Published online 9 August 2007
Published in Soil Sci Soc Am J 71:1469-1472 (2007)
DOI: 10.2136/sssaj2007.0009N
© 2007 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
SOIL PHYSICS NOTE
Automation and Use of Mini Disk Infiltrometers
Matthew D. Madsena and
David G. Chandlera,*
a Department of Plants, Soils and Climate, Utah State Univ., Logan, UT 84322
* Corresponding author (dgc{at}ksu.edu).
 |
ABSTRACT
|
|---|
Measurement replication and objectivity of field soil hydraulic properties can be increased through automation. The goals of this study were to test two automated mini disk infiltrometers (AMDI). Both devices were fitted with differential pressure transducers connected to compact data loggers. Instrument design, method of calculation, and soil moisture condition all affected measured unsaturated hydraulic conductivity [K(h)] and sorptivity (S) at pressure head h = –2.0 cm. We found that the type of AMDI with a capillary tube head control can be operated at inclination angles up to 25° and returned the least variance in K(h) and S if data were not partitioned for calculation. Changing the initial soil moisture content from 0.07 to 0.26 m3 m–3, however, was found to influence K(h) calculation by up to 50% for a silt loam soil. The K(h) measured by the type of AMDI with a bubble chamber head control and larger disk diameter was less dependent on soil moisture content, but more sensitive to inclination.
Abbreviations: AMDI, automated mini disk infiltrometer • AMDIv1, automated mini disk infiltrometer Version 1 • AMDIv2, automated mini disk infiltrometer Version 2 • MDI, mini disk infiltrometer • MDIv1, mini disk infiltrometer Version 1 • MDIv2, mini disk infiltrometer Version 2
|
INTRODUCTION
|
|---|
INFILTRATION OF WATER INTO soil determines hydrologic response and ecological function in many environments and is a critical parameter in understanding land management and ecosystem modeling. Tension infiltrometry is an efficient technique that can provide information on soil hydraulic properties, including hydraulic conductivity, sorptivity, and macroporosity (Reynolds and Elrick, 1991; Zhang, 1997). Several designs for tension infiltrometers have been proposed (e.g., Perroux and White, 1988) and are commercially available. Infiltration rate measurements for large disk tension infiltrometers have been successfully automated with pressure transducers (Ankeny et al., 1988; Casey and Derby, 2002).
Field research in remote, nonagricultural settings places several constraints on the use of large disk infiltrometers, including manual transportation of all research equipment, limited access to water, sloping soil surfaces with locally extreme microtopography, and low, dense branch architecture of desert shrubs. We devised a scheme for automating a small, commercially available instrument, a mini disk infiltrometer (MDI), to avoid the tedium and discomfort of lying on the desert floor, to allow instrument placement in visually obscure locations, and to increase measurement replications and accuracy.
The purpose of this study was to test the response of two models of the automated mini disk infiltrometer (AMDI) to field conditions including use in sloping terrain, variable voltage supply, and wet and dry initial soil moisture conditions.
|
MATERIALS AND METHODS
|
|---|
Design
The MDI has been manufactured (Decagon Devices, Pullman, WA) in two models, each of which requires slightly different automation schemes. The original MDI design (Version 1; MDIv1), having a disk diameter of 3.2 cm, was automated by fitting the two ports of a differential pressure transducer (SenSym ASCX01DN, Honeywell, Freeport, IL) into a pair of holes bored in the rubber stopper at the top of the reservoir tube (Fig. 1a
). A 20-cm section of polyethylene tubing was inserted in the stopper below the transducer pressure port B and extended to about 25 mm above the air inlet tube near the bottom of the infiltrometer. The dimensions of the air inlet tube determine operating tension. We present the results of 2.0-cm supply tension for AMDIv1 infiltrometers, since this is recommended as the most broadly applicable model by the manufacturer.
View larger version (29K):
[in this window]
[in a new window]
|
Fig. 1. Diagram of (a) original style automated mini disk infiltrometer (AMDIv1) and (b) current style automated mini disk infiltrometer (AMDIv2).
|
|
Head control for the current MDI model (Version 2; MDIv2) is provided by a bubbling chamber on the top of the instrument and necessitated a different setup (Fig. 1b). The reservoir of the MDIv2 was connected to a differential pressure transducer with external tubing (1.59-mm [1/16-inch] diameter Nalgene) and threaded steel fittings (10–32 X.170 barb 5/16 O-ring fitting straight connector, Pneumadyne, Plymouth, MN) 9.5 cm from the top and 3 cm from the bottom of the infiltrometer. The supply tension for the MDIv2 is determined by the depth of submergence of the suction control tube in the bubble chamber, minus the distance from the bottom of the Mariotte tube to the porous disk. We present results of 2.0-cm supply tension for the AMDIv2 infiltrometers. The MDIv2 is filled by removing the elastomer housing that retains the 4.4-cm-diameter sintered stainless disk to the bottom of the water reservoir. We fit a small valve near the bottom of the infiltrometer to release pressure during reassembly.
Infiltrometer support stands were fabricated from 10-cm-long. 3.2-cm-diameter polyvinyl chloride tubes. Three support legs, made of 3-mm rod stock, are held in grooves in the tubing by a metal hose clamp (Fig. 2a
). A 6- by 9-mm wing screw was installed through the side of the stand to suspend the infiltrometer preceding infiltration experiments and support the infiltrometer during measurements.
View larger version (11K):
[in this window]
[in a new window]
|
Fig. 2. (a) Infiltrometer stand and (b) schematic wiring diagram to connect four ASCX01 DN pressure transducers (PT1–PT4) with input power (vi) from a 6-V direct current (DC) battery. Output voltage from the transducers (vo) recorded with a Hobo U12 data logger (DL).
|
|
Transducer output (vo) was recorded with a U12 4-external channel outdoor/industrial Hobo data logger (Onset Computer Corp., Bourne, MA). To avoid out-of-range signal voltage, transducers were powered with an external 6-V battery. Four transducers, allowing measurements with four infiltrometers at one time, one data logger, and the battery were connected with simple wiring manifolds (Fig. 2b) and installed in a plastic tool box, which also accommodated the infiltrometers and their support stands for transport and storage.
Calculations
Conversion of vo to volume of water in the infiltrometer (V) at time t is calculated by scaling the maximum vo (vmax) at the beginning of each trial measurement run and the minimum vo (vmin) at the end of each trial to the total volume of water discharged from the infiltrometer during the total measurement period (Vtot):
 | [1] |
The scales on the AMDIs overrepresent the volume in the reservoir due to displacement of water by the internal tubing. Vtot is calculated as
 | [2] |
where L is the distance between the initial water level in the reservoir and the bottom of the poly tube in AMDIv1 or the lower pressure transducer port in AMDIv2, R is the inside radius of the infiltrometers (1.27 cm), and r is the outside radius of the poly tube (0.32 cm) for the AMDIv1 or the reservoir bubbling tube (0.24 cm) for the AMDIv2.
The values of S and K(h) were calculated following the approach of Zhang (1997), using the form of the two-term cumulative infiltration equation suggested by Decagon Devices:
 | [3] |
where I is cumulative intake per unit area, t is time, and C1 and C2 are parameters related to soil K(h) and S, respectively. The values of C1 and C2 were obtained by plotting cumulative infiltration vs. the square root of time and fitting the data to a second-order polynomial equation using spreadsheet software (e.g., Excel). Note that the first- and second-order coefficients correspond to C1 and C2, respectively. This reverses the order of the polynomial terms from equations commonly presented by other researchers (e.g., Vandervaere et al., 2000a, 2000b; Zhang, 1997; Warrick, 1992), but facilitates the use of a spreadsheet to fit the constants to the terms. The values of K(h) and S were calculated as
 | [4] |
 | [5] |
The dimensionless coefficients A1 and A2 were calculated according to Zhang (1997, Eq. [20–22]). The value of A1 depends on water content and disk diameter. We used 4.17 and 3.97 for dry conditions (initial water content 
i = 0.07) and 3.53 and 3.36 for wet conditions (i = 0.26) for the MDIv1 and MDIv2, respectively. The value of A2 depends on disk diameter and soil texture. For A2, we used 3.33 and 2.48 for the MDIv1 and MDIv2, respectively, for the Moab soil, and 10.88 and 8.09 for Greenville silt loam.
Calibration and Testing
Measurement procedures followed instructions in the MDI user manuals. In addition, the AMDI was briefly suspended over the soil before running the infiltration experiment to aid in identifying vmax and the start time of a trial during data analysis. Data loggers were programmed to record transducer output voltage at a 1-s interval during testing and calibration and at 3-s intervals in the field to conserve data storage space.
Laboratory testing was performed with soil from a field site near Moab, UT. The soil is described as a Rizono–rock outcrop complex (loam; mixed, calcareous, mesic Lithic Ustic Torriorthents). Samples were obtained from the 0- to 10-cm depth; sorted through a 20.3-cm-diameter, 2-mm sieve; mixed thoroughly; and hand compacted to a depth of at least 20 cm in a tiled container. We first measured the sensitivity of the AMDI signal response to transducer input voltage (vi) by suspending an AMDIv1 filled to 89 mL and varying vi from 0 to 9 V. The effect of infiltrometer placement angle on vmax and calculated K(h) was tested for AMDIv1 and AMDIv2 by performing infiltration experiments with the infiltrometers at six different angles from vertical: 0, 5, 10, 15, 20, and 25°. All laboratory trials were conducted at room temperature between 20 and 25°C, at a soil moisture content of 0.35 m3 m–3 with room temperature water. Most trials were completed within 10 min.
Additional tests were conducted to assess the comparability of results from the AMDIv1 and AMDIv2 and what opportunities automation might provide for improvement over manual measurement. Two sets of experiments were performed at the Utah Agricultural Experiment Station Greenville Experimental Farm in North Logan on Millville silt loam (coarse-silty, carbonatic, mesic Typic Haploxerolls). To reduce spatial variability among measurements, we performed all experiments on a 1-m2 plot of recently tilled soil. A thin layer (
2 mm) of previously washed dry sand was used to improve contact between the disk and soil and to prevent disk clogging. Two successive measurements were performed in each set of experiments. The first test was performed on initially dry soil (0.07 m3 m–3) and the successive experiment was performed on initially wet soil (0.26 m3 m–3). In the first experimental set, three research assistants manually recorded visual observations of changes in infiltrometer reservoir volume in an AMDIv2 while electronic data were recorded by the automated technique. The infiltrometers were supported by hand because use of the support stand obscured the reservoir scale from sight. Automated vo measurements were made at 1-s intervals, and manual V(t) records were taken at 5-s intervals for the first minute and 10-s intervals thereafter. Each investigator measured three locations for a total of nine replicates for both dry and wet soil conditions. In the second experiment, the same experimental design was used as above to compare K(h) and S values calculated from AMDIv1 and AMDIv2 measurements. Seven replicates were made under both wet and dry initial conditions. Support stands were used.
We took two approaches in calculating K(h) and S: (i) by fitting the total cumulative infiltration data of each measurement to Eq. [3]; and (ii) by partitioning the cumulative infiltration into the nonlinear period dominated by S and the near-steady-state period dominated by K(h). Measurements were compared with the mean and CV along with Pearson correlation coefficient (R) and statistical significance (P).
|
RESULTS AND DISCUSSION
|
|---|
In the laboratory tests on Moab soils using an AMDIv1 with a full water reservoir, we found that transducer vo is linear to vi from 8.9 V, at which point the maximum data logger input voltage (2.5 V) was exceeded, down to 1.5 V. Additionally, vmax decreased with the cosine of the angle of infiltrometer inclination from 0 to 10° for AMDIv1 and AMDIv2 (Fig. 3a
). The relationship does not hold, however, for greater angles of inclination due to the more complex system trigonometry, and an empirical calibration is required (Eq. [1]). Infiltration trials on an inclined sand bed demonstrated the capacity of the AMDIv1, when properly calibrated, to make consistent measurements of K(h) at inclinations up to at least 25° from vertical (Fig. 3b). For the AMDIv2, both vo and h are dependent on the angle of inclination and the axial orientation of the infiltrometer. Pressure ports oriented upslope will record lower pressure heads and consequently lower vo when compared with pressure ports oriented downslope at the same angle. Whereas the effect on vo could be calibrated for consistent orientation during use of the AMDIv2, we observed that inclination also affected h in the bubbling chamber. We calculated a decrease of 0.5 cm from a nominal h of –2.0 cm as the AMDIv2 was inclined from 0 to 25° from vertical. Accurate field measurements of K(h) with the AMDIv2 would require additional effort to adjust the bubble chamber water level to account for inclination and care to maintain consistent orientation of the pressure ports relative to the slope.
View larger version (14K):
[in this window]
[in a new window]
|
Fig. 3. Laboratory results of (a) maximum output voltage (vmax) for original and current style infiltrometers (AMDIv1 and AMDIv2) suspended at different angles and (b) calibrated unsaturated hydraulic conductivity [K(h)] values calculated from infiltration measurements with an AMDIv1 performed at 0 to 25° from vertical.
|
|
The most significant correlations between the manual MDIv2 and the automated AMDIv2 data collection and between methods of calculation were for K(h) of wet soil (partitioned and complete) and S of dry soil (complete), underscoring the control antecedent soil moisture conditions exert on these two aspects of infiltration calculations (Table 1). The poorer agreement between S calculated from manual and automated records for partitioned data under dry initial conditions is evidently due to the greater influence of more precise automated data for short time measurements on partitioned calculation of S.
View this table:
[in this window]
[in a new window]
|
Table 1. Pearson correlation coefficients (R) plus statistical significance (P) for comparison of manual (M) vs. automated (A) measurements performed with the current style automated mini disk infiltrometer (AMDIv2). Measurements of unsaturated hydraulic conductivity, K(h), and sorptivity, S, for pressure head h = –2.0 cm, obtained by partitioning total infiltration data (partitioned) or through measurement of infiltration data as a whole (complete).
|
|
We found that experimental conditions, instrument design, and the method of calculation all affected the results for K(h) and S. The most consistent results (lowest CVs) for K(h) for both wet and dry initial conditions and for S for dry conditions were obtained by using the complete data records from the AMDIv1 for calculation (Table 2). The difference between K(h) results for wet and dry conditions from the complete records from AMDIv1, however, was nearly 50%, indicating a sensitivity of this approach to initial soil moisture. On the other hand K(h) sensitivity to soil moisture was least for the AMDIv2 using complete records, but the CV for dry conditions was three times that of the AMDIv1. We expect that these differences are due to the change in disk diameter between the two devices. For similar reservoir volumes, a larger disk, and hence greater measurement volume, would require proportionately greater allocation of water to S. The subsequent flux to the wetting front under dry conditions, however, would be less influenced by the suction at the wetting front.
View this table:
[in this window]
[in a new window]
|
Table 2. Results of unsaturated hydraulic conductivity, K(h), and sorptivity, S, mean and CV and comparisons between original style automated mini disk infiltrometer (V1) and current style (V2) for pressure head h = –2.0 cm.
|
|
Field Application
We chose the AMDIv1 for our field studies (Lebron et al., 2007) because it was simple to operate, obviated any concern about placement angle, and gave consistent results. The AMDIv1 proved reliable in the field under a variety of weather conditions and levels of operator expertise. For example, quadratic equations fit to infiltration data vs. time from one field campaign at the Moab site with 100 measurements resulted in an average r2 of 0.995. Also, the number of measurements a person was able to perform in a day increased at least fourfold, up to 80 to 100 infiltration measurements per person-day. Automation and use of the tripod stands also allowed infiltrometer placement in low-visibility locations and minimized disturbance to the study area.
Preliminary field application of the AMDI at the Moab site clarified the importance of scaling Vtot to the voltage difference during the course of each measurement. The differences between vmax and vmin for a single test varied up to 50% among infiltrometers and up to 20% for a single infiltrometer during the course of a day, despite nearly constant infiltrometer recharge volumes. Probable causes for the variability include battery strain, infiltrometer placement angle, large changes in ambient air temperature, and differences in cable resistance. Occasional failure of the automated instruments was caused by wetting the pressure transducers or forcing water into the port A tubing when filling the infiltrometers.
|
CONCLUSIONS
|
|---|
Automation of mini disk infiltrometers is a relatively simple and inexpensive technique and allows detailed in situ characterization of soil surface hydraulic properties in remote and difficult experimental settings. Measurement quality and number are enhanced through the use of a tripod support stand. Variability in transducer output voltage, vo, can arise from a number of factors, including battery drain and infiltrometer inclination, but can be accounted for with a simple calibration. We found the original instrument (AMDIv1) provided consistent K(h) results at moderate angles of inclination using our simple calibration approach. Use of the AMDIv2 on an incline would require additional effort to maintain a constant tension in the supply reservoir and operation at a consistent axial orientation to the slope. The AMDIv1 also returned more consistent values of K(h) than the AMDIv2, perhaps due to the smaller volume of water required to reach steady state by the AMDIv1 in the tested soils. Greater accuracy in S and K(h) can be achieved by taking sequential measurements on initially dry soil to account for the control exerted by the initial soil moisture content on these measurements.
This study did not address other factors that contribute to variability in infiltration, which include diurnal and seasonal temperature fluctuations, hydrophobicity, and secondary porosity. We encourage field researchers using this technique, however, to make complementary measurements of these properties to gain a comprehensive understanding of spatial and temporal variability in infiltration.
|
ACKNOWLEDGMENTS
|
|---|
This work was supported by the Utah Agricultural Experiment Station, Utah State University, Logan, UT. Approved as Journal Paper 7855. Additional support was provided by the Kansas State University Targeted Excellence Program. Decagon Devices kindly provided materials and advice for instrument fabrication.
|
NOTES
|
|---|
Currently at: Dep. of Plant and Animal Sciences, Brigham Young Univ., Provo, UT 84602
Currently at: Dep. of Civil Engineering, Kansas State Univ., Manhattan, KS 66506
All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Permission for printing and for reprinting the material contained herein has been obtained by the publisher.
Received for publication January 5, 2007.
|
REFERENCES
|
|---|
- Ankeny, M.D., T.C. Kaspar, and R. Horton. 1988. Design for an automated tension infiltrometer. Soil Sci. Soc. Am. J. 52:893–896.[Abstract/Free Full Text]
- Casey, F.X.M., and N.E. Derby. 2002. Improved design for an automated tension infiltrometer. Soil Sci. Soc. Am. J. 66:64–67.[Abstract/Free Full Text]
- Lebron, I., M.D. Madsen, D.G. Chandler, D.A. Robinson, O. Wendroth, and J. Belnap. 2007. Ecohydrological controls on soil moisture and hydraulic conductivity within pinyon–juniper woodland. Water Resour. Res. (in press).
- Perroux, K.M., and I. White. 1988. Designs for disc permeameters. Soil Sci. Soc. Am . J. 52:1205–1215.[Abstract/Free Full Text]
- Reynolds, W.D., and D.E. Elrick. 1991. Determination of hydraulic conductivity using a tension infiltrometer. Soil Sci. Soc. Am. J. 55:633–639.[Web of Science]
- Vandervaere, J.-P., M. Vauclin, and D.E. Elrick. 2000a. Transient flow from tension infiltrometers: I. The two-parameter equation. Soil Sci. Soc. Am. J. 64:1263–1272.
- Vandervaere, J.-P., M. Vauclin, and D.E. Elrick. 2000b. Transient flow from tension infiltrometers: II. Four methods to determine sorptivity and conductivity. Soil Sci. Soc. Am. J. 64:1272–1284.
- Warrick, A.W. 1992. Models for disk infiltrometers. Water Resour. Res. 28:1319–1327.[CrossRef]
- Zhang, R. 1997. Determination of soil sorptivity and hydraulic conductivity from the disk infiltrometer. Soil Sci. Soc. Am. J. 61:1024–1030.[Web of Science]
TOP
ABSTRACT
INTRODUCTION
