Runoff Characteristics

Runoff rate and volume generally increase after urbanization (Fig. 1.1). We pave the watershed with asphalt and houses and straighten and shorten flow paths by conveying runoff through drainage systems. If the downstream channel capacity is exceeded, flood will occur over the floodplain. Another related problem is channel erosion which depends on runoff rate and its duration. This brings out the frequency issue of runoff rate and volume. As indicated in Fig. 1.1, the percent of time that runoff rate exceeds a certain value generally increases after urban development. Thus, urbanization not only increases runoff rate and volume but also their frequency. The frequency of runoff rate has a direct impact on erosion and sediment transport of river channel.

Runoff from non-urban areas carries eroded sediments, nutrients from natural and/or agricultural sources, bacteria from animal droppings, and pesticides and herbicides from agricultural practices. Its quality is a function of rainfall, soils, vegetation, and agriculture practices. After urbanization, runoff carries solids particles from automobile wear and tear, dust and dirt, and winter sand, nutrients from residential fertilizers, metals such zinc, copper, and lead, hydrocarbons leaching from asphalt pavement materials, spilled oils and chemicals, and bacteria from domestic animals. This change of runoff quality causes a general degradation of water quality in the receiving waters (Table 1).

Fig. 1 Runoff hydrographs before and after development

Table 1: Typical runoff quality

Runoff Impacts

Urban runoff problems can be divided into the following categories:

Table 2 Causes and impacts of urban runoff problems

SWM Planning

SWM Objectives

In order to address these objectives effectively, it requires a multi-disciplinary team of engineers, scientists, planners, etc. The level of details that needs to be addressed in SWM planning is dependent upon the size of development. For small developments, an engineering firm may be retained by the developer to address all drainage and stormwater issues. Dependent upon the water management issues at the site, additional environmental professionals may be needed. Nevertheless, there may be a lot of resistance to hire all these professionals in SWM. The benefits of SWM planning on a watershed basis become obvious for all these small developments. Once the SWM design criteria are identified through the watershed and subwatershed process, it reduces the burden for each of these small developments to hire its own environmental professionals. Fig. 2 shows some of the SWM objective..

Fig. 2 Stormwater Management Objectives

Major and Minor Drainage System Principles

The interface between the minor and major drainage system is the catch basin. Catch basins should be designed to capture all of the flows up to the design capacity of the sewer (e.g., runoff from 5 year storm). When weepers are connected to the storm sewer, the basins should be designed to capture no more than this amount of runoff to prevent surcharge of the storm sewer. The operation of a catch basin inlet determines the fraction of runoff that enters the sewer system or travel along the street. The amount of runoff captured is controlled by the magnitude of the overland flow, the grating types, and the flow conditions in the storm sewers.

Inlet control devices are often used to control the peak flow entering the storm sewer so that it will not be surcharged. Inlet control devices usually limit the flow to less than the leader capacity so free surface flow is maintained in the leader. Inlet control design is governed by the existing surface grade. Surface flow should not only be restricted from the sewer system at critical points but also be directed to detention storage such as parking lots, parks, or in underground storage.

By providing the major and minor system for urban drainage, a higher level of flood protection can be provided and the chance of basement flooding can be reduced. Figs. 3 and 4 demonstrates the operation of inlet control during normal and heavy rainfall conditions.

Fig. 3 Operation of inlet control during normal frequent storm condition

Fig. 4 Operation of inlet control during severe storm condition.

SWM Quality Controls

Stormwater Lot Level Controls

1. Reduced lot grading (Fig. 5)

Traditionally, lot grading is about 2% to facilitate drainage from the building. For lot areas far away from the building (e.g., 2-4 m), the grade can be reduced to about 0.5%. This design increase the pervious depression storage and encourage natural infiltration.

Fig. 5 Flatter lot grading

2. Rainwater leaders or downspouts

Re-direction of downspouts (or roof leader) from driveway to grassed areas with a splash pad can reduce the runoff volume and its rate. Although the grass may not provide high level of water quality treatment, the reduction in runoff volume has improved the overall water quality of runoff.

For commercial and industrial buildings, restricting the number of roof drains or flow rate can detain stormwater on rooftop and reduce the runoff rate into the storm sewer.

Roof leader can also be discharged to soakaway pits which is essentially an underground infiltration trench.

Roof leader can also be discharged to rain barrel or cistern which temporarily store the rainfall for other uses (e.g., watering lawn).

Fig. 6 shows the favourable and unfavourable situations for downspout disconnections.

Fig. 6 Favourable and unfavourable situations for downspout disconnection

3. Porous pavements (Fig. 7)

Porous pavements are designed to encourage infiltration and withstand the freeze-thaw conditions. Open-design concrete block systems are used to provide the bearing capacity and allow grass growth. Recent research at the University of Quelph has found that runoff pollutants can be trapped in soil beneath the porous pavement. Concern of porous pavement is usually centred around the maintenance requirement for the continuous operation of the system.

Fig. 7 Porous pavement

4. Oil/grit separators (Fig. 8)

Traditionally oil/grit separators are designed to trap large grits and hydrocarbon fuel from parking area. However, the storage volume of these facilities are usually small compared to the runoff volume. As a result, the effectiveness of these devices for stormwater runoff control is questionable. Recent innovations provide bypass mechanism for large runoff and storage chamber for the trapped oil and grit. Nevertheless, they do provide dry spill control of oil.

Fig. 8 Typical 3 chamber oil/grit separators

Stormwater Conveyance Controls

1. Stormwater exfiltration system (Fig, 9)

An innovative stormwater exfiltration system has been developed by the City of Etobicoke to address the stormwater quality from fully urbanized area (Li and Koo, 1994; Tran, 1995). The system consists of two eight inches perforated pipes (plugged at the downstream end) which are laid below the storm sewer at the upstream section of the sewer system. The sewer and the perforated pipes are encased in a granular stone trench wrapped with filter fabric. The first flush of runoff will be directed to the perforated pipes with plugged ends. Together with the surrounding granular bedding, this pipe-trench system can store the runoff of a 15 mm rainfall event and allow exfiltration within 2 or 3 days. Excess runoff will bypass the perforated pipe system and travel through the storm sewer. The perforated pipes can be cleaned by hydraulic nozzle. Pretreatment of storm runoff using screening at the perforated pipe inlet can also reduce the maintenance requirements.

The stormwater exfiltration system can be constructed within the right-of-way and integrated with storm sewer replacement or road rehabilitation projects. In 1993, with the assistance and active participation of the Federal Great Lakes Cleanup Fund staff, the Ontario Ministry of Environment and Energy, and Ryerson Polytechnic University, the City of Etobicoke demonstrated the stormwater exfiltration system by completing 2.5 kilometres of these systems at three fully developed areas which were undergoing reconstruction or rehabilitation of roads and storm sewers. The cost of the exfiltration system was estimated to be about 15% of the sewer replacement costs at these sites.

Although the construction of stormwater exfiltration system is cost-effective when road and/or sewer is rehabilitated or replaced, there are circumstances where stormwater exfiltration system may be possible even for good-conditioned residential road and sewer systems. The cost of new stormwater exfiltration system will be the cost of the system plus the cost of road and sewer reconstruction.

Fig. 9 Etobicoke exfiltration systems

2. Grassed swales

Grassed swales (Fig. 10) are similar to ditches but are usually designed with gentler side slopes and a wider channel bottom. Ideally, they are the most environmentally-sound road and lot drainage systems because the grasses can slow down flow velocity, trap sediment, and encourage infiltration along the swale. In fact, it has been reported that grassed swales of at least 61 metres (200 feet) in length could remove approximately 80% of suspended solids, biological oxygen demand (BOD), and total recoverable lead (Stockdale, 1986).

Fig. 10 Grassed swales

Swales should be landscaped properly so that sediment in stormwater can spread out and settle unnoticeably amongst the grasses. To encourage infiltration, the invert of the swale should not be above any compacted fills used for the road and lots. Swales are appropriate in estate residential developments, industrial parks, and other low density developments.

Grassed swales can also be designed with a gentle dip in the driveway which act as a spillway. This eliminates the need for culverts and the associated maintenance complication. Ice may occasionally form on these low spots and create a slippery surface that potentially allows the exiting vehicles to slide into traffic. However, this option should still be considered if it is used on local streets that have very low traffic volumes (e.g., cul-de-sacs).

Culverts under driveways or entrance ways are common to the traditional ditch system design. By raising the invert of the culvert above the base of the ditch, some storm runoff will be trapped within the swale after storm events and be dissipated through infiltration and evapotranspiration. Additionally, the high invert reduces sedimentation and icing within the culvert which in turns reduces the frequency of the culvert cleaning or thawing. The trapped sediment will settle amongst the vegetation in the swale and facilitate new vegetative growth.

At low points within a subdivision, water and sediment may collect and create a maintenance problem. If water freezes at the low point of a swale or within the lowest culvert, it can take up some capacity of the swale or block the culvert causing the road to be overtopped during storm events. To avoid creating sensitive low points, well planned grading practices should be employed. Additionally, road culverts may be oversized at low-points to alleviate this problem. Sub-drains can be incorporated into the swale design to address the concern of road base drainage.

Swales can also be landscaped to slow down the flow with the establishment of broader leafed vegetation and/or check-dams (Fig. 11). The slowing down of flow reduces downstream stream bed and bank erosion. Moreover, the ponding behind these check-dams can enhance infiltration and evapotranspiration (Kercher, 1983; Ministry of Environment & Energy, 1991).

Fig. 11 Grassed swale with check dams

Before choosing the gentle swale option, the anticipated flow velocity at the swale must be determined. High velocity, which causes erosion of the swale itself or the road's supporting fill, should be controlled by reducing the longitudinal slope (e.g., <5%) and/or increasing the roughness of the swale (e.g., more resistive grass mixture, intermittent rip-rap placement).

3. Pervious catchbasins

Swales can also be equipped with catch basins that lead to deep vertical stone trenches. Runoff resulting from minor storms would pool in the swales until spilling into the slightly elevated catch basins which are positioned close to the driveway on the downstream end of the swale. When the trench capacity is exceeded, during a major storm, the culverts or dipped driveways would be used for stormwater conveyance. Since, these conveyance routes would be infrequently used, the size of culverts can be substantially reduced or the dipped driveways would less likely to be subject to icing. Spacing between catch basins should be as great as possible to provide maximum water quality and infiltration benefits. This alternative drainage system would be most appropriately used in low density developments. An application of this design can be found on "Doulton Drive" in the City of Mississauga, Ontario (Fig. 12).

Fig. 12 Grassed swales with vertical infiltration trenches

End-Of-Pipe Swm Facilities

1. Flood control ponds (Fig. 13)

Flood control ponds are supposed to attentuate the post development flows back to the pre-development conditions in order to protection downstream areas from flooding. They are generally sized to provide quantity control for storm events ranging from 2 to 100 year frequency. Both dry and wet ponds have been used for flood control purposes.

Fig. 13 Stormwater management ponds

2. Extended detention wet ponds (Fig. 14)

These are the most common type of end-of-pipe stormwater quality for quality control. A flow splitter is typically installed at the last manhole to direct only the first flush volume to the water quality facilities. They are generally designed with a sediment forebay for larger size particles and a deeper pool for the sedimentation of smaller size particles. Wetland can be used to reduce nutrients in stormwater. The active storage is detained for extended period of time (e.g., 24 or 48 hours) by installing a rise-pipe control outlet.

Fig. 14 Typical wet ponds

3. Stormwater flow-balancing system (Fig. 15)

A stormwater flow balancing system has been developed by the City of Etobicoke to treat storm runoff from a motel strip re-development (Tran, 1994). This RSWMP is an instream facility which is protected by an embayment. It consists of a series of cells which detain storm runoff for a certain period of time in order to remove sediments and reduce bacterial concentration (Tran, 1994; Stripe, 1995). The cells are separated by plastic curtains which are connected at the top to floating pontoons held in place with piles. Each plastic curtain has an opening below water which is sized for a certain level of volumetric control. Storm runoff enters the first cell and displaces the existing water into cell 2 which in turn displaces water to cell 3 and so on. The last cell discharges into a wetland component before exiting the system to the receiving water. Cell 1 is designed for sedimentation while cells 2 and 3 are designed for futuristic fish/wildlife habitats. Waterfowl is prevented from entering any of the cells by using a singing wire. The wetland component provides nutrient removal potential for stormwater. As the storm runoff is detained over extended period of time, it is anticipated that the storm runoff will be treated at high efficiency. A similar system was constructed at the Scarborough Bluff Park in Scarborough (Fig. 15).

Fig. 15 Scarborough flow balancing system

Integration Of SWM Objectives

Swm Practices Selection

A four step methodology has been proposed by the MOEE (1994) for developing SWM plan in the absence of a subwatershed plan:

1. Subdivision/Site Planning

The subdivision or site should be planned with due regard to the protection of natural function of the land and areas should be reserved for SWM. Natural resources such as ESA, ANSI, Flood plain, significant fish habitat, etc. should be mapped so the subdivision or site can be developed with minimal impact on these resources.

2. Assessment of Receiving Water Coneerns

Receiving water concerns listed below should be identified for the development:

- aquatic habitat, pollutant loading, recreation
- flooding
- in-stream erosion
- groundwater recharge, in-stream baseflow/low flow maintenance

3. Selection of Water Management Criteria

a. Water Quality
Four levels of protection have been specified in the MOEE SWM Practices Planning and Design Manual (1994):
Level 1 Protection
Type 1 fish habitat which limits the overall fisheries productive capacity must be protected ("Fish Habitat Protection Guidelines for Developing Areas", MNR, 1994).

Level 2 Protection
Type 2 fish habitat refers to feeding areas, unspecified spawning habitat, and pool-riffle-run complex.

Level 3 Protection
Type 3 habitat refers to areas which have low capacity for fish production and low potential for restoration or rehabilitation. For example, it may includes municipal drains, highly altered or polluted watercourses or artificial drainage swales.

Level 4 Protection
This level of protection is for retrofit or re-development situations.

Water quality storage requirements for the above four level of protection are listed in the following Table 3

b. Erosion potential

In-stream erosion is a complex process. In a watershed/subwatershed planning process, continuous simulation using an erosion index (based on either tractive force or velocity-duration information) can be used to estimate the erosion potential of a watercourse due to urbanization.

For an individual site, the MOEE/MNR Interim Stormwater Quality Guidelines (1991) stipulate erosion control should be addressed by providing extended storage of runoff from 25 mm storm (4 hour Chicago distribution) for 24 hours.

c. Water Quantity

Traditionally, regulatory agencies recommend that the post-development flows be controlled back to the pre-development flows. However, this requirement does not guarantee water quantity control along the watercourse because the controlled peak flow from an individual site may coincide with the peak flows from other sites which result in higher and longer cumulative peak flows in the watercourse.

Some generic recommendations have been proposed by the MOEE for situation where subwatershed plan is not available (1994):

1. Flow must be controlled if there is a potential flood hazard immediately downstream of the proposed site.

2. The post development peak flow should be controlled to pre-development levels if the site is situated at the headwater areas.

3. Either no quantity control or over-control would be required if the site is situated at the lower reaches of the watershed.

d. Baseflow Maintenance

The MOEE recommends the following level of control:

1. No runoff from a 5 mm storm should occur for any development (excluding roads).

2. If the ares discharge to a first order stream where a Type 1 habitat exists downstream from the site, the level of control should be
- development forms which preserve and protect discharge and recharge areas.
- lot grades (2 to 0.5%) should be reduced to promote depression storage and evapotranspiration/infiltration.
- pervious catchbasins and exfiltration pipes should be promoted.
- the existing topography should be preserved in the development layout.
- use of foundation drain sump pumps to either the surface or soakaway pits.

4. Urban SWM Practices (SWMP) Selection

The goal of SWM is to preserve the natural hydrologic cycle and minimize the potential for cumulative downstream flooding, erosion, and baseflow impacts.

There is a perferred hierarchy of SWMP which is listed as follows:

1. Stormwater lot level controls

- roof-leader disconnection from sewer and re-direction to infiltration areas in backyards.
- roof-leader discharge to subsurface soakaway pits with an overflow discharge to the surface.
- reduction of lot grading from 2 to 0.5%.
- foundation drain sumps to surface discharge or soakaway pits.

Fig. 16 Soakaway pits

2. Stormwater conveyance controls

- exfiltration pipe systems
- pervious catchbasins
- grassed swales

3. End-of-pipe SWM facilities

- wet ponds
- wetlands
- dry ponds
- infiltration basins
- infiltration trenches
- filter strips
- buffer strips
- sand filters

Fig. 17 Infiltration trenches

Fig. 18 Sand filters

Fig. 19 Constructed wetlands

Fig. 20 Typical filter strip

The selection of appropriate SWMP is based on four factors:

1. Physical suitability
Criteria (see the following Table 4) for assessment of physical feasibility are:
- topography
- soils stratification
- depth to bedrock
- depth to seasonally high water table
- drainage area
Physical feasibility is the only constraint that can exclude a SWMP from further consideration.

2. Conformity with development plan

Development layout should be planned to take into consideration SWM requirements. Sufficient area should be reserved for the implementation and maintenance of SWM facilities

3. Cost

Different combination of SWMPS should be investigated and compared with respect to cost-effectiveness.

4. Technical longevity/effectiveness

A technical effectiveness/performance ranking of each type of SWMP has been proposed by MOEE and shown in Table 5. When more than one type of SWMP is proposed at a site of interest, an overall technical effectiveness can be calculated based on an areal weighting of individual SWMPs and the drainage area which they serve.

Stormwater Retrofit Planning for Urban Watersheds

Storm retrofit planning for urban watersheds remains a challenge to Canadian municipalities as the funding and planning mechanisms are not well defined. Municipalities can play an important role in managing storm water quality in urbanized areas. However, they have to address the following physical and financial constraints in urbanized areas:

• lack of space and funding
• integration with existing infrastructure an drainage paths
• lack of proven technologies available for retrofit applications
• lack of an appropriate planning strategy
• safety and liability issues

One of the many challenges that municipalities must face in particular, is the lack of a storm water quality management planning tool for urbanized areas.

This section described a Geographic Information System (GIS) based planning methodology for storm water quality management in urban watersheds. The planning methodology comprises five steps: (1) definition of storm water retrofit goals and objectives; (2) identification of appropriate retrofit storm water management practices; (3) formulation of storm water retrofit strategies; (4) evaluation of strategies with respect to retrofit goals and objectives; and (5) selection of storm water retrofit strategies. A case study of the fully urbanized Mimico Creek watershed in the City of Toronto is used to demonstrate the application of the planning tool. The GIS-based planning methodology will allow municipal planners and engineers to identify appropriate retrofit storm water management practices (RSWMP=s), estimate the potential cumulative effectiveness and costs of RSWMP=s, develop and evaluate alternative storm water retrofit strategies, and select the preferred strategy for short and long term implementation.

A Stormwater Retrofit Planning Tool

Stormwater Retrofit Goals and Objectives

• maintain and rehabilitate the hydrologic cycle;
• maintain and rehabilitate the aquatic habitat of a watercourse and its riparian zone;
• maintain and rehabilitate runoff and surface water quantity and quality; and
• maintain and rehabilitate the physical, chemical, and biological relationships of the water ecosystem.

Economic goals can include:

• minimize the capital, operation, and maintenance cost of RSWMPs;
• integrate a stormwater quality management strategy with municipal capital works and maintenance programs;
• rationalize and streamline the approval process of stormwater related capital works and maintenance programs; and
• reduce the liability associated with inappropriate human access.

These goals can be defined by site-specific measurable objectives listed below:

• reduce runoff volume;
• improve water quality by the removal of pollutants from runoff;
• improve the aquatic habitat of a watercourse;
• reduce runoff temperature after development;
• increase fish populations and diversity;
• maintain the current annual funding for municipal stormwater-related capital works and maintenance activities;
• locate RSWMPs only on sites and right-of-ways which are or can become municipal property; and
• select RSWMPs with minimum ponding of water.

Identification Of Appropriate Rswmps

Downspout disconnection

Disconnection of downspout and redirection of roof runoff to lawn areas is a source control RSWMP (MOEE, 1994). By returning the roof runoff to soils through infiltration, this RSWMP reduces runoff volume and solids loadings. As storm runoff is disposed of at the lot level, it can be implemented gradually. This RSWMP is suitable for a site where the local lot grading is gentle and sufficient lawn areas are available. It is also desirable to have sandy soil and a low groundwater

table on the site so that the diverted runoff will not be detained on the lawn over an extended period of time.

Oil/grit separators

Traditional three chamber oil/grit separators are designed to capture spills and small runoff events (MOEE, 1994). Recent designs have improved the capture of runoff by increasing the storage capacity and providing a washout protection mechanism for large flows. Nevertheless, only rigorous field monitor programs can determine the effectiveness of these new designs. Although this RSWMP is primarily designed to control commercial and industrial parking lot runoff, there is no reason to prevent its use in residential areas. It can be installed under parking lots or along a road and/or sewer system which is undergoing reconstruction or rehabilitation. However, the responsibility of maintenance will fall upon the municipalities if oil/grit separators are installed along local residential roads.

Stormwater exfiltration systems

A drainage system control RSWMP was proposed by the former city of Etobicoke to allow stormwater exfiltration along a storm sewer system (Li et al., 1997a). First flush runoff from catchbasins is diverted to two 200mm perforated PVC pipes which are constructed below the storm sewer. As the perforated pipes are plugged at the downstream end, they store the stormwater and allow it to exfiltrate to the granular stone sewer trench and subsequently to the surrounding soils. This RSWMP can be incorporated in residential road/storm sewer reconstruction and rehabilitation projects when the site characteristics are suitable. Based on a demonstration project in the city of Etobicoke, the construction cost of the perforated pipes was assumed to be about 15% of the sewer reconstruction cost. This RSWMP is not appropriate for sites where there is concern over ground water contamination by urban runoff and spills, and damage to foundations by infiltrated water.

Swales and ditches

Alternative drainage systems such as swales and ditches have been investigated by Li et al. (1998) and Sabourin (1997). Swales and ditches can be designed to increase detention of runoff along the drainage path and provide water quality treatment. Li et al. (1998) proposed a variety of swale and ditch designs where the drainage systems are integrated with check dams and infiltration devices. Sabourin (1997) developed an evaluation procedure to determine the most appropriate drainage systems for new developments and retrofit situations. However, the water quality performance of swales and ditches still require confirmation from field monitoring programs. Ditches and swales can be considered as RSWMPs as a road section is re-constructed. They are appropriate for non-erosive soil, mild slope, wide right-of-way, and large distance between entrances.

Stormwater quantity pond retrofit

Flood control ponds may be retrofitted to provide water quality treatment functions. However, it is important to maintain the existing flood storage at the pond. In order to consider this RSWMP, there should be adequate space for the creation of a water quality cell and convenient road access for construction and regular maintenance. Hipolito (1996) developed a database management program to select stormwater quantity ponds for water quality retrofit.

Stormwater quality ponds

If sufficient land is available, stormwater quality ponds can be a feasible downstream control RSWMP. In order to be cost-effective, new stormwater quality ponds should serve a drainage area of at least 5 hectares (MOEE, 1994). Other suitability considerations include separate drainage system outfalls, road access for construction and maintenance, fish habitats, and public acceptance.

Underground Stormwater Quality Tanks/Tunnels

In highly urbanized areas, where the land suitable for a water quality pond may be very expensive and difficult to find, underground storage facilities such as tanks/tunnels may be an appropriate alternative. Although underground storage tanks have been used for the control of stormwater quantity and combined sewer overflows, they are not commonly used for stormwater quality treatment. In order to be cost-effective, several storm outfalls may be disconnected and then reconnected to an interceptor sewer which discharges to an underground stormwater tank. To enhance the settling capacity, cross flow inclined plate settlers may be installed inside the tank. The treatment efficiency of this proposed RSWMP is not known and should be determined by rigorous field monitoring.

Receiving water-based flow balancing systems

This RSWMP was originally developed by Karl Dunkers of Sweden about 20 years ago. One version of the Karl Dunkers flow balancing system has been implemented by the former city of Scarborough (Aquafor, 1994). The Scarborough version of Karl Dunkers flow balancing system consists of 5 cells which detain runoff for the purpose of removal of sediments and reduction of bacterial and nutrient concentrations. These cells are separated by plastic curtains which are connected at the top to floating pontoons held in place with piles. Each plastic curtain has an opening (2 m2) below water which is sized for a certain level of volumetric control. Runoff enters Cell 1 and displaces the existing water into Cell 2 which in turn displaces water into Cell 3. For large runoff events, the flow which exceeds the capacity of the first three cells discharges directly from Cell 3 to Lake Ontario. The captured runoff in Cells 2 and 3 is subsequently pumped back to Cells 1, 4 and 5, and ultimately discharged to the lake. The first four cells allow settling of contaminants within the cells while the last cell, which is relatively shallower than the previous cells, allows nutrient uptake by aquatic vegetation.

Another version of flow balancing system was also constructed by the former city of Etobicoke for the treatment of storm runoff from a motel strip re-development (Tran, 1994). This RSWMP is also a receiving water facility which is protected by an embayment. Runoff will flow through each cell and finally discharge to the receiving lake. This RSWMP may be suitable for fully urbanized areas where the land-based end-of-pipe RSWMP is not appropriate. However, it must be protected from wave and ice by an embayment or engineered breakwater. An access road for construction and maintenance (e.g., dredging) is important for successful installation and operation. The treatment efficiency of this RSWMP is not known and should be determined by rigorous field monitoring programs.

In order to identify the feasible RSWMPs in the study area, a two-step evaluation procedure has been developed for each RSWMPs (Li, 1997). The first step comprises the most critical screening questions which determine the physical suitability of the RSWMPs for a site of interest. All of these questions must be answered affirmatively without exception in order to continue to the second step. The second evaluation step comprises secondary questions, which further examine the suitability of the RSWMPs. All of the questions in the second step should also be answered affirmatively, either with or without implementation of engineering measures designed to remedy the associated environmental impacts. If there are additional environmental impacts associated with the engineering remediation measures, then the RSWMP is not suitable for the site of interest.

Formulation Of Alternative Strategies

Alternative stormwater quality management strategies can be formulated by combining various mixes and magnitudes of RSWMPs in accord with the preferred hierarchy. However, the cost and effectiveness of emerging RSWMPS such as oil/grit separators, stormwater exfiltration systems, stormwater tanks, flow balancing systems are not known and they must be confirmed by rigorous field monitoring programs. Thus, both short- and long-term strategies should be formulated as follows:

· Short-term strategies (e.g., 5 year implementation) may use conventional RSWMPs such as downspout disconnection, quantity pond retrofit, and water quality ponds.

· Long-term strategies (e.g., 5 to 25 year implementation) may incorporate emerging RSWMPs such as oil/grit separators, stormwater exfiltration systems, stormwater tanks, flow balancing systems. It is hope that the cost and effectiveness of these RSWMPs will eventually be confirmed by rigorous field monitoring programs.

Evaluation Of Alternative Strategies

Average annual runoff volume and solids loading are estimated by the analytical probabilistic models and are given by

where R is the average annual runoff volume in m3/yr, A is the drainage area in hectares, q is the average annual number of rainfall events, f is the area-weighted average runoff coefficient, z is the reciprocal of average rainfall event volume (1/mm), Sd is the area-weighted average depression storage (mm), L is the average annual runoff solids loading in kg/yr, and C is the average runoff solids concentration (mg/L). For preliminary planning level analysis, the analytical models offer a quick and reasonable estimate of annual runoff characteristics [16]. The z of a number of Canadian rain stations was compiled by Kauffman (1987).

A multi-efficiency model is used to estimate the cumulative runoff volume (Nv) and solids loading (Ns) reduction efficiencies of a series of RSWMPs

where i is the ith RSWMP, n is the total number of RSWMPs, hv is the runoff volume reduction efficiency of an RSWMP, and hs is the solids concentration reduction efficiency of an RSWMP. For RSWMPs which reduce solids concentration only (e.g., oil/grit separators, ponds), hv is zero. For RSWMPs which reduce runoff volume only (e.g., downspout disconnection, stormwater exfiltration systems), hs is zero. For newly developed RSWMPs which have not been proven by rigorous field monitoring, the hv or hs used in Equations (3) and (4) may be based on conservative estimate of their control potential or computer model simulation.

A planning area is first divided into a number of subcatchment. The runoff volume and solids concentration reduction efficiencies of RSWMPs at each subcatchment are determined as follows:

$ Downspout Disconnection
Runoff volume reduction efficiency (hv) is given by

where Rs is the average annual runoff volume after the application of downspout disconnection and Re is the existing average annual runoff volume. For downspout disconnection, the existing impervious area of a subcatchment is reduced by the equivalent disconnected roof area. The revised area-weighted runoff coefficient and depression storage are used to calculate the revised annual runoff volume (i.e., Rs) of the subcatchment. The runoff volume reduction efficiency of the subcatchment is then determined using Equation (5).

• Oil/grit separators
  The solids concentration reduction efficiency for a subcatchment (hs) is given by

where hsa is the solids concentration reduction efficiency of oil/grit separators, Ra is the average annual runoff volume from the area served by oil/grit separators, and Rc is the average annual runoff volume from the subcatchment. Both Ra and Rc are determined using Equation (1).

• Stormwater exfiltration systems
  The runoff volume reduction efficiency for a subcatchment (Nv) is given by

where Nva is the runoff volume reduction efficiency of stormwater exfiltration systems, Ra is the average annual runoff volume from the area served by stormwater exfiltration system, and Rc is the average annual runoff volume from the subcatchment. Both Ra and Rc are determined using Equation (1).

$ Ditches and swales
The runoff volume reduction efficiency for a subcatchment (hv) is given by

where Ndva is the runoff volume reduction efficiency of ditches and swales, Ra is the average annual runoff volume from the area served by existing and new ditches and swales, and Rc is the average annual runoff volume from the subcatchment. Both Ra and Rc are determined using Equation (1).

The solids concentration reduction efficiency for a subcatchment (Ns) is given by

where Ndsa is the solids concentration reduction efficiency of ditches and swales, Ra is the average annual runoff volume from the area served by existing and new swales and ditches, and Rc is the average annual runoff volume from the subcatchment. Both Ra and Rc are determined using Equation (1).

• Quantity pond retrofit, new quality ponds, underground stormwater quality tanks/tunnels, and Karl Dunkers Flow Balancing Systems

The solids loading reduction efficiency for a subcatchment is determined in a manner similar to that of oil/grit separators.

The average annual runoff volume (Rn) and solids loading (Ln) after the application of a series of RSWMPs are determined by

where R and L are the existing annual runoff volume and solids loadings. Finally, the cumulative runoff volume and solids loading reduction can be determined by aggregating the Rn and Ln of all the subcatchments and dividing them by the corresponding R and L before any RSWMPs are applied.

Unit cost (without land costs) functions of RSWMPs are used to determine the cost of RSWMPs for each subcatchment and the planning area. The marginal cost of RSWMPs is defined as the cost per percent of runoff volume or solids loading reduction

The evaluation of the runoff control effectiveness and cost of RSWMPs is subject to uncertainties both in model selection and parameter estimation. The uncertainty of model selection was not analyzed explicitly in this study due to a time constraint. However, the analytical probabilistic models (Adams and Bontje, 1983) had been compared extensively with a continuous simulation model (HEC, 1974) and the analysis results of both models were generally found to be in good agreement. The uncertainty of parameter estimation can be addressed by a sensitivity analysis which investigates the change in runoff control cost and performance of RSWMPs with respect to the variation of cost and performance parameters.

A LOTUS 1-2-3 spreadsheet program, based on the analytical probabilistic models (Adams and Bontje, 1983) and a multi-efficiency model (Weatherbe, 1995), has been developed to determine the following items for each subcatchment as well as the planning area:

• the existing average annual runoff volume (m3/yr) and solids loading (kg/yr);
• the runoff volume and solids concentration reduction efficiencies (%) of RSWMPs;
• the cumulative runoff volume and solids load reduction efficiencies (%) of a series of RSWMPs;
• the average annual runoff volume and solids loading after the application of each RSWMP;
• the cost of runoff volume and solids loading reduction for each RSWMP;
• the marginal cost of runoff volume and solids load reduction for each RSWMP; and
• the cumulative cost of runoff volume and solids loading reduction for a series of RSWMPs

Selection Of The Preferred Strategy

The preferred strategy should be selected based upon the following principles:

• The environmental and economic objectives are to be achieved by the least expensive strategy (e.g., the least cost to achieve the required level of runoff volume and pollutant load reduction).
• Strategies which can integrate effectively with sewer/road system rehabilitation and redevelopment should have high priority.
• Proven RSWMPs should be implemented first while emerging RSWMPs should be implemented after their cost and effectiveness are confirmed by rigorous field monitoring programs.
• RSWMPs which can receive subsidies from other levels of government should have high priority.

GIS Planning Tool

GIS is a technology that enables processing of geographically-distributed data in a digital environment. There are many conventions for storing GIS data, with the suitability of any one based on the intended uses of the data. In some municipalities, these are primarily mapping purposes, but in circumstances such as this study, a more analytical approach is taken to geographic data processing, and the data are required to be disaggregated from maps. Rather than being map-based, specific "layers" of data are recognised. Layers such as watersheds, roads, lots, elevation etc are differentiated, and these are linked to database records describing their attributes. These are characteristics of the individual features on a layer (each watershed's runoff coefficient or each road's properties -width, surface material, traffic volumes, etc).

As with most information technology, the onerous tasks of data model design, data capture and data processing enable GIS to provide decision support in situations where alternative methods would be considered unmanageable. The need for GIS processing arises from the ability of this technology to impose a standard for ensuring data compatibility, but more from its use in isolating the results of criteria-based queries, in identifying and summarising associations between disparate data and in facilitating scenario-testing.

The GIS planning tool was developed for the identification of potential site selections for placement of RSWMPs and the export of data required for the LOTUS 1-2-3 spreadsheet model (Li and Banting, 1999). Since this planning tool is only appropriate if enough information and data are available, the first task of identifying a meaningful and manageable case study area is crucial. Selection of a study site in a municipality requires the commitment of the related departments and its data and information-technology resources.

The GIS Data Model demands an understanding of the data needs for selecting an effective and appropriately-located RSWMP. Since the planning tool depends on the locations and attributes of geographically-distributed features and involves multiple constraints on site and technology selections, GIS is recognised as the information technology of choice. Figure 21 illustrates the concept of GIS data model. In order to evaluate each potential RSWMP, a database of layers in which geographic locations of features are linked to tables of their attributes can be adopted as the applicable data model. This required that the specific data layers and attribute fields be first identified, and subsequently that the data be compiled and converted into the format prescribed by the types of processing tasks to be undertaken. The criteria, variables and data needs necessary for RSWMP evaluation are summarised in Table 6. Data were identified representing the storm sewer infrastructure (its subcatchment sewersheds and outfalls), and the influences affecting RSWMP design and siting (watercourses, roads, buildings, lots, parks and open spaces, utilities easements, geology/soils, elevation, the groundwater table, and floodlines).

Table 6 Data needs for potential RSWMPs

Figure 21 GIS Database Model: Thematic Layers

From these data layers, GIS tools such as overlay, spatial and attribute queries and aggregation are used to isolate areas meeting the design criteria. Overlay combines data layers so that the analyst is able to assess interactions such as how each of the areas of different land uses are served by sewersheds. Queries can then isolate, for instance, the sewers associated with residential areas, and aggregated summary statistics (such as land areas isolated by the query) can then be calculated for use in determining an optimal RSWMP scenario.

GIS requires that data be in a standardised format and be carefully scrutinised for quality before the results of the analysis are accepted. The case study below serves as a demonstration of the concept given the reality of actual data collection, compilation and analysis.