The AURORA Nowcasting Platform-Extending the concept of a modifiable database for short range forecasting

Our group’s current development in support of the Canadian Weather Research Program (CWRP) is directed towards interactive analysis and short range forecasting of high impact weather phenomena.  The AURORA Nowcasting Platform has been created to prototype new techniques and procedures, as a research tool for case studies and proof-of-concept development for a number of current and future projects.

The AURORA platform builds on the concept of a database of various types of meteorological information that can be displayed and modified by meteorologists to describe an evolutionary sequence for weather events.  Short term forecasting requires the display of multiple datasets: observations, gridded data (from both analyses and forecasts), imagery (including satellite and radar products), as well as objects such as lines and areas that can be extracted or inferred from the various data sets.  Some of this information can be extracted automatically, yet there are many situations where human pattern recognition skills can guide or enhance such processes.  As well, human skills of decision making in the face of incomplete or contradictory data allows for the addition of information which would not otherwise be available to a fully automated system.

The forecast component of the system can use scientific models or processes to provide a conceptual framework for predicting the occurrence of severe events, but the actual event intensity, timing and evolution can be very unpredictable.  The AURORA platform has been developed with the concept of forecaster interaction as an essential requirement for both the analysis and forecast components of short range forecasting.

The result is an information database with a seamless transition from analyses to forecasts, that can be continually validated and updated based on new data.  The database design also allows information to be accessed at any geographical location for running models or processes, or for automatically generating forecasts in both text and graphics.  Interaction with a number of current and future research projects in hydrology and airport aviation will be discussed.