A Cutting-edge Approach to NWP in the Tropics
The ‘SINGV’ convective-scale numerical weather prediction system has been designed especially for local weather forecasting in the Singapore region. It is developed through a multi-year project in collaboration with the UK Met Office. It adopts the latest version of the UK Met Office’s Unified Model with further adaptations for Singapore and the surrounding region.
Two Significant Future Developments in MSS’ NWP Capability Building
- Data Assimilation: Improves Model Initial Conditions and Forecasts
Observations are processed and quality controlled and combined with the forecasts from the NWP model using advanced (variational and ensemble) data assimilation techniques to produce the best estimate of the current state of the atmosphere as initial conditions. These improved initial conditions will lead to better model forecasts.
- Ensemble Forecasting: Provides Forecast Uncertainty Estimates
A group of forecasts (called an ensemble) is run from slightly different initial conditions and/or different model configurations. These forecasts provide a representative sample of possible future weather conditions. The uncertainty in model forecasts and probabilities of weather events can be estimated quantitatively, which has been established internationally to be the best way to include uncertainty in short term weather forecast. This is particularly the case on the convective-scale and in the tropics.
The new forecast system will lead to improvements in severe weather warnings, provide meteorological inputs to dispersion modelling of pollutants and particulate aerosols, and improve capability in risk management of weather-related activities such as water resource management and air traffic management.
Operational NWP Studies
SINGV is in operation to provide weather forecast guidance to the weather service operational forecasters since 2019. The domain of the current SINGV system is in 1.5 km spatial resolution. The initial and boundary conditions are given by ECMWF data. Regional observations are ingested into the system through 3DVAR data assimilation scheme. The system provides regional forecasts in the Singapore region up to 48 hours and 8 times a day.
Nowcasting System Development
CCRS has been expanding its weather nowcasting capabilities on a very short time range (0-6 hours). We have been using neural network and blending nowcast with SINGV NWP forecasts. Efforts have also been putting into machine learning based algorithms to improve rainfall forecasts.
Ocean and Wave Modelling System
SEAOM-WAM and WAVEWATCH III are the current operational ocean and wave forecasting system in CCRS. Research has been ongoing for a fully coupled atmosphere, ocean, and wave forecasting system in 1.5km spatial resolution.