By Stuart Patton Echols

Stormwater is often managed from a limited perspective based on flood control and cost. In fact, the communitywide impacts go far beyond storm sewers and detention ponds; stormwater management plans should reflect design objectives and values of prospective users and should not be viewed as simply a runoff disposal system. These systems have objective outcomes (such as the amount of water retained or detained or the reduced runoff downstream). However, stormwater management projects, particularly those with community design input, can also have less tangible outcomes, such as people-pleasing landscapes and designs that reveal something about the landscape. We need to be able to discuss the range and nature of these outcomes and how we might approach designing stormwater management landscapes to achieve multiple outcomes.

The decision support system presented here is based on issues that designers need to consider when ranking alternative techniques for stormwater management for a particular development site. It provides a set of ranking tools that can be used for evaluating:

• alternative techniques suitable to a site's physical condition and development status,
• stormwater management benefits provided by each technique,
• the expected pollutant removal capacity of each technique with different designs,
• environmental and human amenity values associated with each technique.

These ranking tools can be used in any order or may be used as an overall summary of alternatives.

During the selection process, the ultimate objectives for managing runoff from the site must be identified and valued. Designers can create their own sets of objectives or rely on existing objectives from other stormwater management plans. There are many plans to draw on around the country, such as the stormwater objectives from the 1996 Environmentally Sensitive Low-Impact Development Proposal from the Department of Environmental Regulation of Prince George's County, MD. The Prince George's County plan is an example of stormwater objectives that can be used confidently in a typical locality. The basic objectives of the plan include:

• restoring the site hydrologic regime to mimic the natural or predevelopment condition,
• maintaining surface-water and groundwater quality and minimizing the generation and offsite transport of pollutants,
• minimizing disturbance of riparian habitat functions,
• preserving terrestrial habitat's ecological functions and maximizing conservation of woodland and vegetative cover.

Such stormwater objectives are used as grounding criteria in generating consistent nominal data based on sound design and planning guidelines in developing the ranking tools shown in Tables 1 through 8.

Data Selection

Data selection is the factor part of a sum-of-weighted-factors method. The second step in building this decision support system is to develop inclusive ranking tools for assessing the strengths and weaknesses of each stormwater management alternative. This process is started by determining which alternatives are suited to the site's physical characteristics. For example, one of the most important physical factors to consider in such an analysis is the total area of the site and, hence, the total amount of stormwater that the management techniques will need to accommodate. The ranking tool shown in Table 1A shows these ranges in schematic fashion: ¡ indicates that the size of the site may preclude use of the alternative,  indicates that the size of the site can be overcome with good site design, and  indicates that the size of the site is generally not a restriction.

Note that the selection of the alternative stormwater management techniques listed in Table 1 are given for demonstration proposes and do not reflect the only stormwater management options that should be considered. For the benefit of consistency in explaining this concept, however, these same alternative stormwater management techniques are used in all the tables in this article.

The ranking tool presented in Table 2 is intended to deal with common site soil restrictions to alternative techniques. These issues include permeability and erosion of the soils and how they influence the effectiveness of each alternative. For example, pond-based stormwater alternatives tolerate a much broader range of soil conditions than do infiltration alternatives for addressing the peak rate of runoff discharge.

The remaining ranking tools address other relevant issues, including common site restrictions (Table 3), management benefits (Table 4), water-quality improvement (Table 5), community amenities (Table 6), and development issues (Table 7). Detailed background information and discussion regarding development, design, application, and assessment of available alternatives, including many of those listed here, are beyond the scope this article and are readily available in professional literature. Designers and planners can also consult such sources to help establish the individual ranking criteria and validate the nominal data (¡   ) assigned to each matrix. A key point of the ranking process, however, is that it should reflect local issues and concerns. The ranking factors can also be assigned different degrees of value (weight) based on the degree of local importance; in an urban context where land may be strictly limited, for example, or where its value may be prohibitive, some uses and practices requiring large land areas are simply not feasible. The three most used control devices for stormwater-quality management–dry ponds, wet ponds, and infiltration devices–are not always suitable for urban retrofit situations because of space constraints or underlying soil conditions. On sites where these types of conventional practices are not feasible, innovative and experimental approaches must be tried.

Data Analyses

Data analyses are the weighting part of a sum-of-weighted-factors method. To ensure that the analysis process for selecting alternative techniques reflects local issues and concerns, techniques from the common sum-of-weighted-factors model are used to develop an analysis process that can be customized to local conditions and planning concerns.

After the ranking tables have been developed, each factor and table can be assigned a value (weight). This value is used as a multiplier in order to ensure that the process reflects local issues and concerns. If, for example, the wildlife habitat factor is more important to the community than the recreation factor, the multipliers for each factor should be adjusted accordingly. Likewise, if the general factors of ranking tool No. 7, development issues, are more important than the general factors of ranking tool No. 5, water-quality improvement, multipliers for each ranking tool or table should also be adjusted. In this manner, the analysis can be fine-tuned to accommodate local issues and concerns.

The strength of the model is its ability to respond to local issues by letting local designers or officials establish the importance each factor without having to reinvent or rewrite the decision support system. This framework allows individual communities to customize the analysis by first adding or removing factors to the ranking tables and categories and then assigning numerical multiplying weights to each table to reflect the importance of each issue. An additional benefit of using a system based on sum-of-weighted-factors is that many communities already have developed or been involved in such programs and are therefore familiar with the process.

Example

The systematic and flexible character of this framework is most easily demonstrated by example. In a hypothetical example, a developer has a 5-ac. lot that will be developed as an office complex and parking lot. The site has low slopes, sandy loam soils, and shallow bedrock. Management objectives for the site are: (1) control the two-year peak discharge rate, (2) allow groundwater recharge, (3) achieve moderate-to-high removal of urban pollutants, (4) accommodate the high land cost, (5) get the project's permits quickly, and (6) provide some type of recreation amenity.

Using Table 1, we find the following information regarding 5-ac. sites: infiltration, water-quality inlet, vegetated filter swales, conveyance system, onsite flooding, and wetlands systems are generally not restricted by the site size, and difficulty with retention wet ponds and capacity transfer can be overcome with good site design.

Using Table 2, we determine that sandy-loam soils work well for extended detention, infiltration, water-quality inlet, vegetated filter swales, conveyance system, onsite flooding, and capacity transfer and that wetlands systems and retention ponds would be more difficult to design.

Using Table 3, we determine that shallow (high) bedrock works fine for vegetated filter swales and onsite flooding and that the limitations to the other stormwater alternatives can be overcome with good site design.

Using Table 4, we determine that controlling for the two-year peak discharge rate can easily be achieved using extended detention, retention wet ponds, capacity transfer, onsite flooding, and wetlands systems; that with good design the two-year peak discharge rate can easily be controlled using vegetated filter swales and infiltration; and that a water-quality inlet and a conveyance system will not control the two-year peak discharge rate.

Using Table 5, we determine that moderate-to-high removal of urban pollutants can be achieved best by using vegetated filter swales, infiltration, and wetlands systems; that extended detention, retention wet ponds, and water-quality inlet will remove fewer urban pollutants; and that a conveyance system, capacity transfer, and onsite flooding will not remove any urban pollutants.

Using Table 6, we determine that some type of recreation amenity can best be provided using retention wet ponds and onsite flooding and that infiltration and wetlands systems can contribute to recreation with good site design.

Using Table 7, we determine that the high land cost can be accommodated using infiltration, a water-quality inlet, vegetated filter swales, conveyance systems, onsite flooding, and capacity transfer. Table 7 also helps determine that the permits can most quickly be obtained for extended detention and vegetated filter swales. Obtaining permits for retention wet ponds, water-quality inlet, conveyance systems, and wetlands systems is moderately easy, and obtaining them for capacity transfer, onsite flooding, and infiltration can be very difficult.

These issues are averaged and summarized in Table 8, allowing us to easily see and understand the raw data relevant to the specific development proposal. The information can be used in this nominal form as a qualitative method to analyze alternative stormwater management techniques based on the classifications listed next to the legend (¡   ) below each table. This nominal form of the data is the simplest form of measurement and involves classification based on one or more distinguishing characteristics. There are no mathematical functions to be performed to derive greater understanding. By using the numerical values assigned to the nominal form, however, we may also regard the information as ordinal data, which allows rank ordering on a given characteristic.

The factors in the ranking tools have been carefully worded so the higher numbers can be regarded as the most desirable and lower numbers as least desirable. For example,  or 2 can be attributed as the most desirable and ¡ or 0 as not desirable. These cannot be considered interval data, however, because they do not contain equal differences between the units (each unit on the scale is not exactly equal to any other unit on the scale). This brings up an important point regarding sum-of-weighted-factors methods. Ordinal data are regularly combined in sum-of-weighted-factors methods using addition and multiplication; however, it must be remembered that the results are still based on the original data, which allow only rank ordering and not proportional ordering. Therefore, a value of 100 is ranked as greater than a value of 10, but it should not be considered 10 times greater. This common misunderstanding is one of the foremost shortcomings of sum-of-weighted-factors methods.

To aid in using the data's rank-ordering characteristics, the classifications listed next to the legend (¡   ) below each table are disregarded in favor of the numerical values (¡ = 0,  = 1,  = 2), which can be added across as total rankings for each alternative, as shown in the last column of Tables 1A, 3, 4, 5, 6, 7, and 9.

The numerical values can also be multiplied as a form of weighting factors and added to compare the effects of different objectives. This is demonstrated by example in Table 10. In this example, control of the two-year peak discharge rate and of land-use efficiencies are deemed the most highly valued factors and are multiplied by 4. Ease of obtaining permits and recreational opportunity are the next most highly valued factors and are multiplied by 2. The remaining values are not modified. The resulting weighted values can be added across as total weighted rankings for each alternative, as shown in Table 10.

In this analysis, vegetated filter swales and onsite flooding clearly outscore the other alternatives; however, the final decision must still rest with experience and judgment of local designers and planning officials. This is where the designers' qualitative analyses of the raw data and their understanding of the weighting effects become a critical part of the decision support system. If a different set of factors were selected and/or different weighting criteria established, the framework would give different results. This level of flexibility is designed into the framework and should be balanced by careful and systematic review, selection, and weighting of factors by local designers and planning officials.

A few critical issues should be addressed at this point regarding the general development of decision support systems, such as a framework for ranking alternative techniques for stormwater management for a given site. These issues are universal to all decision support systems and should be recognized as part of developing such a conceptual ranking framework.

• Such a framework can be a valuable tool, but it cannot make the decisions by itself.
• Decision support systems are only as good as the data going into them, and the obtaining of accurate, bias-free data can be a significant issue.
• It is not always clear much or how often a designer should rely on decision support systems, especially when a knowledgeable and experienced designer might be the most effective support system.
• There are many tradeoffs between complex and simple systems. The simple ones are easier to create, understand, and use but are often less useful; the complex systems are more useful but harder to create, use, and understand and are often used incorrectly. Each locality must find the balance between usability and usefulness that fits its particular needs.

Critique of the Framework

Among its strengths, the framework creates a flexible and systematic process for incorporating many factors in a way that allows individual localities to meet their specific needs. Although the sum-of-weighted-factors process is highly value-laden (e.g., what is the appropriate weight to assign to aesthetics?), the values are chosen by members of the community and can be used to represent the diverse interests of a single locality.

Among its weaknesses, the framework is fairly complex and requires a lot of time and effort by the designers or local individuals to establish the factors and weights. Although it is designed to be flexible, this flexibility might lead to misuse. As a result, the framework could be manipulated to give any preferred answer. The subjective work of selecting and weighting factors can allow preconceived notions to unintentionally bias the results. Although such flexibility might better serve individual localities, it does not work well for regional or statewide interests because the values from one locality might not be appropriate for another. Finally, the linear character of ranking tables and sum-of-weighted-factors processes can lead to a reductionist approach to decision-making. It would be easy to oversimplify the system into a series of prescriptive steps. This would be an improper use of the system and would defeat the benefit of considering and reconsidering relevant factors. According to J. Motloch in Introduction to Landscape Design (1991),

Cyclical design processes facilitate more effective decision-making. Through multiple cycles, the problem becomes better defined and multiple design responses can be developed. These responses can then be weighted and sorted. They can be prioritized and the pros and cons of each solution identified. Finally, the best solution can be determined and pursued. By such a process, the designer can move beyond habitual thought and achieve more effective, responsive design.

Stuart Patton Echols focuses on developing innovative stormwater systems to reduce the negative impact of development. He teaches full-time at West Virginia University while completing his Ph.D. from Virginia Tech in environmental design and planning. He also is an adjunct assistant professor at Penn State, working with graduate students in the Center for Watershed Stewardship.

SW - November/December 2002

Home + About + Subscribe + News + Calendar + Glossary
Talk + Images + Advertise + Contact Us + Search + Register + Services

Erosion Control Magazine | MSW Management Magazine
Grading & Excavation Contractor | ForesterPress | Forester Media

© 2000 - 2002 FORESTER MEDIA, INC.