Making Sensible Weather from NWP Output: An Operational Perspective of MOS (Invited Presentation)

For more than four decades model output statistics (MOS) have been an invaluable tool for forecasters. This talk will highlight how, through statistical methods, MOS objective guidance has provided operational meteorologists with a bridge between raw NWP output and the production of forecasts of sensible weather elements for a wide range of applications. In addition to basic surface variables such as temperature, dew point and winds, MOS guidance has been a vital source for forecasters' efforts in preparing public and aviation forecasts of sky conditions and precipitation occurrence, including precipitation type and visibilities. The ability to account for much of the systematic biases within operational modeling systems has been a valued strength of the MOS approach. The quality of MOS forecasts has long been sufficiently high to provide a historical baseline for comparison to operational forecasts. Improvements over MOS guidance have become a standard measure for assessing the quality of human-produced forecasts.

MOS has also provided operational forecasters with objective probabilistic forecast guidance for many sensible weather elements. The robust statistical calibration of the MOS system has also provided forecasters with a hands-on training tool to better understand and formulate reliable probabilistic forecast information.

Originally derived for specific locations, MOS has expanded in recent years to provide gridded objective guidance to support new digital forecast initiatives. Early efforts to utilize MOS forecasts to generate computer worded forecasts also provided a foundation for an evolution of the forecast process to a modernized database of weather information.