A biomechanical filter for treating fish-pond effluents - Science Direct

ELSEVIER

Aquaculture 152 (1997) 103-117

A biomechanical

filter for treating fish-pond effluents

Muki Shpigel a,* , Avital Gasith b, Eitan Kimmel ’ a The Nalional Centerfor

Man’culture, Israel Oceanographic and Limnological Research, P.O. Box 1212, Eilat 88112, Israel b Institute for Nature Conservation Research, George S. Wise Faculty of Life Sciences, Tel-Aviv Uniuersity, Ramat Avis 69978, Israel ’ Agricultural Engineering Department, Technion, Haifa 32000, Israel

Accepted 8 December 1996

Abstract Cultured bivalves, Crassostrea gigas and Tapes philippinarum effkiently removed particulate matter from fish-pond effluents under two hydrological regimes. Two reactor types, a Plug Plow Reactor and a Continuous Stirred Plow Reactor were tested. Under the experimental conditions, the Plug Plow Reactor was found to be more efficient in depuring aquaculture effluents. A mixture of juvenile bivalves of both species further increased treatment effkiency. Flow rate and reactor length influence vertical settling of the non-planktonic particles. A mathematical model is proposed to predict particle removal by the bivalves under the conditions of the two rector types. 0 1997 Elsevier Science B.V. Keywords:

Biotiltration; Bivalve; Reactor; Treatment; Effluents; Mariculture

1. Introduction Effluents from intensive mariculture may have a deleterious environmental impact by enriching littoral waters with particulate and dissolved wastes (Banner, 1974; Tomascik and Sander, 198.5; Bell et al., 1989). Current techniques for reducing effluent particulate matter (PM) involve mechanical removal by sedimentation and microsieves (Makinen et al., 1988; Gowen et al., 1989; Bergheim et al., 1991; Cripps, 1991). However, removal of PM by sedimentation has proved to be relatively inefficient (Henderson and Bromage,

* Corresponding author. 0044-8486/97/$17.00

0 1997 Elsevier Science B.V. All rights reserved.

PII SOO44-8486(97)00004-5

104

M. Shpigel et al./Aquaculture

152 (1997) 103-117

1988; Hennessy, 1991) while microsieves are more efficient but expensive and require frequent maintenance (Heerfordt, 1991; Hennessy, 1991). The use of suspension feeders (e.g. bivalves) and seaweeds in bio-filtration systems (Goldman et al., 1974; Cohen and Neori, 1991; De-Pauw and Salomoni, 1991; Rosenthal, 1991; Shpigel et al., 1993) can provide an inexpensive option for the biological removal of particulate matter and dissolved nutrients from effluent water. Bivalves improve water quality by removing particles through filter feeding and by increasing particle sedimentation by production of feces and pseudofeces which is the indirect result of removing particles by filtration (Mariojouls and Kusuki, 1987). Both processes reduce turbidity (Tenore and Dunston, 1973; Kasprazak, 1986; Laurenstein, 1986). At high densities, bivalves have been shown to be the main factor controlling seston concentration in natural waters (Haven and Morales-Alamo, 1970; Winter, 1978; Jorgensen, 1990). The concept of developing an ‘environmentally clean’ aquaculture in Israel, based on an integrated fish-mollusk-macroalgal system, was first proposed by Gordin et al. (1981) and later tested at the National Center for Mariculture (NCM) in Eilat (Gordin et al., 1990; Shpigel and Blaylock, 1991; Shpigel et al., 1993). Integrated systems, combining fish and macroalgae (McDonald, 1987) or fish, shrimp and oysters in either land-based (Wang, 1990) or off-shore facilities (Jones and Iwama, 1991) were developed elsewhere, but were primarily aimed at increasing mariculture production rather than improving water quality. The main goal of this study was to test the bio-mechanical filtration concept for treating fish-pond effluents prior to discharge. The treatment process was based on the different and complementary bio-filtration of particulate matter by two species of marine bivalves (Crassostrea gigas, Thunberg and Tapes philippinarum, Adams & Reeve) and by mechanical sedimentation. The study aimed at assessing filtration efficiency of the bivalves in two reactor designs with different flow patterns, a Plug Plow Reactor (PFR) and Continuous Stirred Plow Reactor (CSFR). These reactor types have been successfully used for treating large volumes of waste water (Metcalf and Eddy, 1983). The rationale for testing these types of reactors was based on the fact that the filtration rates and efficiencies of bivalves are species-specific and vary with flow rate and particle concentration.

2. Materials and methods The study was carried out at the National Israel, on the Red Sea.

Center for Mariculture

(NCM) in Eilat,

2.1. Water and e&fluent source Unfiltered sea water from the Gulf PVC-lined fish ponds (100 m 2, 1 m depth of 20 m. The flow rate provided volume. All ponds were stocked with

of Eilat (Aqaba) was pumped into three octagonal depth) from an inlet located 300 m offshore at a a daily water exchange of 150-200% of the pond gilt-head seabream (Sparus aurata L.) which were

hf. Shpigel et al./Aquaculture

152 (1997) 103-117

105

fed daily on a pelleted diet containing 40% protein. The effluents from the fish ponds were drained into two 300-m’ earthen, settling ponds. The water in the settling ponds contained suspended materials such as fine sand, silt, as well as detrital, algal and bacterial particles usually up to 150 Frn in size. The water was pumped from the settling ponds into an 11 m3 header tank which supplied the water to the experimental reactors. 2.2. Experimental

reactors

The PFR is a constant flow-through system in which the flow is in a ‘plug’ form (i.e. with minimal longitudinal dispersion). The filtration activity of the bivalves is expected to vary as a result of the change in concentration of the PM along the longitudinal axis of the reactor. The CSFR is a flow-through system in which the water is completely mixed. In an ideal CSFR the PM is continuously redistributed and the concentration is uniform. Seven rectangular troughs (15 cm wide, 80 cm long, 12 cm high), each containing 14.4 1 of water, were used as experimental reactors. In the PFR the water entered at the head of the reactor and exited at the opposite end via a vertical stand pipe. In the CSFR the water entered through eight inlets along one side of the trough and exited through eight outlets on the opposite side (Fig. 1). In the CSFR, aeration mixed the water to maintain uniform conditions. T. philippinarum and C. gigas were used separately and in combination in the experiments. The bivalves were held in the reactors in plastic mesh trays arranged in a series of four stacks of three trays each. The total biomass was 1 kg, 3-7 g individuals per reactor.

Fig. 1. Schematic diagram of the Plug Plow Reactor and Continuous Stirred Flow Reactor containing stacks of trays with bivalves. The inlet, outlet and three ports represent five sampling locations and arrows indicate direction of flow.

M. Shpigel et al. /Aquaculture

106

152 (1997) 103-117

2.3. Measured parameters The total amount of PM removed by filtration, feces, pseudofeces and physical settling, defined as total particulate removal (TPR), was determined by the differences in turbidity levels (NTU) between the incoming and the outflowing water of the experimental reactors (using a turbidity meter, HACH model XR). Significant linear correlation between turbidity levels and total PM in the sedimentation pond water was shown in an earlier study (P < 0.01; r2 = 0.997; Y = 0.21 X + 1.72; Shpigel, 1994) Filtration rate (FR), which represents the amount of particles removed per unit time by the bivalves alone, was determined by the differences in chlorophyll-a levels between the inflowing and outflowing water (using a Perkin Elmer luminescence spectrometer, model LS 5B). The TPR and FR were calculated as follows: TPR or FR = Q X (C,, - Co,,) where Q is the flow rate (1 h- ’ ), and C, and C,,,, are turbidity or chlorophyll-a the inlet and outlet water, respectively. TPR or FR removal efficiency (%) was calculated as follows: E TPR

Or EFR

2.4. Experimental

=

loo

x

(I

-

levels in

ccmt/cin)

design

2.4.1. Experiment I To examine the flow pattern in the reactors, the flow rate was adjusted to 100 1 h-’ in both reactor types. An amount of 0.5 ml of fluorescent dye was injected simultaneously into the reactors. The turbidity level of the outflowing water was measured every 30 s, over a 15 min period (six replicates for each reactor type). 2.4.2. Experiment 2 The relation between particle and vertical settling of seston from the settling pond was assessed under quiescent conditions. Samples of water with sorted particles (filtered through 150, 64, 45, 25, and 15 t.r,rn nets) were placed in vials and the change in turbidity was measured every 5 s, over 10 min periods. 2.4.3. Experiment 3 The effect of flow rate (20, 40 and 60 1 h- ’ or velocity of 1.1, 3.7 and 6.2 cm min- ‘, respectively) on the settling of PM in tbe absence of bivalves, was tested in both reactor types. Settling rates were evaluated by measuring the change in turbidity and chlorophyll-a levels of the water entering and leaving the reactors. Samples were taken every 10 min over 180-min periods (six replicates for each flow rate). 2.4.4. Experiment 4 Total particulate removal by either species of bivalve was assessed, in both reactor types. A total biomass of 1 kg of bivalves (C. gigas of 7 g and T. philippinarum of 3 g average individual whole wet weight), was exposed to incoming pond water with 48


152 (1997) 103-117

107

NTU and 20 NTU turbidity levels (corresponding to 35 and 11 mg 1-l chlorophyll-a, respectively). Flow was kept at 40 1 hh ’ . The bivalves were acclimated to the experimental conditions in the reactors for 30 h prior to experimentation. To avoid measuring instantaneous filtration rates, the outflowing water was collected into a reservoir and sampled (in triplicate) at 2-h intervals over a 6-h period. The turbidity and chlorophyll-a levels of the outflow were measured and compared with those of the incoming water (six replicates of each experiment were carried out). 2.4.5. Experiment 5 The pattern of distribution of the PM concentration in each reactor type was determined by sampling the water from five locations along the reactor as illustrated in Fig. 1. Flow was maintained at 40 1 hh’ and the turbidity levels were measured every 15 min for 2 h. C. gigas or T. philippinarum (1 kg total biomass each, 7 g and 3 g average wet weight, respectively) and a control containing only 1 kg bivalve shells were tested in both reactor types. The bivalves were acclimated to the experimental conditions as described above (experiment 4; six replicates of this experiment were carried out). 2.4.6. Experiment 6 Total particle removal efficiency (ETPR > for each bivalve species, and for the two together, was examined in the plug flow reactor. The initial stocking biomass for each species was 500 g (0.035 kg 1-l) and 250 g of each species when they were tested together. The average wet weight of each bivalve was 5 g. The flow rate was kept at 40 1 hh’. Turbidity and chlorophyll-a levels were measured in the inflowing and outflowing water as describe above (experiment 4). This experiment was repeated six times. 2.4.7. Experiment 7 The effect of bivalve size on TPR was examined using two size classes of C. gigas (5-10 and 20-25 g) and a mixture of both sizes. TPR was tested in a PFR at flow rate of 40 1 h-‘. The total bivalve biomass of each size group was 500 g, and 250 g of each size group when they were combined. The same sizes and combinations of C. gigas shells were used as controls to measure the physical effects of the bivalves on the settling rate. Six replicates were carried out. ANOVA general linear model (Systat) for unequal sample size was used to compare the performance of the reactor type and/or the bivalves. ANCOVA was used to compare the relation between inflow flux and total particulate removal in both reactor types.

3. Results The dye experiment proved that in the CSFR the incoming water was thoroughly mixed within seconds. The turbidity level increased immediately following spiking of the dye and slowly decreased thereafter. In the PPR the incoming water moved as a front (plug) and the turbidity increased and decreased abruptly (Fig. 2).

M. Shpigel et al./Aqwculture

108

0 Fig. 2. Comparison

1

2

3

” 4

of turbidity dynamics

/ ’ ” 5 6

”

152 (1997) 103-117

”

’ ”

7 8 9 TIME (min)

”

’ ”

”

’ ’

10 11 12 13 15

of sea water spiked with a dye in a PPR and a CSFR.

Under quiescent conditions, the vertical settling rates of PM from a sedimentation pond were positively correlated with particle size (Fig. 3, r2 = 0.94; logy = 0.0481og X - 0.065). The PM in the water containing a particle spectrum up to 150 p,rn exhibited a relatively high average settling rate (0.17 NTU s-‘) with 30% of the total PM remaining after 3 min. The PM of small particles (less than 15 pm) settled at about a third of the above rate (0.05 NTU s-l) and 80% of these particles remained after 3 min. In the first case, the PM consisted mostly of sand and silt particles whereas in the latter case it was mostly algae. In the experimental reactor, the flow rate of 40 1 h-’ corresponds to velocity of 3.7 cm min- I. At this velocity, most of the heavy particles settled within 11 cm from the inlet, retention time in the reactor was 21 min.

?3 0.16 -

r ‘= 0.94 logY=o.o481ogx-o.o65

Fig. 3. Relationship

between sedimentation

rate and particle size (mean + SD).


20 I/h

.

109

152 (1997) 103-117

40 I/h

60 I/h

40 I/h

60 I/h

PPRU CPM

Fig. 4. Comparison of turbidity (A) and chlorophyll-a (B) removal (mean&SD) CSFR containing bivalve shells (n = 6 replicates for each treatment).

at three flow rates in PFR and

In the presence of bivalve shells (controls), similar settling rates of PM were observed in the two reactor types at different flow rates (Fig. 4(A)). At all three flow rates 18-22% of the PM settled. The chlorophyll-a levels dropped by lo-16% in the two reactor types (Fig. 4(B)). Comparison of the performances of the two reactor types with the presence of either bivalve species at either relatively high (48 NTU) and low (15-20 NTU) turbidities ranged between 83 and 97% in the PFR for both (experiment 4) showed that E,,, bivalve species and at both turbidity levels (Table 1). Turbidity reduction by T. philippinarum in the PFR was 88% at the low tested level and 97% at the high turbidity level. In the CSFR it was 69 and 84%, respectively. Chlorophyll-a removal in the PFR was 83% at the low turbidity level and 97% at the high turbidity. In the CSFR it was 57% and 87%, respectively. Turbidity reduction by C. gigas in the PFR was 83% at the low tested level and 91% at the high turbidity. In the CSFR it was 75-79%, respectively. Chlorophyll-a removal in the PFR was 86% at the low level and 93% at the high turbidity. It was 79% and 69% in the CSPR, respectively. Comparison of the pattern of distribution of PM concentration along each reactor type


110

152 (1997) IO3-117

Table 1 Comparison of PM removal (NTU and chlorophyll-a; mean+SD) by T. philippinarum (3 g average wet weight) and C. gigas (7 g average wet weight) in two reactor types and high and low levels (flow rate 40 1 h- ’ , n = 6 replicates for each treatment) Species

Reactor

Inlet

Outlet

Initial concentration

% removal

Turbidity (NTIJ)

Chl-a (kg 1-l )

Turbidity (NTU)

ChlLa (kg l- ‘1

Bivalve Shells

PFR CSFR

61 52.4 61 k2.4

27k2.5 27 + 2.5

25k2.5 23+ 1.9

11+1.5 15+ 1.3

Tapes philippinarum

PFR CSFR PFR CSFR

48+2.3 48 f 2.3 48 jz 2.3 48 * 2.3

35 35 35 35

97 f 2.7 84f 1.4 91+3.7 79 k4.6

97*3.8 87+ 1.4 93k4.5 69+5.6

PFR CSFR PI-X CSFR

20.2+ 20.2+ 15.1+ 15.1+

11+1.7 llk1.7 11+1.7 llf1.7

88 f 3.7 69k4.1 83zk3.7 75k3.1

83 k 3.3 57+4.1 86k4.2 79 f 3.5

Crassosrrea gigas

Tapes philippinarum Crassostrea gigas

1.8 1.8 1.3 1.3

+ 2.5 +2.5 + 2.5 k 2.5

Table 2 Comparison of the PM gradient (NTU units, mean + SD) along the PFR and CSFR using Crassostrea gigas (7 g average wet weight) and Tapes semidecussarus (3 g average wet weight); port order as illustrated in Fig. 1 Inlet 7.8 + 0.5 7.8 + 0.5

Port 1

Port 2

Port 3

Outlet

% removal

6.5rfr0.3 6.6f 1.4

6.4kO.7 6.4+I.O

6.Ikl.3 6.8+ 1.1

6.3ko.6 6.2iO.3

19+2.1 20+ 1.7

PFR CSFR

Shells 1 Shells 2

PFR PFR


9.5+0.6(7.6) 11.4+0.8(9.1)

4.4f0.3 4.8kO.6

3.150.3 3.8kO.4

1.7iO.4 1.6+0.2

1.1*0.2 1.4kO.3

88k3.2 88+3.7

CSFR CSFR


9.5kO.6c7.6) 11.4+_0.8(9.1)

2.8f1.2 2.4k0.3

2.7kO.6 2.6k0.3

2.5kO.3 2.4kO.4

2.3*0.2 2.4kO.5

76+4.1 79+3.5

Table 3 Comparison of PM levels (NTU unit, mean k SD) and chlorophyll-a removal by a mixture of equal biomass of T. philippinarum and C. gigas (5 g average wet weight) and of each species alone (flow rate 40 1 h- ’ , n = 6 replicates per treatment) Species

Inlet

Outlet

Initial concentration

Shells Tapes philippinarum (500 g) Crffssostrea gigas (500 g) Mixed (250 g + 250 g)

% removal

Turbidity (NTU)

Chl.-a (OCRI-’ )

Turbidity (NT@

Chl-a (fin I- ‘)

47k3.8 47k3.8 47i3.8 47k3.8

27k2.5 42k3.7 42 i 3.7 42 + 3.7

23 + 4.3 68 55.7 87+3.3 85+3.1

14+ 2.2 83 +4.9 93*4.1 94+3.7

M. Shpigel et al./Aquaculture Table 4 The effect of bivalve size on PM removal

(NTU, mean+

152 (1997) 103-117

SD)

111

in a PFR (flow rate 40 1 h-‘,

n = 6 replicates

per treatment) Species

Inlet (NTU)

% removal

Bivalve shells Crassostrea gigas (22 + 2.5 g) Crassostrea gigas (7 * 0.5 g) Mixed (1:l)

14.1 L-o.3 14.1*0.3 14.1+0.3 14.1*0.3

2Ok2.2 52+3.7 64k4.6 66 + 6.3

(experiment 5) in the presence of only shells showed that the turbidity reduction in both reactors was the same (19% and 20% reduction, P > 0.05, Table 2). In the PFR 17% of the PM settled between the inlet and the first bivalve batch (port l), whereas at the CSFR settling was similar throughout the reactor. A considerable difference was observed in the dynamics of PM along the two reactor types in the presence of either bivalve species. Both bivalve species produced a decreasing turbidity gradient along the PFR. A sharp drop in turbidity (55%) was observed between the incoming water and after the first batch of bivalves at port 1. Thereafter, the turbidity gradually decreased with increasing distance from the inlet, and the clearance efficiency increased. Turbidity reduction measured between the first stacks of bivalves (port 1) and the second stack (port 2) was 14-20%, whereas between the second and third stack (port 3) it was above 50%. The overall turbidity reduction was 88% (from 9.5 to 1.1 and from 11.4 to 1.4 NTU, for T.philippinarum and C. gigas, respectively; Table 2). In the CSFR, no longitudinal turbidity gradient was detected and the overall turbidity reduction was less than 80% (Table 2). In the PFR, a mixture of equal biomass of the two bivalve species (5 g individual average wet weight) filtered as well as an equal biomass of C. gigas alone (85% and 87%, respectively, P > 0.05; Table 3). The efficiency of T. philippinarum alone, was significantly lower (68%, P < 0.01; Table 3). The same pattern but with higher efficiencies, was observed for chlorophyll-a removal (94%, 93% and 83%, respectively). A total biomass of 500 g of small C. gigas (7 g wet weight) or equal biomass (250 g + 250 g) of a mixture of small and large (22 g wet weight) individuals reduced turbidity by three times more than shells (control) of similar size alone (64% and 52%, respectively). TPR of the small oysters was significantly higher than that of large C. gigas by themselves (52%, P < 0.01; Table 4).

4. Discussion The overall efficiency of the PFR was found to be superior by lo-20% to the CSFR in removing total particulate matter as calculated by either PM (turbidity) or algae (chlorophyll-a) reduction. Shells only did not affect the physical settling of the larger and heavier particles in the two reactors. The difference in the performance of the reactors can thus be attributed to the difference in the biofiltration of smaller particles, mostly algae, under different hydrological regimes.

112


I52 (1997) 103-117

The PFR produces a decreasing concentration of particles with the increasing of the distance from the inlet. About 54% of the PM was removed over the first stack of the bivalves. This is due to physical settling of heavy particles (20%), filtration and probably feces and pseudofeces production by the bivalves (34%). As the water flowed over the remaing three stacks an additional 34% of the initial PM was removed. In the CSFR the removal efficiency was uniform over the entire reactor and was lower overall in comparison with the PFR. On the basis of equal total biomass and equal individual wet weight, C. gigas was consistently more efficient in TPR than T. philippinarum. This is consistent with the findings of Tenore and Dunston (1973) and Walne (1976), who showed that at all levels of food concentration, clams were less efficient at TPR than oysters. However, it was interesting to note that a mixture of equal biomass of both species produced a higher TPR than expected. This synergistic action may be explained by a complementary removal of different particle sizes by the two species. This supports the results of Winter (1978), Shumway et al. (1985) and Stenton-Dozey and Brown (19921, which showed that there are apparently adaptive differences in the size of fine particles which are retained by different bivalve species. It has been shown that the burrowing clam Mercenuria mercenaria has the highest retention efficiency (90%) for 3.5 p,m particles, whereas oysters (e.g. Crussostrea uirginica) best retain particles of 5 km (Jorgensen, 1990). As shown in the present and other studies, small bivalves with their higher filtration rate are more efficient in particle removal than an equal biomass of larger ones (Foster-Smith, 1975; Winter, 1978; Bayne and Newell, 1983). Our results confirm that the total particulate removal of suspended particles by physical settling and biological filtration can be an inexpensive means for treating fish pond effluents. Since high filtration efficiency does not necessarily imply better growth of the bivalve (Rodhouse and O’Kelly, 1981), optimization of the system for both water treatment and biomass production is limited. Factors such as temperature, oxygen levels, water flow rates, particle concentration, animal size, total biomass per unit area and species-specific characteristics have been shown to affect bivalve filtration rate and efficiency (e.g Haven and Morales-Alamo, 1970; Walne, 1976; Winter, 1978; Spencer, 1988). Flow rate, bivalve size, biomass, and species can be manipulated to optimize filtration efficiency in a commercial reactor operation. In all the tested flux levels (Fig. 51, the PFR seems more suitable for treating aquaculture effluents (ANCOVA, F = 17.49, IZ= 38, P < 0.01). It should be noted that in the present study, particle concentration ranged between 5 and 35 mg DW 1-l) and the relationship between TPR and flux is positively and linearly correlated (Fig. 5). At higher flux level, TPR in one of the reactor type is expected to level out, thus its relative efficiency may change. A mixture of juvenile bivalves of the above species should increase the efficiency of total particle removal. The bivalve biomass in the reactor could be adjusted to maintain the efficiency requirement, depending on temperature and particle concentration of the effluent (see Winter, 1978). Flow rate and reactor length should be adjusted to allow vertical settling of the non-planktonic particles. A commercially desirable treatment unit calls for maximum flux and minimum bivalve biomass. Further investigation into the


152 (1997) 103-117

113

2,000

500

J 0

0

5txJ

1,500

1WJ

2,@Jo

2,500

FLUX (NTU h” kg”) Fig. 5. Relation between inflow flux and total particulate removal rate in the PFR and CSFR containing wet weight of C. gigas (the data include all the results for C. gigas, n = 38).

relationship between flow rate, size distribution istics is needed.

and species-specific

filtration

1 kg

character-

4.1. Particle removal model for bivalve reactors The following is a mathematical model with which we propose to simulate particle removal by bivalves under the conditions of the two reactor types. For simplicity, we assume that all of the settling process (removal of 20% turbidity) occurs near the inlet of the reactors. Therefore, the actual particle concentration (expressed in turbidity units) for the calculation of biofiltration removal and efficiency is 20% smaller than that measured in the inlet water. Hereafter, the Ci, value is 80% of the measured inlet concentration. 4.2. CSFR performance In the CSFR the PM balance

is described

by

Q( Gin- Coat)= KV where Q is the water flow rate (1 h- ’ >, Ci, and C,,, (NTU) are the turbidity levels in the inlet and outlet water, respectively, and tK is the specific rate of particle removal by the bivalves, expressed in units of NTU h- ‘. This value represent the activity of 1 kg of biomass uniformly distributed in a tank with volume V = 14.41. The relationship between the specific particle removal rate (K) and the particle concentration is described by K = K,C”

(2)

114


152 (1997) 103-117

where K, is the removal rate when the concentration is 1 NTU, and the species-specific activity, m, can be estimated by curve fitting of the experimental data. Accordingly, when m > 0, bivalve particle removal increases PM concentration. In this reactor the PM concentration is uniform and equal to the PM concentration leaving the reactor ((Tout>, thus

Q(Cin-Cow> =KoC,mUt

(3)

Bivalve removal efficiency ( E,,) sides of the equation by Q X Ci, 1 - Co”,/Ci” = By dividing

K,C,",tV/QCn

the right side of the equation

4~a/lo’3 if g=VCE-

can then be derived from Eq. (3) by dividing both

j

1 - Cout/Cin =

by Cg/Cz

( CJCi”) mVCE-'/Q

l/Q

ETPR/~OCJ * 1 - CoJCi, g = K,VC;-

‘/Q = K,V/(

= g(CJCin)m

(4)

Ci,Q)

(5)

g is a parameter that integrates the physical (e.g. volume, flow-rate, inlet particle concentration), and species-specific features (e.g. K, and m). It can also be viewed as the ratio between a theoretical filtration capacity of the reactor (K,V), where all the bivalves act at the inlet PM concentration (C,,) and their removal rate is K, = K,C;

(6)

and the influx of particles (Ci, Q). g is a useful engineering planning and evaluation of the reactor performance.

parameter

which enables

4.3. PFR petiormance In the plug-flow of length dr

reactor the PM concentration

equation

can be written for a segment

QdC = K, C”‘Adx

(7)

where dC describes the PM concentration difference over the segment G!X, x is the longitudinal coordinate in the direction of the flow, and A is the cross-section area of the PF reactor (A = V/L, D is the length), perpendicular to X. Integration of Eq. (7) along the reactor, yields CKX m, - C:i-“)=

(1 -m)VK,x/(QZ)

(8)

the PM concentration (CcX,) at distance x from the inlet to the reactor. The outlet concentration and the removal efficiency can be then derived by introducing x = 1 into Eq. (8) ETPR/lOO * 1 - C,,,,/C,, g=l/(l-m)([l-(I-C,,,/C,,)]‘-“-1)

= I - [ 1 - (1 - m) g]‘l’(l-m))

(9) (10)


152 (1997) 103-117

C. gigas

115

PFR rZ- 0.97 Y = 0.83X o.43 CSFR

r& 0.94 Y = 069X

T. Phil. PFF? r g 0.95 Y = 0.55X CSFFI 0

3

1

C. gigas/PFR -I!&

0.99 Y = 0.43x

G

C. gigas/CSFR

TapeslPFR

Tapes/CSFR

‘,3

32

___y___

Fig. 6. Relation between g values and total particulate removal efficiencies containing 1 kg wet weight of Crassostrea gigas and Tapes philippinarum.

’ a13 O.” 4

( EpR)

in PFR and CSFR

The parameters m, K, and Ki, for each bivalve species can be extracted from the measured data of the concentration profiles in the reactor (Table 2) and the removal rates (Table 1). This procedure was repeated for each bivalve species. Accordingly, the K to C relationship (IQ. (2)) for Tapes phzlippinarum is characterized by m = 3/2. This m value is associated with small errors of prediction of less than 6%. For Crassostrea gigas, m is characterized by m = 0.25 with errors of prediction of less than 2%.

Table 5 The relationship between g and the particle removal rate (K, and Kin when the particle concentration is 1 NTU and X NTU, respectively). C, values arc 80% of the measured inlet concentration and m is the species-specific activity which was estimated by curve fitting of the experimental data Crassostrea gigas (7 g)

Tapes philippinarum ci, (NTU) g K, (NTU h-‘) Kin (NTIJ h- ‘)

m = 0.25 9.1 0.95 12.9 47.6

ci, cNTu) g K, (NTU h-‘) K,, (NTu h- ‘) (3 g)

m= 1.5 7.6 3 0.59 104

12.1 1.05 14.4 49.9

16.2 3.2 0.59 325

38.4 1.2 26.6 175.3

38.4 8 0.61 1339

116


I52 (1997) 103-117

The model predicts that PFR performs better than the CSFR for all values of g (Fig. 6). For each bivalve species, the value of g determines the efficiency of particle removal and can be useful tool in the design for optimizing the performance of bivalve reactor. Reactor performance is positively related to g. g can be increased by reducing the flow rate (Q), increasing the reactor volume (V) and increasing bivalve biomass. It should be noted that K, is affected by water quality conditions such as temperature and oxygen levels. These effects were beyond the scope of the present study. Table 5 demonstrates the relationship between g values and species-specific activity under the specified m value shown above in PFR. As demonstrated by the values of K,,, filtration capability of the Tapes philippinarum (3 g average wet weight) are greater than Crussostrea gigas (7 g average wet weight) for all inlet PM concentrations, and for effective Ci, = 38.4 NTU, T. philippinurum are almost seven times more. Among each of the bivalve species, the differences in K, values may be attributed to species-specific activity.

Acknowledgements The research was supported by the Israeli Ministry for Energy and Infrastructure and by the R&D Network Negev-Arava. We thank Rafi Fridman, David Ben-Ezra, Norman Rag and Alison Marshall for their technical assistance, and Drs. J.J. Lee, A. Colomi and A. Neori for the critical review of the manuscript.

References Banner, H.A., 1974. Kaneoha Bay, Hawaii: Urban pollution and a coral reef ecosystem. Proc. 2nd Int. Coral Reef Symposium 2. Great Barrier Reef Committee, Brisban, December, pp. 685-702. Bayne, B.L. and Newell, R.C., 1983. Physiological energetics of marine molluscs. In: K.M. Wilbur and A.S.M. Saleudin (Editors), The Mollusca. Vol. 3. Academic Press, New York, 352 pp. Bell, P.R., Greenfield, F.P.E., Hawker, D. and Connel, D., 1989. The impact of waste discharges on the factors affecting the physical sedimentation in the reactors depend on reactor size, particle size distribution, water flow rate and the presence of the bivalve shells coral reef regions. Water Sci. Technol., 21: 121-130. Bergheim, A., Tyvold, T. and Seymour, E.A., 1991. Effluent loadings and sludge removal from land based salmon farming tanks. Eur. Aquacult. Sot. Spec. Publ., 14: 27. Cohen, I. and Neori, A., 1991. UZua lactuca biofilters for marine fishpond effluent. I. Ammonia uptake kinetics and nitrogen content. Bot. Mar., 34: 475-482. Cripps, S.J., 1991. Comparison of methods for the removal of suspended particles from aquaculture effluent. Eur. Aquacult. Sot. Spec. Publ., 14: 80. De-Pauw, N. and Salomoni, C., 1991. Aquaculture systems for wastewater treatment. Eur. Aquacult. Sot. Spec. Publ., 14: 87-88. Foster-Smith, R.L., 1975. The effect of concentration of suspension on the filtration rates and pseudofaecal production for Mytilus edulis L., Creastodem edule CL.1 and Venerupis pullastra (Montagu). J. Exp. Mar. Biol. Ecol., 17: l-22. Goldman, J.C., Tenore, R.K., Ryther, H.J. and Corwin, N., 1974. Inorganic nitrogen in a combined tertiary treatment-marine aquaculture system-I. removal efficiencies. Water Res., 8: 45-54. Gordin, H., Motzkin, F., Huges-Games, A. and Porter, C., 1981. Seawater mar&culture pond-an integrated system. Eur. Aquacult. Sot. Spec. Publ., 6: 1-13.


152 (1997) 103-117

117

Gordin, H., Krom, M., Neori, A., Popper, D., Porter, C. and Shpigel, M., 1990. Intensive integrated seawater fish ponds: fish growth and water quality. Eur. Aquacult. Sot. Spec. Publ., 11: 45-65. Gowen, R.J., Rosenthal, H., Makinen, T. and Ezzi, I., 1989. Environmental impact of aquaculture activities. Eur. Aquacult. Sot. Spec. Publ., 10: 300. Haven, D.S. and Morales-Alamo, R., 1970. Filtration of particles from suspension by the american oyster Crassostrea uirginica. Biol. Bull. Mar. Biol. Lab. Woods Hole, 139: 248-264. Heerfordt, L., 1991. Test of the efficiency of alternative treatment systems on the effluent from traditional trout farms in Denmark. Eur. Aquacult. Sot. Spec. Publ., 14: 140. Henderson, J.P. and Bromage, N.R., 1988. Optimising then removal of suspended solids from aquacultural effluent in settlement lakes. Aquacult. Eng., 7: 167-181. Hennessy, M., 1991. The efficiency of two aquacultural effluent treatment systems in use in Scotland. Eur. Aquacult. Sot. Spec. Publ., 14: 142. Jones, T.O. and Iwama, G.K., 1991. Polyculture of the pacific oyster Crassostrea gigas (Thunberg), with chinook salmon, Oncorhynchus tshawytscha. Aquaculture, 92: 313-324. Jorgensen, C.B., 1990. Bivalve filter feeding: Hydrodynamics, bioenergetics, physiology and ecology. Olsen and Olsen, Fredensborg, Denmark, 140 pp. Kasprazak, K., 1986., Role of Unionidae and Sphaeriidae (Mollusca, Bivalvia) in the eutrophic Lake Zbechy. Int. Rev. Ges. Hydrobiol., 71: 315-334. Laurenstein, D.D., 1986., Filter feeding in Corbiculaflumineu and its effect on seston removal. J. N. Am. Bentol. Sot., 5: 165-172. Makinen, T., Lindgren, S. and Eskelinen, P., 1988. Sieving as an effluent method for aquaculture. Aquacult. Eng., 7: 367-377. Mariojouls, D. and Kusuki, Y., 1987. Appreciation des quantites de biodepots emis par les huitres en elvage suspcndu dans la baie d’Hiroshima. Haliotis, 16: 221-231. McDonald, M.E., 1987. Biological removal of nutrients from wastcwatcr: an algal-fish system model. In: K.R. Reddy and W.H. Smith (Editors), Aquatic Plants for Waste Water Resource Recover. Magnolia Publishing, pp. 959-968. Metcalf, L. and Eddy, H.P., 1983. Wastewater Engineering: Treatment Disposal Reuse. Tata McGraw-Hill, New Delhi, 919 pp. Rodhouse, P.G. and O’Kelly, M., 1981. Flow requirements of the oysters Ostrea edulis L. and Crussostrea gigus Thunberg in an upwelling column system of culture. Aquaculture, 22: I-10. Rosenthal, H., 1991., Water and waste water treatment. Eur. Aquacult. Sot. Spec. Pub]., 14: 282. Shpigel, M., 1994. Annual Report. Israel Oceanographic and Linological Research Ltd. The National Center for Mariculture, Eilat, Israel, 104 pp. (In Hebrew.) Shpigel, M. and Blaylock, R.A., 1991. The Pacific Oyster, Crassosfrea gigas, as a biological filter for a marine fish aquaculture pond. Aquaculture, 92: 187-197. Shpigel, M., Neori, A., Popper, D.M. and Gordin, H., 1993. A proposed model for ‘clean’ land based polyculture of fish, bivalves and seaweeds. Aquaculture, 117: 115-128. Shumway, S.E., Cucci, T.L., Newell, R.C. and Yentsch, C.M., 1985. Particle selection, ingestion, and absorption in filter feeding bivalves. J. Exp. Mar. Biol. Ecol., 91: 77-92. Spencer, B.E., 1988. Growth and filtration of juvenile oysters in experimental outdoor pumpes upwelling systems. Aquaculture, 75: 139-158. Stenton-Dozey, J.M.E. and Brown, A.C., 1992. Clearance and retentiom efficiency of natural suspended particles by rock-pool bivalve Venerupis corrugatus in relation to tidal availability. Mar. Ecol. Prog. Ser., 82: 175-186. Tenore, K.R. and Dunston, W.M., 1973. Comparison of feeding and biodeposition of three bivalves at different food levels. Mar. Biol., 21: 190-195. Tomascik, T. and Sander, F., 1985. Effects of eutrophication on reef building corals: growth rate and reef building coral Montasfrea annularis. Mar. Biol., 87: 143-155. Walne, P.R., 1976. Experiments on the culture in the sea of the butterfish Venerupis decussatus L. Aquaculture, 8: 37 l-38 1. Wang, J.K., 1990. Managing shrimp pond water to reduce discharge problems. Aquacult. Eng., 9: 61-73. Winter, J.E., 1978. A review on the knowledge of suspension feeding in lamellibranchiate bivalves, with special reference to artificial aquaculture systems. Aquaculture, 13: l-33.