Table 1.
BDS registers used to compute Mode-S/ADS-B winds.
Fig. 1.
Wind vector composition. Aircraft air motion (red), aircraft ground motion (blue) and wind (green).
Fig. 2.
Difference in magnetic declination between 2020 and 2010 IGRF-13 lookup tables over France and surrounding countries.
Table 2.
Summary of measured and derived meteorological parameters.
from va (airspeed), α (true heading) and ground trackComputed on ground
from va (airspeed), αm (magnetic heading) and ground trackComputed on ground
from va (airspeed), αm (magnetic heading) and ground trackAircraft positionComputed on board
from ground trackComputed on ground
from ATC radar localizationComputed on ground
from ADS-B location (computed on-board)
Fig. 3.
Flowchart of Météo-France and KNMI Mode-S data fluxes and processing for data assimilation in AROME-France.
Fig. 4.
Météo-France ADS-B antenna network (reception range ∼ 400 km). See Table 3 for exact positioning.
Table 3.
Location of ADS-B antennas from the Météo-France network.
Fig. 5.
Daily density coverage of Météo-France Mode-S/ADS-B active data 2 from 6 October to 6 November 2018 (a) and of KNMI Modes-S/EHS active data from 5 to 24 April 2018 (b) by pixels of 0.5°x0.5° (central panel) or 0.5°x50 hPa (right and top panels).
Fig. 6.
Daily density coverage of AMDAR active data from 6 October to 6 November 2018 by pixels of 0.5°x0.5° (central panel) or 0.5°x50hPa (right and top panels).
Table 4.
Global BGD mean and SD of u and v wind components for raw MF Mode-S/ADS-B, VarBC MF Mode-S/ADS-B, KNMI unbiased Mode-S/EHS and AMDAR data from 5 to 24 April 2018.
Fig. 7.
Boxplots of u (blue) and v (red) BGD for aircraft A1 for raw (boxplots 1 and 3) and VarBC (boxplots 2 and 4) Mode-S/ADS-B data from 5 to 24 April 2018. Dots: standard deviation, circles: mean.
Fig. 8.
Evolution for each observation of a aircraft A1 from 5 to 24 April 2018 of Mode-S/ADS-B u component raw (black) and VarBC (blue) BGD in top panel and VarBC computed biases (total bias black, heading bias blue, airspeed bias red) with aircraft heading (green dots) in bottom panel.
Fig. 9.
Evolution, for a aircraft A1, by analysis where data are assimilated from 5 to 24 April 2018 of u (blue) and v (red) diagnosed heading bias (top panel), airspeed bias (middle panel) and passive 3 data number by analysis (bottom panel).
Fig. 10.
Boxplots of aircraft u (blue) and v (red) BGD mean (light colors) and SD (dark colors) for raw MF Mode-S/ADS-B (left), VarBC MF Mode-S/ADS-B (middle) and AMDAR (right) datasets from 6 October to 6 November 2018.
Fig. 11.
Fractional mean absolute bias amplitude by predictor W p i (solid: heading, dashed: airspeed) and relative total mean absolute bias amplitude B r p (grey dots with linear regression) for u (top panel) and v (bottom panel) by individual aircraft ordered by increasing values of Wp 1 from 6 October to 6 November 2018.
Fig. 12.
(a) u (blue) and v (red) computed heading (top panel) and airspeed (bottom panel) biases means by aircraft ordered by increasing u bias parameter mean from 6 October to 6 November 2018. Grey: percentile 5 and 95 of u and v bias parameters means by aircraft. (b) Distribution of estimated heading biases from u (blue) and v (red) wind components (6 October to 6 November 2018).
Fig. 13.
Evolution of u (blue) and v (red) diagnosed heading bias (top panel), airspeed bias (middle panel) and passive data number by analysis (bottom panel) for aircraft A1 by available analysis from 6 October to 6 November 2018.
Fig. 14.
Evolution of Mode-S/ADS-B u component raw (black) and VarBC (blue) BGD in top panel and VarBC computed biases (total bias black, heading bias blue, airspeed bias red) with aircraft heading pointed in green in bottom panel (from 6 October to 6 November 2018 for each observation of aircraft A2). The vertical dotted line corresponds to the 17 October 2018.
Fig. 15.
Evolution, of u (blue) and v (red) diagnosed heading bias (top panel), airspeed bias (middle panel) and passive data number by analysis (bottom panel) for aircraft A2, by available analysis from 6 October to 6 November 2018. The vertical dotted line corresponds to the 17 October 2018.
Fig. 16.
Boxplots of u (blue) and v (red) BGD for aircraft A3 for raw (boxplots 1 and 3) and VarBC (boxplots 2 and 4) Mode-S/ADS-B data from 6 October to 6 November 2018. Dots: standard deviation, circles: mean.
Fig. 17.
Evolution of Mode-S/ADS-B u component raw (black) and VarBC (blue) BGD in top panel and VarBC computed biases (total bias black, heading bias blue, airspeed bias red and aircraft heading pointed in green in bottom panel) from 6 October to 6 November 2018 for each observation of aircraft A3.
Fig. 18.
Evolution, of u (blue) and v (red) diagnosed heading bias (top panel), airspeed bias (middle panel) and passive data number by analysis (bottom panel) for aircraft A3 by available analysis from 6 October to 6 November 2018.
Fig. A1.
Boxplots of aircraft u (blue) and v (red) for Mode-S minus AMDAR (first from the left), Mode-S VarBC minus AMDAR (second), Mode-S minus AROME (third), Mode-S VarBC minus AROME (fourth) and AMDAR minus AROME (right) collocated datasets from 6 October to 6 November 2018. Star: standard deviation, square: mean.
Table A1.
Representativeness error estimation.
Table A2.
Summary of the parameters of the error model decomposition (Equation (9)) derived from the TC and TCb collocations of AMDAR, Mode-S and AROME 1 h forecasts data for u (top) and v (bottom) wind components.
TC/TCbai bi / ai TC: AMDAR101.2831.259TC: Mode-S VarBC1.0110.0821.1171.098TC: AROME10.987−0.0291.6851.732TCb: AMDAR101.4551.265TCb: Mode-S no BC1.020.2583.645NATCb: AROME10.989−0.0371.5531.771v collocation
TC/TCbaibi/ aiTC: AMDAR101.3251.334TC: Mode-S VarBC1.006−0.1071.1471.178TC: AROME10.985−0.0421.7881.803TCb: AMDAR101.461.347TCb: Mode-S no BC1.0090.1093NATCb: AROME10.986−0.0511.7061.812