Development of a Wet Weather Flow Management Master Plan for the City of Toronto

Transcription

1 Water Qual. Res. J. Canada, 2004 Volume 39, No. 4, Copyright 2004, CAWQ Development of a Wet Weather Flow Management Master Plan for the City of Toronto Michael D Andrea,* William J. Snodgrass and Patrick D. Chessie City of Toronto, 55 John Street, Toronto, Ontario M5V 3C6 The City of Toronto has developed a Wet Weather Flow Management Master Plan incorporating a new philosophy in wet weather flow management where rainwater is recognized as a resource. Wet weather flows are to be managed on a watershed basis, and a hierarchical approach to wet weather flow management is to be used, starting with at source, followed by conveyance and finally end-of-pipe control measures. The study area extended across the City of Toronto, encompassing six major watersheds and the waterfront. The Plan development followed the planning principles of Ontario s Environmental Assessment Act and incorporated broad public and agency consultation. A series of 13 objectives were identified and grouped into four major categories: water quality, water quantity, natural areas and wildlife, and sewer system. An innovative approach was used which integrated hydrologic, hydraulic and water quality predictions from land-based, watershed and lake models, respectively, to assess the effectiveness of various strategies. The receiving water response indicated that source controls and conveyance controls were insufficient to achieve the receiving water objectives of the Plan. This was only possible through the implementation of a comprehensive set of measures consisting of: source controls, conveyance controls, end-of-pipe controls, basement flooding protection works, stream restoration works, shoreline management, enhanced municipal operations and an enhanced public education and community outreach program. Overall benefits expected through the Plan include: swimmable waterfront beaches, control of combined sewer overflows in compliance with provincial requirements, basement flooding protection, protection of the City s infrastructure from stream erosion, restoration of degraded local streams and aquatic habitat, the reduction of algal growth along the waterfront, and improved stream water quality in area watercourses. The implementation cost of the Plan over the 25 years is estimated to be $1.047 billion with an additional $233 million in operational and maintenance costs. Key words: watershed based, master plan, stormwater quality Introduction * Corresponding author; mdandre@toronto.ca Urban development within the City of Toronto and the surrounding regions has resulted in intense pressure on the ecosystem through changes to the hydrologic cycle and the natural environment. The corresponding increase in impervious area, changes in surface grading and pollutant inputs from land use practices have resulted in increases in stormwater runoff and degraded stormwater runoff quality. Wet weather flow also results in combined sewer overflows, infiltration and inflow to the sanitary sewer system causing treatment plant bypasses. These impacts result in the degradation of environmental conditions through increased flooding and erosion, physical destruction of terrestrial and aquatic habitat, reduced stream base flow, postings of unsafe water quality conditions at recreational beaches, nutrient enrichment, contaminated sediments and stressed aquatic communities. In the City of Toronto, these impacts have resulted in Toronto s designation, by the International Joint Commission, as one of 43 polluted areas of concern within the Great Lakes basin (Environment Canada et al. 1989). There have been a number of wet weather flow control initiatives undertaken previously. These actions were driven in large part by the need to address local flooding problems and impacts on recreational beach areas. While source control options have been considered and implemented, to varying degrees, end-of-pipe storage and treatment facilities have been used to achieve higher levels of water quality improvements. Although these initiatives represented significant efforts and provided local environmental improvements, it was recognized that a watershed based strategy was necessary to provide a more comprehensive and consistent approach to mitigating wet weather flow impacts. This paper summarizes the work undertaken in the development of the City of Toronto Wet Weather Flow Management Master Plan and detailed in the supporting technical reports (CH2M HILL Canada Limited, MacViro Consultants Inc. 2003; Aquafor Beech Limited 2003; Marshall Macklin Monaghan Limited 2003; Totten Sims Hubicki Associates 2003; XCG Consultants Limited 2003; Dewey 2003; Snodgrass et al. 2003). 417

2 418 D Andrea et al. Background In 1997, the City of Toronto initiated the development of a watershed-based Wet Weather Flow Management Master Plan (WWFMMP) to address the impacts of wet weather flows defined as: storm sewer discharges, combined sewer overflows, and infiltration and inflow to the sanitary sewer system. The Master Plan was developed in accordance with the planning principles of the Province of Ontario s Environmental Assessment Act incorporating broad public consultation at key decision points. A new philosophy was adopted in the development of the Master Plan which emphasized control of rainwater at source to minimize stormwater runoff. Following the runoff pathways from lot level to receiving waters, a hierarchical approach to stormwater management was used beginning with at source controls (lot level), followed by conveyance controls and then end-of-pipe controls. The study area included the entire watershed area of the six major watersheds which extend across the City of Toronto: Etobicoke Creek, Mimico Creek, Humber River, Don River, Highland Creek and the Rouge River, representing an area of about 2100 km 2 (Fig. 1). However, the Master Plan development focussed on the 640-km 2 area contained within the City of Toronto. The Master Plan development was aimed at achieving 13 objectives, which were validated with input from the public, and grouped into four major categories as follows: Water Quality meet guidelines for water and sediment quality virtually eliminate toxins through pollution prevention improve water quality in rivers and the lake for body contact recreation improve aesthetics Water Quantity preserve and re-establish the natural hydrologic cycle reduce erosion impacts on aquatic habitat and property eliminate or minimize threats to life and property from flooding Natural Areas and Wildlife protect, enhance and restore natural features (e.g., wetlands) achieve healthy aquatic communities reduce fish tissue contamination Sewer System eliminate discharges of sanitary sewage reduce infiltration and inflow to sanitary sewers reduce basement flooding Computer Simulation Modelling The computer simulation models were developed to determine the impact of storm sewer and combined sewer overflow discharges on the rivers and streams and to the waterfront. The models, once calibrated, were used to determine existing conditions and the effectiveness of individual wet weather flow control measures, combinations of wet weather flow control options, and alternative wet weather flow management strategies on the stream flow regime and water quality within the six watersheds, flows and water quality discharges from the combined sewer service area, and water quality impacts along the waterfront of Lake Ontario, respectively. The U.S. EPA s Hydrologic Simulation Model HSP-F was used to continuously simulate hydrologic processes as well as pollutant generation and transport for the separated sewer catchments within the watersheds (Aquafor Beech Limited 2003; Marshall Macklin Monaghan Lim- Fig. 1. Study area.

3 Toronto Wet Weather Flow Management Master Plan 419 ited 2003; Totten Sims Hubicki Associates 2003; XCG Consultants Limited 2003; Bicknell et al. 2001). The Dorsch QQS model (CH2M HILL Canada Limited, MacViro Consultants Inc. 2003; Dorsch Consultants et al. 1989) was used to continuously model urban runoff in the combined sewer areas within the Don River and Humber River watersheds, respectively. The QQS model, along with computing runoff hydrographs and pollutographs at any location in the storm and/or combined sewer system, also provided frequencies, volumes and duration of stormwater and CSO discharges. To provide consistency and to determine the full impact, on a watershed basis, the parameters in the HSP-F model applicable to the combined sewer service area of the Don and Humber watersheds were adjusted so that the hydrologic and water quality responses matched the results from the Dorsch QQS model. The RAND two-dimensional hydrodynamic and water quality simulation model (Leendertse 1970) was used to assess water quality impacts along the nearshore area of the City of Toronto waterfront (Dewey 2003). The RAND model was subsequently replaced with the Danish Hydraulic Institute (DHI) Mike 3D model (Danish Hydraulic Institute 2002) in the waterfront analysis because it could represent the thermal stratification of Lake Ontario and better represent vertical mixing of pollutant plumes from all waterfront discharges. A whole lake version of the model was used to establish boundary conditions for a finer scale model of the Toronto waterfront. This waterfront model was used to predict lake circulation patterns and water quality impacts along the waterfront. The waterfront model utilized the time series of flow and associated water quality concentrations from the HSP-F watershed models and from the combined sewer area Dorsch QQS model as inputs. Model Abstraction for Urban Flow Pathways Considerable effort was required to model the hydrological effects of source (lot level), conveyance and end-ofpipe wet weather flow control measures. Typically, for these types of studies, models are set up using lumped parameters for catchment areas in the watershed. Modelling of the control options would be accomplished by adjusting the lumped parameters to reflect the inclusion of control measures within a subcatchment. As the ultimate goal of the master plan development was to identify a suite of wet weather flow control measures, following the hierarchy of source, conveyance and end-of-pipe controls which met the study objectives, it was concluded that the lumped parameter approach was too empirical, and could not adequately account for the different combinations of control measures and flow paths. For example, at the lot level, there are a number of ways by which rainwater falling on a roof area or the lawn of an individual lot can reach a stream. Rainwater on a lawn may infiltrate to groundwater slowly and be released as baseflow or may infiltrate to the building foundation drains and then flow directly to the storm, sanitary or combined sewer system. Rainwater on a roof can flow through the roof downspouts and then directly to the storm, sanitary or combined sewer systems, or if disconnected from the sewer system, flow onto the lot, ideally a pervious area, where it can infiltrate. Proper representation of urban process runoff needs to account for the fact that any given land use can exist in combination with various native soil types. As well, within any of the land-use categories, there can be variations in internal drainage that can significantly affect local hydrologic response. The models had to be set up to represent the unique combinations of 20 land-use types, 3 types of dominant soil classifications (AB, BC and CD), and the internal drainage connectivity to be able to represent specific control measures applied only in certain conditions. A unique approach was adopted to model source controls based upon unit response functions (URFs). The concept identified a set of representative test catchments of approximately ten hectares consisting of a homogeneous land use density, soil type, and flow pathway (i.e., roof downspouts connected to storm sewers) combination. For each of the test catchments, a URF was developed as a time series of the hydrologic and water quality response of the area to a predetermined series of meteorological inputs. In this study, 1991 was determined to be an average rainfall year and was used to generate the URFs. A total of just under 100 unit response functions (URFs) were created. Geographic information system (GIS) mapping was used to establish subwatershed areas down to 200 hectares in size within the City. A coarser discretization was used in the watershed areas outside the City. Figure 2 shows the subwatershed delineation for the Don River. Through GIS, the corresponding land use make-up, soil type and flow pathway were identified to establish the URFs applicable to each subwatershed and the appropriate percentage of the subwatershed each URF represented. Figure 3 shows the land-use distribution and the number of URFs (land use/soil type/flow pathway combinations) applicable to each land use category for a representative subwatershed within the Don River (shown as the shaded area in Fig. 2). Over 90 potential stormwater management and combined sewer overflow control measures were compiled for consideration in the development of the alternative strategies (Totten Sims Hubicki Associates 2003a,b). They were categorized as source controls, conveyance controls, endof-pipe controls, maintenance and operational practices, and special measures which provide a watershed or stream improvement, such as the removal of fish barriers. Since forty-five of the 90 wet weather flow control measures were source control measures, it was impracti-

4 420 D Andrea et al. Fig. 2. Subwatersheds in Don River watershed. cal to model each source control measure individually. As many of the source control measures function in a similar manner and produce similar results, representative measures were modelled to predict the expected effect of the broader range of options. The selection of representative options was based on the categorization of their primary functions described as follows: Flow attenuation through temporary storage, i.e., using a rain barrel Reduction in impervious area, i.e., roof leader disconnection Enhancement of infiltration capability, i.e., replacing standard paved driveways with pervious pavement, and Pollutant source reduction, i.e., elimination of fertilizer and pesticide use The configuration and setup of the QQS model was not amenable to the method and techniques of the URF approach for simulating source controls in the combined sewer area. The source controls were simulated in the QQS model using three parameters: impervious ratio, initial abstraction and re-distribution of the unit hydrograph to simulate the effects of reduction in transportconnected imperviousness. To produce consistency of results with the HSP-F simulations, the QQS source control modelling results were compared with the HSP-F model URF results for similar land use and soil types. The parameters in the QQS model were adjusted until the HSP-F model results and the QQS model results provided the same level of improvements for the various levels of controls. Modelling of conveyance controls was accomplished by adjusting the infiltration parameters to account for local soil conditions as conveyance controls consisted primarily of infiltration/exfiltration measures along the storm drainage collection system. Annual average BMP pollutant load reduction factors were utilized to represent the effects of introducing end-of-pipe controls within stormwater drainage systems. Water Quality Constituents and Information Sources The computer simulation models were set up to simulate in-stream concentrations within each of the watercourse reaches, and the waterfront of Lake Ontario. Eleven water quality parameters were simulated: total phosphorus, total suspended solids, total Kjeldahl nitrogen, nitrate plus nitrite, E. coli bacteria, total lead, total copper, total zinc, dieldrin, benzo(ghi)perylene, and temperature. For urban areas, the event mean concentration (EMC) approach was used (WEF/ASCE 1998) to establish

5 Toronto Wet Weather Flow Management Master Plan 421 Fig. 3. Land-use and URFs breakdown in a representative subwatershed of the Don River watershed. stormwater runoff and CSO concentrations. To simulate the pollutant load delivered to watercourses by urban stormwater, the modelling methodology utilized the annual average EMCs attributed to stormwater runoff times the event volume of runoff calculated by the hydrological model. A substantial amount of data on wetweather EMCs for stormwater and CSO discharges within the Toronto area was used in conjunction with local dryweather storm sewer discharge data (Paul Theil Associates Ltd., Beak Consultants Ltd. 1995; Maunder et al. 1996; Snodgrass and D Andrea 1993). A summary of the event mean concentrations in storm sewer and combined sewer overflow discharges, collected within the City of Toronto, are presented in Fig. 4 and compared to Ontario s Provincial Water Quality Objectives (PWQOs) (Ontario Ministry of the Environment 1994), where they exist. In Fig. 4, pollutant concentrations in storm sewer discharges are comparable to CSO discharges for most parameters. Furthermore, water quality constituents in most cases, are well above both the PWQOs and the Canadian Council of Resource and Environmental Ministers Guidelines (Canadian Council of Ministers of the Environment ). Appropriate EMC values were applied to surface runoff discharges as well as to interflow and groundwater discharges to simulate the total pollutant loading from each subcatchment. Subcatchment pollutant loadings were then routed through the watercourse reach network. No in-stream pollutant transformation or assimilation processes were simulated for any of the water quality parameters except for E. coli, which was subjected to time-dependent first-order decay. Fig. 4. Constituent concentrations in storm sewer and combined sewer overflow discharges.

6 422 D Andrea et al. Model Calibration/Validation The HSP-F model was calibrated for flow characteristics and water quality constituents based on data from the Water Survey of Canada stream gauging stations for the period 1991 to 1996 and an extensive water quality monitoring program undertaken in the 1990s (D Andrea et al. 1999; Paul Theil Associates Ltd., Beak Consultants Ltd. 1995; Maunder et al. 1996; Snodgrass and D Andrea 1993). The quantity calibration for the watersheds concentrated on both overall volumes of water discharged annually (Table 1), and the flow frequency distribution. A representative plot of the flow frequency for one watershed (Mimico Creek) is presented in Fig. 5. Overall, there is excellent agreement with the total annual volumes discharged, and the flow frequency distribution. The water quality model was calibrated with observed concentration measurements that included complete wet weather events and dry weather inter-event periods. Different criteria can be used to determine goodness of fit in model calibration/validation studies. In this study, a line of evidence approach was used as the criteria for assessing the acceptability of the water quality calibration for the watercourses: the measured EMC s values were adjusted within their measured uncertainty estimates such that the calculated in-stream concentrations at the mouth of the six major tributaries were within their measured uncertainty estimates, defined as the 95% confidence interval. A representative comparison of model predictions to measured annual average in-stream concentrations is presented in Fig. 6 for total phosphorus. The dashed lines represent the 95% confidence interval for mean wet weather and dry weather concentrations, represented by solid lines, respectively (D Andrea et al. 1999; Snodgrass and D Andrea 1993). Model predictions match under dry weather conditions and are in good agreement with measurements for wet weather. The Dorsch QQS model calibration was undertaken using sewer flow data collected by the City of Toronto and event mean concentrations for representative storm sewer and combined sewer overflow discharges (Paul Theil Associates Ltd., Beak Consultants Ltd. 1995; Maunder et al. 1996). A representative plot of the average monthly flow volumes for one sewershed (York- Rockcliffe) is presented in Fig. 7. The total annual pollutant loadings at the CSOs and storm sewer outlets were calculated based on the mix between storm and sanitary flows at every outfall. The model was calibrated using measured water quality data for three CSOs discharging to the lakefront (Parkside Drive, Cowen Avenue, Greenwood Avenue). The model predictions were found to be in good agreement with field measurements as shown for representative parameters total phosphorus and E. coli in Fig. 8. Calibration/validation studies were undertaken for the RAND model and subsequently using the Danish Hydraulic Institute Mike 3D (DHI) model using physical 1993 current meter (current speed and direction) data, water treatment plant intake temperature data, and waterfront water quality data (Dewey 2003). Results from representative physical studies for the DHI model are presented in Tables 2 and 3, and in Fig. 9. Three model setups were assessed using vertical depth layers of 6, 10 and 20 m, respectively, for the surface to 40 m of the water column of Lake Ontario. The directions predicted at the Ashbridges Bay location are compared with the 1993 current meter data in Table 2. The most accurate simulations were observed with the 10 m layer setup in all categories (Dewey 2003). The predicted temperatures were compared to 1993 daily water temperature. Figure 9 shows the temperature response of the model along with the measured data. The observed data show large variations over the summer due to upwelling and downwelling events. There is a long period in late July to early August when the temperature is quite low, and then two shorter episodes in early September and two strong events in late September when the temperature drops down to 5 C. The six-layer simulation produced generally warmer temperatures and did not well represent the observed upwelling/downwelling events. The 10 and 20 layer models replicated the overall seasonal trends and the stronger upwelling/downwelling events. The 20 layer model had a higher degree of accuracy with respect to the temperature changes. The 20 layer model results have a correlation coefficient of 0.77 with the intake temperature data (Dewey 2003). For calibration purposes, a 12 layer model was used. A representative calibration for E. coli bacteria is presented in Table 3. The model predictions are very close to the observed data. TABLE 1. Summary of water quantity results using HSPF (volumes in m 3 x 10 6 ) Mimico Humber Don Highland HSPF HSPF HSPF HSPF Target model Target model Target model Target model Calibration period total volume ( ) Validation period total volume ( ) Typical year total volume (1991)

7 Toronto Wet Weather Flow Management Master Plan 423 Fig. 5. HSP-F model calibration for Mimico Creek flow frequency ( ). Fig. 7. York-Rockcliffe overflow monthly flow volumes. Receiving Water Response to Wet Weather Flow Management Options Due to the plethora of potential combinations of the different technologies and management practices, the benefits of major types of controls were assessed in two major sequences: a bundling analysis which emphasized the implementation of source and conveyance controls and strategies which incorporated all types of controls implemented at varying levels of effort. Bundling of Source Controls and Conveyance Controls The hierarchical approach was adopted in the study for the first set of analyses, whereby source control measures were considered first, followed by conveyance controls and then by end-of-pipe controls to control wet weather flows. Groups of source control and conveyance control management options or bundles were developed to represent increasing levels of effort to reveal how much additional benefit would be achieved as the level of effort increased and to reveal the degree to which the study objectives would be achieved. Three levels of effort; Limited Effort, Moderate Effort and Maximum Effort were assessed and defined as: minimum source control (e.g., roof downspouts and minor lot regrading) maximum source control (e.g., minimum source control plus major lot regrading and foundation drain disconnection and porous pavement driveways, etc.), and maximum source control plus conveyance control (e.g., maximum source control with combined sewer separation and stormwater infiltration systems along the municipal right of way where soil conditions were favourable). The water quantity response of the watersheds (Lura et al. 2002) to the different bundles of control options is presented in Table 4, using the ratio of dry weather flow to total annual flow, as the response indicator of how much water infiltrates relative to the total volume of runoff from a watershed. Controls that reduce runoff and promote infiltration can have a substantial hydrologic benefit by increasing baseflow and reducing flow peaks. Ecologically, the ratio of baseflow to average annual flow provides an assessment of the degree to Fig. 6. Comparison of water quality predictions to measured data by watershed total phosphorus.

8 424 D Andrea et al. Fig. 8. Water quality calibration for total phosphorus and E. coli. which watershed hydrology is regulated by groundwater flow and the potential for providing habitat for sensitive fish species. In this study, the ratio is used more as a hydrological indicator, than an ecological indicator. The results indicate that minimum-effort source controls have minimal benefit in terms of improving baseflow conditions. Maximum-effort source controls have a larger effect because this bundle includes more measures that result in infiltration of runoff. But even for the maximum source control bundle, there is only a marginal effect predicted in the flow regimes of Etobicoke Creek and the Don, Humber, and Rouge Rivers because between 84 to 90% of their watersheds lie outside the City of Toronto boundary. A more significant response is observed for the Mimico and Highland Creek watersheds because 50 and 90%, respectively, of these two watersheds are contained within the City, where source controls were assumed to be implementable. The response to the bundles of control options for two watersheds (Highland Creek and the Rouge River) (Aquafor Beech Limited 2003) are presented in Fig. 10 for two water quality parameters (total phosphorus and E. coli). The predicted wet weather, dry weather, and average concentrations are compared to the corresponding PWQOs. The PWQOs represent the significant enhancement receiving water target. These results show that much greater improvements are gained in the Highland Creek watershed where 90% of the watershed is within the City s jurisdiction compared to the improvements gained in the Rouge River where only the lower 10% of its watershed is within the City s jurisdiction. However, none of the source and conveyance control bundles are capable of achieving the study s in-stream water quality targets (PWQOs). The predicted water quality response across the waterfront for E. coli due to implementation of the source controls and conveyance controls in the watersheds is presented in Fig. 11. The simulation period for E. coli covers the summer swimming season from June 1 to September 30 (2928 hours). The results are presented in terms of percent exceedance of the PWQO of 100 EC counts/dl at locations along the waterfront that are either beaches or areas of ecological concern. As illustrated in Fig. 11, loadings to areas such as the Inner Harbour, where there are direct discharges of CSO, can increase as more source con- TABLE 2. Predicted current direction based on depth of layers Layers 6 m 10 m 20 m Angle % Accurate % Accurate % Accurate TABLE 3. Eastern beaches observed and predicted exceedances of PWQOs simulation period for E. coli June 1 to September 30 (2928 hours) Percent Percent exceedances exceedances Beach observed predicted Woodbine Beaches Park Kew Beach Balmy Beach Coatsworth Cut

9 Toronto Wet Weather Flow Management Master Plan 425 Fig. 9. Predicted and observed temperatures at R.C. Harris intake in trols and conveyance controls (sewer separation) are applied due to the fact that increasing amounts of runoff are being diverted from the combined sewer system to the separate storm system. With runoff diverted from the combined sewer system to the storm sewer system, the wet weather flow that was previously intercepted from the combined sewers and treated at a wastewater treatment plan is now discharged untreated from the storm sewer system. While each of the stormwater management bundles provide some improvement in water quality as shown in Fig. 10 and 11, none of the bundles achieve the target of exceeding the PWQO for E. coli 5% of the time (146 hours). The following are the key findings from the bundling assessment: In many cases, the impacts of wet weather flows are so severe that substantial reductions in pollutant loadings, significant moderation of peak flow and significant increase in baseflow are required to meet the proposed targets for healthy watersheds and the waterfront. Source control and conveyance control measures provide limited improvements, for the watersheds and for the waterfront, at substantial costs (i.e., estimated $12 billion for the maximum source control bundle). A combination of stormwater management options, especially end-of-pipe controls are required in order to achieve the WWFMMP objectives. Improvements in water quality and stream baseflow were more significant within subwatersheds wholly contained within the City. For most substances, inputs upstream of the City from stormwater discharges are as important as a source of pollutants to the rivers and waterfront as inputs from within the City. In watersheds such as the Etobicoke Creek and Rouge River, upstream sources are the dominant source of pollutants. For substances such as E. coli in certain watersheds, sources within the City are shown to be significant pollution sources to the waterfront. The outcome of the bundling analysis determined that source controls alone would not be sufficient to TABLE 4. Change in dry weather flow/annual flow for different bundles of control options Change in ratio for specified bundle Ratio of dry weather flow/annual flow Minimum Maximum Maximum source + Location existing conditions source control source control conveyance controls Humber River % 4% 4% Don River % 7% 8% Highland Creek % 17% 18% Rouge River % 9% 9% Etobicoke Creek % 9% 9% Mimico Creek % 21% 21%

10 426 D Andrea et al. Fig. 11. Percent of time E. coli levels exceeded PWQO of 100 counts/dl along waterfront beach areas (see Fig. 16 for locations). Fig. 10. Highland Creek and Rouge River water quality results for total phosphorus and E. coli. meet all of the targets. The lack of effectiveness of the source control bundles and their costs directed the development of five alternative strategies consisting of combinations of source controls, conveyance controls, end-ofpipe controls, and stormwater treatment technologies capable of achieving the management objectives of the WWFMMP. Alternative Strategies Through a strategic planning process a set of five alternative strategies was formulated. Each of the strategies was developed, in accordance with the adopted hierarchical principle for stormwater management, and varied in terms of three levels of targets: status quo, moderately enhanced and significantly enhanced, to reflect increasing levels of ecological enhancement. The three levels of targets and the set of assumptions used to develop the strategies, across each watershed area, are summarized in Table 5 and described in the following: Strategy 1: (Maintain status quo) involved the implementation of stormwater best management practices (BMPs) to maintain existing environmental conditions with future intensification and additional suburban growth. Strategy 2: (Opportunistic) involved the implementation of opportunistic BMPs (i.e., those BMPs that could be implemented as opportunities arise). Strategy 3: (Strive for moderate targets with an end-ofpipe focus) involved the implementation of BMPs that strive to achieve moderate environmental improvements based on a voluntary uptake for source controls and more emphasis on end-of-pipe controls. Strategy 4: (Strive for moderate targets with a source control focus) involved the implementation of BMPs that strive to achieve the same level of TABLE 5. Elements of the alternative strategies Strategy Target Source Conveyance number level controls controls End-of-pipe Dry weather flow 1 Existing As required As required Existing Existing for infill for infill 2 Not specific Voluntary Time limited Opportunistic Limited improvement (40% uptake) exfiltration green (based on search and and filtration infrastructure destroy program) 3 Moderate Voluntary Time limited Opportunistic Limited enhancement (40% uptake) exfiltration and and aggressive improvement filtration facilities 4 Moderate Enhanced Enhanced Opportunistic Extensive improvement enhancement (70% uptake) exfiltration and green (based on long-term filtration (not infrastructure infrastructure time limited) replacement) 5 Significant Enhanced Enhanced Opportunistic Extensive enhancement (70% uptake) (not time limited) and aggressive improvement facilities

11 Toronto Wet Weather Flow Management Master Plan 427 moderate environmental improvements as Strategy 3 but with a focus on an enhanced level of uptake for source controls and less focus on end-of-pipe controls. Strategies 3 and 4 helped to illustrate the differences in environmental impacts and types of BMPs required when voluntary versus enhanced uptake for source controls are applied. Strategy 5: (Strive for enhanced targets) involved the implementation of enhanced levels of source, conveyance and end-of-pipe control measures that strive to achieve significant environmental improvements, such as achieving Provincial Water Quality Objectives. Consistent with the requirements of the Environmental Assessment Act the alternative strategies were assessed using evaluation criteria that incorporated the effectiveness in meeting the objectives of the WWFMMP as well as social/cultural, environmental and economic factors. The water quantity and quality response to the alternative strategies were determined within the mainstem of each watercourse, their tributaries and along the waterfront (CH2M HILL Canada Limited, MacViro Consultants Inc. 2003; Aquafor Beech Limited 2003; Marshall Macklin Monaghan Limited 2003; Totten Sims Hubicki Associates 2003; XCG Consultants Limited 2003; Dewey 2003; Snodgrass et al. 2003). The change in baseflow as a percentage of mean annual flow provided an indication of how well the whole hydrological cycle was balanced (i.e., to reduce the amount of peak runoff and increase the amount of baseflow). The overall goal was to increase the ratio of baseflow to mean annual flow. Figure 12 summarizes the ratio of baseflow to mean annual flow as a percentage at the mouth of each watercourse for each of the five alternative strategies. The water quality response to the five alternative strategies is represented in Fig. 13 for total phosphorus at the mouth of each watershed and in Fig. 14 for E. coli bacteria levels at the waterfront, respectively. Fig. 13. Comparison of average wet weather total phosphorus response in the mainstem of each watershed (mg/l). The figures reflect the ecological enhancements predicted for the strategies where the most significant improvements are predicted for Strategy 5. The most significant improvements were within the Highland Creek watershed because it is essentially contained within the City of Toronto where the city has jurisdiction to apply the proposed control measures throughout the entire watershed. The smallest improvements were in the Humber River, Etobicoke Creek and Rouge River where only 10 to 16% of the watersheds are contained within the City s boundaries. However, in the subwatersheds wholly contained within the City, similar improvements to those predicted within the Highland Creek watershed were also obtained. Cost Effectiveness of the Strategies A cost-effectiveness relationship for a typical subwatershed (Marshall Macklin Monaghan Limited 2003) is presented in Fig. 15. The graph presents an assessment of each strategy in terms of cost and effectiveness in achieving the study objectives, based on technical scores (e.g., achievement of PWQOs) and total scores (based on the evaluation criteria), respectively. In general, no single strategy stands out as the optimum and nor is a kink in the curve observed to identify a point of diminishing improvements as more funds are invested. Rather, the curve shows a relationship where environmental quality continues to improve as more funds are invested. Fig. 12. Comparison of baseflow to mean annual flow response (%). Fig. 14. Percent exceedance of PWQO for E. coli along waterfront beach areas (see Fig. 16 for locations).

12 428 D Andrea et al. Fig. 15. Cost effectiveness of strategies for a representative subwatershed. Based on these curves, it was concluded that the cost-effectiveness curves for the strategies did not provide a definitive basis for selecting a preferred strategy, from among the five strategies. In addition, the total score based on an evaluation of social/cultural and other environmental and economic factors indicated that the strategies did not have a significant negative impact that would preclude selection of any of the strategies. Overview of the Master Plan In accordance with the requirements of the Environmental Assessment Act, the alternative strategies were assessed using evaluation criteria established with input from the public. Based on the adopted study principles, public feedback and the results of the evaluation process, Strategy 5 was selected as the preferred wet weather flow management strategy. The preferred strategy, aimed at achieving the ambitious goal of meeting Provincial Water Quality Objectives in area surface waters, may take 75 to 100 years to implement, at a cost estimated to be $12 billion. Consistent with the planning horizon for Master Plans, a 25-year implementation plan was developed from the preferred strategy, directed at achieving the City s corporate priorities of health and safety (i.e., eliminate basement flooding and provide swimmable water quality at waterfront beaches), infrastructure protection (i.e., prevent stream erosion) and renewal (i.e., eliminate dry weather discharges), intensification (i.e., accommodate growth projected in the City s Official Plan) and the legislative requirements to eliminate combined sewer overflows (i.e., satisfy the Ontario Ministry of the Environment Procedure F-5-5, Ontario Ministry of the Environment, undated). Figure 16 presents an example of the improvements expected, with the implementation of the 25-year Master Plan, along the waterfront, in terms of the number of hours in which E. coli levels exceed the PWQO for body contact recreation (Dewey 2003). Fig. 16. Comparison of exceedances of E. coli objective for existing conditions and the 25-year master plan.

13 Toronto Wet Weather Flow Management Master Plan 429 The Master Plan also addresses the objectives of the City s Environmental Plan regarding water quality improvements and advances the water quality improvement objectives of the Toronto and Region Remedial Action Plan. Overall benefits projected through the implementation of the Master Plan are substantial across the City and include: swimmable waterfront beaches elimination of combined sewer overflows in compliance with provincial requirements elimination of dry weather discharges basement flooding protection protection of the City s infrastructure from stream erosion restoration of degraded local streams and improved stream water quality reduction of alga growth along the waterfront and in streams, and restoration of aquatic habitat. In support of the Master Plan, an implementation schedule (Chessie et al. 2003) was developed prioritizing the list of projects to be implemented over the next 25 years, with an overall objective of striving to meet the moderate level targets. The Plan will be reviewed and its effectiveness in achieving the targets established will be assessed on a 5-year cycle allowing for adjustments to the Plan, if necessary. Regular updating will also permit a review of new and emerging technologies for their applicability and incorporation into the Plan. Table 6 contains a summary of the proposed measures contained within the Master Plan to be implemented over the next 25 years and their associated costs. TABLE 6. Twenty-five year master plan summary Component of the plan Capital cost ($ million) Public Education city-wide over 25 years 30 Focussed on increasing public awareness Source Controls city-wide over 25 years 112 Existing ~10 15% participation rate Target of 40% participation rate is proposed Municipal Operations city-wide over 25 years 8 Search and eliminate dry weather discharges Enhanced street sweeping and catchbasin cleaning Monitoring of Plan implementation and effectiveness Conveyance Controls city-wide over 25 years 74 Protect existing ditch network Sewer exfiltration systems ( leaky storm sewers) Basement Flooding emphasis in first 5 years 55 Focussed on cluster areas previously identified Sewer system upgrading and home isolation program Shoreline Management implemented in first 10 years 42 Humber River and Etobicoke Creek deflector structures Restoration of Highland Creek and Rouge Park Marshes Stream Restoration emphasis in first 15 years 131 Focus on protecting City s infrastructure Restore aquatic stream habitats 104 km of stream restoration proposed End-of-Pipe Controls implemented over 25 years 105 Green end-of-pipe: stormwater ponds, constructed wetlands opportunistic basis where sufficient open space available 180 facilities proposed Underground storage: space limited considerations 491 necessary to address combined sewer overflows 16 CSO facilities proposed 50 stormwater facilities proposed 4 CSO treatment facilities proposed Total capital cost 1047 Operation and maintenance cost 233 (associated the new stormwater control measure)

14 430 D Andrea et al. The total capital cost and associated operational and maintenance costs for the Master Plan over the 25-year implementation period is estimated at $1.047 billion and $233 million, respectively. To support this expenditure, various funding options including increasing water rates, levying property taxes, implementing user charges apportioned to the percent impervious area of an individual lot, development charges and grants/subsidies, etc., are being assessed as potential sources of revenue to finance the Plan. The financial impact on the average home owner has been estimated to be in the range of $30 to $70 per year depending on the funding option. Conclusion An overview of the process followed in developing a watershed-based, 25-year Wet Weather Flow Management Master Plan and longer term (100 year) strategy to mitigate the impacts of wet weather flow in the City of Toronto has been presented. The overall goal of the Plan is to meet Provincial Water Quality Objectives within the City of Toronto surface waters. The receiving water response indicated that a combination of source and conveyance controls were insufficient to achieve the receiving water objectives of the Plan. This was only possible through the implementation of a comprehensive set of measures consisting of source controls, conveyance controls, end-of-pipe controls, basement flooding protection works, stream restoration works, shoreline management, enhanced municipal operations and an enhanced public education and community outreach program. While achieving Provincial Water Quality Objectives was shown to be possible for watershed areas wholly contained within the City of Toronto, water quality within watersheds extending beyond the City limits was projected to continue to exceed PWQOs, even with the implementation of stormwater best management practices in all new development areas. Overall benefits expected to be derived through the Plan include swimmable waterfront beaches; control of combined sewer overflows in compliance with provincial requirements; basement flooding protection; protection of the City s infrastructure from stream erosion; restoration of degraded local streams and aquatic habitat and the reduction of algal growth along the waterfront and improved stream water quality in area watercourses. The cost of the Plan over the 25 years is estimated to be $1.047 billion with an additional $233 million in operational and maintenance costs. Acknowledgements We gratefully acknowledge the significant contributions from the consultant leads and their teams: Eric MacDonald, Laurie Boyce, Ray Tufgar, George Zukovs, Rob Bishop, Dave Maunder, Ray Dewey and Dave Dilks. The advice and input provided by Jiri Marsalek, particularly during the modelling phases of the project, are sincerely appreciated. The financial contribution from Environment Canada, through the Great Lakes Sustainability Fund, to help support the development of the Plan is appreciated. References Aquafor Beech Limited Highland Creek and Rouge River watersheds. Study Area 5 of City of Toronto. Wet weather flow management master plan. Bicknell BR, Imhoff JC, Kittle Jr JL, Jobes TH, Donigian Jr AS Hydrological Simulation Program Fortran (HSP-F). Version 12 user s manual. In co-operation with Hydrological Analysis Software Division. Office of Information USGS and National Exposure Risk Assessment Laboratory. U.S. EPA. March Canadian Council of Ministers of the Environment Canadian water quality objectives. CCME, Winnipeg, Manitoba. CH2M HILL Canada Limited, MacViro Consultants Inc Combined sewer service area. Study Area 1 of City of Toronto. Wet weather flow management master plan. Chessie P, Snodgrass WJ, D Andrea M City of Toronto. Wet weather flow management master plan. Overview and implementation plan. D Andrea M, Boyd D, Anderton R Assessment of six tributary discharges to the Toronto area waterfront, Volume 2. Technical Appendix and Data Summary. Ontario Ministry of the Environment, Toronto, Ontario. Danish Hydraulic Institute MIKE 3 - A 3D modelling system for rivers, lakes, estuaries, coasts and oceans. Dewey R Toronto waterfront model. Simulation results for bundling options, strategy evaluation, and 25 year plan. Dorsch Consultants Quality - quantity simulation (QQS). User s manual, Volumes 1, 2 and 3. Environment Canada, Ontario Ministry of Environment, Ontario Ministry of Natural Resources, Metropolitan Toronto and Region Conservation Authority Metro Toronto and Region Remedial Action Plan. Stage I. Environmental conditions and problem definition. Leendertse JJ A water quality simulation model for well-mixed estuaries and coastal seas. Volume 1. Principles of computation. Rand Memo RM 6230-RC, Rand Corporation. Lura Consulting, XCG Consultants Limited, Totten Sims Hubicki Associates, Aquafor Beech Limited, MacViro Consultants Inc., CH2M HILL Canada Limited, Marshall Macklin Monaghan Limited, Modelling Surface Water Ltd Progress update, City of Toronto, wet weather flow management master plan, February Marshall Macklin Monaghan Limited Don River watershed, Study Area 4 of City of Toronto, Wet weather flow management master plan.

15 Toronto Wet Weather Flow Management Master Plan 431 Maunder D, Whyte R, D Andrea M Metropolitan Toronto waterfront wet weather flow outfall study, Phase II. Report prepared for the Metropolitan Toronto and Region Remedial Action Plan. Ontario Ministry of the Environment (MOE). Determination of the treatment requirements for municipal and private combined and partially separated sewer systems, Procedure F-5-5. Supporting document for Guideline F- 5 Level of Treatment for Municipal and Private Sewage Treatment Works Discharigning to Surface Waters. Ontario Water Resources Act, RSO Ontario Ministry of the Environment (MOE) Water management policies, procedures, guidelines. Provincial Water Quality Objectives. Paul Theil Associates Ltd., Beak Consultants Ltd Metropolitan Toronto waterfront wet weather outfall study, Phase I. Report prepared for the Metropolitan Toronto and Region Remedial Action Plan. Snodgrass WJ, D Andrea M Dry weather discharges to the Metropolitan Toronto waterfront. Ontario Ministry of Environment Report. Snodgrass WJ, Dewey R, D Andrea M, Chessie P City of Toronto. Wet weather flow management master plan. Effects of watershed management within the City of Toronto on the Toronto waterfront. Totten Sims Hubicki Associates Etobicoke Creek and Mimico Creek watersheds, Study Area 2 of City of Toronto. Wet weather flow management master plan. Totten Sims Hubicki Associates, Aquafor Beech Limited, MacViro Consultants Inc., CH2M HILL Canada Limited, Marshall Macklin Monaghan Limited, XCG Consultants Limited. 2003a. Wet weather flow control measures, the blue book. Toronto, Ont. Totten Sims Hubicki Associates, Aquafor Beech Limited, MacViro Consultants Inc., CH2M HILL Canada Limited, Marshall Macklin Monaghan Limited, XCG Consultants Limited. 2003b. Review of management and operations practices. Toronto, Ont. WEF/ASCE Urban runoff quality management. WEF Manual of Practice No. 23, ASCE Manual and Report on Engineering Practice No p. XCG Consultants Limited Humber River watershed, Study Area 3 of City of Toronto. Wet weather flow management master plan. Toronto, Ont. Received: July 23, 2004; accepted: September 11, 2004.