Characterization of dairy waste streams, current treatment practices, and potential for biological nutrient removal

PII:

Wat. Res. Vol. 32, No. 12, pp. 3555±3568, 1998
# 1998 Published by Elsevier Science Ltd. All rights reserved
Printed in Great Britain
S0043-1354(98)00160-2
0043-1354/98 $19.00 + 0.00

CHARACTERIZATION OF DAIRY WASTE STREAMS,
CURRENT TREATMENT PRACTICES, AND POTENTIAL
FOR BIOLOGICAL NUTRIENT REMOVAL
M
J. R. DANALEWICH1, T. G. PAPAGIANNIS1*
, R. L. BELYEA2,
M
M. E. TUMBLESON3 and L. RASKIN1**
1
Environmental Engineering and Science, Department of Civil Engineering, University of Illinois at
Urbana-Champaign, Urbana, IL 61801, U.S.A.; 2Animal Sciences Department, University of MissouriColumbia, Columbia, MO 65211, U.S.A and 3Department of Veterinary Biosciences, University of
Illinois at Urbana-Champaign, Urbana, IL 61801, U.S.A.

(First received March 1997; accepted in revised form March 1998)
AbstractÐFifteen milk processing plants in the Upper Midwest of the United States participated in a
study to obtain information on general process operation, waste generation and treatment practices,
chemical usage, and wastewater characteristics. Long term data on wastewater characteristics were
obtained for 8 of the 15 dairy plants, and a 24-h composite wastewater sample was characterized in
detail for each plant. Wastewater ¯ow rates and characteristics varied greatly among and within plants
and were not easily predictable even when detailed information on processing operations was available.
In addition, the contribution of milk and milk products to the waste streams was underestimated by
plant operators. The use of caustic soda, phosphoric acid, and nitric acid for cleaning had a signi®cant
impact on wastewater characteristics, despite the implementation of changes in chemical usage practices

during recent years. In particular, the use of phosphoric acid based cleaning products has been reduced
to eliminate or decrease discharge ®nes. It was determined that most of the on site treatment facilities
require renovations and/or operational changes to comply with current and future discharge regulations, especially with respect to nutrient (nitrogen and phosphorus) levels in their waste streams. It
was concluded that biological nutrient removal of dairy wastewaters should be feasible given the relatively high concentrations of easily degradable organics, the generally favorable organic matter to total
phosphorus ratio, and the very favorable organic matter to nitrogen ratio. # 1998 Published by
Elsevier Science Ltd. All rights reserved
Key words: dairy wastewater, enhanced biological phosphorus removal, biological nutrient removal.

INTRODUCTION

Discharging wastewater with high levels of phosphorus (P) and nitrogen (N) can result in eutrophication of receiving waters, particularly lakes and
slow moving rivers. To prevent these conditions,
regulatory agencies in many countries have imposed
nutrient discharge limits for wastewater e‚uents.
Recently, restrictions on P discharge have become
more stringent in some regions of the United States
(U.S.). For example, a P discharge limit of 1.0 mg/l
was introduced for Wisconsin on January 1, 1997
(Wisc. Adm. Code NR 217.04, 1997), and the implementation of P standards is anticipated for other
Midwestern states. These regulations impact U.S.
milk processing industries, many of which are
located in the Midwest, since their waste streams
often contain high nutrient levels (Brown and Pico,
1979).
*Author to whom all correspondence should be addressed.
[Tel: +1-217-3336964; Fax: +1-217-3336968/9464, Email: ].

Enhanced biological phosphorus removal (EBPR)
can be more cost e€ective than chemical precipitation strategies (Reardon, 1994). Therefore, it is
important for the dairy industry to evaluate EBPR,

combined with nitri®cation and denitri®cation (to
remove N), as a treatment option for nutrient
removal. Biological treatment of dairy wastewaters
may not be straightforward due to high variations
in ¯ow and chemical characteristics. Those factors,
combined with low temperatures during several
months of the year in the Upper Midwest, may
make consistent biological treatment dicult.
Consequently, reliable waste treatment is a constant
challenge for many of the more than 5,000 dairy
plants in the U.S. (Blanc and Navia, 1990), especially those in the Upper Midwest.
Publications with chemical characteristics of
dairy wastewater and common treatment practices
are scarce. Harper et al. (1971) conducted a
thorough review of dairy waste characteristics and
treatment during the late 1960s, based on an extensive literature study and a survey of 10% of the
dairy plants in the U.S. They concluded that the

3555


3556

J. R. Danalewich et al.

dairy industry had limited knowledge on the organic strength of their waste streams and that the
concentrations of many wastewater constituents
(e.g., nutrients) generally were not determined.
They also reported that existing on site treatment
systems had relatively low eciencies, and that information for the rational design of treatment facilities generally was not available. In a report that

provides the perspective of the dairy industry
during the 1970s, Brown and Pico (1979) summarized dairy wastewater characteristics and concluded
that waste streams generated by milk processing
plants should continue to be treated in municipal
treatment plants (i.e., publicly owned treatment
works, POTW). This view changed considerably
during the 1980s and 1990s as demonstrated by the
publication of several case studies on dairy wastewater treatment. Most of these case studies, as well
as research e€orts, have been limited to physicochemical or anaerobic and aerobic biological treatment, without taking nutrient removal into
consideration (e.g., Backman et al., 1985; Samson
et al., 1985; Martin and Zall, 1985; Sobkowicz,
1986; Goronszy, 1989; Blanc and Navia, 1990;
Eroglu et al., 1991; Rusten et al., 1992; Rusten et
al., 1993; Orhon et al., 1993; Ozturk et al., 1993;
Borja and Banks, 1994; Kasapgil et al., 1994). To
the best of our knowledge, the full scale application
of EBPR to dairy wastewater is discussed in only
one study (Kolarski and Nyhuis, 1995). The lack of
information on both dairy wastewater nutrient
characteristics and treatment using biological nutrient removal (BNR) motivated us to conduct this
study. Herein, we document current dairy plant
waste generation and treatment practices and
describe common wastewater characteristics to
establish the foundation for further studies of BNR
from dairy wastewater.

MATERIALS AND METHODS

Survey data
Fifteen milk processing plants, located in Minnesota,

Wisconsin, and South Dakota, were visited during the
winter of 1995±96. The plants were chosen to be representative for the dairy industry in the Upper Midwest of the
U.S. Composite wastewater samples were collected, and
information regarding general operation, waste generation
and treatment practices, and chemical usage was obtained
from 14 of the 15 plants via a comprehensive survey. In
addition, we received long term data on wastewater
characteristics from 8 of the 15 plants.
Sample collection
Composite wastewater samples (3±4 liter each) were collected over a 24-h time period from 15 milk processing
plants. Samples were stored, without head space, in 1-liter
Nalgene bottles with airtight screw caps. One liter of each
sample was preserved by adding H2SO4 (36 N) to decrease
the pH below 2 (APHA, 1992). All composite samples
were transported on ice and stored at 48C. Analyses were
performed within 2 to 4 days after sampling.

Analytical methods
Sample fractions were ®ltered through a 0.45-mm ®lter
prior to nitrate, nitrite, orthophosphate, and elemental
analyses. Other analyses were performed using un®ltered
sample fractions. Samples were analyzed for total and bicarbonate alkalinity, pH, 5-day biochemical oxygen
demand (BOD5), total solids (TS), volatile solids (VS), suspended solids (SS), volatile suspended solids (VSS), ammonia, and total Kjeldahl nitrogen (TKN) according to
standard methods (APHA, 1992). Chemical oxygen
demand (COD), nitrate, nitrite, orthophosphate, and total
P were determined according to methods developed by
Hach (Loveland, CO), which are based on standard
methods (APHA, 1992). Volatile fatty acid (acetate, propionate, butyrate, isobutyrate, valerate, and isovalerate)
(VFA) concentrations were measured by gas chromatography (GC) (Model 5830A, Hewlett Packard, Palo Alto,
CA). Samples were prepared by adding 50 ml of 50%

phosphoric acid to 1.5 ml of sample, stored at À48C overnight, and centrifuged for 15 min at 15,000 g. To prevent
volatilization of VFAs, supernatant was transferred to a
glass GC vial and sealed with a crimp cap. Concentrations
of selected metallic elements (K, Na, Ca, Mg, Al, Mn, Ni,
Cu, Co, and Fe) were determined by inductively coupled
plasma±optical emission spectrometry (Perkin-Elmer,
Norwalk, CT) at the Microanalysis Laboratory (School of
Chemical Sciences, University of Illinois).
RESULTS AND DISCUSSION

Survey results
Plant size (expressed as mass of milk processed
per day) varied considerably, but the primary products were similar for most facilities (Table 1).
Twelve of the 14 plants produced one or more
types of cheese and 7 of the plants processed whey
as a secondary product. Plant 11 was a cheese processing operation (e.g., slicing and drying of cheese),
while plant 6 specialized in aseptic canning of dairy
products. To relate wastewater production to the
size of the plant, the wastewater ¯ow rates for each
plant (mean, minimum, and maximum ¯ow rates)
are reported in Table 1. Mean wastewater ¯ow
rates ranged from 170 to 2,081 m3/day (45,000 to
550,000 gallon/d). Most plants reported large
hourly, daily, and seasonal ¯uctuations in wastewater ¯ow rates. Minimum wastewater ¯ow rates
ranged from 4 to 1,703 m3/day (1,000 to
450,000 gallon/d) and maximum wastewater ¯ow
rates varied from 257 to 2,650 m3/d (68,000±
700,000 gallon/d).
Waste generation in dairy processing facilities is
characterized by high daily ¯uctuations often associated with washing procedures at the end of production cycles (Goronszy, 1989; Eroglu et al.,

1991). High seasonal variations also are common
and correlate with the volume of milk received for
processing, which typically is high during summer
months and low during winter months (Eroglu et
al., 1991; Kolarski and Nyhuis, 1995). In their survey of the U.S. dairy industry, Harper et al. (1971)
calculated the amount of wastewater generated per
quantity of milk processed (waste volume coecient). The mean waste volume coecients for the


Dairy waste and biological nutrient removal

3557

Table 1. Plant production and wastewater generation
Milk processed
106 kg/day

Products produced 106 kg/year (106 lbs/year)

(106 lbs/day)

primary

secondary

1

0.9 (2.0)

whey 18 (40)

2

0.5 (1.1)

cheddar and Colby cheese 32
(70)
cheddar and Colby cheese 17
(37)

3

1.0 (2.1)

Plant

4
5
6
7
8
9
10

cheddar, Colby, and Monterey
Jack cheese 34 (75)
0.7 (1.5)
cheddar cheese 24 (54)
0.5 (1.2)
cheddar, Colby, and Monterey

Jack cheese 15 (34)
na
aseptic canning and cheese dips
39 (85)
0.7 (1.5)
cheddar, Colby, Monterey Jack,
and reduced fat cheese 25 (55)
0.7 (1.5)
cheddar cheese 28 (62)
0.8±0.9 (1.8±2.0)
cheddar cheese 30 (66)
0.7±0.8 (1.5±1.8)
cheddar cheese 22 (49)

11
12

na
0.5 (1.1)

13

0.7 (1.5)

14

0.9 (2.0)

Wastewater ¯ow rate m3/day (103 gal/day)
mean

min

max

1,135 (267)

nr

nr

946 (250)

nr

nr

651 (172)

nr

nr

whey 13 (29)

1,105 (292)
992 (262)

643 (170)
568 (150)

1,605 (424)
1,132 (299)

beverages (nr)

526 (139)

nr

nr

whey 26 (58)

681 (180)

307 (81)

1,041 (275)

whey 20 (44)

640 (169)
1,211 (320)
719 (190)

333 (88)
813 (215)
416 (110)

1,173 (310)
1,817 (480)
871 (230)

170 (45)
625 (165)

132 (35)
nr

257 (68)
nr

208 (55)

4 (1)

1,450 (383)

2,081 (550)

1,703 (450)

2,650 (700)

whey 22 (48)
whey 16 (35)

process cheese 91 (200)
dried cheese 10 (22)

mozzarella and provolone cheese
21 (46)
cream cheese and related
non-dairy variety
products 44 (97)
¯avored snack dips
5 (10)
Parmesan, Romano, and
alcohol 5,700 m3/yr
(1.5 Â 106 gal/yr)
cheddar cheese (nr)

other

septic cheese
sauce and
puddings (nr)

dried cheese
(nr)

na = not applicable.
nr = no value was reported.

dairy industry in general, and cheese producers in
particular, were 2.43 and 3.14 m3 wastewater/ton
milk processed, respectively. Their analyses indicated that the waste volume coecients for the
dairy industry varied widely (0.1 to 12.4 m3/ton)
and were not related to plant size or degree of automation. Based on these observations, Harper et al.
(1971) concluded that management planning and

eciency of management supervision were the controlling factors in the amount of wastewater generated. In our survey of cheese producers, waste
volume coecients were signi®cantly lower than
those in Harper's study and varied between 0.31
and 2.29 m3 wastewater/ton milk processed (with a
mean of 1.26 m3/ton). Thus, the increase in plant
size (the mean plant size in our study was four
times larger than the mean plant size in Harper's
survey), automation in product processing, and
introduction of clean-in-place (CIP) systems over
the last few decades have resulted in a signi®cant reduction in volume of wastewater generated per
amount of milk processed. However, the wide variation in waste volume coecients for the plants
included in our study indicates that it remains dicult to predict wastewater ¯ow rates, even if
detailed information on processing operations is
available. This suggests that management strategy is
still the determining factor in waste generation and
underscores the importance of characterizing waste
streams and evaluating wastewater treatability to
determine suitable waste treatment strategies.
In the context of pollution prevention e€orts, it is
important to relate wastewater generation to

speci®c locations or activities in dairy plant operations. Therefore, personnel were asked to rate potential wastewater generating activities as either a
major or minor contributor to total waste volume.
These results were used to assign an overall wastewater generation ranking to each activity (Table 2).
Cleaning of transport lines and equipment between
production cycles, cleaning of tank trucks, and
washing of milk silos appeared to be the largest
contributors to the overall wastewater volume. The
information in Table 2 is consistent with the limited
data on dairy plant wastewater generation available

in the literature (Harper et al., 1971; Goronszy,
1989; Kasapgil et al., 1994). In those studies, most
of the wastewater volume and loading was generated during cleanup of tanks, trucks, transport
lines, and equipment. Other sources of wastewater
were associated with equipment malfunctions or operational errors (milk spills during receiving, over¯ow from silos, milk and milk product spills during
processing, leakage from pipes, pumps, and tanks,
discharge of spoiled milk and milk products, and
loss during packing operations) (Eroglu et al.,
1991). Even though the primary source of wastewater is generated during activities essential to
plant maintenance (i.e., cleaning activities), the
ranking provided in Table 2 can be used to prioritize possible strategies to reduce wastewater volume
and loading. For example, some plants reused ®nal
rinse waters for subsequent initial cleaning activities, and several facilities recovered caustic soda.
All plants reported the presence of milk based
substances in their wastewater (Table 3): of the 14


3558

J. R. Danalewich et al.
Table 2. Summary of wastewater generating activities
Number of plants regarding activity as

Wastewater generation activitya

major

Cleaning of transport lines and equipment between production cycles
Cleaning of tank trucks
Washing of milk silos

Milk and milk product spills during processing
Milk spills during receiving
Milk and milk product discharge during production start up and change over
Leakage from pipes, pumps, and tanks
Over¯ow from tanks
Loss during packing operations
Discharge of cooling water
Discharge of spoiled milk and milk products
Lubrication of casers, stackers, conveyors, and other equipment
Cleaning of whey evaporators
Sterilization of equipment
Vegetable oil leaks
a

minor

4
3
3
0
0
0
0
0
0
0
0
1
1
0

0

10
9
9
12
12
12
9
9
9
4
3
1
1
1
1

Overall rank
1
2
2
4
4
4
7
7
7
10
11

12
12
14
14

The selection of wastewater generation activities is based on information provided by Harper et al. (1971) and Eroglu et al. (1991).

plants that participated in the survey, 11 plants
reported the presence of milk and cheese whey and
4 plants mentioned the presence of cheese. Other
products reported to be present in the wastewater
included: lactose, cream, evaporated whey, and
separator and clari®er dairy wastes. Since previous
studies had indicated that the dairy industry was
not able to construct mass balances on various milk
product constituents and did not know their contribution to wastewater volume and concentrations
(Harper et al., 1971), we asked personnel to estimate the contribution of the various milk products.
Six of the 14 plants estimated the loss of milk and/
or whey and those estimates are given in Table 3.
The contribution of milk based substances to nutrient levels in the waste streams is discussed below.
Harper et al. (1971) reported on chemical usage
practices in the dairy industry during the 1960s.
They also reviewed detergent and sanitizer characteristics and applications in the dairy industry. Key
components in alkaline cleaners are basic alkali
(e.g., soda ash (Na2CO3) and caustic soda
(NaOH)), polyphosphates, and wetting agents.
Complex phosphates are used for emulsi®cation,
dispersion, and protein peptizing. Wetting agents

(e.g., sulfated alcohols, alkyl aryl sulfonates, quaternary ammonium surfactants) are used in relatively low amounts, but are major contributors to

the detergents' BOD5 load. In addition to detergent
action, quaternary ammonium surfactants have
antiseptic and germicidal properties. Acid cleaners
are utilized to clean high-temperature equipment
and blends of organic acids (e.g., acetic, propionic,
lactic, citric, tartaric acids), inorganic acids (e.g.,
phosphoric, nitric, sulfuric acids), or acid salts generally are preferred (Harper et al., 1971; Samson et
al., 1985; Kolarski and Nyhuis, 1995). Sanitizers
typically contain large amounts of chlorine, which
can impact biological wastewater treatment (Harper
et al., 1971). In addition to chlorine compounds
(e.g., sodium hypochlorite), iodine compounds, quaternary ammonium compounds, and acids are used
as sanitizers. Harper and coworkers determined
that wash waters containing sanitizer solutions contributed to 0.2 to 13.8% (average 3.1%) of the
wastewater volume, whereas detergents were responsible for 2.2 to 41.6% of the overall wastewater
volume (average 15%). They also reported that
detergents signi®cantly increased wastewater alkali,
phosphate, and acid concentrations, but calculated,

Table 3. Presence of milk based substances in wastewater as estimated by plant personnel and reported use of nitric and phosphoric acids
Plant
1
2
3
4
5
6
7
8
9

10
11
12
13
14

Milk m3/day
(gal/day)

Whey m3/day
(gal/day)

1.1 (300)

0.4 (100)
[
0.3 (86)
[
[

0.3 (86)
[
[
[
0.2 (50)
0.2±0.5
60±120
[
[
0.2 (60)

[

[
0.4 (100)
0.2±0.5
60±120
[
[
[

Cheese

[
[
[

[

H3PO4 kg/day
HNO3 kg/day
HNO3 coecient
H3PO4 coecient
(lbs/day)
kg HNO3/106 kg milk
(lbs/day)
kg H3PO4/106 kg milk
[
[
[
92 (202)

[
[
[
[
99 (218)

[
[
135

37 (81)
[
[
[

54

109±121

53 (117)

59±65

78 (172)
[
[

96±115

8±10

572 (1,260)

630

7 (15)
[
[
[
40 (88)

44

[ indicates that milk/milk products were present or that nitric and phosphoric acids were used, but that quantities were not speci®ed.


Dairy waste and biological nutrient removal

using data supplied by detergent manufacturers,
that detergents contributed little to the BOD load
of the wastewater (a maximum BOD5 of 200 mg/l
was estimated to be attributed to detergents).
However, their own investigation of detergent usage
practices of milk processing plants indicated that
detergents contributed signi®cantly to BOD, to
refractory COD, and may have been important
with respect to toxicity and poor performance of
dairy waste treatment facilities (Harper et al., 1971).
To evaluate chemical usage in the U.S. dairy
industry today, dairy plant personnel were asked to

list types of cleaning, sanitizing, lubrication, and refrigeration chemicals used in their facilities.
Chemicals used most frequently included: caustic
soda, nitric acid, phosphoric acid, and sodium
hypochlorite. Soda ash and quaternary ammonium
were used by several of the plants, and ammonia,
trisodium phosphate, acetic acid, hydrochloric acid,
sulfuric acid, citric acid, lactic acid, hydroxyacetic
acid, sodium metasilicate, hydraulic oils, propylene
glycol, emulsi®ers, and antifoaming agents were
used occasionally in small amounts by a few plants.
To obtain information on nutrient sources in wastewater, we requested detailed information on quantities of nitric and phosphoric acids used. Some of
the plants provided information which was dicult
to interpret because the exact composition of the
cleaners and sanitizers was not provided. Table 3
lists the plants that used nitric and/or phosphoric
acids, and gives the amounts used for those plants
for which this information was obtained. Nitric and
phosphoric acids were used concurrently in 11
plants. Two plants used only nitric acid in their
cleaning cycles, while 1 plant used only phosphoric
acid. Nitric acid and phosphoric acid coecients
were calculated as the mass of acid used per
amount of milk processed (Table 3). These values
indicate that the amounts of cleaners varied considerably throughout the industry and that management strategy apparently was the determining factor
in chemical usage.
A comparison of cleaning practices today and
during the 1960s (Harper et al., 1971) indicates that
the types of acids used in cleaning operations have
changed considerably during the past decades. The
use of various organic acids and sulfuric and hydrochloric acids was more common, while nitric acid

was not utilized for cleaning during the 1960s. We
also asked plant personnel to describe changes in
cleaning practices. Seven plants reported that
chemical usage had been changed during the last
decade. Plants 7 and 10 switched from phosphoric
acid to a phosphoric/nitric acid blend in their cleaning cycles. Plants 2 and 14 reduced the amount of
phosphoric acid and increased the amount of nitric
acid in the cleaning solution. Thus, there appeared
to be a trend towards using less phosphoric and
more nitric acid. Plant 11 also indicated that the
use of acid cleaners (i.e., non-phosphoric acid based

3559

cleaners) had to be increased to improve equipment
cleaning. Waste minimization practices, such as reclamation of cleaning acids and caustic soda, were
initiated by personnel in plant 4. In an e€ort to
reduce caustic vapor problems, plant 9 began using
less caustic soda and more chlorinated alkali.
The changes in chemical usage practices over the
past few decades appear to relate at least partially
to environmental regulations. The reduced use of
organic acids corresponds to the implementation of
the Clean Water Act (1972), whereas the more
recent switch from phosphoric to nitric acid has
been driven by discharge surcharges based on
amount of P discharged in municipal treatment systems and the recent (1997) implementation of an
overall P discharge limit (1.0 mg/l) for Wisconsin.
Even though several plants indicated that the
reduced use of phosphoric acid resulted in substantial savings in P surcharges and ®nes, the switch to

nitric acid caused an increase in the amount of cleaners used. In addition, some plants indicated that
phosphoric acid based cleaners are preferred from a
cleaning perspective and that further decreases in
the use of phosphoric acid are unlikely. This perspective is consistent with the position of dairy
plants in the 1970s: Brown and Pico (1979) discussed that non-phosphate cleaners are not as e€ective as phosphate based cleaners and that their use
can result in increased cleaning costs because they
require higher concentrations and longer cleaning
cycles.
The use of caustic soda and various acids considerably impacts wastewater pH, as indicated in
Table 4. Of the 12 plants that reported pH data, 11
exhibited extreme pH ¯uctuations. Only 4 plants
provided information on wastewater temperature
(Table 4). The large variations in wastewater temperature indicated that temperature may be a concern if BNR would be implemented.
Current wastewater treatment practices in the
dairy industry vary considerably (Table 4). Four
plants did not practice any wastewater treatment on
site and directed their waste streams to a municipal
treatment system. The remaining 10 plants practiced
some form of on site wastewater treatment. A wide
assortment of treatment systems were described,
ranging from simple (e.g., equalization basin, ridge
and furrow system) to more complex (e.g., dissolved
air ¯otation (DAF), extended aeration, oxidation
ditch) systems. Seven facilities had equalization
basins and were better equipped to handle large
wastewater ¯ow and pH variations.
Whether simple or complex treatment systems
were employed, the ®nal disposal of sludge or biosolids is a major concern to the facilities, in particular when biosolids have the potential to contain
pathogens. Nine plants did not separate domestic
wastewater generated in the dairy facility from process wastewater. Five of these plants pretreated

their wastewater on site and thus generated waste-


3560

J. R. Danalewich et al.
Table 4. Wastewater temperature and pH; wastewater (pre)treatment strategy; sludge treatment and disposal strategy
pH

Temp. 8C

Plant

min

max

min

max

Wastewater (pre)treatment systemb

Sludge treatment strategy

1

3.0

11.0

nr

nr

occasional land application

2

3.0

13.0

32.0

43.0

3

nr

nr

nr

nr

4
5
6

4.7
3.0
4.5

11.5
13.0
12.0

nr
nr
nr

nr
nr
nr

7

7.1

12.5

nr

nr

8

4.0

12.0

nr

nr

9

4.7

12.3

nr

nr

10

7.5

8.1

2.8

21.0

11

1.0

14.0

14.0

32.0

12
13
14

5.3
nr
4.8

10.6
nr
11.3

nr
nr
22.0

nr
nr
38.0

pretreatment of main waste stream in equalization basin
and aerated lagoon; high-strength, low-volume waste
stream is land applied

treatment in equalization basin, DAFa, trickling ®lters,
oxidation ditch, post-treatment in series of two lagoons
before discharge into river, chemical additions include
polymers for dewatering and sulfuric acid for pH
adjustment
treatment of main waste stream in ridge and furrow
system; high-strength, low-volume waste stream is land
applied; whey water is discharged directly in river
no pretreatment
pretreatment in equalization basin
no pretreatment; high-strength, low-volume waste stream
is land applied
pretreatment in equalization basin; high-strength, lowvolume waste stream is land applied
no pretreatment of dilute waste stream (land applied or
treated by city); pretreatment of concentrated waste
stream in equalization basin, activated sludge system
(NH3 is added as N source), and oxidation ditch
treatment in aerated lagoons, e‚uent used for irrigation
in spring
pretreatment in equalization basin and conventional
activated sludge system
pretreatment in equalization basin and completely-mixed
activated sludge system
no pretreatment
no pretreatment
pretreatment in grit chamber, extended aeration activated
sludge system with addition of ferric chloride for
phosphate precipitation, and addition of polymers in
clari®ers

aerobic digester, thickening
tank, ®lter press,
composting, land
application
land application
na
na
na
na
nr

land application
belt ®lter press dewatering
and land application
land application
na
na
aerobic digestion, gravity
thickening, Somat Press
Auger, land application

nr = no value was reported.
na = not applicable.
a
DAF = dissolved air ¯otation; fats, oils, scum, and grease are removed from wastewater using DAF and treated together with stabilized
biosolids in ®lter press.
b
Pretreatment indicates that further treatment of wastewater e‚uent was accomplished in the local municipal wastewater treatment plant;
treatment indicates that no further treatment of wastewater was performed.

water biosolids that contained pathogens of potential concern in biosolids disposal or reuse applications. Since it is easier to ®nd biosolids disposal
or reuse options when domestic waste streams are
kept separate from process wastewaters, all plants
indicated that plans to separate the two waste
streams were being evaluated.
To evaluate the level of satisfaction with current
treatment strategies, we asked questions on problems encountered during wastewater treatment and
potential noncompliance with standards. Plants 2
and 11 disclosed that their treatment systems were
overloaded, while plant 9 attributed o€ensive odor
problems to their treatment system. Plants 11 and
14 reported activated sludge bulking as an occasional problem (a few times per year), while
plants 10 and 11 stated that activated sludge foaming, caused by ®lamentous microorganisms, was a
persistent problem. Furthermore, plants 10 and 11
indicated it was dicult to maintain adequate dissolved oxygen (DO) concentrations in their activated sludge tanks. These observations may suggest
that low DO levels encouraged the growth of ®lamentous organisms in these activated sludge systems. Plant 11 further speculated that elevated
levels of Gordona (formerly Nocardia) species were

responsible for foaming problems in their severely
overloaded plant. This is inconsistent with observations that Nocardia foaming generally is not common in plants with high food to microorganisms
(F/M) ratios (Jenkins et al., 1993). de los Reyes et
al. (1998) determined that levels of Gordona were
relatively low in foam taken from plant 11, which
indicated that other ®lamentous microorganisms
may have been responsible for foaming problems in
this plant.
All plants were subjected to regulations, but regulations varied widely depending on discharge practices and capacities of municipal treatment facilities.
Surcharges were based on wastewater ¯ow rate
and/or mass of BOD5, SS, and total P discharged
per day and commonly were levied according to a

predetermined discharge agreement, either with the
state's natural resources department or with the
municipality if (pretreated) wastewater was directed
to the local sewage treatment facility. If land application was practiced, ¯ow rate, BOD5, total P, N
(TKN), chlorides, and/or potassium concentrations
generally were determined. SS violations or surcharges were reported most commonly; 7 plants frequently failed to comply with SS standards. Plants
10 and 14 occasionally exceeded the allotted maxi-


Dairy waste and biological nutrient removal

3561

Table 5. Wastewater characteristics for extensive time periodsa
Time period

Flow rate (103 gal/day)

1

1/1/95±9/30/95

267 281 (37±527)

4

1/1/92±9/27/95

292 243 (170±424)

8.42 1.6 (4.7±11.5)

6
7

1/1/95±12/31/95
1/1/94±12/31/95

143 2 94 (29±1,444)
111 231 (25±168)

11.3 2 1.3 (7.1±12.5)

9
10

7/23/91±10/26/95
(excluding 1992)
8/29/93±4/21/94

12

1/10/95±12/20/95

158 214 (138±207)

7.72 1.8 (5.3±10.6)

14

12/28/94±8/1/95

508 263 (189±677)

7.02 1.0 (5.0±11.0)

Plant

pH

8.32 1.6 (4.7±12.3)
6.8 2 0.7 (5.2±9.6)

BOD5 (mg/l)

SS (mg/l)

2,1032 1,148
(600±10,000)
709 2 139
(420±1,060)

677 2544 (184±
7,330)

1,212 2684
(200±9,900)
2,2972 1,096
(650±9,600)
1,123 2404

(360±2,200)
1,717 2708
(820±3,900)
1,545 2527
(288±5,200)

Total P (mg/l)

928 2305 (152±
3,570)
1,082 21,023 (293±
13,700)
686 2378 (253±
2,540)
405 2163 (110±
1,050)

782 20 (31±227)
552 25 (28±293)
372 16 (14±104)
57 29 (34±72)
362 14 (18±132)

a

Each parameter is reported as mean 2 SD (min±max) for the indicated time period.

mum wastewater discharge volume, and BOD5 discharge violations were reported by plants 4, 5, and
10. Plants 5, 7, 11, and 14 disclosed that ®nes or
surcharges were levied due to high P discharge

levels and several plants were anticipating further
changes in surcharge levels based on e‚uent P concentrations.
Long term data
Eight of the 15 plants provided data on wastewater characteristics for extensive time periods.
Mean, standard deviation (SD), minimum (min),
and maximum (max) values are given in Table 5
and demonstrate that wastewater ¯ow rates and pH
values varied greatly within and among plants.
BOD5, SS, and P concentrations also were commonly measured and varied considerably. The availability of wastewater characteristics for extensive
time periods is useful for determining seasonal
trends, which should help suggest improved wastewater treatment strategies for the dairy industry.
However, the number of parameters measured on a

regular basis was limited and additional analyses
are necessary to help evaluate the potential for
BNR (e.g., nitrate, nitrite, orthophosphate, VFA).
Composite wastewater samples
Detailed chemical characteristics of the 15 composite wastewater samples are summarized in
Tables 6±9. For comparison, summaries of dairy
wastewater characteristics obtained from studies
published during the 1980s and 1990s are given in
Tables 10 and 11. Since signi®cant fractions of the
organic constituents and nutrients in dairy wastewater are derived from milk and milk products,
some of the characteristics of whole milk are presented in Table 12.
Mean total BOD5 and total COD values
(1,856 mg/l and 2,855 mg/l, Table 6) con®rm that
milk processing wastewaters often have a relatively
high organic strength. These values were in the
same range as the data given for extensive time
periods (Table 5) and those cited in the literature

during the 1980s and 1990s (Table 10). In addition,

Table 6. Chemical characteristics of composite wastewater samples

Plant
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Mean
SD
Min
Max

Total BOD5 Total COD
(mg/l)
(mg/l)
BOD5/COD SS (mg/l)
1,843

5,722
1,298
826
2,738
568
1,466
565
3,269
1,003
2,406
1,887
2,108
1,175
959
1,856
1,335
565
5,722

nd = not determined.

2,447
7,619
2,032
2,309
3,556
785
2,909
2,290
4,895

1,644
3,093
2,817
3,232
1,570
1,625
2,855
1,646
785
7,619

0.75
0.75
0.64
0.36
0.77
0.72
0.50
0.25
0.67
0.61
0.78
0.67
0.65
0.75
0.59
0.63
0.16
0.25
0.78

586
1,533
389
696
730
470
1,910
3,560
885
371
757
853
923
326
655
976
833
326
3,560

VSS (mg/l)

TS (mg/l)

VS (mg/l)

pH

419

1,477
225
567
663
307
1,010
1,935
680
327
699
767
890
284
298
703
479
225
1,935

3,747
6,342
nd
2,925
3,583
1,833
4,180
5,354
4,495
2,023
6,063

3,683
2,863
2,327
14,205
4,545
3,114
1,837
14,205

1,710
5,088
nd
1,848
1,967
562
1,513
2,998
3,060
900
1,243
1,550
nd
nd
11,034
2,790
2,863
562
11,034

10.7

6.2
11.3
6.7
6.9
6.8
9.4
7.9
10.3
7.0
6.9
7.5
10.8
9.8
7.6
8.4
1.8
6.2
11.3

Alkalinity/
Alkalinity BOD5 (mg/l
(mg/l as as CaCO3/
CaCO3) mg/l as O2)
375
225
500
500
400
525
1,550

1,525
775
625
500
650
614
450
400
652
382
225
1,550

0.20
0.04
0.39
0.61
0.15
0.92
1.06
2.70
0.24
0.62
0.21
0.34
0.29
0.38
0.59
0.58
0.65

0.04
2.70


3562

J. R. Danalewich et al.
Table 7. Nutrient levels in composite wastewater samples and estimated levels of P and N required for BOD removal

Plant
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Mean
SD
Min
Max

Total P
(mg/l as P)
60
74
49
51
36
65
134
181
79
29
35
68
97
52
54
71
40
29
181

Orthophosphate
(mg/l as P)
19
20
15
11
9
19

32
35
21
7
22
6
26
13
15
18
8
6
35

N required
P required
for BOD
for BOD
removala
NOÀ
NOÀ
TKN
NH3
Organic N
removala
3
2
(mg/l as P) (mg/l as N) (mg/l as N) (mg/l as N) (mg/l as N) (mg/l as N) (mg/l as N)
9
30

7
4
14
3
8
3
17
5
12
10
11
6
5
10
7
3
30

34.7
3.2
51.0
0.8
1.2
8.6
1.0
14.3
47.3
0.6
23.7
0.9

1.4
80.0
52.8
21.4
25.6
0.6
80.0

2.3
4.1
0.3
0.4
0.8
0.7
0.8
1.8
1.2
0.4
2.1
0.6
1.0
34.0
3.3
3.6
8.5
0.3
34.0

111.0
106.0

140.0
40.1
134.0
14.0
62.0
nd
122.0
83.0
128.0
83.0
nd
74.0
nd
91.4
39.3
14.0
140.0

5.3
11.6
10.6
2.8
9.3
1.0
9.4
3.7
9.4
11.7
7.9
34.1

1.4
5.5
4.8
8.6
7.9
1.0
34.1

105.7
94.4
129.4
37.3
124.7
13.0
52.6
nd
112.6
71.3
120.1
48.9
nd
68.5
nd
81.5
38.4
13.0
129.4

47
148

33
21
71
14
38
14
84
26
62
48
54
30
24
48
35
14
148

a

See text for details on calculations.

the organic strength varied greatly within and
among plants, as demonstrated by wide ranges for
BOD5 and COD values in Tables 5 and 10 and
large standard deviations in Table 6, respectively.
To evaluate the potential biodegradability of the
organic compounds in dairy wastewater, we calculated the BOD5:COD ratio. For all but 2 of the
composite wastewaters (plants 4 and 8), the
BOD5:COD ratio was above 0.5, with a mean of

0.63 2 0.16 (Table 6). BOD5:COD ratios obtained
from literature data ranged between 0.47 and 0.67
with a mean of 0.58 (Table 10). Based on an extensive set of BOD5:COD ratios obtained for milk products, milk constituents, and dairy wastewaters,
Harper et al. (1971) concluded that ratios below
0.60 can be interpreted to suggest a less ecient
biological oxidation of milk wastes compared to
pure milk, probably caused by the presence of nonmilk constituents. They also suggested an apparent
``toxicity'' of dairy plant wastes when ratios were

below 0.40. Low ratios apparently coincided with
major periods of equipment process cleaning, indicating the source of toxicity was related to cleaning
operations. Thus, our results indicate that most of
the organic compounds in dairy wastewaters should
be easily biodegradable.
SS and VSS levels also are used to evaluate
wastewater strength and treatability. SS in dairy
e‚uents may originate from coagulated milk, cheese
curd ®nes, or ¯avoring ingredients such as fruit and
nuts (Brown and Pico, 1979). The nature of these
SS sources makes them predominantly organic.
This is con®rmed by the high mean VSS:SS ratio:
On average, about 76% of the SS were volatile,
even though the ratios varied over a wide range. TS
and VS levels also varied signi®cantly (Table 6). On
average, 52% of the TS were found to be volatile,
indicating that soluble inorganic constituents were
important in these waste streams.

Table 8. Volatile fatty acid (VFA) levels in composite wastewater samples
Plant

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Mean
SD
Min
Max

Total VFAs
(mg/l as HAc)
39
168
39
178
231
22
<1
284

124
246
126
431
90
134
90
147
115
<1
431

Acetate
(mg/l as HAc)
39
162
14
157
199
22
<1
257
65
179
86
356
90
134
70
122

99
<1
356

Propionate
(mg/l as HAc)
<1
6
12
22
17
<1
<1
19
7
66
12
71
<1
<1
<1
15
23
<1
71

Butyrate
(mg/l as HAc)
<1
<1

<1
<1
<1
<1
<1
8
5
<1
18
5
<1
<1
3
3
5
<1
18

Isobutyrate
(mg/l as HAc)
<1
<1
9
<1
8
<1
<1
<1
28
<1

6
<1
<1
<1
8
4
8
<1
28

Valerate
(mg/l as HAc)
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1

<1

Isovalerate
(mg/l as HAc)
<1
<1
4
<1
8
<1
<1
<1
19
<1
4
<1
<1
<1
8
3
5
<1
19


Dairy waste and biological nutrient removal

3563

Table 9. Concentrations of selected elements in composite wastewater samples

Plant
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Mean
SD
Min
Max

K (mg/l)
42.8
155.5
35.8
52.0
30.6
8.6
25.3
40.0

75.1
30.8
58.5
53.5
12.1
41.2
38.0
46.6
34.6
8.6
155.5

Na (mg/l)
735
263
419
306
517
453
713
410
485
310
734
1,265
463
423
660
544
252

263
1,265

Ca (mg/l)

Mg (mg/l)

47.7
58.0
52.3
58.4
58.5
33.6
42.8
46.1
58.7
52.2
54.2
56.7
1.4
54.3
58.0
48.9
14.9
1.4
58.5

11.4
46.3
11.0

24.6
25.7
16.9
47.2
14.1
18.8
17.5
11.8
27.0
6.5
8.3
27.0
20.9
12.4
6.5
46.3

Al (mg/l)

Mn (mg/l)

79
233
168
80
213
71
127
63
230

84
257
164
97
100
119
139
67
63
257

As discussed above, the use of acid and alkaline
cleaners and sanitizers in the dairy industry typically results in highly variable wastewater pH
values. All composite samples had pH values above
6.0, and most had pH values above 7.0 (Table 6).
Literature data indicated that pH values ranged
between 4.4 and 12.0, with an average of 7.2
(Table 10). Thus, wastewater pH values cited in the
literature extended over a larger range than pH
values measured for the 15 composite samples. This
di€erence can be explained because most literature
data were obtained from grab samples which were
analyzed individually rather than from a 24-h composite sample. Our data indicated that pH values of
composite samples collected over a 24-h time period
generally were near neutrality or basic; thus, the
large quantities of caustic soda used for cleaning
apparently had a greater impact on overall wastewater pH than the acids used for cleaning.

Ni (mg/l)

27
835
711
18
91
<1
<1
4
137
32
327
249
<1
9
6
163
268
<1
835

Cu (mg/l)

27
33
60
23
25
38
12
17

30
25
71
64
66
25
19
36
20
12
71

Co (mg/l)

<1
<1
10
<1
10
<1
<1
<1
10
10
30
30
20
<1
<1
8

11
<1
30

7
5
<1
<1
<1
<1
<1
3
<1
2
7
<1
<1
4
1
2
3
<1
7

Fe (mg/l)
720
579
137
520
1,910

91
43
39
439
777
4,329
566
420
247
63
725
1,102
39
4,329

Wastewater pH is a key factor in biological treatment because most microorganisms exhibit optimal
growth at pH values between 6.0 and 8.0 and most
can not tolerate pH levels above 9.5 or below 4.0.
Moreover, low pH wastewaters may cause corrosion of plant equipment, including components of
the treatment facility (Tchobanoglous and Burton,
1991). Equalization basins can be installed upstream
of biological treatment systems to stabilize wastewater pH. However, only 7 of the surveyed plants
had equalization basins. At times, equalization
basins alone are not sucient to compensate for the
extreme pH ¯uctuations in dairy waste streams.
This problem can be solved by collecting concentrated caustic wash water and sending it at a low
¯ow rate to the equalization basin (Samson et al.,
1985).
Brown and Pico (1979) consider slightly alkaline
dairy wastewaters (pH 7.5±8.5) desirable because

Table 10. Dairy plant wastewater characteristics obtained from the literaturea

Plant type

Total BOD5 Total COD Soluble BOD5 Soluble
(mg/l)
(mg/l)
(mg/l)
COD (mg/l) SS (mg/l)

Fluid milk 1,200±4,000 2,000±6,000
and cream
Fluid milk
1,390
2,120
Not given
680±4,500 980±7,500
Not given
830±1,980 1,250±3,160
Fluid milk
500±1,300 950±2,400
Not given
2,300
4,500
Fluid milk 1,670±2,200b,c 1,420±4,73b 640±1,100b,c
1,950
3,190
970
Mozzarella

2,075
4,400
450
cheese
Fluid milk 1,000±3,000 1,500±5,500
Not given
a

350±1,000

VSS (mg/l) FOG (mg/l)
330±940

300±500

pH

Total
alkalinity
(mg/l as
CaCO3)

Ref.e

8.0±11.0

150±300

1

300

650±2,290
1,050
960

b

2
3
3
4
4
5

90±450
820
220±3,000d
820
690

110±260
209
270±1,900d
690

5.0±9.5
7.2
3.5±12.0d
7.0±11.5

6

200±700
90±450
250

300±1,000
110±260
210

10.0±11.0
4.4±9.4

7
8

Data in italics are average values.
Data obtained from weekly composite samples.
Data obtained for BOD7.
d
Data obtained from daily composite samples.
e
1=Kasapgil et al., 1994; 2=Goronszy, 1989; 3=Kolarski and Nyhuis, 1995; 4=Ozturk et al., 1993; 5=Rusten and Eliassen, 1993;
6=Sobkowicz, 1986; 7=Anderson et al., 1994; 8=Eroglu et al., 1991.
b
c


3564

J. R. Danalewich et al.
Table 11. Dairy plant wastewater nutrient levels obtained from the literaturea

Plant type
Fluid milk and cream
Fluid milk
Not given
Not given
Fluid milk
Not given
Fluid milk
Fluid milk
Not given

Total P
(mg/l as P)

Orthophosphate
À
c
(mg/l as P)
NOÀ
3 (mg/l as N) NO2 (mg/l as N) TKN (mg/l as N) NH3 (mg/l as N) Ref.
20±50

32
18±162
15±30

3
20±350
10±110

4±15

16
4.4±12.1b
6.7b

4±6
9±210
48

<0.14

50±60
33
70±80
64
50±60
67±85
76

0.14
5
20

1
2

3
3
4
4
5
7
8

a

Data in italics are average values.
Data obtained from weekly composite samples.
c
References are the same as those in Table 10.
b

they help prohibit the development of hydrogen sul®de, assist in grease emulsi®cation, and aid in buffering biological treatment systems. In addition,
recommendations for an upper limit for wastewater
pH are believed to be unnecessary because neutralization of basic waste streams occurs naturally
through the absorption of CO2 gas into the wastewater, thereby lowering the pH. Neutralization of
dairy wastewater before biological treatment is considered necessary only if the ratio of total alkalinity
to BOD5 (expressed as mg/l CaCO3:mg/l O2) is
greater than 2 (Brown and Pico, 1979). In Table 6,
Table 12. Chemical characterization of whole milk and evaporated
milk

Parameter
pH
Total BOD5
Total COD

Total VFAs (as HAc)
Total Alkalinity (as CaCO3)
Total Nc
NH3 (as N)
NOÀ
3 (as N)
NOÀ
2 (as N)
Orthophosphate (as P)
Total P (as P)
Fat
TS
Total Volatile Solids
SS
VSS
Chloride
Potassium
Sodium
Calcium
Magnesium
Aluminum
Manganese
Nickel
Copper
Cobalt
Iron
a

Whole milk
concentration

(mg/l)
7.0a
102,500a
150,000a
200a
7,200b,d

1,000a
125,000a
117,000a
1,000b
1,500b
400b
1,200b
500b

Evaporated milk
concentrationc
(mg/l)
6.2
nd
364,790
<1
6,875
nd
0.3
95.5
66.0
1,426
5,667

nd
251,520
216,790
14,050
13,333
nd
2,887
485
1,582
180
4.0
0.1
2.3
1.1
0.4
6.9

Blanc and Navia, 1990.
Harper et al., 1971.
Papagiannis, 1996.
d
Assuming that whole milk contains 3% protein (and 88% water,
4% fat, and 5% lactose by weight [Goronszy, 1989]), each g of
protein has 0.24 g N, and assuming that the density of milk is
1 kg/l, it was calculated that whole milk contains 7,200 mg/l as
organic N.
b
c

total alkalinity values, BOD5 levels, and

alkalinity:BOD5 ratios are given for each composite
sample. The majority of the wastewater samples
had alkalinity:BOD5 ratios much below 2; only
plant 8 had a value above 2. Therefore, it is unlikely that neutralization of wastewaters would be
important.
The three common forms of P (orthophosphate,
polyphosphate, and organically bound P) are present in dairy processing e‚uents (Brown and Pico,
1979) and originate from cleaning compounds and
from milk or product spillage during processing.
Many facilities continue to use phosphate based
cleaners, usually in combination with nitric acid
based cleaners (Table 3), resulting in high levels of
P in most dairy wastewaters, as indicated by data
from our study and from the literature (Tables 7
and 11). The total P concentrations in our composite samples ranged from 29 to 181 mg/l, with an
average of 71 2 40 mg/l. Since P in dairy wastewaters is derived from both milk and phosphate
based cleaners, the high standard deviation re¯ects
variable operational procedures among plants in the
dairy industry. Orthophosphate concentrations in
the samples were relatively low, averaging
18 2 8 mg/l as P, and, on average, orthophosphate
P accounted only for 27% of the total P. Thus, the
remaining P was present in the organic and/or polyphosphate forms. P present in the organic and polyphosphate form is likely derived from milk, alkaline
cleaners, and emulsi®ers.
Based on information in Tables 1 and 3, it can be
calculated that plants 1, 3, 8, 9, and 13 estimated
that milk and milk products constituted 0.15%,
0.09%, 0.09%, 0.06%, and 0.11% (vol:vol) of their
waste stream. Assuming that these estimates are
correct and assuming that raw milk contains approximately 1,000 mg/l of total P (Table 12), the P

contribution to the wastewater due to milk should
be about 1 mg/l. Using the mean wastewater ¯ow
rates (Table 1) and the use of phosphoric acid for
cleaning (Table 3), the contribution of P from the
cleaning products to the wastewater can be estimated: cleaning products contributed 11, 14, 3, and


Dairy waste and biological nutrient removal

5 mg/l of P to the wastewater from plants 4, 9, 10,
and 14, respectively. The total P concentration in
the wastewater for these plants was 51, 79, 29, and
52 mg/l, respectively. Given this information, it is
clear that the amount of P attributed to milk
should be larger than 1 mg/l. This indicates that
more milk may have been lost than estimated by
plant operators. Thus, plant operators do not
appear to know the contribution of milk and milk
products to their waste streams.
Four di€erent N analyses were performed on the
15 composite samples: nitrate, nitrite, TKN, and
ammonia (Table 7). The organic N content was
estimated by subtracting the ammonia-N concentration from the TKN concentration (APHA,
1992). Nitrate concentrations in the composite
samples ranged from 0.6 to 80 mg/l as N (Table 7).
As discussed above, nitric acid was used frequently
as an alternative to phosphoric acid in cleaning operations. However, it is dicult to relate the use of
nitric acid to the levels of nitrate in the wastewater
since only limited information is available on nitric
acid usage (Table 3). A limited number of literature

studies also determined nitrate levels in dairy waste
streams (Table 11) and attributed high levels of
nitrate to the use of nitric acid. Nitrite concentrations in the composite samples generally were
low, except in one sample which also demonstrated
the highest nitrate level (Table 7).
TKN concentrations ranged from 14 to 140 mg/l
and averaged 91 2 39 mg/l (Table 7). Ammonia
concentrations were relatively low and ranged from
1.0 to 34.1 mg/l as N, with an average of
8.6 2 7.9 mg/l as N. This indicated that the majority
of TKN was present as organically bound N (e.g.,
proteins) and that the conversion of amino groups
to ammonia was incomplete. Again, the variation in
TKN data indicated variable operational procedures used in the dairy industry. Our results generally corresponded to the limited available data on
TKN and ammonia from the literature (Table 11).
As discussed above, plant operators estimated that
milk and milk products contributed between 0.06%
and 0.15% (vol:vol) to their waste stream.
Assuming that these estimates are correct and
assuming that raw milk contains approximately
7,000 mg/l of total N (Table 12), the N contribution
to the wastewater due to milk spillage should be
between 4 and 11 mg/l. This does not seem realistic
since wastewater TKN levels are greater than
100 mg/l for several of the plants. This information
again indicates that much more milk was being lost
than estimated by plant operators.
To evaluate the need for nutrient removal from
dairy wastewater, we estimated the levels of P and
N required for aerobic biological wastewater treatment assuming that BOD removal would be the

only objective (i.e., the levels of P and N required
for aerobic heterotrophic growth). We used kinetic
parameters derived by Orhon et al. (1993) from a

3565

biological wastewater treatment system treating a
milk processing plant wastewater. We further used
a solids retention time (SRT) of 5 days, an fD value
of 0.2 (i.e., the fraction of active biomass contributing to biomass debris, Grady and Daigger, in
press), and the measured in¯uent BOD5 concentrations (Table 6). Based on this information, we
estimated that P requirements for aerobic biological
treatment of the dairy waste streams ranged from 3
to 30 mg/l and averaged 10 2 7 mg/l (Table 7).
These estimates indicate that P was present at levels
much greater than needed for aerobic biological
treatment and that additional P removal would be
necessary to accomplish acceptable P levels for all
dairy plants included in our study.
Using the same parameters, the N required for
BOD removal was estimated to range from 14 to
148 mg/l and averaged 48 2 35 mg/l (Table 7).
Using these estimates it is apparent that there was
sucient N in the waste streams for biological
growth with the exception of plant 2. Thus, the
wastewater of plant 2 was N limited and would
require addition of N in order to achieve BOD5
and P removal, while all other wastewaters contained excess N.
To further evaluate the potential of BNR for
dairy wastewater treatment, we determined concentrations of acetate, propionate, butyrate, isobutyrate, valerate, and isovalerate in the composite

samples (Table 8). Acetate was the most abundant
of the VFAs in the wastewater samples, with an
average concentration of 122 mg/l (83% of the average total VFA concentration). Propionate was the
second largest contributor to the total VFA load of
the samples (10% of the average total VFA concentration). Butyrate, isobutyrate, and isovalerate were
detected at low concentrations and none of the 15
composite samples contained valerate at concentrations above the detection limit. Since VFAs are
absent from milk (Table 12), and only 1 plant
added acetic acid during processing (plant 11), the
presence of relatively high levels of VFAs is probably due to fermentation during cheese production
or in wastewater lines. The presence of VFAs, especially acetate, is bene®cial in the context of
EBPR, particularly since nitrate is present at high
levels in some of the waste streams (see below).
VFAs are more e€ective than other organic substrates for inducing and maintaining EBPR from
wastewaters (Hong et al., 1982; Abu-ghararah and
Randall, 1991; Okada et al., 1991; Shin and Jun,
1992; Randall et al., 1994; Carlsson et al., 1996).
Concentrations of selected elements (K, Na, Ca,
Mg, Al, Mn, Ni, Cu, Co, and Fe) for the 15 composite samples are listed in Table 9. The high average Na concentration of 544 mg/l is indicative of
the large amounts of Na+ used at these plants (e.g.,
NaCl, NaOH). In addition, substantial amounts of
K, Ca, and Mg were detected in the wastewater
samples. The elements Mg and K have been shown


3566

J. R. Danalewich et al.

to be required for successful EBPR (van

Groenestijn et al., 1988; Toerien et al., 1990;
Randall et al., 1992; Rickard and McClintock,
1992). According to Randall et al. (1992), 0.25 mol
of Mg and 0.23 mol of K are needed per mole of P
to be removed. Based on this information and the
total P, Mg, and K levels in the composite samples
(Tables 7 and 9), it was determined that Mg and K
were present in excess, except in the wastewaters of
plants 1, 8, 13, and 14, and 6, 7, 8 and 13, respectively. Thus, to prevent Mg and/or K limitations, it
will be necessary to add these elements to some of
the wastewaters to accomplish optimal EBPR conditions.
The concentrations of the other elements generally were low, with average values in the mg/l range.
Heavy metals, such as Cu, are inhibitory or toxic
for activated sludge microbial communities at concentrations around 1 mg/l (Hascoet et al., 1985;
Madoni et al., 1996; Beyenal et al., 1997). Since
these elements were only present at very low levels,
they should not be a concern for the biological
treatment of milk processing wastewaters.
Potential for biological nutrient removal
The above calculations show that P removal
beyond the removal due to aerobic heterotrophic
growth will be necessary to accomplish acceptable P
levels in dairy wastewater e‚uents. Various wastewater characteristics are instrumental in evaluating
the potential for success for BNR. Generally, the
potential to accomplish EBPR and to obtain an
e‚uent P concentration below 1 mg/l is considered
good if the in¯uent to the anaerobic zone of the
EBPR system has a BOD5:total P (TP) ratio greater
than 20:1 (Randall et al., 1992). Using the data in
Tables 6 and 7, it was calculated that the BOD5:TP

ratio for the 15 dairy wastewaters ranged from
3.1:1 to 77.3:1, with a mean of 32.3:1. The wastewaters of plants 4, 6, 7, 8, and 15 had BOD5:TP
ratios lower than 20:1. Since the COD:TP ratios for
4 of these 5 plants was also relatively low (below
40:1), it is likely that these wastewaters were COD
limited (Randall et al., 1992). Thus, VFAs or fermentable organics will need to be added to some
dairy wastewaters to accomplish complete P
removal.
Another important parameter is the organic matter to N ratio, which is generally expressed as the
COD:TKN ratio, the BOD5:TKN ratio, or the
BOD5:NH3-N ratio. These ratios are important to
evaluate the potential for successful N removal
since sucient amounts of organic matter should be
present to provide reducing equivalents for denitri®cation. In addition, these ratios are critical to evaluate EBPR potential, since poor P removal has been
attributed to the presence of nitrates in the anaerobic zone of the EBPR activated sludge system. A
number of studies have indicated that inhibition of
EBPR in the presence of nitrate is due to compe-

tition by denitrifying bacteria (e.g., Iwema and
Meunier, 1985; Yamamoto et al., 1990; Takeuchi,
1991; Carucci et al., 1994; Kuba et al., 1994). These
bacteria consume easily degradable organics (e.g.,
acetate) while reducing nitrate, leaving fewer VFAs
for P accumulating organisms (some of which are
also denitri®ers). Other studies have indicated that
the higher redox potential, due to the presence of
nitrate, may restrict P release in the anaerobic zone
(Barnard, 1976; Koch and Oldham, 1985; Schon et
al., 1993; Kortstee et al., 1994). In addition, chemical precipitation of phosphate has been suggested as
an explanation for the inhibition of phosphate

release in the presence of nitrate: phosphate precipitation may be induced by the increase in pH associated with denitrifying conditions (Kortstee et al.,
1994). In any case, controlling the nitrate concentration in the anaerobic zone is key for optimal
EBPR.
Grady and Daigger (in press) summarize general
guidelines for organic matter to N ratios for wastewaters that are amenable to complete biological N
removal. COD:TKN, BOD5:TKN, and BOD5:NH3N ratios greater than 9, 5, and 8 should result in
excellent N removal. Using Tables 6 and 7, the
mean COD:TKN, BOD5:TKN, and BOD5:NH3
ratios were calculated to be 36.3, 23.5, and 342.8,
respectively, and all plants had organic matter to N
ratios exceeding the minimum values necessary to
accomplish excellent N removal.
The presence of nitrate in dairy waste streams
adds a level of complexity commonly not encountered for domestic wastewaters. For most reactor
con®gurations, the wastewater in¯uent enters the
BNR treatment system in an anaerobic compartment. However, the presence of nitrates in the in¯uent
prevents
anaerobic
conditions
until
denitri®cation is complete. Dairy plants have several
options to accommodate for the presence of nitrate
in their wastewaters. One option would be to reduce
or eliminate the use of nitric acid as a cleaning
agent, decreasing the e€ect of nitrate on EBPR.
This option should be seriously considered by milk
processing plants when evaluating EBPR. As discussed above, several plants use nitric acid only
because of P surcharges and ®nes and prefer phosphoric acid based cleaners from a cleaning perspective. Thus, eliminating nitrate from their waste
streams may result in improved EBPR performance,
while resulting in better cleaning performance at a

lower cost. However, for some plants, elimination
of nitric acid may not be practical and wastewater
treatment must deal with the presence of relatively
high nitrate levels.
If nitrate were allowed to enter the anaerobic
compartment of the BNR treatment system, a large
portion, if not all, of the VFAs available may be
consumed rapidly for denitri®cation. It can be calculated that 2.86 g of organic matter (expressed as
COD) needs to be available to reduce 1 g of nitrate-


Dairy waste and biological nutrient removal

N. Assuming that all of these reducing equivalents
are provided by the VFAs in the wastewater, it can
be calculated using data from Tables 6 and 7 that a
surplus of VFAs is present in the wastewaters of
most plants (plants 2, 4, 5, 8, 10, 11, 12, and 13),
while all VFAs present in the in¯uent of the
remaining plants will likely be used for denitri®cation of the nitrate present in the in¯uent. The high
organic matter to N ratios for all wastewaters indicates that it is likely that sucient amounts of
rapidly biodegradable organic compounds are available to also allow for complete denitri®cation for
these wastewaters. However, it is not clear whether
the levels of rapidly biodegradable organic compounds that remain after denitri®cation are sucient to warrant successful EBPR. To answer this
question, laboratory and/or pilot scale evaluations
will be necessary. Such evaluations will determine
whether an external source of VFAs needs to be
supplied. In addition, such studies can evaluate
di€erent process con®gurations, such as assessment
of the value of separating a high-nitrate waste

stream from the rest of the wastewater and directing this stream directly to the anoxic compartment, in stead of to the ®rst, anaerobic
compartment of the BNR system.
CONCLUSIONS

This study showed that current waste volume
coecients in the dairy industry are 2 to 3 times
lower than those in the 1960s. This reduction is
attributed to increased plant size, automation in
product processing, introduction of clean-in-place
systems, and waste minimization practices such as
recovery of chemicals and reuse of rinse waters.
However, the wide variation in waste volume coecients and waste characteristics among plants indicate that it is still dicult to predict wastewater
¯ow rates and characteristics based on plant production volume. These ®ndings suggest that management strategies are still the determining factor in
waste generation. Plant operators also severely
underestimated the amount of milk and milk based
products in their waste streams, indicating that implementation of an accounting system for all waste
sources and chemical additions would greatly bene®t future pollution prevention e€orts. A comprehensive waste accounting system would be helpful,
but not sucient to eliminate waste treatment concerns. Most on site treatment facilities will need
substantial renovations and/or operational changes
to comply with present and future direct discharge
regulations, in particular to meet P limits. EBPR, in
combination with nitri®cation and denitri®cation,
will likely be a successful strategy to remove nutrients from dairy wastewaters. However, due to large
hourly, daily, and seasonal ¯uctuations in wastewater characteristics and ¯ow, BNR treatment systems should allow for ¯exible operation and

3567

extensive online control. Sequencing batch reactors
or newer designs (Goronszy, 1997) would allow for
an integration of online control and ¯exible operation, especially when treatment requires the cycling

of the wastewater through di€erent redox zones
(anaerobic, anoxic, and aerobic) as necessary for
BNR.
AcknowledgementsÐWe are thankful to the plant managers, operators, and environmental engineers of the participating milk processing plants; to Tim Bauchman, Jim
Royer, and Rod Meikamp from the Urbana-Champaign
Sanitary District (Urbana, IL) for access to laboratories
and help with ammonia and BOD5 analyses; to Rod
Mackie and Brian White from the Department of Animal
Sciences (University of Illinois) for access to laboratories
and help with VFA analyses; and to Ron Gerards and
Luc Vriens from Seghers Water (Wespelaar) and Luc
Ceyssens from Seghers Dinamec, Inc. (Austell, GA) for
helpful suggestions. This research was supported by grants
from the U.S. Department of Agriculture (95-37500-1911)
and Seghers Water.

REFERENCES

Abu-ghararah Z. H. and Randall C. W. R. (1991) The
e€ect of organic compounds on biological phosphorus
removal. Water Sci. Technol. 23, 585±594.
Anderson G. K., Kasapgil B. and Ince O. (1994)
Comparison of porous and non-porous media in up¯ow
anaerobic ®lters when treating dairy wastewater. Water
Research 28, 1619±1624.
APHA (1992) Standard Methods for the Examination of
Water and Wastewater, eds A. E. Greenberg, L. S.
Clesceri and A. D. Eaton, 18th edn. Am. Publ. Health.
Assoc., Washington, DC.
Backman R. C., Blanc F. C. and O'Shaughnessy J. C.

(1985) The Treatment of Dairy Wastewater by the
Anaerobic Up-Flow Packed Bed reactor, pp. 361±381.
40th Purdue Industrial Waste Conference, West
Lafayette, IN.
Barnard J. L. (1976) A review of biological phosphorus
removal in the activated sludge process. Water South
Africa 2, 136±144.
Beyenal N. Y., Ozbelge (Baser) Y. A. and Ozbelge H.
O. (1997) Combined e€ects of Cu and Zn on activated
sludge process. Water Research 31, 699±704.
Blanc F. C. and Navia R. (1990) Treatment of Dairy
Wastewater by Chemical Coagulation, pp. 681±689. 45th
Purdue Industrial Waste Conference, West Lafayette.
IN.
Borja R. and Banks C. J. (1994) Kinetics of an up¯ow anaerobic sludge blanket reactor treating ice-cream wastewater. Environ. Technol. 15, 219±232.
Brown H. B. and Pico R. F. (1979) Characterization and
Treatment of Dairy Wastes in the Municipal Treatment
System, pp. 326±334. 34th Purdue Industrial Waste
Conference, West Lafayette, IN.
Carlsson H., Aspegren H. and Hilmer A. (1996)
Interactions between wastewater quality and phosphorus
release in the anaerobic reactor of the EBPR process.
Water Research 30, 1517±1527.
Carucci A., Majone M., Ramadori R. and Rossetti
S. (1994) Dynamics of phosphorus and organic substrates in anaerobic and aerobic phases of a sequencing
batch reactor. Water Sci. Technol. 30, 237±246.
de los Reyes M. F., de los Reyes F., Hernandez M. and
Raskin L. Quantitative assessment of Gordona amarae
strains in foaming activated sludge and anaerobic diges-

3568

J. R. Danalewich et al.

ter systems using phylogenetic hybridization probes.
Appl. Environ. Microbiol. 64, in press.
Eroglu V., Ozturk I., Demir I., Akca L. and Alp K.
(1991) Sequencing Batch and Hybrid Anaerobic Reactors
Treatment of Dairy Wastes, pp. 413±422. 46th Purdue
Industrial Waste Conference, West Lafayette, IN.
Goronszy M. (1997) Industrial application of cyclic activated sludge technology. International Association on
Water Quality Yearbook 1997, 35±38.
Goronszy M. C. (1989) Batch Reactor Treatment of Dairy
Wastewaters: A Case History, pp. 795±805. 44th Purdue
Industrial Waste Conference, West Lafayette, IN.
Grady L. C. P. and Daigger G. T. (in press) Biological
Wastewater Treatment: Principles and Practice. Marcel
Dekker, Inc., New York.
Harper W. J., Blaisdell J. L. and Grosshopf J. (1971)
Dairy Food Plant Wastes and Waste Treatment
Practices. U.S. Environmental Protection Agency,
Washington, DC.
Hascoet M. C., Florentz M. and Granger P. (1985)
Biochemical aspects of enhanced biological phosphorus
removal from wastewater. Water Sci. Technol. 17, 23±
41.
Hong S. N., Krichten D. J., Kisenbauer K. S. and Sell R.
L. (1982) A Biological Wastewater Treatment System
for Nutrient Removal. EPA Worksop on Biological

Phosphorus Removal in Municipal Wastewater
Treatment, Annapolis, MD.
Iwema A. and Meunier A. (1985) In¯uence of nitrate on
acetic acid induced biological phosphorus removal.
Water Sci. Technol. 17, 289±294.
Jenkins D., Richard M. and Daigger G. T. (1993) Manual
on the Causes and Control of Activated Sludge Bulking
and Foaming. Lewis Publishers, Inc., Chelsea, MI.
Kasapgil B., Anderson G. K. and Ince O. (1994) An investigation into the pre-treatment of dairy wastewater prior
to aerobic biological treatment. Water Sci. Technol. 29,
205±212.
Koch F. A. and Oldham W. K. (1985) Oxidation±reduction potential: A tool for biological nutrient removal
systems. Water Sci. Technol. 17, 259±281.
Kolarski R. and Nyhuis G. (1995) The Use of Sequencing
Batch Reactor Technology for the Treatment of HighStrength Dairy Processing Waste, pp. 485±494. 50th
Purdue Industrial Waste Conference, West Lafayette,
IN.
Kortstee G. J. J., Appeldoorn K. J., Bonting C. F. C., van
Neil E. W. J. and van Veen H. W. (1994) Biology of
polyphosphate-accumulating bacteria involved in
enhanced biological phosphorus removal. FEMS
Microbiol. Review 15, 137±153.
Kuba T., Wachtmeister A., van Loosdrecht M. C. M. and
Heijnen J. J. (1994) E€ect of nitrate on phosphorus
release in biological phosphorus removal systems. Water
Sci. Technol. 30, 263±269.
Madoni P., Davoli D., Gorbi G. and Vescovi L. (1996)
Toxic e€ect of heavy metals on the activated sludge protozoan community. Water Research 30, 135±141.
Martin J. M. J. and Zall R. R. (1985) Dairy Processing
Wastewater Bioaugmentation ± An Evaluation of

E€ectiveness, pp. 351±360. 40th Purdue Industrial
Waste Conference, West Lafayette, IN.
Okada M., Murakami A., Lin C. K., Ueno Y. and Okubo
T. (1991) Population dynamics of bacteria for phosphorus removal in sequencing batch reactor (SBR) activated sludge processes. Water Sci. Technol. 23, 755±763.
Orhon D., Gorgun E., Germirli F. and Artan N. (1993)
Biological treatability of dairy wastewaters. Water
Research 27, 625±633.

Ozturk I., Eroglu V., Ubay G. and Demir I. (1993)
Hybrid up¯ow anaerobic sludge blanket reactor
(HUASBR) treatment of dairy e‚uents. Water Sci.
Technol. 28, 77±85.
Papagiannis T. G. (1996) Biological Nutrient Removal
from Dairy Processing Wastewater. M.S. thesis,
University of Illinois at Urbana-Champaign.
Randall C. W., Barnard J. L. and Stensel H. D. (1992)
Design and Retro®t of Wastewater Treatment Plants for
Biological Nutrient Removal. Technomic Publishing
Company, Lancaster, PA.
Randall A. A., Bene®eld L. D. and Hill W. E. (1994) The
e€ect of fermentation products on enhanced biological
phosphorus removal polyphosphate storage, and microbial population dynamics. Water Sci. Technol. 30,
213±219.
Reardon R. D. (1994) Cost Implication of Design for
Nutrient Removal, pp. 551±562. Water Environment
Federation 67th Annual Conference and Exposition,
Chicago, IL.
Rickard L. F. and McClintock S. A. (1992) Potassium
and magnesium requirements for enhanced biological
phosphorus removal for wastewater. Water Sci.

Technol. 26, 2203±2206.
Rusten B., Odegaard H. and Lundar A. (1992) Treatment
of dairy wastewater in a novel moving bed bio®lm reactor. Water Sci. Technol. 26, 703±711.
Rusten B. and Eliassen H. (1993) Sequencing batch reactors for nutrient removal as small wastewater treatment
plants. Water Sci. Technol. 28, 233±242.
Rusten B., Lundar A., Eide O. and Odegaard H. (1993)
Chemical pretreatment of dairy wastewater. Water Sci.
Technol. 28, 67±76.
Samson R., Van den Berg B., Peters R. and Hade C.
(1985) Dairy Waste Treatment Using Industrial-Scale
Fixed-Film and Up¯ow Sludge-Bed Anaerobic Digesters:
Design and Start-Up Experience, pp. 235±241. 40th
Purdue Industrial Waste Conference, West Lafayette,
IN.
Schon G., Geywitz S. and Mertens F. (1993) In¯uence of
dissolved oxygen and oxidation±reduction potential on
phosphate release and uptake by activated sludge from
sewage plants with enhanced biological phosphorus
removal. Water Research 25, 349±354.
Shin H. S. and Jun H. B. (1992) Development of excess
phosphorus removal characteristics in a sequencing
batch reactor. Water Sci. Technol. 25, 433±440.
Sobkowicz A. M. (1986) Celrobic1 Anaerobic Treatment
of Dairy Processing Wastewater, pp. 458±464. 41st
Purdue Industrial Waste Conference, West Lafayette,
IN.
Takeuchi J. (1991) In¯uence of nitrate on the bacterial
¯ora of activated sludge under anoxic conditions. Water
Sci. Technol. 23, 765±772.
Tchobanoglous G. and Burton F. L. (1991) Wastewater

Engineering: Treatment, Disposal and Reuse. McGrawHill, Inc., New York.
Toerien D. F., Gerber A., Lotter L. H. and Cloete T. E.
(1990) Enhanced Biological Phosphorus Removal in
Activated Sludge Systems. Plenum Press, New York,
London.
van Groenestijn J. W., Vlekke G. J. F. M., Anink D. M.
E., Deinema M. H. and Zehnder A. J. B. (1988) Role of
cations in accumulation and release of phosphate by acinetobacter strain 210A. Appl. Environ. Microbiol. 54,
2894±2901.
Yamamoto R. I., Komori T. and Matsui S. (1990)
Filamentous bulking and hindrance of phosphate
removal due to sulfate reduction in activated sludge.
Water Sci. Technol. 23, 927±935.