Blending radar-based nowcast and high-resolution Numerical weather prediction (NWP) model: A demonstration at the CASA DFW testbed
Abstract
With an increasing population in urban areas, the accurate prediction of severe weather events is essential for saving lives and properties. The Collaborative and Adaptive Sensing of the Atmosphere (CASA) is operating a X-band dual-pol radar network over the Dallas Fort Worth (DFW) urban area. In addition CASA provides the real-time high-resolution reflectivity nowcast over DFW radar network. The nowcasting is done using the Dynamic and Adaptive Radar Tracking of Storms (DARTS) system. The current CASA real-time nowcast system can utilize both NEXRAD and CASA DFW network radar observations also provides the merged reflectivity nowcast for 30 minutes lead time. In this study, we have analyzed ten precipitation events for nowcating using DARTS over DFW urban area and classifying them into three categories such as super-cell, line, and multi-cell. The results show that DARTS has very good nowcast skill score in case of supercell events as compared to the other two cases. In the analysis it is found that DARTS nowcast can be used until the lead time of 1 hour in case of super-cell events
To improve the CASA nowcast skills beyond lead time of 1 hour, we blend the DARTS nowcast with high-resolution numerical weather prediction (NWP) model. In this study, we have used the WRF model to provide prediction for 6 hours of lead time. The community GSI system is used to assimilate both the reflectivity and radial velocity data from the dual-pol radar network. The current blending technique uses a hyperbolic tangent function to estimate the weight of the WRF model prediction and the DARTS nowcast. This weighting function gives more weights to the DARTS nowcast till a lead time of 1 hour and then lowers the weights after a lead time of 2 hours. In between the lead times of 1 and 2 hours, the weights of DARTS nowcast decreases exponentially. However limitation of this simple blending technique is the weights are spatially uniform. This research work is focused on improving this blending process, by estimating the event based spatial and temporal weights for DARTS and WRF model. The new blending technique will be compared with other existing methods and its advantages will be presented. This study also focuses on identification and correction of the systematic bias in WRF simulations for different precipitation events over DFW urban area.- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2018
- Bibcode:
- 2018AGUFMNG33B0951R
- Keywords:
-
- 3315 Data assimilation;
- ATMOSPHERIC PROCESSESDE: 1910 Data assimilation;
- integration and fusion;
- INFORMATICSDE: 3275 Uncertainty quantification;
- MATHEMATICAL GEOPHYSICSDE: 4468 Probability distributions;
- heavy and fat-tailed;
- NONLINEAR GEOPHYSICS