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

Abstract—Fifteen 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 24h composite wastewater sample was characterized in detail for each plant. Wastewater flow 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 significant 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 fines. It was determined that most of the on site treatment facilities require renovations andor 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

CHARACTERIZATION OF DAIRY WASTE STREAMS,

CURRENT TREATMENT PRACTICES, AND POTENTIAL

FOR BIOLOGICAL NUTRIENT REMOVAL

J R DANALEWICH1, T G PAPAGIANNIS1 * M, R L BELYEA2,

M E TUMBLESON3 and L RASKIN1** M

1 Environmental Engineering and Science, Department of Civil Engineering, University of Illinois at

Urbana-Champaign, Urbana, IL 61801, U.S.A.; 2 Animal Sciences Department, University of Missouri-Columbia, Missouri-Columbia, MO 65211, U.S.A and 3 Department 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 regu-lations, 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 rela-tively 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

phos-phorus (P) and nitrogen (N) can result in

eutrophi-cation 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

im-plementation 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)

Enhanced biological phosphorus removal (EBPR) can be more cost e€ective than chemical precipi-tation 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), es-pecially 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 exten-sive literature study and a survey of 10% of the dairy plants in the U.S They concluded that the

Printed in Great Britain 0043-1354/98 $19.00 + 0.00

PII: S0043-1354(98)00160-2

*Author to whom all correspondence should be addressed.

[Tel: +1-217-3336964; Fax: +1-217-3336968/9464,

E-mail: lraskin@uiuc.edu].

3555

dairy industry had limited knowledge on the

or-ganic 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

in-formation for the rational design of treatment

facili-ties generally was not available In a report that

provides the perspective of the dairy industry

during the 1970s, Brown and Pico (1979)

summar-ized 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

waste-water treatment Most of these case studies, as well

as research e€orts, have been limited to

physico-chemical or anaerobic and aerobic biological

treat-ment, 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

nutri-ent 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

represen-tative 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

col-lected 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 H 2 SO 4 (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 bi-carbonate alkalinity, pH, 5-day biochemical oxygen demand (BOD 5 ), total solids (TS), volatile solids (VS), sus-pended solids (SS), volatile sussus-pended solids (VSS), ammo-nia, 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, pro-pionate, butyrate, isobutyrate, valerate, and isovalerate) (VFA) concentrations were measured by gas chromatog-raphy (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 ÿ48C over-night, 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 pro-ducts 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 pro-cessing 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 waste-water ¯ow rates Minimum wastewaste-water ¯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 as-sociated with washing procedures at the end of pro-duction 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 sur-vey 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 industry in general, and cheese producers in

particular, were 2.43 and 3.14 m3 wastewater/ton

milk processed, respectively Their analyses

indi-cated 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

auto-mation Based on these observations, Harper et al

(1971) concluded that management planning and

eciency of management supervision were the

con-trolling factors in the amount of wastewater

gener-ated 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

re-duction in volume of wastewater generated per

amount of milk processed However, the wide

vari-ation 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 oper-ations Therefore, personnel were asked to rate po-tential wastewater generating activities as either a major or minor contributor to total waste volume These results were used to assign an overall waste-water 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 gener-ated during cleanup of tanks, trucks, transport lines, and equipment Other sources of wastewater were associated with equipment malfunctions or op-erational 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 waste-water is generated during activities essential to plant maintenance (i.e., cleaning activities), the ranking provided in Table 2 can be used to priori-tize possible strategies to reduce wastewater volume and loading For example, some plants reused ®nal rinse waters for subsequent initial cleaning activi-ties, and several facilities recovered caustic soda All plants reported the presence of milk based substances in their wastewater (Table 3): of the 14

Table 1 Plant production and wastewater generation Milk processed

10 6 kg/day Products produced 10 6 kg/year (10 6 lbs/year) Wastewater ¯ow rate m 3 /day (10 3 gal/day)

1 0.9 (2.0) cheddar and Colby cheese 32

2 0.5 (1.1) cheddar and Colby cheese 17

(37) whey 22 (48) septic cheesesauce and

puddings (nr)

3 1.0 (2.1) cheddar, Colby, and Monterey

4 0.7 (1.5) cheddar cheese 24 (54) whey 13 (29) 1,105 (292) 643 (170) 1,605 (424)

5 0.5 (1.2) cheddar, Colby, and Monterey

6 na aseptic canning and cheese dips

7 0.7 (1.5) cheddar, Colby, Monterey Jack,

and reduced fat cheese 25 (55) whey 26 (58) 681 (180) 307 (81) 1,041 (275)

10 0.7±0.8 (1.5±1.8) cheddar cheese 22 (49) whey 20 (44) dried cheese

(nr) 719 (190) 416 (110) 871 (230)

11 na process cheese 91 (200) dried cheese 10 (22) 170 (45) 132 (35) 257 (68)

12 0.5 (1.1) mozzarella and provolone cheese

13 0.7 (1.5) cream cheese and related

products 44 (97) ¯avored snack dipsnon-dairy variety

5 (10)

208 (55) 4 (1) 1,450 (383)

14 0.9 (2.0) Parmesan, Romano, and

cheddar cheese (nr) alcohol 5,700 m

3 /yr (1.5  10 6 gal/yr) 2,081 (550) 1,703 (450) 2,650 (700)

na = not applicable.

nr = no value was reported.

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

contri-bution to wastewater volume and concentrations

(Harper et al., 1971), we asked personnel to

esti-mate 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

nutri-ent 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

charac-teristics 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, qua-ternary ammonium surfactants) are used in rela-tively low amounts, but are major contributors to the detergents' BOD5load 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 gen-erally 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, qua-ternary ammonium compounds, and acids are used

as sanitizers Harper and coworkers determined that wash waters containing sanitizer solutions con-tributed to 0.2 to 13.8% (average 3.1%) of the wastewater volume, whereas detergents were re-sponsible 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 2 Summary of wastewater generating activities

Number of plants regarding activity as

Cleaning of transport lines and equipment between production cycles 4 10 1

Milk and milk product discharge during production start up and change over 0 12 4

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

Table 3 Presence of milk based substances in wastewater as estimated by plant personnel and reported use of nitric and phosphoric acids Plant Milk m

3 /day

(gal/day) Whey m

3 /day (gal/day) Cheese HNO(lbs/day)3kg/day kg HNOHNO33coecient/10 6 kg milk H3PO(lbs/day)4kg/day kg HH33POPO44coecient/10 6 kg milk

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

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

re-frigeration 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

waste-water, we requested detailed information on

quan-tities 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

con-siderably throughout the industry and that

manage-ment 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

hydro-chloric 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

clean-ing 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

cleaners) had to be increased to improve equipment cleaning Waste minimization practices, such as rec-lamation 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 sys-tems 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 substan-tial savings in P surcharges and ®nes, the switch to nitric acid caused an increase in the amount of clea-ners 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 per-spective is consistent with the position of dairy plants in the 1970s: Brown and Pico (1979) dis-cussed that non-phosphate cleaners are not as e€ec-tive 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 con-siderably 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 tem-perature indicated that temtem-perature may be a con-cern 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 bio-solids is a major concern to the facilities, in particu-lar when biosolids have the potential to contain pathogens Nine plants did not separate domestic wastewater generated in the dairy facility from pro-cess wastewater Five of these plants pretreated their wastewater on site and thus generated

waste-water biosolids that contained pathogens of

poten-tial concern in biosolids disposal or reuse

appli-cations 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

pro-blems 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

oc-casional problem (a few times per year), while

plants 10 and 11 stated that activated sludge

foam-ing, caused by ®lamentous microorganisms, was a

persistent problem Furthermore, plants 10 and 11

indicated it was dicult to maintain adequate

dis-solved oxygen (DO) concentrations in their

acti-vated sludge tanks These observations may suggest

that low DO levels encouraged the growth of

®la-mentous organisms in these activated sludge

sys-tems 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 obser-vations that Nocardia foaming generally is not com-mon 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 regu-lations varied widely depending on discharge prac-tices 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 appli-cation was practiced, ¯ow rate, BOD5, total P, N (TKN), chlorides, and/or potassium concentrations generally were determined SS violations or sur-charges were reported most commonly; 7 plants fre-quently failed to comply with SS standards Plants

10 and 14 occasionally exceeded the allotted

maxi-Table 4 Wastewater temperature and pH; wastewater (pre)treatment strategy; sludge treatment and disposal strategy

Plant min max min max Wastewater (pre)treatment system b Sludge treatment strategy

1 3.0 11.0 nr nr pretreatment of main waste stream in equalization basin

and aerated lagoon; high-strength, low-volume waste

stream is land applied

occasional land application

2 3.0 13.0 32.0 43.0 treatment in equalization basin, DAF a , 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

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

3 nr nr nr nr 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

land application

6 4.5 12.0 nr nr no pretreatment; high-strength, low-volume waste stream

7 7.1 12.5 nr nr pretreatment in equalization basin; high-strength,

low-volume waste stream is land applied na

8 4.0 12.0 nr nr no pretreatment of dilute waste stream (land applied or

treated by city); pretreatment of concentrated waste stream in equalization basin, activated sludge system (NH 3 is added as N source), and oxidation ditch

nr

9 4.7 12.3 nr nr treatment in aerated lagoons, e‚uent used for irrigation

10 7.5 8.1 2.8 21.0 pretreatment in equalization basin and conventional

activated sludge system belt ®lter press dewateringand land application

11 1.0 14.0 14.0 32.0 pretreatment in equalization basin and completely-mixed

activated sludge system land application

14 4.8 11.3 22.0 38.0 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 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.

mum wastewater discharge volume, and BOD5

dis-charge 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

con-centrations

Long term data

Eight of the 15 plants provided data on

waste-water 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

com-monly measured and varied considerably The

avail-ability of wastewater characteristics for extensive

time periods is useful for determining seasonal

trends, which should help suggest improved

waste-water 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 com-posite 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 waste-water are derived from milk and milk products, some of the characteristics of whole milk are pre-sented 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 5 Wastewater characteristics for extensive time periods a Plant Time period Flow rate (10 3 gal/day) pH BOD 5 (mg/l) SS (mg/l) Total P (mg/l)

(600±10,000)

4 1/1/92±9/27/95 292243 (170±424) 8.421.6 (4.7±11.5) 7092139

(420±1,060) 6772544 (184±7,330)

6 1/1/95±12/31/95 143294 (29±1,444)

7 1/1/94±12/31/95 111231 (25±168) 11.321.3 (7.1±12.5) 1,2122684

(200±9,900) 9282305 (152±3,570) 78220 (31±227)

9 7/23/91±10/26/95

(excluding 1992) 8.321.6 (4.7±12.3) 2,29721,096(650±9,600) 1,08221,023 (293±13,700) 55225 (28±293)

(360±2,200) 6862378 (253±2,540) 37216 (14±104)

12 1/10/95±12/20/95 158214 (138±207) 7.721.8 (5.3±10.6) 1,7172708

14 12/28/94±8/1/95 508263 (189±677) 7.021.0 (5.0±11.0) 1,5452527

(288±5,200) 4052163 (110±1,050) 36214 (18±132)

a Each parameter is reported as mean2SD (min±max) for the indicated time period.

Table 6 Chemical characteristics of composite wastewater samples

Plant Total BOD(mg/l) 5Total COD(mg/l) BOD 5 /COD SS (mg/l) VSS (mg/l) TS (mg/l) VS (mg/l) pH

Alkalinity (mg/l as CaCO 3 )

Alkalinity/ BOD 5 (mg/l

as CaCO 3 / mg/l as O 2 )

nd = not determined.

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

calcu-lated 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

0.6320.16 (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

exten-sive set of BOD5:COD ratios obtained for milk

pro-ducts, 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

non-milk 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, indi-cating 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 7 Nutrient levels in composite wastewater samples and estimated levels of P and N required for BOD removal

Plant (mg/l as P)Total P Orthophosphate(mg/l as P)

P required for BOD removal a (mg/l as P) NO3

ÿ (mg/l as N) NO2

ÿ (mg/l as N) (mg/l as N)TKN (mg/l as N)NH3 (mg/l as N)Organic N

N required for BOD removal a (mg/l as N)

a See text for details on calculations.

Table 8 Volatile fatty acid (VFA) levels in composite wastewater samples Plant (mg/l as HAc)Total VFAs (mg/l as HAc)Acetate (mg/l as HAc)Propionate (mg/l as HAc)Butyrate (mg/l as HAc)Isobutyrate (mg/l as HAc)Valerate (mg/l as HAc)Isovalerate

As discussed above, the use of acid and alkaline

cleaners and sanitizers in the dairy industry

typi-cally 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

com-posite 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

waste-water pH than the acids used for cleaning

Wastewater pH is a key factor in biological treat-ment 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 cor-rosion 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 waste-water 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 concen-trated 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 9 Concentrations of selected elements in composite wastewater samples Plant K (mg/l) Na (mg/l) Ca (mg/l) Mg (mg/l) Al (mg/l) Mn (mg/l) Ni (mg/l) Cu (mg/l) Co (mg/l) Fe (mg/l)

Table 10 Dairy plant wastewater characteristics obtained from the literature a

Plant type Total BOD(mg/l) 5 Total COD(mg/l) Soluble BOD(mg/l) 5COD (mg/l) SS (mg/l) VSS (mg/l) FOG (mg/l)Soluble pH

Total alkalinity (mg/l as CaCO 3 ) Ref e Fluid milk

and cream 1,200±4,000 2,000±6,000 350±1,000 330±940 300±500 8.0±11.0 150±300 1

Fluid milk 1,670±2,200 b,c 1,420±4,73 b 640±1,100 b,c 650±2,290 b 220±3,000 d 270±1,900 d 3.5±12.0 d 5

Mozzarella

a Data in italics are average values.

b Data obtained from weekly composite samples.

c Data obtained for BOD 7

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.

they help prohibit the development of hydrogen

sul-®de, assist in grease emulsi®cation, and aid in

buf-fering biological treatment systems In addition,

recommendations for an upper limit for wastewater

pH are believed to be unnecessary because

neutral-ization of basic waste streams occurs naturally

through the absorption of CO2gas into the

waste-water, thereby lowering the pH Neutralization of

dairy wastewater before biological treatment is

con-sidered 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,

total alkalinity values, BOD5 levels, and alkalinity:BOD5ratios 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 unli-kely that neutralization of wastewaters would be important

The three common forms of P (orthophosphate, polyphosphate, and organically bound P) are pre-sent 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 compo-site samples ranged from 29 to 181 mg/l, with an average of 71240 mg/l Since P in dairy waste-waters 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

1828 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 phosphate forms P present in the organic and poly-phosphate 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 ap-proximately 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 esti-mated: cleaning products contributed 11, 14, 3, and

Table 11 Dairy plant wastewater nutrient levels obtained from the literature a Plant type (mg/l as P)Total P Orthophosphate(mg/l as P) NOÿ3 (mg/l as N) NOÿ2 (mg/l as N) TKN (mg/l as N) NH 3 (mg/l as N) Ref c

6.7 b

a Data in italics are average values.

b Data obtained from weekly composite samples.

c References are the same as those in Table 10.

Table 12 Chemical characterization of whole milk and evaporated

milk Parameter

Whole milk concentration (mg/l)

Evaporated milk concentration c (mg/l)

Total Alkalinity (as CaCO 3 ) 200 a 6,875

Total Volatile Solids 117,000 a 216,790

a Blanc and Navia, 1990.

b Harper et al., 1971.

c 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.