Near real-time input to an HF propagation model for nowcasting of HF communications with aircraft on polar routes

Transcription

1 Near real-time input to an HF propagation model for nowcasting of HF communications with aircraft on polar routes E.M. Warrington, A.J. Stocker, D.R. Siddle, J. Hallam N.Y. Zaalov F. Honary, N. Rogers D.H. Boteler, and D.W. Danskin 1

2 Airlines are increasingly using trans-polar routes as these provide more direct travel between some destinations. 2

3 Map of HF stations 6 o N 75 o N 15 Jan, 23UT, Kp=3 12 o W 45 o N Alert Qaanaaq Ny Alesund Tromso Nurmijarvi 4 o E Ottawa Leicester 3 o N 8 o W 4 o W o 3

4 Example measurements in the polar cap , 8.1 MHz Azimuth (deg from N) MHz Azimuth (deg from N) Time (UT) 4

5 6 o W Example simulation output 12 o E 12 o E 6 o E S9+6 6 o E S9+6 S9+4 S o W 6 o N 75 o N 18 o W 6 o N 75 o N S9+2 S9+2 A A 45 o N Q o S9 S8 S7 45 o N Q o S9 S8 S7 S6 S6 S5 S5 S4 S4 3 o N 12 o W S3 S2 S1 3 o N 12 o W S3 S2 S1 6 o W UT 8. MHz UT 8. MHz 5

6 Manually intensive Current state of the model The background ionosphere derived from ionosonde measurements (from simulation time ±12 hours) Intensity, number, and initial location of patches are randomised based on estimates of these parameters from the literature (and ensemble statistics generated) D-region absorption model based on DRAP values 6

7 Assimilation of real-time inputs to the model 1 Background ionosphere o Structure based on IRI o Relatively few ionosondes in polar cap and can be difficult to interpret and derive values 7

8 Example ionograms (Tromsø, 17 October 214) GetArcIonos.cgi 7 6 pixels 29/4/215 17:39 GetArcIonos.cgi 7 6 pixels 29/4/215 12: UT 233 UT 8

9 Assimilation of real-time inputs to the model 1 Background ionosphere o Structure based on IRI o Relatively few ionosondes in polar cap and can be difficult to interpret and derive values o So, adjust IRI to fit to TEC observations derived from GPS measurements 9

10 Example TEC measurements (17 18 October 214) 7 algo: Oct alrt: Oct stec (TECU) 4 3 stec (TECU) Time (UT) Time (UT) Algonquin (mid-latitude) Alert (polar cap) 1

11 Assimilation of real-time inputs to the model 2 Patches o Intensity can be estimated from TEC data o Minimum number can be estimated from TEC data o Position of these can be estimated from TEC data o Cannot be certain that all patches are observed D-region absorption o Based on PCA models o Optimised using real-time riometer measurements 11

12 Concluding remarks Good fit to observations for coverage maps produced using historical data Are currently updating model to assimilate data from realtime sources some challenges remain Aim is to be able to: o forecast HF propagation conditions up to 12 hours ahead o provide this information for airline despatchers 12

13 Cross-polar flights, Number of movements Year 13

14 Absorption Riometer sites 16 o W 3 o N 45 o N 1 Apr 21, 2UT, Kp=5 1 o E 14 o W 6 o N 75 o N 8 o E 12 o W 6 o E 1 o W 4 o E Poleward wall 8 o W 6 o W Equatorward wall 4 o W 2 o W o 2 o E 14

15 Absorption HF and riometer measurements Signal-to-noise ratio on link between Kirkenes and Kiruna (44 km) Riometer absorption scaled to 6.78 MHz. 15

16 Absorption HF measurements and D-RAP predictions Qan Ale, Hz, SNR (db) Absorption (db)

17 TOF (ms) Example HF observations Qaanaaq Alert, 11.1 MHz, Azimuth (deg) snr (db)

18 Area coverage simulations Ray tracing model with realistic high latitude ionosphere To date, the simulations have been undertaken in relation to HF-DF, i.e. have been of the direction of arrival and time of flight of signals received over a fixed link. It is possible to undertake a very large number of ray traces to estimate the coverage area of a transmitter. Currently incorporating absorption models and real-time inputs 18

19 F-region critical frequency (left) and area coverage of 12 MHz signal transmitted from Cambridge Bay (right) at 18 UT 19

20 Signalling effects Multipath spread: Multiple propagation paths from transmitter to receiver. o signal fading and reduced data rates in digital systems. Doppler spread: Ionospheric movements can lead to signals arriving with a range of Doppler shifts o reduced signal quality and data rates. 2

21 Observations as a function of solar activity (21, and March 29 July 212) 5 Sub auroral, 7 MHz, Winter night 1 Sub auroral, 11 MHz, Equinox night 4 8 Dop Spr (Hz) 3 2 Dop Spr (Hz) Dop Spr (Hz) Polar cap, 11 MHz, Equinox night y=.8x+7.1 CMPS (ms) Sub auroral, 11 MHz, Equinox day y=.19x Smoothed SSN Smoothed SSN 21