Nowcasting Experience and Challenges in Hong Kong

Hong Kong Observatory
134A Nathan Road, Tsim Sha Tsui, Kowloon, Hong Kong
The Hong Kong Observatory (HKO) has developed a number of nowcasting systems in the past years for providing very-short-range forecasts and related warning services to the general public, aviation community and special users. The core of the HKO’s nowcasting system, named SWIRLS (Short-range Warning of Intense Rainstorm in Localized Systems), has evolved from extrapolation of radar echoes to blending with high resolution (2 km) NWP model to extend the nowcasting domain from 0-1 to 6 hours ahead. The non-hydrostatic model is initialized by a rapidly updated 3DVAR system assimilating remote sensing data, including radar reflectivity, radar Doppler winds, dual Doppler radar derived winds, GPS precipitable water vapour, and etc. The objective of bringing in rapidly updated NWP data is to supplement the extrapolation forecasts with storm growth and dissipation information. Experience showed that though the overall performance of the blended nowcasts could be improved, the spatial, temporal and intensity accuracy of the QPF still need to be improved.

Meanwhile, in the aviation community, an initiative named MSTA (Meteorological Services for Terminal Area) is under development under the joint collaboration between WMO/CAeM and ICAO. In MSTA, weather centres are required to provide (i) 0-2 hr nowcasting convection forecast; (ii) 2-6 hr short-range convection forecast; and (iii) 6 hr upto 48 hr probabilistic convection forecasts for the terminal area under concern. The goal is to produce high accuracy, trajectory dependent convective forecast for airlines/air traffic offices to assess the potential impacts for the convective development around a terminal on the air-traffic so that the airport authority could make timely decision to reroute the aircrafts if necessary. What the aviation community mostly concerns is the “impact” due to weather on their operation instead of QPF. To meet these high demands, the HKO has been testing a nowcasting system with ultrahigh NWP model (down to 200 m) with a view to providing high resolution information to support the operations of the aviation community. In connection with this, a new set of metric are being defined for verifying the “impact” instead of the primary meteorological forecast parameters.

For the special user, a nowcasting system has been developed for predicting Cloud-to-Ground (CG) lightning to support the operations of a local power company under inclement weather. The system adopts an ensemble nowcasting technique to provide up to 2 hr ahead probabilistic CG forecast over areas where overhead power lines are distributed. Feedback from the company confirmed that the probabilistic nowcast are useful to their operations, though some over-forecast/under-forecast cases would still need investigation.

While most nowcasting systems focus on convection, more and more demands are emerging from various sectors/communities urging nowcasting service to expand beyond convective forecasts. They include, e.g., low level windshear, turbulence, landsea-breeze. To meet these demands, the quality of NWP model in terms of the spatial, temporal and intensity of various forecast parameters needs to be improved further. This talk will highlight the gaps between the users’ needs and the current status of the nowcasting systems developed so far in Hong Kong with a view to pointing out the future directions of NWP research and development in support of providing better nowcasting services.