FIM Performance For Some of The Major Events of The 2009-2010 Winter Season

Abstract

At the last Weather and Forecasting/Numerical Weather Prediction Conference in Omaha in 2009 we showed some examples of forecasts from a new global model known as the FIM, for Flow-following finite-volume Icosahedral Model, that was developed at the Global Systems Division (GSD) of ESRL. In addition to the icosahedral horizontal grid, other differences between the FIM and current operational global models such as the Global Forecast System (GFS) and the European Centre for Medium-Range Weather Forecast model (ECMWF) include an adaptive isentropic-sigma hybrid vertical coordinate that is used in the FIM. At this earlier conference we demonstrated that FIM forecasts compared favorably to those from the GFS and ECMWF for some challenging “dropout” cases, and suggested that the performance was consistent with a potential future role as part of the North American Ensemble Forecast System (NAEFS). Since the last conference there have been a number of improvements to the FIM (these will be detailed in an accompanying paper), and for the conference in January we will show more recent forecast examples from the FIM for some of the challenging storms of last winter. Where possible we will also examine cases for the early part of the upcoming winter, since additional changes have been made since last winter. We will demonstrate that the FIM forecasts, initialized from the identical initial conditions as the operational GFS, can be similar to the GFS or ECMWF, but by 120 hours into the forecasts do not systematically resemble a particular model. In some instances the FIM forecast will be different from either model, but well within the scope of reasonable solutions. Both characteristics of the FIM appear to make it a potential useful addition to a global forecast ensemble. In addition, there are several variants of the FIM that are being tested at GSD, and we will compare the forecasts from these runs as well at the conference.