HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (2) Citing Articles via Google Scholar Google Scholar Articles by Yao, T. Articles by Neuman, S. P. Search for Related Content PubMed Articles by Yao, T. Articles by Neuman, S. P. GeoRef GeoRef Citation Agricola Articles by Yao, T. Articles by Neuman, S. P. Related Collections Water Content Soil Hydrology Field-Scale Studies

ORIGINAL RESEARCH

Neutron Probe Calibration in a Vertically Stratified Vadose Zone

^a Soil Water and Environmental Science Dep., Univ. of Arizona, Tucson, AZ 85721
^b Hydrology and Water Resources Dep., Univ. of Arizona, Tucson, AZ 85721
^* Corresponding author (myao{at}ag.arizona.edu)

Received 30 July 2003.

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS SUMMARY REFERENCES

 
A procedure is described for neutron probe calibration in deep multilayer vadose zones. Calibration equations relating neutron count ratios (CR) to soil water content were developed for the upper 2.5 m of layered soil profiles using soil texture, water content, and neutron probe data. These equations were extended down to a depth of 10 m by relying on neutron probe data to delineate soil texture zones at depth. Data from two constant-flux field infiltration experiments were used to verify the calibration procedure. The water balance for each of nine soil profiles within the infiltrated area was computed using up to four separate calibration equations for each soil profile. The use of two or more texture-based calibration equations greatly improved the agreement between water applied at the surface and water measured within each of the nine layered profiles. Best results were obtained using four such equations, three corresponding to distinct soil horizons and one to interfaces between them (soil texture boundaries). The calibration procedure proposed here is applicable to deep soil profiles when soil texture and bulk density data are only available for the upper soil profile.

Abbreviations: CR, count ratio • MAC, Maricopa Agricultural Center

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS SUMMARY REFERENCES

 
THE NEUTRON PROBE has been widely used in the past 50 yr for measuring water content in field soils. Gardner and Kirkham (1952) and van Bavel et al. (1956) were among the first to do so. Although during the last decade the use of this device may have decreased due to concerns raised about the use of a radioactive source, the neutron probe is still one of the best, most reliable tools for assessing water content in soils (Evett et al., 2003).

To obtain water content from a neutron probe count rate or ratio in a soil, one needs a calibration equation relating these quantities (van Bavel et al., 1961). Count ratio is the count rate measured in soil divided by the count rate measured in a standard, such as a protective shield or a container filled with soil. Count ratios are preferred because their use reduces the effects of temperature and drift from electronics. Ideally one needs different calibration equations for different soil types, access tube materials, and access tube sizes (Tyler, 1988; Klenke and Flint, 1991). The procedures for determining soil calibration equations are well established (Holmes, 1956; Greacen and Hignett, 1979; Greacen et al., 1981; Hignett and Evett, 2002). For layered soils, however, such procedures are rare. Greacen et al. (1981) discussed the effects of layered soils on calibration equations and contributing factors such as bulk density, clay content, and organic matter. More recently, Hignett and Evett (2002) discussed neutron probe calibration in soils with variable clay content and demonstrated the use of a mini calibration method for two extreme soil textures in a field study. However, these procedures were limited to shallow depths and have not been tested for layered soils to greater depth.

Evett et al. (2003) found that using a depth control stand can yield more accurate water content data near the surface. Using the analogy of the air layer affecting neutron counts in the near surface soil, we can expect that two adjacent subsurface layers with significantly different properties will also influence measurements near the layer boundary. This boundary layer effect was studied theoretically by Wilson (1988), but to our knowledge the results were not tested in the field. Probe calibration for layered soils at greater depth is difficult because, as we will show, one may need a different calibration equation for each soil horizon, with additional complications posed by the presence of distinct changes in soil texture (Wilson, 1988). This requires identifying each layer and more detailed sampling, an expensive proposition for deep vadose zones.

It is our hypothesis that separate calibration equations can be developed for each soil horizon in the upper part of the soil profile and that these calibration equations may be used for the deeper soil horizons with similar soil texture. We describe how we developed calibration equations for the upper parts of field soil profiles where soil texture and bulk density data were available and then how we used these equations for neutron probe data from much greater depths, where textural information was not available. We then show that using separate calibration equations for the various soil horizons and for the boundary layers lead to marked improvements in the measurement of water content at our site, as verified by water balance calculations.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS SUMMARY REFERENCES

 
Field Site Description
The research was conducted on a 50 by 50 m plot located within Field F-115 on the University of Arizona Maricopa Agricultural Center (MAC), Maricopa, AZ. The MAC farm is located about 145 km (90 miles) northwest of Tucson and 40 km (25 miles) southwest of Phoenix.

The trickle-irrigated plot is bordered to the south and east by irrigation return canals, to the west by a support zone and access road, and to the north by a flood-irrigated alfalfa field. To apply irrigation water as uniformly as possible, parallel 60-m-long trickle irrigation lines were placed 30 cm apart on the plot surface with drippers on the lines at 30-cm intervals. A heavy-duty pond liner was placed over the plot surface and over the drip lines, resulting in a zero flux upper boundary condition with no evaporation before, during, and after irrigation. As part of the monitoring project, 5-cm-diameter (2-inch, Schedule 40) PVC neutron probe access tubes, tensiometers, and pore water samplers were installed at nine monitoring locations in the 50 by 50 m field site (Fig. 1) . Two water meters (McCrometer, Hemet, CA) with the accuracy ±2% placed in series were used to determine the total amount of water applied to the plot. These two meters were calibrated on site. The mean differences (%) between the meters and the volume changes in the water supply tanks were found to be 1% with a standard deviation of 2%.

View larger version (10K): [in this window] [in a new window]   Fig. 1. Monitoring Locations 1 through 9 for Exp. 1 and 2.

However, because the soil at the site exhibited significant layering during the installation of all neutron probes, we suspected that the use of several calibration equations would be advantageous. We developed texture-based calibration equations by simultaneously collecting neutron probe CRs and soil samples for volumetric water content and texture analysis. During this calibration exercise and during all subsequent neutron probe measurement events, we used a CPN 503DR Hydroprobe (CPN, Martinez, CA) and 16-s count rates. The field standard counts were taken at a near by site at the 2-m depth (5-cm [2-inch]-diameter PVC access tube), and the case counts were taken with the probe resting on dry soil. The site for standard counts was kept dry with a cover consisting of a 3-m-diameter Hypalon pond liner (DuPont, Wilmington, DE), which prevented soil evaporation and prevented rain from entering the plot.

Soil samples were collected at 25-cm depth intervals from the surface down to 2.5 m. Samples representing dry conditions were collected within 0.5 m from the neutron probe access tubes at all nine monitoring locations. At the time of sampling for dry conditions, the soil had not been irrigated for 3 yr. The presence of the pond liner over the main plot prevented rain from entering the soil as well as water loss by surface evaporation. Thus, gravity drainage was the major reason for water content decreases during the 3 yr. Samples representing the wet conditions were collected at seven of nine monitoring locations within 0.5 m of the neutron access pipes after the plot had been irrigated for 15 d at a constant flux of 2.67 cm d^–1.

Count Ratio vs. Texture Class
Volumetric water content data were collected by hand drilling to the desired depth. Subsequently a 200-cm³ (5-cm-diameter, 10.2-cm-long) soil core was collected using a Madera probe (Precision Machine Company, Lincoln, NE; see Evett, 2001 for probe description). Samples were removed from the sampler and immediately sealed in freezer bags to avoid water loss. The samples were taken to the laboratory the same day for water content determination by oven drying. A total of 93 samples were taken for volumetric water content determination and particle-size distribution.

The soil samples were sorted on the basis of their texture as sand, loamy sand, or sandy loam. For each soil sample we also determined the CR measured with the neutron probe at each monitoring site at the time and at the depth the core sample was taken.

Neutron Calibration for the 2.5- to 10-m Depth
Only limited soil texture data were available for the 2.5- to 10-m depth. We hypothesize that soil textures can be differentiated based on neutron CRs in a soil profile after a long period of redistribution, or in a well-drained soil profile with little or no precipitation or irrigation. Therefore we determined the vertical distribution of the sand, loamy sand, and sandy loam for each soil profile from the measured CR values after 3 yr of drainage. Knowing the vertical texture distribution and the appropriate calibration equations from the upper 2.5 m of the soil profile, we were then able to calculate water contents for the full 10-m-deep profiles. For example, as shown in Fig. 2 , there is significant variation in CRs with depth. As will be shown in the Results section, the CR measurements between about 775 and 1000 cm were <1.1. Thus, we postulated that the soil texture was a sand and that the calibration curve of sand could be applied to this layer. If the CR value was >1.25, the calibration equation obtained for sandy loam was applied. Several areas were noted that did not fit the observed patterns due to dramatic transitions of CR in these regions (i.e., in Fig. 2 the depth intervals 400–425, 525–550, and 750–775 cm). In these cases, we subjectively decided on the soil types that best represented these intervals. This same process was repeated for each 25-cm depth interval for each of the nine access tubes at the 9 monitoring locations down to 10 m.

View larger version (42K): [in this window] [in a new window]   Fig. 2. Neutron probe count ratio profile at Monitoring Location 6, 6 d before Exp. 1. The dashed lines represent the texture boundaries by count ratio (CR): for sand, CR < 1.1; for loamy sand, CR 1.1–1.25; for sandy loam, CR > 1.25.

The second experiment was started 26 Mar. 2002 and was terminated 11 Nov. 2002. During the experiment, water was applied at a rate 2.68 cm d^–1. Water was applied 12 times per day for 5 min each time. Neutron probe measurements were taken in a fashion similar to the first experiment. Neutron probe monitoring of the water content was intensive for the first 30 d and continued weekly or monthly until November 2002.

Upon termination of the irrigation experiments, soil water contents were calculated at each 25-cm depth interval for all nine locations and the duration of both infiltration experiments. Soil water storage was then computed by integrating the water contents vs. depth at predetermined time intervals. The depth of water stored in the soil at each of the nine locations was compared at increasing times with the depth of water applied through the drip irrigation system. We judged the performance of the calibration method by the agreement between the two methods of water storage calculation through time.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS SUMMARY REFERENCES

 
Figure 3 and Table 1 present soil texture, neutron probe CRs, and water contents measured down to 2.5 m at the nine monitoring locations. The presence of loose gravel and sand at the 200- to 250-cm depths made sampling difficult and resulted in incomplete data at the 225-cm depth. Mean values, standard deviations, and the values of the half lengths of the 95% confidence intervals, ts, for CRs, water contents, and clay contents are presented in Table 2. These data show that after 3 yr of redistribution, layers classified as sand had a mean CR of 1.05 ± 0.04, while samples taken from these layers had an average volumetric water content of 9% by volume and a clay content of 4.6%. The difference in water contents, clay contents, and neutron probe CRs between these three different textured soils were significant and allowed us to differentiate the soil types at our site based on neutron probe CRs. For the purpose of our study we determined that if a soil had a CR value <1.1 at a given depth after a long time of drainage, it was a sand. All CR values between 1.1 and 1.25 were assumed to be loamy sands, and CR values above 1.25 were sandy loams. The data in Fig. 4 show the water content and CRs measured before irrigation and during redistribution. An increase in water contents with CRs, with significant scatter was observed. We used a linear relationship WC = aCR + b to represent all data in Fig. 4, where a is the slope and b is the intercept. This relationship (dashed line) is quite different from those obtained by Young et al. (1999) and by Thomasson (2001), also shown in Fig. 4. The major reason for these differences is that Young et al. (1999) did not take many samples below 2 m, at which depth sands are more common. They considered the scarce sand data points to be outliers. Thomasson (2001) obtained his calibration equation using the same data set (Young et al., 1999) for the soil in this plot using an iterative analysis, which gave greater weight to the sandy loam near the soil surface. The data in Fig. 4 also show the possibility of dividing three apparent clusters into three linear groups according to their texture groups. This allowed us to develop separate calibration equations for the three textures (i.e., sandy loam, loamy sand and sand), as shown in Fig. 5 .

View larger version (60K): [in this window] [in a new window]   Fig. 3. Soil texture, neutron probe count ratio (CR), and water content (WC) data at nine monitoring locations. There were no samples below 2 m for Sites 8 and 9. Samples were taken after 3 yr of redistribution, before the first irrigation experiment. (Samples taken every 25 cm except at the 225-cm depth.)

View larger version (21K): [in this window] [in a new window]   Fig. 4. Water content vs. count ratio at the nine monitoring locations for the upper 2.5 m of soils. Triangles are sand, squares are loamy sand, and diamonds are sandy loam. The dashed line was fitted through all data. The calibration lines obtained by Young et al. (1999) and Thomasson (2001) are also shown.

View larger version (20K): [in this window] [in a new window]   Fig. 5. Water content vs. count ratio for sandy loam, loamy sand, sand, and their corresponding least squares fitted straight lines.

where C is fractional clay content of the soil (g g^–1), and We (g g^–1) is defined as water content resulting from clay H. Using an average clay content of 13% for sandy loam, 6.7% for loamy sand, and 3% for sand, it follows that the maximum "equivalent water" between sand and sandy loam resulting from the difference in the clay content is only 1.2%. This 1.2% value is much less than the 5 to 10% water content difference between the two calibration equations as shown in Fig. 5. Thus, the clay content is not the sole reason for the different calibration equations for soils with different texture.

Recently Wang et al. (2003) also demonstrated that two neutron probe calibration lines were necessary for satisfactory simulation of water contents during field infiltration experiments. To see if we could just as well have two calibration equations, we combined the loamy sand and sand data, and fitted one calibration line to these data. The corresponding parameters are also listed in Table 3 under "Sand + loamy sand". Note that combining the loamy sand and sand data results in a reduced R² value as compared with separate calibration equations for loamy sand and sand.

Finally we looked at the effects of textural boundaries on neutron probe measurements as proposed by Wilson (1988). The rationale is that neutron CRs near a soil texture boundary are affected by abrupt changes in water content across it. The following procedure was used to identify boundaries with abrupt changes in soil texture. Inspection of the initial neutron probe CR profile showed fairly abrupt changes in CR values at interfaces between soil layers. We postulated that if a CR change is >0.1 for a 25-cm depth increment, then it most probably represents an abrupt change in soil texture, and thus a sharp textural boundary. We then fitted a straight line to CR values vs. water content data collected during infiltration near these sharp textural boundaries. The corresponding best-fit parameters are listed in the bottom row of Table 3.

Validation
The validity of using one or more calibration equations for computing water contents at the site was tested using water balance calculations. For this it is assumed that the same amount of water was applied at each monitoring location. Since there was no evaporation, all the applied water ended up in the soil profile. One should be able to recover this water, except for that which drained below the lowest depth of measurement (10 m). Because the soil had not been irrigated for 337 d and the surface was covered during this time, it is safe to assume that initial percolation rates at the 10-m depth were close to zero. To verify this Graham (2004) calculated the water flux rates at the 10-m depth at each of the nine monitoring sites from the decreases in water content of the soil above the 10-m depth. His data show that the average of the downward fluxes at the nine monitoring locations, 155 d after irrigation of the plot was terminated, was 0.09 cm d^–1, with a standard deviation of 0.03 cm d^–1. This average flux was probably much smaller after 337 d. Even at 0.09 cm d^–1, the downward flux at the 10-m depth is only 3.4% of the surface flux of 2.67 cm d^–1. Once the water from the latest irrigation reached the 10-m depth, water losses below 10 m were considerable and the mass balance approach was no longer valid.

As the water application rate was constant and assumed to be the same at each of the nine locations, the cumulative applied water vs. time relationship is a straight line (bold line in Fig. 6) . Provided the neutron probe was calibrated correctly and measurement errors were negligible, the depths of water present in the soil profile (i.e., the total amounts water stored in the profile) and measured with the neutron probe at each location and time should be equal to the applied amounts (Table 4, Fig. 6).

View larger version (34K): [in this window] [in a new window]   Fig. 6. Total water measured in the soil for the top 10 m using one layer (triangle), two-layer (cross), three-layer (square), and three-layer plus interface (circle) calibrations vs. the cumulative applied water (bold line) at all nine monitoring locations during Exp. 2. Between Days 6 and 20, mean RMSE values for the soil measured, using one-layer, two-layer, three-layer, and three-layer plus interface calibration equations, vs. surface applied water are 0.095, 0.067, 0.071, and 0.039 m water, respectively.

Table 4 and Fig. 6 demonstrate that the use of a single calibration equation (all data in Table 3) leads to very significant differences between applied and measured water. There is a large difference between the water applied (26.6 cm) on Day 10, and the water recovered in the soil (mean of 38.9 cm, and ranging from 21.1 to 52.3 cm) at the nine monitoring locations. With two equations (sandy loam and sand + loamy sand, Table 3), three equations (sand, loamy sand, and sandy loam, Table 3) and four equations (three for each texture and one for layer boundaries) the agreement between applied and recovered water is much better as shown in Fig. 6 and Table 4. The RMSE values for the soil measured, using the calibration equations for one layer, two layers, three layers, and three layers plus interface, vs. the surface-applied water were calculated to be 0.095, 0.067, 0.071, and 0.039 m water, respectively. These RMSE values were calculated using the data from all nine locations between Days 6 and 20, during which time the water balance approach was most valid. The RMSE results confirm that the single equation scenario is the worst, and that the best scenario is the three layers plus interface case (four equations).

The two equations scenario had a similar RMSE value as the three-equations scenario and did offer an excellent alternative for the four-equations scenario. This latter scenario performed even better in terms of the mean storage change than the three equations case (Table 4). Because the water was uniformly applied to the field, we used the CV values in Table 4 as a criterion of showing the dispersion of the calculated water storage change from all monitoring locations. The CV values are slightly better for the three- and four-equations scenarios, with four equations being the best. We also observed that after 10 d, when the average water front is down to about 4 m, the coefficients of variation started to diminish for all four calibration scenarios, with the lowest CV measured for the case of four equations. Considerable reductions in CV were observed for Days 15 and 20. The scenarios using two and three equations result in nearly identical RMSE values of the calculated water storage vs. the applied water. Furthermore the calculated water storage at all nine monitoring locations had similar CV values. Thus, one may conclude that the two-equations case offers a good alternative in estimating the average water storage change. The four-equations scenario provided the best water storage estimates at all nine monitoring locations, as evidenced by the most accurate mean water storage changes, smallest RMSE, and smallest CV values.

	SUMMARY

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS SUMMARY REFERENCES

 
A procedure was described for neutron probe calibration in deep multilayer vadose zones. The procedure combines shallow soil texture and water content data with deep neutron probe readings. We found that using two or more texture-based calibration equations brought about a considerable improvement in agreement between water applied at the surface and water in the soil. Best results were obtained using four calibration equations, three corresponding to distinct soil horizons and one to interfaces between them. Our calibration procedure is applicable to deep soil profiles when soil texture and bulk density data are only available for the upper soil profile.

	ACKNOWLEDGMENTS

 
This work was supported by the U.S. Nuclear Regulatory Commission under contract number NRC-04-97-056. We thank our NRC project manager Thomas J. Nicholson for his support.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS SUMMARY REFERENCES

 

• Evett, S.R. 2001. Exploits and endeavors in soil water management and conservation using nuclear techniques. In Proc. International Symposium on Nuclear Techniques in Integrated Plant Nutrient, Water and Soil Management, Vienna, Austria. [CD-ROM] 16–20 Oct. 2000. International Atomic Energy Agency, Vienna, Austria.
• Evett, S.R., J.A. Tolk, and T.A. Howell. 2003. A depth control stand for improved accuracy. Available at www.vadosezonejournal.org. Vadose Zone J. 2:642–649.[Abstract/Free Full Text]
• Gardner, W.H., and D. Kirkham. 1952. Determination of soil moisture by neutron scattering. Soil Sci. 73:391–401.
• Greacen, E.L., R.L. Correll, R.B. Cunningham, G.G. Johns, and K.D. Nicolls. 1981. Calibration. p. 50–81. In Soil water assessment by the neutron method. CSIRO, East Melbourne, VIC, Australia.
• Greacen, E.L., and C.T. Hignett. 1979. Sources of bias in the field calibration of a neutron meter. Aust. J. Soil Res. 17:405–415.
• Graham, A.R. 2004. In situ characterization of unsaturated soil hydraulic properties at the Maricopa Environmental Monitoring Site. M.S. thesis. The University of Arizona, Tucson.
• Hignett, C., and S.R. Evett. 2002. Neutron thermalization. p. 501–521. In J.H. Dane and G.C. Topp (ed.) Methods of soil analysis. Part 4. SSSA Book Ser. 5. SSSA, Madison, WI.
• Holmes, J.W. 1956. Calibration and field use the neutron scattering method of measuring soil water content. Aust. J. Appl. Sci. 7:45–58.
• Klenke, J.M., and A.L. Flint. 1991. Collimated neutron probe for soil-water content measurements. Soil Sci. Soc. Am. J. 55:916–923.[Abstract/Free Full Text]
• Thomasson, M.J. 2001. Hysteretic, variably saturated, transient flow and transport models with numerical inversion techniques to characterize a field soil in central Arizona. Ph.D diss. The University of Arizona, Tucson.
• Tyler, S.W. 1988. Neutron moisture meter calibration in large diameter boreholes. Soil Sci. Soc. Am. J. 52:890–893.[Abstract/Free Full Text]
• van Bavel, C.H.M., D.R. Nielsen, and J.M. Davidson. 1961. Calibration and characteristics of two neutron moisture probes. Soil Sci. Soc. Am. Proc. 25:329–334.
• van Bavel, C.H.M., N. Underwood, and W. Swanson. 1956. Soil moisture measurement by neutron moderation. Soil Sci. 82:29–41.
• Wang, W., S.P. Neuman, T. Yao, and P.J. Wierenga. 2003. Simulation of large-scale field infiltration experiments using a hierarchy of models based on public, generic, and site data. Available at www.vadosezonejournal.org. Vadose Zone J. 2:297–312.[Abstract/Free Full Text]
• Wilson, D.J. 1988. Uncertainties in the measurement of soil-water content caused by abrupt soil layer changes, when using a neutron probe. Aust. J. Soil Res. 26:87–96.
• Young, M.H., P.J. Wierenga, A.W. Warrick, L.L. Hofmann, S.A. Musil, M. Yao, C.J. Mai, Z. Zou, and B.R. Scanlon. 1999. Results of field studies at the Maricopa Environmental Monitoring Site, Arizona. Rep. NUREG/CR-5694. U.S. Nuclear Regulatory Commission, Washington, DC.

This article has been cited by other articles:

				D. A. Robinson, C. S. Campbell, J. W. Hopmans, B. K. Hornbuckle, S. B. Jones, R. Knight, F. Ogden, J. Selker, and O. Wendroth Soil Moisture Measurement for Ecological and Hydrological Watershed-Scale Observatories: A Review Vadose Zone J., February 25, 2008; 7(1): 358 - 389. [Abstract] [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (2) Citing Articles via Google Scholar Google Scholar Articles by Yao, T. Articles by Neuman, S. P. Search for Related Content PubMed Articles by Yao, T. Articles by Neuman, S. P. GeoRef GeoRef Citation Agricola Articles by Yao, T. Articles by Neuman, S. P. Related Collections Water Content Soil Hydrology Field-Scale Studies