Flight paths over Santa Cruz
In March 2015 NEXTGEN arrived in Santa Cruz. The FAA simultaneously made three changes to SFO-bound flight paths over the county :
- flight concentration over the ground track increased, resulting in a more concentrated stream of traffic,
- airplane altitude dropped, making for noisier planes
- the ground track was shifted a couple of miles East.
The sum effect of these changes was a dramatic increase in complaints, though total flight traffic remained unchanged. As a possible remedy, the FAA is now asking for public comment on a new flight path, DAVYJ, that addresses item 3 but not items 1 and 2. The proposal is controversial. If you live near the current or previous flight path, adoption of DAVYJ will affect you. This site presents information that can help evaluate the DAVYJ proposal. A closing section lists resources from which more information is available and covers technical details. This should be sufficient to allow anyone interested to reproduce and confirm all results presented.
As a start, it’s helpful to visually compare the changes. Here is a heatmap that displays the distribution of total flight traffic in July 2014
and here is total flight traffic in July 2015
With the current ground track, SERFR, flights cross land from the Monterey Bay near Capitola. With the ground track in effect in 2014, BSR(Big Sur), flights crossed land near the Boardwalk. The proposed DAVYJ path would follow the BSR ground track. In comparing the above figures, the West-to-East shift is obvious, as is the increased concentration evident in the reduction of flight traffic west of the ground track. These changes are covered in more detail later.
Which flight path affects more people?
By overlaying the ground tracks over a map of Census Bureau population density it is easy to confirm which track affects more county residents:
The SERFR route clearly travels over less populated areas and thus affects fewer people. Though both overfly densely populated neighborhoods near the coast, BSR continues over residential neighborhoods that flank Graham Hil Rd. including Paradise Park, Ocean St. Extended, Pasatiempo, West Scotts Valley, Woods Cove and Mt. Hermon. SERFR avoids most of Scotts Valley but overflies neighborhoods East of Granite Creek Rd.
The color scheme used by ArcGIS is sufficiently broad to capture population density variation across the entire country. However Santa Cruz county consists largely of sparsely populated rural and residential communities. Here is another map that provides a more locally-focused view. The map uses a 6 color scheme with cutoffs of 100,200, 500, 1000 and 1500 persons/square-mile and relies directly on population counts and block-group polygon boundaries available from the Census Bureau. Both maps support the same conclusion: the SERFR track overflies fewer residents.
The actual number of people directly affected by the two ground tracks can be calculated by overlaying the tracks, widened by a 1.5 mile-wide shadow, over the population density map
The result is 32,766 for SERFR and 34,256 for DAVYJ, an increase of 1490 county residents. Details of the calculation are included in the closing section below. This approach only gives an approximation to the number of people affected since it does not consider vectoring.
Did flight concentration increase with NEXTGEN deployment ? Will DAVYJ fix this ?
The second of these two questions is the easier one to answer : no, DAVYJ includes no provisions to fix the flight concentration that accompanied NEXTGEN.
Answering the first question requires a way to measure flight concentration. This is the approach we took:
- First, we defined a rectangle sufficiently large to cover all parts of the county overflown by SFO-bound flights. This rectangle, which covers about 313 square miles, is shown in the image below.
- We divided this rectangle into 1080 equally spaced “plots” or tiles.
- The total number of flights that flew over each plot in July 2014 and July 2015 was counted. These counts do not take into consideration altitude or noise but measure how much air traffic flew overhead.
The values of flight counts over the plots are displayed in the two heatmaps shown above. Though the total amount of traffic remained largely unchanged, the increased concentration is obvious : the large areas of light blue towards the western part of the county are gone in 2015.
It’s possible to confirm the increase in concentration by looking at the changes in flight counts without reference to a map. Comparing the 2014 and 2015 counts of plots with a large amount of traffic demonstrates the increase. Here are values for the 12 plots that received the most traffic in the two years:
The increase in concentration is also apparent when comparing histograms of the flight counts:
Note that the 2015 histogram includes many plots with flight counts that exceed the maximum value of 2589 in 2014 and that there is a corresponding decrease in plots with lower flight counts. In other words, houses under the flight path experienced much more traffic.
The FAA reports that vectoring remained unchanged after the transition from BSR to SERFR. Doesn’t this statement contradict the evidence of increased flight concentration?
It’s important to note that vectoring and flight concentration measure different things. Vectored flights are those that are directed off the ground track as the plane approaches the airport, primarily to reduce congestion. Over Santa Cruz, vectored flights are invariably directed westward. The FAA estimates that the proportion of flights directed off the ground track on BSR was approximately 50% and that this proportion did not change with SERFR. From Appendix E of the Feasibility Report :
“For both the BSR and the SERFR, approximately 50% of the traffic is vectored off the procedure.”
Flight concentration measures flight counts over the parts of the county that receive the most flight traffic, whether under the ground track or off it. Thus there is no contradiction between these results. In Appendix E, the FAA also included a visualization of individual flight paths, shown below, that closely corresponds to the flight count heatmaps.
Did flight altitude drop with NEXTGEN ? Will DAVYJ fix this ?
At the August 4 meeting of the Select Commitee the FAA presented information on the average altitudes for BSR as flown in July 2014, SERFR as flown in July 2015 and the current estimates of expected flight altitudes for DAVYJ. No actual data for DAVYJ is available and altitudes were given as an estimated range with minimum and maximum values. The values presented in the FAA slides are shown in the charts below (the actual values used are listed in the “Methods” section).
Altitude decreases as airplanes approach the airport; the X axis measures an airplane’s distance from the MENLO waypoint, south of SFO. Flight altitudes are measured as elevation above seal level MSL, or above ground level AGL; both are shown. The dashed vertical black lines mark the point from which DAVYJ traffic is projected to fly lower than current SERFR.
Clearly, introduction of SERFR resulted in lower flights. In conjunction with the increase in flight concentration, the resulting noise is a likely reason for the increase in complaints. Altitudes for the three flight paths can be easily compared by selecting SERFR as a baseline and plotting whether BSR or DAVYJ lie above or below this baseline
Features of these plots help understand how DAVYJ flight traffic can be expected to perform:
- Under the worst-case scenario, DAVYJ flight traffic will be everywhere worse than current SERFR.
- Under the most-likely, mid-value scenario, DAVYJ traffic will be everywhere worse than the older, BSR, flight traffic.
- Also, under the most-likely scenario, DAVYJ traffic will be worse than current SERFR for communities in the vicinity of MENLO. For the AGL case, this occurs at 16.2 miles from the waypoint. The map below shows the neighborhoods that fall within this range.
Projections from a DNL model show no noise difference between DAVYJ and BSR as flown in 2014. Does this mean DAVYJ can be expected to be as quiet as BSR?
At the September 1 working meeting of the Select Committee, the FAA presented the following slide
the key message of the graph is that the FAA’s DNL model (Day-Night Average Sound Level) detects no difference between DAVYJ and flights flown over BSR in 2014 except in the vicinity of MENLO. Comparisons near MENLO are complicated by the fact that flights are now required to drop to an altitude of 4000 feet at that waypoint, whereas, as evident from the altitude plots, in 2014 the average altitude there was about 4800 feet.
Output from DNL modeling needs to be interpreted alongside other available information in gauging the likely impact of DAVYJ. As documented above, post-NEXTGEN flights are lower and more concentrated. DAVYJ will not address either of these issues. It does nothing to mitigate increased concentration and, as evident in the altitude plots above, the DAVYJ mid-point line is everywhere below the BSR line.
DAVYJ will not restore BSR altitudes, only its ground track. Nevertheless, the FAA’s DNL model predicts no noise impact from flying the BSR track at lower altitude and with greater concentration. Is this credible ? There is no definitive answer to this question. However there are a number of reasons to be skeptical of DNL as a reliable indicator of real-world noise impact:
- DNL models, developed in the 1970s, are antiquated and not based on actual noise measurements. The FAA is in the course of overhauling the methods (see FAA press release https://www.faa.gov/news/press_releases/news_story.cfm?newsId=18774 ).
- As affirmed by Mr. Glen Martin in response to a question from the Select Committe, DNL models do not incorporate flight concentration in gauging impact.
- DNL models obviously failed to alert the FAA to the impact of NEXTGEN implementation or to predict the resulting spike in noise complaints.
In short, there are sound, but not conclusive, reasons to expect that the DNL predictions are unreliable. Moving post-NEXTGEN noise to the BSR ground track without addressing increased concentration and lower altitude will simply aggravate different residents.
Summary : will DAVYJ help you? What can you do?
The above data supports the following conclusions:
- DAVYJ will fly over more populated communities in Santa Cruz and affect more residents
- DAVYJ includes no provision for mitigating the increased flight concentration that accompanied SERFR : this noise will be transferred onto new residents.
- DAVYJ does include altitude adjustments , which, under the most optimistic circumstances would match altitudes of the older BSR flight path over parts of Santa Cruz. However, with more realistic mid-range estimates, DAVYJ will be everywhere worse than BSR.
- Under all estimates, DAVYJ will be worse than SERFR for residents who live closer to SFO.
In short, instead of fixing the problems that accompanied NEXTGEN, DAVYJ will simply transfer them onto more people. If this does not seem a good idea, please contact the Select Committee and let them know. Contact information is available here http://eshoo.house.gov/wp-content/uploads/2016/08/Select-Committee-Roster.pdf
Councilman Don Lane and Supervisor Bruce McPherson have already stated they do not favor alternatives that simply move noise. Your support is important. A petition that opposes DAVYJ is available here https://www.change.org/p/faa-stop-the-quiet-skies-norcal-proposal-which-seeks-to-move-an-sfo-flight-path-to-sc-slv-sv
Resources and Methods
- A full list of presentations by local advocacy groups and the FAA maintained by the office of Congresswoman Eshoo: http://eshoo.house.gov/constituent-services/airplane-noise-in-the-18th-congressional-district/
- A petition to oppose the DAVYJ proposal: https://www.change.org/p/faa-stop-the-quiet-skies-norcal-proposal-which-seeks-to-move-an-sfo-flight-path-to-sc-slv-sv
- The web site of the SLV advocacy group: http://www.sanlorenzovalley.info/
All flight path statistics are based on FAA TFMS data obtained through a Freedom of Information Act (FOIA) request. The July 2014 and July 2015 intervals are selected because they correspond to those used by the FAA in the analysis included in Appendix E of the Feasibility report.
Population count data comes from two sources. ArcGIS maintains an online map service of US population density based on 2012 Census Bureau data. We also relied on “block group” population counts, polygon boundaries and polygon areas obtained directly from the Census Bureau. In analyzing population data it is crucial to use high-resolution “block group” units. Distances between the two ground tracks are only a couple of miles and population data aggregated to larger units, such as the “census tracts” used in the FAA’s slides from August 4, obscures differences.
The latitude longitude ranges of the rectangle used to calculate flight counts are (36.95,37.233) and (-122.19, -121.90). Latitude was subdivided into 36 intervals and longitude into 30 intervals yielding 1080 “plots” or tiles. Because of the earth’s curvature, the areas of the 1080 plots are not equal though they are equi-distant. This effect is not expected to be large and in any event should not affect pre/post NEXTGEN comparison since the same subdivision was used for both.
The comparison of flight elevations for BSR,SERFR and DAVYJ is based on the following altitude above ground level (AGL) and above mean sea level (MSL) values obtained directly from slides 10, 11 and 18 of the FAA’s August 4 presentation to the Select Committee.
In calculating the Total Persons Affected/Annoyed/Aggravated (TPA) for SERFR and DAVYJ, polygons downloaded from the Census Bureau for all Santa Cruz county block groups were intersected with a 1.5 mile-wide shadow bordering each of the the two ground tracks. For example, here is a detail of two Scotts Valley blocks groups
All 1487 people in block group 060871212002 off Granite Creek Road to the East of Hwy 17 were included in the SERFR TPA count since that block group is entirely covered by the SERFR shadow. However the TPA contribution for block group 060871209003, which includes Scotts Valley High School West of Hwy 17 and counts 2774 residents, is 1241.7 since less than half (0.448) of that block group intersects the SERFR shadow.
TPA values for block groups that are overflown by SERFR are
and TPA values for block groups overflown by the DAVYJ/BSR track are