Incorporation of CAM Ensemble-Based Assimilation into the Deterministic HRRR for HRRRv4 and Work to Transition from the HRRR to an FV3-Based Rapid Refresh Forecast System (RRFS)

The NCEP operational Rapid Refresh (RAP) and High Resolution Rapid Refresh (HRRR) model systems provide gridded analysis and forecast fields that are widely used for a variety of aviation applications. The long-term record of improvement for these systems has played a key role in overall improvement in aviation weather guidance over the past several years. These RAP and HRRR improvements have continued through the RAPv4 / HRRRv3 that was operationally implemented at NCEP in July 2018.

This presentation will provide an update on the latest developmental work for HRRR system, which is focused on two principal areas: 1) use of ensemble covariance information from the HRRR Ensemble(HRRRE) data assimilation system to improve the initialization of the deterministic HRRRv4 and 2) ongoing work to transition from the HRRR to an FV3-based CAM ensemble assimilation and prediction system that will be known as the Rapid Refresh Forecast System (RRFS). Within the first area, preliminary results have shown that use of the HRRRE covariance information in the hybrid analysis for the HRRR improves the deterministic forecast from the HRRR system. We are currently working on optimizing this configuration to allow for partial cycling of the deterministic HRRR (as opposed to initializing it with a pre-forecast hour spin-up off of the RAP each hour). There is a tradeoff between HRRR cycling to maximize the benefits of the HRRRE small-scale covariance information and maintaining a close tir to the RAP (and ultimate the GFS) to reduce potential model drift. Additionally, we are working to determine the optimal configuration of the HRRRE assimilation system, again with a desire to optimize the tradeoff between small-scale cycled assimilation and potential drift from the larger-scale systems.

Within the second area (transition to the FV3-based RRFS), we have been working to get a stand-alone regional (SAR) version of the FV3 running with use of RAP initial and lateral boundary conditions. This is being following by work to build the various RAP/HRRR physics modules and data assimilation features into the FV3-based SAR configuration. The physics modules include Thompson microphysics, MYNN PBL scheme, and Sminova LSM scheme, while the data assimilation features include the HRRR radar reflectivity procedure and the special modification for surface observations.

This SAR FV3 work is geared to a planned operation transition to the RRFS in the 2021 -22 time frame. At the conference, we will report on the latest developments in this work, as well as provide an update on the work toward the 2020 RAPv5 / HRRRv4 implementation.