hirlam

Hi,

I am inspecting hydrometeor output from a Harmonie run. The GRIB file
tells me that graupel, cloud ice, precipitation (rain) etc. are written
in units of kg/m2. I expected mixing ratios (kg/kg). How should I
interpret these fields if I want to determine the water content (kg/m3) ?

Best regards,

John.

Kristian Pagh Nielsen answers:


I have only done this for cloud water. In general, converting loads in a
grid box in kg/m² to concentrations in kg/m³ is done by dividing with
the thickness of the grid box in question. This you get from the
geopotential heights of the surrounding levels, which are calculated in
the hybrid coordinate scheme.

Best regards,

Kristian.

Thank you for the information.

Thank you Kristian and Jelena,

When you talk about dividing with the thickness of the grid box do you mean the distance between two full levels or between two half levels directly above and below ?

Best regards,

John.

For cloud water, I used the full level layer thickness, i.e. the thickness between the half levels.

Dear Kristian,

when looking at the output of the model and calculating the water load in the way that you describe (by taking the output from the model state file and dividing it by the depth of the layer) I get much lower values for e.g. graupel than when I take the values to be given in kg/kg.

When using the contents of the grib file as kg/kg I also get the signals I would expect from the model for e.g. the chance of summer hail. The last thing that, for me at least, points in the direction of the values being kg/kg is that when I calculate the surface precipitation intensity from the hydrometeor values at the lowest model level (with an assumed fall speed) they correspond much better to the values found in the grib file for the precipitation intensity (e.g. 61 105 456 for the intensity of rain) than when I should divide by the depth of the layer.

Therefore I would like to ask you if you are 100% sure that the hydrometeors given in the grib files are water load in kg/m^2 or if it is the mixing ratio that is given in the files (like is normal in e.g. the HIRLAM output files for cloud water and cloud ice).

Dear Sander,

Sorry if I caused a misunderstanding; I did not make any comments on the units of the hydrometeors in the output GRIB files. I do not know what these are. My comment was only on how to convert grid box loads to grid box concentrations in general. Of course, if the output is not loads in [kg/m²] but mixing ratios in [kg/kg] this approach will give the wrong results, as you write.

Best regards,

Kristian.