Overview

The Weather Research and Forecasting (WRF) model is a state-of-the-art numerical weather prediction system that is highly configurable and suitable for a broad range of weather applications. Given the numerous options available, it is important to rigorously test configurations to assess the performance of select configurations for specific applications. To assess the performance of the new schemes, the Developmental Testbed Center (DTC) peformed testing and evaluation with the Advanced Research WRF (ARW) dynamic core (Skamarock et al. 2008) over several physics suites, at the request of the sponsor, the Air Force Weather Agency (AFWA). For each test conducted by the DTC, two configurations were run over a Contiguous United States (CONUS) domain, spanning a full year. One configuration was based on AFWA's operational configuration and is considered the baseline; the comparison configuration, substitutes a namelist option or input data set.

WRF v3.1.1: Comprehensive testing was performed using WRF v3.1.1 in late 2009 to evaluate the Quasi-Normal Scale Elimination (QNSE; Sukoriansky et al. 2005) plantary boundary layer (PBL) and surface layer schemes.

WRF v3.2.1: Same testing setup as WRF v3.1.1 listed above, but upgraded the WRF system to version 3.2.1; this testing was performed in late 2010.

WRF v3.3.1: Comprehensive testing was performed using WRF v3.3.1 in late 2010 to evaluate the updated Rapid Radiative Transfer Model (RRTMG; Iacono et al. 2008) long-wave and short-wave radiation schemes.

WRF v3.4: Comprehensive testing with WRF v3.4 was conducted in late 2012 in a functionally similar environment to AFWA's operations, including a 6-hour "warm-start" spin up and data assimilation. Impacts of initializing WRF with output from two different versions of AFWA's Land Information System (LIS) were assessed.

WRF v3.5.1: Comprehensive testing with WRF v3.5.1 was conducted in late 2013 in a functionally similar environment to AFWA's operations, including a 6-hour "warm-start" spin up and data assimilation; performance of the Noah land surface model with multi-parameterization options (Noah-MP; Niu et al. 2011) was evaluated.

WRF v3.6.1: Comprehensive testing with WRF v3.6.1 was conducted in late 2014 in a functionally similar environment to AFWA's operations, including a 6-hour "warm-start" spin up and data assimilation; performance of the Asymmetric Convective Model (ACM2) with non-local upward mixing and local downward mixing planetary boundary layer (PBL) scheme (Pleim, 2007), the Pleim-Xiu land surface model (LSM) two-layer scheme with vegetation and sub-grid tiling, and the Pleim-Xiu surface layer scheme (Xiu and Pleim 2001) were evaluated.

WRF v3.8.1: Comprehensive testing with WRF v3.8.1 was conducted in early 2017 to evaluate several upgrades to the v3.5.1 version previously tested using v3.8.1. A few of the major physics parameterization changes included in the testing were changes to the microphysics (Thompson-Eidhammer aerosolo-aware) and radiation (RRTMG). Other modifications (e.g., time step, eta_levels) are described in the model configuration sepcifics.