Developing local model output statistics (MOS) equations to predict temperature and probability of precipitation

Model Output Statistics (MOS) guidance is routinely produced from the Nested Grid Model (NGM) and the Aviation (AVN) Models. However, MOS guidance is not currently available from the higher resolution Eta model. A goal of this project was to develop a set of local, seasonal MOS equations for the calculation of temperature and probability of precipitation (POP) at several locations in the Newport/Morehead City (MHX), North Carolina (NC) County Warning Area (CWA) using the Eta model. An additional goal was to produce a system of equations that would be competitive with the NGM and AVN MOS guidance, and to make this information easily accessible to forecasters on a real-time operational basis.

Due to its timely and easy availability to forecasters, the FOUS message FRH62 data was used as predictor variables in local MOS equations. Data have been collected for over a year and entered into a statistical database. Verification was done using the 00 and 12 UTC runs of the Eta model using data from March 15, 1998 to September 15, 1999. The verification data set included approximately 397 cases.

This paper defines the multiple linear regression methods used to derive these localized Eta MOS equations and examines the verification of the localized equations versus the centrally-produced NGM MOS guidance.