Impacts of Motor Vehicle Operation on Water Quality in the United States - Clean-up Costs and Policies

Abstract

This paper investigates the costs of controlling some of the environmental impacts ofmotor vehicle transportation on groundwater and on surface waters. We estimate thatannualized costs of cleaning-up leaking underground storage tanks range from $0.8billion to $2.1 billion per year over ten years. Annualized costs of controlling highwayrunoff from principal arterials in the US are much larger: they range from $2.9 billion to$15.6 billion per year over 20 years (1.6% to 8.3% of annualized highway transportationexpenditures.) Some causes of non-point source pollution were unintentionally created byregulations or could be addressed by simple design changes of motor vehicles. A reviewof applicable measures suggests that effective policies should combine economicincentives, information campaigns, and enforcement, coupled with preventiveenvironmental measures. In general, preventing water pollution from motor vehicleswould be much cheaper than cleaning it up.

Key Words: non-point source pollution; groundwater pollution; motor-vehicletransportation; economic incentives.

1. Introduction

Most studies of the environmental impacts of transportation focus on air pollution, themain environmental externality associated with road transportation, or noise (Delucchi2000). Currently, there is no good estimate of the aggregate impact of motor vehicletransportation on water pollution (Litman 2002), and a review of the relevant literaturesuggests that many estimates of water externalities resulting from motor vehicletransportation are based on educated guesses. While the emphasis of recent regulationsleads us to surmise that these impacts are substantial, it is still very difficult to quantifythem reliably because motor vehicles are just one of several causes of non-point sourcepollution.

This paper is concerned with the costs of controlling water pollution from motorvehicles. It focuses on two problems that have attracted considerable media attentionover the last few years in the US, and particularly in California: leaking undergroundstorage tanks (LUSTs) and highway runoff.

There is evidence that residues from the operation of motor vehicles contributeheavily to non-point source and groundwater pollution. Pollutants from motor vehicles orfrom transportation infrastructure include sediments (from construction or erosion), oils and grease (from leaks or improperly discarded used oil), heavy metals (from car exhaust,worn tires and engine parts, brake pads, rust, or used antifreeze; Table 1), road salts, aswell as fertilizers, pesticide, and herbicides (used alongside roads or on adjacent land).

The US Environmental Protection Agency (EPA) (1996) estimates that up to halfof suspended solids and a sixth of hydrocarbons reaching streams originate fromfreeways. Vehicle-related particulates in highway runoff come mostly from tire andpavement wear (about a third each), engine and brake wear (about 20%), and exhaust(about 8%). Each year, millions of gallons of improperly discharged used motor oilpollute streams, lakes, and coastal areas. This should be cause for concern since onegallon of used oil can contaminate one million gallons of water. Not all pollutants foundin highway runoff, however, come from transportation activities. Roads collect pollutantsfrom many other sources, including agricultural runoff or wind-blown contaminants frommanufacturing and energy production.

Groundwater quality is also threatened. There have been more than 450,000confirmed fuel leaks from underground storage tanks (USTs) in the US, including 44,000in California (US Environmental Protection Agency 2005a). Because of these, manycommunities need to find alternative sources of freshwater. For example, Santa Monica,California, has lost 80% of its local water supply to MTBE contamination and since 2005oil companies responsible for this pollution have had to purchase replacement water at acost in excess of $3 million per year (US Environmental Protection Agency 2005b).

A comprehensive assessment of how motor vehicle transportation affects waterquality is too complex to be feasible, so the focus here is on leaking USTs, on highwayrunoff, and on water pollution resulting from the improper disposal of used oil, wastecoolant/antifreeze, and metal dust from brake pads either because these sources ofpollution are generally important or because they lead to the consideration of informativepolicy solutions.

2. Literature Review

Interest in storm-water runoff pollution in the US is not new: many engineering andpublic health papers examine pollutants in storm-water runoff, their potential healthimpacts, and the effectiveness of best management practices (BMPs) for removing them.

Storm-water Runoff
Heavy metals in storm-water runoff are of particular concern because of their toxicity,pervasiveness, and persistence. In an early study, Ellis et al. (1987) find that heavy metalscan make highway runoff chronically toxic to receiving waters. In their review, Davis etal. (2001) report that pollutant loads typically follow the pattern: Zn (20-5,000 μg/l) > Cu≈ Pb (5-200 μg/l) > Cd (1 and their empirical study reveals that brake wear is the largest contributor to copper loading (47%) in urban runoff while tire wear contribute 25% of zinc loading, the second largest after buildings.

Kayhanian et al. (2003) study the impact of VMT on highway runoff pollutantconcentrations. Concentrations are two to ten times higher for urban than for non-urbanhighways but non-urban highway runoff shows greater concentrations of total suspendedsolids, pesticides, and ammonia, which points to agricultural sources. They also cautionthat a simple linear relationship between annual average daily traffic and pollutants isunlikely because of weather patterns and land use. Many pollutant loadings exhibitseasonal variations: winter brings high concentrations of chlorides and sulfates fromdeicing salt (Legret and Pagotto 1999) while irregular rainfall complicates runoffmanagement. Over a long, dry season pollutants accumulate on road surfaces and enterreceiving waters during the first storm event (Han et al. 2006). Regular street sweepingcan help, although its effectiveness is still debated (Tobin and Brinkmann 2002).

Storm-water runoff has also generated significant public health concerns. Gaffieldet al. (2003) examine impacts from heavy metals in storm-water, which can often betraced to motor vehicle sources. According to Van Metre et al. (2000), vehicles (throughtire wear, oil leaks, or car exhaust) are a significant source of polycyclic aromatichydrocarbons, a known carcinogen, in water bodies.

There has also been considerable interest in the US in storm-water BMPs. Maestriand Lord (1987) identified vegetative controls (e.g. grassed swales); wet detention basins;infiltration systems; and wetlands as measures for run-off-control. Shutes et al. (1999)extend their work to the effective construction, operation, and maintenance procedures ofconstructed wetlands. Inconsistencies across these US BMP studies, however, havelimited their usefulness. Barrett (2005) relies on the California Department ofTransportation’s (Caltrans) BMP Retrofit Pilot Program to develop a methodology forcomparing BMPs. He finds that the degree of pollutant removal depends on interactionsbetween a BMP and the influent water quality, not just on BMP characteristics.

The cost of complying with US federal and state storm-water regulations has beenthe subject of lawsuits in California. As a result, state agencies have conducted researchto better estimate storm-water management costs. Currier et al. (2005) examine sixCalifornia municipalities that made good progress toward storm-water compliance; theyreport annual storm-water management costs ranging from $18 to $46 per household. Bycomparison, a 2004 survey of Orange County (CA) residents found that nearly 60% ofrespondents would pay at least $5 per month to curb urban runoff (Center for PublicPolicy 2004).

Underground storage tanks
The US Energy Policy Act of 2005 targets leak prevention and expands the use of theleaking underground storage tanks (LUST) Trust Fund. In the US, there were more than450,000 confirmed releases at underground storage tanks as of September 2005 andcleanups had been initiated on more than 421,000 of these (US Environmental ProtectionAgency, 2005a). Marxsen (1999), however, claims that the cost of addressing LUSTsmay exceed benefits; he reports that cleanup costs range from $100,000 to more than $1million when groundwater is involved. Rice et al. (1995) find that LUST contaminationtends to be shallow so it may not affect deeper public drinking water wells. In addition, ifthe source of the leak is removed, passive bioremediation processes may naturally contain the spread of contamination. A well-managed UST program that emphasizes leakdetection can reduce the overall cost of LUST damage.

Considerable attention has been paid in the US to groundwater pollution causedby methyl tertiary butyl ether (MTBE). Until recently, it was widely used as a fueladditive in the US EPA's Reformulated Gasoline and Oxygenated Fuel Programs. In theirrisk analysis of MTBE contamination in California, Williams et al. (2004) find, however,that volatile organic compounds (VOCs) other than MTBE are detected more often and athigher risk levels. In contrast, Moran et al. (2005) conclude from a national survey thatMTBE is detected at or above the rate of other VOCs.

3. Environmental Impacts of Non-point Source Water Pollution

Motor vehicles are a major contributor to non-point source (NPS) pollution, as smallquantities of various pollutants are emitted during vehicle use or improperly disposed ofat many different locations. A number of studies link heavy metals (e.g. Pb, Zn, or Cu) orhydrocarbon loadings of surface water with transportation. Heavy metals in highwayrunoff are not necessarily toxic because toxicity depends on chemical form andavailability to aquatic organisms. However, some heavy metals bioaccumulate in the foodchain and can become toxic to humans over the long run.

Sources of Surface Water Pollution
We consider three sources of NPS pollution for surface waters. Of these, used oil is likelythe main hydrocarbon source to runoff (Latimer et al. 1990).² According to the US Environmental Protection Agency (1996), road runoff carries hundreds of thousands oftons of oil. Additionally, improperly disposed used oil filters may account for 5% of usedoil discarded into the environment. Yet, used oil is the “single largest environmentallyhazardous recyclable material” (MARRC 2001).³

Like crude oil slicks, used oil can have devastative impacts on aquatic life.However, refined products such as motor oil and gasoline are more toxic than crude oils.First, they disperse more readily into water. Second, soft tissues absorb them more easily.Third, used motor oil often contains contaminants, such as chemicals added to boostengine performance, compounds produced during engine operation, or wastes mixed-induring disposal.

The severity of the environmental impacts of used oil depends on weather, watertemperature, geographic features, and characteristics of the oil itself. Whereas waveaction can quickly disperse an oil spill in open waters, oil contamination in calm waterscan persist for years, so natural recovery times can vary considerably.

Another source of non-point source pollution is used coolant/antifreeze, which typically consists of 95% ethylene glycol, a clear, sweet-tasting and highly toxic liquid.Millions of gallons of coolant/antifreeze are sold each year in the US yet only 12% isrecycled (Department of Toxic Substances Control 2001). Used coolant/antifreeze isespecially a problem for Do-It-Yourselfers (DIY) because current engine design makes italmost impossible to avoid spilling some product when it is changed.⁴ Enginecoolant/antifreeze can also contribute high biochemical oxygen demand (BOD) levels toStorm-water.

In addition, operating motor vehicle disc brakes contributes heavy metals to nonpointsource pollution. Interestingly, this source of pollution resulted from technologicalchange and new regulations. Indeed, until the end of the 1960s, most cars had encloseddrum brakes. Pads for these brakes typically contained asbestos but no metals.

In the early 1970s, stricter braking requirements and concerns for workers’ healthrelated to airborne asbestos led manufacturers to adopt disc front – drum rear brakingsystems with semi-metallic brake pads. These pads contain no asbestos, wear out moreslowly, and have good braking properties. Corporate average fuel efficiency standardsreinforced the adoption of semi-metallic pads by favoring front wheel drive cars. Discbrakes, however, are open to the environment, so each time semi-metallic brake padssqueeze against the wheels’ rotors, tiny amounts of metal dust, often copper butsometimes also zinc and lead, are deposited along the roadway and washed to waterbodies by rain or snow.

Releases from brake lining wear add up: a recent study estimates that theycontributed 53.8 metric tons of copper in 2003 (95% confidence interval: [31.9, 75.7]) tothe San Francisco Bay watershed for all motor vehicle classes (Sinclair Rosselot 2006).Unfortunately, national estimates are not available.

Sources of Groundwater Pollution
While used oil and used coolant/antifreeze pollution mostly affects surface waters,gasoline spills from leaking underground storage tanks (LUSTs) are a major source ofgroundwater pollution all over the US. Although severe leaks can create fire or explosionhazards, the primary environmental concerns associated with gasoline releases arevolatile organic compounds such as dissolved-phase benzene, toluene, ethylbenzene, andxylene. More than 1.6 million USTs have been permanently closed since the LUSTproblem first surfaced and the number of confirmed releases exceeds 450,000. Althoughcleanups have been completed on nearly 75% of these, there are still more than 150,000leaking registered and unregistered UST where clean-up has not started (items c + j + n inTable 2).

At the same time, more than half of the US population relies on groundwater forat least a portion of its drinking water and 80% of community drinking water systems aredependent on groundwater (US Environmental Protection Agency 1994). LUSTs aretherefore a significant environmental problem. Table 2 summarizes key UST statistics.

Until the mid-1980s, most gasoline USTs were made of bare steel, whichcorroded over time, although connectors and pipes also caused many leaks (USEnvironmental Protection Agency 2001). With increasing awareness of the costs ofgasoline leaks, Congress banned the installation of unprotected steel tanks and piping in1985. According to the State Water Resources Control Board (2006a), 80% of USTs nowmeet California regulations for both release detection and prevention requirements.However, many leaks remain undetected because monitoring is inadequate and manyUSTs are inactive or abandoned (Farahnak and Drewry 1997).

4. Clean-up Costs Estimates

To quantify the costs of cleaning up LUSTs and of controlling highway runoff, we use asocial discount rate of 7% (nominal annual), as recommended by the Office ofManagement and Budget (OMB Circular No. A-94 Revised), and a real social discountrate of 4%. Unless specified otherwise, amounts are in 2005 dollars.

Highway Runoff Control Costs
In general, highway runoff control costs are difficult to quantify because practicalexperience is still relatively limited. For a given site, these costs depend on precipitation,soil and vegetation characteristics, traffic intensity, land availability, proximity ofmaintenance bases, and of course on the regulatory framework.⁵ To capture the uncertainty surrounding BMP costs, we consider two scenarios and two levels of BMPimplementation for which we report construction as well as operation and maintenance(O&M) costs. Costs are annualized over twenty years to limit the burden on limitedpublic resources.

There has been extensive research in California on quantifying highway runoffcontrol costs. Caltrans’ Storm Water Quality Handbook (2002) estimates costs at$100,000 per lane mile for rural highways and $250,000 per lane mile for urban ones.

Implementing BMP during initial construction may add as little as $15,000 per lane(2002 dollars) in rural areas ($90,000 in urban areas). By contrast, experienceaccumulated in Maryland suggests that BMP costs range from $45,000 to $60,000 perlane mile for rural roads and from $150,000 to $300,000 per lane mile for urban roads,which is comparable to California data (in 2002 $).⁶ In Washington State, the average weighted cost of implementing runoff BMP was $319,000 per lane mile for 18 recenturban and rural projects dealing with 644 lane miles, admittedly a very small sample.⁷ Although $319,000 per lane mile is substantial, it represents only a small percentage of project costs (from 0.45% for large rural projects to 8.99%, for small urban ones).

Maintenance costs also need to be accounted for, as it is essential to insure thatBMPs function properly. A 2001 survey conducted for the Washington Department ofTransportation by Herrera Environmental Consultants provides some data onconstruction as well as O&M costs for storm-water BMPs. Treatment and detentionponds are most common; as a percentage of construction costs, their annual O&M costsvary between 0.2% for larger basins and 5% for smaller ones. Infiltration basins areslightly more expensive (from 4 to 7%), but not as much as infiltration trenches (from 9to 12%). A wider range is observed for swales (from 3.7 to 11.5%) and even much moreso for vegetated filter strips (from 0.9 to 200%) because their construction costs can bevery low. To simplify our analysis, we suppose that necessary right-of-ways are alreadyavailable but we compensate for this assumption by using much more expensivemaintenance and operations costs for urban highways.⁸ Moreover, we assume that it will take 20 years to implement BMPs and that BMPs need to be reconstructed after 20 years.

We consider two scenarios. In the low cost scenario, constructing BMPrespectively costs $16,230 and $97,380 per lane mile for rural and urban highways, andthe corresponding annual O&M costs are 1% and 3% of construction. Targeting onlyprincipal US arterials still represents approximately 126,000 miles of rural roads and88,000 miles of urban roads (at the end of 2005), with an average of 3.26 lanes for theformer and 4.72 lanes for the latter. Key road statistics are summarized in Table 3.⁹

In the high cost scenario, BMPs are now, respectively, $64,920 and $324,599 perlane mile for rural and urban highways, and the corresponding annual O&M costs are 3%and 9% of the construction budget. Moreover, costs increase by 1% per year in realterms. Indeed, the composite index for federal aid highway construction increased by3.34% on average between 1993 and 2004 in nominal terms, while inflation during thatperiod was approximately 3% before Hurricane Katrina (US Department ofTransportation 2006).

To better grasp the magnitude of control costs on public finances, we comparethem to highway transportation expenditures. For the country as a whole, highwaytransportation expenditures reached $121.6 billion in 2001, the last year for which thisstatistic is currently available. To extrapolate these expenditures into the future, weassume they grow at a rate of 3.2% per year in real terms (the average between 1990 and2001). For California, highway transportation expenditures reached $10.6 billion in 2000and are assumed to grow by 4% annually in real terms.

Our results are thus driven by several assumptions. First, BMP construction costsfor rural roads are only one fifth or less of the value for urban roads partly to reflectdifferential land costs; likewise, O&M costs are three times cheaper in rural areas than inurban areas. Second, in our low cost scenario, we suppose that costs are constant in realterms thanks to technological improvements; in our high cost scenario, they increase by1% per year in real terms. Third, when we compare control costs to future highwaytransportation expenditures, we assume these grow annually by 3.2% in real terms (the1990-2001 average) at the federal level and by 4% in California. Tables 3 and 4summarize key road statistics and our cost assumptions.

Key results are presented in Table 5. If runoff is controlled only on principalarterials, annualized costs range from $2.9 billion to $15.6 billion, or 1.6% to 8.3% ofannualized transportation expenditures on highways. Extending control to all arterialsincreases these percentages, respectively, to 2.3% and 12.3% of annualized highwaytransportation expenditures. The difference between these estimates is mostly explainedby much higher O&M costs.

We also see that annualized control costs for California represent a largerpercentage of the state’s annual expenditures on highways because California has beenunder-spending on highways, it has a higher proportion of urban arterials compared to thecountry as a whole, and controlling runoff from urban arterials is much more expensivethan from rural arterials.

Groundwater Cleanup Costs
Groundwater cleanup costs depend on the level of contamination and on cleanupstandards. If only small volumes of soil need to be treated, cleanup costs can be as low as$10,000, but they can quickly exceed $1 million if extensive remediation is necessary.The presence of additives such as MTBE tends to substantially boost cleanup bills.Although costs vary widely across states and over time, they tend to increase becauselightly polluted sites were typically treated first and pollution spreads over time.

Getting a reliable estimate of cleanup expenses is difficult because no single levelof government has jurisdiction over all LUST sites, and nobody seems to be trackingfunds from federal, state, and private sources. Partial information suggests that completedand on-going cleanups already required considerable sums. For example, as of December2006, more than $2 billion of California’s UST Cleanup Fund had been spent (StateWater Resources Control Board, 2006b).

To evaluate cleanup costs, we assume that only half of all unregistered andabandoned USTs will be found (a US EPA assumption), so 181,336 USTs need to betaken care of (item q in Table 2); dealing with this backlog will take approximately 10years so 18,130 USTs are cleaned up every year, in addition to new leaking USTs; andthere are on average 2.61 tanks per site. To evaluate cleanup costs, we then consider twoscenarios.

In the low cost scenario, the cleanup cost at closure of a site is the 1997-2005average or $90,050, and it does not change over time.¹⁰ Moreover, we assume that the number of UST sites remains constant,¹¹ only 2.5% of UST leak every year, and cleaning them up costs a quarter of $90,050 per site because leaks are detected early. In the high cost scenario, the cleanup cost at closure in 2006 is instead the maximum annual value between 1997 and 2005 ($115,345), and it increases by 10% per year thereafter. Moreover, an additional 10% of UST begin to leak every year, and cleaning them up costs a quarter of $115,345 per site. These estimates may be over-conservative, however, if the current trend away from UST in favor of above ground storage tanks (AST)Continues.

Tables 2 and 4 summarize UST statistics and cost assumptions. The annualizedcost of cleaning-up LUSTs in the US is between $0.8 billion and $2.1 billion per year, orbetween 0.5% and 1.3% of annualized highway expenditures. The correspondingpercentages for California are slightly higher because California has been spendingproportionately less on highways per capita than the country as a whole.

Overall Estimate
As shown in Table 5, when we combine groundwater and highway runoff pollutioncontrol costs, we obtain annualized values ranging from $3.7 billion to $17.7 billion ifBMPs are installed only on principal arterials; this corresponds to a range of 2.0% to9.6% of annualized highway transportation expenditures. If BMPs are installed andmaintained on all arterials, this range jumps to between 2.8% and 13.6%. Californiaestimates are substantially higher still for two reasons: under-investment in highways anda larger percentage of urban arterials.

These estimates are driven by highway runoff control costs, which dominategroundwater pollution costs almost by an order of magnitude even though they areannualized over a longer period. The share of highway runoff control costs is even largerafter the backlog of leaking USTs has been cleaned up.

These large costs reflect the reach of the US transportation system, and they resultfrom the inadequate design of most of the current transportation infrastructure forprotecting water quality. Under our scenarios, these estimated control costs wouldrepresent a large drag on public budgets over many years, but cleanups are mandated bylaw and they are consistent with the “polluters pay” principle. It is therefore essential tocarefully weigh policy options.

5. Policy Considerations

Cost is understandably one of the main concerns about controlling highway runoff. Sincenon-point source pollution is linked to the operation of motor vehicles, an increase in the gasoline tax could be considered to finance BMPs. A $0.01 increase in the gasoline taxprovides approximately $1.5 billion in additional annual revenues, so a $0.118 gas taxraise would provide enough funds for cleaning up the backlog of leaking USTs as well asconstructing and maintaining BMPs on principal arterials for the high cost scenario.Gasoline taxes are already financing the federal LUST trust fund, although at a muchmore modest level.

Unfortunately, raising gasoline taxes has been very unpopular with legislators formany years: indeed, fuel taxes would have to increase by 11 cents per gallon on averagejust to go back to their 1957 purchasing power (Wachs 2003).

An alternative would be to rely on use fees, which are more efficient and moreequitable than other financing mechanisms such as bonds or general sales taxes.Electronic tolls, which have benefited from recent technological advances, appearespecially promising. However, increasing their use will take time, and their publicacceptance is not guaranteed.

While financing issues are being discussed, it appears wise to adopt policiesdesigned to reduce the contribution of motor vehicles to non-point source pollution.

Dealing with Non-Point Source Pollution
For non-point source pollution, “standard” instruments such as the establishment ofperformance standards or taxes may not be effective for several reasons.

First, it is by nature complicated to establish the relationship between sources andpollutants. Indeed, non-point source pollution results from a very large number of actionsreleasing small amounts of pollutants, whether voluntarily (used oil) or not (metal dustfrom brake pads). Second, non-point source pollution is not easily cleaned up. Third,there is often substantial uncertainty regarding the environmental and health impacts ofsome pollutants because of random factors such as precipitation, flow conditions,temperature, or insufficient toxicity data. Finally, when some non-point source pollutantstransfer from one medium to another, they undergo chemical transformations that affecttheir toxicity (e.g., Chromium).

Effective policies are thus likely to combine a series of measures including publiceducation, economic instruments (such as deposit refund systems for used oil), andpartnerships with industry. Non-structural BMPs such as street sweeping have also beenrecommended, but their effectiveness for smaller particulates has been questioned (Tobinand Brinkmann 2002).

In spite of limited success in the past, policy makers should also continueexploring the feasibility of water quality trading (WQT) programs including highwayrunoff. Such programs could greatly lower the costs of preserving water quality iftransaction costs can be reduced thanks to better hydrologic models combined withgeographic information systems, and well-designed institutions. This approach hasattracted increasing interest over the last few years. Recently, Farrow et al. (2005)proposed criteria to address common WQT implementation problems, and Obropta andRusciano (2006) presented an approach for evaluating the suitability of WQT trading in awatershed. Fang, Easter and Brezonik’s study (2005) suggests that this approach can besuccessful.

Some Specific Policies

Let us examine how this applies to some aspects of transportation-related non-pointsource pollution, starting with used oil.

In the US, only half of all used oil is recycled, so millions of gallons of used oilare still discharged into the environment each year (US Environmental Protection Agency1996). One way to increase recycling rates would be to target Do-It-Yourselfers (DIY),who are responsible for most of the improperly disposed used oil. In a 2002 survey ofCalifornia DIY conducted by Browning and Shafer, 97% of respondents indicated theywould be more likely to recycle if facilities paid more than $0.16 per gallon of used oil;in fact, 56% of respondents asked for at least $2/gallon. Increasing fees on lubricating oilwould provide dedicated funds to help open more recycling centers, boost publiceducation, and step up enforcement. Indeed, although dumping used oil in theenvironment is illegal, prosecutions are rare. Public-private partnerships could also becost-effective, as illustrated by the Canadian experience (Nixon and Saphores 2002).

Much more could be done for used oil filters. According to the FilterManufacturers Council (FMC), only 50% of used filters were recycled in the US in 2006.By contrast, three Canadian provinces (Alberta, Manitoba, and Saskatchewan) haveachieved 80% recycling rates by implementing economic incentives (Nixon and Saphores2002). Unfortunately, the FMC rejects economic incentives in favor of public educationand landfill bans, even though bans may encourage illegal disposal.

A used oil filter collection pilot program conducted in 1995-1997 in Californiarevealed some of the obstacles encountered by this type of program (California IntegratedWaste Management Board 1998). It suffered from limited public knowledge, a smallnumber of collection facilities, and reimbursements to businesses below hauling costs.Recycling was also impaired by a State law that forbids combining used oil and oil filterreimbursement checks, so check processing costs often exceeded their face value.

Another source of non-point source pollution is used coolant/antifreeze. In spiteof its toxicity, there are currently no programs to promote its recycling. A considerablyless toxic coolant/antifreeze based on propylene glycol (instead of ethylene glycol) ispopular in some European countries, but its US market share is only 10%. Better publicinformation may entice manufacturers to switch to propylene glycol and to modify enginedesigns to limit spills. Environmental NGOs could also facilitate changes, as they havefor metal dust from brake pads.

In the absence of direct regulations or economic incentives, environmentalproblems associated with the metal content of brake pads have been addressed bynegotiation, as discussed by Coase (1960). Along with the Stanford Law School,Sustainable Conservation (a Northern California NGO) created the Brake Pad Partnershipin 1996 to bring together businesses, government regulators, storm-water managementagencies, and environmental organizations. As a result, automobile parts manufacturersare conducting research to reduce metal use in friction materials. Apart from regularstakeholder meetings, ongoing activities of the Partnership include environmentalmonitoring and modeling studies (Brake Pad Partnership 2006).

Proactive versus Reactive Policies

To date, government policies for dealing with transportation-related water pollution havebeen mostly reactive instead of proactive. This is particularly the case for LUSTs. In retrospect, it would have been much cheaper to prevent leaks through enforcement andmonitoring. Indeed, according to Sausville et al. (1998), in the late 1990s, annualadministrative costs for compliance activities were less than $60 per tank (in 1998 $).This compares with approximately $2800 per tank per year for administrative costs ofcompliance activities during a site clean-up (for 5 years on average), not to mentioncleanup costs. By contrast, detection and monitoring costs are small: the conventional testfor USTs (which detects ~0.1 gallon/hour) costs $600 to $700, while enhanced tests,which are 20 times more sensitive, cost between $1500 and $1700.¹²

A case for incorporating environmental concerns during design can also be madefor highway runoff. Experience accumulated in Maryland and other states shows thatdesigning and implementing BMPs is much cheaper for new roads (often by a factor of 3or more) or during repair than it is for retrofitting existing roads if special constructionprojects are required. Reducing the large costs of implementing BMPs for highwayrunoff may thus require altering the design of new infrastructure (incorporating theprinciples of design for the environment, as recommended in Graedel and Allenby 1998)and waiting for road repair to install BMPs.

A similar proactive approach for dealing with transportation related pollutantscontributing to nonpoint source pollution is also likely to be cost effective, althoughenvironmental benefits are difficult to quantify in this case.

6. Conclusions

Our inquiry shows that the costs of controlling the impacts of motor vehicles on waterquality are substantial; we estimate that annualized costs of controlling runoff fromprincipal arterials only could cost between 1.6% and 8.3% of annualized highwaytransportation expenditures, while the annualized cost of cleaning up leaking USTs wouldcost an additional 0.5% to 1.4% per year for 10 years. Gasoline leaks, as well asimproperly disposed used oil, waste coolant/antifreeze, and metal dust from brake padsall contribute to non-point source water pollution. Their impacts on water quality as wellas other aspects of motor vehicle transportation are not yet well understood. This study,however, reveals several interesting stories.

First, a number of current environmental problems caused by the operation ofmotor vehicles are due, at least indirectly, to regulations designed to address otherproblems (so-called “intervention failures”). This is the case for MTBE, which wasoriginally introduced to reduce harmful emissions of ozone, or for heavy metals in brakepads after asbestos was abandoned because of health concerns.

Second, as motor vehicle pollution is often released in tiny amounts at a time bymillions of people, implementing pollution reduction programs can entail substantialtransaction costs, as illustrated by the difficulties encountered by the California oil filtercollection pilot program. Experiences in other countries such as Canada, or in otherindustries (e.g., aluminum containers), indicate, however, that it is possible tosuccessfully implement deposit refund programs to collect and recycle items such as usedoil or oil filters.

Third, NGO could have an important role to play in negotiating with industry inorder to make motor vehicle transportation more environmentally friendly, as illustratedby the Brake Pad Partnership.

Finally, the severity of several environmental problems (e.g. UST leaks) couldhave been limited if environmental considerations had been incorporated at the designstage instead of fixing problems later through costly regulations, economic instruments,or re-designs.

Acknowledgements

The financial support of UCTC for this project is gratefully acknowledged. We are alsograteful to Marc Delucchi, Ken Small, and Lyn Long, as well as to Dan Sperling, KenButton, and two anonymous referees, for very helpful comments. All remaining errors areour responsibility.

References
Barrett, M. (2005). Performance comparison of structural storm-water best managementpractices. Water Environment Research 77, 78-86.

Brake Pad Partnership (2006). Brake Pad Partnership Update. Available from:www.suscon.org/brakepad/pdfs/UpdateFall2006.pdf.

Browning, R. and Shafer, H. (2002). Outreach Research—Survey and Focus GroupsDIYers and Used Oil Disposal Initial Results and Recommendations. CaliforniaIntegrated Waste Management Board, Sacramento.

California Integrated Waste Management Board (1998). Residential Used Oil FilterCollection Pilot Program Report. Publication #333-98-001, Sacramento, CA.

Caltrans (2003). Caltrans Public Education Research Study. Sacramento: Caltrans StormWater Program.

Caltrans (2004). BMP Retrofit Pilot Program Final Report. Report # CTSW-RT-01-050.Division of Environmental Analysis. Sacramento, CA.

Center for Public Policy (2004). Close to 65 Percent of OC Survey Respondents Say TheyWould Pay Extra Money to Battle Urban Runoff. Irvine; California State University,Fullerton, Orange County Business Council.

Coase, R. (1960). The problem of social cost. The Journal of Law and Economics 3, 1-44.

Currier, B. K., Jones, J., M. and Moeller, G. L. (2005). NPDES Storm-water Cost Survey.Sacramento: Office of Water Programs, California State University, Sacramento.

Davis, A. P., Shokouhian, M. and Ni, S. (2001). Loading estimates of lead, copper,cadmium, and zinc in urban runoff from specific sources. Chemosphere 44, 997-1009.

Delucchi, M. A. (2000). Environmental externalities of motor-vehicle use in the US.Journal of Transport Economics and Policy 34, 135-168.

Department of Toxic Substances Control (2001). Antifreeze Recycling. Sacramento.Ellis, J. B., Revitt, D. M., Harrop, D. O. and Beckwith, P. R. (1987). The contribution ofhighway surface to urban storm-water sediments and metal loadings. The Science ofthe Total Environment 59, 339-349.

Fang F., Easter, K. W. and Brezonik, P. L. (2005). Point nonpoint source water qualitytrading: a case study in the Minnesota River Basin, Journal of the American Water Resources Association 41, 645-658.

Farahnak, S. and Drewry, M. (1997). Are Leak Detection Methods Effective in FindingLeaks in Underground Storage Tank Systems? State Water Resources Control Board,Final Draft Report.

Farrow R. S., Schultz, M. T., Celikkol, P. and Van Houtven, G. L. (2005). Pollutiontrading in water quality limited areas: Use of benefits assessment and cost-effectivetrading ratios, Land Economics 81, 191-205.

Gaffield, S. J., Goo, R. L., Richards, L. and Jackson, R. J. (2003). Public health effects ofinadequately managed storm-water runoff. American Journal of Public Health 93,1527-1533.

Graedel, T.E., and Allenby, B.R. (1998). Industrial Ecology and the Automobile. PrenticeHall, New Jersey.

Han, Y., Lau, S.-L., Kayhanian, M. and Stenstrom, M. K. (2006). Characteristics ofhighway storm-water runoff. Water Environment Research 78, 2377-2388.

Kayhanian, M., Singh, A., Suverkropp, C. and Borroum, S. (2003). Impact of annualaverage daily traffic on highway runoff pollutant characteristics. Journal ofEnvironmental Engineering 129, 975-990.

Latimer, J. S., Hoffman, E. J., Hoffman, G., Fasching, J. L. and Quinn, J. G. (1990).Sources of petroleum hydrocarbons in urban runoff. Water, Air, and Soil Pollution52, 1-21.

Legret, M. and Pagotto, C. (1999). Evaluation of pollutant loadings in the runoff watersfrom a major rural highway. The Science of the Total Environment 235, 143-150.

Litman, T. (2002). Transportation Cost and Benefit Analysis: Techniques, Estimates, andImplications. Victoria Transportation Policy Institute, Victoria, Canada.

Maestri, B. and Lord, B. N. (1987). Guide for mitigation of highway storm-water runoffpollution. The Science of the Total Environment 59, 467-476.

Manitoba Association for Resource Recovery Corporation [MARRC] (2001). Dirty Oil.Winnipeg. Manitoba, Canada.

Marxsen, C. S. (1999). Costs of remediating underground storage tank leaks exceedbenefits. Oil and Gas Journal 97, 21-22.

Moran, M. J., Zogorski, J. S. and Squillace, P. J. (2005). MTBE and gasolinehydrocarbons in ground water of the US. Groundwater 43, 615-627.

Nixon, H. and Saphores, J.-D. (2002). Used oil policies to protect the environment: Anoverview of Canadian experiences. In K. Wang, G. Xiao, L. Nie, and H. Yang (eds)Conference proceedings of the 3rd international conference on traffic andtransportation studies; Vol. 1, ASCE, GuilinObropta, C. C. and Rusciano, G. M. (2006). Addressing total phosphorus impairmentswith water quality trading. Journal of the American Water Resources Association 42,1297-1306.

Rice, D. W., Dooher, B. P., Cullen, S. J., Everett, L. G., Kastenberg, W. E. and Grose, R.D. (1995). Recommendations to improve the cleanup process for California's LeakingUnderground Fuel Tanks (LUFTs). Lawrence Livermore National Laboratory, CA.Sausville, P., Spiese, R., Stiller, K., Jordan, P. and Crimaudo, S. (1998). Report Card onthe Federal UST/LUST Program. Association of State and Territorial Solid WasteManagement Officials.

Shutes, R. B. E., Revitt, D. M., Lagerberg, I. M. and Barraud, V. C. E. (1999). The design of vegetative constructed wetlands for the treatment of highway runoff. The Scienceof the Total Environment 235, 189-197.

Sinclair Rosselot, K. (2006). Copper released from brake lining wear in the SanFrancisco Bay area. Report prepared for the Brake Pad Partnership, January.Available from: www.suscon.org/brakepad/pdfs/BrakeSourcesReportFinal01-30-06.pdf.

State Water Resources Control Board (2006a). Underground Storage Tank QuarterlyReport; July 1, 2005 to September 20, 2005. California Environmental ProtectionAgency State Water Resources Control Board (2006b). Underground Storage Tank (UST)Cleanup Fund: Program Statistics. California Environmental Protection Agency.

Tobin, G. A. and Brinkmann, R. (2002). The effectiveness of street sweepers in removingpollutants from road surfaces in Florida. Journal of Environmental Science andHealth, Part A-Toxic/Hazardous Substances & Environmental Engineering A37,1687-1700.

US Environmental Protection Agency (1994). The Quality of Our Nation's Water: 1992Report to Congress. Washington, DC. Available from:www.epa.gov/305b/92report/92summ.pdf.

US Environmental Protection Agency (1996). Indicators of the Environmental Impacts ofTransportation. Publication # EPA 230-R-96-009, Washington, D.C.

US Environmental Protection Agency (2001). Overview of the Federal UST program.Washington, D.C. Available from: www.epa.gov/swerust1/overview.htm.

US Environmental Protection Agency (2005a). UST Corrective Action Measures for Endof Year FY 2005. Washington, DC: Office of Underground Storage Tanks.

US Environmental Protection Agency (2005b). Oil Companies Pay US EPA to SettleSanta Monica MTBE Cleanup Costs, Press release dated 2/16/2005.

US Department of Transportation (2006). Price Trends for Federal–Aid HighwayConstruction, First Quarter 2006.

Van Metre, P. C., Mahler, B. J. and Furlong, E. T. (2000). Urban sprawl leaves its PAHsignature. Environmental Science & Technology 34, 4064-4070.

Wachs, M. (2003). Improving efficiency and equity in transportation finance. TheBrookings Institution Series on Transportation Reform, Washington, DC.

Williams, P. R. D., Benton, L. and Sheehan, P. J. (2004). The risk of MTBE relative toother VOCs in public drinking water in California. Risk Analysis 24, 621-634.

Hilary Nixon
Department of Urban and Regional Planning, San Jose State University, San Jose, CA 95192 USA.

Jean-Daniel Saphores
Civil and Environmental Engineering Department, The Henri Samueli School of Engineering, University of California, Irvine, CA 92697 USA.