Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Evaluating source water protection strategies : a soft computing approach Islam, Nilufar 2010

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

Item Metadata


24-ubc_2011_spring_islam_nilufar.pdf [ 4.35MB ]
JSON: 24-1.0063054.json
JSON-LD: 24-1.0063054-ld.json
RDF/XML (Pretty): 24-1.0063054-rdf.xml
RDF/JSON: 24-1.0063054-rdf.json
Turtle: 24-1.0063054-turtle.txt
N-Triples: 24-1.0063054-rdf-ntriples.txt
Original Record: 24-1.0063054-source.json
Full Text

Full Text

EVALUATING SOURCE WATER PROTECTION STRATEGIES: A SOFT COMPUTING APPROACH  by  Nilufar Islam B.Sc., Bangladesh University of Engineering and Technology (BUET), 2007  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in The College of Graduate Studies (Civil Engineering)  THE UNIVERSITY OF BRITISH COLUMBIA (Okanagan) December 2010  © Nilufar Islam, 2010  ABSTRACT Source water protection is an important step in the implementation of a multi-barrier approach that ensures delivery of safe drinking water cost effectively. However, implementing source water protection strategies can be a challenging task due to technical and administrative issues. Currently many decision support tools are available that mainly use complex mathematical formulations. These tools require large data sets to conduct the analysis, which make their use very limited. A simple soft-computing model is proposed in this research that can estimate and predict a reduction in the pollutant loads based on selected source water protection strategies that include storm water management ponds, vegetated filter strips, and pollution control by agricultural practice. The proposed model uses an export coefficient approach and number of animals to calculate the pollutant loads generated from different land uses (e.g., agricultural lands, forests, roads, livestock, and pasture). A surrogate measure, water quality index, is used for the water assessment after the pollutant loads are discharged into the source water. To demonstrate the proof of concept of the proposed model, a Page Creek Case Study in Clayburn Watershed (British Columbia, Canada) was conducted. The results show that rapid urban development and improperly managed agricultural area have the most adverse effects on the source water quality. On the other hand, forests were found to be the best land use around the source water that ensures acceptable drinking water quality with a minimal requirement for treatment. The proposed model can help decisionmakers at different levels of government (Federal/ Provincial/ Municipal) to make informed decisions related to land use, resource allocation and capital investment.  ii  PREFACE A version of Chapters 2 and 3 in this thesis has been accepted (November 2010) in Environmental Reviews as a journal article titled “Reviewing Source Water Protection Strategies: A Conceptual Model for Water Quality Assessment (Islam et al., 2010)”. The paper was written by Nilufar Islam under the supervision of Dr. Rehan Sadiq and Dr. Manuel Rodriguez. A journal article based on Page Creek Case Study (Chapter 4) is under preparation. One conference paper (CSCE 2010) is published and one (BCWWA 2011) is under progress from this research.  iii  TABLE OF CONTENTS ABSTRACT ............................................................................................................................ II PREFACE .............................................................................................................................. III TABLE OF CONTENTS ..................................................................................................... IV LIST OF TABLES .............................................................................................................. VII LIST OF FIGURES ........................................................................................................... VIII ABBREVIATIONS ................................................................................................................ X ACKNOWLEDGEMENTS .............................................................................................. XIII 1.  INTRODUCTION........................................................................................................... 1  1.1 Background ....................................................................................................................... 1 1.2 Research Objectives .......................................................................................................... 4 1.3 Thesis Organization .......................................................................................................... 5 2.  LITERATURE REVIEW .............................................................................................. 6  2.1 Pollution Sources ............................................................................................................. 6 2.2 SWP Regulations ............................................................................................................... 9 2.2.1 Canadian regulations .................................................................................................. 10 2.2.2 USA............................................................................................................................ 12 2.2.3 European Union ......................................................................................................... 13 2.2.4 Australia ..................................................................................................................... 13 2.3 Source Water Protection (SWP) Strategies .................................................................. 14 2.3.1 Storm water management .......................................................................................... 14 2.3.2 Other SWP strategies ................................................................................................. 25 2.4 Watershed Models .......................................................................................................... 28 2.5 Water Quality Assessment ............................................................................................. 29 2.5.1 Selection of water quality parameters (WQP) ........................................................... 29 2.5.2 Transformation into sub-indices ................................................................................ 35 2.5.3 Aggregation formulations .......................................................................................... 36 2.6 Soft Computing Methods ............................................................................................... 39 2.6.1 Fuzzy rule-based model (FRBM) .............................................................................. 44 2.6.2 Fuzzy measures theorem (FMT) ................................................................................ 49 3.  MODEL DEVELOPMENT ......................................................................................... 53 iv  3.1 Conceptual Framework ................................................................................................ 53 3.1.1 Reduced pollutant load calculations .......................................................................... 53 3.1.2 Estimation of pollutant concentration in the source .................................................. 53 3.1.3 Estimation of water quality index (WQI) .................................................................. 55 3.2 Pollutant Loads based on Land Use ............................................................................ 55 3.3 Reduced Pollutant Loads ............................................................................................. 59 3.4 Pollutant Mass Balance ................................................................................................ 63 3.5 Illustrative Example...................................................................................................... 67 4.  PAGE CREEK CASE STUDY .................................................................................... 76  4.1 Study Area ....................................................................................................................... 76 4.2 Data Preparation ............................................................................................................. 79 4.2.1 Pollutant load calculations ......................................................................................... 79 4.2.2 SWP strategies ........................................................................................................... 82 4.3 Results and Discussions .................................................................................................. 86 4.3.1 Predicted WQI–monthly variation ............................................................................. 87 4.3.2 Predicted WQI–a yearly variation ............................................................................. 97 4.3.3 Sensitivity analysis ..................................................................................................... 97 4.3.4 Comparison of proposed WQI with NSF-WQI ....................................................... 103 5.  CONCLUSIONS AND RECOMMENDATIONS .................................................... 105  5.1 Summary and Conclusions........................................................................................... 105 5.2 Limitations and Recommendations ............................................................................. 106 REFERENCES .................................................................................................................... 109 APPENDICES ..................................................................................................................... 136 Appendix A: Available Canadian regulations regarding SWP strategies in three provinces (Ontario, Alberta, and British Columbia) ............................................................................. 136 Appendix B: Source water protection strategies under USEPA bulletin............................. 139 Appendix C: Main factors in storm water management ...................................................... 141 Appendix D: Common formulations for water quality indices (WQI) ................................ 143 Appendix E: Pollutant load mass generation from manure (ASAE, 2003) ......................... 146 Appendix F: Membership functions and inference rules for TSS removal using SWM ..... 147 Appendix G: Membership functions and inference rules for TSS removal using VFS ...... 150 v  Appendix H: Mass balance equations for water quality parameters ................................... 152 Appendix I: Predicted value of the parameters for models in Appendix H (USEPA, 1985) ............................................................................................................................................... 153 Appendix J: Seasonal creek water flow rate prediction with manning‟s formulation ........ 155 Appendix K: Description of box-and-whisker plot ............................................................. 156 Appendix L: Sub index value for monitored data, predicted initial and final creek water quality parameter .................................................................................................................. 157  vi  LIST OF TABLES Table 2.1: Export coefficient (kg/ha/yr) for different land uses .................................................... 8 Table 2.2: Typical concentrations (mg/l) in runoffs from different land uses ............................... 9 Table 2.3: Basic features of common storm water management strategies (NCDENP, 1999) ... 20 Table 2.4: Percent removal of water quality parameters through storm water management ....... 22 Table 2.5: Basic design parameters for storm water management strategies (NCDENP, 1999) . 24 Table 2.6: Effect of other SWP strategies in pollutant removal .................................................. 27 Table 2.7: Comparison of various watershed models .................................................................. 30 Table 2.8: Water quality standards and guidelines recommended by various agencies .............. 33 Table 2.9: Common issues in formulations for water quality indices (Sadiq et al., 2010) .......... 40 Table 2.10: A comparison of soft computing methods for water quality assessment ................. 43 Table 2.11: Advantages and disadvantages of fuzzy-based methods .......................................... 43 Table 3.1: Typical EC (kg/ha/yr) from USEPA (2001i) .............................................................. 58 Table 3.2: PCAP rule-set for TSS removal .................................................................................. 64 Table 3.3: Required inputs to perform analysis using developed model ..................................... 70 Table 3.4: Step-by-step results of component 1 of the conceptual framework............................ 71 Table 3.5: Results for component 2 and 3 (an illustrative example) ........................................... 74 Table 4.1: Land use area distribution for six segments (m2) ....................................................... 80 Table 4.2: The distribution for six land use types (ha) ................................................................ 80 Table 4.3: Pasture/forage livestock numbers in the study area for year 2007 ............................. 82 Table 4.4: Summary data for pollution control by agricultural practice (PCAP) ........................ 83 Table 4.5: Future storm water management pond options for six segments................................ 84 Table 4.6: Inputs for VFS around the Page Creek-current and future recommendation ............. 86 Table 4.7: Percentage of WQI improvement for five WQI group and for overall WQI (FYs 2008-18) ...................................................................................................................... 98 Table 4.8: Percent parameter contribution for SWM pond, VFS and PCAP efficiency............ 102  vii  LIST OF FIGURES Figure 2.1: Common source water hazards and SWP management ............................................ 10 Figure 2.2: Storm water management ponds (NCDENP, 1999)  .............................................. 17  Figure 2.3: Example of a buffer strip (USEPA, 2001e)............................................................... 26 Figure 2.4: Typical sub-index transformations for water quality parameters.............................. 35 Figure 2.5: Sub-index functions for selected water quality parameters (continued next page) ... 37 Figure 2.6: Source water protection strategies and water quality assessment ............................. 41 Figure 2.7: Two causal concepts connecting to an effect concept ............................................... 47 Figure 2.8: Fuzzy rule-based model-making inference using two „causal factors‟ ..................... 48 Figure 2.9: Fuzzy measures theorem-making inference using two „causal factors‟ .................... 52 Figure 3.1: Proposed model based on conceptual framework ..................................................... 54 Figure 3.2: Typical three zones in a watershed or a sub-watershed ............................................ 60 Figure 3.3: Predicting pollutant reduction using different land uses (step 1-2 in Figure 3.1) ..... 61 Figure 3.4: Decision diagram for storm water management model ............................................ 62 Figure 3.5: Membership functions for input parameters and removal efficiency of TSS in PCAP........................................................................................................................... 64 Figure 3.6: Typical mass balance for a flowing system .............................................................. 67 Figure 3.7: Typical mass balance for a lake system .................................................................... 67 Figure 3.8: Scenarios for illustrative example  a) baseline b) proposed .................................. 68  Figure 3.9: Results of a PCAP fuzzy rule-based model for the removal of TSS......................... 72 Figure 3.10: Illustrative example results-a snapshot ................................................................... 75 Figure 4.1: Average monthly total precipitation (mm) in Abbotsford ........................................ 76 Figure 4.2: Page Creek sub-watershed area in Clayburn watershed ............................................ 77 Figure 4.3: Monthly flow rate prediction in Page Creek ............................................................. 79 Figure 4.4: Six segments in the Page Creek effective area .......................................................... 81 Figure 4.5: Soil type distribution in the study area ...................................................................... 83 Figure 4.6: Storm water collection in Clayburn watershed ......................................................... 85 Figure 4.7: NPS1 loading change for the seven pollutants a) without PCAP & fencing and b) with PCAP and fencing ............................................................................................... 88 viii  Figure 4.8: NPS2 loading change a) without storm water management pond and b) with storm water management pond ................................................................................... 89 Figure 4.9: Final pollutant loading change for the seven parameters a) without vegetated filter strip and b) with vegetated filter strip ................................................................ 90 Figure 4.10: Creek water quality prediction for water quality parameters a) with SWP strategies and b) without SWP strategies .................................................................... 91 Figure 4.11: Seasonal primary-WQI ............................................................................................ 93 Figure 4.12: Seasonal nutrient-WQI ............................................................................................ 94 Figure 4.13: Seasonal microbial-WQI ......................................................................................... 94 Figure 4.14: Seasonal organic-WQI ............................................................................................ 95 Figure 4.15: Seasonal toxic substance-WQI ................................................................................ 95 Figure 4.16: Seasonal overall WQI.............................................................................................. 96 Figure 4.17: Primary WQI prediction (FYs 2008-18) ................................................................. 97 Figure 4.18: Impact of land use on WQI without SWP strategies ............................................... 99 Figure 4.19: Impact of land use on WQI with SWP strategies .................................................. 100 Figure 4.20: Impact of livestock land use on water quality for initial condition ....................... 100 Figure 4.21: Impact of livestock land use on water quality with proposed SWP strategies ...... 101 Figure 4.22: Impact of pasture land use on water quality for initial condition .......................... 101 Figure 4.23: Impact of pasture land use on water quality with proposed SWP strategies ......... 102 Figure 4.24: Tornado graphs describing the contribution of various design parameters in the removal efficiency of selected SWP strategies a) PCAP, b) SWM pond, and c) VFS ........................................................................................................................... 103 Figure 4.25: Monthly variations predicted by proposed WQI and NSF-WQI .......................... 104 Figure 4.26: NSF-WQI vs. proposed WQI ................................................................................ 104  ix  ABBREVIATIONS ANN  Artificial neural networks  B. Fen  Barbed fencing  BMPs  Best management practices  BOD5  5 day biochemical oxygen demand  CC  Cover crop  CCME-WQI Canadian Council of Ministers of the Environment Water Quality Index CE  Complex equation  CR  Crop rotation  CSO  Combined sewer overflow  CSS  Combined sewer systems  CSTR  Continuous stirred-tank reactor  DBD  Dry bulk density  DO  Dissolved oxygen  E. Coli  Escherichia coli  EC  Export coefficient  EMC  Event mean concentration  ER  Evidential reasoning  FC  Faecal coliforms  FCA  Fuzzy clustering analysis  Fen  Fencing  FIS  Fuzzy inference system  FMT  Fuzzy measures theorem  FRBM  Fuzzy rule-based model  FSE  Fuzzy synthetic evaluation  GIS  Geographical Information System  GPS  Global Positioning System  HACCP  Hazard Analysis Critical Control Points  LID  Low impact development  MIMO  Multiple-inputs-multiple-outputs  MISO  Multiple-inputs-single-output x  MUSLE  Modified universal soil loss equation  NH3-N  Ammonia nitrogen  NH4-N  Ammonium nitrogen  NO2-N  Nitrite-nitrogen  NO3-N  Nitrate-nitrogen  NPS  Nonpoint pollution sources  NRCS  Natural Ressources Conservation Service  NSF-WQI  National Sanitation Foundation water quality index  OWA  Ordered weighted averaging  O-WQI  Oregon water quality index  Pb  Lead  PCAP  Pollution control by agricultural practice  PO4  Phosphates  PPS  Point pollution sources  RT  Retention time  S  Summer  SISO  Single-input-single-output  SRC  Small and rural communities  SWC  Soil water content  SWM  Storm water management  SWP  Source water protection  TC  Total coliforms  TDS  Total dissolved solids  Temp  Temperature  TKN  Total Kjeldahl nitrogen  TN  Total nitrogen  TOC  Total organic carbon  TP  Total phosphorous  TS  Toxic substance  TSK  Takagi-Sugeno-Kang  TSS  Total suspended solids xi  USLE  Universal soil loss equation  V. Fen  Vegetated fencing  VFS  Vegetated filter strips  W  Winter  WQI  Water quality index  WQP  Water quality parameter  WWT  waste water treatment  xii  ACKNOWLEDGEMENTS I offer my enduring gratefulness to the staff and students of School of Engineering at UBC (Okanagan). First, I would like to thank Dr. Rehan Sadiq for his supervision and support. Dr. Manuel J. Rodriguez (Laval University) was an integral part of this research and I would like to pay my heartiest gratitude to him for his advice during the development of model. I like to thank Dr. Alex Francisque for his valuable feedback, and Drs. Cigdem Eskicioglu and Deborah J. Roberts for their positive criticism which helped to improve the quality of this thesis. I would also like to thank Rod Shred and Darren J. Brown from The City of Abbotsford for providing valuable information to perform the Page Creek Case Study. Finally, special thanks to my parents, sisters and husband for their continuous emotional and moral support during this study.  xiii  Dedicated to my parents  xiv  1. INTRODUCTION  1.1 Background The United Nations General Assembly asserted that safe and clean drinking water and sanitation is a fundamental human right. The grand challenges of engineering for the 21st century included: „providing access to the clean water‟ among the top priorities. The protection of water supplies is a priority and an integral part of Canada‟s Science & Technology Strategy as well. Advancement in technology and knowledge has improved accessibility to safe drinking water, however increasing urbanization, industrialization and emerging environmental issues are making water protection more challenging in the future. Information on the quality of source waters and associated risks to the human health is imperative to make informed decisions in prioritizing available resources. Sources of water for large cities like Toronto, Ottawa and Vancouver are comparatively well managed and adequate treatment is generally available to ensure „reduced risk‟ to the consumers‟ health. However for small and rural communities (SRC), achieving the same level of reduced risk under limited information and budgeting constraints is taxing. The goal of source water protection (SWP) is to protect against potential pollution or resuscitate contaminated water if economically viable. A watershed may consist of diverse systems such as terrestrial, freshwater and ecosystems and that makes SWP a challenging task. Though source water generally refers to both ground and surface waters, the main focus of this research is on the later. Strategies for SWP refer to watershed-based protection of the water at the source. A watershed is a catchment area that drains into a common water source point, e.g., Lake Winnipeg is fed by a watershed that consists of four provinces in Canada and two U.S. states (Postel and Thompson, 2005). A modern water supply system ensures the delivery of high quality water to homes, industries, agricultural lands, essential facilities (such as hospitals, schools) and recreational facilities. Strategies for SWP organize activities in the watershed to minimize undesirable effects that result in a poor water quality. Generally, it is very difficult to restore or substitute the polluted source waters and if the source water quality is compromised, water treatment often becomes prohibitively expensive (Wilsenach et al., 2003). In some cases, the high variability in 1  the quality of source water (e.g., due to seasonal variability or land use change) increases the challenges for water treatment at affordable costs (Curriero et al., 2001). Protecting water at the source is always a preferred option (preventive action) as compared to subsequent expensive water treatment technologies (mainly corrective actions). Small and rural communities (SRC) are generally incapable of providing sophisticated treatment plants due to lack of financial resources and trained staff (Timmer et al., 2007). Properly executed SWP strategies coupled with conventional water treatment can be an effective way of ensuring safe drinking water supplies not only true for SRC but also for larger municipalities. For example, Catskill/ Delaware watershed (USA) that supplies 90% of the New York drinking water, enhanced source water protection was provided. The city spent $1.5 billion for implementing enhanced SWP strategies for Catskill/ Delaware watershed which approximately saved $3.5 billion to the city by avoiding the cost of advanced water treatment (Garcia, 2004). According to WHO (1993), SWP is found to be a cost effective way to ensure safe drinking water. This is due to the fact that effective SWP can result in not only less treatment cost but also lower microbial re-growth and reduced formation of disinfection by-products (DBPs are formed when disinfectant reacts with the naturally occurring organic matter) in the distribution networks (Ivey et al., 2006; Krewski et al., 2004). The Bonn Charter of Safe Drinking Water (2004) provides a generic framework for an effective management of water quality from source to tap that can be linked to WHO guidelines for drinking water. It emphasizes the need for proactive management of drinking water supplies. The Charter is generic enough to support approaches and methodologies such as a multi-barrier approach, Hazard Analysis Critical Control Points (HACCP) framework (Hulebak and Schlosser, 2002) and total water quality management. Multibarrier approach has been recommended all over the world for the protection of drinking water supplies in various stages that includes: source water protection, water treatment, regular water quality in the distribution network and ensuring desirable free residual disinfectant. After the Walkerton enquiry, a multi-barrier approach has been recommended in Canada to ensure safe drinking water from source to tap (O‟Connor, 2002). Numerous studies have reported the benefits of SWP in terms of reduction of contaminants in receiving water bodies (Arora et al., 2003; Bliss et al., 2009; Borin et al., 2010; Dietz, 2007; Hsieh and Davis, 2005; Hathaway et al., 2009; Muthanna et al., 2007; Scholz and Grabowlecki, 2  2007; Patty et al., 1997; USEPA, 2000). The impacts and benefits of SWP can be linked to three dimensions of sustainability (a triple bottom line) social, economical and environmental. SWP is a complex and time consuming process. In most cases, various ministries at the Federal, Provincial/ Territorial and Municipal levels should coordinate to make SWP programs effective and efficient. Increasing population, rapid urbanization and industrialization increase the likelihood of water contamination through both point pollution sources (PPS) (e.g., municipal and industrial discharges) and nonpoint pollution sources (NPS) (e.g., agricultural runoff and storm water). In United States, PPS have been regulated since 1970 (Ernst, 2004), however the physical and regulatory controls of NPS is much more challenging because of the pollution load distribution (Meixler and Bain, 2010). In addition, source water quality is under increasing pressure due to the impacts of climate change and due to rapidly changing land use (IPCC, 2007). SWP has been practiced for the past 125 years in North America, e.g., the Cedar River in Seattle (Washington) was taken into protection in 1896 after a large fire outbreak which destroyed the whole ecosystem (Ernst, 2004). Around the world, various SWP strategies have been implemented. Two major categories for SWP strategies include implementing activities allowing low impact development (LID) (henceforth we refer it as storm water management) and adopting best management practices (BMPs) (management of industrial, municipal and agricultural areas). Water quality is a vague concept which depends on numbers parameters and the intended use of water. Many studies have reported diverse methods for quality assessment of surface waters (e.g., Banerjee and Srivastava, 2009; Dojlido et al., 1994; Jonnalagadda and Mhere, 2001; Lermontov et al., 2009; Rajankar et al., 2009; Ramakrishnaiah et al., 2009; Sedeno-Diaz and Lopez-Lopez, 2007; Suratman et al., 2009). These studies proposed certain water quality parameters to derive a water quality index (WQI). The water quality index (WQI) for the assessment of water quality at the source after reduction in pollutant loads has never been linked to any human effort (implementation of SWP strategies). Decision support tools can be very useful to make informed decisions related to the selection of source water and implementing SWP strategies that can improve the quality of the water at the source. However, the development of a decision support tool requires an understanding about the impacts of SWP strategies on water quality and the related regulatory regimes. A number of watershed models provide a process for the calculation of pollutant loads based on land use ( 3  Bingner et al., 2001; Bjorneberg, 1999; Borah et al., 2002; Bottcher, 1998; Chen et al., 2003; Chiew and McMahon, 1999; Downer and Ogden, 2002; EMRL, 1998; HSPF, 1985; Irvine et al., 1993; Neitsch et al., 2001; Ovbiebo and She,1995; Pitt, 1993; Roesner et al., 1988; Skaggs, 1980; Sydelko et al., 2000; Walker, 1997; Williams et al., 1983; Woolhiser et al., 1990; Young et al., 1987). These models are based on complex mass balance and empirical relationships (e.g., universal soil loss equation). These models predict the reduction in pollutant loads based on different land use, however, the impact on water quality at the source is generally not estimated. There is a pressing need for linking reduction in the pollution loads to the quantitative assessment of source water quality for informed decision-making. To calculate pollutants loads, export coefficients (land use based pollutant export to the source water) and number of animals‟ calculations (Lahlou et al., 1998; Palmstrom and Walker, 1990; USACE-HEC, 1977) provide a much simpler approach that can be easily integrated with water quality assessment calculations. The developed decision support tool based on this research addresses the limitations of existing tools and provides a framework for water quality assessment at the source. Soft computing-based formulations have been used to develop decision support tools (Chang et al., 2001; Chen et al., 2010; Dahiya et al., 2007; Francisque et al., 2009; Hajkowicz and Collins, 2007; Icaga, 2007; Kuoa et al., 2007; Lu and Lo, 2002; Liou et al., 2003; Li et al., 2009; Ocampo-Duque et al., 2006; Rankovic et al., 2010; Sadiq et al., 2006; Sadiq et al., 2007; Spinella et al., 2008; Singh et al., 2009; Yeon et al., 2008) which can handle uncertainty and vagueness in the data.  1.2 Research Objectives Overall objective of this thesis is to develop a model (decision support tool) which can predict the reduction in pollutant loads by implementing selected SWP strategies and subsequently predict the improvement in surface water quality based on reduced pollutant loads. Specific objectives of this research are to: i)  review existing regulations for source water protection in various parts of the world with a special focus on Canada and the USA, and study state-of-the-art source water protection (SWP) strategies to deal with non-point pollution sources  ii)  investigate mathematical formulations to develop a water quality index (WQI) for water quality assessment at the source 4  iii)  develop a conceptual framework that can relate reduction in pollutant loads with the improvement in WQI, and  iv)  apply the developed model to a case study and demonstrate a proof-of-concept for the proposed approach.  1.3 Thesis Organization This thesis is organized in five chapters. The first chapter provides background and objectives of this research. Chapter 2 presents a detailed literature review for the related topics that includes: (a) source water quality; (b) potential impacts of land use on water quality; (c) water quality regulations for source water; (d) source water protection strategies; and (e) existing watershed models used for estimating pollutants loads. Chapter 3 provides the steps for model development based on the proposed conceptual framework. Chapter 4 discusses a case study of a small creek in British Columbia using the developed model. Finally, Chapter 5 provides the conclusions and makes recommendations for future research.  5  2. LITERATURE REVIEW  2.1 Pollution Sources Microorganisms, nutrients, heavy metals, organic and toxic substances can be found in natural water bodies. Surface waters are generally vulnerable to microbial contamination that may lead to devastating consequences if the water is used for drinking without proper treatment. Undesirable levels of organic matter and nutrients provide a favorable environment for rapid microbial growth in surface waters (Carreiro-Silva et al., 2009). A comprehensive assessment of these pollutants and mechanisms by which they reach source water requires an implementation of SWP strategies to ensure high quality water at the source. The pollutants entering into receiving water bodies from a number of point and non-point pollution sources can adversely affect human wellbeing as well as ecosystem health. For example, the release of nutrients (e.g., phosphorus, nitrogen) from agricultural sources can lead to the growth of toxic algal blooms (Bechmann et al., 2009; Hickey and Gibbs, 2009) and can cause eutrophic conditions which impact the biological and chemical quality of water (Schoumans and Chardon, 2003). Similarly, heavy metals such as lead, mercury and arsenic may ingress into the source waters through industrial discharges, runoff and spills (Zakir et al., 2009). If the water is not properly treated before consumption, the exposure to lead and mercury may have severe consequences for the consumers (Howard, 2002). Arsenic is a proven human carcinogen and exposure to high levels may lead to skin cancer (Otles and Cagindi, 2010). Most pesticides contain a number of toxic substances which can adversely affect human health (Zhang et al., 2010). Harmful pesticides (e.g., atrazine, diazinon) can enter into source water not only from agricultural lands but also from runoffs from gardens and parks (Wittmer et al., 2010). After a heavy rainfall or snowfall event, the runoff from agricultural and urban lands can contaminate source waters. Untreated municipal discharges or improper diversion of wastewater from the municipal sources can cause contamination of surface waters. In addition, if a wastewater treatment plant is not working effectively the toxic substances can contaminate source waters. Combined sewer systems (CSS) designed to carry wastewater from industrial, domestic and commercial sources with storm water (snowmelt/ rain water) can also be a source of pollution. In case of excess water (during heavy rains), the wastewater beyond the capacity of 6  the treatment plant may be released as combined sewer overflow (CSO). Approximately 43 million people in 1,100 communities in the USA are served by a CSS (USEPA, 1995). The pollutant load depends on the land use. For example, forests, agricultural land, farms (livestock), pasture/forage, residential, industrial, commercial areas and roads generate diverse types and the amounts of pollutant loads. The export coefficient (EC) concept is widely used for deriving the potential pollutant loads based on soil erosion and runoff (Amatya et al., 2004; Beaulac and Reckhow, 1982; Bourne et al., 2002; Ding et al., 2010; Dodd et al., 1992; Easton et al., 2009; Kay et al., 2008; Line et al., 2002; Loehr et al., 1989; McFarland and Hauck, 2001; Prepas et al., 2001; Rast and Lee, 1983; Reckhow et al., 1980; Shrestha et al., 2008; Veiga and Dziedzic, 2010). It describes the amount of pollutant or a specific parameter that can be related to water quality (e.g., total suspended solids, total nitrogen, and total phosphorous). Table 2.1 provides export coefficients for different land uses expressed in kg/ha/yr (i.e., flux units), except for coliforms which are described by colony forming unit/ha/yr (CFU/ha/yr). A large variability in the EC values for water quality parameter can be noticed. For example, Prepas et al. (2001) and Reckhow et al. (1980) provided EC values for TSS that varied significantly in case of runoffs from forests. The difference in topography and precipitation is the main reason behind this variation. However, the EC approach is found to be very effective in calculating pollutant transport amount from different land uses. Table 2.2 provides average pollutant concentrations (mg/l) for different land uses. This table describes pollutant concentration (either an average or average ± standard deviation) variability for the same land use. It should be noted that the concentration depends not only on the runoff quantity but also on the type of pollutant. Therefore, the reported land use concentration or sometimes expert judgment can be used to estimate the pollutant loads. To protect source water from pollutant loads various SWP strategies can be used. Selecting a suitable SWP strategy is a key for reducing the amount of pollutants and improving water quality at the source. Next section provides an overview of guidelines and regulatory regimes that directly or indirectly impacts source water quality and help in decision-making related to the selection of SWP strategies.  7  Table 2.1: Export coefficient (kg/ha/yr) for different land uses Reference  Place  TSS  TN  TP  FC  TC  Forests  Bourne et al. (2002)  Manitoba, Canada  -  0.2-3.9  0.01-0.4  -  -  Dodd et al. (1992)  Albemarle Pamlico, USA  -  0.70-3.8  0.09-0.2  -  -  Loehr et al. (1989)  NS  -  1-6  0.007-0.9  -  -  Prepas et al. (2001)  Boreal plain, Western Canada  20.86  0.6  0.04  -  -  Reckhow et al. (1980)  NS  253  1.4-6.3  0.02-0.8  -  -  Kay et al. (2008)  UK  -  -  -  Bourne et al. (2002)  Manitoba, Canada  -  0.3-6.7  0.03-1  -  -  Dodd et al. (1992)  Albemarle Pamlico, USA  -  5-14  0.6-2  -  -  Loehr et al. (1989)  NS  -  2-80  0.06-3  -  -  Rast and Lee (1983)  North America  -  5  0.5  -  -  Reckhow et al. (1980)  NS  -  2-80  0.3-19  -  -  Reckhow et al. (1980)  NS  -  1-7.8  0.1-3  -  -  Reckhow et al. (1980)  NS  -  680~7980  21~790  -  -  Loehr et al. (1989)  NS  -  100~1600  10~620  -  -  Bourne et al. (2002)  Manitoba, Canada  -  0.17-4.3  0.02-0.5  -  -  Loehr et al. (1989)  NS  -  3-14  0.05-0.6  -  -  NS  -  5-7  0.8-2  -  -  Agricultural  Livestock  Pasture/Forage  Urban (residential) Loehr et al. (1989)  Urban (industrial and commercial) Loehr et al. (1989)  NS  -  2-14  0.4-4  -  -  Reckhow et al. (1980)  NS  870  2.3  8  -  -  Kay et al. (2008)  NS  -  -  -  a-Unit in CFU/ha/hr ; NS - not specific; TSS- Total suspended solid; TN- Total nitrogen; TP- Total phosphorous; FC- Faecal coliform; TC-Total coliform; CFU-Colony forming unit  8  Table 2.2: Typical concentrations (mg/l) in runoffs from different land uses Reference  Land use  TSS  BOD5  TN  TP  FC  TC  Pb  Boller et al. (2007)  Roads  100±60  -  2±1  0.3±0.1  -  -  0.02±0.01  Gotvajn and Zagorc-Koncan (2009)  Roads  -  1.2  .4  0.04  -  -  -  Resource Management Factsheet (1994)  Agricultural  75  40  26  -  -  -  Poudel et al. (2010)  Agricultural  1800 ± 550  7 ± 0.60  5 ± 0.70  0.40 ± 0.04  -  -  -  Poudel and Jeong (2009)  Agricultural  680 ±200  7±2  2 ±0.4  0.50±0.10  -  -  -  Udeigwe et al. (2010)  Agricultural  1600  6.2  4.8  0.7  Poudel et al. (2010)  Residential  80±20  5± 0.3  3 ± 0.2  0.6 ± 0.03  -  -  -  Poudel and Jeong (2009)  Residential  30 ±10  4±0.6  2 ±0.3  0.5 ±0.10  -  -  -  Lopez et al. (2006)  Municipal  5  29  13  2.4  40  149  -  a-  Unit in CFU/l, mean Standard deviation, TSS- Total suspended solid; BOD5- 5 day biochemical oxygen demand; TNtotal nitrogen, TP-total phosphorous; FC- Faecal coliform; TC- total coliform; Pb- lead  2.2 SWP Regulations Regulatory regimes related to source water protection are very common in Canada, e.g., Ontario (e.g., Clean Water Act, 2006a, 2006b), Alberta (e.g., Environmental Protection and Enhancement Act, 1993a, 1993b) and British Columbia (Ecological Reserve Act, 1975; Environmental Management Act, 2004a, 2004b). In other parts of the world such as the USA (e.g., State Source Water Assessment and Protection Programs, 1997; USEPA, 2001a-h), Australia and the European Union (Drinking Water Directive, 1980) also have promulgated regulatory frameworks for source water protection. Generally, SWP regulations directly or indirectly affect water quality at the source. These regulations help to control or eliminate potential hazards through effective implementation of selected SWP strategies. These strategies are related to the land use, storm water management, agriculture, farming (livestock), vehicle spills, roads, and municipal and industrial wastewater management. Figure 2.1 provides a schematic relating potential hazards to different management strategies. Landfill management is 9  related more to groundwater however, a landfill in close proximity to the surface source water can also be a hazard that may lead to microbial and chemical pollution (Yu et al., 2010).  Figure 2.1: Common source water hazards and SWP management 2.2.1 Canadian regulations Canada has roughly 9% of the world's fresh surface water (Can. Gov 2001). More than 85% of Canadians are supplied with drinking water from surface water (Davies and Mazumder, 2003). Surface waters can be affected by PPS and NPS and can lead to significant health-related risks if the water is not or poorly treated before consumption. Environment Canada (2001) has reported roughly 90,000 illnesses and 90 deaths due to unsafe drinking water each year. Especially, the E. coli outbreak in Walkerton (Ontario), Cryptosporidium outbreaks in North Battleford (Saskatchewan) and British Columbia (BC) have made drinking water one of the major health concerns in Canada in recent years (Krewski et al., 2004). Appendix A lists various regulations that may directly or indirectly influence source water protection in Ontario (ON), Alberta (AB), and British Columbia (BC). 10  Ontario (ON): In the recent past, a few serious incidences related to water outbreaks have been reported in Ontario. For instance in 2000, more than 2,300 people suffered gastroenteritis, 65 were hospitalized and 27 were affected by haemolytic uraemic syndrome during the Walkerton E. coli outbreak (Hrudey et al., 2003). More tragically, seven among those affected died (OMAG, 2002). After a long judicial investigation, the court concluded and recommended to reduce these types of incidences in the future through enhanced protection of source waters and training of technical staff. In Ontario, source water protection is part of the watershed management program and has been clearly mentioned in the regulatory framework. Ontario regulations 288/07 and 287/07 under the Clean Water Act (2006a, 2006b) describe the role of a source water protection committee (298 members) and also list numerous protected source waters. Public involvement and participation are highly encouraged under the Ontario regulation 385/08 (Clean Water Act, 2006b) for effective SWP. Appendix A shows various recommended SWP strategies and many setback rules to comply with specific regulations. Alberta (AB): Alberta promulgated many regulations that are indirectly related to source water protection. A few cases of high faecal coliforms (FC) levels have been reported in Oldman River and its tributaries in Southern-Alberta during the period of 1998-2000 (Hyland et al., 2003). Since 2007, water management programs have been implemented for province wide source water protection through land use planning. The main target of these programs is to establish a water policy and implement integrated water management. Proper education/ awareness and advanced storm water management have been promoted (Linking Source Water Protection to Land Use Planning in Alberta, 2008). Appendix A highlights SWP strategies such as the conservation of forests and wetlands and effective drainage design. In addition, some setback regulations from pollutant sources are also discussed. British Columbia (BC): Since 1980, more than 28 waterborne disease outbreaks have been reported in BC. In most of the cases, proper implementation of multi-barrier approach was lacking (Christensen and Parfitt, 2003). Numerous regulations have been promulgated over time that indirectly affect source water; however there is a growing need for more stringent and direct regulations for source water protection in BC. The Water Act (2007) encompasses groundwater regulations such as flood proofing. However, no allusion of surface source waters is available in 11  these acts. The land management to reduce impervious layer of soil for improved infiltration was also not properly addressed in the Environment and Land Use Act (1996). In the Environmental Management Act (2004a, 2007, 2008a), source water protection was not clearly defined. However, in the Ecological Reserve Act (1975) under B.C. Reg. 335 number of rivers, streams, and creeks have been mentioned to protect certain eco-systems. Many agricultural management acts have been promulgated in BC with SWP strategies such as forests, vegetative buffer (Appendix A). The control of vehicular spillage is also mentioned in the Environmental Management Act (2004b) with restriction on recreational dumping. However, the regulations in BC (Appendix A) discuss problems of surface source waters in a general context, but do not explicitly discuss the detailed strategies to protect or control them. 2.2.2 USA The United States Environmental Protection Agency (USEPA) is responsible for source water protection, and suggests a number of strategies to protect source water quality. In many major cities of USA, such as Seattle, San Francisco, Boston and New York, source water protection has been encouraged since the early 1800s - well before the availability of sophisticated water treatment technologies (Ernst, 2004). Currently, source water protection programs have been integrated with the Safe Drinking Water Act (SDWA) (State Source Water Assessment and Protection Programs Final Guidance, 1997). Conservation vegetation, e.g., buffer strips, riparian vegetation and wetlands have been recommended as methods to protect source water (USEPA 2001a, 2001d, 2001e). The USEPA has also proposed a number of other strategies, such as crop rotation, cover crop, spot treatment, storm water detention ponds`, fencing, geo-textile or impervious cover use. The basic information related to these strategies has been provided in Appendix B.  12  2.2.3 European Union Kramer et al. (2001) reported more than 260,000 cases of gastrointestinal diseases per year in Europe. In most cases, the incidences of bacterial dysentery can easily be attributed to countries with lower standards of sanitation systems. Worldwide more than 325 parasitic protozoan outbreaks have been reported and of which around 28 percent were from European countries (Kourenti et al., 2007). The water quality regulations in European Union countries mainly emphasize treatment technologies, and there is much less emphasis on source water protection (New Drinking Water Directive, 1998). Surface water regulations were initiated with the 1975 Surface Water Directive and later were updated by the 1980 Drinking Water Directive 80/778/EEC (Drinking Water Directive, 1980). Regulations related to drinking water (New Drinking Water Directive, 1998), sources of harmful substances (Dangerous Substance Directive, 2006), nitrate (Nitrates Directive, 1991) and pesticides (Council Directive, 1991) also emphasize strict maintenance of a threshold value (guideline value) of the water quality parameters rather than focusing on protecting source water. 2.2.4 Australia Australia is mainly a dry place but has variations, e.g., most of the areas have low rainfall while the coast of the eastern seaboard has high rainfall (ABS, 2000). The Murray-Darling River is the only major river which is under great pressure because of increasing agriculture (McKay and Moeller, 2000) that uses large quantities of pesticides and herbicides. Blue-green algae have been identified as a threat to surface waters all over Australia (Atech Group, 2000). In 1991, Australia faced tremendous Cyanobacteria blooms which affected more than 1,000 km of the Murray-Darling River and many rural water supplies (McKay and Moeller, 2000). Australia has one of the most sophisticated water supply management systems which strictly follow the principles of multi-barrier approach with special emphasis on source water protection1. From source to tap, the Drinking Water Quality Guidelines in Australia are based on risk assessment approach (Australian Government, 2004). Sedimentation ponds, artificial wetlands, water infiltration (recharge), source controls, and public education are highly encouraged in Australia for effective source water protection (Davis and Birch, 2009). Australia has promulgated numerous regulations to protect surface waters from industrial and commercial 1  13  sources of pollution for more than three decades, but it is still in the process of developing rules for dealing with nonpoint pollution sources (Davis and Birch, 2009). To make water management process more cost effective in Australia, the planning decisions are based on population density (Pigram, 1993).  2.3 Source Water Protection (SWP) Strategies SWP strategies can directly or indirectly improve the water quality at the source by reducing the pollutant loads entering into source waters. Generally, there are two types of SWP strategies: 1) storm water management, and 2) physical barriers to protect source water contamination. Most of the existing regulations have been focussed on storm water management, and control excess runoff from heavy rainfall or snowfall. However, other strategies including crop rotation, cover crop, fencing, and buffer strip can also be effective for protecting source waters. Examples of indirect strategies are enforcing regulations, increasing public education (and awareness) and regular water quality monitoring. These indirect actions lead to an improvement of source water quality as they promote direct actions. 2.3.1 Storm water management Storm water management (SWM) plays an important role in protecting source water from nonpoint pollution sources (NPS) which is also called low impact development (LID). Effective storm water management limit the untreated runoff (due to rainfall or snowmelt) or from the CSO entering into the source water. This can be achieved through specific arrangements by collecting runoff or CSO during excessive storm events. Storm water can cause problems as it carries many pollutants in large quantities to source water. The main purpose of storm water management is to reduce the pollutant load either through controlling runoff entering into the receiving water bodies or providing desired water treatment before it is discharged into surface waters. For example, places used for storm water management are detention ponds that collect water and provide basic treatment naturally and allow the treated water to go into the source waters at a regulated lower rate. The treatment is naturally performed through processes like sedimentation (Hsieh and Davis, 2005; Hsieh et al., 2007a, 2007b; Hathaway et al., 2009; Scholz and Yazdi, 2009), adsorption (Hsieh and Davis, 2005; Hathaway et al., 2009; Muthanna et al., 2007), infiltration (Muthanna et al., 2007; Passeport et al., 2009; Hsieh et al., 2007a, 2007b), filtration (Hsieh and Davis, 2005; Muthanna 14  et al., 2007; Scholz and Yazdi, 2009; Passeport et al., 2009), decomposition (Muthanna et al., 2007; Scholz and Yazdi, 2009), ion-exchange (Muthanna et al., 2007), oxidation-reduction (Hathaway et al., 2009) and biological uptake (Hsieh et al., 2007a; 2007b; Hathaway et al., 2009; Muthanna et al., 2007). These processes remove TSS, coliform bacteria, nutrients (e.g., phosphorus and nitrogen), heavy metals, and trace metals with varying degrees of success based on the design of the facility. Proper management of storm water also helps in reducing the quantity of storm water through evapotranspiration. Common strategies under storm water management includes: wet detention basin, storm water wetlands, sand filter, bioretention, grass swales, and extended detention basins/ dry ponds/ dry detention ponds/ detention basin (Hsieh and Davis, 2005; Hathaway et al., 2009; Muthanna et al., 2007; NCDENP, 1999; Scholz and Yazdi, 2009; Passeport et al., 2009). Figure 2.2 schematically compares six strategies that include: extended dry detention pond, wet detention pond, storm water wetland, bioretention area, sand filter, and grass swales.  1. Extended detention pond/ dry ponds (Figure 2.2a) are simple facilities and generally less effective. For example, the removal efficiency in case of TSS is as low as ~50% (NCDENP, 1999). 2. Wet detention ponds (Figure 2.2b) can hold storm water in a temporary pond first and then it allows it to go into a permanent pond. It may also contain a fore bay to collect the water that can remove suspended sediments. It may have marsh which can increase the biological uptake of the nutrients. Because of this improved design, wet detention ponds are more efficient in removing TSS (~85%). 3. Storm water wetlands are arrangements that contain both soil and plants and are effective in removing phosphorous, trace metals, and hydrocarbons through physico-chemical as well as biological pathways (NCDENP, 1999). Figure 2.2c shows the specifications of storm water wetlands (the area generally contains 50% high marsh, 40% low marsh and 10% water). 4. Bioretention area (Figure 2.2d) contains a grass buffer strip, a sand bed, plants, ponding area, and planting soil for pollutant removal. Three species of trees and three species of shrubs are essential on the area with certain gradient for the grass buffer strips.  15  5. Sand filters (Figure 2.2e) have sedimentation as well as a sand chamber to remove the particulate matter mainly in addition to moderate bacteria removal. However, sand filters are only suitable for extremely urban and small areas (NCDENP, 1999). 6. Grass swale (Figure 2.2f) is the cheapest and the simplest strategy that consists of a trapezoidal or parabolic pond with low depth containing hard grass. It has limited pollutant removal efficiency as compared to other options.  There are basic differences in the structure and design parameters (media depth, retention time, vegetation level and water depth) of these facilities that lead to various degrees of removal efficiencies for different water quality parameters. The percentage removals of TSS for extended dry detention pond, wet detention pond, bioretention area, storm water wetland, grass swales, and sand filter are 50%, 85%, 85%, 35-85%, 85%, 35% and 85%, respectively. Features of storm water treatment strategies are compared in Table 2.3. Weiss et al. (2007) provided a detailed discussion on operation and maintenance cost of storm water management strategies. Storm water wetlands, grass swales, and dry detention ponds are generally found to be costly due to high land requisition cost. Therefore they may not be very pragmatic in urban dwellings where land is limited and expensive. Apart from land costs, dry detention ponds and grass swales are comparatively cheaper strategies from an operation view point. Generally, a combination of low cost strategies results in an effective system that ensures higher pollutant removal (Middleton and Barrett, 2008). Scholz and Yadzi (2009) designed a system involving layer of gravel filter, detention pond and infiltration tank that has a removal efficiency of 77%, 83%, 32% and 47% for BOD5, TSS, NO3-N and ortho-phosphate-phosphorus, respectively. However, grass swales are considered aesthetically more pleasing because of their landscaping potential (NCDENP, 1999).  16  a)  b)  Figure 2.2: Storm water management ponds (NCDENP, 1999)  a- Dry extended detention basin  b- Wet detention pond 17  c)  d)  Figure 2.2: Storm water management ponds (NCDENP, 1999)  c- Storm water wetland  d- Bioretention areas 18  e)  f)  Figure 2.2: Storm water management ponds (NCDENP, 1999)  e- Sand filters  f- Grass swale 19  Table 2.3: Basic features of common storm water management strategies (NCDENP, 1999) Main purpose and mechanism  Advantages  Disadvantages  Special Requisition  O & M cost requirement and amount  Removes less pollutant (50% TSS) which is done with filtration  Removes less pollutant (50% TSS) which is done with filtration  Removes less pollutant (50% TSS) which is done with filtration  Frequent inspection is needed (e.g., 6 month) to run the system properly.  NR  Operation and maintenance cost is 5 % of the total cost.  Large area is required. Proper plant selection is required. Immediate maintenance after any storm event is necessary.  Special plants are required for pollutant removal1.  Cost is similar to wet detention basin but large land requirement makes it more expensive.  Extended detention basins/ dry ponds/ dry detention pond/ detention basins  Removes less pollutant (50% TSS) which is done with filtration  Removes less pollutant (50% TSS) which is done with filtration  Wet detention basin  Removes sediments, nutrients and heavy metals with infiltration and sedimentation.    It can be used in areas with low infiltration rate. Has an extra temporary pond (above the main pond) to hold 1 inch storm water for 2-5 days.    Storm water wetlands  Eliminates not only TSS but also phosphorus, trace metals, and hydrocarbons that are adsorbed to the surface of suspended particles with a number of physical, chemical and biological processes.     The plants increases aesthetic look with proper landscaping. It has longer detention time which ensures biological nutrient uptake by plants and algae.      20  Bioretention areas  Removes TSS, nutrients, trace and heavy metals by adsorption, filtration, volatilization, ion-exchange, and decomposition. Grass buffer strip of bioretention captures particles, and limits velocity. Sand bed provides aeration and drainage for microbes. Mulch layer ensures biological growth with organic matter, and petroleum decomposition.  Low cost compared to curbs, gutters for traffic area.  Maintenance of all the materials is necessary.    Specification for soil and plants are required2.  Maintenance of all the materials is necessary.  Sand filters Removes mainly particulate matter with moderate bacteria by sedimentation and filtration.  Sedimentation chamber of sand filter ensures sheet flow and limits sediments.  Sand chamber traps sediments with pollutants and provides media for microbial removal.  Cannot remove TSS and NO3-N efficiently.  Annual maintenance is required.  NR  Once in a year is necessary which is moderate.  Grass swales     Removes mainly sediments and trace chemicals Works as bio-filter to filter pollutants. Here, vegetation-lessen the velocity and increase the contact time for filtration     Low construction and maintenance cost Ensures replacement of gutter and curb with a better aesthetical view.     Limited pollutant removal (35% TSS) Use of fertilizer increases the nutrient amounts Can generate odor and mosquito problems.    Permeable and non compacted soil with lower water table (>1ft).  Less costly than gutter and curbs. Extensive sedimentation and erosion repair cost is required.  1- Table 2.2, 4.3, 4.4 (NCDENP, 1999); 2- Table 4.3 and 4.4 based on tolerance and morphology (NCDENP, 1999); 3- NR- Not required; O & M costOperation and maintenance cost  21  Table 2.4: Percent removal of water quality parameters through storm water management Reference  Type of SWM  TSS  TN  TP  Middleton and Barrett (2008)  BOD5  Pb  Extended detention basin  91  58  52  Weiss et al. (2007)  Extended detention pond  53a  Hathaway et al. (2009)  Wet detention basin  Weiss et al. (2007)  Wet detention basin  65a  52a  Weiss et al. (2007)  Storm water wetland  68a  42a  Hathaway et al. (2009)  Wetlands  Rusciano and Obropta (2007)  Bioretention  91.5  Davis (2007)  Bioretention  41  68  86  Davis (2007)  Bioretention (anaerobic sump)  22  74  79  Hsieh and Davis (2005)  Bioretention  >96a  74a  98a  Lucas and Greenway (2008)  Bioretention-vegetative  81a  91a  Lucas and Greenway (2008)  Bioretention  41  73a  Hunt et al. (2008)  Bioretention  32.2  31.4  Passeport et al .(2009)  Bioretention (grass)  56  63  Weiss et al. (2007)  Sand filter  82a  46a  Scholz and Yazdi (2009)  Combined detention and infiltration system  83  47b  FC  E. coli  69  25a >50  >50  59.4  >50 91.6  63  31.4  69  71  77  77  a- Change in load (mean) ; b- PO4-P; TSS- Total suspended sediment; TN- total nitrogen; TP- total phosphorus; BOD5-Biochemical Oxygen demand for 5 days; FC- Faecal Coliform  22  Vegetation allows storm water wetlands, bioretention and grass swales to remove nutrients through adsorption on the plants. Nevertheless, the excessive vegetation and its decomposition may also negatively affect the nutrient content in the source water (Li and Davis, 2009). Because of birds and animals‟ excreta, bacterial growth is possible and that leads to harmful microbial contamination. The information related to their impacts on various water quality parameters is provided in Table 2.4. The percent removals are generally reported by the reduction in event mean concentration (EMC) or by pollutant load reduction. Large variations can be observed for the same strategy due to variability in experimental setups though the basic features were similar. For example, for bioretention, Hunt et al. (2008) reported negative results (i.e., concentration increased) in the removal of NO3-N and NO2-N, whereas Davis (2007), Hsieh and Davis (2005) and Li and Davis (2009) have reported significant reduction. Removal efficiency in case of vegetatated systems largely depends on a number of factors such as the soil topography, available nutrients in the soil, pH, vegetation type, and special arrangements such as an internal storage zone. The variability in these factors may cause variability in the removal efficiency. For example, Hunt et al. (2006) have suggested a saturation zone or internal storage zone (inside the pond and less permeable soil composition) to create anaerobic conditions for improved nitrate or nitrite removal as compared to normal aerobic conditions. A saturation zone can also increase the retention time and can enhance adsorption of the pollutants (Passeport et al., 2009). Appendix C lists basic considerations for storm water management strategies. A minimum infiltration rate for soil is necessary to ensure a desired contact time between the media and the polluted water. Generally, more than two hours of retention time with < 0.25 m/hr infiltration rate is recommended (Hsieh et al., 2007a). Sandy loam or loamy sands have been recommended (NCDENP, 1999) as they provide an infiltration rate of 0.013-0.06 m/hr and ensure a proper retention time. In most cases, the retention time of these soils should be 24-48 hours, which is necessary to achieve effective pollutant removal. For vegetation, local plants that can absorb nutrients from the polluted water should be selected. P-index (phosphorous index) can be helpful to determine available nutrient amounts in soil that can limit excess nutrient release from soil (Hunt et al., 2008; Passeport et al., 2009). Exposure to sunlight with long retention times promotes coliform removal (Hathaway et al., 2009; Passport et al., 2009). A medium level of vegetation can ensure nutrient uptake as well as moderate the amount of sunlight exposure. 23  Generally, the media depth, water depth, vegetation level and retention time are found to be the common design parameters that can effectively improve the pollutant removal if designed properly. Table 2.5 provides the values of these design parameters (i.e., media depth, water depth and vegetation level, and retention time). An additional strategy permeable pavement can also be effective strategy to limit storm water runoff in urban development. Permeable pavements can be made of a matrix of concrete blocks or a plastic web-type structure with voids filled with sand, gravel, or soil (Fujita, 1994). But, they may suffer some serious limitations due to requirement for frequent maintenance, which is generally provided through vacuum suction (Dietz, 2007). They are suitable in specific situations such as in low traffic areas (e.g., parking area, sidewalks) (USEPA, 2000) or in an area with high clay soil to prevent groundwater contamination and have been found to be unsuitable during winter (Dietz, 2007). In some cases, they were found to be in effective after only six months (Barrett and Shaw, 2007). However, these strategies can be effective if the objective is groundwater recharge (Scholz and Grabowlecki, 2007). Green roofs can also be used as an SWM strategy, especially for old urban areas due to high imperviousness (USEPA, 2000). Green roofs consist of a vegetative layer, synthetic drain, geotextile and a media. The green roof reduces the flow rate of runoff but in some cases increases the chemical oxygen demand and phosphorus which is not suitable for all places (Bliss et al., 2009).  Table 2.5: Basic design parameters for storm water management strategies (NCDENP, 1999) SWM pond name  Media depth, m  Water depth, m  Vegetation level (type and intensity)  Retention time, hr  Extended dry detention basin  0  0.6  NIL  24-48  Wet detention basin  0  0.9-1.9  NIL or low  48  Storm water wetlands  0  >0.6  Medium: marshes  48  Bioretention areas  1.5-1.8  0.15  High: Grass, shrubs, trees  16-96  Sand filters  >0.46  0.45  NIL  24  0  0.1-1.5  Grass  15  Grass swales  24  2.3.2 Other SWP strategies Apart from SWM strategies, numerous best management practices (BMPs) have been adopted to limit pollutants entering into source waters. Natural vegetation such as vegetative cover (e.g. forests), plant cover, mulching, vegetative hedges, grass land management, and vetiver (an special type of plant) fencing have the natural ability to prevent soil erosion or runoff which can contain fertilizers, pesticides, and many other pollutants. Forest management with wetland and riparian areas can maintain surface water quality by protecting against soil erosion (Antoniadis et al., 2007; Kennedy and Mayer, 2002). Deforestation can lead to negative impacts such as an increase in sediment of transfer because of accelerated erosion, increase in nutrients after their release from decaying organic matter on the ground or in the water; increase in nutrient organic and inorganic concentrations because of harvesting, fertilization and pesticide application accumulation of slash and other organic debris in water bodies which can lead depletion of dissolved oxygen (DO); and an increase in temperatures because of removal of riparian vegetation; and increase in stream flow due to reduced evapo-transpiration (Ernst et al., 2004; Nunez et al., 2006). A vegetated filter strip can perform the same function as forests around source water. Vegetated strips (buffer strips/ filter strips) are found to be very effective in protecting source waters from eroded soil (Grace, 2002), nutrients (Dillaha et al., 1989), pesticides (Dosskey, 2001), agricultural pollution (Lee et al., 2000; Schoonover et al., 2005; Lowrance et al., 2002) and from other nonpoint sources of pollution as well as from runoff from roads or construction sites. Generally, buffer strips/ vegetated filter strips (VFS) consist of stiff and tall grass ranging from 0.75 to 1.2 m width (Kemper et al., 1992; Yuan et al., 2009) and surround the source water in order to stabilize the soil using vegetation roots. The Natural Resources Conservation Service (NRCS) has recommended an 8-10 m buffer strip width (Yuan et al., 2009). It has been observed that if the width > 5 m, the overall removal efficiency of pollutant loads may approach ~ 80%. Dosskey et al. (2005) recommended the use of Global Positioning System (GPS) and Geographical Information System (GIS) for providing variable widths of buffer strips depending on expected pollutant loads around source water. Figure 2.3 provides three constitutive zones of buffer strips (USEPA, 2001e): (a) four/five rows of trees, (b) one/two rows of shrubs and (c) 6.1m to 7.3m feet of grass. The forests and grass area within 6.1 m parallel to the source water can be used as a buffer to ensure good water quality (Herring et al., 2006). 25  Source water  Trees  Shrubs  Grass  Figure 2.3: Example of a buffer strip (USEPA, 2001e) Long rooted vegetation is preferable with a gentle slope for better efficiency. The USDANRCS recommended combining riparian forest and filtering strip (grassland) for effective water quality management (Herring et al., 2006). However, sometimes more than 100 m of buffer strips are recommended around lakes to avoid large pollution loads (Devito et al., 2000). Franti (1997) has discussed the NRCS slope-width relationship which suggests that high sloppy to land area needs larger buffer strips to slow down the velocity of runoff water gradually to slowly infiltrate the water. Table 2.6 highlights the key factors (such as width, slope, length and type of vegetation) of buffer strip and their impacts on removal efficiency of regular water quality parameters as well as pesticides such as atrazine, metolachlor and chlorpyrifor (Arora et al., 2003; Borin et al., 2010; Patty et al., 1997). Fencing is a simple arrangement to keep livestock away from a source water boundary (physical barrier) to avoid contamination from animal waste and prevent bank erosion caused by trampling of the banks by animals (Bewsell et al., 2007; Kolodziej and Sedlak, 2007). It also prevents, depending on the type of fence, the source water from re-suspension of sediments by disturbance in water and also ensures riparian health with undisturbed wetlands (Larsen et al., 1994). There have been only few studies (Table 2.6) reported on the impacts of fencing on source water quality. However, a limited reduction in nutrients, TSS and coliform bacteria is expected (McDowell, 2008).  26  Table 2.6: Effect of other SWP strategies in pollutant removal Reference  Brief description  Turbidity  TSS  TP  Barbed wire fencing (Both sides), length-800 m, distance between fencing and river-40-80m  35.51  69.31  0.071  34.61  67.72  0.072  98  86  NH4-N/ NO3-N  E Coli  91/ 78  92  Pesticides  Fencing Miller et al. (2010) McDowell (2008)  Fenced off with riparian vegetation  Vegetated filter strip (VFS) Sullivan et al. (2007)  993  Width-0, 1, 3, 8, 15, 25 m, slope-3.8-7% Duchemin and Hogue (2009)  57/ 33  Grass: Planted grain corn paired with a 5 m VFS of 45% red fescue (Festuca rubra L.), 45% redtop (Agrostis alba L.) and 10% perennial ryegrass (Lolium perenne L.)  87  86  48  Grass/poplar tree: Grain corn with a 5 m VFS with grass and eight hybrid poplar trees arranged in three rows in the grassy area  85  57  57  Borin et al. (2010)  Length-35m; width-6m, slope- 1.8%, vegetation- two rows of trees  50  Watts and Torbert (2009)  VFS with various amount of gypsum (CaSO4), width- 0.3048m  36  Mankin et al. (2007)  Natural succession grasses, width-16m, slope-4.1%,  99.9  98.6  A strip of 5 m planted with native grass with three rows of American plum, width-10.6m, slope-3.9%  99.7  93.4  A 5-m strip of natural succession grasses followed by three rows of American plum spaced 1 by 2 m, width-8.3m, slope-4.2%  99.5  92.1  . 75  Bhattarai et al. (2009)  Filter strip: width-14 m, length-113 m, slope-1.5%, vegetation-grass  Clausen et al. (2000)  Width-35m, length-250m  Arora et al. (2003)  Width- 1.52 m wide, length- 20.12 m, vegetation-grass  *  Patty et al. (1997)  Grassed strip, width- 0, 6, 12 or 18 m, slope- 7-15%,  **  70 92  73  1-Upstream concentration (mean ), Turbidity in NTU, TSS (Total suspended solids), DO (Dissolved Oxygen), TN (Total Nitrogen), TP (Total Phosphorous) in mg/l, E coli in #/100ml; 2- Downstream concentration (mean ), Turbidity in NTU, TSS (Total suspended solids), DO (Dissolved Oxygen), TN (Total Nitrogen), TP (Total Phosphorous) in mg/l, E coli in #/100ml; 3- Faecal coliform; VFS- Vegetated filter strip; *- Atrazine-46.8%, Metolachlor-48.1%, Chlropyrifor-76.9% removal; **-Atrazine- 72% removal  27  The agricultural practice of mixed cropping instead of monoculture, strip cropping, cover cropping, crop rotations, cultivation of shrubs and herbs, contour cultivation, conservation tillage, land levelling, use of improved variety of seeds, and horticulture increase the productivity of the soil and also its ability to slowly infiltrate water and protect source waters (Kleinman et al., 2005; USEPA, 2001a, 2001b). These strategies can limit fertilizer use and as a result reduce the probability of nutrient contamination in surface water (Kleinman et al., 2005). Therefore, some effective agricultural practices (e.g., cover crop and crop rotation) along with some parameter representing soil properties such as dry bulk density and soil water content can refer to proper agricultural management. For point sources of pollution, wastewater treatment is required to reduce pollutant loads which in addition to conventional pollution loads may also contain heavy metals and complex organic/ inorganic compounds. Numerous physico-chemical and biological treatment processes have commonly been used for the treatment of industrial, commercial and residential wastewaters before they are discharged into receiving water bodies. Point sources of pollution are not explicitly considered in this research.  2.4 Watershed Models Watershed models are used to predict a change in water quality based on specific land use, soil properties, precipitation patterns, vegetation type and related environmental factors. Many watershed models are available that use complex mathematical mass balance equations. One example is the universal soil loss equation (USLE) (Bingner et al., 2001; Bjorneberg, 1999; Chen et al., 2003; Irvine et al., 1993; Ovbiebo and She, 1995; Roesner et al., 1988; Williams et al., 1983; Young et al., 1987). The USLE determines the pollutant load in terms of erosion in ton/ha/year and depends on rainfall patterns, type of soil, crop type and land management practices (Wischmeier and Smith, 1978). Based on these data, the rainfall erosivity index, soil erodibility, topographical factor, plant cover factor and erosion control factor can be determined (Hacisalihoglu et al., 2010). Each of these factors and indices can be derived either from look-up tables or using prescribed mathematical equations that can be based on empirical relationships or simple mass balances. The modified universal soil loss equation (MUSLE) includes additional factors like runoff volume and peak runoff rate that adds complexity (Noor et al., 2010). These models are data 28  demanding which makes their use very limited compared to the export coefficient (EC) approach (discussed earlier in section 2.1). The USEPA TMDL (Total maximum daily load model)-toolbox provides a platform for a number of models such as Watershed Assessment Model (WAMView), Storm Water Management  Model  (SWMM),  Water  Quality  Model  (QUAL2K)  and  Watershed  Characterization System (WCS). The USEPA TMDL-toolbox is complex and deals with various aspects of hydrologic analysis, water quality and fate & transport modelling (USEPA, 1992). Table 2.7 provides a comparison of commonly used models and tools. The models such as BASINS, P8-UCM, and STORM are based on EC approach and found to be much simpler. Some models such as DRAINMOD, GISPLM, GSSHA, KINEROS2 and WEPP have limited use in dealing with water quality issues. On the other hand BASINS, HSPF, SWMM, TMDLtoolbox, WARMF, and WinHSPF are useful for water quality assessment. Tables 2.7 also compares selected models based on their capability to predict improvement in case of four SWP strategies including SWM-pond (storm water management), VFS (vegetated filter strips), PCAP (e.g., crop rotation, soil properties) and Fen (fencing).  2.5 Water Quality Assessment Quantitative assessment of water quality can be performed using a unit less measure, called water quality index (WQI) based on selected water quality parameters (WQP). The WQI formulation includes the following three steps: (1) selecting representative water quality parameters (2) converting non-commensurable WQP measurements into a monotonic quality scale to obtain sub-indices (unit less measure), and (3) aggregating sub-indices into a unit less number index (WQI). 2.5.1 Selection of water quality parameters (WQP) Generally various physico-chemical and microbial parameters have been used for surface water quality monitoring. Water quality can be compromised at the source due to point or nonpoint pollution sources. To develop a WQI, the selection of water quality parameters depends on the predefined use of water such as bathing, drinking, and agriculture.  29  Table 2.7: Comparison of various watershed models Complexity level Common models  Reference  AGNPS  Export coefficient  Water quality issues  SWP strategies  Complex equation  P  N  Young et al. (1987)  √  √  AnnAGNPS  Bingner et al. (2001)  √  BASINS  Lahlou et al. (1998)  DIAS/IDLMAS  Sydelko et al. (2000)  √  DRAINMOD  Skaggs (1980)  √  DWSM  Borah et al. (2002)  √  √  √  √  √  EPIC  Williams et al. (1983)  √  √  √  √  √  √a  GISPLM  Walker (1997)  √  GSSHA  Downer and Ogden (2002)  √  √  HSPF  HSPF (1985)  √  √  KINEROS2  Woolhiser et al. (1990)  √  √  LSPC  USEPA (2002)  √  √  MUSIC  Chiew and McMahon (1999)  P8-UCM  Palmstrom and Walker (1990)  PCSWMM  Irvine et al. (1993)  SLAMM  Pitt (1993)  STORM  USACE-HEC (1977)  SWAT  Neitsch et al. (2001)  SWMM  √  M  TS  SWM pond  VFS/ Veg.  PCAP  √  √  √  √  √  √  √  √  √  √  √  √  √  √  √  √  √  √  OM  √  √ √  √a  √ √  √ √  √  √  √  √  √  a  √  √  √  √  √  √  √  √  √  √  √  √  √  √  √  √  √  √  √  √  √  √  √  √  √ √  √ √  √  √  √  √  √  √  √  √  √  √  √  √  √  Roesner et al. (1988)  √  √  √  √  √  √  √  TMDL-Toolbox  USEPA (1992)  √  √  √  √  √  √  √  √  √  WAMView  Bottcher (1998)  √  √  √  √  √  √  √  √  WARMF  Chen et al. (2003)  √  √  √  √  √  WEPP  Bjorneberg (1999)  √  √  √  Fen  √ √  √  √  a  √  30  Complexity level Common models  Reference  WMS  EMRL (1998)  a-  Export coefficient  Water quality issues  SWP strategies  Complex equation  P  N  M  OM  TS  SWM pond  VFS/ Veg.  PCAP  √  √  √  √  √  √  √  √a  √  Fen  Wetlands; P-Physical/ aesthetic issues; N-nutrients; M- Microbes; OM-organic matter; TS- toxic substances; VFS-vegetated filter strip; Veg-vegetation; PCAP- pollution control by agricultural practice; Fen-fencing;  AGNPS- Agricultural Nonpoint Source Pollution Model; AnnAGNPS- Annualized Agricultural Nonpoint Source Pollution Model; BASINS- Better Assessment Science Integrating Point and Nonpoint Sources; DIAS/IDLAMS- Dynamic Information Architecture System/Integrated Dynamic Landscape Analysis and Modeling System; DRAINMOD- A hydrological model for poorly drained Soils; DWSM- Dynamic Watershed Simulation Model; EPIC- Erosion Productivity Impact Calculator; GISPLM- GISBased Phosphorus Loading Model; GSSHA- Gridded Surface Subsurface Hydrologic Analysis; HSPF- Hydrologic Simulation Program FORTRAN; KINEROS2- Kinematic Runoff and Erosion Model v2; LSPC- Loading Simulation Program in C++; MUSIC- Model for Urban Storm water Improvement Conceptualization; P8-UCM- Program for Predicting Polluting Particle Passage through Pits, Puddles, and Ponds─Urban Catchment Model; PCSWMM- Storm Water Management Model; SLAMM- Source Loading and Management Model ; STORM- Storage, Treatment, Overflow, Runoff Model; SWAT- Soil and Water Assessment Tool; SWMM- Storm Water Management Model; TMDLtoolbox- Total Maximum Daily Load- toolbox; WAMView- Watershed Assessment Model with an Arc View Interface; WARMF- Watershed Analysis Risk Management Framework; WEPP- Water Erosion Prediction Project; Win HSPF- An Interactive Windows Interface to HSPF;  31  For example for drinking, in addition to microbial indicators (e.g., Faecal coliform, total coliform, E coli and Cryptosporidium), the aesthetic parameters color and turbidity, as well as the nutrients and toxic substances are important. For recreational activities like swimming, a special emphasis is given to the microbial indicators and less on other factors. In case of agricultural use, generally the microbial indicators are not considered as important as heavy metals and nutrients. A flawed assessment of water quality for a specific usage can lead to adverse consequences in terms of human health. Table 2.8 presents guidelines and standards for common water quality parameters recommended by various agencies including WHO, USEPA, Heath Canada, European drinking water standard (New Drinking Water Directive, 1998), Australian drinking water standard and BC MOE (British Columbia Ministry of Environment) in terms of health and non-health (Primary and aesthetics) context. The recommended threshold levels vary significantly for the same parameter. It can also be noted that more than one water quality parameter is recommended for a specific water quality characteristic, e.g., conductivity and total dissolved solids (TDS). However, the general rule is that water quality should be assessed using minimum number of parameters by avoiding redundancy where possible without compromising on the assessment accuracy. It is generally recommended to use one or two WQP with respect to microbial, aesthetics, organic substances, heavy metals, nutrients and toxic substances. Kumar and Alappat (2009) provided a list of nine water quality parameters based on a recommendation by 142 water quality experts. This list includes: dissolved oxygen (DO), Faecal coliform (FC), five day biochemical oxygen demand (BOD5), nitrate nitrogen (NO3-N), phosphates (PO4), temperature (Temp), pH, total solids and turbidity. However, heavy metals and pesticides/ herbicides were not included in this list. In this research, total suspended solids (TSS) have been proposed instead of turbidity (commonly measured for drinking water after treatment) as it has been more frequently used in surface water quality studies (Bordalo et al., 2006; Jonnalagadda and Mhere, 2001; Liou et al., 2003; Ocampo-Duque et al., 2006). Based on detailed literature review, the proposed list may include TSS, Temp, TDS, total organic carbon (TOC), BOD5, total coliform (TC), Faecal coliform (FC), TN, TP, and lead (Pb). It is important to note that other water quality parameters can be added (or removed) from this proposed list if they are justified based on data availability and water use. For example, if the source water is susceptible to a specific toxic substance, say a pesticide, the list should include that pollutant. 32  Table 2.8: Water quality standards and guidelines recommended by various agencies Parameters (units)  WHO HB  USEPA  NHB  MCL2  MCLG  Health Canada MCL3  MAC  Europe1  AO  BC MOE6  Australia HB  AO  Guideline  Guide-Limit  Microorganism Faecal coliform, MPN/100 ml Total coliform, MPN/100 ml Primary /physical  0*  0*  0*  0*  10**  20**  <1  50**  Dissolved Oxygen, %  >85  pH  6.5-8.5  Turbidity, NTU  <5  6.5-8.5 0.5-1  Temperature, °C Total dissolved solids, mg/l Total suspended solids, mg/l Conductivity, μS/cm  1000  500  6.5-8.5  6.5-9.5  6.5-8.5  6.5-8.5  6.5-8.5  <1  <1  5  ≤15  15  22  ≤500  500 25  4.7 -5.84  2,500  < 400***  Nutrients Total phosphorus, mg/l Nitrate- N, mg/l  0.01 10  10  10  10  10  Nitrite-N, mg/l  1  1  1  1  1  Ammonia, mg/l  105  1.5  Nitrate, mg/l  50  Nitrite, mg/l  3  Organic matter 0.054  TOC, mg/l  4  BOD5 , mg/l  3  Heavy metals Arsenic, mg/l  0.01  0.05  Lead, mg/l  0.01  0.015  Mercury, mg/l  0.001  0.002  0.01  0  0  0.03  0  0.01  0  0  0.05  0  0.001  0  0  0  0 0  33  Chromium, mg/l  0.05  0.1  0.1  0.05  0  0.1  Cadmium, mg/l  0.003  0.005  0.01  0.005  0  0  Atrazine, mg/l  0.002  0.003  0  0.005  0  Simazine, mg/l  0.002  0.004  0  0.01  0  TCB, mg/l  0.02  0.07  0.07  Chlorpyrifos, mg/l  0.03  0  Pesticides  0.09  Each pesticides, mg/l  0  HB- Health based guideline NHB- Non- health based guideline MCL- Health-related standards called the Maximum Contaminant Levels (MCLs) which are the maximum permissible level of a contaminant in water delivered to users of a public water system under the Safe Drinking Water Act MCLG- Maximum Contaminant Level Goals (MCLGs) at levels where no known or anticipated adverse effects on health occur. MAC- The maximum acceptable concentration AO- Aesthetic objectives TCB- Trichlorobenzene (1, 2, 4) MPN/ 100ml- Most probable number per 100 ml *-Drinking water related **-CFU/100ml- Colony forming unit per 100 ml ***- Electric conductivity unit 1- European drinking water standard (98/83/EC) 2- Primary MCL 3- Secondary MCL 4- USP 23-Standad for purified water imposed by USEPA regulations for drinking water 5- Instead of nitrate-nitrogen the amount is for total nitrogen 6- Environmental standards  34  2.5.2 Transformation into sub-indices The units of water quality parameters are non-commensurate. In addition, some parameters positively impact „water quality‟ whereas others are negatively correlated with water quality. For example, a higher value of DO is desirable and represents a good quality, whereas, a lower value of TSS is desirable and represents a good quality. Therefore, before aggregation is performed, a transformation (using appropriate functions) is required to convert the real values of each WQP into a monotonic quality scale  [0, 1], where 1 represents the best quality and 0 represents the worst quality. The transformed values of a water quality parameter are referred to as sub-index, where each sub-index  [0, 1] regardless of the original units. Two types of functions are required to convert the real WQP value into sub-index, e.g., quality (parameter representing good water quality, e.g., DO) and pollutant (parameter representing bad water quality e.g., TSS) parameters. Figure 2.4 shows WQP transformations based on these functions denoted as subindex increasing, and sub-index decreasing, respectively.  Figure 2.4: Typical sub-index transformations for water quality parameters  Numerous functions have been proposed for transformation of real values into sub-indices (Banerjee and Srivastava, 2009; Cude, 2001; Swamee and Tyagi, 2000). Figure 2.5 represents the sub-index functions for ten selected water quality parameters. Normalization functions are 35  adapted from different studies (Banarjee and Srivastava, 2009; Brown et al., 1970; Cude, 2001; Liou et al., 2004; Ocampo-Duque et al., 2006) with little modification. However, a simple linear sub-indexdecreasing function can be used for any additional WQP representing pollutants in source water which is stated below: Sub-index Decreasing= 1-  (2.1)  where = Sub-index value of WQP  [0-1] X = WQP concentration, mg/l TV = Threshold value, mg/l  It should be noted that the sub-index value can be different depending on the intended use of water. For example, the sub-index value of FC should be different for drinking water quality from source water quality if the issue is other than drinking (Figure 2.5). 2.5.3 Aggregation formulations After transformation, the final step is an aggregation of sub-indices. There are four common types of aggregation formulations that include additive, multiplicative, logical or based on water quality guidelines. A few common formulations used for developing water quality indices are provided in Appendix D. These include: National Sanitation Foundation (NSF-WQI), Oregon (O-WQI), P-W WQI (Pesce and Wunderlin, 2000), Central Pollution Control Board (CPCBWQI) (Sarkar and Abbasi, 2006), River Pollution Index (RPI) (Liou et al., 2004), Universal water quality index (U-WQI) (Boyacioglu, 2007), S- WQI (Said et al., 2004), Simplified Water Quality Index (ISQA), CCME- WQI (Canadian Council of Ministers of the Environment Water Quality Index) and S-T WQI (Swamee and Tyagi, 2000). Aggregation formulations commonly encounter problems as a result of the abstraction of data. This includes ambiguity, eclipsing, compensation and rigidity. For example, weighted arithmetic mean has an eclipsing problem (i.e., one or more sub-indices show poor quality but the overall index does not reflect it), whereas root sum power addition suffers from an ambiguity problem (i.e., all sub-indices show acceptable quality but the overall index shows unacceptable quality). Minimum and maximum operators are free from ambiguity and eclipsing, but they fail to reflect the change in any sub-index other than the lowest (or highest) sub-index value in the group. 36  1  1  0.9  0.9  Sub-index TSSa  0.8  0.7  0.6 0.5 0.4 0.3  Sub-index TDS  Sub-index  Sub-index Sub-index TSS  0.7  0.6 0.5 0.4 0.3  0.2  0.2  0.1  0.1  0  Sub-index TDSe  0.8  0  0  50  100  150  0  50  100  150  TSS, mg/l  200  250 300 TDS, mg/l  350  400  450  500  TDS, mg/l  TSS, mg/l 1 1  0.9  Sub-index Temp  0.8  a 0.8  Sub-index  Sub-index Temp  Sub-index  0.7 0.6 0.5 0.4  Sub-index TCe  0.6  0.4  0.3 0.2  0.2 0.1 0  0 0 10  0  5  10  15  20 Temp, oC  25  30  35  1  Sub-index FCb  0.9  SWQ DWQ  Sub-index TNc  0.9 0.8  0.8  Source water quality  0.6 0.5 0.4  Drinking water quality  0.7  Sub-index TN  Sub-index  0.7  FC-SI  10  1  1  Sub-index  3  10  TC, CFU/l  Temp, °C  0.3  2  10  40  0.6 0.5 0.4 0.3 0.2  0.2 0.1  0.1 0  0 1 10  2  10  3  10  4  10 FC, CFU/l  FC, CFU/l  5  10  6  10  7  10  0  0.5  1  1.5  2 TN, mg/l  2.5  3  3.5  4  TN, mg/l  Figure 2.5: Sub-index functions for selected water quality parameters (continued next page)  37  1  1  0.9  0.9  Sub-index TPc  0.8  0.8 0.7  Sub-index Sub-index TOC  Sub-index Sub-index TP  0.7 0.6 0.5 0.4 0.3  0.6 0.5 0.4 0.3  0.2  0.2  0.1  0.1  0  Sub-index TOCd  0  0.05  0.1  0.15  0.2 TP, mg/l  0.25  0.3  0.35  0  0.4  0  1  2  3  4  TP, mg/l 1  0.9  0.9  Sub-index BOD5e  8  9  10  Sub-index Pbd  0.8 0.7  Sub-index Sub-index Pb  0.7  Sub-index Sub-index BOD5  7  TOC, mg/l  1  0.8  5 6 TOC, mg/l  0.6 0.5 0.4 0.3  0.6 0.5 0.4 0.3 0.2  0.2  0.1  0.1 0  0  0  5  10  15  20 BOD5, mg/l  25  30  35  0  0.002 0.004 0.006 0.008  40  BOD5, mg/l  0.01 0.012 0.014 0.016 0.018 Pb, mg/l  0.02  Pb, mg/l  Figure 2. 5: Sub-index functions for selected water quality parameters a- Liou et al., 2004; b- Brown et al., 1970; c- Cude, 2001; d- Ocampo-Duque et al., 2006; e- Banarjee and Srivastava, 2009.  The minimum and maximum operators do not recognize the importance of contributing subindices (i.e., lack of compensation) and suffer from insensitivity. Commonly WQI formulations are based on the „value‟ of water quality parameters. However, sometimes it also includes other factors, e.g., CCME-WQI, which is a well accepted index uses three factors2: F1 (scope), F2 (frequency) and F3 (amplitude) for the development of WQI. The CCME-WQI can be based on any number of water quality parameters (Lumb et al., 2006; 2  F1 (scope): percentage of variables that do not meet their objectives at least once, F2 (frequency): % of individual tests that do not meet their objectives F3 (amplitude): amount by which failed tests do not meet their objectives.  38  Rickwood and Carr, 2009). The values of index that may range from 0 to 100 are sub-divided into five water quality categories (excellent, good, fair, marginal, and poor). Recently, Khan et al. (2004) suggested including a „very good‟ category to match with experts‟ judgments. A compensation property helps to account for the contribution of all sub-indices, i.e., WQI should not be biased toward extremes (i.e., highest or lowest sub-index value).This property contradicts the situation where ambiguity-free and eclipsing-free models are desired. For example, maximum (or minimum) operators are skewed to the extremes and show poor properties with respect to compensation. As a result, there is a trade-off between ambiguity (and eclipsing) and compensation. Swamee and Tyagi (2000) proposed a model which minimizes ambiguity and eclipsing but suffers rigidity. Recently, Swamee and Tyagi (2007) have improved their formulation which minimizes ambiguity, eclipsing as well as rigidity. Sadiq et al. (2010) have developed penalty functions to deal with these issues in aggregation formulations. Table 2.9 provides a qualitative comparison of various WQI formulations based on their capacity to deal with various aggregation issues. Common watershed and water quality assessment models do not explicitly consider interconnection or interdependencies among the parameters. In addition, „acceptance‟ is a qualitative concept, which cannot be effectively handled through the mathematical formulations provided in Appendix D. Various advanced statistical/ mathematical methods, soft computing and artificial intelligence techniques (e.g., fuzzy rule-based, neural-network) have been effectively used for developing indices (Chang et al., 2001; Juahir et al., 2010; Icaga, 2007; Ocampo-Duque et al., 2006). In this study, we focus on soft-computing methods which are briefly discussed in the following section.  2.6 Soft Computing Methods Source water protection is a complex problem that requires an integration of watershed modelling with water quality assessment. Figure 2.6 was developed on the basis of the conceptual framework highlighting the impacts on water quality in the surface source and relating it to the selected SWP strategies. Numerous interconnections among water quality parameters and SWP strategies can be described through quantitative or qualitative relationships. Soft-computing methods are a good candidate to describe these interconnections and dependencies. 39  Table 2.9: Common issues in formulations for water quality indices (Sadiq et al., 2010) Unsuitable for different water use  Water Quality Index (WQI)*  Unable to handle missing data  Complexity3  Lack of handling Interconnection  Lack of handling Redundancy  NSF-WQI  H  Nil  H  M  H  Nil  H  H  O-WQI  Nil  H  M  M  H  H  H  H  P-W WQI  L  Nil  H  H  H  M  H  H  CPCB-WQI  H  Nil  H  H  H  M  H  H  RPI  M  Nil  M  M  H  M  H  H  U-WQI  H  Nil  M  L  H  M  H  H  S-WQI  Nil  Nil  H  M  H  Nil  H  H  ISQA  M  Nil  H  H  H  L  H  H  CCME-WQI  M  Nil  Nil  Nil  Nil  M  H  H  S-T WQI  Nil  Nil  Nil  L  H  H  H  H  *definitions are described in the text a-  Discussed in section 2.5.3  L-Low; M- Medium; H- High level presence of that particular problem  3  Complexity in terms of using complex equations  40  Figure 2.6: Source water protection strategies and water quality assessment Soft computing techniques consist of an array of heuristic approaches, such as fuzzy logic, evidential reasoning, and artificial neural networks (ANN). These techniques offer innovative solution for complex and uncertain problems. The ANN is a mathematical technique mainly used in data intensive conditions and has been used in many applications related to water quality management (e.g., Juahir et al., 2010; Kuoa et al., 2007; Rankovic et al., 2010; Singh et al., 2009). The main concept behind ANN is to use a number of input variables and determine their weights through hidden layers to predict the response of the system. For example, Chen et al. (2010) used back propagation ANN to predict downstream total phosphorous, total nitrogen and dissolved oxygen based on monthly river flow, temperature, flow travel time, rainfall, upstream total nitrogen and total phosphorous. Generally, the ANN performs better than multiple regression analysis in handling non-linear responses (Juahir et al., 2010). Other soft-computing methods such as evidential reasoning (ER), rough set theory, ordered weighted averaging (OWA) have also been used for water quality management. Evidential reasoning can handle both aleatory (natural stochasticity) and epistemic (ignorance) uncertainty. 41  Commonly, aleatory uncertainty is dealt with using traditional probability theory, whereas the epistemic uncertainties using Bayesian approach. There is no reported study related to water quality assessment using evidential reasoning approach. However, the evidential reasoning approach was used for predicting water quality failure in distribution networks (Sadiq et al., 2006). Fuzzy-based methods are found to be very effective in handling complex environmental problems (Subbarao et al., 2004; Jaramillo et al., 2009). These methods help to transform the natural language into quantitative values and have been used in developing water quality indices (Chang et al., 2001; Icaga, 2007; Ocampo-Duque et al., 2006) and water management (Liou et al., 2003; Spinella et al., 2008; Hajkowicz and Collins, 2007; Li et al., 2009). Fuzzy synthetic evaluation (FSE) (Francisque et al., 2009; Chang et al., 2001; Dahiya et al., 2007) and fuzzy clustering analysis (FCA) (Kung et al., 1992) have been used extensively in water quality management. In addition, the fuzzy inference system (FIS), e.g., Mamdani, and Takagi, Sugeno and Kang (TSK) algorithms, is also becoming popular in describing water quality status (Icaga, 2007; Lermontov et al., 2009; Ocampo-Duque et al., 2006) because of its capabilities such as easy interpretability and improved capability to represent cause-effect relationships and dependencies. Table 2.10 was developed to compare common soft computing methods such as ANN, evidential reasoning, bayesian network, rough set, and fuzzy-based approach in the context of water quality assessment. This comparison is based on the following criteria: simplicity, interpretability, vagueness, randomness, causality and redundancy. Simplicity refers to less use of complex equations or less number of data set requirements for water quality assessment. Interpretability means expert can interpret certain output (e.g., WQI) based on different parameter  values. Redundancy means the ability to address certain parameter redundancy inside the assessment formulations (for example Faecal coliform is a subset of total coliforms). Among the soft computing experts, there is a general consensus that fuzzy-based techniques are the most versatile and flexible. Table 2.11 summarizes advantages and disadvantages of fuzzy-based models considering different studies (Chang et al., 2001; Hajkowicz and Collins, 2007; Icaga, 2007; Jaramillo et al., 2009; Li et al., 2009; Liou et al., 2003; Ocampo-Duque et al., 2006; Spinella et al., 2008; Subbarao et al., 2004).  42  Table 2.10: A comparison of soft computing methods for water quality assessment Assessment criteria  ANN  Fuzzy sets  Evidential reasoning  Bayesian network  Rough sets  L  H  M  M  M  Interpretability  Nil  H  M  H  H  Vagueness  Nil  H  M  M  H  Randomness  Nil  L  H  H  M  Cause & effect  H  H  H  H  M  Redundancy  Nil  H  Nil  Nil  Nil  Simplicity  ANN- artificial neural network; L- Low; M- medium; H- High The relative scale represents efficiency and effectiveness of soft computing methods to deal with different water quality assessment issues.  Table 2.11: Advantages and disadvantages of fuzzy-based methods Advantages  Disadvantages Not free of eclipsing but can be handled with trial and error process    Easy interpretable by natural language      Can handle complex and vague situation    Cannot incorporate guideline values for water quality parameters    Can incorporate experts opinion with hard data    Suffer rigidity to some extent (careful selection of parameter can reduce it)    Can describe a large number of nonlinear relationships through simple rules    Easy to manipulate or can be biased due to human subjectivity    Provides a transparent mathematical model    Able to account interconnection (interdependencies) among parameters    Capable to handle missing data without influencing the final WQI value    Free of ambiguity and can represent different water quality usage if parameters are selected carefully  This research, focussed on two fuzzy-based methods; fuzzy-rule-based model (FRBM) and fuzzy measures theorem (FMT). FRBM is an inferencing method, whereas fuzzy measures theorem (FMT) is a multi-criteria decision-making technique that considers redundancy and 43  interconnections among aggregating factors (Sugeno, 1974). The details of these methods are given below. 2.6.1 Fuzzy rule-based model (FRBM) Zadeh (1965) first introduced fuzzy sets in his pioneering paper, where he argued that probability is an insufficient form to represent uncertainty because it lacked the ability to model human conceptualizations of the real world. Fuzzy-based techniques introduce robustness into systems by allowing a certain amount of imprecision to exist that paved a way to represent human linguistic terms as fuzzy sets, hedges, predicates and quantifiers. Fuzzy logic has played an important role in the management of uncertainties, especially in the areas of expert systems and rule-based models (Ross, 2004). During the last four decades the practical results of fuzzy systems have led to their general acceptance in various engineering disciplines. Fuzzy Logic is applicable to a problem if an approximate solution is acceptable. Although the input may be crisp, the approximation of the outcome is dependent upon the accuracy of the rule set, the inference technique and the selection of membership functions. Contrary to classical set theory where elements of a set may have „0‟ or „1‟ membership, fuzzy sets allow to define membership in an interval of [0, 1]. Fuzzy-based techniques are applicable where input based on human expertise, judgment, and intuitions are required. Fuzzy-based techniques have been successfully applied to a large number of real world problems, and have gained acceptance in the design and control of a variety of systems (Kosko, 1994; Yager and Filev, 1994). A typical fuzzy rule-based system has four components: fuzzifier, rule-base, inference engine, and defuzzifier. The fuzzifier determines the degree of membership of a crisp input in a fuzzy set through functions known as membership functions. The rule-base represents the fuzzy relationships between input and output fuzzy variables. The output of the rule-base is determined based on the degree of membership specified by the fuzzifier. The inference engine uses membership functions to determine conclusions of rules. Optionally, if needed, a defuzzifier converts fuzzy outputs into crisp values. In the case of multiple inputs, fuzzy rule-based models face „dimensionality‟ issues, which can be overcome by the use of hierarchical structures to reduce the number of rules. The most popular fuzzy rule-based systems include Mamdani (1977) and Takagi-Sugeno-Kang (TSK) ( Takagi and Sugeno, 1985). The main difference between these two models is the consequent part of fuzzy rules. The Mamdani model describes the consequent  44  part using linguistic variables, while the TSK model uses the linear combination of the input variables. Both models use linguistic variables to describe the antecedent part of fuzzy rules. In FRBM, relationships between variables are represented by means of fuzzy if-then rules of the form If antecedent proposition then consequent proposition. The antecedent proposition is always a fuzzy proposition of the type „X is A’ where X is a linguistic variable and A is a linguistic constant term. The proposition‟s truth-value (or membership value), which is a real number between zero and 1, depends on the degree of similarity between X and A. This linguistic model (Mamdani, 1977) has the capacity to capture qualitative and imprecise/uncertain knowledge in the form of if-then rules such as  Ri: If X is Ai then Y is Bj  i = 1, 2, …, L;  j = 1, 2, …, N  (2.2)  where, Ri is the rule number i, X is the input (antecedent) linguistic (fuzzy) variable and Ai is a fuzzy subset, which corresponds to an antecedent linguistic constant (one of L in set A). Similarly, Y is the output (consequent) linguistic (fuzzy) variable and Bj is a fuzzy subset, which corresponds to a consequent linguistic constant (one of N in set B). A fuzzy rule can be regarded as a fuzzy relation, i.e., simultaneous occurrence of values X and Y. For example, Equation 2.2 can be applied as follows, R: If slope level is medium then VFS efficiency is low where X denotes levels of slope, A denotes a fuzzy linguistic constant (a fuzzy subset) medium over the universe of discourse of slope levels (e.g., low, medium, high), Y denotes VFS efficiency, B denotes a fuzzy linguistic constant (or a fuzzy subset) low in the universe of discourse of VFS efficiency, and rules R defines their fuzzy relationship: Things become a little more involved when X is not exactly equal to medium but rather has a membership of, say, A2(x) = 0.5 to low and A3(x) = 0.5 to medium. It is clear that since the slope is less than medium VFS efficiency will likely be less than low. The full relationship between X and Y according to rule i can be computed in two basic ways, either by using fuzzy implications or fuzzy conjunctions (Mamdani, 1977). In the proposed approach, the Mamdani method is used, in which conjunction A  B is computed by a minimum (and type t-norm or conjunctive) operator. The interpretation of conjunction A  B is „it is true that A and B  45  simultaneously hold‟. The relationship is symmetric and can be inverted: Each rule is regarded as a fuzzy relation denoted by Ri (X  Y)  [0, 1].  (2.3)  Ri = Ai  Bj, i.e., Ri (x, y) = Ai (x)  Bj (y)  The minimum operator of Equation (2. 3) is applied to the Cartesian product space of X and Y, i.e., for all possible pairs of X and Y. The union of all fuzzy relations Ri comprises the entire model and is given by the disjunction A  B (union, maximum, or type, s-norm) operator of the L individual relations (rules) Ri (i = 1, …, L): L    R   Ri , i.e.,  R ( x, y )  max  Ai ( x)   B j ( y ) i 1, 2 ,...,L  i 1    (2.4)  Remembering that each relationship Ri is symmetric and can be inverted, the entire rule-set is now encoded in the fuzzy relation (rule) set R. Equation (2. 4) can be restated as (2.5)  y=xoR  where the output of the linguistic model is computed by applying the max-min composition (denoted by the operator „o‟) to the input or antecedent proposition. Suppose that A’ is an input fuzzy number (or a singleton), which is mapped on set A, and B’ is an output fuzzy number which is mapped on a set B, such that:   B ( y)  max  A' ( x)   R ( x, y)  (2.6)  X  Substituting R (x, y) from Equation (2. 6), the above expression can be rearranged as       B ( y)  max  max  A' ( x)   A ( x)   B ( y)  i 1, 2,...,L      X  i  j    (2.7)    Defining  i  max  A' ( x )   Ai ( x ) as the degree of fulfillment of the antecedent of the i-th X  rule, the output fuzzy set of the linguistic model becomes   B ( y)  max  i   B ( y) j  i 1, 2,...,L   (2 8)  The above algorithm is called the Mamdani inference. It is developed in Equations 2. 3 through 2. 8 for a SISO (Single-input-single-output) model. It can be extended to MISO (Multiple-inputs-single-output) model. For example, a two-input model will be:  46  i = 1, 2, …, L Ri,j: If X1 is Ai and X2 is Cj then y is Bk;  j = 1, 2, …, M  (2.9)  k = 1, 2, …, N This model is a special case of SISO, where the antecedent proposition is obtained as the Cartesian product of fuzzy sets A and C, hence the degree of fulfillment is given by:       i , j  max  A x1    A' x1  max  C x2    C ' x2     X1  i  X2  j  (2.10)  The extension of MISO to Q antecedents is straightforward. The algorithm can also be extended to multiple outputs (MIMO) model, which is a set of MISO models. Other conjunctive operators such as product can also be used (Yager and Filev, 1994) to make inferences. Consider an effect node B, which is connected by two causal concepts A and C. A graphical representation of the causal relationships is shown in Figure 2.7.  Figure 2.7: Two causal concepts connecting to an effect concept  The AND action represents that both A and C are simultaneously required for B to occur. The details of this two-input MISO model are graphically shown in Figure 2.8. The process is shown in three distinct steps, namely, fuzzification, inference (a rule base and an inference engine) and defuzzification. Assume that causal concepts A and C are activated at levels of A’ = 0.4 and C’ = 0.6, respectively. The rule set consists of 6 rules (3  2) and input activation signals (A’) and (C’) fire the first 4 rules to determine output B’ which is defined over the universe of discourse Y. The defuzzification step provides a discrete (crisp) value of an output of B, i.e., B’.  47  Step 1: Fuzzification A’ = 0.4 A3  A2  A1  Step 2: Inference  C’ = 0.6 C1  B1  C2  B2  B3  2,1  2,2  B4  0.7 0.6 0.4 0.3  1,1  y  1,2 Y  x1  If  and  x2  Then:  Y  Rule-base (Ri,j): Antecedent 1 = Ai; i = 1,…, L; L = 3 Antecedent 2 = Cj; j = 1,…, M; M = 2  B’  Consequent = Bk; k = 1,…, N; N = 4 Total number of rules = 3  2 = 6 Rule 1  R1,1 : If A1 and C1 then B1  Rule 2  R1,2 : If A1 and C2 then B2  Rule 3  R2,1 : If A2 and C1 then B2  Rule 4  R2,2 : If A2 and C2 then B3  Rule 5  R3,1 : If A3 and C1 then B3  Rule 6  R3,2 : If A3 and C2 then B4  Step 3: Defuzzification Defuzzified value  yo  yo  Defuz. (B’)    x1 has support (membership >0) in A1 and A2, and x2 has support in C1 and C2, consequently only the first 4 rules are „fired‟ (or activated).    Rule 1: A1(x1) = 0.3 , : C1(x1) = 0.4  1,1 = 0.3  0.4 = 0.3  B1(y) = 0.3    Rule 2: A1(x1) = 0.3 , : C2(x1) = 0.6  1,2 = 0.3  0.6 = 0.3  B2(y) = 0.3    Rule 3: A2(x1) = 0.7 , : C1(x1) = 0.4  2,1 = 0.7  0.4 = 0.4  B2(y) = 0.4   Rule 4: A2(x1) = 0.7 , : C2(x1) = 0.6  2,2 = 0.7  0.6 = 0.6  B3(y) = 0.6 Use center or area method to calculate the defuzzified value,    ( y).ydy   Defuz.B      ( y)dy  Defuz. (B’) = 0.54  Figure 2.8: Fuzzy rule-based model-making inference using two „causal factors‟  48  The crisp value approximates the deterministic characteristics of the fuzzy reasoning process based on the output fuzzy set Bk(y), which helps convert the uncertainty into an applicable action when solving real-world problems. The defuzzification method described in Figure 2.8, uses center of area method (Ocampo-Duque et al., 2006). 2.6.2 Fuzzy measures theorem (FMT) A significant aspect of aggregation in multi-criteria decision analysis is the assignment of weights to the different criteria. Until recently, the most often used weighted aggregation operators were averaging operators, such as the quasi-linear means. However, the weighted arithmetic means and, more generally, the quasi-linear means have limitations. None of these operators are able to model interaction between factors (concepts) in some comprehensible manner, which makes them unsuitable when it is important to consider interaction between concepts. Sugeno (1974) first introduced the term fuzzy measure. However, this term referred to a notion that was first introduced by Choquet (1953) and named capacity. Over the years the same notion has been used by many different names, such as confidence measure (Dubois and Prade, 1980), non-additive probability (Schmeidler, 1986, 1989), and weighting function (Tversky and Kahneman, 1992). Complex interactions between factors (i.e., sub- and super-additive) are best introduced by assigning a non-additive set function that permits the definition of weights to a subset of criteria rather than to an individual criterion. It is now widely accepted that additivity is not suitable as a required property of set functions in many real situations, due to lack of additivity in many facets of human reasoning (Ross, 2004). Sugeno (1974) proposed to replace the additivity property by a weaker one – monotonicity – and called these non-additive monotonic measures as fuzzy measures. It is important to note however, that fuzzy measures are not related to fuzzy sets (Sugeno, 1974). For a discrete universal set X = {x1, x2, … , xn}, a fuzzy measure is a set function, such that µ: (2n – 2)  [0, 1] satisfying the following conditions (where n is the cardinality of a set)   µ() = 0, µ(X) = 1, (where  is a null subset)    S  T  µ(S)  µ(T). (monotonicity)  49  For any S  X, µ(S) can be viewed as the weight or strength of the combination S for the particular decision problem under consideration. Thus, in addition to the usual weights on criteria taken separately, weights on any combination of criteria can also be defined. Monotonicity means that adding a new element to a combination cannot decrease its importance (Marichal, 1999). For example, S ={x1} and T = {x1, x2} are the (sub) sets of X= {x1, x2, x3}. The corresponding fuzzy measures, e.g., µ ({x1}) = 0.5 and µ({x1, x2}) = 0.7 fulfill the monotonicity condition. The fuzzy measure µ ({x1, x2, x3}) of a universal discrete set X (or sample space) will always be 1. The assessment of fuzzy measures by human experts is a daunting task, since the nonadditivity property of a fuzzy measure requires (2n −2) subsets. Sugeno (1974) proposed a socalled λ-fuzzy measure, which identifies the fuzzy measure of combined attributes from single attributes, expressed as   (A ∪B) =  (A) +  (B) + λ  (A)  (B); (λ > −1)  (2.11)  The parameter λ is used to describe an interaction between factors that are combined. According to the value of λ, the above equation can be interpreted as If λ > 0, then  (A ∪ B) >  (A) +  (B) (super-additive), if λ = 0, then  (A ∪ B) =  (A) +  (B) (additive), and if λ < 0, then  (A ∪ B) <  (A) +  (B) (sub- additive). For λ > 0, the super-additive relationship arises, which implies a synergy effect or strengthening dependency between factors, meaning that the combined contribution of factors A and B is greater than the sum of their contributions. For λ < 0, the sub-additive relationship arises, which implies a redundancy condition or dependency between factors, meaning that the combined contribution of factors A and B is lower than the sum of their contributions. If λ = 0, Equation (2.11) reduces to an additive measure, meaning that each factor acts independently. Sugeno‟s λ-fuzzy measure can be generalized for X = {x1, x2, … , xn} as follows:   x1 , x 2 , x n    1n   1    xi   1;   0    i 1  (2.12)  The value of λ is obtained through the boundary condition,  (X) = 1, which yields a polynomial equation with respect to λ, given by 50  1     1    xi  n  (2.13)  i 1  As Sugeno (1974) has shown, there exists a unique λ, which is greater than „-1‟ and not equal to zero, satisfying Equation (2. 13). The fuzzy measure over the given set S ⊂ X is computed as   S     1   1    xi   1  xi S   (2.14)  One possible meaning of a fuzzy measure can be defined as the level of importance or the degree of belief of a single factor towards the overall evaluation of the system.. Sugeno (1974) also introduced the idea of fuzzy integrals to develop tools capable of integrating all values of a function in terms of the underlying fuzzy measure ( ). An integral for fuzzy measures in a sense represents an aggregation operator, which contrary to the weighted arithmetic means, describes interactions between factors ranging from redundancy (negative interaction, i.e., sub-additive) to synergy (positive interaction, i.e., super-additive). Several classes of fuzzy integrals exist, among which the most representatives are those suggested by Choquet and Sugeno (Marichal, 1999). The Choquet integral C(X), first proposed by Schmeidler (1986) and later by Murofushi and Sugeno (1989, 1991), is based on an idea introduced in capacity theory by Choquet (1953). C(X) is an aggregation operator, where the integrand is a set of n values X = {x1, x2, …, xn}. The Choquet integral of a function x with respect to µ is defined by       C   X    x (i )  x (i 1)   {x (1) , x ( 2 )  x (i ) n  i 1    (2.15)  where x(1)  x(2)    x(n) represent the order of xi (also called utility values in utility theory) in set X in descending order. The values x1,, xn in our case can be replaced by activation values of causal nodes. Therefore, the Choquet integral can be re-written to make inference as      A j   A( i )  A(i 1)   { A(1) , A( 2 )  A(i )  n  i 1  (2.16)  where  {A(1) , A( 2)  A(i )  are fuzzy measures similar to causal weights (wij). Interested readers should refer to Grabisch (1996) for details. In the Equation (2. 16), A is the activation level in case of WQP sub-index to determine water quality groups and water quality parameter group WQI to determine overall WQI. 51  We use an example for describing the inference procedure in FMT for WQI considering only FC, and TC sub-index. Figure 2.9 shows how the activation values (sub-index) from FC and TC feed into an effect node for WQI. Therefore, the sample space for WQI = {FC, TC}. The power set 2Brequires defining 4 fuzzy measures as given in Figure 2.9, where B is the cardinality of sample space, which is 2. The fuzzy measures here are derived arbitrarily based on semantics (expert judgment). However, alternative objective methods based on data, λ-fuzzy measure and heuristics can be used to derive these measures (Grabisch, 1996). Lattice representation of the power set of B is also shown in Figure 2.9. It can be noticed in our example that the fuzzy measures are sub-additive, because  ({FC}) +  ({TC})   ({FC, TC}). It shows that redundancy exist in the causal nodes. But,  ({FC}) and  ({TC})   ({FC, TC}), which represents the monotonicity of the fuzzy measures. Therefore, under these conditions WQI is activated at a level of 0.46.  Fuzzy measures  i   ()  0.0   ({FC})  0.5   ({TC})  0.3   (Microbial WQI)=  ({FC, TC})  1.0  Lattice representation for the power set B  WQI= {FC, TC}  {FC}  {TC}   The activation values for FC and TC are 0.4 and 0.6, respectively, i.e., A({A})= FC’ = FC1 = 0.4  A({C}) = TC’ = TC2 = 0.6  Re-ordering is required to use Choquet integral. The activation values in descending orders are A(1) = 0.6  A(2) = 0.4  (where parenthesis shows the ordinal position)  Using Equation (2. 16), the activation value for WQI can be determined as follows A{ WQI} = WQI’ = [A({TC}) - A({FC})]   ({TC}) + [A({FC}) ]   ({FC, TC}) B’ = [0.6 – 0.4]  0.3 + [0.4]  1 = 0.06 + 0.4 = 0.46  Figure 2.9: Fuzzy measures theorem-making inference using two „causal factors‟ 52  3. MODEL DEVELOPMENT  3.1 Conceptual Framework A conceptual framework is proposed in Figure 3.1 that can link selected SWP strategies to the improvement in drinking water quality described by WQI. The conceptual framework has three components: 1) Reduced pollutant load calculations based on selected SWP strategies 2) Estimation of pollutant concentration in the source 3) Estimation of water quality index The first two components of the model are related to pollutant load assessment in a given watershed. Third component of the conceptual model is related to water quality assessment in the source water. The proposed framework involves six steps as shown in Figure 3.1. 3.1.1 Reduced pollutant load calculations To determine the reduction in pollutant load the following two steps are performed:   The first step involves estimation of potential land use and pollutants produced using the export coefficient approach for a predefined land use. In the case of pasture and livestock land use, the numbers of animals are used to estimate the pollutant load for each water quality parameter.    The second step involves estimating the percentage of pollutant reduction based on the efficiency of selected SWP strategies. Storm water management (SWM), vegetated filter strip (VFS), fencing, pollution control by agricultural practice (PCAP), and wastewater treatment (WWT) options are used in the proposed model. For each selected SWP strategy, the parameters described in Chapter 2 are used to estimate the removal efficiency of a specific pollutant.  3.1.2 Estimation of pollutant concentration in the source This component of the proposed model is described in step 3 which estimates the pollutant (water quality parameter) concentration using a mass balance over a pre-defined time period.  53  Figure 3.1: Proposed model based on conceptual framework 54  To avoid complexity related to mixing, time period, reactions and other physico-chemical processes occurring in the source water, a simple steady-stead mass balance approach is recommended in this study. 3.1.3 Estimation of water quality index (WQI) Component 3 of the proposed model is related to water quality assessment that involves three steps (Figure 3.1). Step 4 transforms previously calculated concentrations of WQP (step 3) into unitless sub-indices. In step 5, the water quality parameters are classified into five water quality groups: (1) Primary or aesthetic –WQI (2) Nutrient-WQI (3) Microbial –WQI (4) Organic-WQI and (5) Toxic substances-WQI.  Each water quality group may include one or more water quality parameters. For example, the Primary-WQI includes total suspended solids, TDS, and temperature; Nutrient-WQI includes total phosphorous, total nitrogen; Microbial-WQI includes faecal coliform and total coliform; Organic-WQI includes BOD5 and total organic carbon; and finally Toxic substanceWQI may include substances such as heavy metals (e.g., lead) and other user defined substances. Finally, the overall WQI can be derived using these five water quality groups in step 6. The value of WQI is in the range  [0 to 100], where 0 refers to the worst and 100 refers to the best quality water. Based on these values, we define six qualitative levels of water quality: poor (0-40), marginal (41-60), fair (61-70), good (71-80), very good (81-90) and excellent (91100) water quality. Moreover, the WQI for five water quality groups can also be classified using these levels.  3.1 Pollutant Loads based on Land Use Proposed model can account for a maximum of 14 water quality parameters (10 fixed and four user-defined toxic substances). The model recommends using at least one water quality parameter in each water quality group that includes TSS (Aesthetic), TN and TP (nutrient), FC  55  and TC (microbial), BOD5 (organic) and Pb (toxic substances). The export coefficient (EC) concept is used to estimate the concentration of selected water quality parameters for different land use (except livestock and pasture) to represent the pollutants. If the information for EC of a specific water quality parameter is not available, the average seasonal concentrations are recommended. The following phases describe the detailed procedure to estimate the concentration of a selected water quality parameter using the EC concept:  Phase 1: Define the area boundary (watershed or sub-watershed) that contributes to source water. Phase 2: Divide the total land area into a number of segments based on land use. Phase 3: Calculate the area for different land uses such as agriculture, farming (livestock), pasture/forage, urban, and forests. In case of urban (developed) land, define the nonpoint (e.g., highway/ roads) and point pollution sources (e.g., commercial, residential). Phase 4: Collect information for monthly and/ or yearly precipitation for the selected region to estimate the flow rates. Phase 5: Collect a soil variability map for the watershed that help determining the percentage of the precipitation that becomes the direct runoff (a portion of the rain will be infiltrated and will not become the part of runoff). The portion of precipitation that becomes direct runoff can be calculated using a runoff coefficient (k)  [0-1], a concept similar to the „Rational method‟, where higher value means more impervious area and vice versa. For example, k = 80% for concrete and k= 30% for forests (Martinez-Martinez and Campos-Aranda, 2010). Phase 6: Use the USEPA (2001i) formulation for EC approach to obtain EMC for a particular pollutant. Here, the formulation to estimate water quality parameter concentration (mg/l) was developed with the EC formula (USEPA 2001i) combined with runoff formulation as follows: ∑  C=R∑  (3.1)  where C = Pollutant concentration [mg/l] (for coliform CFU/ l) i = Order of land use type 56  = Export coefficient for ith number land use [Kg/ha/yr] or [Kg/ha/month] = Area of the ith land use [ha] k = Runoff coefficient representing amount of precipitation after infiltration (unit less) depending on the land use and land cover [%] P = yearly/ monthly total precipitation [mm] R= conversion factor (100)  Equation (3.1) can be used for both monthly/ yearly data depending on the precipitation and export coefficient. Table 2.1 shows a high variability in the EC values for same land use. Therefore, a careful expert judgement is required to select an appropriate value of EC for a specific land use. The USEPA (2001i) proposed typical EC values for different land uses (Table 2.2) and also corresponding pollutant concentrations. In case of missing information, other studies have been considered to assume respected EC values (Table 3.1). It should be noted that the land use EC values can be user defined and can be replaced if the site-specific values become available. For microbial indicators (FC and TC), the average concentration for different land use has been used, e.g., for FC, the event mean concentration (EMC) can be used (USEPA, 2001i; Mishra et al., 2008). To calculate the numbers of total coliforms from sources like agriculture, urban type 1 and forest land use, a multiplier of 2-4 times of Faecal coliform concentration are recommended. As the precipitation patterns can vary significantly in the reported studies, the monthly precipitation ratio can be used to convert monthly concentration from the reported data to the monthly concentration of a particular area. In case of manure load, the ASAE (2003) approach can be used to calculate monthly or yearly load in terms of kg/month or kg/year. Appendix E provides general statistics for total generated manure in terms of different parameters, e.g., total solids, TN, TP, BOD5, FC, TC and Pb. In addition, the average weights of different livestock have also been reported. The pollutant loads generated from manure are described in kg/(1000 kg body weight-day). Appendix E presents the amount of pollutants in fresh manure which usually contains 88-92% water for non-poultry based livestock and 73-75% for poultry based livestock (Ohio Livestock Manure Management Guide, 2006).  57  Table 3.1: Typical EC (kg/ha/yr) from USEPA (2001i) Land use type  TSS  TN  TP  BOD5  Pb  Agriculture (crop land)  2242  17.6  1.1  18  Urban type 1 (Highway/ roads)  1100  7  2.8  98  Urban type 2: Commercial  660  11  2.3  60.50  *  Urban type 2: Residential  390  8  2.2  47  *  Urban type 2: Industrial  780  11  5.4  52.5  *  Forests  abcde*  Elrashidi, 2007 USEPA, 1976 Badar and Romshoo, 2008 USEPA (2001i) assumed no Pb for forests Monthly concentration used from Boller et al. (2007); for yearly analysis average 0.03 mg/l has been assumed Unavailable and can be replaced by expert judgment  Therefore, the equation for estimating concentration for pasture or forage land use generation was developed with ASAE (2003) formulation combined with runoff equation as follows: =  ∑  (3.2)  where = Concentration of WQP [mg/l] m= Order of livestock type from 1- n = Average body weight of particular livestock [kg] (Appendix E) = Generated WQP in terms of mass for specific livestock type [kg/(1000 kg] 58  = Moisture multiplying factor (12%~8% for non-poultry based and 27%~25% for poultry based livestock) = pasture/ livestock land area [ha] ND = number of days (for monthly or yearly variation) S = multiplication factor (100)  3.2 Reduced Pollutant Loads The SWP strategies shown in step 2 (Figure 3.1) depends on area distribution of land use. A watershed or part of watershed can be divided into three zones (Figure 3.2): zone 1 is generally a source that generates point or non-point pollution; zone 2 is the area where a SWP strategy like storm water management will receive pollutant loads; and zone 3 is the designated place for vegetated filter strip (VFS). Figure 3.3 describes component 1 of the proposed framework that highlights these three zones. In zone 1, the pollutant loads are generated from non-point pollution sources like agricultural (LoadANPS) (crop land), urban (e.g., uncontrolled runoffs from roads) (LoadUNPS), forests (LoadFNPS), pasture/forage (LoadPNPS), livestock farms (LoadLNPS), and point pollution sources like urban (e.g., controlled commercial, industrial and residential discharges) (LoadUPPS). Pollution control by agricultural practice (PCAP) and fencing (Fen) can be used to control non-point pollution from agricultural and pasture land use (LoadANPS and LoadPNPS) (Figure 3.3). Equation (3.1) is used to calculate pollutants from forests, agriculture, urban type 1 and urban type 2 land uses. Moreover, Equation (3.2) is used to calculate the pasture and livestock pollutants (total manure pollutants). Beaulac and Reckhow (1982) have suggested that approximately 2-10% of the total load can potentially enter into the source water. As livestock don‟t graze for the whole day, approximately 1-5 % of the total manure pollutants can be considered as pasture/forage pollutants. These pollutants can be reduced by providing physical barriers like fencing (Fen) (Figure 3.3). In case of fencing, three types of land boundaries around a pasture area are possible, e.g., barbed fencing (B. Fen.), vegetated fencing (V. Fen.) and bare land (BL). The developed fencing efficiency equation using those three conditions are stated as follows:  Fen=  (3.3)  59  where Fen= Total effective fencing efficiency [%] = Percentage perimeter of land area covered by B. Fen. = Percentage perimeter of land area covered by V. Fen. = Percentage perimeter of land area covered by BL = B. Fen pollutant removal percentage, [2% assumed] = V. Fen pollutant removal percentage, [4% assumed]  Figure 3.2: Typical three zones in a watershed or a sub-watershed The main focus in this research is on NPS, however for type-2 urban land (Figure 3.3) where wastewater is usually treated before discharge, a user-defined input for treated wastewater load is allowed in the model. The total nonpoint and point pollutant loads are called LoadNPS1 and LoadUPPS, respectively. The LoadNPS1 can go into the source water with or without going through the storm water management ponds. The value X represents the percent pollutants (LoadNPS1) discharged without SWM (LoadNO_SWM) and Y represents the percent load discharged after passing through SWM (LoadSWM). Figure 3.4 highlights the second scenario.  60  Figure 3.3: Predicting pollutant reduction using different land uses (step 1-2 in Figure 3.1) The strategies of storm water management can be based on a series of ponds. In case of interconnected ponds, the pollution loads will be reduced further because SWM ponds will perform like a plug flow reactor. After passing through SWM ponds, the reduced load (Load4) will be either directly discharged into the source water as PPS (Load PPS_SWM) or as NPS (LoadNPS_SWM) in case of overland flow during heavy rainfall. The LoadNPS_SWM and LoadNO_SWM will define the total NPS load (LoadNPS2) going into the source water. Further reduction is possible through VFS, where the pollutant removal efficiency is defined as the following developed equation:  EfVFS =  (3.4)  61  where EfVFS = Total effective vegetative filter strip (VFS) efficiency [%] = Percentage of source water boundary covered by VFS = Percentage source water boundary covered by BL = VFS pollutant removal percentage based on VFS width, manning‟s n and VFS slope  Figure 3.4: Decision diagram for storm water management model The Load5 can be estimated from LoadNPS2 after reduction through VFS (Figure 3.3). Finally, Load5, LoadPPS_SWM, and LoadUPPS are combined to get the final pollutants to the source water (Figure 3.3). It can be noted that, all the pollutant concentrations are derived using mass balance related to flow rates. For example, LoadNPS2 is estimated from mass balance of LoadNPS_SWM and LoadNO_SWM (Figure 3.3):  62  =  Similarly, for LoadNPS1 and final pollutant concentrations (  (3.5)  ) are estimated as:  =∑  (3.6)  =  (3.7)  where Q = Pollutant flow rate C = Pollutant concentration p = Order of water quality parameter i = Order of land use for forests, agricultural, pasture, livestock and urban type 1 [1-5]  In case of pollution control by agricultural practice (PCAP), the removal efficiency of TSS depends on dry bulk density (DBD), soil water content (SWC), cover crop (CC) and crop rotation (CR). A simple FRBM is proposed to estimate the removal efficiency of TSS (Step 2 in Figure 3.1). The membership functions and inference rules are presented in Figure 3.5 and Table 3.2, respectively. The membership functions and inference rule for SWM and VFS efficiency are provided in Appendix F and G respectively.  3.3 Pollutant Mass Balance The SWP strategies aim at controlling pollutant loads entering into the source water. Pollutant loads discharged into surface water over a certain time period are not only diluted (advected) but also go through chemical reactions, diffusion, dispersion and other attenuation processes. Therefore, the magnitude of the surface water quality parameter varies with time. To account for the effect of SWP strategies on surface water, the fate of selected water quality parameters in source water needs to be determined through mass balance equations. Based on advectiondiffusion equations and chemical reactions many studies have identified mass balance formulations for selected water quality parameters (Wang et al., 2008; Covelli et al., 2002; Jha et al., 2007; Kasih and Kitada, 2004).  63  Figure 3.5: Membership functions for input parameters and removal efficiency of TSS in PCAP Table 3.2: PCAP rule-set for TSS removal Rule no 1  if  DBD L  and  SWC L  and  CC L  and  CR L  then  PCAP L  2  if  L  and  L  and  L  and  H  then  M  3  if  L  and  L  and  H  and  L  then  M  4  if  L  and  L  and  H  and  H  then  H  5  if  L  and  H  and  L  and  L  then  M  6  if  L  and  H  and  L  and  H  then  H  7  if  L  and  H  and  H  and  L  then  H  8  if  L  and  H  and  H  and  H  then  VH  9  if  H  and  L  and  L  and  L  then  L  10  if  H  and  L  and  L  and  H  then  L  11  if  H  and  L  and  H  and  L  then  M  12  if  H  and  L  and  H  and  H  then  M  13  if  H  and  H  and  L  and  L  then  L  14  if  H  and  H  and  L  and  H  then  M  15  if  H  and  H  and  H  and  L  then  M  16  if  H  and  H  and  H  and  H  then  H  DBD- Dry bulk density; SWC- Soil water content (field capacity); CC- Cover crop; CR- Crop rotation; PCAPpollution control by agricultural practice; L- low; M- medium; H- high; VH- very high  64  Park and Lee (1996) proposed the segment travel river ecosystem autograph model (STREAM). This model used the assumption of cells in series for representing the whole stream. Each cell was assumed to be a complete mixed flow reactor and used for BOD5, TSS, TDS, TC, TC, TOC, Pb and nitrogen and phosphorous species mass-balance determination. Appendix H lists common surface water quality models based on mass-balance equations. These models include Enhanced Stream Water Quality Model (QUAL2E) (Cox, 2003; Lin et al., 2010; USEPA, 1985), SIMulation of CATchments (SIMCAT, 2004), Temporally Overall Model for CATchments (TOMCAT) (Bowden and Brown, 1984), QUAlity Simulation Along River systems (QUASAR) (Whitehead et al., 1997) and MIKE 11 (DHI Water & Environment MIKE 11, 2005). Appendix H highlights the level of complexity due to the interactions among water quality parameters, e.g., the case of DO which depends on BOD, NH3-N, and NO3-N (USEPA, 1985). Various constants and coefficients used in these formulations and their descriptions are presented in tabular form in Appendix I. However, the dissolved oxygen (DO) usually depends on the unsteady state mass balance equation involving other parameters such as BOD5 and NH3N. As the focus of this research is on pollutants loads, we did not consider DO in water quality assessment. To estimate the overall impact of selected SWP strategies, we proposed a simple mass balance (only advection process) and assumed water quality parameters are conservative over the time of analysis. For example, we define a time period, say one year, and calculate the concentration of pollutants using selected WQP mass balance. For a large flowing system such as a river, the source water can be divided into a number of segments. Each segment will have an initial (Qo Co), and incoming (Qin Cin) mass flow rates and pollutant concentrations based on selected WQP (Figure 3.6). Under steady state conditions the basic equation for flowing system (e.g., creek or a river) will be (Figure 3.6): +  =  (3.8)  where = Incoming pollutant concentration [mg/l] after implementing a SWP strategy (also  in  Equation 3.7) = Incoming flow [liter/month] or [liter/year]  65  = Initial concentration of the water quality parameter [mg/l] inside the system = Flow of the surface water [liter/month] or [liter/year] C = Concentration [mg/l] of the pollutant or water quality parameter in the surface water, i = Order of source water segment= 1, 2, 3,…, (n-1), n  Equation (3.8) can be re-arranged as C=  (3.9)  The Equation (3.9) can be modified for small flowing system, e.g., creek by assuming one section only. Then, the equation reduces to:  C=  (3.10)  The equation for pond or lake system should consider volume (Volume = flow rate  time) rather than flow rate. Therefore, the modified Equation (3.9) for lake system will be: C=  (for small lake system, Figure 3.7)  (3.11)  For large lake system, similar type of equation (Equation 3.9) can be determined assuming the system a continuous stirred-tank reactor (CSTR).  66  Figure 3.6: Typical mass balance for a flowing system  Figure 3.7: Typical mass balance for a lake system  Finally, Steps 3-5 of the proposed framework requires transformation of pollutant concentrations into sub-index values followed by a two-step aggregations using FMT (Section 2.6.1). An illustrative example is provided below to describe the steps involved in the proposed conceptual framework.  3.4 Illustrative Example The example provides step-by-step calculations using the developed model based. We assume a watershed of 100 ha area which is divided into two segments: segment 1 (S1) – agricultural (crop land) and segment 2 (S2): urban type 1 land use (Figure 3.8a). The source water in this watershed is a small flowing system. Figure 3.8 provides two possible scenarios: baseline (current condition) and proposed (with new SWP strategies). Figure 3.8b shows two newly proposed SWM ponds that collect runoff from the watershed and then allows discharging treated water into the receiving water body. In addition, there is still some possibility that some untreated runoff can enter into the source water as non point pollution source. A vegetated filter strip is proposed to reduce some of these pollutant loads (Figure 3.8b). Additional strategies included cover crop (CC) and crop rotation (CR) to as a part of pollution control by agricultural practice (PCAP) to reduce pollutant loads from agricultural land use.  67  Figure 3.8: Scenarios for illustrative example  a) baseline  b) proposed  68  To estimate the water quality improvement in the receiving water body, the six steps procedure described earlier are followed for both scenarios using a monthly analysis. Table 3.3 provides the required input to perform these two scenarios. Table 3.4 provides step-by-step results of the component 1 of the conceptual framework. For both scenarios, the reduced pollutant loads are estimated using the efficiency of selected SWP with respect to TSS. Figure 3.3 is followed to estimate the pollutant concentrations and loads. The fuzzy rule-based model is used to estimate the efficiencies of selected SWP strategies where applicable. Matlab Fuzzy toolbox is used to perform the analysis. An example of pollution control by agricultural practice (PCAP) efficiency based on DBD, SWC, CC and CR is provided in Figure 3.9. The value for the DBD and SWC depends on the type of soil. For example, the dry bulk density for sand, silt loam and clay 1.52 g/cm3, 1.28 g/cm3 , and 1.2 g/cm3 (Linsley et al., 1982). Water holding capacity (field capacity) also depends on the soil type. The Field Estimation of Soil Water Content report (2008) provides a range of the soil water content (SWC) for different soil types. The average DBD, SWC (weighted average based on area proportion) for segment 1 (S1) is calculated (Table 3.3). Finally, DBD, SWC, CC, CR determine the PCAP efficiency using FRBM. In our case the PCAP efficiency for TSS is estimated as 73.3 % (Figure 3.9). Chapter 2 (Tables 2.4 and 2.6) provided the percent removals of different water quality parameters in terms of event mean concentration (EMC) for different SWP strategies. The inferences in terms of the removal efficiency of TSS using fuzzy-rule-based models for different SWP strategies are used to project the removal efficiency of other water quality parameters as TSS is one of the commonly used parameter to perform case studies related to SWP strategies. Therefore, following formulation is used to estimate other pollutant removal efficiency using TSS efficiency:  Percent removal of a WQP (for given SWP) =  Percent removal of a WQP (for given SWP) =    (3.13)  (3.14)  The above factor is estimated (ratio of percent removal of EMC for given WQP divided by percent removal of EMC for TSS) for each water quality parameter and each SWP strategy using Table 2.4 and 2.6.  69  Table 3.3: Required inputs to perform analysis using developed model Source water info  Land use info  depth (m)  3  Segment no  1  2  Width (b) (m)  3.5  Type of land use  Agricultural (crop land)  Urban type 1  n  0.03  Area (ha)  70  30  s  0.50%  Type of soil  50% sand and 50% clay  30% silt and 70% clay  Runoff coefficient  0.4  0.8  Land area slope (%)  3  4  Initial water quality parameter ( ) TSS, mg/l  100  TN, mg/l  3  TP, mg/l FC, mg/l TC, mg/l  0.5 200 2000  BOD5, mg/l Pb, mg/l Temp, ◦C  10 0.00004 20  Temporal info Month  April  No of days  30 200  Flow depth (d) (m)  2  SWP strategy info Pollution control by agricultural practice ( PCAP) DBD (g/cm3)  Segment no  1  Cover crop (%)  SWC (%)  1.36  0.27  Crop rotation (%)  Base line  Proposed  Base line  Proposed  20  100  30  100  Water depth (m)  Vegetation level (1-5)  Retention time (hr)  Media depth (m)  1.9  1  48  0  1.9  1  48  0  VFS slope (%)  VFS width (m)  n*  Storm water management (SWM) ponds ( proposed) Segment no  Type of SWM pond  Number of ponds  1  Wet detention  1  2  Wet detention  1  Dispose as (Figure 3.3) 50% NPS_SWM and 50% PPS_SWM 50% NPS_SWM and 50% PPS_SWM  Vegetated filter strips (VFS) Base line Segment no 1 2  1000  VFS length (m) 30  3  97  2  5  0.4  990  10  1.01  98.99  2.5  3  0.3  ECL (m)  (%)  (%)  Proposed 1  1000  900  90  10  2  5  0.4  2  990  990  100  0  2.5  3  0.4  ECL- existing creek length; BL- bare land; DBD- dry bulk density; SWC- soil water content; *- manning‟s n ; a- All the parameter except Temp: b- Precipitation represents both snowfall, and rainfall; P- percentage of source water boundary covered by certain facility (e.g., BL or VFS)  70  Table 3.4: Step-by-step results of component 1 of the conceptual framework Step 1: Estimate land use pollutants Segment no  Pollutant flow rate (liter/month)= kPA  TSS (mg/l)  1  5.60E+07  233.54  TN (mg/l) 1.83  2  4.80E+07  57.34  0.36  Pollutant concentration (using Equation 3.1) TP (mg/l) 0.11 50000 100000 0.15  14000  42000  BOD5 (mg/l) 1.88  Pb (mg/l) 0.000146  5.11  Step 2: Estimation of reduced pollutants Pollution control by agricultural practice ( PCAP) (valid for Segment 1) Segment no 1  Removal efficiency (using FRBM) Scenario  TSS (%)  TN (%)  TP (%)  FC (%)  TC (%)  Base line  37.9  30.3  30.3  18.9  18.9  BOD5 (%) 34  Proposed  73.3  58.6  58.6  36.7  36.6  65.9  TC (CFU/l) 81050  BOD5 (mg/l) 1.24  63350  0.6  Pb (%) 7.6 14.7  Load ANPS or LoadNPS2 Segment no 1  Scenario  TSS (mg/l)  TN (mg/l)  TP (mg/l)  Base line  145  1.3  0.08  FC (CFU/l) 40525  Proposed  62.4  0.8  0.05  31675  Pb (mg/l) 0.000135 0.000124  Storm water management (SWM) ponds Removal efficiency (using FRBM)  Segment no  Scenario  1 and 2  Proposed  TSS (%)  TN (%)  TP (%)  FC (%)  TC (%)  92.5  49.9  49  92.5  92.5  BOD5 (%) 83.3  TC (CFU/l) 81050  BOD5 (mg/l) 1.24  6079  0.21  Pb (%) 46.3  Load4 Segment no 1  2 Segment no 1 2  Segment no 1  2  Scenario  TSS (mg/l)  TN (mg/l)  TP (mg/l)  Base line  145  1.3  0.08  FC (CFU/l) 40525  Proposed  10.9  0.6  0.04  3039  Base line Proposed Type of load (Figure 3.3) NPS_SWM NPS_SWM  57.3 4.3 Flow rate (liter/month) 2.8E+07 2.4E+09  Scenario  TSS (%)  Pb (mg/l) 0.000135 0.000072  0.4 0.15 14000 42000 5 0.03 0.18 0.07 1050 3150 0.9 0.016 Segment Type of load Flow rate (liter/month) no (Figure 3.3) 1 PPS_SWM 2.8E+07 2 PPS_SWM 2.4E+09 Vegetated filter strip ( VFS) Removal efficiency (using FRBM and Equation 3.4) BOD5 TN (%) TP (%) FC (%) TC (%) Pb (%) (%) 2 2 1.3 1.3 2.3 0.5  Base line  2.6  Proposed  0.9  0.7  0.7  0.4  0.4  0.8  0.17  Base line  77.5  62  62  38.8  38.8  69.7  15.5  Proposed  86  69  69  43  43  77.5  17  71  Load5 Segment no 1 2  a-  BOD5 (mg/l) 1.2  0.63  0.04  3026  6053  0.2  0.000072  0.14  0.06  8576  25727  1.55  0.025351  0.6 0.06 0.02 598 1794 Final pollutant concentrations (Cin/ ) (using Equation 3.7)  0.19  0.013348  TN (mg/l)  TP (mg/l)  Base line  141.28  1.25  Proposed  10.78  Base line  12.91  Flow rate (Qin) (liter/month)  Base line Proposed  TC (CFU/l) 80003  TSS (mg/l)  Proposed  Condition  0.08  FC (CFU/l) 40002  Scenario  TSS (mg/l)  TN (mg/l)  TP (mg/l)  4.86E+09 36.4 0.3 0.1 4.86E+09 2.6 0.13 0.05 Assumed concentration (discussed in Section 3.2)  Pb (mg/l) 0.000134  FC (CFU/l)  TC (CFU/l)  BOD5 (mg/l)  Pb (mg/l)  11622 850  34402 2513  3.3 0.5  0.03 0.01  Figure 3.9: Results of a PCAP fuzzy rule-based model for the removal of TSS  72  For example for storm water management ponds, the removal percentage of TSS is multiplied by 0.54 for TN, 0.53 for TP, 0.90 for BOD5, 1.0 for FC, 1.0 for TC, and 0.5 for Pb. For the vegetated filter strip efficiency, the multiplication factors are 0.80, 0.80, 0.90, 0.50, 0.50, and 0.20, respectively. Same factors have been assumed for PCAP. Finally, for Fencing, the multiplication factors of 0.9, 0.8, 0.9, 0.9, 0.9 and 0.5 are estimated. The estimated removal efficiency for TN, TP, FC, TC, BOD5 and Pb are 58.64, 58.64, 36.65, 36.65, 65.97 and 14.66%, respectively (Table 3.4). Similarly, the removal efficiency of other SWP strategies are estimated and pollutant concentrations are derived in component 1 (Table 3.4). Table 3.5 summarizes results of steps 3-6. A simple mass balance is applied in step 3 using Equation 3.10. The values assumed for initial concentration of water quality parameters inside the source water are provided in Table 3.3. Manning‟s equation is used to calculate the monthly flow rate. Assuming that the source water has a rectangular cross section (b = 3.5 m, d = 2m and s = 0.5%), the flow rate is calculated (Table 3.5). Water quality parameter sub-indices are estimated (Table 3.5) in step 4 using functions described in Figure 2.5. Finally, fuzzy measures theorem (FMT) is used to estimate WQIs (steps 5 and 6). Figure 3.10 provides a vignette of results for all the steps involved in the proposed conceptual framework. Using enhanced SWP strategies, the WQI was improved to 46.52 (proposed scenario) from 42.47 (baseline scenario). The higher flow rates of the source water were the main cause of this small improvement in WQI. The SWP ponds with higher volumes can improve the results further however such decisions should be based on a detailed economic analysis. To further demonstrate the developed model, the next chapter provides a real life case study of Page Creek in Clayburn watershed, Abbotsford, BC, Canada.  73  Table 3.5: Results for component 2 and 3 (an illustrative example) Mass balance (using Equation 3.10) Scenario  TSS (mg/l)  TN (mg/l)  TP (mg/l)  FC (CFU/l)  TC (CFU/l)  Base line  93  2.7  0.46  1415  5446  BOD5 (mg/l) 9.3  Proposed  89.6  2.7  0.45  269  2055  9  Scenario Base line Proposed  Step 4: Water quality parameter sub index (A ) using functions in Figure 2.5 Sub index Sun index Sub index Sun index Sun index Sub index TN TSS Temp TP FC TC 0.08 1 0.3 0 0.40 0 0.12  1  0.3  0  0.60  Pb (mg/l) 3.2E-03 1.1E-03 Sun index BOD5 0.42  Sub index Pb 0.7  0.13  0.43  0.9  Step 5: Water quality groups (using FMT Equation 2. 16) Assigned weights (µ)  TSS  Temp  TN  TP  FC  TC  BOD5  Pb  0.4  0.2  0.4  0.4  0.5  0.3  1  1  Water quality (issue) WQI Scenario  Primary- WQI  Nutrient - WQI  Microbial-WQI  Organic -WQI  Toxic Substance-WQI  Base line  26.5  12.9  20.3  42  71  Proposed  30  13  36.3  43  90  Step 6: Overall WQI (using FMT Equation 2. 16) Assigned weights (µ) Condition Base line Proposed a-  Primary- WQI  Nutrient - WQI  Microbial-WQI  Organic -WQI  Toxic Substance-WQI  0.3  0.5  0.5  0.3  0.5  WQI 42.47 46.52  Equation 26 needs Qo in source water. Manning‟s equations is used to calculate the flow rate, Q o = A  ; n = 0.03 (for clean and straight natural  channel); s = slope = 0.5%, b = average width = 3.5m; d = 2m (stated in Table 3.3 ); A= b*d; P = b+2*d; and R=A/P  74  Figure 3.10: Illustrative example results-a snapshot  75  4. PAGE CREEK CASE STUDY  4.1 Study Area Page Creek is a sub-watershed of the Clayburn watershed (city of Abbotsford) in the Fraser valley (British Columbia, Canada). Abbotsford is located in the Coastal Western Hemlock Biogeoclimatic zone which is known for a cool mesothermal climate with relatively cool summers. The area is in class C of the Köppen climate classification and maritime temperate with most of the area is a coastal rainforest. The average annual precipitation is around 1,573.2 mm/yr (Environment Canada, 2004). The variation of the monthly total precipitation shows that the area remains comparatively dry during summer from June to September (Figure 4.1) (Statistics, 2010). The location for monitoring Page Creek water quality has been identified in  Total monthly average precipitation, mm/month  Figure 4.2. 300 250 200 150 100 50 0  Figure 4.1: Average monthly total precipitation (mm) in Abbotsford  76  Figure 4.2: Page Creek sub-watershed area in Clayburn watershed  77  The total area of Clayburn watershed is 7,451 ha (City of Abbotsford, 2009) where the Page Creek sub-watershed has area around 798 ha (calculated from a contour map). The subwatershed has been slightly modified to get the effective drainage area for the Page Creek sampling station. The total land area of the effective Page Creek sampling station is approximately 761 ha. The water quality monitoring data has been identified based on a study by Brown (2009). Page Creek is small in width ranging from ~2.5-5m (with an average of 3.75 m) and depth ranging from ~0.30-1.5 m (with an average of 0.9 m) unlike other creeks in this area (Fraser Salmon and Watersheds Program, 2009). In certain periods of the year it may remain dry or have a very low flow. The slopes of the creeks in this area vary between 0-3 percent (Barstead, 2004). The average slope is approximately 0.5% as the Page Creek is located in one of the flatter parts of the Clayburn watershed4. Based on the average monthly precipitation, the creek water flow depth has been assumed. Water depth in November is assumed to be the highest (0.70 m) as the precipitation is the highest during this month. It should be noted that precipitation represents both snowfall and rainfall. Assumed water depths from September to December are 0.221m, 0.422 m, 0.700 m, 0.606 m, 0.576 m, 0.466 m, 0.425 m, 0.349 m, 0.288 m, 0.229 m, 0.146 m and 0.143 m which is determined with precipitation proportion multiplied by the highest depth in November. Manning‟s equation was used to estimate the flow rate of Page Creek from September to December (Appendix J). Figure 4.3 plots flow rates in the creek around the year that are based on precipitation intensity. Page Creek sub-watershed is an agricultural and farming area. The land use map (Figure 4.4) shows the diversity of the land use type, e.g., berry (43ha), swine (7.5 ha), poultry (11 ha), dairy (104 ha), hobby (1 ha) farms, beef livestock (26 ha), forage/pasture (125 ha), forests (97 ha), nursery (14.5 ha) areas, impervious roads (7 ha) and some un-identified area (192 ha which is assumed as pasture for simplicity).  4  Observed from the contour map downloaded from  78  4.2 Data Preparation The main objective of the developed model is to estimate an improvement in the source water quality after implementation of selected SWP strategies. The input data requirements of the developed model are: land use distribution, selected water quality parameters, soil properties, types of SWP strategies, and creek water quality monitoring data. The model can predict the improvement in water quality after the implementation of SWP strategies. The following sections describe the data preparation for analysis. 1.20E+10  Flowrate, liter/month  1.00E+10 8.00E+09 6.00E+09 4.00E+09 2.00E+09 0.00E+00  Figure 4.3: Monthly flow rate prediction in Page Creek 4.2.1 Pollutant load calculations To calculate pollutant loads, the effective area of the creek is divided into six segments (Figure 4.4). These segments have been defined to estimate various factors that may impact the creek water quality. Table 4.1 provides the distributions of different land use within each segment (S1 to S6). Six basic land use categories including agriculture, pasture/forage, livestock farms, forests, urban type 1 (NPS from roads-U1) and urban type 2 (Commercial PPS-U2) are considered in this study (Table 4.1 and 4.2). These land use areas have been estimated using maps collected from the City of Abbotsford (2004). Monthly water quality data were collected from September 2007 to August 2008, therefore the channel flow rate and export coefficient (EC) values are also used in monthly units (Brown, 2009).  79  Table 4.1: Land use area distribution for six segments (m2) Land use  Category  S1  S2  S3  S4  S5  S6  Berry farm  Agricultural  496506  0  374285  0  53414  0  Forage/ pasture  Pasture  174961  336313  384532  288890  155398  84036  Dairy farm  Livestock  432491  171317  320735  296289  151465  102303  Forest  Forests  0  610072  361936  0  0  0  Beef cattle  Livestock  0  28665  8056  220727  6117  0  Poultry farm  Livestock  109759  0  48645  14093  0  46352  Swine farm  Livestock  0  0  0  74822  0  0  Hobby farm  Agricultural  0  0  0  0  9881  0  Nursing  Agricultural  145379  0  0  0  0  0  Roads  Urban type 1  0  0  15304  55975  0  0  Unknown area  NA  0  1566480  355155  0  0  0  Total area (m2)  1377466  2712847  1868648  963627  371612  233652  Total area (ha)  137.75  271.28  186.86  96.36  37.16  23.37  Table 4.2: The distribution for six land use types (ha) Segment  Agricultural area  Pasture area  Livestock area  Urban type 1 area  Forest area  Urban type 2 area  S1  64.19  17.50  54.23  0.00  0.00  0.00  S2  0.00  190.28  20.00  0.00  61.01  0.00  S3  37.43  73.97  37.74  1.53  36.19  0.00  S4  0.00  28.89  60.59  5.60  0.00  0.00  S5  6.33  15.54  15.76  0.00  0.00  0.00  S6  0.00  8.40  14.87  0.00  0.00  0.00  Total area (ha)  107.95  334.58  203.18  7.13  97.20  0.00  The EC approach was used to calculate pollutant concentrations. The EC values for selected water quality parameters were used in this case study. In case of missing values, the land use concentrations have been employed. As the precipitation patterns vary significantly in the reported studies from the Clayburn, the monthly precipitation ratio has been used to convert the monthly concentration from the reported data into the monthly concentration of Calyburn 80  watershed. Equation (3.1) has been developed to determine the concentrations of selected water quality parameters (pollutants). Equation (3.2) is used to calculate the concentration of pollutants for livestock and pasture land use from total manure mass production calculations (ASAE, 2003). City of Abbotsford (2004) has provided the information to estimate the numbers of dairy and beef. However, in case of poultry and swine, the area ratio has been used to estimate the poultry and swine livestock numbers. Based on information available for years 1995, 2000 and 2005 (Abbotsford Agricultural Strategy, 2009), the average livestock number in 2007 was estimated. Table 4.3 provides an estimation of different livestock numbers in six segmental areas in year 2007. In this analysis we assume that there is no pasture livestock from December-February. To perform the yearly analysis, livestock numbers are also estimated for years 2008-18 in the same way.  Figure 4.4: Six segments in the Page Creek effective area  81  Table 4.3: Pasture/forage livestock numbers in the study area for year 2007 Segment  Total livestock area (ha)  Dairy cattle (Number)  Beef cattle (Number)  Swine (Number)  Layer (Poultry) (Number)  S1  54.2  242  0  0  48,864  S2  20  115  6  0  0  S3  37.7  181  2  0  21,767  S4  60.6  167  107  1,377  6,219  S5  15.76  101  1  0  0  S6  14.87  52  0  0  20,434  4.2.2 SWP strategies The model considers four types of SWP strategies including PCAP, fencing (Fen), storm water management (SWM) ponds (six types of ponds) and VFS. The data requirements for each SWP strategy are discussed below: Figure 4.5 provides information related to the type of soil of the study area which consists of sandy silt, peat, granite, volcanic rock and silty clay. The dry bulk densities (DBD) of the basic soils are discussed earlier in Section 3.5. For granite soil and volcanic rock, 2.4 g/cm3 DBD has been assumed. Schwarzel et al. (2002) provided a DBD range of 0.14- 0.36 g/cm3 for the peat soils. In this research, we have assumed a value of 0.30 g/cm3 for peat. The average DBD is calculated based on the weighted average of land area proportion. Soil water content (SWC) is also estimated using average area proportion. The SWC for peat soil is assumed as 0.45 as it can hold the highest amount of water due its structure rich in organic matter (grain of mulches). For silt and granite soil (also for volcanic rock), the SWC are assumed as 0.20 and 0.02, respectively. Due to lack of reliable information, it was also assumed that there are no cover crops (CC) or crop rotation (CR) practices prevalent in the study area. Therefore, 100% CC and 100% CR have been proposed for the agricultural areas (except segments 4 and 6 because there was no crop production area). Table 4.4 provides a summary of the PCAP data derived for the six segments.  82  Figure 4.5: Soil type distribution in the study area  Table 4.4: Summary data for pollution control by agricultural practice (PCAP) Segments  Area proportions ( Sandy silt  Peat  )  Granite rock  Silty clay  DBD (g/cm3)  SWC  Future PCAP recommendations CC (%)  CR (%)  S1  700, 770  14, 051  0  669, 206  1.31  0.24  100  100  S2  325, 328  684, 931  1,176, 210  523, 485  1.53  0.20  100  100  S3  798, 625  524, 698  537, 610  0  1.38  0.21  100  100  S4  963, 627  0  0  0  1.40  0.17  0  0  S5  371, 612  0  0  0  1.40  0.17  100  100  S6  233, 652  0  0  0  1.40  0.17  0  0  DBD- Dry bulk density; SWC- soil water content (field capacity); PCAP- pollution control by agricultural practice  83  Generally fencing is provided for pasture/ forage area to keep animals away from the water source. Because of lack a data, conservative estimates of 2% and 4% load reduction has been assumed for the barbed (  ) and vegetated (  ) fencing. It is very common to have partial  fencing therefore we used Equation (3.3) to estimate the efficiency in case of partial fencing. Figure 4.6 shows the storm water collection in Abbotsford. There is no SWM pond identified in the study area, however, there are a few natural wetlands. Therefore, a SWM system can be proposed to control the storm water entering into the creek at higher flow rates. One SWM pond for each segment is proposed. Wet detention ponds, storm water wetlands, and dry detention ponds are recommended. However, for a smaller area, dry detention pond has been proposed. Table 4.5 provides the design parameters for recommended SWM ponds. It can be assumed that 50% (by volume) SWM treated water will be directly disposed to the creek water through pipeline. The remaining 50% will be discharged in to the creek as NPS.  Table 4.5: Future storm water management pond options for six segments Segments  Land area (ha)  Type of SWM pond  SWM parameters recommendations Water depth (m)  Vegetation level (1-5)  Retention time (hr)  Media depth (m)  S1  137.75  Storm water wetlands  0.7  3  48  0  S2  271.28  Wet detention pond  1.9  1  48  0  S3  186.86  Wet detention pond  1.9  1  48  0  S4  93.36  Storm water wetlands  0.7  3  48  0  S5  37.16  Dry detention pond  0.6  0  24  0  S6  23.36  Dry detention pond  0.6  0  24  0  84  Figure 4.6: Storm water collection in Clayburn watershed  In case of VFS, the parameters required for estimating the efficiency of VFS (  ) are  provided in Table 4.6, which shows both current VFS condition and suggests some improvements in the future. The monitoring data for Page Creek are available only for TSS, temperature, PO4, TN, FC, and Pb. As the minimum requirement of the model is to have monitoring data for TSS, TN, TP, FC, TC, BOD5, and Pb, some assumptions have been made for the missing data. It should be noted that PO4/ TP ratio of 0.455 has been used to derive TP from  5  Correlations between trophic indicators (OECD) available at  85  Table 4.6: Inputs for VFS around the Page Creek-current and future recommendation Current Segment  ECL (m)  Avg. Channel width  VFS length (m)  (%)  (%)  (m)  VFS area slope1  Avg. VFS width  (%)  (m)  n*  S1  1, 458  5.0  448  30.73  69.27  0.5  15  0.2  S2  1, 161  4.5  120  10.34  89.66  0.5  30  0.3  S3  2, 460  2.5  178  7.24  92.76  0.5  19  0.3  S4  2, 376  2.5  32  1.35  98.65  0.5  30  0.4  S5  1, 159  4.5  0  0.00  100.00  0  0  0  S6  1, 455  5.3  0  0.00  100.00  0  0  0  Future recommendations S1  1, 458  5.0  1, 458  100  0  0.5  15  0.4  S2  1, 161  4.5  1, 161  100  0  0.5  15  0.4  S3  2, 460  2.5  2, 460  100  0  0.5  15  0.4  S4  2, 376  2.5  2, 376  100  0  0.5  15  0.4  S5  1, 159  4.5  1, 159  100  0  0.5  15  0.4  S6  1, 455  5.3  1, 455  100  0  0.5  15  0.4  1. Slope was assumed ECL- Effective creek length for each segment; VFS- Vegetated filter strip; BL- Bare land; Equation 3.4. * Manning‟s constant  and  discussed in  4.3 Results and Discussions A detailed analysis has been performed using the model for both monthly and yearly time steps. The WQI has been predicted monthly from September 2007 to August 2008 and yearly for years 2007 to 2018. A sensitivity analysis based on land use variations has been performed to observe the effects of different land use on the creek water quality in comparison with base line condition (present land use and SWP strategy conditions). The sensitivity analysis also allows the prediction of the final conditions from 2008 to 2018 using proposed SWP strategies 86  compared to the base line conditions. Finally, water quality assessment based on monitored creek has been compared with NFS-WQI. 4.3.1 Predicted WQI–monthly variation Monthly analysis has been performed from September 2007 to August 2008. September 2007 water quality monitored data have been considered as the initial creek water quality parameter value (Co) to carry out the analysis. The yearly cycle is divided into summer season (S) from April to September and winter season (W) from October to March. Results are presented for these two seasons using box-and-whisker diagrams for LoadNPS1, LoadNPS2 and the final pollutant loads entering into the creek based on zones 1, 2 and 3 (Figure 3.2 and 3.3). Box-and–whisker is a diagram to represent a set of data using, median, lower quartile (25th percentile), upper quartile (75th quartile), maximum and minimum value (explained in Appendix K). Pollutant load calculations and estimation of SWP (PCAP and Fen) efficiencies are used to predict LoadNPS1 in zone 1. Equations 3.1 and 3.2 are used to calculate the possible pollutant concentration for the land uses (e.g., agricultural, pasture, livestock, forest, and urban type 1). Finally, PCAP and Fen were implemented to reduce pollutants for agricultural and pasture land use in zone 1. Previously identified PCAP parameters were utilized to obtain the PCAP efficiency using FRBM. Finally, Equation (3.3) has been used to calculate the efficiency for Fencing (Fen). Figure 4.7 shows the changes in water quality parameters for LoadNPS1 with and without implementation of PCAP and Fencing in Zone 1. Significant reductions in the concentrations of TSS, moderate in case of TN, FC, and TC and minimal in case of TP, BOD5 and Pb are estimated. The results suggest that PCAP and Fencing are more effective to remove (or to protect the source water from) TSS as compared to other pollutants (water quality parameters). However, more strategies are required as the concentrations of most of the pollutants were above the guidelines. Comparison between summer (S) and winter (W) analyse predict that water quality is generally worse during summer months. Zone 2 receives polluants from zone 1 (LoadNPS1). In this zone, storm water management is used to reduce the pollutant concentrations. Six storm water management ponds have been proposed to reduce LoadNPS1. Finally, LoadNPS2 is calculated based on LoadNPS1 and storm water management pond efficiency. Figure 4.8 shows the LoadNPS2 for zone 2.  87  Figure 4.7: NPS1 loading change for the seven pollutants a) without PCAP & fencing and b) with PCAP and fencing  88  Figure 4.8: NPS2 loading change a) without storm water management pond and b) with storm water management pond  89  Figure 4.9: Final pollutant loading change for the seven parameters a) without vegetated filter strip and b) with vegetated filter strip  90  Figure 4.10: Creek water quality prediction for water quality parameters a) with SWP strategies and b) without SWP strategies  91  It shows significant improvement in water quality due to removal of TSS, FC, TC and comparatively moderate improvement due to reduction in TN, TP, BOD5, and Pb. In zone 3, LoadNPS2 was reduced due to VFS and the final pollutant loads (summation of Load5 and LoadPPS_SWM) going into the creek water were calculated. The final pollutants in terms of pollutant concentration without VFS and possible pollutant removals after VFS implementation are represented in Figure 4.9. The figure shows improvements (pollutant reduction) in all parameters. It suggests also that better water quality is expected during winter as compared to summer. Equation 3.10 was used to estimate the changes in the pollutant concentrations in the source water, Figure 4.10 shows changes in pollutant concentrations in the creek. The initial pollutant concentrations, determined by considering the effect of initial SWP strategies around the creek, are as expected higher than the concentrations calculated based on the pollutant removal efficiency of all the proposed SWP strategies. Figure 4.10 also shows reduction in the concentrations for all the parameters. Lower values have been estimated during winter compared to the summer. Creek water quality assessment has been done using WQI based on fuzzy measure theory (FMT). The weights used in this study are assumed (Table 3.5); however, the model allows userdefined weights as well. The assessment has been done based on monitoring data (Brown, 2009), predicted model initial (base line scenario using the current SWP strategies) and final (using proposed SWP strategies) values. It should be noted that the monitoring data represent a particular day of the month, and not a monthly average concentration. Therefore, a comparison between the monitored and the initial predicted assessment results might show some discrepancies. Appendix L provides the sub-indices values from September 2007 to August 2008 for the monitored, initial and final predicted model data sets. Figures 4.11, 4.12, 4.13, 4.14 and 4.15 exhibit the water quality index for each group: primary, nutrient, microbial, organic and toxic substances, respectively. Figure 4.16 shows overall WQI for summer and winter.   For primary- WQI group, Figure 4.11 shows that the predicted WQI is independent of the season. Proposed SWP strategies (final conditions) improve water quality. Moreover, each of the three cases, the water quality during summer and winter varies only slightly, especially for the final condition where it is always excellent. For monitored data, during 92  summer the WQI was excellent but varied from very good to excellent during winter. In the case of initial conditions, WQI varied from good to excellent in the summer and was always excellent during winter.  Figure 4.11: Seasonal primary-WQI   For the nutrient- WQI group (Figure 4.12), the monitored WQI shows a large variability from poor to fair during summer and poor during winter. Initial predicted WQI for summer and winter were the same from marginal to fair. Finally, the predicted final WQI is good during summer and varied from fair to good during winter.    For the microbial- WQI group (Figure 4.13), the monitored WQI during summer varied from poor to fair and poor during winter, whereas the initial and final predicted WQI was poor in all seasons.    For the organic- WQI group (Figure 4.14), the monitored WQI varied from poor to very good during summer and poor to good during winter. Initial predicted WQI was in marginal range whereas the final WQI raged between marginal to very good during summer and was marginal during winter.  93    For the toxic substance- WQI group (Figure 4.15), the monitored WQI was excellent which showed very little improvements after the implementations of selected SWP strategies.  Figure 4.12: Seasonal nutrient-WQI  Figure 4.13: Seasonal microbial-WQI 94  Figure 4.14: Seasonal organic-WQI  Figure 4.15: Seasonal toxic substance-WQI  95    The WQI for monitored data varies from good to very good during summer months and from marginal to very good during winter. In case of predicted initial value, the WQI is in the good range during both seasons. Predicted final WQI was very good in both seasons.  Figure 4.16: Seasonal overall WQI  Three factors including incoming pollutant flow rate, EC and creek water flow rate are affecting water quality. Incoming pollutant flow rates have two contradicting effects: 1) increase in the transport of pollutants due to erosion; 2) dilution of concentration. Dilution due to a higher flow rate can lead to lower concentration whereas higher erosion can increase the pollutant concentration. Similarly, higher creek water flow rates reduce the pollutant concentration through dilution. It usually takes time to change the existing pollutant concentration in the creek water as the flow rates in the receiving bodies are much higher as compared to incoming pollutants loads. For the monthly WQI assessment, identical monthly EC derived from the annual EC has been used for each month due to lack of monthly data. Therefore, for a better comparison between results of initial and final conditions, a yearly WQI has also been derived. 96  4.3.2 Predicted WQI–a yearly variation Yearly simulations have been performed to predict the water quality indices from 2008-18 based on the monitored data (September 2007). Results for the initial and final conditions have been presented in Figure 4.17 for primary- WQI. The results predicted gradual degradation of WQI over time for initial conditions (with current SWP strategies) and improvement of WQI for the final conditions (with proposed SWP strategies). The yearly WQI improvement for primary, nutrient, microbial, organic and toxic substance WQI along with overall WQI are presented in Table 4.7. Table 4.7 shows WQI improvement of highest 8.09%, 16.84%, 12.40%, 3.63%, 0.97% and 5.39% for primary, nutrient, microbial, organic, toxic substance and overall WQI respectively. Therefore, the selected SWP strategies are more effective for controlling nutrient pollutant than microbial, primary, organic and toxic substance respectively.  Primary WQI  100  90  Initial prediction Final prediction  80 Base line 2008  2009  2010  2011  2012  2013  2014  2015  2016  2017  2018  Figure 4.17: Primary WQI prediction (FYs 2008-18) 4.3.3 Sensitivity analysis A sensitivity analysis to assess the robustness of the model was performed. Sensitivity analysis is the process of estimating the degree to which the output of a model changes as values of input factors are changed. It can help to 1) identify the input factors that have the most influence on model output, 2) identify the processes which have the greatest influence on model output, and 3) quantify the change in output caused by uncertainty and variability in the values of input factors. Various methods exist for sensitivity analysis. These methods include the scatter plot, partial and rank correlation coefficients, multivariate regression, and contribution to variance and probabilistic sensitivity analysis. In this case, sensitivity analysis was performed for 97  different land uses and for determining most important SWP efficiency parameter. Land use sensitivity analysis was done by assuming specific land use scenarios while SWP efficiency part was done with spearman rank correlation. Land use sensitivity will help decision-makers to take adequate actions to protect source water in case of land use variation, specially an increase of urbanisation around source water as it is the tendency around major cities in Canada. Table 4.7: Percentage of WQI improvement for five WQI group and for overall WQI (FYs 2008-18) Year  Primary- WQI  Nutrient- WQI  Microbial- WQI  Organic- WQI  Toxic substanceWQI  WQI  2008  1.23  1.98  1.59  0.44  0.11  0.63  2009  2.37  3.80  3.01  0.83  0.21  1.25  2010  3.52  5.58  4.38  1.22  0.31  1.94  2011  4.58  7.21  5.60  1.57  0.40  2.58  2012  5.12  8.76  6.74  1.91  0.49  3.03  2013  5.70  10.39  7.92  2.26  0.58  3.49  2014  6.21  11.79  8.92  2.56  0.67  3.90  2015  6.71  13.16  9.88  2.85  0.75  4.30  2016  7.18  14.43  10.76  3.12  0.82  4.67  2017  7.65  15.68  11.62  3.39  0.90  5.04  2018  8.09  16.84  12.40  3.63  0.97  5.39  Land use impact has been analyzed by comparing the present land use conditions with five fictitious scenarios: fully agricultural, pasture, livestock, forests and urban type 1 land use. Model has been simulated to assess WQI from 2008-2018 for the six scenarios with the same initial condition (present land use and SWP strategies for 2007) and without any proposed SWP strategies. Figure 4.18 shows that the water quality will be the worst in case of „urban type 1‟ land use. For 100% agricultural land use, the WQI will deteriorate but still better than urban type 1‟. If all the land is used for pasture or livestock, the WQI will be better compared to the baseline WQI. As expected, changing to forest land use will also lead to improved water quality in Page Creek. The WQI values for forest land use showed a slow improvement as flow rate of the creek water is very high compared to the incoming pollutant flow rates. Figure 4.19 showed that an implementation of the proposed SWP strategies can improve the water quality; even if all the land use is only „urban 1‟. The base line in Figure 4.19 refers to the case study area with proposed SWP strategies.  98  A sensitivity analysis was performed based on pasture and livestock land use variations with and without proposed SWP strategies. The impacts of livestock land use are provided Figures 4.20 and 4.21. Similarly, the impacts of Pasture are shown in Figures 4.22 and 4.23. A simulation-based sensitivity analysis is performed with spearman rank correlation for PCAP, storm water management ponds and VFS efficiency models to identify important input factors in case of each of these SWP strategies. Figure 4.24 shows the tornado graphs for SWP strategies including PCAP, storm water management and VFS. These graphs highlight important design parameters that significantly affect the efficiency of these SWP strategies. Figure 4.24 shows that retention time (RT) is the most important parameter for the storm water management efficiency, followed by media depth, vegetation level and water depth. For VFS efficiency (Figure 4.24), the manning‟s n is the most important parameter. It is important to remember that a storm water pond is very different from a VFS. The width of a VFS is the second most important parameter followed by the slope. The percentage of parameter contribution are presented in Table 4.8.  Figure 4.18: Impact of land use on WQI without SWP strategies  99  Final WQI  Figure 4.19: Impact of land use on WQI with SWP strategies 100 95 90 85  WQI  80 75 70 65 60 55 50  Base line  20% Livestock  50% Livestock  100% Livestock  Figure 4.20: Impact of livestock land use on water quality for initial condition  100  Figure 4.21: Impact of livestock land use on water quality with proposed SWP strategies  Figure 4.22: Impact of pasture land use on water quality for initial condition  101  Figure 4.23: Impact of pasture land use on water quality with proposed SWP strategies  Table 4.8: Percent parameter contribution for SWM pond, VFS and PCAP efficiency SWP strategy  Parameter percent contribution  SWM pond  Water depth- 0.63 Retention time- 75.71 Vegetation level- 4.73 Media depth- 18.93  VFS  Manning‟s n- 50.81 VFS width- 42.29 Slope- -6.90  PCAP  Cover crop- 45.67 Crop rotation- 47.21 Soil water content- 1.53 Dry bulk density- -5.59  102  Figure 4.24: Tornado graphs describing the contribution of various design parameters in the removal efficiency of selected SWP strategies a) PCAP, b) SWM pond, and c) VFS 4.3.4 Comparison of proposed WQI with NSF-WQI For the monitored data, the proposed overall WQI results are compared with NSF-WQI (Brown et al., 1970). The NSF-WQI values are obtained from online calculator6 for TSS, phosphorous, FC and BOD5. The same, parameters were also used in the proposed model to estimate overall WQI using steps 4, 5 and 6 (Figure 3.1) only. Figures 4.25 compares monthly WQI variations estimated using proposed WQI and the NSF-WQI. The pattern is very similar in both assessment models, as reflected by high R2 value in Figure 4.26, where the proposed WQI 6  103  values are plotted against NSF-WQI. The proposed WQI is predicting worse surface water quality compared to NSF-WQI, as more weight is given to parameters in the proposed model, that govern drinking water quality whereas the NSI-WQI distributes the weighting more evenly  WQI  to all parameters. 90 80 70 60 50 40 30 20 10 0  Predicted model NSF-WQI  Figure 4.25: Monthly variations predicted by proposed WQI and NSF-WQI  90 80 70  WQI  60 50 40  R² = 0.8883  30 20 10 0 40  50  60  70  80  90  100  NSF-WQI  Figure 4.26: NSF-WQI vs. proposed WQI 104  5. CONCLUSIONS AND RECOMMENDATIONS  5.1 Summary and Conclusions Source Water Protection (SWP) is an important step in a multi-barrier approach that ensures safe drinking water to the consumers in a cost effective manner. The management of source water is a complex undertaking and requires tools to help decision makers at different levels of the governments (Federal/ Provincial/ Municipal) to coordinate and make informed decisions. Most of the existing models are exceedingly complex, data intensive, and found not be useful for decision makers with limited or no technical expertise. Moreover these models mainly focus on specific water quality issues (e.g., microbial, nutrients, aesthetics) and don‟t provide an assessment of water quality, which is a key for making resource allocation and capital investment decisions by the related agencies. A comprehensive literature review has been performed on various aspects of this interdisciplinary research. The main focus of the review is on source water protection strategies that includes agricultural practices (e.g., cover crop, crop rotation), storm water management ponds (e.g., dry detention ponds, storm water wetlands, bioretention areas, wet detention ponds, sand filters and grass swales), fencing and vegetated filter strips have been provided in the context of possible pollutant load reduction. The proposed model based on a conceptual framework links selected SWP strategies to the improvement in water quality index (WQI), a surrogate for water quality. The analysis based on developed model is performed in six steps. The first step estimates the pollutant loads based on land use export coefficient approach. In case of pasture and livestock land use, the numbers of animals are used to estimate the pollutant load for each parameter. The second step estimates the percentage of pollutant reduction based on removal efficiency of the selected SWP strategies using fuzzy rule-based models. Step 3 estimates pollutant concentration using steady-stead mass balance approach over pre-defined time duration. Step 4 transforms the calculated concentrations of the water quality parameters into unit less sub-indices. In Step 5, the water quality parameters are classified as Primary, Microbial, Organic, Nutrient and toxic substances water quality indices. Finally in Step 6, an overall WQI is derived from these water quality groups using fuzzy measures theorem. The WQI provides water quality assessment of a 105  receiving water body using a scale  [0 100], where 0 refers to the worst and 100 refers to the best quality water. The developed model has been applied to a case study of Page Creek water in Clayburn watershed (BC, Canada). The Page Creek sub-watershed is mainly an agricultural and farming area. For summer months, the estimated WQI based on the monitored data suggested that quality ranges from good to very good. Contrarily, during winter months it ranges from marginal to very good. Monthly and yearly analyses are carried out using different scenarios of changing land use with implementation of SWP strategies. An improvement in the water quality of Page Creek can be attributed to specific arrangement of SWP strategies (as proposed in this study). Monthly analysis showed improvements in Page Creek water quality for both summer and winter months (very good). PCAP and vegetated filter strip are found to be more effective for removal of TSS, while storm water management ponds are more effective for the removal of both TSS and coliforms (microbial indicators). Yearly analysis (FYs 2008-18) predicts water quality degradation (from very good to good) if no actions are taken, however, if the proposed SWP are implemented, the water quality will either improve or remain in very good condition. The proposed SWP strategies improve source water quality approximately by 8.1%, 16.8%, 12.4%, 3.6%, 1% and 5.4% for primary, nutrient, microbial, organic, toxic substance and overall WQI respectively. A sensitivity analyses shows that if the land use is of „urban type 1‟ (i.e., highway/ roads) and/or „agricultural‟, the adverse impact on water quality will be paramount. However, land uses such as pasture or livestock can also adversely affect the water quality, especially if no precautionary measures are taken. As expected, the use of forests found to be the best land use with minimal impact on source water quality. The key parameters that contributed to the removal efficiency of SWM pond, VFS and PCAP efficiency are retention time (~76%), manning‟s n (~51%) and crop rotation (~47 %), respectively  5.2 Limitations and Recommendations The main thrust of this research is to develop a simple model that require basic data and can be adapted by decision-makers and engineers as a decision support tool. As data availability is a major issue in planning and management of watersheds, simple assumptions (such as steady state mass balance, export coefficient, water quality indices) have been found useful to predict water 106  quality improvement based on selected SWP strategies. Proposed model has following limitations:   Numerous studies are investigated from the literature to estimate the pollutant loads from a specific land use and estimate the efficiency of SWP strategies. However, these studies vary significantly in terms of their experimental setups and ambient conditions.    For monthly analysis, the average yearly EC (kg/ha/yr) is used by dividing with twelve. This implies that for each month the EC remains the same however there can be significant variations due to land use (e.g., impacts of agricultural practices, antecedent conditions and climatic and weather variations).    In the case of unavailability of EC values, the model uses pollutant concentrations based on values reported in the literature. Extrapolation from these concentrations can lead to uncertainties in the estimates of pollutant loads.    For PCAP efficiency only cover crop and crop rotation are incorporated as effective agricultural practices. However, conservative tillage can also be included.    Many assumptions are made for missing data in case of Page Creek case study. For example, the number and type of livestock animal are assumed due to unavailability of information.    The focus of this research was on pollutants, which negatively affect water quality in the receiving water body. However, the water quality is also a function of factors like dissolved oxygen, which is not included in this study.    Factors related to snow melting temperature, snow water equivalent, antecedent ground conditions, saturation and porosity of soil, etc. are not studied in this research.  Based on this research, we recommend the following for future research:   Develop a database for monthly EC values for different pollutants and different types of land uses; however site specific EC values are preferable where possible    Perform uncertainty analysis using probabilistic or fuzzy-based approaches    Perform cost-benefit analysis for the selected SWP strategies    Integrate the model with a GIS platform to make recommendations for potential land use in a specific watershed  107    Verify the model predictions for a case study that has a long-term data available to evaluate the performance.    Modify and apply the model for other possible land uses (e.g., mining) using relevant pollutant export coefficients    Validate removal efficiency for different SWP strategies as function of design parameters using laboratory or pilot studies.  108  REFERENCES Abbotsford Agricultural Strategy, 2009. Agricultural profile. Agricultural advisory committee, City of Abbotsford, South Abbotsford, BC, Canada. ABS (Australian Bureau of Statistics), 2000. Year Book Australia. Alberta Forest Conservation Strategy, 1997. Alberta forest conservation strategy. Alberta Environmental Protection, Edmonton, Alberta, Canada. Amatya, D.M., Chescheir, G.M., Fernandez, G.P., Skaggs, R.W. and Gilliam, J.W., 2004. DRAINWAT-based methods for estimating nitrogen transport in poorly drained watersheds. Transactions of the ASAE 47(3), 677-687. Antoniadis, A., Takavakoglou, V., Zalidis, G. and Poulios, I., 2007. Development and evaluation of an alternative method for municipal wastewater treatment using homogeneous photocatalysis and constructed wetlands. Catalysis Today 124(3-4), 260-265. Arora, K., Mickelson, S.K. and Baker, J.L., 2003. Effectiveness of vegetated buffer strips in reducing pesticide transport in simulated runoff. Transactions of the American Society of Agricultural Engineers 46(3), 635-644. ASAE, 2003. Manure Production and Characteristics. American Society of Agricultural Engineers. ASAE, D384.1 FEB03. ASAE-The Society for engineering in agricultural, food, and biological systems, St. Joseph, USA. Atech Group, 2000. Cost of algal blooms. Land and Water Resources Research and Development Corporation, and Murray-Darling Basin Commission 8, 24–28. Australian Government, 2004. Australian Drinking Water Guidelines. Australia Government. Available at: Accessed June 2009. Badar, B. and Romshoo, S. A., 2008. Assessing the pollutant load of Dal Lake using Geospatial tools. The 12th World Lake conference. Rajasthan, India, October 2007. Banerjee, T. and Srivastava, R.K., 2009. Application of water quality index for assessment of surface water quality surrounding integrated industrial estate-Pantnagar. Water Science and Technology 60(8), 2041-2053. Barrett, M.E. and Shaw, C.B., 2007. Benefits of porous asphalt overlay on storm water quality. Transportation Research Record (2025), 127-134. 109  Barstead, D., 2004. Baseline Document Maclure Road Wetlands. The Abbotsford Land Trust Society. The Land Trust Alliance of BC, Canada. Beaulac, M.N. and Reckhow, K.H., 1982. Examination of land use - nutrient export relationships. Water Resources Bulletin 18(6), 1013-1024. Bechmann, M., Stalnacke, P., Kvrn, S., Eggestad, H.O. and Oygarden, L., 2009. Integrated tool for risk assessment in agricultural management of soil erosion and losses of phosphorus and nitrogen. Science of the Total Environment 407(2), 749-759. Bewsell, D., Monaghan, R.M. and Kaine, G., 2007. Adoption of stream fencing among dairy farmers in four New Zealand catchments. Environnemental Management 40(2), 201-209. Bhattarai, R., Kalita, P.K. and Patel, M.K., 2009. Nutrient transport through a vegetative filter strip with subsurface drainage. Journal of Environmental Management 90(5), 1868-1876. Bingner, R. L. and Theurer, F.D., 2001. AGNPS 98; a suite of water quality models for watershed use. Proceedings - Federal Interagency Sedimentation Conference 2, VII1VII8, Reno, Nevada, March 2001. Bjorneberg, D.L., Trout, T.J., Sojka, R.E. and Aase, J.K., 1999. Evaluating WEPP-predicted infiltration, runoff, and soil erosion for furrow irrigation. Transactions of the American Society of Agricultural Engineers 42(6), 1733-1741. Bliss, D.J., Neufeld, R.D. and Ries, R.J., 2009. Storm water runoff mitigation using a green roof. Environmental Engineering Science 26(2), 407-417. Boller, M., Langbein, S. and Steiner, M., 2007. Characterization of road runoff and innovative treatment technologies. Highway and Urban Environment 12, 441-452. Bonn Charter of Safe Drinking Water, 2004. Bonn Charter for Safe Drinking Water. International Water Association. Borah, D.K., Xia, R. and Bera, M., 2002. DWSM; a dynamic watershed simulation model. In Mathematical Models of Small Watershed Hydrology and Applications. Ed. By Singh, V.P. and Frevert, D.. Water Resources Publications, Highlands Ranch, CO, United States. Bordalo, A.A., Teixeira, R. and Wiebe, W.J., 2006. A water quality index applied to an international shared river basin: The case of the Douro River. Environmental Management 38(6), 910-920. Borin, M., Passoni, M., Thiene, M. and Tempesta, T., 2010. Multiple functions of buffer strips in farming areas. European Journal of Agronomy 32(1), 103-111. 110  Bottcher, A.B., Hiscock, J, G., Pickering, N.B. and Jacobsen, B.M., 1998. WAM: Watershed assessment model for agricultural and urban landscapes. American Society for Agricultural Engineers’ Seventh International Conference of Computers in Agriculture, Orlando, FL, October 1998. Bourne, A., Armstrong, N. and Jones, G., 2002. A preliminary estimate of total nitrogen and total phosphorus loading to streams in Manitoba, Canada. Manitoba Conservation Report No.- 04. Bowden, K. and Brown, S.R., 1984. Relating effluent control parameters to river quality objectives using a generalized catchment simulation-model. Water Science and Technology 16(5-7), 197-206. Boyacioglu, H., 2007. Development of a water quality index based on a European classification scheme. Water Sa 33(1), 101-106. Brown, D. J., 2009. The effects of non-point source pollution on surface water quality, Clayburn Watershed, Abbotsford, British Columbia, Canada. MASc thesis, Royal Roads University, Victoria, BC. Brown, R.M., McClelland, N.I., Deininger, R.A. and Tozer, R.G., 1970. A water quality index: do we dare? Water Sewage Works 117, 339–343. Can. Gov, 2001. Facts on Canada, Canada’s Environment, Catalogue No. PF3-2/36-2001. Available at: Accessed June 2009. Carreiro-Silva, M., McClanahan, T.R. and Kiene, W.E., 2009. Effects of inorganic nutrients and organic matter on microbial euendolithic community composition and microbioerosion rates. Marine Ecology Progress Series 392, 1-15. Chang, N.B., Chen, H.W. and Ning, S.K., 2001. Identification of river water quality using the fuzzy synthetic evaluation approach. Journal of Environmental Management 63(3), 293305. Chen, C.W., Herr, J. and Tsai, W., 2003. Enhancement of WARMF to Track Mercury Species in a River Basin from Atmospheric Depositions to Fish Tissues. Publication No.1005470. Electric Power Research Institute, Palo Alto, CA. Chen, D., Lu, J. and Shen, Y., 2010. Artificial neural network modeling of concentrations of nitrogen, phosphorus and dissolved oxygen in a non-point source polluted river in Zhejiang Province, southeast China. Hydrological Processes 24, 290–299. 111  Chiew, F.H.S. and McMahon, T.A., 1999. Modelling runoff and diffuse pollution loads in urban areas. Water Science and Technology 39(12), 241-248. Choquet, G., 1953. Theory of capacities. Annales de l’Institut Fourier 5, 131–295. Christensen, R. and Parfitt. B., 2003.Watered Down: A Report on Waterborne Disease Outbreaks in British Columbia 1980-2002. Sierra Legal Defence Fund, Toronto, Ontario, Canada. City of Abbotsford, 2004. Agricultural Land Use Inventory, 2003. Ministry of Agriculture, Food and Fisheries, BC, Canada. Clausen, J.C., Guillard, K., Sigmund, C.M. and Dors, K.M., 2000. Ecosystem restoration - water quality changes from riparian buffer restoration in Connecticut. Journal of Environmental Quality 29(6), 1751-1761. Clean Water Act., 2006a. Ontario regulation - source protection committees. Regulation 288, Canadian Legal Information Institute. Available at: Accessed June 2009. Clean Water Act., 2006b. Ontario regulation – general. Regulation 287. Canadian Legal Information Institute. Available at: Accessed June 2009. Conservation Authorities Act., 2006. Hamilton region conservation authority: regulation of development, interference with wetlands and alterations to shorelines and water courses. Ontario regulation 161. Available at: Accessed June 2009. Conservation Land Act., 1990. Ontario regulation- conservation land act. C- 28. Available at: Accessed June 2009. Council Directive, 1991. Council directive on concerning the placing of plant protection products on the market. Official Journal of the European Union, 91/414/EEC. Covelli, P., Marsili-Libelli, S. and Pacini, G., 2002. SWAMP: A two-dimensional hydrodynamic and quality modeling platform for shallow waters. Numerical Methods for Partial Differential Equations 18(5), 663-687.  112  Cox, B. A., 2003. A review of currently available in-stream water-quality models and their applicability for simulating dissolved oxygen in lowland rivers. Science of the Total Environment 314, 335-377. Crown Forest Sustainability Act., 1994. Ontario regulation - crown forest sustainability act. C25. Available at: Accessed June 2009. Crown Timber Act., 1990. Ontario regulation -crown timber act. C- 51. Available at: Accessed June 2009. Cude, C. G., 2001. Oregon water quality index: A tool for evaluating water quality management effectiveness. Journal of the American Water Resources Association 37(1), 125-137. Curriero, F.C., Patz, J.A., Rose, J.B., and Lele, S., 2001. The association between extreme precipitation and waterborne disease outbreaks in the United States, 1948-1994. American Journal of Public Health 91(8), 1194-1199. Dahiya, S., Singh, B., Gaur, S., Garg, V.K. and Kushwaha, H.S., 2007. Analysis of groundwater quality using fuzzy synthetic evaluation. Journal of Hazardous Materials 147(3), 938946. Dangerous Substance Directive, 2006. Council Directive on pollution caused by certain dangerous substances discharged into the aquatic environment of the Community. Official Journal of the European Union, 2006/11/EC. Available at: Accessed June 2009. Davies, J.M. and Mazumder, A., 2003. Health and environmental policy issues in Canada: The role of watershed management in sustaining clean drinking water quality at surface sources. Journal of Environmental Management 68(3), 273-286. Davis, A. P., 2007. Field performance of bioretention water quality. Environmental Engineering Science 24(8), 1048-1064. Davis, B.S. and Birch, G.F., 2009. Catchment-wide assessment of the cost-effectiveness of stormwater remediation measures in urban areas. Environmental Science & Policy 12(1), 84-91. 113  Devito, K.J., Creed, I.F., Rothwell, R.L. and Prepas, E.E., 2000. Landscape controls on phosphorus loading to boreal lakes: Implications for the potential impacts of forest harvesting. Canadian Journal of Fisheries and Aquatic Sciences 57(10), 1977-1984. DHI Water & Environment. MIKE 11, 2005. A modeling system for rivers and channels—user guide and technical reference manual. Horsholm, Denmark. Dietz, M.E., 2007. Low impact development practices: A review of current research and recommendations for future directions. Water Air and Soil Pollution 186(1-4), 351-363. Dillaha, T.A., Reneau, R.B., Mostaghimi, S. and Lee, D., 1989. Vegetative filter strips for agricultural nonpoint source pollution-control. Transactions of the ASAE 32(2), 513-519. Ding, X., Shen, Z., Hong, Q., Yang, Z., Wu, X. and Liu, R., 2010. Development and test of the export coefficient model in the upper reach of the Yangtze River. Journal of Hydrology 383(3-4), 233-244. Dodd, R.C., McMahon, G. and Stichter, S., 1992. Watershed planning in the Albemarle-Pamlico estuarine system: Report 1- annual average nutrient budgets. U.S. Environmental Protection Agency, Center for Environmental Analysis, Report 92-10, Raleigh, NC. Dojlido, J., Raniszewski, J. and Woyciechowska, J., 1994. Water quality index - application for rivers in Vistula river basin in Poland. Water Science and Technology 30(10), 57-64. Dosskey, M. G., 2001. Toward quantifying water pollution abatement in response to installing buffers on crop land. Environmental Management 28(5), 577-598. Dosskey, M.G., Eisenhauer, D.E. and Helmers, M.J., 2005. Establishing conservation buffers using precision information. Journal of Soil and Water Conservation 60(6), 349-354. Downer, C.W. and Ogden, F.L., 2002. GSSHA-User’s Manual, Gridded Surface Subsurface Hydrologic Analysis Version 1.43 for WMS 6.1. ERDC Technical Report, Engineer Research and Development Center, Vicksburg, Mississippi. Drinking Water Directive, 1980. Council directive- relating to the quality of water intended for human consumption. Official Journal of the European Union, 80/778/EEC. Available at: 0778EEC.pdf. Accessed June 2009. Drinking Water Protection Act., 2001. BC regulation- drinking water protection act, C- 9. Available at:  114 Accessed June 2009. Dubois, D. and Prade, H., 1980. Fuzzy sets and systems. Academic Press, New York. Duchemin, M. and Hogue, R., 2009. Reduction in agricultural non-point source pollution in the first year following establishment of an integrated grass/tree filter strip system in southern Quebec (Canada). Agriculture, Ecosystems & Environment 131, 85-97. Dunnette, D.A., 1979. A geographically viable water quality index used in Oregon. J. Water Pollut. Control Fed. 51, 53–61. Easton, Z.M., Walter, M.T., Schneiderman, E.M., Zion, M.S. and Steenhuis, T.S., 2009. Including source-specific phosphorus mobility in a nonpoint source pollution model for agricultural watersheds. Journal of Environmental Engineering 135(1), 25-35. Ecological Reserve Act., 1975. BC regulation- ecological reserve regulations. Regulation 335. Queen‟s Printer for British Columbia. Available at: Accessed May 2009. Elrashidi, M.A., Hammer, D., Fares, A., Seybold, C.A., Ferguson, R. and Peaslee, S.D., 2007. Loss of heavy metals by runoff from agricultural watersheds. Soil Science 172 (11), 876894. EMRL (Environmental Modeling Research Laboratory), 1998. Watershed modeling system (WMS) reference manual and tutorial. Brigham Young University, Provo, Utah. Environment and Land Use Act., 1996. BC regulation- Environment and Land Use Act. C- 117. Canadian Legal Information Institute. Available at: Accessed June 2009. Environment Canada, 2001. Threats to Sources of Drinking Water and Aquatic Ecosystem Health in Canada. National Water Research Institute, Burlington, Ontario. Environment Canada, 2004. Canadian Climate Normals or Averages 1971-2000. Government of Canada. Environmental Assessment Act., 1996. Ontario regulation- Environmental Assessment Act. Regulation 94.  115  Environmental Management Act., 2004a. BC regulation- Environmental Management Act warranty regulation. Regulation 317. Environmental Management Act., 2004b. BC regulation- motor vehicle emissions control warranty regulation. Regulation 321. Canadian Legal Information Institute Web site: Accessed June 2009. Environmental Management Act., 2007. BC regulation- code of practice for the slaughter and poultry processing industries. Regulation 246. Canadian Legal Information Institute. Available at: Accessed June 2009. Environmental Management Act., 2008a. BC regulation - waste discharge regulation. Regulation 377. Canadian Legal Information Institute. Available at: Accessed June 2009. Environmental Management Act., 2008b. BC regulation- code of practice for the discharge of produced water from coalbed gas operations. Regulation 364. Canadian Legal Information Institute. Available at: Accessed June 2009. Environmental Protection Act., 1990. Govt. of Ontario. General – Waste Management. Regulation 347. Available at: Accessed June 2009. Environmental Protection Act., 1994. Ontario regulation - Recycling and composting of municipal waste. Regulation 101. Available at: Accessed June 2009. Environmental Protection Act., 1998a. Ontario regulation -landfilling sites. Regulation 232. Available at: Accessed June 2009. Environmental Protection Act., 1998b. Ontario regulation - Classification and exemption of spills and reporting of discharges. Regulation 675. Available at: Accessed June 2009. 116  Environmental Protection Act., 2007. Ontario regulation- Environmental Protection Act. Spill prevention and contingency plans. Regulation 224. Available at: Accessed June 2009. Environmental Protection and Enhancement Act., 1993a. Alberta regulation- Potable Water. Regulation 122. Canadian Legal Information Institute. Available at: Accessed June 2009. Environmental Protection and Enhancement Act., 1993b. Alberta regulation- wastewater and storm drainage regulation. Regulation 119. Canadian Legal Information Institute. Available at: Accessed June 2009. Environmental Protection and Enhancement Act., 1996a. Environmental code of practice for landfills. Park Plaza 10611 - 98 Avenue Edmonton, Alberta. Environmental Protection and Enhancement Act., 1996b. Alberta regulation -waste control regulation. Regulation 192. Available at: Accessed June 2009. Environmental Protection and Enhancement Act., 2000. Alberta regulation- environmental protection and enhancement act. C- E- 12. Available at: Accessed June 2009. Environmental Protection and Enhancement Act., 2003. Alberta Regulation- Potable water regulation. Regulation- 277. Available at: Accessed June 2009. Environmental Protection and Enhancement Act., 2006. Alberta regulation- environmental code of practice for forage drying facilities. C- 12. Park Plaza 10611 - 98 Avenue Edmonton, Alberta. Ernst, C., 2004. Protecting the Source, Land Conservation and the Future of American’s Drinking Water. Available at: Accessed June 2009.  117  Ernst, C., Gullick, R. and Nixon, K., 2004. Opflow, Protecting the Source, Conserving Forests to Protect Water. Available at: . Accessed June 2009. Field Estimation of Soil Water Content, 2008. A Practical Guide to Methods, Instrumentation and Sensor Technology. Training course series no. 30. International Atomic Energy Agency, Vienna. Forest and Range Practices Act., 2002. BC regulation - forest and range practices act. C- 69. Available at: Accessed June 2009. Forest and Range Practices Act., 2004. Forest planning and practices regulation. B. C. regulation 14. Available at: Accessed June 2009. Forest and Range Practices Act., 2009. Forest planning and practices regulation. B.C. regulation 106. Forestry Act., 1990. Ontario regulation- forestry act. C- F- 26. Available at: Accessed June 2009. Francisque, A., Rodriguez, M.J., Sadiq, R., Miranda, L.F. and Proulx, F., 2009. Prioritizing monitoring locations in a water distribution network: A fuzzy risk approach. Journal of Water Supply: Research and Technology- AQUA 58(7), 488-509. Franti, T.G., 1997. Vegetative filter strips for agriculture. Nebraska Cooperative Extension NF 97-352. Available at: Accessed May 2009. Fraser Salmon and Watersheds Program, 2009. Prioritization of and rehabilitation considerations for fish migration impediments in lower Fraser river. Fraser Salmon and Watersheds Program, Vancouver, BC. Available at: Accessed June 2009. Fujita, S., 1994. Infiltration structures in Tokyo. Water Science and Technology 30(1), 33-41.  118  Garcia, M. W., 2004. New York City‟s approach to watershed management. Watershed Management and Operations Management 2000. Fort Collins, CO, United states, June 2000. Gotvajn, A.Z. and Zagorc-Koncan, J., 2009. Bioremediation of highway stormwater runoff. Desalination 248(1-3), 794-802. Grabisch, M., 1996. The application of fuzzy integrals in multicriteria decision making. European Journal of Operation Research 89(3), 445-456. Grace III, J. M., 2002. Effectiveness of vegetation in erosion control from forest road sideslope. Transactions of the American Society of Agricultural Engineers 45(3), 681-685. Greenbelt Act., 2005. Ontario regulation- greenbelt act. C- 1. Available at: Accessed June 2009. Hacisalihoglu, S., Mert, A., Negiz, M.G. and Muys, B., 2010. Soil loss prediction using universal soil loss equation (USLE) simulation model in a mountainous area in Aglasun district, Turkey. African Journal of Biotechnology 9 (24), 3589-3594 Hajkowicz, S. and Collins, K. 2007. A review of multiple criteria analysis for water resource planning and management. Water Resources Management 21(9), 1553-1566. Hathaway, J.M., Hunt, W.F. and Jadlocki, S., 2009. Indicator bacteria removal in storm-water best management practices in charlotte, North Carolina. Journal of Environmental Engineering 135(12), 1275-1285. Herring, J.P., Schultz, R.C. and Isenhart, T.M., 2006. Watershed scale inventory of existing riparian buffers in northeast Missouri using GIS. Journal of the American Water Resources Association 42(1) 145-155. Hickey, C.W. and Gibbs, M.M., 2009. Lake sediment phosphorus release management-decision support and risk assessment framework. New Zealand Journal of Marine and Freshwater Research 43(3), 819-856. Howard, H., 2002. Human Health and Heavy Metals Exposure, C- 4, Life Support: The Environment and Human Health, Michael McCally (Ed.), MIT press. Hrudey, S.E., Payment, P., Huck, P.M., Gillham, R.W. and Hrudey, E.J., 2003. A fatal waterborne disease epidemic in Walkerton, Ontario: Comparison with other waterborne outbreaks in the developed world. Water Science and Technology 47(3), 7-14.  119  Hsieh, C. and Davis, A.P., 2005. Evaluation and optimization of bioretention media for treatment of urban storm water runoff. Journal of Environmental Engineering 131(11), 1521-1531. Hsieh, C., Davis, A.P. and Needelman, B.A., 2007a. Nitrogen removal from urban stormwater runoff through layered bioretention columns. Water Environment Research 79(12), 24042411. Hsieh, C., Davis, A.P. and Needelman, B.A., 2007b. Bioretention column studies of phosphorus removal from urban stormwater runoff. Water Environment Research 79(2), 177-184. HSPF (Hydrological simulation program-fortran), 1985. Continuous simulation of surface and subsurface flows in cypress creek basin, Florida, using hydrological simulation programfortran. In Hicks C. N. (Ed.), University of Florida, Water Resources Research Center, Department of Geology, Gainesville, FL, United States. Hulebak, K.L. and Schlosser, W., 2002. Hazard Analysis and Critical Control Point (HACCP) history and conceptual overview. Risk Analysis 22 (3), 547-552. Hunt, W.F., Jarrett, A.R., Smith, J.T. and Sharkey, L.J., 2006. Evaluating bioretention hydrology and nutrient removal at three field sites in North Carolina. Journal of Irrigation and Drainage Engineering 132(6), 600-608. Hunt, W.F., Smith, J.T., Jadlocki, S.J., Hathaway, J.M., Eubanks, P.R., 2008. Pollutant Removal and Peak Flow Mitigation by a Bioretention Cell in Urban Charlotte, N.C. Journal of Environmental Engineering 134 (5), 403-408. Hyland, R., Byrne, J., Selinger, B., Graham, T., Thomas, J., Townshend, I. and Gannon, V., 2003. Spatial and temporal distribution of fecal indicator bacteria within the Oldman river basin of southern Alberta, Canada. Water Quality Research Journal of Canada 38(1), 15-32. Icaga, Y., 2007. Fuzzy evaluation of water quality classification. Ecological Indicators 7(3), 710-718. Integrated Pest Management Act., 2004. BC regulation-Integrated pest management. Regulation 604. Available at: Accessed June 2009. IPCC, 2007. Climate Change 2007: Synthesis Report. Available at:  120  Irvine, K.N., Loganathan, B.G., Pratt, E.J. and Sikka, H.C., 1993. Calibration of PCSWMM to estimate metals, PCBs and HCB in CSOs from an industrial sewershed. In New Techniques for Modelling of Stormwater Quality Impacts. Center for Environmental Research & Education State Univ. College of Buffalo, NY, USA. Ivey, J.L., de Loe, R., Kreutzwiser, R. and Ferreyra, C., 2006. An institutional perspective on local capacity for source water protection. Geoforum 37(6), 944-957. Jaramillo, M., Gonzalez, D.E., Nunez, M.E., and Portilla, G., 2009. Integrated air quality index for Colombian cities. Revista Facultad De Ingenieria-Universidad De Antioquia 48, 97106. Jha, R., Ojha, C.S.P. and Bhatia, K.K.S., 2007. Non-point source pollution estimation using a modified approach. Hydrological Processes 21(8), 1098-1105. Jonnalagadda, S.B. and Mhere, G., 2001. Water quality of the Odzi river in the eastern highlands of zimbabwe. Water Research 35(10), 2371-2376. Juahir, H., Zain, S.M., Yusoff, M.K., Hanidza, T.I.T., Armi, A.S.M., Toriman, M.E. and Mokhtar, M., 2010. Spatial water quality assessment of Langat river basin (Malaysia) using environmetric techniques. Environmental Monitoring and Assessment 1-17. Kasih, G.A.A. and Kitada, T., 2004. Numerical simulation of water quality response to nutrient loading and sediment resuspension in Mikawa bay, central Japan: Quantitative evaluation of the effects of nutrient-reduction measures on algal blooms. Hydrological Processes 18(16), 3037-3059. Kay, D., Crowther, J., Stapleton, C.M., Wyer, M.D., Fewtrell, L., Anthony, S., Bradford, M., Edwards, A., Francis, C.A., Hopkins, M., Kay, C., McDonald, A.T., Watkins, J. and Wilkinson, J., 2008. Faecal indicator organism concentrations and catchment export coefficients in the UK. Water Research 42(10-11), 2649-2661. Kemper, D., Dabney, S., Kramer, L., Dominick, D. and Keep, T., 1992. Hedging against erosion. Journal of Soil & Water Conservation 47(4), 284-288. Kennedy, G. and Mayer, T., 2002. Natural and constructed wetlands in Canada; an overview. Water Quality Research Journal of Canada 37(2), 295-325. Khan, A.A., Paterson, R. and Khan, H., 2004. Modification and application of the Canadian council of ministers of the environment water quality index (CCME WQI) for the  121  communication of drinking water quality data in Newfoundland and Labrador. Water Quality Research Journal of Canada 39(3), 285-293. Kim, H., Seagren, E.A. and Davis, A.P., 2003. Engineered bioretention for removal of nitrate from stormwater runoff. Water Environment Research 75(4), 355-367. Kleinman, P.J.A., Salon, P., Sharpley, A.N. and Saporito, L.S., 2005. Effect of cover crops established at time of corn planting on phosphorus runoff from soils before and after dairy manure application. Journal of Soil and Water Conservation 60(6), 311-322. Kolodziej, E.P. and Sedlak, D.L., 2007. Rangeland grazing as a source of steroid hormones to surface waters. Environmental Science and Technology 41(10), 3514-3520. Kosko, B., 1994. Fuzzy Thinking. Harper Collins. Kourenti, C., Karanis, P. and Smith, H., 2007. Waterborne transmission of protozoan parasites: A worldwide review of outbreaks and lessons learnt. Journal of Water and Health 5(1), 1-38. Kramer, M.H., Quade, G., Hartemann, P. and Exner, M., 2001. Waterborne diseases in Europe 1986-96. Journal of American Water Works Association 93(1), 48-53. Krewski, D., Balbus, J., Butler-Jones, D., Haas, C.N., Isaac-Renton, J., Roberts, K.J. and Sinclair, M., 2004. Managing the microbiological risks of drinking water. Journal of Toxicology and Environmental Health 67 (20-22), 1591–1617. Kumar, D. and Alappat, B.J., 2009. NSF-water quality index: Does it represent the experts' opinion? Practice Periodical of Hazardous, Toxic, and Radioactive Waste Management 13(1), 75-79. Kung, H.T., Ying, L.G., & Liu, Y.C., 1992. A complementary tool to water-quality index - fuzzy clustering analysis. Water Resources Bulletin 28(3), 525-533. Kuoa, J.T., Hsiehb, M.H., Lungc, W.S.and Shed, N., 2007. Using artificial neural network for reservoir eutrophication prediction. Ecological Modelling 200, 171-177. Lahlou, M., Shoemaker, L., Choudhury, S., Elmer, R., Hu, A., Manguerra, H. and Parker, A., 1998. Better Assessment Science Integrating Point and Nonpoint Sources—BASINS Version 1.0. (Computer program manual). U.S. Environmental Protection Agency, 823B-98-006, Office of Water, Washington, DC.  122  Lakes and Rivers Improvement Act., 1990. Ontario regulation- lakes and rivers improvement act. C- L.3. Available at: Accessed June 2009. Larsen, R.E., Miner, J.R., Buckhouse, J.C. and Moore, J.A., 1994. Water-quality benefits of having cattle manure deposited away from streams. Bioresource Technology 48(2), 113118. Lee, K., Isenhart, T.M., Schultz, R.C. and Mickelson, S.K., 2000. Multispecies riparian buffers trap sediment and nutrients during rainfall simulations. Journal of Environmental Quality 29(4), 1200-1205. Lermontov, A., Yokoyama, L., Lermontov, M. and Machado, M.A.S., 2009. River quality analysis using fuzzy water quality index: Ribeira do Iguape river watershed, Brazil. Ecological Indicators 9(6), 1188-1197. Li, H. and Davis, A. P., 2009. Water quality improvement through reductions of pollutant loads using bioretention. Journal of Environmental Engineering 135(8), 567-576. Li, Y.P., Huang, G.H., Huang, Y.F. and Zhou, H.D., 2009. A multistage fuzzy-stochastic programming model for supporting sustainable water-resources allocation and management. Environmental Modelling and Software 24(7), 786-797. Lin, C., Huang, T. and Shaw, D., 2010. Applying water quality modeling to regulating land development in a watershed. Water Resources Management 24(4), 629-640. Line, D.E., White, N.M., Osmond, D.L., Jennings, G.D. and Mojonnier, C.B., 2002. Pollutant export from various land uses in the upper Neuse river basin. Water Environment Research 74(1), 100-108. Linking Source Water Protection to Land Use Planning in Alberta, 2008. Linking Source Water Protection to Land Use Planning in Alberta. Available at: Accessed June 2009. Linsley, R.K., Kohler, M.A. and Paulhus, J.L.H., 1982. Hydrology for engineers. McGraw Hill, New York, 508. Liou, S., Lo, S. and Hu, C., 2003. Application of two-stage fuzzy set theory to river quality evaluation in Taiwan. Water Research 37(6), 1406-1416. Liou, S., Lo, S. and Wang, S., 2004. A generalized water quality index for Taiwan. Environmental Monitoring and Assessment 96(1-3), 35-52. 123  Loehr, R.C., Ryding, S. and Sonzogni, W.C., 1989. Estimating the nutrient load to a waterbody. The Control of Eutrophication of Lakes and Reservoirs 1, 115-146. Lopez, A., Pollice, A., Lonigro, A., Masi, S., Palese, A.M., Cirelli, G.L., Toscano, A. and Passino, R., 2006. Agricultural wastewater reuse in southern Italy. Desalination 187(1-3), 323-334. Lowrance, R., Dabney, S. and Schultz, R., 2002. Improving water and soil quality with conservation buffers. Journal of Soil and Water Conservation 57(2), 36A-43A. Lu, R. and Lo, S., 2002. Diagnosing reservoir water quality using self-organizing maps and fuzzy theory. Water Research 36(9), 2265-2274. Lucas, W.C. and Greenway, M., 2008. Nutrient retention in vegetated and non vegetated bioretention mesocosms. Journal of Irrigation and Drainage Engineering-ASCE 134, 613-623. Lumb, A., Halliwell, D. and Sharma, T., 2006. Application of CCME water quality index to monitor water quality: A case of the Mackenzie River basin, Canada. Environmental Monitoring and Assessment 113(1-3), 411-429. Mamdani, E. H., 1977. Application of fuzzy logic to approximate reasoning using linguistic synthesis. IEEE Transactions on Computers C-26(12), 1182-1191. Mankin, K.R., Ngandu, D.M., Barden, C.J., Hutchinson, S.L. and Geyer, W.A., 2007. Grassshrub riparian buffer removal of sediment, phosphorus, and nitrogen from simulated runoff. Journal of the American Water Resources Association 43(5), 1108-1116. Marichal, J.L., 1999. Aggregation operators for multicriteria decision aid. PhD thesis, University of Liège, Liège, Belgium. Martinez-Martinez, S. I. and Campos-Aranda, D. F., 2010. Algorithms used to estimate physical properties of rural watersheds and their application in the calibration of the rational methods, in hydrological region No. 12 partial (Santiago River). Agriciebcia, 44(4), 393408. McDowell, R.W., 2008. Water quality of a stream recently fenced-off from deer. New Zealand Journal of Agricultural Research 51(3), 291-298. McFarland, A.M.S. and Hauck, L.M., 2001. Determining nutrient export coefficients and source loading uncertainty using in-stream monitoring data. Journal of the American Water Resources Association 37(1), 223-236. 124  McKay, J. and Moeller, A., 2000. Is it time for a new model of water quality laws? Environmental and Planning Law Journal 17, 165–175. Meixler, M.S. and Bain, M.B., 2010. A water quality model for regional stream assessment and conservation strategy development. Environmental Management 45(4), 868-880. Middleton, J.R. and Barrett, M.E., 2008. Water quality performance of a batch-type stormwater detention basin. Water Environment Research 80(2), 172-178. Miller, J., Chanasyk, D., Curtis, T., Entz, T. and Willms, W., 2010. Influence of streambank fencing with a cattle crossing on riparian health and water quality of the lower little bow river in southern Alberta, Canada. Agricultural Water Management 97(2), 247-258. Mishra, A., Benham, B.L. and Mostaghimi, S., 2008. Bacterial transport from agricultural lands fertilized with manure. Water Air Soil Pollution 189 (1-4), 127-134. Murofushi, T. and Sugeno, M., 1989. An interpretation of fuzzy measure and the Choquet integral as an integral with respect to a fuzzy measure. Fuzzy Sets and Systems 29 (2), 201–227. Murofushi, T., and Sugeno, M.1991. A theory of fuzzy measures. Representation, the Choquet integral and null sets. Journal of Mathematical Analysis and Applications 159 (2), 532– 549. Muthanna, T.M., Viklander, M., Gjesdahl, N. and Thorolfsson, S.T., 2007. Heavy metal removal in cold climate bioretention. Water, Air, and Soil Pollution 183(1-4), 391-402. NCDENP, 1999. Storm water best management practices. Department of Environment and Natural Resources Division of Water Quality Water Quality Section, North Carolina. Neitsch, S.L., Arnold, J.G., Kiniry, J.R. and Williams, J.R., 2001. Soil and Water Assessment Tool User's Manual, Version 2000. U.S. Department of Agriculture, Agricultural Research Service, Temple. Available at: Accessed June 2009. New Drinking Water Directive, 1998. Council Directive on the quality of water intended for human consumption. Official Journal of the European Union, 98/83/EC. Available at: Accessed June 2009.  125  Nitrates Directive, 1991. Council directive on concerning the protection of waters against pollution caused by nitrates from agricultural sources. Official Journal of the European Union, 91/676/EEC. Noor, H., Mirnia, S.K., Fazli, S., Raisi, M.B. and Vafakhah, M., 2010. Application of MUSLE for the prediction of phosphorus losses. Water Science and Technology 62(4), 809-815. Nunez, D., Nahuelhual, L. and Oyarzun, C., 2006. Forests and water: The value of native temperate forests in supplying water for human consumption. Ecological Economics 58(3), 606-616. Nutrient Management Act., 2002a. Ontario regulation- Nutrient Management Act. Regulation 447. Available at: Accessed June 2009. Nutrient Management Act., 2002b. Ontario regulation- Nutrient Management Act. Regulation 267. Available at: Accessed June 2009. Nutrient Management Act., 2002c. Ontario regulation- Nutrient Management Act. Regulation 511/05. Available at: Accessed June 2009. O‟Connor, D.R., 2002. A Strategy for Safe Drinking Water. Report of the Walkerton Inquiry 2. Ministry of the Attorney General, Toronto. Ocampo-Duque, W., Ferre-Huguet, N., Domingo, J.L. and Schuhmacher, M., 2006. Assessing water quality in rivers with fuzzy inference systems: A case study. Environment International 32(6), 733-742. Ohio Livestock Manure Management Guide, 2006. Ohio Livestock manure management guide. Bulletin 604, Ohio, USA. OMAG (Ontario Ministry of the Attorney General), 2002. Report of the Walkerton Inquiry: The Events of May 2000 and Related Issues. Queen‟s Printer for Ontario, Canada. Available at: r_03.pdf. Accessed March 2010. 126  Otles, S. and Cagindi, O., 2010. Health importance of arsenic in drinking water and food. Environmental Geochemistry and Health 32(4), 367-371. Palmstrom, N. and Walker, W., 1990. The P8 Urban Catchment Model for Evaluating Nonpoint Source Controls at the Local Level. Enhancing States' Lake Management Programs. U.S. Environmental Protection Agency, Washington, DC. Park, S.S. and Lee, Y.S., 1996. A multiconstituent moving segment model for water quality predictions in steep and shallow streams. Ecological Modeling 89 (1-3), 121-131. Passeport, E., Hunt, W.F., Line, D.E., Smith, R.A. and Brown, R.A., 2009. Field study of the ability of two grassed bioretention cells to reduce storm-water runoff pollution. Journal of Irrigation and Drainage Engineering 135(4), 505-510. Patty, L., Real, B. and Gril, J.J., 1997. The use of grassed buffer strips to remove pesticides, nitrate and soluble phosphorus compounds from runoff water. Pesticide Science 49(3), 243-251. Pesce, S.F. and Wunderlin, D.A., 2000. Use of water quality indices to verify the impact of Cordoba city (Argentina) on Suquia River. Water Research 34(11), 2915-2926. Pesticides Act., 2009. Ontario regulation- Pesticides Act. Regulation 63. Available at: Accessed June 2009. Pigram, J. J., 1993. Property rights and water markets in Australia: An evolutionary process toward institutional reform. Water Resources Research 29(4), 1313-1319. Pitt, R., 1993. Source Loading and Management Model (SLAMM). National Conference on Urban Runoff Management, Chicago, IL, March 1993. Planning Act., 1990. Ontario regulation- planning act. C-P- 13. Available at: Accessed June 2009. Postel, S.L. and Thompson Jr., B.H., 2005. Watershed protection: Capturing the benefits of nature's water supply services. Natural Resources Forum 29(2), 98-108. Poudel, D.D and Jeong, C,Y., 2009. Manual composite sampling in edge-of-field surface runoff for assessing nonpoint source pollution from agricultural and residential areas. Journal of Soil and Water Conservation 64 (5), 324-335.  127  Poudel, D.D., Jeong, C.Y. and DeRamus, A., 2010. Surface run-off water quality from agricultural lands and residential areas. Outlook on agriculture 39(2), 95-105. Prepas, E.E., Burke, J., Allen, E., Holst, M., Gibson, K. and Millions, D., 2001. The Virginia Hills Fire of 1998 and the opportunity to evaluate the impact of fire on water quality in upland stands on the Boreal Plain. Department of Biological Sciences. University of Alberta, Edmonton, AB. Public Health Act., 2009. BC regulation- Public health act. Regulation 49. Rajankar, P.N., Gulhane, S.R., Tambekar, D.H., Ramteke, D.S. and Wate, S.R., 2009. Water quality assessment of groundwater resources in Nagpur region (India) based on WQI. EJournal of Chemistry 6(3), 905-908. Ramakrishnaiah, C.R., Sadashivaiah, C. and Ranganna, G., 2009. Assessment of water quality index for the groundwater in Tumkur Taluk, Karnataka state, India. E-Journal of Chemistry 6(2), 523-530. Rankovic, V., Radulovic, J., Radojevic, I., Ostojic, A. and Comic, L., 2010. Neural network modeling of dissolved oxygen in the Gruza reservoir, Serbia. Ecological Modelling 221(8), 1239-1244. Rast, W. and Lee, G.F., 1983. Nutrient loading estimates for lakes. Journal of Environmental Engineering 109(2), 502-520. Reckhow, K.H., Beaulac, M.N. and Simpson, J.T., 1980. Modeling phosphorus loading and lake response under uncertainty: A manual and compilation of export coefficients. Office of Water Regulations, Criteria and Standards Division, U.S. Environmental Protection Agency, 440/5-80-011, Washington, DC, Resource Management Factsheet, 1994. Water quality evaluation of agricultural runoff in the lower FRASER valley. Ministry of agriculture and food. Order No. 511.300-1, Abbotsford, BC, Canada. Rickwood, C.J. and Carr, G.M., 2009. Development and sensitivity analysis of a global drinking water quality index. Environmental Monitoring and Assessment 156(1-4), 73-90. Roesner, L.A., Aldrich, J.A. and Dickinson, R.E., 1988. Storm Water Management Model User's Manual, Version 4: Addendum I, EXTRAN. U. S Environmental Protection Agency, 600/3-88/001b (NTIS PB88236658/AS), Athens, GA, 203.  128  Ross, T., 2004. Fuzzy Logic with Engineering Applications, 2nd Edition, John Wiley & Sons, New York. Rusciano, G.M. and Obropta, C.C., 2007. Bioretention column study: Fecal coliform and total suspended solids reductions. Transactions of the ASABE 50(4), 1261-1269. Sadiq, R., Haji, S.A., Cool, G. and Rodriguez, M.J., 2010. Using penalty functions to evaluate aggregation models for environmental indices. Journal of Environmental Management 91(3), 706-716. Sadiq, R., Kleiner, Y. and Rajani, B., 2007. Water quality failures in distribution networks - risk analysis using fuzzy logic and evidential reasoning. Risk Analysis 27(5), 1381-1394. Sadiq, R., Najjaran, H. and Kleiner, Y., 2006. Investigating evidential reasoning for the interpretation of microbial water quality in a distribution network. Stochastic Environmental Research and Risk Assessment 21(1), 63-73. Said, A., Stevens, D.K. and Sehlke, G., 2004. An innovative index for evaluating water quality in streams. Environmental Management 34 (4), 406-414. Sarkar, C. and Abbasi, S.A., 2006. Qualidex - A new software for generating water quality indice. Environmental Monitoring and Assessment 119(1-3), 201-231. Schmeidler, D., 1986. Integral representation without additivity, Proceedings of American Mathematical Society 97, 255–261. Schmeidler, D., 1989. Subjective probability and expected utility without additivity. Econometrica 57, 571–587. Scholz, M. and Grabowiecki, P., 2007. Review of permeable pavement systems. Building and Environment 42(11), 3830-3836. Scholz, M. and Yazdi, S.K., 2009. Treatment of road runoff by a combined storm water treatment, detention and infiltration system. Water, Air, and Soil Pollution 198(1-4), 5564. Schoonover, J.E., Williard, K.W.J., Zaczek, J.J., Mangun, J.C. and Carver, A.D., 2005. Nutrient attenuation in agricultural surface runoff by riparian buffer zones in southern Illinois, USA. Agroforestry Systems 64(2), 169-180. Schoumans, O.F. and Chardon, W.J., 2003. Risk assessment methodologies for predicting phosphorus losses. Journal of Plant Nutrition and Soil Science-Zeitschrift Fur Pflanzenernahrung Und Bodenkunde 166(4), 403-408. 129  Schwarzel, K., Renger, M., Sauerbrey, R. and Wessolek, G., 2002. Soil physical characteristics of peat soils. Journal of Plant Nutrition and Soil Science-Zeitschrift Fur Pflanzenernahrung und Bodenkunde 165 (4), 479-486. Sedeno-Diaz, J. and Lopez-Lopez, E., 2007. Water quality in the Rio Lerma, Mexico: An overview of the last quarter of the twentieth century. Water Resources Management 21(10), 1797-1812. Shrestha, S., Kazama, F. and Newham, L.T.H., 2008. A framework for estimating pollutant export coefficients from long-term in-stream water quality monitoring data. Environmental Modelling and Software 23(2), 182-194. SIMCAT, 2004. SIMCAT modeling report. Ashford‟s future- IWM study, Environmental Agency, Southern Region, Kent Area. Singh, K.P., Basant, A., Malik, A. and Jain, G., 2009. Artificial neural network modeling of the river water quality-A case study. Ecological Modelling 220(6), 888-895. Skaggs, R.W., 1980. DRAINMOD Reference Report - Methods for design and evaluation of drainage-water management systems for soils with high water tables. USDA SCS, South National Technical Center, Texas. Spinella, S., Ostoich, M. and Anile, A.M., 2008. River water quality assessment with fuzzy interpolation. Ecological Chemistry and Engineering S-Chemia i Inzynieria Ekologiczna S 15(2), 245-262. State Source Water Assessment and Protection Programs, 1997. Final Guidance. Canadian Legal Information Institute. Available at: Accessed June 2009. Statistics, 2010: Abbotsford, BC, Canada. Available at: Accessed June 2009. Subbarao, V.R., Mujumdar, P.P. and Ghosh, S., 2004. Risk evaluation in water quality management of a river system. Journal of Water Resources Planning and Management 130(5), 411-423. Sugeno, M. (1974). Theory of fuzzy integrals and its applications. Ph.D. Thesis, Tokyo Institute of Technology, Tokyo. Sullivan, T.J., Moore, J.A., Thomas, D.R., Mallery, E., Snyder, K.U., Wustenberg, M., Wustenberg, J., Mackey, S.D. and Moore, D.L., 2007. Efficacy of vegetated buffers in 130  preventing transport of fecal coliform bacteria from pasturelands. Environmental Management 40(6), 958-965. Suratman, S., Awang, M., Ling, L.A. and Tahir, N.M., 2009. Water quality index study in Paka river basin,Tterengganu. Sains Malaysiana 38(2), 125-131. Swamee, P.K. and Tyagi, A., 2000. Describing water quality with aggregate index. Journal of Environmental Engineering 126(5), 451-455. Swamee, P.K. and Tyagi, A., 2007. Improved method for aggregation of water quality subindices. Journal of Environmental Engineering 133(2), 220-225. Sydelko, P.J., Majerus, K.A., Dolph, J.E. and Taxon, T.N., 2000. An Advanced Object-Based Software Framework for Complex Ecosystem Modeling and Simulation. 4th International Conference on Integrating GIS and Environmental Modeling (GIS/EM4): Problems, Prospects and Research Needs, Banff, Alberta, Canada, September 2000. Takagi, T. and Sugeno, M., 1985. Fuzzy identification systems and its applications to modellig and control. IEEE Transactions on Systems, Man and Cybernetics SMC-15(1), 116-132. Technical Standards and Safety Act., 2000. Ontario regulation- liquid fuels. Regulation 217. Available at: Accessed June 2009. Timmer, D.K., de Loe, R.C. and Kreutzwiser, R. D., 2007. Source water protection in the Annapolis valley, Nova Scotia: Lessons for building local capacity. Land use Policy 24(1), 187-198. Tversky, A. and Kahneman, D., 1992. Advances in prospect theory: cumulative representation of uncertainty. Journal of Risk and Uncertainty 5, 297–323. Udeigwe, T.K., Wang, J.J., Viator, H.P. and Gaston, L., 2010. Surface water quality as affected by sugarcane residue management techniques. Water, Air, and Soil Pollution 208(1-4), 119-128. USACE-HEC (U.S. Army Corps of Engineers, Hydrologic Engineering Center),. 1977. Storage, Treatment, Overflow, Runoff Model, STORM. Generalised Computer Program, Davis, California. USEPA, 1976. Areawide Assessment Procedures. U. S Environmental Protection Agency, 600/9-76014. Municipal Environmental Research Laboratory. Cincinnati, Ohio.  131  USEPA, 1985. Computer program documentation for the enhanced steam water quality model QUAL2E. Environmental Research Laboratory, Athens, GA. USEPA, 1992. Compendium of Watershed-Scale Models for TMDL Development. U. S Environmental Protection Agency, 841-R-92-002, Office of Water, Washington, DC. USEPA, 1995. Combined Sewer Overflows- Guidance for nine minimum controls. United States Environmental Protection Agency, 832-B-95-003. USEPA, 2000. Low Impact Development (LID): A Literature Review. United States Office of Water (4203). 841-B-00-005, Environmental Protection Agency, Washington, DC. USEPA, 2001a. Source Water Protection Practices Bulletin Managing Agricultural Fertilizer Application to Prevent Contamination of Drinking Water. U. S Environmental Protection Agency, 916-F-01-028, Office of water. USEPA, 2001b. Source Water Protection Practices Bulletin Managing Large-Scale Application of Pesticides to Prevent Contamination of Drinking Water. U. S Environmental Protection Agency, 916-F-01-030, Office of water. USEPA, 2001c. Source Water Protection Practices Bulletin Managing Small-Scale Application of Pesticides to Prevent Contamination of Drinking Water. U. S Environmental Protection Agency, 816-F-01-031, Office of water. USEPA, 2001d. Source Water Protection Practices Bulletin Managing Livestock, Poultry, and Horse Waste to Prevent Contamination of Drinking Water. U. S Environmental Protection Agency, 916-F-01-026, Office of water. USEPA, 2001e. Source Water Protection Practices Bulletin Managing Storm water Runoff to Prevent Contamination of Drinking Water. U. S Environmental Protection Agency, 816F-01-020, Office of water. USEPA, 2001f. Source Water Protection Practices Bulletin Managing Pet and Wildlife Waste to Prevent Contamination of Drinking Water. U. S Environmental Protection Agency, 916F-01-027, Office of water. USEPA, 2001g. Source Water Protection Practices Bulletin Managing Septic Systems to Prevent Contamination of Drinking Water. U. S Environmental Protection Agency, 816-F-01021, Office of water.  132  USEPA, 2001h. Source Water Protection Practices Bulletin Managing Sanitary Sewer Overflows and Combined Sewer Overflows to Prevent Contamination of Drinking Water. U. S Environmental Protection Agency, 916-F-01-032, Office of water. USEPA, 2001i. PLOAD version 3.0: An ArcView GIS tool to calculate nonpoint sources of pollution in watershed and storm water projects. User‟s Manual. USEPA, 2002. The loading simulation program in C++ (LSPC) watershed modeling systemuser’s manual. Tetra Tech, Inc., USA. Veiga; B.V. and Dziedzic, M., 2010. Estimating nutrient loads in the Passana reservoir with FLUX. Water International 35 (2), 210– 222. Walker, W.W., 1997. GISPLM User’s Guide. LaPlatte River Phosphorus Modeling Project for Vermont Department of Environmental Conservation. Available at: Accessed June 2009. Wang, C., Wang, H.V. and Kuo, A.Y., 2008. Mass conservative transport scheme for the application of the ELCIRC model to water quality computation. Journal of Hydraulic Engineering 134(8), 1166-1171. Water Act., 1996. BC regulation- Water act. C- 483. Available at: Accessed June 2009. Water Act., 2000. Alberta regulation- water act. C- W-3. Available at: Accessed June 2009. Water Act., 2007. BC regulation- water act. Regulation 390. Watts, D.B. and Torbert, H.A., 2009. Impact of gypsum applied to grass buffer strips on reducing soluble p in surface water runoff. Journal of Environmental Quality 38(4), 1511-1517. Weiss, P.T., Gulliver, J.S. and Erickson, A.J., 2007. Cost and pollutant removal of storm-water treatment practices. Journal of Water Resources Planning and Management 133(3), 218229. Wetland Management in the Settled Area of Alberta, 1993. Wetland Management in the Settled Area of Alberta- An Interim Policy. Water Resources Commission, Edmonton, Alberta.  133  Whitehead, P.G., Williams, R.J. and Lewis, D.R., 1997. Quality simulation along river systems (QUASAR): model theory and development. Science Total Environment 194, 447 –456. WHO (World Health Organization), 1993. Guidelines for Drinking Water Quality. Volume 1 Recommendations. WHO, Geneva. Williams, J.R., Renard, K.G. and Dyke, P.T., 1983. EPIC: A new method for assessing erosion's effect on soil productivity. Journal of Soil & Water Conservation 38(5), 381-383. Wilsenach, J.A., Maurer, M., Larsen, T. A. and Van Loosdrecht, M.C.M., 2003. From waste treatment to integrated resource management. Water Science & Technology 48(1), 1-9. Wischmeier, W.H. and. Smith. D.D., 1978. Predicting Rainfall Erosion Losses: A Guide to Conservation Planning. Agriculture Handbook No. 537. U.S. Department of Agriculture, USDA, Washington, DC. Wittmer, I.K., Bader, H., Scheidegger, R., Singer, H., Luck, A., Hanke, I., Carlsson, C. and Stamm, C., 2010. Significance of urban and agricultural land use for biocide and pesticide dynamics in surface waters. Water Research 44(9), 2850-2862. Woolhiser, D.A., Smith, R.E. and Goodrich, D.C., 1990. Agricultural Handbook 537. U.S. Department of Agriculture, KINEROS, A kinematic runoff and erosion model: documentation and user manual. ARS-77. USDA-Agricultural Research Service. Washington, DC. Yager, R.R. and Filev, D.P., 1994. Essentials of fuzzy modeling and control. John Wiley & Sons, Inc., NY. Yeon, I.S., Kim, J.H. and Jun, K.W., 2008. Application of artificial intelligence models in water quality forecasting. Environmental Technology 29(6), 625-631. Young, R.A., Onstad, C.A., Bosch, D.D. and Anderson, W.P., 1987. AGNPS, Agricultural NonPoint Source Pollution Model - A watershed analysis tool. Conservation Research Report 35, 1-80. United States Department of Agriculture, Washington, DC. Yu, R., Gan, P., MacKay, A.A., Zhang, S. and Smets, B.F., 2010. Presence, distribution, and diversity of iron-oxidizing bacteria at a landfill leachate-impacted groundwater surface water interface. FEMS Microbiology Ecology 71(2), 260-271. Yuan, Y., Bingner, R.L. and Locke, M.A., 2009. A review of effectiveness of vegetative buffers on sediment trapping in agricultural areas. Ecohydrology 2(3), 321-336. Zadeh, L.A., 1965. Fuzzy sets. Information and Control 8(3): 338-353. 134  Zakir, S.N., Ihsanullah, I., Shah, M.T., Iqbal, Z. and Ahmad, A., 2009. Comparison of heavy and trace metals levels in soil of Peshawar basin at different time intervals. Journal of the Chemical Society of Pakistan 31(2), 246-256. Zhang, Y., Liu, S., Liu, H. and Liu, Z., 2010. Evaluation of the combined toxicity of 15 pesticides by uniform design. Pest Management Science 66(8), 879-887.  135  APPENDICES Appendix A: Available Canadian regulations regarding SWP strategies in three provinces (Ontario, Alberta, and British Columbia) Management  Land Management  Province Name  Regulation  Ontario  Conservation Land Act (1990) Crown Forest Sustainability Act (1994) Crown Timber Act (1990) Forestry Act (1990) Conservation Authorities Act (2006) Lakes and Rivers Improvement Act (1990) Greenbelt Act (2005) Planning Act (1990)  Alberta  Alberta Forest Conservation Strategy (1997) Wetland Management in the Settled Area of Alberta (1993) Water Act (2000)  British Columbia  Water Act ( 2007) Forest and Range Practices Act (2002) Forest and Range Practices Act (2004) Forest and Range Practices Act (2009) Drinking water protection Act (2001) Environmental and Land Use Act (1996)  Identification of SWP strategies         „Conservation of forest‟ to achieve land management for ensuring better water quality. Restriction in development within 130 m of significant and 30 m from other „wetlands‟ so that wetlands can perform their hydrological function. „Riparian areas‟ are mentioned as a mean of water quality improvement. Land management is declared as a SWP strategy. Agricultural land conservation for ecosystem preservation. Proper „planning‟ of a area after preserving natural lands for SWP.      „Forests‟ are identified as an effective mean to have clean water. „Wetland‟ management is highly encouraged to protect water quality. Integrated approach with proper planning with land and water resources.     „Vegetation‟ is stated as a mean to have stable stream channel. „Forests‟ are also considered as effective protection way of watershed management. „Wetlands and riparian areas‟ are stated as a buffer for ecosystem. Provincial and local govt. strategies are recommended for land management to have safe water. Land use relation with source water is stated. Close relation between ecological conservation with land use is declared.       136  Management  Province Name  Ontario  Agricultural and farming management  Pesticides Act (2009) Nutrient Management Act (2002a, 2002b, 2002c) Environmental Protection Act (1990)  Identification of SWP strategies           „Monitoring‟ of pesticides in surface water by leakage and restriction of washing with pesticides contaminated water. Animals are restricted near surface water. „Vegetative buffer‟ is recommended before nutrient application. Setback maintenance for all type of nutrient application near surface water is stated.  Ontario  Environmental Assessment Act (1996) Environmental Protection Act (2007)      Setback from source water is recommended to apply pesticides and herbicides. Careful management in handing pesticides especially for storm water. Limited application by aircraft is suggested. Restriction for livestock watering for prevention of contamination of water. Setback distance is recommended for domestic, agricultural composting, farming composting, incineration, and gazing animal areas. Setback distance is recommended for fertilizer application for sivilculture treatment. Regulation on the selection of pesticides. Better pesticides management to protect water course. In case of poultry and slaughter, 10 mg/l is the maximum permissible for fat, oil or grease release. Non-structural programs and best management practices are encouraged. Regular „monitoring‟ in case of extreme weather events. Effective „drainage design‟ and planning is stated.  Alberta  Environmental Protection and Enhancement Act (1993a, 1993b)     „Drainage design‟ is announced to perform. No substance can be discharged which is harmful.  Environmental Management Act (2004b)    Careful storage of petroleum chemicals is declared.    Installation of a landfill requires regular surface water „monitoring‟, leachate determination. Turbidity, dissolved oxygen, temperature and the hydrology of the site should be monitored. Setback distance for leaf composting. Cannot dump more than 100 liter of mineral oil. Disposal without adequate „treatment‟ is restricted. Bern and dykes are encouraged in landfill area for protecting water pollution.  Alberta  British Columbia  Storm water management British Columbia  Municipal management  Regulation  Ontario  Environmental Protection and Enhancement Act (1993a, 1993b)  Water Act (1996) Environmental Management Act (2008a) Forest and Range Practices Act (2004) Integrated Pest Management Act (2004) Environmental Management Act (2007)  Environmental Protection Act (1990) Environmental Protection Act (1994) Environmental Protection Act (1998a, 1998b)             137  Management  Province Name  Regulation  Identification of SWP strategies  Environmental Protection and Enhancement Act (2000) Environmental Protection and Enhancement Act (1996a, 1996b) Environmental Protection and Enhancement Act (2003) Water Act (2007)      Restriction in dumping waste to source water and ice. Strict design of the landfill. Leachate should be monitored to maintain the standards.  Environmental Management Act (2008b) Environmental Management Act (2007)     Setback distance from landfill is declared to maintain. Restriction in case of changing the direction of stream flow.  Ontario  Environmental Protection Act (1990)    „Treatments‟ are recommended before dumping.  Alberta  Environmental Protection and Enhancement Act (2006)    Restriction in dumping waste near surface water.    Restriction in sewage disposal in proper way to avoid contamination in source water.    Restriction in dumping vehicular spillage more than specific amount.  Environmental Management Act (2004b)    Recreational dumping is restricted.  Ontario  Environmental Protection Act (1990)    Setback from asphalt pavement construction site.  Alberta  N/A  N/A  British Columbia  N/A  N/A  Alberta Municipal management  British Columbia  Industrial management  British Columbia  Vehicle spill management  Ontario  Environmental Protection Act (1998b) Technical Standards and Safety Act (2000) Environmental Protection Act (1990)  Alberta  N/A  British Columbia  Road management  Public Health Act (2009)  N/A= Not available.  138  Appendix B: Source water protection strategies under USEPA bulletin Management sectors Agricultural management  Reference USEPA (2001a)  SWP strategies             Managing large-scale application of pesticides  USEPA (2001b)       Managing small scale application of pesticides  USEPA (2001c)  Farming Management  USEPA (2001d)            „Yearly soil sampling‟ to evaluate the exact fertilizer demand of the crop. Use of „ammonia nitrogen fertilizer‟ to get rid of immediate leaching. Proper „fertilizer timing‟, e.g., close time of fertilizer application to the period of maximum uptake time. Proper calibration of the fertilizer application equipment to ensure the required amount of fertilizer. Correct placement of the fertilizer  Place near the roots  In case of erosive soil, place the dry fertilizer into the ground in bands.  Never apply the fertilizer to frozen ground on in sloppy areas. Proper application of the irrigation water  Increase efficiency by using lined ditches, gate pipe, and reuse system.  Shorten the use of gravity controlled irrigation. Careful fertilizer storage from any source of water. „Conservation tillage‟ to reduce runoff. Using „buffer strips or filter strips‟, e.g., plating dense vegetable near water bodies to filter fertilizers. Use of „crop rotation‟ to minimize fertilizer need. Use of „cover crops‟ to stop wind and soil erosion. Careful use of Integrated Pest Management (IPM) with chemical and non-chemical ways, e.g., mechanical, cultural, biological, sanitation and plating pest resistant plants. Proper pesticide application (proper setbacks and never start the application before any weather event). Economic and effective use of pesticides by „crop rotation‟, proper placement of the pesticides, use split application procedure (Application before and at the plating time), „spot treatment‟ Careful management of the pesticide storage and handling In case of the large scale- pesticide use manual activities, e.g., spading, hoeing, hand picking weeds and pests, mulching to get rid of pests without pesticides are recommended. Proper plant management to reduce the need for the pesticides.  Maintain proper drainage and aeration to have the microbes to degrade the pesticides.  Using biological control (e.g., birds and bats). Feedlot management such as by using waste storage lagoons, litter storage structures, clean water divisions, composting and runoff treatment. Using poultry liner storage which can keep the rainwater runoff from poultry home waste. The water can be further used for agricultural use. Water diversion especially clean water to keep them away from the pollution. Use of „vegetation buffer‟ for feedlot management. Proper application of manure with proper placement. Pasture management such as by „fencing‟.  139  Management sectors Storm Water or Runoff Management  Reference USEPA (2001e)  SWP strategies             Managing pet and wild life  USEPA (2001f)  Managing septic systems  USEPA (2001g)          Managing Sanitary Sewer  USEPA (2001h)        Plant temporary fast-growing vegetation, grasses and flowers to filtrate water. Covering top soil with „geotextile or impervious covers‟. Proper „planning‟ to minimize directly connected impervious areas (Connect runoff from roofs and sidewalks). Placement of concrete grid pavement placed on a sand or gravel base with a void area filled with pervious materials. Effective structural „design‟ to control runoff. Use of „grass swales‟. „Buffer strip‟ which is made of three zones is recommended. The main parts are Four or five rows of trees closest to the source water.  One or two rows of shrubs.  20/24 foot wide grass zone. Long rooted vegetation is preferred for buffer strip. „Storm water ponds‟ which can settle the solids and with the help of the wetland vegetation zone contaminants can be removed biochemically. „Constructed wetlands‟ whose main function is similar to storm water ponds is recommended. „Swirl type concentrators‟ which can create circular motion to remove oil, and grease can be used for oily substances. Clean up and waste disposal. Bury waste. Keep the pets away from the water bodies. „Long grass‟ which not only attracts the pets but also infiltrate the contaminate particle is used for managing wild life. Proper sitting of septic system  Maintenance of proper setback distances (both horizontal and vertical).  Adequate soil permeability to ensure septic system effluent. Design and construction consideration. Annual inspection of the septic tank. Visual inspection about the proper working of the septic tank system. Monitoring and maintenance. Employee training. Public education. Eliminating direct pathways to source water.  140  Appendix C: Main factors in storm water management Reference (Type of SWM)  SA/DA, %  Brief system description  Depth, m  RT, hr  Vegetation type  Li and Davis (2009) (Bioretention)  6 or 2  0.5-0.8 or 0.9  6-8  Natural vegetation ( e.g., grasses, shrubs, and small trees)  Davis (2007) (Bioretention)  2  0.9 or 0.3  24  Grasses, shrubs, and small trees or no vegetation  Hunt et al. (2008) (Bioretention)  6.00  Sandy loam or sandy clay loam : Sand -80% or 45% Silt- 13%or 26% Clay- 7% or 20% pH -7.3 OM- 5.7% or 12.2% Sand-50% Topsoil- 30% Mulch (Shredded hardwood)20% Or Mixture of newspaper mass :sand=0.017:1 Loamy sand: Soil media P-index 6 (low) Silt and clay fraction 5.7% Permeability 0.43 in./h  1.2  24-36  Hsieh and Davis (2005) (Bioetention)  N/A  1.15 or 1.25  6  Hathaway et al. (2009) (Dry and wet pond)  0.64 to 4.67  Mulch-5cm Filter- 55-75cm, Soil- 10-30cm, Fine sand-5cm Or additionally vegetative layer20-35cm above filter-25-50cm Vegetation-grass Sunlight-high or lor  Blueflag iris (Iris virginica), cardinal flower ( Lobelia cardinalis), common rush (Juncus effusus) , hibiscus (Hibiscus spp.) , red maple (Acer rubrum) , sweet pepperbush (Clethra alnifolia) , Virginia sweetspire ( Itea virginica) , wild oat grass (Chamanthium latifolium ) Not specified  N/A  48  Natural trees around the pond  Passeport et al. (2009) (Bioretention)  3.20  0.75 or 1.05 (0.451 or 0.751)  Longer duration for ISZ  Bermuda grass cover  Scholz and Yazdi (2009) (Combined detention and infiltration system)  2.88  80% state 15% sand 5% topsoil Underlying soil-loamy clay or sandy loam: P-index-5 or 8 Humic matter-0.18% Layer 1- 0.05m gravel; Layer 20.15 to 0.25m gravel; Layer 3Geotextile; Layer 4- 0.1to0.2 m of mixture; Layer 5- plastic liner  0.3 -0.5  0.17  S. Viminalis is used as plantation  141  Reference (Type of SWM)  SA/DA, %  Brief system description  Middleton and Barrett (2008) (Sand filter with extended detention basin) Muthanna et al. (2007) (Bioretention)  1.33  Sand filter followed by extended detention basin  4.75  Hsieh et al. (2007 a, 2007b) (Bioetention)  4-5  10 cm gravel (bottom) 55 cm sandy loam: 92.7% sand 2.6% clay 4.7% silt pH-6.88 OM-8.7% 10 cm Mulch (pH-5.5) 15cm freeboard 2% mulch (mixed of grass and vegetation) 20.5%, Soil I 77.5 % Sand II  4-5  Lucas and Greenway (2008) (Bioretention)  N/A  29.3% homogeneous media (3 kg mulch, 3kg soil IV, 6 kg sand) 56.1%, Sand II 14.6 % Soil IV 80cm sandy loam, 80cm loamy sand 20cm pea gravel with loamy sand  Depth, m  RT, hr  Vegetation type  72  No vegetation  0.9  0.83  Lythrum salicaria., Iris pseuacorus, Vinca minor and Hippophaë rahmnoides with small shrubs, evergreens and perennials.  0.95  1-2  Not specified  0.75-0.80 (0.15)  12-18  Swamp Foxtail Grass (Pennisetum alopecurioides) and Flax Lily (Dianella brevipedunculata), and two woody shrubs, Banksia (Banksia integrefolia), and Bottlebrush (Callistemon pachyphyllus).  SA/DA= surface area of the BMPs/ Drainage area; RT= retention time; OM= Organic matter; N/A= Not available; ISZ= Internal storage zone (with media); Depth= Media depth; Soil I= Sandy loam: sand-71%, clay-17%, silt-12%, pH-6.7, OM-4.4.%; Sand II = Sand-95%, clay-3%, silt-2%, pH-7.1, OM-0.15%; Soil IV = Sandy loam: sand-71%, clay-14%,; silt-15%, pH-7.1, OM-3.5%; 1- Water depth  142  Appendix D: Common formulations for water quality indices (WQI) Index  Parameter, unit  Sub-index,  DO, % NSF-WQI (Brown et al., 1970)  FC, MPN/100ml pH BOD5, ppm Nitrates, ppm  O-WQI (Dunnette, 1979; Cude, 2001)  PW-WQI (Pesce and Wunderlin, 2000)  Aggregation formulation 0.17  142 experts draw curves for raw data and assigned a value ranging from 0 (worst) to 100 (best) and final curves were obtained with the weighting curves for each parameter  Range 0-25 = very bad  0.16  ∑  26-50 = bad  0.11  51-70 = regular  0.11  71-90 = good  0.10  91-100 =excellent  Total Phosphates, ppm  0.10  Temp., °C  0.10  Turbidity, NTU/ JTU  0.08  Total suspended solids, ppm Temp., °C  0.07 1, a  DO, %  1, 2  √ ∑  10-59 = very poor 60-79 = poor  BOD5, mg/l  2  80-84 = fair  pH  2  85-89 = good  Ammonia +Nitrate nitrogen, mg/l  2  90-100= excellent  Total phosphorus, mg/l  1, b  Total solids, mg/l  2  FC, #/100ml  2  DO, mg/l Con, µS/cm Turbidity, NTU  ∑  0= minimum quality 100= maximum quality  143  Index CPCB-WQI (Sarkar and Abbasi, 2006)  River pollution index (RPI) (Liou et al., 2004)  Parameter, unit  Sub-index,  Aggregation formulation  DO, %  0.31  BOD5, mg/l  0.19  pH  0.22  FC, MPN/ 100ml  0.28  DO, mg/l BOD5, mg/l Ammonia nitrogen, mg/l Suspended solids, mg/l Turbidity, NTU Temp., °C. FC, MPN/100ml pH Toxicity  Range <38 =bad to very bad 38–50 =bad 50–63 = medium to good 63–100 = good to excellent  ∑  Value varies from 0-64.8 and are divided into nonpolluted, lightlypolluted, moderatelypolluted, and grossly-polluted  ⁄  (∑  )  (∑  )  ∑ = sub-index for two particulate parameters =Sub-index for FC =sub-index for last three parameter  U-WQI (Boyacioglu, 2007)  S-WQI (Said et al., 2004)  Cadmium, cyanide, mercury, selenium, arsenic, fluoride, nitratenitrogen, DO, BOD5, total phosphorus, pH and total coliform DO, % Con, µS/cm Turbidity. , NTU FC, MPN/100ml Total phosphorus (TP)  N/A  N/A  ∑  N/A  ⁄  ]  < 1= poor < 2 =marginal and remediation 3-2= acceptable 3= very good  144  Index  Parameter, unit Temp.,°C  Sub-index,  Aggregation formulation SITemp (SITOC + SISS+ SIDO + SICon) -  TOC, mg/l  Range 0= minimum quality 100= maximum quality  SS, mg/l DO, mg/l Con, μS/cm Not fixed  √  F1: scope (% of variables that do not meet their objectives at least once) F2: frequency (% of individual tests that do not meet their objectives)  S-T WQI (Swamee and Tyagi, 2007)  Not fixed  F3: amplitude (amount by which failed tests do not meet their objectives) Monotonically decreasing Sub-indices , SI= Non-uniformly decreasing subindices, SI=  ∑  0-44= poor 45-64= marginal 65-79= fair 80-94= good 95-100= excellent  0–0.25= poor 0.26–0.50= fair 0.51–0.70= medium /average 0.71–0.90= good 0.91–1.0= excellent  Unimodal subindices, SI= - Relative weights; DO- Dissolved oxygen ; FC- Faecal coliform; BOD5- Biochemical oxygen demand for 5 day; Con- Electric conductivity; SS- Suspended solids; CPCB- Central Pollution Control Board ; NTU-Nephelometric Turbidity Units; JTU-Jackson Turbidity Unit; MPN- The Most Probable Number; P- Parameter real value ; -Characteristic value of P (Swami and Tyagi 2000); Threshold concentration; 1- Polynomial function 2- Exponential function 3- Linear function 4- Step function 5- Logarithmic function a- 3rd order polynomial function b- 2nd order polynomial function; n, q, r- Exponents (Swamee and Tyagi, 2000) * Catalan water Agency, available at ; ** Canadian Council of Ministers of the Environment (CCME), in: Canadian Environmental Quality Guidelines, CCME, Winnipeg, Canada, 2001  145  Appendix E: Pollutant load mass generation from manure (ASAE, 2003) Factors  Dairy  Beef  Veal  Swine  Layer (Poultry)  Turkey  640  360  91  61  1.8  6.8  86  58  62  84  64  47  12  8.5  5.2  11  16  12  TN, kg/1000kg  0.45  0.34  0.27  0.52  0.84  0.62  TP, kg/1000kg  0.094  0.092  0.066  0.18  0.3  0.23  BOD5, kg/1000kg  1.6  1.6  1.7  3.1  3.3  2.1  FC, CFU/1000kg  16  28  18  7.5  1.4  TC, CFU/1000kg  1100  63  45  110  0.00008 4  0.00074  Avg. Size (  ), Kg  Total manure, kg/day  Pb, kg/1000kg  a-  assumed value based on the similar type of livestock  1- TSS has been assumed 1/3 of total solids  146  Appendix F: Membership functions and inference rules for TSS removal using SWM  Rule-set for TSS removal using SWM Rule no  Water depth  Vegetation  RT  Media depth  1  If  L  and  L  and  L  and  2  If  L  and  L  and  M  3  If  L  and  L  and  M  4  If  L  and  L  and  5  If  L  and  L  6  If  L  and  7  If  L  and  8  If  L  9  If  10 11  SWM Efficiency  L  then  VL  and  L  then  L  and  M  then  L  M  and  H  then  M  and  H  and  L  then  L  L  and  H  and  M  then  L  L  and  H  and  H  then  M  and  M  and  L  and  L  then  L  L  and  M  and  M  and  L  then  M  If  L  and  M  and  M  and  M  then  H  If  L  and  M  and  M  and  H  then  H  12  If  L  and  M  and  H  and  L  then  M  13  If  L  and  M  and  H  and  M  then  H  14  If  L  and  M  and  H  and  H  then  H  15  If  L  and  H  and  L  and  L  then  VL  147  Rule no  Water depth  Vegetation  RT  Media depth  SWM Efficiency  16  If  L  and  H  and  M  and  L  then  VL  17  If  L  and  H  and  M  and  M  then  M  18  If  L  and  H  and  M  and  H  then  H  19  If  L  and  H  and  H  and  L  then  L  20  If  L  and  H  and  H  and  M  then  M  21  If  L  and  H  and  H  and  H  then  H  22  If  M  and  L  and  L  and  L  then  L  23  If  M  and  L  and  M  and  L  then  L  24  If  M  and  L  and  M  and  M  then  M  25  If  M  and  L  and  M  and  H  then  M  26  If  M  and  L  and  H  and  L  then  M  27  If  M  and  L  and  H  and  M  then  H  28  If  M  and  L  and  H  and  H  then  VH  29  If  M  and  M  and  L  and  L  then  L  30  If  M  and  M  and  M  and  L  then  M  31  If  M  and  M  and  M  and  M  then  M  32  If  M  and  M  and  M  and  H  then  H  33  If  M  and  M  and  H  and  L  then  H  34  If  M  and  M  and  H  and  M  then  H  35  If  M  and  M  and  H  and  H  then  H  36  If  M  and  H  and  L  and  L  then  L  37  If  M  and  H  and  M  and  L  then  M  38  If  M  and  H  and  M  and  M  then  M  39  If  M  and  H  and  M  and  H  then  H  40  If  M  and  H  and  H  and  L  then  M  41  If  M  and  H  and  H  and  M  then  H  42  If  M  and  H  and  H  and  H  then  H  43  If  H  and  L  and  L  and  L  then  L  44  If  H  and  L  and  M  and  L  then  L  45  If  H  and  L  and  M  and  M  then  M  46  If  H  and  L  and  M  and  H  then  H  47  If  H  and  L  and  H  and  L  then  VH  48  If  H  and  L  and  H  and  M  then  VH  49  If  H  and  L  and  H  and  H  then  VH  50  If  H  and  M  and  L  and  L  then  M  51  If  H  and  M  and  M  and  L  then  M  52  If  H  and  M  and  M  and  M  then  H  53  If  H  and  M  and  M  and  H  then  H  54  If  H  and  M  and  H  and  L  then  VH  55  If  H  and  M  and  H  and  M  then  VH  56  If  H  and  M  and  H  and  H  then  VH  148  Rule no  Water depth  Vegetation  RT  Media depth  57  If  H  and  H  and  L  and  58  If  H  and  H  and  M  59  If  H  and  H  and  M  60  If  H  and  H  and  61  If  H  and  H  62  If  H  and  63  If  H  and  SWM Efficiency  L  then  L  and  L  then  M  and  M  then  M  M  and  H  then  H  and  H  and  L  then  H  H  and  H  and  M  then  H  H  and  H  and  H  then  H  SWM- storm water management; RT- retention time; L- low; M- medium; H- high; VH- very high  149  Appendix G: Membership functions and inference rules for TSS removal using VFS  Rule- set for TSS removal using VFS Rule no  Slope  Width  n  VFS efficiency  1  If  L  and  L  and  L  then  VL  2  If  L  and  L  and  M  then  M  3  If  L  and  L  and  H  then  H  4  If  L  and  M  and  L  then  L  5  If  L  and  M  and  M  then  H  6  If  L  and  M  and  H  then  VH  7  If  L  and  H  and  L  then  M  8  If  L  and  H  and  M  then  H  9  If  L  and  H  and  H  then  VH  10  If  M  and  L  and  L  then  VL  11  If  M  and  L  and  M  then  L  12  If  M  and  L  and  H  then  L  13  If  M  and  M  and  L  then  L  14  If  M  and  M  and  M  then  M  15  If  M  and  M  and  H  then  H  16  If  M  and  H  and  L  then  M  17  If  M  and  H  and  M  then  H  18  If  M  and  H  and  H  then  VH  150  Rule no  Slope  Width  n  VFS efficiency  19  If  H  and  L  and  L  then  VL  20  If  H  and  L  and  M  then  L  21  If  H  and  L  and  H  then  M  22  If  H  and  M  and  L  then  M  23  If  H  and  M  and  M  then  M  24  If  H  and  M  and  H  then  M  25  If  H  and  H  and  L  then  M  26  If  H  and  H  and  M  then  M  27  If  H  and  H  and  H  then  H  n- Manning‟s n; L- low; M- medium; H- high; VH- very high  151  Appendix H: Mass balance equations for water quality parameters Model name  Reference  Water quality parameters DO  QUAL2E  USEPA  =  (1985)  )+(  Organic nitrogen,  (  = )A- L- -  -  SIMCAT  SIMCAT (2004)  TOMCAT  Bowden and Brown (1984)  QUASAR  MIKE 11  Whitehead et al. (1997)  DHI Water & Environment (2005)  =-  L+  (  (  -  =  Nitrite nitrogen,  -  =  Nitrate nitrogen,  -  = )  + µA  -(1-  Organic phosphorus,  =  µA  A-  Dissolved phosphorous,  = µA  + -  BOD5  Coliform, E  BOD5=L(1) =-  L-  =-  E  L  -DO)-  -  -  =(T)  4.57  = τ( -DO+WEIR)+PR- DO+ ( )4.57 L = P-R+ L-  Temp, T  -DO)  =14.6520.41022T+0.0079910 0.000077774 =  A-  Ammonia nitrogen,  (  )-  = [  ]-  = [ ]+  = [ -  ]-  L+ =-  =  +  -  L(1)+  Chla -  152  Appendix I: Predicted value of the parameters for models in Appendix H (USEPA, 1985) Legend  Parameter name  Unit  Carbonaceous deoxygenating rate constant Re-aeration rate constant Rate of loss of BOD due to settling Sediment oxygen demand rate (benthic oxygen uptake)  g/  -day  Coliform die off rate BOD conversion rate coefficient  Predicted value range  Temperature dependant  0.02-3.4  Yes  0-100  Yes  -0.36-0.036  Yes  N/A  Yes  0.05-4.0  Yes  0.23  Yes  Faction of algal biomass that is nitrogen  mg /mg  0.07-0.09  No  Fraction of algal biomass that is phosphorous  mg /mg  0.01-0.02  No  O2 production /algal growth  mg /mg  1.4-1.8  No  O2 uptake per unit of algae respired  mg /mg  1.6-2.3  No  The rate of oxygen uptake per unit per unit of ammonium –nitrogen oxidation  mg /mg  3-4  No  O2 uptake per unit of NO2 oxidation  mg /mg  1-1.14  No  Rate constant for biological oxidation of NH3-NO2  0.1-1  Yes  Rate constant for biological oxidation of NO2-NO3  0.02-0.4  Yes  Rate constant for hydrolysis of organic nitrogen to ammonia nitrogen  0.02-0.4  Yes  Rate constant for the decay of organic-P to dissolved-P  0.01-0.7  Yes  153  Legend  Parameter name  Unit  Organic phosphorous settling rate Algal growth rate Algal respiration day  Predicted value range  Temperature dependant  0.001-0.1  Yes  1-3  No  0.05-0.5  No  = Saturated DO, mg/l; L= concentration of ultimate carbonaceous BOD, mg/l; d= Mean stream depth, ft; τ = the travel time or residence time= ; L is the reach length; v is the water velocity;  = is the concentration of the solute entering the reach; WEIR= the increase in DO concentration due to a weir or other structure in the reach; P = the rate of (algal)  photosynthetic oxygen production; R = rate of oxygen uptake due to algal respiration; decay rate coefficient for fast BOD (BOD );  = the decay rate coefficient for slow BOD (BOD ); AC = the dry-weight of algal carbon as a concentration;  coefficient for the oxidation of detrital carbon; DC = the concentration of detrital carbon; unit mass of photosynthetic oxygen production; concentration dependence of denitrification;  = nitrification rate coefficient; e4= Concentration dependence of nitrification;  = Denitrification rate coefficient;  = the  = the rate  the amount of ammonium uptake per 3  the amount of ammonium produced per unit mass of oxygen consumed in respiration; e 6= coefficient characterizing the = the heat transfer coefficient;  phytoplankton; A= Algal biomass concentration, mg/l;  = the reaction rate coefficient for hydrolysis; rphtn = the nutrient to carbon ratio for nitrogen for  = Benthos source rate for NH3-N, mg/  -day;  = fraction of algal nitrogen uptake from the available ammonium;  =  the source rate of phosphorous from the sediments; T= Temp;  = Concentration of the NH3-N solute entering the reach;  reach; = Concentration of the L solute entering the reach;  = Rate coefficient for the increase in BOD due to algal death; Chla= concentration of chlorophyll-a representing  the algal (biomass) concentration;  = Rate or re-suspension;  = Concentration of the NO3-N solute entering the  = Ambient air temperature; e6= coefficient characterizing the concentration dependence of nitrification; N/A=  Not available  154  Appendix J: Seasonal creek water flow rate prediction with Manning‟s formulation Month  Days  Water depth (d), m  Flow area (A=b*d),  Perimeter (P=b+2*d), m  Wetted perimeter (R=A/P), m  (Q), m3/s  Creek flow rate, liter/ month  September  30  0.221  0.829  4.192  0.198  0.663  1.72E+09  October  31  0.422  1.583  4.594  0.344  1.833  4.91E+09  November  30  0.700  2.625  5.150  0.510  3.948  1.02E+10  December  31  0.606  2.273  4.962  0.458  3.182  8.52E+09  January  31  0.576  2.160  4.902  0.441  2.948  7.90E+09  February  28  0.466  1.748  4.682  0.373  2.135  5.17E+09  March  31  0.425  1.594  4.600  0.346  1.853  4.96E+09  April  30  0.349  1.309  4.448  0.294  1.365  3.54E+09  May  31  0.288  1.080  4.326  0.250  1.009  2.70E+09  June  30  0.229  0.859  4.208  0.204  0.702  1.82E+09  July  31  0.146  0.548  4.042  0.135  0.340  9.12E+08  August  31  0.143  0.536  4.036  0.133  0.329  8.81E+08  1- Q= A  ; n=0.03 (for clean and straight natural channel); s=slope=1/200, b= average width=3.75m  155  Appendix K: Description of box-and-whisker plot  156  Appendix L: Sub index value for monitored data, predicted initial and final creek water quality parameter TSS  Tem  TN  TP  FC  TC  BOD5  Pb  sub-index  sub-index  sub-index  sub-index  sub-index  sub-index  sub-index  sub-index  Monitoring data 0.9737  1.0000  0.8450  0.4000  0.3883  0.0000  0.5484  0.9910  1.0000  1.0000  0.0000  0.0800  0.3670  0.0000  0.6400  0.9820  1.0000  1.0000  0.3650  0.0000  0.3458  0.0000  0.8000  0.9820  1.0000  1.0000  0.0000  0.3340  1.0000  1.0000  0.7200  0.9730  0.8806  1.0000  0.3000  0.0800  0.3970  0.0000  0.4683  0.5680  0.8673  1.0000  0.0000  0.0000  0.2884  0.0000  0.4569  0.9460  0.5280  1.0000  0.4175  0.0800  0.3057  0.0000  0.2640  0.9280  0.5920  1.0000  0.6350  0.0800  0.5071  0.5995  0.2960  0.9460  0.9205  1.0000  0.6775  0.0000  0.3391  0.0000  0.5027  0.9820  1.0000  1.0000  0.0000  0.0000  0.3381  0.0000  0.7200  0.9640  1.0000  1.0000  0.7200  0.4660  0.3843  0.0000  0.5942  0.9910  1.0000  1.0000  0.8525  0.5320  0.5609  0.7595  0.8800  0.9910  Initial value with present strategies (predicted with the model) 0.9737  1.0000  0.8450  0.4000  0.3883  0.0000  0.5484  0.9910  0.9006  1.0000  0.8224  0.3618  0.3811  0.0000  0.5485  0.9788  0.8889  1.0000  0.8236  0.3797  0.3826  0.0000  0.5551  0.9754  0.8920  1.0000  0.8302  0.4089  0.3858  0.0000  0.5625  0.9740  0.8928  1.0000  0.8372  0.4405  0.3884  0.0000  0.5698  0.9728  0.8926  1.0000  0.8433  0.4690  0.3908  0.0000  0.5764  0.9711  0.8863  1.0000  0.8484  0.4969  0.3921  0.0000  0.5830  0.9682  0.8764  1.0000  0.8471  0.5014  0.3927  0.0000  0.5861  0.9645  0.8578  1.0000  0.8422  0.4961  0.3915  0.0000  0.5876  0.9620  157  TSS  Tem  TN  TP  FC  TC  BOD5  Pb  sub-index  sub-index  sub-index  sub-index  sub-index  sub-index  sub-index  subindex  0.8290  1.0000  0.8316  0.4742  0.3883  0.0000  0.5865  0.9574  0.7767  1.0000  0.8129  0.4306  0.3818  0.0000  0.5819  0.9515  0.6392  1.0000  0.7616  0.2948  0.3660  0.0000  0.5630  0.9386  Predicted final value with proposed SWP strategies 0.9737  1.0000  0.8450  0.4000  0.3883  0.0000  0.5484  0.9910  0.9904  1.0000  0.8531  0.4433  0.3936  0.0000  0.5647  0.9829  1.0000  1.0000  0.8627  0.4837  0.3992  0.0000  0.5757  0.9807  1.0000  1.0000  0.8715  0.5189  0.4044  0.0000  0.5842  0.9800  1.0000  1.0000  0.8802  0.5537  0.4094  0.0000  0.5923  0.9793  1.0000  1.0000  0.8882  0.5858  0.4144  0.0000  0.6002  0.9782  1.0000  1.0000  0.8966  0.6202  0.4197  0.0102  0.6579  0.9764  1.0000  1.0000  0.9017  0.6425  0.4243  0.0481  0.7036  0.9740  1.0000  1.0000  0.9058  0.6617  0.4283  0.0774  0.7479  0.9724  1.0000  1.0000  0.9078  0.6740  0.4312  0.0930  0.7863  0.9693  1.0000  1.0000  0.9072  0.6788  0.4322  0.0889  0.8205  0.9653  1.0000  1.0000  0.8963  0.6520  0.4267  0.0151  0.8374  0.9563  TSS- total suspended solids; TN- total nitrogen; TP- total phosphorous; FC- faecal coliform; TC- total coliform; BOD5- 5 day biochemical oxygen demand; Tem-Temperature; Pb-Lead  158  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items