B.H. George & B.

Minasny Soil Water Content

CHAPTER 13. IN SITU MEASUREMENT OF VOLUMETRIC SOIL WATER CONTENT

B.H. George1 and B. Minasny2

1

State Forests of NSW, Northern Research, Research and Development Division. PO Box J19, Coffs Harbour Jetty, NSW 2450.

Department of Agricultural Chemistry and Soil Science, The University of Sydney, NSW 2006.

1.1

Introduction

The amount of the water present in soil is an important factor determining plant distribution and growth. Changes to the presence of water in soil, for prolonged periods, can alter many soil properties and influence classification (e.g. Hydrosols; Isbell, 1996). In the previous chapter we discussed the principles and practical considerations of measuring the movement of water into and through the soil. Here we will discuss the in situ measurement of soil water content and review some of the practical applications in production agriculture and environmental situations. It has long been known that measuring and understanding soil water status allows plant growth to be manipulated for optimum yields (e.g. Allyn and Work, 1941). To increase agricultural production, irrigation of soil has been carried out over centuries (Hillel, 1991). Between 1950 and 1980, the area of irrigated land increased on average by 3 % per annum. This land development, though now slowed to less than 1 % increase per annum, has resulted in a corresponding increase in fertiliser and chemical use (Stewart and Nielsen, 1990). There are many studies investigating catchments and the effect of land management practices on water use and movement (e.g. forest clearing and reafforestation). By measuring local scale (meters) infiltration and water content and how these parameters vary we can model water movement through catchments and estimate likely impacts of significant land management changes. McBratney (ibid) introduces and discusses the importance of scale in environmental studies. Following the previous chapter, here we concentrate on the principles of measuring in situ soil water content and some of the techniques used in the field. There are many techniques available to determine water content in porous media, especially soil. The in situ status of water in soil can be determined on a gravimetric, volumetric or potential basis. Early methods of measuring soil water content depended on physical grab samples being removed from the soil and then weighed, dried at an arbitrary temperature, and re-weighed (Gardner, 1986). Due to the destructive nature of gravimetric determination, techniques such as neutron moderation (NMM), time-domain reflectometry (TDR) and frequency domain (FD), which measure in situ volumetric water content, are widely used.

S.R. Cattle & B.H. George (Eds) describing, Analysing and managing Our Soil. First Edition. Proceeding of the DAMOS 99 Workshop held at The University of Sydney, November 22nd-26th, 1999. Published jointly by The University of Sydney and the Australian Soil Science Society Inc. (NSW Branch). pp. 205-226.

There are three common ways of measuring the water content of porous materials in the field and the laboratory viz: 1. 2. 3. gravimetric a weight based measurement (kg kg-1); potential based on the energy status of the water in the soil (kPa; J kg-1); volumetric determination of the volume of water present in a given volume of soil (m3 m-3)

The preferred method of determination depends on how the information obtained will be used. For example, a plant biologist may prefer to measure and discuss water content with respect to potential, as this is how a plant responds to water in the soil-plant-air continuum. Wetness (w), is the most widely utilised technique (across disciplines) for soil-water content determination, as it is simple and well understood. Conversely, volumetric soil-water content is most commonly used in irrigated agriculture as the reported figures can readily be converted to required volumes of water required for optimum growth. Our primary aim in this chapter is to introduce the concepts of in situ soil-water determination and ensure the validity and accuracy is optimised by erudite technique selection and sound in situ practice. 1.2 Gravimetric determination of soil water

To allow for a comparison between techniques, parameters such as bulk density and relationships such as the soil moisture characteristic require consideration. The bulk density (b) of the soil is an important consideration in converting gravimetric water content, more correctly termed wetness (w) to volumetric soil-water content (). b is the ratio of the mass of solids to the bulk volume of the soil and is expressed in units of mass per volume (Mg m-3). The simplest method for determination of water in soil is by removing a physical sample from the site in question. The sample is weighed, dried and re-weighed with the wetness calculated (Equation 1). The advantage of this technique is that samples are easily acquired and the soil-water content readily determined. Soil water content determined by the gravimetric method is consistently referred to as the standard technique for soil-water measurement. However, the method is destructive in nature, thus precluding repeated sampling at the same location in the field. For more detail, see description of the method and some applications (Gardner, 1986; Reynolds, 1970).
w=

(soil mass

wet

soil mass dry )

(1)

soil mass dry

1.3

In situ measurement of soil water content

Water status of soil can also be determined on a volume basis, (m3 m-3), where is the volume of the liquid phase per unit bulk volume of soil. This is the most popular method of reporting the water status of soil with respect to repeated in situ measurement, especially regarding irrigation scheduling. can be calculated from wetness (w) if the bulk density of the soil is known (or estimated) as shown in Equation 2:

= w

b w
2

(2)

B.H. George & B. Minasny Soil Water Content

Where b is the bulk density of the soil, (Mg m-3), w is the density of water (assume unity for units of Mg m-3), and w is wetness (Mg Mg-1) as defined in Equation 1. Most in situ techniques are field calibrated to account for bulk density effects and report soil water content on a volume basis. The accuracy of the bulk density and wetness determination dictates the accuracy of the calculated volumetric soil-water content as discussed by Gardner (1986). We will now concentrate on several techniques available for the in situ determination of including the NMM and dielectric based techniques in the frequency and time-domains. 1.4 Neutron Moderation Method (NMM)

The neutron moderation method (NMM) is widely used in soil water content measurement studies in Australia and throughout the world. The neutron moderation technique is based on the measurement of fast moving neutrons that are slowed (thermalised) by an elastic collision with existing hydrogen particles in the soil. The transfer of energy from the emitted fast neutron (where their mass is 1.6710-21kg) is greatest when it collides with particles of a similar size. In the soil matrix H+ is a similar mass yielding elastic collisions with emitted high-energy neutrons. Hydrogen (H+) is present in the soil as a constituent of soil organic matter, soil clay minerals, and water. Water is the only form of H+ that will change in content from measurement to measurement. Therefore any change in the counts recorded by the NMM is due to a change in the water with an increase in counts relating to an increase in soil water content. 1.4.1 Methodology

A particular advantage of the NMM technique is the ability to obtain repeated measurements down the soil profile (Fig. 1). In the field, aluminium (Carneiro and de Jong, 1985) or PVC (Chanasyk and McKenzie, 1986) tubes, are inserted into the soil and stoppered to minimise water entry. The installation process requires the use of either a hand auger (generally 1 2 mm undersize) to remove soil, or a drilling rig. The access tube is then hammered into the soil. Some water applied to the access tube during this process can assist with insertion of the aluminium tube into clay soils. The installation should minimise soil compaction whilst ensuring reasonable contact with the surrounding soil. During the augering process soil can be retained for gravimetric determination (w) to be used for site calibration. For a detailed outline and discussion of installation techniques see Williams et al. (1981). Prebble et al. (1981) showed that an infinitely long air-gap (> 2 mm) surrounding a 51 mm diameter tube when saturated (say immediately after irrigation) decreased the measured water content. However, in field situations with careful installation using a suitably sized auger, air-gaps in excess of > 2 mm should be minimised. Where air-gaps are unavoidable, for example in active shrink-swell clay soil, the addition of sand around access tubes does not improve the measurement of soil water content (Cull, 1979). The addition of slurry, made from a mixture of bentonite and/or other clay materials and cement, along the access tube should be minimised (< 2 mm). Larger thickness can introduce a material with different characteristics to the measured soil (Prebble et al., 1981).

[insert Fig. 1 near here]

Increasing the count time for each measurement increases the instrument precision. However, improving the precision will increase the time for measurements. Table 1 shows the 3

increase in count time (CPN 503DR probe, n = 50; George 1999) and the associated error and precision for two extreme conditions with a NMM. The mean count does not differ with increasing time. However, we see that the variation (expressed as the standard deviation and the coefficient of variation) of the count values decreases with increasing time. The range of values decreases in the sand as count time increases. The range of counts in the water is much higher than the sand but the coefficient of variation decreases with the count time and the larger mean values. From these results we recommend that a count of 16 seconds be employed in irrigation scheduling and a count time of 32 seconds for situations (e.g. hourly readings during an infiltration experiment) where precision is more important.

[insert Table 1 near here]

Readings are taken at depths down the profile with a nominated count time (e.g. 16 seconds). Commonly in irrigated production systems, readings from three aluminium tubes are then averaged and soil water content reported as a single reading. The access tubes are generally located within a 10 15 m radius. This aims to counter the effect of spatial variability reducing the value of the measured soil water content data (Cull, 1979). Readings may be taken as a raw count, or a count relative to a reading in a drum of water or in the instrument shield (Greacen et al., 1981). This reading, called the count ratio, is utilised to minimise potential drift in instrument readings. Improved stability of electronics and reduced drift in counting mechanisms in the past fifteen years has diminished the importance of this process. However, instruments differ in their stability (OLeary and Incerti, 1993) and regular normalisation in a large (> 200 L) water drum on a monthly or seasonal basis should be carried out. 1.4.2 NMM Calibration

The need for calibration of the NMM in different porous materials invariably invokes interesting discussion. Neutron meters are commonly provided with standard calibrations for use in common soil types. In Australia, Cull (1979) established a series of standard calibration and currently these calibrations are extensively used in the irrigation industry (pers. comm., P. Cull; Irricrop Technologies International Pty Ltd, Australia). Other research indicates support for a universal calibration encompassing the difference in neutron scattering due to bulk density and texture (Chanasyk and McKenzie, 1986). In irrigated agriculture, in many soil types, farmers who measure changes in soil water content commonly utilise universal calibrations with reasonable success. Success of the universal calibration in scientific studies is limited, with field studies indicating other influences affecting soil water content determination by the neutron moderation method.

[insert Fig. 2 near here]

The NMM calibration generally involves taking neutron readings in the extremes of wet (field capacity) and dry soil and relating this to current in situ wetness (w). Bulk density (b) is either calculated or estimated to yield a neutron soil-water count to known volumetric water content relationship (from Equation 2). The collection of gravimetric samples can involve either careful removal of samples during access tube installation, destructive sampling around access tubes, or sampling from soil near the installed access tubes (Corbeels et al., 1999). Greacen et al. (1981) detailed, in field and laboratory conditions, a calibration procedure for the neutron moisture method in Australian soil. Briefly, NMM readings are taken and then soil is sampled from the immediate region around the access tube. We recommend that at least 5 gravimetric samples (including determination of b) be taken at each depth. To obtain a reasonable range of data from dry to saturated soil, access tubes should be treated accordingly 4

B.H. George & B. Minasny Soil Water Content

and either covered and soil water content reduced by crop water use, or sufficiently irrigated to achieve field capacity. Field calibration of neutron meters is most commonly carried out with a linear equation (from regression analysis) derived for a particular soil type and/or horizon in the form:

= a +bn

(3)

Where is the volumetric water content (m3 m-3); a is a constant (intercept); b is a constant (slope); n is the neutron count or neutron count ratio. Greacen et al. (1981) indicated that correct regression of count (ratio) on water content (water content as the independent variable) reduced the possibility of introducing a bias to the calibration process. The error associated with deriving the water content, dependent on chemical limitations of soil description, is 1.6 % to 3.5 % (Wilson, 1988). A calibration for a NMM in a Vertosol is shown in Fig. 2a. This is the first component of the calibration where the relationship between the measured count ratio and the known (gravimetrically-measured) is determined. In this example: count ratio = 1.42 known + 0.07. To complete the calibration the inverse (Equation 3) is required. Thus, this calibration shows the predicted soil water content from the measured neutron count ratio in a Vertosol near Deniliquin as: = count ratio 0.704 0.053. Webster (1997) discusses the determination of error of the predicted in some detail. Consideration of bulk density (b, Mg m-3) is the major concern in calibrating the NMM in field studies. And is especially important in duplex soil, where there is the potential for significant change in bulk density in the B horizon. An empirical relationship (Greacen and Schrale, 1976) can be used to correct for bulk density effects:
nc = n

(4)

Where nc is the corrected count ratio; n is the count ratio relating to a bulk density (b); and s is the average bulk density for the site calibration. In limited data sets we have found that the inclusion of bulk density correction in the inverse calibration process reduces calibration error in some cases (George, 1999) and not in others (Fig. 2). From the in situ calibration procedure, the bulk density data will be available and Equation 4 should be considered in the eventual calibration equation to predict from NMM readings.

[insert Fig. 3 near here]

The effect of in situ calibration can be significant when compared to available universal calibrations. Fig. 3 shows the determined by a universal equation (derived from Cull, 1979) used widely in the irrigation industry and the same soil water profile determined from an in situ calibration (George 1999) in a Chromosol near Dubbo NSW. The integrated profile water content (0 0.9 m soil depth) differs depending on the calibration employed (Table 2). The in situ calibration is more sensitive to the change in soil water content (larger range of between field capacity reading (January 12) and driest reading (January 27)). In absolute terms, the difference between the universal and in situ calibrations is greatest when the soil is saturated ( 9 % difference). As the soil profile dries the universal and in situ profile differences are smaller (< 3 %). However if the actual is not important, for example, to an irrigator who is interested in the difference between a given

reading and the field capacity of the soil (i.e. is irrigation required, or how long until irrigation is required?), then the universal calibration can be adequately employed. If actual determination is required (e.g. determining farm water use efficiency or in research trials) we recommend that an in situ calibration be carried out. If the information is collected to schedule general irrigation (e.g. commercial surface irrigation) then the cost of developing an in situ calibration needs to be considered against the increased accuracy of determination.

[insert Table 2 near here]

1.4.3 Potential Limitations

A disadvantage of the NMM technique is the radioactive source. In NSW and other Australian states a licence is required to own, operate and store neutron meters. Gee et al. (1976) reported the radiation hazards associated with neutron fluxes in two neutron meters with an activity of 100 mCu. They indicated that safe operation incorporated awareness with respect to time spent in close proximity to the source (i.e. carrying the meter) and neutron escape through the soil surface (especially near surface measurements). Neutron meters are commercially available with differing activities, commonly between 10 and 100 mCu. The activity needs to be considered with respect to the radiation hazard; however greater source activities will yield less variation in recorded neutron counts (Haverkamp et al., 1984). An alternative action is to increase the count time of the meter, however, economically, this is often difficult to justify. Also, radiation-licensing regulations limit the ability to log soil water measurements in many field situations. Another concern regarding wide-spread and continued use of NMM technology is the time taken for readings. As shown in Table 1 the increasing count-time will improve confidence in the recorded soil water content through improving the instrument precision. However, the longer count time will obviously increase the total time for measurement. Finally, the need for calibration is a limitation with NMM technique, as with most soil water measurement procedures. In general irrigation, the need for calibration is reduced due to the managers ability to improve efficiencies in other components of the irrigation system. For example, in surface irrigation large amounts of water (> 1 Ml ha-1) are added with each irrigation, so an error of 5 % in soil water content determination due to calibration will not significantly alter the managers decision as to when to irrigate given ordering time, delivery and volume of water applied. However, in scientific studies we are interested in minimising error and calibration in some form is required. Argument regarding which parameters should be considered continues. Ideally, for a given soil type and conditions (e.g. range of bulk density) calibrations should be available and used. No single database is available for this purpose and site calibration is recommended in long-term and significant research applications. 1.4.4 Positive attributes

The neutron moderation technique is very robust in operation and the field technique is well established. A good standard procedure for installation allows rapid deployment of access tubes and relatively straightforward data collection. There are many NMM instruments in use in Australia for agriculture and other enterprises. Calibration equations for many soils have already been developed (e.g. OLeary and Incerti, 1993; McKenzie et al., 1990; 6

B.H. George & B. Minasny Soil Water Content

Jayawardane et al., 1983) and this background information should assist in ready application in many instances. The neutron technique measures a large volume of soil compared to dielectric techniques in particular. The integration over a large volume of soil can be viewed as a positive aspect of the technique with respect to soil heterogeneity. However, in duplex soil, or where there is a sharp wetting front, the large measured volume can lead to difficulty in data interpretation (Williams et al., 1981). The NMM technique is especially suited for (non-intensive) temporally based measurement through the soil profile, particularly at depth (> 2 m). Where time costs are minimal then the use of the NMM is very cost effective once the equipment is purchased. We envisage continued widespread use of the NMM technique for some years. 1.5 1.5.1 Time-domain reflectometry Introduction

The time-domain reflectometry (TDR) and frequency domain (FD) techniques exploit the change in dielectric properties in soil as the water content changes. The three physical components present in soil are air, water and soil solid. The dielectric (K, unitless) properties of these three components (at 20 C) differ from air (Kair = 1, by definition), to solid (Ksolid 25), and water (Kwater = 80.18) dominating the total dielectric (Weast, 1975). It is this different property of the Kwater that enables the use of the dielectric technique for moisture determination in many porous media, and especially soil. The electromagnetic properties of soil and other porous materials have been studied at length in the past seventy years. Early work of Drake et al. (1930) and Wyman (1930) investigated the measurement of dielectric constants in aqueous solutions. During the 1970s advances in electronics allowed Hipp (1974) to determine the effect of bulk density, soil water content and excitation frequency (30 MHz to 4 GHz) on the measured dielectric of two prepared soil samples in an air-filled coaxial line. The determination of the dielectric constant (K) and other parameters affecting the apparent dielectric (Ka) are detailed by others (Topp et al., 1980; White and Zegelin 1995; George 1999) and are not discussed in detail here. The key equation for the relationship between the travel of the EM wave through the soil is given by:

ct Ka = 2L

(5)

where the apparent dielectric (a unit less ratio of energies) is related to the velocity of propagation (c is the velocity in a vacuum, 3 108 m s-1) multiplied by the time of travel (t, ns) divided by the length of travel (L) along probes embedded in the soil. The travel time of the EM wave along probes buried in the porous media is measured and the Ka calculated. The Ka is then related to either empirically (after Topp et al., 1980) or via various physically-based mixing models (e.g. Whalley, 1993; Ferre et al., 1996). Topp et al. (1982) and Topp and Davis (1985) pioneered field use of the technique with much development occurring in the following two decades concerning hardware and software design and applications of the TDR technique. Instruments may be adapted cable testers (e.g. Zegelin et al., 1992) or dedicated instruments (e.g. Skaling, 1992) operating in a portable or stationary capacity (Baker and Allmaras, 1990; Heimovaara and Bouten, 1990; Herkelrath et al., 1991). 7

1.5.2

TDR calibration

The adaptation of the TDR technique for soil water content measurement occurred in the late 1970s with the seminal work of Topp, Davis and Annan published in 1980. This research linked the measured travel time of an EM wave (Equation 5) with the volumetric soil-water content (, m3 m-3) of different soil types. The relationship, a third-order polynomial, is still the most widely used calibration:
2 = -5.3 10 -2 + 2.92 10 -2 K a - 5.5 10 -4 K a + 4.3 10 -6 K 3 a

(6)

The universal calibration predicted the ( 0.025 m3 m-3) from measured Ka for mineral soil between 10 C < T < 36 C for the range of soil-water contents 0 < < 0.55 m3 m-3 with a variation in b from 1.14 to 1.44 Mg m-3. This equation still forms the basis of most reported by the TDR technique (Topp and Davis, 1985; Zegelin et al., 1989; Zegelin et al., 1992; and Topp et al., 1994). The empirical relationship Ka() is limited by conditions such as dry soil ( < 0.05) where the Ksoil dominates (Zegelin et al., 1992) and in other porous media such as grain and ore (Zegelin and White, 1994). Further questions relating fine-textured soil types and the effect of bound water (Dirksen and Dasberg, 1993), especially in Australian conditions (Bridge et al., 1996), have focused research towards determining a physically-based relationship between measured Ka and reported . A physically-based calibration is preferred in determining the Ka() relationship in soil (Whalley, 1993). However, until now, the extra parameters required (Roth et al., 1990; Bohl and Roth, 1994; Malicki et al., 1996) have deterred most users from employing physicallyderived mixing models and the use of refractive index models for calibration. White et al. (1994), though acknowledging the benefit of such an approach, suggest that most physicallyderived models are in fact semi-empirical. The majority of reported measurements by the TDR technique are determined by the Topp et al. (1980) universal empirical equation (6) or derivatives thereof (e.g. Skaling, 1992). 1.5.3 The effect of bulk soil electrical conductivity: solute transport and salinity

There are soil factors that can affect the use of the Ka in determination of . For example, the bulk soil electrical conductivity (, S m-1) can affect the determination of . If increases, for example in saline soil or soil with increasing clay content, the TDR technique is then susceptible to over-estimation of the (Topp et al. 1988). Perdok et al. (1996) and Bridge et al. (1996), show that increasing can lead to underestimation of . The effect of on the EM wave is complicated and therefore difficult to quantitatively determine. Nadler et al. (1999) show that data presented in the literature is ambiguous and they found that large values (0.8 2.6 S m-1) did not affect the determination of by the TDR technique. To date there is no comprehensive study of this limitation in Australian soil. One clear result of increasing is the eventual total attenuation of the EM wave and the inability of the TDR instrument (or driving software) to determine the end-point (i.e. the point where the EM wave is reflected). If the EM wave is not reflected (Fig. 4) then the cannot be calculated, as determination of t is not possible. This limitation of sensitivity to changing on the TDR waveform can be utilised in other studies, especially solute movement. The TDR technique can, for example, be used in measuring fertiliser movement through soil. Much study is being undertaken to develop a better understanding and application of this phenomenon. For examples, see Topp et al. 8

B.H. George & B. Minasny Soil Water Content

(1988); Kachanoski et al. (1992); Vanclooster et al. (1993; 1995); and Kim et al. (1998). Discussion here, however, is limited to the determination of by the TDR technique, and not application of the technique to solute studies.

[insert Fig. 4 here]

1.5.4 Cable length

An important component in the TDR system is the extension cable (transmission cable) transferring the EM signal from the TDR unit to the embedded probe. The observed waveform actually travels out to the probe and back. The favoured extension cable between the TDR instrument and the probe is the coaxial cable (Topp and Davis, 1985; Heimovaara and de Water, 1993). Commonly, RG-58 type cable is used in multiplexed arrays to a distance of < 30 m. Some systems prefer RG-8 (or similar) lower-loss cable in signal transfer. However, the cost of low-loss cable often limits use, especially in large arrays. To simplify the analysis and reduce impact of increasing cable length on reported (George 1999; Logsdon 2000), we recommend that cables of the same length are employed. 1.5.5 Potential limitations

The TDR technique has limited applications in conditions where the EM wave is attenuated and no reflection obtained. Such conditions can exist in saline soil, soil with large surface areas and where long extension cables are employed in connecting the probe to the TDR unit. The complexity of the K() relationship confuses the application of TDR in saline conditions. Minimal soil disturbance (i.e. small effect of compaction) and good contact between the soil and the probes is essential for accurate in situ measurement. 1.5.6 Positive attributes

The TDR technique is very flexible and is used in many laboratory and field studies. The ability to use the same equipment to measure many points by multiplexing (connecting many probes to one unit) or by carrying from one measurement point to the next allows for temporal and spatial measurement. The technique can be applied to the measurement of other porous materials and the application to solute transport studies is increasing. Finally, the data is easy to transfer to computer by direct downloading, or remotely accessing information via digital cell phone. 1.6 Frequency domain technique (Capacitance)

The frequency-domain technique is similar to that of TDR in that the apparent dielectric (Ka) relationship to soil-water content () is exploited. Malicki (1983), following the work of Thomas (1966), developed a working in situ capacitance-based sensor. Widespread use of frequency-domain (FD) sensors followed development of a down-hole portable instrument (Dean et al., 1987). Other FD sensors utilise short probes in a similar design to the TDR technique (Robinson and Dean, 1993; Hilhorst and Dirksen, 1994). Capacitance (C), measured in Farads (F), is defined as the amount of charge required to increase the voltage by one volt between two plates separated by a known distance containing an insulating material. The relationship to the dielectric can be shown by:

C=

K o A S

(7)

where K is related to C via the interaction of the total electrode area (A) and spacing of the electrodes (S), while o, the permittivity of free space, is constant (White and Zegelin, 1995). Ideally, the soil is situated between two plates (electrodes) forming a capacitor, however in a field situation the design of the capacitance probe is not ideal with either two annular rings (electrodes) placed within a plastic access tube in the soil, or a single electrode inserted into the soil. The measured area is now removed from between the electrodes to outside the access tube. And instrument sensitivity to change in soil water content is reduced (Nadler and Lapid, 1996). The measurement area is considered a fringe field (Thomas, 1966) as the main influence on the oscillating wave is the uniform interior of the access tube. As the soil dielectric inside the access tube should not change, the soil water content will influence the measured capacitance. Thus in a field situation the measured C is determined as:
C = gK

(8)

where C is related to K via a geometrical constant (g) measured in Farads (F) (Whalley et al., 1992). The constant g depends upon electrode spacing, area and orientation of the electrodes in the soil and o. Frequency is, technically, a relatively easy and cheap parameter to measure and the relationship between C and frequency (F, measured in Hz) is exploited. For a down-hole sensor, Whalley et al. (1992), give the simplified relationship as:

F=

1 2 LC

(9)

where L is the inductance. To account for the sensor design a normalised frequency is determined (Paltineanu and Starr, 1997):
Fu =

( Fa Fs ) ( Fa Fw )

(10)

where Fu is the normalised universal frequency, Fa is the frequency measured in air (inside a PVC tube), and Fs is the measured frequency in soil (again inside a PVC tube). This process is used with portable (Dean et al., 1987) and stationary (Starr and Paltineanu, 1998) down-hole FD sensors. 1.6.1 FD calibration

Thomas (1966) indicated a two-step process of calibration with a linear relationship at small values of (< 0.10 m3 m-3), and a semi-logarithmic relationship for larger values (4.5 m3 m-3 < < 45 m3 m-3) of the form:

= 37.7 log C 20.9

(11)

where C is the measured fringe capacitance change to that measured in air. In a laboratory calibration of a FD sensor operating at lower frequency (see Hilhorst and Dirksen (1994) for description), Perdok et al. (1996) found the technique sensitive to changes in b and soil type. Their calibration estimated permittivity with a power relationship dependent on the b. Others have found satisfactory linear relationships (e.g. Mead et al. 1994) and non-linear relationships (e.g. Paltineanu and Starr, 1997) depending on soil type and instrument employed. Calibration 10

B.H. George & B. Minasny Soil Water Content

of the FD technique is especially dependent on the sensor configuration and operating frequency. 1.6.2 FD precision

Dean et al. (1987) and others claim that the FD technique offers an improved precision in determination of compared to other in situ techniques, especially NMM. The theory underlying the frequency shift and relationship to supports this claim. Field studies have indicated however, that the FD technique is susceptible to noise (i.e. error) in determination of . In a fine-loamy mixed thermic Aridic Paleustalf, Evett and Steiner (1995) estimated the root mean square error (RMSE) of determination to be three times that of NMM measurement. This lead Evett and Steiner (1995) to question the applicability of the FD technique for soil water content measurement in heterogeneous soil. Ould Mohamed et al. (1997), using a FD probe (in a fine loamy, mixed, mesic, Typic Eutrochrept) to account for similar textures, considerably reduced the average 95 % confidence interval to 0.0140.019 m3 m-3. This was approximately ten times smaller than errors found by Evett and Steiner (1995). Ould Mohamed et al. (1997) found the FD technique to be more accurate in determination of compared to NMM. Notably, the calibration techniques for the two methods differed, not accounting for the effective NMM measurement area. This issue requires further investigation, especially concerning the different FD systems available. 1.6.3 Influence of air-gaps on measured dielectric

The largest probable source of error in determination with the FD technique is related to airgap occurrence near access tubes that are introduced during installation (Evett and Steiner, 1995; Bell et al., 1987). This could be a particular problem with installation of tubes at drier sites (Evett and Steiner, 1995) or in swelling soils. This corresponds with Tomer and Andersons (1995) comparison of a FD probe and NMM, where the capacitance probe measurement varied near the soil surface. In their discussion regarding the error analysis of , Ould Mohamed et al. (1997) conclude that the difference between FD and NMM determination decreases with an increase in soil , especially at shallow (< 250 mm) depth. This could be accounted for by the escape of neutrons matched by an increase in air-gaps along installed PVC tubes. Air-gaps present in close proximity to the access tube are accentuated by the weighting function of sphere-of-influence (Bell et al., 1987). 1.6.4 Field application of FD technique

Measurement with the FD technique is undertaken by either lowering a sensor into the access tube (Bell et al., 1987; Tomer and Anderson, 1995), placing an array of sensors into the access tubes and logging the output frequency (Buss, 1994; Paltineanu and Starr, 1997), or by insertion of the instrument directly into the soil (Robinson and Dean, 1993; Hilhorst and Dirksen, 1994). Field studies by Robinson and Dean (1993), Starr and Paltineanu (1998) and Wu (1998) indicate that the FD technique is suitable for in situ studies with considerations such as careful tube installation. Tube installation and contact between probe and the soil has been identified as the major constraint in FD implementation in field studies (Malicki, 1983; Bell et al., 1987; Evett and Steiner, 1995; Tomer and Anderson, 1995). 1.6.5 Potential limitations

11

Obtaining a robust calibration for determination is the major limitation for FD systems. Simplification by use of the universal frequency (Equation 10) is an important step. However, given the scope for different FD system configurations and frequencies employed (often proprietary knowledge), instrument and soil calibration is currently required. The small measurement area and sensitivity to bulk density and air gaps can lead to the need for sitespecific calibration. However, with time there will be significant development of calibration curves for specific FD sensors (a good Australian example is the research reported on the enviroSCAN system in current journals). Also, the effect of increasing bulk soil electrical conductivity is not well understood with many instruments and the choice of frequency is important to minimise such sensitivity. In the field, the most pressing concern regarding the acceptance of FD is the installation process and maintaining contact with the immediate surrounding soil. 1.6.6 Positive attributes

The FD technology is simple and flexible in design. This allows for lower cost systems, especially when compared to the current TDR systems. The FD technique is better suited to profile-based measured than the TDR approach and is very well suited to continuous measurement of soil-water content. Also, there are fewer limitations to extension cable length. Finally, there is some scope that the FD technology can be adapted to estimation of soil salinity. 1.7 Data handling and interpretation

Once the soil water content is obtained from the respective instrument the data can be interpreted in: 1. a general spreadsheet (e.g. Excel or an adapted version such as Watsked (CSIRO Forest and Forest Products)): 2. instrument specific software (e.g. EnviroSCAN; Sentek SA, Australia); 3. soil-water specific software (e.g. the Probe (Research Services New England)). Information can be readily displayed down the soil profile (Fig. 1) indicating the amount of water available and activity in the root zone where water is extracted; or temporally, identifying the soil water content at nominated depths or an integrated profile soil water content (Fig. 5). Data in Fig. 5 was continuously collected (EnviroScan) on a 15-minute basis and downloaded to a PC. Initial irrigation was followed by significant rainfall ( 30 mm) and then no further effluent irrigation or rain occurred till the end of the month. The immediate increase in the water content in the soil profile (00.7 m) is evident. Also, the measurement frequency shows the diurnal effect of water extraction and profile water content between January 15 and January 23. Soil-water use during the day is greater (larger negative slope) than night-time. 1.8 Summary

There are three basic ways to measure the presence of water in soil on a weight (wetness), volumetric or potential basis. The method of measurement is best determined by the intended use of the collected information. For in situ measurement of three techniques, the neutron moderation method, time-domain reflectometry and frequency domain, are widely employed. The relative strengths and limitations of each technique need to be considered for 12

B.H. George & B. Minasny Soil Water Content

the task of obtaining . The correct installation, data collection and calibration are essential in providing accurate and precise information. With wise instrument selection and sound field practice, data can be readily presented and it is the correct interpretation of these data that is the ultimate goal of people measuring in situ soil water content. 1.9 References

Allyn, R.B., Work, R.A., 1941. The availameter and its use in soil moisture control: I. the instrument and its use. Soil Science, 51:307-321. Baker, J.M., Allmaras, R.R., 1990. System for automating and multiplexing soil moisture measurement by time-domain reflectometry. Soil Science Society of America Journal, 54:1-6. Bell, J.P., Dean, T.J., Hodnett, M.G., 1987. Soil Moisture Measurement by an Improved Capacitance Technique, Part II. Field Techniques, Evaluation and Calibration. Journal of Hydrology, 93:79-90. Bohl, H., Roth, K., 1994. Evaluation of dielectric mixing models to describe the ()-relation. Symposium and Workshop on Time-domain Reflectometry in Environmental, Infrastructure, and Mining Applications. United States Bureau of Mines, SP 19-94, pp309-319. Bridge, B.J., Sabburg, J., Habash, K.O., Ball, J.A.R., Hancock, N.H., 1996. The dielectric behaviour of clay soils and its application to time-domain reflectometry. Australian Journal of Soil Research, 23:825-835. Buss, P., 1994. Continuous monitoring of moisture in hardwood plantations irrigated with secondary treated effluent. AWWA Recycled Water Seminar, Newcastle Australia. pp180-189. Carneiro, C., De Jong, E., 1985. In situ determination of the slope of the calibration curve of a neutron probe using a volumetric technique. Soil Science, 139:250-254. Chanasyk, D.S., McKenzie, R.H., 1986. Field calibration of a neutron probe. Canadian Journal of Soil Science, 66:173-176. Corbeels, M., Hartmann, R., Hofman, G., Van Cleemput, O., 1999. Field calibration of a neutron moisture meter in Vertisols. Soil Science Society of America Journal, 63:11-18. Cull, P.O., 1979. Unpublished PhD thesis. University of Armidale, NSW Australia. pp241. Dean, T.J., Bell, J.P., Batey, A.J.B., 1987. Soil Moisture Measurement by an Improved Capacitance Technique. Part I. Sensor Design and Performance. Journal of Hydrology, 93:67-78. Dirksen, C., Dasberg, S., 1993. Improved calibration of time-domain reflectometry soil water content measurements. Soil Science Society of America Journal, 57:660-667. Drake, F.H., Pierce, G.W., Dow, M.T., 1930. Measurement of the Dielectric Constant and Index of Refractions of Water and Aqueous Solutions of KCL at High Frequencies. Physical Reviews, 35:613-622. Evett, S.R., Steiner, J.L., 1995. Precision of neutron scattering and capacitance type soil water content gauges from field calibration. Soil Science Society of America Journal, 59:961968. 13

Ferre, P.A., Rudolph, D.L., Kachanoski, R.G., 1996. Spatial averaging of water content by time-domain reflectometry: implications for twin rod probes with and without dielectric coatings. Water Resources Research, 32(2):271-279. Gardner, W.H., 1986. Water content. In: (ed. A. Klute) Methods of Soil Analysis, Part 1. Physical and Mineralogical Methods. Agronomy Monograph No.9 (2nd edn). pp493-544. Gee, G.W., Silver, J.F., Borchert, H.R., 1976. Radiation hazard from Americium-Beryllium neutron moisture probes. Soil Science Society of America Journal, 40:492-494. George, B.H. 1999. Comparison of techniques for measuring the water content of soil and other porous media. Unpublished Master of Science in Agriculture thesis. University of Sydney, NSW Australia. pp 191. Greacen, E.L., Correll, R.L., Cunningham, R.B., Johns, G.G., Nicolls, K.D., 1981. Calibration. In: (ed. E.L. Greacen) Soil Water Assessment by the Neutron Method. CSIRO, Melbourne, pp 50-72. Greacen, E.L., Schrale, G., 1976. The effect of bulk density on neutron meter calibration. Australian Journal of Soil Research, 17:159-169. Haverkamp, R., Vauclin, M., Vachaud, G., 1984. Error analysis in estimating soil water content from neutron probe measurements: 1. local standpoint. Soil Science, 137:78-90. Heimovaara, T.J., Bouten, W., 1990. A computer controlled 36-channel time-domain reflectometry system for monitoring soil water contents. Water Resources Research, 26:2311-2316. Heimovaara, T.J., de Water, E., 1993. A computer controlled TDR system for measuring water content and bulk electrical conductivity of soils. Report No. 41, Laboratory of Physical Geography and Soil Science, University of Amsterdam, pp27. Herkelrath, W.N., Hamburg, S.P., Murphy, F., 1991. Automatic, real-time monitoring of soil moisture in a remote field area with time-domain reflectometry. Water Resources Research, 27:857-864. Hilhorst, M,A., Dirksen, C., 1994. Dielectric water content sensors: time-domain versus frequency domain. Symposium and Workshop on Time-domain Reflectometry in Environmental, Infrastructure, and Mining Applications. United States Bureau of Mines Special Publication SP 19-94, pp23-33. Hillel, D.K., 1991. Out of the Earth. The Free Press. New York, USA. Hipp, J.E., 1974. Soil electromagnetic parameters as functions of frequency, soil density, and soil moisture. Proceedings IEEE, 62:98-103. Isbell, R.F., 1996. The Australian Soil Classification. Australia. pp 143. CSIRO Publishing, Collingwood,

Jayawardane, N.S., Meyer, W.S., Barrs, H.D., 1983. Moisture measurement in a swelling clay soil using neutron moisture meters. Australian Journal of Soil Research. 22:109-117. Kachanoski, R.G., Pringle, E., Ward, A., 1992. Field Measurements of Solute Travel Times Using Time Domain Reflectometry. Soil Science Society of America Journal, 56:47-52. Kim, D.J., Vanclooster, M., Feyen, J., Vereecken, H., 1998. Simple linear model for calibration of time domain reflectometry measurements on solute concentration. Soil Science Society of America Journal, 62:83-89.

14

B.H. George & B. Minasny Soil Water Content

Logsdon, S.D. 2000. Effect of cable length on time domain reflectometry calibration for high surface area soils. Soil Science Society of America Journal, 64:54-61. Malicki, M., 1983. A capacity meter for the investigation of soil moisture dynamics. Zesty Problemowe Postepow Nauk Rolniczych, pp 201-214. Malicki, M.A., Plagge, R., Roth, C.H., 1996. Improving the calibration of dielectric TDR soil moisture determination taking into account the solid soil. European Journal of Soil Science, 47:357-366. McKenzie, D.C., Hucker, K.W., Morthorpe, L.J., Baker, P.J., 1990. Field calibration of a neutron-gamma probe in three agriculturally important soils of the lower Macquarie valley. Australian Journal of Experimental Agriculture, 30:115-122. Mead, R.M., Paltineanu, I.C., Ayars, J.E., Liu, J., 1994. Capacitance probe use in soil water measurements. ASAE Summer Meeting, Kansas City MO, USA. June 20-22, 1994. ASAE paper 942122. Mead et al, 1995 - check Nadler, A., Lapid, Y., 1996. An improved capacitance sensor for in situ monitoring of soil moisture. Australian Journal of Soil Research, 34:361-368. Nadler 1999 - check O'Leary, G.J., Incerti, M., 1993. A field comparison of three neutron moisture meters. Australian Journal of Experimental Agriculture, 33:59-69. Ould Mohamed, S., Bertuzzi, P., Bruand, A., Raison, L., Bruckler, L., 1997. Field evaluation and error analysis of soil water content measurement using the capacitance probe method. Soil Science Society of America Journal, 61:399-408. Paltineanu, I.C., Starr, J.L., 1997. Real-time soil water dynamics using multisensor capacitance probes: laboratory calibration. Soil Science Society of America Journal, 61:1576-1585. Perdok, U.D., Kroesbergen, B., Hilhorst, M.A., 1996. Influence of gravimetric water content and bulk density on the dielectric properties of soil. European Journal of Soil Science, 47:367-371. Prebble, R.E., Forrest, J.A., Honeysett, J.A., Hughes, M.W., McIntrye, D.R., Schrale, G., 1981. Field installation and maintenance. In: (ed. E.L. Greacen) Soil water assessment by the neutron method. CSIRO, Melbourne Australia. pp 8298. Reynolds, S.G., 1970. The gravimetric method of soil moisture determination, part I. A study of equipment, and methodological problems. Journal of Hydrology, 11:258-273. Robinson, M., Dean, T.J., 1993. Measurement of near surface soil water content using a capacitance probe. Hydrological Processes, 7:77-86. Roth, K., Schulin, R., Fluhler, H., Attinger, W., 1990. Calibration of time-domain reflectometry for water content measurement using a composite dielectric approach. Water Resources Research, 26:2267-2273. Skaling, W., 1992. Trase: a product history. In: (eds. G.C. Topp, W.D. Reynolds & R.E. Green) Advances in Measurement of Soil Physical Properties: Bringing Theory into Practice. SSSA Special Publication No 30, pp169-185. Starr, J.L., Paltineanu, I.C., 1998. Soil water dynamics using multisensor capacitance probes in nontraffic interrows of corn. Soil Science Society of America Journal, 62:114-122. 15

Stewart, B.A., Nielsen, D.R., 1990. Scope and objective of monograph. In: (eds. B.A. Stewart & D.R. Nielsen) Irrigation of agricultural crops. Agronomy Monograph No. 30: pp1-4. Thomas, A.M., 1966. In situ measurement of moisture in soil and similar substances by 'fringe' capacitance. Journal of Scientific Instrumentation, 43:21-27. Tomer, M.D., Anderson, J.L., 1995. Field evaluation of a soil water-capacitance probe in a find sand. Soil Science, 159(2):90-98. Topp, G.C., Davis, J.L., 1985. Measurement of soil water content using time-domain reflectometry (TDR): A field evaluation. Soil Science Society of America Journal, 49:1924. Topp, G.C., Davis, J.L., Annan, A.P., 1980. Electromagnetic determination of soil water content: measurements in coaxial transmission lines. Water Resources Research, 16:574582. Topp, G.C., Davis, J.L., Annan, A.P., 1982. Electromagnetic determination of soil water content using TDR: I. Applications to wetting fronts and steep gradients. Soil Science Society of America Journal, 46:672-678. Topp, G.C., Yanuka, M., Zebchuk, W.D., Zegelin, S., 1988. Determination of electrical conductivity using time-domain reflectometry soil and water experiments in coaxial lines. Water Resources Research, 24:945-952. Topp, G.C., Zegelin, S.J., White, I., 1994. Monitoring soil water content using TDR: an overview of progress. Symposium and Workshop on Time-domain Reflectometry in Environmental, Infrastructure, and Mining Applications. Bureau of Mines, SP 19-94, pp56-65. Vanclooster, M., Mallants, D., Diels, J., Feyen, J., 1993. Determining local-scale solute transport parameters using time-domain reflectometry (TDR). Journal of Hydrology, 148:93-107. Vanclooster, M., Mallants, D., Vanderborght, J., Diels, J., Van Orshoven, J., Feyen, J., 1995. Monitoring solute transport in a multi-layered sandy lysimeter using time domain reflectometry. Soil Science Society of America Journal, 59:337-344. Weast, R.C. (ed.), 1975. General physical constants. Handbook of Chemistry and Physics, CRC Press, Cleveland, USA. ppE-61. Webster, R., 1997. Regression and functional relations. European Journal of Soil Science, 48:557-566. Whalley, W.R., 1993. Considerations on the use of time-domain reflectometry (TDR) for measuring soil water content. Soil Science, 44:1-9. Whalley, W.R., Dean, T.J., Izzard, P., 1992. Evaluation of the capacitance technique as a method for dynamically measuring soil water content. Journal of Agricultural Engineering Research, 52:147-155. White, I., Knight, J.H., Zegelin, S.J., Topp, G.C., 1994. Comments on 'Considerations on the use of time-domain reflectometry (TDR) for measuring soil water content' by W.R.Whalley. European Journal of Soil Science, 45:503-508. White, I., Zegelin, S.J., 1995. Electric and dielectric methods for monitoring soil-water content. In: (eds. L.G. Wilson, L.G. Everett & S.J. Cullen) Handbook of Vadose Zone Characterization and Monitoring. CRC Press, Boca Raton, pp343-385. 16

B.H. George & B. Minasny Soil Water Content

Williams, J., Holmes, J.W., Williams, B.G., Winkworth, R.E., 1981. Application in agriculture, forestry and environmental science. In: (ed. E.L. Greacen) Soil Water Assessment by the Neutron Method. pp3-15. CSIRO, East Melbourne. Wilson, D.J., 1988. Neutron moisture meters: the minimum error in the derived water density. Australian Journal of Soil Research, 26:97-104. Wu, K., 1998. Measurement of soil moisture change in spatially heterogenous weathered soils using a capacitance probe. Hydrological Processes, 12:135-146. Wyman, J., 1930. Measurements of the Dielectric Constants of Conducting Media. Physical Review, 35:623-634. Zegelin, S.J., White, I., 1994. Calibration of TDR for applications in mining, grains, and fruit storage and handling. Time-domain Reflectometry in Environmental, Infrastructure, & Mining Applications, US Department of Interior Bureau of Mines, SP 19-94:115-129. Zegelin, S.J., White, I., Jenkins, D.R., 1989. Improved field probes for soil water content and electrical conductivity measurement using time-domain reflectometry. Water Resources Research, 25:2367-2376. Zegelin, S.J., White, I., Russell, G.F., 1992. A critique of the time-domain reflectometry technique for determining field soil-water content. In: (eds G.C.Topp, W.D.Reynolds & R.E.Green) Advances in Measurement of Soil Physical Properties: Bringing Theory into Practice. SSSA Special Publication No 30. pp187-208.

17

 (m3 m-3)
0.25 0.0 0.10 0.20 0.30 0.30 0.35 0.40 0.45 0.50 0.55

depth (m)

0.40 0.50 0.60 0.70 0.80 0.90

12/1 14/1 17/1 19/1 22/1 25/1 27/1

Fig. 1.

An example of a typical soil-water profile in an effluent irrigated Chromosol soil determined by the NMM technique. is determined by an in situ calibration. Measurements are taken from field capacity and after evaporation and soil-water extraction by Eucalyptus globulus spp. globulus trees during the subsequent drying cycle.
(a)
corrected count ratio
0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0.10 0.15 0.20 0.25 0.30 0.30 0.40 0.45

0.7 0.6

(b)

count ratio

0.5 0.4 0.3 0.2 0.1 0 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45

y = 1.42x + 0.07 r2 = 0.926

y = 1.39x + 0.08 r2 = 0.905

known

(m3

m-3)

known

(m3

m-3)

Fig. 2.

Calibration of a NMM in a Vertosol soil (near Deniliquin NSW). (a) is the calibration (with inverse estimation of from NMM count) and (b) calibration accounting for change in bulk density (Equation 4). The inclusion of b does not significantly reduce the error associated with estimation of from the NMM measurement.

18

B.H. George & B. Minasny Soil Water Content

(m3 m-3)
0.25 0.0 0.10 0.20 0.30 0.30 0.35 0.40 0.45 0.50 0.55

depth (m)

0.40 0.50 0.60 0.70 0.80 0.90

12/1 - in situ 12/1 - univ 19/1 - in situ 19/1 - univ 27/1 - in situ 27/1 - univ

Fig. 3.

Soil-water profile with determined by a universal (univ) and in situ calibrations from 0.2 m to 0.8 m depth. The in situ calibration is more sensitive to change in from near field capacity through a drying-cycle.

3000

relative voltage output

2500 2000 1500 1000 500 0 0 200 400 600 800 1000 1200 1400

time (ns)
Fig. 4. Total attenuation of a generated TDR EM wave in a saturated Chromosol soil at Dubbo, NSW. No reflection of the EM wave at the end of the probes means measurement of the travel time (t) is not possible. Thus, in this case was not able to be determined.

19

210

soil water content (mm)

170

rainfall

diurnal variation in soilwater use by the trees

130 90 irrigation

50 10/1 15/1 20/1 25/1 30/1

date
Fig. 5. Soil water content (mm) measured over a 0.7 m soil profile initially receiving effluent irrigation and then significant rainfall. The increased water extraction during the day and the subsequent redistribution of water during the night is evident from January 15 to January 23.

20

B.H. George & B. Minasny Soil Water Content Table 1 Influence of NMM count time on the reported raw counts by a CPN Hydroprobe in a 200 l drum filled with water and another 200 l drum filled with dry sand.

Count Time (seconds)

Mean Count 384.6 390.3 393.8 393.5 396.5 36784.0 36673.4 36722.7 36737.6 36677.0

Standard Deviation 68.6 35.5 19.9 15.7 9.4 825.7 311.5 198.4 142.0 97.0

Range

Coefficient of Variation (%) 17.8 9.1 5.0 3.9 2.4 2.2 0.8 0.5 0.4 0.03

Sand
1 4 16 32 64 304 204 87 67 37 3360 1341 841 672 383

Water
1 4 16 32 64
Table 2

Soil water content profiles (0 0.9 m) determined by a universal NMM calibration and in in situ calibration in an effluent irrigated plantation.

Day of year

Soil profile determined by the in situ calibration (mm H2O) 404.7 350.1 300.6

Soil profile determined by the universal calibration (mm H2O) 372.2 329.7 292.0

Difference (mm)

Difference (%)

121 19 27
1

32.5 20.4 8.6

8.7 6.1 2.9

field capacity following irrigation and rainfall

21