A Comprehensive Development Plan for a Multimodal Noise and Emissions Model (2010)

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.

C-1 APPENDIX C. CURRENT STATES OF THE ART This appendix reports the results of the examination of detailed information on the current domestic and foreign states of the art in modeling of transportation noise and emissions; building upon the initial list of models that had been identified in the proposal. A model evaluation protocol was prepared to assist in the gathering information concerning what the models do, how they do it, and for whom. Appendix I contains detailed descriptions on 47 models that are currently in use around the world for noise and air quality assessments. The models are grouped as follows: ● Air Quality Emissions and Dispersion Models ● Noise Models ● Models that do Both Noise and Air Quality The following sections of this chapter present the results of the gap analyses to identify important capabilities not currently met by existing models or analytical tools. The gap analysis evaluated the models against a desired condition, in this case, the end state of a multimodal tool, which is defined in Exhibits E-1 and E-2 of Appendix E. The gap analysis grouped the models by discipline for each of the elements, identified the gap between the end state and the collective abilities of the models within the particular discipline. Section C.1 covers air quality emissions and dispersion models. Section C.2 addresses noise models. C.1. Gap Analysis to End State – Air Quality Emissions and Dispersion Appendix I provides detailed descriptions of several air quality models in use with tables that summarize each model’s capabilities. The body of this section is devoted to the gap analysis for the aviation and ground noise models. C.1.1. Summary Table C-1 contains a capability gap assessment for emissions modeling and Table C-2 contains a capability gap assessment for dispersion modeling The “End Capability” categories listed in the tables are meant to provide overall summaries of key areas that could further be separated into finer components. For example, much of the emissions capabilities could include sub-categories on equipment characteristics and operational considerations. Also, the chemical transformation capabilities could include sub-categories on nitrate, sulfate, HAP, etc. chemistries. In addition, the different modes (e.g., aviation, highway, etc.) include many sub-categories of sources including different types of aircraft (e.g., jet, turboprop, and piston), Ground Service Equipment (GSE), etc. for aviation. Therefore, it should be understood that these higher-level categories encompass various sub-categories and are used to help facilitate the overall comparisons and gap analysis. Overall, the Transit and Maritime sectors have the biggest gaps because of the lack of models. Emissions and dispersion modeling capabilities will either need to be developed or adapted from existing models. Emissions data will need to be collected and incorporated into a model. Dispersion modeling for these modes can probably be handled through existing EPA models such as AERMOD or CALPUFF.

C-2 The most significant capabilities missing from all or most of the modes are: ● Second-by-second emissions ● Second-by-second concentrations ● Chemical Transformations The first two referring to second-by-second results are indicative of a lack of time-varying simulation capabilities. That is, the second-by-second activities of each source are generally not available at this time. Second-by-second results would provide the ultimate starting resolution that could be used to generate various aggregated results. These simulation results would also provide for the ability to better understand the interactions between the source and the emissions or concentrations. Currently, chemical transformation modeling capabilities are in their infancy, especially with regards to microscale and regional modeling. Although some capabilities exist in Gaussian models like AERMOD, CALPUFF, and CALINE4, they use first order approximations. As is currently being done with EDMS research under the FAA’s PARTNER program, methods from grid-based models (e.g., CMAQ) may need to be adapted for use with these smaller-scale Gaussian models. This gap for chemical transformation capabilities applies to both criteria pollutants and HAPs. Although there are some data for HAP emissions, most models generally cannot model these emissions at the moment. The capability to report emissions of speciated HAPS was recently incorporated into EDMS but only for aircraft and in a limited fashion. Comprehensive health impact studies will not be possible without quantifying these emissions (as well as dispersion effects). The other missing items are associated with dispersion modeling capabilities (i.e., building wake effects, urban canyon effects, complex terrain, and tunnels). These are all important but should be considered secondary to the aforementioned gaps. Dispersion modeling capabilities such as building wake effects will not always be necessary (i.e., depends on the nearby presence of buildings), and as a result, are not considered as important as the other core features. Other secondary features not specifically cited include plume rise, initial plume characteristics, etc. All of these secondary effects will need to be considered as the model development plan is established.

C-3 Green = Little or No Gap Yellow = Some Gap Red = Huge Gap TABLE C-1 Gap Assessment Matrix – Emissions Models End Capability Aviation Highway Off-Road Transit Marine Criteria emissions Yes, mainly from EDMS and the AEDT efforts Yes in models like MOBILE6.2 and EMFAC Yes, in models like NONROAD and OFFROAD Limited data (EFs) Limited data (EFs) HAP emissions Recently added to EDMS – can speciate all HAPs for aircraft only Limited data (EFs) Limited data (EFs) No No GHG emissions Only for some GHGs (CO2 For some pollutants (e.g., CO ) and aircraft only in EDMS 2 For some pollutants (e.g., CO) 2 Limited data (EFs) ) Limited data (EFs) Second-by-second No, only steady-state emissions CMEM can provide these for some vehicle types No/few data for other than motor vehicles available No No Modal emissions Yes for some sources (aircraft) Yes, these can be generated Some semblance of this from power requirements No No Fuel consumption Yes for some sources (aircraft) Yes, these can be generated (e.g., MOVES) Limited data Limited data Limited data Equipment specificity Very specific Very specific Very specific No No TABLE C-2 Gap Assessment Matrix – Dispersion Models End Capability Aviation Highway Off-Road Transit Maritime Transport (includes meteorology and terrain) Yes Yes Yes, since models like AERMOD have been applied No, but existing dispersion models can be adapted No, but existing dispersion models can be adapted Chemical transformation No, but research is undergoing No No No No Second-by-second concentrations No, only steady-state concentrations TRAQSIM can do this, but limited to research at this time No No No Building wake effects Not currently implemented in EDMS, but can be exercised externally through AEDT- Prime No No No No Urban canyon effects N/A No N/A No N/A Tunnels N/A No N/A No N/A Note: N/A = Not Applicable

C-4 C.2. Gap Analysis to End State – Noise Appendix I provides detailed descriptions of noise models in use with tables that summarize each model’s capabilities. The body of this section is devoted to the gap analysis for the aviation and ground noise models. C.2.1. Aviation Noise Model Gap Analysis In the following gap analysis for aviation noise assessment models, the Advanced Acoustic Model (AAM), the Aviation Environmental Design Tool (AEDT), the Heliport Noise Model (HNM), the Integrated Noise Model (INM), the Noise Integrated Routing System (NIRS), and NOISEMAP have been investigated and compared. Although a detailed discussion of their algorithms is not provided, the commercial noise models, CadnaA, IMMI, and LimA have also been considered. These models are, initially, investigated separately and, subsequently, included in the general discussion. C.2.1.1. Summary for Aviation Noise Models. Table C-3 summarizes the strengths and weaknesses of each of the three models whose algorithms have been explored. TABLE C-3 Strengths and Weaknesses of INM, NOISEMAP, and AAM Source Propagation* Model Database Characterization Ground Impedance Terrain Effects Meteo Effects Nonlinear Effects INM Extensive Fair Minimal Minimal Minimal No NOISEMAP Extensive Fair Fair Fair Minimal No AAM Limited hemisphere representations Detailed Good Fair Fair Yes *All models include spherical spreading and atmospheric absorption C.2.1.2. Discussion o f C ommercial Mo dels. It is common for the commercial models to be fully integrated methods of calculating both noise levels produced by many types of sources, and air quality. The commercial models can accommodate projects of virtually any size and are often limited only by the memory of the user’s computer. They are designed to be user-friendly, accepting different types of data input structures, providing different types of data output structures, and accommodating parallel processing for quicker and larger calculations. They provide an aesthetically pleasing user- interface, which often includes a 3-dimensional visualization of the considered landscape. Because users may come from different countries, a large number of standard algorithms, both national and interim EU calculation methods, are available for selection. The databases used for aviation noise sources may originate from different places. However, they may not be as extensive as the NPD database. While commercial models do offer many options regarding the noise computation standards used, the types of noise source, and the scale of calculation, their algorithms are not new and their advances seem to be in their efficiency, parallel computing capabilities, flexibility in size of the map, and refined user interface. C.2.1.3. E ase of U se. The most pervasive of the considered models is INM. Because it was designed to be used by a large number of people who are not as familiar with the complexities of sound production and propagation from aircraft, it is a relatively straightforward program that does not place large demands on the user. The commercial models go one step further than INM. Unlike INM, which is the standard model used by the FAA, the commercial models compete for business and must, therefore, be more conscious of how at ease a user is with their software. Therefore, the user-interfaces are often more aesthetically pleasing and, perhaps, a little less clunky.

C-5 C.2.1.4. Local versus Global. INM is meant to be used for noise assessments around a single airport. Its global counterpart, NIRS, is meant to be a large-scale model, involving noise assessments around multiple airports. AEDT is anticipated to include both the local and global capacities, as a union of these two (among other) models. NOISEMAP, however, is a local model and has no global counterparts. The author hypothesizes that this is a consequence of the density of commercial airports and the relative scarcity of military bases and supersonic aircraft operations. As previously mentioned, the size of the maps produced by commercial models is often limited only by the memory size of the user’s computer. Therefore, they can be applied to both local and global calculations. C.2.1.5. Conclusions. This gap analysis has investigated six U.S. noise assessment models, paying particular attention to three whose algorithms are representative of all considered models. In the comparison of INM, NOISEMAP, and AAM, it was found that many of the necessary aviation noise production and propagation effects have been addressed. However, terrain and meteorology were the effects most often neglected. The comparison also revealed different degrees of accuracy in the models’ source representations. These included a simplified representation, with a compact, extensive database and smaller computational requirements, a detailed source representation with a limited database, and a more theoretical source representation, requiring a user equip to supply the many necessary inputs. If an integration of the strengths of each model is sought, these different representations must somehow be reconciled. The merging of local and global aviation models is (or is anticipated to be) accomplished successfully in both the commercial models and AEDT. Finally, while INM was designed to be relatively easy to use, advances have been made by commercial models that provide a more refined user-interface, more calculation capabilities, and less computational limitations. The U.S. models could benefit from similar enhancements. C.2.2. Ground Noise Model Gap Analysis In the following gap analysis for ground noise assessment models, the Traffic Noise Model (TNM), the Roadway Construction Noise Model (RCNM), the Highway Construction Noise Computer Program (HICNOM), the Chicago Rail Efficiency and Transportation Efficiency (CREATE) model, the High-Speed Rail Initial Noise Evaluation (HSRNOISE) model, and the Horn Model have been investigated and compared. Although a detailed discussion of their algorithms is not provided, the commercial noise models, CadnaA, IMMI, and LimA have also been considered. These models are, initially, investigated separately and, subsequently, included in the general discussion. C.2.2.1. Source R epresentation. TNM, CREATE, HICNOM, and RCNM have extensive available source types. Therefore, the variety of sources used in highway, road construction, and railway noise may be satisfactory. TNM includes a relatively complex source representation with sub-sources and consideration of a 1/3-octave band sound spectrum. HICNOM can include different source geometries and uses frequency and source height to calculate barrier effect. However, frequency is, most often, not considered, and there are built-in limitations on the complexity of the source geometry. All other models provide a simplified model of the source representation. Therefore, source representation in highway noise may be satisfactory, could be improved upon for road construction noise, and should be better addressed in railway noise. C.2.2.2. Propagation Algorithms. Incorporation of divergence is universal among the different models and handled to similar end depending on the representation of the source as a point or line. No gap is identified.

C-6 C.2.2.3. Ground Effects. TNM includes a ground effect that is based on standard and accepted empirical ground impedance models. However, all other models provide a very simplified adjustment for ground impedance, often only providing for a soft ground. Therefore, the incorporation of ground effects in highway noise may be satisfactory and should be better addressed in road construction and railway noise. C.2.2.4. Uneven Terrain and Barriers. TNM includes many different features of terrain and the effects of diffraction, reflection and scattering those terrain obstacles can produce. However, it does simplify the terrain to decrease computation time for very complicated geometries. HICNOM also provides some consideration of terrain effects, though less thoroughly than TNM. All other models provide a simplified model of terrain effects. Therefore, the incorporation of terrain effects in highway noise may be satisfactory, could be improved upon for road construction noise, and should be better addressed in railway noise. C.2.2.5. Meteorological E ffects. As only an atmospheric absorption correction is applied in any considered model, the incorporation of meteorological effects should be better addressed in highway, road construction, and railway noise. C.2.2.6. Summary of S ource R epresentations and P ropagation A lgorithms. Table C-4 summarizes the strengths and weaknesses of each of the four models whose algorithms have been explored above. Table C-4 Strengths and Weaknesses of TNM, CREATE, HICNOM, RCNM, HSRNOISE, and the Horn Model Source Propagation* Model Source Database Source Characterization Ground Impedance Terrain Effects Meteo Effects TNM Extensive Detailed Good Good Minimal CREATE Extensive Minimal Fair Minimal Minimal HICNOM Extensive Fair Fair Fair Minimal RCNM Extensive Minimal Minimal Minimal Minimal HSRNOISE Limited Minimal Fair Minimal Minimal Horn Model Limited Minimal Fair Minimal Minimal *All models include divergence C.2.2.7. Outputs. TNM is the only program capable of producing detailed contours over extensive areas of the community. Therefore, one of the largest gaps in the ground noise models is the contouring capability. The other models should be interfaced with a contouring program in order to provide the necessary detail of noise levels throughout a community. C.2.2.8. Conclusions. This gap analysis has investigated six U.S. noise assessment models that calculate, collectively, highway noise, road construction noise, and rail noise, including train horn noise. These models are split between the relatively simple spreadsheet models (CREATE, RCNM, HSRNOISE, and the Horn Model) and more complex models (TNM and HICNOM). The spreadsheet models use basic source representations and simple propagation effect corrections to calculate levels at a limited number of receiver points. The more complex models use a detailed source representation and calculate (some) propagation effects based on empirical models. However, only TNM can calculate detailed contours over an extensive area of the community. In the comparison of the spreadsheet and complex models, it was found that the spreadsheet models left gaps in most of the important categories of noise prediction capabilities. However, TNM employed a sophisticated source representation and more accurate propagation effect calculations. Still, it left a large gap in the incorporation of meteorological effects and room for some improvement in uneven terrain effects.

C-7 The considered models were relatively easy to use. However, sometime the ease existed because the models were basic. Therefore, because they are more complex, the commercial models are less straightforward to use. Another large gap in the U.S. ground noise prediction models, not previously discussed, is the absence of an industrial noise model (excluding road construction noise). Industrial noise is incorporated in the commercial models. However, it is not addressed by the considered U.S. models. In conclusion, the largest gaps in the ground noise prediction models are caused by the models’ inability to calculate detailed contours, the neglect of meteorological effects, and the omission of industrial noise prediction.