Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Water-energy-carbon nexus : a system dynamics approach for assessing urban water systems Chhipi Shrestha, Gyan Kumar 2017

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2017_september_chhipi shrestha_gyan.pdf [ 5.44MB ]
Metadata
JSON: 24-1.0348074.json
JSON-LD: 24-1.0348074-ld.json
RDF/XML (Pretty): 24-1.0348074-rdf.xml
RDF/JSON: 24-1.0348074-rdf.json
Turtle: 24-1.0348074-turtle.txt
N-Triples: 24-1.0348074-rdf-ntriples.txt
Original Record: 24-1.0348074-source.json
Full Text
24-1.0348074-fulltext.txt
Citation
24-1.0348074.ris

Full Text

WATER-ENERGY-CARBON NEXUS: A SYSTEM DYNAMICS APPROACH FOR ASSESSING URBAN WATER SYSTEMS by  Gyan Kumar Chhipi Shrestha  M.Sc., Tribhuvan University, Nepal 2005  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in  THE COLLEGE OF GRADUATE STUDIES  (Civil Engineering)    THE UNIVERSITY OF BRITISH COLUMBIA  (Okanagan)  June 2017   © Gyan Kumar Chhipi Shrestha, 2017 ii Examination Committee The undersigned certify that they have read, and recommend to the College of Graduate Studies for acceptance, a thesis entitled:  Water-Energy-Carbon Nexus: A System Dynamics Approach for Assessing Urban Water Systems  Submitted by  Gyan Kumar Chhipi Shrestha  in partial fulfillment of the requirements of  the degree of  Doctor of Philosophy .   Dr. Kasun Hewage, School of Engineering / University of British Columbia Okanagan   Supervisor and Associate Professor  Dr. Rehan Sadiq, School of Engineering / University of British Columbia Okanagan   Co-Supervisor, Professor and Associate Dean   Dr. Sumi Siddiqua, School of Engineering / University of British Columbia Okanagan  Supervisory Committee Member, Assistant Professor    Dr. Abbas Milani, School of Engineering / University of British Columbia Okanagan  University Examiner, Professor    Dr. Tarek Zayed, Building, Civil, and Environmental Engineering Department/ Concordia University  External Examiner, Professor    June 1, 2017  (Date Submitted to Grad Studies)     iii Abstract  Water, energy, and carbon emissions of Urban Water Systems (UWSs) are intertwined and have complex interactions forming a water-energy-carbon (WEC) nexus. A comprehensive methodology to quantify dynamic WEC nexus is required. The main objective of this research is to develop a decision support system (DSS) for assessing the WEC nexus for sustainable planning and management of UWSs.   This research has been accomplished in five distinct steps. In the first step, key Sustainability Performance Indicators (SPIs) of small to medium-sized UWSs have been identified. The SPIs related to water consumption, energy use, carbon emissions, and cost were used for developing the DSS. In the second step, a WEC DSS has been developed for an operational phase of an UWS using system dynamics and then applied to the City of Penticton. The highest energy consumer was found to be indoor hot water use in the city. In the third step, a framework has been developed to study the impacts of neighbourhood densification on the WEC nexus. A higher net residential density will result in lower per capita water demand, energy use, net carbon emissions, and life cycle cost of water distribution system. The proposed framework provides an optimal residential density and energy intensity of water distribution, which can be used as inputs to the WEC DSS. In the fourth step, microbial water quality guidelines for reclaimed water have been developed for various non-potable urban reuses. Moreover, the FitWater tool has been developed for evaluating fit-for-purpose wastewater treatment and reuse potentials based on cost, health risk, and the WEC nexus. The outputs of FitWater can be used as inputs to the WEC DSS. In the last step, the economics of the WEC nexus of net-zero water communities has been analyzed using the WEC model.  The DSS developed based on this research is capable of quantifying dynamic water consumption, energy use, carbon emissions, and the cost of UWSs. The DSS can analyze different WEC-based interventions. The DSS can be used by utilities, urban developers, and policy makers for long-term planning of urban water in communities.   iv Preface  I, Gyan Kumar Chhipi Shrestha, conceived and developed all the contents in this thesis under the supervision of Dr. Kasun Hewage and Dr. Rehan Sadiq.  I wrote all the manuscripts and both supervisors have reviewed them and provided feedback to improve the manuscripts and the thesis. Altogether eight journal papers and one conference paper have been published or submitted for publication in peer-reviewed scientific journals and a conference proceeding as follows: 1. A version of Chapter 3 has been published in Clean Technologies and Environmental Policy journal entitled “‘Socializing’ sustainability: A critical review on current development status of social life cycle impact assessment method” (Chhipi-Shrestha et al. 2015a). 2. A version of Chapter 4 has been published in Water Environment Research journal entitled “Sustainability performance indicators for small to medium sized urban water systems: A selection process using Fuzzy-ELECTRE method” (Chhipi-Shrestha et al. 2017a).  3. A version of Chapter 5 has been published in the ASCE Journal of Water Resources Planning and Management entitled “Water-Energy-Carbon nexus modelling for an urban water system: A system dynamics approach” (Chhipi-Shrestha et al. 2017b). 4. A version of Chapter 6 has been published in the Journal of Cleaner Production entitled “Impacts of neighbourhood densification on water-energy-carbon nexus: Investigating water distribution and residential landscaping system” (Chhipi-Shrestha et al. 2017c). 5. A version of Chapter 7 has been published in the Science of the Total Environment journal entitled “ Probabilistic risk-based investigation on microbial quality of reclaimed water for urban reuses” (Chhipi-Shrestha et al. 2017d). 6. A version of Chapter 8 has been submitted and is under review in the Science of the Total Environment journal entitled “Fit-for-purpose wastewater treatment: Conceptualization and development of decision support tool (I)” for possible publication (Chhipi-Shrestha et al. 2017e).  v 7.  A version of Chapter 8 has been submitted and is under review in the Science of the Total Environment journal entitled “Fit-for-purpose wastewater treatment: Testing and implementation of decision support tool (II)” for possible publication (Chhipi-Shrestha et al. 2017f). 8. A version of Chapter 9 has been published in the proceeding of Canadian Society for Civil Engineering (CSCE) conference entitled “System dynamics modelling for an urban water system: net-zero water analysis for Peachland (BC)” (Chhipi-Shrestha et al. 2015b). 9. A version of Chapter 9 has been submitted and is under review in the ASCE Journal of Sustainable Water in the Built Environment entitled “Economic and energy efficiency of net-zero water communities: A system dynamics analysis” for possible publication (Chhipi-Shrestha et al. 2017g).    vi Table of Contents  Examination Committee ......................................................................................................... ii Abstract ................................................................................................................................... iii Preface ..................................................................................................................................... iv Table of Contents ................................................................................................................... vi List of Tables ........................................................................................................................ xiii List of Figures ........................................................................................................................ xv List of Symbols ................................................................................................................... xviii List of Abbreviations ........................................................................................................... xix Acknowledgements ............................................................................................................. xxii Dedication ........................................................................................................................... xxiv  Introduction........................................................................................................... 1 1.1 Background and motivation .................................................................................................. 1 1.1.1 UWS metabolism .............................................................................................................. 1 1.1.2 UWS sustainability ........................................................................................................... 3 1.2 Research gap ......................................................................................................................... 5 1.3 Research objectives ............................................................................................................... 7 1.4 Thesis organization ............................................................................................................... 8 1.5 Meta language ....................................................................................................................... 9  Research Methodology ....................................................................................... 10 2.1 Objective 1 .......................................................................................................................... 11 2.2 Objective 2 .......................................................................................................................... 12 2.3 Objective 3 .......................................................................................................................... 13 2.4 Objective 4 .......................................................................................................................... 14 2.5 Objective 5 .......................................................................................................................... 14  Literature Review ............................................................................................... 16 3.1 Life cycle sustainability assessment (LCSA) ...................................................................... 16  vii 3.1.1 Life cycle assessment (LCA) .......................................................................................... 16 3.1.2 Life cycle cost analysis (LCCA) ..................................................................................... 17 3.1.3 Social life cycle assessment (S-LCA) ............................................................................. 18 3.1.4 Sustainability Performance Indicators for UWSs ........................................................... 19 3.2 Water-energy-carbon (WEC) nexus .................................................................................... 21 3.2.1 Energy for water ............................................................................................................. 21 3.2.2 Energy from water .......................................................................................................... 21 3.2.3 Water for energy ............................................................................................................. 22 3.2.4 GHGs from energy and water ......................................................................................... 23 3.2.5 WEC models ................................................................................................................... 23 3.2.5.1 Static models .......................................................................................................... 23 3.2.5.2 Dynamic models .................................................................................................... 25 3.3 System dynamics modelling (SDM) ................................................................................... 25 3.3.1 System dynamics model construction ............................................................................ 26 3.3.1.1 Stock and flow diagram ......................................................................................... 26 3.3.1.2 Feedback ................................................................................................................ 27 3.3.2 SDM for urban water management ................................................................................. 28 3.4 Net-zero water (NZW) ........................................................................................................ 30 3.4.1 Impact of neighbourhood densification on WEC nexus ................................................. 35 3.4.2 Reclaimed water use ....................................................................................................... 36 3.4.2.1 Global water reuse status and trend ....................................................................... 37 3.4.2.2 Microbial quality of reclaimed water ..................................................................... 40 3.4.3 Fit-for-purpose wastewater treatment ............................................................................. 45  Identification of Sustainability Performance Indicators................................. 48 4.1 Background ......................................................................................................................... 48 4.1.1 Small to medium-sized urban water systems (SMUWSs) .............................................. 48 4.1.2 Fuzzy sets and fuzzy numbers ........................................................................................ 50 4.2 Methodology ....................................................................................................................... 51 4.2.1 Initial screening of SPIs .................................................................................................. 52 4.2.2 Development of selection criteria ................................................................................... 52 4.2.3 Fuzzy-ELECTRE I ......................................................................................................... 55 4.3 Results ................................................................................................................................. 59 4.3.1 Technical ........................................................................................................................ 60 4.3.2 Environmental ................................................................................................................ 62  viii 4.3.3 Economic ........................................................................................................................ 62 4.3.4 Social .............................................................................................................................. 64 4.3.5 Institutional ..................................................................................................................... 64 4.4 Discussion ........................................................................................................................... 65 4.4.1 Technical ........................................................................................................................ 65 4.4.2 Environmental ................................................................................................................ 67 4.4.3 Economic ........................................................................................................................ 69 4.4.4 Social .............................................................................................................................. 70 4.4.5 Institutional ..................................................................................................................... 71 4.5 Summary ............................................................................................................................. 72  System Dynamics Modelling of Water-Energy-Carbon (WEC) Nexus ......... 73 5.1 Background ......................................................................................................................... 73 5.2 Methodology ....................................................................................................................... 74 5.2.1 Water module ................................................................................................................. 74 5.2.1.1 Water consumer growth sub-models ...................................................................... 74 5.2.1.2 Water and wastewater sub-models ......................................................................... 75 5.2.2 Energy module ................................................................................................................ 75 5.2.3 Carbon module ............................................................................................................... 76 5.3 Results and Discussion........................................................................................................ 79 5.3.1 Model calibration and validation .................................................................................... 79 5.3.2 Data requirements ........................................................................................................... 80 5.3.3 WEC model for Penticton ............................................................................................... 81 5.3.3.1 Water use and wastewater generation .................................................................... 83 5.3.3.2 Energy use .............................................................................................................. 83 5.3.3.3 Carbon emissions ................................................................................................... 84 5.3.3.4 Quantitative WEC nexus ........................................................................................ 85 5.3.3.5 Sensitivity analysis ................................................................................................. 87 5.3.4 Scenario analysis ............................................................................................................ 89 5.3.4.1 Business as usual scenario ..................................................................................... 90 5.3.4.2 Indoor water demand management ........................................................................ 92 5.3.4.3 Outdoor water demand management ..................................................................... 92 5.3.4.4 Source water alternatives ....................................................................................... 92 5.3.4.5 Water heating energy alternatives .......................................................................... 93 5.3.4.6 WEC nexus analysis of interventions .................................................................... 93  ix 5.4 Summary ............................................................................................................................. 95  Investigating Impacts of Residential Density on WEC Nexus ........................ 97 6.1 Background ......................................................................................................................... 97 6.2 Methodology ....................................................................................................................... 98 6.2.1 Water demand ................................................................................................................. 99 6.2.2 Energy use .................................................................................................................... 100 6.2.3 Carbon emissions and sequestration ............................................................................. 101 6.2.4 WEC aggregation.......................................................................................................... 102 6.2.5 Life cycle cost analysis ................................................................................................. 103 6.3 Results ............................................................................................................................... 104 6.3.1 Application ................................................................................................................... 104 6.3.1.1 Study area ............................................................................................................. 104 6.3.1.2 Data ...................................................................................................................... 105 6.3.1.3 Residential density and WEC nexus .................................................................... 106 6.3.1.4 Life cycle cost and residential density ................................................................. 110 6.3.1.5 Uncertainty and sensitivity analysis ..................................................................... 110 6.3.1.6 Two-dimensional analysis for WEC nexus scenarios .......................................... 113 6.4 Discussion ......................................................................................................................... 114 6.5 Summary ........................................................................................................................... 117  Development of Microbial Water Quality Guidelines for Reclaimed Water............................................................................................................................................... 118 7.1 Background ....................................................................................................................... 118 7.2 Methodology ..................................................................................................................... 120 7.2.1 Quantitative microbial risk assessment......................................................................... 121 7.2.2 Hazard identification .................................................................................................... 121 7.2.3 Exposure assessment .................................................................................................... 121 7.2.4 Dose-response assessment ............................................................................................ 123 7.2.5 Risk characterization .................................................................................................... 124 7.2.6 Data variability and uncertainty .................................................................................... 126 7.3 Results ............................................................................................................................... 127 7.3.1 Microbial water quality investigation and guideline values ......................................... 127 7.3.2 Sensitivity analysis ....................................................................................................... 131 7.3.3 Required treatment levels ............................................................................................. 132  x 7.3.4 Uncertainty analysis...................................................................................................... 133 7.3.5 Application ................................................................................................................... 134 7.4 Discussion ......................................................................................................................... 137 7.4.1 Microbial water quality guideline values...................................................................... 137 7.4.2 Model uncertainty and limitations ................................................................................ 139 7.5 Summary ........................................................................................................................... 141  Development of Decision Support Tool for Fit-For-Purpose Wastewater Treatment and Reuse .......................................................................................................... 142 8.1 Background ....................................................................................................................... 142 8.2 Methodology ..................................................................................................................... 144 8.2.1 Estimation of microbial concentration in reclaimed water ........................................... 145 8.2.2 Quantitative microbial risk assessment (QMRA) ......................................................... 147 8.2.3 Life cycle cost analysis (LCCA) ................................................................................... 148 8.2.4 Estimation of energy use and carbon emissions ........................................................... 149 8.2.5 Estimation of reclaimed water quantity and its distribution ......................................... 152 8.2.6 Multi-criteria decision analysis (MCDA) ..................................................................... 152 8.3 Results ............................................................................................................................... 154 8.3.1 Testing of FitWater ....................................................................................................... 156 8.3.1.1 Testing with existing data .................................................................................... 156 8.3.1.2 Tool demonstration .............................................................................................. 159 8.3.2 Implementation of FitWater ......................................................................................... 160 8.3.2.1 Study area ............................................................................................................. 160 8.3.2.2 Alternative treatment trains .................................................................................. 160 8.3.2.3 Design of wastewater collection and reclaimed water distribution system.......... 162 8.3.2.4 Ranking of treatment trains .................................................................................. 164 8.3.3 Trade-off analysis ......................................................................................................... 166 8.3.4 Cost and energy use of reducing health risk in varying plant capacities ...................... 167 8.4 Discussion ......................................................................................................................... 170 8.5 Summary ........................................................................................................................... 174  System Dynamics Analysis of Economic and Energy Efficiency of Net-Zero Water Communities ............................................................................................................ 176 9.1 Background ....................................................................................................................... 176 9.2 Methodology ..................................................................................................................... 177  xi 9.2.1 Cost module .................................................................................................................. 177 9.2.2 Energy module .............................................................................................................. 178 9.2.3 Water module ............................................................................................................... 180 9.2.4 Uncertainty analysis...................................................................................................... 181 9.2.5 Data requirements ......................................................................................................... 182 9.3 Results ............................................................................................................................... 183 9.3.1 Economics of WEC nexus of NZW communities ........................................................ 183 9.3.2 Economic and energy use impacts of environmental factors on NZW ........................ 187 9.3.2.1 Impact of precipitation amount on NZW ............................................................. 188 9.3.2.2 Impact of source water proximity on reclaimed water use .................................. 191 9.3.2.3 Impact of elevation head on reclaimed water distribution ................................... 193 9.4 Discussion ......................................................................................................................... 195 9.5 Summary ........................................................................................................................... 198  Conclusions and Recommendations .............................................................. 199 10.1 Conclusions ....................................................................................................................... 199 10.2 Originality and contribution .............................................................................................. 201 10.3 Limitations ........................................................................................................................ 203 10.4 Recommendations for future work.................................................................................... 204 References ............................................................................................................................ 206 Appendices ........................................................................................................................... 245 Appendix A: Additional Information on SPI Identification ........................................................... 245 A.1 Delphi method .............................................................................................................. 245 A.2 Screening of SPIs for SMUWSs ................................................................................... 247 A.3 Application of Fuzzy-ELECTRE I method for ranking SPIs in economic dimension . 251 Appendix B Additional Information on WEC Modelling .............................................................. 257 B.1 Stock and flow diagrams of the WEC model ............................................................... 257 B.2 Data requirements ......................................................................................................... 263 B.3 Validation figure for water consumption ...................................................................... 274 B.4 Sensitivity analysis framework and input data ............................................................. 274 Appendix C Additional Information on Water Distribution and Residential Landscaping System ........................................................................................................................................................ 278 C.1 Alternative Designs and their Characteristics ............................................................... 278 C.2 Water Demand, Energy Use, Carbon Emissions, and Cost Estimation ........................ 279  xii C.3 Carbon Sequestration of Residential Landscaping ....................................................... 282 C.4 Xeriscaping Design ....................................................................................................... 282 Appendix D Additional Information on Microbial Water Quality Guidelines Development ........ 284 D.1 Sensitivity analysis results ............................................................................................ 284 D.2 Raw wastewater microbiology ..................................................................................... 285 D.3 Log removal of treatment processes ............................................................................. 285 Appendix E Additioal Information on FitWater Development ...................................................... 286 E.1 Fuzzy data on microbial concentration of wastewater .................................................. 286 E.2 Fuzzy data on microbial concentration of greywater .................................................... 286 E.3 FitWater Data Entry and Flow ...................................................................................... 287 Appendix F Additional Information on NZW Analysis ................................................................. 288 F.1 Stock and flow diagram of cost module ....................................................................... 288 F.2 Input data for Monte Carlo Simulations ....................................................................... 289   xiii List of Tables  Table 1.1    Domestic water consumption in the developed countries ..................................... 3 Table 3.1    Major literature used for the screening of the SPIs for SMUWSs....................... 20 Table 3.2    Strengths and weaknesses of system dynamics based urban water models ........ 29 Table 3.3    Definitions of net-zero water and its nuances ...................................................... 31 Table 3.4    DSSs for NZW analysis ....................................................................................... 33 Table 3.5    Status of global water reuse ................................................................................. 39 Table 3.6    Reclaimed water quality guidelines for urban reuses in various regions of the world .................................................................................................................... 42 Table 3.7    Existing DSTs related to fit-for-purpose wastewater treatment .......................... 47 Table 4.1   Classification of UWSs based on US EPA (2009a) ............................................. 49 Table 4.2    Criteria for the selection of SPIs .......................................................................... 54 Table 4.3    Final Ranks of the Selected SPIs of UWSs ......................................................... 66 Table 5.1    Equations of the WEC model .............................................................................. 76 Table 5.2    Percent contribution of parameters to the variability of the WEC model ........... 88 Table 5.3    Scenarios developed for the UWS of Penticton................................................... 90 Table 6.1    Major factors affecting water demand and their distribution parameters .......... 111 Table 6.2    Scenario features ................................................................................................ 113 Table 7.1    Exposure factors for different urban water uses ................................................ 122 Table 7.2    Parameter values of dose-response models ....................................................... 124 Table 7.3    Morbidity, disease burden per case and susceptibility fraction ......................... 125 Table 7.4    Microbial water quality guidelines for different reuses (E. coli in cfu/100 mL) 130 Table 7.5    Log removal required for reclaimed water in different water reuses ................ 132 Table 7.6    Uncertainty in model and input parameters ....................................................... 134 Table 7.7    Features of wastewater treatment plants ............................................................ 135 Table 7.8    E. coli concentration in WWTP effluents in Okanagan from 2012-2014 ......... 135 Table 7.9    Human health risk of public park irrigation by WWTP effluents ..................... 136 Table 7.10    Health risk of other pathogens by WWTP effluents ........................................ 137 Table 8.1    Alternative technologies and pathogen removal efficiency (log removal in TFNs) ............................................................................................................................ 146  xiv Table 8.2    Equations used for life cycle costing of treatment technologies ....................... 150 Table 8.3    Energy consumption equations for treatment technologies ............................... 151 Table 8.4    Applications of FitWater to existing wastewater treatment plants .................... 158 Table 8.5    Treatment train alternatives for various water reuse purposes and their ranking ............................................................................................................................ 159 Table 8.6    Minimum water quality required for urban reuses ............................................ 161 Table 8.7    Alternative treatment trains ............................................................................... 162 Table 8.8    Scenarios with varying importance of selection criteria .................................... 164 Table 8.9    Ranks of alternative treatment trains in different scenarios .............................. 165 Table 8.10    Equations developed for unit annualized LCC and energy intensity per unit log removal............................................................................................................... 170 Table 9.1    Scenarios for net-zero water analysis ................................................................ 183 Table 9.2    Average annual net water in different scenarios for 2016 to 2035 .................... 186   xv List of Figures  Figure 1.1    Major inputs and outputs of urban water systems ................................................ 1 Figure 1.2    Thesis structure and organization ......................................................................... 8 Figure 2.1    Research methodology framework ..................................................................... 11 Figure 3.1    Representation of a stock, flow, converter, and connector................................. 26 Figure 3.2    Positive feedback in growing rabbit population ................................................. 27 Figure 3.3    Negative feedback in declining skunk population.............................................. 28 Figure 3.4    Trend of water reuse in different regions of the world ....................................... 40 Figure 4.1    Membership function of  A ................................................................................. 51 Figure 4.2    Methodological framework used for the selection of SPIs ................................ 52 Figure 4.3    Outranking relations of the technical SPIs with DMB ....................................... 61 Figure 4.4    Outranking relations of the environmental SPIs with DMB .............................. 63 Figure 4.5    Outranking relations of the economic SPIs with DMB ...................................... 63 Figure 4.6    Outranking relations of the social SPIs with DMB ............................................ 64 Figure 4.7    Outranking relations of the institutional SPIs with DMB .................................. 65 Figure 5.1    Causal loop diagram of the WEC nexus of a SMUWS ...................................... 79 Figure 5.2    Screenshot of the WEC DSS interface ............................................................... 82 Figure 5.3    Annual direct water use, energy use and GHG emissions in various stages of Penticton UWS .................................................................................................. 86 Figure 5.4    WEC nexus of Penticton UWS........................................................................... 87 Figure 5.5    Monthly water footprint, energy use, and carbon emissions under Scenario 1 from 2015 – 2034 .............................................................................................. 91 Figure 5.6    Change in average annual water footprint, energy use and carbon emissions compared to Scenario 1 ..................................................................................... 91 Figure 5.7    WEC nexus analysis of interventions: water-based (a & b) and energy-based (c & d) .................................................................................................................... 94 Figure 6.1    Impacts of neighbourhood densification on the WEC nexus of WDRLS (per capita) ................................................................................................................ 98 Figure 6.2    Conceptual framework to study the impacts of neighbourhood densification on the WEC nexus of water distribution and residential landscaping system ........ 99  xvi Figure 6.3    Lot coverage, landscaping, and population variation over net residential density ......................................................................................................................... 107 Figure 6.4    WEC nexus in various residential densities: Interaction plot ........................... 108 Figure 6.5    Ecological footprint of various residential densities D1 to D11 ...................... 109 Figure 6.6    LCC of the WDS and its rate of change in different residential densities ....... 110 Figure 6.7    Water demand variability (5th and 95th percentiles) in different residential densities ........................................................................................................... 112 Figure 6.8    Two-dimensional WEC nexus: Varying scenario results in different densities 114 Figure 7.1    Research framework for developing and applying microbial water quality guidelines for reclaimed water ........................................................................ 120 Figure 7.2    CDFs of E. coli concentrations with acceptable risks in different water reuse applications ...................................................................................................... 128 Figure 7.3    RME with 90% confidence interval of E. coli concentration for acceptable risk ......................................................................................................................... 129 Figure 7.4    Effects of input parameters over their range on output mean in lawn irrigation ......................................................................................................................... 131 Figure 8.1    Conceptual model for evaluating fit-for-purpose wastewater treatment and reuse potential ........................................................................................................... 144 Figure 8.2    Sceenshot of the FitWater interface ................................................................. 155 Figure 8.3    Flowchart for preparing alternative treatment trains ........................................ 161 Figure 8.4    WECCo triangle for Alternative 7 in Scenario 1 .............................................. 164 Figure 8.5    Variation of water, health risk, energy use, and life cycle cost in different treatment alternatives ....................................................................................... 167 Figure 8.6    Unit annualized LCC per unit log removal for different treatment technologies ......................................................................................................................... 169 Figure 8.7    Energy intensity per unit log removal for different treatment technologies .... 169 Figure 9.1    Median monthly net water and its energy use, cost, and carbon emissions ..... 185 Figure 9.2    Median annual freshwater saving, energy use, cost and carbon emissions in different scenarios ............................................................................................ 187 Figure 9.3    NZW and its energy intensity and unit cost in different precipitation ............. 189 Figure 9.4    Energy intensities of wastewater system and RWH in different precipitation . 189  xvii Figure 9.5    Unit cost of wastewater system and RWH in different precipitation ............... 190 Figure 9.6    Energy intensities of DWSs in varying conveyance length ............................. 191 Figure 9.7    Unit LCC of DWSs in varying conveyance lengths ......................................... 192 Figure 9.8    Energy intensities in different net elevation heads ........................................... 193 Figure 9.9    Unit LCC of wastewater reuse with different elevation heads of secondary distribution ....................................................................................................... 194      xviii List of Symbols   α   Safety factor (in pumping power equation)  α and r Parameters referring to pathogen infectivity constant (in dose-response relationships) ?̃?   Tilde (~) on “A “meaning a fuzzy number “A” Ap and Aq A pair of alternative with p, q = 1, 2, ..., m and,  p ≠q  Cd  Cadmium C*pq   Concordance index Cu  Copper d   Pathogen dose;   D*pq   Discordance index dy/dx  Differentiate y with respect to x E  Energy consumed f  Friction loss fs   Susceptibility fraction Hg  Mercury l  Lowermost or lower value m  Most probable or middle value N ~ (μ, σ)  Normal distribution with mean (μ) and standard deviation (σ)  N50  Median infective dose, i.e., the dose required to infect 50% of population Pill|inf   Risk of disease given infection, i.e., morbidity;  Pinf (d)  Probability or risk of infection to an individual exposed to a single pathogen dose “d” Pinf(A) (d)  Annual probability or risk of an infection from “n” exposures per year due to a single pathogen dose “d” Pb Lead T~ (l, m, u) Triangular distribution with lower most, most probable, and uppermost value u  uppermost or upper value yr  Year γ   Specific weight of water η   Efficiency  ηt   Mechanical transmission efficiency  ρ  Spearman’s rank correlation coefficient $  Dollar ∫  f(t) dt𝒕0 Definite integral of f(t) from time 0 to t    xix List of Abbreviations  2-D MCA Two-dimensional Monte Carlo Analysis  ABM  Agent-based Modelling Alt.  Alternative AS  Activated sludge Avg.  Average  BC  British Columbia BNR  Biological nutrient removal BWA  Boil water advisory CDF  Cumulative density functions CE  Carbon emissions CF   Coagulation and flocculation (in wastewater process in FitWater)  CF  Carbon footprint (in WEC model) CI  Commercial and institutional CII  Commercial, institutional, and industrial Cl2   Chlorination CLD  Causal loop diagram CO2e  Carbon dioxide equivalent  Consvn. Conservation CoP  City of Penticton DALYs Disability-Adjusted Life-Years DBPC   Disease burden per case  DBPs  Disinfection Byproducts DMB  Decision Maker’s Boundary DSS  Decision Support System DST  Decision Support Tool DU  Dwelling units DWS  Drinking Water System  E. coli  Escherichia coli EC   Economic sustainability performance indicator EE  Embodied energy EF  Ecological footprint EI  Energy intensity ELECTRE  ELimination Et Choix Traduisant la REalite´ (Elimination and Choice Translating Reality) EN   Environmental sustainability performance indicator F-AHP  Fuzzy-Analytical Hierarchical Process FBR  Fed Batch Reactor g  Gram  xx gha  Global hectare GHG  Greenhouse gases GIS  Geographic Information System ha  Hectare IN   Institutional sustainability performance indicator Insti.  Institutional  kg  Kilogram kWh  Kilowatt-hour L  Litre L/p/day Litres/person/day LCA  Life cycle assessment  LCC  Life cycle cost LCCA  Life cycle cost analysis  LCI  Life cycle inventory LCIA  Life cycle impact assessment LCSA  Life cycle sustainability assessment MBR   Membrane Bioreactor MCS  Monte Carlo Simulation MF   Microfiltration (in wastewater process) MF  Multi-family (in neighbourhood densification) ML  Million Litres Mm3  Million cubic metre MWh  Meghwatt-hour MWR   Municipal Wastewater Regulation ND  Not Detected NPV  Net present value NPW  Net-Positive Water NR  Non-residential NWWBI National Water and Wastewater Benchmarking Initiative  NZW  Net-Zero Water O3  Ozonation Oper.   Operating  PPCPs  Pharmaceuticals and Personal Care Products PS  Primary Sedimentation PSR    Pressure-State-Response (framework) Q   Discharge R  Rank r2  Coefficient of determination REUM  Residential End-Use Model  RME  Reasonable Maximum Estimate  xxi RO  Reverse osmosis RW  Reclaimed water SBR  Sequencing Batch Reactor SDM  System dynamics model SF   Surface filtration (in wastewater process) SF  Single-family (in neighbourhood densification) S-LCA  Social Life Cycle Assessment SMUWS Small to Medium-sized Urban Water System SO   Social sustainability performance indicator SP  Sludge processing SPI  Sustainability Performance Indicator sqft  Square feet SWS  Storm Water System tCO2e  Ton carbon dioxide equivalent TE   Technical sustainability performance indicator TFN  Triangular Fuzzy Number UF  Ultrafiltration UK  United Kingdom US EPA United States Environment Protection Agency US  United States UV  Ultraviolet disinfection UWS  Urban Water System W   Water  WDRLS Water Distribution and Residential Landscaping System WDS  Water Distribution System WEC nexus Water-Energy-Carbon nexus WECCo Water-Energy-Carbon-Cost (triangle) WF  Water footprint WHO  World Health Organization WQ   Water quality  WRCC  Water Resource Carrying Capacity WT   Water treatment  WW  Wastewater WWS  Wastewater System WWT   Wastewater Treatment  WWTP Wastewater treatment plant  xxii Acknowledgements  First and foremost, I offer my enduring gratitude to my supervisors Dr. Kasun Hewage and Dr. Rehan Sadiq for their constant support, encouragement, and invaluable guidance. This research would not have been possible without their guidance. I really appreciate their open door policy. Both professors were sources of inspiration in this academic journey. In addition, they are always emotionally and morally supportive, helping me to flourish in my academic and real-world pursuits. Sirs, you are the best supervisors, I ever had!   I am very thankful to my committee member, Dr. Sumi Siddiqua, for her guidance and insightful feedback. Her constructive criticisms helped me to mature the research ideas. I thank Mark Holland and James Kay of New Monaco Enterprise; Elsie Lemke, Corine Gain, Joe Mitchell, and Shawn Grundy of the District of Peachland; Peter Gigliotti of Urban Systems; Brent Edge of the Penticton water treatment; and Randy Craig of Penticton wastewater treatment plant for providing valuable data for case studies and feedback on the research.   I acknowledge the Natural Sciences and Engineering Research Council of Canada (NSERC) for awarding me the competitive Alexander Graham Bell Canada Graduate Scholarship (CGSD). I also thank NSERC for providing me partial financial assistance through NSERC Collaborative Research and Development (CRD) Grants. I am also thankful to the University of British Columbia (UBC) for providing me partial financial support through University Graduate Fellowships (UGF) and a Graduate Student Travel Grant.   I would like to acknowledge my research group especially Dr. Bahareh Reza, Rajeev Ruparathna, Dr. Husnain Haider, and Adil Umer for their constructive and joyful support throughout this research journey. I appreciate the administrative staff of the School of Engineering, UBC, especially Shannon Hohl, Angela Perry, and Karen Seddon for their  xxiii generous support. I offer my sincere gratitude to Dr. Carolyn Labun and Amanda Brobbel for guiding me in the technical writing of this research.  My special thanks to the dedicated library staff of the UBC Okanagan Library for helping me to get the necessary documents on time. Even in the age of e-library, they proved themselves invaluable in searching and getting rare documents on time through inter-library loans. This service is highly praiseworthy and was very helpful in my research. I would also like to thank Patty Wellborn and Chris Bowerman for disseminating my research findings across the globe through UBC media.  Finally, I would like to acknowledge my exceptional parents for their patience and constant support throughout my studies including this research work. Special thanks are owed to my beloved wife Ambika Shrestha for her constant support, caring, and inspiration to overcome the stressful environment in order to accomplish this work.   xxiv Dedication  To my father Jagat Hari,  mother Sagat Maya, and  wife Ambika   1  Introduction 1.1 Background and motivation The world’s urban population is more than half (~54%) of the total population and is rapidly increasing (UN DESA 2014). In Canada, the urban population is very high (~ 81%) and is continuing to grow (Statistics Canada 2014a). The growing population requires a large volume of water served by urban water supply systems. Urban water processes, such as water abstraction, treatment, distribution, wastewater treatment, disposal, and storm water drainage are essential in any urban area. These processes are necessary for the human consumption of safe water and reduction of environmental impacts due to wastewater discharge (Termes-Rifé et al., 2013). These human regulated urban water processes constitute a human hydrologic cycle (Bagley et al. 2005), or simply an urban water system (UWS). 1.1.1 UWS metabolism An urban water system consists of a drinking water system (DWS), a wastewater system (WWS), and a storm water system (SWS) (Figure 1.1).    Note: DWS: Drinking Water System, WWS: Wastewater System, SWS: Storm Water System  Figure 1.1    Major inputs and outputs of urban water systems UWSs consume resources (inputs), such as water, energy, materials (e.g., water treatment chemicals), and financial resources to produce outputs like water and wastewater services. Similar to any other built infrastructure, UWSs release greenhouse gases (GHGs), discharge effluents (potentially containing heavy metals and other harmful materials), generate solid waste  2 (biosolids), and pose hazards like storm water flooding in some instances. This phenomenon is referred to as the urban metabolism (Novotny 2012) for a water sector.   Globally, total freshwater withdrawals are estimated to have increased by around 1% per year from 1987 to 2000 based on the FAO AQUASTAT database, and the present rate is expected to be the same considering a similar overall trend (UN-WWAP 2014). Municipal water accounts for 12% of the total withdrawals with industrial and agricultural water withdrawals accounting for 19% and 69% respectively (FAO 2014). Similar to the global outlook, Canada has 13% of total water withdrawals by municipal sector; however, the thermal power generation, manufacturing, agriculture, mining, and oil and gas account for 69%, 10%, 6%, 2%, and 1% of total water withdrawals respectively (Environment Canada 2014a). Residential water use is a major component of municipal water use and accounts for more than 50% of all municipal water use in Canada (Environment Canada 2004).  Water consumption is highly dependent on the geography and the development status of a country. For instance, domestic water consumption varies largely from 135 L/p/day in Israel to 343 L/p/day in Canada, 490 L/p/day in British Columbia (Canada) and 675 L/p/day in the Okanagan Valley (BC, Canada) as shown in Table 1.1. Even domestic water use reaches upto 1,000 L/p/day in the summer in the Okanagan Valley (OBWB 2011). Canada has a very high domestic water consumption rate that is also referred to as overconsumption (Renzetti, 1999; Ma, 2014). Although Canada overall has abundant freshwater supplies, water supply shortages exist due to water quality and/or quantity issues in some communities. Approximately 26% of municipalities with water supply systems in place experienced water supply shortages from 1994 to 1999 due to droughts, infrastructure problems, and increased consumption (Environment Canada 2004).    3 Table 1.1    Domestic water consumption in the developed countries  Country Consumption (L/p/day) Israel 135 France 150 Sweden 200 Italy 250 United States 382 Canada 343 British Columbia (BC) (Canada) 490 Okanagan average (BC, Canada) 675 Okanagan summer average (BC) 1,000 Source: (Environment Canada 2014b; OBWB 2011)  1.1.2 UWS sustainability Modern urban water systems are designed to provide clean drinking water, remove wastewater, and manage storm water without posing harm to the environment. Although UWSs achieve these first three fundamental requirements to a high degree, this sector is criticized from the sustainability perspective (Hellstro 2000). Moreover, Canadian communities have been facing pressing concerns of carbon mitigation and adaptation to global climate change in different sectors including urban water management (Maas 2009; Environment Canada 2014c; Government of BC 2012). Current UWSs are usually linear in terms of urban metabolism, sometimes called the “take, make, waste approach”. This approach discourages water reuse and has become unsustainable (Daigger 2009). Linear systems result in a higher level of pollution, and also consume a greater amount of resources, such as water, air, and soil for the dilution and assimilation of residuals compared to closed systems based on recycling (Novotny 2012; Joustra and Yeh 2014a). Water reuse provides additional water that would otherwise be discarded from the system. Closed or semi-closed UWSs can be developed for achieving net-zero water with some water input from rainfall (Englehardt et al., 2013).  Net-zero water refers to the balance of water demand and supply within a given areal boundary (Holtzhower et al., 2014). “Net-zero water limits the consumption of freshwater resources and returns water back to the same watershed so not to  4 deplete the groundwater and surface water resources of that region in quantity or quality over the course of a year” (US Army 2011). However, the reclaimed water use in net-zero water may result in the exposure to various pathogens, such as virus, bacteria, and parasitic protozoa posing increased health risks to reclaimed water consumers (Schoen et al., 2014). Urban water systems consume a significant amount of energy. The energy consumption of the treated water supply and wastewater management is about 3% of city energy use in the USA, but it can be as high as 20% in some states, for example in California (Novotny 2012). Energy is required in almost all stages of UWSs: water abstraction, treatment, distribution, use, wastewater collection, treatment, disposal and/or reuse.  Usually, water and wastewater utilities have the largest expenditure in energy cost (Tuladhar et al., 2014). Moreover, energy use directly contributes to GHG emissions, or simply, carbon emissions. Energy use in an UWS can  be significantly reduced by employing the use of efficient water appliances, water conservation strategies, and efficient water and wastewater management practices (Barry, 2007; Novotny, 2011). These strategies can be combined with renewable energy use, including wastewater energy recovery and onsite solar and wind energy generation for achieving net-zero energy and carbon emissions (Novotny 2011). Urban water systems are financially sustainable when their revenues equal or exceed expenses (Rehan et al. 2011). At the minimum, financial resources (revenues) for operation and maintenance costs should be available to make UWSs functional (World Bank 2003). An UWS, particularly water reuse and energy efficiency improvement projects for net-zero water and net-zero energy should be economically sustainable. However, there are a number of challenges for the economic sustainability of these projects. These challenges can be classified as (a) high capital and operation costs, (b) unavailable or inadequate incentives associated with the conservation of water resources and the reduction of pollution, (c) no  reward  for  avoided  headwork in water abstraction, (d) relatively long payback period, and (e) low level of revenue from recycled water services (Listowski et al., 2013). Also, current water reuse strategies do not consider important social and environmental benefits and costs traditionally considered as intangible (Novotny 2012).     5 1.2 Research gap Conventional centralized UWSs, specifically wastewater systems, have less flexibility in associated facilities to adapt to changes (e.g., high population fluctuation), a high and long-term capital investment (Bieker et al. 2010), and longer distance between a recovery station and potential users (Wang et al. 2008). On the other hand,  household level wastewater treatment might be appropriate in low-density households, but not in densely populated urban areas due to operational risk and limited space availability (Bieker et al. 2010). To overcome the limitations of centralized wastewater treatment and decentralized household-level wastewater treatment, an intermediate scale of UWS could be appropriate (Asano et al. 2007; Bieker et al. 2010; CCME 2002; Zarski and Ancel 2012). This level is referred to as the small to medium-sized urban water system (SMUWS). Moreover, the number of studies defining sustainability performance indicators (SPIs) for an entire UWS is very limited  and are most studies are confined to a performance assessment (of service) of individual water and wastewater utilities (CWWA 2009; Sydney Water 2013; Water UK 2011). On the other hand, available SPIs are established mainly for large UWSs (Foxon et al. 2002; Van Leeuwen et al. 2012; Van Leeuwen and Marques 2013) and cannot be adopted as is for a sustainability assessment of SMUWSs. The decision makers of UWSs can have multiple alternatives for each stage in urban water planning and management. These alternatives may vary in their sustainability performance. In addition, the decision makers face increasing challenges as UWSs are under pressure from increasing populations (Lallana et al. 2001), lower household occupancy (Inman and Jeffrey 2006) , increasingly severe droughts and floods (IPCC 2014), higher prices of water and energy (Fagan et al., 2010), lifestyle changes related to technology, personal habits and affluence (Princen, 1999; Lallana et al., 2001), and lower overall sustainability (Hellstro 2000). Moreover, climate change will increase the variability of water availability in many regions of the world (IPCC 2014). These factors ultimately affect the sustainability of UWSs.  In particular, water, energy, and carbon emissions can be considered as major elements of urban water sustainability. These elements are intertwined and have complex interactions (Nair et al., 2014; PMSEIC, 2010; Maas, 2009; Kenway, 2013). This interconnection results in a complex web called the water-energy-carbon (WEC) nexus. Because of the tight linkages in a WEC nexus, decisions for one area could have inadvertent consequences on the other. These pervasive  6 interactions require integrated solutions (PMSEIC, 2010; Nair et al., 2014). However, a comprehensive methodology and decision support system (DSS) to quantify the WEC nexus and its dynamic behavior in an UWS at the community level is lacking (Nair et al., 2014; Rothausen & Conway, 2011; Arora et al., 2013; Kenway, 2013; Kenway et al., 2011). Water Distribution Systems (WDS) can consume significant amounts of energy and release GHGs (Hellstro 2000). Neighbourhood densification can reduce per capita water distribution infrastructure, land resources, and water demand (Duncan 1989; Filion 2008; Frank 1989; Gleick et al. 2003). Reduced water demand itself is associated with decreased upstream energy use. These interlinkages suggest that higher residential densities can have reduced water consumption, energy use, and carbon emissions per capita. However, a lesser amount of landscaping in dense residences results in reduced carbon sequestration too. Furthermore, the life cycle cost of WDSs may be lower for neighbourhoods of high residential density (Speir and Stephenson 2002). An integrated study of the WEC dynamics of water distribution and residential landscaping under neighbourhood densification could not be found in the published literature. Reclaimed water use reduces freshwater withdrawal, enhancing water sustainability. However, reclaimed water use pose human health risks. These risks are primarily associated with pathogenic microorganisms (EPHC/NHMRC/NRMMC 2008). Unlike drinking water, no globally accepted standard guidelines exist for reclaimed water. National guidelines for many urban reuses are yet to be developed in numerous countries, including Canada. Based on the long-term goal of the Canadian federal government (Health Canada 2010), and the recommendations of WHO (WHO 2006a), further research is required for investigating and developing reclaimed water use guidelines for specific reuses in non-potable purposes other than toilet and urinal flushing in Canada. In addition, several decision support tools (DST) are available and are in practice for the planning and operation of wastewater treatment plants, such as ECAM tool (GIZ/MENCBNS/IWA 2015), WEST tool (Stokes et al. 2011), QMRAspot (Schijven et al. 2011), and QMRAcatch (Schijven et al. 2015). However, no DSS is flexible and capable enough to evaluate the potential of wastewater treatment and reuse for different purposes simultaneously  7 based on cost, health risk, energy use, carbon emissions, and the amount of reclaimed water production. Furthermore, decision support systems for analysing the site-specific potential of net-zero water incorporating cost are limited due to a newer concept (Joustra and Yeh 2014a). Net-zero water development may result in higher costs (Englehardt et al. 2013; Gassie et al. 2016; Wang and Zimmerman 2015) and energy use (Vieira et al. 2014; Wang and Zimmerman 2015). 1.3 Research objectives The overall goal of this research is to improve the sustainability of urban water systems by developing a WEC-based decision support system (DSS). The proposed DSS can assist municipalities, urban developers, and policy makers to optimize UWSs in terms of water consumption, energy use, carbon emissions, health risk, and cost. In this research, a WEC-based DSS has been demonstrated using SMUWSs. The specific objectives of the research are as follows: 1. Conduct a state-of-the-art review of the existing sustainability performance indicators (SPIs) of small to medium-sized urban water systems (SMUWSs) and identify key SPIs. 2. Conceptualize, build, and validate a model and decision support system for optimizing the water-energy-carbon (WEC) nexus of SMUWSs.  3. Develop a framework to assess the impacts of neighbourhood densification on the WEC nexus of water distribution and residential landscaping system. 4. Develop microbial water quality guidelines for reclaimed water use in various non-potable urban reuses and propose a tool to optimize life cycle cost and human health risk in conjunction with the WEC nexus for fit-for-purpose wastewater treatment. 5. Assess the economic and energy efficiency of SMUWSs in developing net-zero water in different climatic and topographic regions based on the developed WEC model.  8 1.4 Thesis organization Research Objectives 1, 2, 3, 4, and 5 have been achieved in Chapters 3, 4, 5, 6, 7, 8, and 9. The conclusions and recommendations of the thesis are provided in Chapter 10. The organization of the chapters is shown in Figure 1.2.  Chapter 8: Development of Decision Support Tool for Fit-for-Purpose Wastewater TreatmentChapter 6: Impacts of Densification on WEC Nexus of Water Distribution and Landscaping SystemChapter 5: WEC Nexus Modelling of Urban Water Systems (UWSs)Chapter 10: Conclusions and RecommendationsObjective 1Objective 2Objective 3 Objective 4Objective 5Chapter 3: Literature ReviewChapter 4: Identification of Sustainability Performance IndicatorsChapter 1: IntroductionChapter 9: Economic Analysis of WEC Nexus for Assessing Economic and Energy Efficiency of Net-Zero WaterChapter 7: Development of Microbial Quality Guidelines for Reclaimed Water UsesChapter 2: Research Methodology Figure 1.2    Thesis structure and organization  9 1.5 Meta language This thesis has used specific, technical vocabularies that have widely accepted definitions in the scientific and engineering community. However, certain principles and terminologies used in the thesis can have broad meanings. Such terms are specifically defined for the purpose of this thesis to ensure consistent understanding of the work as follows.  Decision support system (DSS): Represents a system of model, tool, and/or framework to aid decision making (e.g., WEC decision support system). It may comprise models, modules, sub-models, calculation tools (e.g., FitWater), and frameworks and is executable.   Framework: Refers to holistic methods (e.g., framework for estimating WEC impact of neighbourhood densification). A framework can be a part of a DSS for a certain component of an urban water system.   Tool : It is an executable model, framework, or a set of methods (e.g., FitWater tool) for a component/s of an UWS. A tool can be a part of a DSS for a certain component of an UWS.   Model: It is a mathematical representation of a system or its component(s) (e.g., WEC model).    Technique, method, and methodology: These terms have  been used interchangeably for applying mathematical and statistical procedures.   Reclaimed water, recycled water, and reused water: These terms have been used interchangeably referring to treated wastewater, rainwater, and/or storm water, which meet specific water quality criteria for beneficial uses.   Wastewater reuse: This  term has been used specifically to refer the use of reclaimed water produced by the treatment of wastewater.   Community: This term has been used to describe the spatial scale served by small to medium-sized urban water systems.   Dollar values ($): Refer to Canadian dollars.     10  Research Methodology This chapter contains a brief description of research methodology, which was used to achieve the research objectives. An overview of methodology to achieve each objective is given in the following sections and the detailed methodology is presented in individual chapters. A research framework (Figure 2.1) shows various methods and their interconnections under different objectives of this research. Several sustainability performance indicators (SPIs) have been identified to assess the sustainability of SMUWSs (Objective 1) and then a system dynamics-based WEC model for SMUWSs has been developed (Objective 2). The SPIs related to WEC nexus, cost, and health risk have been used in the WEC model. As important inputs to the WEC model, optimal residential density, energy use by a water distribution system, and water consumption by an entire community, including residential landscaping, were estimated (Objective 3). The microbial quality guidelines for reclaimed water in urban reuses have been proposed and a tool, called FitWater, has been developed for evaluating fit-for-purpose wastewater treatment and reuse potential (Objective 4). FitWater provides specific input data: reclaimed water quantity, life cycle cost, energy use, and carbon emissions of wastewater treatment and reclaimed water distribution for the WEC model. Finally, an economic analysis of the WEC nexus has been performed by adding a cost module in the developed WEC model. The cost-embedded WEC model has been used for assessing the economic and energy efficiency of developing net-zero water (Objective 5) to enhance the sustainability of SMUWSs.     11 WATER ●  River water● Lake water● Groundwater● PrecipitationDrinking water treatmentAbstractionUrban Water StageTreatment Distribution Use WW treatmentWater distribution● Domestic● Industrial● Commercial● Institutional● Agriculture● WW transport● WW treatment● Disposal                                                       MATERIALS  & ENERGY         Reclaimed water useCARBON EMISSIONSEMBODIED ENERGY, WATER FOOTPRINTRecycle● Water use● Energy use● Carbon emissions● Life cycle costDensification and Water Distribution & Landscaping SystemAssessment of  economic and energy efficiency for developing net-zero water communities● Reclaimed water production● Health risk● Life cycle cost● Energy use● Carbon emissionsWQ Guidelines & Fit-for-Purpose WW TreatmentWater-Energy-Carbon (WEC) Nexus Model of Small to Medium-sized Urban Water System● Journal papers● Government reports● Technical reports● Product specificationsInput DataResidential density, water demand, & water distribution dataWastewater treatment & reuse dataRequired data on UWSEconomic Analysis of WEC Nexus in Planning Net-Zero Communities● Technical● Environmental● Economic● Social● InstitutionalIdentification of Sustainability Performance IndicatorsWater, energy, carbon emissions & cost related SPIsObjective 1 Objective 2 Objective 4Objective 5Objective 3Papers 1, 2 Paper 3 Papers 5,6, 7Papers 8, 9Paper 4Note: Objective # and paper # refer to the part of the framework with the similar background colour. Paper #  corresponds to the list provided in preface. Figure 2.1    Research methodology framework  2.1 Objective 1 The SPIs for assessing the sustainability of SMUWSs were initially identified from the critical review of indicators (SPIs). The SPIs belong to one of the sustainability dimensions: technical, environmental, economic, social, and institutional dimensions. These SPIs were further evaluated based on four criteria: relevance (importance) to sustainability, measurability, data availability,  12 and comparability. Each SPI was rated using the Likert type scale. The relevance criteria was rated using a 5-point linguistic scale (very high, high, medium, low, and very low), whereas the measurability, data availability, and comparability criteria were measured using a 3-point linguistic scale (high, medium, and low). The relevance criteria was classified into five categories to capture the wide variability of rating provided by experts, whereas other three criteria were assssed more objectively having less variability in rating for which three categories were used. A multi-criteria decision analysis (MCDA) technique, called Fuzzy- ELECTRE I (Elimination and Choice Translating Reality I) was applied based on the four criteria in order to rank and select key SPIs. The identified SPIs are specific to the technical, environmental, economic, social, and institutional dimensions of sustainability.  The identified key SIPs can be used to develop an urban water sustainability index. However, the  SPIs related to water consumption, energy use, carbon emissions, and life cycle cost of SMUWSs have been used in modelling the WEC model. The SPIs are variables, which are used in the consuctrion of the WEC model under Objective 2.       2.2 Objective 2 A WEC model for SMUWSs has been developed using system dynamics to assist municipalities, urban developers, and policy makers for neighbourhood water planning and management. The WEC model used several parameters and variables in modelling, including the SPIs related to water consumption, energy use, and carbon emissions indentified under Objective 1. The model comprises urban water stages: water abstraction, treatment, distribution, use, wastewater treatment, and water recycling if any. Water can be abstracted from river, lake, groundwater, and/or harvested from rainfall. The water consumption, energy use, and carbon emissions from each urban water stage have been included in the model. The water users comprise residential, commercial, institutional, and industrial sectors in the model. The dynamic WEC model and decision support system is for an operational phase of an UWS.  The required input data were collected from journal papers, government reports, technical reports, and product specifications. The model was validated and applied to an existing SMUWS. The necessary data can be classified into three levels: regional (R) containing municipal (site-specific) and regional data; national (N), and global (G). For example, required data (R, N, and  13 G) for the application of the developed DSS to the City of Penticton is given in in Appendix B.2 (Table B.1). The developed WEC model is a broad framework consisting many variables affecting the WEC nexus of urban water systems. For the WEC model, an optimal residential density of a neighbourhood can be estimated using the framework developed under Objective 3. Also, the reclaimed water quality guidelines proposed under Objective 4 can be used to guide users in finding an appropriate level of wastewater treatment required for various reuse applications. Specifically, the developed tool under Objective 4, called FitWater, can be used to estimate the values of variables, such as quantity of reclaimed water, energy use of wastewater treatment, and the realted carbon emissions. These data can be used as inputs for the WEC model developed under Objective 2. 2.3 Objective 3 A framework has been proposed to study the impacts of neighbourhood densification on the WEC nexus of water distribution and residential landscaping system. For this study, water demand was estimated for each neighbourhood design and a water distribution system (WDS) was designed for the corresponding neighbourhood design. The energy use by WDSs was estimated using the prepared WDS design. The related carbon emissions of energy use were estimated using the carbon emission factor of energy source (electricity). The carbon sequestration by residential landscaping, especially by trees, shrubs, and soil was estimated based on their annual carbon sequestration rates. The net carbon emissions were estimated as the balance of the carbon emissions and carbon sequestration. The water consumption, energy use, and net carbon emissions were aggregated by converting them into a common measurement unit – ecological footprint. In addition to the WEC nexus, the study has included the life cycle cost (LCC) of WDSs. The LCC was estimated as the sum of capital cost, operation cost, and repair and replacement cost of WDSs. The WEC nexus-based results were compared with the LCC.  A framework has been proposed to estimate the WEC-based optimal density. The framework was applied to a planned neighbourhood to demonstrate the effectiveness of the methodology. The framework can be used to estimate optimal residential density to be used as an input to the  14 WEC model of Objective 2. In addition, the estimated water demand, energy use, and net carbon emissions of water distribution can be used as inputs to the WEC model of Objective 2.    2.4 Objective 4 Probabilistic risk-based guidelines have been proposed for microbial quality of reclaimed water in various non-potable urban reuses. Health risk was estimated using quantitative microbial risk assessment. The estimation used two-dimensional Monte Carlo simulation to characterize variability and uncertainty in input data. The proposed guidelines were successfully applied to existing wastewater treatment effluents in the Okanagan Valley (BC, Canada). The guidelines help to identify an appropriate level of wastewater treatment required for various reuse applications. The level of treatment determines energy use and related carbon emissions from wastewater treatment plants.   In addition, a tool, called FitWater, has been developed for the evaluation of fit-for-purpose wastewater treatment and reuse potential in communities. The evaluation is based on the criteria: life cycle cost, health risk of reclaimed water, reclaimed water quantity, energy use, and the related carbon emissions. Uncertainty analysis was performed using probabilistic and fuzzy-based methods. The proposed FitWater tool was tested with the existing wastewater treatment plants and then implemented to an actual community. The tool can be used to develop and rank alternative wastewater treatment train and reuses for planning reclaimed water use in a community.  The proposed reclaimed water quality guidelines guide users to find an appropriate level of wastewater treatment for various reuse applications. In particular, FitWater provides the quantity of reclaimed water production, energy use, carbon emissions, and LCC of wastewater treatment and reuse. These data can be used as inputs to the WEC model of Objective 2. 2.5 Objective 5 The economic analysis of the WEC nexus has been conducted by adding a cost module in the WEC model developed under Objective 2 using system dynamics. The cost embedded WEC model also incorporated uncertainty anlysis using Monte Carlo simulations. The cost module includes the LCC of water conveyance, water treatment, distribution, wastewater collection,  15 treatment, and reclaimed water distribution. The LCC was estimated as the sum of capital cost, operation cost, and repair and replacement cost of drinking water system, wastewater treatment system, and reclaimed water use system. The cost embedded WEC model was applied to a community to assess the economic and energy efficiency of net-zero water in different climatic and topographic regions. Also, the impacts of annual precipitation amount (climate), conveyance length (source water proximity), and net elevation head (topography) on energy use and cost of NZW development were assessed in detail in this study. The cost embedded WEC model, i.e., the extended WEC model than that under Objective 2 can be used to estimate water consumption, energy use, carbon emissions, and LCC of SMUWSs. This objective included an extended WEC nexus analysis incorporating LCC and advanced WEC nexus analysis including uncertainty (Monte Carlo simulations).     16  Literature Review A version of this chapter has been published in Clean Technologies and Environmental Policy journal with the title “‘Socializing’ sustainability: A critical review on current development status of social life cycle impact assessment method” (Chhipi-Shrestha et al., 2015a). This chapter contains the state-of-the-art literature related to this research and identifies drawbacks and limitations of current practices in SMUWSs. The chapter includes sustainability assessment approaches, namely life cycle sustainability assessment and sustainability performance indicators; water-energy-carbon nexus; system dynamics modelling; and net-zero water comprising impact of neighbourhood densification, reclaimed water use, and fit-for-purpose wastewater treatment. 3.1 Life cycle sustainability assessment (LCSA) Sustainable development and sustainability are ideas that have been widely used since the 1980s in response to the negative impacts of development, policies, and strategies on the environment and society (UNEP/SETAC 2011; Turcu 2013; Fiksel et al. 2014). Sustainability has three main pillars, namely environment, economic, and social (Valdivia et al. 2011), which are referred to as the triple-bottom-line (TBL) (Sikdar 2007; Vinodh et al. 2012). Integrating life cycle thinking in product or process development with the TBL approach challenges the conventional waste management and pollution prevention mindset that mainly focuses on the factory site (UNEP/SETAC 2011). This new perspective avoids shifting the problem from one phase to another and from one geography to another (UNEP/SETAC 2009). This integrated approach is referred to as life cycle sustainability assessment (LCSA) (UNEP/SETAC 2011). LCSA is “the evaluation of all environmental, social, and economic negative impacts and benefits in decision-making processes towards more sustainable products throughout their life cycle” (UNEP/SETAC 2011). LCSA has three components: life cycle assessment (LCA), life cycle cost analysis (LCCA), and social life cycle assessment (S-LCA) (Klopffer 2003; UNEP/SETAC 2011). 3.1.1 Life cycle assessment (LCA) Life cycle assessment is a technique that assesses potential environmental impacts of a product or service over its life cycle. This technique identifies opportunities to improve the  17 environmental performance of products at different points in the life cycle, from raw material extraction to use, end-of-life treatment, recycling, and final disposal. LCA is performed in four stages: 1) Goal and scope definition, 2) Life cycle inventory (LCI) analysis, 3) Life cycle impact assessment (LCIA), and 4) Interpretation. Stage 1 includes the definition of purpose, functional unit (reference unit to which inputs and outputs are related), and system boundary. Stage 2 consists of data collection and calculation to quantify related inputs and outputs. Stage 3 involves the evaluation of the significance of potential environmental impacts based on the LCI results. Stage 4 includes the interpretation of findings from the LCI analysis and impact assessment. LCA is an environmental management technique that considers the entire life cycle of a product (ISO 2006a) (ISO 2006b). 3.1.2 Life cycle cost analysis (LCCA) Life cycle cost analysis is an economic assessment method in which all costs arising from owning, operating, maintaining, and ultimately disposing of a project or product are considered (US DOE 1996). LCCA is defined “traditionally as the estimation of the total cost associated to an asset over time, including investment, operation, maintenance and major repairs and disposal” (Termes-Rifé et al. 2013). LCCA is a powerful economic tool and is particularly suitable for the evaluation of alternatives. LCCA assesses the long-term cost effectiveness of a project better than other economic methods that focus only on initial costs or on short-term operating costs (US DOE 1996). UNEP/SETAC (2008) identified three types of life cycle cost (LCC): conventional, environmental, and societal LCC based on the 2004 survey of 33 LCC studies from 1984 to 2003. Conventional LCC is “the assessment of all costs associated with the life cycle of a product that are directly covered by the main producer or user in the product life cycle. The assessment is focused on real, internal costs, sometimes even without end of life or use costs if these are borne by others” (UNEP/SETAC 2008). It is a quasi-dynamic method and neglects external costs. The reference flow is mostly one unit of product (e.g., a building), which is easier for computation but may not be the most appropriate for sustainability assessment. Environmental LCC is “an assessment of all costs associated with the life cycle of a product that are directly covered by one or more of the actors in the product life cycle (supplier,  18 manufacturer, user or consumer, and/or end of life actor), with the inclusion of externalities that are anticipated to be internalized in the decision relevant future” (UNEP/SETAC 2008). It is LCA driven and steady-state in nature. Taxes and subsidies are included in environmental LCC if relevant (UNEP/SETAC 2008), whereas, societal LCC is “an assessment of all costs associated with the life cycle of a product that are covered by anyone in the society, whether today or in the long-term future. Societal LCC includes all of environmental LCC plus additional assessment of further external costs, usually in monetary terms (e.g., based on willingness-to-pay methods)” (UNEP/SETAC 2008). This method uses an expanded system boundary unlike other LCC types and comprises more costs including damage costs that will or could occur in the long term. It is quasi-dynamic in nature (UNEP/SETAC 2008). The conventional LCC has been in practice throughout history for many government offices, public organizations, and firms  (UNEP/SETAC 2008). The LCCA method is applicable to urban water systems. The stages of UWSs from water abstraction to wastewater disposal have several alternatives. These alternatives usually vary in water and energy performance, life span, initial investment cost, and operation and maintenance cost. LCCA can be applied to identify a cost effective decision for long-term scenarios. 3.1.3 Social life cycle assessment (S-LCA) Social life cycle assessment (S-LCA) is a technique that assesses the potential social impacts of a product or service caused throughout its life cycle. S-LCA refers to the assessment of the real and potential social and socio-economic impacts of products or services including positive and negative impacts along their life cycle (Dreyer et al. 2010; Feschet et al. 2012; UNEP/SETAC 2009). Primarily, social impacts are related to human capital, human wellbeing, cultural heritage, socio-economy, and social behaviour (UNEP/SETAC 2009). S-LCA complements both LCA and LCCA in terms of sustainability assessment (UNEP/SETAC 2009). S-LCA has similar applications to LCA, such as sustainability labelling, sustainability management, and assessment of technology alternatives considering social aspects. In S-LCA, the area of protection is human dignity and wellbeing (Hauschild et al. 2008). More specifically, the area of protection is autonomy, well-being-freedom, and fairness based on a capability approach (Reitinger et al. 2011). The ultimate goal of S-LCA is the wellbeing of stakeholders over a product’s life cycle  19 (UNEP/SETAC 2009). Similar to LCA, the general framework for S-LCA consists of the same four stages (ISO 2006a; UNEP/SETAC 2009).  3.1.4 Sustainability Performance Indicators for UWSs Urban water utilities provide water to city dwellers through the following processes: water abstraction, treatment, distribution, use, wastewater treatment, and disposal. These human regulated urban water processes constitute a human hydrologic cycle or an urban water system (UWS) (Bagley et al., 2005). UWSs need to resolve emerging water issues related to population growth, urbanization and climate change (Inman and Jeffrey 2006; Lim et al. 2010; IPCC 2014). On the other hand, although UWSs achieve the fundamental requirements of providing clean drinking water and removal of wastewater to a higher degree, this sector is criticized from the sustainability perspective (Hellstro 2000). Sustainability, or sustainable development, in the 21st century is guided by Agenda 21, established by the UN Conference on Environment and Development in 1992 (UN 1992). Based on the Agenda 21 (UN 1992) and other literature such as  Hellstro (2000) and Engel-yan et al. (2005), the major objectives of a sustainable UWS are to: a) provide clean and safe drinking water; b) reduce environmental impacts; c) develop an economically efficient system; and d) optimize water and other natural resource uses. These sustainability objectives can be viewed through five interrelated sustainability dimensions and are briefly described below (Daigger 2009; Harmancioglu et al. 2013; VanLeeuwen and Marques 2013; World Bank 2003). i. Technical: The technical dimension refers to the reliable and proper functioning of UWS technologies and neighbourhood design. This dimension includes sustainability evaluation criteria, such as “neighbourhood location and design” and “water infrastructure and fixtures”.  ii. Environmental: Water resources face many threats, such as pollution and resource depletion due to climate change. These threats ultimately affect the reliability of the resource, i.e., quality and quantity of drinking water supply. On the other hand, water supplies and wastewater facilities themselves threaten the environment through the unsafe disposal of wastewater, emissions of pollutants, and over-consumption of resources. This  20 dimension includes sustainability evaluation criteria such as “resource utilization”, “environmental impacts”, and “resource recovery”.  iii. Economic: UWSs can only function if financial resources are available to meet the operation and maintenance costs, at a minimum. The economic dimension includes evaluation criteria, such as “water economics” and “wastewater economics”. iv. Social: UWS services need to satisfy consumers’ needs and expectations and also promote their health.  This dimension includes sustainability evaluation criteria such as “service provision” and “public health”.  v. Institutional: A community needs an institution in order to keep its UWS operational in order to serve consumers. This dimension includes the “governance and progress criteria”.  The performance or achievement of SMUWSs can be assessed using indicators (Murray et al. 2009). A sustainability performance indicator (SPI) is a parameter, or a value derived from parameters, which provides information about the sustainability achievement of an activity, a process or an organization (CWWA 2009). In particular, the literature outlined in Table 3.1 was found to be more relevant for the sustainability assessment of SMUWSs and was reviewed in detail. Table 3.1    Major literature used for the screening of the SPIs for SMUWSs SN UWS and Water Utility Services Sustainability SN Neighbourhood and City Sustainability 1 City Blueprints (Van Leeuwen et al. 2012) [24] 10 LEED-ND (USGBC 2013) [19] 2 SI-UWS (Popawala and Shah 2011) [20] 11 BREEAM Communities (BREGL 2012) [11] 3 SCDS (Foxon et al. 2002)[62] 12 CASBEE-UD (IBEC 2008) [13] 4 ESI (Lundin and Morrison 2002) [15] 13 ECC (EarthCraft 2014) [22] 5 UWOT (Makropoulos et al. 2008) [22] 14 SCR (SCR 2009) [6] 6 UWCSS (Van Leeuwen and Marques 2013) [35] 15 Asian Green City Index (Siemens AG 2011) [7] 7 SI (Water UK 2011) [21] 16 Global City Indicators (World Bank 2008) [10] 8 PI (CWWA 2009) [30] 17 European Green City Index (Siemens AG 2009) [4] 9 BSS (Sydney Water 2013) [22] 18 IoS (SCI 2012) [4] Note:  Number of indicators used is mentioned in square brackets [no. of indicators].    21 3.2 Water-energy-carbon (WEC) nexus The criteria and indicators related to the WEC nexus for SMUWSs are discussed in the sections below: 3.2.1 Energy for water Energy is required for anthropogenic water use. The energy requirement for each urban water process significantly differs based on topography, technology (Tuladhar et al. 2014; Venkatesh et al. 2014), source water quality (Santana et al. 2014; Tuladhar et al. 2014), social factors (Venkatesh et al. 2014), and operational conditions, especially in water distribution (Cabrera et al. 2010; Giustolisi et al. 2016; Iglesias-Rey et al. 2016; Nardo et al. 2014). For example, in California (US), water supply and conveyance has a very wide range of energy intensity, ranging from zero to 3646 kWh/ML, indicating a higher dependency of energy use on supply methods (like ocean water desalination, and surface water and groundwater withdrawal) and conveyance distance. Similarly, wastewater collection and treatment also has a higher energy intensity, ranging from 291 to 542 kWh/ML, with variation mainly attributed to a variety of treatment methods (CEC and NC 2006). From the life cycle perspective of an UWS, the operational phase has been identified as the most energy intensive phase (Friedrich 2002; Nair et al. 2014). The operational phase of water treatment consumes 94% of total energy use and is responsible for 90% of total GHG emissions (Racoviceanu et al. 2007).  3.2.2 Energy from water Energy can be generated from water. Water and wastewater flow contain kinetic energy, potential energy (Fontana et al. 2012), thermal energy, and chemically bound energy, especially in wastewater, all of which can be harnessed. The amount of kinetic and potential energy depends on topographical conditions. For instance, even at a height of 50 m, the potential energy content of water or wastewater is only 6 kWh per capita per year (Meda et al. 2012). Thermal energy in wastewater is mainly stored due to warm wastewater generation, from hot water showers, laundry, and dishwashing, and contains more energy than potential energy (Meda et al. 2012). Therefore, the greatest potential for heat recovery is from greywater, i.e., wastewater from baths, showers, laundry machines, and possibly kitchen sinks. For example, in Germany, the typical household’s greywater generation rate is 40L/p/day with a temperature difference of  22 300C, which can generate 509 kWh/capita/yr (Meda et al. 2012). This energy is higher than potential energy (6 kWh per capita per year, even at a height of 50 m height).  Chemically bound energy can be estimated from the carbon content, i.e., chemical oxygen demand (COD). On the basis of a daily COD load of 110 to 120 g/capita, the maximum theoretical energy content is approximately 146 kWh/capita/yr, under the assumption that all COD could be transferred to methane and be utilized (Meda et al. 2012). For comparison purposes, energy consumption of wastewater treatment plants ranges from 27 to 37 kWh/capita/yr in Canada (AECOM 2012), indicating the potential that wastewater treatment plants could be energy self-sufficient. However, the practically recoverable energy will be lower than the theoretical recoverable energy.  3.2.3 Water for energy Water is required directly and indirectly for energy generation, which is the water footprint of energy. The water footprint of a product refers to the volume of freshwater used to produce the product, measured over the full supply chain (Hoekstra et al. 2011). Water is used directly for electricity generation by hydropower; however, a large amount of indirect water is used for the exploration, extraction, and beneficiation of fossil fuels, depending on the specific method used (Meda et al. 2012). Indirect water is also used for renewable energy production, such as energy crop cultivation (for biofuel production). Similarly, thermal power plants also use steam (water) directly for driving turbines and indirectly for cooling purposes, i.e., heat dissipation. The quality and quantity of used water, i.e., wastewater from energy generation is an important factor (Meda et al. 2012). Some processes, such as hydraulic fracturing of oil and gas wells and power plant operation (e.g., ash handling) pollute water more than others (e.g., only the temperature increases in the water used in cooling towers). Additionally, the water used for crop cultivation is not directly available to further reuse, whereas the water used in cooling towers is available after its primary use. These processes indicate that the selection of energy types for urban water activities have different water implications.    23 3.2.4 GHGs from energy and water Energy use and wastewater treatment processes release greenhouse gases (GHG). The GHG emissions or simply carbon emissions of energy uses differ by energy source and its generation method, such as fossil fuel, hydropower, thermal power, etc. Energy (e.g., grid electricity) generation methods are location specific and their carbon emissions vary. For instance, carbon emissions from grid electricity generation is 57 times higher for Alberta electricity (824.4 kg CO2e/MWh) than for BC Hydro (14.4 kg CO2e/MWh) (Ministry of Environment 2013). This is because 83% of the grid electricity is produced from fossil fuels (coal and natural gas) in Alberta (Alberta Energy 2014), whereas BC hydro produces 95% of its electricity from hydropower (BC Hydro 2015). In addition, GHGs are also released from wastewater treatment processes due to the microbial degradation of organic matter present in wastewater (IPCC 2006).  The WEC nexus is complex. Energy is needed for water production and energy can also be harnessed from wastewater. Water is needed for energy generation. Both energy use and wastewater processes emit GHGs. Because of this tight inter-linkage in the WEC nexus, decisions in one area could have inadvertent consequences in another (Rothausen and Conway 2011). Thus, only a holistic and generic model can capture the variability and dynamics of UWSs (Nair et al. 2014). 3.2.5 WEC models The WEC model can broadly be categorized into static and dynamic models as follows:  3.2.5.1 Static models Researchers have been working on WEC nexus and some have proposed static models and tools. GIZ/MENCBNS/IWA (2015) has developed an Excel-based Energy performance and Carbon emissions Assessment and Monitoring (ECAM) Tool for evaluating the energy performance and carbon emissions of water and wastewater utilities. However, the tool is only for utilities, and excludes energy performance in indoor water use. Similarly, Venkatesh et al. (2014) extensively studied the WEC nexus of four water utilities of Nantes (France), Toronto (Canada), Turin (Italy) and Oslo (Norway). The study lacks the inclusion of water consumption, which is a very important component of a UWS.  Gu et al.  (2016) investigated the WEC nexus of nine wastewater treatment plants in China, but they are limited only to a specific component of an  24 UWS, i.e., wastewater utility. Similarly, Stillwell et al. (2010) examined the WEC nexus of Texas and estimated potential water and energy savings by implementing water conservation and reuse practices. Arora et al. (2013) researched the life cycle energy use and GHG emissions of alternative urban water supply strategies for the Melbourne Metropolitan region in Australia using static models. Furthermore, the WEC nexus of UWSs including ten water and wastewater utilities in seven cities in Australia and New Zealand was studied by Kenway et al. (2008). Similarly, the carbon cost model was proposed by Reffold et al. (2008) for estimating GHG emissions and their cost for different water supply and demand options in the UK. Some researchers only studied the water-energy nexus component of the WEC nexus for several community elements. Cutter et al. (2014) examined the water-energy nexus of the UWS of California as a means to evaluate the cost effectiveness of strategies for water utilities. Cheng (2002) investigated the water-energy nexus of residential buildings in Taiwan to evaluate the energy savings due to water conservation measures. Similarly, Abdallah and Rosenberg (2014) studied the water-energy linkages of residential indoor water use to determine the implications for water and energy conservation and management in the United States. Malinowski et al. (2015) investigated the energy-water nexus of integrated water management (IWM) measures, i.e., rainwater harvesting and greywater reuse to estimate the energy and cost saving at the national and local scale in the United States. Likewise, Pacetti et al. (2015) explored the water-energy nexus of biogas production from three energy crops: maize, sorghum, and wheat in Italy to estimate the water requirements for biogas from these energy crops. UDWR (2012) researched the water-energy nexus in Utah to determine the interconnection of these two resources at the state level.  EPRI (2002) and Gu et al. (2014) broadly explored the water-energy nexus at the national scale, respectively for the United States and China in order to determine the interdependency and sufficiency of these two resources in the future. In addition, Stillwell (2015) researched three energy-water nexus bills of the U.S. Congress to assess the sustainability of these public policies. None of the above mentioned frameworks and models can perform dynamic analysis and all lack  25 system feedbacks. Also, they have limited flexibility for practical application and many of them do not include all the components of UWSs. 3.2.5.2 Dynamic models An integrated urban water system was dynamically modelled by Fagan et al. (2010) to develop sustainability assessment framework. However, the model was developed by using rigorous mathematical equations. Other dynamic approaches used for urban water modelling are agent-based modelling (ABM) and system dynamics. Agent-based modelling has some disadvantages compared to system dynamics, e.g., higher data requirement for calibration (Gebetsroither-geringer 2014). In addition, the results of agent-based modelling are more difficult for evaluation, resulting in greater efforts required for model validation (Gebetsroither-geringer 2014) as revealed by researchers, such as Fagiolo (2006) and Werker and Brenner (2004). System dynamics has been used in only a few urban water studies (Zarghami and Akbariyeh 2012).  3.3 System dynamics modelling (SDM) System dynamics is a well-established methodology to quantify complex feedbacks in system interactions (Forrester, 1961; Forrester, 1968). The methodology was initially developed by Forrester (1961). A system refers to “a collection of elements that continually interact over time to form a unified whole”. Dynamics means change over time, where the values of variables and parameters change over time. Therefore, system dynamics is a methodology used to understand how a system changes over time (Martin 1997a). The system dynamics model (SDM) is often used to quantify system behaviors with feedback loops for more accurate projections (Qi and Chang 2011). The model allows for the effective trade-off analysis of multi-scenarios and the multi-attributes of the WEC nexus over time (Sehlke and Jacobson 2005). System dynamics involves the construction of “stock and flow diagrams” to mimic a dynamic system.    26 3.3.1 System dynamics model construction 3.3.1.1 Stock and flow diagram System dynamics computer simulation programs such as STELLA provide a framework and easy-to-understand graphical interface to study the quantitative interaction of variables within a system. The interaction can be modelled using four building blocks as given in Figure 3.1.  Figure 3.1    Representation of a stock, flow, converter, and connector  A stock is used to “represent anything that accumulates or drains over time” (e.g., water accumulating in a bucket). A flow is “the rate of change of a stock”. A converter is used to “take input data and manipulate or convert input into some output signal(s)”. “A connector is an arrow that allows information to pass between two converters, stocks and converters, stocks and flows, and converters and flows” (Martin 1997a). Mathematical equations are developed by combining two fundamental ideas (Roberts 2001). First, a stock at present time equals the stock at a certain previous time, plus the change in the stock (net flow) that occurred over the specified time interval. Second, the change during a certain time interval equals the length of the interval, multiplied by the rate of change per time interval. The combination of these two ideas produces the following equation of the present stock at time ‘t’ (Roberts, 2001; Porwal, 2013): Stock (at present time, t) = Stock (at certain previous time) + (Length of the time interval) * Rate of stock change  = ∫ (𝑺𝒕𝒐𝒄𝒌 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆) 𝒅𝒕𝒕𝟎      Equation 3.1    27 3.3.1.2 Feedback “Feedback is a process whereby an initial cause ripples through a chain of causation ultimately to re-affect itself” (Martin 1997b). Feedback occurs when an output of a system is fed back into the system as an input (Martin 1997a). Feedback can occur in an open-loop system or closed-loop system; however, the latter one is more common. Feedback systems can be classified as positive or negative. Positive feedback system moves in the same direction to produce compounding or reinforcing behaviour. These systems drive growth and change (Martin 1997b). For instance, as shown in Figure 3.2, reproduction increases the rabbit population. The growth occurs by the birth of rabbit. The number of births per time depends directly on how many rabbits are already in the area considered and increases with the growth of the population size.  The shaded arrows show the causal links and not material links or information links.    Figure 3.2    Positive feedback in growing rabbit population On the other hand, a negative feedback system “moves in opposite directions to produce balancing or stabilizing behaviour”. These systems negate change and stabilize systems (Martin 1997b). In positive feedback, a variable is eventually increased as a result of the increase in that variable. Whereas, in a negative feedback system, an increase in a variable eventually result in a decrease in that variable. For instance, Figure 3.3 shows a declining skunk population. Every year, a fraction of the total skunk population declines. The number of deaths per time (year) depends directly on the initial skunk population and decreases gradually with the decline of the population.      28    Figure 3.3    Negative feedback in declining skunk population A causal loop diagram (CLD) is developed before developing a complete SDM. A CLD is the foundation of a SDM, and is used to identify relationships between individual system components and show feedback loops that affect system regulation (Nasiri et al. 2013). In the CLD, a ‘‘+’’ sign indicates a positive (reinforcing) relationship, whereas a ‘‘-’’ sign indicates a negative (balancing) relationship between two variables (Nasiri et al. 2013).  3.3.2 SDM for urban water management A critical review of system dynamics-based urban water models was performed, and the strengths and weaknesses of the developed models are given in Table 3.2.  This critical review shows system dynamics has been applied to different aspects of urban water from county to city level but not applied to the WEC nexus for neighbourhoods or a community. Also, urban water processes perform differently in various geographic regions with respect to energy and carbon emissions. This requires a holistic and generic model to capture the variability and dynamics of UWSs (Nair et al., 2014). The WEC nexus model comprising interacting problems can be developed using system dynamics (Nasiri et al., 2013; Nair et al., 2014). The systems dynamics model can assist decision makers in understanding the implications of investment decisions and actions on a SMUWS (Kenway et al., 2011).   29 Table 3.2    Strengths and weaknesses of system dynamics based urban water models Reference Scale Objectives/Strengths Weaknesses Zarghami and Akbariyeh (2012) City Determine the most effective policy to manage water demand & supply; considers different water sources, population, and cost Not considered carbon emissions & energy analysis Willuweit and O’Sullivan (2013) City Model effects of urban development and climate change on urban water cycle; considers social, environmental, economic, & functional indicators Lacks a feedback loop between land use and water balance models Zhang et al.  (2008); Zhang et al. (2009a) City Identify an optimal plan; considers population, industry, agriculture, and water resource Not considered energy analysis, carbon emissions, and water reuse Nasiri et al. (2013)  County Plan and manage reclaimed water use; considers cost, technology, and environmental factors Focused on water, not considered energy, carbon emissions, & other benefits of water reuse, e.g., reduced energy use Karamouz et al. (2012) City Assess reliability of water quantity and quality; considers daily time step Not considered energy and carbon emissions and energy analysis; not calculated model accuracy Qi and Chang (2011) County Estimate domestic water demand under changing macro-economy; considers socio-economy, population, and water Not considered energy, carbon emission, and water reuse Wang (2014) Province Effect of water price and wastewater purification ratio on the water demand; considers population, water demand and supply; based on World Water Model Not mentioned accuracy, not considered energy and carbon emission Zhang et al. (2009b) City Predict the WRCC*; considers water resource; industrial, agricultural and residential water; and WWT & reuse Not considered energy analysis and carbon emissions; not validated Tong and Dong (2008)  City Analyze structure and functions of socio-economic-environmental system; considers domestic, industrial, agricultural, ecological water, and pollution control; used PSR** framework for SDM  Not considered energy, carbon emissions and water reuse Nawarathna et al. (2009) Catch-ment Predict future water supply and demand under changing land use and climate; considers irrigation, urban and environmental water demand Not considered energy and carbon emissions; model not validated Wang (2013) Country Water and energy nexus; implications of biofuel development on water and energy Model at national scale; preliminary model; focus on biofuel *Water Resource Carrying Capacity;** Pressure, State, and Response Framework    30 3.4 Net-zero water (NZW) Historically, the net-zero concept evolved from the building energy budget and its popularity has extended the concept to waste generation, carbon emissions, and water consumption (Joustra and Yeh 2014a). The concept of net-zero water (NZW) is similar to the carrying capacity of a system (Holtzhower et al. 2014). NZW is the balance of water demand and supply within a given areal boundary (Holtzhower et al., 2014). The US Army states “net-zero water limits the consumption of freshwater resources and returns water back to the same watershed so not to deplete the groundwater and surface water resources of that region in quantity or quality over the course of a year” (US Army 2011). The central theme of NZW  emphasizes a balance so that the sum of all input water is offset by comparable output water (Joustra and Yeh 2014a). NZW presupposes that a community system can secure an adequate water supply within its boundaries, typically from surface water, groundwater, reclaimed water, and rainfall (Holtzhower et al. 2014). Achieving net-zero water similar to the natural cycle requires both the conservation of water and the creation of balanced water feedback loops (Joustra and Yeh 2014a). The nuances of NZW are given in Table 3.3. This study has considered the widely accepted definition proposed by the US Army (2011) for NZW. NZW can also be interpreted as a sustainable tolerance of comfortable minimum use. The sustainable minimum use of water is approximated as 70 L/person/day, of which 20 L/p/d is for potable use and 50 L/p/d is for human sanitation and disease prevention according to the World Water Council and the United Nations Development Program (Ma 2014). A minimum use of water would lead to the development of NZW conveniently for communities. Moreover, Net-Positive Water (NPW) can also be developed with the generation of positive water balance. NZW or NPW explores the capabilities of communities for sustainable planning and use of water.      31 Table 3.3    Definitions of net-zero water and its nuances Definitions Source Net-Zero water “Net-zero water limits the consumption of freshwater resources and returns water back to the same watershed so not to deplete the groundwater and surface water resources of that region in quantity or quality over the course of a year.”  US Army (2011) “Annual potable water use is no greater than annual rainfall” Olmos and Loge (2013) “A sustainable tolerance to comfortable use of minimum water. The sustainable minimum use of water is approximated as 70 L/person/day.” Ma (2014) Zero water Zero-water compliance requires that the building water cycle operates independently from water and wastewater municipal systems. On-site wastewater recycling is crucial to zero-water success, and alternative water supplies are limited by the regional climate. However, the preservation of the local ecosystem must be considered when collecting precipitation for on-site use.  Joustra and Yeh (2014a) Life-cycle zero water It requires that the embodied water required for the manufacture and transport of materials be considered over the building lifetime. Achievement of net-zero water over the building lifetime may be an unachievable objective without innovative techniques for on-site renewable water generation.  Joustra and Yeh (2014a) Net-Positive water “One hundred percent of the project’s water needs [except for regulated potable uses] must be supplied by captured precipitation or other natural closed -loop  water  systems  and/or by recycling  used  project  water, and must be purified as needed without the use of chemicals.”  International Living Future Institute (2016) Net-positive water balance in buildings as a result of restorative impacts. Joustra and Yeh (2015)  A report published by the US National Research Council stated that “The use of reclaimed water to augment potable water supplies has significant potential for helping to meet future needs, ….” and also recommended potable reuse with or without an environmental buffer as an alternative water management approach (National Research Council 2012). Similarly, water recycling for the augmentation of drinking water supplies has been promoted by the Australian government, who has been extensively applying reclaimed water and published guidelines for reclaimed water quality management (EPHC/NHMRC/NRMMC 2008). Also, in Canada, the provincial government of British Columbia (BC) has planned for the mandatory construction of dual water-plumbing (additional purple pipes for reclaimed water flow) in new buildings (BC Ministry of Environment 2008). Moreover, the BC government has endorsed a BC Wastewater Regulation that allows reclaimed water use in non-potable and potable reuses after treatment with the  32 approval of local health authorities (BC Ministry of Environment 2013; MWR 2012). These initiatives show an increasing aspiration for reclaimed water use.  Actual NZW has been practised in several places in the world. For instances, the commercial building “The Bullitt Center” in Seattle, Washington is a NZW building that has been operational since 2013 (Crosson 2016; Killough 2016). Another pilot scale NZW building is a four-bedroom university residence hall unit that was built in 2012 (Englehardt et al. 2013; Gassie et al. 2016). At a larger scale, Namibia (Asano et al. 2007; Crook et al. 2005; WABAG 2016) and Singapore (Angelakis and Gikas 2014; Asano et al. 2007) have been utilizing reclaimed water for drinking. Namibia has been applying such practice since 1968 (Asano et al. 2007). Similarly, Cyprus reuses 100% of their wastewater (EU 2015, 2016), whereas Israel (Angelakis and Gikas 2014; Crook et al. 2005) and Malta (Crook et al. 2005; EU 2016) reuse approximately 80% of their wastewater. NZW or NPW could be achieved by using wastewater recycling, rainwater harvesting, and storm water harvesting (Englehardt et al. 2013). However, NZW development may have higher associated costs (Englehardt et al. 2013; Gassie et al. 2016; Wang and Zimmerman 2015) and energy (Vieira et al. 2014; Wang and Zimmerman 2015). Therefore, economic viability and environmental sustainability of NZW is of high concern. In fact, these features are highly location specific, and the national Australian water reuse study recommended to evaluate water reuse project individually using a DSS (DSEWPaC 2012). There are a number of studies that have proposed methodologies or DSSs for evaluating NZW potential. The studies are summarized in Table 3.4.     33 Table 3.4    DSSs for NZW analysis DSSs Applications Limitations Reference ZeroNet decision support system (DSS)  for the San Juan River Basin Focus on drought planning and economic analysis at a watershed level; analyze critical water supply and demand information and assist water utilities in planning for water management in shortages Not included energy consumption by water use; not included carbon emission & features such as water recycling and rainwater harvesting Rich et al. (2005) NZW alternative analysis methodology Assist Department of the Army (US) in evaluating alternatives to achieve NZW in six pilot installations; multi-criteria-based; included criteria cost, environmental variables, and water quality, etc. Methodology only at a conceptual stage; criteria are broadly presented  Payosova and Deason (2012) Urban/suburban NZW treatment process for buildings Based on laboratory experiments, mass balance, & kinetic model; experiments on four-bedroom university residence hall unit; NZW development with 10-20% rainwater make-up; capital cost of $ 6.20/m3 (20 years of life); operational and maintenance cost of $1.83/m3 Initial design than a decision support tool; only at a building level Englehardt et al. (2013) Offsetting of water conservation costs to achieve NZW Reveal potential cost savings in both utility energy conservation and energy reductions in decreased hot water use; selling of carbon offset credits is feasible; NZW at single-family building and community level is feasible. Conceptual case study; lacks dynamic analysis of water; not included wastewater treatment and its cost, energy use, & carbon emissions Olmos and Loge (2013) NPW matrix Proposed guidelines for clean and dirty water management in buildings to develop NPW quality and quantity; NPW matrix helps to identify the means to develop domestic architecture for NPW in buildings Only at a building level and does not include cost, energy use, and carbon emissions Ma (2014)  Integrated Building Water Management (IBWM) Model NZW decision support tool using system dynamics; flexible; and dynamically tracks potential water flows into, within, and from buildings Only at a building level and do not include cost, energy use, and carbon emissions. Joustra and Yeh (2014a); Joustra and Yeh (2014b) NZW potential  assessment Analyze and map the NZW potential of the US; based on urban area clusters (UAC)   Preliminary analysis; not included surface and groundwater sources, temporal variation, cost, & carbon emissions; not an executable tool Holtzhower et al. (2014) Conceptual framework for NPW Develop NPW buildings Only at preliminary stage Joustra and Yeh (2015)  34  Almost all of the listed studies were conducted at the building scale. Among the existing methodologies, some are conceptual frameworks, whereas others lack one or more components such as energy, cost, or carbon emissions. For instance, Ma (2014) proposed a NWP matrix to identify the means to develop domestic architecture for NPW in buildings and Joustra and Yeh (2014a) developed a NZW decision support tool, Integrated Building Water Management (IBWM) Model. Both of the proposed methods are only at a building scale and do not include cost, energy use, and carbon emissions. Also, Joustra and Yeh (2015) proposed a framework for NPW buildings, which is a conceptual framework and at a preliminary stage. Englehardt et al. (2013) proposed an urban/suburban NZW treatment process for buildings based on laboratory experiments and Gassie et al. (2016) presented the results of two years of operation of the treatment process showing the effective NZW system. The proposed treatment process was a “first design” and at a building level rather than a decision support tool at a community scale.  ZeroNet Water-Energy Initiative  developed the ZeroNet DSS for the San Juan River Basin with a focus on drought planning and economic analysis (Rich et al. 2005). The DSS does not include energy use by water consumption. Also, the DSS neither estimates carbon emissions nor includes the features such as water recycling and rainwater harvesting (reuses) that are important for developing NZW. Guo et al. (2016) developed a model for the analysis of economic feasibility Urban/suburban NZW treatment process for buildings Present results of two years of operation of the system showing the effective NZW system with 85% recycling rate; system was projected to be capable of energy-positive operation A “first design” rather than a decision support tool; only at a building level Gassie et al. (2016) NZW at a building level  in severe drought prone areas  Achieve NZW at a building level in severe drought prone areas of Los Angeles, California; based on an office building of 250 Full Time Equivalent employees Although in-depth analysis, not included cost, energy use, and related carbon emissions; only at a building level Crosson (2016) Model for economic feasibility of large-scale NZW management Minimize cost and energy use for NZW systems; minimize the cost for recycling wastewater at the rate of $2.95/m3 Not included  carbon emissions and dynamic interaction of urban water components Guo et al. (2016)  35 of large-scale NZW management; however, it does not include carbon emissions and dynamic interaction of urban water components. The cost, energy, and health risk of NZW are affected by various components of SMUWSs, including neighbourhood density. These components are reviewed as given below. 3.4.1 Impact of neighbourhood densification on WEC nexus A Water Distribution System (WDS) comprises transmission mains, distribution pipelines, and pumping stations. WDSs are affected by residential landscaping practices that consume a significant amount of water. For instance, the water demand of residential landscaping ranges from 30% of domestic demand in coastal areas to 60% in hot inland areas in California, US (Gleick et al. 2003). This value has been shown to be as high as 77% in the Okanagan Valley, BC, Canada (OBWB 2016). WDSs have significant effects on energy use and related greenhouse gas emissions (Hellstro 2000). Energy use is the primary cost factor in the operation of water supply systems consuming approximately 80% of municipal water processing and distribution costs (EPRI 2002). For groundwater systems, almost all of the energy cost is associated with pumping except where ion exchange and physical or chemical treatment is required (Energy Center of Wiscosin 2003). In particular, water conveyance and distribution only consumed 1.6% of the total energy use, i.e., 386 GWh/year in the City of Toronto in 1998, indicating a high energy use for pumping (Cuddihy et al. 2005). Moreover, trees, shrubs, and soil of urban residential landscaping have significant carbon sequestration potential (Lal and Augustin 2012; Zirkle et al. 2011). These features show WDS and residential landscaping are connected in terms of water consumption, energy use, and net carbon emissions forming Water Distribution and Residential Landscaping System (WDRLS).  Energy requirements for WDSs depend on several factors, including service area topography (Bolognesi et al. 2014; Teixeira et al. 2016; Tuladhar et al. 2014), source water (Energy Center of Wiscosin 2003), urban form (Filion 2008; Speir and Stephenson 2002), population density (Filion 2008; Speir and Stephenson 2002), and adopted management strategies (Bolognesi et al. 2014; Teixeira et al. 2016). Numerous researchers have investigated the reduction in energy use in WDSs and many of them focussed on the optimization of distribution system operations. For example, energy cost optimization (Alighalehbabakhani et al. 2013); energy metrics of WDSs  36 (Dziedzic and Karney 2015); effects of different management strategies on energy consumption (Cherchi et al. 2015); life cycle energy use of water distribution pipes (Filion, Maclean, & Karney, 2004); and effects of raw water source on energy use by water conveyance (EPRI 2002).  Some researchers studied the influence of topography on energy use (Guo and Englehardt 2014), while others investigated the effects of housing patterns (i.e., single-family) on public water and sewer costs (Speir and Stephenson 2002) and effects of urban form (i.e., configuration) on water distribution energy (Filion 2008). All these studies assumed a constant rate of water use. The constant rate of water use in various residential densities is very different than reality, where landscaping water demand is high and differs much with residential densities. Residential density is affected by the mix of single-family (SF) and multi-family (MF) buildings. A maximum lot coverage (meaning the lot area covered by buildings) is typically lower in low density SF buildings than high density MF buildings. For instance, a maximum  lot coverage in SF buildings is typically around 23% in the US (Zirkle et al. 2012; Zirkle 2010), around 30% in the most areas in Toronto (City of Toronto 2016); 40% in Penticton (City of Penticton 2015a), Kelowna (City of Kelowna 2007), and District of Peachland (District of Peachland 2014a), and 60% in Saskatoon (City of Saskatoon 2016). Similarly, a maximum lot coverage in high density MF buildings is 70% in Calgary (City of Calgary 2012) and Kelowna  (City of Kelowna 2007), and 100%  in Peachland (District of Peachland 2014a) and Penticton (City of Penticton 2015a) with landscaping on unbuilt land. These regulatory requirements indicate that per capita water use varies highly with residential densities. 3.4.2 Reclaimed water use The challenges in water supply, e.g., growing population, variability in source water, etc. result in increasing water demands and competition among water utilities even across the provincial and national boundaries (Schaefer et al. 2004). In recent times even in Canada, seasonal water shortages have been experienced in various regions. Several cities in BC and Alberta, such as Vancouver, the Cowichan Valley, Penticton, and Calgary experience water restrictions in summer (Gulerian 2015; Water Conservation Company 2015). Water restriction is a municipal regulation to restrict water use in relatively less important activities. Some cities may even reach the severe  37 Stage 3 restriction prohibiting certain water uses, such as lawn irrigation, park irrigation, residential vehicle washing, street cleaning, and outdoor decorative water features.  Water resource management requires careful planning to address urban water shortages and the associated uncertainties. Supply-side and demand-side management are core strategies for water resources management (Kanta and Zechman 2014; Schaefer et al. 2004).  Supply-side management includes water availability augmentation, water infrastructure expansion related to water, and new water source development, whereas demand-side management incudes water conservation activities, leakage control, and price setting (Kanta and Zechman 2014). Under supply-side management, reclaimed water use is an option. Reclaimed water refers to the municipal wastewater that is treated to meet specific water quality criteria, especially intended for beneficial uses. The term recycled water is also synonymously used for reclaimed water (Asano et al. 2007). Reclaimed water is an on-site water resource that can be generated at or near the vicinity of urban water consumption. Reclaimed water can be used for various purposes after treatment. 3.4.2.1 Global water reuse status and trend Reclaimed water is the treated municipal wastewater that meets specific water quality criteria, primarily intended for beneficial uses. The major drivers triggering water reuse are lack of water, management of drought impacts, freshwater saving for first-use that demands high water quality, use of cheaper water sources, water reuse as low cost disposal option for wastewater, and water restoration to the environment (EU 2016; Jiménez Cisneros 2014). Globally, 7000 Mm3/year of reclaimed water was used after treatment in 2011, which comprised 0.59% of the total water use (EU 2016). More than 60 countries have applied reclaimed water for different uses (Angelakis and Gikas 2014). The amount of reclaimed water use in different regions and countries was reviewed and outlined in Table 3.5. Based on the total annual volume, China, Mexico and the United States (primarily, California, Florida, Texas, and Arizona) use the highest amount of reclaimed water in the world (Angelakis and Gikas 2014). However, China and Mexico reuse wastewater with little or no treatment similar to Pakistan (Table 3.5). The intensity of water reuse per capita was highest in Cyprus, Qatar, Israel, and Kuwait (EU 2016; Jiménez Cisneros 2014). Kuwait, Israel, and Singapore ranked first in terms of proportion of reuse with respect to total freshwater use volume (Jiménez Cisneros 2014). Furthermore,  38 California, Singapore, and Japan are probably pioneers with respect to technological advancement in water reuse  (Angelakis and Gikas 2014). Also, Table 3.5 shows major types of water reuse applications comprising non-potable and potable uses in different countries. Non-potable reuse is common although potable reuse has been in practice in Namibia and Singapore. In particular, agricultural irrigation is a primary application reusing 32% of reclaimed water in the world. Other possible water reuses are landscape irrigation (20%), industrial uses (19%), urban uses (8%), environmental enhancement (8%), recreational uses (7%), groundwater recharge (2%), indirect potable use (2%), and others (2%) (EU 2016; Lautze et al. 2014). However, groundwater recharge and indirect potable reuse have high potential for future reuse (EU 2016). People have a historical practice of reclaimed water use across the world. The trend of reclaimed water use has been increasing worldwide and Global Water Intelligence estimated that the world market of water reuse is expected to surpass desalination in the future. The estimation shows that water reuse will represent 1.66% (26,000 Mm3/year) of the total global water use by 2030 (EU 2016). For example, the trend of reclaimed water use in some countries is given in Figure 3.4. Generally, the quantity of reclaimed water use has been increasing in the United States (California), Australia, and Europe since the 2000s or before.     39 Table 3.5    Status of global water reuse Country Reuse# (Mm3/yr) % of WW reused Major applications Reference World 26,000       in 2030 1.66 in 2030*  EU (2016) North & Latin America     United States 3850 - - Angelakis and Gikas (2014) California (US) 1271 - AI, LI, GWR, IU RI, WH Asano et al. (2007) Florida (US) 834 54 Asano et al. (2007) Canada (BC) - 3 AI Schaefer et al. (2004) Mexico 350,000 ha  AI@ Crook et al. (2005) EU 1100 2.4  EU (2016); EU (2015) Spain 347 ~10 AI, IU, TF, LI, PIPU, RI Crook et al. (2005); EU (2015, 2016) Italy 233 ~8 Crook et al. (2005); EU (2015, 2016) Cyprus 20 100 EU (2015, 2016) Germany 42 ~1 Crook et al. (2005); EU (2015, 2016) Malta ~4 ~78 Crook et al. (2005); EU (2016) Australia 300 16.8  DSEWPaC (2012) New South Wales 63 9.8 LI, TF, AI, SI, IU, VW, CU, ENV DSEWPaC (2012) Victoria 100 24.1 ” Queensland 71 23.7 ” South Australia 22 28.1 ” Western Australia 19 12.0 ” Tasmania 3 6.2 ” Northern Territory 1.5 6.0 ” Austr. Capital Territory 3.5 13.3 ” Middle East     Israel 300 ~80 AI, GWR Crook et al. (2005); Angelakis and Gikas (2014) Qatar 760 - AI, LI MDPS (2016) Iran 70 5 AI Crook et al. (2005) Kuwait 52 - AI, LI Crook et al. (2005) United Arab Emirates 500 20 AI, LI Crook et al. (2005) Saudi Arabia 657 10 AI, LI, IU Drewes et al. (2012); WHO (2005); Crook et al. (2005)  Asia     China 7373 9.2 IU, LI, AI, TF @ Zhou et al. (2011) Japan 187 - TF, IU, ENV, AI, Crook et al. (2005) Korea 157 4** IU, TF,CL Crook et al. (2005) Singapore 27 - DW (2.5%) & NPW Angelakis and Gikas (2014) Pakistan - 80 AI@ Crook et al. (2005) Southern Africa     South Africa >45 3  Crook et al. (2005) Namibia 7.67 4** DW blending WABAG (2016); Crook et al. (2005) #Mm3/year unless stated  *% of total water use  **% of water supply @ Little or no treatment AI: Agricultural irrigation, LI: Landscape irrigation, GWR: Groundwater recharge, IU: Industrial use, RI: Recreational impoundment, WH: wildlife habitat, TF: Toilet flushing, PIPU: Planned indirect potable use, SI: Silviculture, VW: Vehicle washing, CU: Constructional use; ENV: Environmental applications (streamflow augmentation, dune stabilization, etc.), CL: Cleaning, DW: Drinking water, NPW: Non-potable water    40  Figure 3.4    Trend of water reuse in different regions of the world 3.4.2.2 Microbial quality of reclaimed water Reclaimed water use poses human health risks, primarily associated with pathogenic microorganisms, disinfection byproducts (DBPs), and pharmaceutical and personal care products (PPCPs). Pathogenic microorganisms in water primarily originate from sewage (feces) contamination (EPHC/NHMRC/NRMMC 2008) and also from natural freshwater bodies containing pathogens, such as Lagionella and Aeromonas (Health Canada 2013a). DBPs are created during water disinfection, primarily by the reaction of natural organic matter contained in water and chemical disinfectants (Tian et al. 2013). PPCPs may be present in treated water due to their presence in wastewater, which may not have been effectively removed during wastewater treatment (Kosma et al. 2014). This research is focused only on the human health risk associated with pathogenic microorganisms. Several groups of wastewater microorganisms have been identified as being pathogenic (EPHC/NHMRC/NRMMC 2008): a) Bacteria, e.g., Campylobacter, pathogenic Escherichia coli, Shigella, Lagionella, Salmonella, and Vibrio cholera; b) Viruses, e.g.,  adenovirus, rotavirus, norovirus, enterovirus, and Hepatitis A; c) Protozoa, e.g., Cryptosporidium and Giardia; and d) Helminths, e.g., Taenia (tapeworm), Ascaris (roundworm), Trichuris (whipworm), and Ancylostoma (hookworm). The human health  41 risks posed by wastewater microorganisms have been estimated by quantitative microbial risk assessment (QMRA) since the 1980s (Haas et al. 2014). Globally, standard guidelines do not exist for reclaimed water use. Indeed, the development of a practical guideline is complex. The complexity can be understood from the historical development of the reclaimed water use guidelines by the leading health organization – World Health Organization (WHO). The WHO published Health guidelines for the use of wastewater in agriculture and aquaculture in 1989 as a 76-page report and prescribed microbiological quality guideline values for wastewater reuse in agriculture (WHO 1989). The same organization published WHO guidelines for the safe use of wastewater, excreta and greywater in 2006 in four volumes, with some of them above 200 pages in length for agriculture and aquaculture (WHO 2006a). However, the risk-based four-volume guidelines have not prescribed any guideline value, rather procedures for developing guideline values suitable to local circumstances (WHO 2006a), indicating the practical complexity involved.  Reclaimed water has been used in various urban purposes across the world. The reclaimed water quality guidelines prescribed in different regions of the world could be a practical reference for developing new guideline values. The existing reclaimed water quality guidelines in different regions of the world were critically reviewed and are presented in Table 3.6. The review reveals that different countries and even provinces or states within a country, environmental organizations (e.g., US EPA), and health organizations (e.g., WHO) have proposed their own guidelines. Unlike  drinking water, no internationally accepted standard guideline values exist for reclaimed water.  42 Table 3.6    Reclaimed water quality guidelines for urban reuses in various regions of the world Country Unrestricted urban reuse Restricted urban reuse Urban agriculture: food crops Source North America       Canada (federal) Toilet and urinal flushing: E. coli or thermotolerant: ND (med), max <=200   -  - Health Canada (2010) Canada (BC) Fecal coliforms: med < 1 or < 2.2 MPN; max 14 (Greater exposure potential) Fecal coliforms: Moderate expo- median 100, max 400; Low expo- med 100, max 1000 E. coli (for crops eaten raw): < 1 or < 2.2 MPN MWR (2012) Canada (Alberta)  -  - Tot col. < 1000 (geom of wk samples (if storage provided) or daily samples (if storage not provided); Fecal coliforms <200 Alberta Environment (2000) US Fecal coliforms: not detectable (Med); max 14 Fecal coliforms: med<=200; max <=800 Fecal coliforms: not detectable (med); max 14 US EPA (2012a) US (California) Tot. coli: 2.2 (7-day med); 23 (not more than 1 sample exceeds this value in 30 d); max 240 Total coliforms: 23 (7-d med); 240 (not more than one sample exceeds this value in 30 d) Tot. coli.:  2.2 (7-day med); 23 (not more than 1 sample exceeds it in 30 d); 240 (max) US EPA (2012a) US  (Florida) Fecal coliforms:  75% of samples ND; max 25; Giardia and Cryptosporidium: sampling once each 2-yr  period for plants ≥1 mgd; once each 5-yr period for plants ≤ 1 mgd Not specified Fecal coliforms: 75% of samples ND; max 25; Giardia, Cryptosporidium: sampling once per 2-yr period for plants ≥ 1 mgd; once per 5-yr period for plants ≤ 1 mgd US EPA (2012a) US  (Hawaii) Fecal coliforms: 2.2 (7-day med); 23 (not more than one sample exceeds this value in 30 d); 200 (max) (R1) Fecal coliforms: 23 (7-day med); 200 (not more than one sample exceeds this value in 30 d) (R2) Fecal coliforms: 2.2 (7-day med); 23 (not more than one sample exceeds this value in 30 d); 200 (max) (R1) US EPA (2012a) US  (Nevada) Total coliforms: 2.2 (30-d geom); 23 (max) (Category A) Fecal coliforms: 2.2 (30-d geom); 23 (max) (Category B) Total coliforms: 2.2 (30-d geom); 23 (max) (Category A) US EPA (2012a) US  (New Jersey) Fecal coliforms: 2.2 (wk med); 14 (max) (Type 1 RWBR) Fecal coliforms: 200 (mon geom); 400 (wk geom) (Type 2 RWBR) Fecal coliforms: 2.2 (wk med); 14 (max) (Type 1 RWBR) US EPA (2012a) US (North Carolina) Fecal coliforms or E. coli: 14 (mon mean); 25 (max) (Type 1) Fecal coliforms or E. coli: 14 (mon mean); 25 (daily max) (Type 1) Processed: Type 1; Non-processed: Type 2: Fecal coli. or E. coli: 3 (mon mean); 25 (daily max); Coliphage (virus): 5 (mon mean); 25 (daily max); Clostridium: 5 (mon mean); 25 (daily max) US EPA (2012a)   43 US (Texas) Fecal coliforms or E. coli: 20 (30-d geom); 75 (max); Enterococci: 4 (30-d geom); 9 (max) (Type 1) Fecal coli. or E. coli: 200 (30-d geom); 800 (max); Enterococci: 35 (30-day geom); 89 (max) (Type 2) Fecal coli. or E. coli: 20 (30-d geom); 75 (max); Enterococci: 4 (30-d geom); 9 (max) (Type 1) US EPA (2012a) US (Virginia) Fecal coliforms: 14 (mon geom), CAT > 49; E. coli: 11 (mon geom), CAT > 35; Enterococci: 11 (mon geom), CAT > 24m (Level 1) Fecal coliforms: 200 (mon geom), CAT > 800; E. coli: 126 (mon geom), CAT > 235; Enterococci: 35 (mon geom), CAT > 104 (Level 2) Fecal coliforms: 14 (mon geom), CAT > 49; E. coli: 11 (mon geom), CA> 35; Enterococci: 11 (mon geom), CAT > 24 (Level 1) US EPA (2012a) US (Washington) Tot. coli: 2.2 (7-d med); 23 (max) (Class A) Tot. coli.: 23 (7-d med); 240 (max) (Class C) Total coli.: 2.2 ( 7-d med); 23 (max) (class A) US EPA (2012a) Arizona - Fecal coli: < 200 in last 4 of 7 samples; 800 (max) (Class B) Fecal coliforms:  ND in last 4 of 7 samples (Class A) US EPA (2012a) Australia     National National level guidelines (2006-2009) implemented but lacks specific water quality value recommendation (EPHC/NHMRC/NRMMC (2006, 2008) Western Australia High exposure: indoor, irrigation (lawn), toilet flushing, cold tap washing machines: < 1  Medium expo: Urban irrigation (restricted access), firefighting, water features, dust suppression:  < 10  Irrigation (unprocessed foods):< 1 WA DoH (2011) Queensland Class A (toilet flushing, lawn & golf course irrigation,  water features): <10 (med) Class B (washdown of hard surfaces in agri industries): <100 (med) Class A (food crops with raw consumed foods irrigation: <10  (med) QS EPA (2005) Southern Australia Class A (dual recirculation): <10 (med); unrestricted municipal irrigation: < 10 (med); landscape irrigation: < 1000 Class B (municipal with restricted access): <100 (med) - SA DHA (2012) Victoria Class A (non-potable urban use: toilet flushing, lawn & golf course irrigation,  fountains & water features): <10 (med); viruses: 7-log reduction, Protozoa: 6-log reduction Class C: Urban (non-potable,, controlled public access): < 1000 Class A (food crops with raw consumed foods irrigation : <10; Class C (Processed/cooked food): <1000 (med) EPA Victoria (2003, 2015) Europe     European Union Lack of coherent and comprehensive legislative although some countries have own standards European Commission (2016) Spain  Garden irrigation: 0; landscape irrigation, street cleaning, fire hydrants and car washing: 200 - Food crops eaten raw: 100; crops not eaten raw: 1000 Royal Decree 1620/2007 (DoET and SERI 2014)   44 France Public green spaces (parks, golf courses): <= 250 (sanitary level A ) Crops, vegetables processed by industrial heat; sold cut flowers:  <= 10,000 (sanitary level B) Crops, vegetables not processed by industrial heat: <= 250 (sanitary level A) Decree of France (2016) Greece - Restricted irrigation (no public access) & crops (processed before consumption): <= 200 (med) Unrestricted irrigation for all crops such as vegetables (raw eaten), vines: <= 5 (for 80% of samples) and <=50 (for 95% of samples) Ilias et al. (2014) Italy - - Vegetable crops: 10 Lonigro et al. (2015) Middle East and Asia       UAE (Abu Dhabi) Unrestricted: 100 Restricted: 1000  Food crops: 100 UAE Guideline 2010 (DoET and SERI 2014) Jordan Public parks and road sides: 100; ground water recharge: <2.2 Landscape irrigation: 1000 Cooked food crops: 100 Jordanian Standards (JS:893/2002) (WHO 2005) Kuwait - - Crop irrigation with raw eaten (100 coliforms); Crops not eaten raw (10,000 coliforms)  WHO (2005) Saudi Arabia - - Unrestricted irrigation: coliforms 2.2  WHO (2005) Mediterranean regions Residential (toilet flushing, vehicle washing, gardening) & urban reuses (parks, golf courses, firefighting & recreation impoundments (pond & stream except bathing)): <=200 - Irrigation of vegetables, fruit trees, landscape impoundments without public contact: <=1000 Mediterranean guidelines (proposed) EMWater (2001) Japan Toilet flushing: ND; sprinkling water: ND; recreational water: ND Landscape irrigation: <=1000 as coliforms groups - Tajima (2007) WHO (Global) Safe Use of Wastewater 2006 for agriculture and aquaculture; complex for practical application, not recommended specific water quality values WHO (2006b) Unit: cfu/100 mL and is for E. coli unless stated; mon is monthly, wk: weekly, med: median, max: maximum ND: Not detectable, geom: geometric mean; mgd: million gallons daily, RWBR: Reclaimed Water for Beneficial Reuse, expo: exposure, Tot.: Total, coli.: coliforms, CAT (Corrective Action Threshold) = A bacterial, turbidity or total residual chlorine standard for reclaimed water at which measures shall be implemented to correct operational problems of the reclamation system within a specified period.  45  3.4.3 Fit-for-purpose wastewater treatment Different water reuse applications require various grades of water quality, resulting in a number of required treatment levels. The production of higher quality water than required can result in overtreatment, leading to unnecessary cost and over use of resources such as energy. DSEWPaC (2012) suggests that a water reuse project cost must be determined on a case by case basis. A wastewater treatment train for a water reuse project can be selected based on the end use of reclaimed water for achieving economic efficiency and environmental sustainability (US EPA 2012a). Such treatment is referred to as fit-for-purpose wastewater treatment. It aims to avoid overtreatment, and obviously under-treatment as it is legally prohibited. Water quality depends on the level of water and wastewater treatment, which is dictated by the end use of reclaimed water.  Wastewater treatment technologies differ mainly in terms of cost (Guo et al. 2014), treatment efficiency (Health Canada 2010), energy consumption (Chang et al. 2008), and the related carbon emissions. Wastewater treatment technologies also affect the efficiency of water recycling, i.e., the volume of reclaimed water produced, especially when influent wastewater is highly polluted. All above factors determine the required level of fit-for-purpose wastewater treatment, which requires a decision support tool (DST) for the evaluation of treatment trains for a community. The DST helps in ranking and identifying a cost-effective, risk-acceptable, and energy efficient treatment train to meet the water quality for an intended use. Several DSTs are available and are in practice for the planning and operation of wastewater treatment plants. The DSTs related to fit-for-purpose wastewater treatment were reviewed and summarized in Table 3.7. The review reveals that various DSTs have their own objectives and applications. For example, an Excel-based ECAM tool was developed for evaluating the energy performance and carbon emissions of water and wastewater utilities (GIZ/MENCBNS/IWA 2015) and the WEST tool was developed to assess the environmental effects including water use, energy use, and carbon emissions of water and wastewater infrastructure (Stokes et al. 2011). Both of these tools lack the capability to estimate health risk associated with the treated water used for a specific purpose and rank the corresponding wastewater treatment chains. Some researchers developed QMRA tools, such as QMRAspot (Schijven et al. 2011) and QMRAcatch 46  (Schijven et al. 2015). The tools were developed only to assess the health risks associated with water use. However, these tools have included either drinking water or recreational water only and they cannot be used to screen various treatment processes. Therefore, no DST is flexible and capable enough to evaluate the potential of wastewater treatment and reuse for different purposes simultaneously based on cost, health risk, energy use, carbon emissions, and amount of reclaimed water production. The decisions with regards to the planning of reclaimed water use projects for a specific reuse application should consider these major factors: quantity, quality, cost, energy, and carbon emissions (NASEM 2016; Nasiri et al. 2013; Zarghami and Akbariyeh 2012). This requires a DST to evaluate alternative wastewater treatment trains and reuses. However, such a tool is not available in the publically accessible literature. The national water reuse assessment report of Australia also revealed the need of a similar DST for a high-level evaluation of reclaimed water reuse projects, called hotspot analysis, across the country (DSEWPaC 2012).    47  Table 3.7    Existing DSTs related to fit-for-purpose wastewater treatment     DSTs Applications Limitations Reference Energy performance and Carbon emissions Assessment and Monitoring (ECAM) Evaluate energy performance and carbon emissions of water and wastewater utilities    Lacks health risk estimation for specific reuse; not capable to rank alternative wastewater treatment chains GIZ/MENCBNS/IWA (2015) Water-Energy Sustainability Tool (WEST) Evaluate environmental impacts of life cycle of water and/or wastewater infrastructure ” Stokes et al. (2011) water-energy nexus of UWS Evaluate cost effectiveness strategies for water utilities in California ” Cutter et al. (2014) Carbon cost model Estimate GHG emissions and associated cost in water supply and demand options in UK ” Reffold et al. (2008) Online design tool Automatically designs a preliminary WWTP (activated sludge or food chain reactor) using the provided input: country (region); hydraulic capacity or population equivalent, and generic effluent criteria; design provides an overall energy intensity and an equipment list Limited effluent criteria options; not specific to a particular water reuse; does not estimate cost Organica Water Inc. (2016) QMRAspot Assess microbial risk of a drinking water production chain from surface water to potable water;  determine drinking water treatment efficiency related to the legislative health-based target Does not rank specific treatment processes to overcome the target risk Schijven et al. (2011) QMRAcatch Catchment model  to assess the health risks associated with E. coli, enterovirus, norovirus, Campylobacter and Cryptosporidium in water resources in a catchment Includes only recreational and drinking water; does not rank treatment processes  Schijven et al. (2015) Recycled water Irrigation Risk Analysis (RIRA) Evaluate human health risks of reclaimed water use in irrigation; users can choose pathogens from the given list in the tool and input their concentrations in water or foods; deterministic model Only for irrigation water; not able to recommend corresponding treatment units; does not estimate energy use and cost  Hamilton et al. (2007) QMRA Wiki Provides fundamental information, steps, and online calculators to conduct QMRA;  intended to be a reference source for the QMRA community Only for risk assessment; users should know background knowledge on risk assessment; calculators cannot recommend treatment units and cost for a specific water reuse CAMRA (2016) Energy Use Assessment Tool Evaluate energy and cost of small and medium sized water and wastewater utilities Not applicable to select an optimum treatment chain for safe water reuse in a specific reuse application US EPA (2012b) 48   Identification of Sustainability Performance Indicators A version of this chapter has been published in Water Environment Research journal with a title “Sustainability performance indicators for small to medium sized urban water systems: A selection process using Fuzzy-ELECTRE method” (Chhipi-Shrestha et al., 2017a). 4.1 Background Urban Water Systems (UWSs) can be viewed at different spatial scales: building, neighbourhood, community, city, and metropolis. Sustainability issues of UWSs can vary with these scales. In particular, the recovery of resources, such as energy, water, and nutrients, from wastewater is important from the urban water sustainability perspective. The recovered resources are required to be distributed efficiently to households to meet their needs. Therefore, the distance between a recovery station and potential users, especially for the reclaimed water use is a critical factor (Wang et al. 2008). However, such a distance is longer in centralized UWSs. In addition, conventional centralized UWSs, specifically wastewater systems, have poor flexibility in associated facilities, a high and long-term capital investment (Bieker et al. 2010). On the other hand,  household level wastewater treatment may be appropriate in low-density households, but not in densely populated urban areas due to operation risk and limited space availability (Bieker et al. 2010). The limitations of the centralized level and that of decentralized level (household level) wastewater treatments could be overcome using an intermediate scale of UWS. CCME (2002), Asano et al. (2007), Bieker et al. (2010), and  Zarski and Ancel (2012) have also indicated the suitability of UWSs at such scale.  This level is referred to as the “urban community” scale or small to medium-sized urban water systems (SMUWSs). 4.1.1 Small to medium-sized urban water systems (SMUWSs) The population size in small to medium-sized urban water systems (SMUWSs) depends on its location and should be guided by the principle “as small as possible, as big as necessary”  (Bieker et al. 2010). For example, Bieker et al. (2010) proposed a community size of 50,000 to 100,000 or even lower population, whereas  Böhm et al. (2011) proposed a population of 20,000 for the community in China. Since, drinking water is a major water input to an UWS and also 49  primarily determines the magnitude of wastewater amount, the drinking water system can be considered as a basis of size classification of UWSs as given in Table 4.1. Table 4.1   Classification of UWSs based on US EPA (2009a) System size Population served Small <3,300 Medium 3,300 – 100,000 Large >100,000  Broadly, the population size of a SMUWS can be up to 100,000; however, the community should be compact as a lower dwelling density in a neighbourhood and a higher dispersion of neighbourhoods increase the cost of providing water and wastewater services (Speir and Stephenson 2002). A SMUWS serves a group of neighbourhoods that can be a town, a small city, a municipality, or a part of any of these. In Canada, the proportion of SMUWs is very high. For example, the municipalities with the population of 5,000 or less are above 80% (FCM and NRC 2005). The SMUWSs are different from large UWSs and have the following characteristics:  Small service area and population  Smaller infrastructure  Limited data availability   Institutional limitations because a SMUWS may cover only a small part of a municipality  Low technical capability in terms of staff and equipment in smaller urban communities   Limited financial resources in smaller urban communities The sustainability of UWSs can be assessed using sustainability performance indicators (SPIs). Available SPIs are established mainly for large UWSs (Foxon et al. 2002; Van Leeuwen et al. 2012; Van Leeuwen and Marques 2013) and cannot be adopted as is for a sustainability assessment of a SMUWS due to different characteristics as listed above. This chapter addresses the gap and aims to develop a set of applied indicators to assess the holistic sustainability of small to medium-sized UWSs. The UWSs can be existing or new.   50  4.1.2 Fuzzy sets and fuzzy numbers Zadeh (1965) introduced fuzzy sets to analyze uncertainty caused by imprecision and vagueness in decision making.  Fuzzy sets are a useful tool for modelling language to approximate a system having fuzzy phenomena (Chhipi-Shrestha et al. 2016; Chu 2011).  The fuzzy set A can be represented as: A = {(x, fA(x)) / x ε U}               Equation 4.1  where U is the universal set,  x is an element in U, A is  a  fuzzy  set  in U,  fA(x) is the  membership function of  A at  x.  The larger fA(x), the stronger the grade of membership for x in A (Chhipi-Shrestha et al. 2016; Chu 2011).  Similarly, a real fuzzy number A is described as any fuzzy subset of the real line R with membership function fA.  The membership function fA of the fuzzy number A can be expressed as:  𝑓𝐴(𝑥) =  {𝑓𝐴𝐿(𝑥), 𝑎 ≤ 𝑥 ≤ 𝑏1,          𝑏 ≤ 𝑥 ≤ 𝑐𝑓𝐴𝑅 (𝑥), 𝑐 ≤ 𝑥 ≤ 𝑑0,        𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒}        Equation 4.2 where fAL(x) and fAR(x) are left and right membership functions of A respectively, a ≤ b ≤ c ≤ d, and A can be represented by (a, b, c, d).  Various fuzzy numbers can be used depending on the condition, but triangular fuzzy numbers (TFNs) are commonly used due to computational simplicity (Sevkli 2010). In this study, TFNs were used. TFNs can be defined as a triplet (p, q, r), where the parameters p, q, and r indicate the smallest possible value, the most promising value, and the largest possible value, respectively,  which describe a fuzzy event (Sevkli 2010). A triangular fuzzy number ?̃? = (p, q, r) is given in Figure 4.1 (Chhipi-Shrestha et al. 2016).    51   Figure 4.1    Membership function of  ?̃? The mathematical operations of TFNs for the two positive triangular fuzzy numbers (a1,b1,c1) and (a2,b2,c2) are given below: A TFN (a,b,c) is said to be a positive TFN if and only if a ≥ 0. (a1, b1, c1) + (a2, b2, c2) = (a1+ a2, b1+ b2, c1+ c2)        Equation 4.3 (a1, b1, c1) * (a2, b2, c2) = (a1* a2, b1* b2, c1* c2)       Equation 4.4 (a1, b1, c1) * k = (a1* k, b1* k, c1* k), where k ≥ 0       Equation 4.5  4.2 Methodology A comprehensive literature review was performed using keywords search, through web-based scientific search engines and online databases. Several researchers have used specific keywords for searching literature for conducting comprehensive reviews (Jørgensen, 2013; Yi and Chan, 2014). The keywords include: urban water sustainability, sustainability indicators of urban water, city water sustainability, water sustainability assessment, neighbourhood sustainability assessment, city sustainability assessment, water sustainability, sustainability performance of water, and community water. These keywords were searched in databases such as Compendex Engineering Village, Web of Knowledge databases, electronic library of the University of British Columbia, Canada, and web-based search engine http://scholar.google.ca/. Since, the sustainability of an UWS at the community scale or SMUWS is to be assessed, three categories of literature have been reviewed: a) sustainability assessment of UWSs, b) performance assessment of water and wastewater services that focus on sustainability, and c) neighbourhood and city sustainability assessment.    52  The SPIs were selected based on the methodological framework (Figure 4.2) consisting the initial screening, development of selection criteria, Delphi method, and multi-criteria decision analysis using Fuzzy-ELECTRE I.  Figure 4.2    Methodological framework used for the selection of SPIs 4.2.1 Initial screening of SPIs  An initial screening of SPIs for SMUWSs was performed by a simple checklist method considering smaller infrastructures, data limitations  (Haider et al. 2014a), small population, small service area, and institutional limitations of urban communities. The screened SPIs were categorized into five sustainability dimensions: technical, environmental, economic, social, and institutional dimensions  (World Bank 2003; Van Leeuwen and Marques 2013).  4.2.2 Development of selection criteria Four selection criteria developed for the selection of initially screened SPIs are relevance (importance) to sustainability, measurability, data availability, and comparability (adapted from Initial screening of SPIs from literature Development of selection criteria for SPIs Evaluation of SPIs based on selection criteria in fuzzy rating scale scale  Form a decision matrix Determination of criteria weight by fuzzy AHP Determine fuzzy concordance and discordance indices Normalize the decision matrix Construct a weighted normalized fuzzy decision matrix Rank alternatives according to their final indices 53  Lundin 2002 and  Haider et al. 2014a) and their ratings are based on Likert type scale (Tveit 2009) as given in Table 4.2. a. Relevance: How much an indicator is relevant and comprehensive to the sustainability of the small to medium-sized urban water systems? It is related to the technical, environmental, social, or economic relevance and the comprehensiveness of many features of the sustainability dimension. b. Measurability: How much a variable is measurable accurately and requires the extent of observations for the calculation of indicators? c. Data availability: How is the availability of the data for indicator calculation? d. Comparability: How much the value of SPI is comparable with the available reference value? It is related to whether the indicator is used for urban water sustainability assessment in the region and/or at the international level. The relevance criteria of SPIs was rated by using a 5-point linguistic scale (very high, high, medium, low, and very low), whereas the measurability, data availability, and comparability criteria were measured by using a 3-point linguistic scale (high, medium, and low). The relevance criteria was categorized into five categories in order to capture the wide variability of rating provided by experts, whereas other three criteria were evaluated more objectively having less variability in rating for which three categories were used. Similar approach was also used by Haider et al. (2014b), Hung et al. (2010), and Tveit (2009). Since these ratings are linguistic and imprecise, their calculation was performed using fuzzy sets. Zadeh (1965) introduced fuzzy sets to analyze uncertainty in decision making caused by imprecision and vagueness. The fundamentals of fuzzy sets, fuzzy number, and their mathematical operations (Chhipi-Shrestha et al. 2016)  are provided in Section 4.1.2.     54  Table 4.2    Criteria for the selection of SPIs Criteria (score1)  Description - Relevance  -  - Very high (0.7,1,1) - Must be included, SPI is highly relevant and more comprehensive indicator of sustainability of SMUWSs. - High (0.5,0.7,0.9) - SPI is highly relevant and of average comprehensiveness. - Medium (0.3,0.5,0.7) - SPI is of average relevance and average comprehensiveness. - Low (0.1,0.3,0.5) - SPI has low relevance to the sustainability of SMUWSs. - Very low (0, 0, 0.3) - SPI seems to be irrelevant for sustainable development for SMUWSs. - Measurability  -  High (0.5,1,1) Variables have absolute values or one annual observation provides the data (e.g., proximity to drinking water system) for the variables. Medium (0,0.5,1) Variables have highly varying values that require a large number of observations in a year. Low (0,0,1) Variables have qualitative data or have estimated values. Data availability   High (0.5,1,1) Data are available in public annual municipal reports (water, wastewater, and financial reports) and official water master plan. Medium (0,0.5,1) Data are available in raw form in internal official records. Low (0,0,1) Data are only available in occasional study reports or rarely available. Comparability   High (0.5,1,1) SPI has been used for urban water sustainability assessment in the region (country). Medium (0,0.5,1) SPI has been used for urban water sustainability assessment outside the region.  Low (0,0,1) SPI has rarely been used for urban water sustainability assessment. (Adapted from: Lundin (2002), Tveit (2009) and  Haider et al. (2014a))  The relevance criteria was evaluated based on a group decision of experts using the Delphi method. The other three criteria were rated based on the reported literature. For the rating of the data availability criteria, five small to medium-sized municipalities were randomly selected from a list of small to medium-sized municipalities by coding and then using a statistical random table (Gibbons et al. 1999)  in each of three large provinces of Canada: Ontario, Alberta, and British Columbia. The selected municipalities are the Cities of Belleville, Brantford, North Bay, St. Thomas, and Stouffville of Ontario; the Cities of Airdrie, Leduc, Lethbridge, Red Deer, and Spruce Grove of Alberta; and the Cities of Parkville, Prince George, Penticton, Vernon, and District of Kitimat of British Columbia. Their public annual municipal reports (water,                                                  1 Triangular Fuzzy Number (TFN) represent by lowest possible(l), middle (m), and highest possible (u) values for all scales 55  wastewater, and financial reports) and official water master plan as far as available were referred to perform the ratings of the data availability criteria.  For the rating of the comparability criteria, national reports were used.  Comparability was considered to be high if a SPI is  available in the national reports, i.e., National Water and Wastewater Benchmarking Initiative (NWWBI) (AECOM 2012) or Municipal Water Use Report (Environment Canada 2011). The NWWBI report was prepared based on the assessment of 44 wastewater utilities, 41 water utilities, and 17 storm water management programs, whereas the municipal water use report was prepared based on the data of 2,779 Canadian municipalities. Similarly, the comparability criteria was considered to be medium if a SPI is used for urban water sustainability assessment by any international literature identified as the major literature in Table 3.1 (Chapter 3). The comparability criteria was considered low if a SPI has rarely been used for urban water sustainability assessment. Furthermore, the weights of the four selection criteria were determined based on a group decision of experts using the Delphi method as explained in detailed in Appendix A.1. The weights of these selection criteria were determined by the fuzzy-Analytical Hierarchical Process (F-AHP) (Kaya and Kahraman 2011) using a group decision method. The calculated weights of the relevance, measurability, data availability, and comparability criteria in terms of TFNs are (0.19, 0.32, 0.49), (0.20, 0.27, 0.35), (0.18, 0.24, 0.35), and (0.11, 0.17, 0.27) respectively. The consistency ratio (CR) of the comparison matrix is 0.0076 that indicates a consistent matrix as the ratio is lower than 0.10 (Alonso and Lamata 2006). In addition, CRs were less than 0.1 for the comparison matrices of all participants. 4.2.3 Fuzzy-ELECTRE I The rating of each SPI for the four criteria was converted to fuzzy scores using TFNs (Hung et al. 2010; Chen et al. 2008) as given in Table 4.2. A decision matrix was formed using these scores and then normalized. An outranking method called fuzzy- ELECTRE I (ELimination Et Choix Traduisant la REalite´, i.e., Elimination and Choice Translating Reality I) was applied for ranking and selecting SPIs. The ELECTRE method was developed by Bernard Roy in the late 1960s. The method uses concordance and discordance indices to determine outranking relations among the alternatives. Concordance and discordance indices can be viewed as satisfaction and 56  dissatisfaction measurements that a decision maker chooses one alternative over the other (Rouyendegh and Erkan 2013). The fuzzy ELECTRE I method is the application of the usual ELECTRE I method to fuzzy data. The main benefits of the fuzzy- ELECTRE technique are as follows: it is highly applicable when criteria are measured in an ordinal scale, has small differences in evaluations, and is non-compensatory (Mousseau and Roy 2014). The method consists of the following steps (Sevkli 2010; Rouyendegh and Erkan 2013). Step 1.  A group of 30 decision makers knowledgeable in the field of urban water management with an experience more than three years was formed. The group was responsible for the evaluation of the relevance criteria and the weights of the four selection criteria. By using the Delphi method, consensus was reached to the rating of all SPIs and the criteria weights. The fuzzy importance weight for each criterion can be described as TFNs ?̃?𝑗 = (𝑙𝑗, 𝑚𝑗 , 𝑢𝑗)  for j =1, 2, 3, and 4, where a tilde (~) represents a fuzzy number. The other three criteria measurability, data availability, and comparability were rated based on the literature.  Step 2. Normalized decision matrix A fuzzy decision matrix was formed for each sustainability dimension and normalized to obtain a normalized decision matrix ?̃? as given below.  ?̃? =  [?̃?𝟏𝟏    ?̃?𝟏𝟐 … ?̃?𝟏𝟒?̃?𝟐𝟏    ?̃?𝟏𝟏 … ?̃?𝟐𝟒…   …   ….     …?̃?𝒎𝟏    ?̃?𝒎𝟐 … ?̃?𝒎𝟒]           Equation 4.6  where   ?̃?𝑖𝑗 = (rlij, rmij, ruij)   and            Equation 4.7 𝑟𝑖𝑗𝑙 = 𝑥𝑖𝑗𝑙√∑ (𝑥𝑢)𝑖𝑗2𝑚𝑖=1  ,   𝑟𝑖𝑗𝑙 = 𝑥𝑖𝑗𝑚√∑ (𝑥𝑚)𝑖𝑗2𝑚𝑖=1  ,   𝑟𝑖𝑗𝑙 = 𝑥𝑖𝑗𝑢√∑ (𝑥𝑙)𝑖𝑗2𝑚𝑖=1    Equation 4.8  where ?̃?𝑖𝑗 = (xlij, xmij, xuij) is an actual rating score and its normalized score is r with  i = 1, 2, ……, m (m= 17, 24, 8, 10, and 9 for technical, environmental, economic, social, and institutional 57  dimensions respectively); j = 1, 2,……4, and the superscripts l, m, and u respectively refer to lower, middle, and upper values of TFNs. Step 3. Normalized weighted matrix A weighted fuzzy decision matrix was computed by multiplying the normalized decision matrix ?̃? with the criteria weights (?̃?𝑗), and then normalized according to Equation 4.8. The normalized weighted matrix ?̃? is shown in Equation 4.9.  ?̃? = [?̃?𝑖𝑗]m× n            Equation 4.9 where i and j are same as previously defined; ?̃?𝑖𝑗= ?̃?𝑖𝑗 × ?̃?𝑗; ?̃?𝑗 = (𝑤𝑗1, 𝑤𝑗2, 𝑤𝑗3), i.e., the relative weight of the jth criterion and  𝑉𝑙 = [   𝑣11𝑙    𝑣12𝑙   …  𝑣14𝑙𝑣21𝑙     𝑣21𝑙  …  𝑣24𝑙…      …     …     …𝑣𝑚1𝑙     𝑣𝑚2𝑙   …  𝑣𝑚4𝑙 ]    ,   𝑉𝑚 = [𝑣11𝑚    𝑣12𝑚   …  𝑣14𝑚𝑣21𝑚     𝑣21𝑚  …  𝑣24𝑚…      …    …     …𝑣𝑚1𝑚     𝑣𝑚2𝑚  …   𝑣𝑚4𝑚] , and  𝑉𝑢 = [𝑣11𝑢    𝑣12𝑢    …  𝑣14𝑢𝑣21𝑢     𝑣21𝑢  …  𝑣24𝑢…       …    …     …𝑣𝑚1𝑢     𝑣𝑚2𝑢  …  𝑣𝑚4𝑢] Equation 4.10  where ?̃?𝑖𝑗 is a positive TFN. Step 4. Concordance and discordance sets The concordance and discordance sets were developed for each matrix 𝑉𝑙, 𝑉𝑚, and 𝑉𝑢  representing lower (l), middle (m), and upper (u) values of TFNs respectively. For each pair of alternative Ap and Aq (p, q = 1, 2, ..., m and  p ≠q), the set of criteria was classified into two distinct subsets. If the alternative Ap was preferred over alternative Aq for all the criteria, then the concordance set was composed and expressed as: C (p,q) = {j| vpj ≥ vqj}             Equation 4.11 where vpj is the normalized weighted rating of the alternative Ap with respect to the jth criterion. In other words, C(p, q) is the collection of attributes where Ap is better than or equal to Aq. The 58  complement of C(p, q) known as the discordance set, contains all the criteria for which Ap is worse than Aq and can be expressed as D(p, q) ={j| vpj < vqj }          Equation 4.12  Step 5. Concordance and discordance indices The concordance and discordance indices were computed for l, m, and u values of each criterion having the weights wj1, wj2, and wj3 respectively. The concordance index Cpq indicates the degree of confidence in pairwise - judgments (Ap→Aq). The concordance index Cpq is defined as 𝐶𝑝𝑞𝑙 = ∑ 𝑤𝑗1𝑗∗ ,  𝐶𝑝𝑞𝑚 = ∑ 𝑤𝑗2𝑗∗ , 𝐶𝑝𝑞𝑢 = ∑ 𝑤𝑗3𝑗∗         Equation 4.13 where j* are attributes contained in the concordance set C (p, q). Similarly, the discordance index measures the power of a discordance set, i.e., the degree of disagreement in (Ap→Aq), which can be expressed as: 𝐷𝑝𝑞𝑙 = ∑ |𝑣𝑝𝑗+𝑙 − 𝑣𝑞𝑗+𝑙 |𝑗+∑ |𝑣𝑝𝑗𝑙 − 𝑣𝑞𝑗𝑙 |𝑗  ,    𝐷𝑝𝑞𝑚 = ∑ |𝑣𝑝𝑗+𝑚 − 𝑣𝑞𝑗+𝑚 |𝑗+∑ |𝑣𝑝𝑗𝑚− 𝑣𝑞𝑗𝑚|𝑗 , and  𝐷𝑝𝑞𝑢 = ∑ |𝑣𝑝𝑗+𝑢 − 𝑣𝑞𝑗+𝑢 |𝑗+∑ |𝑣𝑝𝑗𝑢 − 𝑣𝑞𝑗𝑢 |𝑗    Equation 4.14 where J+ are the criterion contained in the discordance set D (p,q) and vij is the normalized weighted evaluation of the alternative i on the criterion j.  Step 6. Final indices calculation The final concordance (C*pq) and discordance (D*pq) indices are geometric means of l, m, and u values separately of Cpq and Dpq. These indices can be considered as the defuzzification and were computed as: 𝐶𝑝𝑞∗ = √∏ 𝐶𝑝𝑞𝑧𝑧𝑧=1𝑧   and  𝐷𝑝𝑞∗ = √∏ 𝐷𝑝𝑞𝑧𝑧𝑧=1𝑧          Equation 4.15 where Z=3 denoting three values l, m, and u. A larger final concordance index Cpq and a smaller final discordance index Dpq resulted in a stronger dominance relationship of the alternative Ap over the alternative Aq. The outranking relation was obtained by using Equation 4.16 and Equation 4.17. 59  If 𝐶𝑝𝑞∗  ≥ 𝐶̅    and         Equation 4.16 𝐷𝑝𝑞∗  < ?̅?             Equation 4.17 where 𝐶̅ and ?̅?  are averages of Cpq and Dpq respectively.     In this method, Ap outranks (better than) Aq when Equation 4.16 and Equation 4.17 hold true, whereas, alternative Ap is indifferent to Aq when both hold false, and the alternative Ap is incomparable to Aq when one holds true with another false. Based on these relationships, an outranking diagram of SPIs were developed for each sustainability dimension (Yoon and Hwang 1995). Step 7. Ranking SPIs The net outranking relationships can be established using a net concordance index (Cp) and the net discordance index (Dp) for each SPI (alternative). Cp measures the degree to which the dominance of an alternative Ap over competing alternatives exceeds the dominance of competing alternatives over the alternative Ap and can be defined as follows (Yoon and Hwang 1995; Haider et al. 2014b): 𝐶𝑝 = ∑ 𝐶𝑝𝑘𝑚𝑘=1𝑘≠𝑝− ∑ 𝐶𝑘𝑝𝑚𝑘=1𝑘≠𝑝        Equation 4.18 Similarly, the Dp measures the relative weakness of alternative Ap with respect to other alternatives and can be defined as follows (Yoon and Hwang 1995; Haider et al. 2014b): 𝐷𝑝 = ∑ 𝐷𝑝𝑘𝑚𝑘=1𝑘≠𝑝− ∑ 𝐷𝑘𝑝𝑚𝑘=1𝑘≠𝑝        Equation 4.19 For making an overall preference (ranks), a higher Cp and lower Dp will receive a higher rank.  The final ranking was performed based on the values of Cp and Dp of SPIs and their outranking relationships. 4.3 Results A total of 68 potential SPIs were initially screened from the literature for the sustainability assessment of SMUWSs. The screened list is comprised of 17 SPIs in the technical, 24 in the 60  environmental, 8 in the economic, 10 in the social, and 9 in the institutional dimension. These SPIs with their description and measurement method are given in Appendix A.2. Also, the methodological steps involved, including the application of F-AHP, is elaborated in detailed in Appendix A.3 for the ranking of SPIs in the economic dimension as an example. The application of the fuzzy-ELECTRE I method to the initially screened SPIs resulted in outranking relationships. The outranking relationships for each sustainability dimension are given in Figure 4.3, Figure 4.4, Figure 4.5, Figure 4.6, and Figure 4.7. These relationships can be used to identify important SPIs for sustainability assessment of SMUWSs.  The most important SPIs are those positioned at the upper level of outranking diagrams. The decision maker’s boundary (DMB) was used as a cut-off boundary to select the final SPIs (Haider et al. 2014b). A DMB is selected by a decision maker and is based on the relative distance between the indices (concordance and discordance) of SPIs. The final SPI list consists of 38 SPIs including 8 SPIs in the technical, 13 SPIs in the environmental, 4 SPIs in the economic, 7 SPIs in the social, and 6 SPIs in the institutional dimensions as follows:  4.3.1 Technical The selected eight technical SPIs belong to the “neighbourhood location and design” and “water infrastructure and fixtures” criteria (Figure 4.3). The SPIs within the neighbourhood location and design criteria are as follows: proximity to drinking water system/source, proximity to wastewater system, separation of wastewater and storm water, and dwelling density. The SPI proximity to drinking water system/source is one of two top level indicators. Another top level indicator, water leakage, is discussed below. Similarly, the selected SPIs of the water infrastructure and fixtures criteria are as follows: water leakage, water supply reliability, metered connection, and treated water storage capacity.   61   Figure 4.3    Outranking relations of the technical SPIs with DMB    62  4.3.2 Environmental The relationship among environmental SPIs is presented in the outranking diagram in Figure 4.4. Altogether, 13 SPIs were selected in this dimension. The SPIs are grouped into the following criteria: resource utilization, environmental impacts, and resource recovery. The selected SPIs of the resource utilization criteria are as follows: water self-sufficiency, domestic water consumption, non-domestic water consumption, groundwater quality, surface water quality, energy use in water service, energy use in wastewater service, chemical use in water treatment, and chemical use in wastewater treatment. Water self-sufficiency and domestic water consumption are placed at the top level along with water reuse indicator. Similarly, the three selected SPIs of the environmental impacts criteria are as follows: discharged wastewater quality, biosolids quality, and disposal of backwash water. In the resource recovery criteria, the SPI water reuse was selected. In the resource recovery criteria, the SPI water reuse was selected. Water reuse is the third of the top level indicators along with the proximity to drinking water system and surface water quality indicators. 4.3.3 Economic The economic SPIs and their relationships are depicted in Figure 4.5. Four SPIs have been selected in this dimension and were categorized into two criteria “water economics” and “wastewater economics”. In the water economics criteria, the selected SPIs are operating cost coverage ratio for water service, average water fee rate, and non-revenue water. Similarly, in the wastewater economics criteria, the selected SPI is operating cost coverage ratio for wastewater service. The top level indicators were found to be operating cost coverage ratio for water service and that for wastewater service. Similarly, at the second level, two SPIs average water fee rate and non-revenue water are placed. The top level indicators identified are operating cost coverage ratio for water service and that for wastewater service. At the second level, two SPIs average water fee rate and non-revenue water are placed.    63   Figure 4.4    Outranking relations of the environmental SPIs with DMB  Figure 4.5    Outranking relations of the economic SPIs with DMB  64  4.3.4 Social The relationships among social SPIs are presented in Figure 4.6. Altogether, seven SPIs have been selected. These SPIs belong to the “service provision” and “public health” criteria. The SPIs: access to water service, access to wastewater service, and drinking water quality outranked all other indicators. In the service provision criteria, the selected SPIs are as follows: access to water service, access to wastewater service, water restrictions, and public acceptability. Similarly, in the public health criteria, the selected SPIs are drinking water quality, boil water advisories, and safety (from flooding and drought). The SPI drinking water quality was found to be at the top level along with the access to water and wastewater service indicators. 4.3.5 Institutional The institutional SPIs and their relationships are depicted in Figure 4.7. Six SPIs have been selected, and they belong to the “governance and progress” criteria. The top level SPI is urban water policies that is indifferent with the SPI achievement of water demand reduction target. The SPIs institutional capacity and personnel training are placed at the next level.  Figure 4.6    Outranking relations of the social SPIs with DMB   65    Figure 4.7    Outranking relations of the institutional SPIs with DMB The list of selected SPIs are given in Table 4.3.  4.4 Discussion The applicability of SPIs in urban communities is described under each sustainability dimension below and is confined to the final SPIs. 4.4.1 Technical In the neighbourhood location and design criteria, the SPI proximity to drinking water system/source is an important indicator. The further the community is from existing water and wastewater systems, the less sustainable it is as it consumes more resources for the construction, operation, and maintenance of its water and wastewater systems. However, a community should be built  at a certain buffer distance from a water source (USGBC 2013). Similarly, the proximity to wastewater system is an essential indicator. Speir and Stephenson (2002) showed that a longer distance between neighbourhoods (development tracts) and existing water and wastewater services increases the cost of providing these services.   66  Table 4.3    Final Ranks of the Selected SPIs of UWSs Technical   Environmental   Economic   Social    Institutional R SPI*  R SPI  R SPI  R SPI  R SPI 1 Proximity to water (TE1)  1 Self-sufficiency (W) (EN1)  1 Oper. cost (W) (EC2)  1 Access to water (SO1)  1 Demand reduc. (IN8) 2 Water leakage (TE15)  2 Domestic consumption (EN2)  2 Non-revenue water (EC5)  1 Access to WW (SO2)  2 Policies (IN1) 3 Water storage (TE11)  3 Water reuse (EN21)  3 Avg. water fee (EC4)  3 Drinking WQ (SO9)  3 Insti.capacity (IN4) 4 Supply reliability (TE14)  4 Non-domestic consumption (EN3)  3 Oper. cost (WW) (EC7)  4 BWA (SO10)  3 Personnel training (IN5) 5 Separation of WW and SW (TE4)  4 Surface WQ (EN5)  - -  5 Water restrictions (SO3)  5 Consvn. Programs (IN6) 5 Metered connection (TE9)  6 Chemical use WT (EN10)  - -  6 Acceptability (SO4)  6 Public participation (IN7) 7 Dwelling density  (TE5)  7 Backwash water (EN16)  - -  7 Safety from hazards  - - 8 Proximity to WW (TE2)  8 Chemical use in WWT (EN11)  - -  - -  - - - -  9 Bio-solids quality (EN14)  - -  - -  - - - -  10 Energy use (W) (EN7)  - -  - -  - - - -  10 Energy use (WW) (EN8)  - -  - -  - - - -  12 Discharged WW quality (EN13)  - -  - -  - - - -  13 Ground WQ (EN4)  - -  - -  - - Note: * Full description of SPIs with their measurement method is provided in Appendix A.2.  R: rank, WW: wastewater, W: water, WQ: water quality, WT: water treatment, WWT: wastewater treatment, Oper.: operating, Avg.: average, BWA: Boil water advisory, Insti.: institutional, consvn.:  conservation  The next SPI dwelling density (residential building) represents a measure of compact development. A higher value of the SPI indicates a more dense urban community, and represents an efficient use of water infrastructures. Speir and Stephenson (2002) demonstrated that a larger lot size and higher dispersion of neighbourhoods significantly increases the cost of providing water and wastewater services. For a sustainable neighbourhood, the recommended value is a minimum of 8 dwelling units (DU) per acre for residential buildings (USGBC 2013;  EarthCraft 2014). Finally in the neighbourhood location and design criteria, the SPI separation of wastewater and storm water occurs. This indicator is a measure of optimum resource use (Van 67  Leeuwen et al. 2012) because the degree of treatment required for storm water is usually less than that required for wastewater. In the water infrastructure and fixtures criteria, water leakage shows the distribution efficiency of finished water. Its value ranges from 7.6% – 14.9%, with an average of 13.3% in Canada (Environment Canada 2011). The SPI water supply reliability can be expressed as the number of main breaks per 100 km length. Its value ranges from 1 to 19.7 main breaks per 100 km with a median of 5.9 main breaks per 100 km pipe length (AECOM 2012). However, the ratio may be higher in small and medium sized communities because of a shorter pipe length. So, care should be taken when comparing with these values.  The SPI metered connection is a measure of the efficient use of resources. The installment of water meters in homes encourages consumers to conserve water because higher water consumption increases their water bills. Water metering can reduce water consumption by 15 to 25% (DoP 2007). For instance, water consumption was reduced  by 16%  in the District of Peachland, British Columbia in 2007 when water metering began (DoP 2015).  The next SPI, treated water storage capacity, measures the capacity of the water system to meet water demand even during treatment failures. This indicator shows the sustainability of the system (NRC 2009; CSA 2010). In Canada, storage capacity ranges from less than 1 to 96 hours with a median of 29 hours (AECOM 2012).  4.4.2 Environmental  In the resource utilization criteria, the SPI water self-sufficiency is a measure of the availability of the required water in a community territory. Water self-sufficiency can be measured in terms of licensed water or annual renewable water available in the community territory. However, they have different meanings. Water self-sufficiency in terms of licensed water represents the legally available water; however, the licensed amount of water may not necessarily be available in the water source. Water self-sufficiency in terms of renewable water represents the natural water availability in the source water. Usually, licensed water is based on the renewable water availability of the water bodies.    68  Water self-sufficiency, in terms of licensed water in Canadian municipalities, is higher than 130% (i.e., less than 1 to 76% of licensed water withdrawal) (AECOM 2012). The SPI domestic water consumption measures the extent of domestic water use indicating whether an urban community has an over-consumption or under-consumption of water with respect to a benchmark. Canadians have a higher domestic water consumption rate (Renzetti 1999; Ma 2014) with an average of 343 L/capita/day (Environment Canada 2014b), while non-domestic water consumption is 236 L/capita/day in their country (Environment Canada 2011). However, non-domestic water consumption depends on the type and extent of industrial establishments in an urban community. Domestic and non-domestic water may be supplied by groundwater or surface water or both. The SPI groundwater quality is measured in terms of Faecal Coliform, Nitrogen (N), and Phosphorus (P). Another SPI surface water quality is measured in terms of Faecal Coliform, N, P and Biochemical Oxygen Demand (BOD) of the major water body of the urban community or city (Van Leeuwen et al. 2012). Good quality of groundwater and surface water is required for environmental and human health (Van Leeuwen et al. 2012). Since, groundwater and/or surface water are the sources of public water supply, their preservation enhances the water sustainability of an urban community.  The SPI energy use is an important aspect of urban water sustainability (Chang et al. 2014). Energy is required for operating UWSs.  Energy use significantly differs based on site-specific conditions, including distance to the water source, depth in case of groundwater, topography, water quality, and the technology used (Tuladhar et al. 2014).  For example, in California, US, energy use varies from 0.211 to 8.243 kWh m-3 of water supplied and 0.291 to 1.321 kWh m-3 of wastewater treated (CEC and NC 2006). The next SPI, chemical use, in terms of major chemicals such as chlorine and coagulants in water and wastewater treatment is  measured in order to determine the use of resources for water treatment (Lundin and Morrison 2002; Makropoulos et al. 2008; Popawala and Shah 2011; Water UK 2011; Van Leeuwen et al. 2012; Van Leeuwen and Marques 2013). Clean raw water and less polluted wastewater require less chemicals for their treatment, which enhances urban water sustainability.  In the environmental impacts criteria, discharged wastewater quality measures the impact of wastewater disposal on the environment. Discharge wastewater quality is measured as the number of days (or times) out of compliance for BOD, N, P, and heavy metals (Cd, Pb, Hg and 69  Cu). This non-compliance  should be less than 5% (18 days in year) (World Bank 2008). Next, the SPI biosolids quality, measured in terms of heavy metal content, indicates the impacts of the biosolids disposal on the environment. The biosolids can have restricted or unrestricted use based on their heavy metal content (CCME 2010). A higher heavy metal content than a recommended value (for unrestricted use) results in a restricted use of biosolids, such as soil amendments. Furthermore, backwash water is generated by the cleaning (backwash) of the treatment plant equipment, for example filters. Backwash water may contain toxic chemicals, such as aluminum and manganese due to the use of coagulants in water treatment (Haider et al. 2014b). Because of the chemical content, the backwash water should be treated and the discharge of untreated backwash water to natural water bodies should be monitored. In the resource recovery criteria, the SPI water reuse was selected. The wastewater of SMUWSs is a resource and can be recycled to obtain water of usable quality. The water recycling enhances urban water sustainability (Chang et al. 2015) because water reuse saves water that would otherwise be lost from the system. For example, in Canada, water reuse is very low with a value of approximately 3% in British Columbia - one of the water reusing provinces (CCME 2002). A very high use of recycled water, up to 88%, was achieved in Melbourne, Australia, where recycled water of different classes is produced and used for various activities such as toilet flushing, industrial wash down, and municipal watering (Western Water 2013). Similarly, 100% wastewater has been recycled and reused in Cyprus (EU 2015; 2016). Water reuse should be increased in an urban community for its water sustainability.  4.4.3 Economic  The top level indicators identified are operating cost coverage ratio for water service and that for wastewater service. This result is not surprising as any water utility is financially sustainable when water and wastewater revenues equal or exceed expenses for, at least, the operational and maintenance costs (World Bank 2003). In this case, the ratio is 1.0 or higher. For an economically sustainable SMUWS, a lower operating cost of water and wastewater is desired. A 70  dense urban community can reduce water infrastructure requirements and operating costs thereby increasing economic sustainability.   At the second level, two SPIs average water fee rate and non-revenue water are present. The average water fee rate is a measure of affordability of consumers to pay for water and wastewater services delivered. This rate varies from $152 to $489 with a median of $ 366 per 250 m3 water in Canada (AECOM 2012). According to Renzetti (1999), user fees meet only 37% of operational and 66% of capital expenditures. This statistics indicates lower water user fees in Canada. The next indicator, non-revenue water (NRW), measures the water supplied with no revenues collected. NRW includes real losses, apparent losses (customer meter inaccuracies and unauthorized consumption), and unbilled authorized consumption (Kanakoudis and Tsitsifli 2010). NRW is calculated in terms of “liter/connection/day” based on the view that a major water loss occurs at service connections (Hamilton et al. 2006). NRW can be considered as a useful financial indicator (Kanakoudis and Tsitsifli 2010). For an economically sustainable SMUWS, an affordable water fee rate and a lower value of NRW are desired (Zhou et al. 2013).  4.4.4 Social  In the service provision criteria, the SPIs access to water and wastewater services are measured by the percentage of population served by public water supply and wastewater service (with a secondary level or higher treatment) respectively. Water service is required for the development of an individual human being (Van Leeuwen et al. 2012). Wastewater service is required for the safe disposal of wastewater in order to protect human and environmental health (Van Leeuwen et al. 2012).  Therefore, access to these services is crucial for assessing the sustainability of SMUWSs. In Canada, the access to public water service varies from 50% to 98% with an average of 88.9% and the access to wastewater service varies from 37 to 76% with an average of 68% (Environment Canada 2011).  The SPI water restrictions, measured as the number of days per year, are the days that are restricted for water use for specific purposes, such as lawn irrigation. These regulatory measures are imposed by local water utilities to conserve water especially during peak demand. However, this measure restricts consumers from water use. In Canada, this measure is frequently practised. Water restrictions range from zero to 365 days per year with a median of 121 days per year 71  (AECOM 2012). Furthermore, public acceptability is an important aspect of social sustainability. This indicator can be expressed in terms of the number of complaints on water and wastewater services per 1000 population. A low number or no complaint indicates public acceptability to water and wastewater services.  In Canada, the complaints on water and wastewater services range from 0.02 to 24.84 complaints per 1000 people with a median of 4.49 complaints per 1000 people per year (AECOM 2012).  In the public health criteria, the SPI drinking water quality was found to be at the top level along with the access to water and wastewater service indicators. The reason behind this may be – clean and safe water supply is one of the prime objectives of a sustainable UWS (Hellstro 2000; Engel-yan et al. 2005).  This SPI is measured in terms of non-compliance of turbidity, total coliforms, residual chlorine, and nitrates of drinking water. The next SPI, boil water advisories (BWA), is a measure of public health risk due to the contamination of water supply. BWA is calculated as the number of household (HH)-days per year that boil water advisories are in effect as a % of total HH-days. This is an important indicator. For example, a majority of provinces have gone through many BWA and British Columbia has gone through the highest BWA in Canada (Water Chronicles 2014).  The median BWA in Canada is 0.89 days per year with a range of zero to 12 days per year (measured as the number of BWA days x capita affected/total population served). Furthermore, the safety indicator qualitatively assesses plans, measures, and their implementation status in order to protect citizens against flooding and drought. This indicator is also considered by City Blueprints (Van Leeuwen et al. 2012).  4.4.5 Institutional The top level SPI is urban water policies. The urban water policies indicator qualitatively assesses a local government's policies, action plans, and commitments for an integrated urban water management. A similar indicator has also been used by City Blueprints (Van Leeuwen et al. 2012). Moreover, water demand reduction in an existing community is an effective SPI to measure a community’s progress toward the sustainability practice. The reduced demand is primarily achieved by institutional initiatives. Particularly, in high water consuming communities such as in Canada, the reduction of residential water consumption is an important step for 72  achieving water sustainability. The average water demand reduction is 18 litre/capita/day, i.e., 5.5% per year from 2006 to 2009 (Environment Canada 2011).  At the next level, the SPI institutional capacity, in terms of fulltime equivalent (FTE) personnel, measures the strength of a municipality or water purveyor. This indicator is also used by IWA (2006), Government of Canada (2007), CSA (2010), and Sydney Water (2013). Similarly, the SPI personnel training measures the extent of organizational development and is also used by IWA (2006), AWWA (2008), and World Bank (2011). Furthermore, the SPI public participation measures a local community involvement for achieving healthy community activities (Brown and Farrelly 2009;  Siemens AG 2009; Van Leeuwen et al. 2012). These SPIs are also used for assessing the sustainability of UWSs by Van Leeuwen et al. (2012) in City Blueprints. In addition, the SPI conservation program is a measure of conservation efforts of an institution and can be measured by annual expenses for running the program.  A reduction in domestic water consumption can be achieved by effective conservation programs. In Canada, the expenses of conservation programs vary from less than  $1 to $5.64 per person per year with a median value of $0.47 per person per year (AECOM 2012). However, the type of conservation programs required may differ from community to community.  4.5 Summary UWSs are challenged by the sustainability perspective. Certain limitations of the sustainability of centralized UWSs and decentralized household level wastewater treatments can be overcome by managing UWSs at an intermediate scale, referred to as small to medium sized UWSs (SMUWSs). SMUWSs are different from large UWSs, mainly in terms of smaller infrastructure, data limitation, smaller service area, and institutional limitations. Moreover, sustainability assessment systems to evaluate the sustainability of an entire UWS are very limited and confined only to large UWSs. This research addressed the gap and has developed a set of 38 applied sustainability performance indicators (SPIs) by using fuzzy-ELECTRE I outranking method to assess the sustainability of SMUWSs. The developed set of SPIs can be applied to existing and new SMUWSs and also provides a flexibility to include additional SPIs in the future based on the same selection criteria. The SPIs related to water, energy, carbon emissions, cost, and health risk has been used for developing water-energy-carbon (WEC) model in the next chapter.  73   System Dynamics Modelling of Water-Energy-Carbon (WEC) Nexus A version of this chapter has been published in the ASCE Journal of Water Resources Planning and Management entitled “Water-Energy-Carbon nexus modelling for an urban water system: A system dynamics approach” (Chhipi-Shrestha et al., 2017b). 5.1 Background A comprehensive framework and decision support system to quantify the WEC nexus and its dynamic behavior at the neighbourhoods or community level is required (Nair et al., 2014; Rothausen & Conway, 2011; Arora et al., 2013; Kenway, 2013). The extensive review of the water-energy nexus studies by Kenway et al. (2011) also concluded the lack of a unifying framework and consistent methodology for analyzing the WEC nexus. The WEC model comprising interacting problems can be developed by using system dynamics (Nasiri et al., 2013; Nair et al., 2014).  System dynamics has been used in only a few urban water studies (Zarghami and Akbariyeh 2012). The researchers, such as Zarghami and Akbariyeh (2012), Zhang et al., (2008), Zhang et al. (2009), Karamouz et al. (2012), Nasiri et al. (2013), and Zhang et al. (2009) applied system dynamics to identify effective and reliable water resources plan, policy or estimate water resource carrying capacity. Qi and Chang (2011), Wang (2014), Tong and Dong (2008), and Nawarathna et al. (2009) separately studied the dynamic effects, such as of macro-economy, water price, socio-economic-environmental system, or changing land use and climate on water demand and supply. All these system dynamics-based studies lack energy use and carbon emissions.  At the country level, Tidwell et al. (2012) estimated potential impact of water availability on future expansion of thermoelectric power generation, whereas Wang (2013) studied the implications of biofuel development on water and energy. The former lacks carbon emissions and the latter is a preliminary model. Specifically, Willuweit and O’Sullivan (2013) modelled the effects of urban development and climate change on urban water cycle. This critical review shows system dynamics has been applied to different aspects of urban water from county to 74  country level but not applied to the WEC nexus for neighbourhoods, a community, i.e., SMUWS. Also, urban water processes perform differently in various geographic regions with respect to energy and carbon emissions. This requires a holistic and generic model to capture the variability and dynamics of UWSs (Nair et al., 2014). This chapter aimed to develop a dynamic WEC nexus model that can assist municipalities, urban developers, and policy makers in making informed decision for reducing water consumption, energy use, and carbon emissions in UWSs.  5.2 Methodology The operational phase of a SMUWS has been identified as the most energy intensive phase from the life cycle perspective (Friedrich, 2002; Nair et al., 2014). This study has focused only on the operational phase of a SMUWS. The WEC model was based on system dynamics using STELLA® 10.1.3 (ISEE Systems 2016; Karamouz et al. 2012; Qi and Chang 2011). The system dynamics model (SDM) includes water module, energy module, and carbon module as elaborated in the following sections.  5.2.1 Water module Water module is comprised of water consumer growth sub-models and water and wastewater sub-models as follows.  5.2.1.1 Water consumer growth sub-models The water consumers: population; commercial, institutional and industrial (CII) sector; and agriculture were included in the WEC model. The Standard Industrial Classification code numbers 2000 through 3999 (Gleick et al. 2003; US EPA 2009b) was followed to define CII sector. The major commercial sectors included in this study are offices, restaurants, supermarkets and retail, and hotels; major institutions: government institutions, hospitals, and schools as identified by AWWA (2000), and industries in average. The dynamics of water consumers were analyzed by using the growth equation (Nasiri et al. 2013) (Equation 5.1 in Table 5.1).   75  5.2.1.2 Water and wastewater sub-models The water and wastewater sub-models include municipal water use model, wastewater generation model, and water footprint (WF) model for the operational phase of the SMUWS. a) Municipal water model The municipal water use model represents the flow and use of drinking water in neighbourhoods. The municipal water flow occurs through urban water stages: abstraction and conveyance, treatment, distribution, and use. The municipal water use dynamics was modelled in Equation 5.2 in Table 5.1. The equation includes the water consumed by different urban water components over time: residential, CII, public parks, golf courses, and agriculture. Each of these urban water components was modelled by including all their unit water use activities. As an example, residential water is modelled in Equations 5.3, 5.4, and 5.5 in Table 5.1, and the similar equations were used for all other urban water components of Equation 5.2. b) Wastewater generation and water footprint models The wastewater generation model includes wastewater (WW) collection from residential and CII sector as well as infiltration and inflow to sewer network as shown in Equation 5.6. The modelled wastewater includes the indoor water consumed by the respective urban water components except the leakage. The water footprint model is represented by Equation 5.7.  5.2.2 Energy module The energy module includes the operational energy of a SMUWS and embodied energy of major chemicals, such as chlorine, poly aluminum chloride, and polymers. The dynamics of energy use was modelled in Equations 5.8 and 5.9. The hot water energy for residential sector was modelled by using Equation 5.10 (Aguilar et al. 2005) and the similar equation was used for modelling indoor hot water energy of CII sector.    76  5.2.3 Carbon module The carbon module represents carbon emissions of the operational phase of a SMUWS. The module includes direct carbon emissions in terms of CO2e from energy use in a SMUWS, wastewater processes, and carbon footprint of major chemicals. The dynamics of carbon emissions were modelled in Equations 5.11 and 5.12 in Table 5.1. Table 5.1    Equations of the WEC model WEC nexus Aggregation Equations Eqn #  Water module (Water consumer growth sub-model)  W,E,C Nt = N0er t * (Presi)t       [5.1]  Water module (Water and wastewater sub-models)  W (Wdirect)t = (Wresident)t + (Wcomm)t + (Winsti)t + (Windustry)t + (Wparks)t+ (Wgolf)t+ (Wagri)t + (Wdistrib loss)t   [5.2] W (Wresident)t = (Win)t + (Iout)t   [5.3] W (Win)t = (TW)t + (SW)t + (FW)t + (LW)t + (DW)t + Indoor water leakage            = [(fT*ηTW)t + (fS*dS*ηSW*ηSU)t + (fF*ηFW*ηFU)t + (fL*ηLW)t + (fD*ηDW)t]*(1+InLe)*Ict*Nt [5.4] W (𝐼𝑜𝑢𝑡)𝑡 =𝑁𝑡𝐷𝑂∗ (𝑈𝑆 ∗ 𝐿𝑆 ∗ 𝐶𝑆 + 𝑈𝐷 ∗ 𝐿𝐷 ∗ 𝐶𝐷 +𝑈𝑅∗𝐶𝑅∗𝑅𝑈𝐿𝑅∗𝐹𝐴𝑅𝑅+𝑈𝐴𝑠∗𝐶𝐴𝑠∗𝑅𝑈𝐿𝐴𝑠∗𝐹𝐴𝑅𝐴𝑆+𝑈𝐴𝑙∗𝐶𝐴𝑙∗𝑅𝑈𝐿𝐴𝑙∗𝐹𝐴𝑅𝐴𝑙)𝑡∗ 𝐼𝑟𝑡 ∗ 𝐼𝑟𝑐𝑡 ∗ Ip        [5.5] W (WW)t = (Residential WW)t + (Commercial WW)t + (Institutional WW)t + (Industrial WW)t + (Infiltration and inflow to sewer network)t [5.6] W-E (Total WF)t = (Wdirect)t + (WF of direct energy use)t + (WF of major chemicals)t  [5.7]  Energy module  E-W (Edirect)t = (Econvey)t + (EWT)t + (EDistri)t + (Eresi HW)t + (ECII HW)t + (EWW transport)t + (EWWT)t + (EBiosolids)t        [5.8] E-W (Etotal)t = (Edirect)t+ (EEchemicals)t [5.9] E-W (Eresi HW)t = (SE)t + (FE)t + (LE)t + (DE)t + (SL)t       = [(fS*dS*ηSW*ηSU*ηSE*HS)t + (fF*ηFW*ηFU*ηFE*HF)t + (fL*ηLE*WHRL*HL)t +(fD*ηDE*WHRD*HD)t + (SLR)t*(fS*dS*ηSW*ηSU*HS + fF*ηFW*ηFU*HF + fL*ηLW*HL + fD*ηDW*HD)t]*Ect*Nt [5.10]  Carbon module  C-E &W (Cdirect)t= (Cconvey)t + (CWT)t + (CDistrib)t + (CresiHW)t + (CCIIHW)t + (CWWtransport)t + (CWWT)t + (Cbiosolids)t + (CWWprocesses)t [5.11] C- E &W (Ctotal)t = (Cdirect)t + (CFchemicals)t             [5.12] Note:  Water module Water consumer growth sub-model Where for population, Nt is population in a month, N0 is base population, r is population growth rate (monthly), t is time duration in months. Similarly for CII sector, N separately refers to the number of hotel rooms, hospital beds, and school students and N refers to floor area for other CII sectors (restaurants, offices, supermarkets, and industries); for irrigation water, N separately 77  refers to golf course area, neighbourhood and community park land, and agricultural land if present; r refers to their growth rate; Presi refers to the proportion of population residing in that time and is 1 when seasonal migration is not considered.  Water and wastewater sub-models a) In Equation 5.2, Wdirect is direct water use (L), Wresident is residential water use (L), Wcomm is commercial water use (Offices including governmental offices; restaurants and supermarkets) (L), Winsti is institutional water use (Hotels, hospitals, and schools) (L), Windustry is industrial water use (L), Wparks is parks water use (L), Wgolf is golf courses water use (L), Wagri is agricultural water use (L), Wdistrib loss is water loss in distribution (L), and “t” refers to a month.  b) In Equations 5.3 and 5.4, Win is indoor water use (L); Iout is outdoor irrigation water (L); f is frequency of use (per capita/month), η is efficiency (L/min for shower and L/use of others), d is duration of use (min/shower) and their subscripts T is toilet, TW is toilet water, S is shower, SW is shower water, SU is shower use, F is faucet, FW is faucet water, FU is faucet use, L is laundry, LW is laundry water, D is dishwasher, and DW is dishwasher water; InLe is % of indoor water leakage; Ict is indoor water conservation rate (monthly) and for exponential change of this rate, Ict =Ic0e-rt with r as the rate of change of indoor water conservation rate; Nt is population at time t (month).  c) In Equation 5.5, DO is dwelling occupancy (persons/residential unit); U is % of dwelling units, L is average lot size (ha), C is average % of lawn coverage, i.e., (1 - Average % of lot coverage), FAR is floor area ratio, and their subscripts S is single detached home, D is duplex, R is row house, “As” is small apartments, and “Al” is large apartments; RU is residential unit size (ha);  Ir is garden irrigation rate; Irc is irrigation conservation rate (monthly) and for exponential change of this rate, Irct = (Irc)0e-rt with r as the rate of change of irrigation conservation rate; Ip is irrigated garden proportion. The value of r could be different for Ict and Irct; however, they are considered equal in this study due to the lack of data. d) In Equation 5.7, WF of direct energy use is the sum of the products of WF of particular energy source (L/kWh) and total amount of that energy (kWh) for all energy uses, and WF of major chemicals is the sum of the products of WF of a particular chemical (L/kg), rate of chemical use in water or wastewater (kg/L), and total amount of water or wastewater (L) for all chemical types in time “t”  Energy module e) In Equations 5.8 and 5.9, Edirect is direct energy use (kWh), Econvey is raw water abstraction and conveyance energy (kWh), EWT is water treatment energy (kWh), EDistri is water distribution energy (kWh), Eresi HW and ECII HW are the energy for indoor hot water use (kWh) in residential and CII sector respectively, EWW transport is wastewater transport energy (kWh), EWWT is wastewater treatment energy (kWh), EBiosolids is biosolids transportation energy (kWh), Etotal is total energy use (kWh), EEchemicals is embodied energy of major chemicals (kWh), and “t” refers to a month. f) In Equations 5.8 and 5.9, (Econvey)t , (EWT)t , (EDistri)t , (Eresi HW)t , (ECII HW)t , (EWW transport)t , and (EWWT)t are individually estimated as the product of energy intensity of each process (kWh/L) and total amount of water or wastewater flow in the process (L) in time “t”, and (EBiosolids)t  as the product of energy intesnsity of biosolids transportation (kWh/kg), rate of biosolids generation per unit of wastewater (kg/L) and total volume of wastewater (L) in time “t”  and (EEchemicals)t  as the product of unit embodied energy of  a chemical (kWh/kg) estimated from LCA, rate of chemical use in water or wastewater (kg/L) and total amount of water or wastewater (L) for all chemical types in time “t” g) In Equations 5.10, f, η, d, T, TW, S, SW, SU, F, FW, FU, L, LW, D, DW, Nt , and t have same meaning as of Equation 5.4; SE is shower energy; FE is faucet energy; LE is laundry energy; DE is dishwasher energy; H is hot water ratio; WHR is water heating ratio; SL is standby energy loss; R is rate; Ect is hot water-energy conservation rate (per month) and for exponential change of this rate, Ect =Ec0e-rt with r as the rate of change of hot water-energy conservation rate (per month).  Carbon module h) In Equations 5.11 and 5.12, Cdirect is carbon emissions (CE) from direct energy use (kg CO2e), Cconvey is CE from raw water abstraction and conveyance energy (kg CO2e), CWT is CE from water treatment energy (kg CO2e), CDistrib is CE from water distribution energy (kg CO2e), Cresi HW and CCII HW are CE (kg CO2e) from indoor hot water use respectively in residential and CII sector, CWWtransport is CE (kg CO2e) from wastewater transport energy, CWWT is CE (kg CO2e) from wastewater treatment energy, Cbiosolids is CE (kg CO2e) from biosolids transportation energy, CWWprocesses is CE (kg CO2e) from wastewater processes (wastewater treatment), Ctotal is CE (kg CO2e) from total energy use, CFchemicals is carbon footprint (kg CO2e) of major chemicals, and “t” refers to a month i) In Equation 5.11, (Cconvey)t , (CWT)t , (CDistrib)t , (CresiHW)t , (CCIIHW)t , (CWWtransport)t , (CWWT)t , and (Cbiosolids)t are individually estimated as the product of carbon emission factor of the energy source (kg CO2e/kWh) and total amount of energy consumption (kWh) in time “t”, (CWWprocesses)t as the product of organics (BOD) generation rate (kg BOD/person), carbon emission factor of organics (kg CO2e/kg BOD) and total population (persons) in time “t” (IPCC 2006). j) In Equation 5.12, (CFchemicals)t is the product of carbon footprint of a particular chemical (kg CO2e/kg) estimated from LCA, rate of chemical use in water or wastewater (kg/L), and total amount of water or wastewater (L) for all chemical types in time “t”   78  A causal loop diagram (CLD) was developed before developing a complete SDM. A CLD is a graphical representation that enables the visualisation of causal relationships between variables in a causal model. The causal diagram shows how each factor affects others and in turn is affected by other factors. The CLD is given in Figure 5.1, in which “+” indicates a positive relationship and “-” indicates a negative relationship in the UWS. As shown in Figure 5.1, community people use indoor and outdoor water (residential water). Similary, commercial, institutional, and industrial (CII) sector, agriculture, golf courses, and parks consume water (direct water use). The water for residential sector, CII sector, agriculture, golf course irrigation, and park irrigation combinely give water use in a community. Water abstraction supplies water for use and it generates wastewater after use. Raw water collection, water distribution, wastewater transport, and wastewater treatment require energy (water and wastewater conveyance and treatment energy). Indoor hot water use (residential and CII sectors) consumes energy (hot water energy). Energy use has water footprint. Also, chemicals used in water and wastewater treatment have water footprint, embodied energy, and carbon footprint.  The sum of direct water use, water  footprint of water and wastewater conveyance and treatment energy and indoor hot water energy use, and water footprint of chemicals gives total water footprint. Similarly, the sum of water and wastewater conveyance and treatment energy, indoor hot water energy, and embodied energy of chemiclas gives total energy for an UWS. The sum of GHG emissions from conveyance and treatment energy use, GHG emissions from indoor hot water energy use, and carbon footprint of chemiclas provides total GHG emissions from an UWS. The complete SDM in the form of stock and flow diagrams developed based on the CLD is given in Appendix B.1.    79     Figure 5.1    Causal loop diagram of the WEC nexus of a SMUWS 5.3 Results and Discussion A WEC model and its DSS was developed in Stella 10.1.3. The model was calibrated and validated and then applied to the City of Penticton (CoP), Okanagan, British Columbia (BC), Canada as follows. 5.3.1 Model calibration and validation Calibration refers to the estimation of parameter values, e.g., rate of water treatment energy use. The model was calibrated using the historical data for the years other than the validation period (Wang 2014; Willuweit and O’Sullivan 2013). The calibrated model was validated by using the historical data  (Qi and Chang 2011; Willuweit and O’Sullivan 2013). The validation data constituted the monthly data for raw water collection, water treatment, distribution, water consumption, wastewater generation, wastewater treatment, energy use in water and wastewater Direct water useWater abstractionWater distributionenergyRaw watercollection energyWater treatmentenergyWastewater(WW)Water useResidential wateruseReclaimed waterResidentialindoor useResidentialoutdoor usePopulationWaterappliances/faucetsWater and wastewaterconveyance and treatmentenergy useEnergy recoveredWW treatmentenergyWW transportenergy+++++-++++++++++ + +++-+Water footprint (WF) ofconveyance and treatmentenergy use+Total WF of urbanwater system (UWS)+Total energy forUWS+GHG emissions fromconveyance and treatmentenergy useGHG emissions fromhot water energy use+Total GHG emissionsfrom UWS+ +Energy for residentialhot water use++Commercial, institutionaland industrial (CII) floorspaceCII indoor useCII outdoor useDwelling unitsAgricultural landAgricultural waterCII waterGolf and park landGolf and parkswater +++++++++++Energy for CII indoorhot water use+Energy for hotwater use+ ++Infiltration andinflow+Major chemicals forww treatmentMajor chemicals forwater treatment+Embodied energy ofwater treatment chemicals+CF of watertreatment chemicals+++WF of watertreatment chemicals++Embodied of wwtreatment chemicals+Carbon footprint (CF) ofww treatment chemicals++WF of wwtreatment chemicals++++WF of hot waterenergy use+ ++80  transport and treatment, major chemical uses in the water and wastewater treatment of the CoP from 2005 to 2014. The water treatment plant capacity was upgraded in 2008 to 2009 (City of Penticton 2015b), whereas the wastewater treatment plant was upgraded during 2009 to 2012 in CoP (City of Penticton 2014a). The energy consumption rates of both treatment plants after these upgrades are more applicable for energy use modelling and forecasting. Therefore, for the validation of energy use in water and wastewater conveyance and treatment, the simulated results from 2013 to 2014 were used. It is noteworthy that the supplied drinking water was not used for agricultural irrigation in the CoP. The WEC model was further validated with direct structural tests, structure-oriented behavior tests, and behavior pattern tests (Barlas 1996).  5.3.2 Data requirements The required data were collected from various sources. Different periods of data were used for model calibration and validation.   a. Water consumers The data on base population, growth rate, and dwelling occupancy for Penticton were obtained from the census database (Statistics Canada 2015a). The average lot size of residential houses and CII buildings were estimated from the municipal GIS database using ArcGIS. They were verified with the zoning bylaws. In absence of data, the growth rate of CII sector was assumed equal to population growth rate as the growth of CII sector follows population growth. The baseline data on school students, hospital beds, and hotel rooms were obtained from the respective authentic sources as mentioned in detail in Appendix B.2.  b. Water and wastewater The rates and water efficiencies of uses of water fixtures and appliances in residential and commercial and institutional (CI) sectors; industrial water use rate; and irrigation rates were obtained from literature as mentioned in detail in Appendix B.2. Moreover, average lot coverages (%) for different residential houses and CII buildings were estimated by using Google Earth and were verified with the zoning bylaws. The rate of change of indoor water conservation rate, monthly average infiltration and inflow rate, and the average decreasing rate of change of monthly infiltration-flow rate specific to CoP were estimated based on the Penticton data and 81  their detail methods are explained in Appendix B.2. Moreover, the water footprint of major chemicals used in water and wastewater treatments were obtained by conducting a life cycle assessment (LCA) by using SimaPro 8.0.5 (Risch et al. 2014). c. Energy consumption The rate of change of indoor hot water energy conservation rate; monthly average energy consumption rates separately for raw water collection to wastewater treatment; average dosages of major chemical uses; and average rate of increase of monthly energy consumption rate for water and wastewater treatment in Penticton were estimated and explained in detail in Appendix B.2. The embodied energy of the major chemicals was obtained from the same LCA conducted for estimating the water footprint of these chemicals in the earlier section (Risch et al. 2014). As far as data was available, the data of Penticton was used for the estimation of the parameters. The average Canadian values were used for other parameters, for instance, hot water ratios for different water uses were obtained from the national Residential End Use Model (REUM) (Aguilar et al. 2005). The energy efficiency of water fixtures and appliances was obtained from the REUM model (Aguilar et al. 2005) for conventional fixtures and from ENERGY STAR (2014a; b) for efficient fixtures in Canada. d. Carbon emissions The carbon emissions from energy use were estimated using the carbon emission factors of the respective energy sources (Ministry of Environment 2013). The carbon emissions from wastewater processes were estimated based on the IPCC methodology (IPCC 2006).  Moreover, the carbon footprint of major chemicals was obtained from the same LCA conducted for estimating the water footprint of these chemicals in the earlier section (Risch et al. 2014).   The data for all parameters are categorically shown as: regional containing regional and site-specific data (R); national (N), and global (G) in Appendix B.2. In the lack of full data set, the model can be run only by changing the regional data for other cities or communities.  5.3.3 WEC model for Penticton In the developed WEC DSS, the major data can be input from the interface; however, if a community has detailed data in addition to those mentioned in the interface, all the data can be 82  imported from a spreadsheet. The results can be exported to a spreadsheet and the major ones will be displayed on the interface. The interface of the WEC DSS is shown in Figure 5.2. The slider, button, and dropdown list are used for data input. After model simulation, the major outputs are graphically and numerically displayed.    Figure 5.2    Screenshot of the WEC DSS interface The developed WEC model was applied to the SMUWS of the City of Penticton. The city, with an area of 42.1 sq. km, had a population of 32,877 in 2011 with a growth rate of 0.6% per year (Statistics Canada 2015a). The CoP supplies drinking water through 197 km of water mains having three pump stations, two booster stations, and a water treatment plant (City of Penticton 2015b). The generated wastewater is collected by gravity system and then pump to the wastewater treatment plant (Biological Nutrient Removal) by using 10 lift stations (City of Penticton 2014a). The WEC model was simulated for Penticton from 2005 to 2014. Based on the data availability, the developed model was validated by using the historical data of monthly water consumption 83  and wastewater collection from 2005-2014, energy use for water and wastewater conveyance and treatment from 2013 and major chemical use for the treatments from 2010 to 2014.  5.3.3.1 Water use and wastewater generation The simulated monthly water consumption from 2005-2014 has a coefficient of determination (r2) of 0.89 and is compared with its actual data in Appendix B.3. For the 10-year period, the WEC model resulted in an average water consumption of 622 L/capita/day, which is 1 % higher than the actual value. In particular, the WEC model estimated the indoor water consumption rate to be 220 L/capita/day for 2005 to 2006, which is comparable with the metered data of 222 L/cap/day for indoor water in 2006 in Okanagan, including Penticton and Kelowna (Maurer 2010). The WEC model’s simulation resulted in an average of 48.5% of residential water for outdoor irrigation for 2005-2006, which is comparable with the value of ~ 50% for Penticton given by previous studies (Maurer 2010; Neale 2005). From 2008-2014, the r2 of the wastewater sub-model is 0.85 and the predicted average wastewater collection was 370 L/cap/day, a value 1.2% higher than the actual value.  5.3.3.2 Energy use For the predicted energy use in raw water abstraction and conveyance, water treatment, and wastewater treatment, the values of r2 were 0.84, 0.85, and 0.76 respectively. The differences in mean values of energy use by utilities in various urban water stages for actual and modelled data are insignificant ranging from -1.9% to -0.3%. The energy consumed by hot water use (excluding energy for water use in space heating and mechanical work in laundry and dish washing machines) is also an energy use component of a SMUWS. The baseline energy consumption estimated by the WEC model at the start of 2005 is 1906 kWh/capita/year, which is comparable with 1913 kWh/capita/year (based on the national dwelling size of 2.55) reported by the REUM model (Aguilar et al. 2005). From 2005 to 2011, the average hot water energy use at the national level was 2437 kWh/capita/year (Natural Resources Canada 2014), whereas the WEC model estimated 1814 kWh/capita/year for CoP. The value estimated by the WEC model was about 25% lower than the national value because the WEC model was based on the REUM 84  model and the REUM model’s estimation itself was 25% lower than the national value of 2,546 kWh/capita/year for 2005 (Natural Resources Canada 2014).   The WEC model estimated an average energy use of 26 kWh/m2/year for indoor hot water use in the CI sector from 2005 to 2011. For the same period, the survey-based national historical database reported an average of 38 kWh/m2/year (Natural Resources Canada 2014). The value estimated by the WEC model is 30% lower than the national average because the WEC model considered the same pattern of indoor hot water use in the CII sector as of the residential sector and the residential sector estimates a 25% lower value than the national database. However, in spite of different estimation methods of the WEC model (REUM model-based) and the national database (survey-based), the order of magnitude is similar, indicating that the results of the WEC model are reasonable.       5.3.3.3 Carbon emissions The WEC model resulted in an average of 244 kgCO2e/capita/year for hot water energy use in the residential sector in the CoP from 2005-2011, whereas the historical data reported 447 kgCO2e/capita/year in Canada during the same period (Natural Resources Canada 2014). The WEC model estimated 45% lower than the national database. Similarly, the WEC model estimated an average of 3.6 kgCO2e/m2/year in the CI sector in the CoP from 2005-2011, whereas the national database reported 7 kgCO2e/m2/year in Canada during the same period. The estimated value is 51% lower than the national level value.  The lower carbon emissions by hot water use in the residential and CI sector in the CoP has two major reasons. First, the energy use estimation based on the REUM model itself gives a 25% lower value than the national value. Secondly, the grid electricity in the CoP is provided by FortisBC, which has more than 95% of electricity generated from hydropower and renewable energy, such as wood waste (BC Hydro 2015; Ministry of Environment 2013), whereas the national grid electricity contains only 58% hydroelectricity with 28% thermal power-based electricity that has a higher carbon footprint 85  (Canadian Electricity Association 2006). However, some variations may also be resulted by a difference in theoretical estimation and actual data of energy use.   GHGs, i.e., methane and nitrous oxide emissions from wastewater treatment processes were estimated based on the IPCC methodology for a centralized aerobic treatment  system (IPCC 2006), which is valid. For the validation of the carbon footprint (also for the water footprint and embodied energy) of the major chemicals, the simulated results for the amount of chemicals consumed are compared with their actual values. The r2 value is 0.83 for chlorine and that for polymer is 0.88. Moreover, the mean difference in the actual and simulated values in all major chemicals are insignificant ranging from -1.6% to 0.2%. In addition, the difference in average monthly biosolids generation between the actual and modelled data is -1.6% from 2013-2014, indicating a valid estimation. Since, GHG emissions are globally estimated indirectly from energy consumption (Ministry of Environment 2013), the validated energy use, chemicals use, biosolids transportation, and valid emission factors result in a valid carbon module.   Overall, from 2005-2011 the WEC model estimated a reduction of carbon emissions by 1.3%/capita/year for the residential sector, whereas Natural Resources Canada (2014) reported a value of 1.8%/capita/year at the national level. Similarly, the WEC model estimated a reduction of 1.3%/m2/year for CII sector, whereas Natural Resources Canada (2014) reported a national reduction of 1.4%/m2/year in the same period. A slight variation in these trends was mostly due to the fact that the WEC model considered the static proportion of energy sources used for water heating as the model is primarily developed for the planning of water in new neighbourhood developments. Also, location-specific features can result in different emission trends in the CoP compared to the national trend; for example, the emission factor of grid electricity differs regionally.      5.3.3.4 Quantitative WEC nexus All the modules of the WEC model have a complete dataset for the years 2013 and 2014. Therefore, the WEC model provides a completely validated result for all interconnected entities in that period. The annual direct water use, total energy use, and total carbon emissions in absolute values in various stages of the operational phase of the Penticton UWS are given in 86  Figure 5.3. The average annual water footprint of the UWS was 7,625 ML with 92.7% direct water use, 7.2% water footprint of energy use, and ~ 0.03% water footprint of major chemicals used in treatments. The UWS consumed an annual total energy of 83,625 MWh with 89.5% by indoor hot water use (residential 66% and CII sector 23.5%). In the total energy, embodied energy of chemicals is insignificant (0.3%). A similar result of approximately 90% of operational energy consumption by hot water was also obtained by other studies (Graaff and Klaversma 2012; Reffold et al. 2008). Raw waterconveyanceWatertreatmentWaterdistributionWateruseWastewatercollectionWastewatertreatmentBiosolidstransportation7100 ML 7100 ML 6893 ML 3613 ML 4278 ML 8643 tBiosolids665 MLInfiltration/inflowGreenhouse gas emissions48 t CO2e 3.8 t CO2e 9703 t CO2e 3.9 t CO2e2782 MWhEnergy use221 MWh 70,947 MWh 175 MWh 15.7 MWhChemicalsEE: 203 MWh1169 MWh 3695 MWhChemicalsEE: 35 MWh3.5 t CO2e 11 t CO2eChemicalsCF: 91 t CO2eChemicalsCF: 46 t CO2e0.5 t CO2eWW process499 t CO2eNote: EE: Embodied energy, CF: carbon footprint Figure 5.3    Annual direct water use, energy use and GHG emissions in various stages of Penticton UWS The average annual carbon emissions for the operational phase of the Penticton UWS was 11,047 tCO2e with the highest share of 93.5% from the residential (69.2%) and CII indoor (24.3%) hot water use. Surprisingly, the proportion of carbon emissions by indoor hot water use was 99.3% of carbon emissions by direct energy use. The very high proportion of carbon emissions by indoor hot water use is due to the highest energy use and primary use of natural gas for water heating (67%), which has high carbon footprint. Therefore, the energy use and carbon emissions of the UWS can be reduced significantly by using energy efficient hot water systems, behavioural change for reduced hot water use, and clean energy for indoor water heating. A saving of approximately 10% in hot water energy can provide energy for all other operational energy demand in the UWS.   The WEC nexus of the Penticton UWS for 2013-2014 is quantitatively shown in Figure 5.4. The overall correlations among water, energy, and carbon in an UWS are nonlinear. Spearman’s rank correlation coefficient (ρ) was used to estimate their interrelationship. The Spearman’s ρ 87  between water and energy, water and carbon, and energy and carbon were 0.94 (p= 0.000), 0.89 (p = 0.000), and 0.83 (p = 0.000), respectively (Figure 5.4), indicating highly significant interconnections.   Figure 5.4    WEC nexus of Penticton UWS 5.3.3.5 Sensitivity analysis Sensitivity analysis was performed to investigate the sensitivity of input parameters to the variations in the final outputs: total water footprint, total energy use and total carbon emissions. Since the WEC model has a large number of input parameters (more than 200) for sensitivity analysis, the input parameters were first screened and 102 parameters were selected based on the relatively important parameters identified by Venkatesh et al. (2014) and Kenway et al. (2008). A most commonly used sensitivity analysis method was used in which an approximate relative contribution of each parameter to the variance of the final outputs was estimated by squaring the rank correlation coefficients between input parameters and final output and then normalized to 100% (Hammonds et al. 1994; Sadiq et al. 2004b).  The parameters with the highest relative contributions are considered to be the most sensitive input parameters, which would contribute to reduce the largest amount of overall uncertainty in the results (Hammonds et al. 1994).  The WEC model was simulated for the complete validation period 2013-2014 by using Monte Carlo simulations of 10,000 runs in Stella Professional® 1.0.3 by considering uniform distributions for the screened parameters (Sadiq et al. 2004b) as given in Appendix B.4. Since 88  the contributions of input parameters to the final outputs were highly dispersed due to a large number of input parameters (102), the contributions were estimated for the aggregated inputs as given in the sensitivity analysis framework in Appendix B.4. Furthermore, the contributions of basic inputs to the variance of aggregated inputs can be estimated in the same way. The results of sensitivity analysis are shown in Table 5.2.  Table 5.2    Percent contribution of parameters to the variability of the WEC model Parameters % contribution Parameters % contribution Total water footprint  Total energy use  WF_chemicals 33 Hot shower energy (resi) 15 Residential outdoor irrigation 25 NR hot dishwasher energy 14 Shower water (resi) 7 Resi_standby energy loss 14 WF of energy use 7 Hot dishwasher energy (resi) 10 Faucet water (resi) 6 Hot faucet energy (resi) 10 Other municipal water 5 EE_chemicals 8 Toilet water 5 WW transport energy 6 Total carbon emissions  WW treatment energy 6 CF_residential hot water 46   CF_NR hot water 27   CF_chemicals 9             Note: resi= residential, NR= non-residential or CII sector  As shown in Table 5.2, for the variance of the total water footprint of the Penticton UWS, the water footprint (WF) of chemicals (33%) and residential outdoor irrigation (25%) were the largest contributors. Although the WF of chemicals seems to have the highest contribution, in fact the major contributors to the variance of the WF of chemicals were water use (58%), residential wastewater (26%), and infiltration-inflow to sewer network (13%) rather than the unit WF of chemicals. Similarly, another parameter the WF of energy use also contributed over 5%; however, the major contributors to the variance of the WF of energy use were the amount of energy consumption (70%) and partly by unit WF of electricity (26%). For the variance of the total energy use, the largest contributors were hot shower energy (residential) (15%) and residential standby energy loss (14%). All the parameters of energy use given in Table 5.2 are directly related to water use activities except the residential standby energy loss. However, the variance of the residential standby energy loss was also affected two-89  thirds by indoor water use activities (shower, faucet, and laundry water), and one-third by the standby energy loss rate. Similarly, the variance of another parameter the embodied energy of chemicals was mainly affected by water use activities and partly by chlorine consumption and unit embodied energy of chlorine. For the variance of the total carbon emissions of the UWS, the highest contributors were the carbon footprint (CF) of residential hot water (46%) and carbon footprint of non-residential (NR) hot water (27%). The highest contribution by the carbon footprint of residential and NR hot water was obvious as they represent about 93% of the total carbon emissions of the UWS and variances in these inputs would have significant effects on the total carbon emissions. Although the carbon footprint of chemicals seem to affect significantly to the variance of the total carbon emissions, the major contributors to the variance of the carbon footprint of chemicals were water use activities rather than unit carbon footprint of chemicals. Monte Carlo-based based sensitivity analysis has widely been used, such as by  Sadiq et al. (2004), Zio and Pedroni (2012), Veihe and Quinton (2000) and Veihe et al. (2000). The technique removes the difficulties with the traditional linear approach using single-valued inputs that in reality are random variables with the associated distributions (Veihe and Quinton 2000). However, it considers that input are independent (US EPA 1997; Veihe and Quinton 2000). 5.3.4 Scenario analysis Various scenarios can be developed and analyzed by using the WEC model in order to identify an optimum WEC nexus or intervention in the SMUWS. In this study, 10 scenarios in five categories, namely business as usual (Category 1), indoor water demand management (Category 2), outdoor water demand management (Category 3), source water alternatives (Category 4), and water heating energy alternatives (Category 5), were developed for the CoP to improve the sustainability of the UWS (Table 5.3). In order to have a large possible improvement in the WEC nexus, the categories were developed as cumulative from Category 1 to 5; however, the letter suffices indicate alternative scenarios within a specific category.    90  Table 5.3    Scenarios developed for the UWS of Penticton Scenarios Features Challenges/required actions 1 Business as usual -  Indoor water demand management  2 Scenario 1 with efficient (best available) water fixtures: water efficient toilet, showerhead, faucet, waterless urinals (in CII sector) and energy and water efficient cloth washers and dishwashers  Awareness and rebate programs  Outdoor water demand management alternatives  3A Scenario 2 with 50% irrigation demand reduction in lawn and parks Xeriscaping and efficient irrigation 3B Scenario 2 with lawn size reduction; lot coverage increased from 43% to 70% in single family houses and from 45% to 70% in duplex Policy change to increase lot coverage of single family houses and duplex 3C Scenario 2 with high density housing; residents moved from single family houses to small and large apartments equally; single family houses reduced from 51.3% to 5% Awareness to prefer high density housing  Source water alternatives  4A Scenario 3A with treated wastewater reuse for park and lawn irrigation and toilet flushing in residences Secondary distribution pipes for reclaimed water is a challenge  4B Scenario 3A with rainwater harvesting in residences; harvested water use in toilet, lawn, and laundry Awareness and incentives for rainwater harvesting  Water heating energy alternatives  5A Scenario 4A with natural gas increased from 66.8% to 95% and remaining 5% by electricity Majority of residents switch to natural gas for water heating for hot water use 5B Scenario 4A with electricity increased from 26.9% to 95% by replacing all natural gas and by reducing oil from 5.8% to 4.5%; 0.5% propane is as usual Awareness and incentives to use electricity for water heating for hot water use 5C Scenario 4A with 95% solar thermal energy and 5% electricity Awareness and incentives to use solar thermal energy for water heating for hot water use 5.3.4.1 Business as usual scenario  The WEC model was simulated from 2015 to 2034. For Scenario 1, i.e., business as usual scenario, the monthly water footprint, energy use, and carbon emissions of the UWS is shown in Figure 5.5. Figure 5.5.shows a decreasing trend for all three elements primarily due to increasing water conservation and the use of efficient water fixtures and appliances.    91   Figure 5.5    Monthly water footprint, energy use, and carbon emissions under Scenario 1 from 2015 – 2034  The average annual water footprint, energy use, and carbon emissions of the UWS will be 6,539 ML; 72,997 MWh, and 9,644 tCO2e respectively. The results of all scenarios in terms of the change with respect to Scenario 1 are shown in Figure 5.6.   Figure 5.6    Change in average annual water footprint, energy use and carbon emissions compared to Scenario 1    92  5.3.4.2 Indoor water demand management Indoor water demand management strategies in Scenario 2 can reduce the average annual water footprint, energy use, and carbon emissions by 19%, 37%, and 36% respectively compared to Scenario 1. Required actions for Penticton, as given in Table 5.3, include rebate programs for water fixtures and appliances (toilets, washing machines, and showers) although the toilet rebate program was once launched in 2006 (Maurer 2010).  5.3.4.3 Outdoor water demand management Outdoor water demand management can be considered in addition to indoor water demand management as in Scenario 3C. The average annual water footprint, energy use, and carbon emissions can be reduced up to 41%, 38%, and 36% respectively in Scenario 3C compared to Scenario 1. However, it requires a larger behavioural change among the residents, specifically, 95% of them should prefer high density housing, such as apartments and row houses. Alternatively, municipal policy change towards a higher lot coverage as in Scenario 3B, can achieve similar results, especially for total energy use and carbon emissions, but it reduces landscaping area in the community. It is noteworthy that xeriscaping can reduce more than 50% of the water demand of lawn and park irrigation (Boot and Parchomchuk 2009).  By xeriscaping as in Scenario 3A, the average annual water footprint, energy use, and carbon emissions can be reduced up to 34%, 38%, and 36% respectively. In this analysis, the carbon sequestration by landscaping was not considered. 5.3.4.4 Source water alternatives Scenario 4A (wastewater reuse) can reduce the average annual water footprint, energy use, and carbon emissions up to 57%, 39%, and 36% respectively compared to Scenario 1. In Scenario 4A, the rate of energy use for the secondary distribution of reclaimed water was considered to be the same as that of drinking water distribution energy. Since, Penticton has been using reclaimed water for irrigating golf courses and public parks, additional treatment, except chlorination, of reclaimed water may not be required for further use. In Scenario 4B (rooftop rainwater harvesting) can reduce an average annual water footprint, energy use, and carbon emissions up to 47%, 29%, and 36%, respectively compared to Scenario 1, indicating increased energy use by 93  7% from Scenario 3C due to energy intensive rainwater harvesting in Penticton (semi-arid region). 5.3.4.5 Water heating energy alternatives Scenario 5B (increased used of electricity in indoor water heating) reduces the carbon emissions of the UWS up to 87% as a result of a lower carbon footprint of hydro-based electricity than that of natural gas and oil (Ministry of Environment 2013). Alternatively, the increased use of solar thermal energy for water heating (Scenario 5C) can reduce carbon emissions by 86%. It  is noteworthy that the trend of natural gas use for water heating has been increasing in Canada since the past 10 years (Natural Resources Canada 2014). However, Scenario 5A (increased use of natural gas) can reduce carbon emissions only by 22% compared to Scenario 1, indicating an increased carbon emissions of 14% compared to Scenario 4A due to a higher carbon footprint of natural gas (Ministry of Environment 2013). Scenarios 5B and 5C were better scenarios based on lower carbon emissions and energy use; however, Scenario 5C had the best performance with an additional decrease of annual water footprint, but a detailed feasibility study is recommended on the use of solar thermal energy to meet the energy demand of water heating throughout the year, especially in winter. The developed 10 scenarios include extreme cases in addition to the business as usual scenario. The extreme scenarios, such as Scenario 5C will assist a decision maker to approximate a range of extreme values or uncertainty in a particular intervention.    5.3.4.6 WEC nexus analysis of interventions The WEC nexus for important individual interventions in UWSs identified in the previous sections were further analyzed for Penticton in 2015-2034. For this purpose, the individual water-based interventions considered were efficient water fixtures as in Scenario 2; xeriscaping as in Scenario 3A (without additional Scenario 2 features in Scenario 1); and wastewater reclamation as in Scenario 4A (without additional Scenario 3A features in Scenario 1). Similarly, the energy-based individual interventions considered are natural gas dominancy (as in Scenario 5A), electricity dominancy (as in Scenario 5B), and solar thermal energy dominancy (as in 94  Scenario 5C) without the additional Scenario 4A features in Scenario 1 for these interventions. The analysis results are shown in Figure 5.7.   Figure 5.7    WEC nexus analysis of interventions: water-based (a & b) and energy-based (c & d) The unit water saving would result in significantly different energy saving (p = 0.000) and carbon saving (p = 0.000) for various water-based interventions based on repeated measures ANOVA (Figure 5.7).The very high energy saving by efficient water fixtures is due the fact that the reduced water eliminates its energy requirement from raw water collection to wastewater treatment and disposal, whereas the reduced water in irrigation (xeriscaping and water reclamation interventions) eliminates only upstream energy requirement (raw water collection to water distribution). Moreover, efficient water fixtures reduce hot water demand and improve energy efficiency (ENERGY STAR 2014a; c). The similar features as of energy saving were depicted by carbon saving per unit water saved as the carbon emissions are mainly from the energy use. In the energy-based intervention analysis, the replacement of 1 MWh of indoor water heating energy by different interventions have significantly different mean water saving (p = 0.000) and    (a) (b) (c) (d) 95  mean carbon saving (p = 0.000). The natural gas and solar thermal would result positive water saving, whereas it would be negative for electricity because of a higher water footprint of hydro-based electricity (19.7 L/kWh) of Penticton (Okadera et al. 2014) than that of natural gas (0.4 L/kWh) (Okadera et al. 2014)  and solar thermal (3.98 L/kWh) (Fulton and Cooley 2015). Furthermore, carbon emissions would be increased by 0.05 tCO2e for the energy replacement by natural gas, whereas the carbon emissions would be decreased by about 0.16 tCO2e for the replacement either by electricity or solar thermal energy. The carbon footprint of natural gas is higher (Ministry of Environment 2013) than that of electricity (Ministry of Environment 2013) and solar thermal (Menzies and Roderick 2010). This dynamic WEC nexus analysis of interventions would provide better results than a simple unit footprint-based calculation as the dynamic model incorporates the feedbacks in the UWS. The estimated values may be associated with uncertainties, which can broadly be classified as aleatory uncertainty – due to natural variation resulting in uncertain data or parameter values (or parameter uncertainty) and epistemic uncertainty – due to imperfect understanding of the system (model uncertainty) (Dyck et al. 2014). In particular to the present application, exponential growth has been used for human population, CII sector, and other parameters, such as inflow-infiltration and energy consumption. The estimated exponential growth rates were found to be applicable to the historical data and the similar growth pattern with adjusted rates were used for future forecasting (20 years). However, the growth rates may be associated with high uncertainties when forecasted for a very long duration due to variations in dynamics among model variables. The parameter uncertainty can be approximated by Monte Carlo simulations as in sensitivity analysis and to some degree by scenario analysis. Furthermore, uncertainty can be reduced by better quality and region specific data with improved parameter relationships. Such improvement in sensitive parameters will highly reduce uncertainty in the WEC nexus. 5.4 Summary A comprehensive water-energy-carbon (WEC) nexus model for an urban water system (UWS) by using system dynamics is proposed to assist municipalities, urban developers, and policy makers for neighbourhood water planning and management. The proposed model and decision support system was developed for the operational phase of SMUWSs by using Stella®. The 96  model was validated using historical water and energy consumption data (2005-2014) of Penticton (British Columbia). Spearman’s correlation coefficients between water and energy, water and carbon, and energy and carbon were 0.94, 0.89, and 0.83 respectively revealing highly significant interconnections. The energy for water was 11.1 MWh/ML, water for energy was 6512 L/MWh, and carbon emissions were 124.4 kg CO2e/MWh from energy use and 120.8 kgCO2e/ML from wastewater processes. The highest energy consumer was found to be the indoor hot water use in the residential and CII sector consuming approximately 90% of the operational energy demand and contributing about 93% to carbon emissions. Indoor hot water use should be prioritized for the reduction of energy use and carbon emissions. The contributions of residential outdoor irrigation, shower water, faucet water, and non-residential (CII) and residential hot dishwasher energy to the model variability were higher than other parameters. A Monte Carlo-based sensitivity analysis showed residential outdoor irrigation and water heating energy for shower and dishwasher have higher contribution to model variability. The intervention analysis reveals significant differences in savings in water, energy, and carbon for various water and energy-based interventions in SMUWSs and the developed DSS is well capable for analyzing these dynamic savings. The developed decision support system is capable of dynamic analysis of different WEC-based interventions to improve the sustainability of SMUWSs. The decision support system can be used by utilities, urban developers, and policy makers for sustainable urban water planning to reduce water consumption, energy use, and carbon emissions in neighbourhoods. Furthermore, the tool can also be used for operational neighbourhoods to forecast future WEC nexus.   97   Investigating Impacts of Residential Density on WEC Nexus A version of this chapter has been published by the Journal of Cleaner Production entitled “Impacts of neighbourhood densification on water-energy-carbon nexus: Investigating water distribution and residential landscaping system” (Chhipi-Shrestha et al. 2017c). 6.1 Background Water Distribution Systems (WDS) can significantly consume energy and release greenhouse gases (Hellstro 2000; Wu et al. 2015). Per capita energy requirements for WDSs can be reduced by developing high residential density in communities  (Filion 2008). Densification is one of the major reasons to reduce per capita infrastructure and land requirement including water infrastructure (Duncan 1989; Frank 1989; Gleick et al. 2003). Dense residences comprising multi-family (MF) buildings can also have a lesser irrigation demand for landscaping since landscaping requirements for MF buildings are lower (City of Kelowna 2007; City of Penticton 2015a; District of Peachland 2014a). Reduced water demand also has a decreased upstream energy use. These interlinkages suggest that higher density can have reduced water use, energy requirement, and carbon emissions. However, a lesser amount of landscaping due to dense residences results in reduced carbon sequestration. Atmospheric CO2 is photosynthesized and stored as plant biomass by herbs, shrubs, and trees. Shrubs and trees store carbon for a long term in their biomass as carbon stock, whereas herbs (e.g., grasses) decay and some of their biomass-carbon is humified and stored for a long term in soil as soil organic carbon (SOC) (Zirkle et al. 2011, 2012). Furthermore, the life cycle cost of WDSs may be lower in dense residences  (Speir and Stephenson 2002).  Energy is required for water abstraction, conveyance, treatment, and distribution. The required energy varies with several factors, such as topography, source water quality, urban form, population density, and adopted management strategies. Energy can also be harvested from the hydraulic energy of WDSs (Ye and Soga 2011), as well as from the thermal and chemically bound energy of wastewater (Meda et al. 2012; Nowak et al. 2011). In addition, the change of residential density directly affect landscaping that in turn affect the amount of carbon sequestration. Water is required directly and indirectly for energy generation. Direct water is required for hydroelectricity generation,  whereas a large amount of indirect water is necessary 98  for the exploration, extraction, and beneficiation of fossil fuels (Meda et al. 2012) and also for renewable energy crop cultivation, such as biofuel. This interconnection shows a complex WEC nexus of Water Distribution and Residential Landscaping System (WDRLS). The impacts of neighbourhood densification on the WEC nexus of WDRLS are shown in Figure 6.1.  Water Distribution and Residential Landscaping System+                    Outdoor water demand              -+                     Energy use                                    -+                     Carbon emissions                        -+                    Carbon sequestration                  -+                    Life cycle cost                                -Densification Figure 6.1    Impacts of neighbourhood densification on the WEC nexus of WDRLS (per capita) This chapter aims to study the impacts of urban residential density on the WEC nexus of water distribution and residential landscaping system. The results will help municipalities and urban developers to identify water and energy efficient WDSs and landscaping with respect to different residential densities. In addition, the findings will help them to decrease the resulting carbon emissions and water distribution cost. 6.2 Methodology A conceptual framework for studying the impacts of neighbourhood densification on the WEC nexus of water distribution and residential landscaping system has been developed (Figure 6.2). The framework is explained in the following sections.   99  Neighbourhood configurationIndoor water  demandUrban landscapingDensificationLow <----------------> HighWater Energy CarbonCarbon sequestrationWEC nexusOutdoor water demandWEC aggregation based on ecological footprintWater Distribution SystemEnergy footprintChange in water demandPer capita WEC variation over densitypumping Figure 6.2    Conceptual framework to study the impacts of neighbourhood densification on the WEC nexus of water distribution and residential landscaping system 6.2.1 Water demand As the first step, a neighbourhood mainly consisting residential buildings is designed on a given land. The neighbourhood design should follow municipal bylaws for lot coverage, road size, building mix (SF and MF buildings), building height, and public park designation. Also, alternative designs are prepared by changing residential density. The residential density can be increased by increasing MF buildings and decreasing SF buildings, and vice-versa. The developed alternative designs have different total lot coverages, indicating varying per capita residential landscaping size. These designs alter their overall and spatial water demand. The average water demand of each alternative neighbourhood is estimated based on Equations  6.1 to 6.4. Then, WDSs including fire demand are designed as per the concerned municipal bylaw for each alternative by using EPANET 2 (Aydin et al. 2014).   𝑊𝑛𝑒𝑖𝑔ℎ𝑏𝑜𝑢𝑟 = 𝑊𝑟 + 𝑊𝑐𝑖 + 𝑊𝑝𝑎𝑟𝑘                                 Equation 6.1   100  𝑊𝑟 = [ ∑ (𝜂𝑓 ∗ 𝐹𝑓)𝑛𝑖 𝑖] ∗ 𝑃𝑡 + 𝐿 + (𝐴𝑟)𝑙 ∗  𝐼𝑙                      Equation 6.2 𝑊𝑐𝑖 = 𝑁𝑐𝑖 ∗ 𝐶𝑐𝑖 + (𝐴𝑐𝑖)𝑙 ∗ 𝐼𝑙                                          Equation 6.3] 𝑊𝑝𝑎𝑟𝑘 = 𝐴𝑝 ∗  𝐼𝑝                                                          Equation 6.4 where Wneighbour is neighbourhood  water demand, Wr is residential water demand, Wci is commercial and institutional water use (Offices, retails, hotels, and schools), Wpark is parks water demand; η is efficiency of fixtures (L/use), subscripts r, ci, f, l, and p  refers to residential, commercial and institutional, fixture, landscaping, and parks respectively, i refers to fixtures toilet, shower, faucet, laundry, and dishwasher; F is frequency of fixture use (no. of uses/p/d), Pt is total population; L is leakage; A is area (ha); I is irrigation rate (L/ha/d); Nci is number of rooms for hotels and floor space (sqft) for offices, retail, and schools, and C is indoor water consumption rate (L/room/d for hotels and L/sqft/d for offices, retails, and schools).  6.2.2 Energy use The designed WDS is used to estimate the required energy use for water distribution based on the capacity of pump and its duration of use. The design provides utility energy (water mains energy). However, for mid-rise and high-rise residential buildings, i.e., for apartments, additional energy is required for booster pumps at apartments to supply water to elevated levels. The additional energy for a booster pump (house pumping energy) at each apartment can be estimated using Equations 6.5 and 6.6 for buildings under 15 stories (Cheng 2002).  𝑃 =𝛾𝑄𝐻𝑝(1+𝛼)𝜂∗𝜂𝑡∗1000           Equation 6.5 𝐸ℎ = 𝑃 ∗ (1 + 𝑓) ∗ 𝑁         Equation 6.6 By combining Equations 6.5 and 6.6,  𝐸ℎ = 2.23 ∗ 10−3 ∗ 𝛾𝑄𝐻𝑝 ∗ 𝑁 for α =0.2, f =0.3, η = 0.7 and ηt =1    Equation 6.7 where P is power of lift pump (kW), γ is specific weight of water (9806 N/m3), Q is pumping capacity of lift pump (m3/s) which can be estimated from an average water discharge considering a peak factor for hourly demand of  7.4 (Ontario Ministry of Environment 2008), Hp is height 101  from the lift pump to the top of the building (m) and can be estimated as using Hp = 3.1 (F+1) with F as number of floor and 3.1 m as the floor-floor height for residential building (Council on Tall Buildings and Urban Habitat 2016), α is safety factor of pumping power  (0.1 to 0.2), η is pump efficiency (65% to 85%), ηt is mechanical transmission efficiency (92% to 100%),  Eh is Energy consumed (kWh) in a house pump, f is friction loss within pipes (30%), N is number of hours a pump is operated (Cheng 2002). The energy required for WDS (EWDS) is the sum of water mains energy (Em) and house pumping energy (Eh) as in Equation 6.8.  𝐸𝑊𝐷𝑆 = 𝐸𝑚 + 𝐸ℎ         Equation 6.8 The reduced water demand in high density residences also lowers upstream energy use of neighbourhood water demand, such as energy for conveyance and water treatment. The change in total energy requirements in different alternatives due to densification can be estimated by using Equation 6.9.  ∆𝐸𝑡𝑜𝑡𝑎𝑙 = 𝐸𝑊𝐷𝑆 + ∆𝑊 ∗ 𝐸𝐼𝑢𝑝𝑠𝑡𝑟𝑒𝑎𝑚        Equation 6.9 where ∆Etotal is change in total energy requirement (kWh), EWDS is energy required for WDS of a particular neighbourhood design (kWh), ∆W is change in total water demand in the particular design compared to any design, and EIupstream is upstream energy intensity of water (kWh/m3)  6.2.3 Carbon emissions and sequestration The energy related carbon emissions of WDSs are estimated based on the energy use by using Equation 6.10. 𝐶𝐸 = 𝐸𝑊𝐷𝑆 ∗  𝐸𝑓          Equation 6.10 where CE is carbon emissions related to energy, EWDS is energy required for WDS (kWh) and Ef is CO2 emission factor for energy (kgCO2e/kWh). If different sources of energy are used, 102  estimate total carbon emissions for each energy source by multiplying energy use with its respective emission factor and then compute sum to estimate grand total carbon emissions. The carbon sequestration of landscaping constitutes the sequestration of SOC, shrub biomass-carbon and tree biomass-carbon (Lal and Augustin 2012; Zirkle et al. 2011). For soil, the net SOC sequestration rate is the balance of gross carbon accumulation and carbon emission during lawn maintenance practices (mowing, irrigation, fertilizer, and pesticides use). The total carbon sequestration in a residential landscaping can be estimated by using Equation 6.11. 𝐶𝑠 = 𝑆𝑂𝐶𝑠 ∗ ∆𝐴𝑙 + 𝐶𝑠𝑡 ∗ 𝑁𝑡 + 𝐶𝑠𝑠 ∗ 𝑁𝑠       Equation 6.11 where Cs is total carbon sequestration (kg CO2/yard/yr), SOCS is net SOC sequestration (kg CO2/m2/yr), ∆Al is net landscaping area after reducing the landscaping area by trees and shrubs canopy as the landscaping shaded by trees and shrubs is assumed not to be productive (Zirkle et al. 2012), Cst and Css are net carbon sequestration by trees and shrubs,  Nt and Ns are number of trees and shrubs respectively. The net carbon emissions by a WDS and residential landscaping in a neighbourhood can be estimated by using Equation 6.12. 𝐶𝑛 = 𝐶𝐸 − ∑ (𝐶𝑠)𝑖𝑛𝑖=1          Equation 6.12 where Cn is net carbon emissions (kg CO2/yr), CE is carbon emissions related to energy use in a WDS, Cs is total carbon sequestration by an individual landscaping (kg CO2/m2/yr), n is number of all residential landscaping with water supplied by the WDS. 6.2.4 WEC aggregation Water and energy are natural resources, whereas emitted GHGs (carbon) are pollutants from a global warming perspective. Generally, freshwater (streamflow) is produced within a river basin (land area). Energy generation requires landmass too, e.g., hydroelectricity production needs land area for river catchment and reservoir site; a large land area is required for fossil fuel extraction. Furthermore, vegetated land is needed for the sequestration of carbon emitted to the atmosphere. All these three elements – water, energy, and carbon – have a common feature of the use of land resources. This common feature can be used to aggregate them together as WEC nexus. 103  Therefore, water consumption, energy use, and carbon emissions can be integrated by converting them into a common measurement unit – ecological footprint.  The ecological footprint is a well-known resource accounting tool for measuring biologically productive land and water area, an individual or a region requires to produce the resources it consumes and to absorb the waste it generates, using prevailing technology and resource management (Kitzes et al. 2013; Musikavong and Gheewala 2016; Wackernagel and Rees 1996). Global hectares are used as a common unit to express an ecological footprint. A global hectare refers to a hectare that is normalized to have the world average productivity of all biologically productive land and water in a given year (Kitzes et al. 2013). The ecological footprint of water, energy, and carbon emissions can be obtained from the related literature as given in the application section, which can be summed together to estimate aggregated WEC. 6.2.5 Life cycle cost analysis Life cycle cost (LCC) analysis is an economic assessment method in which all costs arising from owning, operating, maintaining, and ultimately disposing of a project or product are considered (US DOE 1996). LCC of WDSs is estimated as the sum of capital cost, operation cost, and repair and replacement cost of water distribution infrastructure, namely pipes, pumps, and valves. The net present value (NPV) of the annualized LCC of WDSs is estimated by using Equation 6.13  (Davis et al. 2005).  𝑁𝑃𝑉 (𝑖, 𝑁) = ∑𝑅𝑡(1+𝑖)𝑡𝑁𝑡=0          Equation 6.13 where i is discount rate, i.e., 4% (C-SHRP 2002; Umer 2015), t is number of years, N is total planning duration (years), Rt  is net cash flow at time t.    104  6.3 Results 6.3.1 Application The proposed conceptual framework was applied as a case study to a newly planned neighbourhood located in the Okanagan Valley, British Columbia (BC), Canada as described in the following sections. 6.3.1.1 Study area The neighbourhood has an area of approximately 51 ha in a rugged topography with a maximum elevation difference of 80 m. The neighbourhood is planned for mixed use comprising residential, commercial, and institutional buildings. The neighbourhood is planned to have approximately 24% area covered by parks and trails. The planned residential population is approximately 4848 with a net residential density of 149 persons per hectare (persons/ha). Net residential density refers to the dwelling units or number of persons living in residential buildings divided by the land area covered by the buildings, private access ways, and local public roads (Landcom 2011). The neighbourhood has three zones: A – low density residential, B – medium density residential, and C – commercial area with high density residential buildings as per the information provided by the neighbourhood developer. The initial neighbourhood plan with lot division and configuration was obtained from the developer. A typical SF building in the neighbourhood has an average lot coverage of 365 m2 for building, garage, driveways, and sidewalk. The remaining area was assumed to be landscaped as per the municipal bylaw and provided neighbourhood plan. A lot coverage of 60% and 100 % was considered for medium and high density MF buildings respectively as per the municipal bylaw (District of Peachland 2014a). Without altering neighbourhood configuration, 11 neighbourhood designs were prepared by increasing residential density from D1 to D11 (Appendix C.1). The residential density was gradually increased by converting SF lots to MF lots by lot reorganization. In the series of design, Design D5 represented initially planned base 105  design. The lot reorganization was performed by maintaining the same average unit residential area and number of stories in SF and MF buildings as of base design (Design D5). A WDS was designed for the entire neighbourhood by following the development and servicing bylaw of the concerned municipality (District of Peachland 2004) and Ontario Ministry of Environment (2008). The maximum daily demand was estimated by multiplying the average daily demand (ADD) and maximum day factor, whereas peak hourly demand was estimated by multiplying ADD and peak hour factor. Also, the needed fire flow was estimated by using Fire Underwriters Survey (2007). Based on the municipal bylaw (District of Peachland 2004), the design flow is the greater value between peak hourly demand and the sum of the maximum daily demand and needed fire flow. The design flow was used to prepare WDS using EPANET 2 (Aydin et al. 2014). The energy intensity (kWh/m3) of water mains was estimated based on the pump capacity used in the WDS. The water mains energy consumption for residential water only was calculated from the estimated energy intensity and total residential water demand. The same process was repeated for all alternatives. In addition, house pumping energy was estimated for multi-family buildings by using Equations 6.5 and 6.6. The number of stories in a MF building can vary from six to ten as per the bylaw and an average of eight stories was considered in this study. Then, the total energy for WDS was estimated by using Equation 6.8. 6.3.1.2 Data  The neighbourhood was planned to be a sustainable community with the use of efficient water fixtures, e.g., efficient toilet, showers, cloth washers, etc. The efficiency and use frequency of various water appliances and fixtures as well as outdoor irrigation for residential and commercial and institutional (CI) buildings were estimated from the literature as given in detail in Appendix C.2. This study is focussed only on the residential water; however, CI water was also estimated to calculate total water demand. The total water demand is required to design a WDS to estimate energy use by residential water as in reality water is distributed by the same WDS to an entire neighbourhood  comprising residential, commercial and institutional buildings, and public parks. The energy related carbon emissions were estimated by using the emission factor of the grid electricity. The cost data of water mains installment and repair/replacement, electricity, water pumps, and valves were obtained from the related literature as given in Appendix C.2 and 106  carbon sequestration of residential landscaping was obtained from the related literature as given in Appendix C.3  The ecological footprint of Canadian freshwater is 1.08 x10-4 gha/m3/yr that was estimated as the reciprocal of annual freshwater availability per unit river basin area  (Kitzes et al. 2013; Meng et al. 2016; Statistics Canada 2003). Similarly, the ecological footprint of Canadian hydroelectricity is 3.9x10-6 gha/kWh/yr that was estimated as the ratio of per capita ecological footprint of hydroelectricity use (0.4 gha/p) (FCM 2005) and per capita hydroelectricity use (10,213 kWh/p/yr) (Environment Canada 2013). The ecological  footprint of carbon emissions is 0.224 gha/tCO2e (Kissinger et al., 2013). For the WDS, a 30-year planning period was considered (Speir and Stephenson 2002). 6.3.1.3 Residential density and WEC nexus The SF residences were changed to MF residences to increase residential density. To show the impact of such changes on water distribution and residential landscaping system, the variation of characteristics: lot coverage, landscaping coverage, SF units, MF units, population, WEC nexus, and LCC with respect to net residential density is presented in the next sections. i) Lot coverage and landscaping: The change in lot coverage and landscaping of residential lots with various net residential densities is presented in Figure 6.3. The number of SF and MF units considered and their population in various densities are also presented in the figure. 107   Note: D1 to D11 refers to the density range of the neighbourhood Figure 6.3    Lot coverage, landscaping, and population variation over net residential density Figure 6.3 shows that initially planned base Design D5 (149 persons/ha) had a lot coverage of approximately 32% and landscaping coverage of 68%. By changing all residential lots to SF lots, the net residential density will be very low with a value of 8.7 persons/ha (D1) with a lot coverage of about 16% and landscaping coverage of 84%. However, when all residential lots were used for MF residence, with a maximum of eight stories as per the bylaw, the net residential density would be very high with a value of 941.6 persons/ha (D11), with a lot coverage of 100%. This means that the neighbourhood can have a landscaping coverage from 0% (all MF residences) to 84% (all SF residences) as per the existing municipal bylaw and the information provided by the developer. Moreover, the change of SF and MF units resulted in the gradual increase of population from 283 (D1) to 30,631 people (D11). The variations of landscaping and population with net residential density have linear relationships with coefficient of determination (r2) of approximately 1 for landscaping and population. ii)    WEC nexus of WDS and landscaping: The per capita water consumption, energy use, and net carbon emissions by WDS and residential landscaping in various densities are shown in Figure 6.4. In Figure 6.4, per capita water consumption, energy use, and net carbon emissions 108  interact over net residential densities. The interaction indicates that all the three elements and respective net residential density should be taken into consideration in decision making.  Note: Negative net carbon emissions at 942 persons per ha (D11) is negative and not shown in the log scale  Figure 6.4    WEC nexus in various residential densities: Interaction plot The three elements were integrated to WEC nexus by converting them to ecological footprint (gha/p/yr). The estimated WEC nexus is given in Figure 6.5. The WEC nexus or simply ecological footprint (EF) sharply decreases from 0.08 gha/p/yr (D1) to 0.008 gha/p/yr (D4) and then gradually decreases to about 0.006 gha/p/yr (D11). A total of 93% of the EF (WEC nexus) was reduced from Design D1 to D11 by increasing net residential density from 8.7 persons/ha (D1) to 941.6 persons/ha (D11) as shown by the characteristic curve in Figure 6.5. The increased residential density would also increase the lot coverage from 16% in D1 to 100% in D11. The decrease in per capita water demand was achieved due to the decrease in landscaping requirements. The curve has a power relationship with r2 of approximately 84%. The power relationship can be used to identify an optimal density. The rate of change of the EF with respect 109  to net residential density (dy/dx) is also shown in Figure 6.5. The value of dy/dx is very low and almost similar beyond D6, indicating the density around 264 persons/ha as an optimal density.  Figure 6.5    Ecological footprint of various residential densities D1 to D11 The per capita water demand decreased from approximately 2,373 L/p/d (D1) to 136 L/p/d (D11) with a total of 94% reduction. Similarly, the energy intensity of water use decreased gradually from 1.2 kWh/m3 (D1) to 0.69 kWh/m3 (D11) as shown in the characteristic curve of energy intensity. Also, the per capita energy use for the WDS decreased from 2.8 kWh/p/d to 0.09 kWh/p/d (D11) with a reduction of 97% (2.5 kWh/p/d). The energy use includes water mains energy (utilities) and house pumping energy (apartment). Moreover, the reduced water demand from Designs D1 to D11 leads to upstream energy saving of 1.3 kWh/p/d resulting in the total energy reduction of 3.8 kWh/p/d from D1 to D11. Energy use emits carbon, whereas residential landscaping sequester it. The net carbon emissions were negative in Design D1 to D10, but positive in D11.  The negative net carbon emission or positive net carbon sequestration decreased from 260.9 g CO2e/p/d (D1) to 1.1 g CO2e/p/d in D10 and -0.3 g CO2e/p/d in D11, indicating positive carbon emissions in D11. All these characteristic curves of water demand, energy intensity, energy use, and negative net carbon emissions have power relationships with r2 ranging from 90% to 99%. Furthermore, the rate of change of per capita water demand, energy 110  use, and net carbon emissions with respect to density (dy/dx) are relatively low and almost unchanged beyond D6 similar to that of the ecological footprint.  6.3.1.4 Life cycle cost and residential density A characteristic curve of per capita LCC of the WDS and net residential density was prepared ( Figure 6.6). The per capita LCC, in terms of net present value, decreased sharply from $96.2/p/d (D1) to $ 9.7/p/d (D4) and then gradually to $ 2.6/p/d (D11), a 97% reduction from D1. The curve has r2 of approximately 99%. The characteristic curve also has a power relationship with residential density similar to that of per capita EF. In addition, the value of dy/dx of the curve is very low beyond D6. Therefore, the density around D6 (264 persons/ha) also represents an optimal density with respect to the per capita LCC of the WDS.   Figure 6.6    LCC of the WDS and its rate of change in different residential densities  6.3.1.5 Uncertainty and sensitivity analysis The uncertainty analysis was performed by estimating probabilistic water demand to approximate the associated uncertainty. The sensitivity analysis was performed to identify sensitive parameters in water demand estimation. i) Probabilistic water demand: The estimated water demand is affected by several factors making the results uncertain. The uncertainty analysis was conducted based on the probabilistic technique using Monte Carlo simulation with 10,000 simulations in each WDS design by using 111  the @Risk 7 (Lee et al. 2011). The simulations predict per capita probable water demand with randomly generated values for input parameters based on the given probability distribution (Lee et al. 2011). The parameters with their distributions are given in Table 6.1. Table 6.1    Major factors affecting water demand and their distribution parameters Factors Distribution Units Remarks Indoor water consumption rate N~ (133, 8.5) L/p/d 133 L/p/d is average for the neighbourhood1 and 150 L/p/d of  the Okanagan and North American2 was considered maximum value  Lot coverage (MF) T~(50, 60, 60) % 60% is maximum lot coverage Lawn irrigation N~ (991, 53) L/m2/yr 991 L/m2/yr is average3 & maximum value was considered as of golf irrigation (1005 L/m2/yr)3 Dwelling occupancy N~(2.5,0.15) Persons/DU 2.5 is average for the neighbourhood4 & 2.2 of a neighbouring city5 was considered as a minimum value. This distribution includes the highest provincial value6 of 2.6. Note: Normal distribution: N ~ (μ, σ) & assumed a truncated distribution for the given min to max range with μ ± 4σ representing 99.997% data; Triangular distribution: T~ (min, most probable, max values); DU= dwelling unit 1. From Appendix C.2  2. OBWB (2016)  3.OBWB (2010)   4. From developers’ plan  5. Statistics Canada (2015)  6. Statistics Canada (2014b)  The results show that per capita water demand band (within 5th and 95th percentile boundary) varies with different net residential density (Figure 6.7). The previously identified optimal density at D6 is consistent with this band. For Design D6, the per capita water demand may vary from 172 L/p/d to 204 L/p/d with the mean value of 188 L/p/d. In another way, to achieve the same average water demand of 188.4 L/p/d, the net residential density can vary from 224 to 316 persons/ha. This range of 224 to 316 persons/ha can be considered as an optimal density range as the characteristic curves of per capita water demand and the ecological footprint are similar.   112    Note: D1 is not shown here as its water demand is very high than others  Figure 6.7    Water demand variability (5th and 95th percentiles) in different residential densities  ii) Sensitivity analysis: Sensitivity analysis was conducted to study the effects of the variation of input parameters on the final output residential water demand. The parameters with the highest relative effects are considered to be the most sensitive input parameters. A reduction in the level of uncertainty (i.e., reducing variance) of the most sensitive parameters would contribute to reduce the largest amount of overall uncertainty in the results (Hammonds et al. 1994). The Monte Carlo simulations for probabilistic water demand were used for sensitivity analysis. The results are almost similar in all 11 neighbourhood designs. The most sensitive input parameters in all designs are indoor water use rate and dwelling occupancy. The effects of indoor water use rate and dwelling occupancy on total residential water use vary from -11% to 11% and -9% to 11% respectively. The effects of input parameters are not high, which may be due to less variation considered in the input parameters. However, the result provides a relative sensitivity of various inputs. 113  6.3.1.6 Two-dimensional analysis for WEC nexus scenarios The increase in residential density reduces per capita water demand to meet a given lot coverage requirement. Alternatively, increasing lot coverage requirement also reduces per capita water demand for a given density, indicating two-dimensional nature of the WEC nexus. The change in bylaws on lot coverage and/or application of xeriscaping would result in a change in water demand. A scenario analysis was performed by considering four scenarios for the same design series (D1-D11). Scenario S1 is with an existing lot coverage bylaw for all 11 designs. Scenarios S2 and S3 are with the modification of bylaw on lot coverage, whereas Scenario S4 is an application of xeriscaping in all 11 designs as shown in Table 6.2. Xeriscaping is low water-use landscaping in place of traditional turf (Sovocool et al. 2006). A xeriscaping of 15% of turf and 85% of water conserving species was designed in a typical SF building lawn of the neighbourhood and its detail is given in Appendix C.4 Table 6.2    Scenario features Scenario  Lot coverage in residence Remarks Single-family Multi-family* S1 348 m2 (10 - 40%) 60% Existing neighbourhood  plan and lot coverage bylaw S2 40% 60% Considered the existing maximum lot coverage guideline as an average coverage to be required S3 70% 80% Bylaw on lot coverage changed from maximum 40% to average 70% in SF and maximum 60% to average 80% in MF buildings S4 348 m2 (10 - 40%) 60% Xeriscaping in residential landscaping in Scenario S1 * Medium density MF buildings  The results of two-dimensional analysis of the WEC nexus are shown in Figure 6.8. The results of scenario analysis show that the reduction in per capita EF (aggregated WEC nexus) of water distribution and landscaping was highest in Scenario S4 (xeriscaping) among four scenarios and Design D1 among all designs. In Scenario S4, the reduction in per capita EF would range from below 1% in Design D11 to 66% (0.02gha/p/yr) in Design D1 compared to Scenario S1. The range of reduction would be less than 1% to 34% in Scenario S3 and less than 1% to 13% in 114  Scenario S2.  Moreover, the reduction of per capita EF of Design D5 would be about 1%, 5%, and 15% in Scenarios S2, S3, and S4 respectively.  Figure 6.8    Two-dimensional WEC nexus: Varying scenario results in different densities 6.4 Discussion Densification of neighbourhoods is generally preferred for sustainable communities, including sustainable water systems (EarthCraft 2014; USGBC 2013). This process will affect the WEC nexus of water distribution and residential landscaping. In this study, a planned neighbourhood with different alternative designs with varying residential densities was considered for the WEC nexus analysis. Although the variations of lot coverage, landscaping coverage, and population with net residential density are linear, the variation of per capita EF (i.e., WEC nexus) with net residential density has a power relationship. The power relationship is due to the decrease of per capita share of landscaping in MF residences coupled with the decrease of landscaping coverage requirements for high density MF buildings. The power relationships of all these parameters with density have a point or zone of inflection, which provides an optimal density.   The net residential density of around D6, i.e., 264 persons/ha or 106 units per ha (units/ha) or gross residential density of 170 persons/ha can be considered as an optimal density based on the per capita EF. It may vary from 224 to 316 persons/ha or 90 to 126 units/ha. LEED-ND and ECC 115  also recommended a residential density above 25 units/ha for a compact and sustainable neighbourhood  (EarthCraft 2014; USGBC 2013). The identified optimal density is lower than the highest densities of Canada and the US. Some of the highest densities of Canada are a gross residential density of 262 persons/ha in Blocks M4Y (Toronto), 212 persons/ha in V6E (Vancouver), and 204 persons/ha in M4X (Toronto) (Urban Toronto 2014) and that of the US are 386 persons/ha in New York, 363 persons/ha in Los Angeles and 297 persons/ha in Miami (Malouff 2013). The estimated optimal density in the present neighbourhood is lower than in other highest densities neighbourhoods, which may be due to lesser number of stories in the present neighbourhood buildings. The reduction in per capita water demand and energy use from 8.7 persons/ha (D1) to 942 persons/ha (D11) (3.5 to 377 units/ha) was 94% and 97% respectively. However, Filion (2008) found a reduction of only 10% energy use in water distribution by increasing residential density from 4 to 110 units/ha. A high reduction in the present study is mainly due to the consideration of increasing lot coverage with increasing density as per the bylaw and also a wide range (108-fold) of density considered. This fact is also supported by the results of scenario analysis, in which the change of lot coverage only in Scenario S3 (lot coverage of 70% in SF and 80% in MF building), the per capita energy use reduction of 16% could be achieved in Design D1. On the other hand, Filion (2008) considered a constant per capita water demand, which deviates highly from an actual condition.  The reduction in per capita energy related carbon emissions is equivalent to the reduction in energy use, i.e., 97%. However, per capita carbon sequestration was reduced by 100% from Design D1 to D11 as D11 lacks landscaping. The per capita negative net carbon emissions or positive net carbon sequestration was reduced by about 99% from Designs D1 to D10 and became net carbon emitter in Design D11. This study has considered only the carbon sequestration benefit of landscaping besides its other benefits such as physical and mental health, economic benefits, and biodiversity (Kabisch et al. 2015). Similarly, the per capita EF was decreased by 93% from 8.7 persons/ha (D1) to 942 persons/ha (D11) and this high reduction is attributed to water component, which dominates the ecological footprint of WEC nexus. In addition, the characteristic curve of per capita LCC of WDSs is also similar with that of the WEC nexus. This means the WEC nexus-based optimal density is supported by the LCC of 116  WDSs.  The findings show that residential density plays an important role in per capita water demand, energy use, net carbon emissions, and also LCC of WDSs. Higher the residential density, lower the per capita water demand, energy use, carbon emissions, and LCC. The reduction in per capita LCC of a WDS in dense neighbourhoods is also revealed by Speir and Stephenson (2002) although the study has mentioned that a major reduction would be in water treatment cost. The reduction of 97% of LCC from D1 to D11 is higher than that of 66% as mentioned by Duncan (1989), most probably due to the consideration of wide variation of density (108-fold increment) in this study. The lot coverage requirements imposed by municipalities affect the landscaping size as the land areas not covered by buildings need to be landscaped (District of Peachland 2014a). This ultimately affects per capita EF. The two-dimensional WEC nexus scenario analysis shows that the reduction in per capita EF is more significant in low density housing as they are composed mainly of SF residences that have higher landscaping requirements. Specifically, the per capita EF was highly reduced in Scenario S4 (xeriscaping) than S3 (lot coverage of 70% for SF and 80% for MF buildings). The xeriscaping would reduce per capita water demand, energy use, and considerable amount of carbon sequestration (~30% reduction in soil organic carbon per unit of landscape area). Xeriscaping can save a high amount of water, such as up to 54%  (Gleick et al. 2003) and 76% of irrigation demand (Sovocool et al. 2006). The estimated water saving of 51% of irrigation demand in xeriscaping in this study is comparable with Gleick et al. (2003) and Sovocool et al. (2006). The reduced water demand will also save the energy use in water distribution and upstream energy. The present study included the impacts of neighbourhood densification on water distribution and residential landscaping system only in terms of water, energy, carbon emissions, carbon sequestration, and water distribution cost. Densification may also affect other neighbourhood elements, such as transportation, open space, etc., which are not considered in this research. Furthermore, this study was conducted in a medium-sized neighbourhood of approximately 51 ha with about 5,000 population. The developed characteristic curves may be site and size 117  dependent, but the methodology is well applicable. The study can be extended by increasing the dimensions of the system, i.e., scale of economies in various neighbourhood configurations. 6.5 Summary Neighbourhood densification is a strategy primarily applied to reduce per capita infrastructure and land requirement. In particular, densification alters residential landscaping that in turn affects water distribution systems. An integrated study of the water-energy-carbon (WEC) dynamics of water distribution and residential landscaping under neighbourhood densification is lacking in the published literature. A conceptual framework was developed and applied as a case study to a planned neighbourhood in the Okanagan Valley (BC, Canada). For this neighbourhood, 11 alternative designs with varying combinations of single-family and multi-family lots representing different residential densities were investigated. Water consumption, energy use, and net carbon emissions by water distribution and residential landscaping system were combined and represented by ecological footprint. The results show that per capita ecological footprint has a power relationship with net residential density despite of a linear relationship between population and net residential density. The power relationship reveals a high dependency of per capita ecological footprint on residential density, which helps to identify an optimal density. Two-dimensional analysis of the WEC nexus scenarios indicates that xeriscaping can reduce per capita ecological footprint ranging from roughly 1% reduction in high density to 66% in low density neighbourhood. Also, the effects of xeriscaping on the WEC nexus are highly density dependent. The results emphasize the importance of amending relevant policies for constructing medium to high-density buildings in urban neighbourhoods to achieve an optimal WEC nexus.   118   Development of Microbial Water Quality Guidelines for Reclaimed Water A version of this chapter has been published in the Science of the Total Environment journal with a title “ Probabilistic risk-based investigation on microbial quality of reclaimed water for urban reuses” (Chhipi-Shrestha et al. 2017d). 7.1 Background Canada has abundant freshwater supplies and is one of the water richest countries in the world based on per capita water availability (Asano et al. 2007; WRI 2001). However, there is a large regional disparity in water availability. The annual precipitation of Canada is approximately 600 mm, ranging from 100 mm in the high Arctic to over 3500 mm along the Pacific Coast. Many agricultural lands in the Prairies and British Columbia (BC) interior receive an average annual precipitation of 300 to 500 mm (Schaefer et al. 2004). In 1994-1999, about 26% of municipalities with water supply systems experienced water shortage due to droughts, deteriorating infrastructure, and increased consumption (Environment Canada 2004). In addition, the water and wastewater infrastructure conditions are anticipated to decline in the future due to inadequate reinvestment (Canadian Infrastructure Report Card 2016). Reclaimed water use is an option to increase water supply. Public perception plays an important role in reclaimed water use. A Canada-wide survey on the public perception on reclaimed water use was conducted by Dupont (2013). The survey results show that at least 80% or more of people are willing to use reclaimed water for toilet flushing and irrigating garden grass and flowers, public parks, and golf courses. In addition, for the irrigation of agricultural crops and garden vegetables respectively 75% and 64% of people are willing to use reclaimed water. Moreover, they are willing to pay an additional annual amount of $142 to $155 per household for using reclaimed water to avoid water restrictions. The willingness to pay is approximately an additional 33% to their annual water bills. The results are consistent with another study on public attitudes on reclaimed water use in several cities in the Lake Simcoe Region in Ontario (LSRCA 119  2010). This public willingness shows that water reuse has a large potential in Canada in non-potable urban purposes.  At the federal level, Canada has the national plumbing code with an installation guide: Design and installation of non-potable water systems/maintenance and field testing of non-potable Water Systems (Canadian Standards Association 2011) and a treatment guide: Performance of non-potable water reuse systems (Canadian Standards Association 2012). The treatment guidelines have been prescribed for very small water use systems with a capacity of 10,000 L/d or less and does not cover custom-engineered systems (AEDA 2013; Canadian Standards Association 2012). In addition, Canada has reclaimed water quality guidelines at the federal level: Canadian guidelines for domestic reclaimed water for use in toilet and urinal flushing (Health Canada 2010). The federal guidelines are prescribed only for toilet and urinal flushing. The federal government has a long-term goal to develop reclaimed water use guidelines for many beneficial purposes besides toilet and urinal flushing (Health Canada 2010). At the provincial level, BC promulgated Municipal Wastewater Regulation (MWR) in 2012, which is a holistic legislation for reclaimed water applications in non-potable and potable uses. The regulation has proposed guidelines for broad water reuse classes (MWR 2012): a) Indirect potable reuse, b) High exposure potential (e.g., agricultural and lawn irrigation, toilet flushing, etc.), c) Moderate exposure potential (e.g., commercially processed agricultural crop irrigation, pasture, nurseries, etc.), and d) Low exposure potential (e.g., industrial process water, dust control, concrete production, etc.). The provincial approach is different from the federal approach that has prescribed guidelines for specific water reuse applications, e.g., toilet and urinal flushing. Moreover, the trend of developing risk-based guidelines on reclaimed water use has increased in several countries, such as Australia (EPHC/NHMRC/NRMMC 2006, 2008), the US (US EPA 2012a), Canada (Health Canada 2010) including the province of Alberta (Canada) (WaterSMART Solutions 2015), and in  the WHO (WHO 2006a). Based on the long-term goal of the Canadian federal government, the recommendations of WHO, existing urban water shortage, and public willingness for water reuse, further research is required for investigating and developing reclaimed water use guidelines for specific reuses in non-potable purposes besides toilet and urinal flushing. Furthermore, a probabilistic approach 120  can be applied to analyze uncertainty in risk estimate. This chapter investigates the risk-based guideline values for microbial quality of reclaimed water in non-potable urban reuses with a case study in the Okanagan Valley, BC. 7.2 Methodology The microbial water quality of reclaimed water for urban reuses was investigated and guideline values were proposed by using the research framework given in Figure 7.1. The framework involve quantitative microbial risk assessment (QMRA) and the application of the guideline values as a case study.  Reclaimed Water Quality Guidelines DevelopmentNoMicroorganisms in raw wastewaterWastewater treatmentMicrobial concentrationDose-response assessmentAllowable Maximum concentrationAcceptable riskReclaimed water quality guideline valuesHazard identificationQuantitative Microbial Risk Assessment (QMRA)Urban water reuse(s)Exposure to pathogensExposure assessmentIdentification of pathogenic microorganismsMeet guideline valuesSafe for water reuseYesGuidelines Application (Case Study)Lab analysis of effluentChange treatment levels Figure 7.1    Research framework for developing and applying microbial water quality guidelines for reclaimed water   121  7.2.1 Quantitative microbial risk assessment The health risk of reclaimed water due to pathogenic microorganisms was estimated by using QMRA (Haas et al. 2014). The QMRA includes four steps - hazard identification, exposure assessment, dose-response assessment, and risk characterization (Haas et al., 1999; Haas, 2002; Environment Canada/Health Canada, 2013). The propagation of variability and uncertainty in risk estimation was represented by using the two-dimensional Monte Carlo technique (Lim et al., 2015; Pouillot et al., 2016; US EPA, 2001). The QMRA steps are elaborated as follows: 7.2.2 Hazard identification The major groups of wastewater pathogens are bacteria, viruses, and protozoans (Haas et al. 2014). In these groups, the pathogenic microorganisms were selected for risk assessment based on the indicator organism, adequacy of literature on the organism, the occurrence of water borne illness, and diseases as reported by health authorities (Katukiza et al. 2014). The selected pathogens were Escherichia coli O157:H7, Salmonella spp., and Campylobacter jejuni from bacteria; rotavirus, adenovirus, and norovirus from viruses, and Cryptosporidium parvum and Giardia spp. from protozoa. These microorganisms were the most relevant pathogens for risk assessment. All the selected pathogens cause gastrointestinal illness (US EPA 2010; WHO 2006b). Among these microorganisms, E. coli is the best available indicator because it does not usually multiply in the environment, is easily detectable even in high dilution due to its excretion in the faeces in large numbers (approximately 109 cells per gram), and has a life span on the same order of magnitude as those of other enteric bacterial pathogens (Health Canada 2013a). The indicator E. coli was used for the development of microbial water quality guideline values (Health Canada 2010, 2013a), whereas all the microorganisms were used for risk assessment in a case study. 7.2.3 Exposure assessment Reclaimed water can be used for various urban purposes based on its quality and intended applications. The unit exposure volume of water in a reuse application and its annual application frequency are given in Table 7.1. Due to the lack of distribution data, a uniform distribution was assumed for most of the exposures similar to that of Mok et al. (2014) and Verbyla et al. (2016) with ±10% variation except for which a data range is available: “golf frequency” and “duration 122  from crop irrigation to consumption”, and a triangular distribution was assumed for “log reduction in natural die off”. Table 7.1    Exposure factors for different urban water uses Reuse Applications Exposure Volume (mL) Frequency/year Source Garden irrigation  Inhalation (aerosol) Unif. (0.09, 0.11) Unif.  (81, 99) 1 Garden irrigation Ingestion (Plant contact) Unif.  (0.9, 1.1) Unif.  (81, 99) 1 Garden irrigation Ingestion (accidental) Unif. (90, 110) Unif.  (0.9, 1.1 1 Public parks Ingestion (Plant contact) Unif. (0.9, 1.1) Unif. (45, 55) 1 Golf courses Ingestion (Plant contact) Unif. (0.9, 1.1) Unif. (26, 40) 1 & 2 Food crop lettuce Ingestion Unif. (4.5, 5.5) Unif. (63, 77) 1 Other raw produce (commercial) Ingestion Unif. (0.9, 1.1) Unif. (126, 154) 1 Fruits consumption (commercial) Ingestion Unif. (1.8, 2.2) Unif. (243, 297) 1 Duration (field to consumption) - Unif. (1, 5) days for fish; Unif. (0.9, 1.1) day for lettuce, other raw produce & fruits 1 Fisheries Ingestion Unif.  (2.7, 3.3) Unif. (22.5, 27.5) 1 Toilet flushing Inhalation (aerosol) Unif. (0.009, 0.011) Unif. (990, 1210) 1 Car washing Ingestion, inhalation Unif. (18, 22) Unif. (45, 55) 1 Laundry machine use Inhalation (aerosol) Unif. (0.009, 0.011) Unif. (90, 110) 1 Cross-connection by dual reticulation systems Ingestion Unif. (900, 1100) Unif. (0.00028, 0.00034)*365 1 Fire fighting Ingestion, inhalation Unif. (18, 22) Unif. (45, 55) 1 Log reduction in natural die off (D)* - Triang. (0.5, 0.5, 1) for cool weather 3 Log reduction in cleaning (D)* - Unif. (0.9, 1.1)  1 Log reduction in cooking (D)* (fisheries) - Unif. (4.95, 6.05)  1 Note: Unif. means a uniform distribution with minimum and maximum values in parenthesis;  Triang. means a triangular distribution with minimum, most likely and maximum values in parenthesis; *Microbial reduction rate is calculated as 10-D/day (WHO 2006b)     Source: 1: EPHC/NHMRC/NRMMC (2008) 2: NAGA (2012) 3: WHO (2006b)  Altogether 12 potential applications of reclaimed water were considered for non-potable urban purposes as follows: a. Lawn (L) irrigation b. Public park (P) irrigation c. Golf course (G) irrigation d. Agricultural (A) irrigation (raw eaten crops)* e. P, L & G irrigation f. P, L, G, & A irrigation 123  g. Toilet & urinal (T&U) flushing h. Vehicle washing i. Laundry machine j. Firefighting k. T&U flushing & laundry machine l. Non-potable urban uses (all above reuses) Note *agriculture in urban semi-urban areas For the identified reuse types, annual exposure volumes were estimated by using the unit exposure and annual frequency (Table 7.1) employing Monte Carlo simulations. 7.2.4 Dose-response assessment Dose-response models were used to estimate the probability of infection, which depend on incubation period (Haas et al. 1999; Katukiza et al. 2014). These models are specific to a microbial species. A Beta-Poisson model was used for E. coli (Health Canada 2010), Campylobacter jejuni (Haas et al. 1999; Katukiza et al. 2014), Salmonella spp. (Haas et al. 1999), and rotavirus (Health Canada 2010; Prez et al. 2015) with species-specific parameter values. Similarly, an exponential model was used for adenovirus (Katukiza et al. 2014; Lim et al. 2015; Vergara et al. 2016), norovirus (Messner et al. 2014; Schmidt 2015), and Cryptosporidium parvum and Giardia spp. (Robertson et al. 2005) with species-specific parameter values. The models and the annual risk estimation equation are as follows (Katukiza et al. 2014): a) Beta-Poisson dose-response model  𝑃𝑖𝑛𝑓(𝑑) = 1 − [1 + (𝑑𝑁50) (21𝛼 − 1)]−𝛼       Equation 7.1  b) Exponential dose-response model  𝑃𝑖𝑛𝑓(𝑑) = 1 − exp(−𝑟𝑑)                Equation 7.2   124  c) Annual risk of infection    𝑃𝑖𝑛𝑓(𝐴)(𝑑) = 1 − [1 − 𝑃𝑖𝑛𝑓(𝑑)]𝑛        Equation 7.3 where Pinf (d) refers to the probability or risk of infection to an individual exposed to a single pathogen dose “d”; d is the pathogen dose;  Pinf(A) (d) is estimated annual probability or risk of an infection from “n” exposures per year due to a single pathogen dose “d”; “α” and “r” are parameters referring to pathogen infectivity constant which characterize dose-response relationships; N50 is the median infective dose, i.e., the dose required to infect 50% of the exposed population. The parameter values of the dose-response models of different pathogens are given in Table 7.2. Table 7.2    Parameter values of dose-response models Pathogens Beta-Poisson Exponential Source α N50 r E. coli O157:H7 0.2019 1120 - Health Canada (2010) Campylobacter jejuni 0.145 8.96E+02 - Haas et al. (1999); Katukiza et al. (2014) Salmonella spp. 0.3126 2.36E+04 - Haas et al. (1999) Adenovirus - - 0.4172 Katukiza et al. (2014); Lim et al. (2015); Vergara et al. (2016) Norovirus - - 0.722 Messner et al. (2014); Schmidt (2015) Rotavirus 0.27 5.6 - Health Canada (2010) Cryptosporidium parvum  - - 0.004 Robertson et al. (2005) Giardia spp. - - 0.0199 Robertson et al. (2005)  7.2.5 Risk characterization Risk characterization was carried out by integrating hazard identification, exposure assessment, and dose-response assessment. Risk characterization results in the determination of a health outcome, such as the risk of infection, illness, and mortality. The final risk was expressed in disease burden, i.e., Disability-Adjusted Life-Years (DALYs) per year. The DALY is a common 125  term to represent health impacts by death and unhealthy life periods. DALY was calculated by using Equations 7.4 and 7.5 (Howard et al. 2006). 𝑅𝑖𝑠𝑘 𝑜𝑓 𝑑𝑖𝑠𝑒𝑎𝑠𝑒 (𝑃𝑖𝑙𝑙) =  𝑃𝑖𝑛𝑓(𝐴)(𝑑) ∗  𝑃𝑖𝑙𝑙|𝑖𝑛𝑓       Equation 7.4 𝐷𝐴𝐿𝑌 = 𝑃𝑖𝑙𝑙 ∗  𝐷𝐵𝑃𝐶 ∗ 𝑓𝑠        Equation 7.5 where Pill|inf is risk of disease given infection, i.e., morbidity; DBPC is disease burden per case (DALY/year); and fs is susceptibility fraction. The values of these parameters were obtained from literature as given in Table 7.3. Table 7.3    Morbidity, disease burden per case and susceptibility fraction Pathogens Morbidity (Pill|inf) Maximum disease burden (DALY/yr) Susceptibility fraction (fs) E Coli O157:H7 Unif. (0.2, 0.6) (US EPA 2010) Unif. (0.0495, 0.0605)* (Health Canada 2010) Unif. (0.8, 1)  (Mok et al. 2014) Campylobacter jejuni Unif. (0.1, 0.6)  (US EPA 2010) Unif. (0.002, 0.0047) (Gibney et al. 2014) Unif. (0.8, 1)  (Mok et al. 2014) Salmonella spp. Unif. (0.18, 0.22)*  (US EPA 2010) Unif. (0.0318, 0.0574)  (Gibney et al. 2014) Unif. (0.8, 1)  (Mok et al. 2014) Adenovirus Unif. (0.45, 0.55)* (Crabtree et al. 1997) Unif. (0.0481, 0.0587)*  (Health Canada 2010) Unif. (0.8, 1)  (Mok et al. 2014) Norovirus Unif. (0.3,0.8) (US EPA 2010) Unif. (0.0004, 0.0008) (Gibney et al. 2014) Unif. (0.8, 1)  (Mok et al. 2014) Rotavirus Unif. (0.61, 0.73)* (US EPA 2010) Unif. (0.0076, 0.0092)*  (Health Canada 2011) Unif. (0.05, 0.07) (Mok et al. 2014) (Health Canada 2010)  Cryptosporidium parvum  Unif. (0.2, 0.7) (US EPA 2010) (Health Canada 2010) Unif. (0.0011, 0.0028)  (Gibney et al. 2014) Unif. (0.8, 1) (Mok et al. 2014) Giardia spp. Unif. (0.2, 0.7) (US EPA 2010) Unif. (0.0015, 0.003)  (Gibney et al. 2014) Unif. (0.8, 1) (Mok et al. 2014) Note: Unif. means a uniform distribution with minimum and maximum values in parenthesis;  * considered ± 10% variation to estimate a range  For the development of guideline values in this research, reverse QMRA was applied. The target risk was considered to be 10-6 DALYs/year (WHO 2006b) and Equation 7.6 was used to estimate the equivalent concentration of E. coli. 𝐷𝑜𝑠𝑒 𝑒𝑞𝑢𝑖𝑣𝑎𝑙𝑒𝑛𝑡 =𝑇𝑎𝑟𝑔𝑒𝑡 𝑟𝑖𝑠𝑘𝐷𝑖𝑠𝑒𝑎𝑠𝑒 𝑏𝑢𝑟𝑑𝑒𝑛 𝑝𝑒𝑟 𝑐𝑎𝑠𝑒 ∗ 𝑃𝑖𝑛𝑓 ∗ 𝑀𝑜𝑟𝑏𝑖𝑑𝑖𝑡𝑦 ∗ 𝑆𝑢𝑠𝑐𝑒𝑝𝑡𝑖𝑏𝑖𝑖𝑡𝑦 𝑓𝑟𝑎𝑐𝑡𝑖𝑜𝑛   Equation 7.6   126  7.2.6 Data variability and uncertainty Quantitative risk assessment should reflect the variability in the risk and take into account the uncertainty associated with the risk estimate (Pouillot et al. 2016; US EPA 2001). The variability in QMRA, also called aleatoric uncertainty, represents the temporal and individual heterogeneity of the risk for a given population. The uncertainty in QMRA, also called epistemic uncertainty, stems from imperfect knowledge about the QMRA model structure and the associated parameters (Pouillot et al. 2016). A two-dimensional (or second-order) Monte Carlo Analysis (2-D MCA) was used to characterize variability and uncertainty in input variables. A 2-D MCA is a Monte Carlo analysis where the distributions reflecting variability and the distributions representing uncertainty are sampled separately in the simulation so that variability and uncertainty in the output may be assessed separately (Pouillot et al. 2016). In this analysis, the input parameters considered were: “exposure factors” as variable parameters, “pathogenic E. coli ratio” as an uncertain parameter, and “morbidity”, “disease burden per case”, and “susceptibility fraction” as variable and uncertain parameters. The 2-D MCA simulations with 10,000 iterations for the inner loop (variability) and 5000 iterations for the outer loop (uncertainty) were performed to make the risk estimates reliable (Ashbolt et al. 2010; Katukiza et al. 2014; Pavione et al. 2013; US EPA 2001) by using the R software (Pouillot et al. 2016). The 2-D MCA produced cumulative density functions (CDFs) of microbial concentration at different quantiles (e.g., 5%, 25%, median, 75%, and 95%). The measures of variability and uncertainty were estimated by using three ratios as shown in Equations 7.7 to 7.9 (Ozkaynak et al. 2009; Pouillot et al. 2016).  𝑉𝑎𝑟𝑖𝑎𝑏𝑖𝑙𝑖𝑡𝑦 𝑟𝑎𝑡𝑖𝑜 =  𝐵𝐴           Equation 7.7 𝑈𝑛𝑐𝑒𝑟𝑡𝑎𝑖𝑛𝑡𝑦 𝑟𝑎𝑡𝑖𝑜 =  𝐶𝐴           Equation 7.8 𝑂𝑣𝑒𝑟𝑎𝑙𝑙 𝑢𝑛𝑐𝑒𝑟𝑡𝑎𝑖𝑛𝑡𝑦 𝑟𝑎𝑡𝑖𝑜 =  𝐷𝐴          Equation 7.9 where A is the median of uncertainty (in 50% CDF) for the median of variability, B is the median of uncertainty (in 50% CDF) for the 97.5th percentile of variability, C is the 97.5th 127  percentile of uncertainty (in 97.5% CDF) for the median percentile of variability, and D is the 97.5th percentile of uncertainty (in 97.5% CDF) for the 97.5th percentile of variability. 7.3 Results 7.3.1 Microbial water quality investigation and guideline values The microbial concentrations of reclaimed water for different reuses were estimated and their CDFs, for different intended uses, are shown in Figure 7.2. The figure shows a wide variation in uncertainty (along the x-axis) and variability (along the y-axis) in the concentration estimate. However, the variability, uncertainty, and overall uncertainty ratios in all water reuse types are similar as also seen from the identical shape of CDFs in Figure 7.2. In these water reuse types, variability ratios vary from 1.96 to 2.06, uncertainty ratios range from 2.31 to 2.34, and overall uncertainty ratios vary from 4.54 to 4.81. Moreover, in this study, the 95th percentile is considered as the Reasonable Maximum Estimate (RME) as used by the US EPA (2001). For example, in lawn irrigation as shown in Figure 7.2 (Plot 1), the median of 0.06 cfu/100 mL at the 50th percentile CDF (i.e., horizontal arrow) could range from 0.04 cfu/100 mL (5th percentile CDF) to 0.13 cfu/100 mL (95th percentile CDF) due to uncertainty. Similarly, the median of 0.06 cfu/100 mL (i.e., vertical a