Within the Avian Alert System of Systems (SoS) initiative of the European Space Agency (ESA), we have explored the potential of operational C-band Doppler weather radar as a bird migration sensor. A bird migration recognition algorithm has been developed, extracting bird density, speed and direction as a function of altitude. The weather radar data have been validated against simultaneous and co-located bird density measurements by a high precision bird radar, provided by the Swiss Ornithological Institute (SOI). This mobile tracking radar has been stationed next to weather radar sites in the Netherlands, Belgium and France during the peak bird migration season in autumn 2007 and spring 2008. The mobile tracking radar is capable of detecting and discriminating bird echoes with a high accuracy, providing additional bird species information by analyzing wing beat frequencies observed in bird echoes, making it an ideal reference for validating the weather radar observations.
We find that Doppler weather radar is highly successful in determining quantitative bird densities as a function of altitude. The detection probability is very high (up to 99%) and the fraction of false alarms is low (down to 2%). We find that weather radar reflectivity can be quantitatively correlated to the bird-densities determined by bird radar. We converted weather radar reflectivity to bird densities using a fixed radar cross section of 10 cm2. For nocturnal migration under the assumption of broad front migration, 74% of the weather radar height specific bird densities are correct within a factor of 2 and 87% within a factor of 3. As the weather radar and bird radar are not surveying exactly the same area and have different time resolutions, the quantitative correspondence in bird density may even higher than these figures suggest.
The current bird detection algorithm meets the requirements for operational implementation. At the Royal Dutch Airforce (RNLAF), the subsequent levels of Bird Strike Warnings (so-called BIRDTAMs) differ in bird density by factors of 2, which is on the order of accuracy obtained by weather radar.
The current study shows that weather radar has a high potential for providing information on the spatial distribution of birds during migration. This is especially important in areas with prominent topographical features, like the Netherlands where large water bodies and coastal areas structure the spatial distribution of birds. Obtaining spatial information from weather radar would therefore be a useful future development.
Further research on the use dual polarization radar (which is rapidly becoming the new operational standard) for bird detection is highly recommended. This study suggests that the combined use of dual-polarization techniques for precipitation filtering and high quality Doppler data for insect filtering is most adequate for high quality bird migration quantification. A field study using a weather radar that combines these characteristics would be valuable to explore the full potential of operational weather radar for bird detection.
For upscaling of the operational bird detection by weather to a European scale, the Operational Program on Exchange of Radar data (Opera) running within Eumetnet, which is the network grouping of 26 European National Meteorological Services, offers an excellent opportunity. Opera’s operational network consists of more than 180 weather radars. An Opera Data Center is currently being developed and the start of operation is planned for January 2011. This data center will provide the kind of infrastructure needed to build a system for the detection of bird migration.
Adriaan M. Dokter, Felix Liechti and Iwan Holleman. Bird detection by operational weather radar
KNMI number: WR-09-06, Year: 2009, Pages: 201