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.

83 C H A P T E R 1 2 12.1 Research Approach 12.1.1 Basic Concepts and Use of the FHWA TNM Parallel Barriers Module Sound levels behind a barrier can increase when there are multiple reflections of the sound between the barrier and a sec- ond barrier parallel to it on the opposite side of the road, form- ing a vertical wall “canyon,” as shown in Figure 55. Figure 56 shows in a schematic overhead view that these multiple reflec- tions are a three-dimensional phenomenon, with reflected sound reaching the receiver before vehicles pass by it and after they pass as well. FHWA TNM 2.5 allows modeling of this phenomenon in a separate two-dimensional parallel barrier analysis module within the program. When the analyst selects a cross section to study and initiates a new parallel barrier design by cutting a sec- tion line through the plan view of the model, a separate parallel barrier view is opened, as shown in Figure 57. Certain input data are passed to the module from the main part of TNM, including the elevations and horizontal offsets of the following: • Roadways. • Analysis locations (receivers with their heights added to their ground elevations). • Barriers (with the barrier input heights added to their ground elevations). Traffic volumes and speeds of the roadways are also passed to the parallel barrier module. The analyst then typically refines this cross section to tailor it more specifically to the actual location, as shown in Figure 58, where • Additional roadways were added to represent individual travel lanes that may not have been modeled in the main part of FHWA TNM. • Barrier heights were adjusted from the input heights to the analyst’s designed heights. • Additional representative analysis locations were added that were not picked up in the initial cutting of the analysis section. • Parallel barrier cross-section surface segments outside of the canyon to be studied were deleted (because analysis locations may not be within the extents of the cross section being calculated). The program then computes sound-level increases over the single barrier “with barrier” LAeq1h computed during the barrier design in the main part of FHWA TNM. The pro- gram does not add these increases to the LAeq1h. That task is left to the analyst. Then, if the increases are considered sig- nificant (typically more than 1 to 2 dB), the analyst can evalu- ate mitigation using the parallel barrier analysis module such as applying a sound-absorbing surface to the inside faces of the parallel barrier or tilting the barrier outward slightly to eliminate the multiple reflections pattern. 12.1.2 Research Steps The objective of this research was to investigate the sen- sitivity of the parallel barrier module to a variety of factors and refine the available guidance on the use of the module. Guidance was needed in a number of areas, in some cases in terms of best modeling practices and in other cases on how to recognize and work around issues with the implementation of the module within FHWA TNM. The areas that were stud- ied relating to the sensitivity of the computed sound-level increase to the input parameters include the following: • Height-to-width ratio for the barriers and receiver posi- tion behind the barrier. • Number of FHWA TNM roadways used to represent the travel lanes. • Source position. • Differences in the heights (top elevations) of the two barriers. Parallel Barriers

84 Figure 55. Highway with parallel noise barriers on opposite sides of the road.58 Figure 56. Plan view of multiple reflections off parallel highway noise barriers. 58 Source: Bowlby & Associates, Inc. Figure 57. Section line through the plan view of the model. Figure 58. Tailored cross section of actual location.

85 Figure 59. Illustration of receiver array at heights of 15, 5, –5 and –15 ft relative to the roadway surface and distances of 25, 50, 100, 150, 200, and 300 ft from the near wall. • Internal vertical reflecting surface. • Vehicle mix (e.g., automobiles only versus heavy trucks only). • Hourly volumes of vehicles. • Vehicle speed. • Noise reduction coefficient of barrier surfaces. The research also evaluated two sets of measured and modeled data for a parallel barrier project before and after the addition of sound-absorbing panels to one of the barriers. 12.2 Outcome of the Research— Best Practices and How to Implement Them for a Noise Study or TNM Model 12.2.1 Height–to-Width Ratio for the Barriers and Receiver Position Behind the Barrier The sound-level increase due to multiple reflections between the parallel barriers is partly a function of the width- to-height ratio for the cross section (distance between the noise barriers divided by their height). In general, the smaller the ratio, the greater the sound-level increase will be. The current FHWA TNM FAQ states: When should I analyze my parallel barriers? Research has shown that the magnitude of the performance deg- radation associated with parallel reflective noise barriers is linked to the ratio of the separation (width) between the barriers and the average height of the barriers. Definitely analyze parallel bar- riers when the cross-section’s width-to-height ratio (W: H) is less than 10:1. When the ratio is between 10:1 and 20:1, you may still want to analyze the cross-section with TNM. If the ratio is greater than 20:1, you do not necessarily have to analyze the cross-section. Such a calculation will yield inconsequential sound-level increases. Please refer to the Parallel Barriers Menu section on Page 103 of the TNM Users Guide for more information. The TNM Users Guide repeats the guidance quoted above and indicates that the maximum expected degradation in the noise reduction provided by a single barrier due to a second barrier (termed “sound level increase” in the FHWA TNM program) will be 0 to 3 dB for W:H ratios of 10:1 to 20:1 and that there will be “no degradation” for a W:H ratio greater than 20:1. The findings of this research were different in terms of the model results from FHWA TNM. Tests were run for an eight-lane cross section with a barrier- to-barrier width of 136 ft consisting of eight 12-ft lanes with 10-ft inside and outside shoulders in each travel direction. As illustrated in Figure 59, an array of receivers was modeled for six distances back from the near wall (25, 50, 100, 150, 200, and 300 ft) and at four heights relative to the roadway surface: 15 ft, 5 ft, -5 ft and -15 ft. These heights represent, respectively, an exterior second-story location in an at-grade cross section, a typical exterior receiver in an at-grade cross section, a receiver alongside a 10-ft roadway embankment, and a receiver alongside a 20-ft roadway embankment. In the FHWA TNM parallel barrier module, the points at which the sound-level increases are predicted are called “analysis loca- tions.” In this chapter, the word “receiver” will be used, where the height of the receiver is the point at which the sound-level increase is predicted. It is known that the parallel barrier module has not been validated for receivers above the top of the near wall, which is the case for some of the studied situations. Barrier heights were varied from 1 to 20 ft, with the result- ing W: H ratios ranging down to 7:1 for 20 ft. The barriers are assumed to be sound reflecting, with a noise reduction coefficient (NRC) of 0.05. Appendix K (available on the NCHRP Project 25-34 web page at http://apps.trb.org/cmsfeed/TRBNetProjectDisplay. asp?ProjectID=2986) contains graphs of the results of mod- eling for automobiles only and heavy trucks only. The sound- level increase is a function of not only barrier height (and thus width-to-height ratio) but also receiver height above or

86 below the road, receiver distance back from the near wall, and vehicle type. For the studied cross section, a 10:1 width-to-height ratio resulted in sound-level increases for automobiles ranging from 1.0 to 6.5 dB. For heavy trucks, the range was 0.3 to 4.1 dB. In some limited testing of medium trucks, the predicted sound- level increases were within 0.1 to 0.3 dB of those for automo- biles. The sound-level increase tends to rise as receiver height increases. For most of the receiver positions, this amount of sound-level increase warrants attention in the barrier design process. For this same cross section, a 20:1 width-to-height ratio resulted in sound-level increases for automobiles from 0.3 to 3.1 dB, being greatest for the 5-ft receiver. For heavy trucks, the range is 0 to 0.7 dB. Depending on the mix of traffic and the receiver location, the sound-level increase even for this 20:1 width-to-height ratio may warrant attention during barrier design. For the 5-ft-high receiver located 25 ft from the near wall, the sound level increases very little as barrier height increases from 11 to 20 ft, with a similar pattern for the same receiver 50 ft from the near wall. For heavy trucks, the sound-level increase is also not particularly sensitive to increasing barrier heights at the closer-in distances. These results counter pre- vailing thought that the sound-level increase rises as barrier height increases because more multiple reflection paths are created as the barriers get taller. This pattern is not consistent across all receiver heights and distances, but suggests that FHWA TNM will show that increasing the barrier heights will overcome the increase in the “with barrier” sound level in a parallel barrier situation, which could lead to increas- ing heights as an alternative mitigation technique to sound absorption. This report is not recommending such a strat- egy because of the lack of field validation. Use of sound- absorbing surfaces on the road side of the walls remains the recommended mitigation strategy for minimizing the sound-level increases. Further tests varied the width-to-height ratio by keeping the barrier heights at 20 ft and increasing the width between the parallel barriers for automobiles-only cases. Even at 20:1, sound-level increases over 2 dB were calculated for the 15-ft- high receiver over all distances and at 100 ft and beyond for the 5-ft receiver. At 20:1, the sound-level increases for the receivers 5 ft and 15 ft below the roadway grade are less than 2 dB. Untested is whether or not the sound-level increases would occur in the real world. A 20:1 width-to-height ratio for 20-ft high barriers means the barriers are 400 ft apart. Meteoro- logical effects on sound propagation, such as wind shear (changing wind speed with altitude) or temperature lapse rate (changing temperature with altitude) could easily have more effect on sound levels over these distances due to refraction than would the reflected paths. 12.2.2 Number of FHWA TNM Roadways Used to Represent the Travel Lanes Within the range of the tested cases described above, modeling the eight-lane cross section by a total of two or four FHWA TNM roadways (one or two in each direc- tion) produced results within a 0.5 dB of the eight-roadway model, with a few exceptions where differences up to 1 dB were computed. 12.2.3 Source Position The finding of insensitivity to the number of modeled roadways across the entire cross section was tested to see how much source position within the canyon between the two barriers affected the parallel barrier sound-level increases. The eight-lane cross section shown in Figure 59 was broken down into cases consisting of the four “far” lanes only being modeled by four FHWA TNM roadways and then by one roadway centered between them, and then the four “near” lanes only being modeled by four roadways and then one roadway. Source position has only a small effect on the sound-level increase for the lower receiver positions (1 dB or less), espe- cially within 150 ft of the near wall. Source position has a larger effect, of 2 dB or more, at the more distant receiver positions and for the 15-ft-high receiver. The results were similar for the automobiles-only and heavy trucks-only cases, with the automobiles-only sound-level increases being generally higher. 12.2.4 Differences in the Heights (Top Elevations) of the Two Barriers As the height of one of the two parallel barriers changes, there is a change in the pattern of sound-level reflection. Con- ceptually, as the height of the far wall decreases, the potential for many multiple reflection paths decreases, a situation that could then reduce the size of the sound-level increase due to reflections. Tests varying the far wall height from 10 to 22 ft while holding the near wall height at 20 ft for the eight- roadway cross section for automobiles only showed, in gen- eral, that the parallel barrier module does compute smaller sound-level increases as the far wall height decreases. The change in the sound-level increase is greater for the higher receivers and the greater distances from the near wall because the actual sound-level increases for the equal wall height cases are larger for these receiver positions. However, even for rela- tively low far wall heights, the sound-level increases can still be substantial enough to warrant investigation and possible mitigation through the use of sound-absorbing surfaces on one or both walls.

87 12.2.5 Internal Vertical Reflecting Surface The FHWA TNM FAQ for parallel barriers cautions ana- lysts about having an internal, vertical, reflecting surface in the analyzed parallel barrier cross-sectional surface. (See Figure 60.) The extent of the effect on the results may depend on the source position, the heights of the noise barriers and the internal vertical surface, the offset of the external wall from the internal vertical surface, and the receiver position. A test was created to illustrate the problem of internal surface reflections. As shown in Figure 61, the cross section on the left consisted of a 20-ft-high near wall and a 19-ft-high far wall. The cross section on the right was the same, except that a 1-ft- high noise barrier was added offset 10 ft to the left beyond the far wall. The 19-ft-high far wall thus went from being an external vertical surface to an internal vertical surface. The parallel barrier module computed sound-level increases in both cases, but there were differences, ranging from 1 dB in close to more than 4 dB farther back. Acoustically, there should be no difference in the calculated sound-level increases. The 1-ft noise barrier offset from the 19-ft verti- cal section is not in a position to reflect sound back across to diffract over the top of the near wall. This fact was tested by making the 19-ft-high wall highly absorptive: all of the sound-level increases became 0 dB. In an additional test, the 1-ft noise barrier was deleted so that the cross section ended with a horizontal segment beyond the top of the 19-ft section, which was reset to being highly reflective. All of the sound-level increases became 0 dB. Because of these inconsistent results, internal vertical reflecting or diffracting surfaces should not be analyzed using the parallel barrier module or included in any parallel barrier analysis cross sections. 12.2.6 Vehicle Mix The FHWA TNM parallel barrier module is not sensitive to changes in vehicle mix (the percentage of automobiles versus trucks in an hourly traffic flow) once trucks are intro- duced into the flow. In general, a ±5% change in percentage Can TNM model more than 2 parallel barriers? Yes, it can be modeled as a single cross section in the Parallel Barriers module. However, keep in mind that when a parallel barrier section contains two separate vertical surfaces offset on the same side of a road (i.e. a retaining wall near the edge-of-pavement and a barrier at the right-of-way), (1) TNM parallel-barrier accuracy is degraded somewhat for receivers on that same side of the roadway (TNM may under-compute or over-compute the noise increase), and (2) TNM may under-compute the noise increase for receivers on the opposite side of the roadway. Please refer to the diagram below: Figure 60. Text and diagram from TNM FAQs on parallel barriers.59 Figure 61. Cross section with 19-ft-high far wall as an external vertical surface (left) and an internal vertical surface with a 1-ft noise barrier offset 10 ft from the top of the internal vertical surface (right). 59 Source: www.fhwa.dot.gov/environment/noise/traffic_noise_model/tnm_faqs/ faq10.cfm#menupara.

88 of automobiles changes the sound-level increase by only a few tenths of a decibel, except in going from 100% automobiles to 95% automobiles, where the change in sound-level increase is on the order of 0.5 dB. 12.2.7 Hourly Volumes of Vehicles The FHWA TNM parallel barrier module is only predict- ing a sound-level increase in the 1-hour Leq and not an actual 1-hour Leq. The module’s calculations are independent of the hourly volumes, but are dependent on the vehicle mix, as was just described in Section 12.2.6. Identical sound-level increases were computed for a run of 1,000 each of auto- mobiles, medium trucks, and heavy trucks compared to a run with just one vehicle of each type. 12.2.8 Vehicle Speed Sound-level increases computed by the FHWA TNM par- allel barrier module are independent of speed for each vehicle type. Results will not change as speed changes. 12.2.9 NRC of Barrier Surfaces When the predicted sound-level increase from reflected sound is determined to be large enough to mitigate, the most common solution has been the use of a sound-absorbing product or material for the surfaces of the walls facing the roadway. Some state highway agencies will also use sound- absorbing barriers in single-wall situations where there are residences on the other side of the road that may or may not be impacted, but do not meet the agency’s noise abatement feasibility or reasonableness criteria. The FHWA TNM parallel barrier module has the capability of testing the effectiveness of changing the NRC of all or parts of one or both of the parallel barriers. The NRC is a frequency- specific quantity, being the average of the sound-absorption coefficients in the 250, 500, 1,000, and 2,000 Hz octave bands. Different products with different sound-absorption coeffi- cients in these bands can have the same NRC, yet perform differently in the field. The FHWA TNM parallel barrier mod- ule computes the diffraction attenuation of the sound passing over the near wall at a frequency of 500 Hz. As such, the appli- cation of an NRC will give an indication of the general effect of the sound-absorbing material, but not a precise calculation for a specific product. To test the parallel barrier module’s application of the NRC, several cases were studied. The basic case was the eight- roadway cross section for automobiles only with 18-ft barriers on either side. The NRC of both walls varied between 0.05 (a typically used value for concrete) and 0.90 in 0.10 incre- ments (starting from 0.10). Then, just the far wall was made sound-absorbing, with the same NRC variation, and then just the near wall in the same manner. Finally, the heights of both walls were varied in tandem to test the effect of the NRC on different height configurations. Appendix K provides graphs of the sound-level increases as a function of the wall NRC for all of the receivers shown in Figure 59. For all three cases, the effectiveness of the increased sound absorption is fairly linear, reducing the sound-level increase as the NRC increases. For absorption on both walls, an NRC of 0.7 or higher brings the reflective barriers’ sound-level increases down to less than 1 dB for all of the receiver posi- tions except at the 15-ft receiver height, for which the maxi- mum sound-level increase is less than 2 dB. Sound absorption on the far wall only is also very effective for this cross section in reducing the sound-level increases. For any given receiver position, the sound-level increases with absorption on just the far wall are 0 to 1.3 dB higher than when there is absorption on both walls. In contrast, absorption on just the near wall is far less effective than absorption on the far wall or both walls. For any given receiver position, the sound-level increases with absorption on just the near wall are up to 3.8 dB higher than when there is absorption on both walls. The results suggest the importance of the single far wall reflections on the total sound level at a receiver, but also show that the program is calculating multiple reflection paths back and forth between the barriers because near wall absorption also reduces the sound level over the fully reflective case. 12.2.10 Comparison of Measured and Modeled Levels Including Parallel Barrier Sound-Level Increases A comparison of measured and FHWA TNM predicted sound levels, including parallel barrier sound-level increases, was made for a study that evaluated traffic noise barriers along both sides of a state highway.60 The walls were both originally sound reflecting. In response to citizen complaints about noise behind one of the barriers, the state highway agency studied the problem61 and then added absorption panels to the wall on the opposite side of the highway to reduce sound reflections back into the community. The follow-up study was then conducted. Included in the follow-up study were noise measurements with concurrent traffic and meteorological data collec- tion, noise modeling with TNM 1.0b, and administration of a follow-up survey of the affected citizens. The data sets 60 Bowlby & Associates, Inc., SUM-8-6.83 Noise Wall “After Absorption” Study, State Route 8, Silver Lake, Ohio, for Ohio DOT District 4, 2000. 61 Bowlby & Associates, Inc., SUM-8-6.83 Noise Barrier Post Construction Study-State Route 8 - Silver Lake Ohio, for Ohio DOT Office of Environmental Services, 1996.

89 from the initial and follow-up studies provide field data and FHWA TNM runs for a reflective parallel wall situation (before absorption) and for a situation with a near side reflec- tive wall and a far-side, sound-absorbing wall (after absorp- tion). FHWA TNM 1.0b runs were converted to run in FHWA TNM 2.5 for use in this research. The project study area had two analysis sections: 1. A “two-wall” area where both walls were essentially at-grade with the road and of nearly equal heights. 2. A “no-wall” area north of both barriers. A reference microphone was deployed in each area (0-Ref in the no-wall area, 2-Ref in the two-wall area), and two individual study sites were chosen within each area (0-A, 0-B, 2-A, 2-B), with a third, more distant, site in each area (0-C, 2-C). The results for the initial measurements with both walls reflective showed FHWA TNM 2.5 predicted well in the no- wall area at sites 0-Ref and 0-B. FHWA TNM 2.5 generally under-predicted the levels at the other sites. However, at 0-A, the FHWA TNM 2.5 over-prediction was 5.4 and 7.0 dB. In the original study in 1996, the FHWA STAMINA 2.0 program over-predicted this same site by 6.1 dB, and when the predic- tions were redone with FHWA TNM 1.0b, the over-prediction was also large. The reasons for all three models’ over-prediction are not clear. The two-wall sites were then studied with the FHWA TNM 2.5 parallel barrier module. The computed parallel barrier sound-level increases were 0.3 dB at 2-Ref, 2.6 dB at 2-A, 3.4 dB at 2-B, and 0 dB at 2-C. Applying the sound-level increases to the main FHWA TNM 2.5 single-wall predic- tions improved the model performance slightly at the 2-Ref site, with it still under-predicting by 0.1 to 1.8 dB. The pre- dicted levels at study sites 2-A and 2-B increased. Site 2-A’s levels became higher than the measured levels by 0.7 to 1.5 dB, whereas they were lower before adding in the calculated sound-level increase. Site 2-B’s over-prediction increased to 4.1 dB. Normalizing the data by the 2-Ref predicted- minus-measured sound-level difference increased the predicted-minus-measured differences at 2-A and 2-B and improved the difference at 2-C. After the sound-absorption installation, new measure- ments were made and the modeling was revisited, using a far wall NRC of 0.80. For this research, the modeling was redone using FHWA TNM 2.5. In the no-wall area, FHWA TNM 2.5 predicted within -0.1 to +1.0 dB of the measured levels at 0-Ref. However, at 0-A, 0-B, and 0-C the results were mixed. The measured levels varied substantially between periods at each site, resulting in both good and poor agree- ment with the modeling. The reasons for the variation in the measured levels were not clear. In the two-wall area, the computed parallel barrier sound- level increase at 2-Ref was 0.3 dB, the same as for the “both walls reflective” case. One would have expected this value to decrease. At 2-A, the sound-level increase dropped from 2.6 dB to 0.2 dB; at 2-B, it decreased from 3.4 dB to 0 dB; and at 2-C, it remained at 0 dB. Overall, after using the par- allel barrier module, FHWA TNM 2.5 predicted within 1 dB of the measured levels at 2-Ref, within 2 dB at 2-A, and within 2.5 dB at 2-B. At 2-C, FHWA TNM 2.5 greatly under- predicted the levels. Overall, the results of the comparisons of the measured and modeled levels in the reflective and far wall absorptive cases were mixed. Agreement was good at the reference microphone sites for both the no-wall and two-wall sites in each case, and at 2-A, the closest site behind the near wall. At the other study sites, agreement ranged from mixed at 0-B to poor at 0-A and 2-C. One issue was with the range in the measured sound lev- els at the sites, especially in the far wall absorptive case. Site 2-C was deep into the community, and while care was taken regarding localized noise sources and meteorological effects on sound propagation, these factors could not be ruled out as possible causes of the sound-level differences. 12.2.11 General Notes and Guidance Finally, several general notes and some guidance are pro- vided here. Because the algorithms in the parallel barrier module have not been calibrated for receivers at elevations above the ele- vation of the near barrier, the guidance from the module’s developer (G. S. Anderson) is to generally require a minimum barrier height of 6 ft for either barrier. Additionally, the algorithms are such that the module should not be used for single-wall reflections. There are also occasions where single reflections off a barrier or a vertical retaining wall on the far side of a highway may be important to receivers on the near side of the highway. Studying single-wall reflections was outside of the scope of this research. FHWA TNM 2.5 has a single reflections routine in the main part of the program— separate from the parallel barrier module—that is currently deactivated in the code because of issues during its develop- ment. The plan for FHWA TNM 3.0, now under development, is to make this single-wall reflections component functional. Until then, FHWA TNM 2.5 modelers should consider the use of “image roadways” in a run to model single-wall reflec- tions without actually having the far wall in the run. Careful addition of TNM roadways to the run to represent the reflected images of the “real” TNM roadways in the barrier is an excel- lent way to study such situations. Care must be taken to ensure that the image roadways in the run represent vehicle sources on “real” roadways, all of which would truly reflect from the barrier to the receivers in the run.

90 In general, a parallel barrier analysis would begin with a review of the highway plans and proposed noise barriers to identify areas where multiple sound reflections might occur, namely, where there are barriers and/or vertical retaining walls on both sides of the road. Several representative sites or cross sections would then be selected for study, such as the following: • Different cross section types (e.g., cut, fill, and at grade). • Different barrier heights (for one or both of the barriers). • Different barrier offset distances from the roadways. In cutting parallel barrier analysis sections in the main plan view of FHWA TNM, the analyst does not have to select or cut through modeled receiver points. The receivers can be added in the parallel barrier module as “analysis locations.” The elevation for an analysis location is not the ground, but the calculation point above the ground (receiver ear height). Analysis locations may not be placed at the edge or within the boundaries of a cross-section surface. In the program’s parallel analysis location input dialog box, the analysis locations may be named by the analyst as other than the default, but the program-assigned “sequence #” needs to be kept in the name for identification on the parallel analysis location table because the parallel barrier view only displays sequence numbers and not names, and the table only displays names and not sequence numbers. The “computed increases in LAeq1h” due to reflections are in this table. The increases are to the “with barrier” LAeq1h values in the main sound-level results table. However, FHWA TNM does not add these increases to those “with barrier” levels. If new analysis locations have been added to the parallel bar- rier case that are not among the main FHWA TNM receivers, these new analysis locations will not have LAeq1h calculated for them in the main TNM run. However, these analysis loca- tions may represent nearby receivers in the main run for which there are results. The analyst should be wary of computed increases of 0.0 dB, especially in sound-reflecting cases. Sometimes, when the analysis location Z coordinate is below the road- way Z coordinate, the computed increases may be incorrectly computed as zero. This problem appears to be random. Sometimes, “grabbing” all of the graphical objects in the parallel barrier view and moving them up or down very slightly will correct the problem. Alternatively, if the anal- ysis location is slightly below the roadway elevation, the analysis location’s elevation could be adjusted to move it slightly above the roadway elevation. The TNM Users Guide suggests using these LAeq1h increases as adjustment factors in the main TNM run for those receiv- ers represented by these analysis locations. If this is done after calculation of levels by the main part of FHWA TNM, the calculated levels would be invalidated and have to be recal- culated to include these factors. Used in this manner, the “no barrier” and “with barrier” LAeq1h will be increased, which could lead to designing taller walls to get back down to the pre-reflections “with barrier” levels. However, as seen in this research, raising the wall heights in a parallel barrier situation may increase the multiple reflections sound-level increase for certain receiver positions, negating the effect of the raised wall heights on overall noise reduction, and likely requiring a re-analysis in the parallel barrier module. The preferred alternative is to not use the parallel barrier sound level increases as adjustment factors, but to use the parallel barrier module as a design tool to analyze the effects of sound-absorbing materials or tilting one or more of the walls outward. To test sound absorption, changes would be made to the NRC values in the parallel cross section input dialog box. However, if the parallel cross section input dialog box is already open, its data will be for the previously remembered case, as indicated by the name in the input dialog box win- dow banner. This input dialog box must be closed and then reopened before changing data such as the NRC. If this is not done, the computed increases for the new case will be based on the old data (and thus will not change from their previous values), even though the parallel analysis location table will show the new design’s name. Note that FHWA TNM will not accept an NRC greater than 0.95 even though some products report higher values. An NRC of 0.05 is typically used for reflective materials such as concrete. An approximate NRC for grassy areas within the cross section would be 0.4. One final note: after an input check in the parallel barrier module, FHWA TNM will show a message box stating that the “Current data is valid. Discard and recalculate?” Actu- ally, the data are not valid because they were just changed; the case must be recalculated to see the effects of the input data changes.