Introduction

Excessive runoff caused by dramatic increase of impervious area is causing flooding in urban areas, and water quality deterioration in receiving water bodies. Best management practices (BMPs) are a means for mitigating the deleterious impact of excessive runoff. However, there are many types of BMPs and potential installation sites available, and the selection process of suitable BMP types is complex in nature given the variability in site conditions, performance, and cost of BMPs. It is therefore very helpful to stormwater management practitioners to have a BMP selection and placement tool, or decision support aid, in the BMP planning and design process so that the most cost-effective BMP installation plan can be obtained (Zhen and Yu 2004; Gitau et al. 2004).

A number of BMP planning/design decision support systems have been developed in recent years. Most notably are the GIS-based Best Management Practice Decision Support System (BMPDSS) (Cheng et al. 2009; Jia et al. 2012) and the System for Urban Stormwater Treatment and Analysis INtegration (SUSTAIN) systems (Shoemaker et al. 2009). Both BMPDSS and SUSTAIN provide multiple-scale (site, region, and watershed) applications, detailed BMP process simulations, and cost optimization analyses. More recently, Viavattene et al. (2010) presented a front-end BMP tool that is includes process-based BMP assessment and a multi-criteria analysis for BMP selection. Such comprehensive and highly technical analysis tools would be very helpful to decision makers in selecting the final BMP planning strategy prior to design and implementation. However, in some situations, it could also be adequate to use a simpler tool for BMP selection, much like the use of a one-dimensional as opposed to a two- or three-dimensional model. A simpler analysis can also be used as a preliminary or “screening” tool that would facilitate the complex modeling process involved with the use of BMPDSS, etc. For example, Young et al. (2010) introduced the analytical hierarchal process (AHP) as an analytical framework for BMP selection, through priority ranking of stormwater control objectives and BMP performance metrics. Using the AHP, a BMP priority ranking is accomplished through the construction of pairwise relative rankings of the selection criteria and then of the BMPs themselves against the respective selection criteria.

This paper describes the development of a multi-criteria index ranking system to facilitate BMP selection. The ranking system established a framework that integrates all aspects of the BMP evaluation processes for a comprehensive screening level analysis that assist the decision making in selecting BMPs (USEPA 2004b; Zhen and Yu 2004; Young et al. 2010).

Development of the multi-criteria index system for BMP ranking

Development of the multi-criteria index system (MCIS) involves four steps. The first step was to establish the basic or key criteria categories that the selection system was based upon. In this study, BMP site suitability, runoff control benefits, and cost and maintenance are the three basic categories considered in formulating the MCIS. The suitable BMPs are first selected according to the site suitability criteria, and then these selected BMPs are further evaluated and ranked according to the indices in runoff control benefits category and cost and maintenance category. The technical route map of MCIS for BMPs ranking is illustrated in Fig 1.

Fig. 1
figure 1

The technical route map of MCIS for BMP ranking

Full size image

The second step was to select a level of index factors within each criteria category. Each of the selected index factors was further broken down into several second-level index factors. The third step was to establish the benchmark of each index, and the fourth step was to develop a ranking mechanism that integrates every index factor. Details of each of the four steps are described in the following sections.

Determine criteria categories

The main intent of the MCIS development is to provide a simple, yet comprehensive tool for selecting BMPs. MCIS can be used as a screening tool for a preliminary siting and BMP implementation plan. The results can then be incorporated into a more rigorous and complete decision support tool such as SUSTAIN for final BMP selection and design. On the other hand, for circumstances under which a simpler approach is acceptable or even preferred, the MCIS can be used as a convenient tool for BMP planning and implementation. It is with the latter objective in mind that the three key criteria categories were determined, i.e., site suitability, runoff control performance, and cost considerations. With these basic considerations, the MCIS could be used as a stand-alone tool for selecting the most cost-effective BMPs.

Of the three key criteria categories, the site suitability criterion is aimed at screening the various BMPs for appropriateness at the site, while the runoff control and the cost criteria categories are assessment tools for comparing the appropriate BMPs so that one or more cost-effective BMPs could be determined.

Select first- and second-level indexes for each criteria category

Site suitability category

In the site suitability category, a total of six first-level and nine second-level index factors were established for the screening of available BMPs. The hierarchal structure of the index factors is described below and is illustrated in Fig. 2.

  1. 1.

    Site conditions

    1. (a)

      Land use type

      Some BMPs are appropriate for certain land use types. Examples include porous pavements for parking lots; underground biofilter boxes for highly impervious downtown commercial districts, etc. Land use types considered in the present study and their assigned codes are listed in Table 1 (The Ministry of Construction, PRC 1991).

    2. (b)

      Pollutant loading

      Pollutant type and loading intensity in development site are important factors in BMP selection.

    3. (c)

      Special requirements

      These refer to recommendations or requirements set forth by authorities regarding BMP implementation. Examples include buffer distance from buildings; proximity from water bodies and alongside of highways, roads, etc.

  2. 2.

    Soil characteristics

    Soil characteristics are classified in accordance to the USDA SCS Hydrologic Soil Group definitions, namely the A, B, C, and D soil groups. The soil classifications and the corresponding soil characteristics are listed in Table 2 (USDA NRCS 1986).

  3. 3.

    Groundwater characteristics

    The depth of the groundwater table is used to characterize groundwater characteristics, which would present limitations of some BMPs use due to pollution concerns.

  4. 4.

    Topography

    Ground slope is the most important factor in BMP selection. It is used as the index factor in the study.

  5. 5.

    Catchment properties

    The main catchment properties considered are drainage area and the imperviousness area in the catchment. Drainage area is a key factor in sizing a BMP. And the ratio of imperviousness area is the main consideration factor for BMP selection. Some BMPs are more suited than others for highly impervious or ultra-urban areas.

  6. 6.

    Space requirement

    The sizes of surface areas required by each type of BMPs are different, such as constructed wetlands require a relatively larger surface area in order to achieve the desired performance goals. So, the area required for BMP is used as the index factor.

Fig. 2
figure 2

Indicators of BMP site suitability

Full size image
Table 1 Land use types
Full size table
Table 2 Soil types
Full size table

Runoff control benefits category

Runoff control performance refers to the mitigating effects of BMPs on runoff water quantity and quality. There are three first-level performance indicators, namely, runoff quantity control, runoff quality control, and additional benefits. The first-level indicators are further divided into 12 second-level indicators as shown in Fig. 3.

Fig. 3
figure 3

Indicators of BMP runoff control benefit

Full size image
  1. 1.

    Runoff quantity control

    The second-level indicators include runoff volume reduction, peak flow delay, and flow rate reduction. Different types of BMPs have different performances on runoff volume reduction, peak flow delay, and flow rate reduction because of the different detention storage characteristics. These data were summarized in Tables 3 and 4 which were obtained by synthesizing information available in the literature (USEPA 1999; Baptista et al. 2001; Li and Huang 2002; Revitt et al. 2003; Scholes et al. 2005). Table 3 lists control effectiveness associated with various BMP processes, or mechanisms. Table 4 was derived from Table 3 based on the processes that occurred in various BMP.

    Table 3 Control effectiveness of urban stormwater BMP processes
    Full size table
    Table 4 Runoff control mechanisms and effectiveness for common structural BMPs
    Full size table
  2. 2.

    Runoff quality control

    The second-level indicators include suspended sediments removal, oxygen-depletion materials removal, bacteria and viruses removal, nutrients removal, heavy metal removal, and toxic organic compounds removal. These indicators all are the pollutants that contained in urban runoff. As shown in Table 3 and 4, different types of BMPs have different performances on these pollutants removal because of the different processes that occurred. These processes include sedimentation, adsorption, infiltration, filtration, plant uptake, evapotranspiration and volatilization, and microbial degradation. And some of the processes which occurred in various BMPs are interlinked when different target pollutants are removed in different types of BMPs.

  3. 3.

    Additional benefits

    The second-level indicators include rainwater capture and reuse, ecological benefits, and aesthetic benefits. Different types of BMPs have different performances on these additional benefits, such as rain barrel has the high benefit of rainwater capture and reuse, and bioretention cells and constructed wetlands also bring about ecological and aesthetic benefits.

BMP cost and maintenance category

In this category, BMP cost includes capital investment cost and operational/maintenance costs. Another first-level indicator is “system reliability,” defined as the robustness in BMP design, reliability of the system in meeting design goals and the risks of system failure, etc. There are three first-level indicators, which are divided into five second-level indicators, as listed below and shown in Fig. 4.

  1. 1.

    Capital cost

    Construction cost is used as the index factor. It should be noted that in many situations, land costs can be an important factor, especially in highly urbanized areas.

  2. 2.

    Operation and maintenance

    It includes operation costs and maintenance costs. Operation costs refer to the personnel as well as material costs required in the operation of the BMPs. And maintenance costs refer to personnel, material and replacement part costs required for maintaining BMP performance.

  3. 3.

    System reliability

    Two second-level indicators are meeting design goals and operational stability and risks. Meeting design goals refer to the robustness in the design approach for various BMPs. It is an indicator whether the BMP will likely fulfill the expectation in its performance. Operational stability and risks indicate the reliability of the BMP in terms of expected performance. For example, a filtration BMP may lose its efficiency due to clogging, etc.

Fig. 4
figure 4

Indicators of BMPs cost and maintenance

Full size image

Methodology to establishing benchmarks for each index

The establishment of benchmarks for the BMP assessment indices is based on an extensive search of literature regarding BMP design approaches, implementation case study results, and reported comments and opinions from professionals in the field. Each assessment index needs to be quantified by assigning appropriate numerical values to all the comparative index factors as described in the previous section. The quantified indices are then integrated into a single benchmark number by applying “weighting” factors. The appropriate weights are determined by considering a combination of literature information and expert opinions.

Quantification of assessment index factors

The most important assessment index for evaluating BMPs is the BMP performance in runoff quantity and quality control. Therefore in this study, BMP performance was used as the primary index category for a quantitative assessment that would lead to the ranking of BMP choices and the eventual selection decisions.

Data presented in Tables 3 and 4 show the control effectiveness associated with various processes and the control effectiveness of common structural BMPs with respect to targets of runoff quantity and quality. However, only qualitative descriptions, i.e., high, medium, and low, were used to describe BMP effectiveness in both Tables 3 and 4. It should be noted that, in terms of microbial degradation of pollutants, both aerobic and anaerobic decomposition were considered. For example, generally BOD5 can be effectively reduced under aerobic conditions. However, under anaerobic conditions, the control effectiveness BOD5 is considered as “medium” to “low.”

To quantify BMP effectiveness, a numbering system was devised. The system assigns numbers from 0 to 5 to represent effectiveness, i.e., 0 = not suitable, 1 = low, 2 = medium low, 3 = medium, 4 = medium high, and 5 = high. The following equation was then used to calculate the overall score for BMP performance:

$$ {X_{ij }}=\sum\limits_{i=1}^m {\sum\limits_{i=1}^n {e \times {e_{ik }}\times {f_{{jk\prime }}}\quad k} } =12\cdots 8 $$
(1)

Where,

x ij :

the effectiveness of the ith BMP in controlling the jth target pollutant.

e ik :

the function of the kth process or mechanism associated with the ith BMP. These processes are presented in Table 3. A total of eight BMP processes were considered.

f jk :

the effectiveness of the kth process or mechanism in controlling the jth target pollutant as shown in Table 3.

The same scoring system as described above was used to quantify the other BMP benefits and also the cost category indicators. It should be noted that the numbers generated in this study do not actually represent an “absolute” quantification. Rather, the numbers only represent the relative scores of various BMPs when they can be compared on the basis of their performance and the required capital and maintenance costs.

Normalizing the numerical indicators

In order to compare the various BMPs under a unified scale, the results from the quantified index factor analysis are normalized to give an assessment index value between 0 and 1. The normalization is made by using the following equation:

$$ {r_{ij }}=\frac{{{x_{ij }}-\mathrm{Min}\left( {{x_{ij }}} \right)}}{{\mathrm{Max}\left( {{x_{ij }}} \right)-\mathrm{Min}\left( {{x_{ij }}} \right)}}, i=1, 2,\ldots 12 $$
(2)

In which:

r ij :

normalized assessment index factor for BMP i and indicator j. Here, the indicator j refers to the second-level index of runoff control benefits category and BMP cost and maintenance category. j k = 1,2,….17.

i :

BMP type, a total of 12 LID-type BMPs are evaluated.

The total assessment score for BMP i is then obtained by calculating the sum of the weighted normalized indices for all indicators:

$$ {I_i}=\sum\limits_{j=1}^{17} {{f_{ij }}\times {r_{ij }}}, i=1, 2,\ldots 12 $$
(3)

Where

I i :

total weighted assessment score for BMP i.

f ij :

weighting factor.

The weighting factors are to be determined by professionals in accordance with site conditions, local regulatory requirements, social and economic considerations, and other pertinent factors. The final ranking of the BMPs can then be obtained and the selection be made.

Developing a BMP ranking mechanism

To develop the ranking mechanism for the MCIS, the first step was to establish the basis for quantitative ranking of the 12 LID-type BMPs considered in the present study. The basis was derived by synthesizing a large amount of literature information on BMP site suitability, runoff control effectiveness, and costs. The basis sets quantitative benchmarks for the 12 BMPs so that a ranking of the BMPs can be obtained by calculating the total assessment score for each BMP.

BMP site suitability analysis benchmarks

Based on an extensive review and of pervious literature (e.g., USEPA 1999, 2004a, b; Revitt et al. 2003; Hunt and Lord 2006; Scholes et al. 2005), benchmarks were assigned for the 12 level 2 index factors pertinent to site suitability for the 12 commonly used structural BMPs. The results are given in Table 5.

Table 5 Benchmark selection for BMP site suitability analysis
Full size table

BMP runoff control analysis benchmarks and ranking

Runoff quantity and quality control processes of various BMPs have been described and presented earlier in this paper and summarized in Tables 3 and 4. Combining the qualitative evaluations listed in Tables 3 and 4, and assigning numerical equivalents ranging from 0 to 3, Tables 6 and 7 were obtained. Table 6 presents quantitatively the comparison of scores for BMP processes in controlling runoff. The control targets selected for demonstrating the MCIS formation are described below.

Table 6 Results on quantity/quality control effectiveness of various BMP processes
Full size table
Table 7 Quantitative results of the roles of control mechanisms in different BMPs
Full size table

Runoff quantity control targets: runoff volume, peak flow, and flow rate.

Runoff quality control targets: Suspended sediments (SS), oxygen-depletion materials (COD and BOD5), bacteria and viruses (Escherichia coli), nutrients (N and P), heavy metals (Cd, Cu, Ni, Pb, Zn, Cr, and Pt), and toxic organic compounds (PAHs and PCB).

The assigned scores for BMP processes in the summarized runoff quality control targets were the average value of the individual scores for BMP processes in its each runoff control target shown in Table 3. For example, the scores for BMP processes in oxygen-depletion materials targets were the average value of the individual scores for BMP processes in COD and BOD5.

Table 7 presents the quantitative results of the roles of control mechanisms in different BMPs. For example, storage is not a major process for reducing runoff volume by infiltration trenches. Therefore, a low/medium rating was assigned, and the corresponding numerical score is 1.5 for process “storage.” By the same token, a score of 3 was given to process “storage” for BMP “constructed wetland” because of its large storage capacity.

To obtain the composite score of BMP effectiveness for both runoff quantity and quality targets, Eq. (1) was used with data listed in Tables 6 and 7. For example, the composite score for infiltration trenches in controlling runoff volume was calculated as:

$$ \mathrm{Composite}\ \mathrm{score} = 1.5\times 3+1.5\times 0+2.5\times 1+3\times 2.5+2\times 0+2.5\times 0+1\times 1 + 0.5\times 1\times 1=16 $$

The results for the composite scores for all BMPs are presented in Table 8. It should be noted that for the above computation, the weighting factors were all taken as 1, except for evapotranspiration volatilization, which was assumed to be 0.5.

Table 8 Composite scores of structural BMPs control effectiveness
Full size table

It was shown from results listed in Table 8 that constructed wetlands and infiltration basins exhibit the best composite scores in general. Infiltration trenches, wet ponds, porous pavements, and bioretention cells show moderate composite scores. Green roofs, dry ponds, vegetative filters, and rain barrels are less effective than the other BMPs.

Other BMP benefits

As described earlier, there are other benefits of BMPs, especially the LID types that could be realized. These benefits include rainwater utilization, erosion control, protection of habitats and other ecological functions, moderating urban heat-island effects, etc. In addition, certain BMPs such as bioretention cells and constructed wetlands also bring about landscape and recreational benefits.

Using the (0, 1, 2, 3, 4, 5) scoring scheme, an assessment was made of the BMP benefits in terms of rainwater utilization potential, ecological services, and landscape values. The results are listed in Table 9.

Table 9 Evaluation of additional BMPs benefits
Full size table

BMP cost benchmarks and ranking

Data on the capital, operational, and maintenance costs of BMPs are still relatively scarce. In this study, information presented in a number of European and US experiences (USEPA 1999; Baptista et al. 2001; Revitt and Ellis 2001; Hunt and Lord 2006) were gathered and summarized in Table 10 below. The data include BMP construction (land cost excluded), operational, and maintenance costs.

Table 10 Capital, operational, and maintenance cost of structural BMPs
Full size table

Quantification of the BMP cost information was made by using the (1, 2, 3, 4, 5) rating scale and giving appropriate numbers to each of the 12 BMPs on the basis of capital costs, management requirement, maintenance cost, meeting design goals, and operational stability, as shown in Table 11.

Table 11 Evaluation of BMPs operational and maintenance cost
Full size table

Integrating and normalization of BMP selection index scores

The assessment scores on BMP runoff quantity/quality control effectiveness, other benefits, capital and operational/maintenance costs, and system stability are first normalized by using Eq. (2) and then are added to give an overall score for each of the 12 BMPs using Eq. (3). The results are listed in Table 12. It should be noted that, when adding the scores for the various assessment categories, weights could be given if one or more categories, e.g., runoff quantity and cost, are deemed more important than the others in the evaluation process. Since the objective of the present paper is to just demonstrate the methodology, an equal weight was given to all the categories.

Table 12 Normalized and integrated BMP selection indicators
Full size table

The results indicate, again, that constructed wetlands, infiltration basins, and bioretention cells are good BMPs to consider on the whole. On the other hand, even though green roofs offer only moderate runoff control benefits, higher scores in other categories raised the overall score and therefore become good choices to be considered.

Development of the BMP selection software, BMPSELEC

In order to facilitate the somewhat complicated steps and processes required for the execution of the MCIS BMP ranking system, a computer software package, BMPSELEC, was developed in the present study. The development environment is described as follows.

The development platform used was Eclipse3.5, and the development tools were Visual Editor and Swing. The programming language used was JAVA. BMPSELEC can be run under Windows XP/7 system.

BMPSELEC was designed with a dropdown menu style and is very easy to use. A user needs to only choose the right item from the dropdown menu based on the local conditions and then the system will generate the results of ranking the various BMPs. The following figures (Figs. 5, 6, 7, and 8) show some interfaces of BMPSELEC.

Fig. 5
figure 5

BMPSELEC interface I

Full size image
Fig. 6
figure 6

BMPSELEC interface II

Full size image
Fig. 7
figure 7

BMPSELEC interface III

Full size image
Fig. 8
figure 8

BMPSELEC interface IV

Full size image

A case study—Foshan, China BMP selection

As a demonstration of the MCIS BMP selection methodology, a case study was conducted on the campus of the Guangdong College of Environment Protection in Foshan City, Guangdong, in Southern China. Figure 9 below depicts the location of the case study.

Fig. 9
figure 9

Location of the case study

Full size image

Foshan City is a metropolis with a population exceeding 6 million. The climate is sub-tropical with an annual rainfall of around 1,600 mm. The college is undergoing an expansion project and is considering the incorporation of a number of LID-BMPs for stormwater management.

According to construction planning of the college campus, the total area of the campus is about 30 ha, and imperviousness ratio is 59 %. The main land use type is commercial and public facilities (Code C), which includes some buildings and roads, a basketball court, and a wet pond. The pollution loading generated at the campus site is considered low. According to onsite survey, the soil types at the site include B, C, and D. The depth of the ground water table ranges between 3.41 and 13.98 m. The ground slope is less than 4 % mostly but increases to15% at the hilly area in the middle. Currently, the overall construction planning for the site has already been approved by the authorities; therefore, BMPs that require large spaces (significant alteration of the site planning) for installation cannot be considered.

Considering the site characteristics and other pertinent information, a preliminary BMP screening analysis was conducted based on Table 5. It was decided that four BMPs, namely, dry ponds, infiltration basins, sand filters, and constructed wetlands were not suitable BMPs for use at the Foshan site. The disqualifying factors were slope and installation space requirements, service area, and imperviousness. Also, the existing artificial lake could be modified as wet pond to enhance its BMP functions.

An MCIS ranking analysis was performed on the remaining eight BMPs. The overall assessment scores after integration and normalization of the initial scores are listed in Table 13 below. It should be noted that because the costs of BMPs are of primary concern to the College, the weighting factors for the cost category were given a value of 1.5, while the other weighting factors were given a value of 1.0.

Table 13 Computation of BMP selection indicators for the study area
Full size table

The results presented in Table 13 indicate that for the Foshan college site, bioretention cells, wet pond, and green roofs are the most preferred LID-type BMPs, while porous pavements, infiltration trenches, and rainwater barrels are the least. Based on the above analysis and the thoughts of the College, the selection and layout of the BMPs for the Foshan college site was determined and is shown in Fig. 10.

Fig. 10
figure 10

Layout scheme of BMPs

Full size image

Conclusions

A watershed BMP selection tool (MCSI) was developed in the study to assist the selection of the most appropriate BMP type(s) for a specific installation site. The MCSI was based on three basic criteria categories, which are BMP site suitability, runoff control effectiveness, and costs. And the three basic criteria categories consist of a total of 12 level 1 and 26 level 2 index factors.

A methodology was then devised to integrate all index factors and to quantify BMP ranking assessment scores. Also, a software program, BMPSELEC, was developed to facilitate the analyses required in the MCIS framework.

The methodology at last was applied to a demonstration site in Foshan City in Southern China. It is shown that the developed MCSI framework is simple, comprehensive, and easy to use. MCSI can be used as a preliminary BMP screening tool, but can also be used as a stand-alone method for BMP selection and planning.