Pollutant Prioritization Project for Water Reuse

Christopher Stacklin, P.E.

Orange County Sanitation District, 10844 Ellis Avenue, Fountain Valley, CA 92708-7018, USA

*Email: cstacklin@ocsd.com.

ABSTRACT

The goal of the Pollutant Prioritization Project for Water Reuse is to develop a prioritized list of

pollutants of concern from an initial list of over 500 candidate constituents. The list will serve as

the basis for OCSD’s Enhanced Source Control Program that is designed to protect the

Groundwater Replenishment (GWR) System from potentially harmful pollutants that may impact

one of Orange County California’s most reliable sources of drinking water supply.

The Pollutant Prioritization Project will determine unit process removal rates and action levels

necessary to ascertain future Source Control restrictions in regards to the GWR System project

and to satisfy the California Regional Water Quality Control Board Order. Pollutants identified

will be used to implement various control strategies first, at an operational level and second at a

source control level. Source control strategies may include voluntary reductions or local limits

for point sources and public outreach or product substitution for non-point sources, etc.

Methodology and results are presented herein.

KEYWORDS: Pollutant Prioritization, Water Reuse, Groundwater Replenishment System,

Predictive Model, Fault Tree Analysis, Removal Rate, Action Level, Emerging Pollutants,

Kalman Filter, Proactive

INTRODUCTION

The GWR System, which is a joint project of Orange County Water District and OCSD, is

considered one of the largest water reuse project of its kind in the world. The project is designed

to ultimately reuse approximately 45,620 million gallons per year of secondary treated

wastewater provided by OCSD Reclamation Plant No. 1. The GWR System employs state-of-

the-art technology including microfiltration, reverse osmosis, and advanced oxidation processes

to produce purified water that will supply approximately 2.5 million people in Orange County

for recharge, irrigation, domestic and industrial uses, and to protect Orange County Groundwater

Basin from further degradation due to seawater intrusion.

Although the GWR System has many technology-based safeguards to reduce the risk of

contamination of the water it produces, source control has been recognized by regulatory

agencies as an additional enhancement to the technology-based safeguards. As a result, OCSD

has been mandated by the California Regional Water Quality Control Board to expand its

existing source control program to include emerging pollutants of concern specified by the

California Department of Public Health as harmful to human health and drinking water supplies

of Orange County.

WEFTEC®.08

1138

With this magnitude of water reuse, the challenge is to execute a comprehensive methodology

that involves evaluating and ranking approximately 500 constituents with about 75,000 water

quality parameters, along with deriving unit level removal rates from each treatment process to

ultimately establish a Priority Pollutant

list and implement an effective source control program.

A Priority Pollutant

is defined herein as a constituent that has the potential to adversely impact

the GWR System product water quality, including those constituents which, while not presently

known to contribute to use impairments or to show increasing loadings or concentrations, have

characteristics that indicate a potential to impact physical or biological integrity. These

characteristics include detectable presence in the product water or watershed, ability to

bioaccumulate, persistence, and toxicity.

Initial List of Priority Pollutant

Candidate Constituents

The list of 500 constituents is composed of constituents with numerical limits or standards

(regulated constituents) and constituents that do not have numerical limits or standards, but may

have toxicity values (unregulated constituents). The basis for the list is shown in Figure 1.

For the GWR System, regulated constituents are derived from all applicable laws, ordinances,

• Local Limits

• Primary and Secondary

• Maximum Contaminant

Levels

• Regulatory Action

Levels

• Maximum Contaminant

Level

• Archived Advisory

Level

• Notification Level

• Title 22 Draft

Groundwater Recharge

Reuse

• Order No. R8-2004-002

or Later

Regional

Water

Quality

Control

Board

California

Department

of Public

Health

Orange

County

Sanitation

District

EPA

Figure 1. Referenced sources for regulated constituents.

WEFTEC®.08

1139

regulations, and statutes. For example, Regional Water Quality Control Board (RWQCB) Order

No. R8-2004-002 – Producer/User Water Recycling Requirements has recycled water quality

specifications which include regulatory limits for about 101 pollutants.

A common attribute of regulated constituents is that toxicity information is highly developed. In

contrast, unregulated constituents have moderate to no toxicity information and are referenced in

a limitless variety of sources.

Unregulated constituents were selected based on their potential for future regulation. Therefore,

the sources that identify constituents of this nature must have a formal screening process,

whether based on peer review or preliminary toxicological data. Inherently, constituents which

go through this process have a higher likelihood of being regulated. This being said, unregulated

constituents are derived from the sources in Figure 2.

• Orange County Water

District

• Orange County

Sanitation District

• Advisory Panel

• Contaminant Candidate

Lists

• Priority Pollutants /

307(a) Toxics

• Public Health Goals

• Order No. R8-2004-002

Monitoring & Reporting

Program

Regional

Water

Quality

Control

Board

Office of

Environment

al Health

Hazard

Assessment

DistrictEPA

Figure 2. Referenced sources for unregulated constituents.

The initial list of candidate constituents is typically large. The initial list developed for the

pollutant prioritization project contained 19 chemical classes of over 500 constituents.

Implementing a source control program at this stage would consume a tremendous amount of

resources to sample, monitor and to establish source control of over 500 constituents for multiple

point and non point sources including industrial, domestic, and commercial entities. Therefore,

methodology for prioritization of this initial list of candidate constituents is the key.

WEFTEC®.08

1140

Process Description

Wastewater from sources including industrial, commercial, domestic users, and urban runoff

throughout Orange County is received by Reclamation Plant No. 1. The flow is diurnal and

ranges from 110 MGD at the daytime peak to 40 MGD at night. The flow undergoes primary

treatment through chemically enhanced primary treatment (CEPT), then secondary treatment

through either trickling filter or air activated sludge systems.

The secondary effluent flows across Reclamation Plant No. 1 battery limits into the GWR

System, through screening facilities where particles 2 millimeters and larger are separated and

removed. The removal of particles reduces fouling of the microfiltration membrane system

located downstream.

OCWD Groundwater

Replenishment System

OCSD Reclamation

Plant No. 1

Chemically

Enhanced

Primary

Treatment

Wastewater

From

Collection

System

Tricking

Filters

Micro-

filtration

Membrane

Reverse

Osmosis

System

Advanced

Oxidation

Process

Air

Activated

Sludge

System

Product

Water to

Settling

Basins &

Seawater

Barrier

Protects RO

0.2 µm Pore Size

85% Recovery

Molecular Weight

Cut-Of f 300-500

Daltons

UV Light

Peroxide

Time Dependent

Ferric Chloride

Anionic Polymer

Mean Cell

Residence Time

0.9-1.4 Days

4 Hour Hydraulic

Retention time

Figure 3. Reclamation Plant No. 1 and the GWR System.

The screened water flows across an 86 MGD microfiltration system. The membrane elements of

the microfiltration system remove particles in the effluent of 0.2 microns or higher which

includes bacteria, protozoa, and suspended solids. The purpose of the microfiltration system is

to reducing fouling of the reverse osmosis system.

The filtered water from the microfiltration system is then pumped across a 70 MGD reverse

osmosis (RO) system with an 85% recovery rate. Cleaned water that passes through the RO

system is called, permeate. Wastes that cannot pass through the RO system are concentrated into

an RO system reject stream.

WEFTEC®.08

1141

The RO permeate flows across a 70 MGD two-step advanced oxidation process (AOP). First,

hydrogen peroxide is added to the RO permeate upstream of an ultraviolet (UV) light treatment

system. Second, UV irradiation is used for disinfection and reduction of light-sensitive organic

contaminants.

The disinfected RO permeate has a high level of carbon dioxide and a low pH after treatment

from the UV system. The disinfected RO permeate is routed through decarbonator towers which

use blown air to strip off carbon dioxide and other dissolved gases in the water. Lime is added to

both raise the pH and add salts back into the water. The added salts prevent the highly purified

water from leaching minerals from the piping used to transport the water to settling basins. The

degassed water is pumped across a dechlorination skid and becomes product water.

Nearly all of the product water from the GWR System will be placed in the Orange County

groundwater basin in two ways. First, some of the product water will be injected into existing

and/or new injection wells designed to prevent seawater intrusion in the Talbert Gap located in

the Huntington Beach, Fountain Valley, and Costa Mesa areas. Second, most of the product

water will be pumped to OCWD’s Santa Ana River water spreading basins in Anaheim and

Orange. From there, the product water will percolate into the groundwater basin.

METHODOLOGY

The methodology for identifying Priority Pollutants

is developed with the intent of establishing

a proactive Source Control program. Proactive means that constituents with the potential to

cause non compliance with regulatory requirements or adverse human health affects will be

ranked high. It also means that the methodology must be predictive. The predictive aspect of

the methodology is achieved using probabilistic fault tree and constituent sensitivity analyses.

After establishing the initial list of Priority Pollutant

candidate constituents, the constituents

need to be strategically ranked using a risk analysis approach. There are two criteria that Priority

Pollutants

must satisfy to achieve high ranking. The constituent must: 1) be detected in the

effluent of the treatment process or have the potential to break through the treatment process and

2) have a toxicological effect, regulatory limit, or regulatory standard. To begin, constituents are

organized into two groups, those that are regulated with numerical limits or standards and those

which are unregulated.

CASE 1: Prioritizing Regulated Constituents

Regulated constituents have existing, risk-based toxicity values. Figure-4 illustrates the method

of what essential steps were utilized to prioritize the regulated constituents. Regulated

constituents which are not detected in the secondary effluent of Reclamation Plant No. 1 or the

GWR System product water are placed on a “watch list” where they are routinely monitored for

detection.

Removal rates are calculated for regulated constituents that are detected in the secondary effluent

of Reclamation Plant No. 1 or the GWR System product water. A removal rate is the fraction or

percentage of the influent constituent mass loading that is removed from the waste stream across

an entire wastewater treatment works or specific wastewater treatment unit within the works

WEFTEC®.08

1142

(U.S. EPA, 2004). Constituents with negative removal rates are grouped together for further

study. Action levels for Source Control (action level

) are determined for constituents with

Prioritize Constituent

Determine Action Levels

Apply Removal, Toxicity

Apply Analytical Results

Apply Regulatory Criteria

Define Sample Panel

Negative

Removal

Rate

Not Detected

Watch List

Further

Study

Below

Action Level

Positive

Removal

Rate

Detected

With

Numerical

Limits

Sample

Panel

Priority

Pollutant SC

Above

Action Level

Watch List

Figure 4. Methodology for regulated constituents.

positive removal rates. An action level

is a level above which triggers a Source Control action,

such as source identification. An action level is calculated from a probabilistic model such as a

fault tree or may be set intuitively. Constituents equal to or above the action level become

Priority Pollutants

. Constituents that are below the action level are placed on a watch list.

CASE 2: Prioritizing Unregulated Constituents

Unregulated constituents typically have moderate to unknown toxicity information, making them

difficult to prioritize. The key to prioritize these constituents is to perform a risk assessment

using simplified or detailed risk-based approaches.

Several detailed risk-based evaluation approaches are available including: Thresholds of

Toxicological Concern (Cramer, Ford, & Hall, 1978) (Kroes, et al., 2004), Calculated Margins of

Exposure, FDA Pregnancy Categories, FDA Threshold of Regulation (U.S. FDA, 2005). The

proper use, application, and refinements of these methods are still a topic of discussion among

the research community (WateReuse Foundation, 2008). Implementation of these approaches

can be a considerable effort both in time and resources and may require a domain expert or panel

of experts.

WEFTEC®.08

1143

While these detailed risk-based evaluation approaches are still being refined by the research

community and regulators, a simplified approach is proposed in the meantime. It is important to

note that this simplified approach is intended to be used only as an interim measure until

researchers and regulators can determine toxicity using detailed risk-based evaluations.

The simplified approach is based on limiting toxicological quires to a small number of on-line

sources, e.g., Integrated Risk Information System (U.S. EPA), Agency for Toxic Substances and

Disease Registry (ATSDR) (U.S. DHHS), National Center for Toxicological Research (U.S.

FDA), Maximum Recommended Therapeutic Dose (MRTD) Database (U.S. FDA) and

Endocrine Disruptor Knowledge Base (EDKB) (U.S. FDA), per Figure 5.

• A factual database focusing on the toxicology of over 5,000 potentially

hazardous chemicals

Hazardous Substances Data Bank HSDB

• A database from the U.S. EPA containing carcinogenic and non-carcinogenic

health risk information on over 500 chemicals

Integrated Risk Information System IRIS

• This database contains data in support of human health risk assessments

compiled by TERA

International Toxicity Estimates for Risk ITER

• A scientifically evaluated and fully referenced data bank, developed and

maintained by the NCI

Chemical Carcinogenesis Research Information System

CCRIS

• The Endocrine Disruptor Knowledge Base consists of a number of scientific

resources, including a biological activity database, QSAR training sets and

computational models to predict estrogen and androgen activity

Center for ToxicoinformaticsEDKB

Figure 5. Partial listing of referenced sources for toxicity data.

Some of these sources can be accessed through the Toxicology Data Network (TOXNET)

(National Library of Medicine). TOXNET is a cluster of databases covering toxicology,

hazardous chemicals, environmental health and related areas. It is managed by the Toxicology

and Environmental Health Information Program (TEHIP) in the Division of Specialized

Information Services (SIS) of the National Library of Medicine (NLM).

Toxicological concentrations which have human or biological effects are collected by querying

the on-line sources for each constituent without regulatory limits or standards, based on oral

consumption, e.g., drinking two liters of water per day for the average adult. In some cases, a

high dose to low dose extrapolation may be required (U.S. EPA, 2005). Each concentration is

WEFTEC®.08

1144

converted to a common unit basis and the minimum of these concentrations selected as the

controlling case.

Prioritize Constituent

Determine Action Levels

Apply Removal, Toxicity

Apply Analytical Results

Apply Regulatory Criteria

Define Sample Panel

Numerical

Limits

Detected

Toxic

Positive

Removal

Above

Action

Level

Below

Action

Level

Watch List

Toxic

Negative

Removal

Further

Study

Not Toxic

Concern

No Toxicity

Data

Concern

Not

Detected

Concern

Sample

Panel

Priority

Pollutant

Figure 6. Methodology for unregulated constituents.

Unregulated constituents are dropped from the prioritization list if they are: 1) not detected in the

effluent stream; 2) detected but are not toxic; 3) or have no available toxicity data.

Removal rates are calculated for unregulated constituents that are detected and have sufficient

toxicity data. Constituents that have negative removal rates are listed for further study. Action

levels

are calculated for constituents with positive removal rates.

Constituents that are below the action level

are added to the watch list. Constituents which are

equal to or greater than the action level

become Priority Pollutants

. This methodology is

summarized in Figure 6.

Water Side Mass Balance Model

The mass balance model is shown in Figure 7 and encompasses only the water side. Note that

sample points were established at the influent flow from the collection system to Reclamation

Plant No. 1, the secondary effluent from Reclamation plant No.1 to the GWR System, and the

WEFTEC®.08

1145

Figure 7. Mass balance for back calculating action levels.

GWR

System

Product

Water to

Recharge

Basins

Secondary Treated

Effluent To

Ocean Discharge

(Via Plant No. 2)

Raw

Sewage

From

Collection

System

Brine Return

(Microfiltration + Reverse Osmosis)

To Treatment Plant No. 2

Reclamation Plant

No. 1

 Primary Treatment

 Secondary Treatment

¾Activated Sludge

¾Trickling Filters

GWR System

 Microfiltration

 Reverse Osmosis

 Advanced Oxidation

GWR System Brine Return

GWR System Product Water

(

)

iREGiEffGWR

CMinC

−

MGDF

EffGWR

70=

−

iEffGWREffGWRiEffGWR

CFm

34.8

−−−

×=

iEffGWRiInfGWRijGWR

mmm

,,,Re −−−

−

EffGWRInfGWRjGWR

FFF

−−−

−=

jGWR

ijGWR

,Re

34.8

−

GWR System Influent

100

−

iGWR

iEffGWR

iInfGWR

InfGWR

iInfGWR

−

34.8

MGDF

InfGWR

86=

−

Reclamation Plant No. 1 Influent

Reclamation Plant No. 1 Secondary Effluent to Ocean

iInfGWRiSEOcean

,, −−

InfGWREffRPSEOcean

FFF

−−−

−=

iSEOceanSEOceaniSEOcean

CFm

34.8

−−−

××=

SEOceanInfGWRInfRP

FFF

−−−

100

,,1

−

avgiRP

iSEOcean

iInfRP

InfRP

iInfRP

−

34.8

listithenCCIf

calculatediInfRPmeasurediInfRP

⇒>

−−

:,:

:,1:,1

Action Level for Priorty Pollutants

WEFTEC®.08

1146

Definition of Equation Variables and Subscripts

C Constituent concentration in milligrams per liter

F Flow Rate in million gallons per day

m Mass flow rate in pounds per day

µ Removal efficiency

Subscripts

GWR,i GWR System of constituent, i

GWR-Inf GWR System influent

GWR-Inf,i GWR System influent of constituent, i

GWR-Eff GWR System effluent

GWR-Eff,i GWR System effluent of constituent, i

GWR-Rej GWR System brine return to Treatment Plant No. 2 across j discrete

sampling events

GWR-Rej,i GWR System brine return to Treatment Plant No. 2 across j discrete

sampling events of constituent, i

Ocean-SE Ocean secondary effluent

Ocean-SE,i Ocean secondary effluent of constituent, i

REG,i Numerical limit or standard of constituent, i

RP1-Eff Reclamation Plant No. 1 secondary effluent

RP1,i,avg Reclamation Plant No. 1 average of constituent, i

RP1-Inf Reclamation Plant No. 1 influent

RP1-Inf,i Reclamation Plant No. 1 influent of constituent, i

RP1-Inf,i:calculated Reclamation Plant No. 1 influent of constituent, i, calculated

RP1-Inf,i:measured Reclamation Plant No. 1 influent of constituent, i, measured

Figure 8. Equation variables and subscripts for the mass balance equations

product water flow from the GWR System to the recharge basins or points 1, 3, and 5,

respectively. Flow rates were recorded for the influent and effluent streams and secondary

effluent (interior) stream identified in the mass balance.

The approach for the mass balance model uses the GWR System regulated contaminant levels or

toxicity data to back calculate the allowable influent wastewater to Reclamation Plant No. 1.

The allowable influent wastewater concentration to Reclamation Plant No. 1 is compared with

the actual average influent wastewater concentration. This approach is formally called, “Back

Calculating Action Levels”.

Start with a mass balance around the GWR System bounded by points 1, 2, and 3 in Figure 7.

The design flow rating, water side removal rates, and regulated contaminant levels or toxicity

data are known. Therefore, the allowable influent concentration to the GWR and brine return

streams can be solved at points 3 and 2, respectively.

After closing the mass balance around the GWR System, a mass balance can be made around

Reclamation Plant No. 1 which is illustrated by points 3, 4, and 5 in Figure 7.

Since the water side removal rate and mass flow rate across Reclamation Plant No. 1 are already

known, the Reclamation Plant No. 1 allowable influent concentration can be determined.

WEFTEC®.08

1147

The allowable influent wastewater concentration to Reclamation Plant No. 1 can be compared

with the actual average influent wastewater concentration. If the actual average influent

wastewater concentration is equal to or above the allowable influent wastewater concentration,

the constituent will be placed on the prioritized list of emerging pollutants.

Fault Tree Analysis

Note that by varying the removal rate across the GWR System and comparing the change in

removal rate with the change in constituents appearing above the action level that constituents

can be assessed.

The removal rate across the GWR can also be adjusted relative to degraded operating states

determined in a fault tree analysis. Fault tree analysis is a probabilistic failure analysis in which

an undesired state of a system is analyzed using Boolean logic to combine a series of lower-level

events (Vesely, W.E., et al. 1981). An undesired state in this case would be contaminant

breakthrough.

DISCUSSION

Project Scope

The scope of project involves the following sampling plan:

• Sample sites include Reclamation Plant No. 1 influent, influent and effluent from the

primary treatment processes, and influent and effluent from the secondary treatment

processes and the GWR System product water;

• Samples are initially collected over a two-year period from 2007 through 2009 and taken

each day of the week, and repeated for three consecutive weeks.

• Although automated samplers collect most samples over a twenty-four hour period,

volatile organics will be collected by hand at four-hour intervals over a twenty-four hour

period.

Sample analysis is performed at an unprecedented scale. The sample analysis group will test for

about 500 chemicals and physical properties of wastewater. This will include: 1) 3 classes of

pesticides, 2) a large variety of industrial chemicals, 3) polychlorinated biphenyls, 4) the

standard group of priority pollutants, 5) flame retardants, 6) several hormones and several

pharmaceuticals, 7) NDMA and 8 other nitrosamines, 8) a suite of 19 metals, 9) mercury, 10)

total suspended solids, 11) turbidity, 12) and total organic carbon. At the end of the project, the

team will have approximately 75,000 separate measurements of water quality parameters.

Sample results will be paired with mass balance flows to determine removal rates across each

unit operation. Influent flow characterization and unit operation removal rates can be compared

with both regulatory and operation –based constituent limits of the GWR System product water

to identify constituent candidates for the prioritized list of pollutants.

In consideration of large scale of the pollutant prioritization project, strategic planning and

method development are required for the success of the expanded Source Control program. Key

WEFTEC®.08

1148

facets are: 1) develop an initial list of constituents to analyze; 2) establish a methodic approach to

prioritize constituents; 3) apply the method to the GWR System; and 4) report the results.

Pollutant Prioritization Timeline

The pollutant prioritization project consists of three initial phases of sampling followed by a

routine sampling program. The phases are based on the changing status of the GWR System

from its construction to routine operation.

• Reclamation Plant No. 1

•In Service

• GWR System

• Under Construction

• Reclamation Plant No. 1

•In Service

• GWR System

• Under Commissioning

•Reclamation Plant No. 1

•In Service

•GWR System

•In Service

•Integrated Plant

•Quarterly Priority Pollutants

•Annual Comprehensive

Figure 9. Pollutant prioritization project timeline sequence.

Note that the GWR System was under construction during the time of the Phase I sampling.

Because of this, representative samples of the GWR System product water were not available to

determine removal rates across the GWR System. However, because the methodology discussed

in this paper uses a probabilistic approach, it was also used to predict Priority Pollutants

during

the construction phase of the GWR System. At the time of this paper that Phases I and II were

completed and Phase III is in progress, per Figure 9.

RESULTS

Priority Pollutants

List

In total, 502 constituents were analyzed in Phases I and II. The results show that there were no

violations of numerical limits or standards and that the GWR System is in 100% compliance of

Order No. R8-2004-002.

WEFTEC®.08

1149

233 constituents were detected in either the secondary effluent of Reclamation Plant No. 1 or the

GWR product water. 117 of the 233 constituents detected had numerical limits or standards. 32

had positive removal rates and 17 constituents were equal to or above action levels. Therefore,

17 regulated constituents are Priority Pollutants

1,4-Dichlorobenzene

1,4-Dioxane

Antimony

bis(2-Ethylhexyl)phthalate

Boron

bromodichloromethane

Formaldehyde

Iron

Manganese

Nickel

Nitrate

Nitrite

n-Nitrosodimethylamine

Perchlorate

Total Dissolved Solids

tert-butyl alcohol

Total Nitrogen

17 Regulated Constituents

Figure 10 - Priority Pollutants

derived from regulated constituents.

23 regulated constituents are on the watch list. 216 constituents had negative removal rates and

are slated for further investigation. 164 constituents are of no concern. 179 unregulated

constituents were detected. At the time of this paper, a toxicity assessment still needs to be

applied to these constituents to prioritize them.

WEFTEC®.08

1150

1,1,2,2-tetrachloroethane

1,1,2-trichlorotrifluoroethane

1,2-dibromoethane

1,3-dichlorobenzene

Aldicarb

Asbestos Fibers >10µm

Baygon

Captan

Carbaryl

Carbophenothion

Diazinon

Dimethoate

Diphenenamid

Ethion

gamma-BHC

isopropylbenzene

Malathion

n-butylbenzene

n-Nitrosodi-n-propylamine

Parathion-ethyl

Parathion-methyl

Pentachloronitrobenzene

tert-butylbenzene

23 Regulated Constituents

Figure 11 – Watch list derived from regulated constituents.

Note that the lists presented in Figures 10 and 11 still need to be finalized based on the Phase III

results and the results of the toxicity assessment and are expected to change.

Evaluation of Quality of Results

The average molecular weight of constituents that were detected downstream of RO and AOP

range from 10.8 to 390 Daltons (about 2 to 13 angstroms) which is consistent with the molecular

weight cut-off of 300-500 Daltons for the RO membrane system. During this time, the RO

permeate average transmittance was 98 percent and the average effluent turbidity was 0.04 NTU.

The AOP UV dose averaged 585 mJ/cm

, which exceeds the 50 mJ/cm

minimum regulatory

requirement.

Phases I and II results are compared with outside data from similar treatment systems shown in

Figure 12. Dataset 1 is from a plant which has microfiltration, reverse osmosis and ultraviolet

light treatment processes. Datasets 2 and 3 are from plants which have microfiltration, reverse

osmosis and advanced oxidation treatment processes. The comparison shows good agreement

between Phases I and II and the three datasets with the exception of Aluminum, diethylphthalate,

and fluorene.

WEFTEC®.08

1151

CNA

CNACNA

CNA

CNACNA

CNA

CNACNA

OK OK OK

OK OK OK OK

OK OK

OK OK OK OK

OK OK OK

OK OK OK OK

OK OK

CNACNA

OK OK

OK OK OK OK

OK OK

CNA

OK OK

CNA

OK OK

CNA

OK OK

CNACNA

OK OK

CNA

OK OK

CNACNA

OK OK

CNA

OK OK

CNA

Figure 12 – Phases I&II results compared with results from other POTW datasets.

WEFTEC®.08

1152

Unit Process Removal Rates

Removal rates for eight constituents are shown in Figure 13. The data shows significant

removals averaging 60 percent occur across the secondary treatment system with the exception

of n-Nitrosodimethylamine (NDMA) in which about 70 percent is removed in the primary

system. Significant removals in the secondary and especially the GWR System are to be

expected.

-20

100

Primary Secondary GWR

Figure 13. Removal rates of unit processes.

1,4-dioxane, acetaminophen, and caffeine have slightly negative removals across the primary

treatment system, but this appears to be error of measurement. Therefore, there is essentially no

removal of these constituents across the primary treatment system. Perchlorate has a negative

removal rate of 15 percent across the GWR, but this is attributed to an inconsistency in method

detection limits as two different laboratories were used at the plant battery limits. Note that

perchlorate is well below the regulatory limit as are the balance of constituents in the figure.

Because the levels of detection are very low, secondary parameters such as removal rates

become more sensitive to change. Also, please keep in mind that the data reflects only Phases I

and II and statistical significance may yet have to be achieved.

Negative Removal Rates

It is important to note that constituents with negative removal rates may indicate that a byproduct

is being formed within the treatment processes or an undesired constituent is being inadvertently

added to the treatment processes via chemical addition. As an example, a polysorbate wetting

agent may be used regularly by operations and maintenance on the RO membranes of the GWR

WEFTEC®.08

1153

System. The specifications of the solution based on a technical bulletin show that up to 0.5 mg/L

of 1,4 dioxane may be present in the solution, per Figure 14.

Figure 14. Specifications for polysorbate wetting agent for membranes.

Sample results from a treatment system using this chemical may show as non detected at the

influent, but detected at the effluent. Therefore, constituents with negative removal rates need to

be further investigated. If a constituent is being formed or added within the treatment processes,

then it is up to Plant Operations and Maintenance to mitigate formation.

Dispersion of Data for Emerging Pollutants

Analytical data in Figure 15 shows that concentrations are more dispersed for emerging

pollutants versus conventional pollutants. As an example, the average influent concentration of

iron, a conventional pollutant is 110 micrograms per liter with a standard deviation of 14

micrograms per liter. Note that the standard deviation represents 13 percent of the average.

In contrast, salicylic acid, a pharmaceutical constituent, has an average of 50 micrograms per

liter and a standard deviation of 54 micrograms per liter. In this case, the standard deviation is

108 percent of the average.

The dispersion of analytical results of this magnitude indicates that the domain space for

pharmaceuticals and personal care products is a stochastic time series, akin to stock market data.

The dilemma is: How does one develop a predictive Source Control program based on data of

this nature?

In the stock market, analysts have applied Moving Average and Kalman filters to model their

data. Both are of a family of techniques used to analyze time series data with the intent of

smoothing out short term data and thereby revealing long term trends.

The benefit of these methods as applied to emerging pollutants is that they are recursive

estimators, meaning that only the estimated state from the previous time step and current

measurement are used to estimate the current state. For example, if a new pharmaceutical is

introduced on the market and does not do well in its first few years, but does extremely well in

subsequent years, the history of observations will be biased and incorrectly reflect lower influent

concentrations.

WEFTEC®.08

1154

100

120

140

160

180

200

Concentration, µg/L

Observations

Acetaminophen Caffeine Salicylic acid

Figure 15. Dispersion of results of select pharmaceuticals and personal care products.

A recursive estimator such as the Kalman filter will more accurately model the trend, allowing

source control to predict whether to implement source control measures for the new

pharmaceutical or other emerging pollutant. Details of the Moving Average and Kalman filter as

applied to this pollutant prioritization effort will be discussed in another paper.

CONCLUSIONS

Overall, this technical risk assessment provided a prioritized list of 17 constituents for source

control starting from an initial list of over 500 constituents in 19 groupings. From this, point and

non point sources will be identified and classified. Next, a source control strategy will be

selected such as targeted public outreach or voluntary reduction. Lastly, the source control

strategy will be implemented and the effectiveness evaluated.

The pollutant prioritization project for water reuse provides a simple, but practical methodology

for ranking constituents with and without numerical limits or standards using a probabilistic

approach. While the methodology may not be ideal, improvements will be made on a continual

basis. This methodology is presented for the benefit of others which may find themselves in

similar situations.

WEFTEC®.08

1155

ACKNOWLEDGMENTS

I gratefully acknowledge the following people for their leadership and hard work which made

this effort possible.

• Mahin Talebi, Source Control Manager

• Jerry Evangelista, Source Control Supervisor

• Ron Wade, Chief Plant Operator, and Staff

• Sam Mowbray, Laboratory Manager, and Staff

REFERENCES

American Chemical Society. (n.d.). CAS Registry

. Retrieved 2008, from Chemical Abstract

System: http://www.cas.org/

Cramer, G., Ford, R., & Hall, A. R. (1978). Estimation of Toxic Hazard-Decision Tree

Approach. Food and Cosmetics Toxicology, 16:255-276.

Kalman, R.E. (1960). "A new approach to linear filtering and prediction problems". Journal of

Basic Engineering 82 (1): 35–45.

Kroes, R., Renwick, A., Cheeseman, M., Kleiner, J., Mangelsdorf, I., Piersma, A., et al. (2004).

Structure-Based Thresholds of Toxicological Concern (TTC): Guidance for Application

to Substances Present at Low Levels in the Diet. Food and Chemical Toxicology, 42:65-

83.

National Library of Medicine. (n.d.). TOXNET. Retrieved 2008, from Toxicology Data

Network: http://toxnet.nlm.nih.gov/

Sorenson, H.W. (1970); Least-Squares estimation: from Gauss to Kalman, IEEE Spectrum, V. 7,

pp. 63-68

Stacklin, C. (2008). Emerging Pollutant Prioritization, Phase I and II Reports. Fountain Valley:

Orange County Sanitation District.

U.S. DHHS. (n.d.). Agency for Toxic Substances and Disease Registry. Retrieved 2008, from

U.S. Department of Health and Human Services: http://www.atsdr.cdc.gov/

U.S. EPA. (2005). BOSC Workshop on EPA Chemical Risk Assessment Principles and

Practices. Board of Scientific Counselors Workshop. Washington DC: U.S. EPA.

WEFTEC®.08

1156

U.S. EPA (2004), Local Limits Development Guidance and Appendices, EPA 833-R-04-002A

and -002B, Office of Wastewater Management 4203.

U.S. EPA. (n.d.). Integrated Risk Information System. Retrieved May 7, 2008, from U.S.

Environmental Protection Agency: http://cfpub.epa.gov/ncea/iris/index.cfm

U.S. EPA. (n.d.). Water Quality Handbook Chapter 4: Antidegradation (40 CFR 131.12).

Retrieved 2008, from U.S. Environmental Protection Agency:

http://www.epa.gov/waterscience/standards/handbook/chapter04.html

U.S. FDA. (n.d.). Endocrine Disruptor Knowledge Base. Retrieved 2008, from Center for

Toxicionformatics EDKB:

http://www.fda.gov/nctr/science/centers/toxicoinformatics/edkb/index.htm

U.S. FDA. (n.d.). http://www.fda.gov/cder/Offices/OPS_IO/MRTD.htm. Retrieved from U.S.

Food and Drug Administration: http://www.fda.gov/cder/Offices/OPS_IO/MRTD.htm

U.S. FDA. (n.d.). National Center for Toxicological Research. Retrieved 05 07, 2008, from U.S.

Food and Drug Administration: http://www.fda.gov/nctr/index.html

U.S. FDA. (2005). Threshold of Regulation for Substances Used in Food. Retrieved from U.S.

Food and Drug Administration: http://www.cfsan.fda.gov/~dms/torguid.html

Vesely , W.E., F.F. Goldberg, N.H. Roberts, D.F. Haasl (1981), Fault Tree Handbook, NUREG-

0492, Systems and Reliability Research, Office of Nuclear Regulatory Research, U.S.

Nuclear Regulatory Commission, Washington, DC 20555

WateReuse Foundation. (2008). WRF Expert Meeting to Recommend Topics for Detailed

Analysis. WRF 06-004: Identifying Health Effects Concerns of the Water Reuse Industry

and Prioritizing Research Needs for Nomination of Chemicals for Research to

Appropriate National and International Agencies. Fountain Valley: WateReuse

Foundation.

WEFTEC®.08

1157