The adverse effects of construction noise upon a community have historically been considered to be an inevitable, short-term, and necessary impact. Construction noise is evaluated in a quantitative manner on relatively few occasions and specific noise commitments are rarely included in a project's environmental documents. Typically, construction noise control commitments only include such common sense actions as ensuring that all vehicles have proper mufflers, trying to schedule work to be least disturbing, possibly erecting project noise barriers early in the construction process, and promising to comply with any and all local noise ordinances (which often exempt construction noise).
In the early 1970s, the EPA released several construction noise reports as well as the often-referenced EPA "Levels Document"ref033. These publications offered the nation's first suggested criteria limits for community noise, which were considered to apply to construction noise as well. The EPA's criteria limits were expressed in long-term noise units of both Leq(24h) and Ldn for various degrees of potential hearing interference or annoyance.
In 1977, a landmark report prepared by Bolt Baranek & Newman, Inc. for the Empire State Electric Energy Research Corporation (ESEERCo) summarized the noise producing sources, phases of work, and potential mitigation options for construction noise. The ESEERCo guide also provided empirically based relationships of construction noise sources and associated noise levels, and contained fairly extensive noise emissions tables for specific pieces of equipment as a function of horsepower. The guide recommended that construction noise be assessed using the Leq(h) and Ln percentile metrics.
FHWA's first attempt to create and distribute a standardized construction noise model resulted in the early-1980s release of the Highway Construction Noise Model (HICNOM)ref091. It consisted of look-up charts, programmable calculator routines, and a computer model. The supporting noise studies were conducted in the late-1970s, concurrent with several of the aforementioned construction noise studies. At the time, FHWA recommended that construction noise should only be evaluated in a general manner for most projects, but that the HICNOM model could be used for any highly complex or controversial project. The model was originally written in FORTRAN and was later simplified to run as an executable file on IBM-compatible computers. The program required the generation of an "input file" in which assumptions, geometries, and equipment specifics were entered. The resulting "output file" provided the user with the overall Leq level at a receptor position as well as the Leq contribution for each piece of equipment.
From the early-1980's until the release of the FHWA Roadway Construction Noise Model (FHWA RCNM)ref083 in 2006, there was very little, if any, substantive investigation of construction noise by the FHWA or any other federal agency. Today's modeling methods typically make use of algorithms promulgated by EPA in the early 1970's, in which the reference noise emission level of each piece of construction equipment is adjusted for distance and usage time and then summed at a discrete receptor location of interest. Modern-day, commercially available computer spreadsheet programs, such as MS Excel or Lotus123, provide users with relatively easy means of creating custom-made construction noise prediction models. However, these spreadsheet models have significant limitations in that they tend to only predict noise at a limited number of discrete receptor locations, and only for a fairly limited configuration of equipment. Also, the accuracy of these spreadsheet models can only be as good as the veracity of the input data used in the cell equations. In addition, these models do not readily allow for computation of noise mitigation effects associated with various ground effects or options such as noise barriers, enclosures, or acoustical window treatments. Consequently, these spreadsheet models can only provide an estimate of construction noise at various locations during snapshots in time.
More recently there have been very sophisticated noise prediction model programs commercially available such as SoundPLAN (by SoundPLAN LLC of Shelton, WA), Cadna/A (by DataKustik of Munich, Germany), and the Environmental Noise Model (ENM by RTA Technology of Australia). These programs are able to display the predicted noise levels in formats that provide much more information, when compared to spreadsheet models, by graphically displaying results as equivalent noise contour lines. In doing so, noise levels at any receptor location of interest can quickly be estimated by interpolating the results between adjacent noise contour lines. Moreover, the construction equipment types and working locations can be changed fairly easily in these models, and new noise results can be computed much more quickly than could be done with discrete receptor point models. These sophisticated models also allow for some evaluation of noise reduction effects from various mitigation measures and/or man-made or natural barriers.
However, these sophisticated models suffer from the same accuracy limitations as the spreadsheet versions. While these models can support input of equipment noise data in the form of sound power levels or sound pressure levels at a reference distance, they too rely on user-defined noise emission levels for each piece or generic type of construction equipment. Consequently, the resulting predicted noise levels and noise contour lines are again only as good as the veracity of the input assumptions.
In creating a construction noise model, there are many issues that must be considered and included in the model if the prediction results are to be meaningful. While some of these variables are more important than others, the overall accuracy of the model must be sufficient for general acceptance. This is particularly relevant when federal and/or State funds may be used to create and mitigate the construction noise.
In general, the larger the number of variables that can be integrated into a noise model, the higher the confidence in its noise level results. While noise levels within close proximity to a construction activity can be predicted sufficiently well by accounting for just a few variables, the accuracy of predicted noise levels at farther distances, or in complex building or terrain conditions, will require a greater number of variables to be taken into account in the model. The trade-off, of course, is one of accuracy versus complexity and cost. The more variables that are taken into account in the model, the more complex the algorithms. Likewise, the cost of the model is proportional to the degree of research and computer programming required in its development.
Nevertheless, there are some fundamental variables that are typically considered in the development of any construction noise model. Those variables include:
A construction noise model should evaluate the severity or acceptability of the resulting predicted construction noise levels. Noise metrics such as Leq, Ldn, or L90 are geared toward evaluating longer-term steady noise levels, while metrics such as Lmax, L10, and SEL are intended to evaluate shorter-term fluctuating noise conditions. While it is understood that no single noise metric or criteria limit will satisfy everyone exposed to the noise, the establishment of criteria limits should attempt to accommodate the vast majority of the affected public and consider both noise hardship and noise annoyance conditions. Additional discussion of noise metrics is contained in Chapter 4.
For those projects where some degree of highway construction noise level prediction is determined to be necessary, the following procedures (discussed in order of their complexity) are available:
Since the publication of the 1977 Handbookref001, substantial noise monitoring has been conducted as part of the transportation project development process. Sufficient noise measurement data associated with construction activities and/or equipment may be available to draw reasonable estimates of expected noise levels for a given construction operation, thus avoiding the need to conduct noise measurements or model noise levels. Such data may be useful in providing estimates of the range of construction noise levels for inclusion in environmental documents, planning reports, etc., during the earlier stages of project development.
The 1977 Handbook contains manual methods for calculating construction noise levels. Similar methods are described in the FTA Transit Noise and Vibration Assessment documentref014 and can be applied, as appropriate, to any type of construction project. While these methods still remain valid, it is recommended that the source input data be reviewed to assure the most current data is reflected. The reader is directed to the variety of source emission data contained and referenced in Chapter 9 of this Handbook.
In lieu of using historical data, the analyst may deem it appropriate to use measured source data or to supplement historical data with measured data in one of the manual calculation processesref001.
This Windows-based noise prediction model has recently been developed in coordination with the preparation of this Handbook. It enables the prediction of construction noise levels for a variety of construction operations based on a compilation of empirical data and the application of acoustical propagation formulas. It enables the calculation of construction noise levels in more detail than the manual methods described above while avoiding the need to collect extensive amounts of project-specific input data. Appendix A of the hardcopy version of this Handbook contains the RCNM User's Guide. The RCNM User's Guideref083 and programref084 are directly accessible from the companion CD-ROM and web-based versions of this Handbook.
The RCNM is largely based on EPA methods from the 1970s in which equipment noise emissions, expressed as Lmax levels at a reference distance of 50 feet, are then adjusted for the actual distance to the receptor as well as for the time (or usage factor) that the equipment is predicted to produce noise on the job site. The current version of the model includes an updated equipment noise emissions database as well as an empirical relationship between energy-averaged (Leq) and percentile (L10) noise levels.
The primary equation used in the RCNM for predicting construction-induced L10 noise levels at receptor locations, when summed over all operating equipment, is as follows:
L10 in dBA = Lmax@50ft - 20 LOG (D/50) + 10 LOG (U.F.%/100) + 3 - ILbar
Lmax@50ft is the emission level for the equipment at 50 feet, expressed in dBA "slow";
D is the distance, in feet, between the equipment and the receptor;
U.F.% is a time-averaging equipment usage factor, expressed in percent; and
ILbar is the insertion loss of any intervening barriers (computed separately) in dB.
The +3 dB adjustment factor was empirically determined by examining the average difference between Leq and L10 noise levels during active construction.
Using RCNM, predicted noise levels can be evaluated at any distance from the construction activities. The selected location(s) analyzed will likely be dictated by the particular noise criteria in effect for the area and logistics in accessing measurement locations.
In addition to predicting the L10 noise level at a receptor location, the output of the model can also yield the predicted Lmax level by eliminating the usage factor term and the +3 dB adjustment term from the algorithm. Similarly, the model can yield the Leq level if just the +3 dB adjustment term is eliminated.
A comparison of the RCNM model and the manual calculation contained in the original 1977 Handbook can be found in Appendix B of this Handbook. Similar results are illustrated for Problem 5 (in the 1977 Handbook) using both analysis techniques.
Developed in 1982, this previously discussed modelref091 enables the calculation of construction noise levels using a variety of methods, including manual (using charts and tables), programmable calculators, and a DOS-based computer program.
While the HICNOM is data intensive and more comprehensive, the RCNMref083 and ref084 is the most up- to-date model in terms of construction noise database information and is most easily applied to projects of varying complexities. In addition, other noise models such as the FHWA TNM, the FTA Transit Noise and Vibration Assessment procedure, and community noise models may be adapted for use in evaluating construction noise levels.
Either the computer version (see Table 10.1 for link to the FHWA TNM data) or the Look-Up versionref090 of the FHWA TNM may be useful for the prediction of truck travel on haul roads, assessing impacts of diverted traffic, comparative modeling of line sources, barrier insertion loss calculations, etc. Either the manualref014 or one of the spreadsheet versionsref073, ref074, or ref075 of the FTA procedure may be employed if construction operations involve rail routes for goods transportation, or if noise from other rail-related activities exists in the study area and needs to be addressed in the evaluation. The FTA model also provides a means of evaluating highway noise, as well as noise from other types of transportation facilities such as stations, maintenance yards, etc.
General environmental noise models and accepted acoustical algorithms may be employed in determining site-specific noise contributions from equipment such as pumps, compressors, demolition equipment, etc.
Table 6.1 below provides a list of the models discussed above plus a quick link to the models and/or their reference material:
Table 6.1 Noise Models and Links.
|Federal Highway Administration (FHWA)|
|1977 FHWA Manual Method||Reference 001|
|Version 2.5 Users Guide Addendum||Reference 086|
|Look-Up Tables||Reference 090|
|FHWA HICNOM||Reference 091|
|Users' Guide||Reference 083|
|Federal Transit Administration (FTA)|
|1995 FTA Manual Method||Reference 014|
|Quattro Pro 6||Reference 074|
|Federal Railroad Administration (FRA)|
|2003 FRA Manual Method||Reference 092|
While the mechanisms associated with the various construction noise prediction techniques and models vary, the underlying prediction process methodologies are generally similar for all techniques, and include the following steps:
It is essential that those individuals performing measurements and analyses associated with the evaluation of construction noise possess the basic knowledge of the noise measurement techniques and prediction methodologies employed. Also essential is an understanding of the input data requirements and their sensitivities. A general knowledge of construction operations typical of the specific project being evaluated is also useful.
The range of items considered and the degree to which such items are addressed vary on a project-by-project basis. Things to consider include: