A growing range of tools is steadily becoming available for weather forecasters as a basis for their forecasts and warnings. These are, however, not always used optimally, because of a lack of interaction between their developers and the forecasters. Additionally, there is a lack of international exchange of “best practices” on forecasting extreme weather events. Therefore, the ESSL Testbed, as a permanent facility, will bring together forecasters and developers from across the world: In a quasi-operational setting with a focus on severe weather, developers will present and explain their tools, forecasters give feedback, and the tools are put to the test. The core activity of the Testbed is the preparation of experimental severe weather forecasts.
At the Testbed, participants take part in real-time testing of new forecasting and nowcasting tools that are based on model, satellite, radar or observational data. Concretely, the participation by both forecasters and developers entails one week of intense hands-on severe weather forecasting and forecasting training, by jointly preparing experimental forecasts for the real-time weather as it occurs across the European continent, supervised by Testbed staff and involving external experts.
In doing so, the forecasters learn how to use the new supporting products (that may be radar-based, model-based, satellite-based, or otherwise) and new forecasting techniques taught to them by international experts. The developers, at the same time, benefit from the feedback to their products that they receive from forecasters. In addition to this interactive component, a daily “classic training” component is foreseen: Participants attend a lecture by an expert forecaster or developer. The Testbed team is arranging for several experts from Europe and the USA to contribute to these trainings.
The paradigm at the basis of the Testbed product testing is the following three step approach: Firstly, experimental forecasts are prepared by participant groups. Secondly, these forecasts are verified against the observed severe weather. Finally, the role of the experimental products in the forecasting process is evaluated. The evaluation takes place in group discussions and is recorded through questionnaires. The collected feedback is condensed into a report that will be provided to the product developers.
Motivation for the Testbed
ESSL think that developers and their products could benefit from more exposure to meteorological operations, and forecasters from a better understanding of the technical background of the model-, radar-, satellite- and observation-based products shown on their workstations. Better international interaction among these people could lead to the establishment of “Best Practices”, a common standard of tools and forecasting techniques that have proven themselves internationally.
This will also help to mitigate the common symptom of “data overload” on part of the forecaster: In an optimal situation, she would only be presented a selection of data products that have proven themselves and that have been optimized trough an iterative process between developers and forecasters. Moreover, she would have a sound conceptual framework in the back of her mind within which to interpret the data, which saves time and improves forecast and warning quality, especially in time-critical severe weather situations. Both aspects are addressed by the Testbed: it intends to foster both the development of optimized forecast-supporting products and to provide forecasters with a conceptual framework for severe weather forecasting based on science.
Within the United States National Oceanographic and Atmospheric Administration (NOAA) there exists the Hazardous Weather Testbed (HWT), a permanent facility for testing experimental products for use in operational forecasting and warning practice, organized primarily by two of its divisions: the Storm Prediction Center and the National Severe Storms Laboratory. The ESSL Testbed was inspired by the HWT and its yearly “Spring Program”. An important difference is that the ESSL Testbed is not tied to one country’s public meteorological service, but that instead forecasters from many different countries are welcome. Moreover, it puts slightly more focus on the aspect of forecaster training.