A Multi-criteria Decision-making Framework for Selecting the Best Low Impact Development Techniques (LIDs)


 Low impact Development (LID) techniques are known as a novel stormwater management approach designed to mitigate the adverse effects of stormwater runoff and nonpoint source pollution. This study proposes a methodology for selecting the best LIDs based on the multi-criteria decision analysis (MCDA) while incorporating different stakeholder priorities. First, we characterize all main stakeholders so that their influences can be assessed. Second, relevant main criteria and sub-criteria are selected. Third, a stakeholder analysis is carried out to obtain social performance values and criteria priorities. Finally, two MCDA methods are applied to obtain the final LID rankings. The results show that the performance of infiltration practices and vegetated filter strips are reasonably good in stakeholders’ opinion. Subsequently, the performance of the soil restoration has a number of positive effects on stormwater quality and quantity function. Most of the ten LID techniques received ratings more than 3 from the Weighted Average Method (WAM) method (performing well) other than green roofs and constructed filters, and no techniques averaged lower than 2.29 out of 5. There is a general consensus on how vegetated filter strip and infiltration practices would perform, while there is little consensus on the bioretention, level spreaders, and porous pavements. The ratings show that the majority of stakeholders thought most LIDs would perform well (score of more than 3). While further investigations into the opinions held by main stakeholders of stormwater management of LID will be required, the results of this study provide an initial insight into how these results can be applied to support effective widespread LID implementation.


Introduction
The urban infrastructure management needs to be integrated, beginning with a de nition of land-use that aims to preserve the natural functions like in ltration and the natural drainage system [1][2][3]. This approach of development has been termed Low Impact Development (LID) in the United States or Water-Sensitive Urban Design (WSUD) in Australia [4]. The LID approach is based on managing stormwater at the source by the use of planning techniques, known as non-structural LID BMPs, and microscale controls, known as structural LID BMPs, that are distributed throughout the site in order to maintain the pre-project hydrologic function of the site [5,6].
LID has the potential to mitigate some negative impacts of urban development that conventional stormwater management techniques have not been able to address adequately [7]. LIDs vary signi cantly in performance across different criteria. Appropriate LID techniques can be selected according to the project objectives, capabilities, and limitations of the study area [7]. Reduction in the peak and volume of runoff, improvement of stormwater quality, and water conservation are the main goals of the LIDs implementation. Besides, basin slope, climate, groundwater depth, operation and maintenance of existing facilities and cost are some in uential factors that should be considered in the selection of proper LID techniques [6].
Additionally, the stakeholders can have different views and priorities, adding to the complexity of the situation [8]. Con icting views from the stakeholders may lead to a complicated selection procedure [9,10]. It is important to note that the effective LID implementation will not occur without the support of all affected stakeholders [10].
Therefore, this study aims to propose a methodology for selecting the best LIDs for urban areas by comparing applying multi-criteria decision analysis (MCDA) under different stakeholder scenarios. The proposed methodology considers every stakeholder separately and to carry out a single evaluation of the LID alternatives. Note that while these kinds of problems are not new in water resource management, and the use of MCDA method, itself, has been investigated in different studies, the direct application of MCDA for selecting best LIDs has been rarely investigated. The proposed methodology comprises analysis of alternatives, stakeholder analysis, application of the MCDA methods.

Materials And Methods
We applied the Weighted Average Method (WAM) and the Preference Ranking Organization We applied the Weighted Average Method (WAM) and the Preference Ranking Organization Method for Enrichment of Evaluations (PROMETHEE) methods as two common scoring mechanism of the MCDA tool. Each the selected MCDA methods was evaluated based on their ability to match the desired criteria. The procedure for evaluating the differing methods of MCDA followed a step-by-step technique as explained in the following sections.

Stakeholder Identi cation
The successful LID implementation requires the involvement of a wide groups of stakeholders such as local government, the local water authority, stormwater practitioners, elected o cials, landscape architects and representatives from the community [8,9]. In this study, the main stakeholder groups within the watershed are considered as below: In this study, we selected a list of sub-criteria based on the requirement that the criteria can be quanti ed according to the literature. The most signi cant factors that in uence the low impact development techniques are listed below [6, 9,11]. Tables S1 to S10 provides the scores of criteria for each LID alternatives according to their importance and impact.

Stormwater quantity function
Volume reduction is the most important factor from the engineers' perspective. After that, planners and citizens are equally concerned about volume reduction [9,10]. Most stormwater systems are on city property or within city setbacks; therefore, municipal governments are the main responsible to control the runoff. Volume reduction is also considered a concern for NGOs. However, NGOs are less concerned about the volume reduction in comparison with planner and municipal governments. For municipal governments, increasing the groundwater recharge can be important factor. This factor is also important for NGOs. LID reduces demand on water supply and encourages natural groundwater recharge [4,12]. Nevertheless, compared to NGOs and governments, groundwater recharge does not have importance for building and land developers, private homeowner and local business owner. Planners are located somewhere between the two aforementioned groups. Peak ow reduction has the same priority for planners, citizens and developers and municipal governments [9,10]. However, NGOs are less concerned about this factor.

Stormwater quantity function
For private homeowner and NGOs, LIDs protect site and regional water quality by reducing sediment, nutrient, and toxic loads to water bodies [9,10,13,14]. TSS removal is an incentive factor for citizens, engineers and municipal governments. Planners are also concerned about this factor, but it is less than three aforementioned groups [9]. As it is mentioned in previous section, stormwater quality improvement is the most important factor for NGOs. Therefore, for all stormwater quality sub criteria, NGOs have the highest score. Planner usually do not care about TP removal. However, citizens are usually more sensitive in this case. Engineers, municipal governments and developers are somewhere between two these groups [9]. Stakeholders' opinion about TP removal is very similar to TN removal [9]. Stormwater temperature reduction is another bene t of LID techniques. As it is mentioned in previous sections, for NGOs, stormwater quality has more importance than other stakeholders. Furthermore, we assumed that temperature of stormwater has the least importance for planner and land developers among all stakeholders [8].

Cost
Quantifying the economic bene ts of LIDs implementation is not a simple task because some bene ts such as aesthetics, recreational opportunities, and environmental impacts are not monetary or cannot be obtained directly [9]. LIDs reduce land clearing, grading, infrastructure costs (streets, curbs, gutters, sidewalks) and stormwater management costs for developers. For municipal governments, it reduces system-wide operations and maintenance costs of infrastructure and costs of combined sewer over ows. For private homeowner, LID provides shading for homes, which decreases monthly energy bills for cooling. It also saves money through water conservation [2,[14][15][16]. Economic NPV has more importance respectively for developers, citizens and planners [10]. For NGOs, cost has the least score.  [5,13,17,18]. Most property owners would prefer and are looking for their municipal government to take on their stormwater drainage systems maintenance and repairs for obvious reasons [19].

MCDA Method
We considered the WAM approach and PROMETHEE analysis in this study as a scoring mechanism to the MCDA tool [4,19]. The Weighted Average Method (WAM) of MCDA is a value-based method where the score assigned to each criterion is a value along a discrete pre-determine range. The overall scores for criteria are determined by following equation [4,19]: where Sj is overall score for technology, W i is weight, and R j is the relative importance of criteria.

Results
Values of 1-5 are assigned by the user for importance of each main criteria with 1 representing little to no importance of a criteria of 5 representing very important criteria [4,9,10,19]. The scale of 1-5 was chosen for the criteria importance since there are ve technologies considered in each analysis and this scale is considered simple for users to understand. Table 7 summarizes the rating for each technology for each criterion in the absence of weighting. As we can see, for stormwater quantity function, in ltration practices have the best score. Bioretention, vegetated lter strip, rainwater harvesting, and soil restoration are the best LID techniques in terms of stormwater quality function. As a result, environmentally, they should be the best alternatives. In addition, vegetated lter strip, permeable pavement, level spreader and soil restoration require less maintenance. According to the below Table, vegetated lter strip, permeable pavement, level spreader and soil restoration have the highest scores for winter performance. In terms of cost, vegetated lter strip, permeable pavement, level spreader are the best LID practices. As a result, they can be acceptable for municipal organizations. However, varying levels of performance across a wide range of criteria and con icting stakeholder views could complicate the selection process of these practices.  The results of this study provide a greater insight into the opinions and concerns of stakeholders of stormwater management.

Weighted Average Method (WAM)
As we can see in the Table 8 and Figures 1 and 2, for municipal governments, in ltration practices and constructed lter are respectively the best and worst techniques. It seems logical because in ltration practices have some distinct advantages mainly around the stormwater quantity function and cost, and it is exactly what is more important for municipal governments. For building and land developers in addition to in ltration practices, vegetated lter strips have also the high score due to the same reasons.
In fact, stormwater quality function has less importance for these groups. For planner, vegetated lter strip and constructed lter have respectively the highest and lowest score. Planners are more sensitive about maintenance and winter performance, and as it is mentioned in previous sections vegetated lter strip needs the less maintenance and better winter performance among LIDs. For private homeowner, in ltration practices are the best techniques. However, vegetated lter strips have importance for NGOs as much as in ltration practices. That is because these techniques have high stormwater quality function, and it is the most important factor for environmental organizations. Finally, the summary of this project for WAM method can be shown in the Table 9 and Figure 3. For these groups of stakeholders, using WAM method, in ltration practices and vegetated lter strip can be chosen as the best techniques. Also, according to stakeholders' preferences and main criteria, green roof and constructed lters are not recommended here. Selection of permeable pavement, level spreaders and bioretention heavily depend on stakeholders' preferences.

Preference Ranking Organization Method (PROM)
As we can see in the Table 10 and Figures 4 and 5, in ltration practice is still the best technique for municipal government. However, from PROM method, green roofs have the lowest score. Vegetated lter strip and in ltration practices have the highest score among LIDs for the most of stakeholders. In other side, green roofs and constructed lters have the lowest score for different group of stakeholders, respectively. Finally, the summary of this project for PROM method can be shown in the Table 11 and Figure 6. For these groups of stakeholders, using PROM method, vegetated lter strip and in ltration practices can be chosen as the best techniques. Using PROM method, rank of vegetated lter strip is more variable compared to WAM method. Also, according to stakeholders' preferences and main criteria, green roof is not recommended here. The installation of the green roofs has very little impact on the overall watershed response.

Summary And Conclusions
As stated previously, the purpose of this project was to investigate the concerns held by each stakeholder group for implementing LID practices. characteristics and project objectives should be considered as the main in uential factors in the planning, design, and construction of LID techniques. Averaged across the group, in ltration practices and vegetated lter strip received the best ratings, while green/living roofs and constructed lters came in last. However, other techniques received ratings more than 3 out of 5 from WAM method (performing well), and no techniques averaged lower than a 2.29. It is also important to note how the stakeholders rated each technique. For example, it appears that there is a general consensus on how vegetated lter strip and in ltration practices would perform here, while there is little consensus on bioretention, level spreaders and porous pavements. The ratings showed that the majority of stakeholders thought most of LIDs would perform well (score of more than 3). Generally, many agreed that all techniques are feasible if used in the intended climate and in proper functions [7]. Overall, there are some concerns and recommendations which should be considered. First, including LID combinations in the analysis can be highly relevant. Second, note that the land-use cover and construction constraints can play an important role as some LIDs are only suited for some speci c urban context [9]. Third, the watershed modeling should be done to calculate the performance of each LID alternative. Finally, several sub-criteria were considered but eliminated due to the inability to nd proper data on the corresponding metrics. The improved understanding of key elements of LIDs provides a background for planners, designers, engineers and stormwater managers to begin the site planning process to achieve stormwater management requirements.  Rating of alternatives (PROM method)