Benoit, Mark David (2016-12). Sensitivity of High-resolution WRF Forecasts to a Single Radiosonde in a Data Sparse Region. Master's Thesis. Thesis uri icon

abstract

  • Radiosonde observations (RAOBs) are relatively sparse over central Texas. The closest RAOB launch site to College Station, Texas is in Fort Worth, Texas, about 250 km away. On-demand soundings were launched by Texas A&M University students in high-impact situations. Both the local NWS offices and the SPC requested RAOBs. These observations had value to forecasters in convective and winter weather situations. The purpose of this research was to find the value of on-demand RAOBs on a high-resolution NWP forecast for College Station. DA was done with 29 RAOBs using WRF, but also incorporated other observations from MADIS. In total, there were 116 simulations since four simulations were done for each RAOB. Using weather model analyses and observations from the EOL, MADIS, and NOAA, the simulations were evaluated. DTC-MET and SPoRT-MET tools utilized the datasets to provide verification. In some cases, DA of a single RAOB produced modest, positive impacts on the WRF forecast. Precipitation characterization and high precipitation amounts were improved in convective cases, while surface and low-level temperature forecast improvements were seen in short-range forecasts for winter cases. Benefit was spatially limited to areas near College Station, and was further limited when MADIS observations were assimilated. Future work supports RAOB launches in high-impact situations; however, real-time DA of these RAOBs is not a high priority.
  • Radiosonde observations (RAOBs) are relatively sparse over central Texas. The closest RAOB launch site to College Station, Texas is in Fort Worth, Texas, about 250 km away. On-demand soundings were launched by Texas A&M University students in high-impact situations. Both the local NWS offices and the SPC requested RAOBs. These observations had value to forecasters in convective and winter weather situations.

    The purpose of this research was to find the value of on-demand RAOBs on a high-resolution NWP forecast for College Station. DA was done with 29 RAOBs using WRF, but also incorporated other observations from MADIS. In total, there were 116 simulations since four simulations were done for each RAOB. Using weather model analyses and observations from the EOL, MADIS, and NOAA, the simulations were evaluated. DTC-MET and SPoRT-MET tools utilized the datasets to provide verification.

    In some cases, DA of a single RAOB produced modest, positive impacts on the WRF forecast. Precipitation characterization and high precipitation amounts were improved in convective cases, while surface and low-level temperature forecast improvements were seen in short-range forecasts for winter cases. Benefit was spatially limited to areas near College Station, and was further limited when MADIS observations were assimilated. Future work supports RAOB launches in high-impact situations; however, real-time DA of these RAOBs is not a high priority.

publication date

  • December 2016