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 (10) Citing Articles via Google Scholar Google Scholar Articles by Kosugi, K. Articles by Katsuyama, M. Search for Related Content PubMed Articles by Kosugi, K. Articles by Katsuyama, M. GeoRef GeoRef Citation Agricola Articles by Kosugi, K. Articles by Katsuyama, M. Related Collections Flow Vadose Zone Processes and Chemical Transport Lysimeter/Rhizosphere Studies

DIVISION S-1—SOIL PHYSICS

Controlled-Suction Period Lysimeter for Measuring Vertical Water Flux and Convective Chemical Fluxes

^a Division of Forest Science, Graduate School of Agriculture, Kyoto Univ., Kyoto 606-8502, Japan
^b Division of Environmental Science and Technology, Graduate School of Agriculture, Kyoto Univ., Kyoto 606-8502, Japan

^* Corresponding author (kos{at}kais.kyoto-u.ac.jp).

	ABSTRACT

TOP ABSTRACT INTRODUCTION THEORY MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSION REFERENCES

 
Unsaturated infiltration of rain and irrigation water and convective chemical transport processes in the vadose zone are poorly understood. This is partly due to the traditional techniques used to sample unsaturated soil water. This study evaluated the performance of a recently developed controlled-suction period lysimeter in the field for more than 400 d. The lysimeter consists of two tensiometers and a porous plate connected to a suction system. The soil matric pressures immediately above the horizontally buried porous plate, and at the same depth in the natural soil profile, are measured at 3-s intervals. The water extraction period is controlled so that the readings of the two tensiometers match. The lysimeter was installed at a depth of 30 cm in a sparse forest. The soil was sandy loam classified as Cambisol. The lysimeter maintained the soil moisture condition in the sampling profile similar to that in the natural soil profile. The mean absolute difference between the two tensiometers was 4.1 cm, with the mean relative difference of 3.5%. Water extraction by the lysimeter did not cause appreciable convergence or divergence in the soil water flow. The water loss due to evapotranspiration (ET) estimated from the water-sampling rate was similar to the evaporation rates measured using water-filled pans and the ET rate estimated by the Thornthwaite method, considering the reduction because of drought. In addition to the water flux, convective fluxes of dissolved silica, nitrate, and ammonium were quantified. As a result, the controlled-suction period lysimeter has great potential to provide quantitative information on water and solute transport in the vadose zone.

Abbreviations: ET, evapotranspiration

	INTRODUCTION

TOP ABSTRACT INTRODUCTION THEORY MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSION REFERENCES

 
THE UNSATURATED INFILTRATION of rain and irrigation water is a major source of ground water recharge, and solute transport in the vadose zone is important for understanding ground water quality. Although the convective chemical transport in the vadose zone can be quantified by measuring both the unsaturated water flux and the solute concentration of flowing water (Brye et al., 1999, 2001, 2002), the number of studies that have accurately measured convective chemical fluxes in the field is still limited. This is partly because the traditional techniques for sampling unsaturated soil water are not necessarily appropriate, considering the mechanism of soil water flow.

One of the traditional techniques for sampling soil water is the tension-free lysimeter. In this method, a horizontally buried pan intercepts infiltrating water, forming a temporary saturated zone over it, and the water drains into a sampling bottle. The tension-free lysimeter collects water only when the soil immediately above the pan has a positive pressure. Therefore, the soil in the sampling profile above the lysimeter is wetter than the soil surrounding the lysimeter. With this matric pressure gradient, water will tend to flow from the lysimeter into the surrounding dry-soil region, resulting in an underestimation of the natural water flux (Chiu and Shackelford, 2000).

In the capillary lysimeter, a wick made of glass or nylon fibers is attached to the base of the water-collecting pan to establish drier conditions above the lysimeter and to lessen the problem of bypass flow around the lysimeter (Holder et al., 1991; Maeda et al., 1999). However, the water-sampling rate by the capillary lysimeter is not necessarily the same as the natural water flux because the capillary suction exerted by the wick is fixed (Gee et al., 2002; Kosugi, 2000). A generalized version of the capillary lysimeter is the tension lysimeter, which is one of the most frequently used water-sampling techniques. In this method, water is collected through a porous cup or plate by applying suction with a vacuum pump instead of a wick. Usually, the suction is fixed at an empirically decided value from about 20 to 40 kPa (Tokuchi, 1999). As a result, the soil moisture in the water-sampling profile depends on the suction applied. In addition, the water-sampling rate of the tension lysimeter is not necessarily the same as the natural water flux. Moreover, it is reported that the solute concentrations in the sampled water depend on the applied suction because water extracted from large pores at low suctions may have composition different from that extracted from micropores (Rhoades and Oster, 1986).

In contrast with these traditional techniques, some recent studies have proposed controlling the suction for extracting unsaturated soil water by referring to matric pressure observations made in the surrounding natural soil profile. The suction control is intended to make the rate of water extraction by the lysimeter the same as the unsaturated water flux in the natural soil profile. Such a controlled-tension lysimeter seems to be the most accurate alternative to methods traditionally used to measure water and convective chemical fluxes in the vadose zone.

In the controlled-tension lysimeter, the suction is manually or automatically adjusted to a target value decided from tensiometer observations in the natural soil profile (Brye et al., 1999; Ozaki, 1999; Ciglasch et al., 2002; Lentz and Kincaid, 2003) or the suction control is made so that the soil matric pressure immediately above the porous plate should be similar to the matric pressure at the same depth in the natural soil profile (Duke and Haise, 1973; van Grinsven et al., 1988; Kosugi, 2000; Kosugi and Katsuyama, 2001, 2002). When the two matric pressures are the same, the rate of water extraction by the lysimeter is expected to be similar to the unsaturated water flux in the natural soil profile, because the water-sampling profile has the same upper (i.e., the flux boundary condition defined by rain and irrigation rate) and lower (i.e., the hydraulic head boundary condition) boundary conditions as the natural profile. Thus, the controlled-tension lysimeter is a method that is sound in accordance with unsaturated-flow theory. Whereas previous studies have conducted statistical analyses on water-sampling efficiency of the new sampling technique, analyses based on long-term field observations involving various storm events were inadequate with respect to whether the lysimeter had maintained a similar matric pressure above the lysimeter to that in the surrounding soil. In addition, previous studies did not report whether the lysimeter had caused convergence or divergence in soil water flow based on intensive matric pressure measurements. The number of studies that have evaluated the seasonal trend in water-sampling amount by a water-balance approach based on a long-term field observation (e.g., Brye et al., 2000, 2001) is still limited.

The purpose of this study is to evaluate the performance of the controlled-suction period lysimeter, developed by Kosugi (2000) and Kosugi and Katsuyama (2001)(2002), by long-term field monitoring with increased number of tensiometers. This device can be regarded as a new type of controlled-tension lysimeter, which controls the water extraction period instead of controlling the suction value. Following an explanatory introduction of the controlled-suction period lysimeter, we show how this lysimeter differs from the controlled-tension lysimeters that have been developed by others. Then, by means of field monitoring in a sparse forest for more than 400 d, we examine the effect of water extraction by the lysimeter on the soil moisture in the sampling profile. We also analyze the spatial variability of the matric pressures between the sampling and natural soil profiles, and the water extraction rate, as compared with the ET loss. The convective chemical fluxes observed for silica, nitrate, and ammonium are also discussed.

	THEORY

TOP ABSTRACT INTRODUCTION THEORY MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSION REFERENCES

 
Controlled-Suction Period Lysimeter
A schematic of the controlled-suction period lysimeter is shown in Fig. 1a . The equipment consists of two tensiometers (DIK-3150, Daiki Rika Kogyo, Tokyo, Japan), and a ceramic porous plate (0675B0.5M2, Soilmoisture Equipment Corp., Santa Barbara, CA) connected to a suction system by a plastic tube. The porous plate is 26.9 cm in diameter, 0.85 cm thick, with an air-entry value of 50 kPa (510 cmH₂O), approximate porosity of 50%, maximum pore size of 6 µm, and a saturated hydraulic conductivity of 3 x 10^–5 cm s^–1. The porous plate is buried horizontally in the water-sampling profile. One tensiometer monitors the soil matric pressure immediately above the plate, _a. The other tensiometer monitors the matric pressure, _b, at the same depth in a natural soil profile, adjacent to the sampling profile. In the sampling profile, infiltrated water is extracted through the porous plate by applying suction so that _a = _b. When _a = _b, the sampling profile has the same upper and lower boundary conditions as the natural profile. As a result, the water-sampling rate is expected to be similar to the unsaturated water flux in the natural profile. In the following subsections, we explain how the lysimeter collects soil water with the minimal disturbance in moisture condition in the sampling profile.

View larger version (22K): [in this window] [in a new window]   Fig. 1. (a) Schematic diagram of the controlled-suction period lysimeter and (b) flow chart of the control system for the suction system by which soil water is extracted (modified from Kosugi [2000]).

Figure 1b is a flow chart of the control system for the suction system. The values of _a, _b, and p_c are continuously monitored at 3-s intervals. In this study, _a, _b, and p_c are described by the water head (1 cm 0.098 kPa) to be consistent with Darcy's equation for soil water flow in which the hydraulic conductivity has the unit of cm s^–1. When _a < _b (i.e., the sampling profile is dryer than the natural profile), the pump is turned off and the valve is opened (only when p_c < –10 cm) to stop water extraction immediately, by releasing the suction in the water-collection container (Fig. 1b). When _a > _b, the sampling profile is wetter than the natural profile, which means that the excess water should be extracted. This operation depends on the air pressure in the water-collection container, p_c. Since the suction applied to the porous plate should be smaller than the air-entry value of the plate, the pump is turned off when p_c is already less than –450 cm. As water is extracted, p_c gradually increases and the water extraction rate gradually decreases. To maintain a high water extraction rate, the pump is restarted when p_c > –400 cm. To ensure smooth pump operation, the release valve is temporarily opened until p_c exceeds –300 cm, after which, the valve is closed and the pump is turned on. Using this method, soil water is not extracted when _a < –450 cm. However, water and solute movement may be negligible under such dry conditions because of the low unsaturated hydraulic conductivity.

Water Extraction by Controlled-Tension Lysimeter
For the most accurate analysis of convective chemical transport processes in the vadose zone, a lysimeter for extracting unsaturated soil water is required to make the water-sampling rate, q_a, equal to the vertical water flux at the same depth in the natural profile, q_b, at all times:

[1]

At the same time, the lysimeter should satisfy the requirement that the matric pressure at the water-sampling point, _a, is equal to the matric pressure at the same depth in the natural profile, _b, at all times:

[2]

The condition described by Eq. [2] is important for avoiding convergence or divergence in the soil water flow, and keeping the soil moisture in the sampling profile similar to that in the natural profile.

The controlled-tension lysimeter usually set the target pressure (i.e., the air pressure applied to the porous plate), p_c(t), at

[3]

where p is a positive constant usually having the value of 2 to 5 kPa (20–51 cm). Based on a simple application of Darcy's low, the water-sampling rate, q_a(t), may be computed as

[4]

where K_s,p and L is the saturated hydraulic conductivity and thickness of the porous plate, respectively. To meet the requirement expressed by Eq. [1],

[5]

Substituting Eq. [3] into [5] yields

[6]

For satisfying Eq. [2], p should meet

[7]

Equation [7] suggests that p cannot be a constant to satisfy the requirements expressed by Eq. [1] and [2] simultaneously, and q_b(t) should be known a priori for the decision of the target pressure p_c(t). The value of q_b(t) depends on the soil hydraulic properties, irrigation intensity, and antecedent soil moisture conditions. Moreover, Eq. [7] suggests that physical properties of the porous plate affect the target pressure. As a result, the vacuum control should be optimized for given soil and porous plate types. The same result is derived when the ratio of p_c(t) to _b(t) is controlled instead of p.

Water Extraction by Controlled-Suction Period Lysimeter
Instead of introducing the target pressure, our method directly compares _a(t) and _b(t) as proposed by Duke and Haise (1973) and van Grinsven et al. (1988). Moreover, to cope with the rapid changes in soil water content that are associated with heavy storms and intensive irrigation, our method uses a strong vacuum (i.e., p_c –450 cm) to extract water when _a > _b. Equation [4] indicates that the smaller p_c attains the greater water extraction rate. Once _a < _b, the water extraction is immediately stopped by releasing suction in the water-collection container. Thus, our method controls the water extraction period instead of controlling the suction itself. Hence, we call the lysimeter developed by Kosugi (2000) and Kosugi and Katsuyama (2001)(2002) the controlled-suction-period lysimeter.

Figure 2 schematically shows the suction control for matching _a(t) with _b(t) during water infiltration (t = 2t to 5t) and redistribution (t = 0 to 2t, and 5t to 10t) processes. Here, t is the interval for monitoring _a(t) and _b(t). When t = t, _a > _b (Fig. 2a), and the vacuum pump is started to reduce p_c to the prespecified minimal value, p_c,min (Fig. 2b). As the water above the porous plate is extracted, _a becomes less than _b at t = 4t, and the release valve is opened to make p_c 0. Because of no water extraction, _a becomes >_b at t = 6t, and the vacuum pump is immediately restarted. At t = 9t, _a again becomes less than _b and the release valve is opened. Through these controls of the suction system, the period for which _a is greater than _b (the Period i indicated by arrows in Fig. 2a) and the period for which _a is smaller than _b (the Period ii in Fig. 2a) are repeated alternately. That is, the water flow divergence from the porous plate toward the surrounding natural soil (i.e., the Period i) and convergence from the surrounding soil toward the porous plate (i.e., the Period ii) are repeated alternately. However, if one can make the difference between _a and _b small, the water-sampling rate may be similar to the vertical water flux at the same depth in the natural profile. This operation of the suction system does not require knowledge of the water flux in the natural profile or physical properties of the porous plate.

View larger version (34K): [in this window] [in a new window]   Fig. 2. Illustration of (a) matching a with b, and (b) controlling air-pressure in the water-collection container, pc. t is the time interval for monitoring a, b, and pc.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION THEORY MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSION REFERENCES

 
Long-Term Monitoring at a Sparse Forest
The controlled-suction period lysimeter was installed at a sampling depth of 30 cm in a sparse mixed broadleaved forest (Fig. 3) , consisting of evergreen (Myrica rubra Sieb.et Zucc. [wax myrtle], Cinnamomum camphora Sieb. [camphor], Lithocarpus edulis Nakai [oak], etc.), and deciduous (Zelkova serrata Makino [zelkova], Styrax japonica Sieb.et Zucc. [storax], Meratia praecox Rehd. et Wils. [Japan allspice], etc.) trees, located in the experimental forest of the Kyoto University (35°02'N, 135°47'E). The study area has a mean annual temperature of 15.1°C, and precipitation of 1465 mm. Although precipitation often occurs as snow during the winter (from December through February), it thaws in the daytime, and the soil does not freeze. The natural soil profile in which _b was measured was about 0.5 m from the center of the sampling profile. Both the sampling and natural soil profiles were located beneath areas without a forest canopy cover. Except during the winter, the ground was covered with grasses, such as Euphorbia humifusa Willd. (spurge), Aster fastigiatus Fisch (daisy), and Oplismenus undulatifolius Roem. et Schult. The soil is sandy loam containing clay of 11.7%, silt of 17.5%, sand of 65.1%, and gravel of 5.7% (maximum diameter is 9.5 mm), and classified as Cambisol (brown forest soil). Water retention characteristics of the sampling profile were measured by the pressure plate method using a soil core sample with a volume of 100 cm³ taken at each depth, and fitted by using a retention model proposed by Kosugi (1994)(1996) (Fig. 4 , Table 1). Table 1 also summarizes the saturated hydraulic conductivity measured with the falling head soil core method (Reynolds and Elrick, 2002), bulk density, and cation-exchange capacity of the soil in the sampling profile.

View larger version (148K): [in this window] [in a new window]   Fig. 3. Cross-sectional view of the porous plate installation.

View larger version (28K): [in this window] [in a new window]   Fig. 4. Observed and fitted water retention curves for soils collected at 5-, 15-, and 25-cm depth in the water-sampling profile. Fitted parameters of the retention model (Kosugi, 1994, 1996) are shown in Table 1.

View this table: [in this window] [in a new window]   Table 1. Parameters s, r, m, and for retention curves plotted in Fig. 4, and measured saturated hydraulic conductivity, Ks, bulk density, b, and cation exchange capacity (CEC).

The lysimeter was pretested from 17 Mar. to 23 Apr. 2000 to evaluate effects of porous plate installation and water extraction on the soil moisture condition in the sampling profile, and examine the performance of the suction control. For the first 18 d of the pretesting period, the suction system was turned off and no water was extracted. For the next 20 d, the suction system was activated and the water-sampling rate was measured. No chemical analysis was conducted on the water sampled during the pretesting period. During the continuous-sampling period from 16 May 2000 to 12 June 2001, precipitation and water accumulated in the water-collection container were sampled 58 times. Water samples were stored in refrigerator with a constant air temperature of 4°C for <24 wk, then filtered through a 0.45-µm-pore cellulose acetate filter. Nitrate and ammonium concentrations were measured by ion chromatography (HIC-6A, Shimadzu, Kyoto, Japan). Dissolved silica concentrations were also measured using an inductively coupled plasma (ICP) emission spectrophotometer (SPS1500VR, Seiko, Chiba, Japan). Silica concentration provides information on soil weathering rate. Moreover, we presumed that the silica concentration may be decided by the residence time of water, hence, can provide information on mean infiltration rate.

Throughout the pretesting period and from 16 May to 19 Sept. 2000 during the continuous-sampling period, the soil matric pressures at the 10-, 20-, and 30-cm depth were monitored in both the sampling and natural soil profiles. The observed data were used to compare vertical distributions of soil matric pressure in the two profiles. From 19 Sept. to 27 Dec. 2000, the soil matric pressures at 20- and 30-cm depths were measured in the sampling profile, the natural profile, and a soil profile between the two (the intermediate soil profile) to analyze how the water extraction by the lysimeter affects the spatial distribution of soil matric pressure (Fig. 5) . After 27 Dec. 2000, the soil matric pressures at 20- and 30-cm depths were measured in both the sampling and natural profiles. All of the matric pressures and the water-sampling rate were recorded at 5-min intervals.

View larger version (20K): [in this window] [in a new window]   Fig. 5. Flat view of the porous plate and tensiometer installation for analyzing convergence or divergence in soil water flow is shown.

To evaluate water loss by ET in the sampling profile, the water storage in the soil layer above the porous plate, S, was computed from the matric pressure data and the soil water retention functions shown in Fig. 4. By subtracting the sampling amount and the change in the soil water storage, S, from the observed precipitation, the ET was estimated. Then, the estimated ET was compared with free-water evaporations from small (20-cm diameter by 10 cm deep) and large (120-cm diameter by 25 cm deep) pans measured at a meteorological station adjacent to, but outside, the experimental forest. The estimated ET was also compared with the ET computed by using the Thornthwaite method (Thornthwaite, 1948), in which ET rate was derived from the observed air temperature and daytime length (Davie, 2002).

Evaluation of Effects of Time Interval for Tensiometer Monitoring
The controlled-suction period lysimeter uses a short time interval for tensiometer monitoring, t. For the purpose of evaluating effects of t on the matric pressure in the sampling profile, a short-term trial was conducted. The site was located in the same sparse forest and 50 m from the long-term monitoring site. Conditions of the lysimeter installation were as explained above; water-sampling depth was 30 cm and a natural soil profile for _b measurement was 0.5 m from the sampling profile. The trial was conducted during water redistribution 11-h after a heavy storm. The duration, total rainfall, and the maximum intensity of the storm were 10 h, 27 mm, and 7.5 mm h^–1, respectively. For the first 15 min, t was fixed at 3 s as we proposed. Then, t was changed to 3 min for the next 50 min. During the trial, _a, _b, and p_c were recorded at 3-s intervals.

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION THEORY MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSION REFERENCES

 
Effects of Time Interval for Tensiometer Monitoring
Figures 6a and 6b show the changes of _a, _b, and p_c observed during the first 15 min of the short-term trial. As soon as _a > _b, the vacuum pump was turned on and p_c fell below –450 cm (Fig. 6b). Due to the strong vacuum, _a became smaller than _b immediately after the start of water extraction (Fig. 6a). Then, the release valve was opened to increase p_c to around 0 cm. These procedures were repeated continuously. As a result, p_c alternated between zero and –450 cm (Fig. 6b). In spite of such cyclic behavior of p_c, the lysimeter succeeded in keeping _a similar to _b (Fig. 6a).

View larger version (55K): [in this window] [in a new window]   Fig. 6. Changes of (a) a and b and (b) pc with the monitoring interval, t, of 3 s, and changes of (c) a and b and (d) pc with t of 3 min.

Effects of Water Extraction on Soil Moisture
During the pretesting period without water extraction, the soil matric pressures at 10- and 20-cm depths in the sampling profile were similar to those at 10- and 20-cm depths in the natural profile, respectively (Fig. 7b,c) . In both profiles, the matric pressures at 10 cm exhibited pronounced diurnal changes, which might be attributable to the effects of soil and air temperature on the matric pressure measurements (van Grinsven et al., 1988) as well as to root water uptake. On Days 91 through 95, the matric pressure at 10 cm was slightly higher in the natural profile than in the sampling profile, which was probably caused by spatial variation in ET. At 30-cm depth, the matric pressure immediately above the porous plate, _a, increased more than did that in the natural soil profile, _b, during every storm (Fig. 7d). For a few days after each storm, _a exceeded _b. That is, in the sampling profile, the infiltrating water formed a temporarily wetter zone over the porous plate.

View larger version (32K): [in this window] [in a new window]   Fig. 7. (a) Precipitation, and matric pressures at (b) 10-, (c) 20-, and (d) 30-cm depths in both the sampling and natural soil profiles for the pretesting period without water extraction.

View larger version (43K): [in this window] [in a new window]   Fig. 8. (a) Precipitation and 1-h moving-average sampling rates, matric pressures at (b) 10-, (c) 20-, and (d) 30-cm depths; and (e) air-pressure in the water-collection container for the pre-testing period with water extraction.

View larger version (72K): [in this window] [in a new window]   Fig. 9. (a) Precipitation and sampling rates, (b) matric pressures at 30-cm depth, and (c) air-pressure in the water-collection container for the storm on 21 Apr. 2000 during the pretesting period with water extraction.

Figure 8c indicates that the soil matric pressure at 20 cm in the sampling profile was similar to that in the natural profile. At 10 cm, the matric pressures in both the sampling and natural profiles underwent pronounced diurnal changes (Fig. 8b). The matric pressure in the natural profile tended to be greater than that in the sampling profile. The tensiometer installed at 10 cm in the natural profile appeared to be affected by the soil and air temperature more than that in the sampling profile. However, the matric pressures resembled each other during each storm. From the results shown in Fig. 8 and 9, we conclude that the controlled-suction period lysimeter succeeded in keeping the soil moisture in the sampling profile at a level similar to that in the natural soil profile.

The 1-h moving-average water-sampling rate had peaks 2 to 12 h after the rainfall peaks (Fig. 8a). The sampling rate was small on Days 100.8 through 102, because the soil was relatively dry before the storm (Fig. 8b,c,d). The large sampling rate on Days 111.7 through 112.5 was attributed to the relatively wet antecedent conditions.

Spatial Variability of the Matric Pressure
Figure 10 shows the changes in the matric pressures in the sampling and natural profiles, and in a soil profile between the two (Intermediate in Fig. 10b and 10c). For the first, second, and fourth storms, the soil matric pressures at both 20 and 30 cm resembled each other in all three positions. The matric pressures in the intermediate soil profile were slightly greater than those in the natural and sampling profiles, indicating slight water divergence from the intermediate profile to both the natural and sampling profiles. Although the matric pressure at 30 cm in the intermediate profile exceeded _a and _b for the third storm (Fig. 10c), the difference was <7 cm. As a result, Fig. 10 indicates that the water extraction by the controlled-suction period lysimeter did not cause appreciable convergence or divergence in the water flow around the lysimeter. The difference for the third storm was attributed to the larger matric pressure values in the intermediate profile, just before the third event (Days 300.7–301.7 in Fig. 10c). When _a and _b decreased to around –100 to –160 cm in drying processes, the matric pressure at 30 cm in the intermediate profile tended to exceed _a and _b by 10 to 30 cm. During the initial stage of infiltration after such drought, the matric pressure at 30 cm in the intermediate profile remained greater than _a and _b.

View larger version (23K): [in this window] [in a new window]   Fig. 10. (a) Precipitation and 1-h moving-average sampling rates, and matric pressures at (b) 20- and (c) 30-cm depths from 23 Oct. to 3 Nov. 2000.

View larger version (38K): [in this window] [in a new window]   Fig. 11. Matric pressures at 30-cm depth in the sampling, a, and natural, b, profiles, and daily precipitation and sampling amount throughout the continuous-sampling period.

During wetting processes after drought, _a tended to be greater than _b (e.g., Days 221.0–222.3, 487.8–488.3, and 508.0–508.6). Similar differences between _a and _b were sometimes found under wetter conditions (e.g., Days 148.0 and 160.3–161.1 in Fig. 11, and Days 100.8 and 110.2 in Fig. 8d). The tensiometer data for the shallower soil layer suggested that, after a drought, the wetting front sometimes advanced more quickly in the sampling profile than in the natural profile. Therefore, we presume that the differences between _a and _b during the rising rim of the matric pressure for some storms were mainly caused by the heterogeneous infiltration.

Similar to the controlled-tension lysimeters proposed by recent studies, the controlled-suction period lysimeter assumes similarity in the infiltration phenomena between the sampling profile and the natural profile for the tensiometer monitoring. Therefore, heterogeneity in the infiltration processes, which can be caused by an existence of preferential flow pathways and soil disturbance during a porous plate installation, might have a big effect on the performance of the lysimeter. To cope with this difficulty, the number of lysimeters might need to be increased so that the average soil water flux can be estimated from the average sampling rates of the lysimeters. Alternatively, the size of the porous plate might need to be increased to capture the representative soil water flux. Radulovich and Sollins (1987) found that variability in sampling amount was reduced with increasing sampling-area of a tension-free lysimeter. Moreover, Brye et al. (1999) concluded that controlled-tension lysimeters produced smaller coefficients of variation for cumulative sampling amount (8.2–36.6%) than tension-free lysimeters tested by Radulovich and Sollins (1987) that produced coefficients of variation between 40.4 and 80.9%. In this study, we examined performances of the new sampling technique based on the detailed measurements of matric pressures. In future research, further investigations based on a measurement replication should be made to analyze the effects of the number and size of the lysimeter.

The comparison of the daily precipitation and sampling amount in Fig. 11 shows that only some of the precipitation was extracted at a depth of 30 cm, except during the winter (i.e., Days 335–426). We discuss the water loss by ET in detail below. During the winter, the cumulative sampling amount for each storm event was similar to, or even slightly greater than, the cumulative precipitation. Errors in measuring precipitation might partly explain the overvalue of the sampling, because the precipitation during the winter often occurred as snow, which might have increased the spatial variation in the water input to the soil surface.

Water Balance and Evapotranspiration Loss
During the continuous-sampling period, the total precipitation and sampling amount were 1594 and 871 mm, respectively (Fig. 12b) . The precipitation was similar to the 30-yr averaged precipitation during the same period (1620 mm). At the start and end of the continuous-sampling period, soil water storage in the water-sampling profile was 95 and 91 mm, respectively. That is, the total loss by ET was 727 mm, which was roughly 80% of the observed evaporation from the water surface in the small and large pans (937 and 898 mm, respectively), and about 70% of the ET (1025 mm) estimated by the Thornthwaite method. The ratio of the total ET to the total precipitation was 46%. For a reference, Suzuki (1980) observed the loss ratio of 43% for a coniferous forest in the study region.

View larger version (39K): [in this window] [in a new window]   Fig. 12. (a) Matric pressures at 30-cm depth in the sampling, a, and natural, b, profiles, (b) cumulative precipitation and sampling amount, (c) 30-d averaged precipitation rate, sampling rate, and change in soil water storage from 0 to 30 cm deep in the sampling profile, S, (d) 30-d averaged loss rate due to evapotranspiration, ET, evaporation rate measured by small and large water-filled pans, Epan, and ET rate estimated by the Thornthwaite method, ETthorn, and (e) 30-d averaged soil water storage throughout the continuous-sampling period.

In Fig 12d, the estimated ET was greater than E_pan and ET_thorn for Days 247 through 321. This period included the heaviest and second heaviest storms during the continuous-sampling period, on Days 251 through 256 (cumulative precipitation 230 mm) and Days 305 through 308 (cumulative precipitation 106.5 mm), respectively (Fig. 12b and Fig. 11). For the heaviest storm, the sampling amount for Days 253 through 255 was underestimated, since the suction system was temporarily turned off on Days 240.8 through 254.4 because of the drought, as discussed above. As a result, ET was overestimated.

The data for the second heaviest storm are shown in Fig. 10. As discussed above, during this event, _a and _b were about the same (Fig. 10c). However, the cumulative sampling amount related to this event was 54.1 mm, which constituted about 50.8% of the cumulative precipitation. This sampling ratio seems somewhat small considering the large amount of precipitation and the relatively wet soil before the storm (Fig. 10b,c). The small sampling amount resulted in the large ET in Fig. 12d. We could not determine the reason for this. The tensiometer measurements suggested that the lysimeter succeeded in maintaining similar soil moisture levels in the sampling, natural, and intermediate profiles (Fig. 10b,c). Intensive observations with an increased number of tensiometers may be required to clarify the water movement around the lysimeter during the heaviest storms.

Convective Chemical Transport
While silica was not detected in the precipitation, its concentration in the sampled water exhibited a clear seasonal trend (Fig. 13a) , which corresponded to the seasonal trend in air temperature (Fig. 13b). The coefficient of determination, R², derived by a linear regression of silica concentration on air temperature was 0.76. This suggests a large effect of temperature on the silica absorption–desorption equilibrium (Drever, 1997; Berner et al., 1998). Moreover, Fig. 13a shows sudden dips in the silica concentration of the water extracted during heavy storms, such as the storms on Days 251 through 256 and 305 through 308. Since large sampling rates were observed for these storms (Fig. 13b), it is likely that infiltrating water with a low silica concentration reached a depth of 30 cm before chemical equilibrium was established. Combining the soil water flux data measured by the controlled-suction period lysimeter with the results of chemical analyses, convective silica flux was quantified (Fig. 13c). Throughout the continuous-sampling period, 6 g m^–2 of dissolved silica was transported at a depth of 30 cm.

View larger version (41K): [in this window] [in a new window]   Fig. 13. (a) Concentration of dissolved silica in precipitation and sampled water, (b) air temperature and daily sampling rate, and (c) cumulative load of dissolved silica throughout the continuous-sampling period.

	CONCLUSION

TOP ABSTRACT INTRODUCTION THEORY MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSION REFERENCES

 
The results of this study indicate that the controlled-suction period lysimeter has great potential as a water-sampling method that provides quantitative information on water and solute transport in the vadose zone. Since the lysimeter minimizes the disturbance of soil moisture in the sampling profile, the physical soil environment in the sampling profile, which affects biological and chemical reactions, is maintained at a similar state to that in the natural soil profile. Therefore, data observed with the controlled-suction period lysimeter could be used effectively to analyze material balances in the vadose zone, and to validate numerical simulations, thereby contributing to the prevention of soil and ground water contamination.

	ACKNOWLEDGMENTS

 
We gratefully acknowledge the support and valuable suggestions of Prof. T. Mizuyama (Lab. of Erosion Control, Kyoto Univ.) for conducting this study. We express our deep gratitude to Prof. T. Mitsuno, Dr. K. Nakamura, and Mr. S. Takeshita (all of Lab. of Hydrological Environment Engineering, Kyoto Univ.), and the staffs of Kyoto University Forests for providing their precious meteorological data. Thanks are due to Mr. H. Ando (Aisin Seiki Co.), who helped us with field observations. This research was partly supported by a grant from the Fund of Monbukagakusyo for Scientific Research (14760101).

Received for publication November 13, 2002.

	REFERENCES

TOP ABSTRACT INTRODUCTION THEORY MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSION REFERENCES

 

• Berner, R.A., J. Rao, and C. Soobum. 1998. Seasonal variability of adsorption and exchange equilibria in soil waters. Aquat. Geochem. 4:273–290.
• Brye, K.R., T.W. Andraski, W.M. Jarrell, L.G. Bundy, and J.M. Norman. 2002. Phosphorus leaching under a restored tallgrass prairie and corn agroecosystems. J. Environ. Qual. 31:769–781.[Abstract/Free Full Text]
• Brye, K.R., J.M. Norman, L.G. Bundy, and S.T. Gower. 1999. An equilibrium tension lysimeter for measuring drainage through soil. Soil Sci. Soc. Am. J. 63:536–543.[Abstract/Free Full Text]
• Brye, K.R., J.M. Norman, L.G. Bundy, and S.T. Gower. 2000. Water-budget evaluation of prairie and maize ecosystems. Soil Sci. Soc. Am. J. 64:715–724.[Abstract/Free Full Text]
• Brye, K.R., J.M. Norman, L.G. Bundy, and S.T. Gower. 2001. Nitrogen and carbon leaching in agroecosystems and their role in denitrification potential. J. Environ. Qual. 30:58–70.[Abstract/Free Full Text]
• Chiu, T.F., and C.D. Shackelford. 2000. Laboratory evaluation of sand under-drains. J. Geotech. Geoenviron. Eng. 126:990–1001.
• Ciglasch, H., K. Aoungsawad, W. Amelung, K. Pansombat, T. Sombuttun, C. Tarn, S. Totrakool, and M. Kaupenjohann. 2002. Fluxes of agrochemicals in an upland soil in Northern Thailand as investigated with novel techniques. Paper No. 1566. In Abstracts—17th World Congress of Soil Science. Bangkok, Thailand. 14–21 Aug. 2002. Paper no. 1566. Kasetsart Univ., Bangkok, Thailand.
• Davie, T. 2002. Fundamentals of hydrology. Routledge, New York.
• Drever, J.I. 1997. The geochemistry of natural waters. Prentice Hall, Upper Saddle River, NJ.
• Duke, H.R., and H.R. Haise. 1973. Vacuum extractors to assess deep percolation losses and chemical constituents of soil water. Soil Sci. Soc. Am. Proc. 37:963–964.
• Gee, G.W., A.L. Ward, T.G. Caldwell, and J.C. Ritter. 2002. A vadose zone water fluxmeter with divergence control. Water Resour. Res. 38:10.1029/2001WR000816.
• Holder, M., K.W. Brown, J.C. Thomas, D. Zabcik, and H.E. Murray. 1991. Capillary-wick unsaturated zone soil pore water sampler. Soil Sci. Soc. Am. J. 55:1195–1202.[Abstract/Free Full Text]
• Kosugi, K. 1994. Three-parameter lognormal distribution model for soil water retention. Water Resour. Res. 30:891–901.
• Kosugi, K. 1996. Lognormal distribution model for unsaturated soil hydraulic properties. Water Resour. Res. 32:2697–2703.
• Kosugi, K. 2000. A new sampling method of vertical infiltration water in unsaturated soil without disturbing soil moisture condition. (In Japanese with English summary.) J. Jpn. Soc. Hydrol. Water Resour. 13:462–471.
• Kosugi, K., and M. Katsuyama. 2001. Measurements of groundwater recharge rate and unsaturated convective chemical fluxes by suction controlled lysimeter. p. 19–24. In H. Gehrels et al. (ed.) Proc. Int. Symp., Impact of human activity on groundwater dynamics. Maastricht, The Netherlands. 18–27 July 2001. IAHS Publ. no. 269. IAHS Press, Oxfordshire, UK.
• Kosugi, K., and M. Katsuyama. 2002. Measurements of convective chemical fluxes in vadose zone by suction controlled lysimeter 965–1-9 In Proc., 17th World Congress of Soil Science. Bangkok, Thailand. 14–21 Aug. 2002. Kasetsart Univ., Bangkok, Thailand.
• Lentz, R.D., and D.C. Kincaid. 2003. An automated vacuum extraction control system for soil water percolation samplers. Soil Sci. Soc. Am. J. 67:100–106.[Abstract/Free Full Text]
• Maeda, M., B.C. Liyanage, and Y. Ozaki. 1999. Water collection efficiency of wick samplers under steady state flow conditions. Soil Sci. Plant Nutr. (Tokyo) 45:485–492.
• Ozaki, Y. 1999. Bury-type lysimeter to collect infiltrated soil water. p. 9–19 In Proc. Symp., study on practical problems on monitoring of NO₃–N leaching from crop field. (In Japanese.) Tsukuba, Japan. 4–5 Nov. 2000. National Agriculture Research Center, Tsukuba, Japan.
• Radulovich, R., and P. Sollins. 1987. Improved performance of zero-tension lysimeters. Soil Sci. Soc. Am. J. 51:1386–1388.
• Reynolds, W.D., and D.E. Elrick. 2002. Falling head soil core (tank) method. p. 809–812 In J.H. Dane et al. (ed.) Methods of soil analysis. Part 4. SSSA Book Ser. 5. SSSA, Madison, WI.
• Rhoades, J.D., and J.D. Oster. 1986. Solute content. p. 985–1006 In A. Klute (ed.) Methods of soil analysis, Part 1. 2nd ed. SSSA, Madison, WI.
• Suzuki, M. 1980. Evapotranspiration from a small catchment in hilly mountains (I) Seasonal variations in evapotranspiration, rainfall interception and transpiration. J. Jpn. For. Soc. 62:46–53.
• Thornthwaite, C.W. 1948. An approach toward a rational classification of climate. Geogr. Rev. 38:55–94.
• Tokuchi, N. 1999. Sampling method of solute in soil. p. 186–188 In Methods of forest environment measurement. (In Japanese.) Hakuyusha, Tokyo, Japan.
• Tsutsumi, T. 1987. Material cycle in forest. (In Japanese.) University of Tokyo Press, Tokyo, Japan.
• van Grinsven, J.J.M., H.W.G. Booltink, C. Dirksen, N. van Breemen, N. Bongers, and N. Waringa. 1988. Automated in situ measurement of unsaturated soil water flux. Soil Sci. Soc. Am. J. 52:1215–1218.[Abstract/Free Full Text]

This article has been cited by other articles:

				J. Mertens, V. Tuts, J. Diels, J. Vanderborght, J. Feyen, and R. Merckx Design and Testing of a Drop Counter for Use in Vadose Zone Water Samplers Vadose Zone J., April 14, 2008; 7(2): 434 - 438. [Abstract] [Full Text] [PDF]

				L. Weihermuller, J. Siemens, M. Deurer, S. Knoblauch, H. Rupp, A. Gottlein, and T. Putz In Situ Soil Water Extraction: A Review J. Environ. Qual., October 24, 2007; 36(6): 1735 - 1748. [Abstract] [Full Text] [PDF]

				J. Mertens, J. Diels, J. Feyen, and J. Vanderborght Numerical Analysis of Passive Capillary Wick Samplers prior to Field Installation Soil Sci. Soc. Am. J., January 1, 2007; 71(1): 35 - 42. [Abstract] [Full Text] [PDF]