Meteorology 5540
Mini-Lecture 5: MOS

I. What is MOS?

  • MOS - Model Output Statistics - a statistical interpretation of output from numerical weather prediction models
  • MOS techniques develop optimal statistical relationships (using multiple linear regression and other techniques) between past model forecasts and "verifying" weather elements (e.g., wind speed, temperature, etc...). For example, statistics could relate model 850 mb temperatures to maximum surface temperatures associated with previously forecast 850 mb temperatures).

    II. What MOS does:

  • Accounts for systematic model bias
  • Accounts for systematic model timing errors
  • Accounts for some regional or local effects
  • Accounts for regional climatic conditions
  • Predicts weather elements that are not explicitly simulated by numerical models
  • Provides reliable probability forecasts

    III. What MOS doesn't do:

  • Doesn't account for random numerical model errors
  • Doesn't handle extreme climatic conditions well (e.g., record highs and lows)
  • Only accounts for some numerical model biases that occur with specific synoptic situations
  • Doesn't account for changes in the modeling system
  • Doesn't account for the full impact of local circulations and orographic effects

    IV. Using MOS for weather forecasting

  • Can be used for time management: Forecaster can focus on "mission critical" forecast issues.
  • Can be used as a forecast "check"
  • Can be used as a first guess
  • Remember that NGM MOS and AVN MOS come from their respective models and may disagree with the picture presented by other models.
  • Forecasters can improve significantly on MOS by identifying rare or extreme events, accurately choosing the model of the day, and considering mesoscale or local effects not accounted for by MOS.
  • Keep in mind that MOS approaches continue to outperform raw model forecasts or "Perfect Prog" techniques that relate model variables to observed weather associated with observed variables. Such techniques thus don't include the effects of systematic timing errors, biases, etc...

    V. References and on-line links

  • Crawford, L., 1996: An analysis of overpredicted high temperatures by NGM MOS guidance at Medford, Oregon. Western Region Technical Attachment No. 95-19.
  • Dallavalle, J. P., 1996: A perspective on the use of Model Output Statistics in objective weather forecasting. Preprints, 15th Conference on Weather Analysis and Forecasting. Norflolk, VA, 479-482.
  • Glahn, H. R., and D. A. Lowry, 1972: The use of model output statistics (MOS) in objective weather forecasting. J. Appl. Meteor., 11, 1203-1211.
  • Jacks, E., J. B. Bower, V. J. Dagostaro, J. P. Dallavalle, M. C. Erickson, and J. C. Su, 1990: New NGM-based MOS guidance for maximum/minimum temperature, probability of precipitation, cloud amount, and surface wind. Wea. and Forecasting, 5, 128-138.
  • Klein, W. H., and F. Lewis, 1970: Computer forecasts of maximum and minimum temperature. J. Appl. Meteor., 9, 350-359.
    Updated May 8, 1997