Hong Kong International AIRPORT AvRDP Phase I Progress Peter PW LI Hong Kong Observatory Hong Kong, China WMO WWRP 4 th International Symposium on Nowcasting and Very-short-range Forecast Hong Kong, China 26 July 2016
Hong Kong International airport (HKG) One of the busiest airports in the World (2015) 2 Runway system Flights: 1112/ day 07L/ 25R 68 flights/hr at peak hours 07R/ 25L Subtropical in Northern Hemisphere 22 ° 18′ 32″ N 113 ° 54′ 52″ E Environment: reclaimed island surrounded by water next to high mountain (~ 1km height) 2
Major impacting weather at HKG Mean Number of Days with Specified Weather Phenomenon at Hong Kong International Airport (January 1998 - December 2010) 10 9 8 7 6 5 Fog 4 Lightning 3 Thunderstorm 2 1 0 Major impacting weather Convection and Thunderstorm 3
Low Level Windshear and Turbulence Frequency distribution by month • Significant windshear: 1 in 500 flights Significant turbulence: • 1 in 2000 flights 2nd peak • Winds blowing across terrain Under non-rainy (terrain-induced) weather • Sea breeze (~90%) • Gust fronts Under convective weather • Microbursts (~10%) • Low-level jets < 1% 1st peak 1st peak -> northeast monsoon 2nd peak -> typhoon season
AMOS RVR Anemometer Forward Scatterer Pressure, temperature sensor, rain gauge
HKG RVR and Forward scatterer
and ATM data Observations, Nowcasting facility, modelling facility HKGA AvRDP IOP data, including Airport HKG Airport Observations Weather Radar (conventional or Doppler) Geostationary Satellite Wind profiler LIDAR Anemometer Visibility sensor AMDAR/ACARS data Other observations model system and Nowcasting Nowcasting system Micro/mesoscale NWP Regional model PIREP Aircraft data ATM data ATM capacity data Air traffic data ADS-B (since 2016)
HKG 1 st IOP for Convection in Northern Hemisphere (May – Sep 15) HKO IOP Data and Data Manual • A number of significant convection cases • A few typhoon cases Observational data • • Nowcasting data Mesoscale model data • • NWC + NWP blended data ATM data • – Airport Capacity data
Real-time airport air traffic data
HKG 2 nd IOP for Convection in Northern Hemisphere (mid May – mid Sep 16) • Completed study: • A number of significant convection cases – Fine-tuned Airport Thunderstorm Nowcasting • One TC case System (ATNS) adjusted thresholds to fit ATC requirements Observational data • • On going Studies: • Nowcasting data – Airport Departure Rate Mesoscale model data • Aircraft delay using Big Data technology – • NWC + NWP blended data – Aircraft Avoidance study using ADS-B data ADS-B within 400km of HKIA • • ATM data – Airport Capacity data – Air traffic data
ADB-S collected (prepared for Phase II) ABS-B overlaid with weather radar and satellite
Effect of Significant Convection to Capacity Types of effect of significant convection Probability t bability to be affected b ected by weather ther + Trajectory Based SigConv F/C + ADS-B flight position + … Capacity Forecast
Air Traffic Data
Maximum Airport Acceptance Rate
Flight Delay using Big Data Time Range Data Granularity Number of records/images Flight 20150506- 13,470 Information 20150531 No. of distinct flights: 890 No. of distinct destination code: 178 Radar Images 20150507- Every 6 minutes 47,541 images with size 480*480 20150531 -- Radar image taken at 3km REF3 -- Vertical Integration Liquid (VIL) Flight Information REF3 Image VIL Image To study the correlation between the convective weather condition and the flight departure status
Departure routes information from HKAIS and divide them into 3 groups
Deep Learning Network • Multilayer Perceptron (MLP) – Train 3 models for 3 departure groups respectively Radar Image REF3 Delay Indicator VIL info Reference (0/1) chart Input layer Hidden Output layer layer
HKO Nowcasting Systems • Radar based nowcasting systems: – General purpose convection radar-based nowcasting – Short-range Warning of Intense Rainstorm for Localized System (SWIRLS) – For Aviation Community – ATLAS (Airport Thunderstorm and Lightning System) and ATNS (Aviation Thunderstorm Nowcasting System) – Community SWIRLS • Mesoscale Models: – Non-hydrostatic model (NHM) – Aviation Model (AVM) • Blended radar-based nowcast + mesoscale model -> 0-6 hr nowcasting system
Aviation Thunderstorm Nowcasting System (ATNS) Specific Forecast for waypoints 0-1 hour Semi-Lagrangian extrapolation scheme (LE)
1-hr convection nowcast for tactical decision-making (prototype) Flight specific weather forecast
RAPIDS-NHM Forecast Mesoscale Diagnostic Parameters Moisture flux + convergence Storm Motion Vector (SMV) Updraft Helicity (UDH) Simulated radar products 3km Reflectivity Max Isothermal reflectivity at -10 degC Vertically integrated ice (VII)
Blending LE with NWP Nowcasting component – LE 0 - 6 hr QPF by extending the linear extrapolation of radar echoes NWP component – Non-hydrostatic Model (NHM) 0 – 6 hr QPF by 2-km non-hydrostatic numerical model 3DVAR, Doppler, dual-radar 3D wind, GPS/PWV, etc. + 3 DAR Spatial & intensity adjusted • • No temporal adjustment • Dynamic-weighting 23
Fine-scale Aviation Model (AVM) for HKIA • Sub- km implementation of WRF-ARW v3.4.1 – Pearl River Delta (PRD): Δ x = 600 m – Hong Kong Int’l Airport (HKA): Δ x = 200 m • Hourly-updated forecasts • Up to T+(7 – 9) Aviation-impact Weather around HKIA Outer domain: 350kmx350km Inner domain: 60kmx60km
Predicting Meso/Micro-scale Systems
Ensemble WRF+NCEP GEFS “Downscaler” 10-km 20-member Twice-daily / T+72 Based on WRF v3.4.1
Probabilistic Assessment of Severe Weather Hong Kong Rainstorm / Sig. Conv. Very Hot Weather ( ≥ 33 ◦ C) (3-hr rainfall percentile) Strong/Gale Winds Domain: 1800kmx1800km (exceeding F6 or F8) dx = 10km
Community SWIRLS (com-SWIRLS) - an radar-based nowcasting system Com-SWIRLS in action (SAWS radar data)
THANK YOU Q & A MRI/ NIED/ JMA 13-14 29 March 2012
Observational Data collected RADAR Sounding SYNOP METAR Local AWS AMDAR Wind profiler Lightning
Nowcasting / Model / ATM data collected MSQ NHM ATNS SIGCONV and Capacity Notification RAPIDS
Recommend
More recommend
