Aaron Johnson- Feb 14

Name:     Aaron Johnson Title:    What is the optimal configuration of convection-permitting ensemble forecasts? Location: NWC 5600 Date:     2020/02/14 Time:     03:00 PM Series:   Convective Meteorology (Mesoscale Dynamics) Abstract: Choosing a configuration of a convection-permitting ensemble forecast system that optimally samples the relevant sources of forecast uncertainty in order to provide skillful probabilistic

Start

February 14, 2020 - 3:00 pm

End

February 14, 2020 - 4:00 pm

Address

120 David L Boren Blvd, Norman, OK 73072   View map
Name:     Aaron Johnson
Title:    What is the optimal configuration of convection-permitting ensemble forecasts?
Location: NWC 5600
Date:     2020/02/14
Time:     03:00 PM
Series:   Convective Meteorology (Mesoscale Dynamics)
Abstract: Choosing a configuration of a convection-permitting ensemble forecast system that optimally samples the relevant sources of forecast uncertainty in order to provide skillful probabilistic forecasts is a non-trivial task.  Such ensembles resolve a broad range of spatial scales, ranging from the convective scale to the synoptic scale.  Consideration of how the ensemble will sample the multi-scale analysis uncertainty, and how this uncertainty interacts with model and physics uncertainty, is important to ensure that the ensemble forecast reflects the true uncertainty of the forecast.  Care must also be taken when choosing verification metrics for the optimization of ensemble configurations.  One of the main advantages of forecasting at a convection-permitting resolution is the ability to obtain guidance on convective storm mode and morphology, regardless of the precise placement of individual storms.  Therefore verification frameworks must also take these features into account.  For some applications, choices about the horizontal and vertical resolutions of the individual ensemble members can also be a key consideration in the ensemble design.
Although no single optimal configuration can be proposed for all users, this seminar will synthesize the highlights of several research projects related to the above issues. The results from these studies can inform the choices made when considering the various aspects of convection-permitting ensemble forecast system design.
This seminar will primarily focus on the methods of generating finite samples of the initial condition uncertainty, an object-based framework for probabilistic verification, and upscale impacts of properly resolving atmospheric bores in the context of predicting nocturnal convection.