ACARS Coverage - May 2001

A study in support of the TAMDAR project by
Brian D. Jamison and Bill Moninger
Forecast Systems Laboratory.

As a part of the TAMDAR (Tropospheric Airborne Meteorological DAta Reporting) project, sponsored by FAA and NASA, FSL is charged with assessing places and times where data from the current aircraft meteorological data system is missing or in short supply, and where TAMDAR data can make a significant contribution.

Summary

To perform this study, we looked in detail at world-wide and CONUS coverage of aircraft data. We stratified the data by day of week, time of day, and altitude. Our conclusions are the folliwing.

Current ACARS data are highly temporally variable; evening peaks and early morning valleys of commercial air traffic cause the total number of reports per hour to vary by more than a factor of 4.
Package carriers (UPS and FedEx) generally do not fly on weekends, and this substantially decreases the number of observations.
The bulk of near-surface data is provided by package carriers (because they use a data format (ARINC-620) that provides high-vertical-resolution data near takeoff).
Most data are gathered at flight levels (25 Kft - 45 Kft).
Below about 25,000 ft, data are concentrated near major hubs.

In order to achieve a reasonable spatial distribution of data below 25 Kft, data from regional airlines and smaller package carrier aircraft must be included. The TAMDAR program, with its emphasis on low-cost sensors and communications, and with its focus on building a buisness case for airline participation is needed.

Details

We looked at data for each hour between Sunday 13 May 2001 and Saturday 2 June 2001. Weekend decreases in data (when package carriers are not flying) are clearly evident, as are large hourly fluctuations.

This shows the mean number of observations by hour for the entire data set. Observations peak at 3UTC (9pm Mountain Daylight Time), and have a minimum at 6 UTC (Midnight MDT). The peak at 8-9 UTC (2-3 MDT) reflects package carrier activity. This drops off by 4 MDT and is followed by a gradual increase until the morning business flyer peak at 8 MDT. The number gradually increases toward the evening hours.

Because of the valley at 6 UTC, we define an experimental "day" as running from 6 UTC until 5:59 UTC the next day, which corresponds to 0 - 23:59 MDT.

This shows total observations by experimental "day," starting at 6 UTC 13 May 2001 and running for 3 weeks, revealing the major effect of data from the package carriers.

For more detailed analysis, Sunday 13 May and Wednesday 23 May were chosen as "typical days." The figures below refer to these days.

This shows altitude coverage for a typical "Sunday." A clear peak in data between 30,000 Kft and 35,000 Kft may be seen. This should be contracted with similar data from our typical weekday, below, which shows much greater coverage near the surface, due to the low-altitude, high-resolution data from the package carriers.

The spatial distribution varies between low and high altitude, as some plan views will show.

This shows the very localized coverage between 20 Kft and 25 Kft. Data below 20 Kft are generally within the "rings" of data around each airport. Only the NE seaboard shows anything approacing uniform coverage.

This may be compared with the rather uniform coverage above 35 Kft:

The entire series of plan views for "Sunday" are here:

A similar series of plan views for our typical "Wednesday" are here:

Within each altitude band, we also broke the data down by 6 h blocks, starting at 6 UTC. These data may be found in this directory. The naming scheme for files in the directory is "MayDD_lNtM.gif", where DD is the day of the month, N is the level number (1 = below 5,001 ft, 9 = above 40,000 ft), and M is the quarter-day number (1 = 0-6 MST, 4 = 18-2359 MST). In addition, files named "MayDD_lvlN.gif" show data for level number N for a 24 h period.