August
11
, 2020
Dear Port of Seattle Commission,
We’re requesting the Port of Seattle make a procedural change in how the data from your system of port owned Larson Davis
831 noise monitors is collected and retained. This is particularly timely given that the contract with the vendor who provides
services associated with the monitors, L3Harris, is due for renewal as per Item 6e in today’s commission meeting.
Currently the Port does not retain the raw second by second measured sound level data but instead contracts the data
collection and analysis to the vendor L3Harris. The vendor periodically downloads this data and then purges it from the
monitors. They use this data to build a list of SELs (Sound Exposure Level) for overflight events at that monitor location along
with a set of noise statistics for the site. Nothing precludes the Port from also downloading a copy of their own raw second by
second data from the Port owned noise monitors before L3Harris purges it. Alternatively, if the Port’s Noise Office wants to
retain a completely hands-off relationship with their own noise monitors, the vendor L3Harris could deliver the raw data with
the other processed data currently being supplied. This raw data amounts to ~62 MB/month (~40 MB compressed) per
monitor, which is tiny by today’s standards.
At the UC Davis Aviation Noise & Emissions Symposium in March I spoke with L3Harris, who was a sponsoring vendor and had
a booth there. Samuel Carter, their Training & Solutions Manager, indicated it would be a simple modification of the
procedure to deliver the raw data with the processed results. The Port simply has to request it. I didn’t ask, nor did he say, if
there would be an extra cost for that. However, again, this is Port owned data on Port owned noise monitors so the Port can
simply download this data themselves periodically before L3Harris purges it.
This raw data is absolutely essential for any rigorous analysis of the aviation noise levels at the site. As a real-world example, I
was looking over the processed data from last December and came across an interesting anomaly. Here is a map of six noise
monitor sites immediately North of the Airport including overflight event counts for December:
Vashon Island Fair Skies
PO Box 1250
Vashon, WA 98070
http://www.vifs.org
info@vifs.org
(206)682-8638
Dedicated to restoring the pre-NextGen dispersed arrival paths and more optimized
profile descents at Seattle/Tacoma International Airport that had been in place since
the introduction of commercial aviation to the Puget Sound region, many decades ago.
What’s odd here is how small the event count at site #12 was. It was near the axis of the runways, as were sites 11 and 13,
yet had a much smaller event count. Site #1, on the grounds of the airport proper also had a low count. In that case the
overall background noise of the airport could drown out the signal from individual operations. Site #15 was laterally
separated from the runways which could explain its low count. However site #12 doesn’t have a good explanation for being
so low.
By looking at the runway data and focusing on the triangle of sites North of the airport (#11, #12, and #13) we can see the key
difference for site #12.
Runway (Southflow)
Runway (Northflow)
Site
16L
16C
16R
34L
34C
34R
Unknown
SEA01
3676
970
2218
74
10
702
232
SEA11
2344
306
31342
0
6
724
0
SEA12
4750
236
304
0
16
778
0
SEA13
228
48
24758
0
10
656
0
SEA14
6446
1186
14692
0
10
790
160
SEA15
308
160
498
2
0
12
0
Unsurprising for December, Southflow was much more common than Northflow. In Northflow, these departures are high
throttle (i.e. loud) and the three sites give similar numbers, or at least the same order of magnitude. In Southflow however
something emerges. The big discrepancy with site #12 is due to 16R Southflow arrivals.
Site #11 is the most centered of the three and its numbers tell the story of runway use for Southflow arrivals. By a large
margin 16R, the “Third Runway”, is the most commonly used. As an aside, during the legal battles over building the third
runway, the community was assured it would only be used in rare circumstances, but that’s another story. So, site #12, being
East of the airport is the furthest from 16R, but we’re talking about a two orders of magnitude reduction in events from Site
#11 to Site #12 even though Site #12 is only ~3300 feet East of Site #11.
That small separation shouldn’t account for such a huge discrepancy. I made Public Records Request 20-19 to get the raw
data from site #12 for December to investigate what was causing this. The data came with this ominous warning:
”The vendor, L3Harris was able to extract the data. Please note that this is not data that the Port keeps in-
house on its Noise Monitoring system. L3Harris has indicated that all future records we obtain from them
are subject to an additional fee that we would have to pass on to you.”
With this data it became clear why site #12 was reporting so many fewer events.
For the purpose of this illustration, I’ll cover a short period on the morning of 12/1/19:
Date/Time
Noise Monitor
Flight ID
Operation
Equipment
Runway
SEL
12/1/2019 8:30:09
SEA12
DAL166
Arrival
A339
16L
85.47776989
12/1/2019 8:30:58
SEA11
ASA85
Arrival
B739
16R
86.59586146
12/1/2019 8:33:11
SEA11
ASA368
Arrival
B738
16R
86.13313923
12/1/2019 8:35:07
SEA11
QXE2535
Arrival
E75L
16R
83.59008407
12/1/2019 8:35:10
SEA12
ASA625
Arrival
B739
16L
83.32092607
12/1/2019 8:36:32
SEA11
ASA169
Arrival
B738
16R
85.4135081
12/1/2019 8:37:20
SEA11
KAL019
Arrival
B77W
16L
77.66649632
12/1/2019 8:37:41
SEA12
KAL019
Arrival
B77W
16L
87.66999824
12/1/2019 8:38:08
SEA11
ASA1017
Arrival
A320
16R
85.72240682
12/1/2019 8:39:33
SEA11
CPZ5782
Arrival
E75S
16R
83.028042
12/1/2019 8:41:02
SEA11
ASA894
Arrival
B738
16R
86.45152272
12/1/2019 8:42:12
SEA12
QXE2433
Arrival
E75L
16L
81.04083749
Note that only KAL019 was recorded by both SEA11 and SEA12 and were separated by about 21 seconds. These events are
attached to the sound level graph below, with the flights detected by SEA12 written in black above the peaks and the flights
only detected by SEA11 written in gray under the smaller peaks. So small in fact, that the algorithm did not consider them to
be events:
The issue here is the very high background noise level at this site, which causes the signal to not be sufficiently above the
ambient noise level to convince the software that it’s an overflight.
Retaining the raw data from the Port owned noise monitors is critical, and instead allowing the data to be purged violates the
spirit, and perhaps the letter, of the Public Records Act (RCW 42.56).
We filed Public Records Request 20-314 asking for a copy of the proposed contract in item 6e, and we request that this item
be removed from today’s Unanimous Consent Calendar until we’ve had a chance to review the changes to the contract.
Thank you,
David Goebel
President, Vashon Island Fair Skies