HYSPLIT Meteorological Data

The following are recommendations for choosing the appropriate meteorological data set:

  • If time allows, make an ensemble using all available recent model runsó the last several HRRR or RAP runs, the last NAM (usually the sigma-pressure hybrid files), and over flat terrain, the last GFS.
  • For a quick first guess, start with the model you feel is the best for the situation of the day.
  • Regardless of the model used, consider that anything affecting the model planetary boundary layer (PBL) will affect the HYSPLIT plume trajectory and pollutant concentrations, including surface initial or forecast characteristics that might impact the model surface energy balance, such as:
    • Discrepancies in soil moisture from reality
    • Errors in snow cover/snow depth
    • Errors in the precipitation forecast that might impact the surface conditions and consequently the surface energy balance
    • Existence of and timing of convection in the model versus what is expected
  • Remember that the accuracy of the synoptic scale forecast also impacts the PBL by affecting lower tropospheric winds, cloudiness and its impact on the surface energy balance, PBL and stable layer development, etc.
HRRR (sigma and pressure)
  • Finer spatial resolution than NAM and produced every hour
  • Runs every hour
NAM (pressure) 
  • Coarser vertical resolution (25 mb) than NAM (sigma-pressure hybrid)
  • Runs every 6 hours
NAM (sigma-pressure hybrid)
  • Finer temporal and vertical resolution than NAM (pressure)
  • Runs every 6 hours
GFS 1 deg.
  • Good in areas of flat terrain and if initial conditions and forecast are better than the NAM
  • Terrain is smoothedówinds generally less useful near terrain
GFS 0.5 deg.
  • Finer horizontal and vertical resolution than GFS (pressure)
  • Terrain following coordinate is better for HYSPLIT calculations
  • Better initialization method in and near ongoing precipitation at the initial time
  • Runs hourly

For more details on the meteorological data sets, see hysp_meteoinfo.html.