Response of a New Soil Water Sensor to Variable Soil, Water Content, and Temperature

Thomas (1966) and later developed by Dean et al. (1987), whose work has lead to commercial soil water The success of time domain reflectometry (TDR) has led to the measuring instruments (Evett and Steiner, 1995; Paltidevelopment of other instruments that use the soil dielectric constant as the basis for determining volumetric soil water content. An example neanu and Starr, 1997). Readers are referred to those is the Water Content Reflectometer (WCR; Campbell Scientific, Lopapers for a more complete description of the measuregan, UT), which is much less expensive than TDR and is used widely, ment principles. Briefly, the basis for the approach is although little has been published concerning its applicability to soil that when a capacitor is subjected to an oscillating curwater content monitoring.1 The primary objectives of this study were rent, the resultant oscillation frequency is related to the to determine the WCR soil water calibration for different soils and capacitance of the circuit, with the oscillation frequency to investigate how it is affected by changing temperature. We found decreasing as the capacitance increases. The exact relathe individual sensors to be very precise (CV # 0.05) under the tionship is specific to the circuitry of the instrument controlled laboratory conditions of this study. Variability among sen(e.g., Dean et al., 1987). In general, the relationship sors, determined in air and ethanol, indicated significant (a 5 0.05) between the capacitance (C) and Ka is sensor differences that were largely accounted for with a simple additive correction. Sensor soil water calibration was investigated in four C 5 gKa [1] soils under varying water contents across a 408C temperature range. We found that (i) soil water calibration was significantly (a 5 0.05) where g is a constant dependent on the spacing and different for each soil tested, (ii) there was a significant (a 5 0.05) geometry of the capacitor and both C and g are meatemperature response for all soils, and (iii) the effect of temperature sured in farads (Dean et al., 1987). Thus, the measured varied with soil water content and soil type. Both the soil type and oscillation frequency is related to the Ka such that Ka temperature sensitivities we observed were probably due to the reladecreases with increasing frequency. Empirical calibratively high electrical conductivity (EC) of the soils tested. tions are used to relate u to frequency because of uncertainty in the value of g and in the complex relationship between u and Ka (Whalley et al., 1992). K of soil water content (u, m3m23) is critiThe WCR consists of a printed circuit board concal for determination of local energy and water nected to two parallel, 30-cm-long, 0.32-cm-diam. stainbalance, transport of applied chemicals to plants and less steel rods 3.2 cm apart that act as waveguides. The groundwater, irrigation management, and precision electronic components within the circuit board are enfarming. Several nondestructive methods have been decapsulated in epoxy at the head of the instrument and vised to measure and monitor u including neutron therare configured as a bistable multivibrator. The output malization (Greacen, 1981), electrical resistance (Colman is a square wave with an amplitude of 6 2.5 V DC. The and Hendrix, 1949; Spaans and Baker, 1992; Seyfried, resultant oscillation frequency, which ranges from ≈15 1993), TDR (Topp et al., 1980; Cassel et al., 1994), and to 45 MHz, is linearly scaled down to the order of kiloelectrical capacitance (Robinson and Dean, 1993; hertz to be read by a data logger, which is how it is reNadler and Lapid, 1996). corded. In the past several years, TDR has been shown to When the WCR is installed, the waveguide and soil provide accurate and precise measurements of u that act as a capacitor. Soil between and along the length of are relatively insensitive to soil texture and chemistry the rods affects the capacitance, but the instruments are variations (Zegelin et al., 1992). With TDR, the travel most sensitive to conditions immediately adjacent to the time of electric pulses traveling along a waveguide, rods (Campbell Scientific, 1996). Changes in Ka, which which is directly related to the apparent soil dielectric are primarily due to changes in u, are recorded as constant (Ka), is measured. Since the dielectric constant changes in the oscillation frequency. Results are generof water (80) is very much greater than that of air (1) ally reported in terms of the wave period (P, ms), which or soil solids (3–5), the measured composite Ka is priis the inverse of the frequency, because it increases with marily a function of u. A fairly robust empirical calibrau. Data can be collected in a continuous mode (or at tion developed by Topp et al. (1980) is commonly used any specified interval) and stored on a data logger. No to compute u from Ka. expensive TDR unit is required and soil water content The high cost of TDR has lead to the development information can be collected at considerable cost of alternative means of using Ka to measure u. The use savings. of soil capacitance to determine Ka was explored by Although the basic operation principles for the WCR are the same as those for other recently described capac1 Mention of manufacturers is for the convenience of the reader itance soil water content sensors, there are at least two only and implies no endorsement on the part of the author or USDA. notable differences: (i) the mode of soil–sensor contact M.S. Seyfried and M.D. Murdock, USDA-ARS, 800 Park Blvd., Boise, ID 83712. Received 4 Jan. 1999. *Corresponding author (mseyfrie@ Abbreviations: EC, electrical conductivity; PVC, polyvinyl chloride; nwrc.ars.pn.usbr.gov). TDR, time domain reflectometry; WCR, Water Content Reflectometer. Published in Soil Sci. Soc. Am. J. 65:28–34 (2001).