The topography of central Israel is too complex for operational models to handle surface temps well. The ECMWF is better than the GFS, however, due to its superior resolution.
Also, note that because Jerusalem's elevation is around 800 m, it is normally around the 925 mb level in winter storms, so temperatures below that level are not relevant. In this particular storm, because pressures will be low initially, Jerusalem will be around the 910-915 mb level for the first half of the storm and close to 900 mb for the second half.
Also, the GFS model has such a coarse resolution that it "thinks" Jerusalem is at a much lower elevation than it is. The elevation increases from sea level to over 800 m in about 35 miles from the coast to the central mountains of Israel.
Having researched past winter storms in Israel and spoken to those who are experts, the most important factors are the 850, 700 and 500 mb temps. Generally in order to get accumulating snow:
850 mb temps: below -1C, and ideally below -2C
700 mb temps: below -10C
500 mb temps: below -25C
Current GFS forecasts indicate 850 mb temps around -1.5C for Sat and early Sun, 700 mb temps around -11.5 to -12C the entire storm, and 500 mb temps of -24 to -25C for Sat, but rising to around -21 to -22C by Sun morning. The 500 mb temps are borderline, but the other indicators are good for snowfall.
The other factor to consider is that the GFS doesn't have the sounding completely saturated at low levels during the time when temps are cold enough for possible snow - Sat into early Sun. It has 900 mb RH around 75 to 85% during that time, implying that temps could drop significantly via evaporational cooling if heavier precip occurs.
donsutherland1, on 14 February 2012 - 10:29 AM, said:
The forecast soundings show too much warm air at 925 mb-950 mb. Surface readings are mainly in the upper 30s or lower 40s (GFS and ECWMF).