1. 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

2. 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)

3. 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

4. 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.

5. 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.

6. 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

7. 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)

8. 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)

9. 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)

10. Interval Management Concept
NASA/TM–
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

11. 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

12. Backups

13. 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.

14. Wx Forecast Performance

15. This material is based upon work supported by the Federal Aviation Administration under Air Force Contract No. FA C-0002 and/or FA D 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 or 7014 (Feb 2014). Notwithstanding any copyright notice, U.S. Government rights in this work are defined by DFARS or DFARS 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.