Lin S.D. Water and Wastewater Calculations Manual

Подождите немного. Документ загружается.

exposures that have taken place in the past, the adverse health effects

of which may or may not have already occurred. It also involves predic-

tion of the likely consequences of exposures that have not yet occurred.

20.2 Risk assessment

Risk assessment is a quantitative evaluation process of health or/and

environmental risks determining the potential risks associated with

exposure to a type of human hazard—physical, chemical, or biological.

There are four components to every (complete) risk assessment (US EPA,

1986):

1. Hazard identiﬁcation:

■

Review and analyze toxic data (animal and human studies, and

negative epidemiological studies).

■

Weigh the evidence that a substance causes various toxic effects.

■

Evaluate whether toxic effects in one setting will occur in other

settings.

2. Dose-response evaluation:

■

Perform an estimate of the quantitative relationship between the

amount of exposure to a hazardous identiﬁed substance and the

extent of toxic injury or disease.

■

Extrapolate from high dose to low dose.

■

Extrapolate test animals to humans.

Public Water Supply 515

TABLE 5.16 Annual Risk of Death from Selected Common Human Activities

Activities Individual risk, per year Lifetime risk*

Automobile accident 1/4500 1/65

Lightening 1/2000,000

Coal mining

Accident 1/770 1/17

Black lung disease 1/125 1/3

Truck driving accident 1/10,000 1/222

Falls 1/13,000 1/186

Home accidents 1/83,000 1/130

Others 1/1,000,000

Smoking

Flying

Drinking diet soda

Living near nuclear power plant

Living in stone or brick building

Use of microwaves

Eating 100 charcoal broiled steaks

*Calculated based on 70-year lifetime and 45-year work exposure

SOURCE: US EPA, 1986; R. Begole, State Farm Insurance, personal communication

3. Human exposure evaluation:

■

Investigate the inhalation, ingestion, and skin contact of a toxic

chemical.

■

Conduct case studies: how many people are exposed and through

which routes; what is the magnitude, duration, and time of exposure.

■

Perform epidemiological studies within different geographic areas.

4. Risk characterization:

■

Integrate the data and analyses of the above three evaluations to esti-

mate potential carcinogenic risk and noncarcinogenic health effects.

Example: In most public health evaluations, it is assumed that an individ-

ual adult or child consumes 2 or 1 L of water, respectively, each day through

all uses. The average body weights for a men, women, and children are

assumed to be 70 kg (154 lb), 50 kg (110 lb), and 10 kg (22 lb), respectively.

A toxic substance is present at 0.7 mg/L in water. Determine the daily dose

of the three groups for toxicological purposes.

solution:

Step 1. Calculate daily intake (DI) of toxic substance

For an adult

DI ⫽ 0.7 mg/L ⫻ 2 L/d

⫽ 1.4 mg/d

For a child

DI ⫽ 0.7 mg/L ⫻ 1 L/d

⫽ 0.7 mg/d

Step 2. Compute daily dose (DD) per unit weight

For a man

DD ⫽ (1.4 mg/d)/70 kg

⫽ 0.02 mg/kg ⋅ d

For a woman

DD ⫽ (1.4 mg/d)/50 kg

⫽ 0.028 mg/kg ⋅ d

For a child

DD ⫽ (0.7 mg/d)/10 kg

⫽ 0.07 mg/kg ⋅ d

Note: A child gets the highest dose.

516 Chapter 5

Dose-response models. The dose-response models for cancer death esti-

mation have been summarized (IUPAC, 1933; Klaassen, 2001). The

most frequently used ones are one-hit, Weibull, and multistage models

shown as follows:

One-hit (one-stage) model. The one-hit mechanistic model is based on the

somatic mutation theory in which only one hit of some minimum criti-

cal amount of a carcinogen at a cellular target for critical cellular inter-

action is required for a cell to be altered (become cancerous). The

probability statement for this model is

P(d) ⫽ 1 ⫺ exp(⫺bd) (5.202)

where P(d ) ⫽ the probability of cancer death from a continuous dose

rate, unitless

b ⫽ constant

d ⫽ dose rate, ␮g/L

bd ⫽ the number of hits occurring during a minute

Weibull model. The Weibull model is expressed as

P(d ) ⫽ 1 ⫺ exp(⫺ bd

) (5.203)

where k ⫽ critical number of hits for the toxic cellular response

Others are the same as the above.

Multistage model. Armitage and Doll (1957) developed a multistage model

for carcinogenesis that was based on these equations and on the hypoth-

esis that a series of ordered stages was required before a cell could undergo

mutation, initiation, transformation, and progression to form a tumor.

This relationship was generalized by Crump (1980) by maximizing the like-

lihood function over polynomials so that the probability statement is

P(d) ⫽ 1 ⫺ exp(⫺ (b

⫹ b

⫹

.........

⫹ b

)) (5.204)

where P(d) ⫽ the probability of cancer death from a continuous dose

rate, unitless

b’s ⫽ constants, (mg/kg ⋅ d)

–1

k ⫽ the number of dose groups (biological stages)

d ⫽ dose rate, ␮g/L

If the true value of b

is replaced with b

* (the upper conﬁdence limit

of b

), then a linearized multistage model can be derived where the expres-

sion is dominated by (bd*)d at low doses. The slope on this conﬁdence

Public Water Supply 517

interval, q

*, is used by US EPA for quantitative cancer assessment. To

obtain an upper 95% conﬁdence interval on risk, the q

* value (risk/⌬

dose in (mg/kg ⋅ d)

⫺1

) is multiplied by the amount of exposure (mg/kg ⋅ d).

Thus, the upper-bound estimate on risk R is calculated as

R ⫽ q

* (mg/kg ⋅ d)

⫺1

⫻ exposure (mg/kg ⋅ d) (5.205)

This relationship has been used to calculate a “virtually safe dose”

(VSD), which represents the lower 95% conﬁdence limit on a dose that

gives an “acceptable level” of risk (e.g. upper conﬁdence limit for 10

⫺6

excess risk). The integrated risk information system (IRIS) developed

by the US EPA gives q* values for many environmental carcinogens

(US EPA, 2000a). Because both the q

* and VSD values are calculated

using 95% conﬁdence intervals, the values are believed to represent

conservative, protective estimates.

The US EPA has utilized the linearized multistage model to calculate

“unit risk estimates” in which the upper conﬁdence limit on increased

individual lifetime risk of cancer for 70-kg human, breathing 1 ␮g/m

contaminated air or drinking 2 L/d of water containing 1 ppm (1 mg/L),

is estimated over a 70-year life span.

The human equivalent dosages are derived by multiplying the corre-

sponding animal dosages by (W

)

1/3

. Where W

and W

are average

body weight of men and test animals, respectively.

When the response and human equivalent dose data are ﬁt to the lin-

earized multistage model, the 95% upper limit on the largest linear

term is q

*. Then,

* ⫽ q

* (W

)

1/3

To estimate the 95% lower level of dose concentration, d, corresponding

to a 95% conﬁdence level of risk R,

R ⫽ 1 ⫺ exp(– dq

*) (5.206)

exp (– dq

*) ⫽ 1 ⫺ R

then d (in mg/kg ⋅ d) ⫽ (1/q

*) ⫻ ln(1 ⫺ R) (5.207)

To solve for d in ␮g/L with a 70-kg man drinking 2 L/d of water, a factor

is applied.

1 mg/kg ⋅ d ⫽1 mg/kg ⋅ d ⫻ 1000 ␮g/mg ⫻ 70 kg ÷ 2 L/d ⫽ 35,000 ␮g/L

then

d (in ␮g/L) ⫽ 35,000 ␮g(mg/kg ⋅ d)

–1

÷ q

*(mg/kg ⋅ d)

⫺1

⫻ ln(1 ⫺ R) (5.208)

518 Chapter 5

Example: The highest value of q

* for a dose-response study for vinyl chlo-

ride in the diet is 2.3 (mg/kg ⋅ d)

⫺1

. Determine the corresponding lowest con-

centration, d, when R is set as 10

⫺4

, 10

⫺5

, and 10

⫺6

solution: Using Eq. (5.208)

Step 1. Setting R ⫽ 10

⫺4

d (in ␮g/L) ⫽ (⫺35,000/2.3) ⫻ ln(1 ⫺ 10

⫺4

)

⫽ 6.09

Step 2. Setting R ⫽ 10

⫺5

d (in ␮g/L) ⫽ (⫺35,000/2.3) ⫻ ln(1 ⫺ 10

⫺5

)

⫽ 0.61

Step 3. Setting R ⫽ 10

⫺6

d (in ␮g/L) ⫽ (⫺35,000/2.3) ⫻ ln(1 ⫺ 10

⫺6

)

⫽ 0.061

Radionuclides in drinking water.

Radionuclides may occur naturally

and may be man-made (about 200% in number more than natural

ones). Natural radionuclides in drinking water supplies include

radium (Ra-226 and Ra-228), radon (Rn-222), uranium (U-238), lead

(Pb-210), polonium (Po-210), and thorium (Th-230 and Th-232). They

omit one or more of the three types of nuclear radiation, i.e. alpha,

beta, and gamma rays in air, food, and drinking water. The radiation

attacks bone, red bone marrow, gonads, breast, kidney, lung, skin

(burn), etc. Most radionuclides have long decay half-lives. For exam-

ple, the half-lives for U-238 and Ra-228 are respectively of 4.5 ⫻ 10

and 6.7 years. The radioactivity is expressed in curies (Ci). One curie

is the number of particles per second from one gram of radium. Dose

is the energy of a particle. One rad of dose is 100 ergs per gram of

energy deposited.

The MCLGs for radionuclides are set as zero. The MCLs for gross

alpha particle activity, combined with Ra-226 and Ra-228, and man-

made (approximately 200) radionuclides are 15 pCi/L, 5 pCi/L, and

4 millirem/yr. respectively. One pCi (pico Curie) is 10

⫺12

Ci.

The drinking water equivalent level (DWEL, in pCi/L) for radionu-

clides is deﬁned as (US EPA, 1986)

(5.209)DWEL 5

sNOAELd sanimalfactor, f

dsBWd

sUFdsWCdshumanfactor, f

Public Water Supply 519

where NOAEL ⫽ no observed adverse effect level

BW ⫽ body weight, 70 kg for adult male

UF ⫽ uncertainty factor, 10, 100, or 100

WC ⫽ daily water consumption, 2L/d for adult

Example: Determine DWEL of natural uranium for a man.

Given: NOAEL ⫽ 1 mg/kg ⋅ d, f

⫽ 0.01, f

⫽ 0.05, and UF ⫽ 100.

solution: Usually, BW ⫽ 70 kg and WC ⫽ 2 L/d are used for assessing risks.

Using Eq. (5.209)

20.3 Risk management

Risk management is deﬁned as decisions about whether an assessed risk

is sufﬁciently high to present a public health concern and about the

appropriate methods for control of a risk judged to be signiﬁcant. In other

words, it is the process of deciding what to do about the problems.

Risk management involves the following:

■

ﬁnd and decide the problems from risk assessment (type and concen-

tration of contaminants),

■

assume knowledge of health risks,

■

factor in the hydrology feasibility study and capital and operating

(site-speciﬁc) cost analyses,

■

decide and evaluate the measures (alternative control strategies) for

solving the problems,

■

reexamine exposure issues previously dealt with in risk assess-

ment, and

■

evaluate and monitor the results for the corrective measures for

meeting drinking water standards control strategies.

The basic categories of alternative control strategies may include source

control, treatment, combination of the above two, and short-term strate-

gies. The short-term control strategies are the use of bottled water, point-

of-use treatment (RO or ion exchange), or issuance of a boil water order.

The disadvantages of point-of-use units (in-home and workplace water

treatment) are maintenance upkeep problems and higher cost.

5 70 ␮g/L

5 0.07 mg/L 3 1000 ␮g/mg

s1 mg/kg

dds0.01ds70 kgd

s100ds2 L/dds0.05d

DWEL 5

sNOAELd sanimalfactor, f

dsBWd

sUFdsWCdshumanfactor, f

520 Chapter 5

Some bottled waters have questionable water quality (although most

meet MCLs) and are not labeled as to sources and treatment processes.

In foreign countries and even in the United States, some are bottled from

the tap water (public drinking water systems).

Several best management practices for performing in the watershed

may be used for controlling raw water source to reduce or eliminate con-

taminants. Water supply utilities may locate new sources of supply. If

so, the new sources can be blended with or replace the existing water

source. Interceptor well(s) upstream of the source water may be used to

protect water supply wells. The wastewaters from the interceptor wells

need proper disposal or treatment.

If the contaminant source is a leaking storage tank, ﬁx or remove the

tank; pump the well until contaminant levels drop, and then treat the

contaminated soil.

Treatment strategies. Treatment involves best available treatment tech-

niques to reduce contaminant to meet MCLs. Conventional treatment

processes combined with additional techniques can reduce contaminant

levels (Table 5.16; see US EPA, 1986):

a. Processes for inorganic removal

■

Conventional treatment

■

Lime softening

■

Iron exchange

■

Reverse osmosis

Public Water Supply 521

TABLE 5.16 Treatment Processes for Removal of Harmful Contaminants

Process Removes

Conventional As (V) at pH <7.5; Cd at pH > 8.5; Cr (III); Pb; Ag at

pH <8.0, pathogens, turbidity

Lime softening As ( V) at pH ⫽ 10 – 10.8; Ba at pH ⫽ 9.5 – 10.8 ; Cd;

Cr(III) at pH >10.5 ; Pb; Ag; F; V

Reverse osmosis All inorganics; commonly used for As (III), As(V); Ba, Cd,

Cr(III)’ Cr(VI); F, Pb, Hg, NO

, Se (IV), Se( VI); Ag, Ra, U

Ion exchange, cation Ba, Cd, Cr(III), Ag, Ra

Ion exchange, anion As(V), Cr(VI); NO

, Se(IV), Se( VI ), U

Activated carbon Volatile organics (benzene, vinyl chloride, carbon

tetrachloride, TCE, PCE, etc.), chlorinated aromatics (PCB,

dichlorobenzene), pesticides (aldicarb, chlordane, DBCP), DBP

Activated alumina As(V ), F, Se(IV)

Aeration Volatile organics, Rn

Air stripping Volatile organics, Rn

Boiling Pathogens, bacterias

Membrane (nano-) Pathogens, bacterias, DBP

■

Activated alumina

■

Electrodialysis

b. Processes for organic removal

■

Conventional treatment

■

Aeration (diffused air, packed column, slate-tray)

■

Adsorption (granular or powder activated carbon, resins)

■

Oxidation (disinfection)

■

Reverse osmosis

■

Biodegradation

■

Boiling

The speciﬁc processes can be added to the existing treatment

processes. An excellent overview and case studies, including capital and

operating costs of the above processes used for risk management and

health advisories, are presented in the US EPA (1986).

Example 1: Based on lack of signiﬁcant decrease in cholinesterase activity

in rats, the NOAEL for aldicarb sulfoxide in rats is 0.125 mg/kg ⋅ d. For HA

calculation purpose, EPA assumes:

Body weight, BW ⫽ 70 kg for a man

BW ⫽ 10 kg for a child

Daily water consumed, WC ⫽ 2 L/d for man

WC ⫽ 1 L/d for 10-kg child

Uncertainty factor, UF ⫽ 100 for use with animal NOAEL

This example illustrates how the US EPA determines the HA numbers for

pesticide aldicarb.

solution:

Step 1. Calculate 1-day health advisory (HA)

For the 10-kg child: Using Eq. (5.13)

Step 2. Determine 10-day health advisory

Since aldicarb is metabolized and excreted rapidly (>90% in urine alone

in a 24-h period following a single dose), the 1- and 10-day HA values would

5 0.012 mg/L or 12 ␮g/L

s0.125 mg/kg

dd s10 kgd

s100ds1 L/dd

One- day HA 5

sNOAELd sBWd

sUFd sWCd

522 Chapter 5

not be expected to differ to any extent. Therefore, the 10-day HA will be the

same as the 1-day HA (12 ␮g/L).

Step 3. Longer-term health advisory

For the 10-kg child:

Step 4. Determine lifetime health advisory

a. Calculate RfD using Eq. (5.13a)

*RRfD ⫽ risk reference dose: estimate of daily exposure to the human

population that appears to be without appreciable risk of deleterious

noncarcinogenic effects over a lifetime of exposure.

The lifetime health advisory proposed above reﬂects the assumption

that 100 % of the exposure to aldicarb residues is via drinking water.

Since aldicarb is used on food crops, the potential exists for dietary

exposure also, lacking compound-speciﬁc data on actual relative source

contribution; it may be assumed that drinking water contributes 20%

of an adult’s daily exposure to aldicarb. The lifetime health advisory for

the 70-kg adult would be 9 ␮g/L, taking this relative source contribu-

tion into account.

Example 2: There are no suitable data available to estimate 1-day, 10-day,

and long-term health advisories. The study of the subacute exposure

to trichloroethylene (TTC) via inhalation 5 days a week for 14 weeks

by adult rats identiﬁed a LOAEL 55 ppm (300 mg/m

). Derive the DWEL

for TTC.

Solution:

Step 1: Determine total absorbed dose (TAD)

where 300 mg/m

⫽ LOAEL

8 m

/d ⫽ volume of air inhaled during the exposure period

TAD 5

s300 mg/m

d s8 m

/dd s5/7d s0.3d

s70 kgd

5 7.35 mg/kg

5 0.00125 mg/kg

RfD* 5

NOAEL

0.125 mg/kg

100

5 0.012 mg/L s12 ␮g/Ld

Longer-term HA 5

s0.125 mg/kg

dd s10 kgd

s100d s1 L/dd

Public Water Supply 523

5/7 ⫽ conversion factor for adjusting from 5 d/week exposure

to a daily dose

0.3 ⫽ ratio of the dose absorbed.

70 kg ⫽ assumed weight of adult

Step 2: Determine RfD

where 7.35 mg/kg ⋅ d ⫽ TAD

100 ⫽ uncertainty factor appropriate for use with data

from an animal study

10 ⫽ uncertainty factor appropriate for use in conversion

of LOAFL to NOAEL

Step 3: Determine DWEL

where 0.00735 mg/kg ⋅ d ⫽ RRfD

70 kg ⫽ assumed weight of protected individual

2 L/d ⫽ assumed volume of water ingested by 70-kg

adult

The estimated excess cancer risk associated with lifetime exposure to

drinking water containing trichloroethylene at 260 ␮g/L is approxi-

mately 1 ⫻ 10

⫺4

. This estimate represents the upper 90% conﬁdence

limit from extrapolations prepared by EPA’s Carcinogen Assessment

Group using the linearized multistage model. The actual risk is unlikely

to exceed this value, but there is considerable uncertainty as to the

accuracy of risk calculated by using this methodology.

Example 3: A subchronic toxicity study of vinyl chloride monomer (VCM)

dissolved in soybean oil was administrated 6 days a week for 13 weeks, by

average to male and female wister rats. The NOAEL in the study was iden-

tiﬁed as 30 mg/kg ⋅ d.

Estimate the values of HA for VCM.

solution:

Step 1. Estimate 1-day HA

There are insufﬁcient data for estimation of a 1-day HA. The 10-day HA is

proposed as a conservative estimate for a 1-day HA.

DWEL 5

s0.00735 mg/kg

dds70 kgd

L/d

5 0.26 mg/L s260 ␮g/Ld

Rfd 5

s7.35 mg/kg

s100ds10d

5 0.00735 mg/kg

524 Chapter 5