Does the 'Showers' variable always represent rain?

[colinnuk]

Question about the showers variable - does it always represent rain rather than snow? Or could showers be either rain or snow, depending on temperature? I want to add showers into either 'rain' or 'snow' so I get the total precip in either form.

I've had a look thru the source code quickly - as far as I can tell you are normally able to source 'showers' directly from the model, so I'm not sure exactly what this represents. Perhaps it differs depending on the model?

[patrick-zippenfenig]

Technically showers could be snow, but convective snow is quite rare. Most models do not even separate between showers and rain in the first place. Also, snowfall data is not very reliable. I just spend 3 days to correct snowfall amount for the ICON weather models which forecast snowfall at 8°C while they even show a snowfall height of 800 metre above ground.

[colinnuk]

OK thanks for replying! Convective snow isn't rare in the mountains in western Canada - particularly in spring.

I think elsewhere you'd mentioned you were thinking of changing some parameters to split rain & snow by temperature rather than taking it from each model - I guess that sort of system would put the precip into the most 'likely' bucket.

[patrick-zippenfenig]

Yes, I noticed that some weather models predict snow at >8°C. This is quite unlikely. I added a post-processing step that converts snow back to rain if temperature is above 1.5°C.

Canada is covered by the HRRR model as well as HRDPS. At least HRRR is a convection allowing model and does not separate between large scale and convective precipitation. HRRR does not contain a showers field in its model output -> showers will always be 0.