Evaluating Methods for Counting Aircraft Operations at Non-Towered Airports (2015)

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.

50 C H A P T E R 4 Conclusions Methods for estimating aircraft operations and methods for counting airport operations were studied under this research. The conclusions are presented separately for each of these research items. Methods for Estimating Annual Airport Operations The methods for estimating annual operations that were studied included the following: 1. Multiplying based aircraft by an estimated number of OPBA, 2. Applying a ratio of IFR flight plans filed to total opera- tions (IFPTO), and 3. Extrapolating from a sample count. OPBA A study was performed to determine if there is a consis- tent number(s) of OPBA that occur at STAD that could then be applied to non-towered airports. (For a discussion on the STAD database, please refer to Chapter 2 of this report.) The study also considered if the OPBA varied by climate or popu- lation and if having a flight school(s) affected this number. Ini- tial analysis revealed that an extremely large range of OPBAs exist for the STAD airports overall and by region, and practi- cal use of any averages would not produce confident results. With this in mind, the research team attempted to actually model total OPBA through regression analysis to determine if an equation could be produced that offered better results. While several different approaches were taken, including full model regression, reduced model regression, and transforma- tion of the data, either the statistical assumptions necessary for the regression to be valid could not be met or there were extremely large variations from actual to estimated operations on the test airports. Accordingly, based on the variables stud- ied there does not appear to be a consistent number of OPBA at small, towered airports that can be applied to non-towered airports for use in estimating airport operations nationally or by climate region. Consequently, the research team cannot recommend using a standard number(s) of OPBA for esti- mating annual airport operations. While FAA Order 5090.3C Field Formulation of the National Plan of Integrated Airport Systems (NPIAS) gives general guidance on OPBA values, applying those values in practice with any degree of confi- dence is difficult at best. Continuing research in this area may be tempting, but this study and historical research has shown that developing statistical models for total operations may be more descriptive than models for OPBA. IFPTO An analysis was performed on small, towered airports to determine if a consistent ratio of IFPTO occur at these facilities that could then be applied to non-towered airports. Overall, the research team concludes that based on the study objectives and data, there are no practical and consistent IFPTOs found at small, towered airports that can then be used to estimate annual operations at non-towered airports nationally or by climate region. Consequently, the research team cannot rec- ommend using a standard ratio(s) of instrument flight plans to total operations for estimating annual airport operations. However, since IFR operations are tracked by the FAA for all airports, an IFPTO could theoretically be computed for a specific airport by sampling all operations, counting all IFR flight plans filed for that same time period from FAA records, and then determining the ratio. The number and times of the samples will impact the results, so care should be taken to ensure operations sampled are truly representative of the actual operations that occur throughout the year. While sam- pling scenarios for IFPTO were not studied in this project, Conclusions and Suggested Research

51 two weeks per season will most likely be needed to adequately cover variations in activity by day of week, weather, and season (see the next section for research completed on sampling sizes and timeframes). These IFPTO ratio(s) could then theoreti- cally be used to project total operations at that specific airport from total FAA tracked IFR operations for that airport. How- ever, the IFPTO ratios would need to be updated regularly to ensure they remained representative. Extrapolating from a Sample Count An analysis was performed on two methods to extrapo- late a sample count of aircraft operations into an annual estimate of aircraft operations. The first method was a sta- tistical extrapolation that follows FAA-APO-85-7, Statisti- cal Sampling of Aircraft Operations at Non-Towered Airports. Based on the analysis using this method and four sampling scenarios, the preferred statistical extrapolating scenario is to sample two weeks per season. This produced the smallest range of estimated operations to actual operations at the test airports. This is consistent with the previous research results discussed in ACRP Synthesis 4: Counting Aircraft Operations at Non-Towered Airports. The second method studied for extrapolating a sample of airport operations into an annual estimate was by the use of regional monthly or seasonal adjustment factors developed from small, towered airports (i.e., STAD). Based on the sta- tistical analyses, the sampling scenario of two weeks each sea- son is still preferred by the research team when using regional monthly or seasonal adjustment factors. While the statistical analyses did not find a significant difference between the sam- pling scenarios except for one month winter and one month spring, summer, or fall, there is a difference in the average per- cent difference and the range of percent differences that may be observed in Table 3-8 in Chapter 3. The two weeks each sea- son sampling scenario has a combination of statistics reported that indicate preference over the other methods. To determine if a statistical significance truly exists for these scenarios, more airports would need to be studied in a future research project. Using the regional monthly/seasonal adjustment factor method makes several assumptions, one being that variations in traffic at the STAD airports adequately represent traffic at non-towered airports. This method also requires calculating the adjustment factors, which are not needed if the statistical extrapolation process outlined in FAA-APO-85-7 is followed. The FAA-APO-85-7 method uses sample counts from two weeks each season from the specific airport where operations are being estimated, rather than depending on external fac- tors. Additionally, use of FAA-APO-85-7 provides a percent sampling error to measure the precision of the annual opera- tions estimate. The monthly/seasonal adjustment factor does not provide a percent sampling error. For all these reasons, the research team believes that statistical extrapolation is the favored process. Consequently, to estimate operations at a non-towered air- port, the research team recommends taking sample counts at the airport for a minimum of two weeks in each season and then using the statistical process outlined in FAA Report No. FAA-APO-85-7, Statistical Sampling of Aircraft Opera- tions at Non-Towered Airports, to extrapolate the samples into an annual estimate. Aircraft Traffic Counters This research involved evaluating four different aircraft traffic counting technologies (that can be used to take sam- ples that are then extrapolated into an annual operations esti- mate for an airport). The technology identified for evaluation included the following: 1. ACC (portable acoustic counter), 2. SMAC (portable acoustic counter), 3. S/TC (portable camera with infrared night vision), and 4. VID (stationary) with ADS-B transponder receiver (stationary). Please refer to Research Approach in Chapter 2 for detailed information on the technology and the evaluation methods. Automated Acoustical Counter Acoustical counters have some inherent limitations, but they can offer a reasonable estimate of operations for a rea- sonable cost if positioned appropriately along the runway and the resulting data is audited correctly. Acoustically based counters record takeoffs and work on the assumption that for every takeoff there is a landing, so the total count is doubled for an estimate of operations. (Note: The original acoustically based counters used analog recordings on cassette tapes, which were then listened to and audited for aircraft and non-aircraft sounds. This form of acoustical counter is no longer manufac- tured. The new generation of acoustical counters are digital and do not offer a listening option for auditing the sounds recorded. While the analog recorders were extremely more labor intensive than the new digital ones, they may have been more accurate in part because of the ability to audit the cassette tape.) The AAC is rugged, dependable, and can be left for months at a time even in below freezing temperatures when a solar panel option is used. On a typical single runway airport, the AAC offers a fairly accurate estimation of annual operations if multiple units are used and positioned properly. The manufacturer of the AAC tested claims that recorded takeoffs will be within 10% of actual takeoffs. This claim was substantiated when at least one counter was within about

52 700 feet of a point perpendicular to the aircraft’s point of rotation (lift-off point) and flight path. The AAC was originally designed for use at small airports with short runways (e.g., approximately 2,000–3,000 feet). Once the runway gets much longer than this, multiple units are needed to achieve the claimed accuracy rate. (The units cost approximately $4,800 each.) However, when more than one counter is used, the raw data will likely need to be manually manipulated to remove duplicate counts because there will be times when more than one counter counts the same takeoff. The manufacturer of the unit tested did not anticipate the use of multiple counters so there was no automated feature to remove double counts from the raw data and the process was cumber- some, time consuming, and prone to human error. Addition- ally, the more counters that are used on a runway, the more susceptible the results are to human error because removing duplicate counts is tedious but requires great attention to detail. The results of the study appear to indicate that the AAC can be positioned as far as 250 feet from the runway centerline and still identify an aircraft takeoff if positioned close enough perpendicularly to the point of aircraft rotation, as described above. However, use of the counters on an airport with mul- tiple runways is very difficult because of double counting, vari- ous wind conditions, and numerous possible rotation points. It is important to note that the AAC has some trouble iden- tifying exceptionally quiet aircraft. The Continental O-300 SER was the engine in a Cessna 172 based at one of the test airports, and the counter missed it the majority of the time. The longest study with the most data was performed at Indianapolis Executive Airport (5,500-foot runway) where three counters together produced results within 8% of actual takeoffs. On average across all the airports when just one counter was used in the middle of the runway, the equipment caught less than 50% of the airport’s traffic. AAC Highlights. Best used at airports with single run- ways with RSAs of 500 feet or less that do not experience sig- nificant traffic by exceptionally quiet aircraft. • 90% accuracy or better can be achieved if located properly. • Locations need to be sufficiently tested to ensure takeoffs are being recorded. • Aircraft lift-off (rotation point) should be within approxi- mately 700 feet of a point perpendicular of the counter to be consistently counted (Figure 4-1). • Multiple counters are required for runways approaching 3,000 ft. or greater. This makes this option more labor intensive. • Exceptionally quite aircraft are missed more often than counted (e.g., Cessna 172 with Continental O-300 SER engine was missed at a distance of approximately 50 feet of the unit). These are typically SEP aircraft. Jets, turbo props, and multi-engine piston aircraft are typically louder and are not missed as often. • Helicopters are harder to count because they do not have a uniform landing path; to be counted they have to fly over the general area of the counter to be detected. • Airports with multiple runways are difficult to count with consistent, acceptable accuracy (Figure 4-2). • No information on aircraft type, make, or model is provided. • The counter achieved 90% or greater accuracy when posi- tioned as far away as 250 feet from the runway centerline. (FAA required submission of FAA Form 7460 and units had to be outside of the RSA.) Sound-Level Meter Acoustical Counter Also acoustically based, the SMAC is a portable unit that counts takeoffs, which are then doubled for an estimation of operations. It is rugged, works in most all weather types, and is fairly easy to use. However, its useful life between battery charges is severely limited, especially in cold weather, and the sound-level meter must be calibrated regularly. Cost is approximately $4,800 per unit. On a typical single runway airport, the SMAC performs similar to the AAC and also offers a fairly accurate estima- tion of annual operations if multiple units are used and posi- tioned properly. The manufacturer claimed accuracy rates of 5–10% of actual operations, and like the AAC, this can likely be achieved but will require multiple units when the runway approaches 3,000 feet in length or greater. Additionally, the SMAC appeared to rely more on false positives to achieve this rate than the AAC did. More specifically, the SMAC appeared Counter can be as far as 250 feet from runway centerline. Aircraft lift-off (rotation point) should be within approximately 700 feet of a point perpendicular of counter, which may require multiple counters. 700 feet AAC 700 feet Figure 4-1. Aircraft lift-off.

53 Also like the AAC, use of the counters on an airport with multiple runways is very difficult because of double count- ing, various wind conditions, and numerous possible rotation points. SMAC Highlights. Best used at airports with single run- ways and RSAs of 150 feet or less that do not experience sig- nificant traffic by exceptionally quiet aircraft. • 90% accuracy or better can be achieved if located properly and generally not more than 75 feet from runway centerline. • Locations need to be sufficiently tested to ensure takeoffs are being recorded. • Aircraft lift-off (rotation point) needs to be within approx- imately 700 feet of a point perpendicular of the counter to be consistently counted (Figure 4-3). • Multiple counters are required for runways approach- ing 3,000 ft. or greater; this makes this option more labor intensive. • Exceptionally quite aircraft are missed at distances greater than approximately 50 feet of the runway centerline (e.g., Cessna 172 with Continental O-300 SER engine). These are typically SEP aircraft. Jets, turbo props, and multi-engine to count a large number of non-events, which were included in the total and thereby significantly improved its computed accuracy level. The SMAC is more impacted by distance from the runway centerline than the AAC (optimal seems to be 75 feet or less). The farther away from the runway centerline, the more dif- ficulty it has detecting takeoffs. For this same reason, it is a bit more likely to miss a touch-and-go than the AAC. However, at closer distances to the runway (e.g., 50 feet), it seems better at detecting takeoffs by the relatively quieter aircraft (Cessna 172 with Continental O-300 SER) than the AAC. At 250 feet from the centerline, the research team was unable to achieve an acceptable level of performance. Like the AAC, when more than one counter is used, the raw data has to be manually manipulated to remove duplicate counts. The use of multiple counters is not addressed in the equipment user manual so there was no automated feature for removing the duplicates and the process was cumbersome, time consuming, and prone to human error. Additionally, the soft- ware included with the counter for performing the statistical extrapolation outlined in FAA-APO-85-7 was not designed to take into account the use of multiple counters, so it could not be used without some reprogramming. Prepared by: Woolpert, Inc. Figure 4-2. Example of configuration conducive for AAC. Counter can be as far as 75 feet from runway centerline. Aircraft lift-off (rotation point) should be within approximately 700 feet of a point perpendicular of counter, which may require multiple counters. 700 feet SMAC 700 feet Figure 4-3. Aircraft lift-off.

54 operations and registration numbers. This is time consuming at a busy airport. Security/Trail Camera Highlights. Best used at airports with centralized terminal and hangar area with limited access points and little touch-and-go activity (Figure 4-5). • Accuracy levels approaching 100% can be achieved for recording aircraft entering or exiting the runway environment. • Unable to count touch-and-goes. • Exceptionally slow moving aircraft may be missed. • As ambient temperature approaches temperature of target aircraft, target may be missed. • Labor intensive because manual tally of images is required. • Information on aircraft type, make, and model can be obtained from aircraft registration number. • Low cost for airports with simple airfield configurations. • Can also be used for detecting wildlife. VID and ADS-B Transponder Receiver Also camera based, the VID system provides a more com- prehensive counting package, but also comes with a much higher price tag. The unit tested (two cameras, ADS-B tran- sponder receiver, and web portal with service contract) cost $36,000 for the seven months it was leased. As the airfield configuration becomes more complex, more cameras are needed and the cost increases accordingly. Once installed, the annual service contract varies depending on the amount of traffic the airport experiences. The service provider also supplements the camera equip- ment with the FAA’s electronic tracking of aircraft known as the ASDI. If positioned correctly on an airfield with a condu- cive configuration, the VID can provide a reasonable level of piston aircraft are typically louder and are not missed as often. • Helicopters are harder to count because they do not have a uniform landing path; to be counted they have to fly over the general area of the counter to be detected. • No information on aircraft type, make, or model is provided. • Airports with multiple runways are difficult to count with consistent, acceptable accuracy (Figure 4-4). • The counter achieved 90% or greater accuracy when posi- tioned as far away as 250 feet from the runway centerline. (FAA required submission of FAA Form 7460 and units had to be outside of the RSA.) Security/Trail Camera A more recent way to count aircraft operations is with the use of motion detection cameras. These can be as simple as a stand-alone S/TC or as sophisticated as a VID system, which is described next. The S/TC that are self-contained are the easiest to use and install since there are no power needs in the field. A stand-alone, solar powered camera with a range of 100 feet can be purchased for under $1,000 (which includes memory cards, batteries, solar panel, and cabling). The S/TC tested had a passive infrared motion detector with a nighttime infrared illuminator. When located cor- rectly on an airport conducive for its use, it can provide an exceptionally high level of accuracy in recording aircraft entering and exiting the runway environment, which can be equated to takeoffs and landings. However, it cannot count touch-and-goes. At less than $1,000 per unit, it can be a cost- effective way to estimate operations for airports with little touch-and-go activity or where the percentage of touch- and-goes is known. Use of this type of equipment gives the added benefit of knowing aircraft registration numbers, but it also requires manual review of the images to determine total Prepared by: Woolpert, Inc. Figure 4-4. Example of configuration conducive for SMAC.

55 ADS-B, this technology will not provide a reasonably accurate way to count an airport’s traffic. VID and ADS-B Transponder Receiver Highlights. Best used at airports with centralized terminal and hangar areas with limited access points and little touch-and-go activity (Figure 4-6). • Accuracy levels as high as 90% were achieved for recording aircraft entering or exiting the runway environment. • Unable to count touch-and-goes. • ADS-B transponder receiver option adds little to no value considering the low equipage rate of the U.S. general avia- tion fleet with ADS-B out. accuracy in recording aircraft entering and exiting the runway environment, which can be equated to takeoffs and landings. It also offers detailed information on the aircraft type, make, model, owner, etc. Like the security/trail cameras, it does not appear capable of counting touch-and-goes. The VID is expensive and requires an annual service contract; however, it can offer a low labor intensive way to estimate operations for airports with little touch-and-go activity or where the percentage of touch-and-goes is known. The transponder receiver, programmed to detect ADS-B transmissions, did not perform well because of the very low number of aircraft equipped with ADS-B technology and because of technical issues with the equipment and software itself. Until the major- ity of the U.S. general aviation fleet becomes equipped with Prepared by: Woolpert, Inc. Example of configuration conducive for S/TC. Example of difficult configuration for S/TC. Figure 4-5. Example of configuration for S/TC. Example of configuration conducive for VID. Example of difficult configuration for VID. Prepared by: Woolpert, Inc. Figure 4-6. Examples of VID configurations.

56 Toward the end of this research project, an inquiry was received from a company with a new type of technology for counting aircraft. The technology monitors the airport Uni- com frequency and uses automated speech recognition to identify and record airport traffic. The company’s system lit- erature purports the ability to provide information about the aircraft, including the date and time, aircraft make and model, aircraft type, operation type (e.g., overflight, takeoff, landing), runway used, and optional registration number and picture of the aircraft. This technology should be monitored for possible future evaluation as a way to provide aircraft operations tallies at non-towered airports. As the deadline for incorporating ADS-B out technol- ogy into all aircraft that operate in specific U.S. airspace gets closer, this technology should be readdressed. While most non-towered airports are typically not inside the airspace where ADS-B enabled avionics will be required, some may be near enough to it that the vast majority of the aircraft using them will be equipped with it. And, as general aviation aircraft owners learn the benefits of ADS-B out technology, more are likely to equip their aircraft accordingly. In short, if the overall fleet equipped with ADS-B capable avionics reaches high enough levels, transponder receivers, with the correct algorithms programmed, could be a viable option for counting aircraft operations at non-towered airports. • Most expensive option. • Least labor intensive option. • Requires service contract. • Can also be used for automated billing of landing fees. Suggested Research While still the best option, the current practice of sam- pling operations with aircraft traffic counting technol- ogy is subject to two types of errors: equipment error and sampling error. If an accurate, but relatively inexpensive, aircraft traffic counter were developed that could be used regardless of the airport configuration and traffic mix, it could be deployed at most all non-towered airports and left in place year-round—alleviating the need for sampling all together. While the existing aircraft traffic counting tech- nology can provide reasonable estimates of operations in certain conditions, better and less expensive equipment is needed if truly accurate and comparable data is desired on a large scale basis. If the existing acoustical counting equipment is ever improved to better detect the quieter aircraft and to automate the use of multiple counters on one runway, additional testing of this equipment would be warranted.