Motivation for the Operational Models Encyclopedia

The availability of numerical guidance from NWP models has been an important component of operational forecasting for decades. For many, the output from this numerical guidance was produced by a mysterious “black box”. Rules for using and adjusting the guidance for operational forecasters were often subjective “Rules of Thumb” based on experience rather than based on quantitative analysis.


To open up this “black box”, we produced this web-based “Operational Model Encyclopedia” linking both generic information on how NWP models work, and specifics on physical parameterizations, dynamics, and data assimilation in operational models. Ensemble Prediction systems and Marine Wave models are included as well.  New model descriptions are updated as operational models are updated.

  • "Deterministic models" refers to NWP model systems with individual, single-valued forecasts. No uncertainty is accounted for from either errors in the forecast starting point (i.e. initial conditions) or from NWP model approximations made in producing the forecast. 
  • "Probabilistic models", also known as Ensemble Forecast or Ensemble Prediction Systems (EFS or EPS), take into account uncertainty in initial conditions and/or  model estimates. In doing so, they generate forecast probabilities for specific events, such as the exceedance of a temperature or precipitation threshold. Probabilistic models make multiple forecasts, each starting from a unique initial condition, using a unique NWP model configuration, and/or randomly varying the change in forecast variables at fixed time intervals. 
  • Finally, "Wave Models" predict the wave spectrum of water waves, which are then broken down into wave height and period at grid points within the wave model domain. Wave models are forced by NWP model winds, and can be run in either deterministic or probabilistic modes.

To view model-specific pages, use the sidebar to navigate to your desired resources.