A presentation on An Economic Analysis of Costs of Bioretention Cells and Stormwater Ponds By Ritu Sharma Clemson University Background Information Urban stormwater runoff, a leading contributor to degradation of water-quality in estuaries, lakes, rivers, and bays was the most important source of impairment of waters along assessed ocean shoreline in the U.S. in the year 2000 (EPA 2002a). The U. S. Environmental Protection Agency (EPA) promulgated in Nov. 1990, Phase I (1987 amendment of the Clean Water Act ) of a comprehensive national program to address storm water discharges, which required: Construction sites disturbing more than 5 acres of land Facilities engaging in 10 other types of industrial activity and Municipal separate storm sewer systems (MS4) that serve at least 100,000 people in urban areas to obtain coverage under NPDES permit for discharge of stormwater runoff (EPA 1999b; EPA 1996). Phase II implemented in Dec 1999 extended the requirement of NPDES permits for: Construction sites disturbing 1 to 5 acres of land and MS4 operators serving less than 100,000 people (EPA 1999b). The NPDES permits requires regulated dischargers to implement SWPPP to reduce pollutants in runoff through a combination of structural and non-structural Best Management Practices (BMPs) (e.g., EPA 1999b). In SC, operators of construction sites that disturb at least five acres of land must install structural BMPs during construction, that remove at least 80% of the average annual load of pollutants in storm water discharges that will occur after construction has finished (Sadler). According to the standard for stormwater runoff set by the Stormwater Management and Sediment Reduction Act of 1991 in SC, “post-development peak discharge rates shall not exceed pre-development discharge rates for the 2- and 10-year frequency 24-hour-duration storm event” (SCDHEC, 2003b). Implementation of these federal and state regulations require the use of BMPs, which are of two types: Non-Structural BMP: consists of administrative, regulatory or management practices that have positive impacts on non-point source runoff (EPA, 2000b) Structural BMP: are designed facilities or modified natural environments that help clean the stormwater-quality. These include, among others, bioretention cells and various types of stormwater ponds, such as dry extended detention ponds and wet ponds (SMRC). Bioretention cells are: Specially designed landscaping area adapted to treat stormwater runoff. Most commonly found along the edges or medians of parking lots and residential streets. Usually built into and under landscapes that serve other purposes, such as beautification and shade. Stormwater ponds are: Basins whose outlets are designed to detain stormwater runoff from a storm for some minimum duration allowing sediments and associated particles to settle out. Require surface area that typically becomes unavailable for other uses. To see which BMP is most suitable in a given area, estimate of the cost of designing, installing and maintaining BMPs and the amounts of pollutant that they can remove should be analyzed. Earlier studies done in this field include: Report submitted to Chesapeake Research Consortium by Brown and Schueler (1997), examining the relationship between storage volume and construction costs of the BMPs. Estimated models of capital and maintenance costs of the most frequently used BMPs by Koustas and Selvakumar (2003). Analyzing construction and annual operating costs of various BMPs and selecting the most effective BMP for the removal of a class of pollutants and its associated cost, by Wossink and Hunt (2003). Additions to previous research made in this paper are: Costs of bioretention cells and stormwater ponds are adjusted for purchasing-power differences in time and space. The addition of the effects on real costs of factors other than water-quality or water-quantity volume like: the cost of land is counted as a cost of a stormwater pond because the surface area of a pond is no longer available for another land use. the effects of three additional input prices--engineering, construction, and landscape wages--on the combined design and construction costs of the two BMPs are estimated. A comparative study of the volume-to-cost relationship for the two BMPs, determining the storage volume below which a bioretention cell is cheaper than a stormwater pond to remove pollutants. Data Sources The dataset compiles data from four different sources: A study conducted by Brown and Schueler at the Center for Watershed Protection (CWP); includes cost data on 37 stormwater ponds, 12 bioretention cells. Report No. 344 (Wossink and Hunt) of the Water Resource Research Institute; includes cost data on 9 stormwater ponds and 13 bioretention cells. Engineering Resource Corporation and Clemson University in South Carolina; includes cost data on a stormwater pond and bioretention cell. Watershed Restoration Program (WRP) of the Montgomery County Department of Environmental Protection; includes data on 3 stormwater ponds. It consists of 29 extended detention ponds and 31 wet ponds. (Extended detention ponds are those which incorporate additional features to improve water-quality along with the usual water-quantity control.) Variables The weekly wage data collected from BLS on: CONWAGE: construction wage data (SIC code 162), which included heavy construction, construction of water and sewer mains, pipelines, power lines and construction of heavy projects which were not specified elsewhere. ENGWAGE: Engineering wages data (SIC code 8711), which consists of engineering services like designing ship boats, industrial, civil, electrical and mechanical engineers, machine tool designers, marine engineering services and petroleum engineering services. LANDWAGE: Landscape wage data (SIC code 078), which included landscape counseling and planning, lawn and garden services, and ornamental shrub and tree services. The national average annual hours worked were also collected from BLS for each of these wages separately. The wage per hour was then calculated by dividing the weekly wage by the calculated hours worked per week. LANDVAL: the value of the particular use of land (residential/commercial) on the outskirts of the city on which the stormwater pond was located. It was calculated as the average of the ten randomly selected land prices from the tax assessors database. QUANVOL: Water-Quanity Volume the runoff from the drainage area for a ten-year storm event for the CWP and the WRP data. Measured as 0.5 times the drainage area for data from the report by Wossink and Hunt. QUALVOL: Water-Quality Volume Responses of survey done for the CWP and the WRP data. Measured as 0.24 inch times the drainage area for the stormwater ponds and was assumed equal to the QUANVOL for the bioretention cells for the data from the report by Wossink and Hunt. The data was classified into three different major land resource areas according to their locations: Piedmont region: have a clayey or loamy subsoil Coastal region: have a sandy surface layer with loamy subsoil. Sandhill Region: sandy subsoil ESTTOTCST: Estimated Total Cost (used for bioretention cells), consisted of design and engineering and construction cost. Construction costs: consisted of excavation and grading cost, cost of materials, cost of the control structures, cost of the sediment control practices, landscaping cost, and the appurtenance cost. ESTTOTCSTLND: Estimated Total Cost including land cost (used for stormwater ponds), calculated by adding land cost (surface area of the pond times the LANDVAL) to the total ESTTOTCST. The estimated total costs, land values and the three wages were appropriately adjusted, using the historical cost indices, to correspond to the year 2003 in Baltimore, Maryland. Baltimore was chosen as the point of reference because of its frequent use as a central location in the study. Pollutant removal data for both the BMPs were collected from six different sources, including the National Best Management Practice Database. Econometric Model and Estimation Procedure The simple model used in this study is specified as follows: where e is the error term, and WQV is used for both water-quantity and -quality volume. Logarithmic transformation of the above equation gives us: LESTTOTCST ln( a) b ln( WQV ) u (Model 1) PROC REG procedure in SAS performed simple linear regression based on the equation above. QUALVOL was used for bioretention cells. QUANVOL was used for stormwater ponds. Values of 0<b<1 indicate the presence of economies of size. Logarithmic transformation of the complicated model incorporating the various inputs and land cost gives us: LESTTOTCSTLND Intercept cLQUANVOL dLCOASTQNV eLQUALVOL fLLANDVAL gLENGWAGE hLCONWAGE iLLANDWAGE u1 (Stormwater Ponds) LESTTOTCST Intercept jLQUANVOL kLQUALVOL lLCOASTQLV mLSANDHILL QLV nLENGWAGE oLCONWAGE pLLANDWAGE u2 (Bioretention Cells) The water-quantity volume of a BMP is assumed to increases by the same amount as the increase in the water-quality volume. The economies of water-quality size for bioretention cells located in the Piedmont region is as follows: LESTTOTCST QUALVOL k j LQUALVOL QUANVOL The average of the ratio of QUALVOL to QUANVOL for those cells located in the piedmont region is considered for the above equation. These models were tested negative for heteroscedasticity (error term is constant and not related to any of the variables in the model) and spatial correlation (error terms are not related to each other due to the presence of location difference in the data). These cost functions are also tested for the homogeneity restriction of the factor prices, giving us a third set of results for each of the BMPs Results for Bioretention Cells Estimates and p-value Variable Name Model 2 Model 3 (Restricted) Intercept -22.50173, 0.0029 -1.12931, 0.4997 LQUALVOL -0.76805, 0.0532 -0.73557, 0.1260 LCOASTQLV 0.14083, 0.0118 0.01586, 0.7075 LSANDQLV -0.16775, 0.0078 -0.23992, 0.0014 LQUANVOL 1.56436, 0.0006 1.70616, 0.0016 LENGWAGE 6.68941, 0.0058 0.64762, 0.6508 LCONSWAGE 0.88943, 0.3909 0.27093, 0.8283 LLANDWAGE -0.04984, 0.9595 0.08146, 0.9466 Adj. R-Square 0.7967 0.6880 Interpretations of Results for Bioretention Cells Both the QUANVOL and QUALVOL are significant determinants of the total adjusted costs of a cell. The effect of the water-quality volume, QUALVOL, on the cost is less by about 0.17% (Model 2) when the cell is located in the Sandhill region. Location of a cell in the Sandhill region can be expected to achieve low transportation cost of sand-one of the materials required for the building of a cell. Every 1% increase in the QUALVOL increases the total costs of the cell by 0.76 % in the coastal region 0.73 % in the Piedmont region 0.63 % in the Sandhill region exhibiting economies of water-quality size in all the regions. The high value of the statistically significant estimate of the Engineering wage (ENGWAGE) can be due to various reasons: A typical bioretention cell can fit into a parking lot or a residential complex, requiring a high level of engineering sophistication for its construction. A highly paid engineer is likely to employ more sophisticated technologies to obtain superior results, causing a rise in the cost. Since the model does not consider material cost separately, a 6.69% increase in the costs maybe due to the better quality materials used by a highly skilled engineer. This high number can be attributed to possible measurement errors in the data coupled with presence of a bias in the engineering wage. Model 3 cannot be used for our analysis as the likelihood ratio test indicated that we reject the null hypothesis of homogeneity restriction. Results for Stormwater Ponds Estimates and p-value Variable Name Model 2 Model 3 (Restricted) Intercept -0.35585, 0.8731 -2.09249, 0.0710 LQUANVOL 0.84162, <.0001 0.84303, <.0001 LCOASTQNV 0.01290, 0.4219 0.01648, 0.2901 LEXTDEQNV -0.05942, 0.0031 -0.06000, 0.0027 LQUALVOL -0.02063, 0.8619 -0.02564, 0.8283 LLANDVAL 0.34192, 0.0016 0.30916, 0.0021 LENGWAGE -0.32427, 0.6896 -0.13904, 0.8592 LCONSWAGE -0.23756, 0.8212 -0.30876, 0.7679 LLANDWAGE 0.48586, 0.6885 1.13865, 0.2450 Adj. R-Square 0.8580 0.8586 Interpretations of Results for Stormwater Ponds Assuming QUALVOL remains constant, 1% increase in the QUANVOL of the pond increases the total costs by 0.84% in the Piedmont region, suggesting economies of water- quantity size. LANDVAL is highly significant. For every 1% increase in the value of a unit of land, total costs of the stormwater pond increase by 0.19%. Total cost would be lower by 0.06% for extended detention ponds compared to wet ponds, for every 1% increase in the QUANVOL, holding QUALVOL constant. They are expected to be deeper with smaller surface area than a wet pond having the same amount of water-quantity volume. As land cost is a major constituent of the stormwater ponds, the decrease in surface area will lower the land cost and hence the total cost of the pond. Comparison between the two BMPs Using model 2 for both the BMPs, we find the average effect of water-quantity volume on the total cost of each of the BMP at the given input prices in each region. Total Cost Bioretention cell Stormwater Pond Water-Quantity Volume 25832 cubic-feet Comparison between both the BMPS Estimated fixed costs are higher for stormwater ponds than bioretention cells. Cost of stormwater ponds increases at a slower rate for every one percent increase in the water-quantity volume compared to the bioretention cells. Finding the cross-over QUANVOL at which bioretention cells and stormwater ponds have the same cost indicate: A bioretention cell is a cheaper management practice than a stormwater pond in the Piedmont region for QUANVOL less than 25,832 ft3. A bioretention cell is a less expensive method of removing pollutants in any feasible volume of water than a stormwater pond in the coastal areas. Calculation of these cross-over volume assumes that both the BMPs remove the same amount of pollutant Table below on the amount of pollutants removed by the two suggests that bioretention cells on average remove more pollutant than a stormwater pond. Average Amount of Average Amount of Type of Pollutant Pollutant Removed Pollutant Removed by Stormwater Ponds by Bioretention Cells (mg/L) (mg/L) Copper 0.0042 17.03854 Lead 0.0110 12.69257 Zinc 0.0493 0.60375 Phosphorus 0.1221 0.59340 Nitrates and Nitrites 0.1385 0.14930 Nitrogen 0.1677 3.6810 Source for stormwater ponds: National Best Management Practice Database (EPA, 1999) Sources for bioretention cells: Inglewood demonstration project (EPA, 2000a) and Maryland’s Greenbelt and Landover field study (Davis) Conclusions Both bioretention cells and stormwater ponds exhibit economies of size after adjusting for time and space and incorporating the input prices. Bioretention cells are likely to be cheaper than stormwater ponds as land price increases. Bioretention cells are cost effective in the coastal region but stormwater ponds are cost effective for most volumes of water treatment in the Piedmont region. Mean QUANVOL for stormwater ponds 336,152ft3 while the cross-over volume over which they become cheaper is 25,832ft3. Notwithstanding lack of sufficient information on maintenance costs, these models can be used by the EPA to improve the accuracy of its estimates of costs of compliance with water-quality regulation.