Smoke v4.6 inputfiles, GRIDMASK file and Layalloc

SMOKE Emissions Inventory
1

I am running smoke v4.6 for emission to be used for research work.

  1. SMOKE 4.6 is not reading the CAMS inventory, it says it cannot be identified
  2. In the arinv.edgar.lst in the inventory directory, how do I list CAMS inventory details so that SMOKE can read them simultaneously. An example is included:

#LIST GRID

#SCC,InvPollName,EDGAR_VarName,Month,FileName

#DESC Sample test case for TRANSPORT EDGAR sector based on HTAP version 2.0

TRANSPORT,CO, ,0,/xxxx/xxx/xxx/xxxx/SMOKE_4.6/data/inventory/nei2005/edgar/CAMS-GLOB-ANT_Glb_0.1x0.1_anthro_co_v2.1.nc_2017.nc4

The CAMS inventory has several VarName which are present in the CAMS inventory are sum, res, shp, slv, swd,tnr,tro, mma,ind, fef,ene, agr - which one should be used or how can they all be fitted into the file arinv.edgar.lst

In addition to this, I have now modified the emissions inventory to only contain the emissions from only one sector but yet smoke doesn’t identify the file or the variable in the file.

3. In the addition to the above question, I made a smoke run using my own GRIDDESC file content for Africa and using Edgar emissions inventory for 2010 with the GRIDMASK file provided for Edgar in smoke v4.6. it runs fine, but when I visualise the results using Verdi , No emissions for Africa on the map are shown. I have specified the grid name under IOAPI_GRIDNAME_1 and done everything else as required for the tutorials and in https://docs.google.com/document/d/1veqEjaTPbDpkqHAIcyReokvv6MaENz0TrwOcFl1ewac/edit#

Do you have any ideals as to why it doesn’t show the emissions on the map for Africa?

  1. Can you kindly provide the run script for Layalloc as this is not distributed with smoke v4.6.

Looking forward to your email - you help its greatly appreciated.

Many thanks John

2

Please see my responses by number:

1 & 2) Although CAMS gridded inventory is in same netCDF format with the same map projection/domain, unfortunately it holds more than a single variable in each file. EDGAR inventory only holds one pollutant per file. I think that is why Smkinven failed to run.

  1. What kind of results are you referring? Are you visualizing the final regridded/speciated hourly emissions in IOAPI/NetCDF fromat? If you tell me what file you are trying to visualize, then I can tell whether it will work with Verdi or not.

  2. I do not have any sample Layalloc run script but it can be created by following the information from SMOKE user’s guide.
    https://www.cmascenter.org/smoke/documentation/4.7/html/ch05s03s09.html

3

Thank you Baek for the response, the problem with the input files has now been resolved.

I did a smoke run using the GRIDMASK file for edgar emissions to process an emission inventory for CAMS which has the same grid size and apparent lat, lon, projection coordinate. When visualization the output file in verdi which is the (.ncf file format file) the emissions seem to have been mapped in the wrong location on the output file. There seem to be have been a shift in the location of the emissions.
Then when I do the same run using the same GRIDDESC information and the HTAP emissions inventory the output file shows the emissions to be mapped into the correct locations.

Do I need a new GRIDMASK file for CAMS emissions inventory or is the Edgar GRIDMASK file meant to be sufficient/adequate for this as both have the same grid resolution and lat, lon coordinate of 90 to -90 and -180 to180?

Thanks

4

John, I am not so familiar with CAMS data file but it probably does not contain the map projection and domain decription (lower-left-corner coordinates, and so on) for Verdi to plot to the right map. Please check out the Verdi user’s guide for the detail on how to visualize the native NetCDF file.

Also, @lizadams can assist you for any further information.