Scheduled Downtime
On Tuesday 24 October 2023 @ 5pm MT the forums will be in read only mode in preparation for the downtime. On Wednesday 25 October 2023 @ 5am MT, this website will be down for maintenance and expected to return online later in the morning.
Normal Operations
The forums are back online with normal operations. If you notice any issues or errors related to the forums, please reach out to help@ucar.edu

CAM5 MG1.0 liquid and ice number concentrations much smaller than in evaluation paper

I

i_h_h_karset@geo_uio_no

New Member
Hi, Andrew. I can see that this link is old, but I'm also struggling with averaging CDNC (and the radius). I wonder if it's enough to divide by FREQL when finding the mean CDNC over time, lat, lev and lon, or if we also have to some how weight by the cloud fraction. My question is whether FREQL only contains information about how many of the timesteps the grid-box contained liquid (cloud), or if it also accounts for how much of the grid-box that is coved by a cloud. Thank you in advance for your help. RegardsInger Helene
 
A

andrew

Member
Dear Inger,The key is to make sure you are using AWNC or AREL when dividing by FREQL. It will not work on the regular history outputs (NUMLIQ, REL). The values for size should be in-cloud values, so they do not need to be weighted by cloud fraction. FREQL does not contain cloud fraction information, only frequency of occurrence. Water contents are specified as in-cloud (variables with *IC*), and if not they are grid box averaged. But microphysical variables like number or size that do not exist outside of cloud are not grid-box averaged.Hopefully that helps clarify.Regards,Andrew
 
A

andrew

Member
Dear Inger,The key is to make sure you are using AWNC or AREL when dividing by FREQL. It will not work on the regular history outputs (NUMLIQ, REL). The values for size should be in-cloud values, so they do not need to be weighted by cloud fraction. FREQL does not contain cloud fraction information, only frequency of occurrence. Water contents are specified as in-cloud (variables with *IC*), and if not they are grid box averaged. But microphysical variables like number or size that do not exist outside of cloud are not grid-box averaged.Hopefully that helps clarify.Regards,Andrew
 
I

i_h_h_karset@geo_uio_no

New Member
Andrew, thanks for your quick answer. I'm using AWNC and AREL. I'm trying to calculate the mean cdnc and the mean effective radius over a larger region (over both time, lev, lat and lon). This is how I do it now:1) take the mean AWNC over time and lev2) divide 1) by the mean FREQL over time and lev (if FREQL is larger than 0. If not, the answer is set to zero)3) weight the values by latitude using information in gw (so grid-boxes near equator count more than the ones near the pole beacuse of larger area covered)4) take the mean over lat and lon, but only for the nonzero values. The results have reasonable sizes (radius around 11 mu m and cdnc around  41 /cm^3), but I wonder if the contribution from each gridbox should be weighted by how much cloud the gridbox contains (so the cdnc/radius in a grid-box totally coverd with a cloud contributes more to the result than a grid-box with only a small cloud fraction). RegardsInger Helene 
 
I

i_h_h_karset@geo_uio_no

New Member
Andrew, thanks for your quick answer. I'm using AWNC and AREL. I'm trying to calculate the mean cdnc and the mean effective radius over a larger region (over both time, lev, lat and lon). This is how I do it now:1) take the mean AWNC over time and lev2) divide 1) by the mean FREQL over time and lev (if FREQL is larger than 0. If not, the answer is set to zero)3) weight the values by latitude using information in gw (so grid-boxes near equator count more than the ones near the pole beacuse of larger area covered)4) take the mean over lat and lon, but only for the nonzero values. The results have reasonable sizes (radius around 11 mu m and cdnc around  41 /cm^3), but I wonder if the contribution from each gridbox should be weighted by how much cloud the gridbox contains (so the cdnc/radius in a grid-box totally coverd with a cloud contributes more to the result than a grid-box with only a small cloud fraction). RegardsInger Helene 
 
A

andrew

Member
Hi Inger,One problem may be averaging over levels, if that is what you are doing. AWNC and FREQL are layer quantities and were not designed to be averaged to get a column. If you wanted to do that, you would definitately want to divide AWNC/FREQL at each level first. Weighting by latitidue is fine for an average, and taking the average of all non-zero values is appropriate. I would not weight by cloud fraction for a global average if you are interested in cloud microphysics. I would probably think about 'compisiting' by cloud fraction: 0-0.5, 0.5-0.75, 0.75-0.95, 0.95-1 and looking at the results that way. The values do seem reasonable, but a bit small: I think the issue might be vertical averaging.If you want a single number (particularly to compare to satellite data like MODIS), then there is also a 'Cloud Top' Field: ACTNL and FCTL.Hope that helps,
Andrew
 
A

andrew

Member
Hi Inger,One problem may be averaging over levels, if that is what you are doing. AWNC and FREQL are layer quantities and were not designed to be averaged to get a column. If you wanted to do that, you would definitately want to divide AWNC/FREQL at each level first. Weighting by latitidue is fine for an average, and taking the average of all non-zero values is appropriate. I would not weight by cloud fraction for a global average if you are interested in cloud microphysics. I would probably think about 'compisiting' by cloud fraction: 0-0.5, 0.5-0.75, 0.75-0.95, 0.95-1 and looking at the results that way. The values do seem reasonable, but a bit small: I think the issue might be vertical averaging.If you want a single number (particularly to compare to satellite data like MODIS), then there is also a 'Cloud Top' Field: ACTNL and FCTL.Hope that helps,
Andrew
 
You must log in or register to reply here.
Top