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.

40 C H A P T E R 5 5.1 Introduction Typical issues encountered in FHWA TNM modeling of roadway sections that contain median barriers include the following: • FHWA TNM Version 2.5 has a component that addresses single reflections; however, this component is “turned off” and not available for use. • The parallel barrier module within the FHWA TNM is not intended for use with lower-height barriers such as median barriers. • It is envisioned that the FHWA TNM Version 3.0 will be capable of modeling single reflections; however, this version is not yet available for use and its limitations and graphic functionality are still being evaluated. Therefore, evaluations using FHWA TNM Version 3.0 were not conducted. It is recognized by some noise practitioners that median barriers can have an effect on noise levels at adjacent receptors. The effect may be related to a variety of factors, including the following: • Horizontal and vertical relationship of the median barrier to adjacent lanes. • Elevation of adjacent receptors with respect to roadways and the median barrier. • Distances between the roadways and median barrier and adjacent receptors. • Height and shape of the median barrier. While other factors such as ground type, topography, and noise barriers affect noise levels, for purposes of testing and evaluating the influences of median barriers, the research team focused on the four bulleted items listed above. There- fore, traffic and noise measurement data associated with loca- tions with relatively simple topography and features were selected for testing and evaluation. While various ground types exist for the selected sites, no attempts were made as part of this investigation to address ground type variability in developing best modeling practices for median barriers. In addition, the team focused on collecting measurement data from sites located at elevations level with the highway, higher than the highway, and lower than the highway. The techniques associated with FHWA TNM Version 2.5 were evaluated and tested using measurement data from the five selected projects described in Section 5.2 of this report. Suggested best modeling practices for median barriers were developed based upon this evaluation and testing. More detailed information is included in Appendix D, which is avail- able on the NCHRP Project 25-34 web page at http://apps.trb. org/cmsfeed/TRBNetProjectDisplay.asp?ProjectID=2986. 5.2 Measurement Locations Evaluated By far, the highest quality noise measurement and vali- dation data exist in recent studies conducted by the Volpe Center. The following three locations in the Volpe measure- ment studies had median barriers and were evaluated in this investigation: • Volpe Site 3C. This measurement location in Arizona was part of Volpe’s Arizona Quiet Pavement Program evaluation project. The location is relatively flat, and the eight-lane divided highway (four lanes in each direc- tion plus a ramp lane in one direction) is relatively level. The median barrier is in the center of a 32-ft wide paved median. • Volpe Site 18PA. In 2001, measurements were taken at this location adjacent to the Pennsylvania Turnpike west of Carlisle, Pennsylvania, for use by Volpe in its FHWA TNM Validation Project. The topography is relatively flat and the four-lane divided highway (two lanes in each direction) Median Barriers

41 is relatively level. The median barrier is in the center of a 10-ft wide paved median. • Volpe Site 22PA. Measurements were taken at this loca- tion adjacent to PA Route 581 in Camp Hill, Pennsylvania, as part of Volpe’s FHWA TNM Validation Project. In this area, an earth berm extends along a length of the highway. It then ends, and an unprotected (no berm) length of highway exists. For the unprotected section, three measurements were taken at distances equal to those in the berm section. With the exception of the berm, the topography for both areas is relatively level. The four-lane divided highway (two lanes in each direction) is also relatively level. The median barrier takes up the majority of the narrow paved median. In addition to the evaluation and testing of the high- quality noise measurement and traffic data available from the Volpe studies described above, the team evaluated and tested each modeling technique identified in Section 5.3 using several measurement data sets from the following two projects: • Pennsylvania Turnpike Noise Analysis Project, Butler, Pennsylvania. In the early to mid-2000s, members of the research team conducted a variety of noise measurements for a section of the Pennsylvania Turnpike in Butler, Penn- sylvania, as part of a noise evaluation for the Turnpike’s pro- posed Warrendale Mainline Toll Plaza project. An existing public park exists within the project area immediately adja- cent to the Turnpike. The majority of the park is approxi- mately 15 to 20 ft lower in elevation than the Turnpike. For this park area, team members used a grid of receivers for analyzing equivalent residential units. During the time of the noise measurements, a median barrier existed in the area and the volume of heavy trucks was significant. • Ohio DOT Interstate 71 (I-71) Noise Analysis, Columbus, Ohio. In 2006, research team members conducted extensive noise measurements for a six-lane (three lanes in each direc- tion) section of I-71 in Columbus, Ohio. During the model validation process, the team evaluated various measure- ment locations where specific attention was directed toward consideration of the effects of median barriers. Locations existed where receptors were located level with, above, and below the roadway elevation. 5.3 Modeling Techniques Evaluated Based on a review of data collected by the team and input from team members and other noise specialists contacted, modeling techniques were identified by the team for evalua- tion and testing. For evaluation of each one of these candidate modeling techniques, the team utilized the measurement and traffic information from the projects described in Section 5.2. This process resulted in the development of methodologies to adjust the basic FHWA TNM output data to appropriately incorporate the effects of median barriers. Receptors located at various distances from the highway were evaluated using each of the techniques discussed below. In its evaluation of each of these modeling techniques, the team utilized the measurement and traffic information asso- ciated with 49 individual measurements taken at distances ranging from 46 to 1,000 ft from the center of the near traffic lane. Measurements were taken at points where the topogra- phy at the measurement site ranged from below the elevation of the highway to near level with the highway to above the highway. While the FHWA TNM Version 2.5 does not have the abil- ity to model low-height reflective surfaces or their shapes, the team did evaluate the relative effects of median barriers using FHWA TNM and varying median barrier heights (from 2.5 to 4.5 ft) for each of the five projects identified in Section 5.2. From this evaluation, the team determined that the relative differences in noise levels associated with the range of median barrier heights evaluated was relatively insignificant. For each modeling technique described below, median barriers were assumed to be the same height as existed at the time of the measurements for each respective project. Vertical median barrier faces that are 100% reflective were assumed. In addition, each roadway lane was input as a separate road- way within the FHWA TNM, with its own geometry, traffic volumes, and speeds. 5.3.1 Image Roadway Technique Approximation Using “Seen” Travel Lanes This technique developed an image roadway to represent the noise reflected from the median barrier from the near travel lanes. This roadway was constructed in an FHWA TNM run by “flipping” the eastbound travel lanes (shown in Figure 31) to the far side of the barrier, as shown in Figure 32. Where differ- ent traffic volumes are assigned to each travel lane in the base FHWA TNM run (such as in the Volpe projects), these travel lane volumes were also flipped to place them at similar dis- tances from the median barrier. When calculating the reflected noise values using this technique, only the image roadways (eastbound flipped travel lanes) were modeled and only the eastbound vehicles that were “seen” (and heard) by a particular receptor were modeled in their “flipped” position. To provide an approximation of which vehicles would be seen, a line was drawn from the receptor to the top of the median barrier. Any vehicle sources falling on or below this line were assumed to be seen by the receptor, unless the line of sight is blocked by some ground or roadway feature. The research team also looked at skew sections representing flanking noise in the identification

42 of seen traffic sources. For the example used, this sight line is shown in Figure 31, with the seen sources indicated by solid circles in Figure 32. In modeling the flipped roadway, the default ground type within the FHWA TNM run was the same as that of the base FHWA TNM run and all other topographic features (ground zones, terrain lines, etc.) were the same as in the base FHWA TNM run. The areas occupied by the roadway median and the eastbound roadways and shoulders were input as ground zones having their respective surface prop- erties. Noise levels generated by this reflected noise run were calculated at the actual receptor locations. Any adjustments determined to be appropriate based on this reflected noise run were applied to values generated by the base FHWA TNM run, which was modeled with the median barrier input as a barrier. Because of the many ray paths and multiple reflections that actually occur between the various sources, the median bar- rier, and the roadway surfaces, the team recognizes that this technique is, at best, an approximation of the reflected noise. However, the use of a more complex ray-tracing technique is beyond the scope of this research. An alternative approach explored by the team was to model the eastbound lanes and shoulder areas but to delete all traffic from the eastbound lanes. In modeling flipped roadways, the team compared reflected noise levels generated by the fol- lowing two approaches for several of the projects described in Section 5.2: Figure 31. Skew section of base FHWA TNM run. Figure 32. Skew section of reflected (flipped) FHWA TNM run.

43 • Modeling flipped roadways and using ground zones to represent the intervening ground occupied by deleted near roadway lanes. • Modeling flipped roadways and retaining the pavement surface geometry of the near roadway lanes, but deleting traffic from these near roadway lanes. The comparison indicated that the “ground zone” approach results in slightly higher values for the reflected noise component than does the “pavement” approach. How- ever, when the reflected component is added to the direct noise component (FHWA TNM modeled noise levels from all lanes with a median barrier in place), the difference in total noise levels ranged from 0.1 to 0.4 dB, which is relatively small. The team chose to use the slightly more conservative ground zone approach in its estimations of reflected noise component values. 5.3.2 Image Roadway Technique Approximation Using All Travel Lanes A more conservative approach evaluated by the team was to model all “flipped” roadway sources as being reflected by the median barrier. This typically resulted in higher reflected noise levels than generated by the “seen” vehicle source tech- nique described in Section 5.3.1. 5.3.3 Ignoring Median Barrier This technique simply ignored the presence of any median barrier in the base FHWA TNM run and did not assume any noise reflections. 5.3.4 Ignoring Median Barrier Reflections This technique included the barrier in the base FHWA TNM run, but made no adjustments to account for reflec- tions off of the barrier. 5.4 Best Modeling Practices The team recognizes that any one of its best modeling practices may not be appropriate for all modeling scenarios. For example, one practice may be appropriate for elevated receptors, but not for receptors located lower than the road- way. One practice may work for nearby receptors, but not for more distant receptors. Review of the trends in the results of the research resulted in suggestions for best management practices for modeling median barriers: • For receptors located within 500 ft of the highway (center of nearest travel lane) and located below the elevation of the highway, model the median barrier and ignore reflec- tions off of the median barrier. • For receptors located beyond 500 ft of the highway and located below the elevation of the highway, model the median barrier and consider reflections off of the median barrier using the appropriate reflected barrier technique. • For receptors that are located from 50 ft to 500 ft from the highway and are level with or less than 6 ft above the highway, model the median barrier and ignore reflections. • For receptors that are located 6 ft or more above the ele- vation of the highway and within 500 ft of the highway, model the median barrier and account for reflections. The majority of the median barriers evaluated were inten- tionally located in areas with relatively simple terrain con- taining no intervening noise barriers or berms. However, the research team recognizes that median barriers often exist in conjunction with these other features, and therefore the team selected one area to test the various modeling techniques against such features. High-quality measurement and valida- tion data were obtained by the team for Volpe Site 22PA. This location offered the opportunity to test and evaluate each of the median barrier analysis techniques at two receptor loca- tions behind an earth berm. The analysis of these sites based on FHWA TNM runs for the receptors indicated that it is probably appropriate to ignore the effects of the median bar- rier. While this can provide some guidance for similar types of projects, the best modeling practice for projects having median barriers located in areas containing noise berms or noise walls should consider the specifics of the project area. 5.5 Conclusions Suggested best modeling practices were developed for adjusting FHWA TNM predictions to account for the effects of noise reflections off of median barriers. The team recog- nizes that any one of its suggested best modeling practices may not be appropriate for all modeling scenarios. However, certain trends were observed that enabled the development of the generalized suggestions, shown in Table 8, related to incorporating the effects of noise reflections off of median barriers. Such suggestions relate to situations where receptors were generally unaffected by intervening objects between the median barrier and the receptor, as well as areas where recep- tors were located behind a noise abatement feature. Based on the team’s evaluation and testing, it is suggested that median barriers be modeled in all cases, even if the effects are slight. An exception to this suggestion could occur where receptors are located behind noise abatement features.

44 Distance from Middle of Near Travel Lane (ft) Height of Receptor with Respect to Roadway Receptor Below to 6 ft Above Roadway Receptor More Than 6 ft Above Roadway 50 Model Median Barrier and Ignore Reflections Model Median Barrier and Consider Reflections 100 200 500 1000 Model Median Barrier and Consider Reflections Table 8. Suggested modeling techniques for median barriers by receptor location.