Predicting saturated hydraulic conductivity from percolation test results in layered silt loam soils.
The disposal of effluent from conventional septic tanks, wastewater treatment systems, and sewage treatment works has potential to cause serious problems in our environment and particularly in the soil ecosystem (Patterson, 1996).
Rural and suburban areas frequently do not have a public sewer system, therefore, the residents of these areas often depend on septic tank systems to dispose of their sewage. The size of these onsite waste disposal systems is usually determined by one or more percolation tests made on the proposed site (Schmidt et al., 1980).
A number of alternative techniques for making a percolation test have been proposed (Mulqueen & Rodgers, 2001; Rodgers & Mulqueen, 2004; Schmidt et al., 1980). The simplest and most commonly used test is performed by digging a borehole in the soil, filling it with water to presoak the soil, and then determining the rate of fall of the water surface in the hole (Fritton, Long, Aron, & Petersen, 1983; Schmidt et al., 1980). Numerous variations of this simple percolation test are available. Standard percolation test procedures have been published by the U.S. Environmental Protection Agency (Schmidt et al., 1980). These procedures are used by individual states throughout the U.S. with some modifications to yield their own percolation test procedure (Schmidt et al., 1980).
Since the percolation test is the tool most widely used to measure water flow capabilities of a field soil, several attempts have been made to find empirical relationships between the saturated hydraulic conductivity (Ks) and the percolation time (PT) with a hope of replacing the percolation test with a hydraulic conductivity test (Fritton, Ratvasky, & Petersen, 1986; Mulqueen & Rodgers, 2001; Winneberger, 1974).
Most of these studies have dealt with homogeneous soils. In reality, however, field soils are inherently heterogeneous and usually consist of distinct layers or horizons of relatively homogeneous materials. Therefore, the objectives of the study described here were to first, develop an empirical relationship between the saturated hydraulic conductivity (Ks) of layered silt loam soils and their percolation times (PT) in order to understand the influence of individual layers; and second, to statistically compare this relationship with the equations developed by Winneberger (1974) and Fritton and co-authors (1986).
Materials and Methods
Soils and Sites Description
Field research was conducted on three soils in Centre County, Pennsylvania. The soils were classified as a fine, mixed, mesic, Typic Hapludalf (Hagerstown silt loam); a fine-loamy, mixed, mesic, Typic Fragiudult (Monongahela silt loam); and a fine-silty, mixed, mesic, Dystric Fluventic Eutrochrept (Nolin silt loam). All of these sites were located in the Ridge and Valley Physiographic Province of Pennsylvania (Braker, 1981). These soils were layered in different ways as follows: (Hagerstown) a silt loam Ap horizon underlain by a high clay content B horizon in the top meter, (Monongahela) a silt loam Ap horizon underlain by a fragipan in the top meter, and (Nolin) a silt loam alluvial soil underlain by a gravel layer in the top meter (Braker, 1981). The Hagerstown silt loam site was located in Ferguson Township on the Pennsylvania State University Agronomy Research Farm at Rock Springs, 30 m E-NE of the Agronomy field laboratory building (40[degrees]42'55" N, 77[degrees]56'15" W). It is about 16 km west of downtown State College, Pennsylvania. The soil is well drained and developed in limestone residuum parent material. The site was in a sod area. The Nolin silt loam site was located 50 m west of the sheep barn and 100 m south of the wooden pedestrian bridge over Spring Creek on Puddintown Road west of Houserville in College Township (40[degrees]50'01" N, 77[degrees]49'40" W). The soil was formed in recent alluvium from limestone and exists on flood plains. The soil at this site was deep and well drained. The site was in a pasture. The Monongahela silt loam site was located 100 m north of the railroad tracks and 150 m east of the Historical Museum at Curtin Village in Boggs Township (40[degrees]53'23" N, 77[degrees]47'40" W). The soil was developed in old alluvium (terrace position) from shale limestone. The soil was deep, moderately well drained, and had a fragipan in the B horizon. The three sites were nearly level to gently sloping.
Field Technique and Site Preparation
Six holes were spaced evenly in two parallel rows of three holes each, approximately 50 m apart. The sampling distance was determined using geostatsitical methods that indicated that samples for the measurements of Ks had to be at least 40 m apart to be independent from each other (Jabro, Fritton, & Dobos, 1991). Field-saturated hydraulic conductivity was measured at three different depths for each soil using a constant head well permeameter (Guelph permeameter [GP]) method (Elrick & Reynolds, 1986; Reynolds, Elrick, & Clothier, 1985).
For each sampling location a 40-mm diameter cylindrical hole was dug using a standard 38-mm diameter screw-type auger to the middle of the top horizon. A rigorous wire brush was used to prepare a clean borehole and to minimize wall smearing, which can cause erroneous unrepresentative [K.sub.s] values (Reynolds, 1993). Two sets of steady flow rate measurements were made at constant pressure heads of 10 cm and 20 cm water for each hole (Elrick, Reynolds, & Tan, 1989; Salverda & Dane, 1993). Steady-state flow rates were assumed when the last three consecutive readings were approximately the same ([+ or -] 5%). Further details about the Guelph permeameter (GP) technique are given in Reynolds (1993).
Soil saturated hydraulic conductivity, [K.sub.s] (L [T.sup.-1]) using a steady-state flow rate of water from a cylindrical borehole augured to a given depth below the soil surface, was calculated using Richards' analysis (Reynolds et al., 1985):
[K.sub.s] = CQ/(2[pi][H.sup.2] + C[pi][r.sup.2] + [2[pi]H/[alpha]]) (1)
where C is a dimensionless shape factor that depends primarily on the H/r ratio and soil texture/structure properties and is a function of both H and r, Q is the steady-state flow rate out of the borehole ([L.sup.3] [T.sup.-1]), H is the steady depth of water in the hole (L), r is the radius of the hole (L), and [alpha] is a soil texture/structure parameter ([L.sup.-1]) (Elrick et al., 1989). The C parameter for each hole of H/a was obtained from those numerically derived by Erick and Reynolds (1986). Further details regarding [alpha] parameter and its estimates for different types of soils are given in Reynolds and co-authors (1985) and Reynolds (1993).
At the end of this measurement, the hole was allowed to drain for three to four days. The hole was then deepened to the middle of the second layer and [K.sub.s] was determined again. After completion of the second [K.sub.s] measurement, the hole was allowed to drain for another three to four days. Then the hole was enlarged using a hand post-hole auger to approximately 0.24 m in diameter and deepened to approximately 0.52 m. After preparing and presoaking the hole, the percolation test was conducted using the standard procedure described by the Pennsylvania Department of Environmental Protection (Fritton et al., 1983). The initial water head used in this procedure was 0.16 m. The head was allowed to drop for half an hour, at which time a reading was taken and the head reestablished at 0.16 m (Jabro & Fritton, 1990).
Once the percolation test was conducted, the hole was allowed to drain for another six to seven days. Then, the hole was deepened to the middle of the third layer using a standard 38-mm diameter screw auger to measure the [K.sub.s] using a constant head well permeameter.
Saturated hydraulic conductivity was calculated using Equation 1 (Reynolds & Elrick, 2002), except for two points at the third depth of the Nolin site, where Laplace's solution of Equation 2 was used because the measurements were taken at this depth in very gravelly soil.
[K.sub.s] = CQ/[2[pi][H.sup.2] + C[pi][r.sup.2]] (2)
Soil samples were collected when holes were being prepared at each site at each depth to measure initial soil water content using the gravimetric method.
Results and Discussion
Measurements of [K.sub.s] at three depths for Hagerstown silt loam, Monongahela silt loam, and Nolin silt loam soils and their descriptive statistics are listed in Table 1.
The corresponding measurements of PT and their descriptive statistics for six test holes at each of the above soils are presented in Table 2. Considerable variations were found in the water flow rates for both [K.sub.s] data and percolation tests results within and between soil sites. These variations were attributed to soil variability (Barbarick, Warrick, & Post, 1976; Derr, Matelski, & Petersen, 1969).
Simple linear regression analyses were performed on the log-transformed data of both Ks and PT. The regression analyses performed between the [K.sub.s] values of each individual layer in all three sites and the corresponding percolation times yielded the following three linear equations for the upper, middle, and lower layers, respectively.
Log([K.sub.s]) = -0.436 - 0.664Log(PT) (3)
Log([K.sub.s]) = 0.369 - 2.472Log(PT) (4)
Log([K.sub.s]) = 1.020 - 3.488Log(PT) (5)
where Log is a base 10 logarithm, [K.sub.s] is soil saturated hydraulic conductivity (cm/hr), and PT is the percolation time (min/cm).
Linear regression lines for the Equations 3, 4, and 5 are shown in Figures 1, 2, and 3, respectively. The coefficients of determination, [R.sup.2], for the three equations were 0.12, 0.45, and 0.58, respectively. The R2 of the three linear regression equations were relatively low, indicating a large variability within and between the measurements of both [K.sub.s] and PT (Tables 1 and 2).
[FIGURE 1 OMITTED]
Poor and nonsignificant linear correlation was found between the logarithms of [K.sub.s] and PT for the upper layer at depth between 28-33 cm, indicating that the upper layer or horizon in the soil profile had no effect on the water flow rate results from the percolation test hole. In contrast, a highly significant negative correlation (p = .002) was found between the measurements of [K.sub.s] and PT of the middle layer at depth of 50-60 cm (Figure 2). This indicated that the middle layer had a direct influence on the flow rate results from the percolation test hole. The lower layer of the soil profile at depth between 73-97 cm also showed a highly significant correlation (p = .0003) between the flow rates of the GP values (Ks) and percolation tests, indicating that the lower layer influenced percolation test results (Figure 3). The slope of Equation 3 was significantly different from the slopes of both Equations 4 and 5 at the .01 level of significance using general linear test (Neter, Wasserman, & Kutner, 1985). A significant difference existed, however, between the slopes of Equation 4 and Equation 5.
[FIGURE 2 OMITTED]
[FIGURE 3 OMITTED]
Significant differences were also found between Equation 5 and similar equations from other studies (Equations 6 and 7 from Winneberger 1974 and Fritton et al., 1986, respectively) at .01 level of significance. In Winneberger's study, the [K.sub.s] was measured using the core technique, which measures the vertical flow only. A linear regression performed upon the [K.sub.s] and PT results yielded the following equation:
Log([K.sub.s]) = 0.68 - 1.21Log(PT) (6)
Fritton and co-authors (1986) used the shallow well pump-in method to measure soil [K.sub.s]. This method required a greater depth and amount of water in the hole than the percolation test. Due to a difference in water levels, more than one soil horizon was in contact with the water in the test hole. The following linear regression equation resulted from their pairs of data:
Log([K.sub.s]) = 0.093 - 0.309Log(PT) (7)
In contrast, Equation 5 was developed using a constant head well permeameter (GP) method, which measures the soil [K.sub.s] in each horizon. In addition, the depth of water in the GP test is approximately the same depth used in the percolation test. Therefore, Equation 5 can be considered an appropriate empirical relation between soil [K.sub.s] and PT for soils similar to those used in this study. It can be concluded that the differences between Equation 5 and other studies were due to procedural variability used to conduct the test (Barbarick et al., 1978; Bouma, 1971; Rodgers & Mulqueen, 2004). The empirical Equation 5 derived for my study is adequate to predict [K.sub.s] from PT in layered or heterogeneous soils. The equation may be used in evaluating the suitability of soils for onsite wastewater disposal systems, in septic tank system design, and in the development of guidelines for these systems.
Fritton and co-authors (1986) developed a set of regression lines for different sets of data with a slope of -1.0 and an intercept as a function of both ??parameter and the radius of the source of water. The ??parameter is soil texture/structure properties, which represents the effect of soil capillarity under steady flow conditions. Fritton's equation is a valid one for different soils to predict [K.sub.s] from PT results when the parameter can be estimated from the GP results. Thus, Fritton's equation has applications across diverse soils but only for homogeneous soil conditions. Further investigations are required to test Equation 5 with other specific soils and site conditions.
The results of this study showed how soil layers (stratification) significantly affect water flow from a percolation test hole compared with those in uniform and homogeneous soils. Results also demonstrated the influence of differing soil hydraulic conductivities on the water flow distribution from a percolation test hole.
Three linear regression equations for the upper, middle, and lower layers were developed between the [K.sub.s] values of each individual layer in all three sites and their corresponding PT. A highly significant negative correlation, r = -0.672 (p = .002) was found between the measurements of [K.sub.s] and PT of the middle layer at depth of 50-60 cm. This indicated that the middle layer had a direct influence on the flow rate results from the percolation test hole. The lower layer of the soil profile at depth between 73-97 cm also showed a highly significant correlation, r = -0.756 (p = .0003) between the flow rates ([K.sub.s]) of the GP values and percolation tests, indicating that the lower layer greatly influenced percolation test results. Significant differences were also found between Equation 5 and similar equations from other studies (Equations 6 and 7 from Winneberger 1974 and Fritton et al., 1986, respectively) at .01 probability level. The empirical equation derived for this study is adequate to predict [K.sub.s] from PT in layered silt loam soils and may be used in evaluating the suitability of soils for onsite wastewater disposal systems as well as in designing the size of these systems.
Disclaimer: Mention of trade names, proprietary products, or specific equipment is intended for reader information only and does not constitute a guarantee or warranty by the USDAARS; nor does it imply approval of the product named to the exclusion of other products.
Barbarick, K.A., Warrick, A.W., & Post, D.R. (1976). Percolation test for septic tank suitability in southern Arizona soils. Journal of Soil Water Conservation, 31(3), 110-112.
Bouma, J. (1971). Evaluation of the percolation test and an alternative procedure to test soil potential for disposal of septic tank effluent. Soil Science Society of America Journal, 35, 871-875.
Braker, W.L. (1981). Soil Survey of Centre County, Pennsylvania. Washington, DC: USDA Soil Conservation Service.
Derr, B.C., Matelski, R.P., & Petersen, G.W. (1969). Soil factors influencing percolation test performance. Soil Science Society of America Journal, 33, 942-946.
Elrick, D.E., & Reynolds, W.D. (1986). An analysis of the percolation test based on three dimensional saturated-unsaturated flow from cylindrical test hole. Soil Science, 142(5), 308-321.
Elrick, D.E., Reynolds, W.D., & Tan, K.A. (1989). Hydraulic conductivity measurements in the unsaturated zone using improved well analyses. Ground Water Monitoring Review, 9, 184-193.
Fritton, D.D., Long, D.A., Aron, G., & Petersen, G.W. (1983). Onsite sewage disposal: Site suitability system selection, and soil absorption area sizing (Research Project, Technical Report Project, B-122 Pa, Bureau of Reclamation). University Park, PA: Pennsylvania State University. (NTIS No. PB 83-21925)
Fritton, D.D., Ratvasky, T.T., & Petersen, G.W. (1986). Determination of saturated hydraulic conductivity from soil percolation test results. Soil Science Society of America Journal, 50, 273-276.
Jabro, J.D., & Fritton, D.D. (1990). Simulation of water flow from a percolation test hole in layered soil. Soil Science Society of America Journal, 54, 1214-1218.
Jabro, J.D., Fritton, D.D., & Dobos, R.R. (1991). Spatial variability of field-saturated hydraulic conductivity in four central Pennsylvania soils. Trends in Soil Science, 1, 253-269.
Mulqueen, J., & Rodgers, M. (2001). Percolation testing and hydraulic conductivity of soils for percolation areas. Water Research 35(16), 3909-3915.
Neter J., Wasserman, W., & Kutner, M.H. (1985). Applied linear statistical models (2nd ed.). Homewood, IL: Richard D. Irwin, Inc.
Patterson, R.A. (1996). Soil hydraulic conductivity and domestic wastewater. In Proceedings of Australian and New Zealand National Soils Conference (pp. 207-208). Melbourne: Australian Soil Science Society, Inc.
Reynolds, W.D. (1993). Saturated hydraulic conductivity: Field measurement. In M.R. Carter (Ed.), Soil Sampling and Methods of Analysis (pp. 599-613). Ann Arbor, MI: Lewis Publishers.
Reynolds, W.D., & Elrick, D.E. (2002). Pressure infiltrometer. In J.H. Dane & G.C. Topp (Eds.), Methods of soil analysis: Part 4, physical methods (pp. 826-836). Madison, WI: Soil Science Society of America.
Reynolds, W.D., Elrick, D.E., & Clothier, B.E. (1985). The constant head well permeameter: Effect of unsaturated flow. Soil Science, 139(2), 172-180.
Rodgers, M., & Mulqueen, J. (2004). Field-saturated hydraulic conductivity soils from laboratory constant-head well tests. Irrigation and Drainage Systems, 18(4), 315-327.
Salverda, A.P., & Dane, J.H. (1993). An examination of the Guelph permeameter for measuring the soil's hydraulic properties. Geoderma, 57(4), 405-421.
Schmidt, C.J., Boyle, W.C., Clements, W.V., Otis, R.J., Bauer, D.H., Seigrist, R.L., Tyler, E.J., Stewart, D.E., & Converse, J.C. (1980). Design manual: Onsite water disposal water treatment and disposal systems (U.S. EPA Report No. 625/1-80-012). Washington, DC: U.S. Environmental Protection Agency.
Winneberger, J.T. (1974). Correlation of three techniques for determining soil permeability. Journal of Environmental Health, 37(2),108-118.
Jay D. Jabro, PhD, CPSSc
Corresponding Author: Jay D. Jabro, Research Soil Scientist, Northern Plains Agricultural Research Laboratory, USDA-ARS, 1500 N. Central Avenue, Sidney, MT 59270. E-mail: email@example.com.
TABLE 1 Saturated Hydraulic Conductivities (Ks) at Three Depths for Three Soils Soil Site Hagerstown Monongahela Nolin Hole number Depth [K.sub.s] Depth [K.sub.s] Depth [K.sub.s] (cm) (m [day. (cm) (m [day. (cm) (m [day. sup.-1]) sup.-1]) sup.-1]) 1 30 0.0100 30 0.00909 30 0.10230 2 28 0.0331 31 0.07535 33 0.20700 3 30 0.0032 30 0.04669 31 0.06332 4 30 0.0111 30 0.01872 30 0.07306 5 30 0.0393 33 0.04066 30 0.15586 6 30 0.0016 33 0.01492 30 0.18510 Mean 0.0164 0.0342 0.1311 CV ([dagger]) (%) 97 73 46 1 50 0.0021 52 0.00066 50 0.33120 2 50 0.0526 55 0.00645 50 0.24350 3 50 0.0025 50 0.00322 52 0.09985 4 50 0.0020 52 0.00002 50 0.15343 5 50 0.0045 54 0.01546 52 0.40183 6 50 0.0034 55 0.00548 50 0.19483 Mean 0.0112 0.0052 0.2374 CV ([dagger]) (%) 182 108 48 1 78 0.0079 85 0.00002 75 2.07002 2 78 0.1381 86 0.00065 73 0.04870 3 77 0.0226 91 0.00016 75 4.14004 4 78 0.0026 92 0.01127 76 0.06332 5 79 0.0021 97 0.00193 76 0.25571 6 78 0.0086 91 0.00014 73 0.08037 Mean 0.0303 0.0024 1.1097 CV ([dagger]) (%) 176 187 151 ([dagger]) CV is a coefficient of variation. TABLE 2 Percolation Times (PT) at Depths 50-55 cm for Three Soils Percolation Time (Minute/cm) Hole Number Hagerstown Monongahela Nolin 1 2.9 7.6 1.6 2 4.3 15.5 2.6 3 6.4 11.7 3.1 4 3.5 8.9 1.6 5 9.1 7.6 1.9 6 5.0 11.4 3.1 Mean (min/cm) 5.2 10.5 2.3 CV ([dagger]) (%) 44.0 29.0 31.0 ([dagger]) CV is a coefficient of variation.
|Printer friendly Cite/link Email Feedback|
|Author:||Jabro, Jay D.|
|Publication:||Journal of Environmental Health|
|Date:||Dec 1, 2009|
|Previous Article:||Assessing consumer awareness about mercury in fish.|
|Next Article:||Consumption of groundwater as an independent risk factor of Salmonella Choleraesuis infection: a case-control study in Taiwan.|