ERAD ERAD 20062006Proceedings ofProceedings of

The primary ATC radar weather channel - ENIGMA III weatherextractor –Z, V, W moments and bird migration detection, conceptand first results with ASR-S

M.Malkomes1, Matthias Toussaint1, Dietmar Veerkamp1, Karl-Josef Müller2, Christoph Neumann2, Patrick Perez2,Michael Weber2

1 GAMIC mbH, Aachen (Germany).2 EADS Defence Electronics, Ulm, (Germany).

1 Introduction

Similar to classical weather radar (WR) data processing, a“weather channel” can be used to extract weather information from primary Air Traffic Control (ATC) radars.The latest generation of ATC radar designed by EADS is theASR-S (Airport Surveillance Radar - S-band). The GermanAir Forces together with EADS and GAMIC have tested andqualified the radar weather channel (extractor) for flightoperations over more than a 1 year period. The ASR-S radaruses a modified version of the standard GAMIC ENIGMAIII signal processor and FROG-MURAN software forweather real time processing and visualization. Additionallyto the weather extraction a new bird migration detectionalgorithm is implemented in this system.

2 The ATC primary radar weather extractor

In contrast to a conventional WR, an ASR- S is a latest radartechnology, a fully solid-state coherent radar using Pulse-Compression and Moving Target Detection associated with amultiple PRF Staggering techniques. Currently ASR-S is inoperation at German Air Force air fields. Up to 25 systemswill be deployed all over Germany within the next 3-4 years.Over the past years a number of approaches have been madeto analyze, upgrade or implement weather extractors onprimary ATC radars, Stern et al. (2003), Basile, M. (1992)

Also Bird detection has always been an important issue forflight operations and to find ways to use the rapid scanningATC radars to detect these. Bruder, J.A. (1997), Larkin, R.P.(1991)

The ATC radar concept handicaps the weather measurementvertical resolution due to a “Fan-beam” antenna, which has a rather wide beam in elevation. Nevertheless good “vertically

averaged estimates”of Z - reflectivity, V- velocity and W -turbulence moments can be achieved in particular in the nearrange of the scanned area up to 60nm or 100km. The Z, Vand W moments are processed to PPI type products andtransferred to the ATC controller workplace in ASTERIXCAT 8 vector format. Additionally to the weather dataextraction, specific bird migration detection algorithms havebeen developed and tested. Using on-line Doppler analysisand statistical algorithms, a distinction betweenprecipitation, clutter and biological targets has been achievedwith low false alarm rate. Update of weather and birdmigration detection information is in the order of 1 min,which is much faster than any conventional WR.Additionally to their role in the ATC environment insuringsafe flight operations near the airports, weather condition aswell as Bird migration, the ASR-S weather channel'sinformation may be also used as backup and gap fillers forthe weather service (- DWD, in Germany) weather radarnetwork.

3 The weather channel

The GAMIC weather channel is designed for a 60nm radarrange, having a range resolution <60m and the major parts ofthe system comprise:

Fig. 1. ASR-S weather channel block diagram

Correspondence to: Martin Malkomes, GAMIC.

malkomes@gamic.com

WSP; Weather Signal Processor (ENIGMA III)

AU: Adaptation unit to the ATC sensor

WDP; Weather data processor with SoftwareFrogRT on LINUX PC

ASTERIX CAT 8 data Interface to the ATCcontroller consoles

4 Weather data and bird migration data results

4.1 Typical weather data display examples

Reflectivity Z and V measurement (with clutter correction)R=100 = 60 nm km range

Fig. 2. Corrected Reflectivity Z, example

Fig. 3. Radial Velocity V, example

4.2 ASTERIX data for the ATC controller consoles

The Category 8 window is used for monitoring contouredtwo dimensional weather images. It is primarily used by theAir Traffic Controllers to have a visualization of the cloudsas underlay in the control area. Additionally the algorithmfor bird detection will generate a special layer contour forwarning in case birds or flocks of birds are detected.

The following screenshot shows a CAT 8 ATC weathercontour display which may be easily exported to ATCcontroller positions

Fig. 4. ASTERIX CAT 8 weather contour image.

4.3 Comparison with DWD weather radar data

The comparison with data recorded at approximately thesame time using German Weather Service DWD radar datashows the following results:

Fig. 5. DWD–PPI data (Z)

Fig. 6. ASR-S radar weather channel data (Z)

It is observed that the phenomena are slightly underestimated at longer range, which can be corrected. Generallyit shows that a good qualitative presentation of the weathersituation is achieved, particularly in the vicinity of the radarsite–which is most important for the flight operations.

4.4 Bird migration detection algorithms and implementation

The EADS European Aeronautic Defence and SpaceCompany in cooperation with GAMIC has developed a newapproach for the detection of bird migration. Someconstraints:

An ATC Primary Surveillance Radar does not haveenough Azimuth/Range/Doppler resolution whichallows a satisfying classification rate between aslow aircraft and a group of birds in real timebecause of a limited amount of dwell hits.

Bird migration is an activity spread over half a daywithout major changes.

Bird flock is similar to widespread discrete movingclutter contrary to an aircraft or a weather front.

A lot of Radar data were recorded and evaluated duringthe year 2005 and 2006. The campaigns ofmeasurements were done in collaboration with the GeoInfo. Off. Biology Sect. of the German Army in order toconfirm the presence of Birds.

Birds

17.10.2005

Ground Clutter

Birds

17.10.2005

Ground Clutter

Fig. 7. ASR-S radar weather channel data with bird data

A new algorithm was developed using image processingtogether with the pattern recognition. The pattern recognitionis a modern and necessary approach for decision making;Aircraft, weather front, or Bird migration. The relevantfeatures from the texture analysis are statically assessed inorder to distinguish information from clear weather, fromweathers front, from ground clutter and from birds.

The radar data flux is processed simultaneously with antennaazimuth rotation as received from WSP. A PPI-detectionmatrix (PDM) comprising 600 x 75 cells is generated. ThePDM first is stored on WDP for post processing.

Fig. 8. ASR-S radar weather channel data flow

Subsequently the bird detection based on the PDM usespattern recognition methods. Image processing algorithmslike texture analysis together with a Doppler analysis areapplied. The relevant features for rain, clutter and birdmigration are separated in the radar data and finally are usedas input data for a statistical classification system.

After optimisation the successive stages of the processingchain (Pattern optimisation, classification tree optimisationand threshold optimisation), a classification better than 80%(Fig.8) could be achieved.

0 10 20 30 40 50 60 70 80 90 1000

500

1000

1500Vogel/Normal: PDF aus QK

0.97 0.0310.029 0.97Anz:4170,25256

0 10 20 30 40 50 60 70 80 90 1000

0.2

0.4

0.6

0.8

1Vogel/Normal: CDF aus QK

0.97 0.0310.029 0.97DW: 0.941

Clearweather

0 10 20 30 40 50 60 70 80 90 1000

200

400

600

800Voegel/Gewitter: PDF aus QK

0.92 0.0780.07 0.93Anz:4170,15704

0 10 20 30 40 50 60 70 80 90 1000

0.2

0.4

0.6

0.8

1Voegel/Gewitter: CDF aus QK

0.92 0.0780.07 0.93DW : 0.849

Birds Birds

Weatherfront

0 10 20 30 40 50 60 70 80 90 1000

500

1000

1500Vogel/Normal: PDF aus QK

0.97 0.0310.029 0.97Anz:4170,25256

0 10 20 30 40 50 60 70 80 90 1000

0.2

0.4

0.6

0.8

1Vogel/Normal: CDF aus QK

0.97 0.0310.029 0.97DW: 0.941

Clearweather

0 10 20 30 40 50 60 70 80 90 1000

200

400

600

800Voegel/Gewitter: PDF aus QK

0.92 0.0780.07 0.93Anz:4170,15704

0 10 20 30 40 50 60 70 80 90 1000

0.2

0.4

0.6

0.8

1Voegel/Gewitter: CDF aus QK

0.92 0.0780.07 0.93DW : 0.849

Birds Birds

Weatherfront

Clearweather

0 10 20 30 40 50 60 70 80 90 1000

200

400

600

800Voegel/Gewitter: PDF aus QK

0.92 0.0780.07 0.93Anz:4170,15704

0 10 20 30 40 50 60 70 80 90 1000

0.2

0.4

0.6

0.8

1Voegel/Gewitter: CDF aus QK

0.92 0.0780.07 0.93DW : 0.849

Birds Birds

Weatherfront

Fig. 9. Classification plots–probability of detection of clutter, rain, birds

The first results from the Years 2005 and 2006 arepromising. It has been demonstrated that a betterclassification of target reports can assist the controller tomake better decisions without decreasing the sensitivity ofthe primary radar.

5 Conclusion

The ASR-S weather channel with GAMIC WSP-WDP istested to improve safety by warning controllers and pilots ofhazardous weather like strong rain, wind and bird migrationevents near runways. The FROG-MURAN can be used topredict the arrival and track weather event motion giving acomplete picture of current and future terminal areahazardous weather conditions which may impact runwayusage for take-off and landing. The weather channel willdetect hazardous weather and bird migration on the ASR-Simproving the capabilities security in air traffic control asneeded by mission sensitive flight operations.

The GAMIC WSP-WDP weather channel can easily be usedwith a variety of operational ATC radars (Magnetron,Klystron, TWT, Solid-State) to upgrade the weatherdetection capabilities to modern technology.

The possibility of using such ASR weather information as asupplement or merged with data from “civilian” weather radars operated by DWD is currently under consideration.

Acknowledgements:

GAMIC wishes to thank the many unmentioned experts at EADS,German Air Force and DWD for the fruitful discussions andcooperation during the project.

References

Basile, M. et al., 1992, An integrated weather channel designed foran up-to-date ATC radar system, Radar 92. InternationalConference

Bruder, J.A.et.al., 1997, Bird hazard detection with airportsurveillance radar, Radar 97, Publication No. 449, p. 160 -163

Larkin, Ronald P., 1991, Sensitivity of NEXRAD algorithms toechoes from birds and insects, 25.AMS, 1991, p.203-205

Stern, A. D. et al. 2003, Exploring the use of FAA radar weatherdata by the NWS: Comparison and observations between the

WSR-88D and FAA weather radars, 19th

Conf. on InteractiveInformation and Processing Systems, Long Beach, CA, Amer.Meteor. Soc., Paper P1.32