Winds Study to Support RTCA SC 206 Subgroup 7 Development of Wind Guidance Document for ATM Principal Investigators – Michael McPartland & Tom Reynolds Presented at 2016 Summer FPAW Eldridge Frazier, FAA Distribution Statement A. Approved for public release: distribution unlimited

NextGen Programs of Interest Wind and temperature forecasts can have a significant affect on aircraft trajectory estimation Ground systems Airborne systems  (WTIC) RTCA SC-206 “Aeronautical Information and Meteorological Data Link Services” Required Time of Arrival (RTA) Wake Vortex Mitigation Interval Management (IM)

Required Time of Arrival Overview Other aircraft forming traffic stream Actual Winds Frequency Time Error 95% time compliance at meter fix Time target and Wind updates Time Target Meter Fix FMS Wind Forecast ATC Systems ATC System Forecast Need for research to inform Minimum Weather Service & 4D-TBO guidance documents

RTCA RTA Performance Research Questions Determine affect of: Forecast source GFS HRRR Perfect forecast (truth) Forecast age: Published at least 2 hours prior to in-air update Number of FMS descent forecasts levels (DFLs) 4 9 How descent forecasts are selected Optimized to match wind magnitude Optimized to match trajectory headwind Equidistant spacing (from cruise to surface) Most FMS system can only accept 3-5 DFLs. Some airlines select decent forecasts over the estimated top of descent or over the destination. Many airlines select decent forecasts at fixed altitudes. Some airlines select decent forecasts along the estimated descent trajectory. We have interesting results for optimized selection of DFLs but in this presentation we will only be showing data where DFLs were selected, along the estimated decent trajectory and at equispaced altitudes from the cruise altitude to the destination altitude.

Introduction to GFS & HRRR GFS (Global Forecast System) 28 km grid (0.5º averaged) 26 pressure levels Published 4 times per day: 00Z, 06Z, 12Z, 18Z Forecast: +03, +06, +12,…+192hrs Global coverage HRRR (High-Resolution Rapid Refresh) 3 km grid 50 pressure levels Published every hour Forecasts: +01, +02, +03,…+18hrs CONUS GFS is probably the most widely used for airlines flight planning systems. Two down-sides of HRRR is that it only covers CONUS (with some surrounding area as well) and that because the resolution is so high, it takes time to download all the forecast data.

Analysis Methodology CONOPS Identify MDCRS flights that stayed on route Use aircraft-measured winds as simulated winds Reproduce flights with simulated B757 and advanced FMS CONOPS Wx Updated Provided 2-hr old Wx updates 10 minutes prior to RTA freeze horizon Assign RTA time and fix 230 NM from destination Descents to ~ 10-15kft RTA Assigned Using ASDE-X, MDCRS and TFMS data, we have established a process that identifies qualified flights which is only a small subset of all flights. Must have valid MDRCS data, must stay on assigned route without flight plan amendments RTA fixes are selected from STARs where the A/C is expected to cross near 10,000 feet For the reproduction of each flight: the A/C are started at roughly mid-cruise Prior to reaching the scheduling freeze horizon (about 10 minutes beforehand) the A/C updates there forecast, which is at least 2-hrs old and may not have changed since departure When 230 miles out, ATC assigns a fix and the aircraft is given a control time that is within their earliest and latest time window for the named fix. Simulation tracks

Aggregated RTA Performance Airports evaluated KATL, KBOS, KDEN, KEWR, KMDW, KORD, KPHX 340 flights Feb 1 – Mar 31, 2016 GFS & HRRR based on 2 hr forecast Truth based on MDCRS ±2 × std (24.5% if no forecast used) RTA Error (secs)

Effect of Speed Constraints Speed constraints on STARs Reduce speed control authority Thus reduced RTA performance FMS honors speed constraints even with RTA operations (SC-214) With speed constraints Without speed constraints 299 with speed constraints, 41 flights did not (KATL, KEWR) RTA Error (secs) RTA Error (secs)

Wake Vortex Mitigation Time sequence of crosswinds over CSPR* FAA looking to wind dependent strategies to increase throughput Use wind forecast system to predict “wake safe” regions Wake Terminal Mitigation System has access to high fidelity wind observations near the ground (ASOS) Numerical Weather Forecast Model predictions (for above ground) WTM System correctly handles forecast model error periods (see right) but at a cost of availability of increased throughput Example wind shift removes availability ~45 minutes System unavailable for hours LIDAR Aloft observations, such as real-time aircraft winds would address this problem RAP ASOS WTM System (Middle of event shown above to illustrate maximum wind shift) *Closely Spaced Parallel Runway (CSPR)

Interval Management Concept NASA/TM–2016-219203 Ahmed, Barmore & Swieringa ATC provides IM clearance to the aircraft near top of descent (left). Pilots follow onboard speed guidance to achieve precise spacing interval at the achieve-by point , Δ behind the target aircraft (right). Δ may be a time or a distance

Conclusions Wind and temperature forecasts can have a significant affect on aircraft trajectory estimation Supporting various RTCA activities Key RTA analysis findings 9 DFLs better than 4, but not significantly for cases examined Performance of GFS nearly as good as HRRR Performance with 2-hr HRRR forecast data nearly as good as “truth” Procedural speed constraints have major impact on RTA performance Next Steps NAS-wide comparison of flights with and without speed constraints Conduct RTA flights down to Initial Approach Fix Analyze impacts of using aircraft-derived winds Modify Mode-S EHS interrogator Generate methods to provide confidence of wind forecasts The use of 9 DFLs is in general better than use of just 4 but not significantly so for descents down to 10,000 feet. In a third of the cases, 9 DFLs actually performed measurably worse, though not by much, than using just 4 DFLs. Perhaps too much trust in bad forecast data? Different wind blending algorithms, which we do have the ability to test may address this issue. We suspect to see a greater separation in performance with 9 versus 4 DFLs as we conduct experiments to lower (2-4,000 ft AGL) altitudes. The hands-down most important finding, and is no surprise mathematically, is that the application of speed constraints on STARs significantly reduced the RTA performance. Respecting speed constraints and speed limits is currently a requirement when conducting RTAs unless otherwise dictated by ATC. When flying STARs that do not have speed constraints, the desired RTA performance requirements are met, even with GFS forecasts. MIT/LL ASR-9 radar

Backups

GFS & HRRR Timelines GFS HRRR +0 +3 +6 +9 +12 +15 +18 00Z +0 +3 +6 +9 Publications times 06Z +0 +3 +6 12Z Forecast lookaheads etc. HRRR +0 +1 +2 +3 +4 +5 +6 +7 +8 +9 +10 +11 +12 +13 +14 +15 +16 +17 +18 00Z 01Z 02Z etc.

Wx Forecast Performance

This material is based upon work supported by the Federal Aviation Administration under Air Force Contract No. FA8721-05-C-0002 and/or FA8702-15- D-0001. Any opinions, findings, conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the Federal Aviation Administration. © 2016 Massachusetts Institute of Technology. Delivered to the U.S. Government with Unlimited Rights, as defined in DFARS Part 252.227-7013 or 7014 (Feb 2014). Notwithstanding any copyright notice, U.S. Government rights in this work are defined by DFARS 252.227-7013 or DFARS 252.227-7014 as detailed above. Use of this work other than as specifically authorized by the U.S. Government may violate any copyrights that exist in this work.