Dear Aseem,

You might want to read the publication by Hock (2003): http://www.sciencedirect.com/science/article/pii/S0022169403002579

Hock discusses the use of temperature for snow(melt) modelling

Some modelers feel that it is usually safe to assume that precipitation is falling as snow when the temperature is below 2 degrees C.

You can reasonably assume that snow can reach the surface some degrees above 0° (suggestion of +1 to +3°C seems acceptable). If you want to make an automatic treatment from daily min and max temperatures, I advise to reconstruct a diurnal cycle of the surface temperature and to consider that in case of rain, you get snow as long as the threshold of 3°C (above 0) is overpassed.

Daily mean temperature based thresholds could be developed. The thresholds would be dependent on physical geographical characteristics, and it may be good to forge them differently for varied climatic zones (types).

I can used numerical model. For example WRF.