CALIBRATION OF A TDR INSTRUMENT FOR SIMULTANEOUS MEASUREMENTS OF SOIL WATER AND SOIL ELECTRICAL CONDUCTIVITY

CALIBRATION OF A TDR INSTRUMENT FOR SIMULTANEOUS MEASUREMENTS OF SOIL WATER AND SOIL ELECTRICAL CONDUCTIVITY N. Ebrahimi-Birang, C. P. Maulé, W. A. Morley ABSTRACT. Time domain reflectometry (TDR) can be used to simultaneously measure in-situ water content and electrical conductivity (EC TDR ). The method used to determine soil electrical conductivity from TDR waveforms requires measurement of the reflection voltages from a special part of the waveform. However, not all TDR instruments have this ability. The purpose of our research was to extend the ability of the MP917 Moisture Point instrument (ESI Environmental Sensors, Inc.) such that bulk soil electrical conductivity could be obtained. Our method involved use of resistors of known resistance to establish a relationship between arbitrary voltage and the reflection coefficient. As we did not have a standard of comparison (another TDR system that could measure both soil moisture content and EC), we used two probes of different lengths (15 and 22 mm). If our method worked, then the two probes should calculate the same EC values. Using our method, we achieved the following: determination of geometric constants for solution EC values between.7 and 25.7 ds m 1, ability to determine reasonably accurate KCl solution EC values for each of the two probes, and ability to determine sand EC TDR that had a 1:1 agreement between the 15 and 22 mm probes for moisture contents less than.17 m 3 m 3 for non-saline and saline conditions, but not for saline conditions near saturation (.28 to.35 m 3 m 3 ). For such conditions, the long probe produced EC TDR values that were 1.4 times that of the short probe. Although the EC TDR values obtained from the MP917 could not be confirmed with an independent measure, there appears to be sufficient response to a range of saline conditions that the relative values of EC TDR provide suitable accuracy such that the usefulness of the MP917 can be extended to relative measurements of EC. Keywords. Soil electrical conductivity, Soil moisture, TDR, Time domain reflectometry. Recent studies have shown that time domain reflectometry (TDR) can be used to simultaneously and reliably measure soil water content and soil salinity. For soil salinity, this provides a simpler method of monitoring than the traditional method involving soil extracts. Topp et al. (198) and Topp and Davis (1985) developed the TDR for measuring the volumetric water content of unsaturated soils. They reported a strong relationship between the volumetric water content and dielectric constant for a wide range of soils. This relationship was found to apply to other soils in later studies (Dalton and van Genuchten, 1986; Zegelin et al., 1989; Heimovaara and Bouten, 199; Wraith and Baker, 1991; Whalley, 1993; Nielsen et al., 1995). However, other authors (Herkelrath et al., 1991; Dasberg and Hopmans, 1992) reported relationships that differed slightly from those reported by Topp et al. (198). In an attempt to Article was submitted for review in July 25; approved for publication by the Soil & Water Division of ASABE in January 26. Presented at the 23 CSAE/ASAE Annual Intersection Meeting as Paper No. RRV3-39. The authors are Nader Ebrahimi-Birang, Graduate Student, Department of Civil and Geological Engineering, Charles P. Maulé, ASABE Member, Professor, Department of Agricultural and Bioresource Engineering, and Wayne A. Morley, Electronics and Instrumentation Technologist (retired), Department of Agricultural and Bioresource Engineering, University of Saskatchewan, Saskatoon, SK, Canada. Corresponding author: Charles P. Maulé, Department of Agricultural and Bioresource Engineering, University of Saskatchewan, Saskatoon, SK, Canada S7N 5A9; phone: 36-966-536; fax: 36-966-5334; e-mail: charles.maule@ usask.ca. standardize the method of calculating moisture content from TDR measurements, Hook et al. (1992) proposed a new TDR technique based on a remote shorting diode technique. They showed that this new technique increased the reliability of soil water content measurement. They also found that there was a linear relationship between volumetric soil water content and time delay when the latter is expressed as T/T air, where T is the time delay in soil and T air is that in air for the same distance. Soil electrical conductivity can be obtained from TDR waveforms by analyzing the change in the voltage signal from the probe reflections of the waveform with travel time (Dalton et al., 1984; Topp et al., 1988; Yanuka et al., 1988; Zegelin et al., 1989). This procedure has become commonly utilized to calculate bulk soil electrical conductivity (Noborio, 21). The voltages can be found with an oscilloscope such as the Tektronix TDR instrument, which has been used by numerous researchers. Another kind of TDR instrument, the MP917 Moisture Point (ESI Environmental Sensors, Inc., Victoria, B.C., Canada), was constructed for the sole purpose of measuring moisture content using the calibration methods introduced by Hook et al. (1992). The instrument has the capability of outputting a waveform (relative voltage of arbitrary units vs. propagation time along the TDR cable and probe); however, as the voltage values are not known, it is not possible to determine soil electrical conductivity using standard analytical methods such as those described by Noborio (21). To our knowledge, no one has utilized the MP917 to determine soil electrical conductivity. Sun et al. (2) and Nichol et al. (22) both utilized MP917 systems Transactions of the ASABE Vol. 49(1): 75 82 26 American Society of Agricultural and Biological Engineers ISSN 1 2351 75

to investigate the effect of electrical conductivity on travel time within saline solutions or soils, but they did not attempt to interpret the MP917 waveform to that of electrical conductivity. In this research, we investigated a method by which the output waveform of the MP917 TDR system can be standardized to voltage units such that the soil bulk electrical conductivity may be measured along with moisture content using an uncoated two-rod moisture probe. Voltage (V) V 2 V 3 V f V dt Range for V f measurement THEORY DETERMINATION OF SOIL MOISTURE CONTENT The measurement of soil water content using time domain reflectometry (TDR) was developed by Topp et al. (198), who showed that a general empirical relationship can be used for many soil types: v =.53 +.29 a (5.5 1 4 ) 2 a (1) + (4.3 1 6 ) 3 a where v is the volumetric moisture content of soil, and a is the apparent dielectric constant of the soil. Hook and Livingston (1995) found that there was a linear relationship between volumetric soil water content and the ratio T/T air : v = (T/T air T s /T air )/(.5 w 1) (2) where v is the volumetric water content of soil, w is water dielectric coefficient, and T, T s, and T air are the travel time of a radio frequency pulse in soil, in oven-dried soil, and in air, respectively. The slope of this linear equation is 1/(.5 w 1), and the intercept at v = is T s /T air. At 2 C, w = 8.362 (Weast, 1986), so the predicted slope is.1256. The volumetric water content of soil ( v ) can be determined by measuring T and substituting this value into equation 2. The theoretical time delay in air is T air = 2Lc 1, where c is the speed of light in free space (2.997 1 8 m s 1 ), and L is the length of the probe in m (Hook and Livingston 1995). DETERMINATION OF SOIL ELECTRICAL CONDUCTIVITY Previous studies have shown that the TDR signal in conductive porous media approaches a constant voltage (V f, fig. 1) after long times such that multi-reflectance interferences (e.g., V 2, V 3, and others before V f ) attenuate (Dalton et al., 1984; Topp et al., 1988; Zegelin et al., 1989; Nadler et al., 1991; Ward et al., 1994; Noborio, 21). The value of V f represents the impedance of the direct current component only and is independent of the probe configuration, transfer efficiency of the pulse energy, or multiple reflections (Nadler et al., 1991). The value of V f, relative to the TDR input signal (V ) can be used to obtain bulk electrical conductivity of the media (EC a ). Although various equations, based on the reflection at long travel times, have been developed for estimation of bulk electrical conductivity (Dalton et al., 1984; Zegelin et al., 1989; Nadler et al., 1991; Noborio, 21), the most commonly used method is the Giese-Tiemann (G-T) method, as presented by Topp et al. (1988) and described by Noborio (21): EC G T K 1 ρ = Zu 1+ ρ (3) T 1 T 2 Time (ns) Figure 1. Schematic of a generalized TDR raw wave scan showing points of measurement for soil water content (T 1, T 2 ) and electrical conductivity (V, V f ). Points V 2 and V 3 are where probe reflection plateaus begin. where EC G-T is the bulk electrical conductivity of the media as determined by the Giese-Tiemann method, K is the geometric constant of the probe (m 1 ), Z u is the characteristic impedance of a cable (ohms), and ρ is the reflection coefficient at a distant point from the first reflection on the waveform. If the measurement system provides true voltage units, then: V f V ρ = (4) V Noborio (21) defines V f as 1 times that of the probe travel time (dt, fig. 1) and V as the signal amplitude from the TDR instrument as measured in the cable before it enters the TDR probe. The reflection coefficient may also be obtained from the cable impedance (Z u ) and the probe impedance (R L ) when embedded in the medium (Nadler et al., 1991; Nissen et al., 1998): RL Z ρ u = (5) RL + Zu Equation 5 can be rearranged to find R L such that the right side of the expression can be used in the Giese-Tiemann method (eq. 3): 1+ ρ R L = Zu (6) 1 ρ Although there is a theoretical representation of K (Topp et al., 1988), it may also be measured by immersing the probe in a solution of known salinity (EC s ), measuring the resistance of the TDR probe (R L ), and then using an equation similar to that of Rhoades and van Schilfgaarde (1976): K = EC s (25 C) R L ƒ 1 t (7) where ƒ t is a correction coefficient for temperature. Once the geometric constant of the probe (K) is known, equation 3 may also be expressed as: EC TDR = K R 1 L (8) where EC TDR is the bulk soil electrical conductivity as measured via TDR. The number of reflections required to achieve V f is affected by the EC of the solution or the soil: the lower the EC, the greater the number of discernable reflections (e.g., V 2, V 3 in fig. 1). Topp et al. (1988) showed a TDR trace 76 TRANSACTIONS OF THE ASABE

for distilled water (their fig. 2a) in which V f was achieved after about 2 reflections (2 times the T 2 -T 1 interval). From figure 5 of Yanuka et al. (1988), V f was achieved at about 8 and 3 reflections for solutions of.2n KCl and.1n KCl, respectively. From figure 1 of Nadler et al. (1991), we estimated that V f was achieved within two reflections for saline soils of varying wet and dry conditions. The definition of where V f is to be measured has been defined differently: as the region on the TDR trace where reflected voltage remains constant, is not zero, and after all multiple reflections have ceased; or at a specified number of T 2 -T 1 intervals, e.g., 2 as defined by Hart and Lowery (1998), or 1 as defined by Noborio (21). MATERIALS AND METHODS TDR INSTRUMENT AND PROBES This study was carried out with an MP917 Moisture Point TDR instrument (ESI Environmental Sensors, Inc., Victoria, B.C., Canada) using a computer attached to the RS232 port to store raw scan data. Raw scan data from the MP917 are in voltage arbitrary units, which are not defined by the instrument or company, and in time counts, which for this instrument is.141 ns per count. The MP917 utilizes TDR technology but also employs a switching diode technique (Hook et al., 1992) to mark the waveform at the start of the probe. Although the MP917 is very appropriate for obtaining volumetric moisture contents, it has two limitations for determining bulk electrical conductivity (EC a ): it does not provide absolute voltage values, and its scan time or time-scaling ability cannot consider long scan intervals such that 1 to 2 reflections can be achieved under all conditions. Absolute voltage values are needed to enable calculation of reflection coefficients (eq. 4), which are required to determine the geometric constant (eqs. 6 and 7). For our measurements, we used a two-rod (stainless steel) single-diode probe (with the diode in an aboveground handle ), with a rod diameter of 3 mm and distance between the two rods of 12 mm. We used two probes of differing lengths: 15 mm (exposed length 136 mm), and 22 mm (exposed length 198 mm). The exposed length is the metal part of the rods inserted into solution or soil for measurement. The probes were fabricated according to the instructions in a Moisture Point technical brief found on the ESI web site (Young, 1998). We kept the probes short (minimum length allowable by the MP917 instrument is 15 mm) so that at least several multiple reflections could be obtained on the raw scan. For the purpose of our study, we used the MP917 in the raw scan mode with the diode turned off. DETERMINATION OF TDR ELECTRICAL CONDUCTIVITY To overcome the limitation of not knowing the absolute voltage values of the MP917 output and to obtain reflection coefficients, we used 11 resistors of known impedance (1.6, 18.4, 27.4, 38.8, 1.2, 147.2, 219.7, 477.8, 824.9, 987.1, and 21 ). The resistors (representing R L of eq. 5) were attached to the end of a TDR cable (Z u of 5.3 ) of similar length used for the two TDR probes. The reflection coefficient for each resistor could then be directly calculated using equation 5 (by inserting the resistor value in ohms as R L ) and compared to measured voltage values obtained by the raw scan of the MP917 as the difference in output voltage from the y-axis scale, dv(v f V ). A regression model was then used to quantify the relationship between dv and ρ such that for any dv measured with the MP917, a ρ value could be predicted and an R L value could be then calculated using equation 6. For obtaining the probe geometric constant, we used the common procedure of immersing the TDR probes into solutions of known strength. This is defined by equation 7 and described by Noborio (21). This method requires a long enough scan time such that V f can be obtained. As this might be a limitation with the MP917, our procedure was to evaluate all reflection plateaus (e.g., V 2, V 3,..., V f ) of the TDR raw scan. We prepared the following KCl solutions with distilled water (8.6 S m 1 ):.18,.75, 1.16, 3.94, 8.16, 12.5, and 25.7 ds m 1. The solutions were equilibrated to room temperature (23 C to 25 C) within covered upright PVC containers (1 mm dia. 25 mm deep). The exposed metal probes of each TDR probe was immersed in the solution in the center of the container. From the obtained raw scans, the values for T 1, T 2, V, V 2, and V 3 to V f (if possible) were determined graphically. For each reflection plateau (V 2 to V f ), the dv value (e.g., V 2 V ) was calculated, and then the reflection coefficient ( ρ ) was found using the regression equation relationship established using the resistors. The ρ value was then used with the cable impedance, Z u (5.3 ), to find the probe impedance (R L ) using equation 6. Once R L was found, the geometric constant could be calculated using equation 7. We then evaluated which reflection plateau, and associated geometric constant, was the most representative among the different solutions by using that which produced the smallest coefficient of variation. As we did not have an appropriate instrument (e.g., a Tektronix cable tester) for accurately obtaining the probe geometric constant, we compared the values of EC TDR for each solution, as obtained by our TDR probes of different lengths (15 and 22 mm). We reasoned that if our method is correct, then the EC TDR results of these solutions obtained by the two different probe lengths should agree. SOIL MEASUREMENTS For a soil testing medium, we used locally collected sand, sieved to remove all particles greater than.84 mm diameter. The sand (95.5% S, 4.5% Si, and % C) was collected from sand deposits just south of Saskatoon, SK, Canada. Most of the sand (79.5%) was in the particle diameter range.1 to.5 mm, as determined by dry sieving. The air-dried sand was then poured into three tared PVC containers of 154 mm inside diameter in 5 to 7 mm thick layers and compacted lightly (by using a 1.1 kg solid plastic rod of.3 m diameter dropped from a height of.2 to.4 m, 5 to 6 times) for each layer such that the final depth was about 3 mm. The columns with air-dried sand were then weighed so that the dry bulk density could be calculated. Raw scan TDR readings were obtained of the dried sand in all columns. The columns were then wet from below, each column with a different solution: distilled water at 25.8 S cm 1, a KCl solution at 3.99 ds m 1, and a KCl solution at 11.65 ds m 1. The water level was allowed to rise to the soil surface in the columns, and the columns stood at room temperature for 48 h (covered to prevent evaporation). TDR readings (at saturation) were then obtained, and the columns Vol. 49(1): 75 82 77

were weighed (so as to calculate saturated moisture content). The columns were allowed to drain for 48 h (again covered to prevent evaporation). After drainage, TDR readings were again obtained, and the columns were weighed. The soil from the columns was then removed into plastic trays and allowed to partially dry with thorough mixing twice a day. The trays were weighed before and after drying. A sample of soil (5 to 8 g) was taken at the start of drying for gravimetric moisture determination. After 2 days of drying, the soils were again well mixed, packed back into the columns (using similar packing techniques as before), allowed to equilibrate for 2 days, and then TDR readings were obtained. This procedure was repeated such that two other moisture contents between draining and air-drying were obtained. After each repacking, the height of soil in the column was measured so that bulk density could be determined. The entire experiment thus measured 12 different conditions (four moisture contents three different salt contents). From the raw wave scans, the parameters T 1, T 2, V, and V 2 to V f (fig. 1) were measured and then used to calculate TDR ( v of eq. 2), and EC TDR (EC G-T of eq. 3, where K had been established with eq. 7 from the KCl solutions and ρ from dv using the resistor regression equation). Values of TDR were compared to actual values of (measured by weighing and volume measurement). As we did not have a method for determination of the absolute value of bulk electrical conductivity (EC a ), we compared the EC TDR values obtained by the probes of two different lengths. If our procedures for determining the ρ vs. dv relationship by resistors and for determining K from the KCl solutions were correct, then the two probes should provide similar EC TDR values. All measurements were done within the departmental soil and water laboratory at room temperature of 23 C to 25 C, corrected to 25 C when necessary. RESULTS VOLTAGE REFLECTION COEFFICIENT ( ) AND PROBE IMPEDANCE (R L ) The voltage reflection coefficients ( ρ ) were calculated using equation 5 for the different resistors, with their measured resistance (using an ohmmeter) in ohms for R L and a cable resistance of 5.3 for Z u. Voltages (V, arbitrary units as from the instrument scale) of the resistor raw scans reached a plateau within 4 to 5 ns regardless of the load, with small changes in voltage for the remainder of the scan (fig. 2). Regression coefficients of ρ vs. dv of the raw scan taken every 1.4 ns from about 9 to 36 ns can be described very well by linear regression equations with coefficients of determination exceeding.999. Cable length has a slight effect upon the linear regression equation, which becomes important when later calculating the geometric constant. Linear regression equations, using the V f - V of the raw scan (where V f is the average of all the voltage values from travel time counts 18 to 255, and V is the average of 1 voltage readings of the cable) for the 1.6 to 21 resistors are:.8 m cable (15 mm TDR probe): ρ =.765dV +.2337 (r 2 =.9992) (9) Voltage (arbitrary units) 25 2 15 1 5 V V f 21 Ω 478 Ω 22 Ω 147 Ω 1 Ω 38.8 Ω 27.4 Ω 18.4 Ω 1.6 Ω 1 2 3 4 Time (ns) Figure 2. Raw wave scans of resistors attached to MP917 single-diode probe interface with a 1.1 m cable. Amplitude of average V f, relative to V, is dv value. 1.1 m cable (22 mm TDR probe): ρ =.759dV +.158 (r 2 =.9994) (1) PROBE IMPEDANCE (R L ) AND GEOMETRIC CONSTANT (K) The wave traces of the TDR probes immersed in KCl solutions of various strengths are shown in figure 3. For solutions of low EC (.75 ds m 1 and lower), distinct reflections after the first plateau are visible. For the shorter probe (15 mm), only three distinct plateaus (V 2, V 3, and V 4 ) could be viewed by the instrument for low EC solutions, whereas for the longer probe (22 mm), only two plateaus could be viewed. The probe end (T 2 at about 11 ns for the 15 mm probe, and T 2 at about 16 ns for the 22 mm probe) becomes difficult to visually differentiate at solution EC values of 12.5 ds m 1 and higher for the short probe and at 3.94 ds m 1 for the long probe. Because the MP917 system cannot extend the raw wave trace for longer time intervals (while maintaining the trace of the cable on screen), we considered which portion of the solution raw wave scan (fig. 3) gave the best match with equations 9 and 1 in terms of producing a geometric constant that held constant for the widest range of solution EC values. The portion of the solution raw wave scan that provided the best geometric constant for solutions of.38 to 25.7 ds m 1 for both probes was the third reflection (V 3 ) of the solutions (fig. 3). The probe geometric constants (K), as calculated for each solution, are relatively constant for each probe from.75 to 25.7 ds m 1 (table 1). This agrees with Rhoades et al. (1989), who found that the relationship between bulk electrical conductivity (using a four-electrode probe) and solution electrical conductivity to be linear between 4 ds m 1 and 2 ds m 1, with little deviation to about 1 ds m 1. The geometric constant for the short probe (15 mm) is 48.8, while that for the 22 mm probe is 39.3. Distilled water is overestimated by an order of magnitude. As most soil solutions will likely be greater than.3 ds m 1, this should not pose a problem. For the 15 mm and 22 mm probes, geometric constants of 48.8 and 39.3, respectively, were used to calculate the EC TDR values. These geometric constants were obtained by averaging the K values from the solutions of.748 to 25.7 ds m 1. EC TDR was calculated using equation 8. 78 TRANSACTIONS OF THE ASABE

15 15 mm probe 2 mm probe V 2, 15 mm V 3, 15 mm V, 15 mm 4.86 ds/m (distilled) 1.185 ds/m Relative signal amplitude (voltage, arbitrary units) 5 5 V V 2, 2 mm V 3, 2 mm 1.16 ds/m 1 3.94 ds/m 25.7 ds/m 15 1 2 3 4 Travel time (ns) Figure 3. Raw trace of the two probes in KCl solutions of different concentrations. Wave scans for 22 mm probe are offset by 1.27 ns such that beginning of the probe (T 1 ) matches with that of the 15 mm probe. Table 1. Actual solution electrical conductivities (EC s ), calculated probe geometric constants (K), and predicted solution electrical conductivities (EC TDR ) in KCl solutions for 15 mm and 22 mm probes. EC s EC TDR15 EC TDR22 (ds m 1 ) K 15 (ds m 1 ) K 22 (ds m 1 ).8 4.8.86 8.8.39.185 37.4.241 35.9.22.748 49.1.743 39.7.742 1.16 49.3 1.15 4. 1.14 3.94 5.5 3.81 39.4 3.94 8.16 49. 8.13 4.2 8.2 12.5 48.1 12.7 38.9 12.7 25.7 46.7 26.8 39.1 26.2 TDR MEASUREMENTS OF SOIL BULK ELECTRICAL CONDUCTIVITY Three sand columns were prepared, each with a different strength solution (26 S cm 1, 3.99 ds m 1, and 11.65 ds m 1 ) added to bring the sand to saturation. After raw wave scans at saturation, the columns were allowed to drain naturally, and the EC values of the drainage waters were measured (ECd). The EC d values averaged.19 ds m 1 (.17,.21, and.18 ds m 1 for the three columns) greater than the EC of the solution added. From raw wave scans of the sand columns, moisture content (eq. 2) and bulk soil electrical conductivity as measured by the TDR (eq. 3) were calculated and compared. Raw scans of the TDR probe immersed in the sand columns at saturation (.33 to.35 m 3 m 3 ) and at a relatively dry value (.7 to.9 m 3 m 3 ) for three different EC d values are shown in figures 4 and 5. The end of the probe (T 2 ) cannot be discerned at the high EC d value of 11.8 ds m 1 for the saturated moisture content of the 22 mm probe. For all scans, only the first reflection plateau (V 2 ) is evident. The effect of salt content in the solution is clearly displayed with the height of the V 2 plateaus. For saturated conditions (fig. 4) and for similar EC d values, the height of the V 2 plateaus are dissimilar between the short and long probes, with the difference increasing with increasing salt content. At low Signal amplitude (voltage arbitrary units) 225 15 mm probe 2 22 mm probe 175.2 ds/m 15 125 1 4.2 ds/m 75 5 25 11.8 ds/m 1 2 3 4 Travel time (ns) Figure 4. Wave traces of soil columns at saturation ( between.33 and.35 m 3 m 3 ) and different soil solution electrical conductivities (EC d ) as measured with the 15 mm and 22 mm probes. Vol. 49(1): 75 82 79

Signal amplitude (voltage arbitrary units) 225.2 ds/m 2 175 4.2 ds/m 15 11.8 ds/m 125 1 75 5 15 mm probe 22 mm probe 25 1 2 3 4 Travel time (ns) Figure 5. Wave traces of soil columns at low moisture contents ( between.7 and.9 m 3 m 3 ) and different soil solution electrical conductivities (EC d ) as measured with the 15 mm and 22 mm probes. moisture contents (fig. 5), the plateaus are similar in voltage; however, there is slight difference with the highest EC d value. The relationship between measured (by weighing) volumetric moisture content and TDR-determined volumetric moisture content of soils from the three different solutions for both probes is linear, with the exception of the saturated values (.33 to.35 m 3 m 3 ), which depart slightly (fig. 6). Sun et al. (2), who also used the MP917 system, found that this relationship was curvilinear, with the departure from linearity increasing with moisture content and saturated soil extract EC. However, our departed values are of low EC d (.2 ds m 1 ) and high EC d (11.8 ds m 1 ). For the long probe for the 11.8 ds m 1 solution, it was difficult to properly determine the location of T 2 for the.33 m 3 /m 3 and for.11 m 3 /m 3 moisture contents; thus, these values were not included in our moisture content analysis. It is apparent from figures 4 and 5 that the EC d and the volumetric moisture content both have strong effects on the height (voltage) of the raw scan following the first plateau (V 2 ) after the end of the probe (T 2 ). Because none of the scans done within the soils show evidence of reflection plateaus other than the first one (V 2 ), we used the portion of the scan between 21 and 36 ns to calculate an average dv and then to calculate EC TDR. The resulting EC TDR values for the short (15 mm) and long (22 mm) TDR probes are shown in figure 7. For EC TDR values of less than.41, there is a 1:1 relationship between the short and long probes (eq. 11), whereas for EC TDR values of greater than.49, for the short probe, the relationship is 1.4 to 1. (the longer probe overestimates the short probe by 1.4, eq. 12). Overestimations occurs for soils that are wet (.28 to.35 m 3 m 3 ) and have EC d values of at least 4.2 ds m 1 (fig. 7). EC TDR22 =.996 EC TDR15.2 (11) r 2 =.987, n = 8, EC TDR15 <.41 EC TDR22 = 1.36 EC TDR15.4 (12) r 2 =.98, n = 4, EC TDR15 >.49 The effect of moisture content and salt content of the solution may be explained if one considers the relationship of EC a as described by Rhoades et al. (1989): EC a = EC s + T EC w (13) where EC s is the apparent electrical conductivity of the soil solid phase, T is a transmission coefficient (<1) that is envi Measured volumetric water content (m 3 m 3 ).4.3.2.1 Length (mm) EC d (ds m 1 ) 15 22.2 4.2 11.8 Drained soils Saturated soils.1.2.3.4 TDR, volumetric water content (m 3 m 3 ) EC TDR22 (ds m 1 ) long probe 2.5 2. 1.5 1..5 3 3 Saturated θ (.38 to.4 m m ) Drained θ (.33 to.35) 1st dried θ (.11 to.18) 2nd dried θ (.7 to.9) EC d > 4.2 ds m 1 and θ>.28 m 3 m 3 EC d =.2 ds m 1 for all θ or EC. d > 4.2 ds m 1 for θ <.17 m 3 m 3..5 1. 1.5 2. 2.5 EC TDR15 (ds m 1 ) short probe Figure 6. Volumetric soil water content measured using balance vs. that measured using MP917. Figure 7. Relationship of EC TDR values from short (15 mm) and long (22 mm) probes. 8 TRANSACTIONS OF THE ASABE

sioned to correct for the effect of tortuosity on current flow through the geometric arrangement of water within the pore space, and EC w is the electrical conductivity of the soil solution. As the moisture content or the EC w increases, bulk electrical conductivity also increases. As the moisture content decreases, the bulk electrical conductivity approaches that of the soil solid phase. We considered whether there was a bulk density effect, as we did not achieve consistent packing with the soils between moisture contents. The first packing, at saturation and drainage moisture contents (figs. 6 and 7) had bulk densities between 16 and 164 kg m 3, whereas subsequent packing for two moisture contents after drying had lower bulk densities of between 14 and 153 kg m 3. We could see no effect of this variation in bulk density on TDR-determined moisture contents (fig. 6) resulting in departure from the 1:1 line. The effect of bulk density on TDR-measured moisture content has been reported as being small, a difference of 1 kg m 3 in bulk density results in a variation of moisture content of.34% (Lediu et al., 1986). Although our data are limited, the higher bulk densities for the wet moisture contents might possibly have affected the difference in slope between the two EC TDR22 relationships described in figure 7 and by equations 11 and 12. SUMMARY In this article, we investigated a method by which a TDR soil moisture instrument (MP917 Moisture Point), originally designed for just moisture measurement, can be used to determine bulk soil electrical conductivity (EC TDR ). Two difficulties were identified in accomplishing this: the arbitrary voltage units used by the instrument, and the limited scan time to enable numerous end-of-probe reflections (more than 1) within solutions. Both of these placed a limitation on obtaining probe geometric constants. To overcome these limitations, we utilized known resistances to obtain voltage reflection coefficients, and from this we were able to obtain stable geometric constants at the third reflection. Because we did not have a standard for comparison, we used two TDR probes of differing lengths (15 and 22 mm) with the premise that if our method worked, then each TDR probe would provide similar EC values. Our method was successful with regards to the following: Obtaining geometric constants for each probe that remained relatively constant between EC solution (KCl) values of.7 and 25.7 ds m 1. Being able to determine accurate electrical conductivity values, relative to the two probes used for the MP917, of the KCl solutions using the geometric constants for both probes. Being able to determine EC TDR values for a sandy soil such that both probes were in agreement for moisture contents less than.17 m 3 m 3 for EC d values of.2 to 11.8 ds m 1, or for all moisture contents (.7 to.38 m 3 m 3 ) for EC d values of.2 ds m 1. The method did not work for saline soils (EC d values of 4.2 and 11.8 ds m 1 ) for wet conditions (near field capacity to saturation,.28 to.38 m 3 m 3 ), as indicated by the EC TDR for the 22 mm probe being about 1.4 times greater than that of the 15 mm probe, rather than 1.. REFERENCES Dalton, F. N., and M. T. Van Genuchten. 1986. The time-domain reflectometry method for measuring soil water content and salinity. Geoderma 38: 237-25. Dalton, F. N., D. S. Helkerrath, D. S. Rawlins, and J. D. Rhoades. 1984. Time-domain reflectometry: Simultaneous measurement of soil water content and electrical conductivity with a single probe. Science 224(4652): 989-99. Dasberg, S., and J. W. Hopmans. 1992. Time domain reflectometry calibration for uniformly and non-uniformly wetted sandy and clayey loam soils. SSSA J. 56(5): 1341-1345. Hart, G. L., and B. Lowery. 1998. Measuring instantaneous solute flux and loading with time domain reflectometry SSSA J. 62(1): 23-35. Heimovaara, T. T., and W. Bouten. 199. A computer controlled 36-channel time domain reflectometry system for monitoring soil water contents. Water Resour. Res. 26(11): 2311-2316. Herkelrath, W. N., S. P. Hamburg, and F. Murphy. 1991. Automatic, real-time monitoring of soil moisture in a remote field area with time-domain reflectometry. Water Resour. Res. 27(5): 857-864. Hook, W. R., and N. J. Livingston. 1995. Errors in converting time domain reflectometry measurements of propagation velocity to estimates of soil water content. SSSA J. 6(1): 35-41. Hook, W. R., N. J. Livingston, Z. J. Sun, and P. B. Hook. 1992. Remote diode shorting improves measurement of soil water by time domain reflectometry. SSSA J. 56(5): 1384-1391. Lediu, J. P. de Ridder, P. de Clerck, and S. Dautrebande. 1986. A method of measuring soil moisture by time-domain reflectometry. J. Hydro. 88(3-4): 319-328. Nadler, A., S. Dasberg, and I. Lapid. 1991. Time domain reflectometry measurements of water content and electrical conductivity of layered soil columns. SSSA J. 55(4): 938-943. Nichol, C., R. Beckie, and S. Leslie. 22. Evaluation of uncoated and coated time domain reflectometry probes for high electrical conductivity systems. SSSA J. 66(5): 1454-1465. Nielsen, D. C., H. J. Large, and R. L. Anderson. 1995. Time-domain reflectometry measurements of surface water content. SSSA J. 59(1): 13-15. Nissen, H. H., P. Moldrup, and K. Henriksen. 1998. Time domain reflectometry measurements of nitrate transport in manure-amended soil. SSSA J. 62(1): 99-19. Noborio, K. 21. Measurement of soil water content and electrical conductivity by time domain reflectometry: A review. Computers and Electronics in Agric. 31(3): 213-237. Rhoades, J. D., and J. van Schilfgaarde. 1976. An electrical conductivity probe for determining soil salinity. SSSA J. 4(5): 647-651. Rhoades, J. D., N. A. Manteghi, P. J. Shouse, and W. J. Alves. 1989. Soil electrical conductivity and soil salinity: New formulation and calibrations. SSSA J. 53(2): 433-439. Sun, Z. J., G. D. Young, R. A. McFarlane, and B. M. Chambers. 2. The effect of soil electrical conductivity on moisture determination using time-domain reflectometry in sandy soils. Canadian J. Soil Sci. 8(1): 13-22. Topp, G. C., and J. L. Davis. 1985. Measurement of soil water content using TDR: A field evaluation. SSSA J. 49(1): 1-24. Topp, G. C., J. L. Davis, and A. P. Annan. 198. Electromagnetic determination of soil water content: Measurement in coaxial transmission lines. Water Resour. Res. 16(3): 574-582. Topp, G. C., M. Yanuka, W. D. Zebchuk, and S. Zegelin. 1988. Determination of electrical conductivity using time domain reflectometry: Soil and water experiments in coaxial liens. Water Resour. Res. 24(7): 945-952. Ward, A. L., R. G. Kachanoski, and D. E. Elrick. 1994. Laboratory measurements of solute transport using time domain reflectometry. SSSA J. 58(4): 131-139. Weast, R.C., ed. 1986. Handbook of Physics and Chemistry. 67th ed. Boca Raton, Fla.: CRC Press. Vol. 49(1): 75 82 81

Whalley, W. R. 1993. Consideration on the use of time-domain reflectometry (TDR) for measuring soil water content. J. Soil Sci. 44(1): 1-9. Wraith, J. M., and J. M. Baker. 1991. High resolution measurement of root water uptake using automated time-domain reflectometry. SSSA J. 55(4): 928-933. Yanuka, M., G. C. Topp, S. J. Zegelin, and W. D. Zebchuck. 1988. Multiple reflections and attenuation of TDR pulses: Theoretical considerations for applications to soil and water. Water Resour. Res. 24(7): 939-944. Young, G. D. 1998. Single diode probes. Moisture Point Technical Brief 13. Victoria, B.C., Canada: Environmental Sensors, Inc. Available at: www.esica.com/support/tech/word/tb13.doc. Accessed 14 Jan. 25. Zegelin, S. J., I. White, and D. R. Jenkins. 1989. Improved field probes for soil water content and electrical conductivity measurements using time domain reflectometry. Water Resour. Res. 25(11): 2367-2376. 82 TRANSACTIONS OF THE ASABE