7TROPICAL Use of an Agile Testing Framework to Analyze Alternative Physics in the Hurricane Analysis and Forecast System (HAFS)

This study demonstrates an agile testing framework while testing alternative physics in the Hurricane Analysis and Forecast System (HAFS).

Testing and validation of new hurricane model configurations can be labor-intensive without proper tools. Whether one is testing hundreds of simulations across three hurricane seasons, or doing a detailed case study of only a few, the mechanics of running such experiments can take a vast majority of the time. This can be made easier if one has a coherent, end-to-end system, that automates the test process, and is connected to both standard and model-specific tools.

Here, we present such an automated system, by using it to study the effects of alternative physics schemes. Our target model, the HAFS, is a tropical cyclone numerical weather prediction model, run operationally by the National Weather Service. It includes a data assimilation system, and coupled atmosphere, wave, and ocean models. HAFS has been tested with a variety of physics schemes, but the two main configurations are similar. We've updated the HAFS to allow more of a diversity of schemes, and use this agile testing framework to examine how the model performs.

Our analysis of the effects of alternative physics schemes will exemplify this agile testing framework, and future features that will be added to it. The framework is freely available for other developers of the HAFS to use. Some aspects of it are tools which could be reused for other models. It is our hope that this framework will allow more science to be done with less labor, streamlining the experimental process.