameter auger hole extending approximately 22 cm below a larger 6.8-cm-diameter hole as shown in Fig. 2. To minimize gas diffusion impedance between the chamber void and the walls of the augered hole, coarse sand backfill, sieved to >1-mm but <2-mm diameter was placed around the chambers to a height just above the screen portion of the chamber assembly and extended 2 to 3 cm below to facilitate drainage. Excavated soil was tamped back into the hole above the sand. A 17-cm-diameter depression was made within the top 4 to 5 cm of the soil surface around each access pipe and filled with hydrated bentonite (Black Hills Bentonite Co., Casper, WY) to further prevent preferential flow of air and water around the access pipe. Plastic sheeting (17-cm diameter) was applied over the bentonite to minimize dehydration, and buried below 1 cm of topsoil. The augured hole for the Type D chamber was the same diameter as the Type B and C chambers; however, the volume of course sand backfill around the Type D chamber was less due to its larger diameter, which might have affected the lateral diffusion rate.

Study Sites
The Type A gas sampling chambers and the KE50 and O2S galvanic cell sensors were installed at two locations: a 5.1-ha private orchard located 7 km south of Las Cruces, NM, and at the New Mexico State University Leyendecker Plant Science Research Center (LPSRC) orchard 14.5 km south of Las Cruces. At the private orchard, 12 gas sampling chambers and four galvanic sensors were installed at 25-, 50-, 75-, and 100-cm depths. The galvanic sensors were arranged between four trees, 4.7 to 6.3 m from the nearest tree. The gas sampling chambers were also located among the same four trees adjacent to the galvanic sensors but placed into three spatial zones with respect to the tree position: Zone 1, 1.4 to 2.0 m from the nearest tree; Zone 2, 3.6 to 4.3 m from the nearest tree; and Zone 3, 5.9 to 6.0 m from the nearest tree. The soil, classified as a coarse silty, mixed (calcareous), thermic Typic Torrifluvent of the Harkey series (Bulloch and Neher, 1980) was comprised of loam in the top 60 cm (21% clay, 36–44% silt, and 43–34% sand), with layers of fine sandy loam and clay stratified below. The orchard was flood irrigated approximately every 9 d during the summer, and the soil surface appeared to be saturated for approximately 21 to 24 h following irrigation.

At the LPSRC site, three galvanic O2 sensors in Type B and C diffusion chambers were installed at 25-, 50-, and 75-cm depths, spaced 1.5 m apart and 3.75 m from the tree row, and the O2S sensor in a Type D diffusion chamber was installed at a depth of 50 cm. Soil at this site was classified as a clayey over loamy, montmorillonitic (calcareous), thermic Typic Torrifluvent of the Belen series (Bulloch and Neher, 1980). The top 90 cm was clay or silty clay (45–60% clay, 35–54% silt, and 1–17% sand). The orchard was flood irrigated every 22 to 31 d, and the soil surface appeared to be saturated for approximately 4 d following irrigation.

Soil temperature probes (T107, Campbell Scientific, Logan, UT) were installed at both sites at 25- and 50-cm depths. Hourly averaged values of both the galvanic O2 sensors and soil temperature sensors were recorded with a datalogger (Model CR10X, Campbell Scientific).


   RESULTS AND DISCUSSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Offset Voltage Drift and Its Effect on Galvanic Sensor Calibration
The galvanic cell O2 sensors produce a voltage output in linear proportion to O2 diffusing through an O2-permeable fluororesin membrane to a Au cathode where it is reduced. When calibrating the sensor at 0% O2 by passing N2 gas over the sensor, the sensor puts out an additional background current that can be partially compensated for by subtracting an offset voltage. Some of this background current is a result of small amounts of O2 in the ambient air diffusing through the acrylonitrile butadiene styrene (ABS) plastic of the sensor body into the electrolyte. In laboratory calibrations, the maximum voltage output of 11 KE50 sensors in ambient air ranged from 48 to 51 mV (average 49.2 ± 1.02 mV). When 99.999% N2 was directed across the sensor opening at 95 mL min–1, the voltage output was about 2.5 mV after 15 min of exposure. When the sensor was enclosed in a chamber exposed to 99.999% N2 bubbled through water for a period 12 h, the minimum voltage output was 1.93 mV, and after 124 h the voltage output was 0.70 mV. When the sensors were placed in diffusion chambers in soil flooded for up to 5 wk, the minimum voltage outputs averaged 0.25 ± 0.18 mV (n = 5). The impact of a maximum possible zero-offset voltage drift of 1.2 mV on the reported O2 concentration in our field measurements could result in a maximum error of about 0.7% O2 in the low concentration range of 0 to 2% O2.This is outside the ±2% full-range linearity accuracy for voltage reported by the manufacturer.

Effects of Trapped Air and Condensation on Galvanic Sensor Output
Sensors housed in flow-through-type diffusion chambers were potentially susceptible to brief periods of inundation by flood-irrigation water, especially at shallower depths and, in high shrink–swell clay soil, through cracks. In laboratory simulations of such events, when either the sensor opening was submerged in a small volume of water or the sensor housed in the head-space type diffusion chamber (Type C) was submerged in water, the decline in voltage output over a 160-h test period was slow (–0.021 ± 0.007 mV h–1, n = 4) (Fig. 4 ). In contrast, when water was applied to all of the inner lid holes, or water was applied directly to the sensor membrane, the decline in voltage output during the submergence period was rapid (–4.63 ± 3.13 mV h–1, n = 7). When only one, two, or three of the inner lid holes were filled with water, the initial rate of voltage decline in a 15-h period was intermediate (–0.405 ± 0.064 mV h –1) but stabilized thereafter to a decline rate of 0.03 mV h–1. Averaged across a 120-h period, the decline was about –0.096 mV h–1 (Table 2 ). Since the galvanic sensor is capable of responding to the full range of O2 concentration within 90 s, these results suggest that when the sensors are deployed in the field, a rapid decline in measured O2 levels could be false, caused by water condensate or water saturating the sensor membrane. Likewise, an unusually slow decline in measured soil O2 levels may not be indicative of the state of soil aeration. Rather, trapped air in the outer lid of the sensor or within a headspace (Type C) diffusion chamber may decrease at a slow rate due to the consumption of O2 by the sensor (approximately 18 nmol h–1 at 20°C) and diffusion of O2 in or out of the trapped air space through a water film. Since both of these conditions fall within the normal sensor response range, we need a method to discriminate between aberrant and normal readings when the sensor is below ground.


Figure 4
View larger version (16K):
[in this window]
[in a new window]

  		Fig. 4. Laboratory simulations of the effect of submergence and condensate formation on the performance of the KE50 galvanic sensor.

 

View this table:
[in this window]
[in a new window]

  		Table 2. The effects of simulated submergence and water condensate formation on the galvanic sensor voltage output.

 
When sensors were installed in the field, we found that the diurnal cyclic O2 concentration due to root respiration was always present and most pronounced at depths of 75 cm and above (Fig. 5 ). Absence of these fluctuations could be considered indicative of an aberrant sensor response. In addition, the observed change in sensor output in response to either irrigation events or daily respiration never exceeded –2.308 mV h–1 (Table 3 ), indicating that direct impedance by water condensate or membrane saturation might not have been a chronic problem.


Figure 5
View larger version (15K):
[in this window]
[in a new window]

  		Fig. 5. Oxygen concentration measured at the private orchard with galvanic sensors (solid line) in buried chambers at 25, 50, 75, and 100 cm. The FOXY O2 sensors readings were taken from chambers located 1.4 to 2.0 m (open circle), 3.6 to 4.3 m (filled square), and 5.9 to 6 m (shaded triangle) from the nearest tree at the respective depths.

 

View this table:
[in this window]
[in a new window]

  		Table 3. Maximum decrease in galvanic sensor voltage output averaged across any 4-h period in response to flood irrigation at sensor depths of 25, 50, 75, and 100 cm.

 
Environmental Effects on FOXY Oxygen Sensor Readings
FOXY fiber-optic O2 sensors are sensitive to temperature, which can affect the energy characteristics of the dye complex, the matrix in which it is embedded, the diffusion rate of O2 through the probe coatings, and the partial pressure of O2 in the gaseous sample. Since the manufacturer's temperature compensation program produced unreliable concentration values for gas standards in the field, the sensor was calibrated at field air temperature before and after each series of measurements. Ambient air temperature during the measurements changed an average of 2.55 ± 1.08°C (n = 31), but flow cell temperature during all measurements changed an average of 0.86 ± 0.42°C (n = 31). In the pecan orchard, the temperature differential between soil (25-cm depth) and ambient air averaged 2.07 ± 1.57°C (n = 31). Based on a sensitivity analysis of the effects of temperature on analyte intensity, a change in temperature of the calibration standard during the measurement period could result in an O2 concentration error of ±1.2%, and the flow cell temperature variation could result in an O2 concentration error of ±0.4% if a single calibration function was used.

These temperature-related errors were minimized by taking measurements early in the morning when the difference in soil and air temperatures was the smallest, and creating calibration curves both before and after the measurement sequence. The first measured O2 values were based on the beginning calibration curve, and the final O2 levels in the measurement sequence were determined using the ending calibration curve. If the difference between the beginning and ending calibration curves was large, an averaged calibration curve was created by linearly interpolating the beginning and ending calibration values and was used for intensity values taken during the middle of the measurement sequence.

Other possible sources of variation were more difficult to control. The increase in humidity from dry gas standards to nearly saturated soil air samples may have changed the permeability of the probe coating such that soil air sampled later in the sequence may have diffused through the probe coating more slowly. The increased humidity also may have affected the ending calibration standard readings. In addition, the 470-nm excitation light intensity appeared to vary daily during a measurement sequence by as much as 6%, even though the computer battery was charged every day. Dust particles could typically be found on the probe tip after a series of measurements, which may have altered the optical characteristics of the probe. With time, "photo-bleaching" of the dye complex resulted in a 0.1% decrease in intensity per hour of continuous use, which is why the sensor should be replaced each year. All the possible errors of the field calibration of the FOXY system require that the user be meticulous in the calibration procedure to get reliable measurements from the equipment.

Agreement between the Galvanic Sensor and the FOXY Sensor Readings
Comparable values were obtained when the O2 concentrations were measured within a single buried diffusion chamber (Type D) utilizing both sensor systems (Fig. 6 ). Since the galvanic sensor voltage recorded by the datalogger is an hourly average of readings taken every 10 s, and the FOXY reading is relatively instantaneous, a statistical comparison of means was not feasible. Among 11 data points during a 23-d measurement period, there was a mean absolute concentration difference of 1.16 ± 0.62% O2 between the two systems.


Figure 6
View larger version (10K):
[in this window]
[in a new window]

  		Fig. 6. Comparison of O2 concentration values measured with the FOXY system and a galvanic sensor within the same diffusion chamber.

 
The most important difference between these two sensor systems is that the quantity and frequency of measurements generated with the galvanic sensors is impossible to obtain with any manual sampling method. While both systems were capable of tracking concentration in response to flooding and depth, continuous data from the galvanic sensors also indicated bimodal concentration fluctuations on a diurnal time scale superimposed on a longer term pattern of O2 depletion and recharge in response to irrigation. This phenomenon was seen prominently at the shallower depths (Fig. 5). The continuous readings of the galvanic sensor were able to detect an O2 level as low as 6% in the major root zone of the pecan trees at a depth of 50 cm, whereas the measurements of the FOXY O2 sensors missed the low reading because of infrequent sampling.

Measured Soil Aeration in Response to Flood Irrigation
Studies have shown widely variable plant survival, growth, and gas exchange responses to periods of root hypoxia or anoxia. Boynton et al. (1938) noted that while apple (Malus domestica Borkh.) root growth decreased when the O2 concentration fell below 10%, roots were able to survive for months at a "subsistence level" when the concentration was maintained at 3%. Furr and Aldrich (1943) reported that date palm (Phoenix dactylifera L.) could withstand 3 wk of continuous soil O2 concentrations below 2% with no measurable injury. Smith et al. (1989) demonstrated that growth and C assimilation in pecan seedlings could be significantly reduced when maintained in soil O2 concentrations as high as 13.5% during a period of 2 wk. In contrast, Huck (1970) reported that 3 h of anaerobiosis could kill the primary taproots of cotton (Gossypium spp.), and Meek et al. (1980) reported alfalfa (Medicago sativa L.) stand losses when soil O2 was 6% or less for >24 h. So, the duration of exposure at a particular concentration may be a more meaningful correlate to plant stress, not concentration by itself. Because of the cost of taking frequent manual measurements, optical O2 sensors such as the FOXY system are not suitable for measuring plant response to O2 stress except for long-term studies under known low-soil-aeration conditions. Also, it is impossible to make measurements with the optical O2 sensor when water is standing on the surface of the field.

The galvanic sensor data indicated a rapid increase in O2 concentration at the 100-cm depth immediately following irrigation (Fig. 5). This is thought to result from O2-enriched soil air from shallower depths being pushed down the profile under hydraulic pressure. According to the manufacturer's specifications, voltage output will increase with increasing barometric pressure at a rate of 0.37 mV kPa–1. An application of 10.2 cm of irrigation water would result in a maximum possible pressure-induced voltage increase of 0.36 mV. This would translate into an increase of 0.15% O2, which may have represented (at most) a minor contribution to the observed increase of 0.44 to 2.52% O2 following the irrigations. Again, this type of response cannot be measured by the FOXY O2 sensors.

The FOXY O2 sensor readings generally followed the depletion and recharge trends recorded by the galvanic sensors (Fig. 5). Differences between the values generated by the single galvanic sensor and the three FOXY O2 sensors at each depth resulted from spatial heterogeneity of O2 concentration in the soil. The increasing variability in concentration correlated with increasing depth, which was probably attributable to the variability in texture and moisture content in the overlying layers. Combining galvanic sensor and FOXY O2 sensor values at each time point and depth, there was an average coefficient of variation of 9.45% at 25 cm, 14.16% at 50 cm, 16.32% at 75 cm, and 30.27% at the 100-cm depth. Consequently, more sensors are needed at lower depths if accurate measurements of O2 levels at those depths are needed to characterize the plant response to O2 stress; however, most roots of plants are in the top 50 cm of a soil profile where spatial variability is the lowest. The increased variability with depth also may be due to the relative concentration of roots. If the sensor is near a root, the O2 level may be less compared with a sensor located some distance from a root.

The absolute difference in measured concentration in the top 25 cm between the FOXY system and the galvanic system when the sensors were in different holes was 2.26 ± 0.62% O2, compared with 1.16 ± 0.62% O2 when the sensors were in the same hole. Consequently, at 25-cm depth, about half of the difference between absolute readings of the two sensor systems could be attributed to the sensor variability and calibration error and the other half of the difference could be attributed to spatial variability. As the depth increases, more of the difference can be attributed to spatial variability.

At the end of the season, when the diffusion chambers were excavated, there was evidence that the flow-through-type chambers had been partially flooded. A film of colloidal material had been deposited on the balloon bladders of most of the Type A chambers and on some of the galvanic sensors in the Type B chambers. Small roots were found growing into, around, and through the screened chambers. There was little evidence from the data, however, that suggested that the galvanic sensor openings were blocked or that membrane permeability had been compromised for extended periods.

Four to 12 h following irrigation, the galvanic sensor recorded a decrease in voltage in the range of –3.9 to –4.2 mV h–1 at the 50-cm depth (Fig. 7A ). Slow rates of change, in the range of –0.026 mV h–1 or slower, were observed from the sensor at the 75-cm depth in clayey soil at the LPSRC site after a normal irrigation (Fig. 7B). This high rate of change at the 50-cm depth was similar to the rate of change that occurred when the KE50 inner lid holes were plugged or the sensor membrane was saturated during the simulations. The voltage and O2 levels never approached zero, however, and the rise of the O2 level within 24 h indicated that the rate of drop was probably due to plant root respiration in a decreasing volume of soil atmosphere as irrigation water filled the soil voids. The slower rate of fall at the 75-cm depth suggests that fewer roots existed at that depth and recharge of O2 to that depth after an irrigation was slower than at the 50-cm depth. While this slow rate was similar to the rate observed in the trapped-air simulations, the presence of diurnal fluctuations in concentration suggest that the sensor was functioning normally.


Figure 7
View larger version (12K):
[in this window]
[in a new window]

  		Fig. 7. Voltage output after a normal irrigation in the private orchard: (A) rapid decline (–3.9 mV h–1) at 50-cm depth 12 h after irrigation; (B) a slow decline (–0.026 mV h–1) at 75-cm depth 144 h after irrigation.

 
Before the growing season began, galvanic sensors buried at the 25-cm depth in the private orchard were installed in Type B diffusion chambers. After a heavy rainfall, one of the sensors failed to function correctly after being flooded. Subsequently, all galvanic sensors buried at 25 cm were replaced in Type C chambers. The function of the flooded sensor was restored after it dried out in the laboratory, but its calibration value at 20.9% O2 had decreased approximately 8 mV, which may have been caused by the sensor electrolyte being contaminated by soil solution moving through the membrane. It is important, therefore, that sensors installed in a zone of soil saturation be protected from submergence.

When O2 measurements are determined using the galvanic sensor, the time that O2 levels are below a putative threshold level in the soil can be measured and correlated with measurement of the photosynthesis rate to determine the effect of O2 levels on plant growth and yield. In a related study, our results showed that O2 levels needed to be near 0% for >5 d at the 50-cm depth before photosynthesis started to decrease (Kallestad et al., 2007). In this study, the O2 levels at both sites never reached this extent or duration of depletion during a normal irrigation. In addition, the continuous data recorded by galvanic O2 sensors facilitated a time series correlation with changes in soil temperature and moisture content (Kallestad et al., 2008). Savage and Davidson (2002) pointed out that for soil seasonal CO2 flux estimates, manual individual measurements may give the same answer as continuous measurements but short-term changes measured with the automated system permitted better empirical modeling of the effects of soil temperature and moisture on soil respiration than could be achieved with the manual sampling system. When measuring the onset of acute O2 stress in orchard trees, the time sequence of depletion is more important than average seasonal values.

Four GS-type galvanic O2 sensors, a data logger, and battery is approximately one-third the cost of the FOXY system with annual probe replacement and calibration supplies. Both systems require a computer to download and analyze data. Considering the additional costs incurred in labor, time, and transportation for doing field calibrations and manually collecting data, the FOXY system is less cost effective and provides a significantly smaller volume of information with greater inherent variability.


   CONCLUSIONS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
The FOXY O2 sensor is relatively expensive and unstable to be mounted permanently within a chamber for even periodic measurements. Photo-bleaching of the light-emitting complex from continuous use shortens its life span, making long-term measurements infeasible. The probes are easily compromised by dust or water, and the instrument is difficult to calibrate under conditions of changing temperature and humidity. Even in applications where soil O2 needs to be sampled only periodically at widely dispersed sites, the time and effort spent in calibration precludes this instrument from being efficient even as a mobile detection system.

Galvanic O2 sensors are a promising innovation for continuous in situ monitoring of agricultural soil aeration. Connected to a datalogger, the sensor's capability for continuous output facilitates measurements of small-scale diurnal changes in concentration that are logistically difficult to detect by manually collecting data. These characteristic fluctuations also enable the user to detect abnormal sensor output that may result from flooding.

The design considerations for buried diffusion chambers housing the galvanic sensors need to balance the trade-offs between minimizing overall volume, providing ample diffusion surface area, ensuring specificity with respect to depth in the soil profile, protecting the sensor from water fouling, and providing a means of temporarily removing the sensor from the chamber for recalibration after several months of operation without disturbing the sampling environment.


   ACKNOWLEDGMENTS
 
This research was supported by New Mexico University Agricultural Experimentation Station, ARS Pecan Initiative and the USDA Rio Grande Basin Initiative


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Permission for printing and for reprinting the material contained herein has been obtained by the publisher.

Received for publication May 8, 2007.


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Kallestad, J. C.
Right arrow Articles by Mexal, J. G.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Kallestad, J. C.
Right arrow Articles by Mexal, J. G.
Agricola
Right arrow Articles by Kallestad, J. C.
Right arrow Articles by Mexal, J. G.
Related Collections
Right arrow Soil Methods/Instrumentation
Right arrow Plant and Soil Interactions
Right arrow Irrigation

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