The Development and Testing of a Bias Correction Algorithm for Mesonet Wind Observations in 3D-RTMA

NOAA's Environmental Modeling Center (EMC) and Global Systems Laboratory (GSL) are collaborating on a joint project to develop a 3D hybrid ensemble/variational Real-Time Mesoscale Analysis (3D-RTMA) suite with a 15-min update cycle. This new 3D-RTMA suite is intended to replace NOAA’s 2D variational RTMA suite, which produces hourly analyses of sensible weather elements over domains encompassing the National Digital Forecast Database (NDFD) grids. The first 3D-RTMA implementation is currently expected in early 2025.

The quality control (QC) of observations has presented a major developmental challenge for the operational 2D-RTMA suite, owing to heterogeneous observing systems. Thus, improving upon the existing QC of the 2D-RTMA suite is a priority for developing the future 3D-RTMA suite. To achieve this, an automated QC (AutoQC) package is under development for 3D-RTMA, with the goal being to remove poor-quality observations from the analyses more efficiently than in the operational system by reducing the reliance on fixed reject lists. Although initial development efforts have focused on surface observations, the AutoQC package will soon be expanded to include aircraft data. While the AutoQC package will be applied to various fields, such as temperature and moisture, the focus of this presentation will be on efforts to improve the near-surface wind speed and wind gust analyses. Observations of wind speed and wind gust from mesonets are particularly challenging to QC owing to heterogeneities in the station siting (e.g., different anemometer heights, nearby obstructions). Previous work was performed to more accurately account for variations in anemometer height; however, further improvements are possible by more directly accounting for the biases in each individual instrument. Thus, the AutoQC package will include a bias correction algorithm for mesonet wind observations, structured as a Kalman-Bucy filter. This presentation will provide a brief overview of the proposed bias correction algorithm, as well as results comparing the performance of the algorithm to existing QC techniques.