Improving AEDT Noise Modeling of Mixed Ground Surfaces (2017)

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.

9-1 CHAPTER 9. APPENDICES APPENDIX A: Annotated Bibliography APPENDIX B: Logan Study Data File Formats APPENDIX C: Gatwick Study Data File Formats APPENDIX D: Heathrow ANOMS Data File Format APPENDIX E: Washington National Airport Studies Data File Formats APPENDIX F: Detailed Documentation of Model Improvements APPENDIX G: AEDT User Guide Supplement

A-1 APPENDIX A: ANNOTATED BIBLIOGRAPHY The papers and reports reviewed here cover the theories that will be addressed in this work and the data that will be used for validation in the following work on this project. Groups of papers that address the same theory will be annotated together. Ahearn, M., Boeker, E., Rosenbaum, J., Gerbi, P., and Roof, C., The Analysis of Modeling Aircraft Noise with the Nord2000 Noise Model, U.S. Department of Transportation Federal Aviation Administration, Washington D.C., DOT-VNTSC-FAA-12-07, October, 2012. This report compares the propagation methods in Nord2000 with INM as a function input parameters. It is a thorough delineation of the effect of varying each type of input to the two models and how capabilities in Nord2000 could be implemented in INM. No comparison with measurements is made. Attenborough, K., Bashir, I., and Taherzadeh, S., Outdoor Ground Impedance Models, J.Acoust. Soc. Am. 129 (5), May 2011. This paper compared various models to characterize ground impedance. The complexity of the models varied from needing only one input parameter, like Delany and Bazley’s effective flow resistivity, to models needing as many as seven input parameters. Types of modeling varied from only characterizing the flow resistivity to including the depth to another layer below the surface. One conclusion in the paper states that models requiring more than two parameters should be discounted for routine prediction of ground impedance. While it was shown that higher fidelity models fit measured data with less error the information needed to characterize the ground (porosity shape, depth to hard-backing layers, etc.) will not be available in general for modeling terrain in the common studies. Boulanger, P., Waters-Fuller, T., Attenborough, K., and Li, K.M., Models and measurements of sound propagation from a point source over mixed impedance ground, J. Acoust. Soc. Am. 102 (3) 1997. This paper reviews how the theories of de Jong and Nyberg compare with a Fresnel zone approximation and boundary element code when modeling sound propagation over surfaces of mixed impedance. The results are compared to measurements. The authors used a modified version of the Fresnel zone method where the prediction of the ground effect over a mixed impedance surface interpolated between the predicted excess attenuation compared to the free field level above uniform boundaries of each impedance. This method performed well on a qualitative basis when compared to level, but did not perform as well in predicting the frequency position of the first dip in measured spectrum due to destructive interference of the direct and the reflected/diffracted sound. The conclusion states that the best agreement between theory and measurement was obtained by using Nyberg’s theory and the boundary element method. Furthermore, the authors state that the Nyberg theory along with the measurements show that the ground effect due to the mixed ground impedance represented by stripes of different impedance material overlaid on a continuous surface of different impedance may be determined from that predicted by using the area-averaged impedance. Brooks, L., Morgans, R., and Hansen, C., Learning Acoustics Through the Boundary Element Method: an Inexpensive Graphical Interface and Associated Tutorials, Proceedings of Acoustics meeting, Busselton, Western Australia, November, 2005.

A-2 This paper describes the utility of a graphic user interface that exercises the Helm 3D code developed by T. W. Wu. Helm 3D runs the boundary element method to solve problems in acoustics. Downing, M., James, M., and Hobbs, C., The Effect of Bodies of Water on Aircraft Noise Propagation, Wyle Research Report No. WR 04-17, August, 2004. This report summarizes measurements made at National Airport (DCA) at locations on either side of the Potomac River which runs next to the airport. It is expected that the data from this report will be used in this study for verification of the algorithms found to model mixed ground impedance effects on aircraft noise propagation. The study found that the treatment of hard ground impedance by the NOISEMAP model was correct for aircraft with tail-mounted engine. The sound exposure level (SEL) metric was the basis of the study. Embleton, T. F. W., Piercy, J. E., and Daigle, G. A., Effective Flow Resistivity of Ground Surfaces Determined by Acoustical Measurements, J. Acoust. Soc. Am. 74 (4), June, 1983. This paper presents a validation of use of treating the ground impedance with a single parameter – flow resistivity. The same theory is reviewed in Attenborough’s paper based on Delaney’s work. The treatment in this paper is noted here because the authors wrote the propagation code used in making the INM database of noise power distance curves. In particular, it was the author’s program, mentioned in Chapter 3 that was used for representing propagation over soft ground while matching noise levels measured from aircraft to create the database. Fleming, G., Burstein, J., Rapoza, A., Senzig, D., and Gulding, J., Ground Effects in FAA’s integrated noise model, Noise Control Eng. J. 48 (1), Jan-Feb 2000. The authors execute the model by Embleton, Piercy, and Daigle that was used to build the original INM data base for hard and soft ground for distances and elevation angles using INM aircraft spectra to calculate the A-weighted ground effect. A regression analysis was performed to show how the results could be used in INM without the burden of the full calculation required by the model. Two ground impedances were used: hard 20,000 kPa s/m2 and soft 150 kPa s/m2. The application of the results for mixed impedance ground made use of the Fresnel ellipse for calculating the ratio of hard to soft ground for a representative frequency in the spectrum. Comparison with measurements was not made. Fleming, G., Plotkin, K., Roof, C., Ikelheimer, B., and Senzig, D., Assessment of Tools for Modeling Aircraft Noise in the National Parks, Federal Interagency Committee on Aviation Noise, March, 2005. This study showed how the relationships between simulation noise modeling and integrated noise modeling could be exploited to compare their results and incorporate improvements from simulation modeling into the integrated model. Granoien, I. and Randeberg, R., Corrective Measures for Aircraft Noise Models, New Algorithms for Lateral Attenuation, Joint Baltic-Nordic Acoustics Meeting, Mariehamn, Aland, June 2004. The authors use the algorithms in Nord2000 to define engine installation effects on directivity for fuselage and wing-mounted engines. The directivity for fuselage-mounted engines is notably different from what is in INM. The results are based upon measurements at Oslo airport in 2001. While the propagation is over nominally soft ground, the results are remarkable for the quantity of measurements used and the directivity pattern found.

A-3 Lam, Y., and Monazzam, M., On the Modeling of Sound Propagation Over Multi-Impedance Discontinuities Using a Semiempirical Diffraction Formulation, J. Acoust. Soc. Am. 120 (2), August, 2006. The authors show that the formulation of De Jong’s model for propagation over a single impedance strip violates the principal of acoustic reciprocity and have shown a correction to the model allowing it to be used for propagation over multiple strips of one impedance material over another. Nyberg, C., The Sound Field from a Point Source Above a Striped Impedance Boundary, Act. Acust. (China) 3, August, 1995. Nyberg simplifies the solution of a point source above a mixed impedance surface consisting of narrow strips of alternating impedance and equivalent width. His solution of the wave equation is done in Cartesian coordinates and a Fourier transform that reduces to a Fredholm equation of the second kind. He proves that the total acoustic wave field will be the same as for a homogeneous surface with the mean value of the two impedances given conditions on the width of the strips. Pirinchieva, R., Model Study of Sound Propagation over Ground of Finite Impedance, J. Acoust. Soc. Am. 90 (5), November 1991. This paper is a classic example that shows how well the one-parameter model of the ground matches measurements using the theory of a point source above a flat boundary condition. Plotkin, K., Page, J., Gurovich, Y., and Hobbs, C., Detailed Weather and Terrain Analysis for Aircraft Noise Modeling, Wyle Research Report WR 13-01, Wyle Laboratories, Arlington, VA, April, 2013. This study performed an updated analysis of measurements at Denver International Airport using versions 5, 6, and 7 of INM and the Advanced Acoustic Model in order to develop algorithms for incorporating detailed weather and terrain modeling into INM. The effects studied all utilized methods for incorporating relationships found using simulation modeling into integrated noise models. Tools for incorporating weather and terrain data were detailed and included as part of the report. Senzig, A., Fleming, G., and Clarke, J.P., Lateral Attenuation of Aircraft Sound Levels Over An Acoustically Hard Water Surface: Logan Airport study, Report No. NASA/CR-2000-210127, National Aeronautics and Space Administration, 2000. This report details the measurements made at Boston Logan Airport to examine the applicability of then-available mathematical models of lateral attenuation to aircraft noise when applied to overwater propagation. It was found that the agreement between prediction and measurement was a function of aircraft geometry. Predictions for aircraft with tail-mounted engines agreed with SAE 1751 when corrected for an acoustically hard surface; whereas, this was not true for aircraft with wing-mounted engines. It was posited by the authors that the difference was the directivity for aircraft with wong- mounted engines was different than the lateral attenuation method specified in SAE 1751. Stusnick, E., Plotkin, K., and Sutherland, L., Short-Range Acoustic Propagation Model, Wyle Research Report No. WR 85-19, July, 1985. The appendices in this report give a detailed discussion of the development and usage of the theory for predicting sound from a point source above an impedance plane. The development by Sommerfeld solving the problem of spherical radiation above a flat boundary condition is the basis for subsequent studies by Rudnick, Chessell, Chien and Soroka, Attenborough, and others. Their

A-4 formulations all involve a version of Fresnel integrals in the solutions with a numerical distance defined in slightly different ways. This report is an excellent summary of the methods used by many authors to solve the same problem and shows the relationships between the different formulations. The review contained here will stand in place of a review of the individual works. Wu, T., Boundary Elements in Acoustics, WIT Press, 2000. The author describes how the boundary element method (BEM) is used to solve acoustic problems using numerical techniques. Provided with the book is a set of codes known as Helm 3D that invoke the solutions described in the book.

B-1 APPENDIX B: LOGAN STUDY DATA FILE FORMATS B.1. Position Data Format One example of a position data file format for the Logan study is included in Table 10. The rows in blue provide descriptive labels for the example data in the file. Microphone location and aircraft position are referenced to a single point in Boston Harbor using a relative coordinate system. TABLE 10 Position Data Format from Logan Study for an Example Event Row in the File Date 1 62399 Event ID Aircraft Type Operation Type (D = Departure, A = Arrival) 2 2 MD80 D 3 * Microphone Locations Number of Microphones 4 6 X- location Y-location Z-location Surface Impedance (cgs rayls) 5 1046 -0.7 21.6 150 6 1046 -0.7 31.6 150 7 0 721 7 20000 8 0 721 17 20000 9 0 3060 7 20000 10 0 3060 17 20000 11 * Aircraft Positions Number of Position Points in Aircraft Event 12 10 Time: Hours Time: Minutes Time: Seconds X- location Y-location Z-location 13 12 53 0 -2394.4 27.8 289.4 14 12 53 1 -2110.3 41.7 331.9 15 12 53 2 -1797.9 64.7 368.5 16 12 53 3 -1504.3 72.1 413 17 12 53 4 -1210.6 72.9 444.9 18 12 53 5 -921.9 87.5 489.5 19 12 53 6 -630.4 108.3 533.6 20 12 53 7 -326.8 124.8 573.3 21 12 53 8 -52.3 143.3 638.9 22 12 53 9 229.3 177.4 731.6

B-2 B.2. Noise Data Format One example of a noise data file format for the Logan study is included below. The rows in blue provide descriptive labels for the example data in the file. The acoustic data were collected with a Larson Davis Laboratories (LDL) Model 2900, two-channel, one-third octave band analyzer (LDL2900) set up at each measurement location. Each channel of the LDL2900 was set up to continuously measure and store, at 1/4-second time intervals, the unweighted, slow, linearly averaged one-third octave band spectral time history of the sound pressure level. Information on the complete LDL2900 file format can be found in the LDL2900 User’s Manual. A specific event of interest is first identified by locating the date and time stamp associated with an aircraft event (e.g., “6/23/1999 12:52:58 PM”). This is immediately followed by detailed data on the LDL2900 settings, followed by the one-third octave band acoustic data in the format below. Channel 1 represents the 1.5 m (5 ft) microphone, and Channel 2 represents the 4.5 m (15 ft) microphone. TABLE 11 Noise Data Format from Logan Study for a Portion of an Example Event Column in Data File 1 2 3 4 5 … 17 18 19 … 43 Number of Frames: 127 run time = 100000 delta time = 0.25 Data type Time (s) Overload Sum … 25 Hz 31.5 Hz 40 Hz … 10000 Hz Channel 1 0.25 No 72 63.6 59.1 61.3 41.1 Channel 2 0.25 No 72.6 65.3 59.6 61.8 42 Channel 1 0.5 No 74.2 56.7 64.6 62.5 41.2 Channel 2 0.5 No 74.7 60 64.8 63.1 42.1 Channel 1 0.75 No 74.9 56.2 62.2 60.4 41.1 Channel 2 0.75 No 75.7 53.1 63.1 60.4 42.1 Channel 1 1 No 76 61.6 56.1 61.4 41.3 B.3. Meteorological Data Format One example of a meteorological data file format for the Logan study is included in Table 12. The rows in blue provide descriptive labels for the example data in the file.

B-3 TABLE 12 Meteorological Data Format from Logan Study for a Portion of an Example Event Column in Data File 1 2 3 4 5 … 12 13 … 18 19 20 21 22 Event ID Vector Mean Wind Speed (m/s) Mean Wind Direction (deg) Standard Deviation of Wind Direction Mean Air Temperature (C ) … Relative Humidity (%) Barametric Pressure (millibars) … Magentic Declanation (deg) Hour Minute Second Wind Gusts (m/s) 150 3.1 100 3.4 22.9 57 1020.1 0 11 26 35 3.9 150 3.1 100 3.4 22.9 57 1020.1 0 11 26 36 3.9 150 3.1 100 3.4 22.9 57 1020.1 0 11 26 37 3.9 150 3.1 100 3.4 22.9 57 1020.1 0 11 26 38 3.9 150 3.1 100 3.3 22.9 57 1020.1 0 11 26 39 3.9 150 3.1 100 3.3 22.9 57 1020.1 0 11 26 40 3.9 150 3.1 100 3.3 22.9 57 1020.1 0 11 26 41 3.9 150 3.1 100 3.3 22.9 57 1020.1 0 11 26 42 3.9

C-1 APPENDIX C: GATWICK STUDY DATA FILE FORMATS C.1. Noise and Position Data Format One example of a noise and position data file format for the Gatwick study is included in Table 13. Data from individual monitors for select aircraft (B767 and DC9) was measured in 1996 to 1998. The individual monitors were located at centerline, 540 m lateral, 750 m lateral, and 1450 m lateral to the flight track. The data consist of operation number (Opnum), date, aircraft type (Type), LAmax, SEL, elevation angle (beta), aircraft altitude in feet (height), and slant distance in feet (Slant_dist) for each individual operation. An example of the centerline data is shown below. TABLE 13 Noise and Position Data Format from Gatwick Study for Example Events Centre Line Opnum Date Type Lamax SEL Beta Height(ft) Slant_dist (ft) 1837209 12/17/1998 MD80 94.4 102.2 75.80 2513 781 1836660 12/16/1998 MD80 95.6 105 71.80 2261 716 1837493 12/17/1998 MD80 89.6 97.5 79.20 1911 584 C.2. Processed Engine Installation Data Format The data from the Gatwick study were used to evaluate the attenuation of an aircraft due to engine installation location. One example of the processed engine installation data file format for the Gatwick study is included in Table 14. Data of a more processed form were organized by engine location: underwing or rearmount (fuselage-mounted engines). There were 13994 operations for the underwing engines and 1,047 operations for the rear mount engines. These data consisted of only aircraft type (Ancon_type), elevation angle (beta), and decibel difference in attenuation from a normalized level (atten). An example is shown below. TABLE 14 Processed Engine Installation Data Format from Gatwick Study for Example Events Ancon_type Beta Atten MD80 12.6 4.5 MD80 30.4 1.2 MD80 21.1 1.7 MD80 41.6 0.7 MD80 21.6 2.1

D-1 APPENDIX D: HEATHROW ANOMS DATA FILE FORMAT The data from the ANOMS query at Heathrow (LHR) contains the information shown in Table 15. The row in blue describes the data columns contents. The pertinent information to define lateral attenuation are in the columns noted PCA (Point of Closest Approach). Using the noise metrics SEL and Max Level (Lmax), the lateral attenuation can be isolated by normalizing the metrics to a fixed distance. The monitor locations are included in the electronic archive. Meteorological information for this time period at LHR is available at weatherunderground.com.

D-2 TABLE 15 Single Operation from Database Query of ANOMS at LHR P C A O p e r a t i o n N u m b e r D a t e T i m e O p e r a t i o n T y p e ( A / D / O / T ) R u n w a y P a t h N a m e T a i l N u m b e r A i r c r a f t T y p e C a l l s i g n L o c a t i o n I D S t a r t D a t e T i m e M a x D a t e T i m e M a x L e v e l S E L D u r a t i o n H o r i z D i s t a t L M a x ( m ) H e i g h t a t L M a x ( f t ) S l a n t D i s t a t L M a x ( m ) P C A A l t i t u d e ( f t ) P C A S l a n t D i s t a n c e ( f t ) P C A G r o u n d R a n g e ( f t ) P C A E l e v a t i o n A n g l e 200835 7925 04- 07- 2015 15:07 :34 D 27R 27RC PT EIDE M 320 EIN1 6N 18 04- 07- 2015 15:08 :09 04- 07- 2015 15:08 :21 75.4 85 24 66 2242 686 2221 2238 270 83 200835 7925 04- 07- 2015 15:07 :34 D 27R 27RC PT EIDE M 320 EIN1 6N 19 04- 07- 2015 15:08 :10 04- 07- 2015 15:08 :25 77.1 84.3 25 661 2316 967 2241 3005 2003 48 200835 7925 04- 07- 2015 15:07 :34 D 27R 27RC PT EIDE M 320 EIN1 6N 17 04- 07- 2015 15:08 :18 04- 07- 2015 15:08 :34 72.5 83.2 28 884 2472 1162 2359 3323 2341 45

E-1 APPENDIX E: WASHINGTON NATIONAL AIRPORT STUDIES DATA FILE FORMATS E.1. Part 150 Study Measurements The measurements completed during the Part 150 study at DCA utilized Larson Davis Model 820 (LD820) sound level meters (SLMs). The meters were set up to collect noise metrics during events when the measured A-weighted sound level exceeded 70 dB. An example of the record output of an SLM is shown in Table 16. The event records are translated to ASCII format as shown in Table 16 using the freely available software package ‘SLMutil.exe’ from Larson Davis. The data archive includes the binary download files from the SLMs; thus, more or less descriptors may be translated by the user, as needed. TABLE 16 Example of LD820 SLM Event Records Measurement Time of Site Location Number Date Time Lmax Duration Leq SEL Lmax Peak Uwpk Sym ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- 3 A-1 Towing Lot 0 07- Oct- 10 16:34:3 4 16:34:3 4 3.2 71.9 76.9 74.9 107.3 106 0.8 3 A-1 Towing Lot 0 07- Oct- 10 16:34:5 1 16:34:5 5 15.5 71.8 83.7 76 91.3 97.4 26.6 3 A-1 Towing Lot 0 07- Oct- 10 16:35:3 3 16:35:4 5 20.6 71.4 84.5 74.2 87.6 98.3 62.1 3 A-1 Towing Lot 0 07- Oct- 10 16:36:2 9 16:36:3 1 7.3 79.1 87.7 84.1 98 101 37.5 3 A-1 Towing Lot 0 07- Oct- 10 16:37:2 4 16:37:3 1 10.7 70.9 81.2 74.6 90.1 97.4 67.2 3 A-1 Towing Lot 0 07- Oct- 10 16:38:5 9 16:39:0 2 8.8 75.8 85.3 80.5 91.1 104.7 42.2 3 A-1 Towing Lot 0 07- Oct- 10 16:39:3 4 16:39:4 5 20.2 74.3 87.3 77.9 91.8 100.1 55.9 3 A-1 Towing Lot 0 07- Oct- 10 16:41:1 2 16:41:1 7 15.4 71.7 83.5 74.4 89.1 97.4 37.5 3 A-1 Towing Lot 0 07- Oct- 10 16:42:5 5 16:42:5 5 12.8 70.6 81.7 72.9 88 96.9 2.7 3 A-1 Towing Lot 0 07- Oct- 10 16:54:1 1 16:54:1 4 16.8 72.2 84.4 75.6 90.2 99 20.7 3 A-1 Towing Lot 0 07- Oct- 10 16:57:0 1 16:57:0 2 9 75.6 85.1 81.2 95.7 99.4 18.8

E-2 Radar tracking data is part of the data archive. The radar tracks have been reduced to the RAT (RAdar Tracking) file format. Wyle Laboratories created RAT files for use with civilian and military airports. The format was developed as part of an effort to distill backbone tracks from radar data for the military using the Noise Data Acquisition and Display System program (NDADS) (Page, 1996). The binary tracking data required by NDADS is in the binary RAT file format. Original radar data are reduced to RAT files via a pre-processor specific to each radar system. The RAT file is a sequence of data blocks, where each block is a sequence of three elements: keyword, count, and data, where keyword is a 4-character string defining the data type count is the number of 4-byte data elements in the file data is the data itself: must be 4*count bytes long All data blocks can be read with a pair of FORTRAN read statements of the form: read(lun)keywrd,icount read(lun)(idata(i),i=1,icount) where lun = FORTRAN logical unit of data file keywrd = character*4 icount = integer*4 data count; must be 1 or greater idata = integer*4 array large enough to hold icount elements The file must be created (and opened) as an unformatted sequential file with a record length of 4 bytes. For systems which allow pure binary access to files (such as C, BASIC, or extensions in some FORTRANs), the file may be pure binary. Pure binary is the preferred data format. When using FORTRAN the file is opened via the statement: open(unit=lun, file=filnam, form=’keyword’, recl=4) where filnam is the name of the file. For standard FORTRAN 77, the form keyword that should be used is “unformatted” along with a zero record length (recl = 0). Actual data elements do not have to be integers; the requirement is that any data block be readable by the above. It is typical that the second read statement will be replaced by an input/ouptut list which matches the actual data types. The following keywords and data elements (and their FORTRAN edit descriptors) are listed in Table 17. An example plot of the radar tracks is shown in Figure 70.

E-3 TABLE 17 Keywords and Data Description of RAT File Format Keyword Data Count Data element Type Description OGEO 6 longdeg I4 Coordinate origin longitude degrees longmin I4 ..... minutes seclong R4 ..... seconds latdeg I4 Coordinate origin latitude degrees latmin I4 ..... minutes seclat R4 ..... seconds OUTM 3 easting I4 UTM easting of coord. origin, meters northing I4 UTM northing of coord. origin, meters izone I4 UTM zone number ANGL 1 rotangle R4 Angle of Y axis: degrees re geo north ZULU 1 deltime R4 Time, seconds, add to T to obtain local time OPER 15 beacon A4 4-digit beacon code (squawk) actype A4 Aircraft type optype A4 Arrival, departure, etc. date A8 Date of operation, yyyymmdd runway A4 Runway name callsign A8 Call sign numac I4 Number of aircraft in flight userinfo A16 User-defined text information iicode I4 Reserved code ncount I4 Count of size of remaining elements T as req time R4 Times along flight path, seconds X as req X R4 X-position, feet Y as req Y R4 Y-position, feet Z as req Z R4 Z-position, feet MSL V as req V R4 Aircraft speed, knots TEXT as req text A Arbitrary text string ENDF 1 anything --- End of file The earlier version of this file format (July, 1995) used keyword CODE instead of OPER. The data count for CODE was 3, and data elements were the beacon, actype, and ncount. The new OPER data block includes these elements, plus others which would have appeared in the now-obsolete FLP file. Use of a single file increases efficiency and simplifies data archiving. Within the OPER data block, element USERINFO is available for the user to add any special information about a flight. Element IICODE is reserved for future use and should not be appropriated by users. A typical file will have repeated sequences of OPER, T, X, Y, Z, and V blocks, each sequence representing the full identification and track data for a given sortie. Each OPER block must be followed by T, X, Y, Z and V blocks, in that order. Each T, X, Y, Z, V block is a sequence of values corresponding to the number of points in that track; the count must obviously match across all five of these for a given OPER block. The quantity NCOUNT represents the total number of 4-byte data units in the T, X, Y, Z and V blocks, including their keywords and data counts. The primary purpose of NCOUNT is to allow rapid skipping of unwanted data by adding it to the current file record pointer. Note

that if the NCOUNT T block sho time must night peri times on t D the RAT Data bloc origin. T latitude ar normally NOISEMA rotation re then ROT T used to co Note that T E.2. Nav T time histo reduced to Air Tour the projec still in use example f data; thus re are N po will be 5*(N ime (T block uld be equal be put in th ods are keye he proper dat ata may orig file to proper k OGEO or O he standard s e positive. D have the Y a P analysis i lative to geo ANGLE is th he TEXT key nvey inform the length of he ENDF key y Overwate he SLMs dep ries; thus, th a format du Management t was to defin around airp ile header wi , it can be op ints in a giv +2). ) is in secon to 0; otherw e ZULU bloc d to this. If e. inate from di ly merge dat UTM docum ign conventi ata are also xis aligned s usually alig graphic north e magnetic d word allows ation require this string mu word marks FIGURE 70 r Study at loyed for the e information bbed zweek Plan (ATMP e the natural orts (Hobbs, th the explan ened with a s en track, the ds past midn ise, the numb k. Local tim a different ti fferent radar a from differ ents the coo ons are follow to be written with magneti ned with ge at the coord eclination. the file crea d for special st be in 4-by the end of the . Example o DCA Data F overwater s the meter co for ZION we ) work cond soundscape 2014) and m ation of the f imple editor data count ight. If the er of hours t e calculatio me system is sites. It is the ent sites, con rdinates (ge ed for geog in alignmen c north. Ra ographic coo inate origin. tor to insert a user-defined te units. data file. f DCA depa ile Format tudy were se llects is con ekly data fil ucted by the (Hobbs and D ilitary install ormat embed to read the h for each of local time is hat the time n ability is re used, care m responsibili verting coord ographic or U raphic coord t with the loc dar data are rdinates. Da If the Y axis ny arbitrary analysis, or rture tracks s tup to captur tinuous whil es. This file National Pa owning, 200 ations (Deme ded in it. W eader and w T, X, Y, Z, used for thi in T block i quired, since ust be taken ty of the pre- inates to a c TM, respect inates: east lo al NDADS c typically in ta block AN is aligned wi text into the to insert aud off Runway e sound pres e it is turned format was rk Service (N 3). This com rs, 2015). F hat follows th ith a compute V will be N s, then the Z s offset from the day-eve to place adj processor cre ommon refer ively) of the ngitude and oordinates, w this system, GL documen th magnetic n file. This m it document 01. sure level sp on. The dat used as part PS). The g pact file form igure 71 sho e header is b r program fo E-4 and ULU local ning- usted ating ence. local north hich while ts the orth, ay be ation. ectral a was of the oal of at is ws an inary r fast

E-5 reading. As noted above, SLMs were deployed on both sides of the river adjacent to the threshold of Runway 01 at DCA. Figure 72 shows the A-weighted time histories at the SLMs during a portion of time. The vertical line in the graphs indicates the departure of a DC-9 off Runway 19. The SLM records for S1, S2, and S3 show the levels on the opposite side of the river while the DC SLM measures the sound on the airport side. Weather stations were deployed near the SLMs. The airfield data included hourly values of temperature, wind, relative humidity, and atmospheric pressure. During observations on the other side of the river, a weather station collected data at a 1-second rate on the rolling 5-minute average of the temperature, wind speed and direction, and relative humidity. Like the measurements made during the Part 150 study, tracking data for aircraft were reduced to the RAT file format noted above.

FIGURE 71. Example zweek file header. E-6

FIGURE 72. Example A-weighted time histories at DCA measurement Sites. E-7

F-1 APPENDIX F: DETAILED DOCUMENTATION OF MODEL IMPROVEMENTS The following appendix describes the recommendations for implementing the findings of this project in the Federal Aviation Administration’s (FAA) Aviation Environmental Design Tool (AEDT). It is organized into the following sections: background; AEDT input requirements and data; terrain and ground type processing method; AEDT segmentation updates; mixed ground impedance modeling method; AEDT implementation; and additional recommendations. F.1. Background The objective of this research is to develop a method to account for impedance variability of ground surfaces in a manner suitable for model implementation to improve the noise prediction accuracy of AEDT; therefore, the recommendations in this document are structured as supplemental methods to be used in conjunction with the noise computations currently in AEDT. AEDT is an integrated noise model. Integrated modeling relies on pre-propagated noise-power- distance results from infinite-length, constant-condition flight segments. The noise fraction determines, based on segment-receiver geometry, the percentage of the infinite segment noise exposure that results from the finite length segment. The lateral attenuation adjustment accounts for the effects of lateral aircraft directivity and for acoustic propagation over soft ground. Within AEDT these contributions are summed for each modeled flight segment to obtain the total noise exposure for an entire flight path for a specific aircraft. The recommended method for computing the effects on noise propagation over mixed ground types would be represented as an additional adjustment to the total noise exposure in AEDT. F.2. AEDT Input Requirements and Data In order to compute mixed ground effects in AEDT, ground type data needs to be imported into AEDT. A good source of ground type data is the National Land Cover Database (NLCD) which includes classifications of land cover in the continental United States. Their website allows the downloading of this land cover data and allows for downloading National Elevation Data (NED) for the same requested area free of charge. These data resemble coordinates with the associated land cover designation. These designations have been evaluated and mapped to the flow resistivity values in Table 5 (Section 5.1). This could then be mapped to a simplified ground type designation list, such as the one provided in Table 1 of the main report. In studies with very limited ground type data, a default setting for water or hard ground of 100,000 kPa s/m2 could be used, along with a default soft ground value in the range of 150 to 300 kPa s/m2. F.3. Terrain and Ground Type Processing Method The two-dimensional profile along the ground on the line below source and receiver should be used to calculate the ground impedance between source and receiver. The minor difference in finding the average impedance using the profile versus the whole ellipse along with the utility of identifying major terrain features along the profile for other types of modeling make it the best choice. The ‘active portion’ of this line in determining the average flow resistivity to be used in the calculation of ground impedance is the extent of the major axis of the Fresnel zone ellipse for each frequency considered. Figure 73 shows a diagram of the source – receiver geometry in relation to the profile cut along the ground. Only the portion of the profile cut inside the ellipse is considered in calculating the average flow resistivity. In the case of low elevation angles and frequencies, the active portion of the ground path can extend beyond the source and receiver.

T be used b receiver. T zone as w average fl boundarie U representi terrain fil ground ty imagery. data to the any grou specificat these prop approach instead of cut by int FIGU Fr he algorithms y AEDT to he formulae as done in th ow resistivit s between wa nlike elevati ng terrain da e is classifie pe in this are While it may next, there i nd classifica ions are for t erties betwe should be us interpolating errogating co RE 73. Diag esnel zone e used by NO identify the in Boulange is project. A y is estimate ter and land on data the ta should not d as a groun a is an estim be reasonab s no basis for tion and the he individua en grid point ed when det between gri ntours of flow ram of sour llipse as the average flow ISEMAP to i impedance p r et al.’s pape s mentioned d by using th with no inter assigned par be interpola d type with ate of the ave le to expect the same beh parameters l grid points, s can lead to ermining the d points; oth resistivities ce – receiver active portio resistivity a nterrogate ter rofile along r (1997) wer in the body e profile cut polation betw ameters to c ted. In the ex an area of 1 rage type of a smooth tra avior with th for estimati and they app unrealistic r value for a erwise, progr . Because th geometry d n of the gro long profile rain data are the ground f e adapted to of the repor routine, the een flow res alculate grou ample of the 0,000 square ground foun nsition from e flow resist ng ground i ly to the ed esults; theref point not ex amming a ro e hypsograph etailing maj und path use cut. very straight rom beneath calculate the t, it is recom y need to be istivities. nd impedan NLCD, eac feet. The cl d within in it one grid poi ivity data. Al mpedance f ge of the are ore, a “close actly mappin utine that ca y layer in ge or axis of d for forward and the source t size of the Fr mended that altered to re ce for grid p h grid point assification o based on sa nt in the elev l that is know or it are tha a. Interpolati st point” ma g to a grid lculates the p ographic dat F-2 could o the esnel if the spect oints in the f the tellite ation n for t the on of pping point, rofile a sets

F-3 is vector based, the boundary definitions of bodies of water, for example, will be more accurate than the 100 ft x 100 ft resolution found in the NLCD. If there are only two flow resistivity values in a data set, then the area each occupies in the ellipse or the length of each along the active portion of the ground path, should be recorded when interrogating the terrain data. One example when this would be appropriate would be if the ground cover consisted of only water and land of uniform ground cover. F.4. AEDT Segmentation Updates Capturing changes in geometric acoustic propagation features, such as varying ground impedance, requires a reduction in the segment length commensurate with the geometric feature to be modeled. For uniform ground impedance, infinite segments are sufficient; however, AEDT models noise propagation based on propagation from a single point on each flight path segment. In order to model varying ground impedance, one must consider the modeled fidelity of the ground impedance. For example, a longer flight path segment may miss the varying ground impedance associated with a small lake or a cove on the shoreline. There is a direct geometric link between the desired output fidelity and the maximum segment length that must be used in the integrated model. In the limit, as flight segment lengths are reduced, errors are introduced due to such pragmatic issues as numerical precision, discretization, algorithm implementation and fundamental assumptions in the noise fraction formulation. There are four things that should be considered when selecting the AEDT segment length: 1. the possible influence (in dB) that the changes in ground impedance might have on the acoustic result from a single segment; 2. the uncertainty (in dB) introduced by the noise fraction when one reduces segment length; 3. the geospatial precision over which impedance changes are defined; and 4. the geometric fidelity for which the acoustic results are desired. In the study by Plotkin et al. (2013), little error was found when applying the lateral attenuation for three points on a line segment (both ends and the closest point of approach) for segments up to 2,000 feet in length. It is reasonable to assume that AEDT may have to be changed in a way that flight path segments are reduced from what is input to the model. This recommendation could be applied as initial user guidance, to be followed by additional analyses and possible future automatic flight path segmentation limits in AEDT. F.5. Mixed Ground Impedance Modeling Method The method investigated for calculating lateral attenuation for this effort utilized straight ray theory. Treatment of surfaces with mixed impedances was done by averaging the flow resistivities in the Fresnel zone ellipse or along the active portion of the ground path between source and receiver inside the ellipse (its major axis). The areas around airports are not going to have finely delineated boundaries between surfaces of different flow resistivities normal to the sound propagation path; therefore, this approach is well suited for the likelihood that the different ground surfaces around the airport being modeled will be spread out like the fractal patterns used to model random terrain in video games. In the singular case where only two flow resistivity values are identified in the terrain, it is recommended that the lateral attenuation be calculated for each of the resistivities and the resulting pressures be apportioned according to the ratio of the two.

F-4 The straight ray method can be used to calculate the excess ground attenuation in decibels for a point source of one-third octave band noise using the EGA model (Stusnick, et al., 1985). Inputs to the model include the actual geometry between source and receiver. This will have to be worked in with some assumption about the receiver height when doing grid calculations in AEDT. Also, this theory assumes the area around the receiver is flat. This should not be a problem as the most likely place to apply this methodology is around airports which are generally placed in flat places. F.6. AEDT Implementation It is recommended that the mixed impedance adjustment be applied as a supplement to the existing impedance method currently implemented in AEDT, which is a level-based adjustment defined in SAE-AIR-5662 “Method for Predicting Lateral Attenuation of Airplane Noise”. SAE-AIR-5662 assumes propagation over soft ground only. Working off of the understanding that the current AEDT lateral attenuation method (SAE-AIR- 5662) is sufficient for modeling soft ground effects, an adjustment can be implemented to compute ground effects from other, non-soft ground types (including mixed ground types) using a computational method similar in structure to that used to compute the atmospheric absorption adjustment in AEDT (based on SAE-ARP-5534 “Application of Pure-Tone Atmospheric Absorption Losses to One-Third Octave Band Data”). This computation works off of the rationale that although AEDT only uses level- based NPDs to model propagation (instead of one-third octave band spectra), and soft ground attenuation is already included in the data sets and the lateral attenuation adjustment, the difference between the new ground effects method computed for propagation over mixed ground and soft ground can be applied as an adjustment to the AEDT NPDs. The following is a step-by-step description of the logic needed to implement this adjustment on a segment-by-segment basis in AEDT. 1. The spectral class data for a particular aircraft and operation type are used as the aircraft noise source in this adjustment. The spectral data in AEDT have been corrected to reference day conditions, using the SAE-AIR-1845 standard atmosphere, at a distance of 1000 feet. This value does not need to be normalized to the NPDs, as long as the same spectral class is used consistently throughout this computation. The spectral class data should be un-weighted. 2. The aircraft spectral class is corrected back to the source (from the 1000 foot reference), removing the SAE-AIR-1845 atmosphere absorption effects on a one-third octave band basis. This is the source spectrum (SSOURCE). 3. The new mixed ground impedance method is used to calculate the propagation effects over hard or mixed ground on a one-third octave band basis. This is computed over a distance from the source to the receiver using the two-dimensional average flow resistivity of the major Fresnel ellipse axis along the propagation path. This is the spectrum propagated to the receptor over the mixed ground type (SREC, MIXED). 4. The new mixed ground impedance method is used to calculate the propagation effects over default soft ground (150 kPa s/m2) on a one-third octave band basis over the same propagation distance. This is the spectrum propagated to the receptor over the soft ground type (SREC, SOFT). 5. Each propagated spectrum is A-weighted (or C-weighted, depending on the noise metric being calculated). 6. The 24 one-third octave band values of each propagated spectrum are logarithmically summed to yield the overall levels (LREC, MIXED and LREC, SOFT). 7. The difference between the two propagated overall levels (LREC, MIXED and LREC, SOFT) is computed (ΔLLA, MIXED-SOFT). This value represents the additional ground effects for the

F-5 study, once the soft ground effects have been applied. ΔLLA, MIXED-SOFT is added to the current AEDT lateral attenuation adjustment (LLA) for soft ground based on SAE-AIR- 5662. In addition, it is recommended that a switch between using the results from the mixed impedance surface and the current lateral attenuation adjustments (SAE-AIR-5662) be implemented, so the new functionality could be turned off, to allow for comparison testing. Furthermore, this procedure is to be used as a supplement to the existing lateral attenuation calculation in AEDT; therefore, if the existing cutoff for its application in terms of elevation angle applies. F.7. Additional Recommendations As part of the incorporation of a new lateral attenuation adjustment based upon mixed-surface impedance, AEDT should be allowed to switch between using the results from the mixed impedance surface and turning it off and using the current lateral attenuation adjustments (SAE-AIR-5662) to allow for comparison testing. Future modifications to AEDT could include the implementation of the mixed impedance adjustment as a complete replacement to the current lateral attenuation adjustment in AEDT. This would require that the SAE-AIR-5662 adjustment be removed and engine installation directivity adjustment be re-implemented independent of lateral attenuation. In this case the spectral-based ground impedance adjustment can be computed independently for any ground type (soft, hard, or mixed), summed across the individual one-third octave-bands into an overall sound pressure level, and then applied to the total noise exposure along with the other adjustments for each segment.

APPEND T ground su of AEDT used in co T Environm “Improvin into four detailed t recommen Survey M Incorpora T and the A as a refere the AEDT G.1. AED In ambient f Guide. F recommen IX G: AED he objective rfaces in a m ; therefore, th njunction wi he following ental Design g AEDT No sections: AE echnical des dations are ethods Use te Improvem his document EDT softwar nce througho Version 2c T Input R AEDT 2c, t iles is describ igure 74 show ded supplem F T USER G of this resea anner suitabl e recommend th the noise c presents a s Tool (AED ise Modeling DT input re cription of t presented in d in Curren ents into AED presents rec e graphical u ut this docum Service Pack equirement errain and am ed in Sectio s a recomm ent to Sectio IGURE 74. R UIDE SUP rch is to dev e for model ations in thi omputations upplement to T) User’s G of Hard, Sof quirements a he recomme “Review Cu t Noise Mo T” (publicat ommendation ser interfaces ent (AEDT, 1 User’s Gui s and Data bient files ar n 9.7 “Terrai ended updat n 9.7 to accou ecommend account PLEMENT elop a meth implementati s document a currently in A the Federal uide for imp t, and Mixed nd data; AE nded method rrent Propaga dels, Design ion pending) s for potenti (GUI). The 2016). Whe de are identif e defined in t n and Time e to the AED nt for groun ed updates t for ground od to accoun on to improv re structured EDT. Aviation Ad lementing th Ground Surf DT interface along with tion Theory Sensitivity . al modificati AEDT Vers re appropriat ied. he “Definitio Above Ambi T Definition d cover files o the AEDT cover files. t for impeda e the noise p as suppleme ministration e findings f aces” in AED ; and model the softwar over Mixed Studies, an ons to the AE ion 2c Servic e, the corresp ns” tab. The ent Files” in s Tab. The (see Figure 7 definitions t nce variabil rediction acc ntal methods ’s (FAA) Av rom ACRP T. It is orga ing guidance e implemen Ground Sur d Detail Ho DT User’s G e Pack 1 was onding sectio use of terrai the AEDT U following tex 5). ab to G-1 ity of uracy to be iation 02-52 nized . The tation faces, w to uide used ns in n and ser’s t is a

T and inclu resistivity values ra represente these non FIGURE 75. he NLCD is a des 16 land c value recom nging from d in this tab -standard valu Recommen Landsat-bas over classifi mendations f 10 to 100,0 le may be m es. ded updates for ed land cove cations. The or use in AE 00 kPa s/m2 odeled in A to AEDT 2c ground cove r database fo se classifica DT are inclu are allowe EDT, provid sp1 User’s r files. r the United S tions and the ded in Table d for in AE ed that there Guide Sectio tates. It has correspondi 18. Estimat DT, so add is appropria n 9.7 to acco a 30 m resol ng estimated ed flow resis itional value te justificatio G-2 unt ution, flow tivity s not n for

G-3 TABLE 18. NLCD Land Cover Classifications and the Equivalent Estimated Flow Resistivity Value Recommendations for use in AEDT NLCD2011 Land Cover Classifications Estimated Flow Resistivity (kPa s/m2) 1. Open water 100000 2. Perennial ice/snow 20000 3. Developed, open space 225 4. Developed, low intensity 10500 5. Developed, medium intensity 19500 6. Developed, high intensity 25500 7. Barren land 3000 8. Deciduous forest 50 9. Evergreen forest 50 10. Mixed forest 50 11. Shrub/scrub 50 12. Grassland/herbaceous 225 13. Hay/pasture 225 14. Cultivated crops 200 15. Woody wetlands 100000 16. Herbaceous wetlands 90000 G.2. AEDT Interface In AEDT 2c, the use of terrain for noise calculations is defined in the “Metric Results” tab under “Define Metric Results” as “Set Processing Options”. The use of this feature is described in Section 5.2.4 “Step 4: Set Processing Options” in the AEDT User’s Guide. The incorporation of Mixed Ground Effects Option in AEDT will result in modifications of sub-section 5.2.4.3 “Noise Modeling Options” and the addition of a new sub-section 5.2.4.5 “Ground Cover Options”, which would immediately follow sub- section 5.2.4.4 “Terrain”. Figure 76 shows a recommended update to the Set Processing Options Menu. This includes the addition of the Ground Cover section. The “Calculate mixed ground impedance effects” option turns on the mixed ground impedance calculation method in AEDT. Once selected, it gives the user options to either use a “Uniform ground cover” value or “Geospatially referenced ground cover data”, the later referencing the user-supplied ground cover file. If “Geospatially referenced ground cover data” is selected, a “Fill ground cover (kPa*s/m2)” option is also made available, which fills any gaps in study coverage in the ground cover file with a user specified estimated flow resistivity value. In addition, if “Calculate mixed ground impedance effects” is selected, the “Use hard ground attenuation for helicopters & propeller aircraft” option should be deactivated (“grayed out”).

T and the ad ground co FIGURE 76 he following dition of a n ver on noise FIGURE 7 . Recomme text is a reco ew sub-sectio propagation ( 7. Recomm account for nded update mmend supp n 5.2.4.5 “G see Figure 77 ended upda ground cove s to the AED options men lement to su round Cover and Figure tes to AEDT r effects on n T define me u. b-section 5.2 Options” to a 78). 2c sp1 User oise propag tric results – .4.3 “Noise M ccount for th ’s Guide Sec ation model set process odeling Opt e effects of m tion 5.2.4.3 t ing. G-4 ing ions” ixed o

F G.3. Mod It existing im SAE-AIR assumes computati taken wh ground. If potential associated other area IGURE 78. account fo eling Guid is recomme pedance me -5662 “Meth propagation onal runtime en implemen study comp solution is to with water s are modele Recommen r ground co ance nded that th thod currentl od for Predic over soft gr ) that is anti ting ground lexity becom create a cus is modeled w d with soft gr ded new sec ver effects P e mixed imp y implemente ting Lateral ound only. cipated from cover data es an issue, tom land cov ith an estim ound using 1 tion to AEDT rocessing Op edance adju d in AEDT, Attenuation o Due to the using the m that does no or the appro er file (base ated flow res 50 kPa*s/m2. 2c sp1 Use tions for no stment be ap which is a lev f Airplane N increased c ixed impedan t differ sign priate land c d on the NL istivity valu r’s Guide (S ise propagat plied as a el-based adj oise” (2006 omplexity (a ce adjustme ificantly from over data ar CD format), e of 100,000 ection 5.2.4.5 ion modelin supplement t ustment defin ). SAE-AIR nd correspo nt, care shou the defaul e not availa where all the kPa*s/m2, a G-5 ) to g o the ed in -5662 nding ld be t soft ble, a area nd all