MEMORANDUM
TO: LFA/Eric Bernhardt, Deputy Project Manager
FROM: Navigation Committee/Dorn Crawford, Chair
SUBJECT: Runway Use Estimates/ARTS Data
The Navigation Committee has met for an initial reading of your runway use memo of August 27, and is continuing to study it for a follow-on meeting on September 14. Meanwhile, here's some early feedback.
The "assumed use" chart for modeling forecasts zero operations on Runway 11. Is this a result of its absence from the list of runway use configurations in the memo, or assumed weather factors, or both? We note that a couple of operations were in fact observed on this runway in the actual data.
The actual data also indicate deviations from the preferred runway configuration during the day at nearly double the rates estimated for modeling (e.g. on 35R, arrivals @ 12.1% observed, vs 6.3 estimated; departures @ 17.8%, vs 9.8; on 35L, arrivals @ 11.3%, vs 6.2; departures @ 7.4%, vs 4.0). This pattern seems large and systematic enough to demand a close look at its two possible sources:
Design problems in compiling the estimates. A particular concern, for example, arises in the reduction of runway use factors other than average wind speed and direction to a naive decrement in the decision thresholds for tailwinds and crosswinds. Is there a logical basis for this substitution? What's the model's sensitivity to the specific values proposed? More to the point, are there no data available, through airport, government agency, or academic archives, on average incidence of wind shear, precipitation, runway maintenance requirements and the like, so that these factors could be represented directly for modeling purposes?
Anomalies in the observed data. The committee had anticipated in this memo a further elaboration of the specific conditions prevailing in the flight track data presented at the last Study Group meeting, and summarized in the "actual use" table. These data may have been skewed by unusual weather, maintenance, or other factors that would help justify using significantly different estimated values for modeling. We therefore remain hopeful of learning more about the ARTS sample.
By the same token, we hope to gain additional insights on the sub-sample of observations portrayed in the ARTS plots presented to the Study Group. The memo suggests, for example, that aircraft approaching visually would not appear on these plots; but there are many tracks far enough outside the prescribed pattern to be difficult to explain otherwise. We should seek clarification, however, on the memo's treatment of this point, in the third bullet under "ARTS Analysis Results", which we found very difficult to interpret.
More generally, operational questions persist about the behavior of the sub-sample itself. On departures to the north, for example, if all flights on 35L are supposed to diverge left, why don't they? And if all on 35R should use runway heading, why don't they? Arriving flights from the north should be converging on runway heading by the time they reach the river, we think, but in the daytime plot look more like Eastern Parkway. The actual use table gives nighttime arrivals from the north less than 5% frequency on runway 17R, but on the plot it looks more like 50%. And at the most general level, how is it that such a large portion of arriving flights enter the area from the south, but almost none of the departing flights go that way?
There's still a great deal still to be learned, in short, from what's been presented so far, some whimsical, some fundamental to further analysis. We aim to continue a careful study of the material provided, and look forward to our ongoing dialogue.