Appendix B
Details of State Roadway Segment Selection

Luke E. Riexinger
Virginia Tech
Blacksburg, VA

Douglas J Gabauer
Bucknell University
Lewisburg, PA

September 30, 2023

List of Tables

Table 1. Roadway Types Considered in the HSM Predictive Methodology

Table 2. Roadway, Traffic and Roadside Characteristics Use in the Roadway Selection Process

Table 3. State Roadway Miles and Number of Routes by HSM Classification and State

Table 4. Available Routes (State Jurisdiction Only) and Total Mileage by State Agency Partner

Table 5. Available Washington State Rural Undivided Roadway Segments

Table 6. Operational Characteristics of the Top 20 Longest RU2L2W Routes in Washington State

Table 7. Geometric Characteristics of the Top 20 Longest RU2L2W Routes in Washington State

Table 8. Available Tennessee Rural Undivided Roadway Segments

Table 9. Operational Characteristics of the Top 20 Longest RU2L2W Routes in Tennessee

Table 10. Geometric Characteristics of the Top 20 Longest RU2L2W Routes in Tennessee

Table 11. Operational Characteristics of the Top 20 Longest RUMLH Routes in Tennessee

Table 12. Geometric Characteristics of the Top 20 Longest RUMLH Routes in Tennessee

Table 13. Available Iowa Rural Undivided Roadway Segments

Table 14. Operational Characteristics of the Top 20 Longest RU2L2W Routes in Iowa

Table 15. Geometric Characteristics of the Top 20 Longest RU2L2W Routes in Iowa

Table 16. Operational Characteristics of Iowa RUMLH Route

Table 17. Geometric Characteristics of Iowa RUMLH Route

Table 18. Selected RU2L2W Routes and Associated Characteristics

Table 19. Selected RUMLH Routes and Associated Characteristics

Table 20. Available Washington State Urban/Suburban Undivided Roadway Segments

Table 21. Operational Characteristics of the Top 20 Longest 2U/3T Routes in Washington State

Table 22. Geometric Characteristics of the Top 20 Longest 2U/3T Routes in Washington State

Table 23. Operational Characteristics of the Top 20 Longest 4U/5T Routes in Washington State

Table 24. Geometric Characteristics of the Top 20 Longest 4U/5T Routes in Washington State

Table 25. Available Tennessee Urban/Suburban Undivided Roadway Segments

Table 26. Operational Characteristics of the Top 20 Longest 2U/3T Routes in Tennessee

Table 27. Geometric Characteristics of the Top 20 Longest 2U/3T Routes in Tennessee

Table 28. Operational Characteristics of the Top 20 Longest 4U/5T Routes in Tennessee

Table 29. Geometric Characteristics of the Top 20 Longest 4U/5T Routes in Tennessee

Table 30. Available Iowa State Urban/Suburban Undivided Roadway Segments

Table 31. Operational Characteristics of the Top 20 Longest 2U/3T Routes in Iowa

Table 32. Geometric Characteristics of the Top 20 Longest 2U/3T Routes in Iowa

Table 33. Operational Characteristics of the Top 20 Longest 4U/5T Routes in Iowa

Table 34. Geometric Characteristics of the Top 20 Longest 4U/5T Routes in Iowa

Table 35. Selected 2U/3T Routes and Associated Characteristics

Table 36. Selected 4U/5T Routes and Associated Characteristics

Table 37. Available Washington State Rural Divided Roadway Segments

Table 38. Operational Characteristics of the Rural Divided Multilane (R4D) Routes in Washington State

Table 39. Geometric Characteristics of the Rural Divided Multilane (R4D) Routes in Washington State

Table 40. Operational Characteristics of the Rural Divided Freeway (R4F) Routes in Washington State

Table 41. Geometric Characteristics of the Rural Divided Freeway (R4F) Routes in Washington State

Table 42. Operational Characteristics of the Rural Divided Freeway (R6F) Routes in Washington State

Table 43. Geometric Characteristics of the Rural Divided Freeway (R6F) Routes in Washington State

Table 44. Available Tennessee Rural Divided Roadway Segments

Table 45. Operational Characteristics of the Top 20 Rural Divided Multilane (R4D) Routes in Tennessee

Table 46. Geometric Characteristics of the Top 20 Rural Divided Multilane (R4D) Routes in Tennessee

Table 47. Operational Characteristics of the Rural Divided Freeway (R4F) Routes in Tennessee

Table 48. Geometric Characteristics of the Rural Divided Freeway (R4F) Routes in Tennessee

Table 49. Operational Characteristics of the Rural Divided Freeway (R6F) Routes in Tennessee

Table 50. Geometric Characteristics of the Rural Divided Freeway (R6F) Routes in Tennessee

Table 51. Available Iowa Rural Divided Roadway Segments

Table 52. Operational Characteristics of the Top 20 Rural Divided Multilane (R4D) Routes in Iowa

Table 53. Geometric Characteristics of the Top 20 Rural Divided Multilane (R4D) Routes in Iowa

Table 54. Operational Characteristics of the Rural Divided Freeway (R4F) Routes in Iowa

Table 55. Geometric Characteristics of the Rural Divided Freeway (R4F) Routes in Iowa

Table 56. Operational Characteristics of the Rural Divided Freeway (R6F) Routes in Iowa

Table 57. Geometric Characteristics of the Rural Divided Freeway (R6F) Routes in Iowa

Table 58. Selected R4D Routes and Associated Characteristics

Table 59. Selected R4F Routes and Associated Characteristics

Table 60. Selected R6F Routes and Associated Characteristics

Table 61. Available Washington State Urban/Suburban Divided Roadway Segments

Table 62. Operational Characteristics of the Top 20 Urban/Suburban Divided Arterial (U4D) Routes in Washington State

Table 63. Geometric Characteristics of the Top 20 Urban/Suburban Divided Arterial (U4D) Routes in Washington State

Table 64. Operational Characteristics of the Top 20 Urban/Suburban Divided Freeway (U4F) Routes in Washington State

Table 65. Geometric Characteristics of the Top 20 Urban/Suburban Divided Freeway (U4F) Routes in Washington State

Table 66. Operational Characteristics of the available Urban/Suburban Divided Freeway (U6F) Routes in Washington State

Table 67. Geometric Characteristics of the available Urban/Suburban Divided Freeway (U6F) Routes in Washington State

Table 68. Operational Characteristics of the available Urban/Suburban Divided Freeway (U8F) Routes in Washington State

Table 69. Geometric Characteristics of the Available Urban/Suburban Divided Freeway (U8F) Routes in Washington State

Table 70. Available Tennessee Urban/Suburban Divided Roadway Segments

Table 71. Operational Characteristics of the Top 20 Urban/Suburban Divided Arterial (U4D) Routes in Tennessee

Table 72. Geometric Characteristics of the Top 20 Urban/Suburban Divided Arterial (U4D) Routes in Tennessee

Table 73. Operational Characteristics of the Top 20 Urban/Suburban Divided Freeway (U4F) Routes in Tennessee

Table 74. Geometric Characteristics of the Top 20 Urban/Suburban Divided Freeway (U4F) Routes in Tennessee

Table 75. Operational Characteristics of the Top 20 Urban/Suburban Divided Freeway (U6F) Routes in Tennessee

Table 76. Geometric Characteristics of the Top 20 Urban/Suburban Divided Freeway (U6F) Routes in Tennessee

Table 77. Operational Characteristics of the Urban/Suburban Divided Freeway (U8F) Routes in Tennessee

Table 78. Geometric Characteristics of the Urban/Suburban Divided Freeway (U8F) Routes in Tennessee

Table 79. Operational Characteristics of the Urban/Suburban Divided Freeway (U10F) Routes in Tennessee

Table 80. Geometric Characteristics of the Urban/Suburban Divided Freeway (U10F) Routes in Tennessee

Table 81. Available Iowa Urban/Suburban Divided Roadway Segments

Table 82. Operational Characteristics of the Top 20 Urban/Suburban Divided Multilane (U4D) Routes in Iowa

Table 83. Geometric Characteristics of the Top 20 Urban/Suburban Divided Multilane (U4D) Routes in Iowa

Table 84. Operational Characteristics of the Top 20 Urban/Suburban Divided Freeway (U4F) Routes in Iowa

Table 85. Geometric Characteristics of the Top 20 Urban/Suburban Divided Freeway (U4F) Routes in Iowa

Table 86. Operational Characteristics of the Top 20 Urban/Suburban Divided Freeway (U6F) Routes in Iowa

Table 87. Geometric Characteristics of the Top 20 Urban/Suburban Divided Freeway (U6F) Routes in Iowa

Table 88. Operational Characteristics of the Urban/Suburban Divided Freeway (U8F) Routes in Iowa

Table 89. Geometric Characteristics of the Urban/Suburban Divided Freeway (U8F) Routes in Iowa

Table 90. Operational Characteristics of the Urban/Suburban Divided Freeway (U10F) Routes in Iowa

Table 91. Geometric Characteristics of the Urban/Suburban Divided Freeway (U10F) Routes in Iowa

Table 92. Selected U4D Routes and Associated Characteristics

Table 93. Selected U4F Routes and Associated Characteristics

Table 94. Selected U6F Routes and Associated Characteristics

Table 95. Selected U8F Routes and Associated Characteristics

Table 96. Selected U10F Routes and Associated Characteristics

List of Figures

Figure 1. Graphical Depiction of Rural Undivided Roadway Classification Scheme

Figure 2. Top 20 Longest Rural 2-Lane, 2-Way Undivided (RU2L2W) Routes in Washington State

Figure 3. Top 20 Longest Rural 2-Lane, 2-Way Undivided (RU2L2W) Routes in Tennessee

Figure 4. Top 20 Longest Rural Multilane Highway (RUMLH) Routes in Tennessee

Figure 5. Top 20 Longest Rural 2-Lane, 2-Way Undivided (RU2L2W) Routes in Iowa

Figure 6. Top 20 Longest Urban/Suburban Undivided Routes (2U/3T) in Washington State

Figure 7. Top 20 Longest Urban/Suburban Undivided Routes (4U/5T) in Washington State

Figure 8. Top 20 Longest Urban/Suburban Undivided Routes (2U/3T) in Tennessee

Figure 9. Top 20 Longest Urban/Suburban Undivided Routes (4U/5T) in Tennessee

Figure 10. Top 20 Longest Urban/Suburban Undivided Routes (2U/3T) in Iowa

Figure 11. Top 20 Longest Urban/Suburban Undivided Routes (4U/5T) in Iowa

Figure 12. 4-Lane Rural Divided Multilane (R4D) Routes in Washington State

Figure 13. 4-Lane Rural Divided Freeway (R4F) Routes in Washington State

Figure 14. 6-Lane Rural Divided Freeway (R6F) Routes in Washington State

Figure 15. Top 20 4-Lane Rural Divided Multilane (R4D) Routes in Tennessee

Figure 16. 4-Lane Rural Divided Freeway (R4F) Routes in Tennessee

Figure 17. 6-Lane Rural Divided Freeway (R6F) Routes in Tennessee

Figure 18. Top 20 4-Lane Rural Divided Multilane (R4D) Routes in Iowa

Figure 19. 4-Lane Rural Divided Freeway (R4F) Routes in Iowa

Figure 20. 6-Lane Rural Divided Freeway (R6F) Routes in Iowa

Figure 21. Top 20 4-Lane Urban/Suburban Divided Arterial (U4D) Routes in Washington State

Figure 22. Top 20 4-Lane Urban/Suburban Divided Freeway (U4F) Routes in Washington State

Figure 23. 6-Lane Urban/Suburban Divided Freeway (U6F) Routes in Washington State

Figure 24. 8-Lane Urban/Suburban Divided Freeway (U8F) Routes in Washington State

Figure 25. Top 20 4-Lane Urban/Suburban Divided Arterial (U4D) Routes in Tennessee

Figure 26. Top 20 4-Lane Urban/Suburban Divided Freeway (U4F) Routes in Tennessee

Figure 27. Top 20 6-Lane Urban/Suburban Divided Freeway (U6F) Routes in Tennessee

Figure 28. 8-Lane Urban/Suburban Divided Freeway (U8F) Routes in Tennessee

Figure 29. 10-Lane Urban/Suburban Divided Freeway (U10F) Routes in Tennessee

Figure 30. Top 20 4-Lane Urban/Suburban Divided Multilane (U4D) Routes in Iowa

Figure 31. Top 20 4-Lane Urban/Suburban Divided Freeway (U4F) Routes in Iowa

Figure 32. Top 20 6-Lane Urban/Suburban Divided Freeway (U6F) Routes in Iowa

Figure 33. 8-Lane Urban/Suburban Divided Freeway (U8F) Routes in Iowa

Figure 34. 10-Lane Urban/Suburban Divided Freeway (U10F) Routes in Iowa

Introduction

The NCHRP 17-88 encroachment database includes data on individual encroachments, including reported and unreported encroachments/crashes determined from available state agency partner crash and maintenance data. The approach was to analyze a representative sample of roadway sections from each of the three state agency partners.

The purpose of this document is to summarize the selection of representative roadway segments for inclusion in the NCHRP 17-88 encroachment database. The representative segments were selected from each of the three state agency partners, Washington State (WA), Iowa (IA), and Tennessee (TN).

State Roadway Segment Selection Process Summary

The research team selected representative roadway segments for each state agency partner using the following procedure:

1. Categorize all inventoried state-jurisdiction roadway segments based on the Highway Safety Manual (HSM; American Association of State Highway and Transportation Officials [AASHTO, 2010]) classification scheme presented in Table 1—e.g. using land use and roadway configuration (primary variables listed in Table 2). Determine the count of different routes and associated mileage of roadways that fall into each possible category (excluding ramps). A single route designation may span multiple classifications, so the mileage associated with each classification will provide additional context for the selection process.

Table 1. Roadway Types Considered in the HSM Predictive Methodology

2. Assess the ranges of secondary variables (see Table 2)—e.g. all other widely available variables not used to determine HSM classification—for the longest mileage routes in each category. Tertiary variables will not be used in the initial selection process, as they are not readily available across all state agency partners, and in some cases, only have limited availability.

Using the available roadway information from each state, the rural undivided roadways were selected and further classified into the appropriate subcategories (listed in Table 1). The two rural undivided roadway segment subtypes are:

1. 2-lane, 2-way roadways (RU2L2W), and
2. Multilane highways (RUMLH).

For each rural undivided subcategory, the range of roadway and traffic characteristics—from the secondary variables shown in Table 2—were then examined for the routes with the longest subcategory-classified length in each of the three states. Representative rural undivided routes were then selected considering the associated roadway and traffic characteristics across all three states.

The available rural undivided roadway segments were first classified into the two subtype categories based on the information provided in HSM Chapter 10.3.1 (RU2L2W) and Chapter 11.3.1 (RUMLH; AASHTO, 2010). Note that ramps will be handled separately in the selection process. Although RU2L2W rural roadways generally have 2 lanes, the HSM indicates that these roadways include segments with one passing lane or a center two-way left turn lane (i.e. 3 lanes total) or short 4 lane segments that provide additional passing opportunities (i.e. side-by-side passing lanes). No maximum length is provided by the HSM to determine if a 4-lane rural undivided roadway section should be classified as a RU2L2W instead of a RUMLH. As the Highway Capacity Manual (HCM) specifies the upper bound for optimal passing lane length to be 2 miles (TRB 2016), 4-lane segments that are 2 miles or shorter were classified as RU2L2W and 4-lane segments that are longer than 2 miles were classified as RUMLH. This classification scheme is depicted graphically in Figure 1.

Washington State Rural Undivided Roadways and Associated Characteristics

Table 5 summarizes the classification results of the available rural undivided roadways in Washington State. The vast majority of the available sections are RU2L2W with only two routes

Table 6 and Table 7 present information on the secondary variables (see Table 2) for the 20 routes shown in Figure 2. Table 6 summarizes the operational characteristics while Table 7 summarizes the geometric characteristics. For each route, the range of average annual daily traffic (AADT), the range of large truck percentage, and the range of posted speed limit are shown in Table 6. Note that for state highways with no posted speed limit (i.e. missing posted speed in Table 6) Washington State legislation specifies a maximum speed of 60 mph (Revised Code of Washington [RCW], 46.61.400).

The operational and geometric characteristics of these 20 routes are shown in Table 9 and Table 10, respectively. For each route, the range of AADT, large truck percentage, and posted speed limit are shown in Table 9. For the geometric characteristics, minimum and mean lane width, minimum, minimum, mean and maximum shoulder width, and proportion of roadway length with shoulder rumble strips are shown in Table 10. For undivided roadways, a single outside shoulder width is reported in the Tennessee roadway data; this width was used to determine the shoulder width values reported in Table 10. Unlike Washington State, the minimum shoulder width for the longest 20 RU2L2W roadways exceeded zero and was typically either 1 or 2 feet wide. Note that Tennessee only reports shoulder width to the nearest foot. The minimum and mean lane widths are reported based on a combination of the pavement width data element (travel lanes only) and number of lanes data in the roadway description table and road geometry tables, respectively. The Tennessee maintenance inventory table includes shoulder rumble strip data, which was used to generate the rumble strip presence proportions shown in Table 10. These values reflect the proportion of RU2L2W length for each route that has shoulder rumble strips present. This was computed by dividing total shoulder rumble strip length for each route by the corresponding total shoulder edge length—twice the RU2L2W length reported in Figure 3.

These observations, along with the available Tennessee rural undivided roadway characteristics, will be combined with corresponding data from the other two state agency partners. This combined data will be used to select representative RU2L2W and RUMLH routes from all three state agency partners.

Iowa Rural Undivided Roadways and Associated Characteristics

Similar to Washington State and Tennessee, Iowa rural undivided roadways (state jurisdiction) were examined in more detail. Table 13 summarizes the results of the classification of the available rural undivided roadways in Iowa. The vast majority of the available sections are RU2L2W and only a single route was classified as RUMLH.

Table 13. Available Iowa Rural Undivided Roadway Segments

Figure 5 summarizes the top 20 RU2L2W roadways in Iowa based on total RU2L2W length available. The roadway inventory data was processed similar to Washington State and Tennessee, combining adjacent segments with the same RU2L2W classification to determine the number of contiguous segments for each route. The maximum contiguous length for each route was tallied and the average contiguous length was computed by dividing the total RU2L2W length by the number of contiguous segments for each route. For the other two states, segments were only combined if the ending milepost of the preceding segment exactly matched the beginning milepost of the next segment. The milepost designations in Iowa, however, were reported to a precision of up to six decimal places. An initial processing of the data with exact milepost matching resulted in an unusually high number of segments as a result of very small discrepancies in a number of milepost values: 0.00001 miles or less, which equates to approximately 0.6 inches. For the Iowa data, the mileposts were considered a match if the absolute value of the difference was 0.00001 miles or less. The data shown in Figure 5 uses this slightly less stringent convention. All of the routes shown in Figure 5 have a total RU2L2W length in excess of 100 miles. Note that approximately 23% (26 of 111 routes) have 10 miles or less total RU2L2W length available (not shown in Figure 5).

The operational and geometric characteristics of these 20 routes are shown in Table 14 and Table 15, respectively. For each route, the range of AADT, large truck percentage, and posted speed limit are shown in Table 14. For the geometric characteristics, minimum and mean lane width, mean and maximum shoulder width, and proportion of roadway length with shoulder rumble strips are shown in Table 15. Similar to Washington State, both left and right shoulder data available for the RU2L2W roadways and are nearly identical. Table 15 combines both left and right shoulder width variables. Similar to Washington State, the minimum shoulder width for the longest 20 RU2L2W roadways was zero, so a minimum shoulder width column was excluded from the table. The minimum and mean lane widths are reported based on a combination of the pavement width information and number of lanes present in the roadway table. The Iowa roadway data includes both right and left shoulder rumble strip data. The approximate rumble strip presence proportions in Table 15 do reflect the proportion of RU2L2W length for each route that has shoulder rumble strips present. This percentage was computed by dividing the total shoulder rumble strip length for each route by the corresponding total shoulder edge length available—twice the RU2L2W length reported in Figure 5 for the same route.

IA 4 was the only RUMLH roadway identified in Iowa. The portion of IA 4 classified as RUMLH consisted of a single 20.8 mile contiguous segment. The operational and geometric characteristics for IA 4 are summarized in Table 16 and Table 17, respectively. Although the traffic volume is relatively low, there is reasonable variation in posted speed, modest truck traffic, some variation in shoulder width and no shoulder rumble strips present.

Based on the data from the top 20 RU2L2W segments in Iowa, the following observations were made:

• The routes with the largest traffic volume variation are SR 30 and SR 6. The maximum traffic volumes are less than the available routes in Washington State and Tennessee, with all but three of the longest routes having AADT in excess of 10,000 vehicles per day.
• Similar to Tennessee, Iowa routes have a wider variation in posted speed limit compared to those in Washington State.
• Similar to Washington State, most of the longest RU2L2W routes have maximum truck percentages in excess of 25%.
• The variation in shoulder width on the longest RU2L2W routes is less than observed in Washington State and Tennessee.
• Compared to Washington, Iowa has several higher volume RU2L2W routes with smaller percentages of rumble strips: 25% or less.

These observations, along with the available Iowa rural undivided roadway characteristics, will be combined with corresponding data from the other two state agency partners. This combined data will be used to select representative RU2L2W and RUMLH routes from all three state agency partners.

Rural Undivided Roadway Selection

RU2L2W Roadway Selection

The available RU2L2W roadway characteristics (secondary variables from Table 2) from each state agency partner combined with the general observations were considered to select

Urban/Suburban Undivided Roadway Characteristics

The urban/suburban undivided roadways, or arterials, are examined in more detail. The four subtypes of urban undivided arterials identified in the HSM (Chapter 12.3.1; AASHTO, 2010) are:

1. 2-lane undivided arterials (2U)
2. 3-lane undivided arterials with center TWLTL (3T)
3. 4-lane undivided arterials (4U)
4. 5-lane undivided arterials with center TWLTL (5T)

As not all states have readily accessible two-way left turn lane (TWLTL) presence data, the four subcategories were merged into the following two categories for the purpose of roadway selection:

1. 2-lane undivided arterials, with or without TWLTL (2U/3T)
2. 4-lane undivided arterials, with or without TWLTL (4U/5T)

For states where information is available on TWLTL presence, this was considered as an additional data element in the selection process. For states where this information is not readily available, TWLTL presence will be determined for each selected route after the roadway selection process. Using the available roadway information from each state, the urban/suburban undivided roadways were selected and further classified into the two subcategories. For each urban/suburban undivided subcategory, the range of roadway and traffic characteristics—from the secondary variables shown in Table 2—were then examined for the routes with the longest subcategory-classified length in each of the three states. Representative urban/suburban undivided routes were then selected considering the associated roadway and traffic characteristics across all three states. Similar to the rural undivided roadways, ramps will be handled separately in the selection process.

Washington Urban/Suburban Undivided Roadways and Associated Characteristics

Table 20 summarizes the results of the available urban/suburban classification of the undivided roadways in Washington State. Note that Washington State does not have TWLTL data readily available. The majority of the available sections are 2U/3T but there are a reasonable number of 4U/5T sections present.

Table 20. Available Washington State Urban/Suburban Undivided Roadway Segments

Figure 6 summarizes the top 20 2U/3T roadways in Washington State based on total 2U/3T length available. Similar to the rural undivided roadways, the available segment data was processed to combine adjacent 2U/3T segments from the same route to determine the number of contiguous segments for each route. The number of contiguous 2U/3T segments for a given route

For the geometric characteristics, minimum and mean lane width, mean and maximum shoulder width, and approximate proportion of roadway length with rumble strips are shown in Table 22. As Washington State roadway inventory data captures both directions of a roadway in a single entry, right and left shoulder width is available for undivided roadways. For the 2U/3T roadways, the left and right shoulder data is nearly identical. The shoulder width data shown in Table 22 considers both left and right shoulder width variables. The minimum shoulder width for all of the routes in Table 22 was zero, so a minimum shoulder width column was not included in the table. The minimum and mean lane widths are reported. Rumble strip presence data is not available in the HSIS roadway inventory data, but Washington State maintains this data separately. For each of the top 20 routes, the proportion of route length with any rumble strips present was computed using this separate dataset. Note that this data is based on the entire length of the roadway and is not necessarily specific to the 2U/3T portions of the route. Nonetheless, the data provides a general indication of whether or not any rumble strips are present along a particular route. The available rumble strip data (not included in Table 22) also indicates the location of the rumble strips: centerline only, shoulders only, or both centerline and shoulder locations. Once specific routes have been selected, the exact details on the rumble strips will be merged into the corresponding HSIS roadway inventory data.

Based on the data from the top 20 2U/3T segments in Washington State, the following observations were made:

• The routes with the largest traffic volume variation are SR 202 followed by SR 305 and SR 9. All have a maximum daily traffic volume in excess of 25,000 vehicles per day or more.
• As expected, the 2U/3T routes generally have a higher minimum and maximum traffic volume than the RU2L2W routes in Washington State.
• Most routes have little variation in posted speed limit. A notable exception is SR 12.
• The 2U/3T routes have maximum truck traffic percentages of 25% or lower.
• Minimum and mean lane widths are generally larger than those observed for the Washington RU2L2W routes and appear less influenced by traffic volume.
• SR 202, SR 12, and SR 169 have the largest variation in shoulder width.
• Approximately half of the routes have more than 50% presence of rumble strips.
• SR 2 was already selected to be included in the study, as the RU2L2W route from Washington so the analogous 2U/3T sections could be included as well.

Based on the data from the top 20 4U/5T segments in Washington State, the following observations were made:

• The routes with the largest traffic volume variation are SR 524 followed by SR 900 and SR 161. All have a maximum daily traffic volume in excess of 40,000 vehicles per day or more.
• As expected, the 4U/5T routes generally have a higher minimum and maximum traffic volume than the 2U/3T routes in Washington State.
• Most routes have little variation in posted speed limit. Notable exceptions are SR 2 and SR 101.
• The 4U/5T routes have maximum truck traffic percentages of about 20% or lower.
• Minimum and mean lane widths are generally larger than those observed for the Washington RU2L2W routes and appear less influenced by traffic volume.
• SR 12 and SR 527 have the largest variation in shoulder width.
• The majority of the routes have less than 50% presence of rumble strips.
• SR 2 was already selected to be included in the study as the RU2L2W route from Washington, so the analogous 4U/5T sections could be included as well.

These observations along with the available Washington urban/suburban undivided roadway characteristics will be combined with corresponding data from the other two state agency partners. This combined data will be used to select representative 2U/3T and 4U/5T routes from all three state agency partners.

Tennessee Urban/Suburban Undivided Roadways and Associated Characteristics

The available Tennessee urban/suburban undivided roadway segments were classified as 2U/3T or 4U/5T. The Tennessee roadway data was split among several different data tables, each with potentially disparate roadway segment boundaries, as opposed to Washington State which has the majority of the basic roadway data integrated into a single data table. While this initially resulted in a more complicated and time-consuming process to combine the roadway data across these tables, the overall classification process was essentially the same between the states once the Tennessee road data was combined.

Table 25 summarizes the results of the classification of the available urban/suburban undivided roadways in Tennessee. Similar to Washington State, the majority of the available sections are 2U/3T but there are a reasonable number of 4U/5T routes available. TWLTL data is available in the Tennessee data in the road description table.

Table 25. Available Tennessee Urban/Suburban Undivided Roadway Segments

Figure 8 summarizes the top 20 2U/3T roadways in Tennessee based on total 2U/3T length available. The roadway inventory data was processed similar to Washington State, combining adjacent segments with the same 2U/3T classification to determine the number of contiguous segments for each route. One important difference in the Tennessee data is that the route mileage resets to zero at every instance the route enters a new county and the data in Figure 8 has not been adjusted accordingly. As a result, each route will generally have a larger number of contiguous segments compared to those in Washington State. Note that approximately 35% (91 of 260 routes) of segments have 3 miles or less total 2U/3T length available (not shown in Figure 8).

The operational and geometric characteristics of these 20 routes are shown in Table 26 and Table 27, respectively. For each route, the range of AADT, large truck percentage, and posted speed limit are shown in Table 26.

Based on the data from the top 20 4U/5T segments in Tennessee, the following observations were made:

• The routes with the largest traffic volume variation are SR 15 followed by SR 10 and SR 71. All have a maximum daily traffic volume in excess of 45,000 vehicles per day or more.
• As expected, the 4U/5T routes generally have a higher minimum and maximum traffic volume than the 2U/3T routes in Tennessee.
• Similar to the 2U/3T routes, most 4U/5T routes have a range of posted speed limit between 30 and 55 mph.
• The majority of the longest 4U/5T routes have maximum truck traffic percentages of 20% or higher.
• Mean lane widths are generally the same or slightly narrower than those observed for the Tennessee 2U/3T routes.
• SR 1 and SR 35 have the largest variation in shoulder width.
• The majority of the routes have less than 30% presence of rumble strips.
• The majority of the 4U/5T roadways have a TWLTL present and, in general, the length is more than one fourth of the total 4U/5T length.
• SR 33 and SR 15 were already selected to be included in the study as representative RUMLH routes. These routes are in the top 20 longest 4U/5T routes so the analogous 4U/5T sections could be included.

These observations along with the available Tennessee urban/suburban undivided roadway characteristics will be combined with corresponding data from the other two state agency partners. This combined data will be used to select representative 2U/3T and 4U/5T routes from all three state agency partners.

Iowa Urban/Suburban Undivided Roadways and Associated Characteristics

Table 30 summarizes the results of the classification of the available urban/suburban undivided roadways in Iowa. The majority of the available sections are 2U/3T but there are also a significant number of 4U/5T sections present. Note that Iowa roadway data has lane type information present to allow for identification of sections with TWLTL.

Table 30. Available Iowa State Urban/Suburban Undivided Roadway Segments

Figure 10 summarizes the top 20 2U/3T roadways in Iowa based on total 2U/3T length available. The available segment data was processed to combine adjacent 2U/3T segments from the same route to determine the number of contiguous segments for each route. The maximum contiguous length for each route was tallied and the average contiguous length computed by dividing the total 2U/3T length by the number of contiguous segments for each route. Due to the high

For the geometric characteristics, minimum and mean lane width; mean and maximum shoulder width; and proportion of roadway length with shoulder rumble strips or a WLTL are shown in Table 34. Similar to Washington State, both left and right shoulder data available for the 4U/5T roadways are nearly identical. Table 34 combines both left and right shoulder width variables. Similar to Washington State, the minimum shoulder width for the longest 20 4U/5T roadways was zero, so a minimum shoulder width column was excluded from the table. The minimum and mean lane widths are reported based on a combination of the pavement width information and number of lanes present in the roadway table. The Iowa roadway data includes both right and left shoulder rumble strip data. The approximate rumble strip presence proportions in Table 34 do reflect the proportion of 4U/5T length for each route that has shoulder rumble strips present. This percentage was computed by dividing the total shoulder rumble strip length for each route by the corresponding total shoulder edge length available—e.g. twice the 4U/5T length reported in Figure 11 for the same route. For the three-lane sections identified, the lane type variable available data was used to determine the sections with a TWLTL present—e.g. the lane type of the second lane equal to “5-center turn lane.” The TWLTL percentage was computed by dividing the length of roadway with a TWLTL present by the total 4U/5T length for the same route, as reported in Figure 11.

Based on the data from the top 20 2U/3T segments in Iowa, the following observations were made:

• The routes with the largest traffic volume variation are SR 69 followed by SR 6. Both have a maximum daily traffic volume in excess of 20,000 vehicles per day or more.
• As expected, the 2U/3T routes generally have a higher minimum and maximum traffic volume than the RU2L2W routes in Iowa.
• Most routes have a relatively wide variation in posted speed limit.
• The majority of the 2U/3T routes have maximum truck traffic percentages of 25% or lower.
• Minimum and mean lane widths are generally larger than those observed for the Iowa RU2L2W routes.
• SR 3, SR 6, and SR 65 have the largest variation in shoulder width.
• Most of the routes have less than 25% presence of rumble strips.
• Less than half of the 2U/3T roadways have a TWLTL present and, in general, the length is less than one fourth of the total 2U/3T length.
• SR 30 was already selected to be included in the study as the RU2L2W route from Washington, so the analogous 2U/3T sections could be included as well.

Based on the data from the top 20 4U/5T segments in Iowa, the following observations were made:

• The routes with the largest traffic volume variation are SR 6 followed by SR 461 and SR 67. All have a maximum daily traffic volume in excess of 20,000 vehicles per day or more.
• As expected, the 4U/5T routes generally have a higher minimum and maximum traffic volume than the 2U/3T routes in Iowa.
• Most routes have a relatively wide variation in posted speed limit.
• The 4U/5T routes have maximum truck traffic percentages of about 20% or lower.
• Minimum and mean lane widths are generally larger than those observed for the Iowa RU2L2W routes.
• The mean shoulder width varies from 0 to approximately 7 feet but the maximum shoulder width is essentially the same for all of the 4U/5T routes.
• All of the routes have less than 35% presence of rumble strips.
• Half of the 4U/5T roadways have a TWLTL present and, in general, the length is less than half of the total 4U/5T length.
• SR 30 was already selected to be included in the study as the RU2L2W route from Iowa, so the analogous 4U/5T sections could be included as well.

These observations along with the available Iowa urban/suburban undivided roadway characteristics will be combined with corresponding data from the other two state agency partners. This combined data will be used to select representative 2U/3T and 4U/5T routes from all three state agency partners.

Urban/Suburban Undivided Roadway Selection

2-lane undivided arterials (2U/3T)

The available 2U/3T roadway characteristics (secondary variables from Table 2) from each state agency partner combined with the general observations were considered to select representative 2U/3T routes from each state. The selected 2U/3T routes are shown in Table 35. Note that the routes denoted with an asterisk (*) indicate that particular route had portions classified as 2U/3T but was selected to represent a different roadway type category.

Due to the relatively short lengths of 4U/5T roadways in Washington (only four routes had lengths in excess of 10 miles), two routes were selected: SR 99 and SR 522. SR 99 has similar traffic volume and posted speed limit variation to SR 2 (which was already selected through the rural undivided selection process) but has no rumble strips present. SR 522 has a higher maximum traffic volume as well as a lower posted speed limit, a smaller proportion of trucks, and narrower shoulders. Washington State SR 2 was previously selected as a representative RU2L2W roadway; the associated 4U/5T characteristics of SR 2 are also listed in Table 36. SR 2 was the second longest 4U/5T route in Washington State.

Three (SR 1, SR 15 and SR 33) previously selected Tennessee routes were also in the top 20 4U/5T routes. Collectively, these routes provide a relatively wide range of 4U/5T characteristics. SR 1 was the longest 4U/5T route and SR 15 has the largest traffic variation of the 4U/5T routes. All three routes have reasonable variation in posted speed limit and reasonable TWLTL lengths (> 70 miles in total between the three routes). Given this, no additional Tennessee 4U/5T routes were selected.

Two previously selected Iowa routes, SR 6 (2U/3T representative roadway) and SR 30 (RU2L2W representative roadway), were also in the top 20 Iowa 4U/5T routes. Similar to Washington State, the 4U/5T roadway lengths were relatively short with only three Iowa 4U/5T routes in excess of 10 miles. Given this, SR 18 was selected as one additional representative 4U/5T route. SR 18 has the third longest 4U/5T length, provides a wide range of posted speed limit, and has some rumble strips present (SR 6 and SR 30 have no rumble strips). In general, the Iowa 4U/5T routes have few TWLTLs but the combination of SR 6 and SR 30 provide nearly 3 miles of TWLTL presence.

Rural Divided Roadway Characteristics

The rural divided roadways were examined in more detail. The four subtypes of rural divided roadway segments identified in the HSM (AASHTO, 2010) are listed below. Note that the HSM does not currently include freeway segments; i.e., subtypes 2, 3, and 4. Safety performance functions for these roadway types were developed under NCHRP Project 17-45 (Bonneson et al., 2012).

1. Rural 4-lane divided multilane highway (R4D)
2. Rural 4-lane divided freeway (R4F)
3. Rural 6-lane divided freeway (R6F)
4. Rural 8-lane divided freeway (R8F)

In addition to the rural classification and number of through lanes, all of the freeway segment subtypes have fully-restricted access control—R4F, R6F, and R8F—while the R4D subtype does not. The roadway data from each state was used to determine the level of access control for each section to determine the rural divided roadway subtype. While the access control data for each state varied slightly, each state had a category identifying segments with full access control and only these segments were considered for the R4F, R6F, and R8F subtypes. The R4D routes could have any other access level control other than full access control.

For each rural divided subcategory, the range of roadway and traffic characteristics from the secondary variables shown in Table 2 were then examined for the routes with the longest subcategory-classified length in each of the three states. Representative rural divided routes were then selected considering the associated roadway and traffic characteristics across all three states.

Washington Rural Divided Roadways and Associated Characteristics

Table 37 summarizes the results of the classification of the available rural divided roadways in Washington State. The vast majority of the available sections are R4D and R4F with only two unique routes having portions classified as R6F and one route with a small section classified as R8F. Given the lack of R8F segment length available in Washington State, the research team proposes selecting an additional R8F route from Tennessee or Iowa, depending on availability.

Table 37. Available Washington State Rural Divided Roadway Segments

As all rural divided subtypes in Washington have less than 20 unique routes available, all of the available routes were examined for the R4D, R4F, and R6F subtypes. Figure 12 summarizes the available R4D roadways in Washington State ranked by total R4D length available. The available segment data was processed to combine adjacent R4D segments from the same route to determine the number of contiguous segments for each route. The number of contiguous R4D segments for a given route is the left number shown in the Figure 12 data label. The maximum contiguous length for each route was tallied and the average contiguous length was computed by dividing the total R4D length by the number of contiguous segments for each route; these are the middle and right numbers in the Figure 12 data label, respectively.

Table 38 and Table 39 present information on the secondary variables (see Table 2) for 9 of the 11 R4D routes shown in Figure 12. Since routes 501 and 9 had only a single segment less than a tenth of a mile long classified as R4D, these two routes were excluded from further consideration. Table 38 summarizes the operational characteristics while Table 39 summarizes the geometric characteristics. For each route, the range of AADT, the range of large truck percentage, and the range of posted speed limit are shown in Table 38. Note that for state highways with no posted speed limit Washington State legislation specifies a maximum speed of 60 mph (RCW, 46.61.400).

Geometric characteristics shown in Table 39 include lane width, outside shoulder width, median width, as well as percentage of positive median barrier and rumble strips present. The minimum and mean lane widths are reported. The shoulder width data in Table 39 considers the outside shoulder width for both directions of the divided roadway. In general, however, the outside shoulder widths were nearly identical in most cases. All routes had a minimum shoulder width of zero (not shown in Table 39) except for routes 12 and 395, which had a 4 foot and 10 foot minimum shoulder width, respectively. The minimum and maximum median widths are reported for each route and, in Washington State, this width includes the inside/left shoulder widths. Positive barrier presence was determined from the median barrier type variable. Median barrier types included cable, guardrail, and NJ type barriers; only the depressed, unprotected, and curb categories were considered to have no positive barrier present. For each route, the total length of sections with positive barrier were divided by the corresponding total R4D length for that route, as shown in Figure 12. Rumble strip presence data is not available in the HSIS roadway inventory data but Washington State maintains this data separately. For each of the routes, the proportion of route length with any rumble strips present was computed using this separate dataset. Note that this data is based on the entire length of the roadway and is not necessarily specific to the R4D portions of the route. Nonetheless, the data provides a general indication of whether or not any rumble strips are present along a particular route. The available rumble strip data (not included in Table 39) also indicates the location of the rumble strips: centerline only, shoulders only, or both centerline and shoulder locations. Once specific routes have been selected, the exact details on the rumble strips will be merged into the corresponding HSIS roadway inventory data.