Intelligent Use of Model-Derived Products - version 2

Intelligent Use of Model-Derived Products - version 2

This module, part of the "NWP Training Series: Effective Use of NWP in the Forecast Process", discusses three aspects of forecast guidance developed from raw NWP model data:

  1. Post-processing
  2. Statistical guidance
  3. Model assessment tools

Post-processing methods, including a new section of downscaling of coarser resolution data, bias correction, and post-processing of ensemble forecast system data, are introduced. Interpolation of raw model data to produce the data seen by operational meteorologists is also described.

Next, we present information on statistical guidance methods and techniques, including perfect-prog and Model Output Statistics (MOS). Strengths and limitations of each technique are described.

Finally, we present model assessment tools for verification of NWP model data. The effects of aggregating the data over space and time are discussed, including

  1. Point verification versus area verification
  2. Short-term versus long-term verification
  3. The effect of analysis methods on verification scores
  4. Statistics used in verification

and more.

Back in 2000, the subject matter expert for this module was Dr. Ralph Petersen of the National Centers for Environmental Prediction, Environmental Modeling Center (NCEP/EMC). Revisions to the module were made in 2009 by Drs. Bill Bua and Stephen Jascourt, from the NWP team at UCAR/COMET.