1. Introduction

Challenges for accurate weather predictions still exist in numerical models. To facilitate the development and improvement of the NOAA operational model, NOAA’s Global Forecast System (GFS) is shared with the broad community through the Unified Forecast System (UFS) applications. The UFS Medium-Range Weather (MRW) Application (App) provides the source code of GFS, including physics schemes for GFS version 15 (CompSet GFS.v15p2) and additional developmental schemes for GFS version 16 (CompSet GFS.v16beta). The most up-to-date source codes for GFS.v16 can be accessed from the ufs-weather-model GitHub repository. GFS.v15 went into operations in June 2019, the operational implementation fo GFS.v16 is currently planned for February 2021 (Yang 2020). The known biases and development priorities related to GFS.v15 (Stan et al. 2019) and GFS.v16 (recently presented by Yang 2020) have been identified and listed below:

  • Less skillful hurricane track forecasts for strong storms in the Atlantic basin

  • Progressive with synoptic patterns

  • Extreme 2-m temperature biases in the mid-west region in the warm season

  • Cold bias in the lower troposphere and near the surface in the winter season

  • Precipitation dry bias for moderate rainfall

  • Struggle to capture boundary layer inversions

This document provides case studies evaluation results for GFS using the UFS Medium Range Weather (MRW) App with the target audience being a broad community. These cases are also run using the NOAA Environmental Model Center (EMC) Global Workflow, which points to the most up-to-date GFS codes, and are expected to provide insights for the on-going model development. Brief descriptions for MRW App and Global Workflow are below.

The UFS Medium Range Weather (MRW) App uses the Common Infrastructure for Modeling the Earth (CIME) workflow that incorporates pre-processing software, forecast model, and post-processor. The app serves as a useful tool to conduct the UFS WM runs. The latest evaluation results are based on physics compsets of GFSv15p2 and GFSv16beta employed in UFS Medium Range Weather App v1.0 (MRW.v1.0), hereafter referred to as MRW_GFSv15p2 and MRW_GFSv16beta, respectively.

The goal of this ongoing effort is to provide the community, as well as the physics development team, with a model testing platform where they can use the resources to conduct model runs and evaluate the model performance for representative meteorological cases. These case studies will provide insights for future model developments and aim at improving NOAA numerical weather forecasts. It should be noted that this is an ongoing effort that is aligned with the model public release and model development. Namely these evaluation results only apply to specific model versions. Timeline of physics frozen date in different model versions, including both UFS MRW APP CompSets and ufs-weather-model GitHub tags, are shown below:

_images/timeline_Apr2021.png

Timeline of physics frozen date

The case catalogue in this documentation are created based on the known biases of GFS model. This is a list that we are updating diligently. Please come back to check updates anytime.

References

Stan C., Yang F., and Harris L. (2019). UFS Development Goals and Priorities for Medium-Range and S2S Applications. Unified Forecast System Community. [Link]

Yang F. (2020). Development and evaluation of NCEP’s Global Forecast System Version 16. Unified Forecast System Community Webinar, Oct 22, 2020. [Link]