Andrés Perez Hortal;
I. Zawadzki;
M-K Yau
McGill University
The 8th EnKF Data Assimilation Workshop
A heuristic approach to radar data assimilation
Talk:
Perez_etal_Heuristic_DA_Method.pdf
We introduce a new ensemble DA technique in which an analysis is constructed by selecting, for each vertical column in the model, the ensemble member with precipitation at the ground locally closest to the corresponding radar observations. The proximity between the modeled and observed precipitation can in principle be determined by the mean absolute difference of intensity or by the closest CSI, ETS, etc. measured over a rectangular window centred at each grid point of the model. The new analysis has discontinuities due to the mosaic of the closest ensemble members. To reduce the impact of possible imbalances, the initial conditions for the new forecast are obtained by nudging the background states towards the analysis over a time interval. The underlying basic hypothesis of the approach is that the ensemble members that are locally closer to the observed precipitation are also closer to the truth in the other state variables.
The potential of the method was studied using Observing System Simulation Experiments employing a small ensemble of 20 members. The ensemble is produced by the WRF model, run at a resolution of 20 km with the Kain-Fritsch convective parametrization. The experiments lend support to the validity of the basic hypothesis and allow the determination of the optimal parameters for the approach, such as the state variables to be nudged, the nudging time interval, and the size of the rectangular window associated with each ensemble member.
Preliminary results indicate that this new DA technique is able to produce forecasts with considerable and long-lived error reductions in the fields of precipitation, temperature, humidity, and wind.
We introduce a new ensemble DA technique in which an analysis is constructed by selecting, for each vertical column in the model, the ensemble member with precipitation at the ground locally closest to the corresponding radar observations. The proximity between the modeled and observed precipitation can in principle be determined by the mean absolute difference of intensity or by the closest CSI, ETS, etc. measured over a rectangular window centred at each grid point of the model. The new analysis has discontinuities due to the mosaic of the closest ensemble members. To reduce the impact of possible imbalances, the initial conditions for the new forecast are obtained by nudging the background states towards the analysis over a time interval. The underlying basic hypothesis of the approach is that the ensemble members that are locally closer to the observed precipitation are also closer to the truth in the other state variables.
The potential of the method was studied using Observing System Simulation Experiments employing a small ensemble of 20 members. The ensemble is produced by the WRF model, run at a resolution of 20 km with the Kain-Fritsch convective parametrization. The experiments lend support to the validity of the basic hypothesis and allow the determination of the optimal parameters for the approach, such as the state variables to be nudged, the nudging time interval, and the size of the rectangular window associated with each ensemble member.
Preliminary results indicate that this new DA technique is able to produce forecasts with considerable and long-lived error reductions in the fields of precipitation, temperature, humidity, and wind.
Thanks to McGill's Atmospheric and Oceanic Sciences department for hosting this web-site.