ARTICLE IN PRESS

WAT E R R E S E A R C H

40 (2006) 175 181

Available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/watres

A Bayesian method to estimate proportional payments of

users in a wastewater treatment plant
Victor M. Gomez-Munoza, Miguel Angel Porta-Gandarab,, Jaime Cervantes de Gortaric
a

Centro Interdisciplinario de Ciencias Marinas, Instituto Politecnico Nacional, Apartado Postal 592, La Paz, Baja California Sur. Mexico 23090,
Mexico
b
Centro de Investigaciones Biologicas del Noroeste, PO Box 128, La Paz, Baja California Sur. Mexico 23000, Mexico
c
Facultad de Ingeniera, UNAM, Coyoacan, Mexico 04510, Mexico

art i cle info

A B S T R A C T

Article history:

A simple Bayesian method to estimate proportional payments for using or constructing a

Received 15 January 2003

wastewater treatment plant is presented. The method addresses the usually unknown

Received in revised form

individual participation in global pollution, of municipal and industrial wastewaters that

28 January 2005

are normally discharged into the same collecting pond. The individual contributions are

Accepted 8 November 2005

calculated from the discharged volume of each source and its pollution percentage in the

Available online 20 December 2005

wastewater volume previous to the disposal. According to the method presented here, the

Keywords:

costs in building and/or running the plant can be covered by proportional individual

Wastewater

payments based on the individual contributions, in a more even way. The simplicity of the

Pollution

model allows to assess different strategies for the participants in a wastewater treatment

Bayes

system. A real case study in the Izta-Popo National Park, located in the Mexican States of

Water remediation

Puebla, Mexico and Morelos, is analyzed.

& 2005 Elsevier Ltd. All rights reserved.

1.

Introduction

Wastewater treatment has become a most important issue in

recent years due to the dangerous presence of industrial,
agricultural and urban polluting materials in water bodies.
Removal or destruction of these pollutants has become a
governmental priority around the world. The design, operation and performance of wastewater treatment systems,
depends on the sources of pollution and the technology used,
the latter usually representing very high costs. In order to
sustain any potential waste treatment process it is necessary
that the financial support does not depend exclusively on
public funds.
Water is also one of the main issues of the environmental
movements around the world, demanding an effective
cleansing of wastewater. A large protest occurred in the
Corresponding author. Tel.: +52 612 1238484; fax: +52 612 1255070.

E-mail address: maporta@cibnor.mx (M.A. Porta-Gandara).

0043-1354/$ - see front matter & 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.watres.2005.11.005

beginning of the eighties against the pollution in land waters

caused by the Finnish forest industry (Konttinen, 1998).
Waste water management continuously requires new
strategies and useful tools to establish the degree of private
and public participation. Potential solutions that consider
multicriteria evaluation of waste management options are
available considering financial effectiveness and environmental impact. For instance, Powell (1996) studied this kind
of evaluation for some waste disposal possibilities that could
be adapted to the problem of sludge and dirty waters, in order
to compare several scenarios and cost criteria.
Merret (2000) on the other hand, studied some European
cases, relevant from the regulatory perspective of the wastewater management. The proposal of wastewater policies,
mainly from industrial and municipal sources has been
presented in recent years. Authorities of different countries

ARTICLE IN PRESS
176

WA T E R R E S E A R C H

have introduced more strict regulatory industrial wastewater

policies.
Pollutants that directly affect public health, like high
concentrations of heavy metals or organic matter, have a
special consideration in any wastewater management system. New technologies to facilitate detoxication, nutrient
removal and ultrafiltration of wastewater, including degradation of phenolic waste waters, are encouraged by governmental policies in order to find reasonable, economic and
efficient solutions.
Huang et al. (1997) in an integral waste management
system case study was formulated by means of the concepts
of gray systems and gray decisions, in order to optimize
processes and their resulting solutions in the long-term
planning of a water treatment facility. The gray continuous
variables represent waste flows along the routes between the
municipalities and the waste management facilities. The
results indicate that reasonable solutions can be generated
through mathematical modeling and useful criteria for
decision-makers in wastewater generation and management,
can be proposed.
Bayesian analyses have been employed in water treatment
processes modeling. For instance, Haag et al. (2003), derived a
criterion for maximizing the likelihood of measured data with
respect to given parameters with no a priori knowledge of
them, in a anaerobic waste treatment process, where a large
number of unknowns were present.
Borsuk and Stow (2000), on the other hand, described a
generalized version of the biochemical oxygen demand decay
model, useful in water quality management plans that
include the impact of municipal and industrial wastes on
the dissolved oxygen concentration of receiving waters. Their
approach used Bayes theorem to develop a joint probability
distribution for all parameter values conditional on the
observed data. They considered it as a more appropriate
representation of the aggregation of underlying processes that contribute to overall oxygen consumption in
organic wastes.
Qian et al. (2003) compared Bayesian Monte Carlo and
Markov Chain Monte Carlo methods using a low-dimensional
biochemical oxygen demand decay model as an example, and
demonstrated that the former is extremely inefficient because the prior parameter distribution is generally a poor
substitution for the posterior parameter distribution.
The general experience with appropriate collection systems
for urban areas of several countries has demonstrated that
systems must be designed regarding each particular condition. Detailed data from some case studies show that solid
waste collection systems, for instance, are not automatically
transferable from one community to another, like the study in
South Africa presented by Smith-Korfmacher (1997).
The costs of building and operation of a wastewater plant
depends on its size and the type of treatment used, like in the
case of the suburbs of Paris where the three main treatment
plants differ of each other (Servais et al., 1999).
Whatever is the cost of the operation of a treatment plant,
or its maintenance and/or construction, it must be proportional to the demand of service by each user. This paper
contributes directly to the design of costs and fee policies for
water discharging that can be used in wastewater treatment.

40 (2006) 175 181

The method herein presented, can be employed to include

economic participation of all the involved actors in the
pollution problem of a common water discharge situation,
prior to treatment.
The most accessible data for this problem are the volume of
water discharged by each participant and the individual
proportion of pollution in the water, previous to the discharge. Nevertheless, the contribution to pollution in a global
context is more difficult to estimate, when the discharges are
mixed before the treatment of the water. This is the case, for
instance, of municipal water collectors.
A simple Bayesian method is proposed to calculate the
individual payments of all contributors, including the municipality, that discharge wastewater in a common pond before
its treatment. System dynamics complexities or real-time
variations are not considered. Nevertheless, a real case study
was tested revealing the appropriateness of this method, by
using the original raw datawithout any manipulation
measured in the Izta-Popo National Park in Mexico.

2.

Description of the method

The approach presented here is based on Bayes theorem of

probability theory. This theorem establishes that given a
sample space associated to a certain statistical experiment,
which is partitioned in n events, called the causeseach of
these with a known non-zero probabilityand given a certain
effect of interest which its conditional probability is known as
a function of the occurrence of each one of the causes, then,
when that effect is observed, it is possible to calculate the
probability of each one of the causes.
Symbolically, if U is the sample space (universe), Ai i
1; . . . ; n are the events that represent the causes, and E is the
event that represents the effect of interest, it is assumed that
U

n
[

Ai

with Ai \ Aj + when iaj

i1

and also the probabilities P Ai 40 and P E j Ai are known for

each i 1; . . . ; n. Then, Bayes theorem establishes that (Berger,
1985)
PE j Ak PAk
.
PAk j E Pn
i1 PE j Ai PAi
In the case of a wastewater treatment plant, the universe is
the total of wastewater flow to be processed, Ai is the
wastewater discharged by a participant or contributor i, and
effect E the polluted water. The probability value P(Ai)
corresponds to the proportion of participant i in the overall
discharged water and P(E | Ai) is the proportion of polluted
water discharged by participant i, as a function of the specific
water treatment that is carried out by this participant, before
it discharges. Therefore, according to the theorem, P(Ak | E) is
the probability that the polluted water comes from participant k. This probability is proposed here as the evenly
proportional payment of the cost of the treatment plant for
the participant k.

ARTICLE IN PRESS
WAT E R R E S E A R C H

3.

Applications

The method provides useful information to the wastewaterutility management of the participating contributorsincluding the municipality if it participateswhich wastewater
contain a specific pollutant proportion to be processed in a
global treatment plant. Their individual contributions are
estimated from samples taken from individual discharges
prior to the sewage treatment plant.
Diverse strategies to treat wastewater previous to the
discharge can be evaluated using this Bayesian method.
Some strategies are analyzed by using a hypothetical example
and a real case study, in a sequential order to estimate
changes, improvement techniques or any new technology to
treat wastewater prior to discharge.
Results are applicable to any economic issue related to the
use of a wastewater treatment plant: service, maintenance or
construction, under the general assumption that water
sampling from the collecting pond represents pollutants
presence from some discharge contributors.
For the sake of clarity in the following example only three
contributors are initially considered. It is assumed that
contributors do not mix their wastewaters before discharging,
so the causes of pollution among participants are independent of each other. In this way, the discharges from these
three participants represent a partition of the total water
discharged, and the theorem conditions are satisfied. The
following percentages of discharged wastewater from these
contributors are given as
Participant 1: 50%,
Participant 2: 30%,
Participant 3: 20%.
Strategy 1. Water treatment before discharging: It is assumed
that participant 1 possesses a treatment plant and as a result,
its discharged water is only 20% polluted. The preliminary
process in participant 2 is not so efficient so it discharges 40%
polluted water. Participant 3 does not have any prior
treatment of wastewater, so it discharges water 90% polluted.
These data represent the following probabilities:
PA1 0:5;

PA2 0:3;

PE j A2 0:4;

PA3 0:2;

PE j A1 0:2,

40 (20 06) 17 5 181

177

contributes the most to the pollution. Therefore, it would

have to participate economically in a larger proportion than
the other participants, to the payment of the common
treatment service. In Table 1, the used data are shown in
the first three columns and the result of the method is
presented in the last one.
Strategy 2. Diminishing the pollutant concentration: When a
contributor dilutes the pollutant density, adding fresh water
prior to the dischargefor instance, if contributor 3 discharges twice to diminish the concentration of the pollutant
in halfits contribution remains the same according to this
method (Table 2). The water misused to diminish the
pollutant concentration does not represent any benefit for
the dishonest participant, and there is an evident social
damage.
Strategy 3. Using again contaminated water: Suppose that with
the initial conditions of Strategy 1, participant 3 maintains its
waters untreated before the discharge. Moreover, suppose
that it prefers to use again the already contaminated water,
resulting in a reduction of its discharge in half, and increasing
the level of contamination, lets say from 90% to 100%. In such
a case its contribution is the same as participant 1, and
greater than participant 2 (Table 3).
Strategy 4. Waste waters free of pollution: Strategies 2 and 3 are
certainly inconvenient as they give misleading and unreliable
results. If instead of these options, participant 1the largest
oneinstalls a perfect system, by means of which its wastewaters are free of pollution, the method demonstrates that
the contributions of the other two companies rise, drastically
for the participant 3 (Table 4).
Strategy 5. A new participant with zero-pollution: This scenario
has the same initial conditions of Strategy 1, but includes a
new participant that does not contaminate, and has a flow
discharge equal to participant 3. The results show that the
contributions remain the same for the original contributors,
nevertheless the more distributed percentages of water
consumption (Table 5). This result is theoretically predictable,
due to a zero conditional probability does not modify the
theorem results, since only the pollutant contributors participate in the processes. It is possible to include any number of
non pollutant contributors and the result will not be affected
for the pollutant ones.

PE j A3 0:9.

Applying Bayes theorem to the first participant

PA1 j E 0:2
0:5=0:2
0:5 0:4
0:3 0:9
0:2
0:10=0:40 0:25.
In the same way,
PA2 j E 0:12=0:40 0:30,
PA3 j E 0:18=0:40 0:45.
Therefore, this probability model provides the following cost
percentages for the use of the treatment plant:
Participant 1: 25%,
Participant 2: 30%,
Participant 3: 45%.
The application of the theorem demonstrates in this case,
that a small participating company that does not have a good
system of wastewater treatment, prior to the discharge,

3.1.

Real world case study

The Izta-Popo National Park located between 181 590 and 191
16.50 North latitude and between 981 350 and 181 16.50 West
longitude, has important hydraulic resources, mainly from
the thaw of the Popocatepetl volcano (altitude of 5482 m over
sea level) and the Iztaccihuatl mountain (altitude of 5280 m
over sea level). The park comprises four important river
basins where several parameters were determined by GomezMarquez et al. (2003), according to the standards of APHA,
AWWA and WPCF and SARH: flow, oxide reduction potential
(ORP), pH, conductivity, salinity, sediment solids (SS), total
sediment solids (TSS), total volatile solids (TVS), total
suspended solids (TSs), total suspended volatile solids (TSVS),
dissolved solids (DS), alkalinity, total hardness, sulfates,
sulfurous, organic matter (OM), particulate organic carbon
(POC), heavy metals (Fe, Cu, Cd, Pb, Zn), biochemical oxygen

ARTICLE IN PRESS
178

WA T E R R E S E A R C H

40 (2006) 175 181

Table 1 Different water treatment levels before discharging (strategy 1)

Participant

Discharge (%)

Pollution (%)

Contribution (%)

1
2
3

50
30
20

20
40
90

25
30
45

Table 2 Adding freshwater, doubling the discharge to diminish the pollutant concentration (strategy 2)
Participant

Discharge (%)

Pollution (%)

Contribution (%)

1
2
3

41.67
25.00
33.33

20
40
45

25
30
45

Table 3 One participant diminishes its water consumption by reusing the polluted water (strategy 3)
Participant

Discharge (%)

Pollution (%)

Contribution (%)

1
2
3

55.56
33.33
11.11

20
40
100

31.25
37.50
31.25

Table 4 One participant does not discharge polluted water (strategy 4)

Participant

Discharge (%)

Pollution (%)

Contribution (%)

1
2
3

50
30
20

0
40
90

0
40
60

Table 5 A new non-polluting participant (strategy 5)

Participant

Discharge (%)

Pollution (%)

Contribution (%)

1
2
3
4

41.67
25.00
16.67
16.67

20
40
90
0

25
30
45
0

demand (BOD), chemical oxygen demand (COD), total coliforms, fecal coliforms, nitrites, nitrates and orthophosphates
(Table 6).
Each river basin contributes to the total flow volume of
the Izta-Popo Park: Atoyac (5.099%), Chalco-Texcoco (1.071%),
Nexapa (87.832%), and Yautepec-Cuautla (5.998%). The
Bayes theorem provides estimates of the possible percentage
water remediation costs considering the individual contribution of each river basin for each parameter (Table 7),
compensating the flow and the measured contaminant in
river basin.
For each parameter, the individual participation of each
basin was calculated as a percentage regarding the maximum
observed value in the system.

4.

Discussion

The proposed method calculates the real contribution of the

participants in the remediation costs of a wastewater, in more
even way, depending of the contribution to the water
pollution, before it is processed in the treatment plant. The
method predicts of course, the trivial cases when all
participants contaminate equally, therefore paying in direct
proportion to the each ones discharge, as well as when they
discharge the same water volumes, then paying in direct
proportion to the pollution caused by each one.
The sensitivity of the method to the mutually
dependent effects among specific water management actions

ARTICLE IN PRESS
WAT E R R E S E A R C H

179

40 (20 06) 17 5 181

Table 6 Maximum parameter values of each river basin of National Park Izta-Popo from Gomez-Marquez et al. (2003)
Parameter

Atoyac

Chalco-Texcoco

Nexapa

Yautepec-Cuautla

Flow (m3/s)
ORP (mV)
pH
Conductivity (mS/cm)
Salinity (o/oo)
SS (ml)
TSS (mg/l)
TVS (mg/l)
TSs (mg/l)
TSVS (mg/l)
DS (mg/l)
Alkalinity (mg/l)
Total hardness (mg/l)
Sulfates (mg/l)
Sulfurous (mg/l)
O.M. (%)
POC (mg/l)
Fe (mg/l)
Pb (mg/l)
Cd (mg/l)
Zn (mg/l)
Cu (mg/l)
BOD (mg/l)
COD (mg/l)
Total coliforms (NMP/100 ml)
Fecal coliforms (NMP/100 ml)
Nitrites (mg/l)
Nitrates (mg/l)
Orthophosphates (mg/l)

0.238
279
8.5
2560
5.0
6.5
2.1
0.3
0.3
0.3
2.1
448.5
278.9
23.2
15.3
3.0
3.5
0.53
0.12
0.09
0.17
0.005
310
665
2712
1500
0.068
0.164
2.39

0.05
40.0
8.6
5780
0.0
3.2
2041
0.02
0.01
0.01
0.23
1246
39.8
6.7
10.8
0.9
0.5
6.46
0.96
0.69
0.97
0.47
191
571
0.0
8256
0.032
0.210
1.33

4.1
316.0
8.2
1095
5.0
30.0
11.9
1.16
3.09
1.49
10.6
360
717.1
123.5
21.1
6.3
7.0
0.34
0.53
0.16
0.9
0.25
245
736
5500
1000
0.079
0.370
1.9

0.28
228
8.6
306
4.0
0.6
0.49
0.08
0.05
0.02
0.48
103.5
199.1
2.6
19.4
2.7
1.2
0.19
0.0
0.0
0.21
0.09
35
66.4
70
0.0
0.0320
0.119
1.17

emphasizes its relevance, because depending on the strategy

adopted by each participant, it not only impacts itself,
but also influences directly the fees of the remaining
participants.
In case of new participants, the results of the method may
encourage them to design and build non-polluting industrial
plants, because of the economic advantage obtained by a
company free of these expenses over its competitors. The
other contributors attempt to diminish their costs avoiding
the installation of equipment, but they actually increase their
expenses in wastewater management fees. This erroneous
strategy will lead them to considerable economical losses due
to natural resources protection regulations and the social
conflicts with public health policies and conservationist
organizations actions.
Water treatment regulation in many countries is based only
on the pollutants concentration figures, meaning that it is
possible to fulfill the standards by applying Strategy 2. In
these cases the method is especially useful.
The focus of the model is to assess economic consequences
from technical factors in a simple and direct way in order
to provide better information to the decision-support
system for an integrated waste water management. The
Bayes theorem permits to decision makers an effective
control of the interaction between particular data (emissions
parameters) and global information (water flows or

waste water discharges to general system), based on technical

and scientific principles applied in several conditions and
cases.
This method contributes to establish the even operation
and maintenance costs, especially for small communities
(Garca et al., 2001) in which authorities and particulars can
share the expenses for the wastewater treatment.
Thanks to the advances in science and technology, future
works in water and wastewater treatment systems will
involve the upgrading of existing facilities for better performance, higher capacity and new ways of regulation (Rosen
and Morling, 1998). For the efficient implementation of any
project, improved strategies should be used, based on careful
studies of the local conditions, wastewater characteristics,
nature of pollution problems, and future demands to be
satisfied. The method presented in this paper may contribute
to these goals.
In the real case study presented, the mean remediation
contribution cost (%) for each river basin (Table 7, last row) is
different to its corresponding water flow percentage. For
instance, the Chalco-Texcoco basin has a remediation contribution of 4.2% with 1.1% of the total water flow. For the
Nexapa basinthe largesthas a remediation cost of 90.5%,
with 87.8% of the total water flow. Meanwhile for the Atoyac
basin, has a remediation cost of 3.5%, with 5.1% of the
total water flow, and finally the Yautepec-Cuautla basin, a

ARTICLE IN PRESS
180

WA T E R R E S E A R C H

40 (2006) 175 181

Table 7 Maximum parameter contribution (%) of each river basin to the Izta-Popo basin, resulting from Bayes theorem
applied individually
Parameter

Atoyac

Chalco-Texcoco

Nexapa

Yautepec-Cuautla

ORP (mV)
pH
Conductivity (mS/cm)
Salinity (o/oo)
SS (ml)
TSS (mg/l)
TVS (mg/l)
TSS (mg/l)
TSVS (mg/l)
DS (mg/l)
Alkalinity (mg/l)
Total hardness (mg/l)
Sulfates (mg/l)
Sulfurous (mg/l)
O.M. (%)
POC (mg/l)
Fe (mg/l)
Pb (mg/l)
Cd (mg/l)
Zn (mg/l)
Cu (mg/l)
BOD (mg/l)
COD (mg/l)
Total coliforms (NMP/100 ml)
Fecal coliforms (NMP/100 ml)
Nitrites (mg/l)
Nitrates (mg/l)
Orthophosphates (mg/l)

4.651
5.257
11.132
5.217
1.239
0.330
1.472
0.560
1.154
1.133
6.376
2.166
1.076
3.788
2.611
2.787
6.652
1.270
3.009
1.054
0.111
6.722
4.911
2.780
7.331
4.616
2.440
6.499

0.140
1.117
5.280
0.000
0.128
67.370
0.021
0.004
0.008
0.026
3.722
0.065
0.065
0.562
0.165
0.084
17.033
2.134
4.846
1.264
2.186
0.870
0.886
0.000
8.477
0.456
0.656
0.760

90.738
87.368
82.023
89.873
98.498
32.210
98.046
99.327
98.747
98.536
88.171
95.950
98.716
89.999
94.460
96.006
73.510
96.597
92.145
96.150
95.359
91.515
93.627
97.135
84.192
92.373
94.821
88.999

4.471
6.258
1.565
4.910
0.135
0.091
0.462
0.110
0.091
0.305
1.731
1.819
0.142
5.651
2.765
1.124
2.805
0.000
0.000
1.532
2.344
0.893
0.577
0.084
0.000
2.555
2.083
3.743

3.512

4.226

90.539

1.723

Average

remediation cost of 1.7% is obtained with a 6% of the total

water flow.
For comparative purposes, the mean remediation cost can
be normalizing as a function of the flow of each basin. Then,
the maximum remediation cost per flow corresponds to the
Chalco-Texcoco basin, with 3.9% per flow unit, calculated
from the ratio of the mean remediation cost of 4.2% and the
flow 1.1 m3/s. Meanwhile for the other three basins values of
0.69%, 1.03%, and 0.28% per flow unit, were calculated
(Atoyac, Nexapa and Yautepec-Cuautla, respectively).
Analyzing the remediation cost for each parameter individually (rows in Tables 7), the great flow in Nexapa basin
dominates most calculations, originating larger values of
costs remediation for almost all parameters. Only the
particular cases of parameters TSS and Fe, the remediation
costs (%) estimated from the Bayes theorem for the Nexapa
basin, was significantly lower than the average (Table 7). This
is due to the larger values corresponding to Chalco-Texcoco
for both parameters, where its remediation costs are 67.4%
and 17%, respectively (Table 6).
Before any specific application, the method outlined in this
paper needs to be adapted to the conditions involved in the
application as in the real world case study presented in this
paper. Any reliable calculation of the percentage of pollution
depends not only on the concentration of the pollutant, it is

also necessary to consider proper contamination indexes for

each kind of pollutant, like in the previously real case
analyzed.

5.

Concluding remarks

The approach herein presented can be helpful to promote and

encourage particulars to participate in the construction,
regulation and operation procedures of a wastewater treatment plant. This can bring many social advantages, mainly
when wastewater is recycled, for instance, for agriculture and
gardens irrigation, helping to the sustainability of the water
policy management.
The presented model allows simulating several scenarios
for the water users, demonstrating the benefit of water
treatment previous to the discharge, including the case of
pollution-free waters.

Acknowledgements
The authors appreciate the financial support from the
National Council of Science and Technology of Mexico
(CONACyT), Grant no. G35167-U.

ARTICLE IN PRESS
WAT E R R E S E A R C H

R E F E R E N C E S

Berger, J.O., 1985. Statistical Decision Theory and Bayesian

Analysis, second ed. Springer, Berlin NY 617pp.
Borsuk, M.E., Stow, C.A., 2000. Bayesian parameter estimation in a
mixed-order model of bod decay. Water Res. 34 (6), 18301836.
Garca, J., Mujeriego, R., Obis, J.M., Bou, J., 2001. Wastewater
treatment for small communities in Catalonia (Mediterranean
region). Water Policy 3 (4), 341350.
Gomez-Marquez, J.L., Pena-Mendoza, B., Arteaga-Meja, M., Guerra-Hernandez, E., Mendoza-Vallejo, E., Arcos-Ramos, R., 2003.
Water quality in the Izta-Popo National Park basin and areas of
influence. Third Latinamerican Congress on Hydrographic
Basins, Arequipa, Peru. FAO, REDLACH, INRENA (www.
congresocuencas.org.pe) (In Spanish).
Haag, J.E., Vande Wouwer, A., Queinnec, I., 2003. Macroscopic
modelling and identification of an anaerobic waste treatment
process. Chem. Eng. Sci. 58, 43074316.
Huang, G.H., Baetz, B.W., Patry, G.G., Terluk, V., 1997. Capacity
planning for an integrated waste management system under

40 (20 06) 17 5 181

181

uncertainty: a North American case study. Waste Manage. Res.

15 (5), 523546.
Konttinen, E., 1998. From industrial consensus to environmental
regulation: the coming of the Finnish industrial waste-water
policy. Water Policy 1 (3), 305319.
Merret, S., 2000. Industrial effluent policy: economic instruments
and environmental regulation. Water Policy 2 (3), 201211.
Powell, J.C., 1996. The evaluation of waste management options.
Waste Manage. Res. 14 (6), 515526.
Qian, S.S., Stow, C.A., Borsuk, M.E., 2003. On Monte Carlo methods
for Bayesian inference. Ecol. Modeling 159, 269277.
Rosen, B., Morling, S., 1998. A systematic approach to optimal
upgrading of water and waste water treatment plants. Water
Sci. Technol. 37 (9), 916.
Servais, P., Garnier, J., Demarteau, N., Brion, N., Billen, G.,
1999. Supply of organic matter and bacteria to aquatic
systems through waste water effluents. Water Res. 33 (16),
35213531.
Smith-Korfmacher, K., 1997. Solid waste collection systems in
developing urban areas of South Africa: an overview and case
study. Waste Manage. Res. 15 (5), 477494.