Sous-produits de la désinfection dans l'eau potable des petits …€¦ · Figure 3.2: Variations...
Transcript of Sous-produits de la désinfection dans l'eau potable des petits …€¦ · Figure 3.2: Variations...
Sous-produits de la désinfection dans l’eau potable des petits réseaux municipaux: variabilité spatio-temporelle,
modélisation et stratégies de suivi
Thèse
Stéphanie Guilherme
Doctorat en aménagement du territoire et développement régional
Philosophiae doctor (Ph.D.)
Québec, Canada
© Stephanie Guilherme, 2014
iii
Résumé
Les trihalométhanes (THM) et les acides haloacétiques (AHA) constituent les seules familles réglementées de
sous-produits de la désinfection (SPD). Les SPD sont des composés issus de la réaction de la matière organique
naturelle présente dans l’eau et du désinfectant lors du traitement de l’eau potable. La plupart de ces composés
ne sont pas réglementés, même si plusieurs études ont montré que certains SPD peuvent présenter un risque
toxicologique plus important que les THM et les AHA. De nos jours, très peu d’informations sont disponibles sur
l’occurrence des SPD non-réglementés dans l’eau potable, en particulier dans les petits réseaux municipaux.
Paradoxalement, les petits réseaux approvisionnés en eau de surface ont souvent des difficultés à mettre en
place des traitements adéquats pour enlever les précurseurs de SPD dans l’eau soumise à la désinfection. L’eau
potable des petits réseaux est ainsi plus vulnérable aux SPD et leur suivi y est d’autant plus important. Cette
thèse s’est donc consacrée à améliorer les connaissances sur l’occurrence des SPD (et en particulier, les SPD
non-réglementés) dans les petits réseaux de distribution d’eau potable, en étudiant notamment leur évolution
spatiale et temporelle. Pour ce faire, deux campagnes d’échantillonnage ont été réalisées sur deux ans (entre
2010 et 2012) dans 25 petits réseaux du Québec et de Terre-Neuve-et-Labrador. Les THM, les AHA et trois
autres familles de SPD non-réglementés, à savoir les haloacétonitriles, les halonitrométhanes et les
haloacétones, ont été étudiés. Les résultats obtenus ont permis de mieux comprendre leur patron de variabilité
spatio-temporelle, de modéliser leur présence et de développer un outil d’aide à la décision pour la mise en
place d’une stratégie de suivi des SPD réglementés et non-réglementés.
v
Abstract
Trihalomethanes (THMs) and haloacetic acids (HAAs) constitute the only regulated disinfection by-products
(DBPs) in various countries. DBPs are compounds formed during drinking water treatment, from the reaction
between natural organic matter and the disinfectant. Most DBPs are not regulated, even if they may have more
pronounced toxicological effects than regulated ones. There is currently very little information about the
occurrence of non-regulated DBPs, particularly in small water systems (SWS). Paradoxically, in many cases,
SWS supplied by surface waters lack adequate treatment processes to remove DBP precursors in water
subjected to the disinfection process. Their tap water may be more vulnerable to the presence of DBPs. This
thesis is dedicated to improving the knowledge of the occurrence of DBPs (especially non-regulated DBPs) in
SWS by studying their spatial and temporal variability. To do that, two sampling programs were carried out in 25
SWS during two years (between 2010 and 2012) in Canada. Small systems in the provinces of Newfoundland
& Labrador and Quebec were considered. The following DBPs were measured during the study: THMs, HAAs,
haloacetonitriles, haloketones and halonitromethanes. The obtained results contribute to a better understanding
of the DBP spatio-temporal variation patterns, to establishing models to evaluate their levels and to developing
decision-making schemes for simultaneously monitoring various families of DBPs, including non-regulated
DBPs.
vii
Table des matières
Résumé ............................................................................................................................................................... iii
Abstract ............................................................................................................................................................... v
Table des matières ............................................................................................................................................. vii
Liste des tableaux .............................................................................................................................................. xi
Liste des figures ................................................................................................................................................ xiii
Liste des abréviations........................................................................................................................................ xv
Remerciements ................................................................................................................................................ xvii
Avant-Propos .................................................................................................................................................... xix
Introduction ......................................................................................................................................................... 1
Bibliographie de l’introduction ........................................................................................................................... 11
Chapitre 1: Occurrence of regulated and non-regulated disinfection by-products in small drinking water
systems ............................................................................................................................................................. 19
1.1. Introduction ...................................................................................................................................... 20
1.2. Methodology .................................................................................................................................... 21
1.2.1. Case studies ................................................................................................................................ 21
1.2.2. Sampling and analysis................................................................................................................. 21
1.2.3. Data Analysis .............................................................................................................................. 23
1.3. Results and discussion .................................................................................................................... 23
1.3.1. Portrait of DBP occurrence in small system ................................................................................ 23
1.3.2. Temporal variations of DBP levels .............................................................................................. 26
1.3.3. Spatial variations of DBP occurrence within the distribution systems ......................................... 30
1.3.4. Spatio-temporal variations of DBP occurrence ............................................................................ 31
1.4. Conclusions ..................................................................................................................................... 33
1.5. References ...................................................................................................................................... 34
Chapitre 2: Models for estimating non-regulated disinfection by-product occurrence in small drinking water
systems ............................................................................................................................................................. 39
2.1. Introduction ...................................................................................................................................... 41
2.2. Methodology .................................................................................................................................... 42
2.2.1. Case studies ................................................................................................................................ 42
2.2.2. Sampling and analysis................................................................................................................. 42
2.2.3. Modelling ..................................................................................................................................... 43
viii
2.3. Results ............................................................................................................................................. 44
2.3.1. Influence of treatment conditions on DBP speciation .................................................................. 44
2.3.2. Influence of water quality characteristics on non-regulated DBP levels in the DS ....................... 45
2.3.3. Correlations between regulated and non-regulated DBPs ........................................................... 46
2.3.4. Modelling levels of non-regulated DBPs in small water systems ................................................. 50
2.3.5. Validation of HAN and HK models ............................................................................................... 53
2.4. Discussion and conclusions ............................................................................................................. 54
2.5. References ....................................................................................................................................... 55
Chapitre 3: Short-term spatial and temporal variability of disinfection by-product occurrence in small drinking
water systems.................................................................................................................................................... 61
3.1. Introduction ...................................................................................................................................... 63
3.2. Methodology .................................................................................................................................... 64
3.2.1. Case studies ................................................................................................................................ 64
3.2.2. Sampling and analysis ................................................................................................................. 65
3.3. Results and discussion .................................................................................................................... 66
3.3.1. Short-term temporal variability of DBP occurrence ...................................................................... 66
3.3.2. Spatial variability of DBP occurrence ........................................................................................... 73
3.3.3. Impact of the location on the short-term variability of DBPs ........................................................ 75
3.4. Conclusions ..................................................................................................................................... 77
3.5. References ....................................................................................................................................... 77
Chapitre 4: Decision-making scheme for disinfection by-product monitoring intended for small drinking water
systems ............................................................................................................................................................. 83
4.1. Introduction ...................................................................................................................................... 84
4.2. Methodology .................................................................................................................................... 86
4.2.1. Case studies ................................................................................................................................ 86
4.2.2. Sampling and analysis ................................................................................................................. 87
4.3. Results and discussion .................................................................................................................... 88
4.3.1. Sampling period identification ...................................................................................................... 88
4.3.2. Sampling location identification ................................................................................................... 99
4.3.3. Decision-making scheme for DBP monitoring ........................................................................... 103
4.4. Conclusions ................................................................................................................................... 105
4.5. References ..................................................................................................................................... 106
Conclusions et recommandations ................................................................................................................... 109
Annexe 1 : Répartition géographique des 25 petits réseaux étudiés dans les provinces a) de Québec et b) de
Terre-Neuve-et-Labrador ................................................................................................................................ 115
Annexe 2 : Caractéristiques des 25 petits réseaux étudiés des provinces de Québec et Terre-Neuve-et-
Labrador .......................................................................................................................................................... 116
Annexe 3 : Informations sur les méthodes analytiques utilisées pour l’analyse des SPD ............................... 117
Annexe 4: Comparison of water characteristics between SWS in QC and NL ................................................ 121
Annexe 5: Spatial evolution of regulated DBP concentrations in SWS in a) NL and b) QC ............................ 122
Annexe 6: Spatial evolution of non-regulated DBP concentrations in SWS in a) NL and b) QC ..................... 123
Annexe 7: Distribution of daily levels (average of the six SWS under study in DS3) of a) free chlorine, b) THMs,
c) HAAs, d) HANs, e) CPK and f) HKs ............................................................................................................ 124
Annexe 8: Evolution of average DBP and free residual chlorine levels along the DS in a) NL1, b) NL2, c) NL3,
d) QC1, e) QC2 and f) QC3 (average of 5 measurements for NL1 and 12 for the others) .............................. 125
Annexe 9: Summary of annual average levels of a) THMs, b) HAAs, c) HANs, c) CPK, and e) HKs, based on
the various scenarios ...................................................................................................................................... 130
Annexe 10: Exemple de rapport de synthèse de données envoyé aux petits réseaux ................................... 135
xi
Liste des tableaux
Tableau 0.1: Normes et recommandations internationales concernant les THM et les AHA
Table 1.1: Parameters measured during the sampling campaign
Table 1.2: Regional variation of regulated and non-regulated DBP occurrence in SWS of NL and QC
Table 1.3: Levels of regulated and non-regulated DBPs observed in the literature
Table 1.4: Spatial variation of residual disinfectant, THMs and HAAs in NL and QC
Table 1.5: Spatial variation of non-regulated DBPs concentration in NL and QC
Table 2.1: Parameters measured during the sampling campaigns
Table 2.2: Average levels of non-regulated DBP in location DS2 during summer (July, August and September)
and in winter (January, February and March) in four SWS using different types of treatment (number of
observations per season = 3)
Table 2.3: Spearman correlation matrix between water quality characteristics of treated water at the treatment
plant (after filtration and before disinfection) and non-regulated DBP levels in the DS in all SWS (number of
observations for each parameter = 300)
Table 2.4: Spearman correlation matrix between DBP average levels in location DS2 (number of observations
for each parameter = 300)
Table 2.5: Spearman correlation matrix between DBP variation ratios a) DS1 and DS2 winter/summer; b) DS2
and DS3 winter/summer
Table 2.6: Classification of DOC levels at the WTP (number of observations = 300)
Table 2.7: Multivariate regression models for non-regulated DBP levels
Table 3.1: General characteristics of SWS under study
Table 3.2: Parameters measured during the sampling campaign
Table 3.3: Weekly average levels of DBPs and free chlorine in DS3 (with coefficients of variation-CV) in the SWS
under study
xii
Table 3.4: Specific locations for the maximum levels of DBPs and free residual chlorine in the SWS under study
Table 4.1: Description of the different scenarios to calculate the annual average levels of DBPs
Table 4.2: Correlation (Spearman) between different results of scenarios for annual average levels of a) THMs,
b) HAAs, c) HANs, d) CPK and e) HKs
Table 4.3: Months presenting the maximum DBP level observed in all SWS (QC and NL included) for each
trimester between fall 2010 and summer 2011 (based on 12 monthly DBP measurements)
Table 4.4: Summer average levels of a) THMs, b) HAAs, c) HANs, d) CPK and e) HKs based on different
scenarios (levels in µg/L)
Table 4.5: Values of free residual chlorine decrease in each location with maximum DBP level obtained in four
SWS (second campaign in summer 2012)
xiii
Liste des figures
Figure 1.1: Temporal evolution of regulated DBP concentrations in SWS in a) NL and b) QC
Figure 1.2: Temporal evolution of non-regulated DBP concentrations in SWS in a) NL and b) QC
Figure 1.3: Variation of average DBP occurrence along the DS in NL and QC in: a) to f) summer (July-
September); g) to l): winter (January-March)
Figure 2.1: Identification of the most correlated regulated DBPs (Spearman correlation factors indicated) with
non-regulated DBPs according to DOC levels before disinfection, season and location in the DS
Figure 2.2: Validation of HAN models: correlation between observed and estimated values in a) all systems, b)
only small systems
Figure 2.3: Validation of HK models: correlation between observed and estimated values in a) all systems, b)
only small systems
Figure 3.1: Variations from day to day within the week of levels of a) Free chlorine, b) THMs, c) HAAs, d) HANs,
e) CPK and f) HKs in NL2 in DS3 (number of observations per day: 4)
Figure 3.2: Variations from day to day within the week of levels of a) Free chlorine, b) THMs, c) HAAs, d) HANs,
e) CPK and f) HKs in QC3 in DS3 (number of observations per day: 4)
Figure 3.3: Daily evolution of raw water characteristics and DBP occurrence in DS3 a) and b) in NL2 and c) and
d) in QC3
Figure 3.4: Location of the maximum levels of DBPs and free residual chlorine along the DS for all SWS under
study
Figure 3.5: Temporal variability of the DBP levels in two locations along the DS of QC3: a) and b) in DS1; c) and
d) in DS6 (SWS presenting an average level of free chlorine higher than 0.3 mg/L in DS6
Figure 3.6: Temporal variability of the DBP levels in two locations along the DS of NL2: a) and b) in DS1; c) and
d) in DS6 (SWS presenting an average level of free chlorine lower than 0.3 mg/L in DS6)
Figure 4.1: Sampling locations during the campaigns
xiv
Figure 4.2: Comparison of results of the various scenarios for the annual average level of a) THMs, b) HAAs, c)
HANs, d) CPK, and e) HKs
Figure 4.3: Variations of average a) regulated and b) non-regulated DBP levels along the DS in winter (January
- March) in all SWS in NL and QC (number of observations for each DBP in each location = 75)
Figure 4.4: Variations of average DBP levels along the DS in summer, a) and c) in two SWS presenting a free
chlorine level higher than 0.3 mg/L at the end of the DS, and b) and d) in two SWS presenting a free chlorine
Figure 4.5: Decision-making scheme for regulated and non-regulated DBP monitoring
xv
Liste des abréviations
ADBA: acide dibromoacétique
ADCA: acide dichloroacétique
AHA: acide haloacétique
AMBA: acide monobromoacétique
AMCA: acide monochloroacétique
ANOVA: analyse de variance
ATCA: acide trichloroacétique
BCAN: bromo-chloroacétonitrile;
BDCM: bromodichlorométhane
CEAEQ: centre d'expertise en analyse environnementale du Québec
CPK: chloropicrin (trichloronitrométhane)
COD: carbone organique dissous
CV: coefficient de variation
DBAN: dibromoacétonitrile
DBCM: dibromochlorométhane
DBP: disinfection by-products
DCAN: dichloroacétonitrile
DCBM: dichlorobromométhane
DCP: 1,1-dichloropropanone
DS: distribution system
xvi
HAN: haloacétonitrile
HC: halocétone
HNM: halonitrométhanes
MDDELCC: ministère du développement durable, de l’environnement et de la lutte contre les changements
climatiques
MON: matière organique naturelle
NL: Newfoundland and Labrador
QC: Québec
RW: raw water
SUVA: absorbance UV spécifique
SPD: sous-produits de la désinfection
SWS: small water system
TBM: tribromométhane
TCAN: trichloroacétonitrile
TCM: trichlorométhane
TCP: 1,1,1-trichloropropanone
THM: trihalométhane
TN: Terre-Neuve et Labrador
USEPA: United States environmental protection agency
UV-254: absorbance UV à 254nm
WTP: water treatment plant
xvii
Remerciements
Je voudrais remercier vivement mon directeur de thèse, Pr. Manuel Rodriguez, pour son soutien, sa disponibilité
et sa confiance en moi tout au long de ma thèse. Ses connaissances illimitées et ses précieux conseils m’ont
permis de réaliser ce beau projet et de surmonter les difficultés de ce doctorat. Je tiens aussi à remercier les
membres de mon jury, Mr Patrick Drogui, Mr. Caetano Dorea, Mr. Jean-Baptiste Sérodes pour leurs
commentaires et remarques qui m’ont permis de bonifier ce manuscrit.
Aussi, je tiens à remercier les partenaires financiers sans qui la réalisation de ce projet n'aurait pas été possible :
RES’EAU-WaterNET, le Conseil de recherches en sciences naturelles et en génie du Canada (CRSNG) et la
Chaire de recherche en eau potable de l'Université Laval.
Je remercie aussi tous les opérateurs des 25 petits réseaux étudiés, pour leur aide, leur disponibilité et leur
accueil durant toutes les campagnes d’échantillonnage. En particulier, je tiens à remercier, Mr. Mario Demers,
Mr. Harvey Henstridge, Mr. Jean Claude Lapointe, Mr James Jr. Peckford, Mr. Philip Smith et Mr. Alain
Tousignant, pour leur grande gentillesse et leur bonne humeur à toute épreuve. Je tiens à remercier les autorités
locales et provinciales qui ont permis la mise en place de ces deux campagnes, en particulier Mme. Anouka
Bolduc et Mr. Haseen Khan.
Je voudrais remercier tous les membres du laboratoire, Sabrina Simard, Michel Bisping et tous les étudiants qui
ont participé aussi bien à l’échantillonnage qu’aux analyses en laboratoire lors des deux campagnes. Sans leur
aide précieuse, ce projet n’aurait pas pu se réaliser. Je voudrais également remercier tout particulièrement
Sabrina Simard pour son soutien et sa grande disponibilité. Aussi, je souhaite aussi adresser un merci particulier
à Catherine Mercier-Shanks, dont ses travaux d’une grande qualité sur les SPDE m’ont permis d’analyser ces
composés.
Je souhaite également remercier tous les étudiants du CREPUL et du CRAD pour leur soutien et leur gentillesse
tout au long de ce doctorat et en particulier mes deux colocataires de bureau qui m’ont supporté pendant quatre
ans : Christelle Legay et Anna Scheili. Je voudrais aussi remercier tout le personnel et les professeurs du CRAD
et de l’ÉSAD et plus particulièrement Francine Baril, Lyne Béland, Marie-Pier Bresse, Willem Fortin et Francis
Rioux pour leur aide et leur bonne humeur communicative.
Finalement, j’aimerais profondément remercier ma famille et mes amis pour leur soutien, leur gentillesse et leurs
encouragements. Tout d’abord, ma mère, Maria, mon père, Luciano, ma sœur, Christelle et ma belle Léa qui
m’ont toujours encouragé dans tous mes projets et en particulier celui-ci de l’autre côté de l’océan. Leur
incroyable soutien ont permis le succès de ce doctorat. Aussi, je tiens à remercier mes amis qui m’ont toujours
xviii
soutenu et encouragé dans ce doctorat Orphé, Jérémy, Léo, Maman Caro, Juliane, Marion, Fabien et Rose-
Marthe. Je voudrais remercier tout particulièrement Orphé pour son génie, son optimisme, sa bonne humeur et
sa motivation, merci d’avoir toujours été là pour moi durant toutes ces années. Je tiens à remercier aussi ceux
qui m’ont encouragé à distance, Delphine, Laure, Yacine, Jeff, Nils, Bertrand, Gary, Joel, Pierrick, Etienne,
Charlotte et Aurélie. Aussi, un grand merci à Sylvie et Catherine pour nos pauses Yoga qui m’ont permis de
rester zen en période de stress.
Enfin, je ne sais comment remercier mon beau Seb pour m’avoir soutenu et encouragé durant mon doctorat.
Son énergie, sa motivation, sa confiance en moi ainsi que ses farces, ses surprises et ses poutines m’ont permis
de surmonter les obstacles de ce doctorat. Je tiens aussi à remercier sa famille Marie-Lise, Fernand, Fred, Marc,
Claudine et Mia pour leur gentillesse.
xix
Avant-Propos
Cette thèse est composée de quatre chapitres. Chaque chapitre est constitué d’un article publié ou soumis à
des revues scientifiques avec comité de lecture. Ces articles ont été rédigés en anglais, en accord avec les
directives des revues scientifiques. Le premier article a été publié récemment dans la revue « Chemosphere ».
Le troisième a été soumis à « Science of the Total Environnement » et les deux autres sont prêts à être soumis
à d’autres revues scientifiques. Les articles sont présentés dans l’ordre chronologique du développement des
travaux de recherche de la thèse. Bien que chaque article puisse être lu séparément, l’ensemble des articles
constitue un cheminement organisé pour une caractérisation par étapes de l’évolution spatio-temporelle des
sous-produits de la désinfection (SPD) dans l’eau potable de petits réseaux municipaux. Ainsi, des rappels de
notions ou de méthodes peuvent être observés dans les articles constituant la thèse. Aussi, entre les chapitres
2, 3 et 4, un texte a été ajouté afin d’expliquer le lien entre chaque article et la logique de développement suivie.
Enfin, le manuscrit contient un résumé de la thèse et quatre chapitres en français et en anglais. Il est aussi
composé d’une introduction générale, d’une conclusion générale ainsi que d’une bibliographie et des annexes.
Pour chaque article, les auteurs sont l’auteur de la thèse ainsi que le directeur de recherche (Manuel J.
Rodriguez). Les titres et références de chaque article constituant cette thèse sont présentés ci-dessous.
CHAPITRE 1
Occurrence of regulated and non-regulated disinfection by-products in small drinking water systems
Stéphanie Guilherme and Manuel J. Rodriguez, 2014, Chemosphere, vol. 117, pp.425–432
CHAPITRE 2
Models for estimating non-regulated disinfection by-product occurrence in small drinking water systems
Stéphanie Guilherme and Manuel J. Rodriguez
Prêt à être soumis
CHAPITRE 3
Short-term spatial and temporal variability of disinfection by-product occurrence in small drinking water
systems
Stéphanie Guilherme and Manuel J. Rodriguez
Soumis à « Science of the Total Environment »
CHAPITRE 4
Decision-making scheme for disinfection by-product monitoring intended for small drinking water systems
Stéphanie Guilherme and Manuel J. Rodriguez
Prêt à être soumis.
1
Introduction
La qualité de l’eau potable est un des plus importants déterminants de la santé, et sa gestion reste un élément
essentiel dans la prévention et le contrôle des maladies hydriques (Organisation Mondiale de la Santé, 2011).
En cela, la désinfection est une étape indispensable à la salubrité de l’eau potable. Elle permet d’inactiver les
micro-organismes pathogènes présents dans l’eau mais aussi de prévenir leur croissance dans le système de
distribution. Ayant l’avantage d’être peu coûteux et facile à utiliser, le chlore est devenu le désinfectant le plus
communément utilisé dans le monde, principalement sous forme d’hypochlorite de sodium (Cedergren et al.,
2002; Villanueva et al., 2007). Cependant, en 1974, J. J. Rook (Rook, 1974) découvre que le chlore utilisé pour
désinfecter l’eau, réagit avec la matière organique naturelle (MON) présente dans l’eau pour générer des sous-
produits chlorés potentiellement toxiques que l’on appelle usuellement sous-produits de la désinfection (SPD).
On sait maintenant que les désinfectants sont, à l’exception du rayonnement U.V., des oxydants (chlore, bioxyde
de chlore, ozone ou chloramines) qui peuvent générer la formation de SPD. Plusieurs facteurs influencent la
formation ainsi que la spéciation (la variété des composés présents) des SPD. Le type de traitement joue un
rôle important dans la formation des SPD, surtout en sa capacité à enlever la matière organique de l’eau brute
avant la chloration (Rodriguez & Sérodes, 2001; Bull et al., 2009). D’autres paramètres physico-chimiques de
l’eau influencent leur formation comme la dose de désinfectant, le temps de contact de l’eau avec le désinfectant
(Singer, 1994; Rodriguez & Sérodes, 2001; Liang & Singer, 2003; Rodriguez et al., 2004; Speight & Singer,
2005; Bull et al., 2009), le pH (Bove et al., 2002; Liang & Singer, 2003; Bull et al., 2009) et la température de
l’eau traitée (Singer, 1994; Liang & Singer, 2003; Rodriguez et al., 2004; Bull et al., 2009). De plus, les propriétés
du réseau de distribution, comme la présence de biofilms dans sur les parois des conduites (Singer, 1994), la
nature et les dimensions des canalisations, les conditions hydrauliques et la présence de réservoirs de stockage
et de stations de rechloration au sein du réseau (Rodriguez et al., 2004) vont, elles aussi, jouer un rôle dans la
formation de SPD. Aussi, le type de MON présente dans l’eau va influencer la formation des SPD. En effet, une
eau riche en substances humiques comme les acides fulviques et humiques (composés multi-aromatiques) va
potentiellement former plus de SPD (Crittenden et al., 2005). Ainsi, le SUVA (rapport de l’absorbance UV à
254nm et de la mesure de carbone organique dissous) représente un bon indicateur de ces précurseurs
spécifiques car il indique la proportion de composés aromatiques présents dans la MON de l’eau brute
(Crittenden et al., 2005). De même, à la suite d’un traitement équivalent, une eau de surface plus riche en MON
présentera une contamination en SPD plus importante qu’une eau souterraine (Singer, 1994; Mouly et al., 2010).
À ce jour, bien que plus de 700 SPD ont été répertoriés dans la littérature, l’occurrence et les effets sur la santé
de peu d’entre eux ont été étudiés (Richardson, 2011; Villanueva et al., 2014). Notre étude se concentre
principalement sur les SDP formés lors de la chloration, en particulier par la réaction de l’acide hypochloreux et
la MON, dont la majorité sont des sous-produits organochlorés. Certains de ces SPD, comme les
2
trihalométhanes (THM) et les acides haloacétiques (AHA) sont relativement bien documentés et sont
réglementés dans de nombreux pays. Les THM et AHA sont les principaux SPD présents dans l’eau potable,
en concentration massique (µg/L), et les plus étudiés depuis les 30 dernières années (Singer, 2002; Rodriguez
et al., 2004). Les THM sont un groupe de composés organiques volatils (Pérez Pavon et al., 2008) dont les plus
communément observés dans l’eau potable sont le trichlorométhane ou chloroforme (TCM), le bromo-
dichlorométhane (BDCM), le dibromo-chlorométhane (DBCM) et le tribromométhane (TBM). Concernant les
AHA, les plus communément observés dans l’eau potable sont l’acide monobromoacétique (AMBA), l’acide
monochloroacétique (AMCA), l’acide dibromoacétique (ADBA), l’acide dichloroacétique (ADCA) et l’acide
trichloroacétique (ATCA) (Rodriguez et al., 2004). La particularité des SDP est associée aussi aux modes
d’exposition multiples dus à leur présence dans l’eau potable. En effet, trois principales voies d’exposition
humaine aux SPD sont recensées globalement dans la littérature. La principale voie d’exposition étudiée est
l’ingestion orale, spécialement pour les SPD non ou peu volatils, comme les AHA par exemple (Richardson et
al., 2007; Nieuwenhuijsen et al., 2009; Villanueva et al., 2014). Les autres voies alternatives d’exposition pour
les SPD plus volatils, comme les THM par exemple, sont par inhalation et absorption cutané lors d’activités
telles que la douche, le bain et la natation (Richardson et al., 2007; Nieuwenhuijsen et al., 2009; Villanueva et
al., 2014). En plus de présenter une variété de voies d’exposition, les SPD présentent des types et des niveaux
de toxicité très variables selon leurs propriétés chimiques.
Différentes études ont montré que les THM étaient cancérigènes (Richardson et al., 2007; Villanueva et al.,
2014), notamment sur les rats. En effet, des études toxicologiques sur le TCM et le BDCM ont révélé que,
administrés dans l’eau potable, ces composés pouvaient causer des tumeurs sur le foie et le rein lors de tests
réalisés sur des animaux (National Cancer Institute, 1976; Jorgenson et al., 1985; Department of health & human
services USA, 2006). Par conséquent, les rapports du centre international de recherche sur le cancer (CIRC)
ont déclaré le TCM et BDCM comme possiblement cancérigènes chez l’homme (2B) (International Agency for
Research on Cancer, 1991; International Agency for Research on Cancer, 1999). Cependant, il reste encore
une incertitude sur la toxicité pour l’homme du TCM. En effet, l’USEPA a déclaré le TCM à la fois possiblement
cancérigènes chez l’homme (2B) et peu probablement cancérigène pour l’homme (USEPA, 2001). Aussi, les
THM ont été associés à une certaine toxicité pour le développement (Bull et al., 2009). Concernant les AHA,
des études épidémiologiques sur l’ADBA, l’ADCA, l’AMBA et l’AMCA ont montré qu’ils étaient mutagènes, avec
une plus forte toxicité des AHA bromés (Giller et al., 1997; Kargalioglu et al., 2002; Plewa et al., 2002;
Richardson et al., 2007). Plusieurs études ont aussi montré une certaine cancérogénicité sur les animaux de
l’ADBA, l’ADCA et l’ATCA principalement sur le foie (Bull et al., 1990; DeAngelo et al., 1996; DeAngelo et al.,
1999; Melnick et al., 2007; Richardson et al., 2007; Bull et al., 2009; Villanueva et al., 2014)
3
Due à leur potentielle toxicité et leur forte exposition, des règlementations sur les THM et AHA ont été mises en
place dans différents pays dans le monde (Tableau 0.1). En Amérique du nord, les premières règlementations
sur les SPD sont apparues aux États-Unis, par l’intermédiaire de l’Environmental Protection Agency (USEPA).
En 1979, l’USEPA avait établi une concentration maximale de THM4 (somme des concentrations des TCM,
BDCM, DBCM et TBM) de 100 μg/L pour les systèmes desservant une population supérieure à 10 000 habitants
(Environmental Protection Agency, 1998). En décembre 1998, USEPA a publié l’étape 1 du plan de
réglementation des désinfectants et des SPD fixant de nouveaux standards sur les THM et les AHA et une
obligation de suivi de ces composés. Ainsi, les THM4 ne doivent pas dépasser une concentration moyenne
annuelle maximale de 80 μg/L. La concentration d’AHA5 (somme des concentrations des MCAA, DCAA, TCAA,
MBAA et DBAA) ne doit, elle, pas dépasser une concentration moyenne annuelle maximale de 60 μg/L. Les
normes ont été mises en place dans les grands réseaux publics de distribution d’eau potable (plus de 10 000
consommateurs) alimentés en eau de surface, puis dans les petits réseaux publics de distribution d’eau potable
(moins de 10 000 consommateurs) alimentés par des eaux de surface ou souterraine (Environmental Protection
Agency, 1998). Les données collectées lors du suivi des SPD dans les réseaux de distribution ont permis
d’approfondir la réglementation sur les SPD et de mettre en place par la suite, l’étape 2 du plan de
réglementation des désinfectants et des SPD. La nouvelle norme établit des points de suivi spécifiques dans le
réseau pour les différents composés réglementés par l’étape 1 (Environmental Protection Agency, 2006b). Ainsi,
la concentration des SPD ne se base plus sur une moyenne de mesures sur différents points dans le réseau,
mais à des points où la concentration des composés réglementés est la plus élevée. Pour cela, les municipalités
doivent réaliser une auto-évaluation dans leur réseau afin d’identifier les sites les plus appropriés pour le suivi
des THM et des AHA. Ce programme, intitulé « évaluation du système de distribution initiale (Initial Distribution
System Evaluation (IDSE)) » est une étude ponctuelle durant laquelle les concentrations de THM et de AHA
sont mesurées à plusieurs localisations le long du système (de deux à quatre sites) d’une à six fois au cours
d’une année. Le nombre de points d'échantillonnage, le nombre d'échantillons prélevés à chaque emplacement
et la fréquence d'échantillonnage dépendent du type de la source d'eau ainsi que de la taille de la population
desservie par le réseau (Environmental Protection Agency, 2006a). Finalement, ce programme permet
d’identifier les localisations présentant les niveaux les plus élevés de THM et de AHA dans le système de
distribution, correspondant aux sites d'échantillonnage où le suivi réglementaire de ces composés devrait être
fait.
Au Canada, il existe une recommandation relative aux concentrations des THM et des AHA dans l’eau potable.
La concentration maximale acceptable (moyenne annuelle sur quatre trimestres) pour les THM4 est fixée à 100
μg/L et celle des AHA5 à 80 μg/L (Health Canada, 2012). Cependant ces valeurs ne sont que des
recommandations. En effet, la règlementation sur l’eau potable au Canada est sous la juridiction des provinces.
Le Québec fait partie des rares provinces canadiennes à s’être dotées d’une réglementation exigeante sur la
4
qualité de l’eau potable. La concentration annuelle maximale acceptable des THM4 est à 80 μg/L, concentration
moyenne maximale calculée sur quatre trimestres consécutifs (MDDELCC, 2012). De plus, une obligation de
suivi des THM au moyen de un à huit prélèvements par trimestre (selon la taille de la population desservie par
le réseau) a été établie. Les AHA, quant à eux, sont règlementés depuis peu au Québec. La norme sur les AHA5
est fixée à 60 µg/L, concentration moyenne maximale calculée sur quatre trimestres consécutifs (MDDELCC,
2012) mais il n’existe pas encore d’obligation de suivi. Par comparaison, la province de Terre-Neuve-et-Labrador
a mis sur place une recommandation sur la concentration maximale en THM dans l’eau potable en suivant les
recommandations de Santé Canada. Ainsi, la concentration annuelle maximale acceptable des THM4 est fixée
à 100 μg/L, concentration moyenne maximale calculée sur quatre trimestres consécutifs (Health Canada, 2012).
Cependant, aucune de ces provinces n’a mis en place d’auto-évaluation du système (comparable à l’IDSE) afin
d’identifier les sites les plus appropriés pour le suivi des SPD dans les réseaux.
Tableau 0.1 : Normes et recommandations internationales concernant les THM et les AHA
Lieux SPD Normes
Québec (MDDELCC, 2012)
THM 80 µg/L (Concentration moyenne maximale calculée sur 4 trimestres)
AHA 60 µg/L (Concentration moyenne maximale calculée sur 4 trimestres)
USA (Environmental Protection Agency, 2006b)
THM 80 µg/L (Concentration moyenne maximale calculée sur 4 trimestres)
AHA 60 µg/L (Concentration moyenne maximale calculée sur 4 trimestres)
Europe (Conseil de l’Union Européenne, 1998)
THM 100 µg/L (Norme réglementaire)
Canada (recommandation) (Health Canada, 2012)
THM 100 µg/L (Concentration moyenne maximale calculée sur 4 trimestres)
AHA 80 µg/L (Concentration moyenne maximale calculée sur 4 trimestres)
OMS (recommandation) (Organisation Mondiale de la Santé, 2011)
THM TCM: 300 µg/L; BDCM: 60 µg/L; DBCM: 100 µg/L; TBM: 100 µg/L
AHA DCAA: 50 µg/L; TCAA: 200 µg/L
HAN DBAN: 70 µg/L ; DCAN: 20 µg/L
Les SPD organiques réglementés (THM et AHA) ne représentent que 30 à 60% de tous les composés
organiques halogénés formés dans l’eau (Karanfil et al., 2008). La majorité des SPD restent encore non-
réglementés. De ces contaminants, trois familles ont été retenues pour cette étude: les haloacétonitriles (HAN),
5
les halonitrométhanes (HNM) et les halocétones (HC). Cette étude se concentre sur les quatre haloacétonitriles
suivants: le trichloroacétonitrile (TCAN), le dichloroacétonitrile (DCAN), le dibromoacétonitrile (DBAN) et le
bromo-chloroacétonitrile (BCAN) et sur un seul HNM, le trichloro-nitrométhane (la chloropicrine (CPK)), qui est
d’ailleurs le HNM le plus étudié dans la littérature (Merlet et al., 1985). De plus, cette étude se porte aussi sur
les deux HC suivant : le 1,1-dichloro-2-propanone (DCP) et le 1,1,1-trichloro-2-propanone (TCP). Ces SPD non-
réglementés sont encore peu étudiés. Leurs concentrations étant plus faibles que celles des THM et des AHA,
l’étude des SPD non-réglementés n’a pu se développer qu’avec l’amélioration des techniques analytiques
(Krasner et al., 1989). De plus, peu d’études épidémiologiques ont été consacrées aux SPD non-réglementés.
Or, certaines indiquent une toxicité des HAN et des HNM (SPD azotés), plus forte que celle des AHA (Muellner
et al., 2007; Hu et al., 2010). Certaines études ont montré une certaine génotoxicité des HAN étudiés ici
(Muellner et al., 2007). Aussi, une toxicité pour le développement ainsi qu’une certaine toxicité pour la thyroïde
(à forte dose sur les animaux) ont été reportées pour les HAN (International Agency for Research on Cancer,
1999; Muller-Pillet et al., 2000; Bull et al., 2009). Des études sur la CPK ont conclu à une génotoxicité du
composé sur les animaux (Giller et al., 1995; Schneider et al., 1999; Kundu et al., 2004; Richardson et al., 2007;
Bull et al., 2009). Et enfin, les HC ont été reportés, eux aussi, comme génotoxiques (Robinson et al., 1989; Bull
et al., 2009). Les diverses voies d’exposition, types et niveaux de toxicité font de ces SPD non-réglementés un
réel problème de santé publique. De plus, peu d’études se sont consacrées à étudier l’évolution temporelle et
spatiale de leur occurrence dans l’eau potable (Golfinopoulos et al., 2003; Krasner et al., 2006; Goslan et al.,
2009; Mercier-Shanks et al., 2013). Il existe donc un manque essentiel de données sur ces contaminants dans
l’eau potable. Afin d’assurer une eau potable de qualité, il est essentiel d’enrichir notre connaissance sur ces
SPD non-réglementés.
Cette thèse de doctorat s’intéresse tout particulièrement à l’étude de l’occurrence des SPD dans les petits
réseaux d’eau potable, qui desservent des petites municipalités ou communautés de moins de 5 000 habitants.
Environ 20% de la population canadienne vit dans des petites municipalités (Statistics Canada, 1851 to 2006),
or, ces petits réseaux ont des difficultés à maintenir la qualité de leur eau potable à des niveaux adéquats
(Coulibaly & Rodriguez, 2004; Hrudey, 2008). Les petits réseaux desservis en eau de surface sont
particulièrement vulnérables aux contaminations microbiennes (Davies & Mazumder, 2003; Edwards et al.,
2012). En particulier, les systèmes d’eau potable des Premières nations et les petits réseaux ruraux ont des
difficultés à assurer une eau de bonne qualité et à répondre aux normes en matière d’eau potable en
comparaison avec des plus grands réseaux municipaux (Edwards et al., 2012). En 2006, au Canada, environ
30% des systèmes d’eau potable de communautés de Premières nations étaient classés à risque élevé (un
système d'eau potable à risque élevé est un système présentant des lacunes majeures dans plusieurs
domaines, notamment un problème lié à la source, à la conception, à l'exploitation ainsi qu'à la formation ou à
la certification de l'opérateur). Aussi, beaucoup de ces petits réseaux sont sujets à des avis d’ébullition qui
6
peuvent rester en vigueur pendant plusieurs années (Patrick, 2011). En effet, une étude de Santé Canada sur
les avis d’ébullition dans les Premières nations a montré que 25% de ces avis sont des avis à long terme (en
application pendant plus d’un an) (Health Canada, 2008). Les avis d’ébullition sont 2,5 fois plus souvent
appliqués à des systèmes d’eau potable de communautés de Premières nations qu’à des systèmes d’autres
communautés (Patrick, 2011).
Les petits réseaux ruraux et des Premières nations de l’Amérique du nord sont particulièrement vulnérables en
raison des coûts d’opération du système de traitement de l’eau potable trop importants pour leurs propres
capitaux, bien que quelques juridictions offrent des subventions pour couvrir les coûts d’investissement (Dore
et al., 2013). Ainsi, les coûts de construction et de maintenance d’un système de traitement d’eau représentent
un budget très important pour les petits réseaux qui ont alors davantage de difficultés à moderniser leur système
de traitement (Davies & Mazumder, 2003; Dore et al., 2013). De plus, la plupart des petits systèmes n’ont pas
de programme de protection des sources, ce qui pourrait possiblement réduire les frais de traitement. En effet,
ces programmes sont plus difficiles à mettre en place dans les petits réseaux car souvent les bassins versants
où se trouvent les prises d’eau sont partagés entre plusieurs juridictions (Davies & Mazumder, 2003).
En plus de leurs limitations financières, les petits réseaux manquent de ressource humaine pour satisfaire aussi
bien aux réglementations sur la qualité de l’eau potable qu’aux attentes des consommateurs (Hrudey 2008; Kot
et al., 2011). En effet, les coûts de formation des opérateurs d’eau peuvent être difficiles à gérer pour les petits
systèmes (Hrudey, 2009; Kot et al., 2011). En plus du coût, la durée de la formation du personnel peut être
problématique pour les petits réseaux puisqu’elles se retrouvent sans opérateur durant cette période (Kot et al.,
2011). Plusieurs études rapportent aussi le fardeau de la conformité aux réglementations pour les opérateurs
des petits réseaux (Kot et al., 2011). En effet, les opérateurs sont souvent accablés par la conformité aux
règlementations, ce qui peut entraîner un stress au travail et une relation négative avec la communauté qu’ils
servent (Kot et al., 2011). Aussi, une étude a révélé que le contexte culturel et politique, particulièrement dans
les réseaux des Premières nations rendait difficile le développement d’un sentiment de responsabilité de la part
de l’opérateur et de la prise en charge par celui-ci de la protection de la santé de sa communauté (Smith et al.,
2006).
Finalement, les contraintes économiques et logistiques des petits réseaux impliquent que leurs infrastructures
de traitement de l’eau sont souvent moins complètes que celles des plus grands réseaux. En particulier, leurs
filières de traitement sont souvent moins efficaces pour enlever les précurseurs de SPD présents dans l’eau
brute. Les petits réseaux seraient donc plus vulnérables aux SPD que les grands réseaux. Cependant, peu
d’études ont été menées sur l’occurrence des SPD (réglementés ou non) dans l’eau potable des petits réseaux
des États-Unis et du Canada (Charrois et al., 2004; White et al., 2007; Tung & Xie, 2009), la plupart des études
7
ont été réalisées sur des grands réseaux (Hebert & al, 2010; Mouly et al., 2010; Ye et al, 2009). De plus, les
seules études effectuées aux États-Unis et Canada dans les petits réseaux se concentrent sur les SPD
réglementés (Charrois et al., 2004; White et al., 2007; Tung & Xie, 2009). La vulnérabilité des petits réseaux
face aux SPD n’est donc pas assez documentée. Les études récentes sur de grands réseaux ont permis
d’observer que les variations saisonnières de l’eau brute ainsi que les variations du temps de résidence de l’eau
le long du réseau ont une influence importante sur les niveaux de SPD observés dans ces réseaux (Lebel et al.,
1997; Rodriguez & Sérodes, 2001; Mouly et al., 2010; Mercier-Shanks et al., 2013). Or, peu d’informations sur
la variabilité spatiale et temporelle des SPD dans les petits réseaux sont présentement disponibles. Il n’existe
pas de données sur la variabilité de l’occurrence des SPD basée sur des programmes d’échantillonnage
structurés et réalisés sur un nombre important de petits réseaux représentatifs des petits réseaux du nord-est
de l’Amérique du Nord.
Si les données ne sont pas disponibles dans la littérature, les petits réseaux disposent cependant de bases de
données réglementaires afin d’estimer le niveau d’exposition de la population aux SPD. Là encore, seules des
données sur les SPD réglementés sont disponibles, et ce uniquement dans les juridictions où des normes sur
les SPD sont en vigueur. D’ailleurs, aucune étude spécifique aux petits réseaux n’a été réalisée afin d’estimer
les niveaux de SPD non-réglementés à partir des seules données disponibles (notamment les données
provenant du suivi réglementaire).
Il a été également observé dans les grands réseaux que la saison estivale est une période particulièrement
difficile pour la gestion des SPD car les niveaux de ces contaminants sont souvent plus élevés et les conditions
environnementales de cette saison (en particulier la température et l’activité microbienne) peuvent favoriser la
formation, la décomposition1, la biodégradation2 ou la transformation de certains SPD (Lebel et al., 1997;
Nikolaou et al., 2000; Nikolaou et al., 2001; Rodriguez et al., 2007). Cependant, aucune étude intensive et
concentrée sur la saison estivale concernant l’occurrence des SPD dans les petits réseaux n’a été réalisée.
Finalement, il existe des outils pour les petits systèmes afin d’estimer les besoins et les coûts liés à l’installation
et à l’entretien des systèmes de traitements de l’eau et qu’il y a des indicateurs permettant d’évaluer la
performance des petits réseaux de distribution (Coulibaly & Rodriguez, 2004; Dore et al., 2013). Cependant,
aucun outil permettant d’orienter les décisions quant aux stratégies de suivi des SPD dans le temps et dans
l’espace n’a été développé.
1 Dans cette thèse, le terme décomposition indique une décomposition chimique d’un composé, par hydrolyse par exemple, qui peut être affectée par les propriétés physico-chimiques de l’eau (pH, température, mesure de chlore résiduel libre, etc.) 2 Dans cette thèse, le terme biodégradation indique une décomposition d’un composé par l’activité de micro-organismes, par exemple par l’activité d’un biofilm.
8
Ainsi l’objectif de cette thèse de doctorat est de répondre à tous ces manques de connaissances sur l’évolution
spatiale et temporelle des SPD, particulièrement les SPD non-réglementés, dans les petits réseaux de
distribution d’eau potable. Pour arriver à cet objectif, cette étude se concentre successivement sur les quatre
axes de recherche suivants:
- L’étude de la variabilité spatio-temporelle à long terme des SPD réglementés et non-réglementés dans
l’eau potable de petits réseaux de distribution;
- le développement de modèles afin d’estimer l’occurrence des SPD non-réglementés dans l’eau potable de
petits réseaux de distribution;
- l’étude de la variabilité spatio-temporelle à court terme des SPD réglementés et non-réglementés dans
l’eau potable de petits réseaux de distribution; et,
- le développement d’un outil d’aide à la décision pour le suivi des SPD dans les petits réseaux de distribution
Pour ce faire, deux campagnes d’échantillonnage ont été organisées entre 2010 et 2012 dans 25 petits réseaux,
dont 14 dans la province de Québec (QC) et 11 dans la province de Terre-Neuve-et-Labrador (TN) (Annexe 1).
Les principales caractéristiques des 25 petits réseaux étudiés sont répertoriées dans l’annexe 2.
Les réseaux ont été sélectionnés sur la base de ces six critères de sélection :
- la taille du réseau : les réseaux étudiés sont considérés comme des petits réseaux (caractérisé par une
population desservie par le réseau de moins de 5 000 habitants pour Santé Canada et moins de 10 000
habitants pour USEPA), desservant des populations de moins de 6 200 habitants dans les municipalités
québécoises et de moins de 2 200 dans celles de Terre-Neuve-et-Labrador,
- la nature de la source d’eau : tous les réseaux sont desservis par des eaux de surface,
- la nature du désinfectant : le chlore (ou la chloramine) doit être utilisé comme désinfectant primaire ou
secondaire,
- la distance et l’accessibilité : les réseaux de Québec doivent se trouver à moins de deux heures de voiture
de l’Université Laval. Les réseaux de Terre-Neuve-et-Labrador doivent être le plus proche possible de la
route Transcanadienne afin que l’envoi des échantillons se fasse le plus rapidement possible,
- l’autorisation des autorités locales pour prélever de l’eau dans leur réseau, et
9
- la disponibilité de l’opérateur de l’usine de traitement de l’eau, en particulier à Terre-Neuve-et-Labrador où
ils étaient en charge de l’échantillonnage.
Lors de la première campagne, l’échantillonnage a été réalisé mensuellement entre septembre 2010 et octobre
2011 afin d’observer l’évolution temporelle des SPD sur l’année. Les SPD ont été mesurés à trois différentes
localisations dans le réseau (début, milieu et fin de réseau). Cette première campagne se concentrait sur
l’évolution à long terme des concentrations en SDP. Afin de compléter cette campagne et d’observer l’évolution
à court terme des concentrations en SPD, une deuxième campagne a été organisée. Durant cette deuxième
campagne (dans uniquement les trois réseaux de chaque province présentant les niveaux en SPD les plus
élevés), un échantillonnage quotidien sur une durée d’un mois (en juillet 2012 à QC et août 2012 à NL) a été
réalisé dans six localisations le long du réseau. Cette deuxième campagne a permis, de plus, d’étudier de façon
détaillée l’évolution spatiale des SPD le long du réseau. Durant les deux campagnes, plusieurs indicateurs de
précurseurs de SPD et d’autres paramètres de la qualité de l’eau ont été mesurés à l’eau brute et à chaque
point d’échantillonnage le long du réseau. Au total, cinq familles de SPD ont été mesurées dans l’eau des
réseaux à l’étude: les THM, les AHA, les HAN, les HC et les HNM. Les SPD étudiés sont principalement des
sous-produits issus de la chloration car la majorité des réseaux étudiés utilisent le chlore comme désinfectant
primaire et/ou secondaire. Les méthodes d’échantillonnage et d’analyses utilisées lors des deux campagnes
sont fournies en annexe 3.
Cette thèse se compose de quatre chapitres correspondant chacun à un article scientifique. Le premier chapitre
se concentre sur l’étude de la variabilité spatiale et temporelle des SPD réglementés et non-règlementés au
cours de l’année dans les 25 petits réseaux à l’étude. Les variabilités spatiales des niveaux de SPD ont été
étudiées. C’est-à-dire, les différences de niveaux de SPD mesurés entre les deux régions à l’étude (QC et TN)
et entre les réseaux d’une même région ont été analysées ainsi que l’évolution temporelle au cours de l’année
(selon les saisons). Pour la première fois à notre connaissance, l’occurrence spatiale et temporelle des SPD
non-réglementés a été étudiée dans l'eau potable de plusieurs dizaines de petits réseaux sur la base d’un
programme d'échantillonnage structuré.
La difficulté de développer et d’appliquer des méthodes analytiques capables d’analyser les SPD non-
réglementés ainsi que les coûts d’analyse élevés sont en partie responsables du peu d’informations disponibles
sur l’occurrence des SPD non-réglementés dans les petits réseaux d’eau potable. Le deuxième chapitre se
concentre donc sur une méthode alternative aux analyses de laboratoire souvent coûteuses afin d’évaluer les
niveaux en SPD non-réglementés dans l’eau potable des petits réseaux. Dans ce chapitre, des modèles sont
développés afin d’estimer les niveaux des SPD non-réglementés. Pour cela, les paramètres les plus corrélés
avec les concentrations des SPD non-réglementés ont tout d’abord été identifiés par des analyses de
10
corrélations bivariées. Ensuite, tous les paramètres pouvant influencer l’occurrence des SPD ont été réunis dans
des modèles de régression pour estimer les concentrations de SPD non-réglementés. Pour la première fois, à
notre connaissance, des modèles pour estimer les concentrations des SPD non-réglementés dans l’eau potable
de petits réseaux ont été développés.
À la lumière de l’importante variabilité temporelle et spatiale des SPD dans les petits réseaux observée dans le
premier chapitre, le troisième chapitre se concentre sur l’étude de cette variabilité mais, cette fois-ci, à court
terme. Ainsi, le chapitre se concentre sur l’étude de la variabilité des SPD non-réglementés au cours de l’été,
saison durant laquelle les concentrations et la variabilité des SPD sont globalement les plus importantes. Pour
ce faire, la variabilité temporelle journalière des SPD (réglementés et non-réglementés) a été étudiée sur une
période d’un mois. De même, la variabilité spatiale, sur de multiples points le long du réseau, de ces SPD a été
analysée. A notre connaissance, c’est la première fois que la variabilité spatio-temporelle des SPD non-
réglementés a été étudiée sur le court terme et à haute fréquence dans les petits réseaux de distribution d’eau
potable.
Le quatrième chapitre résume toutes les connaissances accumulées sur l’étude de la variabilité spatiale et
temporelle des SPD en un outil d’aide à la décision pour le suivi des SPD destiné aux petits réseaux. Cet outil
constitue un guide dans la mise en place de la stratégie de suivi des SPD pour les petits réseaux. Il peut être
utilisé pour identifier les meilleurs périodes et sites d’échantillonnage pour le suivi réglementaire des THM et
des AHA (si un suivi est mis en place) et peut aussi être utilisé pour identifier les périodes et lieux dans le réseau
de distribution où l’exposition de la population aux SPD non-réglementés est maximale. L’outil développé, peut
être facilement utilisé par des petits réseaux car il ne nécessite qu’un suivi du chlore libre résiduel le long du
réseau.
11
Bibliographie de l’introduction
Bove, F., Shim, Y., Zeitz, P., 2002. Drinking water contaminants and adverse pregnancy outcomes: a review.
Environmental Health Perspective 110, 61-74.
Bull, R.J., Rice, G., Teuschler, L.K., 2009. Determinants of Whether or Not Mixtures of Disinfection By-Products
are Similar. Journal of Toxicology and Environmental Health, Part A 72, 437-460.
Bull, R.J., Sanchez, I.M., Nelson, M.A., Larson, J.L., Lansing, A.J., 1990. Liver tumor induction in B6C3F1 mice
by dichloroacetate and trichloroacetate. Toxicology 63 (3), 341-359.
Cedergren, M.I., Selbing, A.J., Löfman, O., Källen, B.A.J., 2002. Chlorination byproducts and nitrate in drinking
water and risk for congenital cardiac defects. Environmental Research 2 (89), 124-130.
Charrois, J.W.A., Graham, D., Hrudey, S.E., Froese, K.L., 2004. Disinfection By-products in Small Alberta
Community Drinking Water Supplies. Journal of Toxicology and Environmental Health, Part A 67, 1797-1803.
Conseil de l’Union Européenne, 1998. Directive 98/83/CE relative à la qualité des eaux destinées à la
consommation humaine.
Coulibaly, H.D., Rodriguez, M.J., 2004. Development of performance indicators for small Quebec drinking water
utilities. Journal of Environmental Management 73, 243–255.
Davies, J.M., Mazumder, A., 2003. Health and environmental policy issues in Canada: the role of watershed
management in sustaining clean drinking water quality at surface sources. Journal of Environmental
Management 68, 273-286.
DeAngelo, A.B., Daniel, B.F., Most, B.M., Olson, G.R., 1996. The carcinogenicity of dichloroacetic acid in the
male Fischer 344 rat. Toxicology 114 (3), 207-221.
DeAngelo, A.B., George, M.H., House, D.E., 1999. Hepatocarcinogenicity in the male B6C3F1 mouse following
a lifetime exposure to dichloroacetic acid in the drinking water: dose – response determination and modes of
action. Journal of toxicology and environmental health 58 (8), 485-507.
Department of health & human services USA, 2006. National Toxicology Program, Toxicology and
carcinogenesis studies of bromodichloromethane in male F344/N rats and female B6C3F1 mice (drinking water
studies). National toxicology program, Technical report series, Issue 321.
12
Dore, M.H.I., Singh, R.G., Achari, G., Khaleghi-Moghadam, A., 2013. Cost scenarios for small drinking water
treatment technologies. Desalination and Water Treatment 51, 3628-3638.
Edwards, J.E., Henderson, S.B., Struck, S., Kosatsky, T., 2012. Characteristics of small residential and
commercial water systems that influence their likelihood of being on drinking water advisories in rural British
Columbia, Canada: a cross-sectional study using administrative data. Journal of Water and Health 10 (4), 629-
649.
Environmental Protection Agency, USEPA, 1998. National Primary Drinking Water Regulations: Disinfectants
and Disinfection Byproducts; Final Rule. Federal Register 63 (241).
Environmental Protection Agency, USEPA, 2001. Toxicological review of chloroform.
Environmental Protection Agency, USEPA, 2006a. Initial distribution system evaluation guide for systems
serving fewer than 10,000 people for the final stage 2 disinfectants and disinfection byproducts rule, Washington,
DC, US.
Environmental Protection Agency, USEPA, 2006b. National Primary Drinking Water Regulations: Stage 2
Disinfectants and Disinfection Byproducts Rule. Federal Register 71 (2).
Giller, S., Le Curieux, F., Erb, F., Marzin, D., 1997. Comparative genotoxicity of halogenated acetic acids found
in drinking water. Mutagenesis 12 (5), 321-328.
Giller, S., Le Curieux, F., Gauthier, L., Erb, F., Marzin, D., 1995. Genotoxicity assay of chloral hydrate and
chloropicrine. Mutation research letters 348 (4), 147-152.
Golfinopoulos, S.K., Nikolaou, A.D., Lekkas, T.D., 2003. The Occurrence of Disinfection By-Products in the
Drinking Water of Athens, Greece. Environmental Science and Pollution Research 10 (6), 368-372.
Goslan, E.H., Krasner, S.W., Bower, M., Rocks, S.A., Holmes, P., Levy, L.S., et al., 2009. A comparison of
disinfection by-products found in chlorinated and chloraminated drinking waters in Scotland. Water Research
43, 4698-4706.
Hansen, K.M.S, Zortea, R., Piketty A., Rodriguez Vega, S., Rasmus Andersen, H., 2013. Photolytic removal of
DBPs by medium pressure UV in swimming pool water. Science of the Total Environment 443, 850–856.
Health Canada, 2008. Drinking water advisories in First Nations communities in Canada, Ottawa, Ontario,
CANADA.
13
Health Canada, 2012. Guidelines for Canadian Drinking Water Quality.
Hebert, A., Forestier, D., Lenes, D., Benanou, D., Jacob, S., Arfi, C., et al., 2010. Innovative method for
prioritizing emerging disinfection by-products (DBPs) in drinking water on the basis of their potential impact on
public health. Water research 44, 3147-3165.
Hrudey, S.E., 2009. Public health wonder or unwitting vector of disease: the challenge of delivering safe supplies.
Reviews in Environmental Science and Bio/Technology 8 (3), 235-237.
Hrudey, S. E., 2008. Safe water? Depends on where you live! Canadian Medical Association Journal 178 (8),
975.
Hu, J., Song, H., Addison, J.W., Karanfil, T., 2010. Halonitromethane formation potentials in drinking water.
Water research 44, 105-114.
International Agency for Research on Cancer, 1991. IARC Monographs on the evaluation of carcinogenic risks
to humans: chlorinated drinking-water; chlorination by-products; some other halogenated compounds; cobalt
and cobalt compounds, Lyon, FRANCE
International Agency for Research on Cancer, 1999. Monographs on the evaluation of carcinogenic risks to
humans: some chemicals that cause tumours of the kidney or urinary bladder in rodents and some other
substances, Lyon, FRANCE
International Agency for Research on Cancer, 1999. Re-evaluation of Some Organic Chemicals, Hydrazine and
Hydrogen Peroxide. ARC Monograph on the Evaluation of the Carcinogenic Risks to Humans 71.
Jorgenson, T.A., Meierhenry, E.F., Rushbrook, C.J., Bull, R.J., Robinson, M., 1985. Carcinogenicity of
chloroform in drinking water to male Osborne–Mendel rats and female B6C3F1 mice. Fundamental and applied
toxicology 5 (4), 760-769.
Karanfil, T., Krasner, S.W., Westerhoff, P., Xie, Y., 2008. Disinfection By-Products in Drinking Water: Occurence,
Formation, Health Effects and Control. Washington, DC: American Chemical Society.
Kargalioglu, Y., McMillan, B.J., Minear, R.A., Plewa, M.J., 2002. Analysis of the cytotoxicity and mutagenicity of
drinking water disinfection by-products in Salmonella typhimurium. Teratogenesis, carcinogenesis, and
mutagenesis 22 (2), 113-128.
14
Kot, M., Castleden, H., Gagnon, G.A., 2011. Unintended consequences of regulating drinking water in rural
Canadian communities: examples from Atlantic Canada. Health & Place 17, 1030-1037.
Krasner, S.W, McGuire, M.J., Jacangelo, J.G., Patania, N.L, Reagan K.M., Aieta, E.M., 1989. The occurrence
of disinfection by-products in US drinking water. Journal American Water Works Association 81, 41-53.
Krasner, S. W., Weinberg, H. S., Richardson. S., Pastor, S.J., Chinn, R., Sclimenti, M.J., et al., 2006. Occurrence
of a new generation of disinfection byproducts. Environmental Science & Technology 40, 7175-7185.
Kundu, B., Richardson S.D., Swartz P.D., Matthews P.P., Richard A.M., DeMarini, D.M., 2004. Mutagenicity in
Salmonella of halonitromethanes: a recently recognized class of disinfection byproducts in drinking water.
Mutation Research/Genetic Toxicology and Environmental Mutagenesis 562 (1-2), 39-65.
Lebel, G.L., Benoit, F.M., Williams, D.T., 1997. A one-year survey of halogenated disinfection by-products in the
distribution system of treatment plants using three different disinfection processes. Chemosphere 34 (11), 2301-
2317.
Liang, L., Singer, P.C., 2003. Factors influencing the formation and relative distribution of haloacetic acids and
trihalomethanes in drinking water. Environmental Science & Technology 37 (13), 2920-2928.
Ministère du Développement Durable, de l'Environnement et de la Lutte contre les Changements Climatiques,
MDDELCC, 2012. Réglement sur la qualité de l'eau potable, Annexe 1.3.
Melnick, R. L., Nyskaa, A., Fostera, P.M., Roycrofta. J.H., Kisslingb, G.E., 2007. Toxicity and carcinogenicity of
the water disinfection byproduct, dibromoacetic acid, in rats and mice. Toxicology 230 (2-3), 126-136.
Mercier-Shanks, C., Sérodes, J.-B., Rodriguez, M.J., 2013. Spatio-temporal variability of non-regulated
disinfection by-products within a drinking water distribution system Water Research 47, 3231-3243
Merlet, N., Thibaud, M., Dore, M., 1985. Chloropicrin formation during oxidative treatments in the preparation of
drinkig water. The Science of the Total Environment 47, 223-228.
Mouly, D., Joulin, E., Rosin, C., Beaudeau, P., Zeghnoun, A., Olszewski-Ortar A., et al., 2010. Variations in
trihalomethane levels in three French water distribution systems and the development of a predictive model.
Water Research 44, 5168-5179.
15
Muellner, M.G., Wagner, E.D., McCalla, K., Richardson, S.D., Woo, Y.-T., Plewa, M.J., 2007. Haloacetonitriles
vs. regulated haloacetic acids: are nitrogen-containing DBPs more toxic? Environmental Science & Technology
41, 645-651.
Muller-Pillet, V., Joyeux, M., Ambroise, D., Hartemann, P., 2000. Genotoxic activity of five haloacetonitriles:
comparative investigations in the single cell gel electrophoresis (comet) assay and the ames fluctuation test.
Environmental and Molecular Mutagenesis 36 (1), 52-58.
National Cancer Institute, 1976. Report on the carcinogenesis bioassay of chloroform.
Nieuwenhuijsen, M.J., Martinez, D., Grellier, J., Bennett, J., Best, N., Iszatt, N., et al., 2009. Chlorination
disinfection by-products in drinking water and congenital anomalies: review and meta-analyses. Environmental
Health Perspective 117 (10), 1486-1493.
Nikolaou, A.D., Golfinopoulos, S.K., Kostopoulou, M.N., Lekkas, T.D., 2000. Decomposition of
dihaloacetonitriles in water solutions and fortified drinking water samples. Chemosphere 41, 1149-1154.
Nikolaou, A.D., Lekkas, T.D., Kostopoulou, M.N., Golfinopoulos, S.K., 2001. Investigation of the behaviour of
haloketones in water samples. Chemosphere 44, 907-912.
Organisation Mondiale de la Santé, OMS, 2011. Guidelines for drinking water quality, Geneva, SWITZERLAND
Patrick, R.J., 2011. Uneven access to safe drinking water for First Nations in Canada: Connecting health and
place through source water protection. Health & Place 17, 386-389.
Pérez Pavon, J.L., Herrero Martin, B., Garcia Pinto, C., Moreno Cordero, B., 2008. Determination of
trihalomethanes in water samples: A review. Analytica Chimica Acta 629, 6-23.
Plewa, M. J., Wagner, E. D., Jazwierska, P., 2004. Halonitromethane Drinking Water Disinfection Byproducts:
Chemical Characterization and Mammalian Cell Cytotoxicity and Genotoxicity. Environmental Science and
Technology 38, 62-68.
Richardson, S.D., 2011. Disinfection byproducts: formation and occurrence in drinking water. The Encyclopedia
of Environmental Health 2, 110-136.
Richardson, S.D., Plewa, M.J., Wagner, E.D., Schoeny, R., DeMarini, D.M., 2007. Occurrence, genotoxicity, and
carcinogenicity of regulated and emerging disinfection by-products in drinking water: A review and roadmap for
research. Mutation Research 636, 178–242.
16
Robinson, M., Bull, R.J., Olson, G.R., Stober, J., 1989. Carcinogenic activity associated with halogenated
acetones and acroleins in the mouse skin assay. Cancer Letters 48 (3), 197-203.
Rodriguez, M.J., Sérodes, J.-B., 2001. Spatial and temporal evolution of trihalomethanes in three water
distribution systems. Water Research 35 (6), 1572-1586.
Rodriguez, M.J., Sérodes, J.-B., Levallois, P., 2004. Behavior of trihalomethanes and haloacetic acids in a
drinking water distribution system. Water Research 38, 4367-4382.
Rodriguez, M.J., Sérodes, J.-B., Levallois, P., Proulx, F., 2007. Chlorinated disinfection by-products in drinking
water according to source, treatment, season, and distribution system. Journal of Environmental Engineering
and Science 6, 355-365.
Rook, J.J., 1974. Formation of haloforms during chlorination of natural waters. Water Treatment Examination
23, 234-243.
Samadi, M.T., Nasseri, S., Mesdaghinea, A. R., Alizadehfard, M. R., 2009. Comparison of nanofiltration
efficiency with GAC adsorption and air stripping processes for CHCl3 removal from Tehran drinking water.
Journal of Water Supply: Research and Technology—AQUA 58 (4).
Schneider, M., Quistad, G.B., Casida, J.E., 1999. Glutathione activation of chloropicrin in the Salmonella
mutagenicity test. Mutation Research/Genetic Toxicology and Environmental Mutagenesis 439 (2), 233-238.
Singer, P.C., 1994. Control of disinfection by-products in drinking water. Journal of Environmental Engineering
120 (4), 727-744.
Singer, P.C., 2002. Occurrence of haloacetic acids in chlorinated drinking water. Water Science and Technology:
Water Supply 2 (5-6), 487-492.
Smith, D.W., Guest, R.K., Svrcek, C.P., Farahbakhsh, K., 2006. Public health evaluation of drinking water
systems for First Nations reserves in Alberta, Canada. Revue du génie et de la science de l'environnement, 5
(S1), S1-S17.
Speight, V., Singer, P.C., 2005. Association between residual chlorine loss and HAA reduction in distribution
systems. Journal of American Water Works Association 97 (2), 82.
Statistics Canada, Census of Population 1851 to 2006.
17
Tung, H.-H., Xie, Y.F., 2009. Association between haloacetic acid degradation and heterotrophic bacteria in
water distribution systems. Water Research 43, 971-978.
Villanueva, C., Cantor, K.P., Grimalt, J.O., Malats, N., Silverman D., Tardon, A., et al., 2007. Bladder cancer and
exposure to water disinfection by-products through ingestion, bathing, showering and swimming in pools.
American Journal of Epidemiology 165, 148-156.
Villanueva, C.M., Kogenivas, M., Cordier, S., Templeton, M.R., Vermeulen, R., Nuckols, J.R., et al., 2014.
Assessing exposure and health consequences of chemicals in drinking water: current state of knowledge and
research needs. Environmental Health Perspectives 122 (3).
White, D.M., Garland, S.D., Woolard, C.R., Narr, J., 2007. Investigations into disinfection by-product formation
and control for small drinking water systems in Northern Regions. Cold Region Engineering, 991-998.
Wobma, P., Pernitsky, D., Bellamy, B., Kjartanson, K., Sears, K., 2000. Biological filtration for ozone and chlorine
DBP removal. Ozone Science & Engineering 22, 393-413.
Ye, B., Wang, W., Yang, L., Wei, J., Xueli, E., 2009. Factors influencing disinfection by-products formation in
drinking water of six cities in China. Journal of Hazardous Materials 171, 147-152
19
Chapitre 1
Occurrence of regulated and non-regulated disinfection by-products in small drinking water systems
Abstract
The occurrence of regulated and non-regulated disinfection by-products (DBPs) was investigated in the drinking
water of small systems in two provinces in Canada, Newfoundland and Labrador (NL) and Quebec (QC), through
an intensive sampling program. Sixteen DBPs were studied: four trihalomethanes (THMs), five haloacetic acids
(HAAs), four haloacetonitriles (HANs), one halonitromethane, chloropikrin (CPK) and two haloketones (HKs).
Average measured concentrations of these compounds were much higher than those reported in the literature
for medium and large systems. The measured average value for THMs was 75 µg/L (Stdv = 69 µg/L); HAAs,
77 µg/L (Stdv = 75 µg/L); HANs, 2.5 µg/L (Stdv = 1.8 µg/L); CPK, 0.4 µg/L (Stdv = 0.3 µg/L) and HKs, 6.0 µg/L
(Stdv = 4.5 µg/L). The gap (some 10 times difference) between the average levels of regulated DBPs (THMs,
HAAs) and non-regulated DBPs (HANs, CPK and HKs) is comparable to that observed in large systems where
the occurrence of the same compounds has been reported. Generally, investigated DBPs followed a comparable
seasonal evolution during the year: they decreased between the fall and winter and then increased to eventually
reach a maximum in late summer. This trend was less observable in NL than in QC. However, observed seasonal
fluctuations of DBPs were less considerable than those observed in medium and large systems located in similar
temperate environments reported in the literature. Spatial variations from the plant to the extremities were high
and comparable to those observed in large systems, which is surprising, considering the smaller size of
distribution networks supplying small communities. Generally speaking, the results support the premise that
problems associated with implementing treatment that removes DBP precursors in water submitted to
chlorination can increase population exposure to these contaminants in small systems.
Keywords: Small systems, disinfection by-products, drinking water, haloacetic acids, trihalomethanes; non-
regulated DBPs
Résumé
L’occurrence des sous-produits de la désinfection (SPD) réglementés et non-réglementés a été étudiée dans
l'eau potable de petits réseaux de deux provinces du Canada, Québec (QC) et Terre-Neuve-et-Labrador (TN),
grâce à un programme d'échantillonnage intensif. Seize SPD ont été étudiés: quatre trihalométhanes (THM),
cinq acides haloacétiques (AHA), quatre haloacétonitriles (HAN), un halonitrométhane, la chloropicrine (CPK)
20
et deux halocétones (HC). Les concentrations moyennes mesurées de ces composés étaient beaucoup plus
élevées que celles rapportées dans la littérature pour les grands réseaux et les réseaux de taille moyenne. La
valeur moyenne mesurée pour les THM était de 75 µg/L (E.T. = 69 µg/L); pour les AHA, 77 µg/L (E.T. = 75
µg/L); HAN, 2,5 µg/L (E.T. = 1,8 µg/L); CPK, 0,4 µg/L (E.T. = 0,3 µg/L) et HC, 6,0 µg/L (E.T. = 4,5 µg/L). L'écart
(une différence de facteur 10) entre les niveaux moyens des SPD réglementés (THM, AHA) et des SPD non-
réglementés (HAN, CPK et HC) est comparable à celui observé dans les grands réseaux où l’occurrence de ces
mêmes composés a été rapportée. Généralement, les niveaux des SPD étudiés ont tous suivi une évolution
saisonnière similaire au cours de l'année: les niveaux ont diminué entre l'automne et l'hiver, puis ont augmenté
afin d’atteindre un maximum à la fin de l’été. Cette tendance est moins observable à TN qu'au QC. Cependant,
les fluctuations saisonnières observées des SPD sont moins importantes que celles observées dans les grands
réseaux et les réseaux de taille moyenne situés en milieu tempéré. Les variations spatiales des SPD entre
l'usine de traitement et les extrémités du réseau étaient élevées et comparables à celles observées dans les
grands réseaux, ce qui est surprenant, compte tenu de la petite taille des réseaux de distribution desservant ces
petites municipalités. D'une manière générale, les résultats confirment l'hypothèse que les problèmes associés
à la mise en place de traitements efficaces pour l’enlèvement des précurseurs des SPD dans l'eau soumise à
la chloration peuvent augmenter l'exposition de la population à ces contaminants dans les petits réseaux.
Mots-clés : Petits réseaux, sous-produits de la désinfection, eau potable, acides haloacétiques,
trihalométhanes, SPD non-réglementés
1.1. Introduction
Disinfection by-products (DBPs), generated by the reaction between a chemical disinfectant usually chlorine
with organic matter, are an important concern for water supply, especially surface water supply, as they are
generally rich in natural organic matter (Cedergren et al., 2002; Mouly et al., 2010).
DBPs constitute a large family of compounds presenting various levels of toxicological effects: more than 600
DBPs have been detected, but few have been identified (Richardson, 2011). Trihalomethanes (THMs) and
haloacetic acids (HAAs) are the most prevalent DBPs in drinking water. Their formation is relatively well
understood and their levels are regulated in various countries (in particular for THMs) (Singer, 2002; Richardson,
2011).
Recently, there has been an increased interest in investigating the presence of other DBPs, for example,
haloacetonitriles (HANs), haloketones (HKs) and halonitromethanes (HNMs). In fact, nitrogen DBPs (like HANs
and HNMs) may have greater toxicological effects than HAAs and THMs especially concerning their cytotoxicity
and genotoxicity (Muellner et al., 2007; Richardson et al., 2007), that can become a public health problem with
21
the increasing use of alternative disinfectants such as chloramines as a way to reduce concentrations of
regulated DBPs (Adams et al., 2005).
Most studies on the occurrence of regulated and non-regulated DBPs have been conducted in large systems
(Rodriguez et al., 2004; Krasner et al., 2006; Goslan et al., 2009; Ye et al., 2009; Mouly et al., 2010). However,
small water systems (i.e., serving 5,000 or fewer people) using surface waters may be more vulnerable to DBPs
because of financial constraints, a relatively low capacity to implement adequate treatment technologies to
remove contaminants, in particular DBP precursors, and an inability to hire qualified operators (Charrois et al.,
2004; Coulibaly et al., 2004; Edwards et al., 2012). There is presently very little information on the spatio-
temporal variability of DBPs in the water of small communities. Only a few studies on the occurrence of regulated
DBPs are available (Charrois et al., 2004; Tung et al., 2009). In many countries, as is the case in Canada,
available data are particularly inexistent for non-regulated DBPs.
The purpose of this study is to improve knowledge on the occurrence of regulated and non-regulated DBPs in
the drinking water of small communities in Canada. Accordingly, spatial evolution (inter-regions, intra-regions
and along the distribution system) and temporal evolution (seasonally) of DBPs in water were investigated. For
the first time (to the best of our knowledge), the spatial and temporal presence of non-regulated DBPs was
investigated in the drinking water of small communities based on intensive and structured sampling programs.
1.2. Methodology
1.2.1. Case studies
Twenty-five small municipal systems were selected and studied in two provinces of Canada: Newfoundland and
Labrador (NL) and Quebec (QC). Sampling campaigns in the systems were conducted monthly for one year
between September 2010 and October 2011 (from September 2010 to September 2011 in NL and from October
2010 to October 2011 in QC). All systems were supplied by surface water sources and used chlorine as their
main disinfectant (for primary and secondary disinfection). Systems in NL served a population varying from 330
to 2,120 inhabitants. In QC, systems served a population varying from 1,000 to 6,220 inhabitants. Systems in
NL did not present any prior treatment to chlorination, whereas in QC, systems had implemented one or more
treatment processes prior to disinfection.
1.2.2. Sampling and analysis
During this study 1,500 samples were collected representing over 21,000 data for numerous parameters. Water
was sampled at source (RW) and in the water treatment plant (WTP) just after filtration and before chlorination.
Various points were identified along the distribution system (DS) in order to collect water samples at different
22
residence times (Table 1.1). Water was sampled at the beginning (DS1), middle (DS2) and end of the DS (DS3).
In NL, systems had no treatment prior to chlorination, chlorination being the main treatment process. Thus, in
NL, RW and WTP were represented by the same point. Samples were collected by water operators (in NL) and
by U. Laval personnel (in QC). Samplers were trained to follow equivalent sampling processes for both regions.
Following field collection, samples were sent to the University Laval laboratory for analysis.
Table 1.1: Parameters measured during the sampling campaign
Physico-chemical parameters DBPs
pH T° Turb. Cond. UV-254 DOC Bromide Free Cl. HAAs, THMs, HANs,
HNM, HKs
Raw water (RW)
X X X X X X X - -
WTP* X X X X X X - O O
DS1 - - - - - - - X X
DS2 - - X X X X - X X
DS3 - - - - - - - X X
X Measured - Non measured O Measured if water is treated before chlorination * Only for QC
Five families of DBPs were considered: THMs, HAAs and three families of non-regulated DBPs (HANs, HNM,
HKs). Four THMs (chloroform (TCM), bromodichloromethane (BDCM), dibromochloromethane (DBCM) and
tribromomethane (TBM)), five HAAs (monochloroacetic acid (MCAA), monobromoacetic acid (MBAA),
dichloroacetic acid (DCAA), trichloroacetic acid (TCAA) and dibromoacetic acid (DBAA)), four HANs
(dichloroacetonitrile (DCAN), trichloroacetonitrile (TCAN), bromochloroacetonitrile (BCAN) and
dibromoacetonitrile (DBAN)), one HNM (chloropicrin (CPK)) and two HKs (1,1-dichloropropanone (DCP) and
1,1,1-trichloropropanone (TCP)) were analyzed during the study. The THM quantification limit was 3.6 µg/L for
TCM, 2.0 µg/L for BDCM, 3.4 µg/L for DBCM, 2.6 µg/L for TBM. The HAA quantification limits were 1.0 µg/L
for all HAAs. The HAN, HNM and HK quantification limit were 0.01 µg/L for all compounds. If the concentration
of a DBP was lower than its quantification limit, concentration was considered as null.
Other water characteristics were also measured. The sample strategy allowed us to obtain some 800 measures
of each physical chemical parameter (turbidity, conductivity, UV absorbance at 254 nm (UV-254) and dissolved
organic carbon (DOC)), some 300 for bromide, 490 for pH, about 470 for temperature and about 1,000 for free
chlorine and each THM, HAA and non-regulated DBP under study. Bromide was analyzed by the MA.303-3.1
23
method (Centre d'expertise en analyse environnemental du Québec, 2009). Physico-chemical parameters were
not sampled in every location because we only wanted to gain a general overview of water characteristics in the
DS. Table 1.1 summarizes parameters measured at each sampling point. Details about analytical methods used
are provided elsewhere (Mercier-Shanks et al., 2013).
1.2.3. Data Analysis
Data were collected in a detailed Excel database for which all descriptive analyses were carried out. SYSTAT
13 Software Version No.13.1 was used for statistical analyses for this paper. Statistical analyses included
Student’s t-test (for means comparison of water characteristics measurements between NL and QC) and
ANOVA (for comparison of DBP mean levels between all systems in each region in order to detect a significant
statistical difference in DBP levels between all systems belonging to a region), followed by a Games Howell test
(for quantification of statistical differences between DBP levels in each system belonging to a region and
identification of systems most or least correlated between each other).
1.3. Results and discussion
1.3.1. Portrait of DBP occurrence in small system
Because there are regulations in QC that mandate water utilities supplied by surface waters to remove turbidity
and NOM (mainly through filtration), the water submitted to chlorination (WTP) is of much higher quality in QC
than in NL (data available in Appendix 4). The gap in levels of DBP precursor indicators (UV-254, DOC, SUVA
and bromide) in water before disinfection explains why the DBP occurrences in small water systems (SWS)
under study are significantly different between the two regions. In fact, the levels of these precursors have an
impact on disinfectant demand and the potential for DBP formation. Table 1.2 presents the concentration
distribution of all studied DBPs in SWS for both regions during the study period, as well as their 5% and 95%
percentiles and coefficient of variation (CV).
24
Table 1.2: Regional variation of regulated and non-regulated DBP occurrence in SWS of NL and QC
Compounds
Newfoundland and Labrador Quebec
t Mean (µg L-1)
5% Percent.
95% Percent.
CV Mean
(µg L-1) 5%
Percent. 95%
Percent. CV
TCM 122 20 272 0.62 39 0 97 0.78 **
BDCM 3.5 0 10 1.03 0.7 0 4.0 3.9 **
THMs(1) 125 25 281 0.61 40 0 99 0.78 **
MCAA 2.4 0 7.6 1.12 0.9 0 3.4 1.48 **
DCAA 52 0 113 0.69 19 1.8 50 2.16 **
TCAA 75 3.2 190 0.73 21 0 63 0.97 **
HAAs(2) 129 7.0 277 0.65 40 4.7 113 1.07 **
DCP 2.2 0.2 4.2 0.59 0.9 0.2 2.5 0.82 **
TCP 7.2 0.6 14 0.57 3.0 0.2 7.0 0.69 **
HKs(3) 9.4 1.0 19 0.54 3.9 0.8 8.4 0.63 **
CPK 0.5 0.07 1.1 0.66 0.4 0.1 1.2 0.85 **
TCAN 0.1 0.0 0.2 4.70 0.04 0.0 0.1 0.75 **
DCAN 2.7 0.4 6.1 0.67 2.0 0.3 4.7 0.69 **
BCAN 0.2 0.0 0.4 2.50 0.1 0.01 0.3 0.92 **
DBAN 0.1 0.01 0.1 3.80 0.01 0.0 0.1 2.02 **
HANs(4) 3.1 0.6 6.6 0.69 2.1 0.4 5.0 0.67 **
Brominated DBPs(5)
3.7 0.04 11 0.97 0.9 0.01 4.5 3.57 **
* : Significantly different at the 5% level of significance according to Student’s t-test with SYSTAT ** : Significantly different at the 1% level of significance according to Student’s t-test with SYSTAT (1) Sum of TCM, BDCM, DBCM and TBM: DBCM and TBM levels were under their limits of quantification (data not shown). (2) Sum of MCAA, DCAA, TCAA, MBAA and DBAA: MBAA and DBAA levels were under their limits of quantification (data not shown). (3) Sum of DCP, TCP (4) Sum of TCAN, DCAN, BCAN, DBAN (5) Sum of BDCM, DBCM, TBM, MBAA, DBAA, BCAN, DBAN
It is important to note the gap between levels of regulated DBPs (THMs and HAAs) and non-regulated DBPs
(HANs, HKs and CPK) in both regions. Both THM and HAA average concentrations were more than 10 times
higher than each family of non-regulated DBPs. Such marked differences are observable in other studies.
Regulated DBP were 13 times higher in water from U.S. systems (Krasner et al., 2006) and 65 times higher in
water from systems in Scotland (Goslan et al., 2009).
25
Table 1.3 summarizes the DBP levels reported elsewhere in several studies dealing mainly with large and
medium systems in China (Ye et al., 2009), Taiwan (Chang et al., 2010), Athens (Golfinopoulos et al., 2003),
France (Mouly et al., 2010), the U.S. (Krasner et al., 2006) and Canada, in the greater Québec City area
(Rodriguez et al., 2003). Results show that the occurrence of THMs and HAAs in SWS in NL or in QC was higher
than those reported in larger systems (Table 1.3). Also, THMs and HAAs (except MBAA for which the level was
basically lower than the quantification limit) presented significantly different concentrations in NL and QC. In fact,
the annual average levels for THM and HAA were about three times higher in NL systems than in QC systems.
In small systems of both provinces, the main THM observed was TCM and the two main HAAs observed were
DCAA and TCAA. These compounds are also the compounds most observed in larger systems (for example,
Golfinopoulos et al., 2003; Rodriguez et al., 2003; Ye et al., 2009 and Chang et al., 2010). Also, the annual
average level of TCAA was higher than the DCAA level in both regions.
Table 1.3: Levels of regulated and non-regulated DBPs observed in the literature
Average
THM level (µg L-1)
Average HAA level (µg L-1)
Average HAN level
(µg L-1)
Average CPK level (µg L-1)
Average HK level (µg L-1)
Average level in NL 125 129 3.1 0.5 9.4
Average level in QC 40 40 2.1 0.4 3.9
Chang et al., 2010 14 8.3 - - -
Goslan et al., 2009 74 20 1.4 0.1 -
Ye et al., 2009 12 7.4 - - -
Krasner et al., 2006 31 34 3.0 0.2 2.0
Rodriguez et al., 2004 44 38 - - -
Golfinopoulos et al., 2003 22 19 0.1 0.1 0.5
Concerning the non-regulated DBPs, HANs, CPK and HKs also presented significantly different concentrations
between the two regions (Table 1.2). The annual average levels for HAN and CPK were around 40% and 20%
higher in NL than in QC, respectively. HAN levels in SWS of both regions were higher than the levels reported
in other systems in Athens (Golfinopoulos et al., 2003) and Scotland (Goslan et al., 2009) and comparable to
occurrences found in the U.S. (Krasner et al., 2006) (Table 1.3). CPK occurrence in both regions was higher
than in Scotland (Goslan et al., 2009) and the U.S. (Krasner et al., 2006) (Table 1.3). Also, the annual average
level for HK was two times higher in NL than in QC. The levels of regulated and non-regulated DBPs in SWS
were higher than those observed in larger systems in Athens (Golfinopoulos et al., 2003) and others in the U.S.
26
(Krasner et al., 2006) (Table 1.3). Finally, as shown in Table 1.2, the significant difference between brominated
DBPs between SWS of the two regions, due mainly to BDCM, might be explained primarily by the higher levels
of bromide in raw waters in NL than in QC.
In addition to the high inter-regional differences in DBP levels described above, there were also considerable
intra-regional disparities between SWS within each region. In Table 1.2, these disparities are represented by the
high values of CV for each family (all higher than 54%) and the 5% and 95% percentile values illustrating the
large range of DBP levels. Moreover, statistical analyses on intra-regional differences in DBP levels based on
an ANOVA, followed by a Games Howell test with SYSTAT were carried out (not presented in this paper) on all
DBP measurements made in each system within a region. These analyses provided information on the statistical
differences between DBP levels in each region and identified systems presenting comparable distributions of
DBP levels within a region. Results showed that very few systems presented a comparable distribution of DBP
levels in each region. Such differences were related to characteristics of water quality before disinfection.
These results suggest that the influence of water characteristics before chlorination on DBP levels in the DS is
particularly meaningful in SWS supplied by surface waters. They may experience problems implementing
treatment technologies to remove contaminants (DBP precursors in particular). The difficulty to remove DBP
precursors may explain why the levels of regulated and non-regulated DBPs in small systems supplied by
surface waters are higher than those observed in larger systems.
1.3.2. Temporal variations of DBP levels
Physical and chemical characteristics of raw water usually vary within a year, especially in NL and QC, where
winters are long and very cold, and summers are comparatively short and hot (Rodriguez et al., 2003;
Environment Canada, 2014). Consequently, seasonal variations of water temperature are considerable. And
because seasonal changes occur very rapidly, there is sudden watershed runoff associated with snowmelt in
the spring and relatively rapid decay of vegetation during the fall, a source of NOM in water (Rodriguez et al.,
2003).
Figures 1.1 and 1.2 illustrate the influence of seasonal climate variation on regulated and non-regulated DBP
occurrence in SWS of NL and QC. In general, THM and HAA levels followed water temperature variations in
both regions (temperature of water from RW in NL and from treated water of the WTP in QC). Globally, levels
decreased between the fall and winter (January-March), then increased until summer (July-September) to
eventually reach a maximum in late summer. However this trend was less observable in NL than in QC and less
observable for several DBPs, for example average HAA levels appeared highest in January in NL due to high
levels in a few systems. Temporal variations of THM and HAA levels in each region were considerable. However,
27
fluctuations (calculated by the ratio (Max-Min)/Min*100) appeared higher in QC than in NL. Average THM
concentrations fluctuated by 90% in 13 months in QC and by 66% in NL. HAA concentrations mostly doubled in
QC and fluctuated by 48% in NL. These temporal fluctuations were lower than those observed in a large system
of the Québec City region (Rodriguez et al., 2004). In fact, using our formula, the temporal fluctuations of THMs
and HAAs in Rodriguez’s study were respectively 980% and 770% for levels measured at the end of the system.
Higher temporal variations of THM and HAA levels in QC than in NL might be explained by higher variations of
both water temperature and DBP precursor indicators in WTP water in QC than in RW in NL. Temperature and
DOC level variations in both regions are presented in figures 1.1 and 1.2, respectively, in order to simplify those
figures. In fact, average water temperatures varied from 6 °C to 20 °C in NL and from 2 °C to 19 °C in QC.
Moreover, THM and HAA levels also followed DOC level (DOC level in water from RW in NL and from WTP in
QC) variations in both regions (Figures 1.1 and 1.2). In fact, average levels of DOC in water before chlorination
varied during the sampling campaign from 5.5 mg/L to 8.8 mg/L (variation of 60%) in NL systems and from 1.5
mg/L to 3.4 mg/L in QC systems (variation of 130%). The higher variation in QC WTP waters might be explained,
in part, by the diversity of treatment processes in place in QC. It may also be worth noting that the DCAA level
was higher than the TCAA level in QC in winter (data not shown), as already observed in a previous study with
a large system (Rodriguez et al., 2004). There could be different possible explanations. NOM characteristics of
water may change over the year, as DCAA and TCAA have different precursors (Reckhow et al., 1985;
Rodriguez et al., 2004). Also, the pH of the treated water was slightly higher in winter (5.3-7.6, 5% - 95%
percentiles) than in summer (5.0-7.5, 5% - 95% percentiles) in QC (difference not significant). In fact, TCAA
formation in chlorinated waters is higher at lower pH, whereas DCAA formation is not as affected by pH (Stevens
et al., 1989). However, as the difference in pH between seasons is low, it is unlikely that pH difference is the
explanation. Finally, the possible biodegradation of DCAA (discussed later in this paper) could change the
preponderance of these two HAAs in the DS.
28
a)
b)
Figure 1.1: Temporal evolution of regulated DBP concentrations in SWS in a) NL and b) QC
0
5
10
15
20
25
0
50
100
150
200
250
300
350
Tem
pera
ture
of
RW
(°C
)
Av
era
ge c
on
cen
trati
on
in
all D
S i
n
NL
(μ
g L
-1)
THMs
HAAs
Temperature
0
5
10
15
20
25
0
20
40
60
80
100
120
Tem
pera
ture
of
WT
P (
°C)
Av
era
ge c
on
cen
trati
on
of
all D
S i
n Q
C
(μg
L-1
)
THMs
HAAs
Temperature
29
a)
b)
Figure 1.2: Temporal evolution of non-regulated DBP concentrations in SWS in a) NL and b) QC
Seasonal variations of HANs, CPK and HKs in SWS under study in both regions were also considerable (Figure
1.2). Non-regulated DBPs followed the same temporal evolution as THMs and HAAs. During the sampling
campaign, average HAN concentrations almost doubled in QC and almost tripled in NL from winter to summer.
0
2
4
6
8
10
0
2
4
6
8
10
12
14
16
18
DO
C m
easu
re i
n R
W (
mg
L-1
)
Av
era
ge c
on
cen
trati
on
in
all D
S i
n N
L
(μg
L-1
) HANs
CPK
HKs
DOC
0
2
4
6
8
10
0
1
2
3
4
5
6
7
8
9
10
DO
C m
easu
re i
n W
TP
(m
g L
-1)
Av
era
ge c
on
cen
trati
on
in
all D
S i
n Q
C
(μg
L-1
)
HANs
CPK
HKs
DOC
30
CPK concentrations almost doubled in QC and fluctuated by 63% in NL. Finally, HKs concentrations fluctuated
by 86% in QC and by 73% in NL. Seasonal fluctuations were higher than those observed during a Canadian
study that investigated DCAN (no fluctuation), CPK (no fluctuation), DCP (fluctuation of 25% in the system, using
our formula) and TCP (fluctuation of 8% in the system, using our formula) in mostly medium and large systems
(Williams et al., 1997).
1.3.3. Spatial variations of DBP occurrence within the
distribution systems
Table 1.4 and Appendix 4 present the portrait of the spatial variations for average concentrations of
free residual chlorine, THMs and HAAs within the DS. In Appendix 4, these variations were calculated using the
ratio of DBP concentrations in DS1. On average, free chlorine levels decreased regularly along the DS in SWS
of both regions. For QC systems (Table 1.4), THM concentration increased regularly along the DS. This type of
evolution has also been observed in different studies with medium and large distribution systems (Chen et al.,
1998; Williams et al., 1998; Rodriguez et al., 2001; Mouly et al., 2010). In NL (Table 1.4) average THM
concentrations increased and then stabilized at the end of the DS. HAA concentrations in QC and NL increased
at the beginning of the DS and then decreased. This evolution pattern along the DS has also been observed in
different studies (Rodriguez et al., 2004; Speight et al., 2005). This decrease was due mainly to the drop in
DCAA concentration along the DS (annual average decrease of 12% in QC and 21% in NL between DS1 and
DS3). This drop was particularly important in summer when the concentration of DCAA decreased by an average
of 23% in QC and 40% in NL between DS1 and DS3. This profile is most likely due to the biodegradation of
DCAA by biofilm (Bayless et al., 2008). In fact, between DS1 and DS3, the biodegradation of DCAA might
counterbalance and even exceed the potential formation of this compound along the DS when free chlorine is
decreasing. This can occur even though the residual disinfection level is relatively high in DS1, especially in NL.
Concerning non-regulated DBPs (Table 1.5 and Appendix 5) in both regions, HAN average levels increased
regularly along the DS.
Table 1.4: Spatial variation of residual disinfectant, THMs and HAAs in NL and QC
Region Location
Free Chlorine
annual average
(mg L-1)
Standard
deviation
(mg L-1)
THM annual
average
(µg L-1)
Standard
deviation
(µg L-1)
HAA annual
average
(µg L-1)
Standard
deviation
(µg L-1)
NL
DS1 1.2 0.8 106 62 125 77
DS2 0.7 0.9 136 85 133 87
DS3 0.4 0.5 132 79 130 86
QC
DS1 0.7 0.4 36 26 39 38
DS2 0.5 0.3 41 33 41 38
DS3 0.3 0.3 44 33 37 31
31
The CPK average concentration stabilized at the end of the DS in NL, whereas it slowly decreased in QC (Table
1.5). Even if this result might suggest CPK decomposition in QC exceeds the potential formation of this
compound along the DS, as already observed in the literature (Lebel et al., 1997), its decrease is too low to
clearly interpret a trend here, due to the low level measured.
Table 1.5: Spatial variation of non-regulated DBPs concentration in NL and QC
HK levels evolved differently along the DS in NL and QC. In fact in NL, the average HK concentration decreased
slowly between DS2 and DS3, whereas in QC, it increased. The difference was due to dissimilar variations of
specific HK species, namely DCP and TCP (data not shown). In both regions, average levels of DCP decreased
along the DS, whereas TCP levels increased. The main explanation is that DCP would be gradually oxidized
into TCP (Bougeard et al., 2010; Mercier-Shanks et al., 2013), of which chloroform is one hydrolysis product
(Yang et al., 2007). Since the average decrease of DCP in NL small systems was higher than in QC, DCP
oxidized more easily into TCP due to a higher residual chlorine level (Table 1.4). TCP would be gradually
hydrolyzed into chloroform and induce a global HK decrease. On the other hand, DCP oxidation in QC would be
lower due to a lower residual disinfectant level than in NL (Table 1.4), and possibly a lower TCP hydrolyze which
would induce a global HK increase.
1.3.4. Spatio-temporal variations of DBP occurrence
Values of standard deviations in Tables 1.4 and 1.5 suggest that spatial variations of DBP occurrence within a
region varied seasonally and intra-regionally (from system to system) during the study period. Figure 1.3
illustrates the seasonal and spatial evolution of free chlorine and regulated and non-regulated DBPs during
summer and winter in NL and QC. Variations in both regions differed according to seasons. First, the drop in
free chlorine in both regions is higher in summer than in winter, due to a higher water temperature and chlorine
demand. Also, the decrease in HAAs is only observable in summer, under low chlorine conditions and mostly
due to the DCAA decrease. It reinforces the hypothesis of a possible biodegradation of DCAA by biofilm in
Region Location
HAN annual
average (µg L-1)
Standard deviation (µg L-1)
CPK annual
average (µg L-1)
Standard deviation (µg L-1)
HK annual average (µg L-1)
Standard deviation (µg L-1)
NL
DS1 2.7 1.8 0.5 0.3 9.3 5.2
DS2 3.1 2.1 0.5 0.3 9.5 4.9
DS3 3.4 2.4 0.5 0.4 9.2 5.0
QC
DS1 2.0 1.3 0.4 0.4 3.7 2.5
DS2 2.2 1.4 0.4 0.3 3.9 2.6
DS3 2.3 1.5 0.4 0.3 4.2 2.4
32
summer. This influence of season (temporal) and the location of the system (spatial) has already been observed
in a previous study in Canada (Rodriguez et al., 2004).
Summer Winter
a) g) b)
h)
c) i)
d) j)
e) k) Figure 1.3: Variation of average DBP occurrence along the DS in NL and QC in: a) to f) summer (July-September); g)
to l): winter (January-March)
0
0,5
1
1,5
2
2,5
DS1 DS2 DS3
Fre
e c
hlo
rin
e le
ve
l (m
g L
-1)
NLQC
0
0,5
1
1,5
2
2,5
DS1 DS2 DS3
Fre
e c
hlo
rin
e le
ve
l (m
g L
-1)
NLQC
0
100
200
300
DS1 DS2 DS3
TH
Ms le
ve
l (µ
g L
-1) NL
QC
0
100
200
300
DS1 DS2 DS3
TH
Ms le
ve
l (µ
g L
-1) NL
QC
0
50
100
150
200
250
DS1 DS2 DS3
HA
As le
ve
l (µ
g L
-1) NL
QC
0
50
100
150
200
250
DS1 DS2 DS3
HA
As le
ve
l (µ
g L
-1)
NLQC
0
2
4
6
8
DS1 DS2 DS3
HA
Ns level (µ
g L
-1) NL
QC
0
2
4
6
8
DS1 DS2 DS3
HA
Ns level (µ
g L
-1) NL
QC
0
0,5
1
DS1 DS2 DS3
CP
K le
vel (µ
g L
-1) NL
QC
0
0,5
1
DS1 DS2 DS3
CP
K level (µ
g L
-1) NL
QC
33
f) l) Figure 1.3: Variation of average DBP occurrence along the DS in NL and QC in: a) to f) summer (July-September);
g) to l): winter (January-March) (suite)
HK variations are different between summer and winter. In summer, HK level decreases and in winter it
increases, especially in NL. This observation strengthens the idea that once the majority of the HK precursors
have reacted, especially in summer when the temperature is high and in NL where free chlorine level is low,
DCP would gradually be oxidized into TCP using residual free chlorine. This spatio-temporal evolution has
already been observed in a previous study in Canada (Mercier-Shanks et al., 2013). At the same time, TCP
would be hydrolyzed into chloroform in SWS, which decreased the total HK level and increased the level of THM
at the end of the network. On the contrary, HAN and CPK variations are comparable between summer and
winter.
Finally, all these observations are made regionally. In fact, standard deviations (represented by errors bars in
Figure 1.3) within regions are high for all seasons. In addition to temporal and spatial variations, it is important
to take into account specific SWS characteristics which have a distinct influence on DBP occurrences. In fact,
DBP variations are different, depending on site-specific distribution system conditions, especially for HAAs, as
DCAA biodegradation depends on biofilm presence in the SWS.
1.4. Conclusions
This study succeeded in enhancing knowledge of the spatio-temporal occurrence of regulated DBPs and HANs,
HKs, and CPK in SWS. This knowledge is necessary to better appraise the particular characteristics of such
systems regarding their high DBP levels in comparison to medium and larger systems (for which said issues
have been documented extensively, particularly for regulated DBPs).
Also, potential biological degradation of DCAA, widely observed in large systems, appears surprising for SWS
considering the size of the networks. It suggests that water residence times may be comparable, in some cases,
to residence times in large systems (perhaps due to stagnation or to a poor hydraulic management).
Generally speaking, observations made in various selected SWS concurred with those made regionally.
However, analyses of data for the SWS under study demonstrated that local characteristics of water quality,
climate, treatment, operations and distribution systems make the occurrence of DBPs in SWS site-specific.
Because of the limited human and technical resources of SWS, the variability of such local characteristics is
0
5
10
15
20
DS1 DS2 DS3
HK
s level (µ
g L
-1) NL
QC
0
5
10
15
20
DS1 DS2 DS3
HK
s le
vel (µ
g L
-1) NL
QC
34
higher than in medium and large systems. The results may prove useful to authorities for establishing investment
priorities to improve SWS capacities to deliver safe water. They may also serve in epidemiological or risk analysis
studies to better assess population exposure to potentially harmful DBPs.
1.5. References
Adams, C., Timmons, T., Seitz, T., Lane, J., Levotch, S., 2005. Trihalomethane and haloacetic acid disinfection
by-products in full-scale drinking water systems. Journal of Environmental Engeneering 131(4), 526-534.
Bayless, W., Andrews, R.C., 2008. Biodegradation of six haloacetic acids in drinking water. Journal of Water
and Health 6(1), 15-22.
Bougeard, C.M., Goslan, E.H., Jefferson, B., Parsons, S.A., 2010. Comparison of the disinfection by-product
formation potential of treated waters exposed to chlorine and monochloramine. Water Research 44, 729-740.
Cedergren, M.I., Selbing, A.J., Löfman, O., Källen, B.A., 2002. Chlorination byproducts and nitrate in drinking
water and risk for congenital cardiac defects. Environmental Research 89, 124-130.
Centre d’expertise en analyse environnemental du Québec, 2009. Détermination des anions en faible
concentration dans l’eau de consommation: méthode par chromatographie ionique. Méthode MA 303-Ions 3.1,
Rév. 3, Ministère du Développement durable, de l’Environnement et de la Lutte contre les Changements
Climatiques, Québec.
Chang, H.H., Tung H.H., Chao C.C., Wang G.S., 2010. Occurrence of haloacetic acids (HAAs) and
trihalomethanes (THMs) in drinking water of Taiwan. Environmental Monitoring and Assessment 162, 237-250.
Charrois, J.W.A., Graham D., Hrudey S.E., Froese K.L. 2004. Disinfection by-products in small Alberta
community drinking water supplies. Journal of Toxicology and Environmental Health, Part A: Current Issues. 67
(20-22), 1797-1803.
Chen, W.J., Weisel, C.P., 1998. Halogenated DBP concentrations in a distribution system. Journal of American
Water Works Association 90(4), 151-163.
Coulibaly, H.D., Rodriguez, M.J., 2004. Development of performance indicators for small Quebec drinking water
utilities. Journal of Environmental Management 73, 243–255.
Edwards, J.E., Henderson, S.B., Struck, S., Kosatsky, T., 2012. Characteristics of small residential and
commercial water systems that influence their likelihood of being on drinking water advisories in rural British
35
Columbia, Canada: a cross-sectional study using administrative data. Journal of Water and Health 10(4), 629-
649.
Environment Canada, 2014. Canadian Climate Normals 1971-2000.
Environment Canada, 2010. Regional water issues. Water Survey of Canada: Hydrology of Canada.
Golfinopoulos, S.K., Nikolaou, A.D., Lekkas, T.D., 2003. The occurrence of disinfection by-products in the
Drinking Water of Athens, Greece. Environmental Science and Pollution Research 10(6), 368-372.
Goslan, E.H., Krasner, S.W., Bower, M., Rocks, S.A., Holmes, P., Levy, L.S., et al., 2009. A comparison of
disinfection by-products found in chlorinated and chloraminated drinking waters in Scotland. Water Research
43, 4698-4706.
Karanfil, T., Krasner, S.W., Westerhoff, P., Xie, Y., 2008. Disinfection by-products in drinking water: occurrence,
formation, health effects and control. American Chemical Society, Washington, DC.
Krasner, S. W., Weinberg, H. S., Richardson. S., Pastor, S.J., Chinn, R., Sclimenti, M.J., et al., 2006. Occurrence
of a new generation of disinfection byproducts. Environmental Science & Technology 40, 7175-7185.
Lebel, G.L., Benoit, F.M., Williams, D.T., 1997. A one-year survey of halogenated disinfection by-products in the
distribution system of treatment plants using three different disinfection processes. Chemosphere 34, 2301-
2317.
Mercier-Shanks, C., Sérodes, J.-B., Rodriguez, M.J., 2013. Spatio-temporal variability of non-regulated
disinfection by-products within a drinking water distribution system. Water Research 47, 3231-3243
Mouly, D., Joulin, E., Rosin, C., Beaudeau, P., Zeghnoun, A., Olszewski-Ortar A., et al., 2010. Variations in
trihalomethane levels in three French water distribution systems and the development of a predictive model.
Water Research 44, 5168-5179.
Muellner, M.G., Wagner, E.D., McCalla, K., Richardson, S.D., Woo, Y.-T., Plewa, M.J., 2007. Haloacetonitriles
vs. regulated haloacetic acids: are nitrogen-containing DBPs more toxic? Environmental Science & Technology
41, 645-651.
Reckhow, D.A., Legube, B., Singer, P.C., 1985. Mechanisms of organic halide formation during fulvic acid
chlorination and implications with respect to preozonation, in: Joley, R.L., Condie L.W., Johnson J.D., Katz S.,
36
Minear R.A., Mattice, J.S., Jacobs V.A. (Eds.), Water Chlorination: Chemistry, Environmental Impact and Health
Effects, Vol. 5, Lewis Publishers Inc., Chelsea.
Richardson, S.D., 2011. Disinfection by products: formation and occurrence in drinking water, in: Nriagu. J.O.
(Ed.), Encyclopedia of Environmental Health 2, Elsevier, Burlington, 110-136.
Rodriguez, M.J., Sérodes, J.-B., 2001. Spatial and temporal evolution of trihalomethanes in three water
distribution systems. Water Research 35, 1572-1586.
Rodriguez, M.J., Vinette, Y., Sérodes, J.-B., Bouchard, C., 2003. Trihalomethanes in drinking water of greater
Quebec region (Canada): occurrence, variations and modelling. Environmental Monitoring and Assessment 89,
69-93.
Rodriguez, M.J., Sérodes, J.-B., Levallois, P., 2004. Behavior of trihalomethanes and haloacetic acids in a
drinking water distribution system. Water Research 38, 4367-4382.
Simpson, K.L., Hayes, K.P., 1998. Drinking water disinfection by-products: an Australian perspective. Water
Research 32, 1522-1528.
Singer, P.C., 2002. Occurrence of haloacetic acids in chlorinated drinking water. Water Science and Technology:
Water Supply 2 (5-6), 487-492.
Speight, V., Singer, P.C., 2005. Association between residual chlorine loss and HAA reduction in distribution
systems. Journal of American Water Works Association 97(2), 82-91.
Stevens, A.A., Moore, L.A., Miltner, R. J., 1989. Formation and control of non-trihalomethane disinfection by-
products. Journal of American Water Works Association 81(8), 54-60.
Tung, H.H. & Xie, Y. F., 2009. Association between haloacetic acid degradation and heterotrophic bacteria in
water distribution systems. Water Research 43, 971-978.
Williams, D.T., Benoit, F.M., Lebel, G.L., 1998. Trend in levels of disinfection by-products. Environmetrics 9,
555-563.
Williams, D.T., Lebel, G. L., Benoit, F.M., 1997. Disinfection by-products in Canadian drinking water.
Chemosphere 34, 299-316.
Yang, X., Shang, C., Westerhoff, P., 2007. Factors affecting formation of haloacetonitriles, haloketones,
chloropicrin and cyanogen halides, during chloramination. Water Research 41, 1193–1200.
37
Ye, B., Wang, W., Yang, L., Wei, J., Xueli, E., 2009. Factors influencing disinfection by-products formation in
drinking water of six cities in China. Journal of Hazardous Materials 171, 147-152.
39
Chapitre 2
Models for estimating non-regulated disinfection by-product occurrence in small drinking water systems
Le chapitre précédent a mis en lumière les variations spatiales et temporelles assez importantes de l’occurrence
des SPD dans les petits réseaux. Dans des études précédemment réalisées dans des grands réseaux ou en
laboratoire, il a été observé que plusieurs facteurs influencent l’occurrence des SPD et donc leurs variations. En
effet, la concentration en matière organique de l’eau brute joue un rôle très important et donc le type de
traitement employé et sa capacité à enlever la matière organique de l’eau brute (Rodriguez & Sérodes, 2001;
Bull, et al., 2009). Aussi, plusieurs paramètres physico-chimiques de l’eau influencent la formation des SPD
comme la dose de désinfectant, le temps de contact de l’eau avec le désinfectant (Rodriguez & Sérodes, 2001;
Rodriguez, et al., 2004; Bull, et al., 2009), le pH (Liang & Singer, 2003; Bull, et al., 2009) et la température de
l’eau traitée (Liang & Singer, 2003; Rodriguez, et al., 2004; Bull, et al., 2009). Mais les paramètres qui influencent
l’occurrence des SPD dans les petits réseaux n’ont pas été étudiés.
Ainsi l’objectif de ce deuxième article est d’identifier les paramètres qui influencent l’occurrence des SPD dans
les petits réseaux, et en particulier les SPD non-réglementés, encore peu étudiés. Une fois ces paramètres
identifiés, des modèles d’estimation des niveaux des SPD non-réglementés dans les petits réseaux sont
développés à partir des paramètres déterminants. Ces modèles présentent diverses applications pour les petits
réseaux, aussi bien opérationnelles dans l’usine de traitement de l’eau, qu’épidémiologiques pour estimer
l’exposition de la population aux SPD non-réglementés.
Abstract
Among all disinfection by-products (DBPs) produced by chlorine, only trihalomethanes (THMs) and haloacetic
acids (HAAs) are regulated in drinking water in various countries. Most DBPs are not regulated. Very little
information exists on the occurrence of non-regulated DBPs, particularly in small water systems. Paradoxically,
small systems are more vulnerable to DBPs because of a relatively low capacity to implement adequate
treatment technologies to remove DBP precursors. The purpose of this study was to develop models to estimate
non-regulated DBP levels in small systems. Since no information on non-regulated DBP levels in small systems
is available, an intensive sampling program was carried out in 25 small systems in two provinces of Canada,
Newfoundland and Labrador, and Quebec. Sixteen DBPs were investigated: four THMs, five HAAs, four
haloacetonitriles (HANs), one halonitromethane (HNM) and two haloketones (HKs). In order to develop the
models, the variables most correlated with non-regulated DBP levels were identified by bivariate correlations
40
between non-regulated DBPs and various water quality parameters. This analysis showed that levels of non-
regulated DBPs are affected primarily by dissolved organic carbon, UV absorbance at 254 nm, temperature and
pH in water before disinfection. Also, the type of treatment, season and location in the distribution system
influence their formation. Multivariate linear mixed regression models were developed to estimate HAN, HK and
HNM levels from water characteristics in the water treatment plant, THMs, HAAs and residual disinfectant levels.
The models obtained have a good explanatory capacity since R2 varies from 0.77 to 0.91 according to the
compound and condition for application (season and type of treatment). Model validation with an independent
database suggests their high capacity for generalization.
Keywords: Small systems, drinking water, non-regulated disinfection by-products, haloacetonitriles,
halonitrometanes, haloketones
Résumé:
Parmi tous les sous-produits de la désinfection (SPD) produits par le chlore, seuls les trihalométhanes (THM) et
les acides haloacétiques (AHA) sont réglementés dans l'eau potable dans plusieurs pays. La plupart des SPD
ne sont pas réglementés. Très peu d'informations existent sur l’occurrence des SPD non-réglementés, en
particulier dans les petits réseaux d'eau potable. Paradoxalement, les petits réseaux sont plus vulnérables aux
SPD en raison d'une capacité réduite à mettre en place des technologies de traitement de l’eau appropriées à
l’enlèvement efficace des précurseurs des SPD. Le but de cette étude était de développer des modèles pour
estimer les niveaux de SPD non-réglementés dans les petits réseaux. Comme peu d’informations sur les
niveaux des SPD non-réglementés dans les petits réseaux sont disponibles, un programme d'échantillonnage
intensif a été réalisé dans 25 petits réseaux dans deux provinces du Canada, Québec et Terre-Neuve-et-
Labrador. Seize SPD ont été étudiés: quatre THM, cinq AHA, quatre haloacétonitriles (HAN), un
halonitrométhane (HNM) et deux halocétones (HC). Afin de développer les modèles, les variables les plus
corrélées avec les niveaux de SPD non-réglementés ont été identifiées par des corrélations bivariées entre les
SPD non-réglementés et différents paramètres de qualité de l'eau. Cette analyse a montré que les niveaux de
SPD non-réglementés sont influencés principalement par le carbone organique dissous (COD), l’absorbance
UV à 254 nm, la température et le pH de l'eau avant la désinfection. En outre, le type de traitement, la saison et
la localisation dans le système de distribution influencent la formation des SPD à l’étude. Des modèles de
régression linéaire mixte multivariés ont été développés pour estimer les niveaux des HAN, HC et HNM à partir
des caractéristiques de l'eau à l'usine de traitement de l'eau ainsi que les niveaux de THM, de AHA et de
désinfectant résiduel dans le réseau. Les modèles obtenus sont assez efficaces car le R2 varie de 0,77 à 0,91
en fonction du composé et des conditions de l'application (la saison et le type de traitement). La validation du
modèle avec une base de données indépendante suggère une grande capacité de généralisation.
41
Mots-clés: Petits réseaux, eau potable, sous-produits de la désinfection non-réglementés, haloacétonitriles,
halonitrométhanes, halocétones
2.1. Introduction
Chlorination of water to prevent microbiological contamination results in the formation of a wide range of organic
compounds known as disinfection by-products (DBPs) (Rook, 1974; Richardson, 2011). More than 600 DBPs
have been detected, but few have been identified (Richardson, 2011). These compounds have potential adverse
effects on human health. Trihalomethanes (THMs) and haloacetic acids (HAAs) are the most prevalent DBPs in
drinking water. Their formation is relatively well understood and their levels are regulated in various countries
(especially THMs) (Singer, 2002; Rodriguez et al., 2004; Richardson, 2011) and various models have been
developed to estimate their levels (Sadiq & Rodriguez, 2011).
However, most DBPs are not regulated. For example, haloacetonitriles (HANs), haloketones (HKs) and
halonitromethanes (HNMs) are non-regulated DBPs present in water treated with chlorine or in combination with
alternative disinfectants (Krasner et al., 1989; Plewa et al., 2004; Hua & Reckhow, 2007; Richardson, 2011).
However, some non-regulated DBPs may have higher toxicological effects than THMs and HAAs (Muellner et
al., 2007). In fact, HNMs and HANs both present cytotoxicity and genotoxicity levels higher than THMs and
HAAs (Muellner et al., 2007).
Small water systems (i.e., serving 5,000 or fewer people) supplied by surface waters are generally vulnerable to
high DBP levels. In fact, these systems present a relatively low capacity to implement adequate treatment
technologies to remove DBP precursors, and an inability to hire qualified operators to manage operational
conditions (Coulibaly & Rodriguez, 2004; Edwards et al., 2012). Our previous study showed that average
measured concentrations of DBPs in small water systems (SWS) were much higher than those reported in the
literature for medium and large systems (Guilherme & Rodriguez, 2014).
The purpose of this investigation was to develop models to estimate the occurrence of non-regulated DBPs in
SWS based on information of relevant parameters easily and regularly monitored, including regulated DBPs.
DBP formation and speciation are affected by various parameters. For example, the nature and amounts of
organic matter (Karanfil et al., 2008) and disinfectant concentration and type (Adams et al., 2005; Crittenden et
al., 2005; Bull et al., 2009; Bougeard et al., 2010) are known to impact DBP formation. In a previous study
(Guilherme & Rodriguez, 2014), results showed that like in large systems, DBP variability in SWS is also
influenced by season and location within the distribution system (DS). In the present paper, based on information
generated during an intensive sampling program carried out in a large diversity of SWS, parameters most
correlated with non-regulated DBP concentrations were identified through bivariate correlation analyses. All
42
parameters influencing DBP occurrence were then brought together in multilinear regression models to estimate
the concentrations of non-regulated DBPs in specific locations of the DS. In this study and according to our
knowledge, this is the first time that the occurrence of non-regulated DBPs in small municipal systems has been
modelled from information obtained through a robust sampling program.
2.2. Methodology
2.2.1. Case studies
An intensive and structured sampling program was carried out in 25 SWS in two provinces of Canada,
Newfoundland and Labrador (NL), and Quebec (QC). System sampling campaigns were conducted monthly
over one year from September 2010 to October 2011. All systems were supplied by surface water sources and
used chlorine as the main disinfectant (for primary and secondary disinfection). In NL, systems served a
population varying from 330 to 2,120 inhabitants. In QC, systems served a population varying from 1,000 to
6,220 inhabitants. Systems in NL did not present any prior treatment to chlorination, whereas in QC, systems
mostly implemented conventional treatment processes prior to disinfection.
2.2.2. Sampling and analysis
Water was sampled at the source (RW) and in the water treatment plant (WTP) just after filtration and before
disinfection. Various points were identified along the DS in order to collect water samples at different residence
times (Table 2.1). Water was sampled at the beginning (DS1), middle (DS2) and end of the DS (DS3). In NL,
systems had no treatment prior to chlorination which was the main treatment process. Thus, in NL, RW and
WTP were represented by the same point. Samples were collected by water operators (in NL) and by University
Laval personnel (in QC). Samplers were trained to follow equivalent sampling processes for both provinces.
Following field collection, the samples were sent to the University Laval laboratory for analysis.
Table 2.1: Parameters measured during the sampling campaigns
Physico-chemical parameters DBPs
pH T° Turbidity UV-254 DOC Free
Chlorine HAAs, THMs, HANs,
HNM, HKs
Raw water (RW)
X X X X X - -
WTP* X X X X X o o
DS1, DS2, DS3
- - - - - X X
X Measured - Non measured o Measured if water is treated before chlorination * Only for QC
43
In RW and WTP, indicators for precursors of DBPs were estimated using various parameters (e.g., ultraviolet
absorbance at 254 nm (UV-254) and dissolved organic carbon (DOC)). DOC primarily represents total dissolved
organic carbon (humic substances and non-humic substances) and UV-254 is an indicator of aromatic
compounds in water. Other parameters were measured (pH and temperature) due to their importance in DBP
speciation or formation kinetics (Liang & Singer, 2003; US Environmental Protection Agency, 2006; Bull et al.,
2009; Zhang et al., 2013). Residual disinfectant levels (free chlorine) and DBP concentrations were measured
in treated water after chlorination. The sample strategy allowed us to obtain some 800 measurements for
turbidity, UV-254 and DOC, some 500 measurements for pH and temperature and about 1,000 measurements
for free residual chlorine. Some parameters were not sampled in every location because we sought only to gain
a general overview of water characteristics in the DS. Table 2.1 summarizes parameters measured at each
sampling point.
Five families of DBPs were considered in the study: THMs, HAAs and three families of non-regulated DBPs,
HANs, HNMs and HKs. Four THMs (chloroform (TCM), bromodichloromethane (BDCM), dibromochloromethane
(DBCM) and tribromomethane (TBM)), five HAAs (monochloroacetic acid (MCAA), monobromoacetic acid
(MBAA), dichloroacetic acid (DCAA), trichloroacetic acid (TCAA) and dibromoacetic acid (DBAA)), four HANs
(dichloroacetonitrile (DCAN), trichloroacetonitrile (TCAN), bromochloroacetonitrile (BCAN) and
dibromoacetonitrile (DBAN)), one HNM (Chloropikrin (CPK)) and two HKs (1,1-dichloropropanone (DCP) and
1,1,1-trichloropropanone (TCP)) were analyzed during the study. The quantification limit for THM species was
1.8 µg/L for TCM, 1.0 µg/L for BDCM, 1.7 µg/L for DBCM, and 1.3 µg/L for TBM. The quantification limit for all
HAA species was 1.0 µg/L. The quantification limit for HANs, HNMs and HKs was 0.01 µg/L. The sample
strategy allowed us to obtain 900 measurements of each DBP under study. Complementary information about
analytical methods used is provided elsewhere (Mercier-Shanks et al., 2013).
2.2.3. Modelling
The software product used for statistical analyses was SYSTAT 13 Software Version No.13.1. The modelling
approach aimed at estimating non-regulated DBP concentrations at specific locations of the DS using data
collected easily and regularly at the treatment plant and within the DS. We chose to propose models for various
types of treatments and seasons to improve the conditions of model applicability. The mixed linear regression
analysis was chosen. This method maximizes the covariance between independent variables and the dependent
variable (that is, non-regulated DBP concentrations) to obtain optimal estimations. Also, linear mixed regression
models include additional random-effect terms, and are appropriate for representing dependent data, for
example when data are gathered over time on the same individuals (distributions systems here) (Minalu et al.,
2011). Physico-chemical parameters of water (pH, DOC, UV-254, temperature, turbidity) at the treatment plant
in QC and in raw water in NL, and THM and HAA levels in the DS were considered as independent variables.
44
Dependent variables included concentrations of HANs, HKs and HNMs in water of the DS. Explanations on
sample location choices are provided in section 2.3.4.
2.3. Results
Because of the diversity of factors influencing DBP occurrence, we conducted bivariate correlation analyses
between the observed levels of non-regulated DBPs and the following factors: treatment conditions, raw water
quality parameters and the levels of regulated DBPs (THMs and HAAs). By taking into account all bivariate
analyses, multivariate models were finally developed to represent the simultaneous influence of the most
correlated factors. Except for treatment conditions influence (presented in the next section), the data of all
systems were considered for the statistical analyses.
2.3.1. Influence of treatment conditions on DBP
speciation
In order to illustrate the influence of treatment conditions (type of treatment and disinfectant used), only four
SWS with different treatments were selected among all SWS studied in QC and NL in order to compare the
same number of data in each category of treatment. Table 2.2 presents the average concentrations of non-
regulated DBPs in the selected SWS during summer (July - September) and winter (January - March). Results
show that non-regulated DBP levels are influenced by type of treatments and disinfectant used. Indeed, in SWSa
and SWSb, HAN, CPK and HK levels were almost 60% lower in systems using chloramines than in systems
using chlorine. Previous studies have shown that monochloramine is less reactive than free chlorine, and forms
DBPs at much lower concentrations than free chlorine (Carlson & Hardy, 1998; Crittenden et al., 2005). However,
this observation concerning non-regulated DBPs has never been observed in small systems. Also, the type of
treatment influences non-regulated DBP levels in SWS. Indeed, the higher the number of treatment processes,
the lower the DBP levels (Table 2.2). HAN, CPK and HK levels decrease as treatment steps increase in SWS,
especially for HKs. Thus, DBP levels in SWSb were less important than in SWSc and even less than in SWSd.
This is due mostly to the presence of treatments prior to disinfection that reduce natural organic matter (NOM)
levels before chlorination. Also, it is important to note that non-regulated DBP levels were higher in small systems
like SWSb, SWSc and SWSd than levels already measured in the literature in larger systems (Golfinopoulos et
al., 2003; Krasner et al., 2006).
45
Table 2.2: Average levels of non-regulated DBP in location DS2 during summer (July, August and September) and in winter (January, February and March) in four SWS using different types of treatment (number of observations per season = 3)
HANs (µg/L) CPK (µg/L) HKs (µg/L)
Summer Winter Summer Winter Summer Winter
SWSa Ozone + Fil.2 +
Chloramine 0.52 0.84 0.13 0.17 1.64 2.20
SWSb Ozone + CFS1 + Fil.2 + Chlorine
1.54 1.75 0.35 0.42 4.88 5.99
SWSc CFS1 + Fil.2 +
Chlorine 5.09 2.13 0.66 0.42 5.58 2.28
SWSd Chlorine only
5.99 2.48 0.69 0.42 20.8 9.75
1 CFS: Coagulation – Flocculation – Sedimentation 2 Fil.: Filtration
Table 2.2 also shows that non-regulated DBP levels were higher in summer than in winter, as already observed
in a previous study (Guilherme & Rodriguez, 2014). In fact, such differences between seasons are explained by
the difference in temperatures and DBP precursor levels between seasons in QC and NL. Average raw water
temperatures varied from 6 °C to 20 °C in NL and from 2 °C to 19 °C in QC between winter and summer, and
DOC levels in water before chlorination varied from 5.5 mg/L to 8.8 mg/L in NL systems and from 1.5 mg/L to
3.4 mg/L in QC systems between winter and summer. Also, seasonal variations in DBP levels were particularly
observable in drinking waters from systems with simple treatments like SWSc and SWSd, where treated water
can be more easily influenced by raw water quality variations.
2.3.2. Influence of water quality characteristics on non-
regulated DBP levels in the DS
Table 2.3 presents Spearman correlation factors between characteristics of water before disinfection and non-
regulated DBP concentrations in the treated waters of all SWS in NL and QC. Results show that HKs are the
non-regulated DBPs most significantly correlated with organic matter indicators. In fact, UV-254, DOC, SUVA
and turbidity are strongly correlated with HKs. SUVA represents the ratio UV-254/DOC*100 and constitutes an
indicator of carbon aromaticity. The literature reports a good correlation between the SUVA-value and DBP-
formation potential (Matilainen et al., 2011). Results also show that turbidity, pH and temperature are not strongly
correlated with DBP levels. In fact, the results of Table 2.3 show that pH is weakly and negatively correlated with
DBP levels especially HKs. Previous studies have shown pH affects the stability of non-THM DBPs (Croue &
Reckhow, 1989; Fang et al., 2010). Indeed, increasing pH may enhance decomposition rates of DCAN and
DCP, thereby reducing their levels (Yang et al., 2007). Finally, the results show that non-regulated DBPs are not
46
strongly correlated with temperature. Usually, reaction rates increase with increasing temperature. If the
compounds are relatively stable (e.g., chloroform), their formation increases with rising temperatures (Yang et
al., 2007). But, at the same time, rising temperatures can enhance the decomposition rates of several unstable
compounds like DCP, TCP, DBAN and DCAN (Nikolaou et al., 2000; Nikolaou et al., 2001; Zhang & Minear,
2002). In summary, organic matter indicators like DOC, UV-254 and SUVA were the most significantly correlated
parameters with the occurrence of non-regulated DBPs in the SWS under study.
Table 2.3: Spearman correlation matrix between water quality characteristics of treated water at the treatment plant (after filtration and before disinfection) and non-regulated DBP levels in the DS in all SWS (Number of observations for each parameter = 300)
DOC UV-254 SUVA Turbidity pH Temp.
DCAN 0.36** 0.30** 0.067* 0.096** -0.055 0.33**
HANs 0.36** 0.31** 0.082* 0.10** -0.050 0.33**
CPK 0.26** 0.25** 0.12** 0.15** -0.062 0.12**
DCP 0.78** 0.76** 0.53** 0.56** -0.20** 0.13**
TCP 0.75** 0.72** 0.46** 0.45** -0.20** 0.27**
HKs 0.80** 0.77** 0.50** 0.50** -0.21** 0.25**
*: Significant correlation at 5% level according to two-tailed test with SPSS© **: Significant correlation at 1% level according to two-tailed test with SPSS©
2.3.3. Correlations between regulated and non-regulated
DBPs
Results presented in the last section show that levels of non-regulated DBPs in the SWS under study depend
on treatment type and NOM indicators. It is also useful to evaluate whether levels of regulated DBPs can be
used as surrogates for evaluating the levels of non-regulated DBPs. Table 2.4 presents correlations between
DBP levels in location DS2 with a Spearman correlation matrix. Correlations were calculated in one DS location
in order to control the locational (spatial) variations along the DS. In a review based on large systems (Bond et
al., 2011), HANs were strongly correlated with regulated DBPs, whereas HNMs were only correlated with DCAA.
For the small systems under study, the results show that HAN and HK levels are strongly correlated with levels
of regulated DBPs and that CPK is moderately correlated with both regulated DBPs. This observation is
encouraging, in that regulated DBPs can, in some way, be used as surrogates for non-regulated DBPs in small
systems.
47
Table 2.4: Spearman correlation matrix between DBP average levels in location DS2 (number of observations for each parameter = 300)
THMs HAAs HANs CPK HKs
HANs 0.61** 0.65** 1.0 - -
CPK 0.38** 0.53** 0.71** 1.0 -
HKs 0.74** 0.76** 0.58** 0.40** 1.0
*: Significant correlation at 5% level according to two-tailed test with SPSS© **: Significant correlation at 1% level according to two-tailed test with SPSS©
Tables 2.5a and 2.5b present correlations between variations of DBP concentrations along the DS in summer
and winter. Spatial variations are represented by ratios, i.e., the ratio between DBP levels in DS2 and DS1 and
ratio between DBP levels in DS3 and DS2. Thus, the correlations are also based on the ratio values. Results
show that correlations between regulated and non-regulated DBP variations were slightly different depending
on location in the system (between DS1 and DS2 or between DS2 and DS3). However, the results of Table 2.5
reveal that HAA variations within the DS are more correlated with non-regulated DBP variations than THM
variations, especially in summer. This observation may be explained by the decomposition or biodegradation of
some species of HAAs and HANs as well as CPK in summer, in particular, when water approaches the extremity
of the DS, as already observed in the past for larger systems (Lebel et al., 1997; Chang et al., 2010; Mercier-
Shanks et al., 2013). Thus, seasons seem to influence not only global non-regulated DBP levels (as mentioned
in section 2.3.1), but also their spatial variability in small systems. Finally, it is important to note that levels
correlation between each couple of non-regulated DBP families were comparable.
Table 2.5: Spearman correlation matrix between DBP variation ratios a) DS1 and DS2 winter/summer; b) DS2 and DS3 winter/summer a)
THMs HAAs HANs CPK
HANs 0.57**/0.21 0.58**/0.51** - -
CPK 0.53**/0.27* 0.65**/0.65** 0.87**/0.74** -
HKs 0.43**/-0.03 0.55**/0.39** 0.83**/0.79** 0.83**/0.54**
*: Significant correlation at 5% level according to two-tailed test with SPSS© **: Significant correlation at 1% level according to two-tailed test with SPSS©
b)
THMs HAAs HANs CPK
HANs 0.60**/0.01 0.54**/0.49** - -
CPK 0.48**/-0.07 0.61**/0.64** 0.82**/0.77** -
HKs 0.52**/-0.07 0.57**/0.38** 0.90**/0.79** 0.81**/0.76**
**: Significant correlation at 1% level according to two-tailed test with SPSS©
48
To better understand the relationships between non-regulated and regulated DBPs, correlations between these
two types of compounds were investigated according to parameters known to influence these relationships
(levels of organic matter indicators, locations within the DS and the seasons). To do so, four categories of DOC
values were selected (Table 2.6). Figure 2.1 presents the conditions at which regulated DBP levels (THMs or
HAAs) are the most correlated with HAN, CPK and HK levels. Accordingly, when DOC is lower and at the end
of the DS and during the warm semester (summer and fall), HANs are more correlated with HAAs. This supports
the fact that variations of HANs are more correlated with HAA variations than THM variations (Table 2.5) due
especially to their decomposition or biodegradation potential. Conversely, HANs seem mostly correlated with
THMs when the DOC level is high and at the beginning of the DS, or during the cold period. This can be explained
by the fact that in cold semester (winter and spring) or in the presence of a high DOC level, it is unlikely that
HAN would be degraded and would be probably be formed continuously along the DS, like THMs.
Table 2.6: Classification of DOC levels at the WTP (number of observations = 300)
DOC Level Very high High Moderate Low
DOC (mg/L) DOC≥ 8.0 8.0 <DOC≤ 4.0 4.0 <DOC≤ 2.0 DOC< 2.0
Number of measurements
44 90 93 73
Note: The central value of 4.0 mg/L was chosen in reference to the TOC threshold value of USEPA DBP stage 2
(USEPA, 2006). Accordingly, if the source water annual average TOC concentration before any treatment, exceeds 4.0
mg/L at any treatment plant treating surface water or ground water under the direct influence of surface water, United
States Environmental Protection Agency (USEPA), DBP stage 2 requires that the system must resume routine
monitoring.
Figure 2.1 also shows that CPK is correlated mostly with HAAs, especially at the end of the DS and during the
warm semester. This observation in small systems is supported by previous investigations that observed that
HAAs (especially DCAA) and CPK biodegradation or decomposition were both observable especially in summer,
in the DS or under laboratory conditions (Lebel et al., 1997; Bayless & Andrews, 2008). Finally, also according
to Figure 2.1, HKs are mostly correlated with HAAs except when the DOC level is very high or when the DOC
level is low and in cold semester. This supports the fact that CPK and HKs variations are more correlated with
HAA variations than with THM variations (Table 2.4).
49
*: Significant correlation at 5% level according to two-tailed test with SPSS© **: Significant correlation at 1% level according to two-tailed test with SPSS©
Figure 2.1: Identification of the most correlated regulated DBPs (Spearman correlation factors indicated) with non-
regulated DBPs according to DOC levels before disinfection, season and location in the DS
HANs CPK HKs
50
THMs are the most regulated and monitored DBPs in drinking water worldwide. However, the results in Figure
2.1 suggest that they cannot be used alone as surrogates for non-regulated DBPs in the small systems under
study (in most conditions of precursor levels, season and location). On the contrary, HAA variability appears a
better indicator for levels of non-regulated DBPs.
2.3.4. Modelling levels of non-regulated DBPs in small
water systems
The next step in this study was to consider together the various parameters that have moderate or high
correlations with regulated DBP occurrence and variation (identified above) in order to develop multivariate
models for estimating the levels of these substances in the SWS. To achieve this, linear mixed regression models
for non-regulated DBP levels were developed based on measurements of raw and WTP water quality
parameters and regulated DBP levels, as well as other conditions (seasons, type of treatment and location within
the DS). The Akaike Information Criterion (AIC) was used for model selection (Akaike et al., 1973). The models
were optimized by selecting only the most significant explanatory variables in the model through backward
elimination. Step-by-step, variables were rejected from the model based on the optimization of AIC values (as
low as possible). The fitness and performance of the regression models were estimated by the coefficient of
determination (R2), normal probability plot of residuals, residuals versus predicted and data order plots and the
validity of Shapiro-Wilk test on residuals (Razali & Wah, 2011). To simplify interpretation, the coefficient of
determination, R2, is a pseudo-R2 defined as the squared correlation coefficient between the observed and the
predicted response (Pearson’s correlation), based on the definition of R2 in linear regression (Xu, 2003).
To take into account treatment, seasons and locations in the DS, the models were calibrated according to
specific conditions for these characteristics. Treatment influence was taken globally. To simplify, all different
forms of treatments were classified in two categories, with or without treatment that can remove NOM prior to
disinfection (only systems using chlorination as secondary disinfectant were selected). Thus, all NL systems
were included in the “without treatment prior to disinfection” category, as well as one system from QC (twelve
systems in total). In the “with treatment prior to disinfection” category, only systems from QC were included
(thirteen systems in total). In case of a presence of a treatment prior to disinfection, differences in treatments
used were not taken into account in models. Seasons were also taken into account by separating the warm
semester (summer and fall) and the cold semester (winter and spring). DBP levels in the DS were chosen in
specific locations that typically correspond to their maximum level measurement. Thus, regulated DBP levels
correspond to levels observed at the end of the DS (DS3) for THMs and at the middle of the DS (DS2) for HAAs.
These locations for regulated DBPs were chosen on the basis of regulations for THMs (Health Canada, 2012;
Ministère du Développement Durable, de l'Environnement et de la Lutte contre les Changements Climatiques,
MDDELCC, 2012) and previous studies conducted by our team that concluded that the HAA maximum level is
51
measured near the middle of the DS (Rodriguez et al., 2004; Guilherme & Rodriguez, 2014). Non-regulated DBP
levels were also estimated at their maximum level locations of the DS, corresponding to the levels measured at
the middle of the DS (DS2). In fact, our previous study showed that non-regulated DBP level variability along
the DS is more associated with HAAs than THMs (Guilherme & Rodriguez, 2014). Considering specific locations
for non-regulated DBPs improves the statistical power and conditions of application of the models. In WTP,
DOC, UV-254, turbidity, SUVA, pH and temperature were taken into account. Also, free chlorine (Free_Cl) levels
in DS3 were included in the models.
Table 2.7 summarizes the results of the linear mixed regression models that associate non-regulated DBPs with
regulated DBPs and WTP water characteristics. The models generated present high R2 values from 0.77 to 0.91.
HAN models present R2 values from 0.77 to 0.91, CPK models from 0.79 to 0.88 and HK models from 0.79 to
0.91 depending of the season and the treatment employed in the system. These values are comparable to R2
values resumed in a review on THM and HAA models (Sadiq & Rodriguez, 2011). Generally speaking, the
models highlighted the main correlations brought to light previously in section 2.3.3. Indeed, HAAs are the most
significant variable for most models estimating HAN, CPK and HK levels in SWS (fifth column in Table 2.7).
Also, THMs are a significant variable for non-regulated DBP models in systems without treatment prior to
chlorination (presenting probably higher DOC levels in WTP), as revealed in Figure 2.1. Finally, the relatively
high correlations between regulated DBPs and non-regulated DBPs in the bivariate statistical analysis in Table
2.5 and between UV-254 and DOC with non-regulated DBPs in Table 2.3 explain the presence of these
parameters in all the models.
52
Table 2.7: Multivariate regression models for non-regulated DBP levels
*: The variable presenting the highest level of significance α in each model
Additional models were developed excluding DOC (and consequently SUVA) from the explanatory variables. In
fact, DOC is not a regulated parameter and its cost for analysis is relatively high compared to other operational
parameters such as UV-254, pH, temperature and residual chlorine. The results show that the performance of
models that exclude DOC as an explanatory variable (sixth column of Table 2.7) are comparable to the original
Tre
atm
ent
Season R2 Models Most
significant variable*
R2 without DOC or SUVA
R2 without THMs and
HAAs
With
out t
reat
men
t prio
r di
sinf
ectio
n
Warm semester
0.88
HANs (µg/L) = 2.07 + 8.05.10-3*HAAs + 8.67.10-2*Temp. + 3.91.10-3*THMs -
11.7*UV-254 + 0.583*UV-254*DOC – 4.02.10-2*SUVA
HAAs 0.88 0.81
0.86 CPK (µg/L) = -0.417+8.83.10-4*HAAs –
2.54*UV-254 + 0.103*DOC + 0.189*SUVA HAAs 0.81 0.81
0.88
HKs (µg/L) = 2.74 + 1.63.10-2*HAAs + 12.8*UV-254 + 0.109*Temp. + 9.20.10-
3*THMs + 0.454*Turb. – 1.50*Free_Cl – 0.352*SUVA + 5.19.10-2*pH
HAAs 0.87 0.89
Cold semester
0.77
HANs (µg/L) = 1.62 + 9.59.10-3*HAAs – 0.306*SUVA + 4.80.10-2*Temp. –
0.531*Turb. + 8.39.10-2*pH + 0.588*UV-254 + 2.01.10-2*Free_Cl
HAAs 0.74 0.65
0.88
CPK (µg/L) = -0.639 + 1.72.10-3*HAAs + 2.55*UV-254 +
0.136*DOC – 0.437*UV-254*DOC – 1.38.10-2*Temp. + 4.69.10-2*pH
HAAs 0.83 0.70
0.91
HKs (µg/L) = 2.11 + 2.88.10-2*THMs – 0.996*pH + 0.139*Temp. + 1.24*Turb. +
1.02*SUVA + 0.788*DOC – 0.521*Free_Cl – 0.582*UV-254*DOC + 0.519*UV-254
THMs 0.84 0.91
With
trea
tmen
t prio
r di
sinf
ectio
n Warm
semester
0.89 HANs (µg/L) = 2.96 + 3.23.10-2*HAAs –
0.337*pH + 0.606*Free_Cl + 23.8*UV-254 – 0.282*SUVA – 2.29*UV-254*DOC
HAAs 0.89 0.82
0.79 CPK (µg/L) = (0.481 +
1.35.10-3*HAAs + 0.412*UV-254*DOC – 7.37.10-3*DOC + 0.637*UV-254)2
HAAs 0.79 0.73
0.87 HKs (µg/L) =3.68 + 3.18.10-2*HAAs +
1.33*Free_Cl + 46.0*UV-254 – 0.372*pH - 0.660*SUVA – 0.541*UV-254*DOC
HAAs 0.86 0.82
Cold semester
0.91 HANs (µg/L) = 0.590 +
5.16.10-2*HAAs – 0.584*UV-254 – 1.36*UV-254*DOC
HAAs 0.91 0.77
0.88 CPK (µg/L) = 9.14.10-2 +
8.80.10-3*HAAs + 7.82.10-2*Turb. – 0.716*UV-254*DOC + 1.09*UV-254
HAAs 0.83 0.67
0.79 HKs (µg/L) = 5.43 + 4.30.10-2*HAAs +
0.730*Turb. – 0.557*pH – 6.06*UV-254 + 1.43*UV-254*DOC
HAAs 0.79 0.75
53
models. Also, the results show that models are less efficient when THMs and HAAs are not considered as
explanatory variables (represented by the lower R2 values in column 7 of Table 2.7). Excluding regulated DBPs
has a greater impact on model performance than excluding DOC. Fortunately, regulated DBP levels information
is generally available for small systems through regulatory monitoring.
2.3.5. Validation of HAN and HK models
The models developed were validated using a different database generated in a previous investigation
conducted with the Quebec Environment Ministry (Développement durable, de l’Environnement et de la Lutte
contre les changements climatiques, MDDELCC). In 2010, a sampling campaign was carried out in ten municipal
water systems in Quebec. Five of these systems were small systems also investigated in this research. Others
were medium water systems (MWS). In the 2010 study, two sampling campaigns were conducted: one in winter
and one in summer. Each system was sampled in RW, in the WTP and in six locations along the DS. Four THMs
(TCM, BDCM, DBCM and TBM), five HAAs (MCAA, MBAA, DCAA, TCAA and DBAA), four HANs (DCAN, TCAN,
BCAN and DBAN) and two HKs (DCP and TCP) were measured along the DS. Also, various water quality
parameters (DOC, UV-254, pH and temperature) were measured at every location, from RW to the end of the
DS. Sampling and analyzing protocols were similar to the protocols presented above. Only one SWS used
disinfection as only treatment, all the other SWS and MWS had treatments to remove NOM prior disinfection.
Using the database generated in these two previous campaigns, data of physico-chemical parameters in water
before chlorination, THMs at the end of the DS and HAAs at the middle of the DS were incorporated into the
models presented in Table 2.7 depending of the season and treatment employed (models included DOC and
regulated DBPs). Data for one SWS were used for models corresponding to systems without treatment prior to
disinfection. Data for the rest of the systems were used in models for systems with treatment prior to disinfection.
The results of models with the validation database were compared to average observed levels of non-regulated
DBP levels in the middle and end of the DS. Figures 2.2 and 2.3 present the correlation levels between observed
values of HANs and HKs and the model estimated values. Globally, the predictions are very good. For all
systems, and although the estimations are generally lower than observed values, the correlation is strong (R2 =
0.68 for HANs and R2 = 0.92 for HKs). Moreover, if we consider only the five SWS of the validation database,
the correlations are even stronger (R2 = 0.85 for HANs and R2 = 0.95 for HKs). These results confirm that the
models developed are effective at estimating non-regulated DBPs in SWS. The underestimation of model values
in comparison with observed values (an average of + 10%) may be associated with the fact that the limit level
of quantification (LQ) for DBPs was lower in the study used to develop models (LQ = 0.01 µg/L per DBP) than
in the study used to validate the models (LQ = 0.2 µg/L per DBP). Thus, levels in the validation database may
be overestimated, especially because BCAN and BDAN levels are often close to the LQ. Generally speaking,
54
differences between estimated and observed non-regulated DBP are lower for HANs (between ± 0.03% and ±
84%) than for HK (between ± 2.0% and ± 110%).
a) b)
Figure 2.2: Validation of HAN models: correlation between observed and estimated values in a) all systems, b) only
small systems
a) b)
Figure 2.3: Validation of HK models: correlation between observed and estimated values in a) all systems, b) only
small systems
2.4. Discussion and conclusions
This study reveals that variability of non-regulated DBP levels in small systems is statistically influenced by the
variability of several parameters and conditions: type of treatment, raw water quality characteristics at the source
and at the treatment plant (especially DOC and UV-254 levels), location in the DS and seasons. The levels of
R² = 0,68
0
2
4
6
8
0 2 4 6 8
Est
imat
ed H
AN
leve
l (µ
g/L)
Observed HAN level (µg/L)
R² = 0,85
0
2
4
6
8
0 2 4 6 8E
stim
ated
HA
N le
vel (
µg/
L)Observed HAN level (µg/L)
R² = 0,92
0
2
4
6
8
0 2 4 6 8
Est
imat
ed H
K le
vel (
µg/
L)
Observed HK level (µg/L)
R² = 0,95
0
2
4
6
8
0 2 4 6 8
Est
imat
ed H
K le
vel (
µg/
L)
Observed HK level (µg/L)
55
non-regulated DBP levels are also correlated relatively highly with HAA and THM levels. Depending on the type
of treatment, the season and location in the DS, regulated DBPs could be used as surrogates for non-regulated
DBPs. Linear mixed regression models were developed to estimate levels of non-regulated DBPs based on
levels of HAAs, THMs, free chlorine and several physico-chemical water quality levels. Model estimations were
quite good and their validation with data from other campaigns demonstrated the capacity for generalization of
the models developed.
The models can be used for various applications. They can be used for routine operational purposes, for example
to estimate the benefits of handling pH or residual chlorine levels regarding non-regulated DBP occurrence.
They can also be used for infrastructure planning purposes, in particular for evaluating the potential impacts on
non-regulated DBPs of removing organic matter by treatment. Finally, the models can be used to estimate
population exposure to non-regulated DBPs occurring in drinking water of small systems, an important issue for
regional risk analysis regarding potential health impacts of DBPs in drinking water.
This study has some limitations. Specific characteristics of operational conditions (e.g., precise chlorine doses)
and distribution system hydraulics (e.g., residence time of water) were not documented. Also, the models
developed do not provide optimum sample location for non-regulated DBPs. In fact, future studies might focus
on a better understanding of spatial variations of these DBPs along the DS (with several sampling locations
distributed in various locations of the DS) to better understand the impact of residence time. Finally, future
studies should also consider other compounds as other HNMs and HKs, as well as iodinated and nitrogenous
DBPs that have high toxicological relevance and for which there is still very little information for SWS.
2.5. References
Adams, C., Timmons, T., Seitz, T., Lane, J., Levotch, S., 2005. Trihalomethane and haloacetic acid disinfection
by-products in full-scale drinking water systems. Journal of Environmental Engeneering 131(4), 526-534.
Akaike, H., Petrov, N. B., Csaki, F., 1973. Information theory and extension of the maximum likelihood principle.
Budapest: Akademia Kiado, 267-281.
Bayless, W., Andrews, R.C., 2008. Biodegradation of six haloacetic acids in drinking water. Journal of Water
and Health 6 (1), 15-22.
Bond, T., Huang, J., Templeton, M. R., Graham, N., 2011. Occurrence and control of nitrogenous disinfection
by-products in drinking water - A review. Water Research 45, 4341-4354.
56
Bougeard, C.M., Goslan, E.H., Jefferson, B., Parsons, S.A., 2010. Comparison of the disinfection by-product
formation potential of treated waters exposed to chlorine and monochloramine. Water Research 44, 729-740.
Bull, R. J., Rice, G., Teuschler, L. K., 2009. Determinants of Whether or Not Mixtures of Disinfection By-Products
are Similar. Journal of Toxicology and Environmental Health, Part A. 72, 437-460.
Carlson, M., Hardy, D., 1998. Controlling DBPs with monocholoramine. Journal of American Water Works
Association 90 (2), 95-106.
Chang, H. H., Tung, H. H., Chao, C. C., Wang, G. S., 2010. Occurrence of haloacetic acids (HAAs) and
trihalomethanes (THMs) in drinking water of Taiwan. Environmental Monitoring and Assessment 162, 237-250.
Coulibaly, H. D., Rodriguez, M. J., 2004. Development of performance indicators for small Quebec drinking water
utilities. Journal of Environmental Management 73, 243–255.
Crittenden, J. C, MWH, Inc., 2005. Water treatment: Principles and design. Hoboken, N.J: John Wiley & Sons,
Inc.
Croue, J.-P., Reckhow, D. A., 1989. Destruction of chlorination byproducts with sulfite. Environmental Science
& Technology 23 (11).
Environmental Protection Agency, USEPA, 2006. Initial distribution system evaluation guidance manual for the
final stage 2 disinfectants and disinfection byproducts rule.
Edwards, J. E., Henderson, S. B., Struck, S., Kosatsky, T., 2012. Characteristics of small residential and
commercial water systems that influence their likelihood of being on drinking water advisories in rural British
Columbia, Canada: a cross-sectional study using administrative data. Journal of Water and Health 10 (4), 629-
649.
Fang, J., Ma, J., Yang, X., Shang, C., 2010. Formation of carbonaceous and nitrogenous disinfection by-
products from the chlorination of Microcystis aeruginosa. Water Research 44, 1934-1940.
Golfinopoulos, S. K., Nikolaou, A. D., Lekkas, T. D., 2003. The Occurrence of Disinfection By-Products in the
Drinking Water of Athens, Greece. Environmental Science and Pollution Research 10 (6), 368-372.
Guilherme, S., Rodriguez, M. J., in press. Occurrence of regulated and non-regulated disinfection by-products
in small drinking water systems. Chemosphere
Health Canada, 2012. Guidelines for Canadian Drinking Water Quality
57
Hua, G., Reckhow, D. A., 2007. Comparison of disinfection byproducts formation from chlorine and alternative
disinfectants. Water Research 41, 1667-1678.
Karanfil, T., Krasner, S. W., Westerhoff, P., Xie, Y., 2008. Disinfection by-products in drinking water: occurrence,
formation, health effects and control. American Chemical Society, Washington, DC.
Krasner, S. W., Mcguire, M. J., Jacangelo, J. G., Patania, N. L., Reagan, K. M., Aieta, E. M., 1989. The
occurrence of disinfection by-products in US drinking water. Journal American Water Works Association 81, 41-
53.
Krasner, S. W., Weinberg, H. S., Richardson. S., Pastor, S.J., Chinn, R., Sclimenti, M.J., et al., 2006. Occurrence
of a new generation of disinfection byproducts. Environmental Science and Technology 40 (23), 7175-7185.
Lebel, G. L., Benoit, F. M., Williams, D. T., 1997. A one-year survey of halogenated disinfection by-products in
the distribution system of treatment plants using three different disinfection processes. Chemosphere 34 (11),
2301-2317.
Liang, L., Singer, P. C., 2003. Factors influencing the formation and relative distribution of haloacetic acids and
trihalomethanes in drinking water. Environmental Science and Technology 37 (13), 2920-2928.
Matilainen, A., Gjessi, E. T., Lahtinen, T., Hed, L., Bhatnagar, A., Sillanpää, M., 2011. An overview of the
methods used in the characterisation of natural organic matter (NOM) in relation to drinking water treatment
Chemosphere, 83, 1431-1442.
Mercier-Shanks, C., Sérodes, J.-B., Rodriguez, M. J., 2013. Spatio-temporal variability of non-regulated
disinfection by-products within a drinking water distribution system. Water Research 47, 3231-3243
Minalu, G., Aerts, M., Coenen, S., Versporten, A., Muller, A., Adriaenssens. N., Beutels. P., Molenberghs, G.,
Goossens H., Hens, N., 2011. Application of mixed-effects models to study the country-specific outpatient
antibiotic use in Europe: a tutorial on longitudinal data analysis. Journal of Antimicrobial Chemotherapy 22, 79-
87.
Ministère du Développement Durable, de l'Environnement et de la Lutte contre les Changements Climatiques,
MDDELCC, 2012. Réglement sur la qualité de l'eau potable, Annexe 1.3.
Muellner, M. G., Wagner, E.D., McCalla, K., Richardson, S.D., Woo, Y.-T., Plewa, M.J.,2007. Haloacetonitriles
vs. regulated haloacetic acids: are nitrogen-containing DBPs more toxic?. Environmental Science and
Technology 41, 645-651.
58
Nikolaou, A. D., Golfinopoulos, S. K., Kostopoulou, M. N., Lekkas, T. D., 2000. Decomposition of
dihaloacetonitriles in water solutions and fortified drinking water samples. Chemosphere 41, 1149-1154.
Nikolaou, A. D., Lekkas, T. D., Kostopoulou, M. N., Golfinopoulos, S. K., 2001. Investigation of the behaviour of
haloketones in water samples. Chemosphere 44, 907-912.
Plewa, M. J., Wagner, E. D., Jazwierska, P., 2004. Halonitromethane Drinking water disinfection byproducts:
chemical characterization and mammalian cell cytotoxicity and genotoxicity. Environmental Science and
Technology 38, 62-68.
Razali, N. M., Wah, Y. B., 2011. Power comparisons of Shapiro-Wilk, Kolmogorov-Smirnov, Lilliefors and
Anderson-Darling tests. Journal of Statistical Modeling and Analytics 2 (1), 21-33.
Richardson, S. D., 2011. Disinfection by products: formation and occurrence in drinking water. The Encyclopedia
of Environmental Health 2, 110-136.
Rodriguez, M.J., Sérodes, J.-B., 2001. Spatial and temporal evolution of trihalomethanes in three water
distribution systems. Water Res. 35, 1572-1586.
Rodriguez, M. J., Sérodes, J.-B., Levallois, P., 2004. Behavior of trihalomethanes and haloacetic acids in a
drinking water distribution system. Water Research 38, 4367-4382.
Rook, J. J., 1974. Formation of haloforms during chlorination of natural waters. Water Treatment Examination
23, 234-243.
Sadiq, R., Rodriguez, M. J., 2011. Empirical Models to Predict Disinfection By-products (DBPs) in Drinking
Water. Encyclopedia of Environmental Health 282-295.
Singer, P. C., 2002. Occurrence of haloacetic acids in chlorinated drinking water. Water Science and
Technology: Water Supply 2 (5-6), 487-492.
Xu, R., 2003. Measuring explained variation in linear mixed effects models. Statistics in Medicine 22, 3527-3541.
Yang, X., Shang, C., Westerhoff, P., 2007. Factors affecting formation of haloacetonitriles, haloketones,
chloropicrin and cyanogen halides, during chloramination. Water Research 41, 1193 – 1200.
Zhang, X., Minear, R. A., 2002. Decomposition of trihaloacetic acids and formation of the corresponding
trihalomethanes in drinking water. Water Research 36, 3665-3673.
59
Zhang, X.-l., Yang, H., Wang X., Fu J., Xie, Y. F., 2013. Formation of disinfection by-products: Effect of
temperature and kinetic modeling. Chemosphere 90, 634-639.
61
Chapitre 3
Short-term spatial and temporal variability of disinfection by-product occurrence in small drinking water systems
A la suite des deux premiers chapitres, la variabilité de l’occurrence des SPD à long terme dans les petits
réseaux ainsi que les facteurs influençant cette variabilité ont été grandement étudiés. Cependant, peu
d’informations sont disponibles sur la variabilité à court-terme des SPD dans l’eau potable de petits réseaux.
L’objectif de ce troisième article est donc de compléter l’information présentée dans les articles précédents à
l’aide d’une seconde campagne d’échantillonnage basée essentiellement sur l’étude de la variabilité spatiale et
temporelle à court terme (sur un mois) des SPD, et particulièrement des SPD non-réglementés, dans l’eau
potable de petits réseaux. Pour ce faire, une campagne d’échantillonnage intensive a été organisée dans les
trois petits réseaux de chaque province (QC et TN) de la première campagne présentant les niveaux de SPD
les plus élevés. Cette nouvelle base de données complète la première base de données car elle permet d’étudier
les variations temporelles sur une très courte période et les variations spatiales de manière plus précise. Cette
campagne a été réalisée en été car cette saison représente une saison critique pour les petits réseaux pour le
suivi des SPD, notamment à cause des températures élevées et une forte prédisposition de certains SPD à se
de dégrader ou à se transformer.
Abstract
Disinfection by-products (DBPs) constitute a large family of compounds. Trihalomethanes (THMs) and
haloacetic acids (HAAs) are already regulated in various countries, but most DBPs are not. Monitoring DBPs
can be delicate, especially for small systems, because various factors influence their formation and speciation.
Short-term variations of DBPs can be important and particularly difficult for small systems to handle because
they require qualified operators to manage operational conditions. According to our knowledge and for the first
time, our study covers the short-term variability of regulated and non-regulated DBP occurrence in small systems
in the summer. In the course of our study, an intensive sampling program was carried out in six small systems
in Canada. Small systems in the provinces of Newfoundland and Labrador and Quebec were sampled daily at
the water treatment plant and at six different locations along the DS. Five DBP families were studied: THMs,
HAAs, haloacetonitriles, halonitromethanes and haloketones. Results show there were considerable variations
in DBP levels from week to week during the month of study and even from day to day within the week. Variations
were greatly influenced by the differences in characteristics (water quality, infrastructure, operations, etc.) that
existed from system to system. Likewise, the large number of sampling locations within the DS served to
62
observe, in detail, DBP variations along the DS. Observations revealed some decomposition of non-regulated
DBPs never before studied in small systems. Finally, this study reveals that the temporal variability of DBPs is
also influenced by spatial location along the DS. Thus, spatial and temporal variations of DBPs in the short term
may make it difficult to select representative locations and periods for DBP monitoring purposes in SWS.
Keywords: Small systems, short-term variability, non-regulated disinfection by-products, trihalomethanes,
haloacetic acids.
Résumé
Les sous-produits de la désinfection (SPD) constituent une grande famille de composés. Les trihalométhanes
(THM) et les acides haloacétiques (AHA) sont déjà réglementés dans plusieurs pays, mais la plupart des SPD
ne le sont pas. Le suivi des SPD peut être délicat, surtout pour les petits réseaux, car plusieurs facteurs
influencent leur formation et leur spéciation. Les variations à court terme des SPD peuvent être importants et
particulièrement difficiles à gérer pour les petits réseaux, car ils exigent des opérateurs qualifiés pour gérer les
conditions d’opération. Selon notre connaissance, pour la première fois, une étude se penche sur la variabilité
à court terme de l’occurrence des SPD réglementés et non-réglementés dans les petits réseaux. Dans le cadre
de notre étude, un programme d'échantillonnage intensif a été réalisé dans six petits réseaux au Canada. Les
petits réseaux dans les provinces de Québec et Terre-Neuve-et-Labrador ont été échantillonnés tous les jours
à l'usine de traitement de l'eau et à six endroits différents le long du système de distribution. Cinq familles de
SPD ont été étudiées: les THM, les AHA, les haloacétonitriles, les halonitrométhanes et les halocétones. Les
résultats ont montré qu'il y a eu des variations considérables dans les niveaux de SPD de semaine en semaine
durant le mois d'étude et même de jour en jour dans la semaine. Les variations ont été fortement influencées
par les différences des caractéristiques (qualité de l'eau, infrastructure, exploitation, etc.) d'un système à l’autre.
De même, le grand nombre de sites d'échantillonnage dans les systèmes de distribution a permis d’observer en
détail les variations spatiales des SPD le long du système de distribution. Les observations ont révélé la
dégradation et la transformation de certains SPD non-réglementés jamais étudiées auparavant dans les petits
réseaux. Enfin, cette étude a révélé que la variabilité temporelle des SPD est également influencée par la
localisation spatiale le long du système. Ainsi, les variations spatiales et temporelles des SPD à court terme
peuvent rendre difficile le choix des périodes et de lieux d’échantillonnage à des fins de suivi des SPD dans les
petits réseaux.
Mots-clés : Petits réseaux, variabilité à court-terme, sous-produits de la désinfection non-réglementés,
trihalométhanes, acides haloacétiques.
63
3.1. Introduction
Disinfection by-products (DBPs) are potentially toxic substances generated by the reaction between a
disinfectant, usually chlorine, and naturally organic matter (NOM) present in water (Rook, 1974).
Trihalomethanes (THMs) and haloacetic acids (HAAs) are the most prevalent DBPs in drinking water. Their
formation is relatively well understood and their levels are regulated in various countries (in particular THMs)
(Singer, 2002; Richardson, 2011). But the DBP family is quite large: more than 600 DBPs have been detected
and most DBPs such as haloacetonitriles (HANs), haloketones (HKs) and halonitromethanes (HNMs) are not
regulated. Recently, there has been increased interest in investigating the presence of non-regulated DBPs, due
to the fact that some DBPs such as nitrogenated DBPs (HANs and HNMs) and brominated DBPs (Krasner et
al., 2006; Bove et al., 2007; Muellner et al., 2007; Richardson, 2011) may have higher toxicological effects than
THMs and HAAs, especially concerning their cytotoxicity and genotoxicity.
To reduce the formation of DBPs in drinking water systems, two strategies may be applied: the reduction in the
amount of NOM present in the water before disinfection and the appropriate management of operational
conditions related to disinfection (Milot et al., 2000). However, because of financial constraints, small water
systems (i.e., serving 5,000 or fewer people) may experience some difficulty implementing adequate treatment
technologies to remove DBP precursors and hiring qualified operators to manage operational conditions
(Coulibaly & Rodriguez, 2004; Edwards et al., 2012). Thus, small systems using surface waters may be more
vulnerable to DBPs. In fact, a previous study showed that average measured concentrations of DBPs in small
Canadian systems using surface waters were much higher than those reported in the literature for medium and
large systems (Guilherme & Rodriguez, 2014).
Understanding the variability of DBPs is challenging because various factors influence their speciation and
evolution in water distribution systems. Factors include: the nature and the amounts of NOM and ions present
in raw water (Uyak & Toroz, 2007; Karanfil et al., 2008), disinfectant concentration and contact time (Singer,
1994; Rodriguez & Sérodes, 2001; Liang & Singer, 2003; Rodriguez et al., 2004; Speight & Singer, 2005; Bull
et al., 2009) and type of disinfectant (Adams et al., 2005; Crittenden et al., 2005; Bull et al., 2009; Bougeard et
al., 2010). Moreover, pH and temperature influence DBP speciation and formation kinetics (Croue and Reckhow,
1989; Yang et al., 2007; Fang et al., 2010). Consequently, important temporal and spatial variations are
commonly observed in the measured concentrations of DBPs in drinking water. Unfortunately, there is currently
little information on the spatio-temporal variability of DBPs in the water of small communities. With the exception
of our recent study consisting of a one-year monthly sampling campaign in 25 small systems (Guilherme and
Rodriguez, 2014), only a few studies are available and they mainly investigate the occurrence of regulated DBPs
(Charrois et al., 2004; Tung & Xie, 2009). Most studies have focused on the temporal and spatial variations of
DBP occurrence in large systems and placed the emphasis on regulated DBPs such as THMs and HAAs (Lebel
64
et al., 1997; Rodriguez & Sérodes, 2001; Summerhayes et al., 2011; Mercier-Shanks et al., 2013). Moreover,
only long-term variability (annual, monthly or seasonal) is evaluated in these studies. Furthermore, in most
studies only a few locations are sampled along the distribution system. Currently, there is a lack of information
on the variability patterns of regulated and non-regulated DBPs in the short term, on a daily basis, and within
small distribution systems. Such information could help utility managers improve their routine operations in order
to reduce the levels of DBPs (through treatment adjustments) and identify representative sampling moments
and locations for monitoring purposes (i.e., regulatory compliance).
Thus, the aim of this study was to investigate the short-term temporal and spatial variability of DBP occurrence
in small systems. The study focused on both regulated and non-regulated DBPs. It was based on sampling
campaigns carried out within a short time period during the summer in various locations along the DS. In this
research and according to our knowledge, this is the first time that the short-term variability of regulated and
non-regulated DBP occurrence in small municipal systems has been studied.
3.2. Methodology
3.2.1. Case studies
An intensive sampling program was carried out during the summer of 2012 in six small water systems (SWS) of
two provinces of Canada, Newfoundland and Labrador (NL) and Quebec (QC). The study focused on the
summer, since various studies have shown that during this season, DBP levels are generally at their highest
and environmental conditions (temperature, microbiological activity) may favour increased decomposition and
biodegradation of some DBP species (Lebel et al., 1997; Nikolaou et al., 2000; Nikolaou et al., 2001; Rodriguez
et al., 2007; Guilherme & Rodriguez, 2014).
All the systems were supplied by surface water sources and used chlorine as their main disinfectant (for primary
and secondary disinfection). The three systems in NL (NL1, NL2 and NL3) served a population varying from 320
to 1,020 inhabitants. In QC, the three systems (QC1, QC2 and QC3) served a population varying from 1,500 to
3,800 inhabitants. All the systems in NL and one in Quebec (QC3) did not present any prior treatment to
chlorination, whereas in QC, QC1 and QC2 used conventional treatment processes prior to disinfection
(coagulation, flocculation, sedimentation and filtration). In fact, in QC, there are regulations that mandate water
utilities supplied by surface waters to remove turbidity and NOM, mainly through filtration, before the water is
subjected to chlorination. Also, QC2 was the only SWS characterized by a rechlorination process in its
distribution system (located just before DS2). Table 3.1 presents information on the populations served and
treatments used in the six SWS studied.
65
Table 3.1: General characteristics of SWS under study
SWS Type of treatment Population
NL1 Chlorination 321
NL2 Chlorination 1,020
NL3 Chlorination 450
QC1 Activated Carbon, Coag.-Floc.-Sed.-Filt.*,
UV, Chlorination 3,220
QC2 Coag.-Floc.-Sed.-Filt.*, Chlorination,
Rechlorination in the DS between DS1 and DS2 1,528
QC3 Chlorination 3,826
*Coag.-Floc.-Sed.-Filt.: Coagulation, Flocculation, Sedimentation, Filtration
3.2.2. Sampling and analysis
In each system, water was sampled at the source uptake (RW) and within the water treatment plant (WTP) just
after filtration and before disinfection. Various points were selected along the DS in order to collect samples at
different residence times of the water (Table 3.2). Water was sampled at six locations along the DS from the
beginning (DS1) to the end of the DS (DS6). Because NL systems had no treatment prior to chlorination, RW
and WTP were represented by the same point. Sampling campaigns were conducted daily (from Monday to
Friday) during one month in July 2012 for QC and August 2012 for NL. On Mondays, Wednesdays and Fridays,
each system was sampled in RW, WTP and all along the DS (from DS1 to DS6) in order to study the spatial
evolution of water quality. But on Tuesdays and Thursdays, water was sampled only in RW, WTP and DS3 to
maintain a follow-up of the temporal variability of water quality during the month. All samplings were conducted
at the same time every day.
Table 3.2: Parameters measured during the sampling campaign
Physico-chemical parameters DBPs
UV-254 DOC Free Chlorine HAAs, THMs, HANs, HNM, HKs
Raw water (RW)
X X - -
WTP* X X 0 0
DS1, DS2 - - X X
DS3 X X X X
DS4, DS5, DS6
- - X X
X Measured - Non measured 0 Measured if water is treated before chlorination * Only for QC
66
Table 3.2 summarizes the parameters measured at each sampling point. In RW, WTP and DS3, indicators for
precursors of DBPs were measured using ultraviolet absorbance at 254 nm (UV-254) and dissolved organic
carbon (DOC). In treated water after chlorination, residual disinfectant levels (free chlorine) and DBP
concentrations were measured. Five families of DBPs were considered: THMs, HAAs and three families of non-
regulated DBPs. Four THMs (chloroform (TCM), bromodichloromethane (BDCM), dibromochloromethane
(DBCM) and tribromomethane (TBM)), five HAAs (monochloroacetic acid (MCAA), monobromoacetic acid
(MBAA), dichloroacetic acid (DCAA), trichloroacetic acid (TCAA) and dibromoacetic acid (DBAA)), four HANs
(dichloroacetonitrile (DCAN), trichloroacetonitrile (TCAN), bromochloroacetonitrile (BCAN) and
dibromoacetonitrile (DBAN)), one HNM (Chloropicrin (CPK)) and two HKs (1,1-dichloropropanone (DCP) and
1,1,1-trichloropropanone (TCP)) were analyzed during the study. Details about analytical methods used are
provided elsewhere (Mercier-Shanks et al., 2013). During this study, more than 300 samples were collected,
representing almost 11,000 pieces of information for numerous parameters. This sampling strategy based on
numerous points within the distribution system and high-frequency sampling served to evaluate the short-term
variability of the mentioned regulated and non-regulated DBPs.
3.3. Results and discussion
3.3.1. Short-term temporal variability of DBP occurrence
The results from the sampling campaigns demonstrated that there were considerable variations of DBP levels
from week to week and from day to day within the week. Table 3.3 summarizes the variations of weekly average
DBP and free chlorine levels in a middle location within the distribution system (DS3). Levels were compared at
the same sampling location in order to compare DBP temporal variability without taking into account their spatial
variability. It is important to highlight the considerable variability of the DBP levels from week to week. For
example, weekly average levels of THMs fluctuated (based on the ratio (Max-Min)/Min) from 130% in NL3 to
34% in QC2, and HAA levels from 260% in NL2 to 21% in QC2 within the month of the study. Similar observations
could be made for non-regulated DBPs. In fact, weekly average concentrations of HAN fluctuated from 230% in
NL2 to 20% in QC1; CPK levels from 550% in NL2 to 24% in NL3; and HK levels from 200% in NL2 to 27% in
QC3 within the month (Table 3.3). Such considerable temporal variations of DBPs appear surprisingly high for
a short-term period like a month. Also, high differences in DBP weekly fluctuations were observed between
SWS, indicating an important site specificity of DBP variations. In fact, a previous study of small systems
(Guilherme & Rodriguez, 2014) revealed that local characteristics of water quality, climate, treatment, operations
and distribution system characteristics make the occurrence of regulated and non-regulated DBPs in SWS site-
specific, especially in small systems where human and technical resources are limited. In addition, previous
studies conducted in small systems in Alberta and the United States have shown great diversity between
systems in regulated DBP levels measured (Charrois et al., 2004; Tung & Xie, 2009).
67
Appendix 7 presents the daily variability of DBPs and free residual chlorine for all SWS taken together. Although
temporal differences in average DBP and residual chlorine levels were observable from one day to the next in
the same week, these trends were very much influenced by the differences in characteristics (water quality,
infrastructure, operations, etc.) that exist from system to system (differences represented in Appendix 7 by the
box distribution with percentiles). Patterns of temporal variations are presented separately hereafter for individual
selected SWS.
Temporal day to day variations of DBPs were investigated for two selected SWS in Figures 3.1 and 3.2. NL2
and QC3 were selected for comparison purposes since they represent systems with no treatment before
disinfection and considerable differences in free residual chlorine patterns (levels and variability). Residual
chlorine is an important parameter to consider, because its variability is affected simultaneously by the quality
of water (e.g., NOM), water temperature, chlorine dose and the residence time of water within the distribution
system. The two systems selected presented average free chlorine levels much higher than 0.3 mg/L (QC3) and
much lower than 0.3 mg/L (NL2). The threshold of 0.3 mg/L was chosen according to the USEPA and Canadian
guidelines that recommend maintaining a free residual chlorine higher than 0.2 mg/L (USEPA, 2007; Health
Canada, 2012), and based on the study by Tung et al. (2009) carried out in SWS that concluded that free residual
chlorine levels lower than 0.3 mg/L, may favour the biodegradation or decomposition of HAAs.
68
Table 3.3: Weekly average levels of DBPs and free chlorine in DS3 (with coefficients of variation-CV) in the SWS under study
Week Free Cl CV THMs CV HAAs CV HANs CV CPK CV HKs CV
NL1
1 1.09 33% 92.1 7.9% 86.5 12% 3.14 27% 0.85 19% 8.92 13%
2 0.39 95% 48.0 82% 58.9 66% 1.56 79% 0.50 80% 4.80 79%
3 0.04 53% 5.79 0.0% 4.09 66% 0.07 123% 0.04 183% 0.33 194%
4 0.03 85% 5.79 0.0% 2.50 0.0% 0.02 0.0% 0.01 0.0% 0.01 0.0%
NL2
1 0.05 40% 166 33% 61.0 62% 4.35 46% 0.26 52% 15.2 32%
2 0.03 26% 134 6.9% 17.3 77% 1.31 29% 0.04 52% 5.08 50%
3 0.04 52% 195 18% 29.2 93% 2.08 16% 0.10 56% 8.46 17%
4 0.04 61% 222 15% 57.6 19% 3.05 29% 0.21 33% 12.0 18%
NL3
1 1.49 9.8% 211 12% 176 30% 8.37 9.7% 1.07 9.9% 13.0 7.9%
2 0.64 44% 252 8.4% 240 6.4% 6.88 8.8% 0.86 6.5% 8.85 50%
3 0.43 91% NA NA 263 21% 7.87 17% 0.97 13% 14.8 12%
4 0.14 86% 481 8.8% 337 11% 7.62 22% 0.86 15% 16.0 15%
QC1
1 0.10 94% 64.2 30% 57.4 34% 3.94 13% 1.57 26% 6.88 24%
2 0.13 92% 100 22% 40.3 23% 3.82 13% 1.54 4.1% 4.66 36%
3 0.44 19% 72.9 27% 36.9 10% 3.29 11% 1.30 10% 2.84 44%
4 0.35 59% 61.7 11% 29.8 13% 3.55 21% 1.10 24% 4.82 76%
QC2
1 0.02 111% 69.9 29% 30.8 32% 3.08 27% 0.22 51% 4.66 27%
2 0.01 40% 74.7 27% 29.9 32% 3.26 29% 0.37 73% 5.02 16%
3 0.04 104% 93.4 20% 36.3 27% 3.87 18% 0.28 43% 5.96 19%
4 0.02 77% 86.1 25% 33.2 30% 3.88 21% 0.32 47% 5.45 18%
QC3
1 0.84 92% 117 20% 143 9.3% 4.03 28% 0.70 11% 11.2 14%
2 1.71 39% 128 7.1% 179 53% 3.60 28% 0.87 31% 8.85 57%
3 1.14 49% 76.6 46% 123 11% 3.20 13% 0.58 11% 9.83 7.1%
4 0.96 70% 122 79% 123 33% 3.53 20% 0.75 49% 8.86 27%
CV: coefficient of variation
NA: Data non available
Note: DBP and free chlorine levels were especially low in NL1 from the Wednesday of the second week due to a break in
the chlorine pomp in the WTP. After that moment, water was not chlorinated.
69
a) b)
c) d)
e) f)
Note: error bars represent standard deviations
Figure 3.1: Variations from day to day within the week of levels of a) Free chlorine, b) THMs, c) HAAs, d) HANs, e)
CPK and f) HKs in NL2 in DS3 (number of observations per day: 4)
0
0,02
0,04
0,06
0,08
0,1
Fre
e c
hlo
rine level in
NL2
(mg/L
)
0
50
100
150
200
250
300
TH
M level in
NL2 (
µg/L
)
0
20
40
60
80
100
120
HA
A level in
NL2 (
µg/L
)
0
1
2
3
4
5
6
HA
N level in
NL2 (
µg/L
)
0
0,1
0,2
0,3
0,4
CP
K level in
NL2 (
µg/L
)
0
3
6
9
12
15
18
HK
level in
NL2 (
µg
/L)
70
a) b)
c) d)
e) f)
Note: error bars represent standard deviations
Figure 3.2: Variations from day to day within the week of levels of a) Free chlorine, b) THMs, c) HAAs, d) HANs, e)
CPK and f) HKs in QC3 in DS3 (number of observations per day: 4)
Figure 3.1 shows that, generally speaking in NL2, the free residual chlorine level increased from Monday to
Friday, even though the free chlorine level on Tuesday was overestimated due to an extreme value (median free
chlorine level on Tuesday was 0.028 mg/L). By taking into account the fact that free residual chlorine level was
0
0,5
1
1,5
2
2,5
3F
ree c
hlo
rine level in
QC
3
(mg/L
)
0
50
100
150
200
250
300
TH
M level in
QC
3 (
µg/L
)
0
50
100
150
200
250
300
HA
A level in
QC
3 (
µg/L
)
0,0
1,0
2,0
3,0
4,0
5,0
6,0
HA
N level in
QC
3 (
µg/L
)
0
0,2
0,4
0,6
0,8
1
1,2
1,4
CP
K level in
QC
3 (
µg
/L)
0
3
6
9
12
15
HK
level in
QC
3 (
µg/L
)
71
overestimated on Tuesday, DBPs and free residual chlorine basically seemed to follow the same evolution from
Monday to Friday. It was also noted that THM daily fluctuation was low compared to HAA, HAN, CPK and HK
daily fluctuations. In fact HAAs, HANs, CPK and HKs fluctuated more than free residual chlorine. In NL2, free
residual chlorine level fluctuated from day to day from 13% to 86% (range of coefficient of variations (CV)
calculated each day: based on four measurements made on separate weekdays during the month), THMs from
22% to 42%, HAAs from 37% to 96%, HANs from 34% to 70%, CPK from 51% to 89% and HKs from 32% to
57%. The high fluctuations of HAAs, HANs, CPK and HKs in comparison with residual free chlorine indicate that
the fate of these compounds might be influenced by conditions other than the residual disinfectant level at their
location. In fact, some of these compounds are known to biodegrade in the presence of biofilm on pipe surface,
favoured by the residence time of water, especially in summer, with high temperature and low residual
disinfectant levels (Bayless & Andrews, 2008; Pluchon et al., 2013; Berthiaume et al., 2014). Indeed, the
difference between free residual chlorine levels in DS1 and DS3 fluctuated every day (CV values from 16 to
32%, data not shown). Variations in difference in free residual chlorine levels can be a partial indicator of
variations of the residence time of water, even though they can also be affected by the NOM, temperature and
the chlorine dose. Thus, the residence time of water fluctuating each day would influence daily variations in DBP
levels. Conversely, system QC3 presented high free residual chlorine levels within the distribution system (higher
than 0.3 mg/L), and lower HAA, HAN, CPK and HK fluctuations (Figure 3.2). In QC3, free residual chlorine levels
fluctuated daily from 16% to 100%, THMs from 12% to 70%, HAAs from 10% to 53%, HANs from 5% to 34%,
CPK from 15% to 36% and HKs from 8% to 56%. Contrary to NL2, in QC3, THMs and free residual chlorine
fluctuated the most (Figure 3.2). Also, differences between free residual chlorine levels in DS1 and DS3
fluctuated enormously every day (CV values from 58% to 66%, data not shown), indicating that the residence
time of water may fluctuate greatly each day. Thus, although the residence time of water fluctuated a lot in QC3
(more than in NL2), HANs, CPK and HKs did not fluctuate more than in NL2, possibly indicating that the
residence time of water has less influence if free residual chlorine levels are higher than 0.3 mg/L. These
compounds might not be degraded due to the relatively high free residual chlorine level. Also, THMs, HAAs,
HANs and CPK basically seemed to follow free residual chlorine variations in QC3 (Figure 3.2). The higher the
level of free chlorine, the higher the levels of these DBPs. However, HK fluctuations during the week did not
follow the same pattern as free chlorine in QC3 (Figure 3.2). These differences in fluctuations might be explained
by the fact that DCP can be oxidized into TCP in the presence of free chlorine (Bougeard et al., 2010; Mercier-
Shanks et al., 2013) inducing no variation in the total HK levels.
The daily variability of DBPs could also be affected by the variability of raw water quality (Figure 3.3). Results
showed that, generally speaking, DBP variations along the month followed the variations of the indicators of
NOM in raw water (DOC and UV-254), especially in NL2. Figures 3.3a and 3.3b show that NOM indicators levels
decreased and then increased in NL2 during the month in raw water, inducing a comparable evolution pattern
72
for DBP levels in DS3 during the same period. At the same time in QC3, NOM indicators levels in raw water
remained relatively constant along the month, like DBP levels in the DS (Figures 3.3c and 3.3d). Thus, variations
in raw water quality could contribute to DBP level variations during the month in systems with no treatment prior
to disinfection. However, since no information on the chlorine dose was available during the period of the study,
the impact of operational practices on DBP variations could not be evaluated. In addition to the factors mentioned
above, the daily temporal variations of DBPs and free residual chlorine may be influenced by the residence time
of water. In fact, hydraulic conditions of the distribution systems and water demand can change from day to day,
affecting residence time in the short term and, thus, water quality, including DBPs and residual chlorine (Mouly
et al., 2010). However, information about the residence time of water was not available during the study.
a) b)
c) d)
Figure 3.3: Daily evolution of raw water characteristics and DBP occurrence in DS3 a) and b) in NL2 and c) and d) in
QC3
The results reflect that regulated and non-regulated DBPs present high and complex temporal variations in the
short term (daily and weekly). This information is relevant in a context of DBP monitoring considering the
regulatory framework, in particular for regulated DBPs. Because their levels can differ significantly from week to
6
7
8
9
0
50
100
150
200
250
300
30
-07
31
-07
01
-08
02
-08
03
-08
06
-08
07
-08
08
-08
09
-08
10
-08
13
-08
14
-08
15
-08
16
-08
17
-08
20
-08
21
-08
22
-08
23
-08
24
-08
DO
C le
vel in
NL2 (
mg/L
)
DB
P le
vel in
NL2 (
µg/L
)
THMs HAAs DOC
0,2
0,22
0,24
0,26
0,28
0,3
0
4
8
12
16
20
30
-07
31
-07
01
-08
02
-08
03
-08
06
-08
07
-08
08
-08
09
-08
10
-08
13
-08
14
-08
15
-08
16
-08
17
-08
20
-08
21
-08
22
-08
23
-08
24
-08
UV
-254 in
NL2 (
cm
-1)
DB
P le
vel in
NL2 (
µg/L
)
HANs CPK
HKs UV_254
3
3,5
4
4,5
5
5,5
6
0
40
80
120
160
200
240
280
320
360
26
-06
27
-06
28
-06
29
-06
03
-07
04
-07
05
-07
06
-07
09
-07
10
-07
11
-07
12
-07
13
-07
16
-07
17
-07
18
-07
19
-07
20
-07
23
-07
DO
C le
ve
l in
QC
3 (
mg
/L)
DB
P le
ve
l in
QC
3 (
µg/L
)
THMs HAAs DOC
0,1
0,12
0,14
0,16
0,18
0,2
0
2
4
6
8
10
12
14
26
-06
27
-06
28
-06
29
-06
03
-07
04
-07
05
-07
06
-07
09
-07
10
-07
11
-07
12
-07
13
-07
16
-07
17
-07
18
-07
19
-07
20
-07
23
-07
UV
-25
4 in
QC
3 (
cm
-1)
DB
P le
ve
l in
QC
3 (
µg/L
)
HANs CPK
HKs UV_254
73
week and even day to day, seasonal DBP values used as representative levels for regulatory purposes will also
be different according to the sampling period when the samples are collected. Fortunately, since their variations
can be partially linked to raw water characteristics and free chlorine variations, DBPs fluctuations can be
evaluated by the fluctuations of these parameters.
3.3.2. Spatial variability of DBP occurrence
The different families of investigated DBPs did not behave similarly within the distribution systems. The average
maximum level location differed according to the DBP family (Figure 3.4). As might be expected, free residual
chlorine level was highest at the beginning of the DS. Then it decreased, contributing to the formation of further
DBPs. Concerning the DBP levels, THM levels were highest at the end of the DS (from DS4) and HAA, HAN,
CPK and HK levels were highest in the middle of the DS. This type of evolution has been observed before for
these compounds (Lebel et al., 1997; Rodriguez et al., 2004; Mercier-Shanks et al., 2013). However, these
global portraits vary depending on the SWS, as shown in Table 3.4. Locations with the highest concentrations
differed highly between systems, indicating that the maximum level locations of DBPs were site-specific and
influenced by the individual characteristics of each SWS. The diversity of DBP variation patterns is illustrated in
Appendix 8. Trends in this figures validate some observations made in larger systems (Lebel et al., 1997;
Rodriguez et al., 2004; Mercier-Shanks et al., 2013) and in our previous study of 25 SWS, including these six
SWS in NL and in QC and based on only few sampling locations along the DS (Guilherme & Rodriguez, 2014).
For example, Appendix 8 shows that on the whole, THM levels increased regularly along the DS and that DCAA
and TCAA did not present the same behaviour. Generally speaking, DCAA levels increased at the beginning of
the DS and then decreased (particularly, in systems where the free chlorine level is under than 0.3 mg/L at the
end of the DS) or at least stabilized at the end of the DS, and TCAA levels increased along the DS. In fact, DCAA
is known to biodegrade in the presence of biofilm on pipe surface (Rodriguez et al., 2007; Bayless & Andrews,
2008; Berthiaume et al., 2014). Also, CPK levels decreased, or at least stabilized, at the end of the DS indicating
a possible decomposition (Lebel et al., 1997; Zhang et al., 2013; Guilherme & Rodriguez, 2014).
Figure 3.4: Location of the maximum levels of DBPs and free residual chlorine along the DS for all SWS under study
/ / / / THMs Free chlorine
HAAs
HANs, CPK, HKs
End of
the DS WTP DS1 DS4 DS3
74
Table 3.4: Specific locations for the maximum levels of DBPs and free residual chlorine in the SWS under study
NL1 NL2 NL3 QC1 QC2 QC3
Free Chlorine
DS2 DS1 DS1 DS1 DS2 DS1
THMs DS5 DS4 DS4 DS6 DS2 DS6
HAAs DS4 DS3 DS6 DS6 DS2 DS6
DCAA DS4 DS3 DS6 DS5 DS2 DS3
TCAA DS5 DS3 DS6 DS6 DS2 DS6
HANs DS5 DS3 DS5 DS6 DS2 DS6
DCAN DS5 DS3 DS5 DS6 DS2 DS6
BCAN DS2 DS3 DS2 DS6 DS2 DS6
CPK DS4 DS3 DS4 DS6 DS2 DS6
HKs DS4 DS3 DS4 DS6 DS2 DS6
DCP DS4 DS1 DS1 DS2 DS1-DS6 DS6
TCP DS5 DS3 DS5 DS6 DS2 DS6
Note: the maximum number of observations is 20 in DS3, and 12 in the other locations.
Appendix 8 also highlights that HAN variation patterns differed from system to system. In some systems (such
as QC1 and QC3), HAN levels increased along the DS. But in NL2 and QC2, their levels decreased at the end
of the DS, indicating a possible decomposition of these compounds, especially DCAN and BCAN, as they are
more easily degradable than other DBPs (Nikolaou et al., 2000; Zhang et al., 2013). Finally, concerning HKs,
the profiles of DCP and TCP can differ depending on the system, as may be noted in Appendix 8. In QC3, DCP
and TCP followed the same evolution, with their levels regularly increasing along the DS. Possibly due to the
high level of residual free chlorine present along the DS, HKs would form all along the DS. Conversely, in NL,
DCP and TCP levels decreased at the end of the DS when the free chlorine level was below 0.3 mg/L, revealing
the decomposition of these compounds (Nikolaou et al., 2001).
However in NL2, the DBP variability portraits differed from those described above due to the specific
characteristics of the system. Thus, in NL2, DBP variations indicated a long residence time of water in this
system (Appendix 8). In fact, DCAA and TCAA level decreases were observable at the end of the DS. Since
TCAA is much more resistant to biodegradation than DCAA, a longer residence time of water is necessary to
observe TCAA biodegradation (Wang et al., 2009). This example provides a good illustration of the site specificity
of DBP variations and the influence of the characteristics of the SWS on DBP occurrence.
75
3.3.3. Impact of the location on the short-term variability
of DBPs
In order to study the influence of the spatial location on the temporal variability of DBPs, a comparison of the
DBP temporal evolution patterns within the month between the beginning of the DS (DS1) and the end of the
DS (DS6) in NL2 and QC3 was performed. Results show that the temporal evolution pattern of DBPs is
influenced by spatial location (Figures 3.5 and 3.6). In system QC3 (Figure 3.5), the temporal fluctuations of
DBPs during the month were globally higher at the end of the DS. In fact, temporal fluctuations during the month
of THMs, HAAs, HANs, CPK and HKs were respectively 340%, 126%, 110%, 82%, 63% in DS6 compared to
68%, 106%, 113%, 56% and 38% in DS1.
a) b)
c) d) Figure 3.5: Temporal variability of the DBP levels in two locations along the DS of QC3: a) and b) in DS1; c) and d) in
DS6 (SWS presenting an average level of free chlorine higher than 0.3 mg/L in DS6)
0
0,5
1
1,5
2
2,5
3
3,5
4
4,5
0
50
100
150
200
27
-06
29
-06
01
-07
03
-07
05
-07
07
-07
09
-07
11
-07
13
-07
15
-07
17
-07
19
-07
21
-07
23
-07 Fre
e c
hlo
rin
e le
ve
l in
DS
1 (
mg
/L)
DB
P le
ve
l in
DS
1 (
µg/L
)
THMs HAAs Free chlorine
0
4
8
12
27
-06
29
-06
01
-07
03
-07
05
-07
07
-07
09
-07
11
-07
13
-07
15
-07
17
-07
19
-07
21
-07
23
-07
DB
P le
ve
l in
DS
1 (
µg/L
)
HANs CPK HKs
0
0,5
1
1,5
2
2,5
3
3,5
4
4,5
5
0
50
100
150
200
27
-06
29
-06
01
-07
03
-07
05
-07
07
-07
09
-07
11
-07
13
-07
15
-07
17
-07
19
-07
21
-07
23
-07
Fre
e c
hlo
rin
e le
ve
l in
DS
1 (
mg
/L)
DB
P le
ve
l in
DS
6 (
µg/L
)
THMs HAAs Free chlorine
0
4
8
12
27
-06
29
-06
01
-07
03
-07
05
-07
07
-07
09
-07
11
-07
13
-07
15
-07
17
-07
19
-07
21
-07
23
-07
DB
P le
ve
l in
DS
6 (
µg/L
)
HANs CPK HKs
76
a) b)
c) d) Figure 3.6: Temporal variability of the DBP levels in two locations along the DS of NL2: a) and b) in DS1; c) and d) in
DS6 (SWS presenting an average level of free chlorine lower than 0.3 mg/L in DS6)
In NL2 (Figure 3.6), a system presenting a free chlorine level lower than 0.3 mg/L in DS6, the average levels of
some DBPs during the month were higher in DS6 than in DS1, and for others, lower in DS6 than in DS1. In fact,
average levels of THMs and HANs were respectively 123 and 2.6 µg/L in DS6 compared to 62 and 1.4 µg/L in
DS1. Conversely, HAA, CPK and HK levels were respectively 29, 0.12 and 9.7 µg/L in DS6 compared to 82,
0.26, 12 µg/L in DS1. Indeed, in NL2, the residual free chlorine level at the end of the DS was low, so some
DBPs might be degraded or decomposed. In addition, the temporal fluctuations of DBPs differed depending on
the spatial location and according to the DBP family. Indeed, temporal fluctuations of THMs and HANs within
the month were respectively 61% and 75% in DS6 compared to 148% and 100% in DS1. On the other hand,
temporal fluctuations of the levels of HAAs, CPK and HKs were respectively 287%, 360% and 71% in DS6
compared to 38%, 116% and 63% in DS1. Thus, the highest temporal fluctuations of the DBPs studied in the
system and presenting a free chlorine level lower than 0.3 mg/L in DS6 occurred at the spatial locations where
DBP levels were the lowest, compared to the system presenting a free chlorine level higher than 0.3 mg/L in
DS6 where the highest temporal fluctuations appeared at the spatial locations where DBP levels were the highest
(except for HANs, for which temporal fluctuations do not differ between DS1 and DS6). This information reflects
0
1
2
3
4
0
40
80
120
160
30
-07
01
-08
03
-08
05
-08
07
-08
09
-08
11
-08
13
-08
15
-08
17
-08
19
-08
21
-08
23
-08
Fre
e c
hlo
rin
e le
ve
l in
DS
1 (
mg
/L)
DB
P le
ve
l in
DS
1 (
µg
/L)
THMs HAAs Free Chlorine
0
0,1
0,2
0,3
0,4
0
4
8
12
16
30
-07
01
-08
03
-08
05
-08
07
-08
09
-08
11
-08
13
-08
15
-08
17
-08
19
-08
21
-08
23
-08
CP
K le
ve
l in
DS
1 (
µg
/L)
DB
P le
ve
l in
DS
1 (
µg
/L)
HANs HKs CPK
0
1
2
3
4
0
40
80
120
160
30
-07
01
-08
03
-08
05
-08
07
-08
09
-08
11
-08
13
-08
15
-08
17
-08
19
-08
21
-08
23
-08 Fre
e c
hlo
rin
e le
ve
l in
DS
6 (
mg
/L)
DB
P le
ve
l in
DS
6 (
µg/L
)
THMs HAAs Free Chlorine
0
0,1
0,2
0,3
0,4
0
4
8
12
16
30
-07
01
-08
03
-08
05
-08
07
-08
09
-08
11
-08
13
-08
15
-08
17
-08
19
-08
21
-08
23
-08
CP
K le
ve
l in
DS
6 (
µg
/L)
DB
P le
ve
l in
DS
6 (
µg/L
)
HANs HKs CPK
77
the strong influence of residual free chlorine levels on DBP spatio-temporal variations. These results also show
that the spatial and temporal variations of DBPs in the short term may make it difficult to select representative
locations and periods for monitoring purposes.
3.4. Conclusions
This research documents the short-term variability of regulated and non-regulated DBP occurrence in small
municipal systems. The knowledge obtained through this study will be useful to support small systems in DBP
monitoring. With regard to temporal variations of DBPs, this study noted that regulated and non-regulated DBPs
presented a high variability pattern within a month and even a week. In a context of DBP monitoring for regulatory
compliance (in particular for THMs and HAAs), this information indicates a possible high degree of variation of
DBP levels measured in each quarter (trimester) depending of the sampling period chosen. On the issue of
spatial variability, the results for SWS mostly agree with patterns observed in large systems in other studies. Our
study also served to highlight some site specificities of regulated and non-regulated DBP occurrence in small
systems. In various systems, the biodegradation and decomposition of some DBPs were observed thanks the
high number of sampling locations. However, this study has some limitations. Observations made were specific
to the summer season when DBP biodegradation and decomposition are more probable than in the other
seasons (trends in the other seasons could be different). Also, the impact of site-specific operations, including
chlorine adjustments and hydraulic management (that influence residence time of water within the distribution
system), were not taken into account (as information was not uniformly available in all systems).
DBP short-term variations are important in SWS. In the future, it would be appropriate to study whether DBP
compliance monitoring strategies used by SWS are compatible with DBP variations during a short-term period
(weekly, daily and even hourly). Indeed, few studies have focused on the representativeness of the compliance
monitoring strategies according to observed temporal and spatial variations of DBPs.
3.5. References
Adams, C., Timmons, T., Seitz, T., Lane, J., Levotch, S., 2005. Trihalomethane and haloacetic acid disinfection
by-products in full-scale drinking water systems. Journal of Environmental Engeneering 131(4), 526-534.
Bayless, W., Andrews, R.C., 2008. Biodegradation of six haloacetic acids in drinking water. Journal of Water
and Health 6 (1), 15-22.
Berthiaume, C., Gilbert, Y., Fournier-Larente, J., Pluchon, C., Filion, G., Jubinville, E., et al, 2014. Identification
of dichloroacetic acid degrading Cupriavidus bacteria in a drinking water distribution network model. Journal of
Applied Microbiology 116 (1): 208-221.
78
Bougeard, C.M., Goslan, E.H., Jefferson, B., Parsons, S.A., 2010. Comparison of the disinfection by-product
formation potential of treated waters exposed to chlorine and monochloramine. Water Research 44, 729-740.
Bove, G.E.Jr, Rogerson, P.A., Vena, J.E., 2007. Case control study of the geographic variability of exposure to
disinfectant byproducts and risk for rectal cancer. International Journal of Health Geographics 6 (18).
Bull, R.J., Rice, G., Teuschler, L.K., 2009. Determinants of Whether or Not Mixtures of Disinfection By-Products
are Similar. Journal of Toxicology and Environmental Health, Part A. 72, 437-460.
Charrois, J.W.A., Graham D., Hrudey S.E., Froese K.L. 2004. Disinfection by-products in small Alberta
community drinking water supplies. Journal of Toxicology and Environmental Health, Part A: Current Issues 67
(20-22), 1797-1803.
Coulibaly, H.D., Rodriguez, M.J., 2004. Development of performance indicators for small Quebec drinking water
utilities. Journal of Environmental Management 73, 243–255
Crittenden, J. C, MWH, Inc., 2005. Water treatment: Principles and design. Hoboken, N.J: John Wiley & Sons,
Inc.
Croue, J.-P., Reckhow, D. A., 1989. Destruction of chlorination byproducts with sulfite. Environmental Science
& Technology 23 (11).
Edwards, J.E., Henderson, S.B., Struck, S., Kosatsky, T., 2012. Characteristics of small residential and
commercial water systems that influence their likelihood of being on drinking water advisories in rural British
Columbia, Canada: a cross-sectional study using administrative data. Journal of Water and Health 10 (4), 629-
649.
Environmental Protection Agency, USEPA, 2007. Complying with the stage 2 disinfectant and disinfection
byproducts rule: small entity compliance guide. Office of Groundwater and Drinking Water, United States, EPA
815-R-07-014
Fang, J., Ma, J., Yang, X., Shang, C., 2010. Formation of carbonaceous and nitrogenous disinfection by-
products from the chlorination of Microcystis aeruginosa. Water Research 44, 1934-1940.
Guilherme, S., Rodriguez, M.J., 2014. Occurrence of regulated and non-regulated disinfection by-products in
small drinking water systems. Chemosphere 117, 425-432.
Health Canada, 2012. Guidelines for Canadian Drinking Water Quality
79
Karanfil, T., Krasner, S.W., Westerhoff, P., Xie, Y., 2008. Disinfection by-products in drinking water: occurrence,
formation, health effects and control. American Chemical Society, Washington, DC
Krasner, S.W., Weinberg, H.S., Richardson. S., Pastor, S.J., Chinn, R., Sclimenti, M.J., et al., 2006. Occurrence
of a new generation of disinfection byproducts. Environmental Science and Technology 40 (23), 7175-7185.
Lebel, G.L., Benoit, F.M., Williams, D.T., 1997. A one-year survey of halogenated disinfection by-products in the
distribution system of treatment plants using three different disinfection processes. Chemosphere 34 (11), 2301-
2317.
Liang, L., Singer, P.C., 2003. Factors influencing the formation and relative distribution of haloacetic acids and
trihalomethanes in drinking water. Environmental Science and Technology 37 (13), 2920-2928.
Mercier-Shanks, C., Sérodes, J.-B., Rodriguez, M.J., 2013. Spatio-temporal variability of non-regulated
disinfection by-products within a drinking water distribution system. Water Research 47, 3231-3243
Milot, J., Rodriguez, M.J., Sérodes, J.-B., 2000. Modeling the susceptibility of drinking water utilities to form high
concentrations of trihalomethanes. Journal of Environmental Management 60, 155-171.
Mouly, D., Joulin, E., Rosin, C., Beaudeau, P., Zeghnoun, A., Olszewski-Ortar, A., et al., 2010. Variations in
trihalomethane levels in three French water distribution systems and the development of a predictive model.
Water Research 44, 5168-5179.
Muellner, M.G., Wagner, E.D., McCalla, K., Richardson, S.D., Woo, Y.-T., Plewa, M.J.,2007. Haloacetonitriles
vs. regulated haloacetic acids: are nitrogen-containing DBPs more toxic?. Environmental Science and
Technology 41, 645-651.
Nikolaou, A.D., Golfinopoulos, S.K., Kostopoulou, M.N., Lekkas, T.D., 2000. Decomposition of
dihaloacetonitriles in water solutions and fortified drinking water samples. Chemosphere 41, 1149-1154.
Nikolaou, A.D., Lekkas, T.D., Kostopoulou, M.N., Golfinopoulos, S.K., 2001. Investigation of the behaviour of
haloketones in water samples. Chemosphere 44, 907-912.
Pluchon, C., Sérodes, J.-B., Berthiaume, C., Charrette, S.J., Gilbert, Y., Filion, G., et al., 2013. Haloacetic acid
degradation by a biofilm in a simulated drinking water distribution system. Water Science and Technology: Water
Supply 13 (2), 416-447.
80
Richardson, S.D., 2011. Disinfection by products: formation and occurrence in drinking water. The Encyclopedia
of Environmental Health 2, 110-136.
Rodriguez, M.J., Sérodes, J.-B., 2001. Spatial and temporal evolution of trihalomethanes in three water
distribution systems. Water Research 35, 1572-1586.
Rodriguez, M.J., Sérodes, J.-B., Levallois, P., 2004. Behavior of trihalomethanes and haloacetic acids in a
drinking water distribution system. Water Research 38, 4367-4382.
Rodriguez, M.J., Sérodes, J.-B., Levallois, P., Proulx, F., 2007. Chlorinated disinfection by-products in drinking
water according to source, treatment, season and distribution location. Journal of Environmental Engineering
and Science 6, 355-365.
Rook, J.J., 1974. Formation of haloforms during chlorination of natural waters. Water Treatment Examination
23, 234-243.
Singer, P.C., 1994. Control of disinfection by-products in drinking water. Journal of Environmental Engineering
120 (4), 727-744.
Singer, P.C., 2002. Occurrence of haloacetic acids in chlorinated drinking water. Water Science and Technology:
Water Supply 2 (5-6), 487-492.
Speight, V., Singer, P.C., 2005. Association between residual chlorine loss and HAA reduction in distribution
systems. Journal of American Water Works Association 97 (2), 82-91.
Summerhayes, R.J., Morgan, G.G., Lincoln, D., Edwards, H.P., Earnest, A., Rahman, M.B., et al., 2011. Spatio-
temporal variation in trihalomethanes in New South Wales. Water Research 45, 5715-5726.
Tung, H.H., Xie, Y.F., 2009. Association between haloacetic acid degradation and heterotrophic bacteria in water
distribution systems. Water Research 43, 971-978.
Uyak, V., Toroz, I., 2007. Investigation of bromide ion effects on disinfection by-products. Journal of Hazardous
Materials 149, 445-451.
Wang, G.S., Lai, S.P., Huang, Y.T., 2009. Biodegradation of haloacetic acids in water treatment processes.
Water Science and Technology: Water Supply 9 (5), 557-564.
Yang, X., Shang, C., Westerhoff, P., 2007. Factors affecting formation of haloacetonitriles, haloketones,
chloropicrin and cyanogen halides, during chloramination. Water Research 41, 1193–1200.
81
Zhang, X.-l., Yang, H., Wang X., Fu J., Xie, Y. F., 2013. Formation of disinfection by-products: Effect of
temperature and kinetic modeling. Chemosphere 90, 634-639.
83
Chapitre 4
Decision-making scheme for disinfection by-product monitoring intended for small drinking water systems
Dans les chapitres précédents, les variabilités des SPD à long et à court terme ont été étudiées et les résultats
ont révélé que ces variabilités étaient très importantes. Le but de ce chapitre est donc de développer un outil
d’aide à la décision pour le suivi de l’occurrence des SPD dans l’eau potable des petits réseaux, qui tient compte
des patrons de variabilité des SPD et des caractéristiques propres des systèmes. Pour ce faire, ce chapitre
évalue tout d’abord si le cadre réglementaire actuel prend en compte les variabilités à long et à court terme des
SPD aussi bien spatiales que temporelles. Par la suite, les meilleures périodes d’échantillonnage pour les SPD
sont identifiées en comparant différents scénarios de calculs de la concentration moyenne annuelle en SPD.
Puis, le chapitre se concentre sur l’identification des sites d’échantillonnage optimaux pour le suivi réglementaire
des SPD. Finalement, un outil simple et facilement utilisable par les petits réseaux est développé. Cet outil peut
aussi être utilisé pour identifier les périodes et les sites dans le réseau où l’exposition aux SPD non-réglementés
est maximale, ce qui pourrait être utile dans le cas d’une future réglementation.
Abstract
Trihalomethanes (THMs) and haloacetic acids (HAAs) are the most prevalent disinfection by-products (DBPs)
in drinking water and their occurrence is regulated in several countries. However, most DBPs are not regulated,
even though some may have greater toxicological relevance than regulated DBPs. Small water systems (SWS)
supplied by surface waters are vulnerable to high levels of DBPs due, primarily, to a lack of adequate treatment
processes that remove DBP precursors. Moreover, monitoring DBPs is difficult for SWS because it requires
good knowledge of the system and qualified operators to select locations and periods for sampling. This study
focuses on the development of a decision-making scheme for DBP monitoring in SWS, and the identification of
appropriate periods and locations for DBP sampling in particular. The study is based on information generated
in 25 small systems in Canada. The following DBPs were considered: four THMs, five HAAs, four
haloacetonitriles, two haloketones and one halonitromethane. The comparison of various sampling scenarios
that took account of the temporal variability of DBPs served to identify appropriate periods for DBP sampling.
Free residual chlorine demand was used as an indicator of the spatial variability of DBPs helped to identify
appropriate locations for sampling. The scheme developed provides SWS operators with a high cost-benefit tool
for decision making to select sampling periods and locations in order to comply with current regulations
concerning THMs and HAAs (and non-regulated DBPs for eventual future regulations).
84
Keywords: Small systems, disinfection by-products, drinking water, monitoring, regulation compliance.
Résumé
Les trihalométhanes (THM) et les acides haloacétiques (AHA) sont les sous-produits de la désinfection (SPD)
les plus répandus dans l'eau potable et leur concentration est réglementée dans plusieurs pays. Cependant, la
plupart des SPD ne sont pas réglementés, même si certains peuvent présenter une plus grande toxicité que les
SPD réglementés. Les petits réseaux d’eau potable approvisionnés en eaux de surface sont vulnérables à des
niveaux élevés en SPD en raison, principalement, d'une difficulté à mettre en place des procédés de traitement
de l’eau capables d’enlever les précurseurs des SPD avant la désinfection. En outre, le suivi des SPD est difficile
pour les petits réseaux, car il nécessite une bonne connaissance du système et des opérateurs qualifiés pour
sélectionner les périodes et les sites d'échantillonnage adéquats. Cette étude se concentre donc sur le
développement d'un outil d’aide à la décision pour la mise en place d’une stratégie de suivi des SPD dans les
petits réseaux, et l'identification des périodes et sites adéquats pour le suivi des SPD. L'étude se base sur des
données collectées dans 25 petits réseaux au Canada. Les SPD suivants ont été mesurés: quatre THM, cinq
AHA, quatre haloacétonitriles, deux halocétones et un halonitrométhane. La comparaison de différents
scénarios d'échantillonnage qui prennent en compte la variabilité temporelle des SPD a servi à identifier les
périodes appropriées pour le suivi des SPD. Le chlore libre résiduel a été utilisé comme un indicateur de la
variabilité spatiale des SPD et a permis d'identifier les sites appropriés pour le suivi des SPD. L’outil développé
ici présent un ratio coûts-avantages particulièrement intéressant pour les petits réseaux et permet aux
opérateurs de sélectionner facilement les périodes et sites d'échantillonnage pour le suivi réglementaire des
THM et des AHA (et des SPD non-réglementés en cas d’éventuelles futures réglementations).
Mots-clés: Petits réseaux, sous-produits de la désinfection, eau potable, suivi réglementaire, conformité
réglementaire.
4.1. Introduction
Disinfection by-products are formed in water when the disinfectant (usually chlorine) reacts with naturally
occurring organic matter (Rook, 1974). This family of compounds is large: more than 600 DBPs have been
detected (Richardson, 2011). However, potential risks for cancer and reproductive and developmental effects
have been associated with some DBPs (Cedergren et al., 2002; Richardson et al., 2007; Villanueva et al., 2007).
Thus, DBPs constitute an important concern for water systems, especially systems supplied by surface waters,
because they are generally rich in natural organic matter (Cedergren et al., 2002, Mouly et al., 2010).
Trihalomethanes (THMs) and haloacetic acids (HAAs) are the most prevalent and studied DBPs, and their levels
are regulated in various countries (especially THMs) (Singer, 2002; Richardson, 2011). Most DBPs are not
85
regulated. These include haloacetonitriles (HANs), haloketones (HKs) and halonitromethanes (HNMs). On the
other hand, some nitrogen DBPs (HANs and HNMs for example) and brominated DBPs may have higher
toxicological effects than THMs and HAAs (Krasner et al., 2006; Bove et al., 2007; Muellner et al., 2007;
Richardson, 2011), in particular concerning their cytotoxicity and genotoxicity.
In the United States, monitoring of THMs and HAAs are required under their respective maximum acceptable
concentration (MAC) of 80 µg/L for THMs and 60 µg/L for HAAs (USEPA, 2007). Compliance with the MAC is
based on the annual average of the maximum concentrations from samples collected each trimester at specific
locations along the distribution system (DS) corresponding to their higher level location (number of samples and
frequency depending on source water type and population served by the system). In Canada, the MAC for THMs
is 100 µg/L (Health Canada, 2006) and 80 µg/L for HAAs (annual average of the maximum concentrations from
samples collected each trimester) (Health Canada, 2008). However, these levels do not represent mandatory
standards. In Canada, drinking water regulations fall under provincial jurisdiction. The Province of Quebec has
established a standard of 80 µg/L for THMs and 60 µg/L for HAAs (annual average of the maximum
concentrations from samples collected each trimester) (MDDELCC, 2012). Moreover, in February 2012, Quebec
established the obligation to monitor THM4 (sum of the four regulated THMs) at specific frequencies during the
year. Frequencies range from one to eight samples per quarter depending on the population served by the
system. Sampling should be collected at the point where the concentration of THMs is the highest, which is
typically at the extremity of the DS. However, there is no obligation to monitor HAA5 (sum of the five regulated
HAAs) (MDDELCC, 2012). Also, several other provinces in Canada, including Newfoundland and Labrador,
have set guidelines for THM4 maximum concentration in drinking water at 100 µg/L, as recommended by Health
Canada (based on the average of a minimum of four measurements, one per quarter) (Health Canada, 2006).
Small water systems (i.e., serving 5,000 or fewer people) using surface waters may be more vulnerable to DBPs
because they have more difficulty implementing adequate treatment technologies to remove DBP precursors.
Indeed, our previous study showed that average measured concentrations of DBPs in small water systems
(SWS) were much higher than those reported in the literature for medium and large systems (Guilherme &
Rodriguez, 2014b). Selecting strategies for DBP monitoring can be a delicate task, because natural raw water
variations throughout the seasons have an impact on DBP levels, as well as the water residence time variations
along the DS (Lebel et al., 1997; Rodriguez & Sérodes, 2001; Mouly et al., 2010; Mercier-Shanks et al., 2013).
Consequently, there is an important degree of temporal and spatial variations in measured concentrations of
DBPs in the drinking water of SWS (Guilherme & Rodriguez, 2014b). Thus, identifying representative locations
and periods for DBP sampling for compliance purposes is a challenge for SWS operators. Several compliance
guides have been developed in Canada and the USA (MDDELCC, 2006; USEPA, 2007). Specifically, the United
States Environmental Protection Agency (USEPA) compliance guide offers several tools to help SWS better
86
monitor DBPs and the identification of sampling locations for THMs and HAAs. In Canada, however, the Initial
Distribution System Evaluation (IDSE) program proposed in the Stage 2 disinfectant and disinfection by-product
rule in the USA (USEPA, 2006) is not provided by Health Canada. The IDSE program is a one-time study
conducted during one year by the water system in order to identify distribution system locations with high
concentrations of THMs and HAAs. THMs and HAAs are measured in several locations along the DS (from 2 to
4 locations) from one to six times during the year of the campaign. The number of sampling locations, numbers
of samples collected at each location, and sampling frequency depend on the source water type and population
served (USEPA, 2006). Finally, this program provides system locations with high levels of THMs and HAAs,
corresponding to the sampling locations where the regulatory monitoring of THMs and HAAs should be carried
out. However, this program is difficult to implement in SWS. In fact, the cost of the program is quite high for SWS
because it requires many samplings and analyses. Moreover, this program requires in-depth knowledge of the
system (hydraulic characteristics and estimation of water residence time) and qualified operators. Yet SWS
experience some difficulty hiring qualified operators (Coulibaly & Rodriguez, 2004).
The objective of this paper was to develop a decision-making scheme for SWS that provides sampling periods
and locations for the regulatory monitoring of DBPs in small systems. This decision-making tool is relatively
inexpensive and easily implemented in SWS and takes account of their constraints and limitations. Moreover,
this tool can be used to evaluate maximum level locations and periods of non-regulated DBPs in SWS. The
development of this tool is based on data generated through two sampling campaigns organized in two provinces
in Canada (Newfoundland and Labrador and Quebec). The first campaign resulted in a good overview of the
spatial and temporal variability of DBP occurrence in one year in 25 SWS. Thereafter, the second campaign
completed the first by studying, in greater detail, the short-term variability of DBPs during a specific period within
the year (summer) in selected systems. According to our knowledge and for the first time in this research, a
decision-making tool for SWS has been developed to monitor regulated and non-regulated DBPs.
4.2. Methodology
4.2.1. Case studies
Two sampling campaigns were organized in the SWS of two provinces of Canada, Newfoundland and Labrador
(NL) and Quebec (QC). During the first sampling campaign, twenty-five SWS were sampled. All systems were
supplied by source surface waters and used chlorine as their main disinfectant (for primary and secondary
disinfection). Systems in NL (11 in total) served a population varying from 330 to 2,120 inhabitants. In QC,
systems (14 in total) served a population varying from 1,000 to 6,220 inhabitants. Thereafter, a second sampling
campaign was carried out in six SWS selected from the first campaign. For this campaign, the three systems
presenting the highest DBP levels in each province were chosen. Systems in NL did not present any prior
87
treatment to chlorination, whereas in QC, systems had implemented one or more treatment processes prior to
disinfection.
4.2.2. Sampling and analysis
During the two sampling campaigns, water was sampled in various sites along the DS in order to collect water
samples at different residence times. Figure 4.1 resumes sampling locations of the two campaigns. During the
first campaign, water was sampled at the beginning (DS1), middle (DS3) and end (DS6) of the DS. Sampling
was conducted monthly during a one-year period between September 2010 and October 2011 (from September
2010 to September 2011 in NL and from October 2010 to October 2011 in QC). During the second campaign,
the sampling locations of the first campaign were kept and three more sampling locations were added. One was
located between the beginning and the middle of the DS (DS2) and two were located between the middle and
the end of the DS (DS4 and DS5). Sampling was conducted daily (from Monday to Friday) during one month in
July 2012 for QC and in August 2012 for NL. During the sampling campaigns, samples were collected by water
operators (in NL) and by Université Laval personnel (in QC). Samplers were trained to follow equivalent sampling
protocols in both regions. All samples were drawn at the same time every day. Following field collection, the
samples were sent to the Université Laval laboratory for analysis.
Figure 4.1: Sampling locations during the campaigns
Residual disinfectant levels (free chlorine) and DBP concentrations were measured at each sampling location.
Five families of DBPs were considered: THMs, HAAs and three families of non-regulated DBPs (HANs, HNMs
and HKs). Four THMs (chloroform (TCM), bromodichloromethane (BDCM), dibromochloromethane (DBCM) and
tribromomethane (TBM)), five HAAs (monochloroacetic acid (MCAA), monobromoacetic acid (MBAA),
dichloroacetic acid (DCAA), trichloroacetic acid (TCAA) and dibromoacetic acid (DBAA)), four HANs
(dichloroacetonitrile (DCAN), trichloroacetonitrile (TCAN), bromochloroacetonitrile (BCAN) and
dibromoacetonitrile (DBAN)), one HNM (Chloropicrin (CPK)) and two HKs (1,1-dichloropropanone (DCP) and
1,1,1-trichloropropanone (TCP)) were analyzed during the study. During this project, a total of some 2,000
88
samples were collected. Details about analytical methods used are provided elsewhere (Mercier-Shanks et al.,
2013).
4.3. Results and discussion
The principal purpose of this decision-making scheme is to provide a strategy to identify sampling periods and
locations most appropriate for DBP monitoring. For DBP regulatory monitoring purposes, the most appropriate
sampling periods should represent the periods presenting the highest levels observed along the year. Similarly,
the most appropriate sampling locations should represent the locations presenting the highest levels observed
along the DS.
4.3.1. Sampling period identification
In this section, different scenarios were developed to calculate annual average levels of DBPs in order to identify
the most representative sampling periods for DBP monitoring. These scenarios present different selection
criteria for the sampling periods. First, the scenarios were compared to a reference scenario, which summarizes
the actual highest levels observed during the year. This reference scenario corresponds to the regulatory levels
obtained if more than one sample per trimester were measured. Then, scenarios most correlated with the
reference scenario and practical for the SWS were selected. A total of five different scenarios were established
to calculate the annual average levels of DBPs and one scenario to calculate the quarterly average levels of
DBPs. Table 4.1 presents the characteristics of these scenarios.
89
Table 4.1: Description of the different scenarios to calculate the annual average levels of DBPs
In all scenarios, DBP levels were selected at specific locations. THM levels were those measured at the end of
the DS (DS6), as recommended by Health Canada (Health Canada, 2006). The other DBP levels were those
measured in the middle of the DS (DS3), based on our previous study on the temporal and spatial variability of
Name of the scenario
Description Calculation process Database used
Reference scenario
Annual average of the four monthly levels observed in each trimester from fall 2010 to summer 2011
1. Selection of the maximum DBP levels observed each quarter (from three monthly measurements) 2. Calculation of the average of these four quarter maximum levels.
Université Laval database from the first sampling campaign: A one-year monthly sampling campaign organized in 25 SWS
Scenario 1
Regulatory DBP annual average based on a four-quarter average from fall 2010 to summer 2011 (only available for THMs)
1. Selection of the maximum DBP levels observed each quarter 2. Average of the four maximum levels measured at each quarter between fall 2010 and summer 2011
Regulatory databases from the provincial authorities responsible of DBP regulatory monitoring in Newfoundland & Labrador and Quebec
Scenario 2
DBP annual average based on a four-quarter average from fall 2010 to summer 2011. The four sampling periods starting respectively in October 1st, January 1st, April 1st and July 1st should be spaced out by at least 2 months (as recommended by QC guidelines).
1. Random selection of one DBP level each quarter (from three monthly measurements). Each selection is spaced out by at least 2 months. 2. Calculation of the average of these four quarter DBP levels.
Université Laval database from the first sampling campaign
Scenario 3
DBP annual average based on a four-quarter average from fall 2010 to summer 2011. There is no restriction on period between the four sampling periods starting at the same dates as scenario 2 (as recommended by NL guidelines).
1. Random selection of one value at each quarter (from three monthly measurements). 2. Calculation of the average of these four quarter DBP levels.
Université Laval database from the first sampling campaign
Scenario 4
DBP annual level based on a twelve-month average from October 2010 to September 2011
1. Calculation of the average levels from 12 monthly measurements.
Université Laval database from the first sampling campaign
Scenario 5
DBP annual average of levels observed on selected months from fall 2010 to summer 2011
1. Selection of DBP levels observed in specific months in each quarter for each DBP (months are listed in Table 4.3) 2. Calculation of the average of these four quarter DBP levels.
Université Laval database from the first sampling campaign
Scenario 6
Quarterly average level of DBPs based on a twenty-day average in one month in July and August 2012
1. Calculation of the average level from twenty daily measurements
Université Laval database from the second sampling campaign
90
DBPs in SWS, revealing that HAA, HAN, CPK and HK levels were maximized in the middle of the DS (Guilherme
& Rodriguez, 2014).
Figure 4.2 compares the annual averages of DBP levels for the same period 2010-2011 based on the different
scenarios (specific values are provided in Appendix 9). According to the scenarios shown in Table 4.1, the
optimal scenario for the sampling period should be the one most correlated with DBP values obtained with the
reference scenario. Concerning THMs, the results show that globally, levels obtained with scenario 1 represent
the lowest levels of THMs compared to the other scenarios (Figure 4.2). Average THM levels obtained with
scenario 1 (79.6 µg/L on average for all SWS) are 30% lower than those obtained with the reference scenario
(113 µg/L on average for all SWS). However, scenario 1 sampling and analytical methods differed from the other
scenarios because they were not measured by our laboratory. In order to preclude sampling and analytical bias,
scenario 2, scenario 3 and scenario 4 were also compared to the reference scenario. Scenario 2 and scenario
3 are comparable to scenario 1, but they are based on data generated during our first campaign (fall 2010 to
summer 2011), meaning with identical sampling and analytical methods to the reference scenario. Scenario 4,
on the other hand, presents a different sampling frequency. Results show that average THM levels obtained
with scenario 2 (86.5 µg/L on average for all SWS) and with scenario 3 (87.0 µg/L on average for all SWS) were
both lower than those obtained with the reference scenario. Also, as revealed by Figure 4.2, in NL more SWS
exceed the maximum acceptable concentration (MAC) of 100 µg/L when the annual average levels of THMs
are based on a twelve-month average with scenario 4 (7 SWS out of 11) than on a four-quarter average with
scenario 2 or scenario 3 (6 SWS out of 11). Also, most levels obtained with the reference scenario observed in
SWS in NL exceed the MAC (10 SWS out of 11).
91
a)
b)
Figure 4.2: Comparison of results of the various scenarios for the annual average level of a) THMs, b) HAAs, c)
HANs, d) CPK, and e) HKs
92
c)
d)
Figure 4.2: Comparison of results of the various scenarios for the annual average level of a) THMs, b) HAAs, c)
HANs, d) CPK, and e) HKs (suite)
93
e)
Figure 4.2: Comparison of results of the various scenarios for the annual average level of a) THMs, b) HAAs, c)
HANs, d) CPK, and e) HKs (suite)
Concerning the other DBPs (HAAs, HANs, CPK and HKs), levels obtained globally with scenario 2, scenarios 3,
and scenario 4 are all lower than the reference scenario. In fact, more SWS in NL exceed the MAC guideline on
HAAs (80 µg/L) when scenarios used are scenario 2 or the reference scenario (all SWS exceed MAC in this
case) compared to scenario 3 or scenario 4 (10 SWS on 11 in total exceed MAC). Table 4.2 summarizes the
Spearman correlation factors between the different scenarios for the calculation of the DBP level annual
average. Results revealed that values obtained with scenario 2 and scenario 4 are the most correlated with
those obtained with the reference scenario (representing the highest levels observed along the year). According
to these results, scenario 2 and scenario 4 represent the most adequate calculation scenarios in a context of
regulatory monitoring purposes compared to scenario 3. However, an annual average based on twelve monthly
samples (scenario 4) is more difficult to implement in small systems because it requires more samples, more
analyses and a bigger budget for SWS. Thus, an annual average based on a four quarter-average method
(scenario 2) is more convenient for SWS in a context of DBP monitoring.
94
Table 4.2: Correlation (Spearman) between different results of scenarios for annual average levels of a) THMs, b) HAAs, c) HANs, d) CPK and e) HKs
a)
**: Significant correlation at 1% level according to two-tailed test with SPSS©
b)
**: Significant correlation at 1% level according to two-tailed test with SPSS©
Spearman correlation factor
TH
M values w
ith
scenario 1
TH
M values w
ith
scenario 2
TH
M values w
ith
scenario 3
TH
M values w
ith
scenario 4
TH
M levels w
ith
scenario 5 THM values with
reference scenario 0.88** 0.97** 0.92** 0.98** 0.95**
THM values with scenario 1
1.00 0.90** 0.82** 0.89** 0.83**
THM values with scenario 2
- 1.00 0.90** 0.98** 0.95**
THM values with scenario 3
- - 1.00 0.95** 0.92**
THM values with scenario 4
- - - 1.00 0.98**
Spearman correlation factor
HA
A values w
ith
scenario 2
HA
A values w
ith
scenario 3
HA
A values w
ith
scenario 4
HA
A values w
ith
scenario 5
HAA values with reference scenario
0.93** 0.89** 0.95** 0.92**
HAA values with scenario 2
1.00 0.93* 0.97** 0.96**
HAA values with scenario 3
- 1.00 0.96** 0.96**
HAA values with scenario 4
- - 1.00 0.97**
95
Table 4.2: Correlation (Spearman) between different results of scenarios for annual average levels of a) THMs, b) HAAs, c) HANs, d) CPK and e) HKs (suite)
c)
**: Significant correlation at 1% level according to two-tailed test with SPSS©
d)
**: Significant correlation at 1% level according to two-tailed test with SPSS©
Spearman correlation factor
HA
N values w
ith
scenario 2
HA
N values w
ith
scenario 3
HA
N values w
ith
scenario 4
HA
N values w
ith
scenario 5
HAN values with reference scenario
0.87** 0.91** 0.90** 0.94**
HAN values with scenario 2
1.00 0.91** 0.96** 0.92**
HAN values with scenario 3
- 1.00 0.93** 0.91**
HAN values with scenario 4
- - 1.00 0.96**
Spearman correlation factor
CP
K values w
ith
scenario 2
CP
K values w
ith
scenario 3
CP
K values w
ith
scenario 4
CP
K values w
ith
scenario 5
CPK values with reference scenario
0.90** 0.79** 0.87** 0.83**
CPK values with scenario 2
1.00 0.94** 0.98** 0.93**
CPK values with scenario 3
- 1.00 0.92** 0.84**
CPK values with scenario 4
- - 1.00 0.90**
96
Table 4.2: Correlation (Spearman) between different results of scenarios for annual average levels of a) THMs, b) HAAs, c) HANs, d) CPK and e) HKs (suite) e)
*: Significant correlation at 1% level according to two-tailed test with SPSS©
Thereafter, we wanted to know if the most adequate scenario at the time, scenario 2, could be better correlated
with the reference scenario if the months of sampling were specified in each trimester (scenario 5, presented in
Table 4.1). To achieve this, months presenting maximum levels of DBP within each trimester were identified
using monthly DBP measures between October 2010 and September 2011 (Table 4.3). Figure 4.2 and Appendix
9 reveal that scenario 5 values are higher on average than scenario 2, scenario 3 and scenario 4 values for all
DBPs and much lower than reference scenario values. Table 4.2 also reveals that even if scenario 5 values are
much less correlated with reference scenario values than scenario 2 values, they are more correlated than
scenario 3 values. Thus, scenario 5 represents a good alternative to scenario 2 and scenario 3 which present a
random bias possibly due to the choice of sampling month within the trimester.
Table 4.3: Months presenting the maximum DBP level observed in all SWS (QC and NL included) for each trimester between fall 2010 and summer 2011 (based on 12 monthly DBP measurements)
DBPs (sampling location) Fall Winter Spring Summer
THMs (DS6) October January June September
HAAs (DS3) October January June August
HANs (DS3) October January May September
CPK (DS3) October March June September
HKs (DS3) October January June September
To move further ahead, we sought to evaluate whether a higher DBP sampling frequency would be necessary
in summer (scenario 6, presented in Table 4.1). In fact, various studies have shown that during the summer
Spearman correlation factor
HK
values with
scenario 2
HK
values with
scenario 3
HK
values with
scenario 4
HK
values with
scenario 5
HK values with reference scenario
0.94** 0.95** 0.95** 0.95**
HK values with scenario 2
1.00 0.96** 0.97** 0.96**
HK values with scenario 3
- 1.00 0.98** 0.98**
HK values with scenario 4
- - 1.00 0.99**
97
season, DBP levels are generally at their highest and environmental conditions (temperature, microbiological
activity) may favour increased formation, decomposition and biodegradation of some DBP species (Lebel et al.,
1997; Nikolaou et al., 2000; Nikolaou et al., 2001; Rodriguez et al., 2007; Guilherme & Rodriguez, 2014b). To
achieve this, levels obtained with scenario 1 (for THMs) and scenario 6 (in summer) were compared to levels
obtained with the comparison scenario in summer, scenario 2. First, the values of scenario 1 in summer and
scenario 2 taken in summer (one value chosen randomly from daily campaign in July and August 2012) were
compared (Table 4.4). Results showed that scenario 1 values in summer are relatively close to scenario 2 values
(difference of ±14% on average for all SWS under study) in spite of laboratory and analytical process differences.
Also, the values of scenario 2 and scenario 6 in summer are very close (difference of only ±12% on average for
all SWS under study). Thus, based on our campaign, intensive daily measurements during one month do not
provide a better overview of the population’s exposure to DBPs than one sample per trimester, as established
by DBP regulatory institutions. Finally, scenario 5 represents the most appropriate scenario for DBP monitoring.
Table 4.4: Summer average levels of a) THMs, b) HAAs, c) HANs, d) CPK and e) HKs based on different scenarios (levels in µg/L) a)
SWS Summer 2012
regulatory level of THMs (Scenario 1)
One value chosen randomly from daily campaigns in July and August 2012
(Scenario 2)
Daily average level of THMs based on a twenty-day average during one month in
July and August 2012 (Scenario 6)
NL3 80.9 82.8 63.9
NL5 130 118 123
NL10 308 229 212
QC8 86.1 94.4 81.4
QC11 77.1 79.6 77.4
QC13 88.3 117 104
98
Table 4.4: Summer average levels of a) THMs, b) HAAs, c) HANs, d) CPK and e) HKs based on different scenarios (levels in µg/L) (suite) b)
SWS
One value taken randomly from daily campaigns in July and August 2012
(Scenario 2)
Daily average level of HAAs based on a twenty-day
average during one month in July and August 2012
(Scenario 6)
NL3 82.3 85.4
NL5 188 179
NL10 218 271
QC8 37.7 33.2
QC11 29.6 30.6
QC13 113 118
c)
SWS
One value taken randomly from daily
campaigns in July and August 2012 (Scenario 2)
Daily average level of HANs based on a twenty-day average during one
month in July and August 2012
(Scenario 6)
NL3 2.95 2.85
NL5 4.08 4.28
NL10 7.14 8.07
QC8 3.66 3.22
QC11 3.20 3.46
QC13 2.92 3.30
d)
SWS
One value taken randomly from daily
campaigns in July and August 2012 (Scenario 2)
Daily average level of CPK based on a twenty-day average during one
month in July and August 2012
(Scenario 6)
NL3 0.79 0.85
NL5 0.43 0.50
NL10 0.87 0.97
QC8 1.52 1.28
QC11 0.75 0.31
QC13 0.53 0.58
99
Table 4.4: Summer average levels of a) THMs, b) HAAs, c) HANs, d) CPK and e) HKs based on different scenarios (levels in µg/L) (suite) e)
SWS
One value taken randomly from daily
campaigns in July and August 2012 (Scenario 2)
Daily average level of HKs based on a twenty-day average during one
month in July and August 2012
(Scenario 6)
NL3 8.96 8.35
NL5 19.3 16.3
NL10 12.5 14.0
QC8 5.24 3.15
QC11 4.46 5.25
QC13 9.60 9.91
Note 1: Numbers in bold represent values presenting difference (ratio: (Regulatory level - Calculated level)/Regulatory
level*100) with annual regulatory average upper than 20%.
Note 2: Grey boxes represent values higher than MAC for THMs and HAAs (respectively 100 and 80 µg/L in NL and 80
and 60 µg/L in QC).
4.3.2. Sampling location identification
In this section, the spatial variation of DBP levels in SWS was studied in order to identify the most representative
sampling locations for DBP monitoring (i.e., locations presenting the highest levels observed along the DS).
Results obtained from our first study (Guilherme & Rodriguez, 2014b) showed that DBP spatial variation patterns
depend on the season. Figures 4.3 and 4.4 present the spatial variation of DBPs within the DS in winter and in
SWS in summer. DBP levels in winter were obtained from the first sampling campaign carried out in 25 SWS in
2010 and 2011 (three sampling locations per system). DBP levels in summer were obtained from the second
sampling campaign carried out in 6 SWS in the summer of 2012 (six sampling locations per system). Generally
speaking, in winter, all DBPs seem following the same spatial evolution pattern along the DS. Their levels appear
to increase (or at least stabilize) at the end of the DS (Figure 4.3). Thus, maximum level location for all DBPs
would be at the end of the DS. However, in summer, our recent study showed that some HAAs (DCAA) and
some non-regulated DBP (DCAN, BCAN, CPK and HKs) levels decrease at the end of the DS in several SWS
(Guilherme & Rodriguez, 2014a). This study was based on data obtained from the second campaign with more
sampling locations per system in order to gain a better overview of the spatial variation of DBPs in this strategic
season. In fact, we observed that the biodegradation or decomposition of some DBPs appeared to occur more
particularly in SWS presenting a low free chlorine level at the end of the DS, usually under 0.3 mg/L (Guilherme
& Rodriguez, 2014a). This is in agreement with the Tung study carried out in SWS which concluded that free
residual chlorine levels lower than 0.3 mg/L, may favour biodegradation or decomposition of HAAs (Tung & Xie,
2009). It is also in agreement with DBP regulation guidelines that recommend maintaining a free residual chlorine
100
level greater than 0.3 mg/L in Quebec and 0.2 mg/L in the United States and Canada (USEPA, 2007; Health
Canada, 2012). However, the free chlorine level is easier to control in winter than in summer. In fact, we observed
that free chlorine demand is lower in winter (an average 50% decrease) than in summer (an average 70%
decrease) in our SWS under study. Decrease was calculated by the following ratio: (Level in DS1 – Level in
DS6)/Level in DS1*100. In summer, SWS experience greater difficulty ensuring adequate free chlorine levels at
the end of the DS due to higher temperatures, chlorine demand and the possible presence of biofilm in the DS.
Thus, under low chlorine conditions at the end of the DS, the decomposition of compounds (DCAA, DCAN,
BCAN and CPK) by biofilm might be possible.
a) b)
Figures reflect results obtained during the first campaign in 2010-2011, when samples were collected at three sampling locations along the DS.
Figure 4.3: Variations of average a) regulated and b) non-regulated DBP levels along the DS in winter (January -
March) in all SWS in NL and QC (number of observations for each DBP in each location = 75)
0
0,5
1
1,5
2
0
25
50
75
100
DS1 DS3 DS6
Fre
e c
hlo
rine level (µ
g/L
)
DB
P level (µ
g/L
)
THMs HAAs Free_Cl
0
2
4
6
DS1 DS3 DS6
DB
P level (µ
g/L
)
CPK HANs HKs
101
a) b)
c) d)
THM level in DS6 in Figure 4.4 a) is particularly low due to a lack of data at this point (30% less data at this point in one
SWS).
Figures reflect results obtained during the second campaign in 2012, when samples were collected at six sampling locations along the DS.
Figure 4.4: Variations of average DBP levels along the DS in summer; a) and c) in two SWS presenting a free
chlorine level higher than 0.3 mg/L at the end of the DS; and b) and d) in two SWS presenting a free
chlorine level lower than 0.3 mg/L at the end of the DS (number of observations for each DBP in each
figure = 40)
Figure 4.4 presents separate DBP spatial variations in summer between systems presenting a free residual
chlorine level higher (in two SWS) and lower (in two SWS) than 0.3 mg/L at the end of the DS. Two SWS under
study were not taken into account in this figure because one SWS carried out a rechlorination process, thus
presenting a DBP spatial variation pattern different from the other SWS. The second SWS experienced a break
in the chlorine pomp in the WTP during the campaign. Finally, in systems presenting a free chlorine level higher
102
than 0.3 mg/L at the end of the DS, DBP spatial variations in summer followed the same evolution pattern as in
winter. DBP levels increased regularly along the DS (except for THMs level in DS6 in Figure 4.4a due to a lack
of data at this point, 30% less of data at this point in one SWS). Thus, maximum DBP level locations remained
similar to winter locations at the end of the DS. But, in systems presenting a free residual chlorine concentration
lower than 0.3 mg/L at the end of the DS, DBP variations were different. HAA, HAN, CPK and HK concentrations
decreased at the end of the DS. In a context of DBP monitoring, this observation is relevant. It indicates that
depending on the free chlorine level decrease along the DS, it is possible to approximately identify the maximum
level location (MLL) for the different DBPs, in winter or in summer.
Since free chlorine levels in the DS differed greatly between systems, we chose to use the Δ free residual
chlorine concept, that is Δ free Cl (% of free residual chorine decrease compared to the beginning of the DS: Δ
free Cl DBP= (Free Cl in DS1 – Free Cl in MLL)/Free Cl in DS1 *100). Table 4.5 summarizes the values of Δ
free Cl for each DBP family (HAAs, HANs, CPK and HKs) at their MLL. Δ free Cl for THMs was not calculated
because no decomposition of THMs was observed in our SWS or in the literature (Rodriguez & Sérodes, 2001;
Mouly et al., 2010). MLL of THMs is usually located at the end of the DS. Results in Table 4.5 show that globally,
Δ free Cl values for each DBPs (HAAs, HANs, CPK and HKs) in all SWS are comparable and fluctuate from
45% to 80% (average Δ free Cl for all SWS and all DBPs equals to 70% ± 20% (average value ± standard
deviation)). Thus, Δ free Cl for HAAs, HANs, CPK and HKs is approximately between 50% and 90%. Based on
this information we were able to conclude that MLL for HAAs, HANs, CPK and HKs corresponds to a location
where Δ free Cl is between 50% and 90%, and as close as possible to 70%.
Table 4.5: Values of free residual chlorine decrease in each location with maximum DBP level obtained in four SWS (second campaign in summer 2012)
Δ free Cl HAA Δ free Cl HAN Δ free Cl CPK Δ free Cl HK
NL3 (average value ± standard deviation) 68% ± 11% 72% ± 17% 67% ± 12% 67% ± 12%
NL10 (average value ± standard deviation)
75% ± 14% 77% ± 15% 67% ± 12% 80% ± 18%
QC8 (average value ± standard deviation)
71% ± 22% 78% ± 15% 75% ± 18% 77% ± 21%
QC13 (average value ± standard deviation)
45% ± 17% 61% ± 22% 52% ± 15% 57% ± 25%
All SWS (average value ± standard deviation)
65% ± 20% 73% ± 18% 68% ± 17% 71% ± 21%
Note 1: Δ free Cl DBP= (Free Cl in DS1 – Free Cl in MLL )/Free Cl in DS1 *100 MLL: Maximum DBP level location Note 2: two SWS from the second campaign were excluded because one of them uses a rechlorination system and the other was subjected to a break in the chlorination pump at the WTP during the campaign.
103
Likewise in the context of DBP monitoring, the selection of sampling locations for THMs and HAAs should concur
with guidelines of DBP regulatory authorities. For example, the USEPA guidelines recommend the identification
of THM and HAA monitoring sites for SWS to avoid a dead-end site where there are no customers and a site
prior to booster disinfection (USEPA, 2007). Moreover, concerning the HAA site, USEPA guidelines recommend
avoiding sites where no disinfectant residual exists and sites with biofilm problems (USEPA, 2007). Thus, it
would be easy for SWS to select sampling locations for DBP monitoring using calculations of Δ free Cl values in
several locations along the DS. Locations presenting Δ free Cl values between 50% and 90% and as close as
possible to 70% could be potential sampling locations for DBP monitoring. Compliance with provincial authority
DBP regulations guidelines would help in the selection of final sampling locations.
4.3.3. Decision-making scheme for DBP monitoring
Results and observations regarding DBP variations presented in the previous sections were combined in a final
decision-making tool for DBP monitoring for SWS that provide maximum level periods and locations for regulated
and non-regulated DBPs. This tool can be used for the identification of the most adequate sampling periods and
locations for the regulatory monitoring of THMs and HAAs (if regulated). The tool can also be used to identify
periods and locations in the system where exposure of the population to non-regulated DBPs would be at its
maximum.
This decision-making scheme is presented in Figure 4.5. To apply this scheme, it was necessary to conduct an
initial investigation of the characteristics of the SWS. For example, the presence of a rechlorination facility or a
water storage tank will strongly influence the MLL of DBPs (Rodriguez et al., 2004). Thereafter, various potential
sampling locations must be selected along the DS according to DBP regulation guidelines. For example, in
Figure 4.5, site E should not be considered as a good sampling location because it conflicts with USEPA
guidelines: it is a dead-end site where there are no customers.
In the pre-selected sampling locations, a field campaign should be organized to measure free residual chlorine
along the DS. Such a campaign should provide relevant information, including an approximate estimation of the
water residence time all along the DS. From the information of free residual chlorine levels, SWS will be able to
calculate the seasonal Δ free Cl of each location. The SWS should select a sampling location for HANs, HAAs,
CPK and HKs with a Δ free Cl value between 50% and 90%, and as close as possible to 70%. Typically, THMs
will be sampled at the end of the DS (according to current DBP regulation and guidelines).
104
Figure 4.5: Decision-making scheme for regulated and non-regulated DBP monitoring
105
Usually, chlorine demand is lower in winter and in the cold periods of spring and fall than in summer (except
during the snowmelt in spring, which can be a source of NOM in water increasing chlorine demand). So, if Δ free
Cl is lower than 50% until the end of the DS (D in Figure 4.5), site D should be selected as the sampling location
for the regulatory monitoring of THMs and HAAs and would be the MLL for the non-regulated DBPs. However,
in summer and in the warm periods of spring or fall, maintaining an adequate free chlorine level might be more
difficult, because of higher temperatures and chlorine demand (also due to the decay of vegetation during the
fall). Thus, if Δ free Cl in D is higher than 90%, site C would be the sampling location for the regulatory monitoring
of HAAs and would be the MLL for the non-regulated DBPs. But, if Δ free Cl in C is higher than 90%, site B
would be selected, etc. Samples should be collected at least once per quarter for SWS. The final step of the
decision-making scheme identifies the months presenting the maximum DBP levels in each trimester. Samples
of each DBP should be taken during the corresponding months. Finally, an annual average could be calculated
based on a four-quarter average (of the maximum level observed each quarter), as recommended by Health
Canada (Health Canada, 2012).
4.4. Conclusions
This investigation served to develop an easy and practical decision-making scheme for DBP monitoring intended
for SWS. SWS experience some difficulties monitoring DBPs, especially SWS supplied by surface waters. In
fact, various factors in raw water and treated water influence DBP formation or speciation. There is also an
important degree of temporal and spatial variation of DBP levels within a year and along the DS. DBP monitoring
is difficult for SWS because it requires good knowledge of the DS and qualified operators to oversee monitoring.
The study of the spatial and temporal variability of DBPs served to develop scenarios to identify the most
appropriate sampling periods and locations for monitoring regulated and non-regulated DBPs. The results show
that quarterly sampling of DBPs made at specific months along the year allow the measurement of DBPs at their
highest level periods.
The study also highlights the fact that a simple free residual chlorine measure at several locations along the DS
can serve to identify locations where DBP levels are the highest. Based on this information, it would be easy for
SWS to identify the most adequate sampling periods and locations for the regulatory monitoring of THMs and
HAAs (and also for non-regulated DBPs in case of potential future regulation).
The decision-making scheme could also be combined with models that we have developed estimating non-
regulated DBP levels at their maximum level location depending on the season and treatment employed by the
system (Guilherme & Rodriguez, 2014a). Models can easily applied because they use measurements of
parameters already regulated or monitored in the WTP and in the DS, including residual disinfectant and
regulated DBPs. A combination of models and scenarios could be used for operational purposes at the WTP,
106
monitoring purposes using regulatory databases, and for assessing population exposure to non-regulated DBPs
for future epidemiological studies that associate drinking water contaminants with human health.
4.5. References
Bove, G.E.J., Rogerson, P.A., Vena, J.E., 2007. Case control study of the geographic variability of exposure to
disinfectant byproducts and risk for rectal cancer. International Journal of Health Geographics 6 (18).
Cedergren, M.I., Selbing, A.J., Löfman, O., Källen, B.A.J., 2002. Chlorination byproducts and nitrate in drinking
water and risk for congenital cardiac defects. Environmental Research 2 (89), 124-130.
Coulibaly, H.D., Rodriguez, M.J., 2004. Development of performance indicators for small Quebec drinking water
utilities. Journal of Environmental Management 73, 243–255
Environmental Protection Agency, USEPA, 2006. Initial distribution system evaluation guide for systems serving
fewer than 10,000 people for the final stage 2 disinfectants and disinfection byproducts rule, Washington, DC,
US.
Environmental Protection Agency, USEPA, 2007. Complying with the stage 2 disinfectant and disinfection
byproducts rule: small entity compliance guide, Washington, DC, US.
Guilherme, S., Rodriguez, M.J., 2014a. Monitoring and Modeling of Non-Regulated Disinfection-by Products in
Small Systems. AWWA Annual Conference and Exposition, Boston, Massachusetts, USA
Guilherme, S., Rodriguez, M.J., 2014b. Occurrence of regulated and non-regulated disinfection by-products in
small drinking water systems. Chemosphere 117, 425-432.
Health Canada, 2006. Guidelines for Canadian Drinking Water Quality: Guideline Technical Document,
Trihalomethanes. Water Quality and Health Bureau, Healthy Environments and Consumer Safety Branch.
Health Canada, 2008. Guidelines for Canadian Drinking Water Quality: Guideline Technical Document —
Haloacetic Acids., Ottawa, Ontario, CANADA
Health Canada, 2012. Guidelines for Canadian Drinking Water Quality.
Krasner, S. W., Weinberg, H. S., Richardson. S., Pastor, S.J., Chinn, R., Sclimenti, M.J., et al., 2006. Occurrence
of a new generation of disinfection byproducts. Environmental Science & Technology 40, 7175-7185.
107
Lebel, G.L., Benoit, F.M., Williams, D.T., 1997. A one-year survey of halogenated disinfection by-products in the
distribution system of treatment plants using three different disinfection processes. Chemosphere 34 (11), 2301-
2317.
Ministère du Développement Durable, de l'Environnement et de la Lutte contre les Changements Climatiques,
MDDELCC, 2006. Guide de conception des installations de production d'eau potable.
Ministère du Développement Durable, de l'Environnement et de la Lutte contre les Changements Climatiques,
MDDELCC, 2012. Réglement sur la qualité de l'eau potable.
Mercier-Shanks, C., Sérodes, J.-B., Rodriguez, M.J., 2013. Spatio-temporal variability of non-regulated
disinfection by-products within a drinking water distribution system Water Research 47, 3231-3243
Mouly, D., Joulin, E., Rosin, C., Beaudeau, P., Zeghnoun, A., Olszewski-Ortar A., et al., 2010. Variations in
trihalomethane levels in three French water distribution systems and the development of a predictive model.
Water Research 44, 5168-5179.
Muellner, M.G., Wagner, E.D., McCalla, K., Richardson, S.D., Woo, Y.-T., Plewa, M.J., 2007. Haloacetonitriles
vs. regulated haloacetic acids: are nitrogen-containing DBPs more toxic? Environmental Science & Technology
41, 645-651.
Nikolaou, A.D., Golfinopoulos, S.K., Kostopoulou, M.N., Lekkas, T.D., 2000. Decomposition of
dihaloacetonitriles in water solutions and fortified drinking water samples. Chemosphere 41, 1149-1154.
Nikolaou, A.D., Lekkas, T.D., Kostopoulou, M.N., Golfinopoulos, S.K., 2001. Investigation of the behaviour of
haloketones in water samples. Chemosphere 44, 907-912.
Richardson, S.D., 2011. Disinfection byproducts: formation and occurrence in drinking water. The Encyclopedia
of Environmental Health 2, 110-136.
Richardson, S.D., Plewa, M.J., Wagner, E.D., Schoeny, R., DeMarini, D.M., 2007. Occurrence, genotoxicity, and
carcinogenicity of regulated and emerging disinfection by-products in drinking water: A review and roadmap for
research. Mutation Research 636, 178–242.
Rodriguez, M.J., Sérodes, J.-B., 2001. Spatial and temporal evolution of trihalomethanes in three water
distribution systems. Water Research 35 (6), 1572-1586.
108
Rodriguez, M.J., Sérodes, J.-B., Levallois, P., 2004. Behavior of trihalomethanes and haloacetic acids in a
drinking water distribution system. Water Research 38, 4367-4382.
Rodriguez, M.J., Sérodes, J.-B., Levallois, P., Proulx, F., 2007. Chlorinated disinfection by-products in drinking
water according to source, treatment, season, and distribution system. Journal of Environmental Engineering
and Science 6, 355-365.
Rook, J.J., 1974. Formation of haloforms during chlorination of natural waters. Water Treatment Examination,
23, 234-243.
Singer, P.C., 2002. Occurrence of haloacetic acids in chlorinated drinking water. Water Science and Technology:
Water Supply 2 (5-6), 487-492
Tung, H.-H., Xie, Y.F., 2009. Association between haloacetic acid degradation and heterotrophic bacteria in
water distribution systems. Water Research 43, 971-978.
Villanueva, C., Cantor, K.P., Grimalt, J.O., Malats, N., Silverman D., Tardon, A., et al., 2007. Bladder cancer and
exposure to water disinfection by-products through ingestion, bathing, showering and swimming in pools.
American Journal of Epidemiology 165, 148-156.
109
Conclusions et recommandations
Cette thèse de doctorat a permis d’améliorer les connaissances sur la variabilité de l’occurrence des SPD, en
particulier les SPD non-réglementés, dans les petits réseaux d’eau potable. Ce projet s’est principalement basé
sur deux importantes campagnes d’échantillonnage dans 25 petits réseaux de deux provinces du Canada
(Québec et Terre-Neuve-et-Labrador) étalées sur deux années entre 2010 et 2012. A l’aide de l’importante base
de données générée lors de ces deux campagnes d’échantillonnage, une étude précise de la variabilité des
SPD (THM, AHA, HAN, HNM et HC) a pu être réalisée, aussi bien spatiale que temporelle, et à long et court
terme.
Tout d’abord, le premier chapitre a révélé les différences de niveaux de SPD entre les petits et grands réseaux
alimentés par de l’eau de surface. Les niveaux de SPD sont globalement plus importants dans les petits réseaux,
attestant de leur vulnérabilité aux SPD. Le premier chapitre a surtout étudié les variabilités spatiales et
temporelles à long terme des SPD (sur une base mensuelle et saisonnière). L’article a soulevé une grande
différence des niveaux mesurés de SPD entre les deux provinces du Canada à l’étude, qui est principalement
liée aux différences d’obligations règlementaires concernant les filières de traitement de l’eau à employer dans
les réseaux alimentés par des eaux de surface. Aussi ce premier chapitre a permis d’observer des variabilités
spatiales étonnantes pour les petits réseaux, comme par exemple la potentielle dégradation du DCAA le long
du réseau indiquant fort probablement un temps de séjour de l’eau étonnamment long dans les petits réseaux.
Bien que les observations sur la variabilité temporelle des SPD concordent avec les observations faites dans
les plus grands réseaux (niveaux maximaux de SPD observables en été), les résultats obtenus ont révélé la
forte influence des caractéristiques locales sur les niveaux de SPD mesurés, d’où l’importance d’étudier les
petits réseaux de façon spécifique.
Le deuxième chapitre s’est focalisé sur l’étude des facteurs associés à la variabilité spatiale et temporelle des
SPD et en particulier des SPD non-réglementés. A l’aide d’analyses de corrélations bivariées, les paramètres
les plus déterminants dans les niveaux des SPD ont pu être identifiés. Ces paramètres étaient principalement
le type de traitement, les caractéristiques de l’eau soumise à la chloration tel que l’absorbance UV, le COD, le
pH et la température ainsi que le chlore libre résiduel. Afin de prendre en compte tous ces paramètres, des
modèles de régression linéaires à effets mixtes ont été développés. Ces modèles permettent d’estimer la
concentration en SPD non-réglementés dans l’eau potable de petits réseaux à partir de mesures de certains
paramètres de l’eau dans l’usine de traitement et dans le réseau de distribution. La particularité de ces modèles
est qu’ils intègrent dans leurs variables explicatives les SPD réglementés comme indicateurs des SPD non-
réglementés. Ces modèles prennent en compte les variabilités spatiales et temporelles spécifiques des SPD.
En effet, les modèles diffèrent entre les périodes chaudes et froides pour prendre en compte la variabilité
110
saisonnière des SPD. Aussi, les SPD réglementés sont mesurés à des sites spécifiques correspondant aux
localisations où leur niveau est maximal. De même les niveaux des SPD non-réglementés sont estimés à des
endroits spécifiques le long du réseau, où leur niveau sont maximaux. Finalement les modèles développés dans
ce chapitre peuvent être utilisés pour différentes applications, par exemple, pour des fins opérationnelles afin
d’évaluer l’influence d’une variation d’un paramètre à l’usine de traitement sur l’occurrence finale des SPD non-
réglementés. Ces modèles peuvent aussi être utilisés à des fins épidémiologiques, afin d’évaluer l’exposition
maximale de la population aux SPD non-réglementés.
Dans le troisième chapitre, les variabilités spatiale et temporelle à court terme des SPD, et en particulier des
SPD non-réglementés, ont été étudiées. Pour cela, une deuxième campagne d'échantillonnage a été organisée
dans six petits réseaux au Canada (QC et TN) en été 2012. Les systèmes ont été échantillonnés tous les jours
à l'usine de traitement de l'eau ainsi qu’à six points différents le long du réseau de distribution. Pour la première
fois, la variabilité à court terme des SPD a été observée dans les petits réseaux à partir de données récoltées
sur le terrain. Les résultats obtenus ont permis de mettre en lumière une importante variabilité temporelle et
spatiale à court terme. Par exemple, dépendamment de la période d’échantillonnage, un réseau peut présenter
des niveaux en SPD réglementés supérieurs aux moyennes annuelles admissibles et quelques jours plus tard,
présenter un niveau tout à fait acceptable. De plus, cette étude a permis d’étudier plus en détails la variabilité
des SPD le long du réseau. Ainsi, selon les petits réseaux étudiés, la dégradation de plusieurs composés de la
famille des AHA, des HAN, des HNM ou des HC a été détectée renforçant l’hypothèse d’un temps de séjour
plus long et complexifiant de ce fait le suivi des SPD dans ces réseaux. Enfin, ce chapitre a mis en évidence
que le chlore libre résiduel peut être un bon indicateur du temps de séjour de l’eau dans les petits réseaux. De
plus c’est un paramètre couramment mesuré par les opérateurs et dont la mesure est peu coûteuse et de
manière générale bien maîtrisée par les opérateurs des petits réseaux.
Finalement, ces observations ont mis de l’avant la nécessité de mettre en place une stratégie de suivi des SPD
qui tient compte des fortes variations temporelles et spatiales. Ainsi le quatrième chapitre a été axé sur le
développement d’un outil d’aide à la décision afin de faciliter le suivi des SPD réglementés. L’outil développé,
peut facilement être utilisé par des petits réseaux car il ne nécessite qu’une simple campagne de chlore libre
résiduel, un paramètre peu coûteux à analyser. L’outil permet d’identifier les périodes et sites d’échantillonnage
les plus représentatifs pour le suivi réglementaire des THM et des AHA (si un suivi est mis en place), ainsi que
les périodes et les lieux dans le réseau où l’exposition aux SPD non-réglementés est maximale.
L’originalité de cette étude réside dans son intérêt à étudier la qualité de l’eau potable desservant des
populations parmi les plus vulnérables, soit les habitants des petites municipalités. En s’intéressant
particulièrement à ces réseaux, notre étude a voulu mettre en évidence les conséquences de leurs contraintes
111
financières et logistiques sur la qualité de leur eau potable. La première conséquence découlant de leur système
de traitement, souvent beaucoup plus simple que dans les grands réseaux, est des concentrations en SPD bien
supérieures à celles observées dans les grands réseaux. De plus, les problèmes de stagnation de l’eau dans
les canalisations ou les difficultés de gestion du système hydraulique entraînent des temps de séjour de l’eau
dans les conduites des petits réseaux souvent supérieurs à ceux des plus grands. Des temps de séjour
particulièrement longs peuvent favoriser la transformation ou la dégradation de certains SPD et rendent le suivi
des SPD particulièrement compliqué.
En plus de son originalité, cette thèse se distingue par l’applicabilité des résultats. Pour pallier aux difficultés
des petits réseaux dans le suivi des SPD, notre étude s’est consacrée à développer deux approches pour les
guider. La première se veut une méthode alternative aux analyses de laboratoire souvent coûteuses afin
d’évaluer les niveaux en SPD non-réglementés dans l’eau potable des petits réseaux à l’aide de modèles de
régression. Ainsi, les petits réseaux peuvent évaluer les concentrations de ces composés et se préparer en cas
de future réglementation. La deuxième est un outil d’aide à la décision, qui se veut un guide de base pour la
mise en place d’une stratégie de suivi des SPD. L’outil développé pourrait être généralisé à tous les petits
réseaux qui présentent un climat similaire à celui des régions étudiées, typiquement le nord-est de l’Amérique
du nord.
Malgré la quantité des résultats et des applications mis en valeur lors de cette thèse, les nombreuses données
générées par ce projet pourront être utilisées par la suite à d’autres fins. Par exemple, dans un contexte plus
pratique, les deux campagnes d’échantillonnage réalisées vont permettre aux petits réseaux étudiés d’évaluer
leurs concentrations en SPD, en particulier de AHA et de SPD non-réglementés. En effet, un suivi des THM a
été mis en place dans tous les SPD étudiés mais ce n’est pas le cas encore pour les AHA. Au Québec, le suivi
obligatoire des AHA a de grandes chances d’être instauré par le gouvernement dans un avenir proche. À Terre-
Neuve-et-Labrador ce suivi pourrait être obligatoire dans l’avenir. Ces campagnes d’échantillonnage ont donc
permis aux municipalités échantillonnées d’avoir une estimation des niveaux de AHA présent dans leur réseaux
ainsi qu’un rapport complet sur les données. Un exemple de rapport envoyé aux petits réseaux étudiés à la suite
de la première campagne est présenté en annexe 10. Cette information leur permet de se préparer à être
conforme avec la réglementation (surtout au Québec) et de prévoir à l’avance, si nécessaire, des modifications
aux infrastructures de traitement de l’eau ou des stratégies de distribution de l’eau. De même pour les SPD non-
réglementés, les petits réseaux étudiés ont pu évaluer les concentrations de ces composés et se préparer en
cas de future réglementation. De plus, les deux grandes bases de données réalisées sur ces 25 petits réseaux
pourront être utilisées dans le futur dans un contexte d’évaluation en santé publique. Par exemple, ces bases
de données pourraient être utilisées pour évaluer l’exposition de la population des petits réseaux aux SPD
présents dans l’eau potable, et en particulier les SPD non-réglementés.
112
En plus des données originales générées, cette thèse a réussi, à chaque chapitre, à répondre à des problèmes
spécifiques des petits réseaux. Chaque connaissance et chaque outil peuvent être utilisés de façon
indépendante. Cependant, puisque ces chapitres sont liés ensemble, il serait intéressant par la suite de
combiner les outils majeurs développés lors de ce doctorat en un guide complet pour le suivi des SPD pour les
petits réseaux. L’idée serait de simplifier les modèles développés lors du deuxième chapitre afin de n’avoir
besoin que des données disponibles dans les bases de données réglementaires pour évaluer les niveaux des
SPD non-réglementés. Les modèles ainsi simplifiés seraient combinés à l’outil d’aide à la décision développé
dans le quatrième chapitre. Ce guide final permettrait aux gestionnaires des petits réseaux d’identifier les
périodes et les lieux d’échantillonnage pour le suivi réglementaire des SPD réglementés (et non-réglementés,
en cas de future réglementation) tout en leur permettant d’évaluer le niveau en SPD non-réglementés dans leur
eau potable. Ce guide pourrait être facilement utilisé par les petits réseaux car il ne nécessite que des mesures
de chlore libre résiduel et des bases de données provenant du suivi réglementaire. Il pourrait être généralisé à
tous les petits réseaux du nord-est de l’Amérique du nord et permettrait une meilleure gestion de l’eau dans ces
réseaux. En effet, tous ces petits réseaux seraient en mesure d’estimer leur niveau en SPD (THM, AHA, HAN,
HNM et HC) et donc de se préparer pour de futures normes, et de prévoir à l’avance, si nécessaire, des
modifications à leur usine de traitement de l’eau et aux stratégies de distribution de l’eau.
Cette thèse comporte néanmoins certaines limites. Par exemple, les bases de données générées sont basées
sur des campagnes de terrain réalisées dans 25 petits réseaux de deux provinces du Canada (Québec et Terre-
Neuve-et-Labrador). Il aurait été intéressant dans ce projet d’observer peut-être plus de provinces en Amérique
du nord. Cette idée aurait engendré des coûts et de la logistique bien supérieurs mais aurait permis possiblement
une plus grande généralisation géographique. Aussi, il aurait été intéressant de réaliser des campagnes
similaires à la seconde campagne en d’autres saisons au cours de l’année en plus de l’été. Par exemple, dans
le nord-est de l’Amérique du nord, lors de la fonte des neiges pendant la saison printanière et lors de la
décomposition de la végétation lors de la saison automnale, ces deux périodes de l’année entraînent des
problématiques pour le contrôle de la qualité de l’eau, notamment en raison de l’importante quantité de matière
organique dans les eaux de surface. Il aurait donc été intéressant de réaliser des campagnes intensives lors de
ces événements afin de mieux suivre l’évolution des niveaux de SPD à ces moments particuliers. Cela aurait
aussi permis de guider les petits réseaux dans le suivi des SPD dans ces périodes stratégiques. Aussi, les
modèles développés dans le deuxième chapitre prennent en compte globalement le type de traitement utilisé.
Il aurait été intéressant de développer des modèles spécifiques à chaque type de traitement utilisé dans les
petits réseaux. Cependant, vu la multitude de traitements utilisés au Québec, la validité des modèles aurait été
compromise par la quantité réduite de données disponibles pour développer chaque modèle. De plus, dans les
petits réseaux, la gestion des usines de production d’eau influence beaucoup la qualité de l’eau traitée, voire
plus que le traitement lui-même. Or aucune information sur la gestion des usines de production n’a été récoltée
113
lors de cette campagne. En effet, ces informations sont très difficiles à obtenir dans les petits réseaux et, de
manière générale, ne sont pas disponibles.
Les programmes d’échantillonnages organisés lors de ce projet pourraient être améliorés. Tout d’abord, il est
important de préciser que les bases de données générées ne comportent pas de renseignements sur les
paramètres opérationnels, notamment la dose de chlore ajoutée dans l’eau au poste de chloration ou à l’usine
de traitement. Ce paramètre est rarement enregistré de manière précise par les opérateurs des petits réseaux.
Notre étude s’est donc focalisée sur le chlore libre résiduel, car ce paramètre est couramment mesuré et
documenté dans tous les petits réseaux. En effet, les opérateurs doivent s’assurer qu’un certain niveau minimal
de chlore libre résiduel soit présent en tout temps dans le réseau. Dans des futures études à réaliser dans les
petits réseaux, il serait néanmoins intéressant de documenter les doses de chlore ajouté et d’étudier leurs
influences sur l’occurrence des SPD. Le fait que seulement quelques SPD non-réglementés ont pu être analysés
lors des programmes d’échantillonnages représente une autre limite. La méthode analytique employée nous
restreignait aux familles de composés sélectionnés (HAN, HNM et HC). Cependant, il aurait été intéressant
d’analyser d’autres SPD non-réglementés qui présentent des risques toxicologiques importants comme par
exemple les nitrosamines (et notamment le N-nitrosodiméthylamine (NDMA)) ainsi que les SPD iodés.
Cependant, cela aurait nécessité le développement de plusieurs méthodes d’analyse assez onéreuses et aurait
été limité par les capacités et disponibilités des dispositifs analytiques de notre laboratoire. Heureusement, les
outils développés dans notre projet peuvent s’adapter à des futurs ajouts de familles de composés dans le cas
où d’autres espèces de SDP non-réglementés seraient étudiées dans les petits réseaux. Enfin, lors de notre
collecte de données, aucune information sur les temps de séjour de l’eau dans les réseaux de distribution n’a
pu être collectée car très peu de petits réseaux ont à leur disposition d’études hydrauliques précises. Cette
information aurait été très utile pour mieux comprendre la dégradation de plusieurs SPD (dont plusieurs SPD
non-réglementés) et faciliter leur suivi dans les petits réseaux.
Toutes ces limites nous révèlent l’importante multidisciplinarité des solutions à apporter aux problématiques des
petits réseaux. Une aide complète aux petits réseaux ne pourra se faire qu’en associant des connaissances en
gestion de la qualité de l’eau, en modélisation hydraulique, en chimie analytique et en santé publique.
115
Annexe 1 : Répartition géographique des 25 petits
réseaux étudiés dans les provinces a) de Québec
et b) de Terre-Neuve-et-Labrador
a)
b)
116
Annexe 2 : Caractéristiques des 25 petits réseaux
étudiés des provinces de Québec et Terre-Neuve-
et-Labrador
Région
Réseau Population (2006) (1)
Type de source
Type de traitement (2)
Coag. /Floc.
Séd. Filtr. Ozon. Charbon
Act. UV Chloration Chloram.
Qu
ébec
QC1 6 226 Rivière X X X - - - X -
QC2 3 006 Rivière X X X - - - X -
QC3 5 287 Rivière X X X - - - X -
QC4 1 005 Rivière X X X - X (été) - X -
QC5 5 564 Rivière - X X - - - X -
QC6 1 312 Lac - - - - - X X -
QC7 4 045 Lac X X X X - - X -
QC8 3 220 Rivière X X X - X X X -
QC9 1 528 Rivière X X X - - - X -
QC10 1 574 Lac X X X - - - X -
QC11 3 363 Rivière - - X X X - - X
QC12 1 183 Lac - - X X - X X -
QC13 3 826 Lac - - - - - - X -
QC14 5 021 Rivière X X X - - - X -
Terre-N
euve-et-L
abrad
or
TN1 753 Petit Lac - - - - - - X -
TN2 2 072 Lac - - - - - - X -
TN3 321 Rivière - - - - - - X -
TN4 2 062 Petit Lac - - - - - - X -
TN5 1 020 Petit Lac - - - - - - X -
TN6 1 500 Rivière - - - - - - X -
TN7 670 Petit Lac - - - - - - X -
TN8 571 Rivière - - - - - - X -
TN9 841 Petit Lac - - - - - - X -
TN10 329 Petit Lac - - - - - - X -
TN11 450 Rivière - - - - - - X -
TN12 1 029 Petit Lac - - - - - - X -
(1) Statistique CANADA, recensement de 2006 (2) Gouvernement du Québec, Suivi règlementaire des THM, 2009 Note : Petit lac correspond à la traduction française du terme « pond » utilisé par la réglementation de Terre-Neuve et Labrador. Coag./Floc. : Coagulation / Floculation, Sed : Sédimentation, Filt : Filtration, Ozon. : Ozonation, Charbon act. : Charbon actif, Chloram. : Chloramination
117
Annexe 3 : Informations sur les méthodes
analytiques utilisées pour l’analyse des SPD
Protocole d’échantillonnage
L’eau est échantillonnée au robinet des sites sélectionnés. Un écoulement d’eau froide d’une période de cinq
minutes précède la collecte des échantillons pour s’assurer que l’échantillon est bien représentatif de l’eau du
réseau et non l’eau qui aurait pu stagner dans les canalisations du robinet. Les échantillons des THM et AHA
sont prélevées en duplicata dans des vials de 40ml contenant une solution de NH4Cl afin d’éviter la formation
de SPD après l’échantillonnage. Les échantillons des SPD non-réglementés sont prélevées en duplicata dans
des vials de 60ml contenant une solution de NH4Cl afin d’éviter la formation de SPD après l’échantillonnage
mais aussi une solution de tampon pH.
Lieux d’analyse
Les échantillons de AHA et SPD non-réglementés ont été analysés dans le laboratoire de la Chaire de recherche
en eau potable de l’université Laval. Les échantillons de THM ont été analysés dans le laboratoire du service
de l'environnement de la ville de Québec. Les analyses de bromures ont été réalisées par le laboratoire du
CEAEQ du Ministère du Développement Durable, de l’Environnement et des Parcs de Québec (MDDEP).
Méthodes analytiques (basées sur le mémoire de Maîtrise de Catherine Mercier-Shanks : Variabilité spatio-
temporelle des sous-produits de la désinfection émergents (Haloacétonitriles, Halonitrométhane et Halocétones)
dans un réseau de distribution d’eau potable, 2012, Université Laval)
SPD non-réglementés
Le protocole d’extraction des SPD non-réglementés a été mis en place par le laboratoire de l’Université Laval.
Il a été développé par une étudiante lors de sa maîtrise (Mercier-Shanks, 2012). Les SPD non-réglementés ont
été extraits par une procédure d'extraction liquide-liquide principalement basée sur la méthode EPA - 551.1
(Environmental Protection Agency, 1995a). Pour la mesure des SPD non-réglementés, l’eau a été prélevée
dans des vials de 60 mL en présence de deux agents de conservation: du chlorure d'ammonium (NH4Cl) et un
tampon phosphate. Le NH4Cl est utilisé comme agent déchlorant permettant de neutraliser le résiduel de chlore
présent dans l'échantillon. Le tampon phosphate utilisé est un mixte solide de 1% de Na2HPO4 pour 99% de
KH2PO4 en proportions poids/poids (Mercier-Shanks, 2012). Ce tampon permet d’obtenir un pH entre 4,5 et 5,5.
Une fois prélevés, les échantillons étaient entreposés à l'abri de la lumière à environ 4°C en attendant leur
extraction, réalisée au plus tard dans les 48 heures. L’extraction des SPD non-réglementés a été réalisée dans
118
du MTBE. Afin de favoriser l'extraction des composés, un sel (du Na2S04) a été ajouté lors de l'extraction de
manière à diminuer la solubilité des composés dans la phase aqueuse et de favoriser leur transfert vers la phase
organique (MTBE). Afin de s’assurer de la validité et de la qualité des résultats obtenus, des blancs, des
duplicatas d'échantillons et des contrôles de concentration connue (0,20 µg/L et 4,00 µg/L) ont été ajoutés à
chacune des séquences d'échantillons extraits (Mercier-Shanks, 2012). De plus, un analogue d'extraction
(surrogate), du 1-chloro-2-bromopropane, a été ajouté à tous les échantillons afin de suivre l'efficacité de
l'extraction ainsi qu’un standard interne, du 1,2,3-trichloropropane, afin de s'assurer de la qualité des analyses
chromatographiques de chacun des échantillons traités. L'analyse des extraits a été effectuée à l'aide d'un GC-
ECD de marque Perkin Elmer (Clarus 500, muni de deux détecteurs ECD avec sources radioactives au Ni). Une
colonne DB-1 (100% diméthylpolysiloxane) a été utilisée pour la quantification, et une colonne DB-5 (5%
diphényl 95% diméthylpolysiloxane) a été utilisée pour la confirmation des résultats pour s'assurer qu'il n'y ait
pas de coélutions entre nos composés cibles et d’autres composés présents dans l’échantillon (Mercier-Shanks,
2012). La quantification des concentrations des SPDE a été effectuée à l'aide de courbes d'étalonnage allant
de 0,01 µ/L à 15,00 µg/L. Sept SPD non réglementés ont été analysés: le trichloroacétonitrile (TCAN), le
dichloroacétonitrile (DCAN), le dibromoacétonitrile (DBAN) et le bromo-chloroacétonitrile (BCAN), le
trichloronitrométhane (la chloropicrine (CPK), le 1,1-dichloro-2-propanone (DCP) et le 1,1,1-trichloro-2-
propanone (TCP). La limite de quantification de la méthode (LQM) pour tous les SPD non-règlementés analysés
a été fixée à 0,01 µg/L. Les incertitudes sur les mesures des concentrations de SPDE sont de : ± 10% pour le
TCAN, le DCAN, la DCPone et la TCPone; ± 15 % pour le DBAN et la CPK; et ± 20 % pour le BCAN. Plus de
détails sur la méthode d’analyse des SPD non-réglementés sont disponibles dans le mémoire de Maîtrise de
Catherine Mercier-Shanks (Mercier-Shanks, 2012).
SPD réglementés
L’analyse des THM a été réalisée suivant la méthode EPA 524.2 (Environmental Protection Agency, 1995b).
Les concentrations des THM ont été déterminées par microextraction sur phase solide (SPME) suivi d'une
analyse par chromatographie en phase gazeuse combinée à un spectromètre de masse à trappe ionique (GC-
MS). Le tout est effectué de façon automatisée à l'aide d'un PAL COMBI-xt et d'un GC-MS de marque Varian
(GC : Varian 3900; MS : Varian 2100T). L'extraction SPME est effectuée en mode « headspace » à l'aide d'une
fibre PDMS 100 pm (Supelco, # cat. 57341-U). Elle est effectuée dans des microvials de 2 mL contenant 800
µL d'échantillon en présence de 200 µg/L de fluorobenzene, de 4-bromofluorobenzene et de 1,2-
dichlorobenzene-d4, qui agissent à la fois à titre d'analogue d'extraction (« surrogate ») et à titre de standard
interne. Les THM étudiés sont le chloroforme (trichlorométhane, TCM), le bromodichlorométhane (BDCM), le
dibromochlorométhane (DBCM) et le bromoforme (tribromométhane, TBM). Les limites de quantification (LQM)
pour chacun des THM sont respectivement de 3,7 pg/L, 2,0 pg/L, 3,3 pg/L et 2,7 pg/L pour le TCM, le BDCM,
119
le DBCM et le TBM. Les incertitudes sur les mesures des concentrations des THM sont de ± 25 % pour les
quatre composés analysés. L’extraction des THM a été réalisée à l’intérieur dans un délai d’un mois après
l’échantillonnage. Plus de détails sur la méthode d’analyse des THM sont disponibles dans le mémoire de
Maîtrise de Catherine Mercier-Shanks (Mercier-Shanks, 2012).
La méthode d’analyse des AHA concorde avec la méthode EPA 552.2 (Environmental Protection Agency,
1995c). Les concentrations des AHA ont été déterminées par extraction liquide-liquide incluant une dérivation
pour obtenir les esters de méthyle correspondant, suivi d'une analyse par chromatographie en phase gazeuse
combinée à un détecteur à capture d'électron (GC-ECD). L'acide 2-bromopropionique a été utilisé à titre
d'analogue d'extraction (« surrogate ») et du 1,2,3-trichloropropane a été utilisé comme standard interne.
L'analyse des extraits a été effectuée à l'aide d'un GC-ECD de marque Perkin Elmer (Autosystem XL, muni d'un
détecteur ECD avec source radioactive au Ni) avec une colonne capillaire DB-1701 (30 m x 0,32 mm DI, film de
0,25 pm, J&W # cat. 123-0732). Les AHA étudiés sont l’acide monobromoacétique (AMBA), l’acide
monochloroacétique (AMCA), l’acide dibromoacétique (ADBA), l’acide dichloroacétique (ADCA) et l’acide
trichloroacétique (ATCA). La limite de quantification de la méthode (LQM) pour tous les AHA analysés a été
fixée à 1,0 pg/L. Les incertitudes sur les mesures des concentrations des AHA sont de : ± 15% pour l'AMBA,
l'AMCA et l'ADCA; ± 25 % pour l'ATCA; et ± 30 % pour l'ADBA. L’extraction des AHA a été réalisée dans un
délai de deux semaines après l’échantillonnage. Plus de détails sur la méthode d’analyse des AHA sont
disponibles dans le mémoire de Maîtrise de Catherine Mercier-Shanks (Mercier-Shanks, 2012).
L’étude des chromatogrammes de AHA, SPD non-règlementés et THM obtenus a été réalisée à l’aide du logiciel
TotalChrom de Perkin Elmer.
Paramètres physico-chimiques
Les bromures ont été analysés à l’aide de la méthode MA. 303 - Ions 3.1 du CEAEQ (Centre d’expertise en
analyse environnemental du Québec, 2009). Le COD a été mesuré à l’aide de l’analyseur de carbone de marque
Sievers 5310C de Général Electric, muni d’un passeur automatique Sievers 900. L’absorbance UV a été
mesurée à l’aide du spectrophotomètre DR5000 de la marque HACH et de la cellule de Quartz de 50 mm de la
marque Starna. La turbidité a été mesurée à l’aide du turbidimètre 2100N, de la marque HACH. Les paramètres
de de physico-chimie ont être mesurés dans les 48 à 72 heures après échantillonnage, selon le paramètre
étudié.
Note : Lors du premier chapitre, suite à la demande d’un relecteur, si la concentration mesurée est en dessous
de la limite de quantification, la concentration est considérée comme nulle. Dans les trois autres articles, si la
concentration mesurée est en dessous de la limite de quantification, la moitié de cette limite est considérée.
120
Bibliographie de l’annexe 3:
Centre d’expertise en analyse environnemental du Québec, 2009. Détermination des anions en faible
concentration dans l’eau de consommation : méthode par chromatographie ionique. Méthode MA 303 - Ions 3.1,
Révision 3. 18.
Environmental Protection Agency, USEPA, 1995. Method EPA 551.1, Determination of chlorination disinfection
byproducts, chlorinated solvents and halogenated pesticides/herbicides in drinking water by liquid-liquid
extraction and gas chromatography with electron capture detection, Revision 1.0, National Exposure Research
Laboratory, Office of Research and Development, Cincinnati, Ohio
Environmental Protection Agency, USEPA, 1995. Method EPA 524.2, Measurement of purgeable organic
compounds in water by capillary column gas chromatography/mass spectrometry, Revesion 4.0 National
Exposure Research Laboratory, Office of Research and Development, Cincinnati, Ohio
Environmental Protection Agency, USEPA, 1995. Method EPA 552.2, Determination of haloacetic acids and
dalapon in drinking water by liquid-liquid extraction, derivatization and gas cmomatography with electron capture
detection, Revision 1.0, National Exposure Research Laboratory, Office of Research and Development,
Cincinnati, Ohio
Mercier-Shanks, C., 2012. Variabilité spatio-temporelle des sous-produits de la désinfection émergents
(Haloacétonitriles, Halonitrométhane et Halocétones) dans un réseau de distribution d’eau potable, Mémoire de
Maîtrise, Université Laval
121
Annexe 4: Comparison of water characteristics
between SWS in QC and NL
Type of water
Parameters
Newfoundland and Labrador Quebec
t Meana
Coefficient of variation
Meana Coefficient of
variation
Raw water (RW)
pH (5%-95% percent.)
5.7-8.5 0.14 5.9-7.9 0.09 NS
Turbidity (NTU) 0.78 1.15 1.96 1.54 **
Conductivity (µS cm-1)
59 0.60 61 0.80 NS
UV-254 (cm-1) 0.26 0.50 0.25 0.86 NS
DOC (mg/L) 7.27 0.41 5.78 0.59 **
SUVA (L mg-1 m) 3.60 0.31 3.98 1.10 NS
Bromide (µg L-1) 12.4 0.79 3.8 0.82 **
Water before
disinfection (WTP)
pH idemb idemb 5.1-7.7 0.12 *
Turbidity (NTU) idemb idemb 0.34 0.97 **
Conductivity (µS cm-1)
idemb idemb 83 0.61 **
UV-254 (cm-1) idemb idemb 0.05 0.88 **
DOC (mg L-1) idemb idemb 2.44 0.53 **
SUVA (L mg-1 m) idemb idemb 1.93 0.54 **
Water from distribution
system (DS2)
Turbidity (NTU) 0.57 0.61 0.46 1.79 NS
Conductivity (µS cm-1)
69 0.46 89 0.57 **
UV-254 (cm-1) 0.21 0.48 0.05 0.83 **
DOC (mg L-1) 6.87 0.38 2.51 0.55 **
SUVA (L mg-1 m) 3.19 0.35 1.91 0.50 ** a: Annual average on thirteen measures b: Same values as raw water *: Significantly different at the 5% level of significance according to Student’s t-test with SYSTAT ** : Significantly different at the 1% level of significance according to Student’s t-test with SYSTAT
122
Annexe 5: Spatial evolution of regulated DBP
concentrations in SWS in a) NL and b) QC
a)
b)
123
Annexe 6: Spatial evolution of non-regulated DBP
concentrations in SWS in a) NL and b) QC
a)
b)
124
Annexe 7: Distribution of daily levels (average of
the six SWS under study in DS3) of a) free chlorine,
b) THMs, c) HAAs, d) HANs, e) CPK and f) HKs
a)
b)
c)
d)
e)
f)
Notes: Error bars represent the 5th and 95th percentiles. In each box, the 25th percentile, the median and the 75th
percentiles are represented. Mean values are represented by black points.
0
0,5
1
1,5
2
Fre
e C
hlo
rine level (m
g/L
)
0
50
100
150
200
250
300
350
TH
M level (µ
g/L
)
0
50
100
150
200
250
300
350
HA
A level (µ
g/L
)
0
2
4
6
8
10
HA
N level (µ
g/L
)
0
0,4
0,8
1,2
1,6
CP
K level (µ
g/L
)
0
5
10
15
20
HK
level (µ
g/L
)
125
Annexe 8: Evolution of average DBP and free
residual chlorine levels along the DS in a) NL1, b)
NL2, c) NL3, d) QC1, e) QC2 and f) QC3 (average of
5 measurements for NL1 and 12 for the others)
a)
Note: TCAA and THM decreases between DS5 and DS6 could be explain by a mix of water appearing in DS6 from a close
junction with pipeline containing a water with a shorter residence time (indicating by a free chlorine level increase).
0
0,2
0,4
0,6
0,8
1
0
20
40
60
80
100
120
DS1 DS2 DS3 DS4 DS5 DS6
Fre
e C
hlo
rin
e level in
NL
1
(mg
/L)
DB
P level in
NL
1 (
µg
/L)
HAAs THMs Free Chlorine
0,0
2,0
4,0
6,0
8,0
DS1 DS2 DS3 DS4 DS5 DS6
Le
vel in
NL
1 (
µg
/L)
DCP TCP
0,00
0,05
0,10
0,15
0,20
0,25
0,0
1,0
2,0
3,0
4,0
DS1 DS2 DS3 DS4 DS5 DS6
BC
AN
level in
NL
1 (
µg
/L)
DC
AN
level in
NL
1 (
µg
/L)
DCAN BCAN
0
5
10
15
20
25
30
35
40
45
50
DS1 DS2 DS3 DS4 DS5 DS6
HA
A l
evel in
NL
1 (
µg
/L)
TCAA DCAA
126
b)
Note: DS1 to DS5 are all located on the same main water pipe, DS6 is located in a junction of this pipe (junction starting at
DS2). Also, between DS4 and DS5, a water pipe goes through a drinking water storage tank before returning in the main
water pipeline around DS5. Volatilization of chloroform, the most volatile THMs, could happen in the water storage tank. In
fact, the water storage tank is located at a high height, on the hoof, so water need to be pump to arrive in the tank. Thus,
swirls might be observable during the tank filling, inducing chloroform volatilization. This could induce a THM level decrease
from DS4. Also, in DS6, the difference in THM levels can be explained that the fact that DS6 is located on a junction of the
main water pipeline (junction starting at DS2).
c)
0,0
0,5
1,0
1,5
2,0
2,5
0
50
100
150
200
DS1 DS2 DS3 DS4 DS5 DS6
Fre
e C
hlo
rin
e level in
NL
2
(mg
/L)
DB
P level in
NL
2 (
µg
/L)
HAAs THMs Free Chlorine
0,0
4,0
8,0
12,0
DS1 DS2 DS3 DS4 DS5 DS6
Le
vel in
NL
2 (
µg
/L)
DCP TCP
0,00
0,04
0,08
0,12
0,16
0,0
1,0
2,0
3,0
4,0
DS1 DS2 DS3 DS4 DS5 DS6
BC
AN
level in
NL
2 (
µg
/L)
DC
AN
level in
NL
2 (
µg
/L)
DCAN BCAN
0
20
40
60
80
100
DS1 DS2 DS3 DS4 DS5 DS6
HA
A l
evel in
NL
2 (
µg
/L)
TCAA DCAA
0,0
1,0
2,0
3,0
4,0
0
100
200
300
400
DS1 DS2 DS3 DS4 DS5 DS6
Fre
e C
hlo
rin
e level i
n N
L3
(m
g/L
)
DB
P level in
NL
3 (
µg
/L)
HAAs THMs Free Chlorine
0,0
4,0
8,0
12,0
DS1 DS2 DS3 DS4 DS5 DS6
Le
vel in
NL
3 (
µg
/L)
DCP TCP
127
d)
0,00
0,10
0,20
0,30
0,40
0,0
3,0
6,0
9,0
DS1 DS2 DS3 DS4 DS5 DS6
BC
AN
level in
NL
3 (
µg
/L)
DC
AN
level in
NL
3 (
µg
/L)
DCAN BCAN
0
40
80
120
160
DS1 DS2 DS3 DS4 DS5 DS6
HA
A l
evel in
NL
3 (
µg
/L)
TCAA DCAA
0,0
0,2
0,4
0,6
0,8
1,0
0
30
60
90
DS1 DS2 DS3 DS4 DS5 DS6
Fre
e c
hlo
rin
e level in
QC
1(m
g/L
)
DB
P level in
QC
1 (
µg
/L)
HAAs THMs Free Chlorine
0,0
1,0
2,0
3,0
4,0
DS1 DS2 DS3 DS4 DS5 DS6
Le
vel in
QC
1 (
µg
/L)
DCP TCP
0,00
0,10
0,20
0,30
0,40
0,0
1,0
2,0
3,0
4,0
DS1 DS2 DS3 DS4 DS5 DS6
BC
AN
level in
QC
1 (
µg
/L)
DC
AN
level in
QC
1 (
µg
/L)
DCAN BCAN
0
5
10
15
20
25
30
DS1 DS2 DS3 DS4 DS5 DS6
HA
A l
evel in
QC
1 (
µg
/L)
TCAA DCAA
128
e)
0,0
0,3
0,6
0,9
0
30
60
90
120
DS1 DS2 DS3 DS4 DS5 DS6
Fre
e C
hlo
rin
e level in
QC
2
(mg
/L)
DB
P level in
QC
2 (
µg
/L)
HAAs THMs Free Chlorine
0,0
2,0
4,0
6,0
DS1 DS2 DS3 DS4 DS5 DS6
Le
vel in
QC
2 (
µg
/L)
DCP TCP
0,00
0,04
0,08
0,12
0,16
0,0
1,0
2,0
3,0
4,0
5,0
DS1 DS2 DS3 DS4 DS5 DS6
BC
AN
level in
QC
2 (
µg
/L)
DC
AN
level in
QC
2 (
µg
/L)
DCAN BCAN
0
5
10
15
20
25
30
35
DS1 DS2 DS3 DS4 DS5 DS6
HA
A l
evel in
QC
2 (
µg
/L)
TCAA DCAA
129
f)
0,0
1,0
2,0
3,0
0
40
80
120
160
DS1 DS2 DS3 DS4 DS5 DS6
Fre
e C
hlo
rin
e level in
QC
3(m
g/L
)
DB
P level in
QC
3 (
µg
/L)
HAAs THMs Free Chlorine
0,0
3,0
6,0
9,0
DS1 DS2 DS3 DS4 DS5 DS6
Le
vel in
QC
3 (
µg
/L)
DCP TCP
0,00
0,04
0,08
0,12
0,0
1,0
2,0
3,0
4,0
DS1 DS2 DS3 DS4 DS5 DS6
BC
AN
level in
QC
3 (
µg
/L)
DC
AN
level in
QC
3 (
µg
/L)
DCAN BCAN
0
30
60
90
DS1 DS2 DS3 DS4 DS5 DS6
HA
A l
evel in
QC
3 (
µg
/L)
TCAA DCAA
130
Annexe 9: Summary of annual average levels of a)
THMs, b) HAAs, c) HANs, c) CPK, and e) HKs,
based on the various scenarios
a)
THMs Reference scenario
Scenario 1
Scenario 2
Scenario 3
Scenario 4
Scenario 5
NL1 149 119 134 134 130 139
NL2 114 79.8 99.8 88.5 95.7 111
NL3 151 107 82.4 82.4 92.2 81.0
NL4 153 94.4 109 108 112 122
NL5 334 127 236 252 232 322
NL6 121 76.8 88.4 50.2 77.9 103
NL7 169 54.9 138 87.2 119 152
NL8 135 103 97.8 123 113 115
NL9 177 91.6 122 145 138 165
NL10 297 309 297 262 257 281
NL11 193 162 141 171 160 130
QC1 78.7 68.0 52.9 55.6 61.9 82.2
QC2 35.5 37.3 34.8 27.7 31.7 39.0
QC3 38.2 9.33 18.5 30.2 25.3 32.6
QC4 73.3 44.6 64.3 64.2 63.6 79.4
QC5 70.1 49.5 50.2 37.1 53.4 68.5
QC6 90.6 34.9 44.4 104 76.5 88.9
QC7 41.1 34.0 36.0 32.9 35.7 44.4
QC8 79.7 64.1 71.6 60.2 64.7 67.8
QC9 30.4 NA 17.6 5.8 14.0 5.82
QC10 15.5 11.7 10.7 9.6 11.4 12.2
QC11 90.9 38.5 66.6 58.6 59.2 57.6
QC12 58.5 42.7 37.3 60.4 46.0 54.9
QC13 92.4 72.9 74.4 71.6 65.0 69.0
QC14 48.5 NA 39.2 36.3 37.9 44.5
NA: Data not available
131
b)
HAAs Reference scenario
Scenario 2
Scenario 3
Scenario 4
Scenario 5
NL1 172.64 116.45 116.45 117.65 106.55
NL2 167.37 116.55 114.15 103.52 105.96
NL3 250.59 197.93 197.93 153.92 198.03
NL4 85.38 112.88 137.60 127.82 112.86
NL5 237.33 204.13 138.28 159.64 118.03
NL6 104.74 81.70 37.20 69.77 79.75
NL7 235.81 123.10 117.63 119.70 122.56
NL8 162.78 131.63 157.55 132.18 134.77
NL9 155.30 117.65 119.00 91.84 107.81
NL10 299.90 388.50 363.53 304.77 412.22
NL11 216.06 128.68 152.63 148.13 128.89
QC1 61.39 52.28 47.20 49.97 59.09
QC2 22.33 15.28 15.95 17.86 16.20
QC3 37.98 17.75 23.85 26.37 33.04
QC4 70.42 59.95 62.50 56.87 79.96
QC5 55.34 36.23 51.90 48.61 63.09
QC6 93.69 38.13 80.73 51.18 76.41
QC7 31.62 22.18 31.08 23.50 24.42
QC8 49.59 35.40 49.63 43.10 50.27
QC9 23.03 18.55 9.50 16.44 10.03
QC10 18.46 7.60 8.38 9.93 11.10
QC11 52.57 46.23 43.38 40.17 51.90
QC12 43.83 29.80 53.10 39.43 53.06
QC13 76.53 138.65 132.63 123.44 133.58
QC14 28.32 23.23 23.50 23.46 28.22
132
c)
HANs Reference scenario
Scenario 2
Scenario 3
Scenario 4
Scenario 5
NL1 7.23 4.53 4.53 4.26 5.26
NL2 3.54 2.08 2.43 2.04 2.76
NL3 6.08 3.39 3.39 4.16 6.34
NL4 1.33 1.79 1.47 1.61 2.21
NL5 3.86 2.40 3.93 2.44 3.81
NL6 2.01 1.70 0.94 1.34 1.61
NL7 2.04 1.30 1.15 1.16 1.32
NL8 5.68 4.46 4.61 4.60 4.48
NL9 5.93 3.73 4.23 3.37 4.56
NL10 6.74 5.56 6.39 4.94 4.85
NL11 4.23 3.34 3.11 3.43 4.33
QC1 4.08 3.63 3.11 3.38 4.15
QC2 1.26 0.96 0.94 1.03 1.04
QC3 1.79 1.15 1.52 1.43 1.56
QC4 5.45 4.32 4.44 4.39 4.73
QC5 3.44 2.05 2.40 2.43 2.27
QC6 3.44 0.83 1.96 1.23 1.74
QC7 1.78 1.73 1.97 1.51 1.30
QC8 3.73 2.46 3.74 3.07 3.48
QC9 1.38 0.91 0.46 0.82 0.93
QC10 1.20 0.83 0.82 0.91 1.01
QC11 4.51 3.45 3.50 3.24 3.36
QC12 2.73 1.57 2.10 1.59 2.11
QC13 2.34 2.76 3.11 2.90 3.03
QC14 1.80 1.57 1.48 1.56 1.63
133
d)
CPK Reference scenario
Scenario 2
Scenario 3
Scenario 4
Scenario 5
NL1 0.88 0.69 0.69 0.69 0.65
NL2 0.54 0.38 0.33 0.36 0.33
NL3 0.90 0.83 0.83 0.57 0.83
NL4 0.14 0.25 0.25 0.25 0.29
NL5 0.26 0.22 0.33 0.21 0.28
NL6 0.84 0.48 0.27 0.44 0.79
NL7 0.45 0.27 0.27 0.27 0.21
NL8 0.38 0.30 0.35 0.33 0.30
NL9 1.25 0.99 1.03 0.96 0.83
NL10 0.64 0.75 0.80 0.67 0.73
NL11 0.66 0.57 0.58 0.59 0.54
QC1 0.71 0.60 0.43 0.55 0.78
QC2 0.14 0.13 0.13 0.14 0.18
QC3 0.30 0.17 0.22 0.19 0.19
QC4 0.71 0.53 0.55 0.57 0.75
QC5 0.33 0.23 0.25 0.26 0.26
QC6 0.33 0.16 0.24 0.19 0.21
QC7 0.43 0.37 0.44 0.37 0.34
QC8 1.46 1.08 1.22 1.36 1.54
QC9 0.24 0.15 0.09 0.13 0.22
QC10 0.38 0.26 0.25 0.29 0.31
QC11 0.43 0.28 0.32 0.28 0.24
QC12 0.42 0.31 0.45 0.32 0.31
QC13 0.35 0.59 0.70 0.66 0.68
QC14 0.26 0.21 0.20 0.21 0.22
134
e)
HKs Reference scenario
Scenario 2
Scenario 3
Scenario 4
Scenario 5
NL1 9.12 7.45 7.45 6.87 6.66
NL2 9.47 7.59 7.39 7.56 7.63
NL3 13.03 10.30 10.30 9.16 11.76
NL4 10.08 12.17 11.18 11.72 12.43
NL5 18.57 16.02 21.86 17.04 21.00
NL6 4.09 2.45 1.72 2.62 3.36
NL7 12.54 8.34 8.74 8.27 8.83
NL8 9.06 6.67 7.49 7.49 7.91
NL9 10.48 7.31 7.77 7.16 8.22
NL10 17.92 11.26 12.12 13.85 12.31
NL11 13.73 10.95 10.47 11.37 11.65
QC1 5.31 4.49 3.72 3.95 5.21
QC2 3.24 1.97 1.66 1.83 2.19
QC3 3.25 1.91 2.60 2.36 2.73
QC4 5.50 3.84 4.14 4.03 4.53
QC5 6.28 3.80 4.37 4.56 5.29
QC6 6.80 3.63 5.11 4.17 4.79
QC7 6.19 5.87 6.20 5.31 5.42
QC8 6.00 3.59 5.34 4.23 5.82
QC9 3.42 2.65 2.03 2.56 2.34
QC10 1.69 1.06 1.20 1.18 1.11
QC11 8.56 5.43 6.00 5.25 5.98
QC12 4.69 2.57 2.13 2.12 2.31
QC13 8.03 9.16 9.86 9.86 10.64
QC14 2.55 2.33 1.98 2.25 2.12
135
Annexe 10: Exemple de rapport de synthèse de
données envoyé aux petits réseaux
RAPPORT SUR LA QUALITÉ DE L’EAU RÉSUMÉ DES TREIZE CAMPAGNES D’ÉCHANTILLONNAGE DANS
QC 1
OCTOBRE 2010 - OCTOBRE 2011
RAPPORT RÉALISÉ PAR STÉPHANIE GUILHERME ET ANNA SCHEILI,
ÉTUDIANTES AU DOCTORAT À L’UNIVERSITÉ LAVAL
JUIN 2012
Notre laboratoire n’est pas accrédité. Nos données sont uniquement à titre informatif.
136
Introduction:
Cette étude fait partie des activités de recherche du regroupement canadien RESEAU-WaterNet duquel fait
partie la Chaire de recherche en eau potable de l’Université Laval.
Notre étude s’est intéressée à la qualité de l’eau dans les réseaux de distribution d’eau potable des petites
municipalités, et en particulier à la présence de certains sous-produits de la désinfection (SPD), comme les
trihalométhanes (THM), les acides haloacétiques (AHA) et d’autres SPD encore non règlementés au Canada.
Les SPD se forment lorsque le désinfectant réagit avec la matière organique (MO) naturelle et/ou avec d’autres
substances inorganiques naturellement présentes dans l’eau.
Afin de caractériser votre eau, des échantillons ont été prélevés à la source, en cours de traitement et dans le
réseau de distribution. Les principales mesures concernent les caractéristiques chimiques de votre eau, mais
d’autres mesures physiques et microbiologiques ont également été effectuées.
En effet, il est nécessaire de considérer la température de l’eau, les doses de chlore, la quantité de matière
organique (MO) naturelle et le pH, car ces paramètres sont liés à la formation des SPD dans le réseau.
Les résultats présentés dans ce rapport sont uniquement à titre informatif. Nous vous rappelons que notre
laboratoire n’est pas accrédité.
137
RÉSULTATS DES ANALYSES
Des échantillons d’eau ont été prélevés mensuellement entre octobre 2010 et octobre 2011. Les points
d'échantillonnage nommés EB et ET représentent respectivement l’eau brute et l’eau traitée (prélevée après
l’étape de filtration et avant l’étape de chloration, directement à l’usine de filtration située au 325 de la route
108). Les points d'échantillonnage appelés R1, R2 et R3 correspondent à l'eau traitée au début, au milieu et à
la fin du réseau. Les points d'échantillonnage sont: R1 au poste de rechloration, R2 à l’Hôtel de Ville et R3 à
l’usine d’épuration.
I. CARACTÉRISTIQUES DE L’EAU BRUTE
En ce qui concerne l'eau brute, plusieurs paramètres physico-chimiques ont été mesurés tels que la
température, le pH, la conductivité, la turbidité, la MO, l’absorbance UV-254 (à 254 nm). De plus, certains
paramètres microbiologiques ont été également analysés: Coliformes totaux, E. Coli, entérocoques et des
bactéries hétérotrophes aérobies et anaérobies facultatives (BHAA).
La conductivité donne des informations sur la quantité d’ions présents dans l'eau, considérés comme
précurseurs des SPD.
La MO est l’ensemble des composés organiques présents dans l’eau brute (acides humiques et fulviques, par
exemple). Elle est un précurseur principal des SPD.
Toutefois, la MO ne peut pas être mesurée directement; d’autres paramètres s’y rapportant doivent être évalués
tels que le carbone organique dissous (COD), l’absorbance UV-254 et la chlorophylle.
La turbidité mesure la quantité de matières en suspension constituées de particules de différentes tailles
généralement invisibles à l'œil. La turbidité est habituellement le résultat de fines particules organiques et
inorganiques qui ne décantent pas.
Le COD représente, le carbone organique dissous présent dans l'eau, le CI le carbone inorganique et le CT, le
carbone total.
L’absorbance UV-254 est un indicateur de substances organiques ayant au moins un groupe aromatique dans
leur structure chimique.
La chlorophylle est utilisée comme indicateur de biomasse phytoplanctonique présente dans les eaux brutes.
Les bromures sont liés à la formation des SPD, car ils sont des précurseurs de SPD bromés.
Pour caractériser la qualité microbiologique de l’eau potable, la méthode du Nombre le Plus Probable (NPP) a
été utilisée. Celle-ci détermine le nombre le plus probable de colonies d’organismes (provoquant des résultats
positifs) par unité de volume de l'échantillon original. Lors des campagnes, un volume de 100 ml a été considéré
pour les bactéries coliformes, la bactérie E. coli, et les entérocoques; et 10 ml pour les bactéries hétérotrophes.
Les coliformes totaux sont utilisés comme indicateurs de la qualité microbienne de l'eau car ils sont
indirectement liés à une pollution d'origine fécale. Pratiquement aucune espèce de coliformes n’est pathogène.
Ils ne représentent donc pas de risque direct pour la santé, à l'exception d'Escherichia Coli (E. Coli), ainsi que
de rares bactéries pathogènes opportunistes (Groupe de recherche scientifique sur l'eau (2003), Coliformes
totaux dans Fiches synthèses sur l'eau potable et la santé humaine, Institut national de santé publique du
Québec, 4 p.). C'est pourquoi, la présence de colonies d’E. Coli a aussi été analysée. Les entérocoques sont
une famille de microorganismes dont la résistance contre les agents désinfectants est nettement supérieure à
celle d’autres indicateurs. Les entérocoques sont analysés afin d’évaluer l’efficacité du traitement.
138
Les résultats concernant les bactéries hétérotrophes donnent une indication sur la qualité globale de l'eau mais
ne sont pas des indicateurs de la salubrité de l'eau. Le niveau de la flore hétérotrophe est estimé avec une
méthode microbiologique qui utilise la formation de colonies bactériennes sur les milieux de culture. Cette
méthode ne donne pas d’indication sur les types d'organismes présents ou de leurs sources. Cependant, la
majorité d'entre eux ne sont pas pathogènes. Les BHAA sont analysées uniquement dans l’eau du réseau.
Tableau 1: Résumé des mesures physico-chimiques de l’eau brute
Date Campag
ne
Point d’échantillonna
ge pH
Temp
(°C)
Turbidité
(UTN)
Conductivité (µS/cm)
UV-254 (cm-
1)
COD
(ppm)
CI (ppm)
CT (ppm
)
Chlorophylle (µg/l)
Bromure (µg/l)
20/10/2010
1 QC1-EB 7,79
10,5 1,06 76 0,344
0
11,7
0 0,02
11,7
0
Pas de donnée
s
Pas de donnée
s 5,4
16/11/2010
2 QC1-EB 8,66
12,8 1,57 72 0,289
3 8,31 4,73
13,0
5 1,09 0,39 5,2
07/12/2010
3 QC1-EB 7,12
1,5 1,08 76 0,240
8 3,98 2,43 6,41 4,59 0,48 4,9
06/01/2011
4 QC1-EB 7,15
2,9 2,63 92 0,243
0 6,09 5,67
11,8
0 1,67 0,54 <4
15/02/2011
5 QC1-EB 7,11
0,0 1,28 99 0,171
0 4,09 5,79 9,88 0,64 0,74 <4
09/03/2011
6 QC1-EB 6,93
2,6 2,13 94 0,251
6 6,81 5,29
12,1
0 0,21 0,26 4,4
06/04/2011
7 QC1-EB 7,25
0,0 2,08 134 0,227
4 3,88 3,92 7,79 0,39 0,27 7,8
11/05/2011
8 QC1-EB 6,87
8,0 1,92 58 0,261
8 5,72 3,52 9,24 1,21 0,40 <4
08/06/2011
9 QC1-EB 7,33
16,0 2,37 82 0,256
0 7,42 5,12
12,5
0 1,95 1,39 5,6
04/07/2011
10 QC1-EB 6,97
19,0 2,07 96 0,270
0 7,13 7,10
14,2
0 1,27 1,28 <4
08/08/2011
11 QC1-EB 7,10
20,0 1,57 122 0,223
2 5,40 8,70
14,1
0 0,76 1,08 9,7
06/09/2011
12 QC1-EB 6,68
18,0 20,20 52 0,702
0
15,7
5 3,64
19,3
5 1,34 1,47 11,0
12/10/2011
13 QC1-EB 6,48
12,0 0,82 102 0,309
4 5,89 5,38
11,3
0 0,97 0,80 6,1
Les valeurs <4 indique des concentrations de bromure inférieure à la limite de détection (4 µg/l)
Note : Lors de la campagne 1, la chlorophylle n’a pas pu être mesurée pour des raisons analytiques.
139
Tableau 2: Résumé des mesures microbiologiques de l’eau brute
Date Campagne Point
d’échantillonnage Coliformes totaux
(UFC/100ml) E. Coli
(UFC/100ml) Entérocoques (UFC/100ml)
20/10/2010 1 QC1-EB 816,4 16,1 18,9
16/11/2010 2 QC1-EB 285,1 33,6 4,1
07/12/2010 3 QC1-EB 307,6 31,3 2,0
07/01/2011 4 QC1-EB 42,6 8,5 1,0
15/02/2011 5 QC1-EB 59,4 15,8 3,1
09/03/2011 6 QC1-EB 866,4 62,7 7,5
06/04/2011 7 QC1-EB 866,4 61,6 35,9
11/05/2011 8 QC1-EB 145,0 9,8 <1
08/06/2011 9 QC1-EB 980,4 16,1 2,0
04/07/2011 10 QC1-EB 2359,0 115,0 58,2
08/08/2011 11 QC1-EB 1203,3 21,6 66,3
06/09/2011 12 QC1-EB *12997 *1203,3 1119,9
12/10/2011 13 QC1-EB 980,4 26,5 8,5
Une concentration de <1 indique qu’aucune colonie bactérienne n’a été observée selon la limite de détection d’une colonie bactérienne
* Échantillons dilués
II. CARACTÉRISTIQUES DE L’EAU TRAITÉE
Les mêmes caractéristiques physico-chimiques ont été mesurées dans l’eau traitée après sa filtration et avant sa chloration. Tableau 3: Résumé des mesures physico-chimiques de l’eau traitée
Date Campagne Point
d’échantillonnage pH
Temp (°C)
Turbidité (UTN)
Conductivité (µS/cm)
UV-254 (cm-1)
COD (ppm)
CI (ppm)
CT (ppm)
20/10/2010 1 QC1-ET 6,19 10,4 0,289 90 0,0612 3,48 0,01 3,49
16/11/2010 2 QC1-ET 6,42 10,3 0,290 83 0,0430 Pas de
données Pas de
données Pas de
données
07/12/2010 3 QC1-ET 7,05 4,0 0,065 93 0,0378 2,21 2,80 5,01
06/01/2011 4 QC1-ET 6,79 4,1 0,213 107 0,0390 2,01 3,82 5,83
15/02/2011 5 QC1-ET 6,52 0,5 0,236 114 0,0348 1,38 3,50 4,87
09/03/2011 6 QC1-ET 6,64 3,4 1,670 112 0,0374 1,93 3,79 5,72
06/04/2011 7 QC1-ET 6,76 2,0 0,071 156 0,0294 0,99 3,15 4,15
11/05/2011 8 QC1-ET 6,27 10,0 0,203 78 0,0336 2,79 2,15 4,94
08/06/2011 9 QC1-ET 6,93 17,5 0,115 105 0,0398 3,08 3,80 6,88
04/07/2011 10 QC1-ET 6,49 20,0 0,138 110 0,0468 3,16 3,44 6,60
08/08/2011 11 QC1-ET 6,47 20,0 0,060 136 0,0540 2,93 6,03 8,95
06/09/2011 12 QC1-ET 4,83 22,0 0,327 80 0,1184 5,13 1,19 6,32
12/10/2011 13 QC1-ET 6,22 14,0 0,223 119 0,0498 2,60 3,55 6,15
Lors de la campagne 2, certains paramètres physico-chimiques n’ont pas pu être mesurés pour des raisons analytiques.
140
III. QUALITÉ DE L’EAU DANS LE RÉSEAU
Tableau 4: Résumé des mesures physico-chimiques dans le réseau
Date Campagne Point
d’échantillonnage
Chlore Libre
(mg/L)
Chlore Total
(mg/L)
Turbidité (UTN)
Conductivité (µS/cm)
UV-254 (cm-1)
COD (ppm)
CI (ppm)
CT (ppm)
20/10/2010 1
QC1-R1 1,34 1,57
QC1-R2 0,96 1,21 0,569 106 0,0580 4,05 0,015 4,06
QC1-R3 0,72 0,95
16/11/2010 2
QC1-R1 1,15 1,54
QC1-R2 1,03 1,22 0,578 101 0,0434 3,06 3,21 6,27
QC1-R3 0,80 0,98
07/12/2010 3
QC1-R1 1,01 1,21
QC1-R2 0,87 1,11 0,164 106 0,0340 2,19 3,40 5,58
QC1-R3 0,70 0,80
06/01/2011 4
QC1-R1 0,72 1,09
QC1-R2 0,77 0,91 0,156 125 0,0320 1,74 3,54 5,28
QC1-R3 0,70 0,87
15/02/2011 5
QC1-R1 0,90 1,18
QC1-R2 0,80 1,04 0,140 126 0,0282 1,70 4,76 6,45
QC1-R3 0,75 0,97
09/03/2011 6
QC1-R1 1,10 1,32
QC1-R2 0,96 1,30 0,142 146 0,0278 1,68 4,78 6,45
QC1-R3 0,83 1,06
06/04/2011 7
QC1-R1 0,95 1,17
QC1-R2 0,83 0,95 0,121 167 0,0256 0,988 3,39 4,38
QC1-R3 0,56 0,57
11/05/2011 8
QC1-R1 0,90 1,02
QC1-R2 0,73 0,84 0,167 92 0,0324 2,37 2,82 5,19
QC1-R3 0,65 0,75
08/06/2011 9
QC1-R1 1,10 1,25
QC1-R2 0,75 0,91 0,161 119 0,0242 3,05 4,18 7,23
QC1-R3 0,59 0,71
04/07/2011 10
QC1-R1 0,89 1,08
QC1-R2 0,63 0,74 0,257 123 0,0360 2,99 4,09 7,08
QC1-R3 0,42 0,54
08/08/2011 11
QC1-R1 0,88 1,06
QC1-R2 0,55 0,70 0,180 145 0,0454 3,82 6,52 10,35
QC1-R3 0,33 0,52
06/09/2011 12 QC1-R1 0,51 0,83
QC1-R2 0,33 0,53 0,569 135 0,0702 3,77 3,57 7,34
141
QC1-R3 0,28 0,48
12/10/2011 13
QC1-R1 0,98 1,21
QC1-R2 0,75 0,96 0,322 148 0,0428 2,44 4,94 7,38
QC1-R3 0,33 0,51
Table 5: Résumé des mesures microbiologiques dans le réseau
Date Campagne Point
d’échantillonnage
Coliformes totaux
(UFC/100ml)
E. Coli (UFC/100ml)
Entérocoques (UFC/100ml)
Bactéries Hétérotrophes
(UFC/1ml)
20/10/2010 1 QC1-R2 <1 <1 <1 4,5
16/11/2010 2 QC1-R2 <1 <1 <1 2,0
07/12/2010 3 QC1-R2 <1 <1 <1 <2
07/01/2011 4 QC1-R2 <1 <1 <1 <0,2
15/02/2011 5 QC1-R2 <1 <1 <1 <2
09/03/2011 6 QC1-R2 <1 <1 <1 Pas de données
06/04/2011 7 QC1-R2 <1 <1 <1 <2
11/05/2011 8 QC1-R2 <1 <1 <1 <2
08/06/2011 9 QC1-R2 <1 <1 <1 <2
04/07/2011 10 QC1-R2 <1 <1 <1 <2
08/08/2011 11 QC1-R2 <1 <1 <1 <2
06/09/2011 12 QC1-R2 <1 <1 <1 <2
12/10/2011 13 QC1-R2 <1 <1 <1 <2
Une concentration de <1 ou <2 indique qu’aucune colonie bactérienne n’a été observée selon la limite de détection de deux colonies bactériennes pour les bactéries hétérotrophes et d’une colonie bactérienne pour les autres types. Une concentration de
<0,2 indique qu’aucune colonie bactérienne n’a été observé sur un échantillon dilué selon une limite de détection équivalente à 0,2 colonie bactérienne par ml.
Note : Lors de la campagne 6, les BHAA n’ont pas pu être mesurées pour des raisons analytiques.
IV. OCCURRENCE DES SOUS-PRODUITS DE LA DÉSINFECTION (SPD) 1. Trihalométhanes (THM)
L’indicateur THM4 représente la somme de quatre composés appartenant à la famille des trihalométhanes
(trichlorométhane (Chloroforme), bromo-dichlorométhane (BDCM), dibromo-chlorométhane (DBCM) et
tribromométhane (TBM)). Le tableau suivant résume les concentrations en THM4 mesurées dans le réseau sur
une période de treize mois.
Tableau 6: Résumé des mesures de THM dans le réseau
Date Campagne Point
d’échantillonnage Chloroforme
(µg/l) BDCM (µg/l)
DBCM (µg/l)
TBM (µg/l)
THM4 (µg/l)
Moyenne THM4 (µg/l)
20/10/2010 1
QC1-R1 56,31 <2 <3,32 <2,66 60,30
67,07 QC1-R2 63,11 <2 <3,32 <2,66 67,10
QC1-R3 69,82 <2 <3,32 <2,66 73,81
142
16/11/2010 2
QC1-R1 44,06 <2 <3,32 <2,66 48,05
55,85 QC1-R2 51,90 <2 <3,32 <2,66 55,89
QC1-R3 59,62 <2 <3,32 <2,66 63,61
07/12/2010 3
QC1-R1 27,88 <2 <3,32 <2,66 31,87
36,57 QC1-R2 20,98 <2 <3,32 <2,66 24,97
QC1-R3 48,90 <2 <3,32 <2,66 52,89
06/01/2011 4
QC1-R1 30,92 2,49 <3,32 <2,66 36,40
39,97 QC1-R2 38,72 2,13 <3,32 <2,66 43,84
QC1-R3 35,68 <2 <3,32 <2,66 39,67
15/02/2011 5
QC1-R1 24,96 <2 <3,32 <2,66 28,95
31,29 QC1-R2 25,02 <2 <3,32 <2,66 29,01
QC1-R3 30,87 2,05 <3,32 <2,66 35,91
09/03/2011 6
QC1-R1 18,98 <2 <3,32 <2,66 22,97
25,05 QC1-R2 20,63 <2 <3,32 <2,66 24,62
QC1-R3 23,56 <2 <3,32 <2,66 27,55
06/04/2011 7
QC1-R1 19,29 <2 <3,32 <2,66 23,28
26,63 QC1-R2 21,25 <2 <3,32 <2,66 25,24
QC1-R3 27,39 <2 <3,32 <2,66 31,38
11/05/2011 8
QC1-R1 35,36 <2 <3,32 <2,66 39,35
38,16 QC1-R2 33,85 <2 <3,32 <2,66 37,84
QC1-R3 33,31 <2 <3,32 <2,66 37,30
08/06/2011 9
QC1-R1 43,92 2,10 <3,32 <2,66 47,91
54,30 QC1-R2 47,96 <2 <3,32 <2,66 53,05
QC1-R3 57,95 3,01 <3,32 <2,66 61,94
04/07/2011 10
QC1-R1 51,27 <2 <3,32 <2,66 57,27
58,85 QC1-R2 54,44 <2 <3,32 <2,66 58,43
QC1-R3 56,87 4,95 <3,32 <2,66 60,86
08/08/2011 11
QC1-R1 78,87 4,81 <3,32 <2,66 86,81
98,08 QC1-R2 95,24 5,09 <3,32 <2,66 103,04
QC1-R3 96,31 4,02 <3,32 <2,66 104,39
06/09/2011 12
QC1-R1 114,40 3,91 <3,32 <2,66 121,41
139,83 QC1-R2 137,96 4,74 <3,32 <2,66 144,87
QC1-R3 145,47 2,89 <3,32 <2,66 153,20
12/10/2011 13
QC1-R1 64,62 2,50 <3,32 <2,66 70,50
65,94 QC1-R2 69,52 <2 <3,32 <2,66 75,01
QC1-R3 48,31 3,23 <3,32 <2,66 52,30
Une concentration de <2, <3,32 ou <2,66 indique que la valeur observée est inférieure à la limite de détection correspondante à chaque composé
143
2. Acides haloacétiques (AHA)
L’indicateur AHA5 représente la somme de cinq composés appartenant à la famille des acides haloacétiques
(acide monochloroacétique (AMCA), acide dichloroacétique (ADCA), acide trichloroacétique (ATCA), acide
monobromoacétique (AMBA) et l’acide dibromoacétique (ADBA)). Le tableau suivant résume les concentrations
en AHA5 mesurées dans le réseau sur une période de treize mois.
Tableau 7: Résumé des mesures d’AHA dans le réseau
Date Campagne Point
d’échantillonnage AMCA (µg/l)
ADCA (µg/l)
ATCA (µg/l)
AMBA (µg/l)
ADBA (µg/l)
AHA5 (µg/l)
Moyenne AHA5 (µg/l)
20/10/2010 1
QC1-R1 <1 34,77 52,07 <1 <1 88,34
98,04 QC1-R2 <1 36,46 54,85 <1 <1 92,81
QC1-R3 <1 42,17 69,30 <1 <1 112,97
16/11/2010 2
QC1-R1 <1 20,58 22,83 <1 <1 44,91
52,98 QC1-R2 <1 28,56 30,66 <1 <1 60,72
QC1-R3 <1 26,59 25,22 <1 <1 53,31
07/12/2010 3
QC1-R1 <1 13,62 15,24 <1 <1 30,36
25,71 QC1-R2 <1 14,10 17,32 <1 <1 32,92
QC1-R3 <1 3,37 8,99 <1 <1 13,86
06/01/2011 4
QC1-R1 <1 11,33 15,40 <1 <1 28,23
28,76 QC1-R2 <1 11,23 15,50 <1 <1 28,23
QC1-R3 <1 11,89 16,43 <1 <1 29,82
15/02/2011 5
QC1-R1 <1 8,16 13,18 <1 <1 22,84
25,74 QC1-R2 <1 10,02 17,01 <1 <1 28,53
QC1-R3 <1 8,98 15,38 <1 <1 25,86
09/03/2011 6
QC1-R1 1,18 6,71 8,77 <1 <1 17,66
19,19 QC1-R2 <1 7,34 9,38 <1 <1 18,22
QC1-R3 2,08 8,04 10,57 <1 <1 21,69
06/04/2011 7
QC1-R1 1,40 8,75 10,53 <1 <1 21,68
20,30 QC1-R2 1,08 8,37 9,42 <1 <1 19,87
QC1-R3 <1 7,44 10,09 <1 <1 19,35
11/05/2011 8
QC1-R1 1,81 11,11 14,76 <1 <1 28,68
29,46 QC1-R2 1,53 11,52 15,08 <1 <1 29,13
QC1-R3 1,46 12,19 15,93 <1 <1 30,58
08/06/2011 9
QC1-R1 1,18 15,16 20,86 <1 <1 38,20
42,37 QC1-R2 1,28 16,52 21,76 <1 <1 40,56
QC1-R3 1,77 19,77 25,82 <1 <1 48,36
04/07/2011 10
QC1-R1 2,08 21,67 29,26 <1 <1 54,01
58,18 QC1-R2 2,14 23,31 32,52 <1 <1 58,97
QC1-R3 1,95 24,42 34,18 <1 <1 61,55
144
08/08/2011 11
QC1-R1 2,75 31,20 38,90 <1 <1 73,85
75,76 QC1-R2 2,26 27,83 43,67 <1 <1 74,76
QC1-R3 2,11 26,78 48,78 <1 <1 78,67
06/09/2011 12
QC1-R1 4,17 52,81 54,95 <1 <1 112,93
110,91 QC1-R2 3,75 47,51 62,70 <1 <1 114,96
QC1-R3 3,24 38,95 61,65 <1 <1 104,84
12/10/2011 13
QC1-R1 2,33 29,78 40,68 <1 <1 73,79
63,55 QC1-R2 1,90 24,95 32,31 <1 <1 60,16
QC1-R3 1,75 20,39 33,55 <1 <1 56,69
Une concentration de <1 indique que la valeur observée est inférieure à la limite de détection de 1 µg/l
3. SPD émergents : les Haloacétonnitriles (HAN), la Chlopicrine (CPK) et les Halocétones (HC)
L’indicateur HAN4 représente la somme de quatre composés de la famille des haloacétonitriles
(trichloroacétonitrile (TCAN), dichloroacétonitrile (DCAN), bromo-chloroacétonitrile (BCAN) et
dibromoaéetonitrile (DBAN)).
Le tableau suivant résume les concentrations en HAN4 mesurées dans le réseau sur une période de treize
mois.
Tableau 8: Résumé des mesures de HAN dans le réseau
Date Campagne Point
d’échantillonnage TCAN (µg/l)
DCAN (µg/l)
BCAN (µg/l)
DBAN (µg/l)
HAN4 (µg/l)
Moyenne HAN4 (µg/l)
20/10/2010 1
QC1-R1 0,106 4,114 0,025 0,015 4,260
5,152 QC1-R2 0,112 5,014 0,288 0,019 5,433
QC1-R3 0,113 5,377 0,255 0,019 5,764
16/11/2010 2
QC1-R1 0,059 2,634 0,145 0,040 2,878
3,323 QC1-R2 0,068 3,044 0,159 0,032 3,304
QC1-R3 0,079 3,542 0,160 <0,01 3,786
07/12/2010 3
QC1-R1 0,047 1,823 0,173 0,066 2,109
2,603 QC1-R2 0,049 2,016 0,181 0,069 2,316
QC1-R3 0,033 3,215 0,131 <0,01 3,385
06/01/2011 4
QC1-R1 0,062 1,972 0,025 0,061 2,120
2,288 QC1-R2 0,067 2,079 0,183 0,064 2,392
QC1-R3 0,065 2,206 0,016 0,064 2,352
15/02/2011 5
QC1-R1 0,027 1,618 0,179 0,011 1,836
2,131 QC1-R2 0,030 1,718 0,218 0,011 1,977
QC1-R3 0,033 2,340 0,203 <0,01 2,581
09/03/2011 6
QC1-R1 0,061 1,592 0,135 0,036 1,824
2,034 QC1-R2 0,054 1,789 0,148 0,043 2,034
QC1-R3 0,060 1,999 0,139 0,047 2,245
145
06/04/2011 7
QC1-R1 0,055 1,488 0,164 <0,01 1,711
1,950 QC1-R2 0,053 1,550 0,170 <0,01 1,788
QC1-R3 0,062 2,091 0,193 <0,01 2,351
11/05/2011 8
QC1-R1 0,052 2,369 0,109 <0,01 2,535
3,023 QC1-R2 0,049 2,643 0,116 <0,01 2,812
QC1-R3 0,062 3,519 0,136 <0,01 3,722
08/06/2011 9
QC1-R1 0,058 2,809 0,142 <0,01 3,014
3,194 QC1-R2 0,057 3,046 0,145 <0,01 3,253
QC1-R3 0,057 3,105 0,148 <0,01 3,315
04/07/2011 10
QC1-R1 0,042 3,392 0,162 <0,01 3,601
3,881 QC1-R2 0,044 3,688 0,166 <0,01 3,904
QC1-R3 0,048 3,920 0,165 <0,01 4,138
08/08/2011 11
QC1-R1 0,038 4,686 0,237 <0,01 4,966
5,395 QC1-R2 0,039 5,098 0,254 <0,01 5,396
QC1-R3 0,039 5,509 0,271 <0,01 5,824
06/09/2011 12
QC1-R1 0,080 5,293 0,178 <0,01 5,557
5,788 QC1-R2 0,072 5,712 0,182 <0,01 5,972
QC1-R3 0,064 5,589 0,178 <0,01 5,835
12/10/2011 13
QC1-R1 0,067 3,327 0,199 <0,01 3,598
3,977 QC1-R2 0,068 3,621 0,214 <0,01 3,908
QC1-R3 0,068 4,135 0,217 <0,01 4,425
Une concentration de <0,01 indique que la valeur observée est inférieure à la limite de détection de 0,01 µg/l.
La CPK est un halonitrométhane, le trichloro-nitrométhane, aussi appelée chloropicrine.
L’indicateur HC2 représente la somme de deux composés de la famille des halocétones (1,1-dichloro-2-
propanone (11DCPone) et 1,1,1-trichloro-2-propanone (111TCPone)). Le tableau suivant résume les
concentrations de la CPK et du HC2 mesurées dans votre réseau sur une période de treize mois.
Tableau 9: Résumé des mesures de CPK et de HC2 dans le réseau
Date Campagne Point
d’échantillonnage CPK (µg/l)
Moyenne CPK (µg/l)
11DCPone (µg/l)
111TCPone (µg/l)
HC2 (µg/l)
Moyenne HC2 (µg/l)
20/10/2010 1
QC1-R1 0,874
0,970
1,207 4,661 5,868
7,230 QC1-R2 0,989 1,774 5,922 7,696
QC1-R3 1,048 1,704 6,420 8,125
16/11/2010 2
QC1-R1 0,579
0,626
0,781 3,062 3,843
4,736 QC1-R2 0,627 0,975 3,665 4,640
QC1-R3 0,673 1,505 4,221 5,726
07/12/2010 3
QC1-R1 0,466
0,399
0,363 2,166 2,529
3,056 QC1-R2 0,495 0,380 2,378 2,759
QC1-R3 0,237 1,229 2,651 3,880
146
06/01/2011 4
QC1-R1 0,524
0,535
0,463 1,874 2,337
2,551 QC1-R2 0,534 0,522 2,028 2,550
QC1-R3 0,548 0,556 2,210 2,766
15/02/2011 5
QC1-R1 0,227
0,238
0,431 1,374 1,805
2,157 QC1-R2 0,228 0,450 1,465 1,915
QC1-R3 0,260 0,571 2,178 2,749
09/03/2011 6
QC1-R1 0,483
0,510
0,541 1,489 2,030
2,348 QC1-R2 0,512 0,635 1,728 2,363
QC1-R3 0,536 0,610 2,041 2,652
06/04/2011 7
QC1-R1 0,331
0,351
0,461 1,355 1,816
2,332 QC1-R2 0,335 0,545 1,487 2,032
QC1-R3 0,388 0,692 2,456 3,148
11/05/2011 8
QC1-R1 0,438
0,489
0,397 2,328 2,725
3,287 QC1-R2 0,455 0,445 2,613 3,058
QC1-R3 0,575 0,571 3,506 4,078
08/06/2011 9
QC1-R1 0,393
0,420
0,443 2,878 3,322
3,596 QC1-R2 0,429 0,513 3,182 3,694
QC1-R3 0,439 0,527 3,244 3,771
04/07/2011 10
QC1-R1 0,386
0,403
0,336 3,702 4,039
4,626 QC1-R2 0,408 0,453 4,192 4,645
QC1-R3 0,415 0,470 4,725 5,194
08/08/2011 11
QC1-R1 0,359
0,391
0,339 4,091 4,431
5,109 QC1-R2 0,390 0,556 4,635 5,191
QC1-R3 0,424 0,422 5,282 5,704
06/09/2011 12
QC1-R1 1,213
1,151
0,568 5,568 6,136
6,738 QC1-R2 1,180 0,596 6,304 6,900
QC1-R3 1,059 0,724 6,454 7,178
12/10/2011 13
QC1-R1 0,509
0,537
0,532 3,391 3,923
4,600 QC1-R2 0,528 0,629 3,850 4,479
QC1-R3 0,575 0,684 4,713 5,398
Conclusion:
Bien que les résultats contenus dans ce rapport sont à titre informatif, étant donné que les analyses ont été
réalisées par le laboratoire de l'Université Laval qui n'est pas accrédité, nous souhaitons vous informer que
certains résultats ont révélé un dépassement de la norme prévue RQEP au regard des THM totaux et/ou des
AHA. Nous vous laissons le soin de déterminer quelles mesures sont à prendre en fonction de l'historique des
concentrations de THM totaux dans votre installation ainsi que des actions déjà entreprises pour sa mise aux
normes.
147
Enfin, nous voudrions vivement vous remercier pour votre précieuse collaboration tout au long de notre projet.
La prochaine étape sera de déterminer les facteurs responsables de l’occurrence et de la variation de ces sous-
produits. Avec ces informations, il sera possible d’identifier des solutions pour réduire leur présence et améliorer
leur surveillance.