On the Construction and Calibration of Dual-Probe Heat Capacity Sensors

Transcription

1 On the Construction and Calibration of Dual-Probe Heat Capacity Sensors J. M. Ham* and E. J. Benson ABSTRACT parallel hypodermic-tubing probes extending from a Dual-probe heat-capacity (DPHC) sensors can be used to measure small sensor body (Fig. 1). Hereafter, probe will be used soil heat capacity (C), water content, and temperature. Research to describe the hypodermic tubing components while was conducted to test design factors that affect sensor calibration, sensor will refer to the entire assembly. One probe con- including: (i) calibration media, (ii) diameter and length of the needle tains a heater made of resistance wire and the other probes, (iii) sensor body material, and (vi) duration and total power probe houses a thermistor. Once installed, the heat caof the applied heat pulse. All sensors were calibrated in media with pacity of the soil surrounding the heater is determined known C, including: agar (water), water-saturated glass beads, and by applying a short heat pulse (e.g., 8 s) and then meadry glass beads. Calibration consisted of collecting heat pulse data in suring the subsequent temperature increase at the thera given media and then calculating the apparent probe spacing (r app, mometer probe for about 60 s. As shown by Campbell distance between heater and detector needles) that yielded correct value of C. An ideal sensor would have the same r app regardless of et al. (1991), the heat capacity of the soil can be esti- media type. The r app for all sensor designs increased as C decreased, mated as on average changing by 6.7% between agar and dry beads. This undesirable result was consistent with previous studies that showed DPHC C q/( er 2 T m ) [1] sensors calibrated in agar overestimated C in drier soils. Needle diame- where C is heat capacity (J m 3 K 1 ), q is applied power ter (1.27 vs mm), sensor body material (urethane vs. high-conduc- per length of probe (J m 1 ), r is the distance between tivity epoxy), and shortening of the detector probe had a small effect the hypodermic probes (m), and T m is the maximum on r app. Sensors made with urethane bodies, 1.27-mm diam. needle increase in temperature at the detector probe following probes, and shortened temperature probes showed less sensitivity to the application of the heat pulse ( C). Equation [1] is calibration media and are therefore recommended. The r app for this design only increased by 2.6% between dry and water-saturated wet a solution to the heat equation assuming an infinite line- beads. Apparent probe spacing was not affected by changes in total source heater and an instantaneous heat pulse. Knight applied power ( J m 1 ) or heat pulse duration (2 16 s) when and Kluitenberg (2004) presented an improved heat the correct analytical model was used to compute C. capacity equation that considers the time interval of heating D C q e r 2 T m 1 ε ε 1 3 ε 8 5 7ε ual-probe heat capacity sensors are a promising 2 3 heat-pulse technology for measuring soil thermal [2] properties, soil water content, and soil temperature. Soil where ε is t o /t m, t o is the duration of the heat pulse (s), heat capacity is determined by applying a heat pulse to and t m is the time from the initiation of heating to the a line source and measuring the temperature increase occurrence of the maximum temperature rise (s). about 6 mm from the heater (Campbell et al., 1991). Calibration of a DPHC sensor requires precise deter- Due to their small size and sample volume, DPHC sen- mination of r, the only empirical sensor-specific paramesors are particularly useful near the soil surface. A theo- ter in the calculation of C. Probe spacing is estimated retical analysis by Kluitenberg and Philip (1999) showed by collecting heat pulse data in a media of known C DPHC sensors could be placed within 10 mm of the and then rearranging Eq. [1] or [2] to solve for apparent surface. Sensors are relatively inexpensive to build and probe spacing, hereafter noted as r app. The calculation can be readily automated using data acquisition systems of C is very sensitive to r; for example, a 2% error in r to obtain near-continuous data. For example, DPHC causes a 4% error in C (Campbell et al., 1991). Theresensors are ideal for monitoring heat storage in the soil fore, for a sensor with 6-mm probe spacing, r app must layer above flux plates when measuring soil heat flux be determined to within 0.3 mm to measure C to within by the combination method (Ham, 2001). Because C is 10%. In theory, r app should not change if a sensor is dependent on soil water content, DPHC sensors also calibrated in media with different heat capacities. Howcan be used to monitor soil moisture (Bristow et al., ever, at lower water contents, data suggest r app increases 1993; Tarara and Ham, 1997; Basinger et al., 2003; Heit- when C decreases which causes a progressively increasman et al., 2003; Ochsner et al., 2003) and plant water ing overestimate of C and soil water content as the soil use (Song et al., 1998). The DPHC sensor, as explained here, consists of two Dep. of Agronomy, Kansas State University, Manhattan, KS Contribution no J from the Kansas Agric. Exp. Stn., Manhat- tan, KS. Received 20 Nov *Corresponding author: (jayham@ ksu.edu). dries (Tarara and Ham, 1997; Song et al., 1998; Basinger et al., 2003). This response indicates model errors (i.e., Eq. [1] or [2] do not accurately represent sensor physics) or instrumentation errors (i.e., inaccurate measurement of q, T m,ort m ) are changing as a function of water content. An added concern is that r may be altered during installation in the soil (probe deflection). It might be Published in Soil Sci. Soc. Am. J. 68: (2004). Soil Science Society of America Abbreviations: AWG, American Wire Gauge; C, soil heat capacity; 677 S. Segoe Rd., Madison, WI USA DPHC, dual probe heat capacity. 1185

2 1186 SOIL SCI. SOC. AM. J., VOL. 68, JULY AUGUST 2004 advantageous to use larger diameter or shorter needle bodies (32 mm long, 13 mm diam.) were cast using either of probes that are more rigid (Kluitenberg et al., 1993). two epoxies: RBC4300 (RBC Industries, Inc., Warwick, RI) The objectives of our research were to calibrate DPHC or CR600 (MicroMark, Berkeley Heights, NJ). Thermal propsensors of various designs by measuring r erties of the hardened epoxies were measured using DPHC app in media sensors. RBC4300 is a thermally conductive, black epoxy with with different heat capacities and water contents. Sena thermal conductivity of 0.6 W m 1 K 1 and a heat capacity sors having a better design and operated in the most of 2.9 MJ m 3 K 1, whereas CR600 is a nonconductive white optimal manner should have a more similar r app across urethane epoxy with a thermal conductivity of 0.2 W m 1 K 1 all media. Tests were conducted to examine the effect and a heat capacity of 1.7 MJ m 3 K 1. Thermal conductivity of: (i) calibration media, (ii) needle-probe diameter and of the RBC4300 was about half the value stated by the manufacturer. length, (iii) sensor body material, and (iv) duration and Epoxies with two different thermal conductivities were total power of the applied heat pulse. Also provided is a chosen to determine if heat transfer in or near the sensor body description of fabrication methods suitable for building might affect results. Of the two casting materials tested, the large numbers of DPHC sensors. CR600 urethane was easier to use because the hardener and epoxy mixture was easier to prepare, cured rapidly at room temperature (e.g., 75 min), and was easier to remove from MATERIALS AND METHODS the mold. The basic template for the DPHC sensors was a modified All sensors were fabricated in the same manner but certain version of the design proposed by Campbell et al. (1991) and aspects of the design were modified. As mentioned earlier, similar to that used by Tarara and Ham (1997) and Basinger sensors were built with two different hypodermic probe diame- et al., (2003). Each sensor consisted of two needle probes, ters, two different body materials, and two different temperature made from or 1.65-mm diam. (16 or 18 American wire probe lengths. Sensor designs used for testing included: gauge [AWG]) hypodermic tubing, mounted in a cylindrical (i) urethane body with 1.27-mm diam. probes, (ii) urethane body body (Fig. 1). Hypodermic tubing was custom cut to 36 mm, with 1.65-mm diam. probes, and (iii) RBC4300 body with deburred, and flanged on one end (Small Parts Inc., Miami 1.27-mm diam. probes. After all testing had been completed Lakes, FL). The temperature-sensing probe contained a thermistor with the original sensors, the length of the temperature probes (10K3MCD1, Betatherm Corp., Shrewsbury, MA) po- was shortened to 18 mm by removing the tips with a precision sitioned halfway between the sensor body and the tip of the cutting wheel, leaving the thermistor 4 mm from end (Fig. 1). tubing. Heater probes were made from two loops (four strands Heater probes remained 28 mm long. All tests were then repeated total) of enameled resistance wire (205 ohm m 1, Nichrome with the modified probes. In all, six different sensor 80, Pelican Wire Co., Naples, FL) resulting in an overall heater configurations were tested. In addition to changing sensor resistance of 820 ohms m 1. The resistance of the heater wire design, sensors were calibrated using different levels of q (400 was measured in the laboratory and was slightly different from 1600 J m 1 ) and different heat pulse lengths (2 16 s). Duration the manufacturer s rating. Once the thermistors and heater of the heat pulses was verified by measuring the heater voltage wires were in place, the hypodermic tubing was filled with at 10 Hz using a CR23X datalogger (Campbell Scientific, Logan, high-conductivity epoxy (Omegabond 101, Omega Engineering, UT). One special sensor was constructed (1.27-mm diam., Stamford, CT). Epoxy was loaded into a syringe and 18-mm long temperature probe, urethane body) that had finewire injected into the bore of the hypodermic tubing using a flexible thermocouples (Type-T, 0.08-mm diam.) both inside and tube. Wires exiting the back of the heater and thermistor probes cemented to the surface of the heater probe 14 mm from the were spliced to an extension cable (9L28024-H100-8, Belden sensor body. Thermocouples also were attached to the surface Wire and Cable, Richmond, IN), four conductors to the ther- of the temperature probe at 4 and 14 mm from the sensor body. mistor and two conductors to the heater (Fig. 1). Splices from Sensors were calibrated by collecting heat pulse data in me- the heater wire and thermistor to the copper extension wire dia of known heat capacities and calculating r app using Eq. [1] were made using 7% silver solder (Kapp Alloy and Wire Co., or [2]. For conditions used in this study, sample calculations Oil City, PA). showed that the difference between Eq. [1] and [2] is 0.5% Once the heater- and temperature-probe assemblies were except for very long pulse lengths (e.g., 16 s). Thus, Eq. [1] completed, they were then clamped into a custom-built stain- was used for all calculations except when Eq. [1] and [2] were less steel mold, which was used to hold the needles 6 mm apart explicitly compared to study the effect of heat pulse length. and parallel. Two different molds were used for all fabrication, Media included agar, water-saturated glass beads, and dry one each for the and 1.67-mm probe diameters. A release glass beads with heat capacities of 4.18, 2.82, and 1.23 MJ m 3 agent (MS122DF, Miller-Stephenson Chemical Co., Danbury, K 1, respectively. The heat capacity of agar (6 g L 1 ) was assumed CT) was sprayed on the molds before use. Cylindrical sensor equal to that of water. The glass beads (0.43- to 0.60-mm Fig. 1. Diagram of a dual-probe heat capacity sensor. For some sensors, the length of the temperature probe was reduced to 18 mm.

3 HAM & BENSON: DUAL-PROBE HEAT CAPACITY SENSORS 1187 Table 1. Mean apparent probe spacing (r app ) of different sensor designs in three calibration media. Also included are measurements of probe spacing as determined with a digital caliper. 28-mm temperature probes Sensor design 18-mm temperature probes Media or method A B C Avg. A B C Avg. mm Agar Wet beads Dry beads Caliper Designs tested included sensors with 28- and 18-mm long temperature probes with the following configuration of probe diameters and body types: (A) 1.27-mm diam. probes with urethane bodies, (B) 1.67-mm diam. probes with urethane bodies, and (C) 1.27-mm diam. probes with high thermal conductivity bodies. Values reported are means of four sensors, for each sensor design the average SD 0.02 mm, except for caliper values, which are the average of two measurements. sensor body distorts the temperature regime around the temperature probe, a process not considered by the infi- nite line source model. Data from the sensor equipped with thermocouples on the probe exterior suggests the sensor body is a heat sink in dry beads (Fig. 3). In dry beads, there was a 0.25 C temperature difference be- tween the thermocouples at 4 and 14 mm at the time of maximum temperature increase. Conversely, there is almost no longitudinal gradient along the temperature diam., Agsco Corp., Wheeling, IL) had a specific heat of 794 J kg 1 K 1 at 23 C as determined by differential scanning calorimetry at the Thermal Physical Properties Research Laboratory, Inc., West Lafayette, IN. Heat capacities of the wet and dry beads were calculated on a volumetric basis in the manner used by DeVries (1963). Bulk density of the glass bead media was found to be 1530 kg m 3 (Basinger, 1999). Calibration data were collected in the laboratory using a CR10X data logger and two AM416 multiplexers (Campbell Scientific, Logan, UT). Current flowing through the heater circuit was measured with a four-leg shunt resistor (SM155-4, 1 ohm, 0.1%, Precision Resistor Co., Largo, FL). After a heat pulse, thermistors were sampled at 2 or 4 Hz for 75 s using a four-wire half bridge with a 0.1%, 5000-ohm reference resistor. The thermistor circuit provided a resolution of C, which was about 10 times better than that achieved with a Type-T thermocouple. Once sensors were installed in a given media and the data collection system configured for specific heat pulse length and q, measurements were collected hourly for 1 to 2 d. The last 20 heatpulse measurements from each test were used for analysis. In addition to the indirect determinations of r app, the actual spacing between the heater and detector probes (center to center) for each sensor was measured with a digital caliper. RESULTS AND DISCUSSION Calibration Media Regardless of sensor design, r app increased significantly as C of the calibration media decreased (Table 1, Fig. 2). On average, r app for the full-length probes increased from 5.70 mm in agar to 6.08 mm in dry beads, a 6.7% increase. These results are consistent with field and laboratory analyses that indicate DPHC sensors calibrated in agar tend to overestimate C and soil water content under dry conditions (Tarara and Ham, 1997; Song et al., 1998; Basinger et al., 2003). Of all sensor designs tested, the urethane bodies with shortened 1.27-mm diam. temperature probes showed the least sensitivity to calibration media. Changing from agar to dry beads with this sensor resulted in a 4.4% increase in r app (Table 1). Overall, r app in wet beads was closest to the caliper measurements. The dependence of r app on C indicates there was some form of error that changed with media thermal properties. Most likely, there are measurement errors in T m, or Eq. [1] and [2] are not accounting for some form of heat flow affecting T m (i.e., model error). Greater con- tact resistance between temperature probe and media in the dry beads may cause an underestimation of T m. Another possible problem is that heat flow into the Fig. 2. Effect of calibration media on the estimate of r app. Shown are data from sensors with (a) full-length, 28 mm, temperature probes and (b) shortened, 18 mm, temperature probes. Comparisons are made among sensors constructed from and 1.67-mm diam. stainless steel needles mounted in bodies fabricated from high thermal conductivity (HTC) RBC4300 epoxy and low thermal con- ductivity (LTC) urethane epoxy. Different letters indicate r app is statistically different within a given media type (P 0.10).

4 1188 SOIL SCI. SOC. AM. J., VOL. 68, JULY AUGUST 2004 Fig. 3. Temperature increase at the temperature probe following an 8-s, 858-J m 1 heat pulse in water-saturated and dry glass beads. Data are from an experimental sensor (1.27-mm probes, urethane body) with fine-wire thermocouples attached 4 and 14 mm from the sensor body. Arrows depict the time of maximum temperature increase. probe in wet beads. Initially, the different response in wet and dry beads was attributed to the magnitude of the temperature increase itself (1.2 vs. 2.4 C). However, as will be shown later, changing the applied power and the magnitude of T m had no effect on the estimate of r app. The sensor body was probably absorbing a larger fraction of the heat pulse in the dry case because the Fig. 4. Effect of pulse length on the estimate of r app. Shown are data thermal conductivity of the sensor body was 0.2 W m 2 from sensors with (a) full-length, 28 mm, temperature probes and K 1, which was almost equal to that of the dry beads (b) shortened, 18 mm, temperature probes. Comparisons are made (0.18 W m 1 K 1 ). Furthermore, the low diffusivity of among sensors constructed from and 1.67-mm diam. stainless the dry beads, m 2 s 1, creates a slower moving steel needles mounted in bodies fabricated from high thermal con- pulse providing more opportunity for heat flow into the ductivity (HTC) RBC4300 epoxy and low thermal conductivity (LTC) urethane epoxy. Results from the instantaneous pulse sensor body. In wet beads, with a thermal conductivity (Eq. [1]) and finite pulse (Eq. [2]) models are shown. of 0.87 W m 1 K 1 and a diffusivity of m 2 s 1, the pulse moved more rapidly and the temperature grawet to dry beads, regardless of the elements of sensor dient between the media and sensor body was smaller. While this analysis is somewhat circumstantial, results design. Finally, a shorter temperature probe is less likely suggest sensor performance is affected by the difference to deflect during insertion into soil. It is recommended in thermal properties between the sensor body and that the temperature probe be about 10 mm shorter the soil. than the heater probe in future designs. Probe diameter was evaluated using sensors with urethane bodies only. In wet glass beads and agar, the Probe Length, Probe Diameter, diameter of the hypodermic tubing did not cause a sigand Body Material nificant difference in r app (Table 1, Fig 2). In dry beads; A paired t test showed that shortening the temperaprobes. however, r app was 2.5% larger for the 1.67-mm diam. ture probes from 28 to 18 mm caused a small but statistiprobes Also, sensors made from the 1.67-mm diam. cally significant 0.4 and 1.0% decrease in r app in the wet showed a greater sensitivity to changing the calically and dry beads, respectively (Table 1, Fig. 2). Heat transing bration media from wet to dry beads (Table 1). Averag- fer near the tip of the heater probe will be different over all sensors tested, switching from wet to dry than that of an infinite line source, a fact not accounted beads caused a 3.1 and 5.6% increase in r app in the for in Eq. [1] and [2]. Thus, in the full-length sensors and 1.67-mm diam. probes, respectively. Using larger (i.e., 28 mm), the tip of the temperature probe may not diameter probes could reduce deflection when sensors rise to the same temperature as the rest of the needle are installed in soil, thus minimizing errors in r. How- after a heat pulse. This would cause longitudinal heat ever, results shown here suggest larger diameter probes flow toward the tip of the temperature probe, moving may exacerbate the problem of increasing C at lower heat away from the thermistor located in the center of soil water contents. It is recommended that a probe the probe (14 mm). Indeed, T m was slightly lower in the full-length probes resulting in a larger r app. Also, r app of the shortened probes was less sensitive to the changes from diameter of 1.27 mm or less be used for fabrication. The effect of sensor body material was tested using 1.27-mm diam. probes only, all of which were cast from

5 HAM & BENSON: DUAL-PROBE HEAT CAPACITY SENSORS 1189 Table 2. Internal and external heater temperatures immediately after the heat pulse as recorded by thermocouples imbedded in or mounted on the surface of an experimental sensor (1.27-mm probes with urethane body). Temperatures represent the average of five measurements at each pulse length in water-saturated and dry glass beads. Voltage was adjusted inversely to the pulse length so q was Jm 1 for all tests. Wet beads Dry beads Pulse length Internal External Internal External s C the same mold. Body material did not significantly affect calibration in agar or wet beads. However, in dry beads, r app was significantly larger (P 0.1) in sensors made Fig. 5. Effect of interval between repeated measurements on the defrom the RBC4300 epoxy for both full length and short- termination of r app. ened probes. The RBC4300 epoxy has a relatively high conductivity and heat capacity; thus, it can act as a strong by thermocouples imbedded in or mounted on the surheat sink during a heat pulse. As mentioned previously, face of the experimental sensor (1.27-mm probes with it is probable that some of the heat generated in the urethane body). Voltage was adjusted inversely to the heater probe was conducted into the sensor body, effecpulse length so q was Jm 1 for all tests. At the tively reducing q and T m. Because the effect of the sensor end of the 2-s pulse in dry beads, internal and external body on q is not accounted for in the model, the result temperatures were and C, respectively. is larger r app. When using casting epoxies, it appears that Thus, in drier soils, high temperatures resulting from products with lower thermal conductivities and heat short heat pulses might cause moisture movement near capacities (e.g., urethane) are superior. the heater. Furthermore, the high temperatures inside the heater could damage the enamel on the resistance Total Applied Power, Heat Pulse Length, wire and eventually cause an electrical short. Increasing and Probe Temperatures the heat pulse length to 8 s caused a 62 C reduction in The effect of applied power, q, was examined by using internal and a 22 C reduction in external temperatures. input voltages of 9.2, 13.0, and 18.4 V, which for 8-s Heat pulse lengths of 8sorgreater with a q near 850 J pulses corresponded to heating of 400, 800, and 1600 J m 1 are recommended. For the sensors described here, m 1, respectively. For all sensor designs and all calibraof T m between 1.1 C in wet soils and 2.2 C in dry soils. this protocol would result in an easily measured value tion media, changing q had no statistically significant affect on r app (not shown). Heat pulse length was varied between 4 and 16 s while Sampling Frequency the input voltage was adjusted between 9.2 and 18.4 In most cases, DPHC sensors are used to collect data V, respectively. This procedure kept q almost constant on an hourly or daily basis. However, there could be (approximately 850 J m 1 ). Figure 4 shows r app for full cases where more frequent sampling is required. For length and shortened sensors in wet beads as calculated example, if used in an automated irrigation control sysusing the instantaneous pulse model (Eq. [1]) and the tem, the sensors would need to detect the movement of finite pulse model (Eq. [2]). Heat pulse length had no a wetting front at different depths in the soil profile statistical effect on r app when calculated using either (Bremer and Ham, 2003). Figure 5 shows the effect of equation. However, there was a trend for r app to increase repeatedly sampling sensors at different intervals (4, 8, with pulse length when using Eq. [1], albeit the effect 16, 32, and 60 min). When the interval between repeated was very small. For the 1.27-mm urethane sensor, r app samples was 16 min, r app began to increase in both wet increased by 0.9% between 4- and 16-s pulses. Con- and dry beads, although the impact was more proversely, the finite pulse model (Eq. [2]) showed no trends nounced in the dry case. This indicates that heat from as a function of pulse length. Although the differences the previous heat pulse was affecting the determination between Eq. [1] and [2] are small, Eq. [2] is recom- of T m. Thus, the fastest allowable sampling interval for mended because there is both empirical and theoretical DPHC sensors was about 15 min. evidence that it is a better heat flow model for DPHC sensors. Given that applied power and pulse length have negligible effects on sensor performance, these parameters General recommendations from this study are: the di- CONCLUSIONS should be selected to avoid thermally induced water ameter of the needle probes should be 1.27 mm or less; flow, prolong the life of the sensors, and maximize the precision of data acquisition. Table 2 shows maximum internal and external heater temperatures as recorded the length of the temperature probe should be shorter than the heater probe; sensor bodies should be made from materials with low thermal conductivities and heat

6 1190 SOIL SCI. SOC. AM. J., VOL. 68, JULY AUGUST 2004 capacities (e.g., urethane); applied power should be F.W. Caldwell. G.K. Kluitenberg made many helpful sugges J m 1 ; heat pulse lengths should be 8 2s; tions on the first version of the manuscript. the model of Knight and Kluitenberg (2004) should be REFERENCES used for both calibration of r app and the field measurements of C; and repeated sampling of the same sensor Basinger, J.M Laboratory and field evaluation of the dual- probe heat-pulse method for measuring soil water content. M.S. should occur at intervals of 15 min or longer. Thesis. Kansas State Univ., Manhattan. Apparent probe spacing of the recommended sensor Basinger, J.M., G.J. Kluitenberg, J.M. Ham, J.M. Frank, P.L. Barnes, design (1.27-mm diam. probes, urethane body, shortheat-pulse method for measuring soil water content. Vadose Zone and M.B. Kirkham Laboratory evaluation of the dual-probe ened temperature probe) will still have some sensitiv- J. 2: ity to soil water content. Sample calculations were per- Bremer, D.J., and J.M. Ham Soil moisture sensors can help formed to emulate the drying of a silt loam soil from regulate irrigation. Golfdom. June to 0.1 m 3 m 3 (Tarara and Ham, 1997). At 0.1 m 3 Bristow, K.L., G.S. Campbell, and K. Calissendorff Test of a heat-pulse probe for measuring changes in soil water content. Soil m 3, representing the worst case, a sensor calibrated in Sci. Soc. Am. J. 57: wet beads would overestimate C by 4.2% and predict Campbell, G.S., K. Calissendorff, and J.H. Williams Probe for a water content of m 3 m 3. The difference between measuring soil specific heat using a heat-pulse method. Soil Sci. Soc. Am. J. 55: and 11.5% water content is probably negligible in DeVries, D.A Thermal physical properties of soil. p most studies. Furthermore, other sources of error would In W.R. van Wijk (ed.) Physics of the plant environment. John likely have a larger impact on accuracy, including probe Wiley and Sons, New York. deflection upon insertion, uncertainty in soil specific Ham, J.M On the measurement of soil heat flux to improve estimates of energy balance closure. Eos Trans. AGU, 82(47), Fall heat, spatial variability in soil properties, and poor probe- Meet. Suppl., Abstract B51A American Geophysical Union, soil contact caused by shrinking and swelling. After in- Washington, DC. stallation in the field, it is strongly recommended that Heitman, J.L., J.M. Basinger, G.J. Kluitenberg, J.M. Ham, J.M. Frank, and P.L. Barnes Field evaluation of the dual-probe heatsensor results be compared with independent estimates pulse method for measuring soil water content. Vadose Zone J. of C. Volumetric samples can be collected at the same 2: depth as the sensors and C approximated from graviof the heat pulse method for measuring soil volumetric heat capac- Kluitenberg, G.J., J.M. Ham, and K.L. Bristow Error analysis metric analysis (Basinger, 1999; Heitman et al., 2003). ity. Soil Sci. Soc. Am. J. 57: Although some variation is expected, this procedure Kluitenberg, G.J., and J.R. Philip Dual thermal probes near will provide assurance that the laboratory calibrations plane interfaces. Soil Sci. Soc. Am. J. 63: are valid. Knight, J.H., and G.J. Kluitenberg Simplified computational approach for dual-probe heat-pulse method. Soil Sci. Soc. Am. J. 68: ACKNOWLEDGMENTS Ochsner, T.E., R. Horton, and T. Ren Use of the dual-probe heat-pulse technique to monitor soil water content in the vadose This material is based upon work supported by the Cooper- zone. Vadose Zone J. 2: ative State Research, Education, and Extension Service, USDA, Song, Y., J.M. Ham, M.B. Kirkham, and G.J. Kluitenberg Meaunder Agreement No Any opinions, findpulse technique. J. Am. Soc. Hortic. Sci. 123: suring soil water content under turfgrass using the dual-probe heat- ings, conclusions, or recommendations expressed in this publi- Tarara, J.M., and J.M. Ham Measuring soil water content in cation are those of the author(s) and do not necessarily reflect the laboratory and field with dual-probe heat-capacity sensors. the view of the USDA. Technical support was provided by Agron. J. 89: