Treatment performance of small-scale vermifilter for domestic wastewater and its relationship to earthworm growth , reproduction and enzymatic activity

A vermifilter system packed with quartz sands and ceramsite was studied for domestic wastewater treatment. Results showed that good performance of vermifilter was achieved and removal rates were COD (47.3 – 64.7%), BOD5 (54.78 – 66.36%), SS (57.18 – 77.90%), TN (7.63 – 14.90%), and NH4-N (21.01 – 62.31%), respectively. An increase in hydraulic loading led to a decrease in treatment efficiency and adult earthworm abundance. In addition, activities of protease, alkaline phosphatase (ALP), and cellulase in earthworm body dropped, but superoxide dismutase (SOD) and catalase (CAT) increased with the hydraulic loading. Correlation analysis implied that larger earthworm (>0.3 g) abundance might play more positive role on wastewater treatment in vermifilter, compared to smaller worm. Earthworm enzymatic activities had significant correlation with treatment efficiency of COD and BOD5 by vermifilter. Thus an important relationship exists for earthworm population dynamics and enzymatic activities with COD and BOD5 removal rates of domestic wastewater by vermifilter.


INTRODUCTION
Vermifilter (Lumbrifiltration) was first advocated by the late Professor Jose Toha at the University of Chile in 1992 (Bouché and Qiu, 1998;Aguilera, 2003;Li et al., 2008), which was a low-cost sustainable technology over conventional systems with immense potential for decentralization in rural areas (Taylor et al., 2003;Sinha et al., 2008).It was firstly used to process organically polluted water using earthworms (Li et al., 2008).Introduction of earthworm was a considerable innovation to conventional biofilter of wastewater treatment, and it had created a new method of biological reaction through extending food chains, conversing energy and trans-*Corresponding author.E-mail: lixiaowei419@163.com.
ferring mass from the biofilm to the earthworm.
Vermifilter had been found to be generally good for swine wastewater treatment (Li et al., 2008), municipal wastewater treatment (Godefroid and Yang, 2005;Xing et al., 2005;Yang and Zhao, 2008;Yang et al., 2008), and domestic wastewater treatment (Taylor et al., 2003;Sinha et al., 2008).However, these studies on vermifilter focused on treatment efficiency of wastewater, and few on earthworm population dynamics and enzymatic activity in the vermifiltration wastewater treatment process.
Wastewater had likely led to an influence on earthworm population dynamics and enzymatic activity, because it contained a complex mixture of contaminants, including nutrients, pathogens and toxic compounds (e.g.endocrine disrupting compounds, Hughes et al., 2007).Hughes et al. (2007Hughes et al. ( , 2008Hughes et al. ( , 2009) ) had investigated the risk of pH, ammonia/ammonium, and sodium accumulation to earthworm in vermifiltration wastewater treatment.Further, the kinetic model of conventional biofilter was based mostly on organic matter degradation of biofilm (Dorado et al., 2008), but vermifilter has an important additional decomposition feature involving earthworms.Thus, research on earthworm population dynamics and enzymatic activity would supply vital data for establishing kinetic model of organic matter degradation in vermifiltration wastewater process.
The aims of the present study are: (1) to evaluate treatment efficiency, earthworm population dynamic and enzymatic activity in different hydraulic loadings of vermifiltration wastewater process, and (2) to analyze correlation between treatment efficiency and earthworm characteristics.

Vermifilter system
A pilot-scale vermifilter was set up for treating domestic wastewater in a wastewater plant of Shanghai city, China, according to our 6year studying experience.Figure 1 showed the schematic diagram of vermifilter, with the parameters of vermifilter design outlined in Table 1.Ceramsite vermifilter (CV) was included in the quartz-sand vermifilter (QV) (Figure 1).The influent water was distributed by turning spurt water device.A layer of plastic fibrous filler covered the surface of filter bed.The fibrous filler was used for redistribution of wastewater and was an excellent opaque property for earthworm.

Experimental design
The vermifilter began operation February 28th 2006.Initially, the filter system had undergone a biofilm culturing stage of 40 days.The water load was 4.8 m 3 .m -2 .d -1 in the phase.After 40 days, earthworm (Eisenia foetida) was added evenly in the first filter bed with an initial density of ca.21 000 ind•m -2 , and the total earthworm biomass was ca.30.3 kg in the whole vermifilter.E. foetida was chosen because it was widely used in vermifiltration (Taylor et al., 2003) and had been shown to process organic wastes with the greatest efficiency (Edwards and Bater, 1992).After earthworm inoculation period of 20 days, vermifilter system was taken up to the steady operation.
During May 1st to August 31st 2006, four working conditions (W1, W2, W3, and W4) were used to treat the wastewater from the Quyang wastewater plant in Shanghai, China.Each working condition was operated for 30 -31days.Hydraulic loading of the four working conditions in m 3 •m -2 •d -1 was W1 = 2.4, W2 = 4.8, W3 = 6.0, and W4 = 6.7.The characteristics of influent water in the four working conditions were outlined in Table 2.

Water sampling and analysis
Influent and effluent samples were collected weekly for chemical oxygen demand (COD), five-day carbonaceous biochemical oxygen demand (BOD5), suspended solid (SS), total nitrogen (TN) and ammonium (NH4-N) analysis.COD was measured by a COD analyzer (NOVA 60, Merck, Germany).BOD5 was measured using a WTW oxitop IS 12 BOD analyzer.SS, TN, and NH4-N were analyzed according to the American Public Health Authority (1995).All samples were analyzed in triplicate and the results were averaged during a working condition.

Earthworm sampling and analysis
Earthworms were sampled monthly from the first filter bed of vermifilter.Four sampling points were set up evenly in the filter, and 100 ml of sampling was taken in every point for analyzing earthworm numbers (adults, hatchlings and cocoons) and clitellated development.Earthworms and cocoons were separated from the samples by hand sorting, after which they were counted, examined for clitellated development and weighed after washing with water and drying them by paper towels (Garg et al., 2005).The worms were weighed without voiding their gut content.Corrections for gut content were not applied to any data in this study.The results from four sampling were averaged.
Meanwhile, earthworm enzymatic activities were analyzed monthly.Protein activity was determined according to the method of Lowry et al. (Bradford, 1976).Alkaline phophatase (ALP) activity was measured as described by Li Sui-yan (Li and Li, 2004).Cellulase activity was assayed according to the method of Zhang Dean (Zhang et al., 1991).Superoxide dismutase (SOD) activity was estimated by the pyorgallol auto-oxidation method (325 nm) (Yu et al., 2005).Catalase (CAT) activity was assayed according to the method of Saint-Denis (Saint-Denis et al., 1998).All the samples were analyzed in triplicate and the results were averaged.

Correlation analysis
The data were analyzed using SPSS 13.0 and Origin 7.5.Correlation analysis was used to find out the relationship between treatment efficiencies and earthworm characteristics.

Earthworm population dynamics
Figure 4 showed the weight distribution of earthworm in two kinds of vermifilter.Proportion of less than 0.2 g earthworm increased gradually with the operation time.The less the earthworm's weight was, the higher the   Proportion was.The proportion of less than 0.1 g earthworm increased by 2.64 times, while 0.1 -0.2 g earthworm proportion rose by only 0.55 times in quartzsand vermifilter.However, percentage of more than 0.3 g earthworm decreased from 28.35 to 3.21% with the operation time in quartz-sand vermifilter.There was a marked drop in 0.2 -0.3 g earthworm percentage when the hydraulic loading was more than 6 m . Ceramisite vermifilter had a similar variation to quartzsand vermifilter in earthworm weight distribution.

Earthworm enzymatic activity
Figures 5 and 6 showed variations of earthworm enzymatic activities, including protease, ALP, cellulase, SOD and CAT in the two kinds of vermifilter.In the quartz-sand vermifilter, protein declined from 66.3 to 45.66 mg g -1 dw (dried weight) earthworm, and ALP activity from 74.62 to 19.75 U • mg -1 protein with hydraulic loading.In addition, cellulase had the maximum in the W 2 condition (100.2U • mg -1 protein), and the lowest in the W 4 condition (4.83 U • mg -1 protein).However, SOD and CAT activity increased from 47.35 to 59.74 U • mg -1 protein, and 3.50 to 7.45 U • mg -1 protein with hydraulic loading.Furthermore, cellulase and SOD activity were significantly different in both kinds of vermifilter (T-Test, P= 0.04 and 0.02), but there were no significant differences for protease, ALP and CAT activities (T-Test, P > 0.05).

Treatment efficiency
At present, there is an abundance of ecological and decentralized wastewater treatment technology, such as constructed wetland (Babatunde et al., 2008;Zhang et al., 2009), stabilization pond (Garcia et al., 2000;Heubeck et al., 2007), and land treatment (Li et al., 2005).These technologies had good treatment effect, but they were restrictively applied on wastewater treatment due to large occupied area.Our results showed that vermifilter could achieve good performance; the results were close to or even better than those of conventional decentralized wastewater treatments (Zhang et al., 2009).Further, the hydraulic loading of our vermifilter could reach 2.4 -6.7 m . However, the conventional ecological wastewater treatment, such as constructed wetland, was usually less than 1.55 m 3 • m -2 • d -1 (Zhang et al., 2009).Higher hydraulic loading indicated more processing capacity of wastewater.Therefore, we considered that vermifilter would have desirable application due to less land area, compared to other ecological and decentralized waster treatments.
Our observations showed that hydraulic loadings could affect the removal rates of COD, BOD 5 , SS, TN, and NH 4 - N (Figure 2).We proposed the following reasons: (1) increasing hydraulic loadings means shortening hydraulic retention time, so organic substrates are not fully degraded before discharged from the vermifilter; (2) increasing hydraulic loadings leads to stronger scour for media surfaces, which was also responsible for the decrease in treatment efficiencies of vermifilter (Liu et al., 2008).

Earthworm characteristics
With vermifilter operation, adult and clitellated earthworm abundance decreased, but the densities of hatchling and cocoon increased (Figure 3).The increase of earthworm hatchling and cocoon indicated that earthworm could breed and incubate in vermifilter very well, and were suitable to the vermifilter environment.The weight distribution of earthworm also showed that the proportion of the smaller individuals rose with operation time (Figure 4).However, the decrease in the proportion of adult and bigger individuals may be due to two reasons.It could be the result of normal metabolic process: the adults produced cocoons, then the cocoons became juveniles, and the adults died.It could also be related to the increase of hydraulic loading which increased humidity and scouring of vermifilter, which was not beneficial for earthworm growth.Digestive enzymes existed in the body of earthworm, such as protease, alkaline phosphates, and cellulase.These enzymes had an important relationship with the N and P cycle, and the turnover of carbon (Alef et al., 1995;Paul and Clark, 1996;Chapin et al., 2002;Schimel and Bennet, 2004;Aira et al., 2007).Another kind of earthworm enzymes were the antioxidant enzymes, such as SOD and CAT, which had been often used as biomarkers of environmental stress (Song et al., 2009).These enzymes could protect cells against adverse effects of reactive oxygen species.An increase in the activities of these enzymes indicated deterioration in environmental condition.In our study, the activities of digestive enzyme dropped, and that of antioxidant enzyme rose with the increase of hydraulic loading (Figures 5 and 6).This indicated that the increase of hydraulic loading was not beneficial for the digestion and growth of earthworm, as revealed in the data of earthworm abundances.

Relationship between treatment efficiency and earthworm characteristics
Earthworm played a critical role on the vermicompost (Domínguez and Edwards, 2004).This was because earthworm could improve activity of microorganism and stabilization of organic matter (Arancon et al., 2004(Arancon et al., , 2005(Arancon et al., , 2006;;Aira et al., 2007;Ravikumar et al., 2008;Pramanik et al., 2009).In our study, abundance and enzymatic activities of earthworm had significant correlation with treatment efficiency of vermifilter (Table 3).We supposed that the treatment efficiency was not influenced only by earthworm abundance, but also by earthworm growth state.In order to keep the vermifilter in good treatment efficiency, therefore, we need to maintain sufficient density of earthworm in vermifilter.More importantly, good condition is necessary for earthworm growth in suitable hydraulic loading.The proportion of more than 0.3 g earthworm correlated significantly with removal rates of COD, BOD 5 , SS, and NH 4 -N, which indicated that the bigger earthworm might have a more important role on wastewater treatment of vermifilter, compared to the smaller one.Thus, although hydraulic loading had little influence on earthworm reproduction and increase of juveniles, the decrease in adults and larger earthworms contributed to the drop in treatment efficiency of vermifilter.
In our study, we found that CAT activity had a significant correlation with most of the indicators tested except NH 4 -N reduction and the proportion of 0.2 -0.3 g earthworm.Thus, in comparison to other characteristics of earthworm, CAT activity of earthworm should be a good indicator for the treatment efficiency and earthworm characteristics.

Figure 2 .Figure 3 .
Figure 2. Removal rates of wastewater in four working conditions of both vermifilters.Values are means, bars are S.E., and n = 4.

Figure 4 .
Figure 4. Earthworm percentage of different weights in four working conditions of both vermifilters.Values are means, bars are S.E., and n = 4.

Figure 5 .Figure 6 .
Figure 5. Earthworm digestive enzymatic activities in four working conditions of both vermifilters.Values are means, bars are S.E., and n = 3.

Table 2 .
Characteristics of the influent water for four working conditions of both vermifilters.

Table 3 .
Pearson correlation coefficients of treatment efficiencies and earthworm characteristics in vermifilter a .