Environmental Performance Measures for State Departments of Transportation (2015)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

20 Each of the suggested measures presented in Chapter 3 generally addresses an environmental issue of meaningful significance, while focusing on desired outcomes over which state DOTs have at least some control, and providing value to decision makers and clarity to the public. Without good data, however, none of these measures will be usable. In Chapter 4, the results from proof of concept testing research to validate suggested measures are presented. Proof of concept testing was used to apply real data provided by 27 volunteer state DOTs to demonstrate the degree to which the measures in Chapter 3 are valid in terms of three key quantitative criteria: 1. Can Measure Be Applied Consistently by All or Most States? Does the measure allow fair comparisons among states as determined by a focus on a commonly encountered environ- mental issue that many states experience, availability of a common standard for data collec- tion among states, and avoidance of bias created by factors outside a state DOT’s control, such as highway system size, climate, ecosystems, or extent of urbanization? 2. Can Measure Be Reported Easily with Existing Data Sources or Data that is Easy to Generate? Can the measure easily be reported with existing data sources housed inside or outside the state DOT? This ensures the viability of the measure as one that multiple states are willing to invest in. 3. Is Measure Data Quality Credible and Defensible? Is the measure’s credibility acceptable? Comparative measurement is weakened by use of measures whose credibility in terms of their accuracy, calculation methods, or connection to desired outcomes is in question. Table 3 overleaf shows the states that participated in the testing for each measure. The proof of concept testing phase of the research reflects the disaggregated approaches in states today toward environmental performance measurement. No state could provide data for every one of the measures proposed in Chapter 3. Nonetheless, proof of concept testing demonstrates their viability within a subset of states. Each section in this chapter examines an individual measure and what was learned from the data collected for the states shown. Overview of Testing Results After proof of concept testing, the six measures ended up falling into one of three categories, which are described below. Table 4 provides a summary of testing results. • Air, Energy and Climate, and Materials Recycling Measures. For the measures of air qual- ity, energy and climate change, and recycling, proof of concept testing generally validated the availability of data and viability of measure calculation methods, suggesting these measures C H A P T E R 4 Measure Proof of Concept Testing

Measure Proof of Concept Testing 21 Comprehensive Statewide Data Obtained Experimental Data Obtained Air Energy/Climate Change Recycling Stormwater Wildlife/ Ecosystems Statewide Vehicle Emissions (16 states) Gasoline Consumpon per Capita (All states) DOT Fleet Alternave Fuels Use (15 states) RAP as % of Total Pavement (11 states) % of Roads Receiving Treatment (4 states) Ecosystems Self- Assessment Tool (6 states) CA CO DE FL GA IL IA ME MD MN MO NE NJ NM NC ND OH OR PA SC SD TX UT VT VA WA WY Table 3. Summary of participating pilot states.

22 Environmental Performance Measures for State Departments of Transportation Measure Support Does Measure Consistent Application from State to State? Can DOTs Report Measure Easily with Existing Data or Data That Is Easy to Generate? Quality Is Data Credible and Defensible? Implementation Readiness Change in statewide motor vehicle emissions for NOx, VOC, and PM2.5 Ready for use by many DOTs today State DOT fleet alternative fuel use Ready for use by many DOTs today Statewide on- road gasoline consumption Ready for use by many DOTs today Annual percent by mass of all asphalt pavement materials composed of RAP Ready for use by many DOTs today Impervious surface for which water treatment is provided Suitable for use by DOTs in longer term Ecosystems Self- Assessment Tool Ready for use by many DOTs today Key Measure is fully consistent with criteria Measure is mostly consistent with criteria Measure is somewhat consistent with criteria Measure lacks consistency with criteria Table 4. Summary of proof of concept testing results. can be adopted by several to many state DOTs in the near term. For these measures, compre- hensive data was obtained for many states. • Stormwater Measure. Prior to proof of concept testing, the stormwater measure was rec- ognized to be more experimental than other measures, with the performance measurement capabilities of most state DOTs in this area best described as being nascent. For this reason, testing was confined to a handful of states for the stormwater measure. While the stormwater measure should not be considered implementation ready, testing results suggest it shows strong promise and continued efforts to expand on the potential of the stormwater measure are encouraged with phased adoption over time. • Wildlife and Ecosystems Measure. While proof of concept testing results validated initial measurement suggestions for most of the measures considered in this report, the results for an initial wildlife and ecosystems measure “Percent of wetland and stream mitigation that achieves regulatory approval on, or ahead of, schedule based on the permitted monitoring period” demonstrated that the measure was unlikely to succeed. Proof of concept results for this original measure are included in Appendix A, but Chapter 4 contains additional informa- tion about proof of concept testing conducted to evaluate the ESAT in its place. The complete set of questions that make up the ESAT can be found in Appendix B.

Measure Proof of Concept Testing 23 4.1 Air Quality Measure: Proof of Concept Results Measure: Change in Statewide Motor Vehicle Emissions for NOx, VOC, and PM2.5 Proof of Concept Validation Implementation Readiness Does Measure Support Consistent State-to-State Application?  Ready for Use by Many DOTs Today Can DOTs Report Measure Easily with Existing Data or Data That Is Easy to Generate?  Suitable for Use by DOTs in Longer Term Is Data Quality Credible and Defensible?  Not Suitable for Use by Most DOTs Overview Pilot States: California, Colorado, Delaware, Florida, Illinois, Maryland, Missouri, North Carolina, North Dakota, New Jersey, Pennsylvania, South Dakota, Vermont, Virginia, Washington, Wyoming Motor vehicle emissions are calculated as the product of two components: vehicle activity (reported as VMT) and emission rates. As an example, Figure 2 shows how these two compo- nents have changed over time for Vermont; the reduction in emissions has been almost entirely due to lower fleetwide emission rates, which are generated by the introduction of newer vehicles that meet more stringent EPA-mandated emission standards and the retirement of older, higher emission vehicles. These trends are very similar in other states. The tables that follow show per- formance results for the 16 pilot states in terms of NOx, VOC, and PM2.5, respectively, between 2005 and 2011. Air Quality Performance Results Tables 5, 6, and 7 summarize proof of concept performance results. See Appendix C for full state-by-state air quality data gathered. VMT NOx PM2.5 VOC -40% -60% -20% 0% 2005 2006 2007 2008 2009 2010 2011 Figure 2. Illustration of relationship between VMT and emissions rates over time (Vermont).

24 Environmental Performance Measures for State Departments of Transportation Trend 2005-2011 2005 2006 2007 2008 2009 2010 2011 Change 2005-2011 California 714,712 654,873 609,501 561,268 497,180 453,950 415,170 -41.9% Vermont 22,935 21,152 19,155 16,792 15,371 13,557 12,329 -46.2% Washington 255,658 241,920 227,552 206,155 188,020 176,963 164,130 -35.8% Pennsylvania 279,829 257,313 237,824 217,207 186,708 166,890 152,621 -45.5% Virginia 209,649 194,251 181,234 165,662 144,770 134,851 122,530 -41.6% Wyoming 31,361 29,594 27,338 25,197 22,291 20,007 18,339 -41.5% S Dakota 27,397 26,904 24,374 22,358 19,447 17,928 16,696 -39.1% N Dakota 24,133 22,742 21,104 19,226 17,663 16,468 16,656 -31.0% N Carolina 248,307 228,042 213,917 191,834 171,687 156,710 146,002 -41.2% New Jersey 161,692 152,224 142,294 126,983 111,822 102,632 94,268 -41.7% Missouri 183,778 168,763 153,699 139,805 124,471 117,911 105,347 -42.7% Maryland 129,530 122,642 114,172 102,665 91,373 85,179 78,719 -39.2% Illinois 269,780 242,775 225,901 205,153 180,917 166,311 149,505 -44.6% Florida 471,765 442,237 413,975 368,808 325,919 300,298 271,901 -42.4% Delaware 21,762 19,878 18,476 16,005 14,365 13,042 12,183 -44.0% Colorado 127,526 118,591 109,939 99,331 84,446 78,606 72,385 -43.2% Table 5. Annual change in NOx emissions from 2005 base year. Trend 2005-2011 2005 2006 2007 2008 2009 2010 2011 Change 2005-2011 California 90,493 82,533 75,222 69,558 62,512 57,271 51,251 -43.4% Vermont 3,268 3,011 2,670 2,317 2,171 1,905 1,700 -48.0% Washington 116,229 110,568 103,669 94,947 90,022 85,339 78,180 -32.7% Pennsylvania 48,417 43,824 39,587 35,773 30,873 27,606 24,588 -49.2% Virginia 36,096 32,911 30,112 27,543 24,345 22,827 20,250 -43.9% Wyoming 3,029 2,858 2,548 2,342 2,099 1,909 1,705 -43.7% S Dakota 3,045 2,997 2,628 2,374 2,069 1,903 1,732 -43.1% N Dakota 2,807 2,652 2,341 2,123 1,961 1,834 1,815 -35.3% N Carolina 36,662 33,661 30,744 27,371 24,454 22,412 20,584 -43.9% New Jersey 23,533 21,567 19,513 17,110 15,098 13,825 12,429 -47.2% Missouri 25,136 22,513 20,291 18,025 16,106 15,211 13,267 -47.2% Maryland 16,835 16,072 14,283 12,565 11,149 10,402 9,395 -44.2% Illinois 37,204 34,049 31,161 28,469 25,821 24,223 21,884 -41.2% Florida 80,556 72,810 65,406 57,051 50,106 45,846 40,214 -50.1% Delaware 3,287 2,915 2,614 2,225 2,005 1,815 1,637 -50.2% Colorado 16,097 14,690 13,060 11,581 9,901 9,274 8,300 -48.4% Table 6. Annual change in VOC emissions from 2005 base year.

Measure Proof of Concept Testing 25 Air Quality Measure: Criteria Assessment Does Measure Support State-to-State Comparisons? • Measure Is Most Meaningful in States with Air Quality Problems. Thirty-four have Clean Air Act–designated nonattainment or maintenance areas within their boundaries. Therefore the measure may not be considered relevant in the remaining states and within the affected states, not all pollutants included in the measure are necessarily significant issues. There are 46 ozone nonattainment areas in the United States, involving 21 states. For PM2.5, 32 regions are designated nonattainment areas, involving 18 states. • Use of Nationally Accepted Data Supports Comparative Measurement. The air qual- ity measure’s reliance on a standardized model (EPA’s Motor Vehicle Emissions Simulator [MOVES]) and a consistent source for VMT data (FHWA’s Highway Statistics series) makes it appropriate for comparisons among states. Notably, California uses the EMissions FACtors (EMFAC) model in place of MOVES and parameters in the EMFAC model, such as statewide VMT, may differ from parameters in MOVES and Highway Statistics; thus, California results using EMFAC will not be directly comparable with other states. Do State DOTs Have Easy Access to Data/Is Measure Easily Calculated? • Use of MOVES Drives Ease of Calculation. If a state DOT has in-house expertise in using MOVES, calculation of the measure requires only modest effort; however, not all state DOTs have staff fluent in the operation of MOVES, which may somewhat reduce the simplicity for calculation of this measure. Trend 2005-2011 2005 2006 2007 2008 2009 2010 2011 Change 2005-2011 California 25,094 23,454 21,510 19,633 17,057 15,909 14,616 -41.8% Vermont 754 708 632 555 506 456 415 -44.9% Washington 8,230 7,846 7,269 6,587 6,001 5,727 5,328 -35.3% Pennsylvania 9,241 8,633 7,868 7,209 6,178 5,630 5,171 -44.0% Virginia 6,835 6,414 5,885 5,401 4,688 4,447 4,053 -40.7% Wyoming 1,018 971 880 812 720 658 602 -40.8% S Dakota 861 857 766 704 613 577 540 -37.3% N Dakota 759 726 665 608 560 533 542 -28.7% N Carolina 7,943 7,404 6,854 6,168 5,494 5,117 4,797 -39.6% New Jersey 5,622 5,370 4,952 4,431 3,879 3,639 3,365 -40.2% Missouri 5,893 5,484 4,913 4,472 3,958 3,813 3,420 -42.0% Maryland 4,586 4,271 3,894 3,499 3,094 2,936 2,720 -40.7% Illinois 9,066 8,261 7,573 6,892 6,056 5,684 5,137 -43.3% Florida 14,400 13,567 12,432 11,016 9,513 8,868 8,057 -44.1% Delaware 690 641 589 516 458 423 398 -42.3% Colorado 4,175 3,912 3,572 3,237 2,757 2,613 2,416 -42.1% Table 7. Annual change in PM2.5 emissions from 2005 base year.

26 Environmental Performance Measures for State Departments of Transportation Is Data Quality Credible and Defensible? • Use of Nationally Accepted Data Strengthens Data Quality/Credibility. The air quality mea- sure’s reliance on MOVES and FHWA’s Highway Statistics series lends credibility to this mea- sure. The measurement approach follows EPA guidance for statewide emission inventories; however, the suggested calculation approach is more simplistic than the county-by-county calculations states perform as part of the triennial National Emission Inventory, so results will not match the National Emission Inventory data. • Measure Data Is Not Sensitive to State DOTs’ Congestion Relief Efforts. The measure’s defensibility is somewhat limited because it does not reflect a state DOT’s congestion relief efforts. Emissions rates increase as speeds drop below 35 miles per hour, are relatively flat between 35 to 55 miles per hour, and increase above 55 miles per hour. Efforts to tackle very heavy congestion are recognized to result in emissions reductions at the corridor scale because they move speeds into the range where emission control technologies are most effective. At the state level, however, aggregate average speeds typically fall in the 40 to 50 miles per hour range and vary little from year to year. In that speed range, small changes in average speed will have no significant impact on emissions. 4.2 Fleet Alternative Fuels Measure: Proof of Concept Results Measure: State DOT Fleet Alternative Fuel Use (as percentage of total fuel use) Proof of Concept Validation Implementation Readiness Does Measure Support Consistent State-to-State Application?  Ready for Use by Many DOTs Today Can DOTs Report Measure Easily with Existing Data or Data That Is Easy to Generate?  Suitable for Use by DOTs in Longer Term Is Data Quality Credible and Defensible?  Not Suitable for Use by Most DOTs Key Measure is fully consistent with criteria Measure is mostly consistent with criteria Measure is somewhat consistent with criteria Measure lacks consistency with criteria Overview Pilot States: California, Iowa, Maine, Maryland, Minnesota, Missouri, Montana, Nebraska, New Mexico, North Carolina, Pennsylvania, South Carolina, Vermont, Washington, Wyoming Use of alternative fuels in place of conventional gasoline and diesel generally reduces GHG and criteria (Clean Air Act–regulated) pollutant emissions. Use of alternative fuels can also lessen dependency on petroleum, which has associated costs for society and can subject vehicle fleet owners to volatile fuel prices. Although the fuel consumed by state DOT fleets represents only a small fraction of total transportation fuel use, state DOTs control some of the largest public sector fleets. Their actions can support emerging mar- kets and set a positive example that helps make alternative fuels viable for a wider market.

Measure Proof of Concept Testing 27 “Alternative” transportation fuels include any fuel other than gasoline or diesel such as the following: • Ethanol. Ethanol is a renewable fuel made from various plant materials. As an alternative fuel, ethanol is typically blended at 85 percent with gasoline, known as E-85. • Biodiesel. Biodiesel is a renewable fuel made by reacting animal or vegetable fats with alcohol. Most biodiesel is used in low-level blends, usually 5 percent or 20 percent biodiesel blended with conventional diesel, referred to as B-5 or B-20, respectively. Most diesel vehicles can use low-level biodiesel blends without modification; higher-level blends require vehicle modifications. • Natural Gas. Natural gas is an odorless, gaseous mixture of hydrocarbons, predominantly composed of methane. Natural gas can be used as transportation fuel in CNG or liquefied natural gas (LNG) form. • Propane. LPG is commonly referred to as propane. Propane is mainly used in light-duty pickup trucks as well as some off-road equipment. • Electricity. Electricity can be used to power battery electric vehicles (BEVs) and plug-in hybrid electric vehicles (PHEVs), collectively known as plug-in electric vehicles (PEVs). All PEVs draw electricity from off-board electrical power sources, that is, the electricity grid, and store the energy in batteries. Conventional hybrid electric vehicles are not considered alterna- tive fuel vehicles. Note that virtually all gasoline sold in the United States contains 10 percent ethanol. Ethanol is added to gasoline to boost octane levels, meet air quality requirements, or satisfy mandates such as the EPA’s Renewable Fuel Standard. This ethanol fraction is not considered an alterna- tive fuel for the purposes of this study, because it is considered the conventional and default gasoline blend for all fleets. Figure 3 shows how the pilot test states performed over time in terms of the performance mea- sure. See Appendix C for full state-by-state fleet fuel use data gathered. The following pilot states CA (dashed), 3.5% IA, 1.9% MD (solid), 3.5% MN, 12.8% MO, 5.2% NC, 10.2% NE, 0.4% PA, 2.9% WA, 8.7% WY, 0.2%0.0% 5.0% 2008 2009 2010 2011 2012 10.0% 15.0% Figure 3. Alternative fuels as percentage of total DOT fleet fuel consumption volume basis: 2008 through 2012.

28 Environmental Performance Measures for State Departments of Transportation report that they do not use any alternative fuels in their fleet: Maine, Montana, New Mexico, South Carolina, and Vermont; therefore, they are not shown in Figure 3 or subsequent figures and tables. Other State DOT Fuel Use Performance Measure Formulations In its suggested format, this performance measure is calculated on a volumetric basis, which means that one gallon of gasoline, diesel, ethanol, biodiesel, or propane has the same weight when calculating the percentage of alternative fuel use for a state, regardless of each fuel’s energy content. An exception is natural gas, which is usually reported in terms of gasoline gallon equiva- lents and therefore already accounts for differences in energy content. As part of proof of con- cept testing, several other formulations of this measure were evaluated including formulations based on energy content, GHG intensity, and trends in fuel use: • Alternative Fuels Measure Based on Energy Content. An alternative formulation for this mea- sure is to normalize all fuel use by its energy content. Most alternative fuels have lower energy content per gallon than conventional fuels. For example, pure ethanol has about 33 percent less energy than gasoline and pure biodiesel has about 6 percent less energy than conventional diesel. To calculate the measure on an energy content basis, the volume of each fuel would be multiplied by the energy content of the fuel (in megajoules) per gallon. The energy for con- ventional fuels and alternative fuels would each be summed, and the ratio would serve as the alternative metric. Table 8 compares the alternative fuels measure for the ten pilot states for 2012 on a volume basis versus an energy basis. The alternative fuels fraction is slightly lower in all cases. If this measure were calculated by energy content rather than on a volume basis, states that use large volumes of E-85 would tend to see their performance drop more than states that use large volumes of biodiesel, natural gas, or propane (see Figure 4). This is because ethanol’s energy content is lower than most other alternative fuels. In theory, a measure based on energy content could incentivize state DOTs to focus their efforts more on alternative fuels with higher energy content. Since the fuel options for state DOTs are limited, and because equipment type often dictates fuel choices, the energy content formulation for this measure would likely have little or no practical effect on their actions. • Alternative Fuels Measure Based on GHG Intensity. Fuel use information can also be used to calculate GHG emissions from a state DOT fleet. (One of the primary benefits of using alter- native fuels is their GHG benefit.) To properly compare the climate change impacts of various fuels, GHG emissions should be calculated on a “well-to-wheels” basis that reflects emissions associated with fuel production and distribution as well as vehicle tailpipe emissions. The Greenhouse Gases, Regulated Emissions, and Energy Use in Transportation (GREET) model developed by Argonne National Laboratory can be used to obtain “well-to-wheels” GHG emission factors for each fuel. Fuel use (by type) is multiplied by the appropriate emission factor, and the results summed to obtain total state DOT fleet GHG emissions. The magnitude of each state DOT’s GHG emissions would correlate roughly to the size of the state DOT’s fleet. To enable meaningful comparisons among states, the GHG emissions should be normalized, for example, by dividing GHG emissions by a measure of fuel use, such as total energy content. This approach CA IA MD MN MO NC NE PA WA WY Volume Basis 3.5% 1.9% 3.5% 12.8% 6.2% 10.2% 0.4% 2.9% 8.7% 0.2% Energy Basis 3.1% 1.9% 3.3% 9.7% 5.9% 9.8% 0.2% 2.8% 8.4% 0.2% Table 8. Comparison of alternative fuels measure on volume and energy basis (2012).

Measure Proof of Concept Testing 29 would create a composite GHG emission factor for a state DOT’s fleet that allows state-to-state comparisons. Table 9 shows this value for the 10 pilot states for 2012. A lower number is consid- ered better because it indicates a fuel mix that produces less climate change impacts per unit of energy delivered. Figure 5 shows the trend over time on GHG intensity basis for pilot states. A GHG-based performance measure is attractive because it more closely tracks one of the primary environmental outcomes of interest (reduced climate change impacts). However, the measure based on GHG intensity is less transparent and understandable to the public. Moreover, alternative fuels have benefits other than GHG reduction, such as reduced criteria pollutant emissions and reduced dependency on petroleum and its associated price volatility. There is uncertainty, and in some cases controversy, with regard to the upstream emissions associated with alternative fuels, particularly biofuels. While the GREET model is widely con- sidered the most authoritative source of lifecycle GHG emission factors for motor vehicles, there continues to be research on the indirect land use effects of biofuels. These upstream GHG emis- sions depend heavily on the fuel feedstock and can vary by state. Notably, California has its own version of GREET. While these uncertainties are present in every model to some degree and do not represent a fatal flaw, they are likely to limit the adoption of a measure based on GHG intensity in some states and reduce the measure’s transparency. • Trends in Total Fuel Use. Fuel use data can also be used to track trends in a state DOT’s fleet consumption and fuel efficiency. In addition to environmental implications, fuel is a major operating expense for state DOTs, and agencies seek to minimize fuel use through operational improvement such as less vehicle idling or technology improvements like using more fuel- efficient vehicles. If a state DOT maintains data on the annual mileage of all its vehicles, it can calculate an aggregate fleet fuel economy either for the entire state DOT fleet or separately for CA (dashed), 3.1% IA, 1.9% MD (solid), 3.3% MN, 9.7% MO, 5.0% NC, 9.8% NE (dashed), 0.2% PA (solid), 2.8% WA, 8.4% WY, 0.2%0.0% 5.0% 10.0% 15.0% 2008 2009 2010 2011 2012 Figure 4. Alternative fuels as percentage of total DOT fleet fuel use energy content basis: 2008 through 2012. CA IA MD MN MO NC NE PA WA WY Carbon Intensity 72.3 73.0 72.5 68.9 71.2 66.8 76.7 74.0 69.9 74.5 Table 9. State DOT fleet composite GHG intensity in grams CO2-equivalent per megajoule (2012).

30 Environmental Performance Measures for State Departments of Transportation light-duty and heavy-duty vehicles. This metric would facilitate comparison among states as well as performance tracking within a single agency. Some states do not have accurate records of the mileage of all their vehicles, so calculating aggregate fuel economy is not possible. An alternative is to track total fuel use, which reflects both the efficiency of the fleet and the use of the fleet. The benefit of this metric is that it would capture a state DOT’s efforts to reduce travel. A disadvantage of this metric is that variations in fleet usage may obscure miles per gallon improvements. For example, if a state DOT experiences a year with heavy demands on its fleet, such as a heavy snow fall year, it could see an increase in fuel use even if its efficiency improved. A metric based on total fuel use is also difficult to compare across states. One way to facilitate comparison is to report the percentage change in fuel use in relation to a given base year. Figure 6 shows the trend over time in fleet fuel use for selected pilot states. CA, 72.3 IA, 73.0 MD, 72.5 MN, 68.9 MO, 71.2 NC, 66.8 NE, 76.7 PA, 74.0 WA, 69.9 WY, 74.5 64.0 66.0 68.0 70.0 72.0 74.0 76.0 78.0 2008 2009 2010 2011 2012 Gr am s o f C O 2-E qu iv al en t Figure 5. Composite GHG intensity of DOT fleet fuel use in grams of CO2-equivalent per megajoule (2008 through 2012). Figure 6. Percentage change in total DOT fleet fuel use (base year 2008). California Maryland Minnesota N Carolina Pennsylvania -20% -10% 0% 10% 20% 2008 2009 2010 2011 2012

Measure Proof of Concept Testing 31 Alternative Fuels Measure: Criteria Assessment Does Measure Support State-to-State Comparisons? • Standardized Fuel Records and State DOT Control of Fleets Enables Fair Comparisons. In general, the measure supports state-to-state comparisons because of the consistent and accu- rate state DOT records on fleet fuel use and the relatively high degree of state DOT control over vehicle and fuel purchasing decisions. • Some External Factors Can Influence State DOTs’ Alternative Fuel Use. While the amount of alternative fuels being purchased and used in state DOT vehicles is partly an agency deci- sion, external factors that vary from state to state may affect usage: – Alternative Fuel Mandates Promote Use. Alternative fuel use in some states is driven by state mandates. Minnesota, for example, began requiring a 2 percent biodiesel blend for all diesel sold in the state in 2005, increasing to 5 percent in 2007. Beginning in July 2014, 10 percent biodiesel will be required from May to October of each year. Minnesota DOT scores well on this measure in part because of these state mandates, although Minnesota DOT has also exceeded the state requirement through its own actions. Similarly, since 2009, Pennsylvania has required that all diesel sold in the state contain a minimum of 2 percent biodiesel. – Supply Challenges Can Limit Use. Some state DOTs may also experience limits on their performance as a result of alternative fuel supply issues. In Vermont, VTrans was forced to curtail biodiesel use when their fuel vendor stopped supplying biodiesel in the state. Other states may have individual fleets or districts that lack alternative fuel supply. The remote location of Caltrans District 9, for example, prohibits use of E-85 that is used in most other Caltrans districts. Conversely, biofuels supply and fueling infrastructure is abundant in many Midwestern states. – Climate Can Limit Use. In cold weather states, some fleets have experienced problems with higher biodiesel blends in winter because the fuel can gel at low temperatures; thus, state DOTs in cold weather states may face limits on their ability to use biofuels. – State Purchasing Rules Can Limit Use. State purchasing rules—which are outside the direct control of a state DOT—can constrain the opportunities to use alternative fuels. In some states, state DOTs must purchase new vehicles through a statewide general services purchasing agreement. If these agreements do not include alternative fuel vehicle options, the state DOT will be unable to acquire such vehicles. Despite these limitations, the alternative fuels measure as presented in this report is a good one for all state DOTs to use for reporting. The measure is based on readily available data, enables comparison among states as well as progress tracking over time, reflects environmentally beneficial practices at state DOTs, and is an area of growing interest. Do State DOTs Have Easy Access to Data/Is Measure Easily Calculated? • Measure Based on Reported Fuel Use. State DOT records of fuel purchases are considered highly accurate. The measure is not, therefore, subject to the risk of error or uncertainty that can be caused by modeled or estimated data. • Data on Alternative Fuels Is Readily Available at Many State DOTs. Use of alternative fuels in state DOT fleets is usually simple to measure because state DOTs themselves generally maintain the data required. • Adjustments to Tracking Methods Are Needed in Some States, but Are Surmountable. In some states, including North Carolina and Vermont, the state DOT owns and maintains records only for heavy-duty vehicles. In these instances, light-duty vehicles used by the state DOT are managed by a state “motor fleet” department or general services administration and the state DOT may not have ready access to fuel records for the light-duty vehicles it uses. Some states do not use centralized fueling stations because their vehicles are fueled at commercial stations and their fuel purchase card technology may not support tracking of

32 Environmental Performance Measures for State Departments of Transportation alternative fuel purchases such as biodiesel blends or E-85. In these cases, additional effort may be needed to collect fully accurate data. Is Data Quality Credible and Defensible? • Measure Does Not Easily Capture Use of Electric Vehicles. This measure captures use of all liquid alternative fuels, but is likely to omit use of electricity as a transportation fuel. Among the state DOTs contacted for this pilot, only one reported owning a PEV as of 2013 (a single Chevrolet Volt owned by Washington State DOT). However, more states are expected to purchase and operate PEVs in coming years. Most state DOTs are unlikely to maintain data on grid-derived electricity supplied to PEVs. State DOTs that wish to capture electricity used in transportation can estimate it based on the VMT of BEVs and PHEVs. For BEVs, a simple fuel economy rating can be multiplied by VMT to derive total energy use. For PHEVs, the state DOT would need to estimate the portion of miles driven in all-electric mode, and estimate gasoline and electricity use for these vehicles. • Measure Is Easy to Understand. A volume-based measure is simple to calculate and easy for the public to understand. A measure based on fuel energy content or GHG intensity will be less easy to understand and communicate. 4.3 Gasoline Consumption Measure: Proof of Concept Results Measure: Statewide On-Road Gasoline Consumption per Capita Proof of Concept Validation Implementation Readiness Does Measure Support Consistent State-to-State Application?  Ready for Use by Many DOTs Today Can DOTs Report Measure Easily with Existing Data or Data That Is Easy to Generate?  Suitable for Use by DOTs in Longer Term Is Data Quality Credible and Defensible?  Not Suitable for Use by Most DOTs Key Measure is fully consistent with criteria Measure is mostly consistent with criteria Measure is somewhat consistent with criteria Measure lacks consistency with criteria Overview Pilot States: All 50 States On-road gasoline consumption per capita captures the environmental and energy inten- sity of a state’s highway transportation system. Carbon dioxide, which is the most important GHG, and gasoline use are directly correlated. Transportation agencies can contribute to a reduction in gasoline use through improvements in highway system efficiency, support for alternative modes, and deployment of fuels and vehicle technologies that displace gasoline. This measure generally provides a strong correlation with personal travel in light-duty vehicles. Gasoline engines power the vast majority of personal travel on roadways. The measure does not capture the activity of diesel powered trucks or alternative fuel vehicles. Diesel is primar- ily used for long-haul freight, the demand for which is often created outside of a state. Per- formance results for 2011 are shown in Figure 7. See Appendix C for complete state-by-state gasoline consumption data gathered.

Measure Proof of Concept Testing 33 Gasoline Consumption Performance Results Trend, 2005–2011 Annual Gasoline Consumpon/Capita 2011 (gallons) Figure 7. Annual gasoline consumption per capita (2011). (continued on next page)

34 Environmental Performance Measures for State Departments of Transportation Trend, 2005–2011 Annual Gasoline Consumpon/Capita 2011 (gallons) Figure 7. (Continued).

Measure Proof of Concept Testing 35 Gasoline Consumption Measure: Criteria Assessment Does Measure Support State-to-State Comparisons? • State DOTs Have Limited Control over Gasoline Consumption. State DOTs do not directly control gasoline consumption and have only limited ability to influence this metric. State- wide gasoline use per capita is influenced by a variety of technological and behavioral factors. For example, the introduction of more fuel-efficient vehicles into the fleet will tend to reduce gas consumption over time, as will greater use of alternative fuels. The other major influence on this metric is VMT per capita, which reflects a complex phenomenon affected by demo- graphics, economic conditions, and land development patterns, among other influences. By supporting a multi-modal transportation system, however, state DOTs can affect the energy consumption impacts of current policies, programs, and practices. State DOTs may also par- ticipate in efforts to provide alternative transportation fuels, such as electric vehicle charging infrastructure. The measure is responsive to actions by a broader spectrum of public agencies including state, regional, and local transportation agencies, environmental and energy agen- cies, and local planning departments. • Economic Growth Is Not a Predictor of Growth in Gasoline Consumption. Annual changes in the measure are partially influenced by economic factors, but the research team’s analysis suggests the measure does not closely track a state’s economic growth. The scatter plot in Figure 8 shows all 50 states plotted according to the percentage change in median income versus the percentage change in gasoline use per capita, from 2005 through 2011. North Dakota is an outlier, with high growth in both variables. But for the remainder of states, there is not a strong relationship between income and gasoline use. • Level of Urbanization Is Related to Gasoline Consumption. Differences between states in gasoline use per capita are partly explained by their degree of urbanization. Residents in dense, metropolitan areas tend to drive less, so states with a relatively large share of population in such areas often have lower VMT per capita. For example, Hawaii is the third lowest state in terms of gasoline use per capita. The state is the sixth most urban (92 percent urban), and with most of the population concentrated in Oahu, the opportu- nities for long distance travel on highways are limited. Conversely, the least urban states, -20% -15% -10% -5% 0% 5% 10% 15% -20% -15% -10% -5% 0% 5% 10% 15% 20% Ch an ge in g as ol in e us e pe r c ap ita , 2 00 5- 20 11 (V er c al A xi s) Change in median income, 2005-2011 (Horizontal Axis) Figure 8. Change in gasoline use per capita versus change in median income (2005 through 2011).

36 Environmental Performance Measures for State Departments of Transportation such as Maine (39 percent urban) and Vermont (39 percent urban), also rank among the 10 highest states in terms of gasoline consumption per capita (41 and 42). There are some exceptions, however. Alaska ranks 6th lowest in gasoline use per capita but ranks 38th in level of urbanization. In Figure 9, a scatterplot shows urbanization versus gasoline consumption. Do DOTs Have Easy Access to Data/Is Measure Easily Calculated? • Gasoline Sales Data Are Easily Obtained for All States. This measure relies on FHWA- provided data on gasoline sales by state, which are listed in the publication Highway Statis- tics. Annual on-road gasoline sales for every state are reported in Tables MF-21 and MF-27 of Highway Statistics, making data collection simple, fast, and reliable. In these tables, sales of pure gasoline are combined with sales of gasohol (low-level blends of ethanol with gasoline). Is Data Quality Credible and Defensible? • FHWA Gasoline Sales Data Is Robust. FHWA obtains data on statewide purchases of gaso- line from state motor-fuel tax agencies, which collect taxes at the point of sale. FHWA makes adjustments so that the data is uniform and complete for all states. FHWA’s data for on-road motor fuels are used to allocate highway funding and are generally considered to be more accurate than other sources. • Gasoline Sales Are a Generally Acceptable Proxy for Gasoline Use. Use of FHWA data for this measure assumes that gasoline sales are an acceptable proxy for gasoline use. In many states, the two metrics are nearly identical. In some situations, however, there is potential for discrepancy between the place of fuel sales and the location of travel activ- ity and consumption of fuel. This can be an issue particularly for states that have a lot of traffic across state lines or where fuel tax rates differ widely across state boundaries. For instance, given its size and the large amount of through traffic it experiences, Maryland DOT has found that fuel sales do not provide as accurate a basis for estimat- ing GHG emissions as VMT-based methods. Similarly, New York State DOT determined that fuels sales underestimated VMT in the state because many New York travelers 30% 50% 70% 90% 250 300 350 400 450 500 550 600 Pe rc en t U rb an iz ed Gasoline Used per Capita Figure 9. Percentage urbanized versus gasoline use per capita, 2011.

Measure Proof of Concept Testing 37 purchase fuel in New Jersey. This is an issue in so few states, it should not preclude its use as a measure. 4.4 Materials Recycling Measure: Proof of Concept Results Measure: Annual Percentage by Mass of All Roadway Asphalt Pavement Materials Composed of Reclaimed Asphalt Pavement Proof of Concept Validation Implementation Readiness Does Measure Support Consistent State-to-State Application?  Ready for Use by Many DOTs Today Can DOTs Report Measure Easily with Existing Data or Data That Is Easy to Generate?  Suitable for Use by DOTs in Longer Term Is Data Quality Credible and Defensible?  Not Suitable for Use by Most DOTs Key Measure is fully consistent with criteria Measure is mostly consistent with criteria Measure is somewhat consistent with criteria Measure lacks consistency with criteria Overview Pilot States: Colorado, Delaware, Florida, Illinois, New Jersey, Missouri, North Carolina, North Dakota, Pennsylvania, South Dakota, Utah Reclaimed asphalt pavement is the term given to reprocessed pavement materials con- taining asphalt and aggregates. These materials are generated when asphalt pavements are removed for reconstruction or resurfacing. Although old asphalt pavements can be recycled at central processing plants, in-place recycling processes are increasingly common, which include partial removal of the pavement surface, mixing of reclaimed material with beneficial additives (such as virgin aggregate, binder, and rejuvenating agents to improve binder proper- ties), and placing and compacting the resultant mix in a single pass. This measure examines the annual percent by mass of all roadway asphalt pavement materials composed of RAP used by a DOT. Performance results are shown in Figure 10. See Appendix C for complete state- by-state RAP data gathered. Materials and Recycling Measure: Criteria Assessment Does Measure Support State-to-State Comparisons? • RAP Is Widely Used by State DOTs. All state DOTs allow use of RAP to some extent, and while precise specifications can vary, RAP is used in essentially the same way across the coun- try. A 2010 survey by the American Association of State Highway and Transportation Officials (AASHTO) Subcommittee on Materials collected information on RAP use by state DOTs; it found that of the 36 states that responded to the survey, all reported allowing RAP in some circumstances. Such a high level of acceptance among practitioners makes national use of this measure feasible, and success in its implementation more likely. • Little Variation in Ability to Use RAP among States. All states are able to use RAP. While climate conditions may have some influence on appropriate pavement mixes, the use of RAP is not greatly hindered by climate. States that currently allow high RAP

38 Environmental Performance Measures for State Departments of Transportation percentages include both cold weather states (e.g., Maine) and warm weather states (e.g., Arizona). • DOTs Influence RAP Use. State DOTs have control over the amount of RAP used in pave- ments because they establish specifications for pavement composition. Note, however, that most state specifications establish a maximum RAP percentage for pavements. Interviews with pavement engineers by the research team indicate that RAP availability is not limited—in many cases, DOTs have large surpluses of RAP. These factors point to increased ability for DOTs to influence performance of the measure. • Limits on RAP Content May Affect Performance, but States Are Increasing RAP Limits. The percentage of RAP allowed in new pavement varies among states—typically ranging from 15 percent to 40 percent. In some states, RAP use is approaching the maximum allowed by the state’s specifications. For example, RAP made up 92 percent of the maximum allowable amount in Colorado DOT’s 2012 projects. If state DOTs do not increase RAP limits, usage may flatten in many states. Use of RAP in projects, however, has been increasing steadily in the past decade as materials and best practices have matured. Many state DOTs have increased the maximum amounts of RAP allowed in their specifications and this trend appears to be continu- ing. Furthermore, many asphalt producers do not always use the maximum allowable amounts. For example, when there are different limits for different courses, contractors will sometimes create one mix meeting the lowest minimum amount for multiple pave- ment layers rather than produce two mixes. NJ, 14.3% FL, 19.4% IL, 15.5% CO, 15.7% DE, 21.7% MO, 16.7% SD, 7.9% 0% 10% 20% 2008 2009 2010 2011 2012 UT, 18% NC, 22% ND, 15% PA, 16% 0% 10% 20% 2011 2012 Note: PA is represented by a single data point, not a line. Figure 10. RAP as a percentage of asphalt laid annually (2008 through 2012).

Measure Proof of Concept Testing 39 Do DOTs Have Easy Access to Data/Is Measure Easily Calculated? • Records of Necessary Data Exist, But May Require Additional Effort to Compile. Contrac- tors must track materials usage as part of their contract requirements and DOTs’ specifica- tions monitoring, therefore it is possible to track down the total quantity of RAP versus other pavement inputs in most states. What is not being done universally, however, is compilation of project-level information to allow ongoing monitoring at the statewide level. Implement- ing a reporting system by contractors or charging division engineers to gather this decentral- ized information will thus be necessary in some states. Virtually all state DOTs keep track of the total amount of asphalt laid each year. • Tracking Efforts Appear to Be Increasing at DOTs. Several DOTs contacted by the research team indicated that they just started tracking RAP use, suggesting that the collection of infor- mation needed for this measure is increasing. Is Data Quality Credible and Defensible? • RAP Data Collection Practices Vary from State to State. While use of RAP is widespread, record- keeping practices vary widely from state to state. Many DOTs maintain data on RAP use on indi- vidual projects and thus are able to accurately calculate this measure. However, just as many do not appear to maintain accurate records on use of RAP. For example, the DOTs in Ohio and New York set allowable RAP percentages in their asphalt pavement specifications, but allow the con- tractors to develop their own mix designs; the DOT does not maintain records of the mix design employed by the contractors. Among the 29 states contacted, 11 were able to provide RAP use data. However, there is some contradictory information on the tracking of RAP use; an FHWA survey found that only 3 of the 18 respondents indicated that they track the amount of RAP used. 4.5 Stormwater Measure: Proof of Concept Results Measure: Percentage of State DOT-Owned Impervious Surface for which Water Treatment is Provided Proof of Concept Validation Implementation Readiness Does Measure Support Consistent State-to-State Application?  Ready for Use by Many DOTs Today Can DOTs Report Measure Easily with Existing Data or Data That Is Easy to Generate?  Suitable for Use by DOTs in Longer Term Is Data Quality Credible and Defensible?  Not Suitable for Use by Most DOTs Key Measure is fully consistent with criteria Measure is mostly consistent with criteria Measure is somewhat consistent with criteria Measure lacks consistency with criteria Overview Pilot States: Delaware, Maryland, North Carolina, Ohio This measure quantifies the share of impervious surface area owned by a DOT that is treated. “Treatment,” for the purposes of this measure, refers solely to active treatment using a struc- tural BMP. Runoff from roadways can also be sufficiently treated by natural landscapes and non-engineered structures, but more research must be done to account for these elements. In

40 Environmental Performance Measures for State Departments of Transportation addition, this measure treats all impervious area the same in terms of its need for treatment, when in reality some areas are in less need of treatment than others. Three major data inputs are needed to complete the measure: (1) an inventory of all structural BMPs, (2) drainage areas for each BMP, and (3) the impervious surface area of the road system of interest. Few states currently have all data elements readily available. Research undertaken as part of the pilot testing finds that state DOTs increasingly are under- taking BMP inventories to meet NPDES permitting requirements, so the measure will become feasible at more DOTs over time. The focus of proof of concept testing for this measure centers on four states where some or all of the three major data elements are available: Delaware, Mary- land, North Carolina, and Ohio. Data from each of these states reflects widely differing data col- lection practices and therefore, data for the stormwater measure should not be considered suitable for comparisons from state to state. Maryland (Statewide Results Available) Maryland’s State Highway Administration (SHA) currently tracks a similar performance mea- sure: number of acres of untreated pavement retrofitted for stormwater management controls each fiscal year (see Table 10). The information used for their performance measure is easily translated into the suggested measure through simple GIS operations and because SHA maintains data for the whole state system, the performance measure can be calculated for the whole state. Delaware DOT (Newcastle County Results Available) DelDOT’s impervious surface data includes impervious areas both owned and not owned by the state DOT. To estimate the fraction that is DOT-owned, and therefore the amount DelDOT is responsible for treating, GIS was used to analyze a 15-meter buffer around the DOT’s linear road network (see Table 11). Only impervious area within this buffer was counted in the denominator. Ohio DOT (Cuyahoga County Results Available) Drainage areas associated with structural BMPs were not available from the Ohio DOT, so esti- mated drainage areas were calculated using GIS tools on a digital elevation model for Cuyahoga County (see Table 12). In the future these calculations could be verified by an engineer or vali- dated with fieldwork. Total impervious area was calculated by summing each road segment’s width and shoulder width, which are recorded in the attribute table of the state DOT’s GIS file of the road system. North Carolina DOT (Statewide Primary Roadway Results Available) The North Carolina DOT provided a spreadsheet of BMPs located on the state’s Primary Road System. Thirty newly installed BMPs had estimates of the drainage areas, as well as the amount of impervious surface in the drainage areas (see Table 13). The average size of these drainage areas Maryland (Statewide) Results State highway system impervious area 32,259 acres Treated highway area (by SHA BMPs) 9,025 acres Percent of DOT-owned impervious surface for which treatment is provided* 28.0% *Does not account for surface that is passively treated or that does not require treatment. Table 10. Maryland stormwater results.

Measure Proof of Concept Testing 41 Delaware (Newcastle County Only) (Excludes DelDOT maintenance facilies and rest areas) State system impervious area 4,077 acres Treated area: All (includes locally owned) DelDOT owned only 899 acres 399 acres Percent of impervious surface for which treatment is provided:* All (includes locally owned) DelDOT owned only 22.0% 9.8% *Does not account for surface that is passively treated or that does not require treatment. Note: Much of the impervious area treated by DelDOT’s BMPs is locally owned or private developments. How much this treatment area should count is a queson that will need to be resolved before widespread implementaon. In the meanme, the results are shown with and without this area not owned by the state DOT. Table 11. Delaware stormwater results. and the average share of imperviousness were applied to all other BMPs for which this informa- tion was not available. North Carolina DOT staff indicated there is nothing notably different about these BMPs compared with other BMPs around the state. The results are therefore a pre- liminary estimate, and a more reliable measure would require more data gathering and analysis. Stormwater Measure: Criteria Assessment Does Measure Support State-to-State Comparisons? • The Amount and Type of Stormwater Treatment Needed Varies by State. Not all roads need the same amount of stormwater treatment. Largely rural and arid climate states in the North Carolina (Statewide Primary Roads) Primary roadways impervious area 94,730 acres Treated area 4,644 acres Percent of DOT-owned impervious surface for which treatment is provided 4.9% *Does not account for surface that is passively treated or that does not require treatment. (Drainage areas and impervious area calcula ons are based on a sample of 30 BMPs.) Table 13. North Carolina stormwater results. Ohio (Cuyahoga County Only) (Excludes ODOT maintenance facilies and rest areas) State system impervious area 4,184 acres Treated area 49.9 acres Percent of DOT-owned impervious surface for which treatment is provided 1.2% *Does not account for surface that is passively treated or that does not require treatment. Table 12. Ohio stormwater results.

42 Environmental Performance Measures for State Departments of Transportation southwest, for example, likely have many miles of roadway that require minimal treatment, while population centers in the wetter Pacific Northwest need more stringent treatment. Holding these two areas to the same standards may not give a true sense for how well a state DOT is meeting the need for treatment. Establishing limiting criteria across the board, such as by population or rainfall amounts, can address this issue. • States with More Stringent NPDES Requirements Are Best Equipped to Calculate This Mea- sure. A state DOT’s degree of focus on runoff treatment is often influenced by the requirements of the agency’s NPDES permit. Maryland’s NPDES regulating agency was one of the first to require a full inventory of stormwater treatment BMPs, because of concerns about the Chesapeake Bay, and the Bay states have long had one of the most rigorous treatment programs. State DOTs often undertake an inventory of BMPs only after it becomes a requirement in their permit. Those states with newer BMP inventory requirements—such as Ohio and North Carolina—are less prepared for this measure than those with longer-standing requirements, such as Maryland and Delaware. Do State DOTs Have Easy Access to Data/Is Measure Easily Calculated? • Most States Today Have Limited Access to BMP Data. Conversations with selected state DOTs, suggest that—with the exception of Maryland and Delaware—most are unlikely to have enough data to fully calculate the stormwater measure, as proposed. State DOTs’ storm- water data capabilities are rapidly evolving as stormwater policy requirements and technology in the area of GIS mapping advance swiftly. • Phased Approach for Measure Implementation Is Viable. Most state DOTs today lack suf- ficient data in the area of stormwater to calculate the measure, as proposed. Trends in storm- water policy and data, however, suggest a favorable outlook for phasing in such a measure over time by gradually adopting the following interim measures as more data becomes available: – Phase 1: BMP Inventory-Based Measure. A BMP inventory documents how many high- way-related BMP structures a state DOT has and where they are located. BMP inventory data is needed as the first step for calculating the stormwater measure, but regularly updated BMP inventory data alone could be used for an interim measure such as the number of acres of untreated pavement retrofitted for stormwater treatment annually. State DOTs increasingly are preparing BMP inventories as part of their compliance with statewide NPDES permits. Nebraska, Ohio, and Nevada who were contacted as part of this study, for example, are all in the initial stages of preparing a BMP inventory for their highway systems and the team expects that BMP inventories will become more common over time. – Phase 2: Impervious Surface-Based Measure. With a BMP inventory in place, BMP data can be conflated with data describing the amount of state DOT-owned impervious area, which is the second step in calculating the proposed stormwater measure therefore allowing use of a measure like number of BMPs per acre of impervious area before a state DOT begins calculating the drainage area of each BMP, which is the most analytically complex element of the final stormwater measure. – Phase 3: Complete Implementation. Assignment of drainage areas to each BMP represents the final major step in the data gathering process. Ideally drainage areas would be recorded in a GIS environment to allow for simple spatial analysis—finding the intersection of drainage areas and impervious surface is a standard process in all GIS software. At this point, a state DOT is able to calculate the full performance measure. Even under a phased approach, the stormwater measure would require most states to dedi- cate considerable staff hours to the labor-intensive process of inventorying their treatment network and delineating individual treatment areas. • GIS Tools Are Key to Stormwater Data Accessibility. GIS-derived latitude and longitude data that can be linked to a BMP inventory enables easy delineation of drainage areas based on topography. It also better enables users to analyze the results at varying geographic

Measure Proof of Concept Testing 43 levels (county, district, statewide) and with varying road characteristics (road class, traffic volume, population, urban/rural, proximity to impaired water bodies, etc.). Is Data Quality Credible and Defensible? • Stormwater Measure Implementation Will Require Resolution of Analytic Issues. Full implementation of the stormwater measure will require answers to a variety of nuanced ques- tions on the details and parameters of the measure: – Do All Roads Need to Be Part of the Measure? For example, do arid areas or low-volume roads need to be treated to the same extent as areas with lots of rain and traffic? It may not be efficient to encourage the same treatment in these places. – How Should “Treatment” Be Defined? Large sections of state DOT right of way are sufficiently pervious and vegetated to provide natural water quality and flow treatment to runoff. These areas could be accounted for in the measure, either by counting as treatment or by reducing the total area in need of treatment (the denominator). Similarly, treatment of less engineered “green infrastructure,” such as bio-swales should be resolved as part of the measure. – Under What Conditions Does an Adequate BMP Need to Perform? Some structures are built to accommodate a 2-year storm, but should they be able to withstand a more severe 10-year storm? – Should Non-Structural BMP Activity Be Incorporated, and If So, How? Because the measure looks only at BMPs that are discrete and easily quantified, that is, engineered, structural BMPs, a portion of water quality improvements, which includes all non-structural BMP activities, are not captured. DOTs undertake an array of activities to improve water quality, such as regu- lar street sweeping, vegetation planting, and deicing improvements. These elements are not captured in the current form of the measure. (It is possible to develop a series of “impervious area equivalents” for non-structural BMP activities, as Maryland has started exploring for its program.) – Is It Possible to Include the Quality of Treatment in This Measure? This measure currently only takes into account how much pavement is in an area served by a structural BMP, based on original designs and functionality. It does not take into account whether each BMP is being adequately maintained, or how well the structure is working. These elements are also important to water quality assurance, and should be captured. 4.6 Wildlife and Ecosystems Measure: Proof of Concept Results Measure: Ecosystems Self-Assessment Tool Proof of Concept Validation Implementation Readiness Does Measure Support Consistent State-to-State Application?  Ready for Use by Many DOTs Today Can State DOTs Report Measure Easily with Existing Data or Data That Is Easy to Generate?  Suitable for Use by DOTs in Longer Term Is Data Quality Credible and Defensible?  Not Suitable for Use by Most DOTs Key Measure is fully consistent with criteria Measure is mostly consistent with criteria Measure is somewhat consistent with criteria Measure lacks consistency with criteria

44 Environmental Performance Measures for State Departments of Transportation 0% 20% 40% 60% 80% 100% Wetlands Aqua c, Streams Wildlife movement Invasive Species T&E/ Sensi ve General All Categories Pe rc en t S co re d Ecological Focus Area IL TX GA MD OH OR O utstanding Room for Im provem ent Good Perform ance Figure 11. ESAT proof of concept score comparison for six pilot states. Overview Pilot States: Georgia, Illinois, Maryland, Ohio, Oregon, Texas The wide variations in biological diversity among states make a single metric for wildlife and ecosystems extremely challenging. The ESAT is a tool state DOTs can use to assess their perfor- mance using a standardized questionnaire that attempts to capture the complexity of this topic area in a meaningful way. It is a relatively simple, yet comprehensive, self-administered assess- ment that evaluates performance in terms of avoiding, minimizing, or mitigating impacts in six ecological focus areas. The ESAT has been tested in six pilot states; typically either the director of environmental services or senior biology staff conducted the pilot test in each state. DOT representatives from each pilot state that agreed to participate in proof of concept testing were provided the ESAT and an instructional memo that also included an overview of the project and the wildlife and ecosystems measure. The mean overall score for the six participating states was 47 percent of the maximum points available. Given the small sample size, only basic statistical parameters were evaluated; however, the data suggest a relatively normal distribution of scores around the mean, with a minimum score of 22 percent and a maximum score of 69 percent. Scores were similarly distributed within each ecological focus area, as shown in Figure 11. The range of scores for each individual ques- tion were also well distributed, suggesting that the scoring system was capturing the range of DOT performance. Only one question had an identical response from all states. See Appendix B (Table B-3) for a complete summary of scores from proof of concept testing. Wildlife and Ecosystems Measure: Criteria Assessment Does Measure Support State-to-State Comparisons? • Subjective Interpretation and Scoring for the ESAT Makes Comparisons between States Challenging. Given the wide range of practices and experiences of DOTs across the country, or even between different divisions or individuals within a given program, answers may vary

Measure Proof of Concept Testing 45 because of divergent interpretations of the questions rather than actual differences in practice. Qualitative questions are the most difficult to compare because they draw upon the DOT respondent’s own interpretation and experiences to assess performance. Additionally, the ESAT does not provide explicit instructions for calculating responses that are based on quan- titative data because of the absence or impracticality of using a single, nationwide standard for tracking this information. Therefore, calculation methodologies and associated responses may vary somewhat between programs. Do State DOTs Have Easy Access to Data/Is Measure Easily Calculated? • The ESAT Evaluates Many Practices Not Regularly Tracked by Most DOTs. In cases where data is not readily available and would be burdensome to collect for the purposes of com- pleting the ESAT, DOTs may select an answer based on the best available information on hand. For example, certain DOTs reported that they did not track the proportion of impacted streams or wetlands that have been mitigated over the last 5 years or were unable to calcu- late the average share of the DOT’s annual research budget allocated to ecosystem-focused research over the last 5 years. Is Data Quality Credible and Defensible? • Many Questions Are Subjective and Qualitative in Nature and Are Therefore Susceptible to Divergent Interpretations. Scores for 18 of the 41 questions require DOTs to perform cal- culations or estimate their performance quantitatively; the remaining score ranges are quali- tative. Even when quantitative scoring is applied, DOTs may vary in their interpretation. In addition, for most questions, the DOTs themselves are the authoritative source for informa- tion on their program’s practices. Responses could be externally verified if DOTs choose to provide an online source for each response; however, providing this information is optional and no verification system is currently in place.