Common Operating Picture Example of Forecasts for DFW

1
NAS Weather Index (WITI) vs. Combined WITI-FA and
Delta (forecast goodness)
30-day period ending 06/17/2009
Positive delta Over-forecast of traffic impact
NWX 100 is a normally-impacted day
Negative delta Under-forecast of traffic impact
Delta /-50 may indicate a forecast issue
1
7/17/2009
2
En-route and Terminal WITI-FA30-Day Period
Ending 06/17/2009
En route Convection E-WITI (NCWD) vs. E-WITI-FA (
CCFP)
Positive Delta Over-forecast
Negative Delta Under-forecast
Terminal Weather T-WITI (METARs) vs. T-WITI-FA (T
AFs)
Positive Delta Over-forecast
Negative Delta Under-forecast
2
7/17/2009
3
NAS Wx Index Breakdown by Component Last 7
Days, Ending 06/17/2009
See analysis on Slides 6-13
7/17/2009
3
4
TWITI, 7-Day Period Ending 06/17/2009, by
NWS Region TWITI shows potential operational
impact of IMC, Wind, Winter precipitation, and
Local convective Wx
Terminal Weather T-WITI (METARs) vs. T-WITI-FA (T
AFs)
Positive Deltas (bars) over-forecast
Negative Deltas (bars) under-forecast
4
5
Analysis of Selected Airport / Days
5
7/17/2009
6
Analysis for selected airports/days
Central Region, ORD, June 11, 2009
Under-forecast for most of the day Low ceilings,
light rain/mist, light wind most of the day
forecast called for unlimited ceilings, light
winds, and showers-in-vicinity Also, convective
under-forecast for early morning (see next slide)
7
Analysis for selected airports/days
Convective Forecast Highlights, June 11, 2009
4-hr CCFP shows much less convective impact
affecting ORD and Midwest than actual Wx
8
Analysis for selected airports/days
Eastern Region, June 11, 2009
June 11 was the worst day of the year so far in
terms of operational impact of weather
9
Analysis for selected airports/days
Eastern Region, EWR/LGA, June 11, 2009
Some under-forecast at EWR, LGA, JFK mostly in
the evening low ceilings prevalent
Terminal weather forecast doesnt seem to have
had an impact on airport delays and cancellations
10
Analysis for selected airports/days
Eastern Region, PHL/JFK, June 11, 2009
Some under-forecast for PHL, mostly in the
evening for JFK, morning and midday
Terminal weather forecast doesnt seem to have
had an impact on airport delays and cancellations
11
Analysis for selected airports/days
Southern Region, ATL, June 12, 2009
Some under-forecast stronger winds actual than
forecast Little or no impact of terminal forecast
on ATL operations, except for possibly late
evening
12
Analysis for selected airports/days
Eastern Region, LGA, June 16, 2009
Low ceilings (e.g., 2100 Ft) actual vs. 5000 to
unlimited ceilings forecast Possible impact on
LGA operation
13
Analysis for selected airports/days
Convective Forecast Highlights, June 16, 2009
4-hr CCFP for 0000Z (evening) shows much less
convective impact affecting Midwest and South
than actual Wx
14
AvMet Applications Website
For more detailed drill-down and analysis,
please go to www.avmet.com/CWITI

14
7/17/2009
15
NWX / WITI / WITI-FA Components

• WITI model consists of two principal components
  En-route (E-WITI) and Terminal (T-WITI). The
  NAS-wide WITI metric is called NAS Wx Index
  (NWX).
• En-route WITI (E-WITI) reflects the impact of
  en-route convective weather and en-route traffic
  demand (flows between OEP34 airports) on the
  NAS.
• Terminal WITI (T-WITI) reflects the impact of
  local airport weather and local traffic demand on
  the airports operation
• Airport capacity can decrease due to inclement
  weather (low ceilings, rain, snow, wind etc).
  Arrival and departure rates may be reduced,
  resulting in delays and/or cancellations.
• If scheduled traffic demand exceeds airport
  capacity (be it in good or bad weather), queuing
  delays ensue. These delays can quickly grow
  exponential in some cases, wide-spread
  cancellations are the only way to limit
  non-linear growth of delays.
• T-WITI reflects both the linear increase in
  delays (some impact of inclement weather but
  airports capacity remains higher than traffic
  demand) and, in more severe cases, non-linear
  increase in delays (impact of weather and/or
  traffic demand grows exponentially when demand
  exceeds airports capacity)
• NWX / WITI is computed using actual (recorded)
  weather. WITI-FA (Forecast Accuracy) is
  computed using forecast weather, both en-route
  convective and terminal.

16
NAS Wx Index Breakdown by Cause Explanation
to Slide 3

• NAS Wx Index / WITI software can distinguish the
  following factors
• En-route convective weather. This shows
  convective weather impact on an airports
  inbound/outbound flows within approx. 500-NM
  range. This component does not affect queuing
  delay at the airport.
• Local convective weather. This reflects how
  convective weather in the vicinity (directly over the airport reduces airports
  capacity. It may affect queuing delay.
• Wind. Any time there is a wind greater than 20
  Kt, or there is precipitation and wind greater
  than 15 Kt, the corresponding impact is recorded.
  Airport capacity may decrease, i.e. queuing
  delays may increase.
• Snow, freezing rain, ice etc. The corresponding
  impact is recorded. Airport capacity may
  decrease, i.e. queuing delays may increase.
• IMC. Ceiling or visibility below airport specific
  minima fog and heavy rain. The corresponding
  FAA capacity benchmarks for IMC are used. Queuing
  delays may increase.
• Queuing Delay (No Weather) plus Ripple Effects.
  No particular weather factor recorded locally for
  the given airport / given hour but WITI software
  computed that there would be queuing delays. This
  can be simply due to high traffic demand or in an
  aftermath of a major weather event when queuing
  delays linger on (even as the weather has moved
  out).
• Additionally, Ripple Effects are recorded in this
  component. For example, if ORD experiences
  departure queuing delays, its corresponding
  destination airports will get some additional
  arrival queuing delay.
• Other. Includes minor impacts due to
  light/moderate rain or drizzle but
  ceilings/visibility above VFR minima also
  unfavorable RWY configuration usually due to
  light-to-moderate winds (15-20 Kt or even 10 Kt)
  that prevent optimum-capacity runway
  configurations from being used.

Convective
Non-convective
Other
17
Rolling 4-hr Look-ahead Forecast

• 4-hr TAF is mentioned throughout this slide
  set.
• In actuality, Terminal Area Forecasts (TAFs) are
  issued every 6 hours, with amendments issued at
  irregular time intervals if/as necessary.
• From this TAF stream, the WITI software
  constructs a rolling 4-hr look-ahead forecast.
• If, for instance, it is 1300Z and an operator at
  airport NNN would like to know the expected
  weather situation at 1700Z, what is the TAF
  information available to him/her at 1300Z? It
  could be the standard 1200Z TAF valid through
  1800Z) with perhaps an amendment issued at 1300Z.
  An hour later, at 1400Z, if the operator needs to
  know the forecast for 1800Z, he or she might
  still have the same information as at 1300Z but
  perhaps a new amendment has been issued, and so
  on.
• Rolling 2-hr, 4-hr and 6-hr CCFP (convective
  forecast) is interpreted in a similar fashion.
  There are no amendments as in TAF. CCFP is issued
  every 2 hours at odd hours (1300Z, 1500Z, ) as a
  set of three forecasts. A CCFP forecast for even
  hours is an interpolation of these 2-hr CCFPs.

18
Arrival Rate Charts (Analysis)
Scheduled and actual arrival rates (solid purple
and dashed blue lines on the above sample chart)
are extracted directly from ASPM data. METAR and
Rolling-4hr-lookahead-TAF based rates (red and
yellow lines) are WITI model estimates based on
historical data and FAA airport capacity
benchmarks.

• Things to keep in mind
• WITI model estimated rates show potential airport
  capacity given the perceived or expected weather
  impact.
• Direct comparison between WITI model-estimated
  and actual arrival rates should be made with
  caution the WITI model does not reflect all the
  factors, events and human decisions that are
  behind a specific actual arrival rate. Comparison
  with facility-called rates can help to understand
  these effects.
• Recorded (actual or forecast) weather data is
  discrete for example, wind is recorded in hourly
  intervals and its direction can vary, affecting
  what WITI model selects as the optimal runway
  configuration. Or, snow can start and stop. But
  actual impact of weather can be longer-lasting
  (e.g. snow removal) and an airport cannot react
  to wind changes by changing runway configuration
  in an abrupt manner. The result may be a larger
  variability in WITI model-forecast rates vs.
  actual arrival rates.
• Non-weather factors, as well as weather in other
  parts of the NAS, may impact airport capacity on
  a particular day this is not reflected in WITI
  model-based arrival rates (they are based only on
  local weather).
• Suggested uses for the arrival-rates charts
• Significant differences between METAR- and
  TAF-based arrival rates may be an indication of
  an over- or under-forecast of terminal weather
• In some cases, these significant differences may
  be coupled with actual arrival rates being
  noticeably lower than scheduled. This, in turn,
  may in some instances indicate an impact of an
  inaccurate weather forecast.

19
Snow/Ice Impact Quantification
Moderate snow/ice may in some instances cause
higher impact on airports (delays) than indicated
by NWX/WITI. The reason is that even as the
snowfall stops and winter weather moves out, snow
and ice removal may take a long time this is not
reflected in METAR/TAF data and hence the
NWX/WITI may be lower. Also, it takes time for
airlines to restore their schedules back to
normal, which again leads to higher delays
compared to perceived weather impact. Conversely
, on days with very heavy impact of winter
weather, NWX can be much higher than the
normalized Delay. This is due to massive
cancellations that lower traffic demand. However,
in these cases NWX correctly reflects the overall
weather impact on the NAS. Typically, on the
next day, when the winter weather moves out, NAS
Delay metric is significantly higher than NWX (as
airlines work to restore schedules back to
normal).