National Academies Press: OpenBook

How Weather Affects the Noise You Hear from Highways (2018)

Chapter: Chapter 3 - Conceptual Models and Tools

« Previous: Chapter 2 - Measurement and Modeling of Roadway Noise
Page 42
Suggested Citation:"Chapter 3 - Conceptual Models and Tools." National Academies of Sciences, Engineering, and Medicine. 2018. How Weather Affects the Noise You Hear from Highways. Washington, DC: The National Academies Press. doi: 10.17226/25226.
×
Page 42
Page 43
Suggested Citation:"Chapter 3 - Conceptual Models and Tools." National Academies of Sciences, Engineering, and Medicine. 2018. How Weather Affects the Noise You Hear from Highways. Washington, DC: The National Academies Press. doi: 10.17226/25226.
×
Page 43
Page 44
Suggested Citation:"Chapter 3 - Conceptual Models and Tools." National Academies of Sciences, Engineering, and Medicine. 2018. How Weather Affects the Noise You Hear from Highways. Washington, DC: The National Academies Press. doi: 10.17226/25226.
×
Page 44
Page 45
Suggested Citation:"Chapter 3 - Conceptual Models and Tools." National Academies of Sciences, Engineering, and Medicine. 2018. How Weather Affects the Noise You Hear from Highways. Washington, DC: The National Academies Press. doi: 10.17226/25226.
×
Page 45
Page 46
Suggested Citation:"Chapter 3 - Conceptual Models and Tools." National Academies of Sciences, Engineering, and Medicine. 2018. How Weather Affects the Noise You Hear from Highways. Washington, DC: The National Academies Press. doi: 10.17226/25226.
×
Page 46
Page 47
Suggested Citation:"Chapter 3 - Conceptual Models and Tools." National Academies of Sciences, Engineering, and Medicine. 2018. How Weather Affects the Noise You Hear from Highways. Washington, DC: The National Academies Press. doi: 10.17226/25226.
×
Page 47
Page 48
Suggested Citation:"Chapter 3 - Conceptual Models and Tools." National Academies of Sciences, Engineering, and Medicine. 2018. How Weather Affects the Noise You Hear from Highways. Washington, DC: The National Academies Press. doi: 10.17226/25226.
×
Page 48
Page 49
Suggested Citation:"Chapter 3 - Conceptual Models and Tools." National Academies of Sciences, Engineering, and Medicine. 2018. How Weather Affects the Noise You Hear from Highways. Washington, DC: The National Academies Press. doi: 10.17226/25226.
×
Page 49
Page 50
Suggested Citation:"Chapter 3 - Conceptual Models and Tools." National Academies of Sciences, Engineering, and Medicine. 2018. How Weather Affects the Noise You Hear from Highways. Washington, DC: The National Academies Press. doi: 10.17226/25226.
×
Page 50
Page 51
Suggested Citation:"Chapter 3 - Conceptual Models and Tools." National Academies of Sciences, Engineering, and Medicine. 2018. How Weather Affects the Noise You Hear from Highways. Washington, DC: The National Academies Press. doi: 10.17226/25226.
×
Page 51
Page 52
Suggested Citation:"Chapter 3 - Conceptual Models and Tools." National Academies of Sciences, Engineering, and Medicine. 2018. How Weather Affects the Noise You Hear from Highways. Washington, DC: The National Academies Press. doi: 10.17226/25226.
×
Page 52
Page 53
Suggested Citation:"Chapter 3 - Conceptual Models and Tools." National Academies of Sciences, Engineering, and Medicine. 2018. How Weather Affects the Noise You Hear from Highways. Washington, DC: The National Academies Press. doi: 10.17226/25226.
×
Page 53
Page 54
Suggested Citation:"Chapter 3 - Conceptual Models and Tools." National Academies of Sciences, Engineering, and Medicine. 2018. How Weather Affects the Noise You Hear from Highways. Washington, DC: The National Academies Press. doi: 10.17226/25226.
×
Page 54
Page 55
Suggested Citation:"Chapter 3 - Conceptual Models and Tools." National Academies of Sciences, Engineering, and Medicine. 2018. How Weather Affects the Noise You Hear from Highways. Washington, DC: The National Academies Press. doi: 10.17226/25226.
×
Page 55
Page 56
Suggested Citation:"Chapter 3 - Conceptual Models and Tools." National Academies of Sciences, Engineering, and Medicine. 2018. How Weather Affects the Noise You Hear from Highways. Washington, DC: The National Academies Press. doi: 10.17226/25226.
×
Page 56
Page 57
Suggested Citation:"Chapter 3 - Conceptual Models and Tools." National Academies of Sciences, Engineering, and Medicine. 2018. How Weather Affects the Noise You Hear from Highways. Washington, DC: The National Academies Press. doi: 10.17226/25226.
×
Page 57

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.

42 Four approaches or tools to adjust modeled sound levels for meteorological effects are: • Statistical regression models, • Look-up tables, • Engineering models, and • PE models. Each approach uses meteorological parameters in diverse ways and in varying degrees of com- plexity. This chapter will discuss these parameters and approaches and the costs and benefits associated with each. Meteorological Parameters This project measured and assessed several meteorological parameters. This section discusses the most useful meteorological parameters considered and describes how data for these parameters can be obtained or estimated. As noted previously, the primary driver for determining sound levels—other than distance, ground effects, and obstructions—is the vertical (and horizontal) profile of sound speed (see Chapters 1 and 2). Increasing sound speed with elevation bends sound rays toward the ground, and the opposite bends rays upward. Bending sound toward the ground usually results in greater sound levels, and upward bending usually results in the lowering of sound levels at the receiver, although there are exceptions in certain circumstances. The primary drivers of the sound speed profile are the vertical profiles of wind speed (wind shear), wind direction, and temperature (lapse rates). In addition to the sound speed profile, turbulence in the atmosphere can scatter sound. This is especially important in the shadow zone, where sound levels will tend to increase with greater turbulence. Finally, temperature and humidity will affect the absorption of sound in the atmosphere but will have negligible effects on refraction. Atmospheric absorption is most effective with higher frequency sound (see Figure 1). Wind As discussed in Chapter 1, wind speed alone does not directly affect sound propagation. How- ever, several characteristics of wind speed do. First, the vector wind speed profile directly affects the sound speed profile. The vector wind speed is the component of wind that is moving from the source to the receiver. The profile is how that wind speed changes with the height above ground. Second, the turbulence of the air will affect the sound levels. Turbulence is also generally proportional to wind speed, especially near the ground. Wind shear will also affect turbulence; higher wind shear will generally result in greater turbulence. C H A P T E R 3 Conceptual Models and Tools

Conceptual Models and Tools 43 While the vertical wind speed profile is important for determining the sound speed profile, it is not directly measured at most National Weather Service weather stations. Wind shear can be measured by anemometers at multiple heights, or through Lidar, Sodar, or weather balloon radiosonde measurements. The research team used a combination of anemometers and Lidar for this project. Since wind shear is not directly measurable at a reasonable cost, a proxy is necessary. Many models, such as Harmonoise and CONCAWE, use wind speed alone, since wind speed and wind shear are mostly proportional near the ground. However, the amount of shear can change substantially due to the ground roughness and other site and meteorological characteristics. AERMET Meteorological Data One of the problems in determining atmospheric effects on traffic noise has been access to the meteorological data required. One primary source of these data, which is time and location specific, is from air quality personnel at state environmental and transportation agencies. For industrial source air pollution dispersion modeling, the U.S. EPA guideline preferred model is AERMOD [110]. In addition, the U.S. EPA has been transitioning in recent years to the use of AERMOD for dispersion modeling of highways. The dispersion modeling process requires detailed meteorology and a precursor model to AERMOD to supply this information is called AERMET [96]. While the software is usually used as an input to the AERMOD air quality dispersion model, the output is extensive and lends itself to modeling meteorological effects on highway noise. Since the U.S. EPA requires the use of AERMET data for air dispersion models, data used in AERMET is available for most major U.S. airports and regions. AERMET is used as a general purpose meteorological preprocessor that utilizes available mete- orological data. Hourly surface observations from the National Weather Service (NWS), FAA, and even local data from a data site-specific meteorological measurement program (such as that collected in this study) can be combined with NWS twice-daily upper air soundings to derive location-specific meteorological conditions. The research team has derived a methodology for use of this meteorology data to supply the needs for determining propagation impacts on traf- fic noise. Two output files are written from AERMET: a file of hourly boundary layer parameter estimates (surface output file) and a file of multiple-level observations of wind speed, wind direc- tion, temperature, and standard deviation of the fluctuating components of the wind (profile output). Table 9 shows the information included from the surface output file and Table 10 shows the information from the profile output. Using AERMET data, one can calculate the temperature as a function of height using equations found in references 41, 111, and 118: Equation 2. Similarity theory—calculation of vertical wind speed profile. ( )( ) = + − ϕ * 00u z uk ln z zz zLM where: u(z) = the wind speed at height z; u* = the friction velocity; z0 = the roughness length; k = the von Karman constant (0.41); ( )ϕ zLM = a diabatic momentum profile correction (mixing) function; and L = the Monin-Obukhov length.

44 How Weather Affects the Noise You Hear from Highways As part of this project, the research team created a spreadsheet that uses an AERMET file to estimate the wind speed profile using the formula in Equation 2. This is discussed further in the following section. Temperature Like wind, temperature alone does not substantially affect the refraction of sound, although it does affect the absorption of sound in the atmosphere. (Atmospheric absorption is a well-understood Field Description Units 1–5 Year (two-digit), month, day, Julian day, and hour As noted 6 Sensible heat flux W/m2 7 Surface friction velocity m/s 8 Convective velocity scale (set to -9.0 for stable atmosphere) m/s 9 Potential temperature gradient above the mixing height K/m 10 Convectively driven mixing height (-999. for stable atmosphere) m 11 Mechanically driven mixing height (computed for all hours) m 12 Monin-Obukhov length m 13 Surface roughness length (month and wind direction dependent) m 14 Bowen ratio (month and wind direction dependent) Nondimensional 15 Albedo (month and wind direction dependent; 1.0 for hours before sunrise or after sunset) Nondimensional 16–18 Wind speed, wind direction, and anemometer height that were used in the computations during processing m/s degrees, m 19–20 Temperature and measurement height that were used in the computations during processing K and m 21 Precipitation type code 22 Precipitation amount mm/hr 23 Relative humidity % 24 Station pressure mb 25 Cloud cover Tenths 26 Wind speed adjustment and data source flag 27 Cloud cover and temperature substitution by interpolation Source: RSG for NCHRP Project 25-52. Table 9. AERMET surface file output. Field Description Units 1–4 Year (2-digit), month, day, and hour 5 Measurement height m 6 Indicator flag: 1=last level in profile for the hour, 0=not the last level 7–8 Wind direction and speed m/s, m 9 Temperature °C 10 —standard deviation of the lateral wind direction degrees 11 w—standard deviation of the vertical wind speed m/s Source: RSG for NCHRP Project 25-52. Table 10. AERMET profile file information.

Conceptual Models and Tools 45 phenomenon and already considered in all engineering sound models.) However, the vertical temperature profile will directly affect the sound speed profile. As discussed previously, increasing temperature with elevation will refract sound toward the ground and decreasing temperature with height will refract sound away from the ground. As with the wind profile, there is no straightforward way to measure or obtain temperature profile for a specific site. This study used a temperature profiler along with precision thermom- eters mounted on a 10-meter tower. As a proxy for temperature profile, practitioners can use the following parameters: • Heat flux, which is the amount of heat radiated (or absorbed) by the ground. Heat flux is calculated by AERMET and is a function of cloud cover, solar angle, and albedo (reflectivity of the ground). • Day vs. night with cloud cover. Clear days will tend to have temperature profiles that increase with elevation and clear nights will tend to result in temperature inversions. Cloudy days and nights are closer to adiabatic. The Harmonoise model estimates a stability class based only on these parameters. • Inverse Obukhov Length. The inverse Obukhov Length is a measure of atmospheric stabil- ity and calculated directly by AERMET from NWS data. Negative values represent unstable atmospheres and positive values represent stable atmospheres. References 41, 111, and 118 also provide equations to estimate the vertical temperature profile data that can be obtained through AERMET files. One can calculate the vertical temperature gradient using the following: Equation 3. Similarity theory—calculation of vertical temperature profile. * 0 0 0 T z T T k ln Z z z z L zH ( )( ) = + + + ϕ  + Γ where: T(z) = the temperature at height z; T* = a turbulent temperature scale; ( )ϕ zLM = the diabatic heat profile correction (mixing) function; and Gz = -0.01 °C/m is the adiabatic lapse rate As part of this project, the research team created a spreadsheet that takes an AERMET file and estimates the temperature profile using the above formula. This will be discussed further in the following section. Turbulence Turbulence substantially affects the refraction of sound by scattering sound in the atmo- sphere. Higher turbulence generally increases sound levels in the shadow zone, but has a much lower impact elsewhere. Obstructions, such as noise barriers, will affect both the wind speed profile and turbulence intensity. Thus, higher turbulence will increase highway traffic noise in the shadow zone of a barrier. Turbulence is generally a function of height above ground, with the greatest turbulence near the ground. As with wind and temperature, turbulence intensity (which is the standard deviation of wind speed divided by wind speed) is not directly reported by the NWS. As such, practitioners must use a proxy. Generally, Stability Class is a good proxy for turbulence, as less stable atmospheres result in more vertical mixing and turbulence. How- ever, this is not always the case, especially with turbulence that is associated with high wind

46 How Weather Affects the Noise You Hear from Highways shear in stable conditions. Alternatively, turbulence intensity can be directly measured using relatively inexpensive logging anemometers. Tools to Adjust Highway Sound Levels for Meteorology As noted, there are conceptually four options for creating a tool for estimating the adjusting modeling under neutral conditions to account for meteorological effects. These are described in the following section. AERMET Sound Speed Profiler Spreadsheet Tool As noted, AERMET data are used by state environmental and highway agencies to provide meteorological data as an input to air dispersion models. It can also be used to estimate the prob- ability of adverse meteorological conditions on highway noise at a particular locale. This section describes a conceptual screening tool to calculate these impacts. Air quality analysts commonly use at least one year of hourly data, or multiple years of averaged hourly data, in dispersion analysis to develop a true picture of the local conditions. Short-term measurements are good when monitoring, but these limit prediction because there is no way to be sure the short-term measurements are representative of the local area. The longer-term data reduces the chances of non-representative data. For a highway noise analysis, while each hour of the year could be analyzed, this is not needed if the focus is only on the “worst” traffic hour. If one considers the need for impact analysis of highway noise, then only the impact on the worst hour of the day needs to be considered [97]. The regulations (noise standards) promulgated by FHWA only require the greatest one-hour noise level to determine if impacts occur (described in more detail in Chapter 4). The worst hour is directly related to traffic flow (volume and speed) and the exact hour is determined by the noise analyst or a theoretical worst case. As such, the noise analyst only needs to know the prob- ability of a change due to atmospheric effects in this general time of the day. Since inversions can have a large effect on propagation and occur more often in the early morning and evening periods, this provides a first indication of what hours should be chosen for consideration. Add to this that the worst traffic hour noise is often just before, during, or after the morning peak rush hour period and it becomes apparent that the morning period is the time that should be analyzed. The morning period is the time between sunrise to midmorning, considering both the meteorology and the traffic flow. By reviewing these time periods, three possible scenarios may be considered: • The average annual meteorological effect, • The maximum meteorology effect, and • The probability, or number of hours, of “worst-case” meteorology. Additionally, these scenarios could be calculated by season to account for local receptor impacts. The average impact that occurs during this time period, as described in this methodol- ogy, is not truly average but is instead an average of the maximum effects that occur each day of the year or season during the morning period. The maximum is the largest meteorology effect on propagation that occurs for the entire year during the time frame of the worst hour. The average is probably the most representative of a normal day and could be used in analyses. However, the maximum could be used to determine the largest impact that could also reasonably be expected. Additionally, when dealing with the public, the question of the maximum impact on local noise levels often occurs. This methodology supplies the answers to these scenarios.

Conceptual Models and Tools 47 The methodology is based on the evaluation of the three atmospheric parameters related to noise propagation: wind speed changes with height (wind shear), temperature changes with height (lapse rate), and stability. Practitioners must evaluate each of these three parameters to determine the correct acoustic atmospheric classification to allow the meteorological adjust- ment factor (MAF) to be determined. Once the acoustic atmospheric classification has been determined, the MAF can be determined based on the results of the extensive measurements accomplished for this research. In addition, practitioners may want to look beyond these scenarios to determine the actual impact of meteorology on the traffic noise propagation on a specific day and a specific time. In this case, local data or NWS data, coupled with the use of AERMET, could be used to make the determination of the MAF. For example, when measurements are made to calibrate a model such as TNM, often there are significant differences. By using this methodology to derive the impacts of meteorology, one source of error can be reduced. In these cases, the user would need to locate the appropriate hours and follow the procedure to determine the correct class and adjustment factor. Considerable thought went into determining an accurate yet useable methodology, since there are several ways the three required meteorological parameters (wind shear, lapse rate, and stability) could be determined. For example, stability could be estimated simply using cloud cover and night/day. Additional details could be added by adding in wind speed, the standard deviation of the wind speed, and/or the inverse Monin-Obukhov length, which is a good mea- sure of stability. All are available in the AERMET output. (The height above the ground where turbulence is generated more by buoyancy than by wind shear.) The same principle applies to wind shear and lapse rate with multiple choices being available from the AERMET dataset. The research team has chosen to use similarity theory, defined by Equations 2 and 3. The research team has developed a simple spreadsheet program to assist the noise analyst in estimating the frequency of upward and downward refracting atmospheres at a location using AERMET data. In this “AERMET Sound Speed Profiler” spreadsheet tool, an AERMET surface data (.SFC) file is loaded. Then, the user specifies the wind direction from due north, from the source to receiver. For example, if the receiver is due east of the source, the wind direction would be 90°, and due northeast would be 45°. The workbook then calculates the vertical sound speed profile for every hour of the year. The result is a frequency distribution of sound speed profiles by hour of the day. In the example output in Figure 42, downward refracting conditions are shown in darker colors and upward refracting conditions are shown in lighter colors. In this case, there is a strong tendency for the downward refracting conditions to occur during the morning peak traffic hour and upward-refracting conditions to occur during the PM peak traffic hour. Practitioners can this use this information, in combination with expected sound levels from the highway of interest, to assess what might be the impact on the worst hour from meteorological conditions. Statistical Models In a statistical model, the meteorological effect would be estimated using observed data. An equation is derived that best describes the sound levels measured as a function of the observed meteorological variables. A regression method is used to provide the best fit of the model to the observations. In Appendix D, the research team describes several regression models that worked well for predicting sound level difference from the 15-meter microphone using combinations of distance, vector wind speed, turbulence, heat flux, Monin Obukhov length, or sound speed profile. The

48 How Weather Affects the Noise You Hear from Highways model that was most useful is one that uses data that is readily available from AERMET using NWS data. Note that the effective sound speed gradient can be calculated using the AERMET Sound Speed Profiler spreadsheet tool described in this chapter. The parameters in the model include: • The log base 10 of distance from the center of the closest travel lane of the highway • The effective sound speed gradient near the ground (1.8 to 10 meters), treated as a third order polynomial, in units of s-1 (c_effp) • The effective sound speed gradient at a moderate height (30 to 44 meters), in units of s-1 (c_effp30) Source: RSG for NCHRP Project 25-52. 0% 20% 40% 60% 80% 100% 0: 00 1: 00 2: 00 3: 00 4: 00 5: 00 6: 00 7: 00 8: 00 9: 00 10 :0 0 11 :0 0 12 :0 0 13 :0 0 14 :0 0 15 :0 0 16 :0 0 17 :0 0 18 :0 0 19 :0 0 20 :0 0 21 :0 0 22 :0 0 23 :0 0 Pe rc en t of ti m e Time of Day Annual Frequency Distribution of Refraction Due to Wind Profile Strong upward Upward Neutral Downward Strong downward Strong upward Upward Neutral Downward Strong downward 0% 20% 40% 60% 80% 100% 0: 00 1: 00 2: 00 3: 00 4: 00 5: 00 6: 00 7: 00 8: 00 9: 00 10 :0 0 11 :0 0 12 :0 0 13 :0 0 14 :0 0 15 :0 0 16 :0 0 17 :0 0 18 :0 0 19 :0 0 20 :0 0 21 :0 0 22 :0 0 23 :0 0 Pe rc en t of ti m e Time of Day Annual Frequency Distribution of Refraction Due to Temperature Profile Strong upward Upward Neutral Downward Strong downward 0% 20% 40% 60% 80% 100% 0: 00 1: 00 2: 00 3: 00 4: 00 5: 00 6: 00 7: 00 8: 00 9: 00 10 :0 0 11 :0 0 12 :0 0 13 :0 0 14 :0 0 15 :0 0 16 :0 0 17 :0 0 18 :0 0 19 :0 0 20 :0 0 21 :0 0 22 :0 0 23 :0 0 Pe rc en t of ti m e Time of Day Annual Frequency Distribution of Refraction Due to Effective Sound Speed Profile Figure 42. Example output from the AERMET sound speed profile calculator.

Conceptual Models and Tools 49 • The interaction of distance with the sound speed gradient, since the distance from the highway affects the shape of the sound speed gradient effect (LogMicDist * c_effp30) The model results are shown in Table 11. An absolute value of the T-statistic greater than 2.0 is significant at a 95th percentile level. In this case, the Barrier model is generally a poor fit of the observed data, while the No-Barrier model is a good fit. The overall R-squared value for the No-Barrier model is 0.91, indicating 91% of the variability is explained by these parameters. Figure 43 compares the measured A-weighted sound level difference between the 480-meter and 15-meter No-Barrier microphone positions (1.5-meter height) with the regression model estimate. Results for each microphone position are shown in Model 7 of Appendix C. As shown, the model does well in predicting the observed sound level difference, with the exception of the most negative extremes that tended to occur around the afternoon peak hour. In addition, the research team modeled individual 1⁄3 octave bands, which also showed similarly meaningful results. In this way, the model can be used for any combination of cars and trucks with different spectral characteristics. The spectral model results are shown in Appendix C. The benefit of using a statistical model is that it is relatively simple and makes use of available data. The downsides are that it is not physics based, is only validated for the site measured, and cannot be used to assess the meteorological effect at locations with highway barriers. Look-Up Tables Practitioners can use simple look-up tables to combine several different parameters to estimate the change in sound levels due to meteorology. This approach is used in NCHRP Report 791, where several tables were generated. Each table shows sound levels under strong and weak lapse/inversion, and strong and weak upwind/downwind winds, by distance (15.2 meters to 487 meters), for a total of eight meteorological classes. The tables varied in their ground type (soft and hard), and vehicle types (autos only vs. autos and trucks). In addition, a set of scenarios were run in the presence of a noise barrier. Independent Variable Barrier No-Barrier n = 1,759; R^2 = 0.25 n = 13,955; R^2 = 0.91 Standardized Coefficient T stat Standardized Coefficient T stat Intercept -8.37 -11.1 22.7 172 Log (mic distance) -0.52 -20.7 -1.26 -335 c_effp (2 to 10 m) -0.71 -0.83 16.8 43.6 c_effp2 (2 to 10 m) -12.9 -4.36 -2.26 -2.06 c_effp3 (2 to 10 m) 54.3 3.76 -112 -22.1 c_effp30 (30 to 44 m) -76.7 -4.13 -96.9 -44.9 LogMicDist * c_effp30 2.98 4.81 3.84 62.6 Source: RSG for NCHRP Project 25-52. Table 11. Model developed to calculate the difference between 15 meters and microphone positions.

50 How Weather Affects the Noise You Hear from Highways The two drawbacks of the NCHRP Report 791 tables are that they use the temperature lapse rate, which is not easily measured, and the tables did not show combinations of meteorological conditions. That is, the NCHRP Report 791 tables can only be used where one of the two follow- ing conditions are true: the vector wind speed is zero or the temperature lapse rate is zero. In this study, the research team addressed these issues by modeling combinations of six wind classes with five stability classes with Harmonoise, as in Table 4. Measured or Modeled Data? While the look-up tables using modeled data are useful, the monitored data collected in this study did not correlate well to those models. Both the Nord2000 model from NCHRP Report 791 and the Harmonoise model results show a much greater correction to be applied to account for the meteorological effects. As such, the research team investigated whether it would be best to create look-up tables based on the modeled or monitored meteorologi- cal effect. While merits of each approach (use of measurements vs. modeling) could be argued, the research team concluded that the corrections based on the measurements would be more appli- cable for the purposes of this study. The research team’s conclusion was based on the following reasons. First, the research team’s experience informed the understanding that the application of large corrections could lead to misleading final conclusions. Using the large correction fac- tors from the models would represent sound levels that are unlikely in most areas where TNM would be applied. Second, the research team also concluded that measurements are generally a better source of information for the corrections because of modeling limitations that occur due to Source: RSG for NCHRP Project 25-52. Figure 43. Observed sound levels (in black) compared to predicted (in red) at the 480-meter No-Barrier microphone position.

Conceptual Models and Tools 51 simplifications made during the modeling process. The simplifications occur for the follow- ing reasons: • In the measurements, background levels existed and even though low still represent a “floor” to the measurements like what is expected in TNM modeling. Modeling does not have this limitation. • Models represent a more homogeneous atmosphere than what occurs in practice. Changes with the wind and temperature not only occur with height but laterally during measurements. This is not completely accounted for by the models. • Other simplifications occur in modeling such as ground effects, diffraction, and vegetation that present real effects in the measurements. To further substantiate the limits of the meteorological effect, the research team reviewed data obtained as part of the Volpe TNM validation [112]. While the TNM validation was carefully performed in that study, meteorological effects were not probed to explain variances around average measurements. The research team reviewed measurement periods from this dataset that experienced higher wind speeds. Unfortunately, wind shear or lapse rate was not reported, but the higher wind speeds would tend to include meteorological effects from wind shear. The research team observed that during these higher wind periods, the differences between TNM predictions and measurements increased, further supporting the hypothesis that meteorological effects were occurring. In general, the research team observed that higher wind speeds at one location resulted in differences between TNM predictions and measurements to be 4 to 6 dBA Leq at 15.2 meters, 1.5 meters above the roadway plane and 4 to 5 dBA Leq at 46 meters from the roadway both at 1.5 and 4.6 meters above the roadway plane. At a second location, with greater wind speeds, differences between measurements and TNM modeling increased as expected due to greater meteorological effects. The differences were 6 to 8 dBA Leq for distances ranging from 15.2 meters to 183 meters for both 1.52 meters and 4.6 meters above the ground plane. The difference between modeling and observed sound levels increased with increasing wind speed represent- ing refraction effects due to the local meteorology. The results tend to be more in line with the measured correction factors than the modeled correction factors. Other measurements also support this idea. During an extensive, extended testing along I-10 just west of Houston, Texas, Wayson and Bowlby (1990) measured sound levels along with detailed meteorology at varying distances from the roadway [33]. After normalization for refrac- tive effects, they found that the meteorological conditions caused changes in sound levels as shown in Table 12. Based on the detailed meteorology, Wayson and Bowlby found that the three major causes of refraction (turbulence, wind shear, and lapse rate) sometimes caused increased effects and sometimes tended to negate each other. More importantly for this research, their values are more in line with this research team’s measurement corrections found in this study than is the modeling. Based on experience, modeling limitations, and previous measured differences, the research team concludes that the corrections based on the measured data are more representative of what may occur in typically modeled areas and should be used going forward within the scope of this research. Look-up Tables Using Measured Meteorological Effects To compile the look-up tables included here, the research team normalized all measurement data to 5,000 vehicles per hour (vph) with the project average traffic mix and speed, effectively

52 How Weather Affects the Noise You Hear from Highways eliminating the variation in traffic type and flow. FHWA TNM Reference Energy Mean Emis- sion Level (REMEL) [119] sound powers were calculated for the normalization procedure and applied to the measurement data, Lpmeasured, in the following manner: Equation 4. Calculation of sound pressure levels normalized to 5,000 vph. = + − ( ) ,5,000 5,000Lp Lp Lw Lwvph measured REMEL vph REMEL actual where Lw(REMEL actual) is the sound power level of the actual vehicle mix and volume for that 5-minute period. Table 13 reports the average vehicle mix for the entire measurement period, which was utilized to calculate LwREMEL 5,000vph. In addition to Table 4, the research team created several look-up tables based on the tempera- ture, sound speed profile, and vector wind speed. These tables are like those found in NCHRP Report 791, but consider combinations of these parameters: • Table 14 shows the meteorological effect as a function of the vector wind speed (at 10 meters) and the vertical temperature profile. As noted, while temperature and wind profiles are not readily available, the AERMET parameters and the Excel tool that can be downloaded from the TRB website can help calculate these profiles. • Table 15 shows the meteorological effect only as a function of the temperature profile. It includes all measurements, not just those under zero-wind conditions. Mic C Mic D Mic E Mic F Mic G Mic H Distance from Roadway 61 meters 61 meters 61 meters 1,221 meters 1,221 meters 1,221 meters Height above Roadway 10 meters 3 meters 1.5 meters 10 meters 3 meters 1.5 meters Maximum 3.4 2.3 1.8 4.9 6.5 5.2 Minimum -0.9 -0.9 -1.2 -1.0 -1.2 -2.2 Average 0.7 0.2 0.0 1.3 1.3 0.0 Std. Dev. 1.0 0.7 0.8 1.5 2.0 2.0 Source: Data from Wayson and Bowlby [33], formatting from RSG for NCHRP Project 25-52 Table 12. Measured refraction (meteorology) effects, various cases from Wayson and Bowlby (1990) [33]. Vehicle Type Percentage Passenger Car 90.6 Bus 0.6 Medium Truck 3.8 Heavy Truck 4.4 Motorcycle 0.6 Source: RSG for NCHRP Project 25-52. Table 13. Average vehicle mix on I-17 during field measurements.

Conceptual Models and Tools 53 • Table 16 shows the meteorological effect only as a function of the 10-meter wind speed. It includes all measurements, not just those under zero temperature lapse conditions. • Table 17 shows the meteorological effect only as a function of the effective sound speed pro- file. In other words, this table combines the vector wind speed profile and temperature profile effects into five effective sound speed categories. The categories relate the resulting refraction— upwind, downwind, and neutral. While effective sound speed is not readily available for a loca- tion, the Excel tool on the TRB website can help calculate this based on AERMET parameters. Look-up tables are relatively easy to use. However, they are general results and may not match the characteristics of a specific site. They should not be used with noise barriers, since the sound levels in the shadow zone are highly dependent on the barrier length, height, and relative distance to the road and receiver. Vector Wind Speed at 10 m Temp Profile Distance to Highway (m) 15 30 60 120 240 480 960 Str. lapse – – – – – – – Strong Lapse – – – – – – – Downwind Zero – – – – – – – Inversion – – – – – – – Strong inv. – – – – – – – Str. lapse 0 -1 -2 -2 -4 -6 -6 Weak Lapse 0 -1 -1 -2 -3 -4 -5 Downwind Zero 0 1 1 3 5 5 4 Inversion 0 1 2 4 7 6 2 Strong inv. 1 1 2 5 8 5 -1 Str. lapse 0 -2 -2 -3 -5 -7 -7 Lapse 0 -1 -2 -2 -4 -6 -6 No Wind Zero 0 0 0 0 0 0 0 Inversion 0 1 1 2 4 4 2 Strong inv. 0 1 1 2 3 2 0 Str. lapse 0 -3 -3 -4 -6 -6 -7 Weak Lapse 0 -2 -3 -5 -7 -8 -8 Upwind Zero 0 -1 -2 -3 -4 -7 -8 Inversion 0 0 0 -1 -2 -5 -5 Strong inv. 0 0 1 0 -1 -6 -9 Str. lapse 0 – -4 -6 -6 -4 – Strong Lapse 0 -3 -4 -5 -6 -4 -5 Upwind Zero 0 – -2 -3 -4 -4 -4 Inversion 1 – -2 -3 -3 -3 -3 Strong inv. – – – – – – – * Gray shading shows sound levels decreasing because of meteorology. Unshaded is no change from neutral or an increase in sound levels. Dashes indicate no data. Source: RSG for NCHRP Project 25-52. Table 14. Look-up table of sound levels relative to acoustically neutral conditions based on measurements at No-Barrier location (in dB).*

54 How Weather Affects the Noise You Hear from Highways Temperature Profile Distance to Highway (m) 15 30 60 120 240 480 960 Strong lapse 0 -2 -3 -4 -6 -6 -7 Lapse 0 -2 -2 -3 -5 -6 -6 Zero 0 0 0 1 1 0 0 Inversion 0 1 2 3 5 5 2 Strong inversion 0 1 1 2 4 2 -1 * Gray shading shows sound levels decreasing because of meteorology. Unshaded is no change from neutral or an increase in sound levels. Source: RSG for NCHRP Project 25-52. Table 15. Look-up table of sound levels according to temperature profile relative to acoustically neutral conditions based on measurements at No-Barrier location (in dB).* Vector Wind Speed at 10 M Distance to Highway (m) 15 30 60 120 240 480 960 Strong downwind – – – – – – – Downwind 0 1 1 3 5 3 0 Zero wind 0 0 0 0 1 1 -1 Upwind 0 -2 -2 -3 -5 -7 -8 Strong upwind 0 -3 -3 -4 -5 -4 -4 * Gray shading shows sound levels decreasing because of meteorology. Unshaded is no change from neutral or an increase in sound levels. Dashes indicate no data. Source: RSG for NCHRP Project 25-52. Table 16. Look-up table of sound levels according to vector wind speed relative to acoustically neutral conditions based on measurements at No-Barrier location (in dB).* Effective Refraction Direction Distance to Highway (m) 15 30 60 120 240 480 960 Strong upward 0 -2 -3 -3 -5 -5 -5 Upward 0 -1 -1 -2 -3 -4 -4 Neutral 0 0 0 0 0 0 0 Down 0 1 2 3 6 7 5 Strong downward 0 1 2 4 7 6 3 * Gray shading shows sound levels decreasing because of meteorology. Unshaded is no change from neutral or an increase in sound levels. Source: RSG for NCHRP Project 25-52. Table 17. Look-up table of sound levels according to effective sound speed profile relative to acoustically neutral conditions based on measurements at No-Barrier location (in dB).*

Conceptual Models and Tools 55 Engineering Models Engineering models are generally based on physical principles and can be used to model diverse types of site-specific conditions. Engineering models tend to generalize, which allows for ease-of-use and reasonable calculation times. TNM is an example of an engineering model that models sound under acoustically neutral conditions. Harmonoise and Nord2000 are exam- ples of engineering models that can also consider meteorological conditions. The benefit of engineering models is that they can be reasonably accurate under a wide range of conditions, including terrain, varying ground characteristics, and noise barriers. The main disadvantage of adapting TNM to include meteorology is the level of effort to code and validate the model. In addition, new policies and guidance would need to be developed to make use of the results. The adaption of TNM to consider meteorological effects can be accomplished, using similar approaches to Harmonoise or Nord2000. However, the specific details regarding the algo- rithms are beyond the scope of this report. This is discussed further in the Framework for TNM Update section. PE Models PE models are the most accurate predictors of meteorological effects. However, PE models require detailed data on the range-dependent sound speed profiles. In addition, PE models are computationally intensive, requiring hours of calculation for a single source-receiver pair. As a result, PE models are not appropriate for engineering use. They can, however, be used on a limited basis to form the foundation of engineering models and to validate mod- els. Appendix D shows how a PE model was used in this research to validate the results of Harmonoise. Framework for TNM Update As noted, the FHWA TNM model does not include meteorological effects, other than atmospheric attenuation. It assumes a non-refracting atmosphere. Updating TNM to include meteorological effects will involve additional research not included in the scope of this research project. However, it is useful to consider what potential options could be explored if FHWA chooses to move forward with this. This report discusses four approaches to account for heterogeneous meteorological effects on sound level. These approaches are con- sidered specifically as they relate to modifications to TNM: research models, engineering mod- els, look-up tables, and statistical models. Research Models Research models are models in which the calculations are closely tied to first principles and fundamental physics. These models include PE models, finite element models, bound- ary element models, and hybrid models, among others. While these models can make precise and accurate predictions, they require long run-times due to the computationally intensive nature of the algorithms. Importantly, and in practice, these models also require detailed data on the meteorological conditions as a function of time and detailed data of the terrain’s geometry and acoustical impedance. If these data are not available, then research models will only be able to provide accurate predictions of a condition that is not present at the position of interest. Engineering Models Engineering models address the issue of run-time by utilizing heuristic algorithms that are still based on physics, but that also make reasonable simplifying assumptions. An engineering-based

56 How Weather Affects the Noise You Hear from Highways model represents the practical state-of-the-art direct implementation that can be achieved in TNM. Such a model would be able to handle multiple terrain features more precisely than could be achieved with look-up tables or regressions while at the same time being able to produce results much faster than a research model. This could be particularly important when a practitioner wants to determine the best choice among several barrier designs where trade-offs between time and accuracy are significant. One way to implement an engineering based model to account for heterogeneous meteorological effects is to replace the straight, non-refracting rays used for the homogenous case with curved rays. This approach has been implemented in several other models. If the wind speed profile is assumed to vary linearly, then curved rays can be simplified as circular rays, which makes their paths more predictable. In so doing, the divergence and atmospheric absorption, phase relationships between different ray paths, and reflections and diffractions can be more accurately computed. Additionally, shadow zones due to refractions can also be computed, which cannot be determined using straight rays. Implementing such a model would require modifications to the TNM graphical user interface (GUI) to allow for entry of pertinent meteorological parameters and to report the new results. However, most of the changes to TNM would be in the acoustics, specifically the vertical acous- tics for each leg of each elemental triangle. Most of the equations currently in TNM’s vertical acoustics would need to be modified or the inputs to the equations would need to be modified. Because of this, there is a risk of introducing bugs to the code during development and thorough validation and testing would be required. Further, care would need to be taken to ensure that as meteorological conditions tended toward homogeneity, the results would tend toward the tra- ditional straight ray acoustics. Finally, it is expected that the run-time could increase somewhat, perhaps by as much as a factor of two or three. While such an effort would be worthwhile overall to create more accurate impact determinations, it would take time before such a version of TNM could be implemented and validated. Look-up Tables and Statistical Models While implementing a new engineering model in TNM to account for heterogeneous meteo- rological effects is worthwhile in the long term, some benefits from this current research could be obtained more immediately by implementing methods to utilize look-up or regression based accounting of heterogeneous meteorological effects. Because these modifications would be sta- tistical in nature, they would not provide the same accuracy as a physics-based model; however, these modifications could still be used to either provide better predictions than those obtained from using a homogeneous model alone or they could be used to bound the true value between reasonable limits. Adjustments from statistical regression models and look-up tables both offer relatively simple and minimal risk methods to account for meteorological effects since they require mostly modification of the GUI/business logic of TNM. Only minor additions to the horizontal acoustics would be necessary to return a few needed parameters, such as distance between source and receiver. The applicability of adjustments from regression models or look-up tables is limited to the range of cases from which the regression or look-up tables are generated. Ideally, the implementation would involve creating look-up tables (or regression models) that are a function of frequency and source height for cases with and without the line-of-sight between the highway and receiver blocked. The cases resulting from the combination of six wind and five stability classes that were computed using Harmonoise (described in Chapter 2) or the look-up tables in this chapter provide useful datasets for computing adjustments for cases where there are not significant line-of-sight block- ages due to hills, berms, or wall type barriers and for one condition with a noise wall. To account for meteorological effects for the general cases where there are significant line-of-sight blockages, a more diverse set of cases over a range of distances and blockage heights must be collected.

Conceptual Models and Tools 57 Adjustments from simple regression models offer relatively fast computation times. In addition, changes in the required adjustments due to subtleties in parameter levels can be smoothed out in the statistical model. For this reason, a look-up table is more desirable if the underlying data are free from random variations, which should be the case for modeled tabulated meteorological effects. After computation of the homogeneous meteorological case in the TNM acoustics, the user could adjust the results for each roadway/receiver pair by interpolating between the relevant parameters in the look-up table. The results would then be aggregated for all roadways for each receiver to obtain a final corrected sound pressure level. Because this method is implemented in the GUI, it would also be possible to compute adjustments for a range of conditions without a significant increase in computation time.

Next: Chapter 4 - Implications for Noise Impact and Abatement »
How Weather Affects the Noise You Hear from Highways Get This Book
×
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

TRB's National Cooperative Highway Research Program (NCHRP) Research Report 882: How Weather Affects the Noise You Hear from Highways documents the meteorological effects on roadway noise propagation under different atmospheric conditions. Highway noise changes from day to day and hour to hour—not just because of variations in traffic volumes, vehicle mix, and speed, but also because of the weather. The report develops guidance to identify when atmospheric conditions should or should not be considered in noise analyses.

The report is accompanied a PowerPoint presentation and a tool called the AERMET sound speed profile calculator. The report also includes a brochure designed to communicate the concepts of the research to non-technical audiences. The brochure is made available in MSWord format to enable customization and the ability to insert an official logo and contact information. An Interactive Tool is also available for download. The interactive tool includes audio files that allow the user to hear differences in highway noise under various meteorological conditions.

Disclaimer: This software is offered as is, without warranty or promise of support of any kind either expressed or implied. Under no circumstance will the National Academy of Sciences or the Transportation Research Board (collectively "TRB") be liable for any loss or damage caused by the installation or operation of this product. TRB makes no representation or warranty of any kind, expressed or implied, in fact or in law, including without limitation, the warranty of merchantability or the warranty of fitness for a particular purpose, and shall not in any case be liable for any consequential or special damages.

Original data used to develop NCHRP Research Report 882 are available upon request. Send requests via email to Ann Hartell, ahartell@nas.edu, and include a short explanation of the intended use of the data (for example, name of research project, research sponsor, affiliation and location of research team, and general plan for publication of results).

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!