Effect of filter separators on the clogging of leachate collection systems

674 Effect of filter separators on the clogging of leachate collection systems Reagan McIsaac and R. Kerry Rowe Abstract: This paper reports the results obtained after 6 years operation of nine mesocosm experiments that simulate the 50 cm of the drainage layer closest to the leachate collection pipe in a landfill. Five different design configurations were examined involving a 300 mm thick layer of coarse (38 mm) gravel. The designs differed in terms of the presence, nature, and location of a filter separator layer either at the waste gravel interface or partway through the gravel. A nonwoven geotextile filter separator (GTF/S) is shown to reduce clogging of the gravel relative to the no filter separator or woven GTF/S designs. Some clogging of the geotextiles is reported, with reductions in geotextile hydraulic conductivity of 23% for the woven GTF/S, 74% 89% for the nonwoven GTF/S, and 75% 94% for the nonwoven geotextile partway through the gravel. The clogged nonwoven geotextile filter separator maintained a higher hydraulic conductivity than the extracted woven geotextile. Of the designs with a filter separator between the waste and gravel, the granular filter separator most effectively reduced clogging of the gravel but at the expense of leachate mounding above the sand once the sand layer clogs. The design with a nonwoven geotextile partway through the gravel (GTMF) provides better protection of the underlying gravel from clogging than other designs involving a geotextile. Key words: landfill, waste, leachate, clogging, biofilm, geotextile. Résumé : On présente les résultats obtenus après six années d opération de neuf expériences de mésocosmes qui stimulent les 50 cm de la couche de drainage la plus près du tuyau de collecte du lixiviant dans un enfouissement sanitaire. On a examiné cinq différentes configurations de conception impliquant 300 mm d épaisseur de gros (38 mm) gravier. Les conceptions étaient différentes en termes de la présence, de la nature et de la localisation de la couche filtre séparateur soit à l interface résidus gravier, soit à mi-chemin à travers le gravier. On a montré qu un filtre séparateur de géotextile non tissé (GTF/S) réduit le colmatage du gravier par rapport aux conceptions sans filtre séparateur ou GTF/S tissé. On rapporte qu il yaducolmatage du géotextile avec les réductions de la conductivité hydraulique du géotextile variant de 23 % pour le GTF/S tissé, de 74 %à89%pour le GTF/S non tissé, et de 75 %à94%pour le géotextile non tissé à mi-chemin à travers le gravier. Le filtre séparateur en géotextile non tissé colmaté a maintenu une conductivité hydraulique plus élevée que le géotextile tissé extrait. Parmi les conceptions comportant un filtre séparateur entre les résidus et le gravier, le filtre séparateur granulaire a réduit le colmatage du gravier de la façon la plus efficace, mais aux dépends de la formation de buttes au-dessus du sable lorsque la couche de sable se colmate. La conception avec un géotextile non tissé à mi-chemin dans le gravier (GTMF) fournit une meilleure protection contre le colmatage du gravier sous-jacent que les autres conceptions impliquant un géotextile. Mots clés : enfouissement sanitaire, résidus, lixiviant, colmatage, biofilm, géotextile. [Traduit par la Rédaction] McIsaac and Rowe 693 Introduction Modern landfills typically have a leachate collection system (LCS) to control leachate mounding on the liner and to collect and remove contaminants that would otherwise be available for transport to the environment. Typically, the LCS comprises a layer ( blanket ) of granular material and a Received 25 June 2005. Accepted 10 March 2006. Published on the NRC Research Press Web site at http://cgj.nrc.ca on 16 June 2006. R. McIsaac. Department of Civil and Environmental Engineering, University of Western Ontario, London, ON N6A 5B9, Canada. R.K. Rowe. 1 Department of Civil Engineering, GeoEngineering Centre at Queen s RMC, Queen s University, Kingston, ON K7L 3N6, Canada. 1 Corresponding author (e-mail: kerry@civil.queensu.ca). network of perforated leachate collection pipes. Field experience (Brune et al. 1994; Fleming et al. 1999; Maliva et al. 2000; Bouchez et al. 2003) has shown, however, that the growth of biofilms on the material in the collection layer induces the deposition of inorganic constituents (predominantly calcium carbonate) from the leachate and the accumulation of particulate matter; this process is referred to herein as clogging. A leachate collection system is said to be clogged when its hydraulic conductivity drops to the point where the leachate head on the liner exceeds the design value (typically 0.3 m; Rowe et al. 2004). This buildup of clog material can occur in the granular drainage layer, in geotextiles, and in the leachate collection pipe perforations or the pipes themselves. The clog that develops decreases the pore space available to transmit leachate, reducing the hydraulic conductivity and consequently the efficiency of the leachate collection system. The physical blocking of the granular drainage layer due to intrusion of waste material Can. Geotech. J. 43: 674 693 (2006) doi:10.1139/t06-030

McIsaac and Rowe 675 Fig. 1. Schematic of experimental mesocosm cells and the prescribed interval spacing and location over which wet mass measurements were made within the mesocosms at termination (modified from Fleming and Rowe 2004). from above is also important. Since leachate collection systems may be required to collect and remove leachate for extended periods of time, it is important to be able to design these systems to minimize the clogging process and prolong the service life of these systems. The work of Brune et al. (1994) and Rowe et al. (2000a) has clearly shown that the particle size of the granular material in the drainage layer is important, with smaller grain size material resulting in faster clogging. These findings demonstrate the desirability of using coarse, uniform gravel. This said, although there are many different opinions regarding what is the most suitable design (as manifest by the many different LCS design configurations used in practice), there is a paucity of data regarding the relative performance of different systems and their susceptibility to clogging. For example, there have been conflicting opinions expressed regarding the suitability of placing a geotextile or granular filter between the waste and underlying granular material. Likewise, there are different opinions regarding whether the geotextile should be placed directly between the waste and gravel or at some location within the gravel layer. Some researchers have opined that because geotextiles experience a reduction in hydraulic conductivity due to clogging, they should not be used in landfill leachate collection systems. While acknowledging clogging, Rowe (1993, 1998) has indicated that a filter between the waste and underlying granular material may be beneficial if used correctly. In support of the latter opinion, Fleming et al. (1999) reported observations from a field exhumation of a leachate collection system with a 50 mm gravel drainage blanket drain at a large municipal landfill site. After 4 years of exposure to municipal landfill leachate, the drainage gravel was found to contain a considerable mass and volume of clog materials. At one location where a sand fill and geotextile were placed between the waste and the drainage gravel, the observed clogging of the underlying gravel was considerably less than at locations where there was no filter present. This suggests a potential beneficial effect of the geotextile in terms of acting as a separator, minimizing intrusion of particulate material into the drainage gravel and the potential for reducing clogging due to leachate treatment that may occur as the leachate passes through the geotextile (or granular filter). To investigate the effect of a number of different design configurations with and without the use of filter separator layers, a series of experiments was initiated as described by Fleming and Rowe (2004). These experiments were designed to examine in real time and at real scale the twodimensional leachate flow conditions adjacent to a leachate collection pipe in a primary leachate collection system involving a continuous granular drainage layer. Details regarding the design and early operation of these cells (called mesocosms) are given by Fleming and Rowe. Relevant to this paper are five different design configurations involving a 300 mm thick layer of coarse gravel (38 mm). Design 1 (denoted herein as no F/S ) had no filter separator layer. Three designs had a filter separator between the waste material and gravel: the filter separator was a nonwoven geotextile for design 2 (nonwoven GTF/S), a woven geotextile for design 3 (woven GTF/S), and a graded granular filter for design 4 (graded granular F/S). Design 5 (nonwoven GTMF) had a nonwoven geotextile partway through the gravel. The primary objective of this paper is to examine the effect of the different design configurations on the clogging of the gravel at the time of the termination of these tests after 6 years operation. Particular attention is paid to the distribution of clog material in the gravel (especially over the last year of operations). A secondary objective is to examine the clogging of the geotextiles used as filter separators. Methodology Mesocosm fabrication and materials The mesocosm collection system cells (Fig. 1; Table 1) were fabricated from welded 9 mm thick PVC sheeting and were built at a large enough scale (internal dimensions measuring 565 mm in length, 235 mm in width, and 574 mm in height) to simulate the last 0.5 m of a leachate collection system adjacent to a leachate collection pipe at full scale and with materials typically used in practice (Fleming and Rowe

676 Can. Geotech. J. Vol. 43, 2006 Table 1. Summary of mesocosm design variables. Mesocosm Filter Average initial gravel porosity Test duration Distance from waste to top of filter (cm) Days Years C-04 None 0.41 2207 6.05 C-01 Nonwoven geotextile 0.43 0 2250 6.16 C-02 Nonwoven geotextile 0.45 0 2182 5.98 C-21 Woven geotextile 0.41 0 2168 5.94 C-22 Woven geotextile 0.42 0 2243 6.15 C-07 Granular filter; 4 cm each of pea gravel and sand 0.46 0 2231 6.11 C-08 Granular filter; 4 cm each of pea gravel and sand 0.44 0 1534 4.20 C-05 Nonwoven geotextile between two gravel layers 0.39 16.3 2245 6.15 C-06 Nonwoven geotextile between two gravel layers 0.48 16.3 2196 6.02 Table 2. Properties and placement in the mesocosms of the Polyfelt TS650 and Terratrack 24-15 geotextiles. Polyfelt TS650 Terratrack 24-15 ASTM test method Manufacturer Polyfelt Geosynthetics Terrafix Geosynthetics Inc. Physical characteristic and polymer Continuous filament, nonwoven, Woven polypropylene needle-punched polypropylene Application Between waste and gravel; between Between waste and gravel two gravel layers Thickness at 2 kpa (mm) 2.3 0.5 D5199 AOS (µm) 110 700 D4751 Mass per unit area (g/m 2 ) 235 140 D5261 Grab tensile strength (N) 755 750 D4632 Permittivity (s 1 ) 1.60 0.04 D4491 2004). The collection system design generally consisted (Fig. 1) of waste material (a mix of refuse and cover soil with a moisture content of 40%) overlying a 300 mm thick gravel drainage layer of crushed dolomitic limestone, overlying a nonwoven geotextile sand cushion graded at 1.5% to a half section of perforated PVC pipe (internal pipe diameter of 102 mm, two rows of perforations, perforation diameter of 15.9 mm, and perforation spacing along the pipe of 127 mm). The crushed dolomitic limestone had a nominal size of 38 mm (D 10 = 20 mm, D 60 = 27 mm, D 85 = 33 mm). Mesocosm C-04 had waste directly over the gravel. The filter separator layers between the waste material and drainage gravel (Fig. 1) were (1) a nonwoven needle-punched polypropylene geotextile (Polyfelt TS650, Table 2) for mesocosms C-01 and C-02, (2) a woven slit-film geotextile (Terratrack 24-15, Table 2) for mesocosms C-21 and C-22, and (3) a graded granular filter consisting of 40 mm each of well-graded concrete sand (D 10 = 0.2 mm, D 60 = 0.7 mm, D 85 = 1.8 mm) and pea gravel (D 10 = 5 mm, D 60 = 8 mm, D 85 = 10 mm) for mesocosms C-07 and C-08. For the mesocosms (C-05 and C-06) with a geotextile in the gravel layer (16.3 cm below the waste), the same nonwoven geotextile was used as in case (1). Mesocosm operation In a typical modern leachate collection system, leachate percolates vertically down through the waste and into the drainage layer. The drainage layer then conveys the leachate horizontally to collection pipes. The mesocosms were designed to mimic these flow conditions adjacent to a collection pipe. As described by Fleming (1999) and Fleming and Rowe (2004), the mesocosms were permeated under anaerobic conditions with leachate collected from the Keele Valley Landfill (Toronto, Ontario) at flow rates representative of field conditions. The vertical infiltration rate was specified to correspond to an annual average percolation of 0.2 m 3 a 1 m 2 through the waste and was approximately 73 ml/day. To simulate horizontal flow, leachate was pumped into the mesocosms at one end at a rate of 2.4 ml/min (1.26 m 3 /a) and, after it flowed horizontally through the drainage gravel, entered the collection pipe and exited out the constant leachate level outflow at the downstream end of the mesocosm. The horizontal flow rate was selected to simulate the average horizontal flow in the drainage layer near the collection pipe for a 25 m drainage path to the pipe. Between collection and use, the leachate was kept stirred in a series of storage tanks in a room maintained at a temperature of 7±2 Cbefore being delivered by gravity as required to a manifold for temperature equalization to that of the mesocosms (i.e., 27 ± 2 C). The tests were conducted at 27±2 Ctosimulate the conditions anticipated in an active leachate collection system (Rowe et al. 2004). Variablespeed peristaltic pumps, each with dedicated pump channels, were used to deliver the leachate from the manifold to the mesocosms. The bottom 100 mm of the drainage gravel in each mesocosm remained saturated (this was the design leachate outflow elevation) and the top 200 mm unsaturated during operation. Experimental analysis A testing program was implemented to monitor and quantify the amount of clogging and changes in leachate compo-

McIsaac and Rowe 677 sition both temporally and spatially. Water-quality testing was performed on leachate samples collected from before the influent valve and after the effluent valve (McIsaac 2006). These samples were tested immediately to measure chemical oxygen demand (COD), calcium (Ca 2+ ) concentration, and ph. Tests were performed to follow the change in drainable porosity (and hence the change in void volume) with time as clogging developed. The procedure adopted involved periodically saturating the entire drainage layer with leachate and then measuring the volume of the leachate removed from each prescribed interval. The measured drainable porosity is the ratio of the volume of the leachate removed to the total volume of the drained interval. The drainable porosity will be lower than the actual porosity because of incomplete draining of the water under gravity due to fluid adhering to the drainage medium and clog material. Termination of nine mesocosms allowed for visual inspection of the degree of clogging that occurred over a 6 year operational lifespan. The clogged drainage gravel was removed and stripped of clog material and mass measurements were used to obtain the amount of wet solids within different intervals (Fig. 1) in the mesocosms. Clog material was stripped from the gravel by first placing the gravel in a sequence of sieves and sieving for 15 min, with the remaining clog material from each stone removed by hand with a hammer and large knife. Wet solids is a measure of the mass of total wet clog material per sample section measured immediately after collection from the mesocosm. It represents the combined mass from biomass, chemical precipitates, and transported fines. The bulk density of the clog material was measured using a modified American Society for Testing and Materials (ASTM) method D854 (ASTM 1998). Based on the mass of clog removed from the disassembled mesocosms, the bulk density of the clog material, and the initial void volume in each sample section, the volume of clog within each section and the resulting void volume occupancy (VVO) were calculated, where VVO is the ratio of the volume of pore space occupied by clog material to the initial void volume. Thus a VVO of 100% would indicate that the initial void volume is completely filled with clog material. Clog samples were sent for elemental analysis to ascertain the composition of the clog material. Hydraulic conductivity testing (ASTM 1992) was conducted on both clean and extracted clogged samples of the geotextile (both woven and nonwoven). The exhumed geotextiles were inspected for clog development using a scanning electron microscope combined with energy dispersive X-ray analysis (SEM EDX) (Hitachi S-4500 field emission SEM combined with a Phoenix EDAM III EDAX system). EDX analysis provides a quantitative method for detecting elements above atomic number five, with a minimum detection limit of approximately 0.5 wt.%, for most elements. An accelerating voltage of 10 kv was used, which has a probing depth of approximately 1.5 µm in a carbon-based substrate. Leachate characteristics The leachate used in the mesocosms was collected from the Keele Valley Landfill (KVL) located near Toronto, Ontario, Canada. This is a 99 ha landfill with a design capacity of approximately 33 000 000 m 3 that received municipal solid waste from 1984 to 2002. The underdrain system is a mix of french drains in the older stages and a full leachate underdrain system in the newer sections. The leachate was collected from a manhole on the main header line at the downstream end of the leachate collection system. The leachate generated at the KVL has been monitored regularly since landfilling began. The analytical composition of the raw KVL leachate used in the mesocosms during their operation is reported in Table 3. It is generally accepted that waste progresses through four stages of decomposition: aerobic, anaerobic acid, initial methanogenic, and stable methanogenic (Christensen and Kjeldsen 1995). The general trends with respect to ranges of concentrations and time dependency of selected parameters of KVL leachate coincide with the numerous landfills studied in Germany by Ehrig (1983), in the US (Wisconsin) by Krug and Ham (1997), and in Denmark by Kjeldsen and Christophersen (2001). During the operation of the mesocosms (1993 1999) the leachate had a high organic strength (dissolved organic carbon (DOC), biochemical oxygen demand (BOD 5 ), and COD values) indicative of a young leachate. The organic strength was mainly in the form of volatile fatty acids (McIsaac 2006). The BOD to COD ratio and ph from 1993 to 1999 indicate that the Keele Valley leachate was initially acetogenic with a gradual transition into the initial methanogenic stage after 1996. A drop in the organic strength of the leachate in 1994 and 1995 with a subsequent increase in 1996 was attributed to the placement of fresh waste over old in conjunction with lower total precipitation between 1992 and 1994 (Armstrong and Rowe 1999). Both the influent and effluent leachate were monitored with time but they are not discussed herein because the focus of this paper is on the effect of the filters separators on the clogging of the gravel. The interested reader is referred to McIsaac (2006). Drainable porosity results The initial drainable porosities in the mesocosms ranged between 0.39 and 0.48. Drainable porosity measurements made within the mesocosms for the first 1500 days were reported by Fleming and Rowe (2004). Figures 2 and 3 show the variation in the average drainable porosity in the unsaturated and saturated zones from inception to termination. In the present paper, particular attention is focused on the difference in clogging with position in the gravel. Figure 4 illustrates pictorially a vertical profile through each mesocosm showing the material filling each mesocosm and the interval spacing and location over which the drainable porosities were measured within the mesocosms. Narrow intervals were used in the saturated gravel layer and in the unsaturated gravel directly above and below the design leachate outflow level. The measured interval height and locations were selected to monitor the clog development directly above and below the nonwoven GTMF in the gravel layer. The variation in drainable porosity measured within each interval within the mesocosms and its change with time between 1600 days and test termination (2168 2252 days) are shown in Figs. 5 and 6.

678 Can. Geotech. J. Vol. 43, 2006 Table 3. Characteristics of landfill leachate collected from the Keele Valley Landfill during the operation of the mesocosms. Parameter Year a 1993 (10) 1994 (11) 1995 (12) ph 6.5 7.4 7.1 7.0 7.1 7.3 7.4 DOC 5 470 2 140 1 760 5 510 4 510 3 750 3 630 BOD 5 12 210 3 910 3 510 11 350 8 110 6 810 5 130 COD 16 150 5 990 5 050 17 060 13 370 11 380 9 520 BOD 5 /COD 0.75 0.95 0.71 0.81 0.60 0.59 0.53 Conductivity 16 350 14 780 10 120 18 450 17 830 17 610 19 700 Calcium 1 250 430 470 1 190 630 610 470 Sulphate 1 9 16 227 153 164 93 Iron 185 18 50 222 225 192 122 Manganese 10 410 970 3 630 9 460 6 680 4 590 3 180 Chloride 2 230 2 420 1 380 2 380 3 050 2 960 3 020 Ammonia-N 500 540 400 880 970 1 010 1 130 Potassium 760 720 430 760 930 1 000 970 Sodium 1 500 1 660 910 1 680 2 010 2 100 2 200 Zinc 1.11 0.35 0.71 2.77 7.91 7.36 8.35 Cadmium 0.027 0.004 0.007 0.010 0.026 0.026 0.003 Chromium 0.239 0.061 0.118 0.237 0.166 0.140 0.159 Lead 0.027 0.016 0.013 0.008 0.011 0.022 0.026 Mercury 0.0003 0.0096 0.0003 0.0001 0.0001 0.0001 0.0002 Magnesium 450 420 210 410 430 470 450 TKN 600 620 440 980 1 110 1 100 1 240 Arsenic 0.041 0.007 0.004 0.002 0.031 0.030 0.058 Cobalt 0.039 0.000 0.009 0.051 0.051 0.056 0.075 Copper 0.023 0.009 0.035 0.015 0.014 0.021 0.020 Nickel 0.308 0.149 0.216 0.806 0.685 0.576 3.322 Barium 0.442 0.142 0.195 0.273 0.296 0.278 0.267 Selenium 0.390 0.003 0.002 0.006 0.030 0.006 0.016 Silver 0.017 0.011 0.005 0.007 0.006 0.000 Aluminum 1.20 0.33 0.55 0.85 1.03 1.49 1.81 Note: All parameters are in mg/l, except ph. TKN, total Kjeldahl nitrogen. a Number of years of operation is given in parentheses. 1996 (13) 1997 (14) 1998 (15) 1999 (16) Clogging tended to decrease as one moved up from the bottom of the gravel, with the greatest clogging being in the initially saturated zone (Figs. 5, 6). Once the upper portion of the saturated gravel ( 2 to 0 cm) reached a drainable porosity of approximately 0.25 0.20, the drainable porosity in the interval above it (0 2 cm) began to decrease. This may be associated with clogging of the lower gravel and the lower perforations in the leachate collection pipe to the point where the leachate level rose to allow flow through the uppermost perforation in the pipe, causing saturation of the interval from 0 to 2 cm. This resulted in a consequent increased rate of clogging in this zone (attributed to the increased mass loading as leachate preferentially flowed in this less clogged zone toward the upper perforations). The clog within the pipes was soft black runny biological ooze (Fig. 7). A hard clog deposit only existed as a thin layer on the inside surface of the pipes that was constantly submerged saturated. These results show the importance of keeping the leachate level low (minimizing the thickness of the saturated zone) and the desirability of keeping the pipe and pipe perforations from clogging by regular cleaning of the pipes. The much greater rate of clog development within the saturated gravel layers compared with the unsaturated gravel (Figs. 5, 6) is attributed to the higher mass loading in the saturated zone due to the higher flow rate. This is consistent with the findings of Rowe et al. (2000b). One can also identify a difference in the rate of clogging in the unsaturated zone in the different mesocosms, however, and this can be associated with the different filter separator designs. High drainable porosities (typically greater than 0.40) were measured within the unsaturated gravel layer adjacent to the waste in the mesocosms that incorporated a separator between the waste and the gravel, for example in the 14 18 cm interval within the design incorporating the nonwoven geotextile (Figs. 5b, 5c), the woven geotextile (Figs. 5d, 5e), and the granular filter (Fig. 6a). Reduced drainable porosity measurements (typically around 0.35) were obtained in the same interval in designs without a separator (i.e., Figs. 5a, 6b, and 6c). Thus the filter separator between the waste and the gravel reduced waste intrusion into the gravel. Clogging of the saturated zone was also influenced by the design. The mesocosms with the woven GTF/S (Figs. 5d, 5e) had generally lower drainable porosities in the saturated zone than those with a nonwoven GTF/S (Figs. 5b, 5c). In fact, mesocosm C-21 designed with a woven GTF/S exhibited the lowest drainable porosities of all mesocosms with a

McIsaac and Rowe 679 Fig. 2. Variation in drainable porosity with time: (a) no F/S, C-04; (b) nonwoven GTF/S, C-01; (c) nonwoven GTF/S, C-02; (d) woven GTF/S, C-21; (e) woven GTF/S, C-22 (data up to 1400 days from Fleming and Rowe 2004). filter separator, with values of less than 0.21 (even within the 0 2 cm interval) by 1625 days (4.6 years). Decreased drainable porosity values were measured in the 2 6 cm interval as early as 1750 days. Although not as severe as C-21, its duplicate C-22 had drainable porosities that were generally less than 0.22 in the saturated gravel after 1625 days. The design with the graded granular filter (Fig. 6a) had the lowest clogging of the saturated zone, and the drainable porosity never dropped below 0.20. Disassembly results (discussed later in the paper) indicate that the granular filter may be the most effective at reducing the leachate mass loading to the gravel and the amount of fines entering the gravel layer as the leachate percolates vertically through the system, resulting in the higher observed drainable porosity within the gravel drainage layer. For the designs with a nonwoven GTMF (Figs. 6b, 6c) in the gravel, the measured drainable porosities indicate that the geotextile was a site for clog development. Lower drainable porosities were measured within the interval that contained the geotextile and decreased with time, indicating that the geotextile was clogging. This can be seen in Figs. 6b and 6c for the 2 6 cm interval, which contained the geotextile. The interval directly above the geotextile (6 10 cm) had lower drainable porosities than the overlying unsaturated

680 Can. Geotech. J. Vol. 43, 2006 Fig. 3. Variation in drainable porosity with time: (a) graded granular F/S, C-07; (b) graded granular F/S, C-08; (c) nonwoven GTMF, C-05; (d) nonwoven GTMF, C-06 (data up to 1400 days from Fleming and Rowe 2004). Fig. 4. Vertical profiles through the mesocosms showing the interval spacing and location over which drainable porosities were measured within the mesocosms: (a) no F/S (C-04), nonwoven GTF/S (C-01, C-02), woven GTF/S (C-21, C-22); (b) graded granular F/S (C-07); (c) nonwoven GTMF (C-05, C-06). gravel, consistent with observations of higher amounts of clog above the geotextile than elsewhere in the unsaturated zone. The drainable porosities above the geotextile were still significantly greater than that measured in the saturated gravel, however, and would likely still provide an alternate means for leachate flow should the geotextile clog sufficiently to prevent the free flow of leachate through it with leachate perching.

McIsaac and Rowe 681 Fig. 5. Drainable porosities in the mesocosms designed with (a) no F/S (C-04), (b) nonwoven GTF/S (C-01), (c) nonwoven GTF/S (C-02), (d) woven GTF/S (C-21), and (e) woven GTF/S (C-22). Disassembly results No filter separator design findings Without a separator between the waste and the gravel drainage layer, the waste material physically intruded (due to compaction and settlement of the waste) into the relatively large open voids to a depth of up to 60 mm, partially filling the voids (Fig. 8a) and reducing the effective thickness of the drainage layer. Approximately eight times more wet solids (Fig. 9) were measured within this interval (260 300 mm) than in mesocosms designed with a separator. The filter separator between the waste and the gravel drainage layer in the other mesocosms prevented this physical intrusion of waste material and the subsequent reduction in void volume within the top 60 mm of the gravel drainage layer. The absence of a filter separator resulted in more clog mass throughout the drainage layer, higher VVO values (Table 4), and more severe cementation of the gravel within the saturated layer than was observed for designs incorporating a filter separator between the waste and the gravel drainage layer. There was a distinct difference in the visual quantity of clog (Figs. 8d, 8e), and generally more clog mass

682 Can. Geotech. J. Vol. 43, 2006 Fig. 6. Drainable porosities in the mesocosms designed with (a) graded granular F/S (C-07), (b) nonwoven GTMF (C-05), and (c) nonwoven GTMF (C-06). (Fig. 9) was measured within all of the sample intervals in the gravel for the designs without a filter separator than for those with a filter separator between the waste material and the drainage gravel. Of all the designs, the one with no filter separator had the highest VVO values within the saturated gravel layer (40% within the upper (50 100 mm) and 76% within the lower (0 50 mm) saturated intervals) (Table 4). The use of a nonwoven GTF/S reduced the VVO within the upper and lower saturated gravel layers to 28% and 69%, respectively, and the use of the nonwoven GTMF in the gravel layer reduced the VVO to 35% and 56% within the upper and lower saturated gravel layers, respectively. For the designs without a waste gravel separator, visibly more fines accumulated on the top surfaces of the unsaturated gravel, and biofilm development was more abundant (i.e., more of the gravel surface was covered with biofilm), was thicker, felt grittier, and had more of a grey tan colour (compared to black) than for the design with a filter separator. A screwdriver was required to pry apart and remove all the gravel within the saturated gravel layer, indicating a higher degree of cementation of the gravel particles to each other within the saturated gravel layer in mesocosms without a filter separator. In contrast, the gravel could be removed by hand when a filter was used. Filter separator designs Woven geotextile filter separator design (woven GTF/S) The design incorporating a woven GTF/S exhibited more biological, physical, and chemical clogging, within both the unsaturated and saturated layers of the drainage gravel, than was observed for the nonwoven GTF/S designs, and the clogging approached what was observed when there was no filter separator (Fig. 9). Progressing down through the unsaturated gravel layer below the woven geotextile, the amount of biofilm increased and covered the flat top portion of the gravel where fluid and fines could accumulate. The biofilm that accumulated on the top of the gravel was less than 3 mm thick, soft, and viscous with a gritty feel, suggesting the accumulation of chemical precipitation and (or) particulates. The leachate flowing vertically through the woven geotextile and unsaturated gravel was of sufficient strength to support more biological growth on the unsaturated gravel than was the case with the nonwoven geotextile and granular filter designs. The degree of cementation (or the force required to separate and remove the cemented gravel) within the saturated gravel layer was greater for the woven GTF/S design than for the designs with a filter (nonwoven GTF/S, graded granular F/S, or nonwoven

McIsaac and Rowe 683 Fig. 7. Soft black runny ooze that accumulated in the bottom of the half section of perforated pipe in the mesocosms. This, in conjunction with the high degree of clogging within the base saturated gravel layer, was sufficient to prevent the free flow of leachate through the bottom perforations of the pipe. The pipe was essentially pristine where unsaturated (i.e., see top section of the pipe above the perforations). GTMF). The soft clog was more dense, less slimy and runny, and grittier than the soft clog in the saturated gravel layer of the nonwoven GTF/S design. Higher amounts of silicon and aluminum associated with the accumulation of fines from the waste material were measured in the clog (Table 5) for the woven GTF/S design than for the designs with a nonwoven or granular filter. Consistent with the foregoing observations of the filter separator designs, the design incorporating the woven geotextile had the greatest VVO throughout the drainage gravel at termination (Table 4). The differences in the clog texture and composition (Table 5), the amount of clog (Table 4), and the degree of cementation within the saturated gravel layer for the woven GTF/S compared with that for those having other filter separator designs suggest that the migration and accumulation of fines (originating from the waste material) into the gravel and pipe and the lower degree of leachate treatment within the woven geotextile (compared with the nonwoven geotextile and granular filter designs) influenced the clog mechanism and as a result the drainage gravel below the woven geotextile design experienced increased clog compared to the nonwoven geotextile and granular filter designs. Thus, while the woven GTF/S provided better performance than that without a filter separator, this was the least desirable filter separator design. There was no appreciable biological, chemical, or physical clogging of the woven geotextile, the waste material above it, or the underlying gravel in contact with the geotextile. There was no biofilm visually evident where the woven geotextile made direct contact with the underlying gravel. Biofilm did grow in the capillary fringe of retained leachate between the gravel and the woven geotextile. It is hypothesized that capillary action between the woven geotextile and the gravel provided a moist environment for microbial activity that caused the accumulation of biofilm to occur at these locations. There was no cementation of the geotextile to either the waste or gravel, however. When removed from the mesocosm, after 6 years, there was very little clog material on the woven geotextile. SEM images of the top and bottom surfaces (Fig. 10) of the extracted woven geotextile show the presence of clog on its surface, but not an abundant accumulation. Only a very thin, relatively uniform coating of clog material covered the surface of the woven geotextile and caused a partial occlusion of the openings between the fibres. EDX analysis of the material on the surface of the woven geotextile indicates the presence of silicon and aluminum due likely to fines (siltation) and calcium, oxygen, and carbon from chemical precipitate; however, there was no evidence of a filter cake on the surface of the woven geotextile that would have impeded the migration of fines from the waste material into the gravel. This is likely due to the high apparent opening size (AOS = 700 µm) of the woven GTF/S. There was no pooling of leachate on the surface of the woven geotextile, indicating that the permittivity was still sufficient to readily allow the flow of leachate through the geotextile. The extracted woven geotextile experienced only a 23% reduction in hydraulic conductivity from its original measured clean value of 2 10 5 m/s (Fig. 11). The absence of biological and chemical clogging above, directly below, or within the woven geotextile itself indicates that there was very little passive treatment of the leachate in these areas and that, consequently, essentially all the inorganic and organic constituents in the leachate entered the underlying gravel drainage layer with negligible attenuation. Nonwoven geotextile filter separator design (nonwoven GTF/S) The nonwoven GTF/S reduced the amount of clog material in all sections of the underlying drainage layer (Fig. 9) by a factor of 1.2 (saturated gravel) to 2.4 (unsaturated

684 Can. Geotech. J. Vol. 43, 2006 Fig. 8. (a c) Waste gravel interface (260 300 mm). (d f) Unsaturated gravel layer (100 180 mm). Waste material occludes the voids when no separator is used (a). The use of a separator (nonwoven GTF/S (b) and woven GTF/S (c)) prevented occlusion of voids with intruded waste material. The amount of clog within the gravel layer protected by a woven GTF/S ( f ) approached that observed in the designs with no F/S (d). The presence of a nonwoven GTF/S reduced the amount of clog material in the underlying gravel drainage layer (e). Length of scale is 5 cm. gravel) relative to that of the design without a geotextile. The nonwoven GTF/S was also more effective than the woven GTF/S but not as effective as the granular F/S design in terms of reducing clogging of the underlying gravel. Examination of the unsaturated gravel layer between elevations of 100 and 300 mm for the mesocosms incorporating the nonwoven GTF/S revealed fewer fines (i.e., less siltation) than was observed with a woven GTF/S. Correspond-

McIsaac and Rowe 685 Fig. 9. Distribution of wet solids within the mesocosms. Table 4. Void volume occupancy (VVO) within the mesocosms. Interval (mm) No F/S (C-04) Nonwoven GTF/S (C-01) Woven GTF/S (C-22) GGF/S (C-07) 260 300 56 7 8 7 32 180 260 9 7 9 7 8 100 180 7 9 11 8 14 50 100 40 28 40 27 35 0 50 76 69 75 53 56 Nonwoven GTMF (C-05) Note: F/S, filter separator; GGF/S, graded granular filter separator; GTF/S, geotextile filter separator; GTMF, geotextile partway through the gravel. ingly, the biofilm was less gritty and there was visibly less soft black biological film for the mesocosms with a nonwoven GTF/S than for those with a woven GTF/S (Fig. 8). Although the composition of the clog material collected from the bottom saturated gravel interval (Table 5) indicates only a small difference in fines content from that of either the nonwoven or woven geotextile designs, the silicon and aluminum contents in the clog material removed from the nonwoven GTF/S design were lower (Si = 1.22%/dry, Al = 0.25%/dry) than those for the woven GTF/S design (Si = 1.97%/dry, Al = 0.43%/dry), indicating that the nonwoven geotextile may have been slightly more effective as a filter than the woven geotextile by reducing the amount of fines (clay particulates, siltation) mobilized and transported by the vertical flow of leachate to the gravel layer from the overlying waste material. SEM and EDX analysis (Fig. 12) of clog from the nonwoven GTF/S shows traces of Si and Al, indicating that fines (clay particles) originating from the waste material likely became embedded in the clog that developed in the nonwoven geotextile. There was no evidence of a blinding layer on the top surface of the nonwoven GTF/S nor was there a uniform development of biofilm or a significant accumulation of fines throughout the internal structure of the nonwoven geotextile. Clogging of the nonwoven GTF/S was observed, however, primarily due to the accumulation of precipitates. SEM photographs of the clog within the nonwoven GTF/S are shown in Fig. 13. In general, clog that developed in the nonwoven geotextile did so by attachment to the geotextile fibres and seems to have propagated away from the intersection of fibres. Clog development in the nonwoven GTF/S was relatively uniform and not exclusive to the gravel geotextile contacts; however, intense clog developed within the geotextile at the locations where the gravel made contact with the nonwoven geotextile (Fig. 14a) which cemented the gravel particles to the geotextile. The geotextile could be held from one end, and the gravel would not fall off (Fig. 14b); substantial force was required to tear the gravel away from the nonwoven geotextile. This is attributed to the fact that the gravel below the geotextile prevented vertical movement of leachate through the geotextile at those loca-

686 Can. Geotech. J. Vol. 43, 2006 Table 5. Composition of clog removed from the mesocosms. Parameter No F/S (C-04), lower saturated gravel layer Nonwoven GTF/S (C-01), lower saturated gravel layer Woven GTF/S (C-22), lower saturated gravel layer GGF/S (C-07), lower saturated gravel layer Nonwoven GTMF (C-05) Lower saturated gravel layer Water content (%/wet) 74.5 66.9 61.1 72.2 65.6 Organic matter (TVS; %/dry) 14.9 12.6 12.0 15.5 12.6 Carbonate as CO 3 (%/dry) 45.7 48.7 47.8 49.0 51.0 51.2 Calcium, Ca (%/dry) 25.6 27.0 26.4 25.0 26.4 29.7 Magnesium, Mg (%/dry) 2.24 2.52 2.70 4.06 3.91 2.20 Silicon, Si (%/dry) 0.79 1.22 1.97 0.43 0.61 2.19 Iron, Fe (%/dry) 5.10 4.16 4.22 4.40 3.80 3.97 Sodium, Na (%/dry) 0.85 0.59 0.57 0.72 0.53 0.34 Aluminum, Al (%/dry) 0.19 0.25 0.43 0.14 0.19 0.49 Potassium, K (%/dry) 0.42 0.33 0.38 0.36 0.30 0.30 Phosphorus, P (%/dry) 0.11 0.12 0.07 0.07 0.07 0.04 Titanium, Ti (%/dry) 0.01 0.01 0.02 0.01 0.01 0.03 Manganese, Mn (%/dry) 0.02 0.02 0.03 0.01 0.02 0.04 Strontium, Sr (mg/kg) 1000 1020 820 840 890 770 Barium, Ba (mg/kg) 120 140 120 110 120 190 Ca/CO 3 0.56 0.56 0.55 0.51 0.52 0.58 Note: TVS, total volatile solids. Other abbreviations as in Table 4. Geotextile surface Fig. 10. SEM photographs of clogged woven GTF/S: (a) top surface; (b) bottom surface. tions, and hence any movement of leachate in the geotextile had to be lateral to the edge of the gravel. This gave rise to greater retention of leachate on the top lateral surfaces of the gravel than in the adjacent geotextile where vertical flow could occur, and consequently allowed the chemical reactions involved with biochemical precipitation to occur to completion. EDX analysis illustrates that the clog is mainly comprised of calcium, carbon, and oxygen likely in the form of calcium carbonate. The SEM EDX analysis of the clog did indicate that the leachate was being treated, albeit at the expense of some geotextile clogging. The hydraulic conductivity results from the clean nonwoven geotextile and the partially clogged geotextile removed from the mesocosms indicate a 74% 89% reduction in hydraulic conductivity from its initial measured clean value of 4.4 10 4 m/s (average of three samples; Fig. 11). The nonwoven geotextile experienced a greater loss in hydraulic conductivity than the woven geotextile because its structure was more prone to clog formation. The hydraulic conductivity was still sufficiently high that ponding did not occur on the geotextile even after 6 years exposure and after drainable porosities within the base of the gravel had dropped below 0.20. The nonwoven GTF/S was not a critical location with respect to clog development (which was in the underlying saturated zone). Although the nonwoven geotextile experienced greater clogging and thus greater reductions in hydraulic conductivity than the woven geotextile, the measured hydraulic conductivities from the clogged samples of nonwoven geotextile from the mesocosms after 6 years exposure to leachate were all

McIsaac and Rowe 687 Fig. 11. Initial and final measured hydraulic conductivities of the geotextile filter separators: (a) nonwoven geotextile; (b) woven geotextile. Fig. 12. SEM photograph (a) and EDX trace (b) of particulate embedded in clog material within the nonwoven GTF/S. greater than those measured across the woven geotextile samples (Fig. 11). Granular filter design (graded granular F/S) Clogging occurred within the 40 mm thick layer of wellgraded concrete sand of the granular filter. The entire top layer of the sand was completely cemented to a depth of approximately 10 mm at the sand waste interface. Within the remainder of the sand layer approximately 50% of the sand volume was made up of hardened concreted zones. The sand was also cemented to the pea gravel (Fig. 15). The concreted zones varied in size from 15 to 40 mm in length, 20 to 40 mm in width, and about 10 to 15 mm in depth. The void space between the sand grains was generally less than 2 mm. It is hypothesized that capillary action within the sand allowed the retention of leachate and provided an environment in which increased biological activity and clogging could occur. Fines and suspended solids migrating with the vertical flow of leachate through the granular filter could physically be retained through filtering and straining in the small pore throat openings of the sand, an effect that would be further enhanced by material adhering to the biofilm within these zones. The clog material removed from the gravel drainage layer below the filter had the lowest amount of silicon and aluminum (Table 5) of all the designs. Due to the small pore size of the sand layer, the clog development and cementation of the sand are more extensive than in the large-diameter gravel used above the nonwoven GTMF design. In general, the pea gravel layer of the granular filter was uncemented and easily removed. The exceptions were at the sand pea gravel interface where it appears that the sand had partially infiltrated into the pea gravel layer and created an environment favourable for chemical clogging that cemented the pea gravel together. Zones of blackening due to biofilm growth within the pea gravel were restricted to the sand gravel interface. The treatment of the leachate within the granular filter and the possible filtering and straining of the fines and suspended solids from the waste material served to minimize clog formation within the underlying gravel. The mesocosm with the granular filter separator design had the least amount of wet clog solids accumulated in either the saturated or unsaturated layers of the drainage gravel (1.4, 1.4, and 1.2 times less than that for the case with no F/S, woven

688 Can. Geotech. J. Vol. 43, 2006 Fig. 13. SEM photographs of clog formation within nonwoven GTF/S: (a) clog formation on fibres; (b, c) clog development at intersection of fibres. GTF/S, and nonwoven GTF/S, respectively, in the saturated gravel layer and 2.5, 1.3, and 1.0 times less, respectively, than that in the unsaturated gravel layer). The drainage gravel was not as occluded with clog material as in the other designs, as indicated by the lower VVO values (Table 4) for the granular filter than for all other designs. The clog material removed from the saturated layer of the granular filter mesocosm was notably more slimy and less gritty than that removed from the other mesocosms. Although the granular filter design reduced the risk of clogging of the gravel drainage layer (and consequent buildup of head on the base of the landfill), the sand layer experienced significant particulate, chemical, and biological clogging. A zone of reduced permeability (as indicated by the clogged sand) was created above the pea gravel in the sand layer of the granular filter system. This would allow the development of a perched leachate mound within the waste once clogging of the sand layer occurs. Nonwoven geotextile partway through the gravel design (nonwoven GTMF) This design featured a 163 mm thick gravel layer between the waste and a nonwoven geotextile filter. Due to the placement of the nonwoven geotextile in the drainage gravel layer (approximately 165 mm from the base of the gravel layer), there was only 65 mm (or less than two particle diameters) of unsaturated gravel below the geotextile. The gravel within the top portion of this layer was influenced by the clog development at the particle geotextile interface, and the gravel near the bottom of this layer was influenced by partial submersion in leachate, thus the wet mass data in Fig. 9 are omitted. This made it difficult to quantitatively assess (with measured wet mass values) the impact of this design on the clog development within the underlying unsaturated gravel. Visually, however, the surfaces of the gravel particles were essentially pristine, with no appreciable accumulations of biofilm or fines. Also, the clog material that accumulated within the saturated gravel layer was a soft, moist, thick, slimy, and viscous film covering the gravel in contrast to the gritty clog material that essentially filled the voids from the 0 50 mm sample interval in the mesocosms designed with no F/S and with a woven GTF/S designs. The amount of silicon in the saturated gravel of the nonwoven GTMF design was less than that in the nonwoven GTF/S, woven GTF/S, or no F/S designs, indicating that the nonwoven GTMF did prevent most of the fines from entering the underlying gravel. The percentage of the void space occluded with clog (VVO = 56%) in the lower saturated gravel layer (0 50 mm)

McIsaac and Rowe 689 Fig. 14. Contact between gravel and nonwoven GTF/S. (a) Location where gravel was removed from the nonwoven geotextile illustrates that the clogging of the nonwoven geotextile itself was not uniform and was most severe where there was intimate contact of the gravel with the geotextile. (b) Gravel cemented to the geotextile. Gravel stuck to the geotextile when hung vertically on end, and substantial force was required to remove gravel from the geotextile. Fig. 15. The sand layer of the granular filter system experienced significant clogging. Concreted zones of cemented sand (a) and cemented pea gravel (b) created a zone of reduced permeability and perched leachate mounding within the waste once the sand layer clogged. was significantly less than that for the nonwoven GTF/S (69%), woven GTF/S (75%), and no F/S (76%) designs. As in the mesocosms designed without a separator, the waste material physically intruded into and partially occluded the voids within the 260 300 mm interval (Fig. 9) for the nonwoven GTMF design, reducing its effective drainage thickness, and more clog material was observed and measured in the 180 260 mm interval than for the nonwoven GTF/S and granular F/S designs. As with the no F/S design, some gritty clog accumulated on the top surfaces and at contact points between the gravel in the unsaturated layer. The gravel layer below the zone affected by waste intrusion and above the nonwoven GTF/S did not experience significant clog development (i.e., occlusion of its voids with clog was less than 14% in the 100 180 mm and 180 260 mm intervals), however, and remained very permeable. Unlike in the sand layer of the graded granular F/S design and the nonwoven GTF/S, there would need to be a very substantial buildup of clog material in the large voids of the sacrificial gravel layer above the geotextile before there was a sufficient reduction in the hydraulic conductivity to impact on the lateral flow capacity to leachate collection pipes. A design configuration that would allow leachate to enter a manhole from both above and below the geotextile layer has been described in Rowe et al. (2000c). Clogging in the unsaturated gravel focused primarily on

690 Can. Geotech. J. Vol. 43, 2006 Fig. 16. The clog mechanism within the nonwoven GTMF design involved the formation of a layer of black gritty ooze (filter cake) on the top of the nonwoven geotextile (a), pooling of the leachate (b), and biofilm growth and the eventual cementation of the geotextile itself (c). Length of scale in (a) and (b) is5cm. the top surface of the nonwoven GTMF. A layer of black gritty ooze (filter cake) formed on the nonwoven geotextile (Fig. 16a). This filter cake varied in thickness from 2 to 6 mm over the surface of the geotextile. At one location on the geotextile surface, clogging of the geotextile itself was sufficient to cause pooling (4 mm deep and 120 mm in diameter, Fig. 16b) of leachate. At this location the filter cake was thicker, the geotextile itself was hard and cemented, and the gravel located in this pool of leachate required substantially more force to remove from the surface of the geotextile than the other gravel. Biofilm did not accumulate on the bottom side of the geotextile except for a thin coating (1.5 mm thick) covering the hard clog hanging from the geotextile coinciding with the location of the pooled leachate. SEM images of the ooze (Fig. 17) show an accumulation of spherical sand-sized material, mat-like deposits of biofilm, crystal forms precipitation, and biologically induced cementation (chemical clog). Elemental analysis (Table 5) showed that the ooze had higher silicon (2.19%/dry) and aluminum (0.49%/dry) content than the clog removed from below the geotextile in the saturated layer (Si = 0.61%/dry, Al = 0.19%/dry). The accumulation of this mixture of biofilm and fines was more abundant around the perimeter of the contacts of the gravel with the geotextile (indicating fines were being transported in the leachate as the leachate percolated down through and over the surface of the gravel). The entire exhumed geotextile was essentially saturated (moisture content of 180%). Leachate squeezed from the geotextile when lightly compressed. In comparison, in the nonwoven GTF/S design (the same geotextile was used in both designs) the moisture content was much lower, at 117%. X-ray diffraction (XRD) powder pattern analysis from samples of air-dried clog material from the top and bottom surfaces of the nonwoven GTMF and from the saturated gravel layer (McIsaac 2006) indicated (i) low Mgcalcite, or chemical clog, was dominant in the clog removed

McIsaac and Rowe 691 Fig. 17. SEM photographs of (a) ooze covering top surface of nonwoven GTMF, (b) sand-sized particulates, (c) gypsum crystal growth, and (d) development of hard biochemical clog on the underside of the nonwoven GTMF. from all locations; (ii) quartz was significantly more abundant from the clog on the upper surface of the geotextile than from the clog on the bottom surface, and no quartz was present in the saturated layer sample; and (iii) gypsum was also more abundant on the upper surface of the geotextile than on the bottom of the geotextile, with none in the saturated gravel layer. Silicon and aluminum content in the clog removed from below the nonwoven GTMF design was much lower than that for the nonwoven GTF/S design (Table 5), although both designs incorporated the same type of nonwoven geotextile, indicating that the nonwoven GTMF was more effective as a filter (i.e., preventing the migration of fines through the geotextile) and provided better protection to the drainage layer below, resulting in less clogging of the underlying gravel than with other designs except the graded granular filter. The greatest reduction in hydraulic conductivity was measured across a sample from the nonwoven GTMF design. In one sample, an order of magnitude drop in hydraulic conductivity was measured between the initial clean value of 4.4 10 4 m/s and the clogged value of 2.8 10 5 m/s. Despite the presence of a filter cake (ooze) and some clogging of the geotextile itself, the geotextile remained sufficiently permeable to transmit leachate without significant ponding of leachate over the geotextile. The lowest hydraulic conductivity measured within the clogged nonwoven geotextile was greater than that measured across the extracted woven geotextile (Fig. 11). Practical implications Higher amounts of clog and higher rates of clog formation were measured in the saturated gravel than in the unsaturated gravel. This was due to the combination of higher mass loading, a longer residency time, and a more favourable environment for biological growth within the saturated gravel. These results are consistent with the findings of Rowe et al. (2000b) and confirm the importance of minimizing mass loading (e.g., by closer spacing of leachate collection pipes) in terms of extending the service life of leachate collections systems. The increased clogging of the gravel that was evident with an increase in the leachate level (due to clogging of the lower gravel and (or) perforations in the collection pipe in the mesocosm tests) highlights the importance of minimizing the thickness of the saturated gravel (e.g., by regularly pumping leachate and not allowing it to accumulate within the leachate collection system). The clog within the pipes was predominantly a soft, black, runny ooze (mainly biological) (Fig. 7). A hard clog deposit only existed as a thin layer on the inside surface of the pipes where they were constantly submerged saturated. Thus, regular cleaning of the leachate collections pipes would be able