GEOS-Chem nested grid simulations

From Geos-chem
Jump to: navigation, search

This page contains some basic information pertaining to the GEOS-Chem nested grid simulations. We also invite you to visit the GEOS-Chem Nested Model wiki page for information current projects.

Setting up a nested-grid simulation

Please see the following wiki pages:

Frequently asked questions

What is involved in running a nested grid simulation?

Running a nested grid simulation in GEOS-Chem requires the additional step of first running a coarse resolution GEOS-Chem simulation to save out boundary conditions that will be used to initialize species concentrations at the boundaries of your nested grid region.

Is it possible to run nested GEOS-Chem simulations on the AWS cloud?

Yes, you can run the nested grid simulations on AWS cloud. Please see the Running GEOS-Chem on AWS cloud online tutorial and contact the GEOS-Chem Support Team with any questions.

If we perform a nested grid simulation, is it possible to save out output in netcdf format?

Yes, you can output netcdf diagnostics for nested grid runs starting in GEOS-Chem 12 by modifying the settings in HISTORY.rc.

What nested grid domains are supported?

You can run GEOS-Chem nested grid simulations for the following regions:

  1. 0.25° x 0.3125° China region (GEOS-FP only)
  2. 0.25° x 0.3125° North America region (GEOS-FP only)
  3. 0.25° x 0.3125° European region(GEOS-FP only)
  4. 0.5° x 0.625° Asia region (GEOS-FP and MERRA-2)
  5. 0.5° x 0.625° North America region (GEOS-FP and MERRA-2)
  6. 0.5° x 0.625° European region(GEOS-FP and MERRA-2)

Starting in GEOS-Chem v10-01 you do not need to use separate emissions files with resolutions matching those of the your nested grid. The HEMCO emissions component can read the global emissions data and regrid/crop them on-the-fly to the size of the nested grid.

Starting in GEOS-Chem 12.4.0, you can run GEOS-Chem on any nested-grid region by specifying the boundaries in the Grid Menu of input.geos. The high-resolution global meteorology fields are read in and regridded/cropped by HEMCO to the size of your nested grid. If you are specifying a smaller domain within one of the predefined nested domains listed above, then you can provide the high-resolution met fields for the regional domain instead to avoid having to read in the larger met files. You may also use a coarse resolution restart file (e.g. one from the global simulation used to save out your boundary conditions) to initialize your nested grid simulation. The restart files are also read in and regridded/cropped by HEMCO. As always, we recommend spinning up your GEOS-Chem to remove the impact of the initial conditions on your results.

Can you save out boundary conditions for more than one nested grid in the same global run?

Yes, a single global simulation can output boundary conditions for multiple nested grid regions. In GEOS-Chem 12.3.2 and earlier, simply select multiple regions you would like to output boundary conditions for within the Nested Grid Menu of input.geos. In GEOS-Chem 12.4.0 and later, you can choose to either save out global boundary conditions that can be used for any region or you can create additional BoundaryCondition collections in HISTORY.rc.

How can I find which data are available for nested grid simulations?

Nested grid meteorolgy fields may be downloaded from Harvard, Dalhousie, or Compute Canada. Please see our Downloading GEOS-Chem data directories wiki page. Starting in GEOS-Chem v10-01 you do not need to use separate emissions files with resolutions matching those of the your nested grid. The HEMCO emissions component can read the standard recommended emissions data and regrid/crop them on-the-fly to the size of the nested grid. You should ensure you have the necessary HEMCO data directories on your system.

What if I can't find the meteorology data that I need?

If you cannot find the nested grid meteorology data for your particular geographic region of interest or time period of interest, then you have a few options.

  1. If you are using GEOS-Chem 12.4.0 or later, then you can use the high-resolution global met fields as input in your nested grid simulations and HEMCO will automatically regrid/crop them to your region. If you want to reduce the file size on your system, you may also choose to download the high-resolution global met fields and crop them to your region of interest before using in GEOS-Chem.
  2. If you are using GEOS-Chem 12.3.2 or earlier, then you may be required to download and process the met data yourself. Please see our

Nested grid data protocol document for more information. It is always a good idea to contact the GEOS-Chem Nested Model email list before processing met fields for a new nested grid domain to ensure no one has done this work already.

Where can I find out more info about nested grid errors?

Please see our Guide to GEOS-Chem error messages for more information.

I noticed abnormal concentrations at boundaries of the nested region. Is that normal?

If you see high tracer concentrations right at the boundary of your nested grid region, then this may be normal.

For nested grid simulations, we have to leave a “buffer zone” (i.e. typically 3 boxes along each boundary) in which the TPCORE advection is not applied. However, all other operations (chemistry, wetdep, drydep, convection, PBL mixing) will be applied. Therefore, in the “buffer zone”, the concentrations will not be realistic because the advection is not allowed to transport the tracer out of these boxes.

In any case, the tracer concentrations in the “buffer zone” will get overwritten by the 2° x 2.5° or 4° x 5° boundary conditions at the specified time (usually every 3h).

Therefore, you should exclude the boxes in the “buffer zone” from your scientific analysis.

The following diagram illustrates this:

 <-------------------------------------- IGLOB ---------------------->

 +-------------------------------------------------------------------+   ^
 | GLOBAL REGION                                                     |   |
 |                                                                   |   |
 |                       <-------------- IIPAR ------------->        |   |
 |                                                                   |   |
 |                       +=================================[Y]  ^    |   |
 |                       |  WINDOW REGION (met field size)  |   |    |   |
 |                       |                                  |   |    |   |
 |                       |      <------- IM_W ------->      |   |    |   |
 |                       |      +--------------------+  ^   |   |    |   |
 |                       |      |  TPCORE REGION     |  |   |   |    |   |
 |                       |      |  (advection is     |  |   |   |    |   |
 |<------- I0 ---------->|<---->|   done in this     | JM_W | JJPAR  | JGLOB
 |                       | I0_W |   window!!!)       |  |   |   |    |   |
 |                       |      |                    |  |   |   |    |   |
 |                       |      +--------------------+  V   |   |    |   |
 |                       |        ^                         |   |    |   |
 |                       |        | J0_W                    |   |    |   |
 |                       |        V                         |   |    |   |
 |                      [X]=================================+   V    |   |
 |                                ^                                  |   |
 |                                | J0                               |   |
 |                                V                                  |   |
[1]------------------------------------------------------------------+   V

Diagram notes:

  1. The outermost box (GLOBAL REGION) is the global grid size. This region has IGLOB boxes in longitude and JGLOB boxes in latitude. The origin of the "Global Region" is at the south pole, at the lower left-hand corner (point [1]).
  2. The next innermost box (WINDOW REGION) is the nested-grid window. This region has IIPAR boxes in longitude and JJPAR boxes in latitude. This is the size of the trimmed met fields that will be used for a "nested-grid" simulation.
  3. The innermost region (TPCORE REGION) is the actual area in which TPCORE advection will be performed. Note that this region is smaller than the WINDOW REGION. It is set up this way since a cushion of grid boxes is needed TPCORE Region for boundary conditions.
  4. I0 is the longitude offset (# of boxes) and J0 is the latitude offset (# of boxes) which translate between the "Global Region" and the "Window Region".
  5. I0_W is the longitude offset (# of boxes), and J0_W is the latitude offset (# of boxes) which translate between the WINDOW REGION and the TPCORE REGION. These define the thickness of the Buffer zone mentioned above.
  6. The lower left-hand corner of the WINDOW REGION (point [X]) has longitude and latitude indices (I1_W, J1_W). Similarly, the upper right-hand corner (point [Y]) has longitude and latitude indices (I2_W, J2_W).
  7. Note that if I0=0, J0=0, I0_W=0, J0_W=0, IIPAR=IGLOB, JJPAR=JGLOB specifies a global simulation. In this case the WINDOW REGION totally coincides with the "Global Region".
  8. In order for the nested-grid to work we must save out concentrations over the WINDOW REGION from a coarse model (e.g. 2° x 2.5° or 4° x 5°) corresponding to the same fine-resolution WINDOW REGION. These concentrations are copied along the edges of the 1x1 WINDOW REGION and are thus used as boundary conditions for TPCORE.

What should be the resolution of the global boundary conditions data?

Please see this section.