Supplemental Guidance on the Application of FHWA’s Traffic Noise Model (TNM) (2014)

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.

65 C H A P T E R 8 This chapter focuses on guidance on the use of topographic features within FHWA TNM. Appendix G (available on the NCHRP Project 25-34 web page at http://apps.trb.org/cmsfeed/ TRBNetProjectDisplay.asp?ProjectID=2986) provides sub- stantial detail on the methods and test cases that were used to develop the guidance. 8.1 Outside Edge of Pavement: Horizontal Precision 8.1.1 Guidance for TNM Input The research team suggests the following best practices for entering the near edge of pavement (or “equivalent” terrain line) into TNM: • First, analysts should be on the lookout for intervening ground that is flat and level (no intervening hills or ridges) out to the nearest receivers (or “equivalent” barrier) or that slopes gently up or down (±1 to 2 degrees or so) toward them. • Where such situations exist, analysts should determine the vertical angle in degrees, at the near edge of roadway (or “equivalent” terrain line), subtended by the receiver (or “equivalent” barrier) height. • The reciprocal of that angle is an upper bound on the change in Leq produced by a 4-ft horizontal shift in the position of the edge of pavement (or “equivalent” terrain line). • If that sound-level uncertainty is too large for modeling purposes, then extra input effort should be spent to model that edge with more horizontal precision: – For edge of pavement—a “shoulder” roadway is sug- gested, with width that overlaps the nearest travel lane and weaves left and right to precisely position the edge of shoulder. – For “equivalent” terrain line—a closer look at road- way plans to more precisely locate the terrain line is suggested. This guidance applies to (1) traffic of all mixes, (2) roadways of all widths, (3) receiver distances up to 1,000 ft, (3) receivers or “equivalent” barriers up to 15 ft above the terrain. The research team suggests that analysts not model the near edge of pavement with a pavement ground zone. Instead, it should be modeled with a “shoulder” roadway—that is, a roadway without traffic that weaves right/left to best match the near edge of pavement. 8.1.2 Further Explanation To develop the guidance above, the actual analysis was augmented with other “equivalent” situations (based upon an understanding of roadway noise acoustics) beyond the exact cases computed. In particular, the augmentation takes the following into account: • The likelihood that an “equivalent” terrain line—that is, one located somewhat outside the roadway pavement and at pavement height—would experience the same location sensitivity as the near edge of pavement actually computed. This equivalence is based upon the belief that the sensitive TNM behavior discussed above is due to very small grazing angles when sound diffracts from the near edge of pavement toward the receiver. • The belief that this same sensitivity would accrue when an “equivalent” barrier substitutes for the computed receiv- ers. In this situation, the sensitivity “trigger” is the barrier top. The resulting sensitivity would likely accrue to most receivers in the barrier’s shadow zone. 8.2 Required Terrain Lines along Elevated Roadways 8.2.1 Background When roadways are elevated, getting the most accurate sound-level predictions requires a terrain line along the Topography

66 roadway—either (1) along the toe of slope for roadways on fill or (2) at ground level just off the edge of structure, for roadways on structure. That terrain line serves to pull the ground downward to its proper elevation, thereby properly modeling the height of lines of sight above the ground. For example, with a roadway on 20-ft fill or on 20-ft structure, omission of such a terrain line can result in under-prediction of sound levels by the amounts shown in Table 11. 8.2.2 Resulting Guidance for TNM Input When modeling roadways on fill, analysts should always include a terrain line along the toe of slope of the roadway fill. Similarly, when modeling roadways on structure, a ter- rain line should always be included at ground level just off the edge of the structure. 8.3 Minimum Terrain Line Spacing 8.3.1 Background The diffraction mathematics within TNM assumes that sound waves are spherically shaped when approaching a dif- fraction edge. This is normally true; however, when two dif- fracting edges are spaced very closely together, the first of these edges distorts the wave shape so that it is no longer spheri- cal when it approaches the second edge. As a result, terrain line spacing of less than 4 ft produces an abrupt, anomalous increase in sound level of • Approximately 6 dB when the terrain lines are near the top of an intervening hill or berm. • As large as 6 dB when the terrain lines are on intervening flat ground. • Between 0 dB and 6 dB when the terrain lines lie in an intervening gully. When digital terrain models approximate undulating ter- rain, they often divide that terrain into a large collection of triangles. If the edges of those triangles are used as terrain lines within TNM, then the terrain line spacing reduces to 0 ft near the vertex of all those triangles. Although not tested in this research, such a set of terrain lines could produce these 6-dB anomalies throughout. 8.3.2 Resulting Guidance for TNM Input Analysts should never input terrain lines less than 4 ft apart, especially on an intervening hill or intervening flat ground. In addition, terrain lines should not be input to duplicate the triangular topography regions that are produced by digital terrain models. 8.4 Terrain Lines: Vertical Precision 8.4.1 Background As part of NCHRP Project 25-34, the research team solic- ited and received a number of noise studies and/or TNM runs for actual highway projects around the United States. Input for two of these included an interesting assortment of terrain lines. Of concern to this research project is TNM’s sensitivity to the input Z coordinates of these modeled terrain lines. To that end, these two TNM cases were re-run with all the terrain lines moved upwards by 2 ft. In addition, a sensitivity analysis was performed with offset terrain line elevations under three geometries: (1) interven- ing flat ground, (2) intervening 40-ft hill, and (3) intervening 20-ft gully. 8.4.2 Resulting Guidance for TNM Input Resulting guidance for TNM input is the following: • Guidance from the Highway Projects. Analysts should attempt to keep the vertical precision of all terrain lines to ±1 ft—especially for barrier design projects, for which accuracy of ±1 to 2 dB is generally the goal. • Guidance from the Sensitivity Analysis. Table 12 pro- vides the appropriate guidance. No additional guidance is needed for situations not shown in Table 12. In particular, no guidance is needed when the terrain lines are in inter- vening gullies of significant depth. 8.5 Barrier Tops: Vertical Precision 8.5.1 Background When a barrier just grazes the source-receiver line of sight, the resulting path length difference for the barrier is nearly Under- Prediction Conditions Receiver Height Receiver Distance Predominant Vehicle(s) Height of Roadway-Edge Barrier 2 to 3 dB 5 and 15 ft 100 ft and greater Automobiles and medium trucks 7 ft or less, or no barrier 3 to 4 dB 5 and 15 ft 200 ft and greater Automobiles and medium trucks 3 ft or less 4 to 5 dB 5 ft 300 ft and greater Automobiles 3 ft or less Table 11. Approximate under-predictions with omitted terrain line.

67 zero. For this condition, the barrier attenuation can be highly sensitive to barrier height. More specifically, when the path length difference (from the upper subsource height) is less than 0.04 ft: • A 2-ft shift in barrier height can result in 2-to-8 dB shifts in barrier attenuation and therefore in receiver Leq. • Within this range, the shift is worse for small source- receiver distances: – 4-to-6 dB shifts are possible for source-receiver distances less than 300 ft. – 6-to-8 dB shifts are possible for source-receiver distances less than 100 ft. • This Leq sensitivity occurs for all vehicle types. • Over flat ground, such small path length differences occur only for low barrier heights (generally 8 to 10 ft). However, rolling terrain might lower barrier tops of tall barriers rela- tive to source and receiver elevations, thereby producing this high sensitivity even for taller barriers. 8.5.2 Resulting Guidance for TNM Input When any lines of sight from upper vehicle subsources to receivers closely graze a barrier top or berm top, the research team suggests taking extra care with TNM barrier input so as to precisely match (within 1 ft) barrier heights with physical reality (for existing barriers) and with intended construction heights (for future barriers). In addition, where uniform-height barriers are planned on undulating terrain, the same input care is suggested for the terrain just under the barrier, that is, for the Z coordinates of the barrier’s baseline input points. When providing guidance on barriers to roadway designers, it is better to recommend specific “barrier-top elevations” than to recommend “barrier heights above the ground.” Also, the thousands of test case comparisons conducted with TNM in this research have shown very large Leq sen- sitivity to the exact location of diffracting edges, whenever sound paths just graze across those edges. For those grazing situations, Leq is also very sensitive to the slightest wind in the direction of propagation, which TNM does not account for. Chapter 11 provides information on the effects of wind on sound levels behind barriers. 8.6 Flat-Top Berms 8.6.1 Background During TNM validation studies, the U.S. DOT Volpe National Transportation Systems Center acoustics group determined that TNM sometimes miscomputes sound levels behind flat-top berms by 5 dB or more. To avoid this miscom- putation, TNM 2.5 currently prevents entry of a berm object’s top width—thereby restricting berm objects to “wedges,” with- out flat tops. Nonetheless, TNM users can bypass berm objects entirely by using terrain lines to manually input berm shapes, including shapes with flat tops, and, unfortunately, such man- ually input berms produce the same miscomputations. 8.6.2 Resulting Guidance for Highest Precision (Generally for Project “Design Phase”) Recently the Volpe Center has devised (but not published) the following work around for TNM’s flat-top berm problem. To avoid miscomputation, the top edges of flat-top berms should be “rounded-off” (see Figure 47), as follows: 1. The original top-edge terrain line should be moved toward the center of the berm top by 1/10th of the Intervening Terrain Dominant Vehicle Type Receiver Heights Receiver Distances Roadway Width Guidance: Match Actual Terrain Elevation within This Amount Flat within ± 10 ft Gullies less than 10 ft deep tf 2 ± llA llA tf 5 skcurt yvaeH Medium trucks Automobiles 5 ft tf 2 ± llA tf 054 naht sseL 450 to 750 ft More than 50 ft ± 0.5 ft 30 to 50 ft ± 1 ft Less than 30 ft ± 2 ft 750 to 1000 ft More than 50 ft ± 1 ft Less than 50 ft ± 2 ft tf 2 ± llA tf 0001 naht eroM Hills more than 10 ft high All All tf 2 ± llA llih eht no yllautcA tf 5.1 ± llA llih eht dniheb tf 001 nihtiW Farther than 100 ft behind the hill tf 1 ± llA Table 12. Guidance for elevation of intervening terrain lines.

68 berm-top width (W), keeping its original elevations from point to point. 2. A second terrain line should be added down the berm slope, positioned outward by 1/10th the berm-top width and downward by the amount needed to keep it approxi- mately on the original berm slope. 3. As the toe-of-slope terrain line moves in/out and up/down along the length of the berm, that might change the slope along the berm. For such situations, the new terrain line can be positioned vertically so the new piece’s slope (the thick line in Figure 47) is approximately one-half the orig- inal berm slope. 4. This process should be repeated for the other top edge of the berm as well. 8.6.3 Resulting Guidance for Moderate Precision (Generally for Project “Location Phase”)35 For a flat-top berm, a wedge-shaped berm of the same height can be used as a substitute. Such substitution might be slightly conservative, that is, it might compute slightly lower noise reduction than actually achieved by the flat-top berm. 35 This recommendation is a paraphrased condensation of the TNM FAQ on the web at www.fhwa.dot.gov/environment/noise/traffic_noise_model/tnm_faqs/ faq07.cfm. Figure 47. Section view of flat-top berm shape including the suggested “rounding off” of the top edge.