How to output snow water content for each glacier_mec column?

[wadewei]

Hi all,

I am using CESM1.2.2 with `GLC_MEC` set to 10 levels.

Section 10.2 Multiple elevation class scheme in CLM4.5 Tech Notes says "Each glacier_mec column within a grid cell has distinct ice and snow temperatures, snow water content, surface fluxes, and SMB." However, the default `SNOWICE` output seems to be grid cell avarage. How can I configure the model to output SWE on each mec level?

Best regards,
Wade

 

[oleson]

There is a SNOWICE_ICE variable that outputs snow ice over just ice landunits, however, that will be an average of all ice landunits in the gridcell.
For information about individual columns you'll need to request 1d output in your simulation. So, for example, to generate a secondary 1d monthly history file stream for SNOWICE at the column level:

hist_mfilt(2) = 1
hist_dov2xy(2) = .false.
hist_nhtfrq(2) = 0
hist_type1d_pertape(2) = 'COLS'
hist_fincl2 = 'SNOWICE'

Some explanation of these namelist items can be found here:

CESM2 Series Public Release

www2.cesm.ucar.edu

There will be some supplementary information in that history file to help you map those columns to individual grid cells and to individual GLC_MEC classes:

double cols1d_lon(column) ;
cols1d_lon:long_name = "column longitude" ;
cols1d_lon:units = "degrees_east" ;
double cols1d_lat(column) ;
cols1d_lat:long_name = "column latitude" ;
cols1d_lat:units = "degrees_north" ;
int cols1d_ixy(column) ;
cols1d_ixy:long_name = "2d longitude index of corresponding column" ;
int cols1d_jxy(column) ;
cols1d_jxy:long_name = "2d latitude index of corresponding column" ;
double cols1d_wtgcell(column) ;
cols1d_wtgcell:long_name = "column weight relative to corresponding gridcell" ;
double cols1d_wtlunit(column) ;
cols1d_wtlunit:long_name = "column weight relative to corresponding landunit" ;
int cols1d_itype_col(column) ;
cols1d_itype_col:long_name = "column type (see global attributes)" ;
int cols1d_itype_lunit(column) ;
cols1d_itype_lunit:long_name = "column landunit type (vegetated,urban,lake,wetland,glacier or glacier_mec)" ;
int cols1d_active(column) ;
cols1d_active:long_name = "true => do computations on this column" ;
cols1d_active:flag_values = 0, 1 ;
cols1d_active:flag_meanings = "FALSE TRUE" ;
cols1d_active:valid_range = 0, 1 ;

 

[wadewei]

Thank you Keith! This is exactly what I need.
Could you kindly elaborate more about mapping? I am somewaht confused about the difference of columns and GLC_MEC classes. How do I determine which GLC_MEC class each specific column belongs to?

Best,
Wade

 

[oleson]

The global attributes in the history file can be used, in particular:

ctype_landice_multiple_elevation_classes = "4*100+m, m=1,glcnec"

So I expect you could identify each GLC_MEC class using cols1d_itype_col. The GLC_MEC classes should be 401-410

 

[wadewei]

The global attributes in the history file can be used, in particular:

ctype_landice_multiple_elevation_classes = "4*100+m, m=1,glcnec"

So I expect you could identify each GLC_MEC class using cols1d_itype_col. The GLC_MEC classes should be 401-410

These global attributes does not seem to exist in CLM4's output. Looking at cols1d_itype_lunit, it looks like each grid corresponds to 11 columns, one with cols1d_itype_lunit=1, ten with cols1d_itype_lunit=7, which I assume is GLC_MEC columns. However, I haven't been able to find an index (pointer) to easily associate the columns with respective grids.

Additionally, shall I use SNOWICE+SNOWLIQ as the snow water equivalent (snow+ice+water, i.e. all H2O contentes)?

 

[oleson]

Sorry, I didn't see that you were using cesm1.2.2. My response was based on the supported release version of the model (release-cesm2.1.3). I'm not sure how much I can help. You mentioned you are looking at CLM4 output but are looking at the CLM4.5 technical note. Which version are you actually using?
Regardless, I don't see that cols1d_itype_col is present in that version of the model unfortunately, so I'm not sure how you can get information for individual columns.

 

[wadewei]

Sorry, I didn't see that you were using cesm1.2.2. My response was based on the supported release version of the model (release-cesm2.1.3). I'm not sure how much I can help. You mentioned you are looking at CLM4 output but are looking at the CLM4.5 technical note. Which version are you actually using?
Regardless, I don't see that cols1d_itype_col is present in that version of the model unfortunately, so I'm not sure how you can get information for individual columns.

The history says 4.0 `:source = "Community Land Model CLM4.0" ;`
Although I vaguely remember a certain piece of CESM support document says it's CLM4.5 with 4.0 physics. Are CLM4.5 and CLM4.0 interchangable in CESM1.2.2?

The reason I referred to CLM4.5's technical description is that 4.0's tech notes did not say anything about GLC_MEC class. Which added to my suspicion that the CLM4.0 in CESM1.2.2 is kind of a fusion.

I added the lines to user_nl_clm to request SNOWICE output at column levels as you suggested. Each GLC_MEC grid has 11 columns in the output. Without cols1d_itype_col, there should still be a way of determining information of individual columns. It would be extremely helpful if you can recall anything about it. I've pasted the "ncdump -h" output. Otherwise, I'll try to comb through the code.

Also, shall I use SNOWICE+SNOWLIQ as the snow water equivalent (snow+ice+water, i.e. all H2O contentes), or SNOWICE alone?

 

[oleson]

I believe you can run either CLM4 or CLM4.5 from cesm1.2.2, it is specified by the compset you are using. What compset and resolution are you using, i.e., what is your create_newcase command?