Abstract
The Hurricane Analysis and Forecast System (HAFS) is the next-generation operational modeling system for predicting tropical cyclones in support of the mission of the National Hurricane Center. HAFS is a specific application of the Unified Forecast System (UFS), which represents an effort by NOAA, NSF NCAR, the Developmental Testbed Center (DTC), and other partners to improve weather and earth system prediction by consolidating disparate models, software packages, and workflows into a system of applications sharing unified dynamical cores, physics, and other components. Maintained and developed by the NOAA Environmental Modeling Center (EMC), HAFS undergoes continual development as both scientific and technical improvements are introduced by HAFS developers. Simultaneously, the communal nature of the various components of HAFS, such as the UFS Weather Model and the Common Community Physics Package (CCPP), means that new modifications introduced by external groups must be constantly tested and evaluated to ensure they do not degrade the existing HAFS configurations. This rapid pace of development necessitates a verification system that is agile and robust enough to allow for quick comparisons between these various prototypes of HAFS and the operational versions.
In the DTC we have previously created an Agile Benchmark Testing Framework (ABTF) based on the advanced Python-based Model Evaluation Tools (METplus) for comparing prototypes of the Rapid Refresh Forecasting System (RRFS). Leveraging this previous effort, we have embedded a verification workflow within the HAFS Application, both to implement an agile testing framework within HAFS for developmental use, as well as to benefit the UFS community for research and scientific purposes. In the spirit of a unified approach to earth system modeling, this verification capability makes heavy use of the existing scripts and workflow developed in the context of the UFS Short-Range Weather (SRW) Application, as well as increased use of Unified Workflow Tools (uwtools) which further applies shared tools and code across UFS workflows.
This presentation will describe the implementation of an agile testing framework for verification within HAFS, and detail ongoing and future efforts to unify model verification across different applications of the UFS.
In the DTC we have previously created an Agile Benchmark Testing Framework (ABTF) based on the advanced Python-based Model Evaluation Tools (METplus) for comparing prototypes of the Rapid Refresh Forecasting System (RRFS). Leveraging this previous effort, we have embedded a verification workflow within the HAFS Application, both to implement an agile testing framework within HAFS for developmental use, as well as to benefit the UFS community for research and scientific purposes. In the spirit of a unified approach to earth system modeling, this verification capability makes heavy use of the existing scripts and workflow developed in the context of the UFS Short-Range Weather (SRW) Application, as well as increased use of Unified Workflow Tools (uwtools) which further applies shared tools and code across UFS workflows.
This presentation will describe the implementation of an agile testing framework for verification within HAFS, and detail ongoing and future efforts to unify model verification across different applications of the UFS.
| Original language | American English |
|---|---|
| Title of host publication | American Meteorological Society Annual Meeting |
| State | Published - Jan 16 2025 |