Line 147: Line 147:
 
===2005 MnDOT Report===
 
===2005 MnDOT Report===
  
In 2005, the Minnesota Department of Transportation (MnDOT) published work by Peter T. Weiss and John S. Gulliver titled The Cost and Effectiveness of Stormwater Management Practices. In examining the cost effectiveness of various BMPs, collected data were used to derive relationships between cost and water quality volume to estimate construction and maintenance costs. This study’s
+
In 2005, the Minnesota Department of Transportation (MnDOT) published work by Peter T. Weiss and John S. Gulliver titled The Cost and Effectiveness of Stormwater Management Practices. In examining the cost effectiveness of various BMPs, collected data were used to derive relationships between cost and water quality volume to estimate construction and maintenance costs. This study’s cost relationships are examined for comparison with the results here, and suggested as a surrogate for missing data on maintenance found in readily available data.
 +
 
 +
Tables 3, 4 and 5 compare the results for the estimator models described. These tables add two BMPs not found in Tables 1 and 2; sand filters and dry ponds. They also do not differentiate between large and small wet detention basins, and biofiltration, underground infiltration, and pervious pavement BMPs are not in Tables 3 through 5.
 +
 
 +
[[Summary of BMP Construction Costs from Estimator Models]]

Revision as of 17:01, 20 December 2017

Best Management Practices Construction Costs, Maintenance Costs, and Land Requirements

Introduction and Project Purpose

In this MIDS task, Barr was asked to summarize a typical range of low-impact development stormwater management best management practices (BMPs) costs. Barr identified a range of typical construction and operating costs for eight1

In order to develop a basis for estimating the life cycle costs of stormwater BMP implementation, readily available data from construction projects and other studies were examined. Barr project files and other public information were used to compile a list of project data that included cost and basic design information. Few data sets included maintenance or land costs. The data sources used vary considerably in where and when they occurred. A major element of this effort was to normalize the data for 2010 Minnesota costs.

In addition to summarizing construction cost data, the data was compared to available cost models that have been developed by the U.S. Environmental Protection Agency (USEPA), the Minnesota Department of Transportation (MnDOT), and the University of North Carolina. This provided a method of benchmarking the data collected. Use of predictive models was also used for maintenance costs and land area requirements. Because of the paucity of data for maintenance costs and land area, these models provide the greatest source of information for developing life-cycle cost estimates.

Land costs are not included in the life-cycle cost estimates. BMP land costs are dependent on parcel-specific land costs and land costs vary widely throughout the state and with zoning classification. Instead of including land costs with the BMP life-cycle cost estimates, Barr identified and summarized land requirements for the different structural BMP types. Land requirements depend not only on the size of the BMP, but also on the easement requirements of the permitting authority. Based on a review of the regulatory requirements and interviews with six Minnesota cities, Barr determined that the land area required for easements is a small component of the overall land area needed for each BMP type.

Construction cost data were collected and evaluated for eight structural Best management Practices (BMPs) categories. These types are described in this section. Each BMP is described using generally accepted definitions found in the literature. The BMPs on which data were collected are consistent with these definitions; however, there may be some variation on the size of the BMPs compared to the typical size ranges included in the definitions below. Water quality treatment volumes are also discussed for each BMP. These water quality treatment volumes were used to compare BMP costs to one another.

Bioretention Basin/Rainwater Garden without Drain Tile

A bioretention basin is a natural or constructed impoundment with permeable soils that captures, temporarily stores, and infiltrates the design volume of stormwater runoff within 48 hours (24 hours within a trout stream watershed). These facilities typically include vegetation. For the purposes of this study, the water quality treatment volume of a bioretention basin is considered to be the total holding capacity below the outlet or overflow elevation of the basin.

Biofiltration Basin/Rainwater Garden with Drain Tile

Biofiltration basins are nearly identical to bioretention basins. The only difference is the addition of a drain tile below the designed filtration media. Filtration basins are often used in areas of potential stormwater “hot-spots2," where groundwater recharge is undesirable, or areas with very low infiltration rates in the underlying soil. As with bioretention basins, the water quality treatment volume is considered the total holding capacity below the outlet or overflow elevation of the basin

Wet Detention Basin

These facilities capture a volume of runoff and retain that volume until it is displaced in part or in total by the next runoff event. Wet detention basins maintain a significant permanent pool of water between runoff events. Wet detention basins that conform to National Urban Runoff Pollution (NURP) criteria have permanent pools with average depths of four to ten feet and volumes below the normal pond outlet that are greater than or equal to the runoff from a 2.5-inch 24-hour storm over the entire contributing drainage area. These basins utilize gravity settling as the major pollutant removal mechanism but nutrient and organic removal can be achieved through aquatic vegetation and microorganism uptake. For the purposes of this study, the water quality treatment volume of a wet detention basin is considered to be the total holding capacity below the permanent pool (dead storage). Wet detention basins are not considered stormwater volume control devices.

Constructed Wetlands

Constructed wetlands are similar to wet detention basins, except they are shallower and the bottom is planted with wetland vegetation. Constructed wetlands remove pollutants through contact time with the permanent pool of water and vegetation uptake. Constructed wetlands typically require large areas to allow for adequate storage volumes and long flow paths. The Minnesota Stormwater Manual recommends that a minimum of 35% of the total wetland surface area should have a depth of 6 inches or less and 10% to 20% of the surface area should be a deep pool (1.5 to 6 foot depth). For the purposes of this study, the water quality treatment volume of a constructed wetland is estimated as the surface area of the wetland multiplied by 18 inches. This estimate is needed to develop a water quality treatment volume for many of the projects samples.

Infiltration Trench/Basin

An infiltration trench is a shallow excavated trench, typically 3 to 12 feet deep, that is backfilled with a coarse stone aggregate, allowing for the temporary storage of runoff in the void space of the material. Discharge of this stored runoff occurs through infiltration into the surrounding naturally permeable soil. Trenches are commonly used for drainage areas less than five acres in size.

An infiltration basin is a natural or constructed impoundment that captures, temporarily stores and infiltrates the design volume of water over several days. Infiltration basins are commonly used for drainage areas of 5 to 50 acres with land slopes that area less than 20 percent. Typical depths range from 2 to 12 feet, including bounce in the basin.

For the purposes of this study, the water quality treatment volume of an infiltration basin or trench is considered the total holding capacity below any outlet or overflow.

Underground Infiltration

In underground infiltration, storage tanks are either incorporated directly into or before the storm sewer system. If the storage systems are completely enclosed, stormwater is released at a controlled rate to a sewer system or open water course, and no stormwater volume is lost. If the storage systems are bottomless or perforated, they will allow infiltration and reduce stormwater volume leaving a site. For the purposes of this study, the water quality treatment volume of an underground infiltration system is estimated as its hold capacity before discharging to a sewer system or open water course.

Pervious Pavement

Pervious pavements can be subdivided into three general categories:

  1. Porous Pavements – porous surfaces that infiltrate water across the entire surface (i.e., porous asphalt and pervious concrete pavements);
  2. Permeable Pavers – impermeable modular blocks or grids separated by spaces or joints that water drains through (i.e., block pavers, plastic grids, etc.);
  3. Reinforced Soil – soil reinforced with a system of modular cells added to the surface soil to increase the bearing capacity of soil, maintain soil structure, and prevent compaction. Modular cells are typically concrete or plastic and are filled with either topsoil to support turf grass or gravel. They are most commonly used for seasonal (summer) parking and fire lanes. There are many different types of modular systems available from different manufacturers.

For the purposes of this study, the water quality treatment volume of a pervious pavement is the void space of the engineered base below the paving surface. This base is typically uniformly sized crushed rock.

Grass Swale/Channel

Grass channels are designed primarily to convey stormwater runoff. Typical specifications include a runoff velocity target of 1 foot per second small storms and the ability to handle the peak discharge from a 2-year, 10-year, or 100-year design storm. Estimating a treatment volume for grass channels is problematic because most channels are built to meet flow rate needs and available data does not include sufficient detail to estimate treatment volume. Grass swales are typically considered a water quality BMP, not a volume control BMP. The velocity in the swale must be low enough to allow sediment to drop out. There can be some infiltration along the length of the swale but this is highly dependent on surface soils and the duration of flow in the swale, which is generally too short for appreciable infiltration. Significant stormwater volume reductions can be created by placing check dams across the swale. A grass swale with check dams functions similar to a series of bioretention basins and should be viewed accordingly.

BMP Cost Factors & Methodology

Construction Costs

Actual construction costs were used to calculate construction cost per water quality volume, as defined in Section 2 for each BMP. Construction cost information from an assortment of locations and owners was used.

Data Uncertainty

The construction data collected varies considerably in its detail and comprehensiveness. Costs for design, geotechnical testing, legal fees, and other unexpected or additional costs are not usually included in the reports and are not included in the construction costs listed in this memo. Uncertainty in these construction cost estimates can come from this variable project related data and from factors such as complexity of design details, variation in local regulatory requirements,nreported soil conditions, and other site specifics. For example, variable design parameters that could affect the total construction cost include pond side slopes, depth and free board on ponds, total wet pond volume, outlet structure configuration, the need for retaining walls, and other site specific variables. These details are generally not reported in the data collected.

Another source of uncertainly is a relatively few data sources for some of the BMP categories. For example, biofiltration devices are lacking in readily available project specific data and are only represented by two data sources. Any use of the data set or derivations of it should consider the high level of uncertainty involved.

Approach to Normalizing and Reporting Data

The considerable spread in time, space, and size of the projects reporting data, leads to the need for some normalization of the data. The data were adjusted to account for these factors as described below.

Unit Construction Costs=

With an eye toward the potential of producing a calculator for developing cost estimates for BMPs, the various data points have been “normalized” for project size and scope by dividing the cost by the water quality treatment volume. This results in a construction cost per cubic foot of water quality treatment. This unit cost accounts for the size of the project and might provide a convenient basis for cost estimation. For example, a developer could develop an estimate of water volume to be treated based on one of many watershed runoff models and local regulations, and then apply it directly to these normalized estimates for individual BMPs.

Regional Adjustment Factors

The data for this study was normalized by region using regional cost factors reported in Weiss, P.T., J. S. Gulliver and A. J. Erickson, (2005), U.S. Environmental Protection Agency (1999), and first published by the American Public Works Association in 1992. All of the data were normalized to the region that includes Minnesota. All of the data statistics then are in “Minnesota” dollars.

Price Adjustment Factors

The construction costs reported have also been translated to 2010 dollars. This was done using the Consumer Price Index (CPI) history as reported by the U.S. Bureau of Labor Statistics. The CPI is a wide spectrum index that closely parallels the various construction price indices available.

Data Standard Deviation

The uncertainty and the small sample size of some of these data categories make statistical analyses suspect. The standard deviation for each data sample is reported here to indicate the level of variation within the individual data sets. For data populations with a normal distribution, one standard deviation above and below the average would encompass about 68% of the data. The figure below shows a plot of a normal distribution (or bell curve). Each colored band has a width of one standard deviation.

This image shows shows a plot of a normal distribution (or bell curve).
A plot of a normal distribution (or bell curve).

The construction cost data collected for this study are probably not normally distributed. Weiss et al. suggested that a log normal distribution would better fit construction cost data. In that case, confidence interval calculations are not straight forward.

Research Results Discussion

Data for 69 projects were normalized and statistics calculated as described in Section 3.1.2 and shown in Table 1 below. The diverse data is regionalized to the Midwestern US and converted to 2010 dollars. The averages are shown as cost per water quality volume, where that information was available. Water quality volume refers here to the raw volume of water that is treated by the BMP in terms of a straight forward and regularly reported characteristic.

As discussed in Section 2, the total volume of the BMP below the outlet was used for the water quality volume of bioretention basins, biofiltration basins, infiltration trenches/basins and underground infiltration structures. For wet detention basins, the dead storage volume was used. For constructed wetlands, the surface area of the wetland multiplied by 18 inches was used due to the lack of detail regarding the wetland projects sampled. To calculate water quality volume for pervious pavement, the void space of the base aggregate below the pavement was used. In the case of grass swales, the data reviewed did not report characteristics amenable to a treated volume estimate and assumptions regarding the watersheds and design would need to be made to estimate a treatment volume. This data was not available. Due to both a lack of good background information and highly variable information, a cost analysis of grass swales is not included in the report.

In some cases an economy of scale is clearly shown in the data. Wet detention basins exhibited the strongest apparent economy of scale as reported in Appendix A: BMP Cost Survey Data Tables. The cost difference between very small basins and large basins is several orders of magnitude. This is a significant difference and is hard to explain. Detailed project data related to the three small basin projects was not available but based on the project name, two of the projects appear to be vaults. Another explanation is that these could be decorative ponds serving more as a landscape feature than a stormwater function. For these reasons, the small wet detention basins were separated from the large basins in Table 1. Underground infiltration BMPs also exhibited an apparent economy of scale but not as clearly as wet detention basins and with some individual project exceptions. There was no apparent economy of scale in the data collected for bioretention basins.

Summary of Construction Cost Data Collected
Link to this table
BMP name Number of BMPs Cost per Average cost (💲) Sample standard deviation
Bioretention Basins 11 Water quality volume/ft3 15 9
Biofiltration Basins 2 Water quality volume/ft3 58 61
Large Wet Detention Basins treating more than 100,000 ft3 5 Water quality volume/ft3 2 2
Small Detention Basins treating less than 10,000 ft3 3 Water quality volume/ft3 145 42
Constructed Wetlands 4 Water quality volume/ft3 1 1.5
Infiltration Trenches 8 Water quality volume/ft3 11 30
Infiltration Basins 6 Water quality volume/ft3 21 15
Underground Infiltration 8 Water quality volume/ft3 213 372
Pervious Pavement 7 Water quality volume/ft3 16 8



Annual Maintenance Costs

Of the 69 BMPs presented in Table 1, 25 included annual maintenance costs. The data are regionalized to the Midwestern U.S., converted to 2010 dollars, and summarized below.

Summary of Annual Maintenance Cost Data Collected
Link to this table
BMP name Number of BMPs Cost per Average cost (💲) Sample standard deviation
Bioretention Basins 8 Water quality volume/ft3 1.25 1.18
Biofiltration Basins 0 Water quality volume/ft3 No data -
Large Wet Detention Basins treating more than 100,000 ft3 4 Water quality volume/ft3 0.07 0.10
Small Detention Basins treating less than 10,000 ft3 0 Water quality volume/ft3 No data -
Constructed Wetlands 0 Water quality volume/ft3 No data -
Infiltration Trenches 8 Water quality volume/ft3 0.39 0.11
Infiltration Basins 6 Water quality volume/ft3 No data
Underground Infiltration 4 Water quality volume/ft3 1.26 2.16
Pervious Pavement 0 Water quality volume/ft3 No data -


Data Limitations and Uncertainty

The data collected varies considerably in its detail and comprehensiveness. This leads to undocumented variability in the data from factors such as design detail, variation in local regulatory requirements, unreported soil conditions, and other site specifics. For example, variable maintenance parameters that could affect maintenance costs include soil conditions, land use within the tributary watershed, plant selection, precipitation patterns, and other site specific variables. With few exceptions, these details are generally not reported in the data available. As a whole, the maintenance cost data collected lists the maintenance costs as a lump sum without detailed breakdown or discussion.

Another source of uncertainly is a relative few data sources for some of the BMP categories. For example, constructed wetlands are lacking in readily available project-specific data and are represented by four data sources. Any use of the data set or derivations of it should consider the high level of uncertainty involved.

Approach to Normalizing and Reporting Data

The data for annual maintenance cost was normalized for region and for the date it was reported. Regional bias was adjusted using regional cost factors reported in Weiss, P.T., J. S. Gulliver and A. J. Erickson, (2005), U.S. Environmental Protection Agency (1999), and first published by the American Public Works Association in 1992. All of the data were normalized to the region that includes Minnesota.

The maintenance costs reported have also been translated to 2010 dollars. This was done using the Consumer Price Index (CPI) history as reported by the U.S. Bureau of Labor Statistics. The CPI is a wide spectrum index that closely parallels the various construction price indices available.

Estimator Models

1999 USEPA Study

The U.S. Environmental Protection Agency’s (USEPA) Engineering and Analysis Division conducted a study on stormwater best management practices during 1997 and 1998. The report: Preliminary Data Summary of Urban Storm Water Best Management Practices, (EPA-821-R-99-012) was published in August 1999. In addition to summarizing existing information and data regarding the effectiveness of BMPs to control and reduce pollutants in urban stormwater, the report provides a synopsis of the expected costs and environmental benefits of BMPs and identifies information gaps as well. It includes simple methods for estimating costs for construction and maintenance of stormwater BMPs.

This study has often been referred to and built upon in subsequent studies, including those surveyed here. The USEPA study cost estimation methods were examined for comparison with the results here.

2003 UNC Study

Ada Wossink and Bill Hunt of the University of North Carolina (UNC) examined the costs of BMPs including both installation (construction and land) and annual operating costs (inspection and maintenance) in The Economics of Structural Stormwater BMPs in North Carolina (UNC-WRRI-2003-344) in 2003. For the UNC study, construction costs and annual operating costs are statistically analyzed for effects of scale by means of BMP specific nonlinear equations relating the costs to watershed size. Annual costs were related to the area treated and to the removal effectiveness of the specific BMP for an economic evaluation. The cost relationships were given in terms of watershed area, which then requires assumptions regarding runoff characteristics to arrive at a treatment volume. For this reason, a comparison with the results of the UNC study was not done for construction or maintenance cost data collected for this study. However, the UNC study also provides land area requirement estimates that are compared with the data collected for this study.

2005 MnDOT Report

In 2005, the Minnesota Department of Transportation (MnDOT) published work by Peter T. Weiss and John S. Gulliver titled The Cost and Effectiveness of Stormwater Management Practices. In examining the cost effectiveness of various BMPs, collected data were used to derive relationships between cost and water quality volume to estimate construction and maintenance costs. This study’s cost relationships are examined for comparison with the results here, and suggested as a surrogate for missing data on maintenance found in readily available data.

Tables 3, 4 and 5 compare the results for the estimator models described. These tables add two BMPs not found in Tables 1 and 2; sand filters and dry ponds. They also do not differentiate between large and small wet detention basins, and biofiltration, underground infiltration, and pervious pavement BMPs are not in Tables 3 through 5.

Summary of BMP Construction Costs from Estimator Models

Privacy policy

Template:Footer