Innovative Lightning Detection System by The New Met Office

1. The new Met Office ATDNET lightning detection system J. Nash, N.C. Atkinson, E. Hibbett, G. Callaghan, P. L.Taylor, M.Turp, P. Odams, D. Jenkins, S. Keogh, C. Gaffard, E. Walker Met Office, Exeter/Reading TECO-2006, 4-6 December 2006, Geneva

2. ATDNet/ATD • The flash location process in ATDNET is similar to the existing ATD in principle, but ATDNET uses updated technology throughout. • The system is designed and operated entirely by Met Office staff and does not rely on support from any other commercial lightning detection manufacturer. • ATDNET is designed to run autonomously under the supervision of a network manager, with much less intervention expected from Met Office network and Engineering than with the current ATD • Data acquisition is PC based • Communications will be transitioned to Internet systems • ATDNET will have more outstations than the current ATD to cope with situations where severe local storms reduce the sensitivity of some outstations • In the immediate future ,the Met Office is committed to extending the main service area from Europe to the whole of Africa and Arabia and the western Indian Ocean

3. ATD Sferics Lightning Location Whip antenna sensitive to vertical component of electromagnetic field from lightning discharge narrow band observations of lightning emission centred at 10 kHz , but new system will probably be changed to 13.6 kHz, because of problems at 10 kHz in the Indian ocean Frequency can be selected using software from the operations centre

4. Propagation of electromagnetic waves at 10-14 kHz • The strong emissions from lightning at these frequencies are caused by rapid neutralisation of charge in the lowest few hundred meters of cloud-to ground strokes. • Atmospheric attenuation at 10 to 14 kHz is very low and the electromagnetic discharge (SFERIC) can propagate over thousands of kilometres along the earth-atmosphere waveguide. • Propagation is best over the sea or moist surfaces, and differs between day and night because of the different height of the reflecting layers in the ionosphere between day and night. • ATDNET outstations Fourier analyse the SFERIC wave. • The waves from the different outstations are correlated in the central Flash Location Processor to obtain time differences between the stations. • A minimum of 3 pairs of time differences are required for a location, but it is preferable to have at least 4 if possible to guard against error in an individual time difference [ possibly 110 µs]

5. New outstation allows much higher throughput than old outstations, essential to meet user requirements in European summer

6. Arrival Time Differencefixing process • Accurate time stamping , 0.1s - rubidium oscillators, checked by GPS every second. • Waveforms are Fourier analysed and sent to the central control station within 30s (will use cheap internet based communication, transition for all outstations should be achieved during 2007 ) • Waveforms from different outstations are correlated to estimate time differences • Arrival Time Differences are then used to calculate lightning position by iterative method • Distribution of data messages every five minutes, but potential to extract messages within 2 minutes for new products. • Speed of data delivery very much higher than with current system.

7. Expected disposition of out-stations in Europe

8. Purple lines try to implement Before mid- 2007 Outstation in Reunion, observes at 13.6 kHz, VPN communications established and working well for several months, after initial problems

9. Examples from recent testing of ATDNET

10. Existing ATD depends on NOS since October 2004

11. Comparison of number of flashes reported by ATD and ATDNET ATDNET much better than ATD over Europe in summer and also at long range

12. Extent of long range service area

13. ATDNET Climatology for September 2006

14. ATD Climatology for September 2006

15. Completion of operational link from flash location processor to logical data store and then to product generation system

18. Operational computers now all configured , and NOS feeding directly to Flash Location processor

19. Check of ATDNET output in South America

21. Rindat lightning detection system, southern Brazil

22. Limitations in Africa requiring better communications to Mediterranean sensors and new sensors located south of the Sahara

23. 11-12 UTC 731 locations

24. MSG + old ATD,12 UTC for preceding 6 hours

25. Current ATDNET geometry not good for flashes in this area 11-12 UTC 1712 poor locations

26. Superimposition of lightning on MSG cloud pictures

27. For the future • Various countries have indicated willingness to collaborate further in extending ATDNET towards a global system • Met Office is currently setting up the infrastructure to deal with the interested parties , so that suitable arrangements can be made • The extension of the system across Africa and Arabia will be primarily funded by the Met Office, with suitable memoranda of understanding agreed with countries hosting sensors. • Individual negotiations will be pursued with countries in other regions, as policy for expansion is agreed with Met Office board. • The main limitation on speed of expanding the network is setting up the communications. Here the main difficulties have now been overcome, but the complexity of dealing with various network operators often limits the rate of expected progress.

28. Are you interested in collaboration? • Please contact: • J. Nash • C. Gaffard • T. Oakley

29. 70% 60% 50%(good/ total)

30. future work • compare efficiency/ accuracy with meteo france • investigate why poor ratio good/total fixes • introduce a wave propagation model • evaluate more accurately the accuracy • simulate different design.

32. comparison ATD EA technology ATD more sensitive than EA during day time

33. Questions & Answers