e c o l o g i c a l e n g i n e e r i n g 3 2 ( 2 0 0 7 ) 1–10

available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/ecoleng

Improving multi-soil-layer (MSL) system

remediation of dairy effluent

R. Pattnaik a , R.S. Yost a,∗ , G. Porter a , T. Masunaga b , T. Attanandana c

a Department of Tropical Plant and Soil Sciences, University of Hawai‘i, 3190 Maile Way, 102 St. John, Hawaii 96822, USA
b Faculty of Life and Environmental Science, Shimane University, Matsue 690-8504, Japan
c Department of Soil Science, Faculty of Agriculture, Kasetsart University, Jatujak, Bangkok 10900, Thailand

a r t i c l e i n f o a b s t r a c t

Article history: Dairy effluent disposal is a serious problem in the Hawaiian Islands. Dairies often estab-
Received 23 December 2006 lish multiple settling lagoons to accumulate and store effluent. Occasionally, the overflow
Received in revised form 5 July 2007 of lagoons leads to the transfer of nutrients, such as nitrogen (N) and phosphorus (P), and
Accepted 16 August 2007 other contaminants, to hydrologically associated surface, subsurface, and coastal waters.
This study was conducted to assess the removal of inorganic N and phosphate in dairy
effluent using multi-soil-layer (MSL) systems. Four MSL systems were constructed with two
Keywords: replications of two treatments, which were Perlite and the Leilehua soil. Both materials were
Dairy effluent used separately for forming an aerobic layer in the MSL systems, whereas an anaerobic layer
Dairy wastewater treatment was formed from a mixture of charcoal, sawdust, iron filings and Honouliuli soil. The results
Multi-soil-layer system of this study revealed that the removal of inorganic N was similar for the Leilehua and Perlite
Soil dynamics MSL system, which was 22–93% and 21–96%, respectively. Phosphate removal was higher
Tropical soils in the Leilehua MSL system (64–99%) compared to the Perlite MSL system (9–97%). Addi-
Inorganic nitrogen removal tional aeration increased the removal of phosphate by the Leilehua MSL system. Sucrose
effectiveness application with a constant rate of aeration increased the removal of inorganic N both
Phosphate removal effectiveness in the Leilehua and Perlite MSL systems and increased phosphate removal in the Perlite
MSL system. The study demonstrated that MSL systems have the potential to remove high
percentages of inorganic N and phosphate in dairy effluent enabling reuse of the water.
© 2007 Elsevier B.V. All rights reserved.

1. Introduction ents in Hawaii and in the Continental U.S. (USDA-SCS, 1992).

The dairy industry generates wastewaters characterized by
high concentrations of nutrients, organic contents, and
pathogens (USDA-SCS, 1992). The organic and nutrient con-
tent of dairy wastewaters depends upon the size, lactation,
and diet of the cow. In addition, dairy wastewater composition
is significantly influenced by the wastewater management,
climate, operating conditions, and types of flushing. Table 1
shows the levels of major nutrients in dairy wastewater. The
dairy industry is one of the major sources of waste efflu-
Dairy effluent disposal is a serious problem in Hawaii and
other Pacific Islands (Farrell-Poe, personal communication,
2007). The problem is due to the confined aquifers and lim-
ited availability of water in Pacific Island environments. The
current method used in Hawaii to dispose dairy effluent is
large settling lagoons. Dairies often establish multiple lagoons
to accumulate and store effluent. Occasionally, the lagoons
overflow, leading to the transfer of nutrients, such as nitro-
gen (N) and phosphorus (P), and other contaminants, which
can pollute surface, subsurface, and coastal waters. Effluents

∗
Corresponding author. Tel.: +1 808 956 7066; fax: +1 808 956 3894.
E-mail address: rsyost@hawaii.edu (R.S. Yost).
0925-8574/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.ecoleng.2007.08.006
2 e c o l o g i c a l e n g i n e e r i n g 3 2 ( 2 0 0 7 ) 1–10

Table 1 – Dairy wastewater characteristics (Wright, 1996)

Potential pollutant Biochemical oxygen Nitrogen Phosphorus Volume gallons
source demand (mg L−1 ) (mg kg−1 ) (mg kg−1 ) (100 cows y−1 )

Milking center waste 400–10,000 80–900 25–170 73,000

Silage leachate 12,000–90,000 4400 500 105,000
Barnyard runoff 1000–10,000 50–2500 5–500 80,000
Dairy manure 20,000 5600 900 660,000

high in N and P concentration can cause eutrophication of
the receiving waters, degrading water quality (Smith et al.,
1999). The Environmental Protection Agency (EPA) and State
Department of Health (DOH) have rules and regulations for
the disposal of dairy effluent (Hawaii State Department of
Health Wastewater Branch, 1996). Proper management of dairy
effluent is currently a serious problem in Hawaii, which has
increased the operation costs and reduced profitability of
many island dairies. The inability of many dairy operators
to properly manage the effluent has forced more than 50%
of them to close during the last 10 years (C.N. Lee, personal
communication, 2006).
With rising environmental concerns and tighter gov-
ernmental regulations, managing animal wastes in an
environmentally responsible and economically feasible way
can be a challenge. It is becoming imperative that new ways
of waste treatment be found that reduce excessive nutri-
ents from dairy effluent and yet are efficient and reliable.
Some of the methods of dairy waste treatment include land
application (Caro-Costas et al., 1972; Valencia-Gica et al.,
2004), vegetative filter strips (Ikenberry and Mankin, 2000),
constructed wetlands (Schaafsma et al., 2000), aerobic and
anaerobic processes (Manariotis and Grigoropoulos, 2003), and
bioremediation (Prochaska and Zouboulis, 2003) which have
performed well, but their widespread use is limited because
they are either costly, require regular maintenance, require
large areas of land, or the wastewater must be pre-treated. The
multi-soil-layer (MSL) system is a promising alternative with
potential for reducing contamination associated with dairy
effluent.
The MSL system is a technology that uses natural soil in
a unit to facilitate wastewater treatment (Wakatsuki et al.,
1993). This has been successfully developed in Japan and Thai-
land to treat domestic and restaurant wastewater as well as
polluted river water (Wakatsuki et al., 1993; Luanmanee et
al., 2001). The system reduces levels of inorganic contami-
nants such as nitrate, ammonium, and phosphate, as well
as organic contaminants as measured by high COD (chem-
ical oxygen demand) and BOD (biological oxygen demand).
This is a biphasic layered system that uses locally available
materials such as soil, iron particles, jute or sawdust, char-
coal, and zeolite or alternative materials (Attanandana et al.,
2000; Luanmanee et al., 2001). Two layers that comprise MSL
systems are aerobic and anaerobic. Aerobic layers consist of
zeolite or Perlite alternated with anaerobic layers of soil mix-
ture blocks. The efficiency of the MSL system in purifying
wastewater depends on the relative effectiveness of aerobic
and anaerobic layers (Wakatsuki et al., 1993; Attanandana et
al., 2000). The aerobic layer enhances nitrification, oxidation
and precipitation of mobile ferrous iron to high-surface area
ferric oxide, enhancing phosphorus sorption (Wakatsuki et al.,
1993). In the anaerobic layer of the soil mixture block, nitrate
is transformed into nitrous oxide and nitrogen gas (denitri-
fied) and ferric iron is reduced to the more mobile ferrous
iron, which moves out of the anaerobic layer (Wakatsuki et
al., 1993). Although an appropriate amount and timing of aer-
ation is necessary (Luanmanee et al., 2002), the maintenance
of an MSL system is simple and the effective life of such sys-
tems was estimated to be longer than 10 years (Luanmanee et
al., 2002).
Although various types of wastewater treatments have
been treated successfully using the MSL system in Japan and
Thailand, to-date, no MSL system has been tested or adapted
for the remediation of dairy effluent. In addition, there is
not much information available on the reliability, consistency,
and nutrient removal efficiency of MSL systems. Thus, it is of
interest to determine whether the MSL system can remedi-
ate dairy effluent. This study was conducted to (a) investigate
the potential of the MSL systems in remediating dairy efflu-
ent, (b) compare the removal of inorganic N and phosphate,
between MSL systems with the aerobic layers made from
Leilehua soil or Perlite, and (c) evaluate the effect of aeration
and sucrose additions on inorganic N and phosphate removal
efficiency.

Table 2 – Selected physical properties of Leilehua and Honouliuli soils

Series Clay (<0.002) Silt (0.002–0.05) Sand (0.05–2) Water holding Bulk density Particle
(% of <2 mm (% of <2 mm (% of <2 mm capacity (% of <2 mm (g cm−3 ) density
mineral soil) mineral soil) mineral soil) mineral soil) (g cm−3 )

33 kPa 1500 kPa

Leilehua 58.5 33.6 7.9 n/a 33.2 0.97 2.88

Honouliuli 58.4 34.8 6.8 30.2 22.3 1.31 2.93

Source: Soil Survey Staff (2006).

 e c o l o g i c a l e n g i n e e r i n g 3 2 ( 2 0 0 7 ) 1–10
2. Materials and methods
2.1. Experimental site and design
The experimental site was located in Waianae, latitude 21◦ 27 ,
longitude 158◦ 11 on the west shore of the island of O’ahu,
Hawaii. Average maximum and minimum daily temperatures
of the area are 28 ◦ C (83◦ F) and 16 ◦ C (61◦ F) (Hobo Weather
Station, 2001–2003). The experiment was conducted using
dairy effluent from the third settling lagoon of an effluent
waste management system. Four MSL systems were con-
structed, comprising two treatments with two replications
each, arranged in a completely randomized design (CRD). Per-
lite or Leilehua soil (A horizon, Typic Kanhaplohumult) was
used for the aerobic layer in the two treatments. The anaerobic
layer for both treatments consisted of a mixture of charcoal,
sawdust, iron filings and the Honouliuli soil (A horizon, Typic
Chromustert). The physical and chemical properties of both
the Leilehua and Honouliuli soils used for the experiment are
given in Tables 2 and 3.
2.2. MSL systems and operations
Cross-section composition of the overall MSL systems is pre-
sented in Fig. 1. Each of the MSL system consisted of a
high-density polyethylene (HDPE) corrugated sewage pipe
with 45.7 cm interior diameter by 1 m in height with a cross-
sectional area of approximately 0.1648 m2 (Fig. 1). A 25.4 mm
PVC pipe was installed at the base of each of the upright HDPE
pipe to discharge the MSL treated effluent from the system.
A layer of gravel (≈5 cm) was placed at the bottom of the
upright pipes to facilitate system discharge. Each system was
assembled from seven alternating layers of soil mixture blocks
(anaerobic layers) and eight layers of Leilehua soil or Perlite
(aerobic layers) (Fig. 1). Each of the soil mixture blocks con-
sisted of Honouliuli soil mixed with finely ground charcoal,
fine sawdust, and approximately 1 mm diameter iron filings at
the ratio of 7:1:1:1 by dry weight. The soil mixture was evenly
mixed using an concrete mixer and packed into two sizes of
pre-stitched burlap bags, approximately 5 cm × 10 cm × 22 cm
and 5 cm × 10 cm × 38 cm. The particle sizes of both the Leile-
hua soil and the Perlite filler were less than or equal to 4 mm.
An aeration pipe was installed approximately 50 cm from the
bottom for the subsequent infusion of air whenever it was nec-
Table 3 – Chemical properties of Leilehua and Honouliuli soils
Soil pH OCa (% of
(H2 O, 1:1) <2 mm)
Leilehua 4.8 2.61
Honouliuli 6.9 0.74
Source: Soil Survey Staff (2006).
a
Organic carbon.
b
Total nitrogen.
c
Dithionite-citrate extractable iron.
d
Ammonium oxalate iron.
e
Guo and Yost (1998).
3
essary (Fig. 1). An array of effluent emitters was installed on
the top (≈80 cm from the bottom) of the aerobic and anaer-
obic layers through which the dairy effluent was discharged
into the system (Fig. 1).
2.2.1. Application rates
The dairy effluent was directly pumped from the lagoon, fil-
tered using a 0.0254 m plastic disc filter (140 mesh) to remove
the larger particles, and discharged into the MSL system.
Three application rates of effluent were applied to the sys-
tem according to the performance of the system. An initial
flow rate of 80 L day−1 (505 L m−2 day−1 ) was applied to each
of the system from 18 April to 3 November 2005. Effluent
was applied through drip irrigation emitters during approx-
imately 20 h day−1 . The flow rate was reduced to 40 L day−1
(252 L m−2 day−1 ) on 3 November 2005 and continued until 20
April 2006. Then from 20 April to 10 July 2006 the flow rate was
again reduced to 28 L day−1 (178 L m−2 day−1 ).
2.2.2. Aeration rates
Different aeration rates were applied to the systems. Aera-
tion was not applied to the systems until the 10th month of
the study. The systems were aerated at a rate of 28 L min−1
from 10 February to 13 April 2006. The aeration was increased
to 31 L min−1 from 14 to 27 April 2006. The aeration rate was
decreased to 17 L min−1 for 1 week, from 28 April to 4 May 2006
followed by a rate of 11 L min−1 until 18 May 2006. The aera-
tion rate was increased again to 23 L min−1 from 19 May until
10 July 2006.
2.2.3. Sucrose additions
An additional source of carbon in the form of a sucrose solu-
tion was applied to the MSL system beginning at the end
of the 12th month of the study in attempt to improve sys-
tem performance. The sucrose solution was calculated based
on the amounts needed for a stoichiometric reduction of the
expected oxygen content of the MSL system.
The percent pore space was calculated first from the bulk
densities and particle densities of Leilehua and Honouliuli soil,
and Perlite. Then the amount of air space was calculated from
the volume of each system. The amount of oxygen was calcu-
lated from the amount of air space and the amount of oxygen
in the air. The amount of sucrose was calculated based on
the stoichiometric reaction equation, which shows how much
sucrose is needed for the microorganisms to consume the spe-
TNb (% of Dithionitec Oxalated (% of P sorbede
<2 mm) (% of <2 mm) <2 mm) (mg kg−1 )
0.233 6.4 1.04 1400
0.11 7.5 n/a 100
4 e c o l o g i c a l e n g i n e e r i n g 3 2 ( 2 0 0 7 ) 1–10

Fig. 1 – Cross-sections of the MSL systems (Leilehua and Perlite).

cific amount of oxygen: The calculated concentration of sucrose solution was 19 g

(0.055 moles per 500 mL) for the Leilehua system and 22 g
(0.064 moles 500 mL) for the Perlite system.
C12 H22 O11 + 12O2 → 12CO2 + 11H2 O

2.3. Analytical methods

Finally, the application of sucrose was made as a solution
mixed with the incoming effluent and applied based on the Samples were taken every week except during the period
effluent retention time of the MSL systems. December 2005 to 14 January 2006. After filtering samples
 e c o l o g i c a l e n g i n e e r i n g 3 2 ( 2 0 0 7 ) 1–10 5

Table 4 – Dairy effluent used in this experiment, in comparison with data from other dairy lagoons in Hawaii
Source pH EC TSS TN NH4 + -N NO3 − -N TP IP COD
(mS cm−1 ) (mg L−1 ) (␮g mL−1 ) (␮g mL−1 ) (␮g mL−1 ) (␮g mL−1 ) (␮g mL−1 ) (mg L−1 )

Dairy aa 8.2 6.4 n/a 395.7 183.8 n/a 14.5 n/a n/a
Dairy bb 8.1 3.0 ∼1000 119 108 5.76 16.8 n/a n/a
This experimentc 8.4 3.6 400 44.6 43.16 1.30 21.39 13.21 482
This experimentd 7.77 3.22 n/a 20 17.05 2.86 17.76 5.72 447
This experimente 8.90 3.62 320 5.39f 5.26 0.12 NA 6.08 710

EC: electrical conductivity; TSS: total suspended solid; TN: total nitrogen; TP: total phosphorus; IP: inorganic phosphate; COD: chemical oxygen
demand.
a
Analysis of lagoon effluents from various nutrient streams (Fukumoto et al., 2000).
b
Valencia-Gica et al. (2004).
c
One month before running the experiment, 3 March 2005.
d
The beginning of the experiment, 2 May 2005.
e
The end of the experiment, 10 July 2006.
f
Total inorganic N (summation of NH4 + -N and NO3 − -N).

were analyzed for the following: ammonia nitrogen (NH4 + -N),
nitrate nitrogen (NO3 − -N), and inorganic phosphate. Ammo-
nia nitrogen was measured using the salicylate method
(Mulvaney, 1996a). Nitrate nitrogen was measured using the
cadmium reduction method (Mulvaney, 1996b). Total effluent
nitrogen consisted of 98% ammonium and about 1% nitrate.
The total inorganic nitrogen (Inorganic N) was approximated
as the summation of NH4 + -N and NO3 − -N. The ascorbic acid
method was used to measure total inorganic phosphate (Kuo,
1996).
2.4. Statistical analysis
The percentage removal of inorganic N and phosphate
between the Leilehua and Perlite MSL systems were com-
pared using Sigma Plot version 9 (Sigma Plot, 2004). Data for
selected intervals of time corresponding to specific treatments
were also analyzed using the Statistical Analysis Software, SAS
PROC MIXED Repeated Measures ANOVA and Least Square
means (LSmeans) (SAS, 2004) (Littell et al., 1996, 1998; SAS,
2004).
(3) the combination of these two improvements.
3.1. Characteristics of dairy effluent
The analysis of the effluent was compared with other dairy
effluents in Hawaii (Fukumoto et al., 2000; Valencia-Gica et
al., 2004) (Table 4). The concentration of total N, NH4 + -N, and
NO3 − -N was lower in the effluent use in this experiment than
the other dairy effluent. This might be a result of using effluent
from the third and last settling of the lagoon system, which
was more diluted than that from the first lagoon.
3.2. First phase (year 2005)
The effectiveness of the MSL systems in removing inorganic N
and phosphate was compared over a 6-month period (5 May–5
October) (Figs. 2 and 3).
The MSL systems were not significantly different in per-
centage removal of inorganic N (P > 0.1) (Table 5). However,
the percentage removal of inorganic N was significantly differ-

3. Results

The percentage removal of inorganic N and phosphate are

discussed in two phases. The first phase data occurred from
May to October 2005, when the system was operated at con-
stant conditions and the second phase data from January to
July 2006, where specific treatments were applied. Although
samples were collected in the first phase from October to
December 2005, the data were not included in the analysis
because of system malfunction. There was a 6-week pause (2
December 2005 to 14 January 2006) between the two phases
due to mechanical problems and also due to a suspected build
up of biofilms. In the second phase aeration and sucrose addi-
tions were compared in an attempt to increase the efficiency
of the MSL systems. Three possible improvements were tested
in the second phase:
Fig. 2 – Removal of inorganic N in the Leilehua and Perlite
(1) effect of increased aeration, MSL systems as affected by time. In this figure “2D”
(2) effect of sucrose addition with a constant rate of aeration, indicates a 2-day pause; “3D” a 3-day pause; “6D” a 6-day
and pause; “8D” a 8-day pause; “12D” a 12-day pause.
6 e c o l o g i c a l e n g i n e e r i n g 3 2 ( 2 0 0 7 ) 1–10

Fig. 3 – Removal of phosphate in the Leilehua and Perlite Fig. 4 – Removal of inorganic N in the Leilehua and Perlite
MSL systems as affected by time. In this figure “2D” MSL systems as affected by sucrose addition and different
indicates a 2-day pause; “3D” a 3-day pause; “6D” a 6-day rates of aeration.
pause; “8D” a 8-day pause; “12D” a 12-day pause.

3.3. Second phase (year 2006)

Table 5 – A comparison of the effect of time and MSL
system on inorganic N removal as analyzed by SAS Proc 3.3.1. The effect of aeration
Mixed repeated measures analysis A comparison was made between no aeration and two differ-
Factor Probability of significance (P) ent rates of aeration (28 L min−1 and 31 L min−1 ) in removal
of inorganic N and phosphate during a sampling period of 19
MSL systems 0.8303
Time <0.0001 January to 27 April 2006 (Figs. 4 and 5).
MSL systems × time 0.9474 The percentage removal of inorganic N was not signifi-
cantly different between the MSL systems (P > 0.1) (Table 7).
There was no significant difference in percentage removal of
ent over time for both the MSL systems (P < 0.0001) (Table 5).
inorganic N with aeration for both the MSL systems (P > 0.1)
The non-significant interaction indicates that the MSL sys-
(Table 7). The non-significant interaction indicates that the
tems behaved similarly in percentage removal of inorganic
two MSL systems behaved similarly in percentage removal of
N (P > 0.1) (Table 5). The inorganic N removal by the Leilehua
inorganic N (P > 0.1) (Table 7). The removal of inorganic N by the
MSL system and the Perlite MSL system ranged from 22 to 93%
Leilehua system and the Perlite system ranged from 8 to 61%
(LSmean of 61.94) and 21 to 96% (LSmean of 63.40), respectively
(LSmean of 29.34) and 10 to 73% (LSmean of 33.10), respectively
(Fig. 2).
(Fig. 4).
The percentage removal of phosphate was significantly dif-
ferent by both the MSL systems (P < 0.05) (Table 6). The Leilehua
MSL system was more effective in removing phosphate than
the Perlite MSL system. There was also a significant differ-
ence in percentage removal of phosphate over time (P < 0.001)
(Table 6). The significant interaction indicates that there was
a decrease in percentage removal of phosphate in the Perlite
MSL system (P < 0.1) (Table 6). The percentage removal of phos-
phate by the Leilehua MSL system (64–99%) (LSmean of 92.70)
was greater than the Perlite MSL system (9–97%) (LSmean of
59.41) (Fig. 3).

Table 6 – A comparison of the effect of time and MSL

system on phosphate removal as analyzed by SAS Proc
Mixed repeated measures analysis
Factor Probability of significance (P)

MSL systems 0.0292

Time 0.0004 Fig. 5 – Removal of phosphate in the Leilehua and Perlite
MSL systems × time 0.0997 MSL systems as affected by sucrose additions and different
rates of aeration.
 e c o l o g i c a l e n g i n e e r i n g 3 2 ( 2 0 0 7 ) 1–10 7

Table 7 – A comparison of the effect of aeration and MSL Table 10 – A comparison of the effect of sucrose and MSL
system on inorganic N removal as analyzed by SAS Proc system on phosphate removal as analyzed by SAS Proc
Mixed repeated measures analysis Mixed repeated measures analysis
Factor Probability of significance (P) Factor Probability of significance (P)

MSL systems 0.8008 MSL systems 0.0342

Aeration 0.8214 Sucrose 0.1325
MSL systems × aeration 0.4330 MSL systems × sucrose 0.0980

Table 8 – A comparison of the effect of aeration and MSL Table 11 – A comparison of the effect of different rates of
system on phosphate removal as analyzed by SAS Proc aeration and MSL system on inorganic N removal as
Mixed repeated measures analysis analyzed by SAS Proc Mixed repeated measures analysis
Factor Probability of significance (P) Factor Probability of significance (P)
MSL systems 0.0088 MSL systems 0.8640
Aeration 0.0782 Aeration 0.3546
MSL systems × aeration 0.0141 MSL systems × aeration 0.3249

The percentage removal of phosphate was significantly

of phosphate was increased in the Perlite system (P < 0.1)
different between the MSL systems (P < 0.01) (Table 8). The
(Table 10). The removal of phosphate by the Leilehua MSL sys-
Leilehua MSL system was more effective in percentage
tem ranged from 59 to 93% (LSmean of 76.31) and was more
removal of phosphate than the Perlite MSL system. Changes
effective than the Perlite MSL system (11–75%) (LSmean of
in aeration resulted in a significant difference in percentage
46.13) (Fig. 5).
removal of phosphate (P < 0.1) (Table 8). The significant inter-
action indicates that the MSL systems behaved differently in
the removal of phosphate (P < 0.05) (Table 8). The removal of 3.3.3. The effect of different rates of aeration
phosphate by the Leilehua MSL system ranged from 42 to 91% Three different rates of aeration (11 L min−1 , 17 L min−1 , and
(LSmean of 66.64) and was greater than the Perlite MSL system 23 L min−1 ) were compared in removal of inorganic N and
(11–41%) (LSmean of 27.22) (Fig. 5). phosphate between a sampling period of 4 May to 10 July 2006
when sucrose was added (Figs. 4 and 5).
The MSL system performance was not significantly differ-
3.3.2. The combined effect of sucrose with aeration
ent in the percentage removal of inorganic N (P > 0.1) (Table 11).
A comparison was made between the non-sucrose and
There was no significant difference in percentage removal of
sucrose applications with constant aeration in removal of inor-
inorganic N with different rates of aeration between the MSL
ganic N and phosphate from 16 February to 13 April and 25 May
systems (P > 0.1) (Table 11). The non-significant interaction
to 10 July 2006 (Figs. 4 and 5).
indicates that the MSL systems behaved similarly in percent-
The MSL systems were not significantly different in per-
age removal of inorganic N (P > 0.1) (Table 11). The removal of
centage removal of inorganic N (P > 0.1) (Table 9). However,
inorganic N by the Leilehua and the Perlite systems ranged
there was a significant increase in percentage removal of inor-
from 31 to 89% (LSmean of 60.33) and 43 to 92% (LSmean of
ganic N with sucrose additions for both the MSL systems
62.24), respectively (Fig. 4).
(P < 0.1) (Table 9). The non-significant interaction indicates that
There was a significant difference observed in percent-
the MSL systems behaved similarly in percentage removal of
age removal of phosphate between the MSL systems (P < 0.1)
inorganic N (P > 0.1) (Table 9). The inorganic N removal by the
(Table 12). There was no significant difference in removal of
Leilehua and the Perlite MSL system ranged from 9 to 89%
phosphate with different rates of aeration (P > 0.1) (Table 12).
(LSmean of 48.77) and 10 to 92% (LSmean of 53.36), respectively
The non-significant interaction indicates that the MSL sys-
(Fig. 4).
tems behaved similarly in removal of phosphate (P > 0.1)
The MSL systems differed significantly in the percentage
(Table 12). The percentage removal of phosphate by the Leile-
removal of phosphate (<0.05) (Table 10). There was no over-
hua MSL system ranged from 59 to 93% (LSmean of 73.71) was
all significant difference in percentage removal of phosphate
greater than the Perlite MSL system (17–75%) (LSmean of 46.77)
with the sucrose application (P > 0.1) (Table 10). However, the
(Fig. 5).
significant interaction indicates that the percentage removal

Table 9 – A comparison of the effect of sucrose and MSL Table 12 – A comparison of the effect of different rates of
system on inorganic N removal as analyzed by SAS Proc aeration and MSL system on phosphate removal as
Mixed repeated measures analysis analyzed by SAS Proc Mixed repeated measures analysis
Factor Probability of significance (P) Factor Probability of significance (P)

MSL systems 0.6583 MSL systems 0.0937

Sucrose 0.0539 Aeration 0.1373
MSL systems × sucrose 0.8913 MSL systems × aeration 0.2538
8 e c o l o g i c a l e n g i n e e r i n g 3 2 ( 2 0 0 7 ) 1–10
Fig. 6 – The relationship between P sorbed and soil solution
P in Leilehua and Honouliuli soil.
4. Discussion
4.1. Performance of the MSL system in removing
inorganic N
The efficiency of the MSL systems in removing inorganic N
from dairy effluent was not significantly different through-
out the study (years 2005 and 2006). Both MSL systems used
for the aerobic layers, whether Leilehua or Perlite, were sim-
ilarly effective in removing inorganic N. However, the MSL
systems were significantly different in removal of inorganic
N over time in 2005. The removal rate of inorganic N in efflu-
ent decreased over time. We hypothesized the decrease might
be due to inadequate aeration in the aerobic layer or decreased
microorganism-available carbon in the anaerobic layer and
tested this by adding supplemental aeration and sucrose (as
a carbon source) in 2006. The systems were not significantly
different in removal of inorganic N with supplemental aera-
tion. However, the removal of inorganic N was significantly
increased in both the MSL systems with the application of
sucrose. The increased removal rate was likely due to the
additional carbon provided by sucrose applications, which
enhanced microbial activity and thus increased the denitri-
fication in both MSL systems. The removal of inorganic N was
not significantly different with different rates of aeration in
2006. There were some pauses in effluent delivery by the MSL
systems in 2005 and sudden drops in removal of inorganic N
seemed to related to these pauses.
4.2. Performance of the MSL system in removing
phosphate
The efficiency of the MSL systems in removing phosphate from
dairy effluent varied significantly during the study (years 2005
and 2006). The Leilehua MSL system was consistently more
effective in removing phosphate than the Perlite MSL system.
This was probably because of the high P sorption capacity of
the Leilehua soil in the Leilehua MSL system (1600 ␮g P g−1 soil)
(Fig. 6) which adsorbs phosphate from the effluent. The
removal of phosphate was significantly decreased over time
by the Perlite MSL system in 2005. We hypothesized that the
decrease in removal of phosphate by the Perlite system might
be a result of decreased microorganism-available carbon in
the anaerobic layer related to reduced iron movement into the
aerobic layer. Supplemental aeration and carbon (as sucrose)
were applied in 2006 to increase the efficiency of the MSL
systems in removing phosphate. The percentage removal of
phosphate was significantly increased with the application of
sucrose in the Perlite MSL system. This might be because the
additional sucrose carbon increased the activity of microor-
ganisms resulting in more oxygen consumption and enhanced
reducing conditions in the anaerobic layer chemically reduc-
ing and moving iron into the aerobic layer where it could
precipitate as ferric iron and sorb the phosphate in effluent.
The hypothesized decrease in microorganism-available car-
bon in 2005 seems to be supported by the sharp increase
in percentage removal of phosphate by the Perlite MSL sys-
tem with sucrose applications observed in 2006. The removal
of phosphate was significantly increased with supplemental
aeration by the Leilehua MSL system. This might be because
of the sufficient aeration in the Leilehua MSL system oxi-
dized ferrous iron to ferric iron in the aerobic layer, leading
to higher adsorption of phosphate by the soil colloids. The
sucrose application did not increase the already high removal
of phosphate in the Leilehua MSL system. The removal of
phosphate was not significantly different with different rates
of aeration when comparing the two MSL systems in 2006.
The systems consistently removed phosphate with different
rates of aeration. Thus, from the results of supplemental aer-
ation and sucrose applications it appears that the phosphate
removal mechanism is likely different between the two MSL
systems. The removal of phosphate in the Leilehua MSL sys-
tem was mainly due to sorption by iron in the aerobic layer,
whereas in the Perlite MSL system it appears to be due to three
steps, solubilization in the anaerobic layer, movement into the
aerobic layer, and precipitation as ferric oxide.
4.3. Use of MSL-treated effluent
The Hawaii State Department of Health has three different
categories of recycled water—R-1, R-2, and R-3 water—which
are listed in Table 13 with specific criteria (Hawaii State
Department of Health, 2002). R-1 is the highest quality recy-
cled water. It has been filtered and disinfected. It can be used
in any form of irrigation served by fixed irrigation systems sup-
plied by buried piping for turf and landscape irrigation of golf
courses, parks, elementary schools, roadsides, and residential
property where managed by an irrigation supervisor (Hawaii
State Department of Health, 2002). R-2 is a slightly lower
quality recycled water. It is secondary (biologically) treated
wastewater that has also been filtered and disinfected (Hawaii
State Department of Health, 2002). Its use requires more cau-
tion and restrictive controls than R-1 water. R-3 is the least
pure class of recycled water. R-3 quality water is wastewater
that has been treated to the secondary level. It can only be
used for irrigation at places where people rarely go (Hawaii
State Department of Health, 2002).
The average concentration of NO3 − -N and phosphate, and
fecal coliform colonies in MSL-treated effluent of our study
is given in Table 14. If we compare our study with the recy-
cled water requirements of State Department of Health in
Hawaii, the MSL-treated effluent comes in as R-3 water. MSL-
 e c o l o g i c a l e n g i n e e r i n g 3 2 ( 2 0 0 7 ) 1–10 9

Table 13 – Recycled Water Standards (Hawaii State Department of Health, 2002)

Type of recycled water Treatment Recycled water quality Recycled water monitoring

R-1 Oxidizeda ≤23 fecal coliform/100 mL Coliform: no more than one sample in any 30-day period
Filteredb Nitrate ≤ 10 mg L−1
Disinfectedc Total phosphorus ≤ 1.0 mg L−1

R-2 Oxidized ≤200 fecal coliform/100 mL Coliform: no more than one sample in any 30-day period
Filtered Nitrate ≤ 10 mg L−1
Disinfected Total phosphorus ≤ 1.0 mg L−1

R-3 Oxidized
Secondary
Undisinfected

a
Wastewater in which the organic matter has been stabilized.
b
The passing of wastewater through natural undisturbed soils or filter media such as sand.
c
The destruction, inactivation, or removal of pathogenic microorganisms by chemical, physical, or biological means. Disinfection may be
accomplished by chlorination, ozonisation, other chemical disinfectants, UV radiation, membrane processes, or other processes.

Table 14 – Concentrations of MSL-treated effluent

NO3 − -N (␮g mL−1 ) Phosphate (␮g mL−1 ) Fecal coliform (cfu/100 mL)
a
May–October 2005
Leilehua MSL system 2.15 ± 3.25 0.46 ± 0.56 658 ± 1321
Perlite MSL system 3.81 ± 5.35 2.16 ± 1.72 459 ± 674

May–July 2006b
Leilehua MSL system 2.48 ± 2.68 2.83 ± 1.38 64 ± 95
Perlite MSL system 5.04 ± 9.24 5.25 ± 2.74 36 ± 53

a
First phase of data without aeration and sucrose addition (mean ± S.D., n = 21).
b
Second phase of data with different rates of aeration and constant rate of sucrose, considered as the optimal management of the system
(mean ± S.D., n = 9).

treated effluent meets the criteria of nitrate and fecal coliform
(May–July 2006) of R-2 water and approaches the criteria for R-
1 water. Improvements in efficiency of the type examined in
this study are needed to meet the phosphate criteria. In addi-
tion a process, such as chlorination is needed to disinfect the
treated effluent.
5. Conclusions
Results of this study showed that both MSL systems have the
potential to remediate dairy effluent. The percentage removal
of inorganic N was high and similar in both the MSL systems.
The percentage removal of phosphate was high to very high
in the Leilehua MSL system and it removed considerably more
phosphate than the Perlite MSL system. The supplemental
aeration, which was applied in the second phase of the study,
did not significantly improve the removal of inorganic N. The
removal of phosphate, however, increased in the Leilehua MSL
system with additional aeration. Application of sucrose with
constant aeration was crucial for removing inorganic N and
phosphate. It appears that sucrose additions increased the
microbial activity in the MSL systems which helped to increase
the removal of inorganic N and phosphate. The sucrose appli-
cations have the potential to improve MSL systems treatment
efficiency. The installation of MSL systems is simple and basi-
cally requires only electricity, freshwater, a constant supply
of effluent and a very small amount of land. The materials
used in the system are inexpensive and easily obtainable.
The MSL-treated effluent approaches R-1 water criteria, with
improvements in P removal still needed.
Acknowledgement
We gratefully acknowledge the USDA T-STAR Program, Univer-
sity of Hawaii for the support of this research.
references
Attanandana, T., Saitthiti, B., Thongpae, S., Kritapirom, S.,
Luanmanee, S., Wakatsuki, T., 2000. Multi-media-layering
system for food service wastewater treatment. Ecol. Eng. 15,
133–138.
Caro-Costas, R., Abruna, F., Figarella, J., 1972. Effect of nitrogen
rates, harvest interval and cutting heights on yield and
composition of stargrass in Puerto Rico. J. Agric. Univ. Puerto
Rico 56, 267–279.
Fukumoto, G.K., Duponte, M.W., Lee, C.N., 2000. Livestock
Industry Partnering for Education and Program
Implementation: Nutrient Management Alternatives and
Pollution Prevention Planning. Phase I. A report submitted to
the State of Hawaii, Governor’s Agriculture Coordinating
Committee (GACC).
Guo, F., Yost, R.S., 1998. Partitioning soil phosphorus into three
discrete pools of differing availability. Soil Sci. 163 (10),
822–833.
Hawaii State Department of Health Wastewater Branch, 2002.
The Health State Guidelines for the Treatment and Use of
10 e c o l o g i c a l e n g i n e e r i n g 3 2 ( 2 0 0 7 ) 1–10
Recycled Water. Available on-line at http://hawaii.gov/health/
environmental/compliance/sb library/livestock.pdf (Verified
January 5, 2006).
Hobo Weather Station, 2001–2003. Onset Computer Coporation.
Procasset, MA.
Hawaii State Department of Health Wastewater Branch, 1996.
The Health State Guidelines for Livestock Waste
Management. Available on-line at www.state.hi.us/health/
environmental/water/wastewater/pdf/reuse-final.pdf (Verified
May 5, 2006).
Ikenberry, C.D., Mankin, K.R., 2000. Review of Vegetative Filter
Strip Performance for Animal Waste Treatment. Presented at
the 2000 ASAE Mid-Central Meeting, Paper No. MC00128.
American Society of Agricultural Engineers. St. Joseph,
MI.
Kuo, S., 1996. Phosphorus. In: Sparks, D.L. (Ed.), Methods of Soil
Analysis, Part 3. Chemical Methods—SSSA Book Series No. 5.
SSSA, Madison, WI, pp. 908–909.
Littell, R.C., Milliken, G.A., Stroup, W.W., Wolfinger, R.D., 1996.
SAS System for Mixed Models. SAS Institute, Cary, NC.
Littell, R.C., Henry, P.R., Ammerman, C.B., 1998. Statistical
analysis of repeated measures data using SAS procedures. J.
Anim. Sci. 76, 1216–1231.
Luanmanee, S., Attanandana, T., Masunaga, T., Wakatsuki, T.,
2001. The efficiency of a multi-soil-layering system on
domestic wastewater treatment during the ninth and tenth
years of operation. Ecol. Eng. 18, 185–199.
Luanmanee, S., Boonsook, P., Attanandana, T., Saitthiti, B.,
Panichajakul, C., Wakatsuki, T., 2002. Effect of intermittent
aeration regulation of a multi-soil-layering system on
domestic wastewater treatment in Thailand. Ecol. Eng. 18,
415–428.
Manariotis, I.D., Grigoropoulos, S.G., 2003. J. Environ. Sci. Health
A; Toxic Hazard. Subst. Environ. Eng. 38, 2057–2068.
Mulvaney, R.L., 1996a. Nitrogen-Inorganic Forms, Ammonium. In:
Sparks, D.L. (Ed.), Methods of Soil Analysis, Part 3, Chemical
Methods—SSSA Book Series No. 5. SSSA, Madison, WI, pp.
1153–1155.
Mulvaney, R.L., 1996b. Nitrogen-Inorganic Forms, Nitrate. In:
Sparks, D.L. (Ed.), Methods of Soil Analysis, Part 3, Chemical
Methods—SSSA Book Series No. 5. SSSA, Madison, WI, pp.
1155–1159.
Prochaska, C.A., Zouboulis, A.I., 2003. Performance of
intermittently operated sand filters: a comparable study,
treating wastewaters of different origins. Water Air Soil Pollut.
147, 367–388.
SAS Institute, 2004. The SAS system for Windows. Release 9.1.
SAS Institute, Cary, NC.
Schaafsma, J.A., Baldwin, A.H., Streb, C.A., 2000. An evaluation of
a constructed wetland to treat wastewater from a dairy farm
in Maryland, USA. Ecol. Eng. 14, 199–206.
Sigma Plot, 2004. Systat Software. Version 9.0. Systat Inc.,
Richmond, CA, USA.
Smith, V.H., Tilman, G.D., Nekola, J.C., 1999. Eutrophication:
impacts of excess nutrient inputs on freshwater, marine, and
terrestrial ecosystems. Environ. Pollut. 100, 179–196.
Soil Survey Staff, 2006. National Soil Survey Characterization
Data. Soil Survey Laboratory, National Soil Survey Center,
USDA-NRCS-Lincoln, NE. Available on-line at http://
ssldata.nrcs.usda.gov/querypage.asp (Verified 5 April 2005).
U.S. Department of Agriculture-Soil Conservation Service
(USDA-SCS), 1992. Agricultural Waste Management Field
Handbook. Washington, DC.
Valencia-Gica, R.B., Wilcox, V., Yost, R.S., Evensen, C.I., 2004.
Nutrient uptake and effluent clean-up potential of tropical
pasture grasses—Preliminary Report. University of Hawaii at
Manoa (unpublished).
Wakatsuki, T., Esumi, H., Omura, S., 1993. High performance and
N & P removable on-site domestic waste water treatment
system by multi-soil-layering method. Water Sci. Technol. 27,
31–40.
Wright, P.E., 1996. Prevention, collection and treatment of
concentrated pollution sources on farms. In: Animal
Agriculture and the Environment: Nutrients, Pathogens and
Community Relations. Northeast Regional Agricultural
Extension Service, Ithaca, NY, pp. 142–158.