Chih-Chien Tsai
Taiwan Typhoon and Flood Research Institute, National Applied Research Laboratories, Taipei, Taiwan
Youngsun Jung
Center for Analysis and Prediction of Storms, University of Oklahoma, Norman, Oklahoma
The 8th EnKF Data Assimilation Workshop
Quantitative Precipitation Forecasting with Polarimetric Radar Data Assimilation: Typhoon Soudelor (2015)
Poster:
20180507_CCTsai.pdf
Polarimetric radars provide observation variables such as differential reflectivity (
), specific differential phase (
) and co-polar correlation coefficient (
) that probe more microphysical characteristics of hydrometeors in addition to radial velocity (
) and horizontal reflectivity (
). To investigate the effects of polarimetric radar data assimilation on simulating heavy rainfall events in Taiwan, an observation operator for , and in accord with various bulk cloud microphysical schemes is incorporated into a Weather Research and Forecasting (WRF)-local ensemble transform Kalman filter (LETKF) system. The case of Typhoon Soudelor (2015) is selected for its devastating wind and rainfall impinging on northern Taiwan, where Central Weather Bureau’s RCWF S-band polarimetric radar is situated. Comparing five cold-start simulations using different double-moment cloud microphysical schemes with RCWF observations, the one using the WRF double-moment six-class (WDM6) scheme is found to give a clear structure of spiral rainbands with the smallest wet bias. Then, using WDM6, several assimilation experiments are carried out to assimilate various combinations of observation variables for nine 15-minute cycles after a 6-hour spin-up of the same perturbed initial ensemble. The results of deterministic forecasts show that radar data assimilation greatly improves the spatial distribution of predicted rainfall for the first 3 hours within the coverage of RCWF, but deterioration occurs for the second 3 hours beyond the coverage. It is a good choice to assimilate all observation variables while contributes more than which in turn contributes more than .
Polarimetric radars provide observation variables such as differential reflectivity (
), specific differential phase () and co-polar correlation coefficient () that probe more microphysical characteristics of hydrometeors in addition to radial velocity () and horizontal reflectivity (). To investigate the effects of polarimetric radar data assimilation on simulating heavy rainfall events in Taiwan, an observation operator for , and in accord with various bulk cloud microphysical schemes is incorporated into a Weather Research and Forecasting (WRF)-local ensemble transform Kalman filter (LETKF) system. The case of Typhoon Soudelor (2015) is selected for its devastating wind and rainfall impinging on northern Taiwan, where Central Weather Bureau’s RCWF S-band polarimetric radar is situated. Comparing five cold-start simulations using different double-moment cloud microphysical schemes with RCWF observations, the one using the WRF double-moment six-class (WDM6) scheme is found to give a clear structure of spiral rainbands with the smallest wet bias. Then, using WDM6, several assimilation experiments are carried out to assimilate various combinations of observation variables for nine 15-minute cycles after a 6-hour spin-up of the same perturbed initial ensemble. The results of deterministic forecasts show that radar data assimilation greatly improves the spatial distribution of predicted rainfall for the first 3 hours within the coverage of RCWF, but deterioration occurs for the second 3 hours beyond the coverage. It is a good choice to assimilate all observation variables while contributes more than which in turn contributes more than .
Thanks to McGill's Atmospheric and Oceanic Sciences department for hosting this web-site.