LAGOON SEALING AND FILTER CAKES J. S. Tyner, W. C. Wright, J. Lee ABSTRACT. A recent study presented a two-layer (liner and seal) unsaturated model to predict the flux from lagoons. The model predicted that flux is not related to liner thickness, is only weakly related to the saturated hydraulic conductivity of the liner, and is primarily controlled by the hydraulic conductivity of the seal. In this study, we evaluated those predictions by monitoring the flux of dairy waste through eight intact soil monoliths (silt loam) with large macropores. After applying a 2.3 m column of dairy waste (2.3% total solids) to the top of the monoliths, a seal developed within 2 d, and the leakage continued to decrease for the remaining 83 d of the test. The average leakage rate after 85 d was 0.70 mm d−1, and an average of 218 mm of waste infiltrated into the monoliths. Small holes drilled into the sides of the monoliths, just below the soil surface, failed to fill with liquid, which demonstrated that the soil was unsaturated and therefore was not limiting the leakage rate. The amount of waste required to initially seal the monoliths did not correlate to the seal growth rate, which also implies that the soil contributed little to the seal growth rate or leakage rate after seal development. A plot of cumulative waste infiltration versus the square root of time showed a strong linear relationship (R2 = 0.996), which suggests that the phenomenon of dairy waste sealing a soil is analogous to filter cake growth. Long-term studies with other soil and waste types are needed to confirm the findings of this research. Keywords. Animal waste, Clay liner, Confined animal, Dairy, Lagoon, Leachate.

L

eakage from animal waste ponds and lagoons into the underlying soil and groundwater releases unwanted nutrients and pathogens into the environment (Parker et al., 1999). Most U.S. states regulate the new construction of animal waste storage ponds by requiring a liner with a saturated hydraulic conductivity (KL ), when tested with water, to be less than a prescribed level (KL < 10−6 cm s−1 for many states). The assumption underlying this regulation is that animal waste will rapidly seal a soil and will quickly reduce the flux of leachate by at least an order of magnitude. Appendix 10D of the Agricultural Waste Management Field Handbook (AWMH) provides an example in which a 2.74 m head of waste lies atop a 0.30 m compacted soil liner, which results in a hydraulic gradient of 10 (SCS, 1997). Flux through the liner is estimated using Darcy’s law, while including an order of magnitude reduction of KL from waste sealing, as follows:

Submitted for review in September 2005 as manuscript number SE 6058; approved for publication by the Structures & Environment Division of ASABE in January 2006. The authors are John S. Tyner, ASABE Member Engineer, Assistant Professor, Wesley C. Wright, ASABE Member Engineer, Research Associate, and Jaehoon Lee, Assistant Professor, Department of Biosystems Engineering and Soil Science, University of Tennessee, Knoxville, Tennessee. Corresponding author: John S. Tyner, 2506 E. J. Chapman Dr., Knoxville, TN 37996; phone: 865-974-7130; fax: 865974-4514; e-mail: [email protected].

q= =

KL ∆h sealing credit ∆L 10 − 6 cm s −1 (2.74 m + 0.30 m ) 10 0.30 m

= 10 − 6 cm s −1 = 0.86 mm d −1

(1)

where q is flux,
h is the head loss across the liner, and
L is the thickness of the liner. Tyner and Lee (2004) derived an alternative model to describe the flux of waste through a two-layer (seal and liner) system, which predicted that q is only weakly related to KL instead of the linear relationship predicted by equation 1. Their model also predicted that q is not related to L because the soil desaturates beneath the seal until a unit hydraulic gradient is present. A number of other studies have also concluded that the soil liner beneath a seal desaturates (Barrington et al., 1987; Barrington and Madramootoo, 1989; Culley and Phillips, 1982; Miller et al., 1985). A transient model describing seal formation and subsequent flux can be derived by applying Darcy’s law across the seal only: ∆h dl = Ks s (2) dt ∆Ls where l is the cumulative length of waste passing through the seal, t is elapsed time, Ks is the hydraulic conductivity of the seal,
hs is the difference in head across the seal, and
Ls is the length of the seal. If seal growth is proportional to l, such that
Ls =  l, where  is a constant related to the size, concentration, and packing of waste solids, then substitution of
Ls into equation 2 followed by integration yields:

Transactions of the ASABE Vol. 49(2): 527−531

E 2006 American Society of Agricultural and Biological Engineers ISSN 0001−2351

527

1

 2 K s ∆hs t  2 l =  υ  

(3)

when l = 0 at t = 0. Therefore, the cumulative amount of waste that passes through the seal should be proportional to the square root of time when the soil liner provides negligible resistance to flow. Resistance to flow by the soil prior to seal formation is neglected by this model. Forms of equations 2 and 3 are commonly used to describe filtration and the subsequent reduction of flux through a growing filter cake. The objectives of this research were to: (1) evaluate the ability of dairy waste to sufficiently seal intact silt loam monoliths containing large macropores, (2) evaluate the performance of the filter cake model (eqs. 2 and 3) to predict transient seal performance, and (3) evaluate the predictions of the steady-state Tyner and Lee model that soil liner thickness does not affect flux and that the hydraulic conductivity of the soil liner has minimal effect on flux following seal formation.

METHODS

Eight soil monoliths were collected from the Ames Plantation in Grand Junction, Tennessee, on 16 October 2004. A backhoe was used to dig two 1 m deep parallel trenches, leaving a 0.7 m wide standing wall of soil between the trenches. Each monolith was sculpted with hand tools to

a diameter of 0.28 m and height of 0.75 m (fig. 1). A rubberized sealant was sprayed on the circumference of the monoliths for stabilization and to restrict edge effects. Twelve-inch nominal schedule 40 PVC pipes (inner diameter = 0.30 m) were slid over the monoliths, and melted paraffin was poured into the gap between the monolith and PVC pipe. Thirty millimeters of coarse sand was placed into PVC end caps and cemented to the bottom of the pipes. The monoliths were then transported to the laboratory. Particle size analysis (ASTM Standards, 2005) was conducted to determine the soil texture. The results of the analysis were 3% sand, 78% silt, and 19% clay, making this soil a silt loam by the USDA classification scheme (Soil Survey Staff, 1975). To wet the monoliths, 5 mm of water with 0.001 M CaSO4 was gently applied with a watering can every day for eight weeks. The CaSO4 was added to the water to minimize clay dispersion (Klute and Dirksen, 1986). No ponding occurred during the application of the water. A 19 mm drainage hole cut in the base of each monolith allowed excess water to drain. Within the eight-week period, all eight monoliths had reached symmetry between the daily inflow and outflow of water. Monoliths were converted to falling head permeameters by attaching 1 in. nominal schedule 40 clear PVC pipe (inner diameter = 2.6 cm) to the top of the 0.30 m PVC pipe (fig. 2). The ratio of the cross-sectional area of the soil monolith to the falling head pipe was 114 to 1, which enabled precise measurement of the daily flux through the monoliths. Liquid dairy waste with a solids content of 2.3% was collected from

Figure 1. Undisturbed monoliths were hand sculpted, and the sides were coated with a rubberized sealant; 0.3 m PVC pipe was placed around the monoliths, and the gap between the monoliths and the PVC pipe was filled with melted paraffin.

528

TRANSACTIONS OF THE ASABE

a private dairy farm west of Knoxville, Tennessee. The permeameters were filled with 2.45 m of dairy waste and re− filled after 1 h, then every day for five weeks, and then every second or third day for the remaining seven weeks. Because the head decreased between fillings, the time-averaged height of waste above the monoliths was approximately 2.3 m. The head varied by approximately ±0.2 m between fillings.

26 mm

2.3 m

RESULTS

Animal waste

Seal

0.28 m Silt loam

0.75 m

Coarse sand

Figure 2. Schematic of monolith incorporated into a falling head permeameter.

During excavation, inspection of the soil monoliths revealed several 10 mm diameter macropores. We confirmed the macropores were continuous for a distance of at least 150 mm. The deepest macropore was located in the base of the trench, 1.0 m from the ground surface (fig. 3). A beetle was found residing in a macropore 0.7 m from the ground surface. Immediately after filling the columns with waste, the initial fluxes of monoliths 1 to 7 were on the order of 10−3 cm s−1. Flux through monolith 8 was extremely rapid, with dark liquid exiting the base of the column within 60 s of filling the 2.6 cm PVC pipe with waste. After twice refilling monolith 8, the rapid drainage of dark liquid ceased. Figure 4 is a graph of the flux through the monoliths versus time. The first vertical set of points represents the initial 1 h average from each monolith. The line indicates the average flux of all eight monoliths during each time interval. Fluxes decreased rapidly for 2 d and then decreased more slowly for the remaining 83 d. The final average flux was 0.70 mm d−1.

Figure 3. This large macropore (10 mm diameter) was located in the base of the trench, approximately 1.0 m from the ground surface.

Vol. 49(2): 527−531

529

100.0

250

y = 15.9x + 74.5 R2 = 0.996

10.0

l (mm)

q (mm/d)

Average

125

1.0

0 0

0.1 0

30

t (d)

60

90

5

t

1/2

10 1/2

(d)

Figure 4. Leakage rate of combined seal-soil system through time for eight monoliths. The line is the average value from the eight monoliths and is plotted from day 2 through day 85.

Figure 6. Average cumulative leakage (l) versus the square root of elapsed time. The points represent the average values from all eight monoliths, and the linear trend line has an R2 of 0.996.

Because the ratio of fluxes between our tests and those calculated in equation 1 is less than the similar ratio of waste heights, i.e.:

the slope of the trend lines from the eight monoliths (R2 = 0.015), indicating that the amount of waste necessary for initial seal establishment was unrelated to the seal growth rate. Clearly some of the assumptions for equation 3 were not maintained, including: initially the soil provided all the resistance to flow,
hs varied slightly between fillings, potential biological sealing and seal degradation is not included, compaction of the seal may have taken place, and some of the solids in the waste may have lodged in pore spaces within the upper soil layers instead of atop the soil. Nonetheless, the data from all eight monoliths fit the filter cake model (eqs. 2 and 3) well for the 85 days that were measured. At 85 d, while the permeameters were still full with waste, two 10 mm diameter × 100 mm long holes were drilled horizontally into the side of each monolith approximately 50 mm beneath the top of the soil surface. No liquid entered any of the holes, which confirms a negative pore pressure, and therefore unsaturated conditions directly beneath the seals. This leads to the conclusion that the seal and the upper 50 mm of soil consumed all of the pressure head from the column of waste, and the continued movement of the waste through the soil was due to gravity only.

 0.70 mm d −1   2.3 m   < = 81 % = 84%   0.86 mm d −1   2.75 m   

(4)

we conclude that the monoliths were sufficiently sealed. Figure 5 presents the flux versus the cumulative infiltration into each monolith. After two days (the third point of each data series), the slopes abruptly decreased. We believe the rapid change of slope at approximately 2 d indicates the formation of a continuous seal across the tops of the soil monoliths. Monolith 8 required much more waste (0.32 m) to initially seal. Although the amount of waste required to initially seal the monoliths varied greatly, all the monoliths showed a similar rate of flux decrease following seal formation, as noted by the similar slopes after 2 d. Figure 6 is a graph of the cumulative length of waste that flowed into the monoliths (average of all eight monoliths) versus the square root of time. The plot fits the prediction by equation 3 that l ∝ t with a coefficient of determination (R2) of 0.996. Linear trend lines were also fit to the data from each individual monolith (not shown), and the minimum R2 was 0.986. No correlation existed between the intercept and 100.0

q (mm/d)

10.0

1

2

3

4

5

6

7

8

1.0

0.1 0

150

300

450

Cumulative leakage (mm)

Figure 5. Flux versus cumulative leakage from each of the eight monoliths. Although varying amounts of waste were required to initially seal the monoliths, the seal growth rate was similar for each monolith, as demonstrated by the similar slopes.

530

DISCUSSION

Based on the results from this study, it is possible that the process of sealing soil with animal waste is analogous to the construction of a filter cake and is therefore functionally described by equation 3. If this is indeed the case, then three interesting hypotheses can be drawn following the formation of a seal that dissipates the vast majority of the head: S Flux is not dependent on KL . Since seal formation is linked to the waste being filtered by the soil surface and the seal is consuming all the pressure head, attempts to minimize flux by selecting a soil liner with a small KL should be ineffective following seal formation. At later times, the flux from different soil textures should coalesce to a value determined by the properties of the waste. S Flux is not dependent on L. Since the seal consumes all of the pressure head from the column of waste and the continued movement of the waste through the soil is due to gravity only, a unit hydraulic gradient forms within the soil (dh/dL = 1); therefore, q = KL (), where

TRANSACTIONS OF THE ASABE

KL () is the unsaturated hydraulic conductivity at volumetric water content . S A deep pond will produce less leachate than a shallow pond with equivalent volume. Since l ∝ hs and the volumetric flow rate from a waste pond is proportional to its area, a deep pond with small area should produce less leachate than an equivalent volume pond with shallow depth and large area. Initially, the volumetric flow rates of the two cases should be identical, but because the flux through the deep pond would initially be higher, its seal would grow more rapidly, and its volumetric flow rate would also decrease more rapidly. In apparent contrast to the findings of this study, Huffman (2004) found that four-fifths of 34 swine waste lagoons in North Carolina showed elevated downgradient mineral-N concentrations. Yet, Glanville et al. (2001) found that less than 4% of 28 manure storage structures and lagoons in Iowa significantly (p < 0.05) exceeded the state’s flux requirement of 1.6 mm d−1 at a waste depth of 1.8 m. A possible cause of the apparently contrasting findings of these studies can be reconciled by acknowledging that a 0.2 ha (0.5 acre) lagoon with a lawful flux of 0.86 mm d−1 (1/32 in. d−1) releases 1.74 m3 d−1 (460 gal d−1) of leachate to the subsurface.

CONCLUSIONS

Eight large silt loam monoliths with macropores were sufficiently sealed by a 2.45 m column of dairy waste within 85 days. After 85 days, the average leakage rate was 0.70 mm d−1, and an average of 218 mm of waste had infiltrated into the monoliths. The soil just beneath the surface was unsaturated, which demonstrates that all the pressure head was consumed within the seal and the remaining depth of soil was not influencing the leakage rate. No correlation existed between the amount of waste required to initially seal the columns and the rate of additional sealing. All eight monoliths demonstrated a linear relationship (minimum R2 of 0.986) between the cumulative amount of infiltrated waste and the square root of time, which suggests that the pheno−

Vol. 49(2): 527−531

menon of dairy waste seal formation may be analogous to filter cake growth. Long-term studies with other soil and waste types are needed to confirm the findings of this research.

REFERENCES

ASTM Standards. 2005. D422: Standard test method for particle size analysis of soils. Sec. 4: Vol. 4.08, Soil and Rock. West Conshohocken, Pa.: American Society for Testing and Materials. Barrington, S. F., and C. A. Madramootoo. 1989. Investigating seal formation from manure infiltration into soils. Trans. ASAE 32(3): 851-856. Barrington, S. F., P. J. Jutras, and R. S. Broughton. 1987. The sealing of soils by manure: I. Preliminary investigations. Canadian Agric. Eng. 29(2): 99-103. Culley, J. L. B., and P. A. Phillips. 1982. Sealing of soils by liquid cattle manure. Canadian Agric. Eng. 24(2): 87-89. Glanville, T. D., J. L. Baker, S. W. Melvin, and M. M. Agua. 2001. Measurement of leakage from earthen manure structures in Iowa. Trans. ASAE 44(6): 1609-1616. Huffman, R. L. 2004. Seepage evaluation of older swine lagoons in North Carolina. Trans. ASAE 47(5): 1507-1512. Klute, A., and C. Dirksen. 1986. Hydraulic conductivity and diffusivity: Laboratory methods. In Methods of Soil Analysis, No. 9, part 1. 2nd ed. Madison, Wisc.: Soil Science Society of America. Miller, M. H., J. B. Robinson, and R. W. Gillham. 1985. Self-sealing of earthen liquid manure storage ponds: 1. A case study. J. Environ. Quality 14(4): 533-538. Parker, D. B., D. D. Schulte, and D. E. Eisenhauer. 1999. Seepage from earthen animal waste ponds and lagoons: An overview of research results and state regulations. Trans. ASAE 42(2): 485-493. SCS. 1997. National Engineering Handbook: Part 651. Agricultural Waste Management Field Handbook: 651.1080. Appendix 10D: Geotechnical, Design, and Construction Guidelines. Washington, D.C.: USDA Soil Conservation Service. Soil Survey Staff. 1975. Soil taxonomy: A basic system of soil classification for making and interpreting soil surveys. Agric. Handbook 436. Washington, D.C.: USDA Soil Conservation Service. Tyner, J. S., and J. Lee. 2004. Influence of seal and liner hydraulic properties on the seepage rate from animal waste holding ponds and lagoons. Trans. ASAE 47(5): 1739-1745.

531

532

TRANSACTIONS OF THE ASABE