1 Aircraft Data: Geographic Distribution, Acquisition, Quality Control, and Availability Work at NOAA/ESRL/GSD and elsewhere.

1

Aircraft Data: Geographic Distribution, Acquisition, Quality

Control, and Availability

Work at NOAA/ESRL/GSD and elsewhere

2

AMDAR data -- definition

• AMDAR - Aircraft Meteorological DAta Relay

• The generic term for all automated meteorological reports from commercial aircraft

• GSD receives (from a variety of sources) and distributes (with restrictions) all AMDAR data

• Most AMDAR data are used in GSD’s, and NCEP’s weather models

3

Flavors of AMDAR data

• MDCRS data (to be discussed by Al)• E-AMDAR (Europe and Africa)• “AMDAR” (Asia and Australia)• Canadian (Canada, inc. flights into U.S.)• TAMDAR (U. S. experimental)• Other experimental

4

MDCRS data (24 hours)

5

E-AMDAR data

6

“AMDAR” data (China, Japan, Australia, New Zealand, So. Africa)

7

Canadian data (jets and turboprops)

8

TAMDAR data (turboprops, generally below 20 Kft)

9

Experimental Turbulence (EDR) data, from UAL (above 20Kft) and TAMDAR (problematic)

10

Experimental Icing data from TAMDAR and (sometimes) Delta

11

Availability

• Real-time data are restricted to governments, researchers, etc.*

• Graphical displays are available at– http://amdar.noaa.gov/

• Binary and text data are available from MADIS– http://madis.noaa.gov/– data > 48 h old are unrestricted– archival data back to July 2001– many tools for subsectioning data

• GSD is a research, not operational organization– nonetheless, data flows have been pretty steady

12

*Data Restrictions

• Described at http://amdar.noaa.gov/FAQ.html

• Worked out over 15 years between NOAA and the participating US airlines

• Generally, non-commercial research users may have access to real-time data, if their work is of potential benefit to airlines

13

AMDAR QC in the U.S.

• NCEP (Brad Ballish) is the official U.S. data quality center– http://www.nco.ncep.noaa.gov/pmb/qap/

• GSD provides many research QC products based on the Rapid Update Cycle (RUC)

• Another U.S. center doing excellent work: Pat Pauley at the Naval Research Laboratory

• Many non-U.S. centers also perform AMDAR QC

14

AMDAR QC at GSD

• We look at differences between AMDAR observations and 1h forecasts from various versions of the RUC model

15

The Rapid Update Cycle

• An hourly cycle• Ingests all available data each hour

– AMDAR, TAMDAR, RAOBS (0 & 12 Z), Satellite, Surface, etc., etc.

• Covers the CONUS• We run many RUC cycles• For AMDAR QC, we use a 20 km cycle

called “dev2”

16

RUC - AMDAR Comparisons

• Statistical results are available at http://amdar.noaa.gov/ruc_acars/

and include:– Time series for individual aircraft– 3 and 7 day interactive statistics– plan view of data and corresponding RUC

values– (summary statistics are open; real-time data

are restricted)

17

Plan view showing Wind-

rejects for TAMDAR.

Reject reason indicated.

Overall statistics shown

http://amdar.noaa.gov/ruc_acars/

18

Right column shows the RUC reject code for any

rejected variable(s)

“sp=0”: wind speeds of 0 are rejected.

This information is downloaded automatically by some users.

19

7-day summary statistics (sortable). New columns show the fraction of obs rejected by the RUC, for each variable.

We reject many TAMDAR winds because we don’t accept descent winds currently

20

Current AMDAR Data Quality

• Generally as good or better than RAOBs• To check, we compare AMDAR and RAOBs with

RUC 1h forecasts• (RUC is not the “truth”, but, when a long time

period is aggregated, can serve as a common denominator)

• Results to follow are from– 1-31 Dec. 2006– AMDAR data from 0 UTC +/- 30 minutes

• 200 - 800 observations per pressure level• Canadian jets and all turboprops excluded

– In the Great Lakes Region• 13 RAOBs at 0 UTC

21

AMDAR - dev2 1h fcst Temperature RMS difference for obs taken near 0 UTC

Red shows AMDAR ascents and en-route

Blue shows AMDAR descents

22

Red shows AMDAR ascents and en-route

Blue shows AMDAR descents

Green shows RAOB

AMDAR - dev2 1h fcst Temperature RMS difference for obs taken near 0 UTC

AMDAR and RAOBs show very similar statistics

23

T bias is small (< 0.5°C) for AMDAR.

(Generally AMDAR ascents are warmer than descents, but not at low levels during this winter month.)

Red shows AMDAR ascents and en-route

Blue shows AMDAR descents

24

T bias is small (< 0.5°C) for AMDAR and RAOBs.

RUC shows a slight warm bias wrt RAOBs.

Red shows AMDAR ascents and en-route

Blue shows AMDAR descents

Green shows RAOB

25

For RMS of the vector wind difference with the RUC…

Red shows AMDAR ascents and en-route

Blue shows AMDAR descents

26

For RMS of the vector wind difference with the RUC, AMDAR and RAOB show similar values at all levels.

Red shows AMDAR ascents and en-route

Blue shows AMDAR descents

Green shows RAOB

27

Water Vapor

• Most AMDAR measurements don’t include vapor information

• WVSS-II measures it, but recent data is problematic

• TAMDAR measures it– In lower troposphere only (below 500 mb,

18 Kft)– In U. S. Midwest only– We’ll look at recent TAMDAR results

28

Relative Humidity RMS (wrt RUC dev2) for data taken near 0 UTC

Red shows aircraft ascents and en-route

Blue shows aircraft descents

TAMDAR (open circles)

29

TAMDAR (open circles)

RAOBs (green)

TAMDAR RH differences with RUC are generally smaller than RAOB RH differences.

An encouraging result.

Red shows aircraft ascents and en-route

Blue shows aircraft descents

Relative Humidity RMS (wrt RUC dev2) for data taken near 0 UTC

30

Summary

• AMDAR is an extensive, accurate, and timely data source

• We thank the airlines for providing the data• Tools at NCEP and GSD have been very helpful

in evaluating– Aircraft performance– Model behavior

• Feedback to and from airlines, researchers, and other nations has been invaluable in improving quality.