Machine-Learning Based Hourly Probabilistic Forecast of Sea-Breeze Induced Runway Headwind Changes at the Hong Kong International Airport

The Hong Kong International Airport (HKIA) has been built on a land reclamation extended from a small island called Chek Lap Kok in Hong Kong. The HKIA is surrounded by sea at the east, north and west, and a large landmass to its south. The diurnal variation of temperature difference between the sea and the airport landmass occasionally triggers a prominent sea-breeze circulation at the airport. While the prevailing wind direction in Hong Kong is from the east, the runways at HKIA are oriented in the east-west direction and the onset of sea breezes often leads to westerly winds over the runways during the day (Cheng, 1999). Accurately predicting the evolution of the sea-breeze circulation and the timing of wind direction changes are crucial for efficient airport operations, as the runway in use for aircraft landing and take-off is determined by the prevailing wind direction over the runways.

By identifying weather elements that contribute to the development of westerly sea breezes at the HKIA, machine learning techniques are considered a suitable method to provide a quantitative likelihood assessment in hourly intervals. This presentation will briefly discuss the sea breeze effects at HKIA, the machine learning model adopted for sea breeze forecasting, the initial validation results and several case studies. The hourly probabilistic forecast of runway headwind change demonstrates the added-value in forecasting runway wind direction, as well as enabling early preparedness and decision-making on runway in use for air traffic flow management.

Reference:

Cheng, C.M. (1999). Characteristics of sea breezes at Chek Lap Kok. Hong Kong Observatory Technical Note No. 96. Hong Kong Observatory.