- Email: [email protected]
Intermittent filtration of wastewater—removal of fecal coliforms and fecal streptococci
Water Research 36 (2002) 3507–3516
Intermittent filtration of wastewater—removal of fecal coliforms and fecal streptococci G. Auslanda,*, T.K. Stevikb, J.F. Hanssenc, J.C. K hlerd, P.D. Jenssenb a Elkem Research, P.O. Box 8040 Vaagsbygd, 4602 Kristiansand, Norway ( Norway Department of Agricultural Engineering, Agricultural University of Norway, P.O. Box 5065, 1432 As, c ( Norway Department of Biotechnological Sciences, Agricultural University of Norway, P.O. Box 5065, 1432 As, d ( Norway Center for Soil and Environmental Research, 1432 As, b
Received 7 July 2000; received in revised form 14 January 2002; accepted 1 February 2002
Abstract Removal of fecal coliforms and fecal streptococci was monitored over a period of 13 months in 14 buried pilot scale filters, treating septic tank effluent. The effects of grain size, hydraulic dosing rate and distribution method were investigated. Two different natural sands (sorted sand and unsorted sand) and three different types of light weight aggregates (LWA 0–4 mm, LWA 2–4 mm and crushed LWA 0–3 mm) were used. Intermittent dosing rates from 20 to 80 mm/day in 12 doses per day were applied to the filters by uniform pressure distribution or point application by gravity dosing. Removal of fecal coliforms was more than three orders of magnitude higher in the media with the finest grain sizes (unsorted sand) as compared to the coarsest media (LWA 0–4 mm and LWA 2–4 mm) operated under same conditions. Fecal streptococci were determined only in effluent from filters with LWA 0–4 mm and LWA 2–4 mm. Higher removal of fecal coliforms was observed in pressure dosed filters compared to gravity dosed filters. A lower removal was observed by increasing the hydraulic dosing rate. Minimum retention time was found to be a key parameter for predicting removal of bacteria in unsaturated, aerobic filters. At minimum retention times lower than about 50 h, there was a correlation of 0.96 between retention time and removal of fecal coliforms. Retention times longer than 50 h gave almost complete removal of fecal coliforms. r 2002 Elsevier Science Ltd. All rights reserved. Keywords: Biological filters; Wastewater; Fecal coliforms; Fecal streptococci; Rapid infiltration systems
1. Introduction If properly designed, filter and infiltration systems have a capacity to almost completely remove pathogenic microorganisms from wastewater [1,2]. In spite of this, in the US septic tank effluent is the most frequently reported source of groundwater contamination [3]. Contaminated groundwater causes almost half of the outbreaks of waterborne disease [4]. To avoid contamination, pathogenic microorganisms must be removed during percolation of the liquid. The removal mechanism is a combination of physical filtration, chemical reactions and biological transformations [5]. The effi*Corresponding author. Tel.: +47-3801-7543. E-mail address: [email protected] (G. Ausland).
ciency of these processes is related to filter temperature [6] and the physical flow conditions which are controlled by the amount, frequency and distribution of effluent applications and particle size of the infiltration media [7,8]. Increasing particle size and loading rates, may result in preferential flow paths through the filter media and reduced retention time [9–11]. This results in reduced adsorption due to inadequate interaction between the percolating wastewater and the porous media [12,13]. Knowledge on effects of filter design factors like dosing rate, distribution method and media particle size is important to reduce the risk of pollution from on-site wastewater treatment systems. Numerous studies of filtration processes have been conducted in laboratory [14–17]. The main objective of this project was to determine effects of filter design
0043-1354/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 0 4 3 - 1 3 5 4 ( 0 2 ) 0 0 0 6 0 - X
G. Ausland et al. / Water Research 36 (2002) 3507–3516
3508
factors on removal of fecal coliforms and fecal streptococci in buried pilot scale filters operated under cold climate conditions.
2. Material and methods ( This study was performed at a research facility in As municipality, 30 km south of Oslo from November 1996 to January 1998. During the experiment, septic tank effluent (STE) was applied to 14 wastewater filters of different designs (Fig. 1). The filters had been in continuous operation for 13 months prior to the beginning of this study.
2.2. Filter media Five different types of filter media were investigated in the experiment. Two types of natural sand, and three types of light weight aggregates (LWA) (Table 2). The sorted and the unsorted sand were selected for the experiment to represent two different classes of filter media, according to Norwegian guidelines for on-site systems [19]. The sorted sand was taken from a glaciofluvial terminal moraine, while the unsorted sand was taken from a till area. The three different types of LWA have the same chemical composition, while their physical properties differed due to various aggregate sizes and different specific surface area (Table 2). Different types of LWA have been tested as filter media during the last years [20–23]. For further information on LWA, see Zhu et al. [24].
2.1. Experimental design Effects of the design factors: filter media grain size, distribution method and hydraulic dosing rates were investigated by monitoring removal of fecal coliforms and fecal streptococci. Setup of the experimental factors are shown in Table 1.
2.3. Geometry of filters All filters were placed into the natural till, and the surface area for each filter was 4.0 1.0 m2 and the depth 120 cm (Fig. 2). Filter effluent was collected in a
filter efflent sampling ports
1m
4 m filter 1
filter 2
filter 3
filter 4
filter 5
filter 6
filter 7
filter 9
filter 10
filter 11
filter 12
filter 13
filter 14
gravity distribution pipe system
pressure distribution pipe system
gravity distribution reservoir
pump tank
septic tank effluent
Fig. 1. Site schematic of the research facilities.
G. Ausland et al. / Water Research 36 (2002) 3507–3516
3509
perforated under-drain pipe placed in a 10 cm drainage layer below the filter media, and transported to the sampling ports by gravity (Fig. 1). The filters were filled by sequential packing of 15 cm layers of filter media. The distribution layer of gravel (25 mm), LWA (20 mm) or in-drain systems were placed on top of the filter media. The in-drain systems contained prefabricated units, where the applied wastewater prior to soil infiltration were filtered through a folded textile cover with a surface area much larger than the infiltrative surface [25]. The cover was 1 mm thick and had a pore size of 0.5 mm. The purpose of the textile cover was to spread the wastewater over the surface of the filter to increase the total retention time through the filter. Perforated inspection pipes were placed on filter surfaces to monitor ponding in the distribution layer (Fig. 2). On top of the
distribution layer, a geotextile separated the filter from an insulation layer of 40 cm of straw. Plastic membranes were placed on top of the insulation to avoid dilution by precipitation and snow melting. These were open on both sides to ensure free air exchange. Temperature sensors were installed 2 cm below the filter surfaces.
Table 1 Setup of experimental design factors
2.5. Gravity dosing
Filter number 1 2 3 4 5 6 7 9 10 11 12 13 14
Filter media
Distribution system
Loading rate (mm/day)
LWA 0–4 mm LWA 2–4 mm Crushed LWA 0–3 mm Sorted sand Sorted sand Unsorted sand Unsorted sand
Pressure dosed Pressure dosed Pressure dosed
80 80 40
Pressure dosed Pressure dosed Pressure dosed Pressure dosed (in-drain) Gravity dosed Gravity dosed Gravity dosed Gravity dosed Gravity dosed Gravity dosed
24 40 40 80
Unsorted sand Unsorted sand Unsorted sand Sorted sand Sorted sand Crushed LWA 0–3 mm
20 20 20 40 24 40
2.4. Pressure dosing For the pressure dosed filters, STE was applied intermittently by time controlled electrical pumps (250 W), connected to pressure pipe distribution systems (Fig. 1). Two parallel, horizontal distribution pipes over each filter were used for dosing (Fig. 2). Diameter of the distribution pipes were 20 mm. Holes (D=4 mm) were drilled for every 50 cm of the pipes, giving 16 application points.
For gravity dosed filters, STE was pumped intermittently to the gravity distribution reservoir (Fig. 1) and drained out to the filters. The STE was applied to the filters from a single point above the distribution layer in each filter. This was done to simulate the distribution that normally appears in gravity dosed systems, due to clogging of distribution holes [26]. The STE was applied in 12 doses per day. The dosing intervals were programmed to simulate daily variations in water consumption. This was done by adding doses every hour in the morning and in the afternoon, and less frequent doses during the rest of the day. 2.6. Sampling schedule STE was supplied to the experiment from a cluster of 12 households. Samples of the STE were taken and analyzed for different pollution parameters during the study (Table 3). All the chemical parameters were
Table 2 d10 ; d50 ; uniformity coefficient (¼ d60 =d10 ), and specific surface area for the filter media investigated in the experiment. dn represent the particle diameter where n weight % of the media is of smaller size
Sorted sand Unsorted sand Spherical LWAe 2–4 mm Spherical LWAe 0–4 mm Crushed LWAe 0–3 mm a
d10 a (mm)
d50 b (mm)
S0 ð¼ d60 =d10 Þ
Specific surface area (m2/g)c
0.24 0.13 2.05 0.80 0.08
0.65 0.84 2.80 1.90 0.70
3.13 9.85 1.46 2.68 15
0.38
10% of the media is smaller than this grain size. 50% of the media is smaller than this grain size. c Measuring method described by Brunauer et al. [18]. d Not measured. e LECA, filtralite. b
d
1.96 3.74 1.17
G. Ausland et al. / Water Research 36 (2002) 3507–3516
3510
controll pipe
40 cm
insulation layer
30 cm
plastic cover distribution pipes pressure systems
distribution pipe gravity systems
distribution layer
mass separation
water proof plastic tarpaulin
temperature sensor 100 cm 150 cm
moraine soil
filter media 120 cm
10 cm
filter under drain pipe
Fig. 2. Cross section of the wastewater filters, showing placement of distribution pipes used for pressure- and gravity-dosing. Table 3 Compositions of septic tank effluent applied for the research facility. The incoming and the out flowing effluent used for enumeration of bacteria and calculation of removal represent the same water mass Parameter
Number of samples
Mean
Std. dev.
pH El. cond. (mS/cm) Tot-P (mg/l) PO4-P (mg/l) Cl (mg/l) Tot-N (mg/l) NO2+NO3-N (mg/l) NH4-N (mg/l) SS (mg/l) CODcr (mg/l) BOD7 (mg/l) Tot. bact. number (log10 n/ml) Fecal coliforms (log n/100 ml) Fecal streptococci (log n/ml)
23 20 13 6 6 8 9 8 8 15 6 10 10 10
7.3 0.9 6.2 4.8 45.2 50.3 0.2 43.4 77 312.4 177.3 6.8 6.3 2.9
0.3 0.1 1.5 1.2 6.1 13.6 0.04 12.5 18.2 97.6 46.5 0.9 0.4 0.4
a
Typical value for STEa
15
40
75 300 140
Average septic tank effluent quality [27].
analyzed by Center for Environmental Research accredited national analysis laboratory; based on Norwegian standard methods [28]. Every 6 weeks, during a period of totally 60 weeks, filter effluent samples were taken for microbial analysis.
2.7. Pour plate method (CFU/ml) The number of heterotrophic bacteria were determined by pour plate method. A 10 ml sample of wastewater was added to 90 ml of distilled water and
G. Ausland et al. / Water Research 36 (2002) 3507–3516
shaken for 1 min and serial dilutions were made. A sample from undiluted wastewater and from each dilution was pippeted into three replicates plates and then melted agar nutrient was added. The petri plates were inverted and incubated at 251C for 4–7 days [29], and colonies were then counted. Enumeration of fecal coliforms by the most probable number (MPN) method. A 10 ml sample of wastewater was added to 90 ml of sterilized water in a 100 ml Pyrex bottle and shaken for 3 min. Serial dilutions were made from these suspensions and samples analyzed using the standard total coliform test [29]. The dilution samples that tested positive, after incubation for 48 h at 371C in the MPN presumptive test, were investigated for the presence of fecal coliforms. These samples were grown in EC-medium for 24 h at 44.51C. 2.8. Fecal streptococci The numbers of fecal streptococci were determined by plating an appropriate dilution of a sample suspension on KF-Streptococci agar (Oxoid) in a pour plate. A sample of undiluted wastewater and one from each dilution was inoculated into three replicate petri plates followed by additional melted agar. The petri plates were inverted and incubated at 351C for 48 h [29]. Red and pink colonies were counted. 2.9. Tracer measurements Potassium bromide was added to the pump tank in September 1997 at concentrations of 90 (720) mg/l for a period of 10 days during normal operation. The variation around the mean concentration of 90 mg/l was due to dose application of bromide solution to the pump tank. This had a minor effect on effluent breakthrough curves, since the frequency of the variations was short compared to the filter hydraulic retention times. Concentrations of bromide in filter effluent were measured by ion selective electrode (Metrohm). Hydraulic retention times were calculated as time for reaching 10% (minimum retention time) and 50% (mean retention time) of initial bromide concentrations in the filter effluent. Tracer measurements were not run for the filters 8 and 12, since operation of these filters were stopped 5 months before the rest of the experiment. Filter 6 was the only filter with ponding of wastewater over the surface (20 cm) at the time of the tracer measurements. The breakthrough curve for this filter is not comparable to the other breakthrough curves, due to dilution of tracer in the distribution layer. 2.10. Statistical analysis Due to the great number of variables, the experiment could not be run as a complete factorial design.
3511
Treatments means were compared by analyses of variance based on log-normal distribution of bacteria numbers. Pooled standard deviations used in the statistical comparisons were calculated by repeated measurements from each filter and not by replicates of equally designed filters. Stevik et al. [10] showed that effluent bacteria concentrations means from parallel columns with equal design were insignificant compared to the variance between repeated samples taken from the same filter column.
3. Results and discussion The concentration of fecal coliforms in the effluent from different filters has a great variation, but in general the removal efficiency was high (Table 4). The log10 removal for fecal coliform bacteria was in a range of 2.9–6.3. Fecal streptococci were detected only in the effluent from LWA 0–4 mm and LWA 2–4 mm (Table 4). For additional water quality data of effluent samples, see [30]. 3.1. Effects of filter media Concentrations of fecal coliforms in the filter effluent were more than three orders of magnitude higher in LWA 0–4 mm (filter 1) and LWA 2–4 mm (filter 2) compared to unsorted sand (filter 7), at a hydraulic dosing rate of 80 mm/day (Fig. 3). Lower removal in LWA filters can probably be explained by different dynamic behavior of the flow through the filters. This assumption is supported by experiments conducted by Stevik et al. [10] who observed that 6.25 mm doses (50 mm/day) applied in coarse LWA filters caused a fast moving wetting front advancing more than 40 cm into the filter media, before gradually transitioning into a steady-state, unsaturated flow. The flow pattern in the top of the filters is important since the rate of removal is highest in the upper part of filters, caused by better oxygen conditions, higher number of active protozoa and smaller pore sizes due to biological clogging. The hydraulic sorptivity of the unsorted sand and ponding over the filter surface, produce a steady-state unsaturated flow throughout the whole filter volume. This results in 49% and 85% longer mean retention times and 4 and 9 times longer minimum retention times, as compared to LWA 0–4 mm and LWA 2–4 mm, respectively (Table 5). The minimum retention has been shown to be a key parameter for removal because the fastest moving fraction of the flow can be expected to contribute to most of the bacteria transport through the columns [10]. Harvey and Garabedin [31] showed that adsorption is a major bacterial immobilization process, and since this process is time dependent [32], enhanced residence time will increase the likelihood for
3512
G. Ausland et al. / Water Research 36 (2002) 3507–3516
Table 4 Mean, standard mean error and average removal of total bacteria number per ml, fecal streptococci per ml and fecal coliforms per 100 ml (all log10 numbers) Filter number 1 LWA 0–4 mm 80 mm/day 2 LWA 2–4 mm 80 mm/day 3 Cr. LWA 0–3 mm 40 mm/day 4 Sorted sand 24 mm/day 5 Sorted sand 40 mm/day 6 Unsorted sand 40 mm/day 7 Unsorted sand 80 mm/day 9 Unsorted sand 20 mm/day 10 Unsorted sand 20 mm/day 11 Unsorted sand 20 mm/day 12 Sorted sand 40 mm/day 13 Sorted sand 24 mm/day 14 Cr LWA 0–3 mm 40 mm/day
Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log.
total fecal fecal total fecal fecal total fecal fecal total fecal fecal total fecal fecal total fecal fecal total fecal fecal total fecal fecal total fecal fecal total fecal fecal total fecal fecal total fecal fecal total fecal fecal
number coliforms streptococci number coliforms streptococci number coliforms streptococci number coliforms streptococci number coliforms streptococci number coliforms streptococci number coliforms streptococci number coliforms streptococci number coliforms streptococci number coliforms streptococci number coliforms streptococci number coliforms streptococci number coliforms streptococci
adsorption. In addition straining can be an important mechanism for immobilization of bacteria in the small pores of the unsorted sand, according to the straining criteria for bacteria in porous media, suggested by Matthess and Pekdeger [33]. Pore sizes in the LWA 0–4 mm and LWA 2–4 mm is larger than bacteria cell sizes (0.5–1.5 mm), and straining cannot be expected to be an important removal mechanism in these media. An exception could be within the clogging zone, that normally develops due to high amount of biofilm and accumulation of solids in the upper part of the filters. In this layer the probability of bacterial straining could be higher due to smaller pore sizes [34].
Mean
S.E.M.
Log. removal
4.8 2.9 0.2 5.0 3.5 0.4 5.0 0.1 0.0 4.8 0.2 0.0 4.9 0.1 0.0 4.4 0.1 0.0 4.5 0.2 0.0 4.7 0.0 0.0 4.7 0.0 0.0 4.9 0.0 0.0 5.1 0.7 0.0 4.9 0.4 0.0 4.9 0.7 0.0
0.2 0.3 2.2 0.2 0.3 2.2 0.1 0.1 0.0 0.2 0.2 0.0 0.2 0.1 0.0 0.2 0.1 0.0 0.2 0.1 0.0 0.2 0.0 0.0 0.2 0.0 0.0 0.2 0.0 0.0 0.2 0.4 0.0 0.2 0.2 0.0 0.2 0.2 0.0
1.9 3.4 2.7 1.8 2.9 2.5 1.8 6.2 2.9 1.9 6.1 2.9 1.8 6.2 2.9 2.3 6.2 2.9 2.3 6.2 2.9 2.1 6.3 2.9 2.1 6.3 2.9 1.9 6.3 2.9 1.6 5.6 2.9 1.9 5.9 2.9 1.8 5.6 2.9
Higher removal of fecal coliforms was observed in the LWA 0–4 mm media as compared to LWA 2–4 mm (p ¼ 0:03) (Fig. 3). The only difference between the LWA 0–4 mm and LWA 2–4 mm was the aggregate size distribution and the specific surface area (Table 2). Tracer measurements showed that the longitudinal dispersion was increased by increasing the grain sizes. Even though the mean retention time was just 25% longer in LWA 0–4 mm, the minimum retention time was twice as long as in LWA 2–4 mm (Table 5), which was consistent to the measured removal efficiency (Fig. 3). Low effluent concentrations were measured for filters dosed at 40 mm/day. No statistically significant
G. Ausland et al. / Water Research 36 (2002) 3507–3516 0.25
4
sorted sand crushed LWA 0-3 mm unsorted sand
LWA 2-4 mm LWA 0-4 mm unsorted sand
3.5
Log fecal coliforms/100 ml
Log fecal coliforms/100 ml
3513
3 2.5 2 1.5 1
0.20
015
0.10
005
0.5 0
0 Fig. 3. Mean log number per 100 ml and standard mean error of fecal coliforms for a loading rate of 80 mm/day (pressure distribution) for filters containing LWA 0–4 mm, LWA 2–4 mm and unsorted sand.
Table 5 Time to reach 10% (minimum retention time) and 50% breakthrough (mean retention time) of bromide Filter number
10% breakthrough (h) (t10 )
50% breakthrough (h) (t50 )
1 LWA 0–4 mm 2 LWA 2–4 mm 3 Crushed LWA 0–3 mm 4 Sorted sand 5 Sorted sand 7 Unsorted sand 9 Unsorted sand 10 Unsorted sand 11 Unsorted sand 13 Sorted sand 14 Crushed LWA
10 5 180 113 84 44 120 220 185 45 48
51 41 235 171 138 76
a
a a a
96 82
Not measured.
differences were observed between sorted sand, unsorted sand and crushed LWA 0–3 mm for dosing rates of 40 mm/day and pressure distribution (Fig. 4). The minimum retention time was o84 h for these filters. These results show that the filter media characteristics are not important as long as the retention time is adequate. 3.2. Effects of distribution method Filter effluent concentrations of fecal coliforms were generally higher for gravity dosed filters than for pressure dosed filters (Table 4). Calculated p-values for comparisons between gravity and pressure distribution were 0.006, 0.015 and 0.300 for crushed LWA 0–3 mm at 40 mm/day, sorted sand at 40 mm/day and sorted sand
Fig. 4. Mean log number per 100 ml and standard mean error of fecal coliforms for loading rate 40 mm/day (pressure distribution) for filters containing sorted sand, unsorted sand and crushed LWA.
at 24 mm/day, respectively. The lower removal for gravity dosed systems was consistent with the shorter retention times in the filters. For crushed LWA 0–3 mm the mean and minimum retention time were increased by factors of 2.9 and 3.8, respectively, by applying uniform pressure distribution compared to gravity distribution at a loading rate of 40 mm/day (filters 3 and 14). In filters of sorted sand, the mean and minimum retention time were increased by factors of 1.8 and 2.5, respectively, for pressure distribution at a loading rate of 24 mm/day. Blombat et al. [35] reported higher concentrations of fecal coliforms from a gravity dosed infiltration systems compared to pressure dosed infiltration systems in a fine sandy loam at a dosing rate of 31 mm/day (p ¼ 0:19). Biological and physical clogging of filter surfaces and capillary forces in fine grained media has been observed to spread the flow towards a uniform vertical flow through the filters [36]. This results in more equal flow conditions for point loaded versus uniformly loaded filters. Therefore less difference due to the distribution method could be expected of treatment performance when a fine-grained filter media is used. For coarser filter media with high hydraulic capacity and lower sorptivity, wastewater normally infiltrates immediately after dose application [37]. Lower retention times in gravity dosed filters (Table 5) show that nonuniform distribution of the flow, as produced by the point application, was maintained throughout the filters. Only a fraction of the filter volume was then used effectively for percolation and treatment of wastewater. In addition, an increase in flow velocity, a thicker water film due to lower average sorption head and larger relation between minimum and mean retention (Table 5) as a result of gravity dosing. All these factors result in lower removal of bacteria. These results show that the effects of the distribution method on removal of fecal coliforms are significant, and are more important at higher dosing rates and in coarser filter media.
3514
G. Ausland et al. / Water Research 36 (2002) 3507–3516
3.3. Effects of hydraulic dosing rate
3.4. Temperature effects Temperatures 2 cm below the filter surfaces varied from about 21C in February/March up to about 171C in August/September. There was no correlation between filter effluent concentrations of fecal coliforms and filter temperatures observed. This is contrary to Gold et al. [6], who found reduced levels of fecal coliform bacteria during warm weather. The fact that temperature measurements were done 2 cm below the filter surfaces, and seasonal variations are greater close to the surface than for the rest of the filter, may partly explain why no significant correlation between removal and measured filter temperatures was observed. However, the results show that low winter temperatures may not necessarily limit the removal of bacteria in buried filters or infiltration systems. 3.5. Retention time and removal The number of fecal coliforms was generally low in most filter samples, with complete removal in 67% of 129 effluent samples, this averaged X5.2 log10 removal. Filter media grain sizes, distribution method and the hydraulic dosing rates were observed to be important for removal of fecal coliforms and fecal streptococci. All these factors are directly related to the wastewater retention times and dispersion in the filters. Low loading rate, fine grained filter media and uniform distribution of wastewater over the filter surface increase the
log fecal coliforms/100 ml coli
Filter effluent concentrations of fecal coliforms were generally low for unsorted sand, at dosing rates of 20 (filters 9–11), 40 (filter 6) and 80 mm/day (filters 7 and 8) (Table 4). Dosing at 20 mm/day resulted in significantly lower effluent concentrations of fecal coliforms than for dosing at 40 mm/day (p ¼ 0:05), while difference between dosing at 40 mm/day and dosing at 80 mm/day was insignificant (p ¼ 0:60). In all these filters, wastewater was ponded over the filter surface most of the time. Intermittent dosing could then be expected to have no effects on the steady-state unsaturated flow through the filters. Even at dosing rate of 80 mm/day the minimum retention time was as long as 44 h. These results differed from experiments with coarse filter media in which Stevik et al. [10] found that increasing the intermittent loading rate from 25 to 50 mm/day, decreased the removal of E. coli by 3.65 and 3.76 log for LWA 0–4 mm and crushed LWA 0–3 mm, respectively. The results show that the effects of dosing rate are less important for a fine grained filter media. This is probably caused by both longer retention time and the higher extent bacteria straining, which is a mechanism less dependent of retention time.
4 3.5 3 2.5 2 1.5 1 0.5 0 0
50
100 150 200 minimum retention time [hours]
250
Fig. 5. Mean fecal coliforms concentrations per 100 ml in each filter plotted against wastewater minimum retention time. Firstorder model was fitted based on the retention times up to 50 h.
retention time and decrease the amount of longitudinal dispersion of wastewater in the filters. These factors enhance the likelihood of adsorption and elimination, which are the main mechanisms for immobilization and removal of fecal coliforms in porous media. Using minimum retention time instead of mean retention time was reported by Stevik et al. [10] to be more relevant for predicting removal based on breakthrough curves in unsaturated filters. Filter effluent concentrations were found to be correlated with the minimum hydraulic retention time for retention times o50 h (r2 ¼ 0:96) (Fig. 5). For minimum retention times longer than 50 h, the removal was almost complete and independent of retention time.
4. Conclusions Removal of fecal coliforms and fecal streptococci in buried infiltration systems is mainly dependent on the media grain size and the hydraulic dosing rate. Higher removal is observed in pressure dosed filters compared to gravity dosed filters. It was not observed any significant correlation between concentration of fecal coliforms and temperatures in a range of 2–171C. Minimum retention time is a key parameter for predicting removal of bacteria in infiltration filters.
Acknowledgements This research was conducted as a part of the national research program Natural Systems for Wastewater Treatment with funding provided by Ministry of The Environment and The Norwegian Research Council. Galina Waag, Department of Biotechnological Sciences, is acknowledged for her work on microbial analysis.
G. Ausland et al. / Water Research 36 (2002) 3507–3516
References [1] Bouma J, Zibell WA, Walker WG, Olcott PG, McCoy E, Hole FD. Soil adsorption of septic tank effluent. A field study of some major soils in Wisconsin. Information Circular No. 20, University of Wisconsin, 1972. [2] Ziebell WA, Nero DH, Deininger JF, McCoy E. Use of bacteria in assessing waste treatment and soil disposal systems. In: Proceedings of the National Home Sewage Disposal Symposium, American Society for Agricultural Engineering Proceedings no. 175, St. Joseph, MI, 9–10 December, 1974. [3] US Environmental Protection Agency. The report to congress, waste disposal practices and their effects on ground water, Washington, DC, 1977. [4] Craun GF. A summary of waterborne illness transmitted through contaminated groundwater. J Environ Health 1985;48:122–7. [5] Goldstein SN, Wenk VD, Fowler MC, Poh SS. A study of selected economic and environmental aspects of individual home wastewater treatment systems. National Technical Information Service, US Department of Commerce, Springfield, VA, 1972. p. 10–29. [6] Gold AJ, Lamb BE, Loomis GW, Boyd JR, Cabelli VJ, McKiel CG. Wastewater renovation in buried filters and recirculating sand filters. J Environ Qual 1992;21:720–5. [7] Yavuz Corapcioglu M, Haridas A. Transport and fate of microorganisms in porous media: a theoretical investigation. J Hydrol 1984;72:149–69. [8] Ziebell WA, Anderson JL, Bouma J, McCoy E. Fecal bacteria: Removal from sewage by soil. Winter meeting, ASAE publication paper no. 75–2579, Chicago, IL, 15–18 December, 1975. [9] Bouma J, Baker FG, Veneman PLM. Measurement of water movement in soil pedons above the watertable. Information Circular no. 27, Wisconsin, Geological and Natural History Survey, 1974. 116pp. [10] Stevik TK, Ausland G, Jenssen PD, Siegrist RL. Removal of E. coli during intermittent filtration of wastewater effluent as affected by dosing rate and media type. Water Res 1999;33:2088–98. [11] Thomas GW, Philips RE. Consequences of water movement in macropores. J Environ Qual 1979;8:149–52. [12] Huysman F, Verstraete W. Water-facilitated transport of bacteria in unsaturated soil columns: influence of inoculation and irrigations methods. Soil Biol Biochem 1993;25:91–7. [13] Smith MS, Thomas GW, White RE, Ritonga D. Transport of Escherichia coli through intact and disturbed soil columns. J Environ Qual 1985;14:87–91. [14] Bellamy WD, Hendricks DW, Logsdon GS. Slow sand filtration: influences of selected process variables. J Am Water Well Assoc 1985;12:62–6. [15] Goldshmid J, Zohar D, Argaman Y, Kott Y. Effects of dissolved salts on the filtration of coliform bacteria in sand dunes. In: Jenkins SH, editor. Advance in water pollution research. New York: Pergamon Press, 1973. p. 147–55. [16] Martin MJ, Logan BE, Johnson WP, Jewett DG, Arnold RG. Scaling bacterial filtration rates in different sized porous media. J Environ Eng 1996;122:407–15.
3515
[17] Sarkar AK, Georgiou G, Sharma MM. Transport of bacteria in porous media. I. An experimental investigation. Biotechnol Bioeng 1994;44:489–97. [18] Brunauer S, Emmet PH, Teller E. Adsorption of gasses in multi-molecular layers. J Am Chem Soc 1938;60: 309–19. [19] Ministry of the Environment, Forskrift om utslipp fra separate avl psanlegg. T-616 ISBN 82-7243-126-2, 1992 [in Norwegian]. [20] Jenssen PD, M!chlum T, Zhu T. Design and performance of subsurface flow constructed wetland in Norway. Proceedings of Constructed Wetlands in Cold Climate, Operation and Performance Symposium, Niagara-on-thelake, Ont. [21] Mæhlum T, Jenssen PD, Warner WS. Cold-climate constructed wetlands. Water Sci Technol 1995;32(3): 95–101. [22] Rasmussen G, Jenssen PD, Westlie L. Graywater treatment options. Environ Res Forum 1996;5–6:215–20. [23] Zhu T. Nutrient pollutants removal in light-weight aggregate (LWA) constructed wetlands and filter systems. Doctor Scientiarum Theses, 1997:16, Department of Agricultural Engineering, Agricultural University of Nor( Norway, 1998. way, As, [24] Zhu T, Jenssen PD, Mæhlum T, Krogstad T. Phosphorus sorption and chemical characteristics of lightweight aggregates (LWA)—Potential filter media in treatment wetlands. Water Sci Technol 1997;35(5):103–8. [25] Laak R. On-site wastewater drain fields using lightweight in-drains. Proceedings of the Second International Conference on Cold Regions Environmental Engineering, Edmonton, Alta., Canada, 23–24 March, 1987. [26] Converse JK. Distribution of domestic waste effluent in soil adsorption beds. Trans ASAE 1974;17:299–309. [27] Canter LW, Knox RC. Septic tank system effects on ground water quality. Chelsea, MI: Lewis Pub., Inc., 336p. [28] Norwegian Standards Association, Catalog of Norwagian standards, ISS N0800-10567, Oslo, Norway, 1998. [29] Clesceri LS, Greenberg AE, Trussell RR. Standard methods for the examination of water and wastewater. Washington, DC: American Public Health Association, 1989. [30] Ausland G, K hler JC, Jenssen PD, Stevik TK. Intermittent filtration of wastewater under cold climate conditions I—hydraulic performance and removal of nutrients and organic matter. In: Hydraulics and purification in wastewater filters. Doctor Scientiarum Theses, Agricultural University of Norway, Department of Agricultural ( Norway, 1998. 23pp. Engineering, As, [31] Harvey HW, Garabedin SP. Use of colloid filtration theory in modeling movement of bacteria through a contaminated sandy aquifer. Environ Sci Technol 1991;25: 178–85. [32] Van Loosdrecht MCM, Lykleuna J, Norde W, Zehuder AJK. Bacterial adhesion: a physiochemical approach. Microbiol Ecol 1989;17:1–15. [33] Matthess G, Pekdeger A. Survival and transport of pathogenic bacteria and virus in groundwater. In: Ward CH, Giger W, McCarty PL, editors. Ground water quality. New York: Wiley, 1985. p. 472–82.
3516
G. Ausland et al. / Water Research 36 (2002) 3507–3516
[34] McDowell-Boyer LM, Hunt JR, Sitar N. Particle transport through porous media. Water Resour Res 1986;22: 1901–21. [35] Blombat C, Wolf DC, Gross EM, Rutledge EM, Gbur EE. Field performance of conventional low pressure distribution septic systems. In: Proceedings of the Seventh National Symposium on Individual and Small Community Sewage Systems. Atlanta, GE, St. Joseph, MI: ASAE Publishers, 11–13 December, 1994. p. 18–94.
[36] Ausland G, Siegrist RL, Jenssen PD, Gudimov A. Unsaturated flow and prediction of purification in wastewater filters and infiltration systems. In: Hydraulics and purification in wastewater filters. Doctor scientiarum theses, Agricultural University of Norway, Department ( Norway, 1998. 23pp. of Agricultural Engineering, As, [37] Stevik TK, Ausland G, Hanssen JF, Jenssen PD. The influence of Physical and chemical factors on the transport of E. coli through biological filters for wastewater purification. Water Res 1999;33:3701–6.
Intermittent filtration of wastewater—removal of fecal coliforms and fecal streptococci G. Auslanda,*, T.K. Stevikb, J.F. Hanssenc, J.C. K hlerd, P.D. Jenssenb a Elkem Research, P.O. Box 8040 Vaagsbygd, 4602 Kristiansand, Norway ( Norway Department of Agricultural Engineering, Agricultural University of Norway, P.O. Box 5065, 1432 As, c ( Norway Department of Biotechnological Sciences, Agricultural University of Norway, P.O. Box 5065, 1432 As, d ( Norway Center for Soil and Environmental Research, 1432 As, b
Received 7 July 2000; received in revised form 14 January 2002; accepted 1 February 2002
Abstract Removal of fecal coliforms and fecal streptococci was monitored over a period of 13 months in 14 buried pilot scale filters, treating septic tank effluent. The effects of grain size, hydraulic dosing rate and distribution method were investigated. Two different natural sands (sorted sand and unsorted sand) and three different types of light weight aggregates (LWA 0–4 mm, LWA 2–4 mm and crushed LWA 0–3 mm) were used. Intermittent dosing rates from 20 to 80 mm/day in 12 doses per day were applied to the filters by uniform pressure distribution or point application by gravity dosing. Removal of fecal coliforms was more than three orders of magnitude higher in the media with the finest grain sizes (unsorted sand) as compared to the coarsest media (LWA 0–4 mm and LWA 2–4 mm) operated under same conditions. Fecal streptococci were determined only in effluent from filters with LWA 0–4 mm and LWA 2–4 mm. Higher removal of fecal coliforms was observed in pressure dosed filters compared to gravity dosed filters. A lower removal was observed by increasing the hydraulic dosing rate. Minimum retention time was found to be a key parameter for predicting removal of bacteria in unsaturated, aerobic filters. At minimum retention times lower than about 50 h, there was a correlation of 0.96 between retention time and removal of fecal coliforms. Retention times longer than 50 h gave almost complete removal of fecal coliforms. r 2002 Elsevier Science Ltd. All rights reserved. Keywords: Biological filters; Wastewater; Fecal coliforms; Fecal streptococci; Rapid infiltration systems
1. Introduction If properly designed, filter and infiltration systems have a capacity to almost completely remove pathogenic microorganisms from wastewater [1,2]. In spite of this, in the US septic tank effluent is the most frequently reported source of groundwater contamination [3]. Contaminated groundwater causes almost half of the outbreaks of waterborne disease [4]. To avoid contamination, pathogenic microorganisms must be removed during percolation of the liquid. The removal mechanism is a combination of physical filtration, chemical reactions and biological transformations [5]. The effi*Corresponding author. Tel.: +47-3801-7543. E-mail address: [email protected] (G. Ausland).
ciency of these processes is related to filter temperature [6] and the physical flow conditions which are controlled by the amount, frequency and distribution of effluent applications and particle size of the infiltration media [7,8]. Increasing particle size and loading rates, may result in preferential flow paths through the filter media and reduced retention time [9–11]. This results in reduced adsorption due to inadequate interaction between the percolating wastewater and the porous media [12,13]. Knowledge on effects of filter design factors like dosing rate, distribution method and media particle size is important to reduce the risk of pollution from on-site wastewater treatment systems. Numerous studies of filtration processes have been conducted in laboratory [14–17]. The main objective of this project was to determine effects of filter design
0043-1354/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 0 4 3 - 1 3 5 4 ( 0 2 ) 0 0 0 6 0 - X
G. Ausland et al. / Water Research 36 (2002) 3507–3516
3508
factors on removal of fecal coliforms and fecal streptococci in buried pilot scale filters operated under cold climate conditions.
2. Material and methods ( This study was performed at a research facility in As municipality, 30 km south of Oslo from November 1996 to January 1998. During the experiment, septic tank effluent (STE) was applied to 14 wastewater filters of different designs (Fig. 1). The filters had been in continuous operation for 13 months prior to the beginning of this study.
2.2. Filter media Five different types of filter media were investigated in the experiment. Two types of natural sand, and three types of light weight aggregates (LWA) (Table 2). The sorted and the unsorted sand were selected for the experiment to represent two different classes of filter media, according to Norwegian guidelines for on-site systems [19]. The sorted sand was taken from a glaciofluvial terminal moraine, while the unsorted sand was taken from a till area. The three different types of LWA have the same chemical composition, while their physical properties differed due to various aggregate sizes and different specific surface area (Table 2). Different types of LWA have been tested as filter media during the last years [20–23]. For further information on LWA, see Zhu et al. [24].
2.1. Experimental design Effects of the design factors: filter media grain size, distribution method and hydraulic dosing rates were investigated by monitoring removal of fecal coliforms and fecal streptococci. Setup of the experimental factors are shown in Table 1.
2.3. Geometry of filters All filters were placed into the natural till, and the surface area for each filter was 4.0 1.0 m2 and the depth 120 cm (Fig. 2). Filter effluent was collected in a
filter efflent sampling ports
1m
4 m filter 1
filter 2
filter 3
filter 4
filter 5
filter 6
filter 7
filter 9
filter 10
filter 11
filter 12
filter 13
filter 14
gravity distribution pipe system
pressure distribution pipe system
gravity distribution reservoir
pump tank
septic tank effluent
Fig. 1. Site schematic of the research facilities.
G. Ausland et al. / Water Research 36 (2002) 3507–3516
3509
perforated under-drain pipe placed in a 10 cm drainage layer below the filter media, and transported to the sampling ports by gravity (Fig. 1). The filters were filled by sequential packing of 15 cm layers of filter media. The distribution layer of gravel (25 mm), LWA (20 mm) or in-drain systems were placed on top of the filter media. The in-drain systems contained prefabricated units, where the applied wastewater prior to soil infiltration were filtered through a folded textile cover with a surface area much larger than the infiltrative surface [25]. The cover was 1 mm thick and had a pore size of 0.5 mm. The purpose of the textile cover was to spread the wastewater over the surface of the filter to increase the total retention time through the filter. Perforated inspection pipes were placed on filter surfaces to monitor ponding in the distribution layer (Fig. 2). On top of the
distribution layer, a geotextile separated the filter from an insulation layer of 40 cm of straw. Plastic membranes were placed on top of the insulation to avoid dilution by precipitation and snow melting. These were open on both sides to ensure free air exchange. Temperature sensors were installed 2 cm below the filter surfaces.
Table 1 Setup of experimental design factors
2.5. Gravity dosing
Filter number 1 2 3 4 5 6 7 9 10 11 12 13 14
Filter media
Distribution system
Loading rate (mm/day)
LWA 0–4 mm LWA 2–4 mm Crushed LWA 0–3 mm Sorted sand Sorted sand Unsorted sand Unsorted sand
Pressure dosed Pressure dosed Pressure dosed
80 80 40
Pressure dosed Pressure dosed Pressure dosed Pressure dosed (in-drain) Gravity dosed Gravity dosed Gravity dosed Gravity dosed Gravity dosed Gravity dosed
24 40 40 80
Unsorted sand Unsorted sand Unsorted sand Sorted sand Sorted sand Crushed LWA 0–3 mm
20 20 20 40 24 40
2.4. Pressure dosing For the pressure dosed filters, STE was applied intermittently by time controlled electrical pumps (250 W), connected to pressure pipe distribution systems (Fig. 1). Two parallel, horizontal distribution pipes over each filter were used for dosing (Fig. 2). Diameter of the distribution pipes were 20 mm. Holes (D=4 mm) were drilled for every 50 cm of the pipes, giving 16 application points.
For gravity dosed filters, STE was pumped intermittently to the gravity distribution reservoir (Fig. 1) and drained out to the filters. The STE was applied to the filters from a single point above the distribution layer in each filter. This was done to simulate the distribution that normally appears in gravity dosed systems, due to clogging of distribution holes [26]. The STE was applied in 12 doses per day. The dosing intervals were programmed to simulate daily variations in water consumption. This was done by adding doses every hour in the morning and in the afternoon, and less frequent doses during the rest of the day. 2.6. Sampling schedule STE was supplied to the experiment from a cluster of 12 households. Samples of the STE were taken and analyzed for different pollution parameters during the study (Table 3). All the chemical parameters were
Table 2 d10 ; d50 ; uniformity coefficient (¼ d60 =d10 ), and specific surface area for the filter media investigated in the experiment. dn represent the particle diameter where n weight % of the media is of smaller size
Sorted sand Unsorted sand Spherical LWAe 2–4 mm Spherical LWAe 0–4 mm Crushed LWAe 0–3 mm a
d10 a (mm)
d50 b (mm)
S0 ð¼ d60 =d10 Þ
Specific surface area (m2/g)c
0.24 0.13 2.05 0.80 0.08
0.65 0.84 2.80 1.90 0.70
3.13 9.85 1.46 2.68 15
0.38
10% of the media is smaller than this grain size. 50% of the media is smaller than this grain size. c Measuring method described by Brunauer et al. [18]. d Not measured. e LECA, filtralite. b
d
1.96 3.74 1.17
G. Ausland et al. / Water Research 36 (2002) 3507–3516
3510
controll pipe
40 cm
insulation layer
30 cm
plastic cover distribution pipes pressure systems
distribution pipe gravity systems
distribution layer
mass separation
water proof plastic tarpaulin
temperature sensor 100 cm 150 cm
moraine soil
filter media 120 cm
10 cm
filter under drain pipe
Fig. 2. Cross section of the wastewater filters, showing placement of distribution pipes used for pressure- and gravity-dosing. Table 3 Compositions of septic tank effluent applied for the research facility. The incoming and the out flowing effluent used for enumeration of bacteria and calculation of removal represent the same water mass Parameter
Number of samples
Mean
Std. dev.
pH El. cond. (mS/cm) Tot-P (mg/l) PO4-P (mg/l) Cl (mg/l) Tot-N (mg/l) NO2+NO3-N (mg/l) NH4-N (mg/l) SS (mg/l) CODcr (mg/l) BOD7 (mg/l) Tot. bact. number (log10 n/ml) Fecal coliforms (log n/100 ml) Fecal streptococci (log n/ml)
23 20 13 6 6 8 9 8 8 15 6 10 10 10
7.3 0.9 6.2 4.8 45.2 50.3 0.2 43.4 77 312.4 177.3 6.8 6.3 2.9
0.3 0.1 1.5 1.2 6.1 13.6 0.04 12.5 18.2 97.6 46.5 0.9 0.4 0.4
a
Typical value for STEa
15
40
75 300 140
Average septic tank effluent quality [27].
analyzed by Center for Environmental Research accredited national analysis laboratory; based on Norwegian standard methods [28]. Every 6 weeks, during a period of totally 60 weeks, filter effluent samples were taken for microbial analysis.
2.7. Pour plate method (CFU/ml) The number of heterotrophic bacteria were determined by pour plate method. A 10 ml sample of wastewater was added to 90 ml of distilled water and
G. Ausland et al. / Water Research 36 (2002) 3507–3516
shaken for 1 min and serial dilutions were made. A sample from undiluted wastewater and from each dilution was pippeted into three replicates plates and then melted agar nutrient was added. The petri plates were inverted and incubated at 251C for 4–7 days [29], and colonies were then counted. Enumeration of fecal coliforms by the most probable number (MPN) method. A 10 ml sample of wastewater was added to 90 ml of sterilized water in a 100 ml Pyrex bottle and shaken for 3 min. Serial dilutions were made from these suspensions and samples analyzed using the standard total coliform test [29]. The dilution samples that tested positive, after incubation for 48 h at 371C in the MPN presumptive test, were investigated for the presence of fecal coliforms. These samples were grown in EC-medium for 24 h at 44.51C. 2.8. Fecal streptococci The numbers of fecal streptococci were determined by plating an appropriate dilution of a sample suspension on KF-Streptococci agar (Oxoid) in a pour plate. A sample of undiluted wastewater and one from each dilution was inoculated into three replicate petri plates followed by additional melted agar. The petri plates were inverted and incubated at 351C for 48 h [29]. Red and pink colonies were counted. 2.9. Tracer measurements Potassium bromide was added to the pump tank in September 1997 at concentrations of 90 (720) mg/l for a period of 10 days during normal operation. The variation around the mean concentration of 90 mg/l was due to dose application of bromide solution to the pump tank. This had a minor effect on effluent breakthrough curves, since the frequency of the variations was short compared to the filter hydraulic retention times. Concentrations of bromide in filter effluent were measured by ion selective electrode (Metrohm). Hydraulic retention times were calculated as time for reaching 10% (minimum retention time) and 50% (mean retention time) of initial bromide concentrations in the filter effluent. Tracer measurements were not run for the filters 8 and 12, since operation of these filters were stopped 5 months before the rest of the experiment. Filter 6 was the only filter with ponding of wastewater over the surface (20 cm) at the time of the tracer measurements. The breakthrough curve for this filter is not comparable to the other breakthrough curves, due to dilution of tracer in the distribution layer. 2.10. Statistical analysis Due to the great number of variables, the experiment could not be run as a complete factorial design.
3511
Treatments means were compared by analyses of variance based on log-normal distribution of bacteria numbers. Pooled standard deviations used in the statistical comparisons were calculated by repeated measurements from each filter and not by replicates of equally designed filters. Stevik et al. [10] showed that effluent bacteria concentrations means from parallel columns with equal design were insignificant compared to the variance between repeated samples taken from the same filter column.
3. Results and discussion The concentration of fecal coliforms in the effluent from different filters has a great variation, but in general the removal efficiency was high (Table 4). The log10 removal for fecal coliform bacteria was in a range of 2.9–6.3. Fecal streptococci were detected only in the effluent from LWA 0–4 mm and LWA 2–4 mm (Table 4). For additional water quality data of effluent samples, see [30]. 3.1. Effects of filter media Concentrations of fecal coliforms in the filter effluent were more than three orders of magnitude higher in LWA 0–4 mm (filter 1) and LWA 2–4 mm (filter 2) compared to unsorted sand (filter 7), at a hydraulic dosing rate of 80 mm/day (Fig. 3). Lower removal in LWA filters can probably be explained by different dynamic behavior of the flow through the filters. This assumption is supported by experiments conducted by Stevik et al. [10] who observed that 6.25 mm doses (50 mm/day) applied in coarse LWA filters caused a fast moving wetting front advancing more than 40 cm into the filter media, before gradually transitioning into a steady-state, unsaturated flow. The flow pattern in the top of the filters is important since the rate of removal is highest in the upper part of filters, caused by better oxygen conditions, higher number of active protozoa and smaller pore sizes due to biological clogging. The hydraulic sorptivity of the unsorted sand and ponding over the filter surface, produce a steady-state unsaturated flow throughout the whole filter volume. This results in 49% and 85% longer mean retention times and 4 and 9 times longer minimum retention times, as compared to LWA 0–4 mm and LWA 2–4 mm, respectively (Table 5). The minimum retention has been shown to be a key parameter for removal because the fastest moving fraction of the flow can be expected to contribute to most of the bacteria transport through the columns [10]. Harvey and Garabedin [31] showed that adsorption is a major bacterial immobilization process, and since this process is time dependent [32], enhanced residence time will increase the likelihood for
3512
G. Ausland et al. / Water Research 36 (2002) 3507–3516
Table 4 Mean, standard mean error and average removal of total bacteria number per ml, fecal streptococci per ml and fecal coliforms per 100 ml (all log10 numbers) Filter number 1 LWA 0–4 mm 80 mm/day 2 LWA 2–4 mm 80 mm/day 3 Cr. LWA 0–3 mm 40 mm/day 4 Sorted sand 24 mm/day 5 Sorted sand 40 mm/day 6 Unsorted sand 40 mm/day 7 Unsorted sand 80 mm/day 9 Unsorted sand 20 mm/day 10 Unsorted sand 20 mm/day 11 Unsorted sand 20 mm/day 12 Sorted sand 40 mm/day 13 Sorted sand 24 mm/day 14 Cr LWA 0–3 mm 40 mm/day
Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log. Log.
total fecal fecal total fecal fecal total fecal fecal total fecal fecal total fecal fecal total fecal fecal total fecal fecal total fecal fecal total fecal fecal total fecal fecal total fecal fecal total fecal fecal total fecal fecal
number coliforms streptococci number coliforms streptococci number coliforms streptococci number coliforms streptococci number coliforms streptococci number coliforms streptococci number coliforms streptococci number coliforms streptococci number coliforms streptococci number coliforms streptococci number coliforms streptococci number coliforms streptococci number coliforms streptococci
adsorption. In addition straining can be an important mechanism for immobilization of bacteria in the small pores of the unsorted sand, according to the straining criteria for bacteria in porous media, suggested by Matthess and Pekdeger [33]. Pore sizes in the LWA 0–4 mm and LWA 2–4 mm is larger than bacteria cell sizes (0.5–1.5 mm), and straining cannot be expected to be an important removal mechanism in these media. An exception could be within the clogging zone, that normally develops due to high amount of biofilm and accumulation of solids in the upper part of the filters. In this layer the probability of bacterial straining could be higher due to smaller pore sizes [34].
Mean
S.E.M.
Log. removal
4.8 2.9 0.2 5.0 3.5 0.4 5.0 0.1 0.0 4.8 0.2 0.0 4.9 0.1 0.0 4.4 0.1 0.0 4.5 0.2 0.0 4.7 0.0 0.0 4.7 0.0 0.0 4.9 0.0 0.0 5.1 0.7 0.0 4.9 0.4 0.0 4.9 0.7 0.0
0.2 0.3 2.2 0.2 0.3 2.2 0.1 0.1 0.0 0.2 0.2 0.0 0.2 0.1 0.0 0.2 0.1 0.0 0.2 0.1 0.0 0.2 0.0 0.0 0.2 0.0 0.0 0.2 0.0 0.0 0.2 0.4 0.0 0.2 0.2 0.0 0.2 0.2 0.0
1.9 3.4 2.7 1.8 2.9 2.5 1.8 6.2 2.9 1.9 6.1 2.9 1.8 6.2 2.9 2.3 6.2 2.9 2.3 6.2 2.9 2.1 6.3 2.9 2.1 6.3 2.9 1.9 6.3 2.9 1.6 5.6 2.9 1.9 5.9 2.9 1.8 5.6 2.9
Higher removal of fecal coliforms was observed in the LWA 0–4 mm media as compared to LWA 2–4 mm (p ¼ 0:03) (Fig. 3). The only difference between the LWA 0–4 mm and LWA 2–4 mm was the aggregate size distribution and the specific surface area (Table 2). Tracer measurements showed that the longitudinal dispersion was increased by increasing the grain sizes. Even though the mean retention time was just 25% longer in LWA 0–4 mm, the minimum retention time was twice as long as in LWA 2–4 mm (Table 5), which was consistent to the measured removal efficiency (Fig. 3). Low effluent concentrations were measured for filters dosed at 40 mm/day. No statistically significant
G. Ausland et al. / Water Research 36 (2002) 3507–3516 0.25
4
sorted sand crushed LWA 0-3 mm unsorted sand
LWA 2-4 mm LWA 0-4 mm unsorted sand
3.5
Log fecal coliforms/100 ml
Log fecal coliforms/100 ml
3513
3 2.5 2 1.5 1
0.20
015
0.10
005
0.5 0
0 Fig. 3. Mean log number per 100 ml and standard mean error of fecal coliforms for a loading rate of 80 mm/day (pressure distribution) for filters containing LWA 0–4 mm, LWA 2–4 mm and unsorted sand.
Table 5 Time to reach 10% (minimum retention time) and 50% breakthrough (mean retention time) of bromide Filter number
10% breakthrough (h) (t10 )
50% breakthrough (h) (t50 )
1 LWA 0–4 mm 2 LWA 2–4 mm 3 Crushed LWA 0–3 mm 4 Sorted sand 5 Sorted sand 7 Unsorted sand 9 Unsorted sand 10 Unsorted sand 11 Unsorted sand 13 Sorted sand 14 Crushed LWA
10 5 180 113 84 44 120 220 185 45 48
51 41 235 171 138 76
a
a a a
96 82
Not measured.
differences were observed between sorted sand, unsorted sand and crushed LWA 0–3 mm for dosing rates of 40 mm/day and pressure distribution (Fig. 4). The minimum retention time was o84 h for these filters. These results show that the filter media characteristics are not important as long as the retention time is adequate. 3.2. Effects of distribution method Filter effluent concentrations of fecal coliforms were generally higher for gravity dosed filters than for pressure dosed filters (Table 4). Calculated p-values for comparisons between gravity and pressure distribution were 0.006, 0.015 and 0.300 for crushed LWA 0–3 mm at 40 mm/day, sorted sand at 40 mm/day and sorted sand
Fig. 4. Mean log number per 100 ml and standard mean error of fecal coliforms for loading rate 40 mm/day (pressure distribution) for filters containing sorted sand, unsorted sand and crushed LWA.
at 24 mm/day, respectively. The lower removal for gravity dosed systems was consistent with the shorter retention times in the filters. For crushed LWA 0–3 mm the mean and minimum retention time were increased by factors of 2.9 and 3.8, respectively, by applying uniform pressure distribution compared to gravity distribution at a loading rate of 40 mm/day (filters 3 and 14). In filters of sorted sand, the mean and minimum retention time were increased by factors of 1.8 and 2.5, respectively, for pressure distribution at a loading rate of 24 mm/day. Blombat et al. [35] reported higher concentrations of fecal coliforms from a gravity dosed infiltration systems compared to pressure dosed infiltration systems in a fine sandy loam at a dosing rate of 31 mm/day (p ¼ 0:19). Biological and physical clogging of filter surfaces and capillary forces in fine grained media has been observed to spread the flow towards a uniform vertical flow through the filters [36]. This results in more equal flow conditions for point loaded versus uniformly loaded filters. Therefore less difference due to the distribution method could be expected of treatment performance when a fine-grained filter media is used. For coarser filter media with high hydraulic capacity and lower sorptivity, wastewater normally infiltrates immediately after dose application [37]. Lower retention times in gravity dosed filters (Table 5) show that nonuniform distribution of the flow, as produced by the point application, was maintained throughout the filters. Only a fraction of the filter volume was then used effectively for percolation and treatment of wastewater. In addition, an increase in flow velocity, a thicker water film due to lower average sorption head and larger relation between minimum and mean retention (Table 5) as a result of gravity dosing. All these factors result in lower removal of bacteria. These results show that the effects of the distribution method on removal of fecal coliforms are significant, and are more important at higher dosing rates and in coarser filter media.
3514
G. Ausland et al. / Water Research 36 (2002) 3507–3516
3.3. Effects of hydraulic dosing rate
3.4. Temperature effects Temperatures 2 cm below the filter surfaces varied from about 21C in February/March up to about 171C in August/September. There was no correlation between filter effluent concentrations of fecal coliforms and filter temperatures observed. This is contrary to Gold et al. [6], who found reduced levels of fecal coliform bacteria during warm weather. The fact that temperature measurements were done 2 cm below the filter surfaces, and seasonal variations are greater close to the surface than for the rest of the filter, may partly explain why no significant correlation between removal and measured filter temperatures was observed. However, the results show that low winter temperatures may not necessarily limit the removal of bacteria in buried filters or infiltration systems. 3.5. Retention time and removal The number of fecal coliforms was generally low in most filter samples, with complete removal in 67% of 129 effluent samples, this averaged X5.2 log10 removal. Filter media grain sizes, distribution method and the hydraulic dosing rates were observed to be important for removal of fecal coliforms and fecal streptococci. All these factors are directly related to the wastewater retention times and dispersion in the filters. Low loading rate, fine grained filter media and uniform distribution of wastewater over the filter surface increase the
log fecal coliforms/100 ml coli
Filter effluent concentrations of fecal coliforms were generally low for unsorted sand, at dosing rates of 20 (filters 9–11), 40 (filter 6) and 80 mm/day (filters 7 and 8) (Table 4). Dosing at 20 mm/day resulted in significantly lower effluent concentrations of fecal coliforms than for dosing at 40 mm/day (p ¼ 0:05), while difference between dosing at 40 mm/day and dosing at 80 mm/day was insignificant (p ¼ 0:60). In all these filters, wastewater was ponded over the filter surface most of the time. Intermittent dosing could then be expected to have no effects on the steady-state unsaturated flow through the filters. Even at dosing rate of 80 mm/day the minimum retention time was as long as 44 h. These results differed from experiments with coarse filter media in which Stevik et al. [10] found that increasing the intermittent loading rate from 25 to 50 mm/day, decreased the removal of E. coli by 3.65 and 3.76 log for LWA 0–4 mm and crushed LWA 0–3 mm, respectively. The results show that the effects of dosing rate are less important for a fine grained filter media. This is probably caused by both longer retention time and the higher extent bacteria straining, which is a mechanism less dependent of retention time.
4 3.5 3 2.5 2 1.5 1 0.5 0 0
50
100 150 200 minimum retention time [hours]
250
Fig. 5. Mean fecal coliforms concentrations per 100 ml in each filter plotted against wastewater minimum retention time. Firstorder model was fitted based on the retention times up to 50 h.
retention time and decrease the amount of longitudinal dispersion of wastewater in the filters. These factors enhance the likelihood of adsorption and elimination, which are the main mechanisms for immobilization and removal of fecal coliforms in porous media. Using minimum retention time instead of mean retention time was reported by Stevik et al. [10] to be more relevant for predicting removal based on breakthrough curves in unsaturated filters. Filter effluent concentrations were found to be correlated with the minimum hydraulic retention time for retention times o50 h (r2 ¼ 0:96) (Fig. 5). For minimum retention times longer than 50 h, the removal was almost complete and independent of retention time.
4. Conclusions Removal of fecal coliforms and fecal streptococci in buried infiltration systems is mainly dependent on the media grain size and the hydraulic dosing rate. Higher removal is observed in pressure dosed filters compared to gravity dosed filters. It was not observed any significant correlation between concentration of fecal coliforms and temperatures in a range of 2–171C. Minimum retention time is a key parameter for predicting removal of bacteria in infiltration filters.
Acknowledgements This research was conducted as a part of the national research program Natural Systems for Wastewater Treatment with funding provided by Ministry of The Environment and The Norwegian Research Council. Galina Waag, Department of Biotechnological Sciences, is acknowledged for her work on microbial analysis.
G. Ausland et al. / Water Research 36 (2002) 3507–3516
References [1] Bouma J, Zibell WA, Walker WG, Olcott PG, McCoy E, Hole FD. Soil adsorption of septic tank effluent. A field study of some major soils in Wisconsin. Information Circular No. 20, University of Wisconsin, 1972. [2] Ziebell WA, Nero DH, Deininger JF, McCoy E. Use of bacteria in assessing waste treatment and soil disposal systems. In: Proceedings of the National Home Sewage Disposal Symposium, American Society for Agricultural Engineering Proceedings no. 175, St. Joseph, MI, 9–10 December, 1974. [3] US Environmental Protection Agency. The report to congress, waste disposal practices and their effects on ground water, Washington, DC, 1977. [4] Craun GF. A summary of waterborne illness transmitted through contaminated groundwater. J Environ Health 1985;48:122–7. [5] Goldstein SN, Wenk VD, Fowler MC, Poh SS. A study of selected economic and environmental aspects of individual home wastewater treatment systems. National Technical Information Service, US Department of Commerce, Springfield, VA, 1972. p. 10–29. [6] Gold AJ, Lamb BE, Loomis GW, Boyd JR, Cabelli VJ, McKiel CG. Wastewater renovation in buried filters and recirculating sand filters. J Environ Qual 1992;21:720–5. [7] Yavuz Corapcioglu M, Haridas A. Transport and fate of microorganisms in porous media: a theoretical investigation. J Hydrol 1984;72:149–69. [8] Ziebell WA, Anderson JL, Bouma J, McCoy E. Fecal bacteria: Removal from sewage by soil. Winter meeting, ASAE publication paper no. 75–2579, Chicago, IL, 15–18 December, 1975. [9] Bouma J, Baker FG, Veneman PLM. Measurement of water movement in soil pedons above the watertable. Information Circular no. 27, Wisconsin, Geological and Natural History Survey, 1974. 116pp. [10] Stevik TK, Ausland G, Jenssen PD, Siegrist RL. Removal of E. coli during intermittent filtration of wastewater effluent as affected by dosing rate and media type. Water Res 1999;33:2088–98. [11] Thomas GW, Philips RE. Consequences of water movement in macropores. J Environ Qual 1979;8:149–52. [12] Huysman F, Verstraete W. Water-facilitated transport of bacteria in unsaturated soil columns: influence of inoculation and irrigations methods. Soil Biol Biochem 1993;25:91–7. [13] Smith MS, Thomas GW, White RE, Ritonga D. Transport of Escherichia coli through intact and disturbed soil columns. J Environ Qual 1985;14:87–91. [14] Bellamy WD, Hendricks DW, Logsdon GS. Slow sand filtration: influences of selected process variables. J Am Water Well Assoc 1985;12:62–6. [15] Goldshmid J, Zohar D, Argaman Y, Kott Y. Effects of dissolved salts on the filtration of coliform bacteria in sand dunes. In: Jenkins SH, editor. Advance in water pollution research. New York: Pergamon Press, 1973. p. 147–55. [16] Martin MJ, Logan BE, Johnson WP, Jewett DG, Arnold RG. Scaling bacterial filtration rates in different sized porous media. J Environ Eng 1996;122:407–15.
3515
[17] Sarkar AK, Georgiou G, Sharma MM. Transport of bacteria in porous media. I. An experimental investigation. Biotechnol Bioeng 1994;44:489–97. [18] Brunauer S, Emmet PH, Teller E. Adsorption of gasses in multi-molecular layers. J Am Chem Soc 1938;60: 309–19. [19] Ministry of the Environment, Forskrift om utslipp fra separate avl psanlegg. T-616 ISBN 82-7243-126-2, 1992 [in Norwegian]. [20] Jenssen PD, M!chlum T, Zhu T. Design and performance of subsurface flow constructed wetland in Norway. Proceedings of Constructed Wetlands in Cold Climate, Operation and Performance Symposium, Niagara-on-thelake, Ont. [21] Mæhlum T, Jenssen PD, Warner WS. Cold-climate constructed wetlands. Water Sci Technol 1995;32(3): 95–101. [22] Rasmussen G, Jenssen PD, Westlie L. Graywater treatment options. Environ Res Forum 1996;5–6:215–20. [23] Zhu T. Nutrient pollutants removal in light-weight aggregate (LWA) constructed wetlands and filter systems. Doctor Scientiarum Theses, 1997:16, Department of Agricultural Engineering, Agricultural University of Nor( Norway, 1998. way, As, [24] Zhu T, Jenssen PD, Mæhlum T, Krogstad T. Phosphorus sorption and chemical characteristics of lightweight aggregates (LWA)—Potential filter media in treatment wetlands. Water Sci Technol 1997;35(5):103–8. [25] Laak R. On-site wastewater drain fields using lightweight in-drains. Proceedings of the Second International Conference on Cold Regions Environmental Engineering, Edmonton, Alta., Canada, 23–24 March, 1987. [26] Converse JK. Distribution of domestic waste effluent in soil adsorption beds. Trans ASAE 1974;17:299–309. [27] Canter LW, Knox RC. Septic tank system effects on ground water quality. Chelsea, MI: Lewis Pub., Inc., 336p. [28] Norwegian Standards Association, Catalog of Norwagian standards, ISS N0800-10567, Oslo, Norway, 1998. [29] Clesceri LS, Greenberg AE, Trussell RR. Standard methods for the examination of water and wastewater. Washington, DC: American Public Health Association, 1989. [30] Ausland G, K hler JC, Jenssen PD, Stevik TK. Intermittent filtration of wastewater under cold climate conditions I—hydraulic performance and removal of nutrients and organic matter. In: Hydraulics and purification in wastewater filters. Doctor Scientiarum Theses, Agricultural University of Norway, Department of Agricultural ( Norway, 1998. 23pp. Engineering, As, [31] Harvey HW, Garabedin SP. Use of colloid filtration theory in modeling movement of bacteria through a contaminated sandy aquifer. Environ Sci Technol 1991;25: 178–85. [32] Van Loosdrecht MCM, Lykleuna J, Norde W, Zehuder AJK. Bacterial adhesion: a physiochemical approach. Microbiol Ecol 1989;17:1–15. [33] Matthess G, Pekdeger A. Survival and transport of pathogenic bacteria and virus in groundwater. In: Ward CH, Giger W, McCarty PL, editors. Ground water quality. New York: Wiley, 1985. p. 472–82.
3516
G. Ausland et al. / Water Research 36 (2002) 3507–3516
[34] McDowell-Boyer LM, Hunt JR, Sitar N. Particle transport through porous media. Water Resour Res 1986;22: 1901–21. [35] Blombat C, Wolf DC, Gross EM, Rutledge EM, Gbur EE. Field performance of conventional low pressure distribution septic systems. In: Proceedings of the Seventh National Symposium on Individual and Small Community Sewage Systems. Atlanta, GE, St. Joseph, MI: ASAE Publishers, 11–13 December, 1994. p. 18–94.
[36] Ausland G, Siegrist RL, Jenssen PD, Gudimov A. Unsaturated flow and prediction of purification in wastewater filters and infiltration systems. In: Hydraulics and purification in wastewater filters. Doctor scientiarum theses, Agricultural University of Norway, Department ( Norway, 1998. 23pp. of Agricultural Engineering, As, [37] Stevik TK, Ausland G, Hanssen JF, Jenssen PD. The influence of Physical and chemical factors on the transport of E. coli through biological filters for wastewater purification. Water Res 1999;33:3701–6.