Improving weather forecasting in the Philippines through WRF dynamical downscaling and data assimilation

In 2012, the Nationwide Operational Assessment of Hazards (NOAH) program was established by the Philippine Department of Science and Technology (DOST) to address the pressing need for developing an integrative and sustainable system that will mitigate loss associated with natural hazards. One of the projects under the NOAH program is the Weather Information-integration for System Enhancement (WISE), established in early 2013, which is a multi-agency effort that aims to enhance numerical weather prediction through the use of Weather Research and Forecasting (WRF) model in 1) dynamical downscaling of global forecast system to 12-km and 4-km over the Philippines; 2) 3DVAR and 4DVAR data assimilation of surface measurements, Doppler radar rainfall estimates and satellite radiances; and 3) sensitivity testing of microphysics, cumulus parameterization and initial boundary conditions. Metrics for objective forecast verification are used to evaluate the performance skills of the models.

An initial assessment of the performance of WISE shows an improvement to 70 – 80% accuracy for the dichotomous rainfall forecast with the highest accuracy observed in January, February and March period. It was also observed that assimilating expected rainfall from radar has the greatest impact in the model's skill with mostly positive Critical Success Index (CSI) anomalies. The performance is expected to be significantly improved with the availability of more than 800 new rain gauges installed around the country. Sensitivity testing was done for the case of Typhoon Haiyan, and it was found that the use of Grell-Divenyi cumulus physics scheme provided the best rainfall forecast while the WSM 6-class microphysics scheme provided the best forecast for pressure and wind speed and Kain-Fritsch cumulus physics scheme gave the closest to the observed typhoon track.

Future efforts will include assessment of impact of satellite data assimilation, particularly that of MODIS calibrated radiances, brightness temperature and cloud products. Greater utilization of satellite data is also expected in the validation and evaluation of forecast performance skills. Multi-model comparison and ensemble forecasting will be implemented in order to further improve the Philippines' numerical weather prediction capabilities.