Strategic Issues Facing Transportation, Volume 3: Expediting Future Technologies for Enhancing Transportation System Performance (2013)

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.

45 Over 70 percent of the nation’s roads are located in snowy regions, which receive more than five inches (or 13 cm) average snowfall annually. Nearly 70 percent of the U.S. population lives in these snowy regions. Snow and ice reduce pavement friction and vehicle maneuverability, causing slower speeds, reduced roadway capacity, and increased crash risk. Average arterial speeds decline by 30 to 40 percent on snowy or slushy pavement. Freeway speeds are reduced by 3 to 13 percent in light snow and by 5 to 40 percent in heavy snow. Heavy snow and sleet can also reduce visibility. Lanes and roads are obstructed by snow accumulation, which reduces capacity and increases travel time delay. Each year, 24 percent of weather-related vehicle crashes occur on snowy, slushy or icy pavement and 15 percent happen during snowfall or sleet. Over 1,300 people are killed and more than 116,800 people are injured in vehicle crashes on snowy, slushy or icy pavement annually. Every year, nearly 900 people are killed and nearly 76,000 people are injured in vehicle crashes during snowfall or sleet. Snow and ice increase road mainte- nance costs. Winter road maintenance accounts for roughly 20 percent of state DOT maintenance budgets. State and local agencies spend more than 2.3 billion dollars on snow and ice control operations annually. Each year, these road agencies also spend millions of dollars to repair infrastructure damage caused by snow and ice53. Frame Winter operations include plowing, sanding, and salting— the traditional methods for snow removal and ice control. Although agencies are gradually shifting from reactive meth- ods to more proactive strategies, the three classical methods remain a mainstay. Although the traditional methods are relatively cheap and easy to use, they tend to be less efficient and less effective than emerging technologies while having larger adverse effects on the environment, infrastructure, and vehicles. Therefore, states could adopt new technologies to better meet their goals while avoiding the disadvantages of traditional methods. Major inputs to the STREAM analysis of snow removal and ice control are as follows: • Functions: snow removal and ice control • Goals: preservation, safety, mobility, sustainability • Objectives (metrics) – Preservation: less corrosion—less corrosive effect of winter maintenance chemicals on pavements and steel structures – Mobility: less congestion and closure—shorter travel time and faster speed – Safety: better de-icing and anti-icing—fewer crashes due to higher road friction and better maneuverability – Sustainability: less environmental damage—reduced detrimental effects on air and water quality, soil, and vegetation Winter Management Strategies and the Status Quo DOTs must meet their preservation, mobility, safety, and sus- tainability goals while contending with winter storms. Further- more, all state DOTs are facing increasing demand for higher level of service (LOS54), especially during inclement weather, more environmental considerations, and ever tighter winter operation budgets (Rochelle, 2010). Traditional methods such as salting, sanding, and plowing are being challenged because • Salts based on chlorides have corrosive (low preserva- tion) and environmentally detrimental (low sustainability) characteristics; A p p e n d i x B STREAM Applied to Snow Removal and Ice Control Technology 53 Road Weather Management Program, Federal Highway Administration (http:// www.ops.fhwa.dot.gov/weather/weather_events/snow_ice.htm) 54 LOS, in the context of roadway snow and ice control operations, is a set of operational guidelines and procedures that establish the timing, type, and fre- quency of treatments. The maintenance actions are directed toward achieving specific pavement condition goals for various highway sections (Blackburn, Amsler Sr and Bauer, 2004).

46 • Sand has poor performance characteristics and produces particulate matter which leads to low LOS (low mobility and safety) and causes health problems; and • Plowing slows traffic (low mobility) and creates problems with dispatching and routing decisions. DOTs may choose to adopt new technologies that are more environmentally friendly and less corrosive while maintain- ing high LOS and optimizing the use of pre-existing materials and equipment. For example, thermal methods (e.g., bridge deck heating systems) are emerging because they have no nega- tive effects on environment and transportation infrastructures while improving traffic flow and reducing car crashes during inclement weather. In addition, when acetate-based chemi- cals are used as an anti-icing agent, they have less detrimental effects on environment and infrastructures than chloride-based chemicals. As such, by introducing emerging technologies, state DOTs can attain their goals in an efficient and effective manner. Most winter management strategies used by most state DOTs are mechanical removal with or without friction enhance- ments and de-icing55 and anti-icing56 with chemicals or other methods. While these strategies can be used individually, they are more often used in combination with one another (Blackburn et al., 2004). Among those strategies, anti-icing is relatively new and emerging in snow-belt states due to its effi- ciency and effectiveness with the advancement of information technology and more reliable weather forecasting. In 1996 maintenance managers with the Idaho DOT began an anti-icing program on a 29-mile (47-kilometer) section of US Route 12. This highway segment is located in a deep canyon and is highly prone to snowfall and pavement frost (i.e., black ice) due to sharp curves and shaded areas. An anti-icing chemical is applied to road surfaces as an alternative to spreading high quantities of abra- sives. Abrasives are thrown to the roadside by passing vehicles and only improve roadway traction temporarily. . . . Mobility, produc- tivity, and safety enhancements resulted from the anti-icing treat- ment strategy. Mobility was improved, because a single application of magnesium chloride was typically effective at improving trac- tion for 3 to 7 days—depending on precipitation, pavement tem- perature, and humidity. Faster clearing of snow and ice reduced operation costs and enhanced productivity. Safety improvements were realized by reducing the frequency of wintertime crashes (Idaho DOT). Most sources agree that, for anti-icing, the new technolo- gies bring high LOS and cost saving while meeting DOT mis- sions. Despite this, most transportation agencies continue to use traditional methods to maintain highways and bridges in winter. Some states are shifting their approaches for remov- ing snow and controlling ice from reactive (i.e., plowing with chemical follow-up and abrasives after a storm) to more pro- active strategies (i.e., treatments to prevent or weaken the bond between the pavement and snow prior to a storm), but these states still depend heavily on traditional methods. In particular, when choosing winter maintenance chemi- cals, most states still execute de-icing practices with solid chemicals—primarily, sodium chloride (NaCl) and magne- sium chloride (MgCl2)—rather than anti-icing with liquid chemical products (Rochelle, 2010). According to Rochelle, one common reason is that it is believed that anti-icing chemicals will be washed away during wet snow events or during storms in which rain turns into snow. Another is because forecasting algorithms have not been extensively implemented into their winter maintenance practices. This suggests the possibility for improvement by shifts to newer anti-icing methods with devel- oping advanced information technology (IT), more accurate weather forecasting, and more refined decision systems. Some chemicals can be used for de-icing as a solid and for anti-icing practices as a liquid. Common table salt (NaCl) can be used for de-icing as rock salt and for anti-icing as salt brine and this is true for others as well. They may have both less impact on the environment and infrastructures but more positive influence on mobility and safety when they are used for anti-icing as opposed to de-icing purposes. As shown in Idaho DOT’s practice, liquid MgCl2 was used for anti-icing and attained the DOT’s goals better than granular MgCl2 as a deicer. Factors Driving a Shift in Approach When selecting winter management strategies, many states have begun to consider the following factors: LOS, cost, infra- structure and environmental impacts, equipment, and weather. This has led several to conclude that there are limitations in using traditional methods. Some states have enacted legisla- tion for reducing the use of salt and the use of alternative, agricultural-based products57 (ABP) in addition to acetates. Many studies cite the disadvantages of traditional methods, in particular salt usage: One concern regarding reactive maintenance practices is the increased potential for accidents and injuries due to poor road conditions while maintenance crews are being deployed. Another problem with reactive practices is the quantity of materials and labor hours needed to maintain the desired LOS for winter roadways (O’Keefe and Shi, 2006). 55 De-icing breaks the bond between snow/ice and the pavement by chemical and mechanical means after a storm. 56 Anti-icing requires timely application of winter maintenance chemicals before the onset of a storm to weaken or prevent the bond between compacted snow and the pavement surface from forming so as to improve removal efforts. 57 Byproducts from the agricultural industry are often used as additives to inorganic (e.g., chloride-based) winter maintenance chemicals. Some ag-based products are produced by the fermentation and processing of cane or beet sugar syrup, corn barley, or other carbohydrates and milk (Fay and Shi, 2012).

47 The widespread use of rock salt (sodium chloride) to remove snow and ice and facilitate a ‘bare pavement’ LOS has provided for the increased safety of motorists for some time. However, de-icing salt use has some detrimental side effects. The dam- age to the ecosystem from chloride ions has been documented, along with the corrosive effects to metals. Consequently, fre- quent repair and rehabilitation of bridges has resulted (Kahl, 2002). Traditional methods often lead to poor efficiency and low effectiveness of plowing, i.e., LOS: In contrast to anti-icing operations, traditional snow and ice control practice is to wait until the snow accumulates on the pavement before beginning to plow and treat the highway with chemicals or abrasives. A consequence of traditional practice is formation of a compacted snow layer tightly bonded to the pave- ment surface. A subsequent de-icing of the pavement is then nec- essary and usually requires a large quantity of chemical to work its way through the snow pack to reach the pavement and destroy or weaken the bond. Although requiring less information and training than for anti-icing, de-icing may provide less safety as a result of the inherent delay (Kahl, 2002). Another concern is corrosion-related issues when using salt as a deicer: A survey of 200 concrete highway bridges carried out by Maunsell & Partners found that many of the bridges had reinforcement corrosion because of the high chloride content in the concrete caused by the use of salting for winter maintenance. The study confirmed that leakage through bridge joints occurred frequently. Consequently, areas of the abutments, piers, and deck soffits became stained and contaminated with chloride. Other areas were affected by spray from passing vehicles (Burtwell, 2004). Chlorides in de-icing salts can significantly increase concrete scaling, possibly due to increased osmotic pressure in addition to expansion of freezing water and/or when dissolved salts recrys- tallize in the concrete pores (Kahl, 2002). Finally, environmental problems resulting from salt usage are significant: Already at an early stage, it was recognized that the use of salt had not only the desired effect of improved traffic safety and accessibility but also several negative impacts. Numerous inves- tigations of impacts on vegetation, soil, and groundwater have been presented, and the matter is still of great concern in North America, Europe, and Japan (Gustafsson and Blomqvist, 2004). The salts, NaCl, CaCl2, and MgCl2, leave residues of chloride ions that can be swept up in storm water runoff or snowmelt and carried into adjacent drainage ditches to be discharged into downstream surface waters. It is these de-icing compounds that are the focus of the most intense environmental scrutiny. Chlo- ride concentrations from roadway de-icing can be substantial. Although natural background concentrations in water may be only a few parts per million, roadway runoff during de-icing operations has been measured as high as 18,000 mg/l. Resulting chloride concentrations in the environment also can be signifi- cant. Values measured in lakes can vary from 15 to 300 mg/l in rural settings to 2,000 to 5,000 mg/l in urban impoundments. Streams have been documented to carry concentrations as high as 4,300 mg/l. These values become important because of the relatively low thresholds at which chlorides can do harm to fresh- water aquatic species (Davis, 2004). Identify Snow Removal and Ice Control Technologies Three categories of technologies are available (Rochelle, 2010): • Traditional methods: mechanical plowing, de-icing (salt- ing), and sanding; • Proactive approach: anti-icing; and • Emerging technologies: heated bridge deck, pre-wetting, fixed automated spray technology (FAST), surface overlay, information technology, improved removal equipment. This three-part breakdown presents a neater division than exists in practice. Some chemicals are used as both a deicer and an anti-icer, and technological innovations are being made that cross categories so the latter are not mutually exclusive. Thus, it is desirable to look at all three categories in an integrated manner. Incremental innovations as well as revolutionary innovations need to be considered. Table B-1, which shows various technologies employed in winter operations, provides a list of technologies ranging from maintenance management systems and plows to IT and thermal methods. Table B-1 also shows frequency of use based on a sur- vey conducted by the two studies cited in the table. The research team divided these technologies into two categories: Category I contains primary technologies which have direct impact on snow removal and ice control; Category II contains secondary or peripheral technologies which have indirect impact. For the purpose of this case study, the research team put more emphasis on Category I technologies than Category II. Table B-2 shows a wide range of available chemicals for de-icing or anti-icing. The first two categories—traditional methods and pro- active approach—represent the current state of practice used by most states. Examples are as follows: • Mechanical plowing (Ketcham et al., 1996)—Most states use the same basic type of equipment including, dump trucks with plows, rotary plows, and loaders. Some common types of plows are one-way front plows, reversible plows, deform- able mold board plows, underbody plows, and side-wing plows. Some plows can be shifted from side to side using hydraulics, allowing the plow to extend to the side by 9 to 12 feet. • Chemicals used for de-icing and anti-icing (Rochelle, 2010)—NaCl is the most common because it is abundant

48 Category I Ranking by Frequency Specific Technology Plow Configuration 1 (53/54) Front plows 2 (29/54) Underbody plows 3 (18/54) Wing plows 4 (9/54) Rear plows Plow Blades (Types) 1 (47/51) Carbide 2 (24/51) Underbody blade 3 (19/51) Wear plates 4 (9/51) Double/triple edge 4 (9/51) 14+ ft 6 (4/51) Rubber 7 (2/51) Triple-blade 7 (2/51) Tow blade 7 (2/51) Carbide with steel backer 9 (1/51) Steel De-icing & Anti-icing 1 (50/54) Anti-icing with liquids 2 (47/54) De-icing with solids 3 (38/54) De-icing with liquids 4 (22/54) Anti-icing with solids 5 Other—Pre-wet system Application Methods 1 (50/52) Spinner applications with solids 2 (32/52) Stream method with liquids 3 (28/52) Spray method with liquids 4 (12/52) Gravity Feed 5 (9/52) Zero Velocity 6 (4/52) Advanced Placement Thermal Methods (Pavement Heating Methods) N/A Electrically conductive concrete Electrical resistive heating Geothermal heat pumps Infrared heating Microwave and radio frequency power Solar and wind power Category II Ranking by Frequency Specific Technology Maintenance Management System (MMS) 1 (12/29) GPS / Automatic Vehicle Location (AVL) 2 (7/29) TAPER logs 3 (2/29) Work Management System (WMS) 4 (1/29) Resource Management System (RMS) 5 Other—Timesheets, manual and vehicle reports, crew information cards Information Technology 1 (43/48) Road Weather Information System (RWIS) 2 (26/48) GPS 2 (26/48) AVL 4 (24/48) Maintenance Decision Support System (MDSS) 5 Other—Free web-based information provided by the National Oceanic and Atmospheric Administration (NOAA), Full Mobile Data Computing Windshield Wipers 1 (47/51) Standard equipment 2 (8/51) Hot Shot 3 (6/51) SlapMe 4 (4/51) Clear Fast Add on Vehicle Accessories and Training 1 (26/40) Specialized lighting packages 2 (11/40) Back-up cameras 2 (11/40) Vehicle airfoils 4 (9/40) Driver simulator training 5 (8/40) Vehicle deflectors 6 (7/40) Vehicle moldboards 7 (14/40) Other Vehicle Sensors 1 (48/48) Pavement temperature sensors 2 (39/48) Air temperature sensors Note: In the column of “Ranking by Frequency” the number in parentheses means the number of survey respondents. Sources: Veneziano et al., 2010; and Zhang and Peterson, 2009 Table B-1. Technologies employed in winter operations.

49 and inexpensive while performing relatively well. CaCl2 and MgCl2 perform better than NaCl in colder conditions, but they lead to higher cost and possibly greater impacts on infrastructure. Acetates are less expensive and less cor- rosive to metals. One example is CMA and another is KAc which is good for bridge decks because of its much less corrosiveness. ABPs are an additive to other chemicals to improve performance and reduce corrosion but they have higher costs. • Sanding (Blackburn et al., 2004)—Sanding and use of abra- sives is used to enhance friction on a snow or ice surface, include sand, cinders, ash, tailings, and crushed stone. The last category—emerging technologies—includes incre- mental improvements from the first two categories along with innovations for which such technologies are being attempted by a few states but are not yet widespread. Examples of such technologies are as follows: • Thermal methods, including heated road and bridge deck (Rochelle, 2010; Zhang and Peterson, 2009)—Local heat- ing of road segments segregates the ice-substrate interface and allows ice to be removed with little effort. There are three types of technologies: (1) hydronic (heated fluid is pumped through tubing embedded in pavement); (2) heat pipe (a working fluid contained in steel pipes vaporizes and condenses resulting in a passive transfer of heat); and (3) electric (heat is generated by electrical resistance cables buried in the pavement near the surface). The heat pipe type is perceived as having lower cost and causes less dam- age to infrastructure durability than the hydronic or electri- cal types. In addition to these three primary methods, there are also infrared heating, microwave and radio frequency power—both of which can be mounted on a truck or on bridge side structures with their beams directed toward snow and ice—and solar and wind power as a supplement for electricity generation. • Pre-wetting (Shi, 2010) and (Blackburn et al., 2004)— Pre-wetting is the addition of a liquid chemical to an abra- sive or solid chemical before it is applied to the road. The pre-wetting of solids is performed either at the stockpile or at the spreader. Pre-wetting has been shown to increase the performance of solid chemicals or abrasives on the roadway surface and their longevity, thereby reducing the amount of materials required. Most commercially available liquid ice control chemicals can be used for pre-wetting of solid ice control chemicals, abrasives, and abrasive/solid chemical mixtures. The primary function of the liquid in pre-wetting is to provide the water necessary to start the brine genera- tion process for the solid chemicals. When used on abra- sives, pre-wetting helps abrasives adhere to the ice surface and provides some ice control chemical to the roadway that may at some point improve LOS. • Fixed Automated Spray Technology (FAST) (Zhang and Peterson, 2009)—FAST uses active and passive sensors embedded in the road surface to predict surface tempera- ture and activate the spray system. The system continuously monitors conditions on the structure, based on the detec- tion of critical threshold parameters, and sprays the chemi- cal just in advance of icing conditions. Road sensors can be either passive or active. Passive sensors are tuned for the type of de-icing chemical used in order to determine the proper freezing-point depression. Active sensors can accu- rately measure the freezing point independent of the type of chemicals being used. As of 2003, 23 states either have FAST systems or are planning to install them. Chemicals Listed Abbreviation Ranking by Frequency Sodium Chloride (solid) NaCl (s) 1 (20/24) Abrasives (sand) Sand 2 (17/24) Magnesium Chloride MgCl2 3 (14/24) Agricultural-based Product ABP 4 (12/24) Calcium Chloride CaCl2 5 (11/24) Potassium Acetate KAc 6 (6/24) Sodium Chloride (liquid brine) NaCl (l) 7 (4/24) Sodium Chloride & Abrasives NaCl & Sand 8 (3/24) Clearlane® NaCl, MgCl2 8 (3/24) IceSlicer® NaCl, KCl, MgCl2 8 (3/24) Calcium Magnesium Acetate CMA 11 (2/24) Sodium Acetate Nac 11 (2/24) Potassium Formate Kform 13 (1/24) Note 1: In the column of “Ranking by Frequency” the number in parentheses means the number of survey respondents Note 2: ABP included Ice B’Gone® (n=2), Magic by Caliber® (n=1), beet and/or corn based (n=3), unspecified ABP as inhibitor mixed with MgCl2 (n=2), unspecified ABP as inhibitor mixed with CaCl2 and NaCl (l) (n=1), or an unspecified small amount of ABP listed generally as inhibitor (n=3), and Geomelt® (n=1) Source: Fay et al., 2008 Table B-2. Chemicals employed in winter operations.

50 • Surface overlay (Rochelle, 2010)—Special bonded surface overlays set anti-icing chemicals in place and gradually release them onto the surface. This method is primarily used for bridges against frost and ice formation. For example, the Wisconsin DOT applied a thin layer of epoxy covered with a layer of absorptive aggregate to a bridge in Wisconsin. Alaska and Nebraska also attempted to use rubber asphaltic mixes and a conductive concrete overlay for breaking ice and preventing ice formation, respectively. • Information technology (Rochelle, 2010)—ITS approaches are closely related to snow removal and ice control in that ITS approaches can assist winter management person- nel in making informed decisions. Of many information technologies, the following four technologies pertain to snow removal and ice control: (1) MDSS (maintenance decision support system) which provides objective guid- ance to winter control decisions concerning appropriate strategies; (2) RWIS networks which provide relevant road information through non-invasive road temperature and condition sensors; (3) weather observation technology which uses passenger vehicles as weather probes by having automobile manufacturers equip cars with on-board units and receivers, and collects data such as windshield wiper state or outside air temperature and transmits the data to a national communication station; and (4) GIS (global information systems) and AI (artificial intelligence) which prioritize snowplowing routes and improve snowplowing time and personnel dispatch. • Equipment, including snow-blowing vehicles and blade geometry (Rochelle, 2010)—This applies to equipment containing various advanced technologies ranging from specialized vehicles to specific parts. Specialized vehicles include AVL, which consists of on-board computer sys- tems, pavement-sensing devices, multiple material distri- bution systems, increased horsepower, automated activity reporting, and a friction measuring device; and HMCV (highway maintenance concept vehicle) which applies a precise amount of material at a given time and uses a friction meter to adjust the chemical rate. Specific parts include high-speed environmental plow, which is equipped with flexible cutting edges, and blade geometry with a plow angle of 55 degrees rather than the typical 90 degrees. • Advanced chemicals for de-icing or anti-icing (Kahl, 2002)—Advanced chemicals perform better and are envi- ronmentally friendly and less corrosive than conventional anti-icing and de-icing materials. For example, the Mich- igan DOT used them and consequently could provide the required high LOS (i.e., bare pavements) quicker, while reducing the chemical application rate and inhib- iting the corrosive effect of chloride ions. Maturity of Alternative Technologies To determine the maturity of state-of-the-art snow removal and ice control technologies, the research team analyzed rel- evant patent data. Technological maturity frequently follows the general model of patenting activity shown in Figure B-1. On the x-y plane (where the x axis = the number of appli- cants and the y axis = the number of patent applications), technological maturity shows a spiral shape throughout the stages from the introduction of technology to rapid growth to A general model for patterns in patenting activity can be established to understand the stages of development of a particular technology. On the introduction of a new technology, only applicants are involved in patenting in the field and only few applications are filed. Following this growth period, the technology enters a development period, during which the technology develops rapidly as a result of active competition between numerous applicants, who together file many applications. As research and development continues, the growth in the number of applications stagnates or declines as does the number of applicants. This period can be termed a “maturity period.” As new technologies or even entirely new technology paradigms emerge, a period of decline begins for the original technology, at which point the number of applications and applicants in that field declines strongly. It is possible for a revival of interest to occur in the original technology, if a new application can be found for it, leading to resurgence in the number of applications and applicants (WIPO, 2009). Figure B-1. General model of patenting activity (technological maturity).

51 maturity and then the introduction of alternative technology to the re-discovery of technology. Figure B-2 provides two examples of anti-icers and corrosion- inhibiting chemicals. When compared to Figure B-1, both graphs show these two technologies are still at the stage of technology introduction (Stage I) or at the beginning of rapid growth (Stage II). In other words, these two technologies are not fully mature yet and need more time and more R&D efforts to enter the stable stage. In addition to these two technolo- gies, the maturity of other technologies, such as pre-wetting, thermal method, RWIS and blade geometry, are presented in Appendix E. It seems that all these technologies are at Stage I or II, except for blade-related technologies which are at Stage III (Stable technology renovation). Characterize In this phase, characteristics of both state-of-the-practice and state-of-the-art technologies were described with exem- plary practices of some DOTs. For each technology, its per- formance, corrosive and environmental impacts, and cost were roughly characterized according to parameters rele- vant to transportation agency decision-making. Also, based on the anti-icing strategy as a successful case, the drivers or barriers to technology adoption were evaluated in the organizational terms. In order to consider technological aspects of new, advanced technology adoption, the concept of technological monopolization was introduced using the Concentration Ratio and Herfindahl Index. Characteristics are as follows: • Mechanical plowing (Rochelle, 2010)—Mechanical plow- ing is regarded as the most important and widespread technique. Its innovations come with advances in snow- plow technology, in particular snow blade geometry and the mechanics of scraping snow and ice, which reduce the amount of energy needed to remove snow and ice. Some studies show that fairly minor changes in the cutting edge geometry provide substantially improved ice cutting. For example, preferred blade geometry and serrated blades outperform conventional blades, and trucks with under- body plow blades showed performance that improved on that from front-mounted blades. • Chemicals used for de-icing and anti-icing – De-icing (Fischel, 2001)—De-icing is suitable for most weather, locations, and traffic conditions. It also allows for higher traffic speed and volume, reduces the need for abrasives (thus improving air quality), and saves on fuel consumption compared to plowing alone. – Anti-icing (Rochelle, 2010)—Anti-icing uses smaller vol- umes of chemical to achieve effective results, but is limited by lack of established dispersal rates, laboratory studies to verify field studies, and an understanding of the science associated with anti-icing principles. • Abrasives (Nixon and Williams, 2001)—Because they do not lower the freezing point of water, abrasives are not used for de-icing or anti-icing operations. Even though the Montana DOT uses abrasives in its winter maintenance operations, it says that “abrasives are costly to purchase, store, use and clean up. Additionally, they are poor in performance, have a short beneficial life, and are hard on the environment as well as human health, and cause wear to pavement markings.” Due to these concerns, snow and ice control operations in Mon- tana are shifting from abrasives to the use of winter mainte- nance chemicals to maintain desired levels of service. Table B-3 summarizes the main materials in use in the de- icing and anti-icing activities of agencies. Table B-4 provides further information on widely perceived advantages and disadvantages. 0 2 4 6 8 10 0 5 10 15 No o f A pp lic ati on s No of Applicant Technological Maturity Curve (Anti-icers) Figure B-2. Technological maturity of anti-icers and corrosion-inhibiting chemicals.

52 In the Identify step, several new, advanced technologies were introduced. In particular, information technology and equip- ment have plentiful, heterogeneous sets of sub-technologies. In this case study, however, the research team selected advanced RWISs and changes in blade geometry for discussion and did not more fully explore the other potential innovations and technologies. Thermal Method—Pavement temperature in winter has a significant influence on highway maintenance and safety issues concerned with snow and ice management (Adams et al., 2004). Conductive concrete overlay (Tuan and Yehia, 2004)— Unlike conventional concrete, conductive concrete is a cementi- tious admixture containing electrically conductive components to attain stable and high electrical conductivity. Due to its elec- trical resistance and impedance, a thin conductive concrete Abrasive Sodium Chloride (NaCl) Calcium Chloride (CaCl2) Magnesium Chloride (MgCl2) Calcium Magnesium Acetate (CMA) Potassium Acetate (KAc) Performance Eutectic Temp.58 NA -21°C @23% -51°C @29.8% -33°C @21.6% -27°C @32.5% -60°C @49% General <11°C Effectively depresses the freeze point of water Effective at low temperatures; melts ice faster than NaCl Effective at low temperatures; melts ice faster than NaCl Effective as a liquid anti- icer; melts longer than NaCl Effective as a liquid anti- icer; effective at low temperatures Corrosion Impacts Highway Structures and Vehicles Non- corrosive Corrosive Moderately corrosive Moderately corrosive Non- corrosive Non- corrosive Asphalt Concrete Non- corrosive Slightly corrosive Slightly corrosive Slightly corrosive Moderately corrosive Moderately corrosive Environmental Impacts Air Quality Fine particulate material increase in air pollution Net decrease in air pollution from reduced use of abrasives Net decrease in air pollution from reduced use of abrasives Net decrease in air pollution from reduced use of abrasives Net decrease in air pollution from reduced use of abrasives Net decrease in air pollution from reduced use of abrasives Vegetation Can smother roadside vegetation causing mortality Inhibits water and nutrient uptake; vegetation damage and mortality Inhibits water and nutrient uptake; vegetation damage and mortality Inhibits water and nutrient uptake; vegetation damage and mortality Potential mortality from oxygen depletion in soil Potential mortality from oxygen depletion in soil Soil Little effect on soil expected Increases salinity; decreases soil stability and permeability Increases salinity; improves soil structure Increases salinity; improves soil structure and permeability Potential oxygen depletion from breakdown of acetate; improves soil structure Potential oxygen depletion from breakdown of acetate Surface/Gro- und Water Increases turbidity; inhibits photosyn- thesis in aquatic plants Potential increase in water salinity; slight increase in metals Potential increase in water salinity; slight increase in metals Potential increase in water salinity; slight increase in metals Potential oxygen depletion Potential oxygen depletion Cost Initial Low cost Low cost Relatively low cost Relatively low cost High cost High cost Associated High cost High cost High cost High cost Low cost Low cost Source: Fischel, 2001 Table B-3. Summary characteristics of current winter maintenance materials. 58 The eutectic temperature is the lowest temperature at which ice melts.

53 overlay can generate enough heat to prevent ice formation on a bridge deck when connected to a power source. The Nebraska Department of Roads (DOR) examined this method with the University of Nebraska and Western Michigan University and obtained data showing that an average of 500 W/m2 (46 W/ft2) was generated by the conductive concrete to raise the slab temperature about 9°C (16°F) above the ambient temperature. It proved that the conductive concrete overlay had the poten- tial to become the most cost-effective bridge deck de-icing method. Pre-wetting (Burtwell, 2004)—With the Transportation Research Laboratory (TRL) Limited, the UK Highways Agency and the National Salt Spreading Research Group (NSSRG) evaluated the applicability and costs of introducing pre- wetted salt compared with dry salts. Salt is most effective if it can form a solution with the moisture on the road surface. If this moisture has already frozen before the salt is applied, the salt is much less effective in combating the slippery con- ditions. For this reason, dry salt is wetted and, usually, pre- wetting agents such as NaCl or CaCl2 brine are used. Special vehicles and equipment are needed to process pre-wetted salts: vehicles with a traditional hopper for the dry de-icing agent and integral tanks for the storage of brine; and a satura- tor station to produce the brine solution. Information technology—Advanced RWIS—RWIS con- sists of a network of weather stations, forecasting services, and the supporting infrastructure (Ballard, 2004). Given that it refers to the entire system used to obtain and send data, RWIS is making innovations both in hardware and software. • Hardware (Hoffman et al., 2009)—RWIS stations have pro- gressed from being expensively and permanently mounted along the roadside (even on an existing sign post or over- head mast) to being electrically powered and collecting data from a puck embedded in the road. The Utah DOT made a special point to show the team examples of its newer RWIS stations, which are portable, lightweight, solar- or wind- powered, and video-camera equipped. Furthermore, the Colorado DOT pointed out that new technology can now gather surface condition information noninvasively from Ice Control Materials LOS Advantages Disadvantages Abrasives Low They can provide at least some measure of traction enhancement when it is too cold for chemicals to work effectively. They are suitable for use on unpaved roads and on thick snow pack or ice surfaces that are too thick for chemicals to penetrate. When mixed with enough ice control chemical, abrasives will support anti-icing and de-icing strategies; however, this is very inefficient and costly as the abrasives for the most part are “going along for the ride” while the chemical portion of the mix is doing the “work.” Solid Chemicals High They support both anti-icing and de- icing strategies. When anti-icing, they are most effective when applied early in a winter weather event, before ice or pavement bond has a chance to develop. Some snow, ice, or water on the pavement will minimize bouncing and scattering of the chemicals. They may be used as a pretreatment, but only when applied at traffic speeds under about 30 mph and traffic volumes under 100 vehicles/hr. Liquid Chemicals High They support anti-icing and limited de-icing strategies. They are particularly well suited to pretreating for anticipated frost, icing, or black ice situations. Here, the water evaporates, and the residual dry chemical is relatively immune to dispersal by traffic. Liquid chemicals are also used to pretreat roadways before a general snow or ice event. This is an effective way to initiate the anti-icing strategy. At pavement temperatures higher than about 28°F, liquid chemicals are a very effective treatment for thin ice in the absence of precipitation. The ice- melting process in this situation is almost immediate. They are not well suited to general de-icing operations as they have little ability to penetrate thick snow and ice. They may be used for limited de-icing if the treatment is immediately followed by an application of solid chemicals or the process is reversed. Liquid chemicals are probably not a good choice at pavement temperatures below about 20°F. Liquid chemicals, as a within-winter weather event treatment, should be limited to lower moisture content events, pavement temperatures above 20°F, and cycle times less than about 1.5 h. This will minimize the risk of ice/pavement bond formation. It is not advisable, however, to use liquid chemical during moderate or heavy snow, sleet, and freezing rain events. Source: Blackburn et al., 2004 Table B-4. Advantages and disadvantages of state-of-the-practice technologies.

54 the roadside and is able to provide a useful measurement of slipperiness (i.e., friction). • Software (Boselly, 2004)—Since the late 1980s, some RWISs such as Weather Traveller Information Web and Real-time Road became operational. However, they were not enough to provide key forecast information needed by decision- makers. In this vein, the Washington DOT developed a new capability for maintenance operations decision-making, named ARROWS (Automated Real-time ROad Weather System), with the University of Washington. ARROWS takes numerical weather prediction and pavement condition out- puts and presents the forecast information in a format for easy use and understanding by maintenance personnel. To do that, ARROWS requires the high-resolution mod- eled output, the integration of other weather information sources, and developing the presentation format. Equipment—Blade geometry—Cutting edges (Hoffman et al., 2009)—The type of plow cutting edge varies from state to state, mostly based on temperature, weather conditions, and whether the agency is dealing with solid ice, snow, or slush. Blades are made of regular rolled steel, hardened steel, serrated steel, carbide steel inserted in steel, rubber blades, and carbide steel inserted in rubber. The most popular configuration is the standard single blade per plow configuration; however, experimentation is being done with double- and triple-blade configurations. While snow- and ice-removal agencies are interested in the experiences of others, they tend to do their own in-house experimenting and to base purchasing selec- tions on their own research. Much depends on their organi- zation’s own unique culture and their own weather and traffic conditions. Driving forces include extending life (thus reduc- ing frequency of replacement), reducing vibration, minimizing damage due to obstructions on the surface, reducing noise, and balancing pressure to reduce chemical usage versus willingness to achieve goals using more chemicals. Perhaps equally impor- tant is the operator’s degree of interest and willingness to try or use something new. Today’s industry is responsive to user agen- cies in providing options to cover varying needs and conditions. Advanced chemicals for de-icing or anti-icing: • Agriculture-based products (ABP): Research in recent years shows that adding various organic compounds to common winter maintenance chemicals can significantly decrease the freezing point (Koefod, 2008). Nixon suggests that ABP can be combined with winter maintenance materials to act as corrosion inhibitors and increase melting capacity (Nixon and Williams, 2001). ABP have low eutectic and effective temperatures and are relatively benign to the environment and highway infrastructure. • Corrosion-inhibiting chemicals: Corrosion-inhibited deicer products must prove to be at least 70% less corrosive than NaCl to be qualified for the PNS59 (Pacific Northwest Snow- fighters) specification for corrosion60 (Fay et al., 2008). Drivers or Barriers to New Technology Adoption According to the case study presentation arranged by TRB, it is said that anti-icing was one example of a successful distribution of new, advanced technology in the transporta- tion area (even though it is not yet a complete success in that some states still are not practicing anti-icing). The research team therefore examined the critical factors affecting distri- bution (i.e., the drivers or barriers to technology adoption) in the specific areas of snow removal and ice control. These include • Number of Initial Participants. Twelve states were involved in the anti-icing program from the start. There were enough personnel in several agencies to know the “language” and to try the new method. • Attitude toward Failures. When practicing the anti-icing program, early failures were accepted as part of the learning process. • Knowledge Sharing. Experience of anti-icing practices was collected into a manual that is readily available and includes clear guidelines. In particular, the manual was freely avail- able on line and had some user-oriented charts that gave recommended practice in most conditions likely to prevail during winter weather. • Communication. The lead states team focused on com- munication. Even team members who were not themselves technical staff played a significant role in communication. 59 PNS is the association of state members of Washington, Oregon, Montana, Idaho, Colorado, and British Columbia, whose mission is to strive to serve the traveling public by evaluating and establishing specifications for products used in winter maintenance that emphasize safety, environmental preservation, infra- structure protection, cost-effectiveness and performance. PNS developed specifi- cations for chemical products which must pass a series of tests for chemical, fric- tional, toxicological, and corrosion; meet environmental and health standards; and be at least 70% less corrosive than road salt. 60 One of PNS’s functions is to develop anti-icing chemical specifications that all member organizations utilize. The PNS specification for corrosion is that a corrosion-inhibited anti-icing chemical must be at least 70% less corrosive to a given type of metal than sodium chloride is corrosive to that same type of metal. This reduced level of corrosion is determined by a laboratory test. Generally, the lab test consists of immersing and removing separate metal washers in a sodium chloride solution and a corrosion-inhibited chemical solution. Over a 72-hour period, the metal samples are immersed for 50 minutes and removed from the solution for 10 minutes. This immersion and removal process is done hourly for the 72-hour period. After the test period is complete, the metal samples are weighed. If the metal sample exposed to corrosion-inhibited chemicals has at least 70% less weight loss compared with the weight loss of the metal sample exposed to the sodium chloride solution, the corrosion-inhibited chemical meets the PNS specification (Baroga, 2004).

55 • Conservation of Momentum. When the lead states effort came to its conclusion, the SICOP61 (Snow and Ice Pooled Fund Cooperative Program) was present and able to take over. Dissemination costs money, and SICOP provided (via the states) the conduit for that money to keep flowing. Technological Monopolization In addition to organizational issues affecting new technol- ogy adoption as discussed in the previous section, technol- ogy applications themselves also generate drivers or barriers to broader employment of new, advanced technology. One such technology-specific factor is the degree of monopoliza- tion in presenting applications of a technology to the market. As economic theory shows, a monopoly results in a smaller number of products at higher price when compared to a competitive market. A monopoly would therefore incline a vendor to impose a high price on selling or licensing their technologies.62 This situation is enabled by strong and/or numerous patents. Agency officials may be more reluctant to commit to a tech- nology application for which there is only one vendor com- pared to a situation where there are multiple vendors (thus giving the potential adopting agency more market power). To look at the degree of technological monopolization in the area of snow removal and ice control, the research team analyzed patent data again and computed two indicators— Concentration Ratio (CRn) and Herfindahl-Herschman Index (HHI)—to determine how intensely patents in specific tech- nologies are concentrated in a handful of companies and to compare them across technologies. When examining the maturity of technologies in areas of snow removal and ice control before, the research team found these technologies are in the initial phase of develop- ment. This often means there are a few relevant companies. Table B-5, which gives the market share of the three largest firms (CR3) and HHI,63 shows that pre-wetting technology is intensively concentrated and dominated by one company. Environmentally friendly chemicals are moderately concen- trated and other technologies have low levels of concentra- tion. Pre-wetting and environmentally friendly chemicals, therefore, have a certain level of limitation in vendor choice and proprietary relations. Regardless of organizational issues, in this case technology itself might pose a sizable hurdle for adopting new, advanced technology due to the difficulty of technology transfer and implementation and subsequent low accessibility and availability. Compare In this phase, the research team created metrics with both normative and natural units in order to evaluate likely out- comes stemming from each candidate technology. On the basis of given assumptions and initial conditions, the research team sorted out five technology packages: use of either CaCl2 or KAc, delivery of each by vehicle or FAST, and the thermal method— and compared them with respect to created metrics and cost to adopt. When creating metrics, the research team consid- ered four primary DOT mission goals: preservation, safety, mobility, and sustainability. When conducting cost analysis, the research team included not only the easy-to-quantify part (e.g., installation and material costs) but also the hard-to- quantify part (e.g., the cost of corrosion and environment and the benefit from reduced travel time and crashes). The former are used to provide input to the STREAM Decide step. The latter will be used as a check on the benefit metrics discussed immediately below. Creating Appropriate Metrics In order to compare technologies, the research team needed to (1) sort out candidates from various available and emerging Heated Road Pre-wetting RWIS Blade Geometry Advanced Chemicals Environmentally friendly Corrosion- inhibiting CR3 24 100 30 11 57 36 HHI 202 5000 300 47 1950 528 Table B-5. Concentration ratio and Herfindahl-Herschman index. 61 SICOP has the task of demonstrating the effectiveness of the new technol- ogy (and other new tech) rather than creating new research. It is a pooled fund study that requests about $2,000 from each state every 2 or 3 years for ongoing expenses and solicits additional funds for specific projects (such as the soon- to-be-released computer-based training in RWIS and anti-icing). SICOP com- municates via their website (www.sicop.net or http://www.transportation.org/ Default.aspx?SiteID=88) and the snow and ice listserv. 62 The degree of monopoly rents the vendor chooses to extract may be modi- fied by the time frame within which it is willing to operate. We speak now only of pure theory. 63 A small index indicates a competitive industry with no dominant players. HHI scores above 2,500 suggest a high degree of concentration.

56 technologies to be compared; (2) set up standards and create metrics to judge; (3) assess the range of possible outcomes from the use of each candidate technology; and (4) judge likely outcomes according to created metrics. Given that snow removal and ice control strategy depends on several conditions (e.g., LOS and weather), candidate technologies are limited according to those conditions. For example, air or road temperature puts a limitation on the avail- ability of chemicals because of their eutectic temperatures. In a region with sensitive vegetation or national monuments, chloride chemicals are excluded from the list of candidate technologies. When creating metrics, the research team needed to con- sider DOT goals (e.g., preservation, safety, mobility, and sustainability). Then, the research team could think of other metrics as well. For example, using NaCl as a reference point, the following paragraph suggests some idea of how to com- pare alternatives (For example, it contains information about cost, the ease of application, the LOS, corrosive effects, and environmental impacts.) Rock salt is used extensively because it is inexpensive, easy to spread, and effective in keeping pavements safe in the win- ter. Damage to vegetation, soil, water quality, vehicles, and infrastructure is the known negative impact of rock salt, although most deicers have some of these environmental impacts (Burtwell, 2004). Potential outcomes of each candidate technology could be estimated through literature review, laboratory and field experiments, and users’ perception based on their experience and expertise. In this vein, Fay et al. (2008) shows user per- ception on performance of winter chemicals (see Figures B-3 and B-4). Finally, the research team matched likely outcomes with the metrics and evaluated them. The research team needed to consider various aspects—not only DOTs’ goals but also of other metrics such as costs across candidate technologies. Among the current-use and state-of-the-art technologies discussed in the Identify step, several appeared likely candi- dates for consideration under the assumptions and initial conditions presumed in the scenario. Blackburn et al. (2004) suggests the following conditions when selecting appropriate technologies (Blackburn et al., 2004): • Climate conditions: frequency of snow and ice events (low/ moderate/high), severity of winter pavement exposure (mild/moderate/severe), wintertime precipitation (type/ rate), urban influence (small/medium/large/industrial), water influence (minor/river/lake/ocean), elevation/ large-scale topography (plain/rolling/mountainous) • Weather conditions: rain (light/moderate/heavy/freezing), sleet (light/moderate/heavy), snow (light/moderate/heavy/ blowing; powder/ordinary/wet or heavy) • Site conditions: area type (urban/suburban/rural), special highway segment area (hills/curves/grades/intersections/ bridges/sags/ramps/crosslopes/weaving areas/narrowings/ roadway widenings/elevated roadways/pavement surface types/tangents), shadings from solar influence (forest or vegetation/buildings or structures/cuts), pavement condi- Figure B-3. User perception on positive performance.

57 tions (temperature/ice and pavement bond/frost or thin ice/ slush, loose snow, packed snow, thick ice) • Traffic conditions: traffic volume (very low/low/medium/ high/very high), commercial vehicle mix (low/moderate/ high), vehicle speeds (low/moderate/high) Definition of Metrics In the most general case, the research team developed a metric for each of the mission objectives: safety, preserva- tion, mobility, and sustainability. If these were true out- come metrics, each would be measured in its natural units, e.g., “safety” in terms of car crashes (or relative risk), “pres- ervation” in terms of maintenance of road and bridge con- dition (or relative corrosion), “mobility” in terms of travel time (or relative speed), and “sustainability” in terms of environmental impact (or relative detrimental effects on air, soil, water, etc.). The effect of each technology alterna- tive on each of these outcomes would have a distribution that depends on how the technology is implemented and used, as well as the specific characteristics of weather, site, and traffic. Winter maintenance is subject to environmental condi- tions as well as methods or strategies in use as shown in Fig- ure B-5. Although input factors are homogeneous, outputs or outcomes may be heterogeneous. In this vein, the research team used output64 or outcome65 measures for the metrics rather than input66 measurements. Metric for Safety and Mobility For the example of snow removal and ice control, the research team defined a single metric for “safety” and “mobility” in terms of how well each candidate technology can attain the desired LOS. When defining LOS goals, Blackburn et al. suggest pavement snow and ice conditions (PSICs) (Blackburn, Amsler Sr., and Bauer, 2004). For each “Condition State,” the Manual recommends different feasible actions, as shown in Table B-6. Higher LOS are associated with “better” PSICs and more rapid achievement of “better” or “bare” pavement conditions. Given that PSICs include several measures related to “safety” and “mobility,” the research team based the overall metric on them and sought to make an assessment that would Source: Fay et al., 2008 Figure B-4. User perception on negative performance. 64 Output measures quantify physical outputs from the resources that are used in units of work of winter operations. Output specifications primarily deal with defining methods of performing the work and the associated accomplishments (Maze, 2007). 65 Outcome measures reflect the end result of winter maintenance during and after a storm event, usually as perceived by the motorist (Maze, 2007). 66 Input measurements are used to quantify the resources spent on snow and ice control or winter maintenance operations, typically applied to equipment, material, and labor used for winter operations (Maze, 2007).

58 incorporate the following measures that might actually be available to a transportation agency (Maze, 2007): • Measure: Degree of clear pavement – Approach: Manual observation – Approach: Camera-assisted observation • Measure: Traffic flow – Approach: Detectors providing information on speed, volume, and occupancy – Approach: Road closure • Measure: Crash risk – Approach: Friction (or slipperiness) – Approach: Reported crashes Table B-6 shows that road surface conditions are associated with the friction coefficient which has direct bearing on both “safety” (relative risk) and “mobility” (relative speed). The research team used those values as the natural unit for the pur- pose of comparison. According to a Swedish study, for example, sanding could produce an increase in the friction coefficient of around 0.1 from a baseline level of around 0.2–0.3 and, consequently, cause travel speed to be increased on average by 2.4 km/h. It is known that sanding generally could reduce the number of accidents by 62% at a 5% significance level. In terms of PSICs as shown in Table B-6, sanding could change the road surface conditions to Condition 3 from Condition 4. Based on these PSICs, the research team could then define the following metric for “safety” and “mobility” in terms of normative units. These, in turn, could be obtained through either a Delphi technique or a survey of relevant professionals: • Metric Value 1 – Inability to attain Conditions 1, 2, 3, and 4; • Metric Value 2 – Ability to attain Condition 4, but not Conditions 1, 2, and 3; • Metric Value 3 – Ability to attain Condition 3, but not Conditions 1 and 2; • Metric Value 4 – Ability to attain Conditions 1 or 2. Table B-7 shows both normative and natural units in order to compare candidate technologies. Metric for Preservation and Sustainability Because the effects on “preservation” and “sustainability” of snow removal and ice control primarily result from (1) corro- sion and (2) dispersal and runoff of winter maintenance chemi- cals (with chlorides) which lead to structure (including bridge) deterioration and the contamination of air, water, and soil, the research team defined a single metric for “preservation” and “sustainability.” The problem is that these detrimental effects are highly variable depending on location, the type of chemi- cals, application rate, and so on. For example, chemicals applied to a bridge adjacent to the sea might have no actual influence on the environment. On the other hand, chemicals applied to a road and then splashed off to a roadside where chemical- sensitive vegetation and wildlife grow might have a significant adverse effect. The transport of salt from the road to the roadside envi- ronment is the main environmental concern of winter main- tenance. The basic mechanisms determining the salt exposure are salt dose to the road, road conditions, traffic characteris- tics (type, intensity, and speed), and meteorological param- eters, such as wind (Gustafsson and Blomqvist, 2004). In addition, there are several categories of adverse effects from spills (Burkett and Gurr, 2004): • Soils and groundwater: increase in calcium, magnesium, phosphorous, and soil organic matter Source: Maze, 2007 Environmental Conditions = Storm Severity Inputs = Labor, Equipment, Materials, Management, and Information Quality and Quantity Snow Removal and Ice Control - Outputs Desired Outcome = Customer Satisfaction Safety Preservation Mobility Sustainability Performance Measure = Level of Service, Time to Bare Pavement Terrain & Geography Solar Energy Precipitation Wind Speed Temperature Anti-icing Cycle Length Number and Type of Truck Abrasives Salt RWIS Operations Management Figure B-5. Relationship between inputs, outputs, and outcomes in snow removal and ice control.

59 Road surface conditions (Friction coefficient) Safety: Relative risk Mobility: Relative speed Condition 1 All snow and ice is prevented from bonding and accumulating on the road surface. Bare/wet pavement surface is maintained at all times. Traffic does not experience weather-related delays other than those associated with wet pavement surfaces, reduced visibility, incidents, and "normal" congestion. (0.7 – 0.9) 1.0 1.0 Condition 2 Bare/wet pavement surface is the general condition. There are occasional areas having snow or ice accumulations resulting from drifting, sheltering, cold spots, frozen melt- water, etc. Prudent speed reduction and general minor delays are associated with traversing those areas. (0.4 – 0.7) 1.3 0.9 Condition 3 Accumulations of loose snow or slush ranging up to 2 in. are found on the pavement surface. Packed and bonded snow and ice are not present. There are some moderate delays due to a general speed reduction. However, the roads are passable at all times. (0.3 – 0.4) 1.5 0.7 Condition 4 The pavement surface has continuous stretches of packed snow with or without loose snow on top of the packed snow or ice. Wheel tracks may range from bare/wet to having up to 1.5 in. of slush or unpacked snow. On multilane highways, only one lane will exhibit these pavement surface conditions. The use of snow tires is recommended to the public. There is a reduction in traveling speed and moderate delays due to reduced capacity. However, the roads are passable. (0.1 – 0.3) 2.5 0.4 Condition 5 The pavement surface is completely covered with packed snow and ice that has been treated with abrasives or abrasive/chemical mixtures. There may be loose snow of up to 2 in. on top of the packed surface. The use of snow tires is required. Chains and/or four- wheel drive may also be required. Traveling speed is significantly reduced and there are general moderate delays with some incidental severe delays. (smaller than 0.1) 4.4 0.3 Condition 6 The pavement surface is covered with a significant buildup of packed snow and ice that has not been treated with abrasives or abrasives/chemical mixtures. There may be 2 in. of loose or wind-transported snow on top of the packed surface due to high snowfall rate and/or wind. There may be deep ruts in the packed snow and ice that may have been treated with chemicals, abrasives, or abrasives/chemical mixtures. (smaller than 0.1) Greater than 4.4 0.1 The use of snow tires is the minimum requirement. Chains and snow tire equipped four-wheel drive are required in these circumstances. Travelers experience severe delays and low travel speeds due to reduced visibility, unplowed loose, or wind-compacted snow, or ruts in the packed snow and ice. Condition 7 The road is temporarily closed. This may be the result of severe weather (low visibility, etc.) or road conditions (drifting, excessive unplowed snow, avalanche potential or actuality, glare ice, accidents, vehicles stuck on the road, etc.). 0 Sources: Blackburn et al., 2004; and Elvik and Hoye, 2009 Table B-6. Description of PSICs.

60 • Terrestrial vegetation: potential damage by airborne contaminants • Streams: depletion of dissolved oxygen (DO), water quality concern • Air quality: fine particulate material For the normative unit, the research team created metrics based on output or outcome of environmental and corrosive effects with four different scales for measurement in both corrosive and environmental terms: • Metric Value 1 – serious effect; • Metric Value 2 – moderate effect; • Metric Value 3 – small effect; • Metric Value 4 – no/little effect. In creating this metric the research team considered win- ter operation strategies’ environmental effect on all four subjects—air, vegetation, soil, and water—and then used the average value of the effect on these four categories. Based on the natural units in Table B-8 and the norma- tive metric above, the research team created Table B-9, which compares candidate technologies. Metric for POSI The research team based the metric for “POSI” on the severity of the barriers to implementation (e.g., the inertia of DOTs to avoid changing from the current methods of plowing, sanding, and salting; difficulties associated with approvals, acquisition, training, and other necessary actions to enable the use of the alternative technologies; and concerns about its use, for example, uncertainties about the cost and technical viability of the new method, as well as its possible demands on management and training resources.) Based on Table B-10, the research team in Table B-11 defined the following metric for “POSI.” Comparison of Current Methods and Technology Alternatives Assumptions To consider several conditions when comparing can- didate technologies, the research team assumed that the PSICs accurately and fully reflect the conditions of climate and weather. Furthermore, through the initial conditions, the research team took the conditions of site and traffic into account. Another important assumption was about the frequency and regularity of snowfall and inclement conditions. More northern states expect snow during winter and thus prepare for winter storms; however, some states with less severe winters may also anticipate snow at irregular intervals. To account for this inter- mittency of snow, the research team assumed that each state behaves in the economic manner. In other words, that states are rational and try to optimize winter operations as appro- priate. This posture is then reflected in their winter strategies, Chemical Corrosion Rate (mils per year) Relative Corrosiveness Calcium Chloride 63.53 1.2 Sodium Chloride (rock salt) 52.94 1 Magnesium Chloride 17.44 0.33 Distilled Water / Potassium Acetate (KAc) 5.29 0.1 Calcium Magnesium Acetate (CMA) 4.16 0.08 Source: WSDOT and Minnesota Corn Processors Table B-8. Comparative corrosion rates and relative corrosiveness for selected chemicals. Detrimental Impact Normative Unit Natural Unit Preservation Sustainability No/Little 4 Smaller than 0.1 4 Small 3 0.1 – 0.5 3 Moderate 2 0.5 – 1 2 Serious 1 Greater than 1 1 Note: The natural unit of “preservation” follows “relative corrosiveness.” The research team used the same value in “sustainability” across normative and natural units. Table B-9. Metrics table for preservation and sustainability. PSIC Normative Unit Natural Unit Safety Mobility 1 4 1.0 1.0 2 0.77 0.9 3 3 0.67 0.7 4 2 0.4 0.4 5 1 0.23 0.3 6 0.15 0.1 7 0.05 0 Note: The natural unit of “safety” is the inverse of “relative risk” while that of “mobility” equals “relative speed.” Table B-7. Metrics table for safety and mobility.

61 primarily de-icing and anti-icing. Based on that, the research team assumed that snow-belt states execute anti-icing practices while non-snowy region states employ de-icing practices. This seemed reasonable in as much as anti-icing requires a large prior investment. To take effective measures before snow storm events, anti-icing requires more expensive chemicals (e.g., liquid chemicals), information technology (e.g., RWIS and MDSS), relevant equipment, and appropriate training. Thus, it would be economically optimal for less severe region states to prepare for intermittent snow storms with de-icing strategies. Initial Conditions For comparison purposes, the research team set up the fol- lowing initial conditions: • Climate and weather conditions: Snow-belt States. The research team assumed PSIC = 2 before snow storms and expected PSIC = 4 after snow storms without anti-icing. The aim was to attain PSIC = 2 through winter operations. – Anti-icing: Because this area belongs to snow-belt states, in particular the Northeast, the Midwest or Alaska, the research team assumed that this area adopts anti-icing strategies. – Low temperature: The research team assumed that these states want to make certain that they can operate their de-icing and anti-icing missions effectively at any temperatures above the very low temperature of -40°F and so this became the planning scenario. This would limit the availability of candidate technologies in part because of chemicals’ eutectic temperature. For exam- ple as Figure B-6 shows, at such low temperatures only CaCl2 and KAc are appropriate for application. • Site conditions: Bridge – Various available technologies: Bridges are of interest for the present purpose because they permit a greater Category of Impediments Specific Barriers Affecting POSI Technology Unfamiliarity with core or applied technology Uncertainty concerning actual performance Additional implementation requirements (training, standards, etc.) Agency Process or Institutions Need for new or conflict with existing regulations & standards Non-fungibility of funding for required expenditures Extended or problematic approval processes External to Agency Inertia of existing processes and methods Insufficient political or public acceptance Lacking presence of necessary vendor or support base Table B-10. Sources of impediments that reduce POSI. POSI Score Level Conditions for Achieving POSI Score Level 4 Number of Major Concerns = 0 3 Does not meet criteria for POSI score level of 1, 2, or 4 2 Number of Show Stoppers = 1 or Number of Major Concern > 2 1 Number of Show Stoppers > 1 Table B-11. Metric table for POSI. Figure B-6. Eutetic temperature and concentration of chemicals. 67 AADT = annual average daily traffic variety of technologies to be compared than with roads. In addition to chemicals, for example, the research team could test thermal methods (pavement heating technol- ogies), FAST, and so forth. • Traffic conditions: Highway—For conditions, the research team drew on the data shown in Table B-12 from the I-35W bridge over the Mississippi in Minnesota where anti-icing strategies were employed in 2000 (Johnson, 2001). – Traffic volume: AADT67 = 140,000 vehicles with the number of commercial vehicles = 5,000.

62 – Number of car accidents: the number of crashes related to human and property damages dropped from 31 (dur- ing winter season of 1996-1997) to 9 (during 2000-2001). Candidate Technologies to be Compared The research team examined five different technology bun- dles for this demonstration. Four of the bundles were varia- tions based on combinations of chemical type (CaCl2 or KAc) and application method (vehicle or FAST). The fifth technol- ogy alternative was the thermal method, described above. Drawing on the description of the metrics discussed above and a review of the relevant literature, the research team assessed and compared the candidates in Table B-13 (and Figure B-7) based on normative units and Table B-14 (and Figure B-8) based on natural units. It is worth noting that in a full STREAM analysis of this agency activity, the research team would include a richer set of both technology bundles and of alternative sources of inputs from expert sources thus most likely generating a range of assessments that would be treated in the same manner as the bridge deck monitoring and evalu- ation assessment presented above. There did not seem to be much difference between the two approaches to constructing a metric for comparison. The order of preference between technology bundles remained the same between the two views and the absolute magnitude of this ordering remained largely the same, with one excep- tion. When measured in natural units, the drawbacks for Non-dry surface crash before and aer employing the an-icing system Description Unit Before Aer damage Fatal crashes 0 0 Injury Type A only crashes 0 0 Injury Type B only crashes 2 0 Injury Type C only crashes 8 5 Property Damage only crashes 21 4 daily traffic vehicles 140000 auto vehicles 135000 truck vehicles 5000 Note: “Before” = 1996-1997 winter season; “After”= 2000-2001 winter season. Table B-12. Comparison of crash data for I-35W bridge before and after treatment. Thermal Method Calcium chloride Potassium acetate Calcium chloride Potassium acetate (e.g., thermal fluid) Attained PSIC 3 3 3 3 2 Safety 3 3 3 3 4 Mobility 3 3 3 3 4 Corrosive effect on: Highway structures & vehicles 2 4 2 4 4 Asphalt concrete 3 2 3 2 4 Preservation 2.5 3 2.5 3 4 Environmental effect on: Air 4 4 4 4 4 Vegetation 2 3 2 3 4 Soil 2 3 2 3 4 Water 2 3 2 3 4 Sustainability 2.5 3.25 2.5 3.25 4 Probability of Successful Implementation: Technology 4 3 3 3 3 Agency Process or Institution 4 4 3 3 2 External to Agency 4 4 3 3 3 POSI 4 3.67 3 3 2.67 Application by Vehicle Chemical Method Application by FAST Table B-13. Comparison of technology application bundles using normative units.

63 0 0.5 1 1.5 2 2.5 3 3.5 4 Safety Mobility PreservationSustainability POSI CaCl2 (Vehicle) KAc (Vehicle) CaCl2 (FAST) KAc (FAST) Thermal Method Figure B-7. Comparison of technology application bundles using normative units. Thermal Method Calcium chloride Potassium acetate Calcium chloride Potassium acetate (e.g., thermal fluid) Attained PSIC 3 3 3 3 2 Safety 2.68 2.68 2.68 2.68 3.08 Mobility 2.8 2.8 2.8 2.8 3.6 Corrosive effect 1.2 0.1 1.2 0.1 0 Preservation 0.33 4 0.33 4 4 Environmental effect Sustainability 2.5 3.25 2.5 3.25 4 Probability of Successful Implementation: Technology 4 3 3 3 3 Agency Process or Institution 4 4 3 3 2 External to Agency 4 4 3 3 3 POSI 4 3.67 3 3 2.67 Chemical Method Application by Vehicle Application by FAST Table B-14. Comparison of technology application bundles using natural units. 0 1 2 3 4 Safety Mobility PreservationSustainability POSI CaCl2 (Vehicle) KAc (Vehicle) CaCl2 (FAST) KAc (FAST) Thermal Method Figure B-8. Comparison of technology application bundles using natural units.

64 using CaCl2 as the active chemical became starker. In the nor- mative 1 to 4 scaling, by definition there cannot be a greater than four times difference. When using natural units, how- ever, an order of magnitude (or larger) difference becomes possible, as in this case. Both methods can convey useful information to agency decisionmakers and the use of both provides a cross-check on each other. Cost of Current Methods and Technology Alternatives Given that transportation, and therefore snow removal and ice control, is closely related to human lives and activities, it has a significant effect on not only DOTs in terms of preser- vation and sustainability but also users in terms of safety and mobility. All of these, therefore, have costs and benefits asso- ciated with them. Road users, for example, are affected by the cost of accidents (as well as the value of accidents avoided), travel time, fuel cost, corrosion damage, snow tire and equip- ment costs, and the loss of benefit from cancelled journeys. Road authorities face costs of re-asphalting, road mark- ings, washing signs, increased bridge maintenance, as well as the direct costs of sanding, salting, and plowing (Elvik and Hoye, 2009). These money values frequently consist of two parts: easy to quantify and hard to quantify. Table B-15 shows some examples in the area of snow removal and ice control. In what follows the research team considered the easy-to-quantify costs as being representative of the fixed and recurring costs actually incurred by transportation agencies. This is what the research team used when analyzing the cost of each candidate technology adoption. The hard-to-quantify costs largely rep- resent quantifications of some of the mission goal benefits discussed above. The data on which the calculations below are based is pro- vided in Appendix E. The research team presents only the results of analysis here. Table B-16 shows those costs that apply directly to agencies. These include one-time initial costs such as equipment and installation and recurring costs such as materials and labor. In this scenario, it would seem that calcium chloride by vehicle is least costly compared to the use of potassium acetate and other methods for dispersal such as FAST. The thermal measure appears most costly among the five shown. Yet, if different categories of cost are more or Technologies Cost Benefit Easy to quantify Hard to quantify Easy to quantify Hard to quantify Anti-icing and de-icing Materials, labor, maintenance Environmental and societal impacts Fuel and travel time savings, the potential for material and labor savings Plow blades Equipment Potential damages caused by changes to plowing equipment and practices Labor and material savings Efficiency gains, safety improvements, and added equipment versatility RWIS Complete site installations Maintenance, power, and communications Labor and material savings, improved LOS, safety improvements, lower insurance costs, and fuel savings Source: Veneziano et al., 2010 Table B-15. Costs and benefits of technologies. Cost Thermal Method Calcium chloride Potassium acetate Calcium chloride Potassium acetate (e.g., thermal fluid) Net Parcipants' Cost Equipment 20 20 300 300 300 Installaon 5 5 300 300 400 Lifecycle (O&M) 2 2 20 20 30 Ulity Incenve Payments 0 0 0 0 0 Cost of Operaon Materials 60 120 60 120 100 Labor 60 60 1 1 2 Total Cost 147 207 681 741 832 Chemical Method (Vehicle) Chemical Method (FAST) Table B-16. Cost calculation, transportation agency ($000s/bridge).

65 less sensitive to modification or local agency circumstances, it is important to look beyond the total shown. The use of calcium chloride in combination with FAST has the lowest operational cost followed by the thermal method which has the highest net fixed cost. If initial funding support is avail- able from outside sources, such as the federal government, FAST or the thermal method remain viable candidates on this basis. The calculation becomes more complicated if the research team also seeks to quantify net costs when looking at elements of cost and benefit that accrue largely to the road users and society such as corrosive and environmental effects, influence on travel time, and human and property damage from car accidents. The single cursory analysis presented in Table B-17 seems to show that the thermal method is most attractive followed by FAST, given that they bring greater benefit from travel time savings and crash reduction as well as smaller monetary damage from corrosion and to the environment compared to chemical methods by vehicle. In a fuller analysis, these costs elements would also be crafted from several input sources and from surveys of experts in order to craft ranges of possible net cost as was done in the prior bridge deck evalua- tion application of STREAM. Solely for the purpose of obtaining a final cost calculation for this demonstration of STREAM, the research team now combined the two categories of cost through simple addition. The result is shown in Table B-18. Because the elements on the benefit side of the ledger in the hard-to-quantify part are of a larger magnitude than those on the cost side, and this net result is also larger than the net cost from the easy-to-quantify part, these values show a combined net benefit for each technology. By this calculation, the thermal method is the most attractive because of its huge benefit from reduced travel time and crash avoidance. The next most attractive technology is potassium acetate by vehicle. In what follows the research team, however, relied solely on the elements of agency cost to derive insights on the decision facing transportation officials.68 Decide The first three steps of STREAM could be done on a col- laborative basis by agencies or by an external body that would provide the information to agencies. The Compare step would also be performed on this basis but would also require a par- ticular agency to modify it in accord with local conditions and preferences. This could certainly be done after receiving a pre- liminary comparative analysis. The purpose of these steps has been to provide extensive and comparable information that is also framed in the relevant terms for transportation agen- cies and the decisions they face. Such input or evidence is not fully actionable: there are other factors to consider. As a result, the ultimate decision rests with the individual transportation agencies themselves. The Decide step outlines the procedures that may be followed to weigh these alternatives. What Factors to Consider As indicated in the Compare phase, snow removal and ice control technology is dependent on several conditions. Road users care about the LOS, corrosive effect on their vehicles, and health effects. Transportation agencies are interested as well in performance, cost, and the ease of application. In both cases, weather and location are important because they play a key role in the availability of technologies. Table B-19 summarizes these considerations. Cost Thermal Method Calcium chloride Potassium acetate Calcium chloride Potassium acetate (e.g., thermal fluid) Cost of Corrosion and Environment 300 120 300 120 - Benefit of Travel Time Saving - - (394) (394) (590) Benefit of Reduced Crash (756) (756) (756) (756) (869) Total Cost (456) (636) (850) (1,030) (1,459) Chemical Method (FAST)Chemical Method (Vehicle) Table B-17. Cost calculation, user and society perspective ($000s/bridge). Cost Thermal Method Calcium chloride Potassium acetate Calcium chloride Potassium acetate (e.g., thermal fluid) Total Cost (309) (429) (169) (289) (627) Chemical Method (Vehicle) Chemical Method (FAST) Table B-18. Net cost calculation, both parts ($000s/bridge). 68 Not to do so would be a case of double counting since the issue of costs and benefits from the perspective of agency missions and goals has already been rendered in the mission-specific metrics discussed above.

66 Tradeoffs in Metrics There are complex tradeoffs between metrics for pres- ervation, safety, mobility, and sustainability. Regarding envi- ronmental issues, for example, transportation agencies are continually challenged to provide a high LOS and improve safety and mobility in a cost-effective manner while minimizing corrosion and other adverse effects to the environment (Maze, 2007). According to Maze (2007), it is desirable to adopt new, advanced technologies to attenuate tradeoff problems. But, there still exist tradeoff issues even if the magnitude of issues is relatively small as compared to traditional technologies. The research team now used the data developed in the Com- pare phase, in both normative and natural units, to help weigh tradeoffs between metrics. Figure B-9 shows a series of one-on- one direct comparisons between the four summary measures in normative units the research team had developed for the four main metric categories corresponding to four main agency mis- sions and also the POSI metric. Overall, it seems that the ther- mal method is superior to others largely because of its higher scores for sustainability and preservation. Yet, it turns out that CaCl (Vehicle) and KAc (Vehicle) are better technologies in terms of POSI. This is not surprising given that these come closer to the current state of practice. Figure B-10 shows a similar set of simple comparisons for the same metrics, this time using the natural unit measures. They show similar results to those presented using the nor- mative measures. Such views (of which Figures B-8 and B-9 provide only some examples) can be useful in the decision-making at an agency level. But such views can also be difficult to interpret because of the large number. For this reason, it is useful to develop more integrated representations of this information. Figure B-11 is such a representation. As in the bridge deck evaluation exam- ple of STREAM, the research team presents the respective POSI scores along the x-axis, an integrated value metric consisting of an unweighted multiplication of the individual goal metrics70 as a y-axis, and the plotted expected values.71 The research team also shows hyperbolae representing lines of equivalent trades between the combined value metric and potential difficulty of implementation (the POSI measure). By using the aggregate value metric instead of individual metric, the research team can consider all metrics versus POSI at the same time. Figure B-10, using the normative unit measures, suggests the marked attractiveness of the thermal method, specifically when using normative units, as compared to previous matrices. The next- best technologies, although considerably less so, are those using KAc (Vehicle) and KAc (FAST). The former of these two has the virtue of possessing the highest POSI score, thus pre- senting less of an administrative hurdle. Yet, the difference in outcome with the thermal method might in itself be used as a means to perhaps lower some of these hurdles once the consid- erable value compared to other prospective and currently used approaches could be made clear. Figure B-12 shows the same results, this time using the natu- ral unit measures. Tradeoffs in Cost There are tradeoffs on the cost side as well. Table B-20 shows each technology’s ranking by each part of cost. Factors to Consider Source Overall Strategies Specific Chemicals - LOS - Cost - Infrastructure and environmental impacts - Equipment - Weather - Performance: ability to penetrate, undercut and break the bond between the ice and the pavement, or to prevent the ice–pavement bond from forming—eutectic and effective69 temperatures - Cost - Availability - Ease of use - Corrosion impacts - Environmental impacts - Health effects Rochelle, 2010 - Geographical location - Intensity of precipitation - Cost - Applicability: eutectic temperature—air and road surface temperature, concentrations of chemicals - Cost: availability of raw materials, process methods - Size of area - Geographic, economic, and environmental factors Zhang and Peterson, 2009 Table B-19. Factors to consider when deciding on appropriate technologies. 69 The lowest effective temperature for a deicer is defined as the temperature at which the deicer will melt a reasonable amount of ice within a reasonable amount of time. 70 The four individual scores for safety, preservation, sustainability, and mobility could be weighted differently if an agency desired to do so. 71 Expected value = the product of calculated total metric and POSI

67 While chemical dispersal from vehicles is preferred to oth- ers in terms of initial cost (e.g., equipment and installa- tion), the thermal method dominates the others in terms of hard-to-quantify costs (e.g., benefits from reduced travel time and crash). Interestingly, there is a totally reverse relationship between the agency-specific and road user/ social costs. This highlights the importance of making clear what areas of costs are being discussed. Only in this way are fully informed decisions about tradeoffs made possible. Overall Tradeoffs As a way of incorporating the anticipated outcome met- rics with the cost information, the research team reproduces prior figures while including notations for cost in each case. These figures show not only value metric and POSI but also cost information (see Figures B-13 and B-14). The interpretation of these curves must necessarily be based on local agency conditions. Based solely on assumptions and initial conditions given in the Compare phase, it seems that the two preferable approaches are the thermal method and KAc dispersed by vehicles. The thermal method scores well on the mission value measures while being of approximately the same order of magnitude as the other candidates in terms of cost and POSI. The KAc/vehicle approach, on the other hand, provides less but still improved mission value with a quarter of the cost of the thermal method. Yet, if either assumptions or initial conditions change even slightly, as they are bound to do once looking at agency-specific factors, the desirable tech- nology may well change accordingly. In particular, weather and location have critical impacts on which metric should be prioritized. Also, consideration of local (e.g., funding and traffic) will make the cost analysis more valuable to decision- makers. Thus, the ultimate decision rests with the individual transportation agencies themselves. 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 0 1 2 3 4 5 Su st ai na bi lit y Safety Safety - Sustainability Matrix CaCl (Vehicle) KAc (Vehicle) CaCl (FAST) KAc (FAST) Thermal Method 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 0 1 2 3 4 5 PO SI Safety Safety - POSI Matrix CaCl (Vehicle) KAc (Vehicle) CaCl (FAST) KAc (FAST) Thermal Method Figure B-9. Comparison of technology application bundles using normative units. 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 0 1 2 3 4 5 PO SI Sustainability Sustainability - POSI Matrix CaCl (Vehicle) KAc (Vehicle) CaCl (FAST) KAc (FAST) Thermal Method 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 0 1 2 3 4 5 Pr es er va ti on Sustainability Sustainability - Preservation Matrix CaCl (Vehicle) KAc (Vehicle) CaCl (FAST) KAc (FAST) Thermal Method

68 Figure B-10. Comparison of technology application bundles using natural units. 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 2.6 2.7 2.8 2.9 3 3.1 Su st ai na bi lit y Safety Safety - Sustainability Matrix CaCl (Vehicle) KAc (Vehicle) CaCl (FAST) KAc (FAST) Thermal Method 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 2.6 2.7 2.8 2.9 3 3.1 PO SI Safety Safety - POSI Matrix CaCl (Vehicle) KAc (Vehicle) CaCl (FAST) KAc (FAST) Thermal Method 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 0 1 2 3 4 5 PO SI Sustainability Sustainability - POSI Matrix CaCl (Vehicle) KAc (Vehicle) CaCl (FAST) KAc (FAST) Thermal Method 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 0 1 2 3 4 5 Pr es er va ti on Sustainability Sustainability - Preservation Matrix CaCl (Vehicle) KAc (Vehicle) CaCl (FAST) KAc (FAST) Thermal Method Figure B-11. Comparison of technology application bundles by integrating all summary metrics and POSI, using normative units. CaCl (Vehicle) KAc (Vehicle) CaCl (FAST) KAc (FAST) Thermal Method 0 2 4 6 8 10 12 14 16 18 0 1 2 3 4 Sa fe ty *M ob ili ty *P re se rv at io n* Su st ai na bi lit y Probability of Successful Implementation

69 Figure B-12. Comparison of technology application bundles by integrating all summary metrics and POSI, using natural units. 0 2 4 6 8 10 12 14 16 18 0 1 2 3 4S af et y* M ob ili ty *P re se rv at io n* Su st ai na bi lit y Probability of Successful Implementation CaCl (FAST) KAc (FAST) Thermal Method CaCl (Vehicle) KAc (Vehicle) Cost Part Chemical (Vehicle) Chemical (FAST) Thermal Method CaCl2 KAc CaCl2 KAc Easy-to- Quantify Initial Cost 1 1 2 2 3 Recurring Cost 3 5 1 4 2 Subtotal 1 2 3 4 5 Hard-to-Quantify 5 4 3 2 1 Both 3 2 5 4 1 Note: the numbers shown in the table are ordinal rankings with 1 being the best and 5 the worst among the five alternatives. Table B-20. Ranking of technologies by different categories of cost. Figure B-13. Comparison of technology application bundles by integrating all summary metrics and POSI, using normative units, and including notations on agency-specific net cost ($000s/bridge). CaCl (Vehicle) KAc (Vehicle) CaCl (FAST) KAc (FAST) Thermal Method 0 2 4 6 8 10 12 14 16 18 0 1 2 3 4 Sa fe ty *M ob ili ty *P re se rv at io n* Su st ai na bi lit y Probability of Successful Implementation -$832 -$741 -$207 -$681 -$147

70 Figure B-14. Comparison of technology application bundles by integrating all summary metrics and POSI, using natural units, and including notations on agency-specific net cost ($000s/bridge). CaCl (Vehicle) KAc (Vehicle) CaCl (FAST) KAc (FAST) Thermal Method 0 2 4 6 8 10 12 14 16 18 0 1 2 3 4 Sa fe ty *M ob ili ty *P re se rv at io n* Su sta in ab ili ty Probability of Successful Implementation -$832 -$741 -$207 -$681 -$147