Gravity and wick pan samplers

Field methods and materials

Results

Conclusions

References

 

Protection of groundwater quality requires accurate knowledge of chemical transport through the vadoze zone, either through macropores, e.g., cracks and root channels (preferential flow), and through the bulk of soil (matrix flow). To obtain accurate measurements of ground water quality, the contribution of both preferential flow and matrix flow must be determined using a sampling method. There are five sampling methods available to collect water samples from the soil above the ground water table (vadoze zone) or from the shallow groundwater, as summarized in Table 1. In general, methods for continuous sampling of water and solutes involve the acquisition of water drained either by the force of gravity (e.g., gravity pan samplers, agricultural tile lines, and shallow wells), or by applying a “capillary” suction (e.g., porous cup samplers and wick pan samplers). This section focuses on the effectiveness of wick and gravity pan samplers in assessing solute transport in the vadoze zone. Concentrations measured with the pan samplers were compared with those obtained from the porous cup samplers.

Table 1. Sampling methods and their features (Boll et. al., 1992)

Gravity and Wick Pan samplers

Gravity pan samplers are designed to collect water from saturated regions in the vadoze zone. However, in many cases, such regions are not present naturally, and the sampler induces a pressure increase in soil above the pan. This artifact often leads to bypassing of the pan samplers.

A modification to the gravity pan was the alundum tension plate sampler, in which samples were drawn from unsaturated soil by applying a suction across alundum filter discs. Pan samplers with fiberglass wicks are based on the same principle. The wicks are self-priming and act as a hanging water column, providing a suction in the soil above the pan. Many of wick pan samplers used consist of a 30 by 30 cm pan with one large wick in the center. The drawback to this design is that chemicals entering the sampler at the sides need to travel a considerable distance to the center, while solutes near the middle flow without delay, giving rise to a large instrument dispersion coefficient. This design was improved by Boll in 1991. He decreased the pan-induced dispersion by dividing the 30 by 30 cm plate into a 5 by 5 grid, and sampling each 6 by 6 cm cell separately with a fiberglass wick.

Back to top

Field methods and materials

Gravity and wick pan samplers were tested in conjunction with porous cup samplers on Cornell University’s Thompson Vegetable Crops Farm near Freeville, NY (the “Freeville Site”) and in the Cornell University Apple Orchard in Ithaca, NY (the “Orchard Site”). Gravity and wick pan samplers were installed 1 m apart, 0.6 m below the soil surface in 0.9-m long, horizontal tunnels excavated in the side of a trench. The pan consisted of a 5 by 5 grid of individually sampled compartments, and was pressed against the native soil above using screw-jack supports. The difference between the wick and gravity pan samplers was the way effluent was collected. In the 320 by 320 mm gravity pan, the 25-mm high sampling compartments were filled with pea gravel so that water dripped into the bottles from saturated soil. The undisturbed soil above each cell of the wick pan was sampled under continuous suction applied by a 0.4-m long, 9.5-mm diameter fiberglass wick. The upper end of the wick was spread over an acrylic plate, and each plate was seated on a compression spring (65 mm tall, 24 mm in diameter), assuring good contact with the soil. Each wick was encased in 13 mm (in diameter) tygon tubing and was suspended well above a collection bottle to prevent upward movement in the wick. A schematic drawing of one cell in the wick pan sampler is given in Figure 1.

Figure 1. Schematic diagram of one 6 by 6 cell in the wick pan sampler by Boll et. al. (1991)

Porous cup samplers consisted of a ceramic cup having an outside diameter of 48 mm in the Freeville Site and 24 mm in the Orchard Site, and an overall length of 53 mm. The cup was cemented to a PVC pipe plugged with a two-hole stopper. Plastic tubing was put through the stopper holes; a short length to apply the vacuum (around 0.2 bars daily), and a longer section reaching into the cup for sample retrieval. Each sampler was installed through vertical augered holes backfilled with a slurry of the original soil mixed with a small amount of bentonite. All holes were sealed with a 250-mm layer of bentonite to prevent leakage from the surface. All porous cup samplers were installed at a depth of 0.6 m.

Freeville Site
The soil was a well-drained Genesee silt loam characterized by a root zone of 60 cm of dark brown sandy to loamy soil containing worm and root holes, overlaying a dark grayish-brown silt loam to very fine sandy loam from 60 to 200 cm, and a substratum of layers of gravel and sand at a depth of > 200 cm. Two gravity and two wick pan samplers, and four porous cup samplers were installed according to above descriptions. The grass-covered plot (6 by 6 m) received 2 cm of irrigation twice a day. After a steady-state outflow pattern had been reached, a 0.1 M bromide solution was applied for one day in 4 cm of rain, followed by 22 days of irrigation.

Orchard Site
The soil is a Rhinebeck clay loam in which structural cracks are a distinct feature. One gravity and one wick pan sampler, and 24 porous cup samplers were installed in each of two 2 by 6 m plots. Each plot was subjected to a different management practice to suppress weeds under the apple trees: a mowed sod and a moss cover resistant to roundup (glyphosate) (sprayed yearly). Both practices had been in place for five years prior to the experiment. In the plot with mowed sod cover, many very fine roots were evident in the upper 30 cm, with no cracks larger than 1 mm observed to this depth. Below 30 cm, the original soil structure was maintained and water could flow through structural cracks between hexagonally shaped peds 20 to 30 cm in diameter. Large surface-connected cracks (sometimes as large as 1 to 2 cm wide) were visible in the moss-covered plot. These cracks remained open to a depth of at least one meter. Water was applied to both plots daily at a rate of 1 cm/hr for three to four hours. The duration of the irrigation period was three weeks on the mowed grass plots and slightly less than two weeks on the moss-covered plots. A 0.1 M bromide solution was applied with the first irrigation.

Back to top

Results

Freeville site
Figure 2 shows breakthrough curves of bromide for each pan sampler. Each concentration value was obtained from the total mass and volume collected in all 225 cells. The concentration peak for both sampling methods occurred 9 or 10 days after application, with a peak concentration of approximate 650 mg/l for both wick samplers, and 813 and 280 mg/l for the gravity samplers. Although initially, a few cells in wick sampler 1 and gravity sampler 2 collected high bromide concentrations as a consequence of preferential flow, experimental data were fitted closely using the standard convective-dispersive (Gaussian) curve. This indicates that matrix flow was the predominant component for transport of the bulk of the bromide in the upper 60 cm of the soil.

Figure 2. Spatially averaged breakthrough curves for two wick and two gravity pand samplers at the Freeville Site. The missing line segments indicate that no water was collected.

Wick pan samplers collected nearly 100% of the applied water and, on average, 63% of the total bromide, whereas the gravity pan samplers intercepted, on average, 28% of the applied water and only 7% of the total bromide (Table 2). Water and solutes clearly bypassed the gravity pan samplers and, to a much lesser degree (if at all), the wick pan samplers. Evidence of bypass flow also is derived from the distribution of the cumulative loss of bromide in the pan for each cell during the experimental period. Figure 3. shows the total percentage of bromide intercepted by each cell in one gravity and one wick pan samplers. As observed, the highest percentage of bromide was collected in the center of the gravity pan sampler, whereas in the wick pan sampler, a more random distribution was apparent. Water and bromide bypassed the outer edges of the gravity pan sampler due to the hydraulic potential gradient between saturated soil above the sampler and unsaturated soil in its surroundings. The soil above the wick pan sampler was unsaturated and the potential gradient did not manifest itself under the imposed flux, resulting in a more random pattern of high and low fluxes. Accordingly, the distribution of fluxes in this soil was much better represented in the wick pan sampler than in the gravity sampler. Under very low conditions (not tested here), water may also bypass the wick sampler.

Table 2. Average water and total bromide collected by wick and gravity pan samplers installed at 60 cm depth at Freeville site.

Figure 3. Spatial distribution of mass of bromide intercepted by each cell during the experimental period as a percentage of the amount applied for one gravity pan sampler (a) and wick pan sampler (b) at the Freeville site.

As shown in Figure 4, the four porous cup samplers at the 60 cm depth differed not only from the pan samplers, but also from each other. Porous cup 1 had a peak concentration of 968 mg/l which occurred on day 15. Porous cups 2 and 3 had high peak concentrations of 1792 and 1304 mg/l, respectively, which occurred on day 4, three days after application. Porous cup 4 did not collect except on the first day. Only three of the 100 pan sampling cells provided concentration data similar to the patterns of porous cups 2 and 3, strongly suggesting that the porous cup sampling method had distorted the flow pattern. This was confirmed at the end of the experiment when blue dye was applied shortly before the samplers were removed. Blue dye had moved along thin sampling tubes (despite careful installation), directly to the porous cup samplers at the 60 cm depth.

Figure 4. Breakthrough for porous cup samplers at the 60 cm depth at the Freeville site

Orchard Site
Breakthrough curves for the weighted average of all cells show that, in both plots, the highest concentrations arrived with the wetting front, due to preferential flow, while the matrix flow contribution was insignificant (Figure 5). Concentrations in the moss-covered plot almost reached the pulse concentration and less than half that in the mowed sod plot. So, in this clay loam soil, the gravity and wick pan samplers gave almost identical results because most of the water flowed through cracks, and sideways matrix flow movement was minimal due to the dense and almost impermeable soil matrix.

Figure 5. Spatially averaged breakthrough curves for wick and gravity pan samplers for the mowed-grass and moss-covered plots at the Orchard Site.

In both plots, the pan samplers collected approximately all of the applied bromide, but not always all of the applied water (Table 3). Bromide was collected during the first five days of the experiment, when most cracks were still largely open at the surface. Subsequently, overland flow increased as some of the surface cracks closed, resulting in reduced infiltration. Furthermore, the water collection pattern was a function of the micro topography and contributing area of the cracks. The latter is illustrated in Figure 6, which again shows the cumulative percentage of bromide collected in each cell in the gravity and wick pan samplers during the experimental period. Cells with great losses of bromide were observed adjacent to cells with very little loss of bromide. Obviously, the large amounts of bromide were carried preferentially through the cracks and neighboring cells were not affected as lateral movement in the region above the samplers was insignificant due to the extreme low permeability of the soil matrix. It was coincidental that the highest amounts of bromide were collected in the center of the gravity pan sampler. Under the experimental conditions tested, the data from the gravity and wick pan samplers appear to have accurately captured the characteristics of the soil and the difference between the treatments.

Table 3. Total water and bromide collected by wick and gravity pan samplers installed at 60 cm depth at Orchard site

Figure 6. Spatial distribution of mass of bromide intercepted by each cell during the experimental period as a percentage of the amount applied for each cell of the gravity pan sampler (a) and wick pan sampler (b) for the moss-covered plot at the Orchard Site.

The average concentration of the 24 porous cups is depicted in Figures 7a and 7b for the mowed-sod plot and the moss-covered plot, respectively. The most striking feature is that the early bromide breakthrough, as seen in the pan samplers, was not observed with the porous cups. Especially in the moss-covered plots, all of the applied solutes bypassed the porous cup samplers. Bypass of the cups in the mowed-sod also was evident, however, less pronounced.

Figure 7. Comparisons of wick pan samplers with the porous cup samplers for the mowed-grass (a) and moss-covered (b) plots at the Orchard Site.

Back to top

Conclusions

Two case studies showed the relative importance of matrix and preferential flow inferred from the appearance of breakthrough curves obtained with wick and gravity pan samplers. The effectiveness of the wick and gravity pan samplers depended on the type of soil. Wick pan samplers represented flow through the vadoze zone much more accurately than gravity samplers when water and solutes were transported by matrix flow in the loamy soil. The bulk of transport in the clay loam soil followed macropore flow paths, in which case both wick and gravity pan samplers performed equally well. Porous cup samplers indicated a predomincance of matrix flow. In the clay loam soil, the porous cup samplers failed to register flow through existing cracks, which was unmistakably observed with the pan samplers. Hence, under conditions similar to those presented here, use of porous cups in the upper part of the vadose zone is not recommended.

Back to top

References:

Steenhuis, T.S., J. Boll, E. Jolles, and J. Selker. 1995. Field Evaluation of Wick and Gravity Pan Samplers. Chapter 34, In: Handbook of Vadose Zone Characterization & Monitoring, L. Everett, S. Cullen, and L. Wilson, Eds. Lewis Publishers, Ann Arbor, MI. pp. 629-638.

Boll, J., T.S. Steenhuis, and J.S. Selker. 1992. Fiberglass Wicks for Sampling Water and Solutes in the Vadose Zone. Soil Sci. Soc.Am. J. 56(3):701-707.

Boll, J., J.S. Selker, B.M. Nijssen, T.S. Steenhuis, J. Ban Winkle, and E. Jolles. 1991. Water Quality Sampling Under Preferential Flow Conditions. p. 290-298. In R.G. Allen, et al. (ed.) Lysimeters for evapotranspiration and environmental measurement. Proc. ASCE Int. Symp. Lysimetry, Honolulu, HI. 23-25 July 1991. ASCE, New York.

Back to top