Municipal Stormwater System Maintenance

This article provides an overview of challenges in structuring local stormwater system maintenance programs, provides a snapshot of communities’ maintenance practices from a nationwide survey of local governments, and suggests a methodology for understanding and upgrading maintenance programs.

Maintenance Program Overview
Local governments must make a multitude of decisions concerning the construction, operation, and maintenance of their stormwater infrastructures. In terms of maintenance practices, the questions include:

• How does the local government define exactly what makes up the stormwater system?
• For which parts of the stormwater system should a local government be responsible?
• What services should a local government provide to various parts of a stormwater system?
• What are the cost implications of differing maintenance policies?
• How does the local government balance cost, liability, and customer satisfaction?

Stormwater System Components
An urban stormwater system is made up of a number of components, depicted in Figure 1. These components can be categorized in a number of ways, depending on the focus of the concern or analysis. Ordinarily, differences in how these elements are considered have to do with location, category of water that flows in them, regulatory mandates, ownership, and type of structure or system.

• Water supply system : From the water meter to the structure is private and is the responsibility of the property owner. From the meter outward is the responsibility of the local government, and the property owner pays a fee for treatment and delivery of water and construction and operation of the shared system.
• Sewer system : From the habitable structure to a sewer main is considered private and is the responsibility of the property owner. From the main outward is the responsibility of the local government, and the property owner pays a fee for conveyance and treatment of sewage and construction and operation of the shared system.

In a similar way, there is a private” stormwater system and a shared public system. However, in a stormwater system there are some historical, physical, and operational differences from water supply or sewer systems that complicate the process of allocating and assigning maintenance responsibilities to property owners and local governments. These differences include:

• Much of the stormwater system is not manmade but consists of natural surface streams.
• Much of system is governed by state and/or federal regulations and termed waters of the state” even though it conveys local stormwater runoff (effluent”).
• There is no conveniently defined dividing line between the public and private system in stormwater, such as a meter or main.
• Private actions can often greatly impact the ability of the shared system to function.
• Easements and other information about the shared system are often missing or unknown.
• Stormwater maintenance programs, ownership, and responsibility have been ill-defined, fragmented, and mostly reactionary in the past.
• Because stormwater has not been contained organizationally within an enterprise fund in the past [before Government Accounting Standards Board (GASB) 34 impacts], there has been no impetus to define capital value and depreciation, maintenance, operation, or management programs.
• In the past, the prevailing wisdom was that public work on private property, regardless of the size of the system or flow it conveyed, was somehow inappropriate, and was sometimes even legally defined as such.
• Stormwater programs have been chronically underfunded.

To overcome these differences, local governments are adopting an understanding of the stormwater system that is analogous to their understanding of water and sewer systems, at the same time as they are implementing user-fee-based funding and operational aspects of a comprehensively managed public infrastructure system.

Extent and Level of Service
There are two main concepts or policies that, together, address consideration of the features and management of a stormwater system: Extent of Service and Level of Service . These combine to define the construction, operation, and maintenance program. There are three major considerations that these two policies combine to address: (1) local responsibility and customer satisfaction, (2) liability and risk management, and (3) financial ability to fund programs.

Extent of Service
A drainage system, starting from the headwaters (and rooftops) and moving downstream toward the mouth, carries incrementally larger and larger flows. The Extent of Service policy seeks to define the type of structure or the point in this dendritic system that serves as the dividing category or line between local government and private ownership, operation, and/or responsibility. Extent of Service focuses primarily on addressing liability and risk management and customer satisfaction issues. The cost of such services is then modulated through Level of Service policies.

How far into the system should a local government provide service? To satisfy liability issues, it is important to consider that similarly situated systems must be treated in a consistent manner. Thus, it is important to have a classification system that makes sense and fits local management styles and approaches. One system used by AMEC Earth and Environmental in a number of local government maintenance enhancement studies, defining the Extent and Level of Service, is described below.

The extent of responsibility for regulatory activities must go far beyond the rights of way and easements to meet the local government’s stormwater-quantity and -quality control responsibilities that protect the general public health, safety, and welfare.

For physical maintenance and operation, all of the drainage system can be categorized according to location, conveyance, and legal standing:

• Is inside or outside the public right of way
• Does or does not contain significant public” water (i.e., street water or water draining from other public properties)
• Is or is not within a permanent dedicated drainage easement

Thus, there are four “Extent of Service policy” categories of drainage system:

1. In the right of way
2. Outside the right of way, carrying public water, and within an easement
3. Outside the right of way, carrying public water, but not within an easement
4. Totally private systems

Each of these system categories carries different legal liabilities and abilities to service. The narrowest approach for a maintenance program to take in defining Extent of Service responsibilities for stormwater systems would be to accept responsibility only for publicly owned property. This would include the right of way and any other publicly owned land such as local facilities and parks. With this approach, a local government would not be involved with any stormwater systems on private property, except for possible regulatory action.

Nearly every stormwater program surveyed claimed responsibility for stormwater systems contained within the public right of way. Many provided little or no service beyond this line. While this approach may seem most easily defined, there are some significant drawbacks, most notably that the rest of the stormwater system is not maintained.

A second drawback has to do with liability. Most local governments own a system of streets, roads, and bridges, which are constructed, owned, and operated for the public’s benefit. The express power to grade and open streets implicitly carries with it the power of local governments to establish a storm drainage system. This power, however, does not include the right to redirect surface waters onto adjacent private properties to the landowners’ detriment. The owners may sue the government for damages in such situations. Therefore, the duty on the local government is twofold. It must adequately design and construct its drainage system so as not to divert water onto private property in quantities more than its natural flow so as to cause damage, and thereafter it must maintain the drainage system so that its operation does not constitute a nuisance. Many states are developing a body of case law indicating that if public water” (i.e., street runoff) causes a problem, the local government is most often found to be liable for damages.

The Extent of Service or responsibility will almost certainly change over time with experience and program maturation, in terms of both the local government’s policies and the application of those policies. For example, for routine maintenance of the systems, the extent of responsibility might consistently be limited to those components within rights of way and easements that allow adequate access to the facilities. But rights of way and easements will be added over the years, so the practical extent of responsibility will expand even if the policy does not change. It may be that the local government will come to realize that the whole of the public system is a local government responsibility and will take actions for suitable system definition, identification, and expansion.

The difficulty with expanding the responsibility of the local government is in determining where to end local responsibility and how to fund the extra responsibilities. These decisions must be made in a fair and equitable manner. One example of this approach would be for a local government to accept operation and maintenance responsibilities for all residential stormwater systems, but not for any commercial or industrial systems. Similarly situated properties must be treated in a legal and consistent manner. This consideration has led many local governments to consider the public” drainage system as all parts of the system that carry public water. They then moderate this policy through recognition of availability of easements, and through suitably defined Level of Service policies to control program cost, discussed next.

Level of Service
In addition to determining the Extent of Service a local government is willing to assume, a decision must be made about exactly what services it is willing and able to provide for which parts of the stormwater system. This decision determines the Level of Service (LOS) that the system must achieve or that the local government will provide. The LOS is defined two ways: performance Level of Service and maintenance Level of Service.

The susceptibility to flooding or water-quality problems caused by stormwater can be measured by assessing the performance Level of Service available. For example, for flooding issues, a Level of Service can be expressed in terms of the degree of roadway flooding and/or the extent of first floor flooding for a given hypothetical storm event. For example, a level of roadway service may be defined as no less than one open lane on evacuation routes during the largest 24-hour rain event with a 100-year recurrence interval. LOS definitions vary considerably by community and are defined as a design frequency tied to a specified condition (e.g., the 10-year storm design frequency for culvert overtopping).

Compared to a flooding LOS, the concept of a water-quality Level of Service is fairly new to some communities and consists variously of capture and treat criteria, percent removals of key pollutants, effluent limits, and presumptive standards for various kinds of controls. A water-quality LOS system might promote land-use controls, followed by structural treatment measures, and might penalize untreated discharge from urban areas.

A maintenance Level of Service is defined by the types of services a community provides to different parts of the drainage system, by the priorities assigned to it, or by the specific condition of the system. For example, within the right of way and in critical areas highly susceptible to flood damages, the maintenance LOS might include periodic inspection, priority cleaning, and the highest level of emergency response. In similar right-of-way areas not susceptible to flooding, the LOS for maintenance might be much lower. A community might perform maintenance for residential structural stormwater controls but provide only inspection and enforcement of maintenance agreements for structural controls located on nonresidential parcels.

Maintenance Levels of Service can also be defined in terms of the inspected condition of the system. Channel mowing might take place when the grass is greater than 8 inches high, or culverts might be cleaned out when they are, on average, 25% blocked with sediment. In these cases, inspection-derived work orders, rather than flooding complaints, drive maintenance actions.

Extent and Level of Service Matrix
The Extent of Service responsibility and maintenance Level of Service combine to define the remedial capital project program, the operation and maintenance program, and each program’s priorities, as illustrated in Figure 3. The four categories of Extent of Service are given in parallel columns and three Levels of Service (high, medium, and low) are defined in the rows. The intersection of columns and rows defines nine stormwater maintenance policy decisions.

For example, it might be that outside the right of way where there is no easement, a local government is only willing, and only has the resources, to perform emergency response services and to give technical advice. But in the high-priority public right-of-way areas and similar areas where there has been a donated and accepted easement, the local government might provide a much higher Level of Service. In order to qualify for such service, a property owner who would not otherwise qualify must donate an easement and then be reassigned to another column.

Thus, the matrix in Figure 3 can be developed into a comprehensive stormwater maintenance policy approach. Each segment of the drainage system can, over time and through inventory, be classified and placed in a given column and row. Differing LOSs can be defined for each box within the matrix or group of boxes based on a community’s willingness and ability to provide specific services. LOS 1 might be reserved for those segments of the drainage system that, for example, would cause flooding of habitable structures, flood roads to a depth of more than 6 inches, or undermine stream crossing structures through bank and channel erosion. LOS 2 might be for less critical flooding conditions, and so on. The goal is that similarly situated properties are treated in a similar and consistent manner.

Approaches to Maintenance
The approaches to maintenance, or types of maintenance that a stormwater maintenance program might perform, fall naturally into four categories as shown in Figure 4. Each community balances these four types of maintenance and blends them into a coherent program–with due regard to the strengths and weaknesses of each approach.

• Reactive maintenance involves emergency and complaint response. It involves the assumption that damage has already occurred prior to maintenance activity. In the past, many maintenance programs have conducted only reactive maintenance due to funding shortages and lack of a coherently developed public infrastructure program.
• Periodic maintenance involves routine, periodic maintenance activities driven by a predetermined schedule (such as mowing three times in the growing season, cleaning catch basins twice a year, and so on). The success of this type of maintenance depends on having sufficient data and experience to manage the system without inspections or complaints. Periodic maintenance can be very efficient if the data and understanding of the system are good, or it can be inefficient if the cyclic periods do not match the need.
• Predictive maintenance is driven by an inspection report or work order; an inspector periodically checks portions of the drainage system using metrics designed to catch problems before they either become overly expensive to fix (for example, a ditch over 33% full will trigger a cleaning work order) or cause damage to adjacent property. The goal of this type of maintenance is to find the knee in the infrastructure aging curve when repairs are most cost-effective, done by balancing repair frequency and cost against the risk of damage if maintenance is deferred.
• Proactive maintenance is the practice of using study and background research to define the root causes of chronic problems. It might involve changing a development policy, a design standard, an equipment type, or a maintenance procedure. For example, the use of a certain pipe material may be banned after a study shows that it has poor performance or a short life span. This proactive step then saves the local government thousands of dollars in pipe replacement costs. Or an erosion control program may be enhanced to reduce sediment buildup in systems. Proactive maintenance establishes design standards and specifications as an adjunct to levels of service.

Figure 5 depicts a typical mix of these four maintenance approaches for both an immature stormwater program and one that has achieved a greater maturity in maintenance management. As a stormwater program matures, and as an expanding program gains control of its stormwater system, there is a shift toward inspector-generated work orders and away from reactive maintenance where damage-driven complaints generate often efficient reaction and field response. Such reactive maintenance can never, and should never, be eliminated, because this is a system dependent on the vagaries of nature. There will always be a bigger storm. However, when services become consistent and responsive, we have seen a great reduction in politically driven reactionary maintenance and an increase in customer understanding and satisfaction.

Some communities are finding that, when they have real-time rainfall data and models of systems, they have a powerful tool to assess whether flooding complaints are the legitimate result of a poorly designed or maintained system or simply the result of a storm event in excess of the design Level of Service of the system in question. Such data serve to limit unnecessary capital construction, to focus priority construction and rehabilitation programs, and to fend off pressure to do something” unnecessary. The data also provide significant support in the event of post-flood-event legal action.

Life-Cycle Maintenance
Given these four basic types of maintenance, another way to understand a stormwater maintenance program is to consider the life cycle of physical assets, including stormwater assets. In this model there are three types of activities for any element of the stormwater infrastructure: routine maintenance, remedial maintenance, and capital construction (replacement).

• Routine maintenance simply cleans what’s there. It can be complaint-, schedule-, inspection-, or research-driven.
• Remedial maintenance (often called remedial or minor construction) fixes what’s there, restoring it as nearly as possible to its original capacity or condition.
• Capital construction involves planning, design, and replacement of the system, usually resizing it for current or future conditions or improved design criteria standards.

Figure 6 shows the typical program response to a condition assessment of, say, a culvert or catch basin.

• When the system component is rated excellent” or good” by inventory or inspection information, the maintenance approach is routine maintenance.
• When the condition eventually deteriorates to poor,” the response is to rehabilitate the system using the remedial maintenance program.
• Finally, when the segment deteriorates to an unacceptable” rating, either through physical deterioration or increased demand on the system that renders it undersized, it is placed into the capital improvement program–either as part of a system- or neighborhood-wide plan, or as a priority using pay-as-you-go funding.

Program response is actually a continuum of activities from routine to remedial to capital replacement, though various funding sources, manpower resources, or technical assessment approaches are used.

From a life-cycle perspective, the operation and maintenance of the infrastructure looks something like Figure 7. The infrastructure goes through several cycles of repeated routine maintenance and one or more rehabilitations until its useful life is gone or its effectiveness is compromised, and upgrading through replacement is the only option. This approach lends itself nicely to the alternate approach under GASB 34 requirements when paired with a condition inventory of the system, a computerized infrastructure management system, and a coherent set of maintenance management policies through the matrix approach described above.

National Maintenance Survey
To get a snapshot of the maintenance extent and Level of Service that local governments are providing to their stormwater systems, an online survey was conducted in partnership with Forester Communications, the publisher of Stormwater magazine. The following sections describe the methodology used to conduct the survey and the results and conclusions derived from the survey responses.

Methodology: Stormwater professionals nationwide were invited to participate in the survey via e-mail, and the survey reached both public and private sector professionals. In addition to several demographic questions about the local government, respondents were asked to complete a matrix, similar to the one in Figure 4. Each component of the stormwater system (pipes, culverts, ditches, streams, catch basins and inlets, detention ponds, and stormwater-quality structures) was divided into categories indicating Extent of Service: public right of way (public water), in easement (public water), not in an easement (public water), and residential and nonresidential (private water). Respondents indicated the Level of Service they provide (none, emergency, complaint-driven, inspector-/work-order-driven, or routine-/calendar-driven) for each category of service extent for each type of structure. For each infrastructure element, respondents indicated each Level of Service driver for their programs.

For instance, for streams in the right of way, respondents could indicate if maintenance is performed on an emergency and a complaint basis. For the purposes of this survey and data analysis, the levels of service are regarded as follows: Low Levels of Service are none” (the lowest Level of Service) and emergency” (a higher level than none”). The medium Level of Service is complaint-driven.” The high Levels of Service are work-order-/inspector-driven” and routine-/calendar-driven” (see Figure 8).

Survey Responses : There were 282 usable surveys received from across the United States and Canada (Figure 9): 178 municipal governments, 84 county governments, and 29 other agencies or governments from 41 states and two Canadian provinces. Although the survey was intended for municipal and county governments, an unintended result of the wide distribution of the survey was that other governmental agencies and entities that perform stormwater system maintenance, but that are not municipalities or counties, responded to the survey. For instance, survey information was received from army bases and state departments of transportation (Figure 10).

Figure 9. Geographical Distribution of Survey Response

Figure10. Survey Respondents by Government/Agency Type

Limitations: The survey results give a clear picture of the level at which responding governments and agencies conduct stormwater maintenance. Because many of the qualities of the sample are unknown, including the sample size, the extent to which the survey analysis can be generalized to other governments is also unknown.

Survey Findings
The survey responses show every combination of Extent and Level of Service within the Extent and Level of Service matrix. The answers ranged from local governments that provide no maintenance to those that provide routine maintenance on most of the public and even some of the private systems. For the group of respondents, several general trends could be identified:

1. As the Extent of Service increases, from structures carrying public water to structures carrying private water, survey respondents, as a whole, offer lower Levels of Service.
2. Complaint-driven maintenance is the Level of Service offered by the greatest number of respondents on public rights of way and on structures in easements.
3. More municipal governments than counties offer the higher Levels of Service, assuming that work-order-driven and routine maintenance are considered higher Levels of Service.

Table 1 shows the responses to the Extent and Level of Service matrix survey questions for all of the respondents. Many respondents indicated that their government offers multiple levels of service on each type of structure for each Extent of Service. For instance, some governments offer both complaint-driven service and work-order-driven service on catch basins in residential neighborhoods, because they respond to complaints about clogged inlets but also generate work orders for cleaning as crews go out on regular inspections of the stormwater system. Thus, each Extent of Service category for each structure often had multiple answers, which explains why the percentage of governments offering maintenance in each category, such as pipes carrying public water in easement, adds up to more than 100% in Table 1.

The survey responses indicate that as the Extent of Service goes from public to private facilities, the Level of Service decreases. The first row in Table 1 shows the percent of respondents that offer no service (none”) on pipes. The percent of governments that offer no service on pipes in the public right of way is very low, 5.1%, but as you move across the table to the right, the percent of governments offering no service increases, up to 71% for pipes carrying private water in nonresidential areas. Almost three-quarters of governments offer no service to this part of the stormwater system. In addition, it is clear that a majority of respondents provide no maintenance on private systems. The two right-hand columns in Table 1 show that for each type of structure, a majority of respondents offer no service on private residential and nonresidential structures. In the case of detention ponds and stormwater-quality structures, a majority offers no service on commercial structures. The survey shows that many local governments do not extend maintenance activities to private systems, although the survey does not tell us whether it is a deliberate policy decision or whether, as for programs that are just developing, it is an issue that has not been addressed or resolved.

The survey data also indicate that complaint-driven maintenance is the Level of Service conducted by the greatest percentage of respondents for all structures located in the rights of way, except for detention structures. More respondents indicated that they conduct complaint-driven service than any other Level of Service. For instance, in Table 1, examine the row marked with an arrow: 53.6% of respondents conduct complaint-driven maintenance on culverts in the right of way, and 61.6% conduct complaint-driven maintenance on culverts in easements. However, for culverts with no easement, the percent of respondents offering complaint-driven maintenance drops to 31.3%, and for private residential and private nonresidential culverts, the percent of complaint-driven maintenance drops even lower to 15.8% and 14.3%, respectively. What do these numbers tell us? It is likely that for structures with public water, a large percentage of local government not only receives stormwater complaints but also chooses to respond to the complaints. Many governments draw a distinction between their responses to complaints about structures in rights of way and easements and their responses to complaints about other structures.

Tables 2 and 3, data from municipal governments and county governments, show that while the general trends discussed previously are true for both groups, county governments tend to have lower Levels of Service overall than municipalities. For instance, notice that for structures in rights of way, a greater percentage of counties than municipalities offer the lower Levels of Service denoted as none” and emergency.”

The discussion above is based on the survey results in a raw form. Another way to analyze the survey responses is to arrange the levels of service on a scale and assign values to the levels: none, emergency, complaint-driven, work-order-driven, and routine-driven (see Figure 8). Work-order-driven and routine-driven levels do not fit perfectly into this scale because they are similar proactive approaches. However, this scale is useful insofar as it allows us to examine the respondents’ average Level of Service.” Table 4 shows the results when responses are arranged on this scale. Each respondent is assigned a single Level of Service: the highest” Level of Service marked by the respondent in the Extent and Level of Service matrix. The trends described earlier, lower Levels of Service on private systems and the frequency of complaint-driven maintenance, can also be seen here.

Grouping local government respondents and comparing municipal with county respondents, we can average their responses using this same scale, rather than just looking at whether they answered yes or no to a question. When the responses are averaged, we arrive at an average Level of Service for a group. For instance, if all of the counties offer “work-order-driven” maintenance on stormwater-quality structures in the right of way, their average Level of Service would equal four (work-order-driven), which would indicate that on average, counties offer a high Level of Service on those particular structures. It is most helpful to look at the average Level of Service in terms of high, medium, or low service since the scale does not account for the parity between “work-order” and “routine” levels of service. For example, Figure 11 shows the average Level of Service for all respondents on structures in the public right of way. The maintenance Level of Service for all of the structures is high.

When we compare municipal and county data as noted before, the data continue to support the finding that municipalities offer a higher Level of Service, on average, than counties. When we look at the differences between groups, it is important not only that our sample shows that they are different but also that we can predict the same result for other respondents in that same group. Based upon the survey responses, for several structures (pipes in the public right of way, in easements, and not in easements; and culverts in the public right of way, in easements, and not in easements) it is likely that municipalities would conduct a higher Level of Service than counties. That is, the difference is statistically significant. This is important because one can begin to ask what factors influence municipalities to offer higher Levels of Service on average than counties do on these structures. For example, more municipalities than counties might conduct pipe and culvert maintenance on a work-order or routine basis because they own the majority of the roads within their jurisdiction, they own fewer roads within their jurisdiction, or they have stable and dedicated funding requiring a higher Level of Service.

Municipal Population and LOS: Using the same idea that was used to compare municipalities to counties, we can look at the relationship between municipal Levels of Service and their population. Because municipalities that fell under the National Pollutant Discharge Elimination System (NPDES) Phase I regulations were required to develop stormwater programs in the early 1990s, years before Phase II municipalities, we can gauge the differences between older and larger stormwater programs versus newer stormwater programs. The responding municipalities were divided into two groups based on population: municipalities that likely fell under Phase II [“small” municipal separate storm sewer systems (MS4s)] and ones that likely fell under Phase I regulations (“medium and large” MS4s). These groups serve as a proxy for NPDES Phase I and Phase II communities, because the survey did not ask respondents for this information. The difference in the average Levels of Service for these two groups was quite small. However, for water-quality structures, the average Level of Service difference was statistically significant between smaller (assumed to be Phase II) and larger (assumed to be Phase I) municipalities.

The “Phase II” communities offer a higher average Level of Service on stormwater-quality structures in the public right of way. The more mature programs could have adopted a more proactive maintenance approach to these structures, while “Phase II” programs are still developing their stormwater-quality programs.

Program Funding and LOS: Finally, the survey data were analyzed to determine whether there was a relationship between the respondents’ type of program funding and the Level of Service offered. Figure 12 shows maintenance program funding, broken down by type of funding: 32.8% of programs were funded by utilities (90 programs), 49.6% of programs were funded through general funds (136 programs), and 17.5% of programs were funded by other sources (48 programs). Among the programs that used other sources of funding were state departments of transportation, which relied on state and federal highway funds (eight DOTs of 11 listed state and federal funds). The mean Level of Service offered by the three programs was different among the groups.

The differences among the programs were statistically significant for pipes (in rights of way, in easements, and in nonresidential areas), culverts (in rights of way, in easements, and in nonresidential areas), catch basins in nonresidential areas, detention ponds in general residential areas, and water-quality structures (rights of way, homeowners’ association, general residential, and shared commercial).

It is interesting to note that for pipes, culverts, and catch basins/inlets, programs with utilities as their funding sources had a lower average Level of Service on infrastructure that carried private water (nonresidential) than did general fund programs. It might be expected that utility-funded programs would offer a higher average Level of Service on structures carrying private water than programs funded by the general fund, because utilities offer a predictable, dedicated funding source. However, it could be postulated that utilities offer a lower Level of Service on private structures, not because they lack funds, but because the formation of a utility demands a more rigid maintenance policy structure, to which the local government adheres.

Developing Maintenance Policy
Although the purpose of this article is to review maintenance approaches and to take a snapshot of national practice, several observations on bringing about enhancement to local maintenance practices are included in closing.

Local governments should consider it a priority to develop and articulate consistent maintenance program policies. Maintenance program policies help ensure the long-term stability of maintenance programs and provide a measure of defense against both legal challenges and outside pressures to perform services not in line with written policy. Decisions about how a stormwater system maintenance program is conducted must be purposefully thought through and defined rather than happen by default. For a maintenance program to be effective, the roles for each responsible party and for each stormwater system component must be clearly defined by local governments. Infrastructure falls into disrepair when it is unclear who owns and is responsible for the component parts of the system.

A local government must determine which combination of Extent and Level of Service best suits its capabilities, both physically and financially. The program must sustain the stormwater infrastructure while staying within the local government’s resources. The process a local government follows is to define and follow an orderly path of transition from an understanding that stormwater is essentially a private responsibility in which it intervenes only in emergency situations to one where stormwater is considered a public infrastructure management program with both public and private responsibility.

Implementation of the maintenance policy starts with identification of critical systems and implementation of a high-priority maintenance program to keep these systems functional or to restore flow capacity. Through a process of adding information and prioritization based on past history, and/or field inspection and supplemental modeling, certain segments are moved into appropriate categories and priorities.

Any new policy will require time to implement, and will change as experience is gained. The following steps are recommended to help ensure the new maintenance policies are successful and gain public support:

1. Perform an inventory of the drainage system, collecting necessary information for the defined uses of the inventory data. Specific information might include size, type, location, condition, connectivity, ownership, and maintenance category.
2. Place the inventory information into a geographic information system with sufficient programming to allow, at a minimum, for simple queries and searches of the information. Develop an ability to generate maps of areas easily.
3. Over time, define the drainage system, dividing it into the categories contained in the matrix.
4. Implement a work-order system integrating the inventory information as applicable.
5. Ensure that the local government has the manpower and equipment and the institutional, legal, and financial resources necessary to fully implement the policies. For example, set inspection schedules in accordance with available resources.
6. Develop written policies for each of the matrix blocks. Ensure they are legal.
7. Train inspectors and maintenance crews in the different responsibilities and how to articulate them to citizens.
8. Develop a complaint-response procedure that incorporates the policies.
9. Develop policy brochures to support the maintenance policies. For example, develop a brochure that tells citizens what their and the local government’s responsibilities are for a given drainage system type.
10. Advertise the new policies to political leaders, stakeholders, and the general public.

Acknowledgements
The authors thank Beth Chesson of AMEC for her assistance with this paper, Steve Di Giorgi of Forester Communications for making the distribution of the survey possible, and Nadia English for developing the online survey form.