w a t e r r e s e a r c h 5 1 ( 2 0 1 4 ) 2 1 6 e2 2 7
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/watres
Seasonal changes in the invertebrate community of granular activated carbon filters and control technologies Qing Wang a,b, Wei You a,b, Xiaowei Li a,b, Yufeng Yang a,b,*, Lijun Liu c a
Institute of Hydrobiology, Jinan University, Guangzhou 510632, PR China Key Laboratory of Aquatic Eutrophication and Control of Harmful Algal Blooms, Guangdong Higher Education Institutes, Guangzhou 510632, PR China c Shenzhen Water (Group) Co., Ltd, No. 1019, Shennan Middle Road, Shenzhen 518031, PR China b
article info
abstract
Article history:
Invertebrate colonization of granular activated carbon (GAC) filters in the waterworks is one of
Received 22 May 2013
the most frequently occurring and least studied biological problems of water processing in
Received in revised form
China. A survey of invertebrate colonization of GAC filters was carried out weekly from
25 October 2013
October 2010 to December 2011 at a reservoir water treatment works in South China. Twenty-
Accepted 26 October 2013
six kinds of invertebrates were observed. The abundance was as high as 5600 ind. m3 with a
Available online 7 November 2013
mean of 860 ind. m3. Large variations in abundance were observed among different seasons and before and after GAC filtration. The dominant organisms were rotifers and copepods. The
Keywords:
average invertebrate abundance in the filtrate was 12e18.7 times of that in the pre-filtered
Rotifer
water. Results showed that the GAC filters were colonized by invertebrates which may lead
Copepod
to a higher output of organisms in the filtrate than in the pre-filtered water. The invertebrate
GAC filter
abundance in the GAC filters was statistically correlated with the water temperature. Sea-
Drinking water
sonal patterns were observed. The invertebrate abundance grew faster in the spring and
Biological filtration
summer. Copepods were dominant in the summer while rotifers dominated in all other
Waterworks
seasons of the year. There was a transition of small invertebrates (rotifers) gradually being substituted by larger invertebrates (copepods) from spring to summer. Control measures such as backwashing with chloric water, drying filter beds and soaking with saliferous water were implemented in the waterworks to reduce invertebrate abundances in the GAC filters. The results showed that soaking with saliferous water (99%, reduction in percent) was best but drying the filter beds (84%) was more economical. Soaking filter beds with 20 g/L saliferous water for one day can be implemented in case of emergency. In order to keep invertebrate abundance in the acceptable range, some of these measures should be adopted. ª 2013 Elsevier Ltd. All rights reserved.
1.
Introduction
One of the most pervasive problems afflicting people throughout the world is inadequate access to clean water
(Shannon et al., 2008). The significance of multicellular organisms in drinking water is attracting increasing scientific attention as we are starting to better understand their capacity to act as vectors of waterborne pathogens (Bichai et al.,
* Corresponding author. Institute of Hydrobiology, Jinan University, Guangzhou 510632, PR China. Tel.: þ86 20 85221397; fax: þ86 20 85220239. E-mail address: [email protected] (Y. Yang). 0043-1354/$ e see front matter ª 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.watres.2013.10.064
w a t e r r e s e a r c h 5 1 ( 2 0 1 4 ) 2 1 6 e2 2 7
2010). Previous results have shown that nematodes and other invertebrates may protect pathogenic microorganisms from disinfection during surface water treatment (Bichai et al., 2008, 2009, 2010). Furthermore, the bacteria associated with the invertebrates are potential pathogens for human beings (Wolmarans et al., 2005), although there is no evidence of a significant effect on microbial safety yet. The estimated number of bacteria that could be associated with a single invertebrate (as based on average invertebrate numbers) could range from 10 to 4000 bacteria per organism, most of them were isolated from the intestines of invertebrates (Wolmarans et al., 2005). Invertebrates such as rotifers and crustaceans are likely to be found in all drinking water distribution systems and at some stages of full-scale water treatment works (Van Lieverloo et al., 2012), but mostly in granular material filter effluents (Castaldelli et al., 2005; Schreiber et al., 1997). Although most of the invertebrates that pass through the water filter and taps are not easily observed, a few larger invertebrates are visible to the naked
217
eye and lead to consumer complaints (Van Lieverloo et al., 2002). The occurrence of macroscopically visible organisms in the drinking water such as oligochaetes, for example, is certainly an aesthetic problem to say the least. For example, during 1999, the water supplier in Gauteng, South Africa handled 68 consumer complaints of which 19 were due to invertebrates (Shaddock, 2005) and approximately several dozen times of consumer complaints every year in Netherlands (Van Lieverloo et al., 1997). Although there is no direct hygienic relevance to the invertebrates detected in the GAC filtrate, a high abundance of invertebrates in the stagnant pipes can lead to the regrowth of microbes within the distribution system (Schreiber et al., 1997). Tan et al. (2000) reported that Lecane inermis, which infected the urinary system of a person in China was common in the drinking water system. Comparing the number of invertebrates found in raw water samples with those found in drinking, there is a striking rise in all waterworks examined that shows that the biological
Fig. 1 e The technical process in the water works examined and sampling sites shown.
218
w a t e r r e s e a r c h 5 1 ( 2 0 1 4 ) 2 1 6 e2 2 7
granular activated carbon (GAC) filters themselves can be the origin of invertebrates in the drinking water supply (Schreiber et al., 1997). GAC filters are often the penultimate stage of surface water treatment and provide ideal habitats for invertebrates. Proliferation of chlorine-resistant invertebrates in GAC filters may lead to their efflux into distribution systems, possibly resulting in contamination of customers’ tap water (Weeks et al., 2007). Furthermore, rotifers, which are potential predators of (oo)cysts in GAC filters and play a role in the transport of internalized (oo)cysts into filtered water (Bichai et al., 2010), accounted for the majority of the isolated invertebrates (Schreiber et al., 1997). Predation by invertebrates can favour persistence of Cryptosporidium and Giardia (oo)cysts in GAC filter beds and act as a vehicle for them to be released into filtered effluents. Therefore, predation could be seen as increasing the persistence and the transport of these pathogens through filters (Bichai et al., 2010). Van Lieverloo et al. (2002) presented a concept explaining the presence and abundance of invertebrates in the water supply. The bacteria and free-living protozoans in the biofilm and sediments most likely serve as the primary food source for the invertebrate community, considering the absence of other food sources such as algae. The objectives of this paper are to (i) quantify the invertebrates in GAC filters, (ii) identify key factors explaining the seasonal variation of invertebrate abundance, and (iii) implement control measures in the works to reduce invertebrate abundance. The present study employed surface and core sampling of GAC filters at various depths and times to investigate invertebrate distribution.
2.
Methods
The selected waterworks, built in 2006, treat reservoir water. In the works, eight GAC filters are the last treatment before disinfection (see flow diagram, Fig. 1). In this study, invertebrate animals collected from the GAC filter beds and filtrate were isolated, identified and quantified. Also control measures were applied to the filter beds in order to reduce invertebrate abundance in the summer.
2.1.
Sampling sites
Samples were collected weekly at a waterworks in southern China from October 2010 to December 2011, according to the following distribution: total inlet water (site 1), inlet water of filter 1 (site 2), inlet water of filter 2 (site 3), filter bed 1 (site 4), filter bed 2 (site 5), filtrate of filter 1 (site 6), filtrate of filter 2 (site 7) and total filtrate water (site 8) (Fig. 1). At site 4 and site 5, the carbon samples in the filters were collected in three layers: top layer 0e40 cm, middle layer 70e110 cm and bottom layer 140e180 cm. There were 372 samples collected from water and 372 carbon samples from filter beds.
2.2.
Sampling procedure
Water samples were taken with a 35 mm mesh plankton net (from Institute of Hydrobiology, Chinese Academy of Sciences, China). About 1 m3 water was filtered at an approximate flow
rate of 0.4 m3/h. Carbon particles in the GAC filter beds were sampled using a stainless steel stratified sampler with a diameter of 4 cm. Fifty grams of carbon sub-samples were removed from each filter bed fraction. Composite samples were washed 5 times by vigorous shaking with sterile distilled water. All washing water was collected and filtered through a 35 mm net to retain the invertebrates. Samples were fixed with 4% formaldehyde (final concentration) (from Guangzhou Chemical Reagent Factory). Identification was carried out using a stereomicroscope (Olympus SZX16) and a regular compound microscope (Olympus BX51). The monogonont rotifers were identified to genus or species as far as possible (Koste, 1978). Other invertebrate groups, such as copepods, cladocerans, oligochaetes, nematodes, were not determined further. For invertebrate enumeration in the GAC material and the water, samples were transferred onto a counting plate and allowed to settle for 5 min. The entire counting chamber was scanned and organisms were enumerated. Invertebrate abundance was expressed as a monthly average in ind/m3 for water samples and ind./kg for carbon material samples.
2.3.
Backwashing procedure
In this waterworks, the backwashing procedure was 13.7 L/s pressure air for 2 min, combined air/water for 2 min, 8 L/s drinking water for 7 min and carried out every two days in summer. In order to determine the effect of a certain backwashing procedure on the invertebrate abundance, samples were taken before and after backwashing procedure. After before backwashing samples were collected, the procedure was carried out immediately. After half an hour the procedure was stopped and then the after backwashing samples were collected. Water samples were also collected from inlet water and the filtrate of a filter bed, and carbon samples were collected from the filter bed, according to sampling procedure mentioned in Section 2.2 for water and carbon material sampling.
2.4.
Backwashing procedure with chlorine
In order to evaluate the effect of backwashing procedures with different concentrations of chlorine on invertebrates, abundances in filtrate were calculated before and after the treatment for total invertebrate, copepods, cladocerans, nauplius and oligochaetes. Three filter beds in the waterworks were selected and the filtrate samples were collected according to sampling procedure in Section 2.2. Backwashing procedures with three concentrations of chlorine (0, 1.0, 1.5 mg/L) were tested and the removal efficiency of the four kinds of invertebrates in the filtrate was compared. The samples before backwashing were taken close to the backwashing procedure, similar to that in Section 2.3, while the samples after backwashing were collected after 2 h, in order to decrease the effect of high invertebrate abundance in the initial filtrate on the result. The three kinds of backwashing procedures were similar and were 13.7 L/s pressure air for 2 min, combined air/ water for 2 min, 8 L/s drinking water for 7 min. For the test of backwashing with 1.5 mg/L chlorine, another step with 8 L/s drinking water for 5 min was added after the normal procedure was stopped for 2 h.
w a t e r r e s e a r c h 5 1 ( 2 0 1 4 ) 2 1 6 e2 2 7
2.5.
Drying carbon filter beds
For diminishing invertebrate abundance in the filtrate, drying carbon filter beds was carried out in the summer. In order to test the effect of drying carbon filter beds on invertebrate abundance, the inlet valve was closed and the outlet valve with the emptying valve was opened. Water in the filter bed was drained out. The carbon material in the bed was air-dried for several days. The air temperature was 23e28 C during the experiment. The height of dry top layer in the filter beds increased with time and was about 80 cm after five days, accounting for 44% of total depth of the carbon layer which was 180 cm. Drying was affected by the humidity and temperature of the air. Air backwashing could be used to accelerate drying speed. After five days, the emptying valve was closed and the inlet valve was opened. After some water flowed into the carbon filter beds, the manual backwashing process was started. The backwashing procedure was: 13.7 L/ s air pressure for 2 min, combined air/water for 2 min, 8 L/s drinking water for 7 min. After the manual backwashing was finished, an automatic backwashing process was started every two days. After 22 days, the backwashing process cycle was adjusted to a 3-day cycle. Water samples and carbon samples were collected before drying (0.5 h), in the initial day after drying was finished (0.5 h, 1 h, 2 h, 4 h, 24 h), and in the stable period after drying was finished (two days later, every 2e5 days). Sampling, processing, identification and counting methods for water and carbon material samples were the same with that described in Section 2.2 for water and carbon samples.
2.6.
Soaking carbon filter beds with salt water
Soaking carbon filter beds with salt water can reduce invertebrates in a short time in case of emergency. After the water in the filter beds was drained, crystalline NaCl salt was spread evenly over the surface of the filter beds (about 50 kg/ m2). Water was added to filter beds by the backwashing pipe to a depth of 10 cm from the surface of the carbon beds. Air backwashing was implemented to make sure the salt was fully dissolved and uniformly distributed over the filter beds. Conductivity in the water of the beds and filtrate was monitored. The final concentration of the salt solution was about 20 g/L. During the immersing period, air backwashing was implemented for 3 min every 3 h to suspend the filter material and mix it thoroughly with the salt water, which ensured that the invertebrates were fully exposed to the salt solution. No salt water flowed out of the filter beds during air backwashing. The survival of invertebrates was determined in the surface, middle and bottom of the beds every 4 h. After 24 h, the salt solution was drained and water was added to a depth of 10e20 cm from the carbon surface. Air backwashing (3 min) and water backwashing (10 min, 2000 m3/h for flow) were carried out twice, after which the water in the filter beds was drained. Backwashing (2000 m3/h flow) was applied again for 60 min until the conductivity in the head and bottom water of the filter beds recovered to normal (150 ms/cm) that is, the same as before soaking. The filter beds were allowed to operate as usual. Before soaking carbon filter beds with salt water, water backwashing was
219
carried out. The first water samples were collected from backwashing water after it had run for 15 min. After 24 h, the salt solution was drained and water backwashing started. The other water samples were collected from backwashing water after it had run for 15, 90 and 120 min, respectively, in order to evaluate living crustaceans and oligochaetes in the filter beds. Sampling, processing, identification and counting methods for water and carbon material samples were the same with that described in Section 2.2 for water and carbon samples.
2.7.
Water quality parameters
Water quality parameters such as water temperature, turbidity, pH, ozone, dissolved oxygen were the real-time data monitored by online instrumentation (HACH Water Online Analysis System). The potassium permanganate index (CODMn) was determined according to the state standard method (Ministry of Environmental Protection of the People’s Republic of China, 2002).
2.8.
Statistical analyses
The monthly mean abundances of invertebrates were calculated as arithmetic means with standard deviation (SD). The one-sample KolmogoroveSmirnov test of nonparametric test was used to test data distribution of mean abundances. The independent samples t-test was used for comparing the monthly mean abundances of invertebrate in the inlet water and filtrate. Pearson correlations between water quality parameters and log-transformed mean abundance of invertebrate were used to identify significant correlation of the variables with abundances. Regression analysis was used between the identified variables and the mean abundance of invertebrate by a curve estimation of regression. The specific model was selected based on coefficient of determination (R2), the standard error of the estimate, the F value and the significance (P) obtained by curve estimation. Considering the relatively large interval of measurements (a month), we believe that the theoretical possibility of autocorrelation in the time series data is negligible. Statistical tests were performed with SPSS 16.0 and figures were drawn with Origin 7.0.
3.
Results and discussion
3.1.
Water quality
During sampling, water temperature ranged between 14.3 C and 30.7 C. On most sampling occasions, the water temperature was above 20 C. The annual mean water temperature was 23.1 C. Turbidity showed little variation and was generally below 0.1NTU, which was much less than the WHO guideline (1 NTU) (WHO, 2011). Through the procedures of preozone contact tank and primary ozone contact tank, the concentration of ozone was in the range between 0.200 and 0.400 ppm in the inlet water. Dissolved oxygen concentration was adequate for organisms in the GAC filters because of ozone decomposition. The CODMn was relatively low (Table 1).
220
w a t e r r e s e a r c h 5 1 ( 2 0 1 4 ) 2 1 6 e2 2 7
Table 1 e Water quality parameters in the water examined. Parameter Water temperature Turbidity (NTU) pH pH Ozone (ppm) Dissolved oxygen (mg/L) Dissolved oxygen (mg/L) CODMn (mg/L) CODMn (mg/L)
Range
Mean SD
Site
14.3e30.7 0.053e0.102 7.41e8.01 6.95e7.66 0.200e0.400 10e12 5e8 1.20e2.25 0.75e1.65
23.1 5.23 0.0762 0.0135 7.68 0.18 7.28 0.21
1 1 1 8 1 1 8 1 8
6 1.55 1.05
CODMn: chemical oxygen demand (potassium permanganate index). Site 1 represents total inlet water to the GAC filter beds and site 8 represents total filtrate water from the GAC filter beds.
Fig. 2 e Comparisons among mean (±SD) abundances of invertebrates in the inlet water, filtrate and GAC filter beds from October 2010 to December 2011.
3.2. Species composition of invertebrates in the inlet water and filtrate During the survey period, 26 kinds of invertebrates in 3 phyla, 5 classes, 6 orders, 6 families, and 8 genera were observed. The dominant groups were rotifers (20 species), then copepods, both in GAC material and water samples. Most of the rotifers belonged to Lecane, Lepadella, Colurella and Bdelloidea species. Most copepods were Harpacticoida species (e.g. Nitocra pietschmanni) and Cyclopoidea species (e.g. Mesocyclops spp.). In the inlet water, 19 taxa were observed, while in the filtrate, 24 were found. Five Lecane species and Cyclopoida species were only found in the filtrate and cladocerans species were only found in the inlet water. Differences in the dominant species amongst rotifers between the inlet water and filtrate examined were also documented. Colurella spp. and Lecane spp. were dominant in the inlet water, while in the filtrate, only Lecane spp. were dominant. Other invertebrates included oligochaetes (e.g. Nais spp.), nematodes, chironomid larvae and water mites. Schreiber et al. (1997) found that rotifers were the dominant taxonomic group in terms of numbers followed by nematodes which were in general the second important group in the GAC filtrate samples collected by a plankton net of 10 mm mesh width. Nematodes are the most common and abundant colonizers. Castaldelli et al. (2005) reported that they were present in 80% of Italian drinking water treatment plants which treat groundwater or spring water. However, nematodes were in low abundance compared to rotifers and copepods in the present study. Since our experimental protocol targeted invertebrates larger than 35 mm, it is likely that not all invertebrate organisms were extracted from the GAC material and water samples. The species were recorded with specimens in the 35 mm filters.
3.3. Invertebrate abundance in the inlet water and filtrate To check for the general influence of the GAC filters on invertebrate abundance in the water, samples were taken for comparison (Fig. 2). The average abundance of invertebrates found in the total inlet water (inlet water to the eight GAC filters) was 113 79 ind./m3, while the average abundance in
the total filtrate (effluent of the eight GAC filters) was 907 639 ind./m3. The average invertebrate abundance in the filtrate was 8 times of that in the inlet water and the abundances in the filtrate were significantly higher than that in the inlet water (independent Student’s t-test, t ¼ 6.548, P < 0.001). The average invertebrate abundances in the filtrate of GAC 1 and GAC 2 (636 483 ind./m3 and 377 330 ind./m3) were 14.2 and 10.5 times of that in the inlet water (45 19 ind./m3 and 36 17 ind./m3) and were significantly higher (independent Student’s t-test; t ¼ 8.081, P < 0.001, GAC 1; t ¼ 6.952, P < 0.001, GAC 2), respectively. The average invertebrate abundances in the filtrate and inlet water of GAC 1 were slightly higher than that of GAC 2, but there were no significant difference between the two GAC filters. Comparing the invertebrate abundances found in inlet water samples with those found in the filtrate, there was a striking increase in the GAC filters, showing their importance for invertebrate multiplication. This underlines the importance of metazoans in the biological processes of GAC filters. The accumulating biomass within the filters, of which the output of invertebrates was an expression, might give rise to a more complex biocoenosis (Schreiber et al., 1997). The overall changes in the invertebrate output with the water temperature were investigated in the two pre-filtered and filtrate waters (Fig. 3). The invertebrate abundances in spring and summer were usually higher than those in autumn and winter. Invertebrate abundances in the process exhibited a high level in the summer with the maximum abundance of 275 ind./m3 and 1920 ind./m3 in the total GAC inlet water and filtrate, respectively. Seasonal changes in invertebrate abundances in the filtrate of GAC (total, GAC 1 and GAC 2) were considerably larger than those in the inlet water. The invertebrate abundances in the filtrate of GAC (total, GAC 1 and GAC 2) ranged 1300e1920 ind./m3 in the warm season, while those in the inlet water were mostly under 200 ind./m3. Rotifers dominated the taxonomic group in terms of abundance followed by copepods, accounting for w73% and w26% of the total invertebrate number in the total inlet waters and for w60% and w35% released in the total filtrate, respectively (Fig. 4). Abundance of other invertebrates, such as oligochaetes and nematodes, was relatively low. Colurella spp.
w a t e r r e s e a r c h 5 1 ( 2 0 1 4 ) 2 1 6 e2 2 7
221
Fig. 3 e Seasonal changes in invertebrate abundances (monthly means ± SD, n [ 4) in (A) total inlet water and total filtrate for eight GAC carbon filters; (B) GAC carbon filters 1 and 2; (C) inlet water and filtrate of GAC carbon filter 1; (D) inlet water and filtrate of GAC carbon filter 2 from October 2010 to December 2011.
(44% in the filter 1, 38% in the filter 2) and Lecane spp. (25% in the filter 1, 33% in the filter 2) were the most common rotifers in the inlet water samples. Lecane spp. (72% in the filter 1, and 73% in the filter 2) dominated in the filtrate while Colurella spp. just accounted for 16% and 15% in the filter 1 and filter 2, respectively. Other benthic rotifers such as Lepadella spp. (10% in the inlet water, 5% in the filtrate) and Bdelloidea (4% in the inlet water and 3% in the filtrate) were found to a much lesser extent. In another study, Colurella sp. and Bdelloidea dominated the rotifer fauna, and other rotifers like Lepadella sp., Dicranophoridae and Lecane sp. were common (Schreiber et al., 1997).
Regression analysis showed that mean abundance of rotifers and copepods in the filtrate were significantly impacted by water temperature (log Y ¼ 0.369 þ 0.233X 0.004X2, R2 ¼ 0.486, P ¼ 0.019, rotifers; log Y ¼ 2.662 0.171X þ 0.006X2, R2 ¼ 0.682, P ¼ 0.001, copepods). A change in the frequency of the dominant group in the filtrate was detected during the study period (Fig. 4). The percentage of copepods increased in the summer, especially in the total filtrate where the percentage exceeded that of rotifers. Rotifers accounted for 51e100% of total invertebrate abundance in the total inlet water, while 5e98% in the total filtrate. From June to September, the percentage of rotifers was less than 50%, and
Fig. 4 e Seasonal changes in the abundances of rotifers, copepods and proportion in the total inlet water and total filtrate. (A) Abundance in the total inlet water, stacked bar diagrams; (B) abundance in the total filtrate, stacked bar diagrams; (C) percentage in the total inlet water; and (D) percentage in the total filtrate.
222
w a t e r r e s e a r c h 5 1 ( 2 0 1 4 ) 2 1 6 e2 2 7
Fig. 5 e The vertical distribution of invertebrates (monthly means ± SD, n [ 4) from the GAC carbon (ind./kg) in the filter bed 1 (site 4) and filter bed 2 (site 5).
the percentage of copepods was more than 50%. During the period, the mean water temperature was 29.3 1.40 C, while during the other time of the study period, the mean water temperature was 20.9 4.14 C. In the summer, in order to ensure the water quality parameters under consideration and to decrease invertebrate abundance in the filter beds, the backwash cycle is carried out once in every 1e2 days, while in the other times of the year, the backwash cycle is carried out once in every 3e4 days. Overall, from spring to summer, the percentage of rotifers out of the total invertebrate abundance gradually diminished, while the percentage of copepods gradually increased in the same time period. Looking at the relative frequency of rotifers and copepods in the filtrate samples, there was a tendency of copepods to outcompete the rotifers as water temperature increased. The GAC filters increased the relative frequency of copepods as water temperature increased.
3.4.
Invertebrate abundance in the GAC filter beds
Invertebrates were distributed throughout the entire depth of the filter bed. The monthly average concentrations of invertebrates over time in the upper, middle and bottom parts of the filter beds (ind./kg) from GAC 1 and GAC 2 are shown in Fig. 5. The mean invertebrate abundances were comparable in both filter beds: 1021, 1189 and 572 ind./kg in the upper, middle and bottom parts of the GAC 1 and 927, 990 and 462 ind./kg in the GAC 2, respectively. Invertebrate abundance was higher in the summer and reduced when the carbon layer depth increased. The abundances in the upper (0e40 cm) and middle (70e110 cm) parts of GAC were significantly higher than that in the bottom (140e180 cm) (independent Student’s t-test, P < 0.01) while there were no significant differences in abundances between the upper and middle layers. Most aquatic invertebrates have little potential for passing through the entire modern water treatment system. However, a small proportion can survive chemical treatment and colonize the treatment media used in rapid gravity filters and granular activated carbon (GAC) filters, occasionally leading to
an efflux of these organisms into distribution systems, where there can be enough organic matter to sustain populations (Weeks et al., 2007; Pre´vost et al., 1998). Regression analysis showed that water temperature had a significant effect on invertebrate abundance in the water and filter beds (Table 2). The water temperature is above 20 C most of the year and there are enough eggs and larvae. Once food source is sufficient, invertebrates multiply and abundance increases fast. In the warm season, bacteria growth is also fast and the concentration of organic matter increases. Previous studies suggested that water temperature was an important factor influencing bacteria regrowth and that the concentration of organic substrate in the distribution system (Niquette et al., 2001; Francisque et al., 2009; Shaddock, 2005), may be the main food source for invertebrates in the filter beds. Furthermore, water temperature influences invertebrate growth and reproduction (Shaddock, 2005). Higher temperatures support faster growth rates and enable some biota to attain significant populations (Chapman, 1996). Although 62.7e92.9% of the variability of mean abundance in most sites could be statistically explained by differences in water temperature (Table 2), there is a need for further study of other factors. The other water quality parameters under study did not have an observable impact on invertebrate abundance (probably due to their relatively limited variability). Therefore, other factors such as backwashing and flow rate should be taken into account in future studies.
3.5. Control technologies and emergency measures in waterworks Since the occurrence of invertebrates in drinking water can cause complaints from consumers and not all critical aspects concerning the microbiological safety of drinkable water can be dismissed, an extensive removal of invertebrates by treatment processes should be achieved (Castaldelli et al., 2005). Starvation and flushing were used as control measures to inhibit invertebrate development in the drinking distribution system (Van Lieverloo et al., 1997, 1998). Various processes in
223
w a t e r r e s e a r c h 5 1 ( 2 0 1 4 ) 2 1 6 e2 2 7
Table 2 e Regression analysis between mean abundance of invertebrate (Y) and water temperature (X). Invertebrate abundance (ind./m3)/(ind./kg) Total inlet water Total filtrate Inlet water of GAC 1 Inlet water of GAC 2 Filtrate of GAC 1 Filtrate of GAC 2 Filter beds of GAC 1 Filter beds of GAC 2
Site
Model N
1 8 2 3 6 7 4 5
15 15 15 15 15 15 15 15
2
R
0.929 0.64 0.682 0.897 0.636 0.296 0.627 0.770
Equation
Residuals statistics SE
log Y ¼ 0.292 þ 0.07X log Y ¼ 1.422 þ 0.061X log Y ¼ 1.918 þ 0.279X 0.005X2 log Y ¼ 3.151 þ 0.372X 0.007X2 log Y ¼ 0.951 þ 0.07X log Y ¼ 1.364 þ 0.044X log Y ¼ 3.509 0.09X þ 0.003X2 log Y ¼ 2.993 0.048X þ 0.002X2
0.104 0.248 0.156 0.110 0.288 0.370 0.149 0.095
F 171 23.1 12.9 52.4 22.7 5.46 10.1 20.1
Sig. b
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/watres
Seasonal changes in the invertebrate community of granular activated carbon filters and control technologies Qing Wang a,b, Wei You a,b, Xiaowei Li a,b, Yufeng Yang a,b,*, Lijun Liu c a
Institute of Hydrobiology, Jinan University, Guangzhou 510632, PR China Key Laboratory of Aquatic Eutrophication and Control of Harmful Algal Blooms, Guangdong Higher Education Institutes, Guangzhou 510632, PR China c Shenzhen Water (Group) Co., Ltd, No. 1019, Shennan Middle Road, Shenzhen 518031, PR China b
article info
abstract
Article history:
Invertebrate colonization of granular activated carbon (GAC) filters in the waterworks is one of
Received 22 May 2013
the most frequently occurring and least studied biological problems of water processing in
Received in revised form
China. A survey of invertebrate colonization of GAC filters was carried out weekly from
25 October 2013
October 2010 to December 2011 at a reservoir water treatment works in South China. Twenty-
Accepted 26 October 2013
six kinds of invertebrates were observed. The abundance was as high as 5600 ind. m3 with a
Available online 7 November 2013
mean of 860 ind. m3. Large variations in abundance were observed among different seasons and before and after GAC filtration. The dominant organisms were rotifers and copepods. The
Keywords:
average invertebrate abundance in the filtrate was 12e18.7 times of that in the pre-filtered
Rotifer
water. Results showed that the GAC filters were colonized by invertebrates which may lead
Copepod
to a higher output of organisms in the filtrate than in the pre-filtered water. The invertebrate
GAC filter
abundance in the GAC filters was statistically correlated with the water temperature. Sea-
Drinking water
sonal patterns were observed. The invertebrate abundance grew faster in the spring and
Biological filtration
summer. Copepods were dominant in the summer while rotifers dominated in all other
Waterworks
seasons of the year. There was a transition of small invertebrates (rotifers) gradually being substituted by larger invertebrates (copepods) from spring to summer. Control measures such as backwashing with chloric water, drying filter beds and soaking with saliferous water were implemented in the waterworks to reduce invertebrate abundances in the GAC filters. The results showed that soaking with saliferous water (99%, reduction in percent) was best but drying the filter beds (84%) was more economical. Soaking filter beds with 20 g/L saliferous water for one day can be implemented in case of emergency. In order to keep invertebrate abundance in the acceptable range, some of these measures should be adopted. ª 2013 Elsevier Ltd. All rights reserved.
1.
Introduction
One of the most pervasive problems afflicting people throughout the world is inadequate access to clean water
(Shannon et al., 2008). The significance of multicellular organisms in drinking water is attracting increasing scientific attention as we are starting to better understand their capacity to act as vectors of waterborne pathogens (Bichai et al.,
* Corresponding author. Institute of Hydrobiology, Jinan University, Guangzhou 510632, PR China. Tel.: þ86 20 85221397; fax: þ86 20 85220239. E-mail address: [email protected] (Y. Yang). 0043-1354/$ e see front matter ª 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.watres.2013.10.064
w a t e r r e s e a r c h 5 1 ( 2 0 1 4 ) 2 1 6 e2 2 7
2010). Previous results have shown that nematodes and other invertebrates may protect pathogenic microorganisms from disinfection during surface water treatment (Bichai et al., 2008, 2009, 2010). Furthermore, the bacteria associated with the invertebrates are potential pathogens for human beings (Wolmarans et al., 2005), although there is no evidence of a significant effect on microbial safety yet. The estimated number of bacteria that could be associated with a single invertebrate (as based on average invertebrate numbers) could range from 10 to 4000 bacteria per organism, most of them were isolated from the intestines of invertebrates (Wolmarans et al., 2005). Invertebrates such as rotifers and crustaceans are likely to be found in all drinking water distribution systems and at some stages of full-scale water treatment works (Van Lieverloo et al., 2012), but mostly in granular material filter effluents (Castaldelli et al., 2005; Schreiber et al., 1997). Although most of the invertebrates that pass through the water filter and taps are not easily observed, a few larger invertebrates are visible to the naked
217
eye and lead to consumer complaints (Van Lieverloo et al., 2002). The occurrence of macroscopically visible organisms in the drinking water such as oligochaetes, for example, is certainly an aesthetic problem to say the least. For example, during 1999, the water supplier in Gauteng, South Africa handled 68 consumer complaints of which 19 were due to invertebrates (Shaddock, 2005) and approximately several dozen times of consumer complaints every year in Netherlands (Van Lieverloo et al., 1997). Although there is no direct hygienic relevance to the invertebrates detected in the GAC filtrate, a high abundance of invertebrates in the stagnant pipes can lead to the regrowth of microbes within the distribution system (Schreiber et al., 1997). Tan et al. (2000) reported that Lecane inermis, which infected the urinary system of a person in China was common in the drinking water system. Comparing the number of invertebrates found in raw water samples with those found in drinking, there is a striking rise in all waterworks examined that shows that the biological
Fig. 1 e The technical process in the water works examined and sampling sites shown.
218
w a t e r r e s e a r c h 5 1 ( 2 0 1 4 ) 2 1 6 e2 2 7
granular activated carbon (GAC) filters themselves can be the origin of invertebrates in the drinking water supply (Schreiber et al., 1997). GAC filters are often the penultimate stage of surface water treatment and provide ideal habitats for invertebrates. Proliferation of chlorine-resistant invertebrates in GAC filters may lead to their efflux into distribution systems, possibly resulting in contamination of customers’ tap water (Weeks et al., 2007). Furthermore, rotifers, which are potential predators of (oo)cysts in GAC filters and play a role in the transport of internalized (oo)cysts into filtered water (Bichai et al., 2010), accounted for the majority of the isolated invertebrates (Schreiber et al., 1997). Predation by invertebrates can favour persistence of Cryptosporidium and Giardia (oo)cysts in GAC filter beds and act as a vehicle for them to be released into filtered effluents. Therefore, predation could be seen as increasing the persistence and the transport of these pathogens through filters (Bichai et al., 2010). Van Lieverloo et al. (2002) presented a concept explaining the presence and abundance of invertebrates in the water supply. The bacteria and free-living protozoans in the biofilm and sediments most likely serve as the primary food source for the invertebrate community, considering the absence of other food sources such as algae. The objectives of this paper are to (i) quantify the invertebrates in GAC filters, (ii) identify key factors explaining the seasonal variation of invertebrate abundance, and (iii) implement control measures in the works to reduce invertebrate abundance. The present study employed surface and core sampling of GAC filters at various depths and times to investigate invertebrate distribution.
2.
Methods
The selected waterworks, built in 2006, treat reservoir water. In the works, eight GAC filters are the last treatment before disinfection (see flow diagram, Fig. 1). In this study, invertebrate animals collected from the GAC filter beds and filtrate were isolated, identified and quantified. Also control measures were applied to the filter beds in order to reduce invertebrate abundance in the summer.
2.1.
Sampling sites
Samples were collected weekly at a waterworks in southern China from October 2010 to December 2011, according to the following distribution: total inlet water (site 1), inlet water of filter 1 (site 2), inlet water of filter 2 (site 3), filter bed 1 (site 4), filter bed 2 (site 5), filtrate of filter 1 (site 6), filtrate of filter 2 (site 7) and total filtrate water (site 8) (Fig. 1). At site 4 and site 5, the carbon samples in the filters were collected in three layers: top layer 0e40 cm, middle layer 70e110 cm and bottom layer 140e180 cm. There were 372 samples collected from water and 372 carbon samples from filter beds.
2.2.
Sampling procedure
Water samples were taken with a 35 mm mesh plankton net (from Institute of Hydrobiology, Chinese Academy of Sciences, China). About 1 m3 water was filtered at an approximate flow
rate of 0.4 m3/h. Carbon particles in the GAC filter beds were sampled using a stainless steel stratified sampler with a diameter of 4 cm. Fifty grams of carbon sub-samples were removed from each filter bed fraction. Composite samples were washed 5 times by vigorous shaking with sterile distilled water. All washing water was collected and filtered through a 35 mm net to retain the invertebrates. Samples were fixed with 4% formaldehyde (final concentration) (from Guangzhou Chemical Reagent Factory). Identification was carried out using a stereomicroscope (Olympus SZX16) and a regular compound microscope (Olympus BX51). The monogonont rotifers were identified to genus or species as far as possible (Koste, 1978). Other invertebrate groups, such as copepods, cladocerans, oligochaetes, nematodes, were not determined further. For invertebrate enumeration in the GAC material and the water, samples were transferred onto a counting plate and allowed to settle for 5 min. The entire counting chamber was scanned and organisms were enumerated. Invertebrate abundance was expressed as a monthly average in ind/m3 for water samples and ind./kg for carbon material samples.
2.3.
Backwashing procedure
In this waterworks, the backwashing procedure was 13.7 L/s pressure air for 2 min, combined air/water for 2 min, 8 L/s drinking water for 7 min and carried out every two days in summer. In order to determine the effect of a certain backwashing procedure on the invertebrate abundance, samples were taken before and after backwashing procedure. After before backwashing samples were collected, the procedure was carried out immediately. After half an hour the procedure was stopped and then the after backwashing samples were collected. Water samples were also collected from inlet water and the filtrate of a filter bed, and carbon samples were collected from the filter bed, according to sampling procedure mentioned in Section 2.2 for water and carbon material sampling.
2.4.
Backwashing procedure with chlorine
In order to evaluate the effect of backwashing procedures with different concentrations of chlorine on invertebrates, abundances in filtrate were calculated before and after the treatment for total invertebrate, copepods, cladocerans, nauplius and oligochaetes. Three filter beds in the waterworks were selected and the filtrate samples were collected according to sampling procedure in Section 2.2. Backwashing procedures with three concentrations of chlorine (0, 1.0, 1.5 mg/L) were tested and the removal efficiency of the four kinds of invertebrates in the filtrate was compared. The samples before backwashing were taken close to the backwashing procedure, similar to that in Section 2.3, while the samples after backwashing were collected after 2 h, in order to decrease the effect of high invertebrate abundance in the initial filtrate on the result. The three kinds of backwashing procedures were similar and were 13.7 L/s pressure air for 2 min, combined air/ water for 2 min, 8 L/s drinking water for 7 min. For the test of backwashing with 1.5 mg/L chlorine, another step with 8 L/s drinking water for 5 min was added after the normal procedure was stopped for 2 h.
w a t e r r e s e a r c h 5 1 ( 2 0 1 4 ) 2 1 6 e2 2 7
2.5.
Drying carbon filter beds
For diminishing invertebrate abundance in the filtrate, drying carbon filter beds was carried out in the summer. In order to test the effect of drying carbon filter beds on invertebrate abundance, the inlet valve was closed and the outlet valve with the emptying valve was opened. Water in the filter bed was drained out. The carbon material in the bed was air-dried for several days. The air temperature was 23e28 C during the experiment. The height of dry top layer in the filter beds increased with time and was about 80 cm after five days, accounting for 44% of total depth of the carbon layer which was 180 cm. Drying was affected by the humidity and temperature of the air. Air backwashing could be used to accelerate drying speed. After five days, the emptying valve was closed and the inlet valve was opened. After some water flowed into the carbon filter beds, the manual backwashing process was started. The backwashing procedure was: 13.7 L/ s air pressure for 2 min, combined air/water for 2 min, 8 L/s drinking water for 7 min. After the manual backwashing was finished, an automatic backwashing process was started every two days. After 22 days, the backwashing process cycle was adjusted to a 3-day cycle. Water samples and carbon samples were collected before drying (0.5 h), in the initial day after drying was finished (0.5 h, 1 h, 2 h, 4 h, 24 h), and in the stable period after drying was finished (two days later, every 2e5 days). Sampling, processing, identification and counting methods for water and carbon material samples were the same with that described in Section 2.2 for water and carbon samples.
2.6.
Soaking carbon filter beds with salt water
Soaking carbon filter beds with salt water can reduce invertebrates in a short time in case of emergency. After the water in the filter beds was drained, crystalline NaCl salt was spread evenly over the surface of the filter beds (about 50 kg/ m2). Water was added to filter beds by the backwashing pipe to a depth of 10 cm from the surface of the carbon beds. Air backwashing was implemented to make sure the salt was fully dissolved and uniformly distributed over the filter beds. Conductivity in the water of the beds and filtrate was monitored. The final concentration of the salt solution was about 20 g/L. During the immersing period, air backwashing was implemented for 3 min every 3 h to suspend the filter material and mix it thoroughly with the salt water, which ensured that the invertebrates were fully exposed to the salt solution. No salt water flowed out of the filter beds during air backwashing. The survival of invertebrates was determined in the surface, middle and bottom of the beds every 4 h. After 24 h, the salt solution was drained and water was added to a depth of 10e20 cm from the carbon surface. Air backwashing (3 min) and water backwashing (10 min, 2000 m3/h for flow) were carried out twice, after which the water in the filter beds was drained. Backwashing (2000 m3/h flow) was applied again for 60 min until the conductivity in the head and bottom water of the filter beds recovered to normal (150 ms/cm) that is, the same as before soaking. The filter beds were allowed to operate as usual. Before soaking carbon filter beds with salt water, water backwashing was
219
carried out. The first water samples were collected from backwashing water after it had run for 15 min. After 24 h, the salt solution was drained and water backwashing started. The other water samples were collected from backwashing water after it had run for 15, 90 and 120 min, respectively, in order to evaluate living crustaceans and oligochaetes in the filter beds. Sampling, processing, identification and counting methods for water and carbon material samples were the same with that described in Section 2.2 for water and carbon samples.
2.7.
Water quality parameters
Water quality parameters such as water temperature, turbidity, pH, ozone, dissolved oxygen were the real-time data monitored by online instrumentation (HACH Water Online Analysis System). The potassium permanganate index (CODMn) was determined according to the state standard method (Ministry of Environmental Protection of the People’s Republic of China, 2002).
2.8.
Statistical analyses
The monthly mean abundances of invertebrates were calculated as arithmetic means with standard deviation (SD). The one-sample KolmogoroveSmirnov test of nonparametric test was used to test data distribution of mean abundances. The independent samples t-test was used for comparing the monthly mean abundances of invertebrate in the inlet water and filtrate. Pearson correlations between water quality parameters and log-transformed mean abundance of invertebrate were used to identify significant correlation of the variables with abundances. Regression analysis was used between the identified variables and the mean abundance of invertebrate by a curve estimation of regression. The specific model was selected based on coefficient of determination (R2), the standard error of the estimate, the F value and the significance (P) obtained by curve estimation. Considering the relatively large interval of measurements (a month), we believe that the theoretical possibility of autocorrelation in the time series data is negligible. Statistical tests were performed with SPSS 16.0 and figures were drawn with Origin 7.0.
3.
Results and discussion
3.1.
Water quality
During sampling, water temperature ranged between 14.3 C and 30.7 C. On most sampling occasions, the water temperature was above 20 C. The annual mean water temperature was 23.1 C. Turbidity showed little variation and was generally below 0.1NTU, which was much less than the WHO guideline (1 NTU) (WHO, 2011). Through the procedures of preozone contact tank and primary ozone contact tank, the concentration of ozone was in the range between 0.200 and 0.400 ppm in the inlet water. Dissolved oxygen concentration was adequate for organisms in the GAC filters because of ozone decomposition. The CODMn was relatively low (Table 1).
220
w a t e r r e s e a r c h 5 1 ( 2 0 1 4 ) 2 1 6 e2 2 7
Table 1 e Water quality parameters in the water examined. Parameter Water temperature Turbidity (NTU) pH pH Ozone (ppm) Dissolved oxygen (mg/L) Dissolved oxygen (mg/L) CODMn (mg/L) CODMn (mg/L)
Range
Mean SD
Site
14.3e30.7 0.053e0.102 7.41e8.01 6.95e7.66 0.200e0.400 10e12 5e8 1.20e2.25 0.75e1.65
23.1 5.23 0.0762 0.0135 7.68 0.18 7.28 0.21
1 1 1 8 1 1 8 1 8
6 1.55 1.05
CODMn: chemical oxygen demand (potassium permanganate index). Site 1 represents total inlet water to the GAC filter beds and site 8 represents total filtrate water from the GAC filter beds.
Fig. 2 e Comparisons among mean (±SD) abundances of invertebrates in the inlet water, filtrate and GAC filter beds from October 2010 to December 2011.
3.2. Species composition of invertebrates in the inlet water and filtrate During the survey period, 26 kinds of invertebrates in 3 phyla, 5 classes, 6 orders, 6 families, and 8 genera were observed. The dominant groups were rotifers (20 species), then copepods, both in GAC material and water samples. Most of the rotifers belonged to Lecane, Lepadella, Colurella and Bdelloidea species. Most copepods were Harpacticoida species (e.g. Nitocra pietschmanni) and Cyclopoidea species (e.g. Mesocyclops spp.). In the inlet water, 19 taxa were observed, while in the filtrate, 24 were found. Five Lecane species and Cyclopoida species were only found in the filtrate and cladocerans species were only found in the inlet water. Differences in the dominant species amongst rotifers between the inlet water and filtrate examined were also documented. Colurella spp. and Lecane spp. were dominant in the inlet water, while in the filtrate, only Lecane spp. were dominant. Other invertebrates included oligochaetes (e.g. Nais spp.), nematodes, chironomid larvae and water mites. Schreiber et al. (1997) found that rotifers were the dominant taxonomic group in terms of numbers followed by nematodes which were in general the second important group in the GAC filtrate samples collected by a plankton net of 10 mm mesh width. Nematodes are the most common and abundant colonizers. Castaldelli et al. (2005) reported that they were present in 80% of Italian drinking water treatment plants which treat groundwater or spring water. However, nematodes were in low abundance compared to rotifers and copepods in the present study. Since our experimental protocol targeted invertebrates larger than 35 mm, it is likely that not all invertebrate organisms were extracted from the GAC material and water samples. The species were recorded with specimens in the 35 mm filters.
3.3. Invertebrate abundance in the inlet water and filtrate To check for the general influence of the GAC filters on invertebrate abundance in the water, samples were taken for comparison (Fig. 2). The average abundance of invertebrates found in the total inlet water (inlet water to the eight GAC filters) was 113 79 ind./m3, while the average abundance in
the total filtrate (effluent of the eight GAC filters) was 907 639 ind./m3. The average invertebrate abundance in the filtrate was 8 times of that in the inlet water and the abundances in the filtrate were significantly higher than that in the inlet water (independent Student’s t-test, t ¼ 6.548, P < 0.001). The average invertebrate abundances in the filtrate of GAC 1 and GAC 2 (636 483 ind./m3 and 377 330 ind./m3) were 14.2 and 10.5 times of that in the inlet water (45 19 ind./m3 and 36 17 ind./m3) and were significantly higher (independent Student’s t-test; t ¼ 8.081, P < 0.001, GAC 1; t ¼ 6.952, P < 0.001, GAC 2), respectively. The average invertebrate abundances in the filtrate and inlet water of GAC 1 were slightly higher than that of GAC 2, but there were no significant difference between the two GAC filters. Comparing the invertebrate abundances found in inlet water samples with those found in the filtrate, there was a striking increase in the GAC filters, showing their importance for invertebrate multiplication. This underlines the importance of metazoans in the biological processes of GAC filters. The accumulating biomass within the filters, of which the output of invertebrates was an expression, might give rise to a more complex biocoenosis (Schreiber et al., 1997). The overall changes in the invertebrate output with the water temperature were investigated in the two pre-filtered and filtrate waters (Fig. 3). The invertebrate abundances in spring and summer were usually higher than those in autumn and winter. Invertebrate abundances in the process exhibited a high level in the summer with the maximum abundance of 275 ind./m3 and 1920 ind./m3 in the total GAC inlet water and filtrate, respectively. Seasonal changes in invertebrate abundances in the filtrate of GAC (total, GAC 1 and GAC 2) were considerably larger than those in the inlet water. The invertebrate abundances in the filtrate of GAC (total, GAC 1 and GAC 2) ranged 1300e1920 ind./m3 in the warm season, while those in the inlet water were mostly under 200 ind./m3. Rotifers dominated the taxonomic group in terms of abundance followed by copepods, accounting for w73% and w26% of the total invertebrate number in the total inlet waters and for w60% and w35% released in the total filtrate, respectively (Fig. 4). Abundance of other invertebrates, such as oligochaetes and nematodes, was relatively low. Colurella spp.
w a t e r r e s e a r c h 5 1 ( 2 0 1 4 ) 2 1 6 e2 2 7
221
Fig. 3 e Seasonal changes in invertebrate abundances (monthly means ± SD, n [ 4) in (A) total inlet water and total filtrate for eight GAC carbon filters; (B) GAC carbon filters 1 and 2; (C) inlet water and filtrate of GAC carbon filter 1; (D) inlet water and filtrate of GAC carbon filter 2 from October 2010 to December 2011.
(44% in the filter 1, 38% in the filter 2) and Lecane spp. (25% in the filter 1, 33% in the filter 2) were the most common rotifers in the inlet water samples. Lecane spp. (72% in the filter 1, and 73% in the filter 2) dominated in the filtrate while Colurella spp. just accounted for 16% and 15% in the filter 1 and filter 2, respectively. Other benthic rotifers such as Lepadella spp. (10% in the inlet water, 5% in the filtrate) and Bdelloidea (4% in the inlet water and 3% in the filtrate) were found to a much lesser extent. In another study, Colurella sp. and Bdelloidea dominated the rotifer fauna, and other rotifers like Lepadella sp., Dicranophoridae and Lecane sp. were common (Schreiber et al., 1997).
Regression analysis showed that mean abundance of rotifers and copepods in the filtrate were significantly impacted by water temperature (log Y ¼ 0.369 þ 0.233X 0.004X2, R2 ¼ 0.486, P ¼ 0.019, rotifers; log Y ¼ 2.662 0.171X þ 0.006X2, R2 ¼ 0.682, P ¼ 0.001, copepods). A change in the frequency of the dominant group in the filtrate was detected during the study period (Fig. 4). The percentage of copepods increased in the summer, especially in the total filtrate where the percentage exceeded that of rotifers. Rotifers accounted for 51e100% of total invertebrate abundance in the total inlet water, while 5e98% in the total filtrate. From June to September, the percentage of rotifers was less than 50%, and
Fig. 4 e Seasonal changes in the abundances of rotifers, copepods and proportion in the total inlet water and total filtrate. (A) Abundance in the total inlet water, stacked bar diagrams; (B) abundance in the total filtrate, stacked bar diagrams; (C) percentage in the total inlet water; and (D) percentage in the total filtrate.
222
w a t e r r e s e a r c h 5 1 ( 2 0 1 4 ) 2 1 6 e2 2 7
Fig. 5 e The vertical distribution of invertebrates (monthly means ± SD, n [ 4) from the GAC carbon (ind./kg) in the filter bed 1 (site 4) and filter bed 2 (site 5).
the percentage of copepods was more than 50%. During the period, the mean water temperature was 29.3 1.40 C, while during the other time of the study period, the mean water temperature was 20.9 4.14 C. In the summer, in order to ensure the water quality parameters under consideration and to decrease invertebrate abundance in the filter beds, the backwash cycle is carried out once in every 1e2 days, while in the other times of the year, the backwash cycle is carried out once in every 3e4 days. Overall, from spring to summer, the percentage of rotifers out of the total invertebrate abundance gradually diminished, while the percentage of copepods gradually increased in the same time period. Looking at the relative frequency of rotifers and copepods in the filtrate samples, there was a tendency of copepods to outcompete the rotifers as water temperature increased. The GAC filters increased the relative frequency of copepods as water temperature increased.
3.4.
Invertebrate abundance in the GAC filter beds
Invertebrates were distributed throughout the entire depth of the filter bed. The monthly average concentrations of invertebrates over time in the upper, middle and bottom parts of the filter beds (ind./kg) from GAC 1 and GAC 2 are shown in Fig. 5. The mean invertebrate abundances were comparable in both filter beds: 1021, 1189 and 572 ind./kg in the upper, middle and bottom parts of the GAC 1 and 927, 990 and 462 ind./kg in the GAC 2, respectively. Invertebrate abundance was higher in the summer and reduced when the carbon layer depth increased. The abundances in the upper (0e40 cm) and middle (70e110 cm) parts of GAC were significantly higher than that in the bottom (140e180 cm) (independent Student’s t-test, P < 0.01) while there were no significant differences in abundances between the upper and middle layers. Most aquatic invertebrates have little potential for passing through the entire modern water treatment system. However, a small proportion can survive chemical treatment and colonize the treatment media used in rapid gravity filters and granular activated carbon (GAC) filters, occasionally leading to
an efflux of these organisms into distribution systems, where there can be enough organic matter to sustain populations (Weeks et al., 2007; Pre´vost et al., 1998). Regression analysis showed that water temperature had a significant effect on invertebrate abundance in the water and filter beds (Table 2). The water temperature is above 20 C most of the year and there are enough eggs and larvae. Once food source is sufficient, invertebrates multiply and abundance increases fast. In the warm season, bacteria growth is also fast and the concentration of organic matter increases. Previous studies suggested that water temperature was an important factor influencing bacteria regrowth and that the concentration of organic substrate in the distribution system (Niquette et al., 2001; Francisque et al., 2009; Shaddock, 2005), may be the main food source for invertebrates in the filter beds. Furthermore, water temperature influences invertebrate growth and reproduction (Shaddock, 2005). Higher temperatures support faster growth rates and enable some biota to attain significant populations (Chapman, 1996). Although 62.7e92.9% of the variability of mean abundance in most sites could be statistically explained by differences in water temperature (Table 2), there is a need for further study of other factors. The other water quality parameters under study did not have an observable impact on invertebrate abundance (probably due to their relatively limited variability). Therefore, other factors such as backwashing and flow rate should be taken into account in future studies.
3.5. Control technologies and emergency measures in waterworks Since the occurrence of invertebrates in drinking water can cause complaints from consumers and not all critical aspects concerning the microbiological safety of drinkable water can be dismissed, an extensive removal of invertebrates by treatment processes should be achieved (Castaldelli et al., 2005). Starvation and flushing were used as control measures to inhibit invertebrate development in the drinking distribution system (Van Lieverloo et al., 1997, 1998). Various processes in
223
w a t e r r e s e a r c h 5 1 ( 2 0 1 4 ) 2 1 6 e2 2 7
Table 2 e Regression analysis between mean abundance of invertebrate (Y) and water temperature (X). Invertebrate abundance (ind./m3)/(ind./kg) Total inlet water Total filtrate Inlet water of GAC 1 Inlet water of GAC 2 Filtrate of GAC 1 Filtrate of GAC 2 Filter beds of GAC 1 Filter beds of GAC 2
Site
Model N
1 8 2 3 6 7 4 5
15 15 15 15 15 15 15 15
2
R
0.929 0.64 0.682 0.897 0.636 0.296 0.627 0.770
Equation
Residuals statistics SE
log Y ¼ 0.292 þ 0.07X log Y ¼ 1.422 þ 0.061X log Y ¼ 1.918 þ 0.279X 0.005X2 log Y ¼ 3.151 þ 0.372X 0.007X2 log Y ¼ 0.951 þ 0.07X log Y ¼ 1.364 þ 0.044X log Y ¼ 3.509 0.09X þ 0.003X2 log Y ¼ 2.993 0.048X þ 0.002X2
0.104 0.248 0.156 0.110 0.288 0.370 0.149 0.095
F 171 23.1 12.9 52.4 22.7 5.46 10.1 20.1
Sig. b
