1
2
Part I: Cocopeat for wastewater treatment in the developing world:
comparison to traditional packing media in lab scale biofiltration columns
3
4
5
Ashley A. Thomson+1, Courtney M. Gardner2, Carley A. Gwin3, and Claudia K. Gunsch+*4
+
A.M. A.S.C.E.
6
7
Abstract
8
Cocopeat, a by-product of coconut processing plants widely available in Vietnam,
9
Thailand, Cambodia, the Philippines and Indonesia, was studied for its ability to support
10
biological nutrient removal in lab-scale vertical flow columns treating simulated wastewater.
11
Treatment performance for cocopeat was compared to sphagnum peat, a traditional packing
12
medium, and Celite®, an inert clay pellet. Removal efficiencies of nitrogen, phosphorus, and
13
biological oxygen demand (BOD) were measured over a period of 325 days. During the
14
treatment period, varying configurations were tested to determine the effect of varying aerobic,
15
anoxic, and anaerobic zones on nutrient removal. Overall, similar BOD removal profiles were
16
obtained for cocopeat as compared to sphagnum peat. Slightly more efficient anoxic conditions
17
and a less acidic environment developed with the cocopeat. Up to 75% nitrogen removal was
18
obtained; however, phosphorus removal was not accomplished using our experimental setup,
19
likely due to the absence of a completely anaerobic treatment zone. Overall, cocopeat appears to
20
be a promising alternative packing material for on-site wastewater treatment in Southeast Asia in
21
terms of nitrogen removal.
22
23
1 Doctoral Graduate, Department of Civil and Environmental Engineering, Duke University, Box 90287, Durham, North Carolina 27708-0287,
USA
2 Doctoral Student, Department of Civil and Environmental Engineering, Duke University, Box 90287, Durham, North Carolina 27708-0287,
USA
3 Doctoral Candidate, Department of Civil and Environmental Engineering, Duke University, Box 90287, Durham, North Carolina 27708-0287,
USA
4 Associate Professor, Department of Civil and Environmental Engineering, Duke University, Box 90287, Durham, North Carolina 27708-0287,
USA
*Corresponding Author: Phone: 919-660-5208. Fax: 919-660-5219. E-mail: ckgunsch@duke.edu.
24
25
26
27
28
29
Subject Headings: Wastewater treatment, biological process, nitrification, denitrification,
developing countries, international factors
Introduction
Increased attention and effort by the global community has led to an increase in access to
30
water treatment technologies in the developing world over the last 25 years. The Joint
31
Monitoring Program by the World Health Organization (WHO) and the United Nations
32
Children’s Fund (UNICEF) has declared that the world met the Millennium Development Goal
33
(MDG) target for drinking water. However, it is projected that the MDG target for sanitation will
34
likely be missed (WHO/UNICEF, 2014). Over 1.8 billion people have gained access to improved
35
sanitation since 1990, but 2.5 billion people remain without access (WHO/UNICEF, 2014). On-
36
site sanitation technologies can provide an interim solution. One such method is biofiltration.
37
Naturally occurring media, such as peat, can be utilized in fixed film systems, where the attached
38
growth, colonization and reproduction of microorganisms are promoted for the treatment of
39
wastewater streams (Sherman 2006).
40
Sphagnum peat filtration has been used for municipal wastewater treatment for many
41
decades (Gutenspergen et al 1980, Farnham and Brown 1972, Nichols and Boelter 1982), as well
42
as on-site treatment of household wastewater from septic tanks and pit latrines (Nichols and
43
Boelter 1982, Brooks et al 1984). Peat-packed biofilters have been effective at removing
44
nutrients, biological oxygen demand and coliform bacteria from wastewater streams (Corley et al
45
2006, Viraraghavan and Ayyaswami 1987, Kennedy and Van Geel 2001, Kennedy and Van Geel
46
2000, Viraraghavan 1993, Mariappan and Viraraghavan 2006). Naturally occurring media such
47
as peat offer a benefit over artificial media such as foam, textiles and plastic because they utilize
48
naturally occurring flora and microbial fauna. However, often these natural media must be
49
mined, which can have critical negative impacts on the environment.
50
Cocopeat, a readily available product in Southeast Asia, is a by-product of coconut
51
processing plants derived from the outer husk of the coconut (Verdonck 1983) when the coconut
52
shell is shredded and the fibers are removed. Cocopeat contains approximately 30% fibers and
53
70% ground pith, which has a soil-like texture. Although cocopeat is not a traditional peat as it is
54
not a mined material that has aged for hundreds to thousands of years, this material is referred to
55
as cocopeat due to its similar physical properties and uses. Because cocopeat exhibits similar
56
physical properties to traditional peat (Table 1), cocopeat may be a suitable packing medium for
57
biofiltration especially in economies with high coconut production and processing.
58
While cocopeat may be an ecologically sustainable and locally available packing medium
59
suitable for biofiltration, it is necessary to evaluate how it compares to other traditional packing
60
media. Thus, in the present study, cocopeat was compared to sphagnum peat and Celite® in lab
61
scale biofilters treating simulated septic tank effluent. These two media have been widely studied
62
for various biofiltration applications (Corley et al 2006, Gunsch et al 2007). Characteristics of
63
each packing media are shown in Table 1. In addition, microbial fingerprints were obtained and
64
compared between the different packing media to determine if similar community profiles and
65
robustness were established. Finally, the effect of promoting different redox zones in the
66
biofilters was evaluated with the goal of maximizing biological nutrient removal.
67
68
69
70
Material and Methods
Bioreactor Design. Six lab scale column reactors were built based on previous research
(Patterson 1999). Solid white plastic PVC pipe with a 102 mm diameter and 750 mm height was
71
used for the main body of the bioreactors. The height from the top of the reactor to the effluent
72
pipe was 367 mm. Three sampling ports were placed on two sides of the length in order to obtain
73
water and peat samples at different depths. Three sampling ports were placed at 241, 406 and
74
622 mm from the top (Figure 1A). Influent wastewater was sprayed intermittently over the
75
packing media using a timer (Titan controls, Vancouver, WA) and a Masterflex pump (Cole-
76
Parmer, Vernon Hills, IL). The effluent drained out of the bottom through a 4.7 mm inner
77
diameter clear vinyl tube placed at the bottom of the reactor. Each biofilter was first packed with
78
a 10 cm deep base layer of 4.7 – 12.7 mm diameter gravel. The purpose of this layer was to
79
allow for easier drainage from the reactors. The gravel used in this study was obtained from a
80
local hardware store and sieved to ensure uniform distribution. On top of the gravel layer, each
81
packing medium (i.e., cocopeat, sphagnum peat or Celite®) was placed to a height of 610 mm.
82
Each bioreactor was prepared in duplicate. A photograph of the experimental setup is shown in
83
Figure 1B.
84
85
Jar Tests. Baseline sorption and desorption rates for nitrogen and phosphorus were
86
compared for the three packing media. Each test was performed by filling Erlenmeyer flasks
87
with 100 mL of deionized water. One g of each packing medium was added to individual flasks,
88
in duplicate. Flasks were shaken at 60 rpm for 2 h. Liquid samples were collected and filtered
89
through 0.45 µm mixed cellulose ether filters (Whatman, Piscataway, NJ) for analysis
90
(Viraraghavan and Ayyaswami 1987). Ammonium and phosphate concentrations were measured
91
and compared before and after addition of the packing media using methods described below.
92
93
Bioreactor Operation. Activated sludge was obtained from the North Durham
94
Wastewater Treatment plant (Durham, NC) which currently treats 20 MGD and completes
95
biological nutrient removal of BOD, NH4+, and phosphorus. Approximately 4.14 L of activated
96
sludge inoculum was used to seed the bioreactors with an active microbial community. During
97
inoculation, the wastewater was added at a flow rate of 6 mL/min over a period of 11.5 h.
98
Following inoculation, the reactors were charged with simulated septic waste at a rate of 1.4
99
mL/s for 10 s every 2 h, based on previously published work (Corley et al 2006). A modified
100
simulated septic wastewater medium was used in this study and was based on a recipe developed
101
by Zeng et al (2003) and previously used in our laboratory (Alito and Gunsch, 2014). Two L of
102
the simulated wastewater consisted of 100 mL of solution A and 1.9 L of solution B. Solution A
103
contained 17.53 g/L NaCH3CO-2,1 g/L MgSO4, 0.45 g/L CaCl2, 3.3 g/L NH4Cl, 0.5 g/L peptone
104
and 6 mL/L of a nutrient solution (recipe provided below). Solution A was adjusted to pH 5.5
105
with 2 M HCl and autoclaved. Solution B contained 28.5 mg/L KH2PO4 and 32.5 mg/L K2HPO4
106
and was adjusted to pH 10 with 2 M NaOH. The nutrient solution consisted of 1.5 g/L FeCl3,
107
0.15 g/L H3BO3, 0.03 g/L CuSO4, 0.18 g/L KI, 0.12 g/L MnCl2, 0.06 g/L Na2MoO4, 0.12 g/L,
108
ZnSO4, 0.15 g/L CoCl2, and 10 g/L ethylenediamine tetraacetic acid (EDTA). All chemicals
109
were obtained from Sigma Aldrich. The simulated wastewater was prepared daily and added to
110
the column using a metered pump (Masterflex, Cole-Parmer, Vernon Hills, IL).
111
Various hydraulic residence times were tested with the intent of modifying redox
112
conditions along the column depth and promote the development of aerobic, anoxic and
113
anaerobic zones. Hydraulic residence times were modified by moving the effluent tube to
114
different heights thereby increasing the amount of time wastewater remained in the biofilters.
115
Four different water level depths were tested: 0, 140, 356 and 495 mm corresponding to the four
116
phases of the experiment. Phase I consisted of a fully aerobic treatment. Phase II consisted of
117
mostly aerobic treatment with a short anoxic treatment. Phase III consisted of aerobic and a
118
longer anoxic zone residence time while Phase IV consisted of the shortest aerobic zone. The
119
hydraulic residence times for the various phases were 0.1, 1.1, 2.8, and 4.0 d for Phases I, II, III
120
and IV, respectively. After steady state was reached for each Phase, 15 mL of packing media
121
samples were collected at each sampling port as well as the top of the bioreactor to analyze for
122
microbial community structure. In addition, throughout the experimental period, suspended
123
solids (SS), volatile suspended solids (VSS), biological oxygen demand (BOD), dissolved
124
oxygen (DO), pH, nitrogen (ammonium, nitrate, and nitrite) and phosphorus (phosphate)
125
concentrations were measured in the influent and effluent wastewater as described below.
126
127
128
129
Analytical Measurements. SS and VSS were measured according to standard methods as
described in The Standard Methods for the Examination of Water and Wastewater (2005).
BOD was measured in 300 mL glass BOD bottles using standard methods (EPA, Method
130
5210 B); approximately 0.16 g of Hach nitrification inhibitor was added to each 300 mL bottle
131
(Loveland, CO), followed by Hach BOD nutrient buffer pillows (Loveland, CO). Measurements
132
were recorded in duplicate. Dissolved oxygen measurements were also obtained for each
133
columns’ influent and effluent.
134
Ammonium, nitrite, nitrate and phosphate were all measured on 96 well plates using
135
colorimetric assays as previously described in the literature (Holzem et al 2014, Tu et al 2010,
136
Hernandez-Lopez and Vargas-Albores et al 2003). Concentrations of each nutrient were
137
calculated in influent and effluent samples by comparing the optical density of each test to the
138
optical density of known standards. For ammonium, nitrate, nitrite and phosphate, the standards
139
were ammonium sulfate, potassium nitrate, sodium nitrite, and potassium phosphate,
140
respectively. A brief description of each method is provided below.
141
Ammonium concentration was measured using a modified standard methods procedure
142
(Holzem et al 2014). The sample volumes were scaled down for a 96 well plate (from 2 mL to
143
200 µL). A standard curve was constructed by performing serial 50% dilutions of the stock
144
ammonium sulfate solution (25 mg/L). Two hundred µL of each dilution was aliquoted into each
145
well of the first row on a 96 well plate. Two hundred µL of either synthetic wastewater (influent)
146
or bioreactor effluent was also added to the 96 well plate in triplicate. A phenol solution and a
147
nitroprusside solution were added. After 1 h of incubation, the optical density in each well was
148
measured at 620 nm using a Thermo Electron Corporation Multiskan MCC (Waltham, MA) plate
149
reader. Ammonium concentrations in each sample were calculated based on the standard curve
150
(R2 = 0.9839 for concentrations ranging from 0 to 25 mg/L NH3-N).
151
Nitrite was measured using a modified Griess method (Holzem et al 2014). Fifty µL of
152
either sample or standard were added to a 96 well plate. Fifty µL of 1% sulfanilamide in 3N HCl
153
was added to each well. After a ten min incubation at room temperature, 50 µL of 0.1% N-(1-
154
naphthyl)-ethylenediamine dihydrochloride was added to each well and the plate incubated at
155
room temperature for 10 min. Absorbance was measured at 540 nm on a Thermo Electron
156
Corporation Multiskan MCC (Waltham, MA) plate reader. Nitrite concentrations in each sample
157
were calculated based on the standard curve (R2 = 0.9969 for concentrations ranging from 0 to
158
345 mg/L NO2- - N).
159
To measure nitrate, all samples underwent a cadmium reduction to convert nitrate to
160
nitrite (Tu et al 2010) followed by the nitrite colorimetric method. The original nitrite
161
concentration was subtracted from the measured concentration to give the nitrate concentration.
162
Nitrate concentrations in each sample were calculated based on the standard curve (R2 = 0.9942
163
for concentrations ranging from 0 to 722 mg/L NO3-).
164
Phosphate was measured using the microplate method developed by Hernandez et al
165
(2003). Phosphate concentrations in each sample were calculated based on the standard curve (R2
166
= 0.9953 for concentrations ranging from 0 to 570 mg/L PO4-P).
167
168
Microbial Sampling and DNA Extraction. After steady state was reached in each phase,
169
microbial community samples were collected at each sample port and the top of the reactor to
170
gauge redox zone depth. Dissolved oxygen (DO) could not be measured as the DO probe could
171
not be used with the various packing media, especially the celite, as the media would scratch the
172
probe membrane, thus rendering the measurement void. Microbial distribution was key since DO
173
was not measured. Seeing shifts in the microbial community with each phase change is a direct
174
reflection and indication that the redox zones changed. This observation is also supported by
175
production/reduction of nitrite and nitrate. Samples were frozen at -20 oC prior to DNA isolation.
176
DNA was extracted using a method adapted from Gunsch et al. (2005). The
177
phenol/chloroform/IAA mixture was obtained from Life Technologies (Carlsbad, CA).
178
Concentration and purity of DNA was measured on a Nanodrop spectrophotometer (Thermo
179
Fisher Scientific). Only samples with high quality DNA were used for subsequent analysis
180
(260/280 > 1.5).
181
182
T-RFLP. PCR was performed targeting the 16S gene, with the 6 – carboxyfluorescein-
183
labeled fluorescent forward primer (27F) and reverse primer (1392R). Methods were based on
184
Alito and Gunsch (2014) and Ikuma and Gunsch (2013). PCR conditions used were 94 ºC for 5
185
min followed by 35 cycles of 30 s at 94 ºC, 30 s at 50 ºC, and 30 s at 72 ºC, with a dissociation
186
step at the end for quality control. Amplicons were purified utilizing a Qiagen PCR Purification
187
Kit (Qiagen, Hilden, Germany) following the manufacturer’s protocol. Final amplicon
188
concentrations and purity were measured on a Nanodrop spectrophotometer (Thermo Fisher
189
Scientific). One hundred ng of purified PCR product and 10 U of Msp1 (New England Biolabs,
190
Beverly, MA, USA) were used for each T-RFLP digestion reaction. The mixture was incubated
191
at 37 ºC for 2 h. Applied Biosystems GeneScan v3.7.1 software (Foster City, CA) was used to
192
inspect T-RFLP profiles. T-REX (T-RFLP analysis Expedited) software, available online
193
(Culman et al 2009), was used. Nonmetric multidimensional scaling (nm-MDS) analysis was
194
performed and ordination plots are generated using PAST (Hammer et al, 2001).
195
196
Statistical Analysis. The unpaired, two tailed student’s t test was used to identify
197
statistical differences between samples. Results were considered statistically different when the
198
p-value <0.05.
199
200
201
Results and Discussion
A summary of the nitrogen, phosphorus, BOD, DO, pH, TSS, and VSS data for both the
202
influent wastewater and effluent from each bioreactor for all four phases is shown in the
203
supplementary information (Table S1).The average BOD removal efficiencies were 98, 99, and
204
99% for Celite®, cocopeat, and sphagnum peat, respectively. The BOD removal was statistically
205
identical between cocopeat and sphagnum peat (p>0.05) while Celite® had statistically higher
206
BOD effluent concentrations throughout (p<0.05). These data suggest that cocopeat is a suitable
207
packing material for supporting microbial growth.
208
The average pH of the influent wastewater was 7.70 + 0.38. The average effluent pH
209
from Celite®, cocopeat, and sphagnum peat was 6.36 + 0.39, 4.18 + 0.69, 3.95 + 0.33,
210
respectively. Effluent pH was significantly lower for all three packing media, however cocopeat
211
and sphagnum peat were statistically lower than Celite® (p<0.001). Peat has previously been
212
shown to be acidic (Shaw Precast Solutions, Shaw Peat Technical Manual, April 2003). The
213
effluent pH in the cocopeat reactor was statistically higher than in the sphagnum peat (p =
214
0.025), suggesting that cocopeat is less acidic than sphagnum peat. The decrease in pH in all
215
three reactors may also be linked to biological activity, specifically nitrification, which is known
216
to produce protons thereby lowering pH (Xu et al 2006).
217
218
Nutrient removal. The average ammonium removal was 54, 49, and 55% for Celite®,
219
cocopeat, and sphagnum peat reactors during the four phases, respectively (Figure 2A).
220
Generally, cocopeat had the highest effluent nitrate indicating nitrifying activity (Figure 2B).
221
Celite® consistently produced nitrite in the effluent for all phases, but nitrite was below the
222
detection limit in the effluent from both the cocopeat and sphagnum peat reactors (Figure 2C).
223
224
Phase I. The concentration of ammonium for all three packing media dropped
225
immediately after startup which was attributed to sorption (Witter and Kirchmann 1989).
226
Ammonium concentration then increased and stabilized following the microbial establishment of
227
ammonium oxidizing bacteria. Nitrite concentrations were low at the beginning of Phase I, then
228
increased and finally stabilized once steady state was reached suggesting that nitrite oxidizing
229
bacteria may be less efficient than ammonium oxidizing bacteria. Nitrite concentrations were
230
found to be highest in the Celite® reactor. Nitrate concentrations increased steadily throughout
231
Phase I, possibly as a more robust nitrite oxidizing community was established in the bioreactors.
232
In Phase I, nitrification was observed while denitrification was not accomplished, suggesting that
233
the oxygen concentration was too high and that the residence time was not long enough to
234
promote an anoxic zone.
235
236
Phase II. Ammonium concentration was relatively steady throughout Phase II for all
237
packing media. Nitrite concentration was highest in the Celite® reactors while full removal of
238
nitrite was accomplished in the cocopeat and sphagnum peat reactors. Nitrate concentrations
239
were elevated at the beginning of Phase II, but steadily decreased suggesting denitrifying
240
activity. These data suggest that both aerobic and anoxic conditions were promoted during Phase
241
II. Note that the Celite® reactor had higher concentrations of nitrite and lower concentrations of
242
nitrate, suggesting that nitrite oxidizing bacteria were not as active in the Celite® reactors
243
compared to the peat reactors. These data suggest that sphagnum peat and cocopeat can support
244
nitrification and denitrification.
245
246
Phase III. Ammonium concentrations remained steady throughout Phase III. Nitrite
247
concentrations were low throughout Phase III for all packing media. Nitrate steadily increased up
248
to Day 42 and then decreased. This change in nitrate may be due to the re-establishment of the
249
microbial community after the redox zone change. By the end of Phase III, nitrification and
250
denitrification were observed in all packing media. However, no statistical difference between
251
ammonium, nitrite, or nitrate concentrations was observed in any of the packing media types
252
when comparing data from Phases II and III. These data suggest that the longer hydraulic
253
residence time in Phase III did not facilitate further nitrogen removal.
254
255
Phase IV. Throughout Phase IV, ammonium and nitrate concentration profiles were
256
erratic in all bioreactors. This may be due in part to the steady decrease in pH in the cocopeat and
257
sphagnum peat reactors and acid production by nitrifiers (Xu et al 2006). From Phase I to IV, pH
258
decreased from 5.21 to 3.71 in the cocopeat and, from 4.17 to 3.59 in the sphagnum peat.
259
Nitrifiers and denitrifiers have been previously reported to be sensitive to pH (Rittman and
260
McCarty 2001), so as the pH gradually decreases from Phase I to IV, nitrification and
261
denitrification activity may have become stressed and led to the erratic ammonium and nitrate
262
profiles observed. These data suggest that pH control may be needed for the long term operation
263
of peat-packed biofilters.
264
265
While cocopeat and sphagnum peat bioreactors achieved higher levels of nitrification
266
than the Celite® bioreactors (p=2.07 x 10-7 and p=1.24 x 10-4 for cocopeat and sphagnum peat,
267
respectively, as compared to Celite®), in no instance was full nitrogen removal achieved. This
268
suggests that this reactor design may only be able to achieve full nitrification / denitrification
269
with lower influent nitrogen loadings. Furthermore, because nitrate was always present in the
270
effluent, an anaerobic zone was never promoted which is necessary for biological phosphorus
271
removal. The average phosphate concentration of the influent wastewater was 9.59 + 5.79 mg/L
272
(Figure 3). The average effluent phosphate concentration from the Celite®, cocopeat, and
273
sphagnum peat bioreactors were 9.82 + 4.43, 9.38 + 6.01, and 8.09 + 2.85 mg/L, respectively.
274
None of the packing media accomplished statistically significant removal of phosphate (p>0.05).
275
Phase changes did not affect phosphate removal for any of the packing media. This result is not
276
unexpected as anaerobic treatment was not achieved as discussed above. Furthermore, both
277
cocopeat and sphagnum peat were found to leach phosphate, evidenced by the occasional higher
278
concentration of phosphate in the effluent as compared to the influent. To further demonstrate
279
the involvement of peat in phosphate release, jar tests experiments were performed and results
280
showed that cocopeat released 46.5 + 7.5 mg/L of phosphate per g of cocopeat and sphagnum
281
peat released 141.5 + 85.5 mg/L of phosphate per g of sphagnum peat. While problematic in the
282
lab scale setup used herein, if anaerobic treatment could be achieved, the additional phosphate
283
may be removed biologically. In Part II of this study, we do see phosphate removal which may
284
be contributed to microbial uptake or phytoremediation.
285
286
287
Microbial community distribution.
Phase I. Community differences were observed between the three media types in Phase
288
I. Generally, phosphate, nitrate, and ammonium were the most important variables in the
289
cocopeat and sphagnum peat reactors as evidenced by the clustering of the microbial
290
communities along those vectors. pH and nitrite were the most important factors for Celite®
291
communities. This is significant as peat may select for different microbial communities than an
292
inert material such as Celite®, due to differences in surface area, porosity and nutrients.
293
Analyzing the nm-MDS plot shown in Figure 4A with PAST shows that, in Phase I,
294
component 1 explained 93% of the microbial community differences between the three packing
295
media and component 2 explained 2% (Figure S1). Using the coordinates generated by PAST for
296
the nm-MDS plot, we graphed each nm-MDS component 1 and 2 with the ammonium and
297
phosphate concentration, respectively, and there was a linear relationship for each (R2 = 0.9078
298
and 0.5604 for ammonium and phosphate, respectively), indicating that ammonium and
299
phosphate, to a lesser extent, were important parameters affecting microbial community
300
structure. This result is logical as ammonium plays a key role in the development of a nitrifying
301
microbial community in Phase I and phosphate leaching was high in the peat bioreactors.
302
303
Phase II. Nitrate was the most important variable in the cocopeat and sphagnum peat
304
communities indicated by the clustering along the nitrate vector (Figure 4B). However, the
305
community from one of the cocopeat duplicate reactors (reactor 4) was strongly impacted by
306
phosphate. As in Phase I, the Celite® communities were strongly impacted by nitrite and pH.
307
This may suggest that denitrification levels between the two peat types and Celite® were
308
important parameters controlling microbial community distribution. Using the same analysis
309
described above, nitrate was associated with component 1 (explained 59% of microbial
310
community structure differences; R2 = 0.8585) and phosphate was associated with component 2
311
(explained 16%; R2 = 0.6513) (Figure S2). Nitrate was important in Phase II, likely due to the
312
introduction of denitrifying organisms. Phosphate may have also been an important parameter
313
controlling removal due to the effect of leaching.
314
315
Phase III. Nitrate and ammonium were the most important variables for cocopeat and
316
sphagnum peat communities indicated by the clustering around the nitrate and ammonium axes
317
(Figure 4C). The two Celite® reactors showed distinct clusters indicating that the communities
318
between the duplicate reactors were different. Ammonium was associated with component 1
319
(56%; R2 = 0.9875) and nitrate was associated with component 2 (43%; R2 = 0.9867) (Figure
320
S3). Ammonium and nitrate may have been significant factors controlling microbial community
321
distribution due to the effect of nitrification and denitrification. Phosphate may not have been as
322
important in this phase, as compared to Phase II because the effect of leaching may have been
323
less significant. This result is consistent with data to be presented in Part II of this study where
324
phosphate leaching was found to no longer be significant after 110 d in field operated
325
constructed wetlands.
326
327
Phase IV. Cocopeat communities clustered along the ammonium axis. Sphagnum peat
328
communities clustered independently from the five environmental factors. Celite® communities
329
clustered along the phosphate and nitrite axes (Figure 4D). Ammonium was associated with
330
component 1 (90%; R2 = 0.9241) and nitrate was associated with component 2 (4%; R2 =
331
0.7422) (Figure S4). Ammonium and nitrate may have been significant factors controlling
332
microbial community distribution due to the effect of nitrification and denitrification.
333
334
Conclusions
335
In this study, 99% BOD and up to 75% nitrogen treatment was obtained using cocopeat
336
as a packing medium. Cocopeat was able to support nitrification and denitrification of synthetic
337
wastewater. The most efficient nitrification and denitrification was observed in Phase III where a
338
shallow aerobic and longer anoxic zone was present. Significant phosphorus removal was never
339
achieved which may either be a result of the media itself which leaches phosphate or the reactor
340
design used in our laboratory study. Overall, cocopeat was found to be a comparable packing
341
medium to sphagnum peat and provides an attractive alternative for field application. In
342
particular, this packing material may be attractive in applications where complete nutrient
343
removal is not required such as aquaculture which is commonly employed in Vietnam where the
344
effluent is discharged into fish ponds and the remaining wastewater nutrients act as fertilizer for
345
duck weed, a food source for the fish.
346
347
348
Acknowledgements
This material is based upon work supported by the National Science Foundation Graduate
349
Research Fellowship Program, RTI International, and the Bill and Melinda Gates Foundation.
350
Any opinions, findings, conclusions, or recommendations expressed in this material are those of
351
the author(s) and do not necessarily reflect the views of the NSF, RTI International, or the Bill
352
and Melinda Gates Foundation. We would like to thank our collaborators at RTI International,
353
especially David Robbins.
354
355
356
References
357
Arenas, M.; Vavrina, C.S. (2002). “Coir as an Alternative to Peat in Media for Tomato
358
Transplant Production.” HortScience 37(2): 309-312.
359
Alito C.; Gunsch C. (2014). “Assessing the Effects of Silver Nanoparticles on Biological
360
Nutrient Removal in Bench-Scale Activated Sludge Sequencing Batch Reactors.”
361
Environmental Science & Technology. 48(2):970-976.
362
APHA, AWWA, and WEF. (2005). Standard Methods for the Examination of Water and
363
Wastewater. 21st edition. American Public Health Association, Washington, D.C.
364
Brooks, J. L.; Rock, C. A.; Struchtemeyer, R. A. (1984). “Use of peat for on-site wastewater
365
treatment .2. field studies.” Journal of Environmental Quality, 13 (4), 524-530.
366
367
Corley, M.; Rodgers, M.; Mulqueen, J.; Clifford, E. (2006). “The performance of fibrous peat
368
biofilters in treating domestic strength wastewater.” Journal of Environmental Science and
369
Health Part a-Toxic/Hazardous Substances & Environmental Engineering, 41 (5), 811-824.
370
371
Culman, S.; Bukowski, R.; Gauch, H.; Cadillo-Quiroz, H.; Buckley, D. (2009). “T-REX:
372
software for the processing and analysis of T-RFLP data.” BMC Bioinformatics, 10,
373
(1), 171.
374
375
“Data resources and estimates – Introduction.” WHO/UNICEF Joint Monitoring Programme
376
(JMP) for Water Supply and Sanitation. <http://www.wssinfo.org/data-
377
estimates/introduction/> (Jan. 20, 2014).
378
379
Farnham, R. S. and Brown, J. L. (1972). “Advanced wastewater treatment using organic and
380
inorganic material.” Proc. 4th International Peat Congress, 4, 271–285.
381
382
Gunsch, C.; Cheng, Q.; Kinney, K.; Szaniszlo, P.; Whitman, C. (2005). “Identification of a
383
Homogentisate-1,2-Dioxygenase Gene in the Fungus Exophiala lecanii-corni: Analysis and
384
Implications.” Applied Microbiology and Biotechnology, 68(3):405-411.
385
386
Gunsch, C.; Kinney, K.; Szaniszlo, P.; Whitman, C. (2007). “Relative gene expression
387
quantification in a fungal gas-phase biofilter.” Biotechnology and Bioengineering, 98(1):
388
101-111.
389
390
Guntenspergen, G., Kappel, W. and Stearns, F. (1980) “Response of a bog to application of
391
lagoon sewage: the Drummond project – an operational trial.” Proc. 6th International Peat
392
Congress, 559–562.
393
394
Hammer, Ø.; Harper, D.; Ryan, P. (2001). “Past: Paleontological Statistics Software Package
395
for Education and Data Analysis.” Palaeontologia Electronica, vol. 4, issue 1, art. 4: 9pp.,
396
178kb. Accessed online 2014 at <http://palaeo-electronica.org/2001_1/past/issue1_01.htm>.
397
Hernandez-Lopez, J.; Vargas-Albores, F. (2003). “A microplate technique to quantify
398
nutrients (NO2- NO3-NH4+ and PO43-) in seawater.” Aquaculture Research, 34 (13), 1201-
399
1204.
400
Hutomo, Y.; Pinder, K. (2006). “Effects of Residence Time Distribution and Packing on
401
Methanol Oxidation in Biotrickling Filters.” J. Air and Waste Manage. Assoc. 56: 334 – 342.
402
Holzem, R.; Stapleton, H.; Gunsch, C. (2014). “Determining the Ecological Impacts of
403
Organic Contaminants in Biosolids Using a High-Throughput Colorimetric Denitrification
404
Assay: A Case Study with Antimicrobial Agents.” Environmental science & technology,
405
48(3), 1646 – 1655.
406
Ikuma, K; Gunsch, C. (2013). “Successful genetic bioaugmentation with Pseudomonas
407
putida for toluene degradation in soil columns: Engineering in situ microbial adaptation: A
408
case study for genetic bioaugmentation in soil columns.” Environmental Chemistry Letters,
409
11(4), 365 – 370.
410
Kennedy, P. L.; Van Geel, P. J. (2001). “Impact of density on the hydraulics of peat
411
filters.” Canadian Geotechnical Journal, 38 (6), 1213-1219.
412
Kennedy, P.; Van Geel, P. J. (2000). “Hydraulics of peat filters treating septic tank
413
effluent.” Transport in Porous Media, 41 (1), 47-60.
414
Mariappan,V., and Viraraghavan, T. (2006). “Nitrogen removal from on-site effluents: A
415
review.” Fresenius Environmental Bulletin,15(1):3-14.
416
Nichols, D. S.; Boelter, D. H. (1982). “Treatment of secondary sewage effluent with a peat-
417
sand filter bed.” Journal of Environmental Quality, 11 (1), 86-92.
418
419
Patterson, R.A. (1999). “Peat Treatment of Septic Tank Effluent.” Proceedings of On-site '99
420
Conference: Making on-site wastewater systems work. 13-15 July 1999. Held at University
421
of New England, Armidale. Lanfax Laboratories, Armidale. pp273-281.
422
Rittman, B. and McCarthy, P. (2001). Environmental Biotechnology: Principles and
423
Application, McGraw-Hill, New York, NY.
424
425
Sherman, Kevin M. (2006). “Introducing a New Media for Fixed-Film Treatment in
426
Decentralized Wastewater Systems.” Proceedings of the Water Environment Federation,
427
WEFTEC 2006: Session 51 through Session 60 , 4616-4624(9).
428
429
“Standard method for measuring biological oxygen demand (BOD), Method 5210 B.” US
430
EPA. <http://www.epa.gov/region9/qa/pdfs/5210dqi.pdf > (March 13, 2014).
431
Tu, X. H.; Xiao, B. D.; Xiong, J. A.; Chen, X. D. (2010). “A simple miniaturised
432
photometrical method for rapid determination of nitrate and nitrite in freshwater.” Talanta,
433
82 (3), 976-983.
434
435
Verdonck, O. (1983). “Reviewing and evaluation of new material used as substrates.” Acta
436
Horticulture, 150, 467-473.
437
Viraraghavan, T.; Ayyaswami, A. (1987). “Use of peat in water-pollution control – a
438
review.” Canadian Journal of Civil Engineering, 14 (2), 230-233.
439
440
Viraraghavan,T. (1993). Peat-based onsite wastewater systems. Journal of Environmental Science
441
and Health, 28(1): 1-10.
442
443
Witter, E., & Kirchmann, H. (1989). “Peat, zeolite and basalt as adsorbents of ammoniacal
444
nitrogen during manure decomposition.” Plant and soil, 115 (1), 43-52.
445
Xu, J. M., Tang, C., & Chen, Z. L. (2006). “The role of plant residues in pH change of acid
446
soils differing in initial pH.” Soil Biology and Biochemistry, 38(4), 709-719.
447
Zeng, R. J.; Lemaire, R.; Yuan, Z.; Keller, J. (2003). “Simultaneous nitrification,
448
denitrification, and phosphorus removal in a lab-scale sequencing batch
449
reactor.” Biotechnology and Bioengineering, 84 (2), 170-178.
450
Zeng, W.; Li, L.; Yang, Y. Y.; Wang, S. Y.; Peng, Y. Z. (2010). “Nitritation and
451
denitritation of domestic wastewater using a continuous anaerobic-anoxic-aerobic (A(2)O)
452
process at ambient temperatures.” Bioresource Technology, 101 (21), 8074-8082.
453
Zhang, B.; Gao, T. Y. (2000). “An anoxic/anaerobic/aerobic process for the removal of
454
nitrogen and phosphorus from wastewater.” Journal of Environmental Science and Health
455
Part a-Toxic/Hazardous Substances & Environmental Engineering, 35 (10), 1797-1801.
456
457
458