For using sensitivity bc file for nested runs

Hello,I’m using CMAQ_5.2-DDM-3D.When I want to run the run_cctm.csh for using sensitivity bc file for nested runs ,I set as follows:
setenv DDM3D_BCS T # use sensitivity bc file for nested runs [ T | Y | F | N ] (default is N/F)
set S_BCpath =
set S_BCfile =

But,I don not know how to set “S_BCfile”. Would you please help me with the problem?
Thanks for your time

Hello,

You first need to have sensitivity boundary conditions available. Typically, you can create them from a larger domain you have already simulated. Then, you specify them just like any other CMAQ input file, by typing in the file path and file name in place of the blanks in your script. You do need to make sure that the larger domain providing the boundaries for sensitivity has exactly the same sensitivity parameters as the smaller domain reading them.

Normally, this option is rarely used, and the sensitivity boundary conditions are assumed to be zero.

Sergey,
How do I need to setup if I have three nested domains (e.g. 36-km, 12-km, and 4-km)? I would like to get sensitivity BC files for second (12-km) and third domain (4-km) because I am studying how emission reductions of NOx and VOC in both coarse domains (36 and 12-km) influence surface ozone in the target domain (4-km). How do we generate initial and boundary conditions for 12-km (4-km) domain with HDDM run or sensitivity coefficients from 36-km (12-km) domain?
Thank you very much.
Feng

Feng,

You have to run the domain 1 at a time from largest to smallest. First, you will run the 36km, and when it completes you will extract the sensitivity boundaries for the 12km with the same tools you use to extract concentration boundaries. Second, you will run the 12km and extract the boundaries for the 4km. Finally, will run the 4km.

Hello Sergey,
Thank you very much for the reply. The key point is “extract the sensitivity boundaries for the 12km with the same tools you use to extract concentration boundaries.” You mean we can straightforward use CMAQ/PREP/bcon/scrip/run_bcon.csh. The input files is on-hour concentrations CCTM_CONC_, output file from 36-km. I am wondering how this implementation can realize sensitivity boundaries because I could not find any information about sensitivity from the input file "CCTM_CONC_". The boundaries are completely the same as CMAQ runs without DDM option according to my understanding.
On the other hand, do we need to keep all species and all layers in output files CCTM_CONC_* to this purpose because we need to keep the completed information in boundaries for the finer domain? I mention this because there is an option in run_cctm.csh to chose species and layers we interested in for On-Hour Concentrations.
The sensitivity of NOx and VOC emissions in coarse domain (36km) to ozone concentration in finer domain (12-km) should be reflected by boundary conditions for the nested domain (12-km) as my understanding, however, I could not get idea how to realize it technically in CMAQ with HDDM.
Thank you very much for your help.
Feng

Feng,

Yes, you can use the run_bcon.csh script. It doesn’t care what file you point it to. It should extract the boundary of whatever your specify. So, you would specify CCTM_ASENS* to it. If its not clear, you need to run this script once for concentrations and a second time for sensitivities.

Yes, you need to write all of the sensitivity species in order to use them for boundaries. It will has the potential to be a huge file depending on what you are doing.

Note, that this functionality in DDM was submitted by a user and is not invoked very often. There may be issues that I am not aware of. I personally do not use it, since you can just calculate sensitivity to the CONC boundaries in the daughter domains and that is often enough for most applications.

If you run into something strange, I will try to check here periodically.

Sergey

Sergey,
Thank you very much for taking the time to reply. I am sorry for my late response.

We need to run run_bcon.csh script twice according to your message using the following inputs, respectively:

  1. setenv CTM_CONC_1 CCTM_CONC.YYYYMMDD.nc (CMAQ instantaneous concentration outputs for coarse domain)
  2. setenv CTM_CONC_1 CCTM_ASENS.YYYYMMDD.nc (CMAQ sensitivity coefficients outputs for coarse domain)

We do not have any problems to understand the first run. Bu I doubt if the second one works because run_bcon.csh script need to define chemistry mechanism first and then to pick up defined chemical species from CMAQ instantaneous concentration files. However, the variables in CCTM_ASENS.YYYYMMDD.nc are first order, second order, and cross-term sensitivity coefficients (e.g. of O3 to NO2: O3_1stNO2, O3_2ndNO2, O3_csNO2). They are not defined by the chemistry, they are not instantaneous concentrations, they are sensitivity coefficients even though they have the same concentration unit in ppm or ppb.
As my understanding of how to create sensitivity boundaries, we conduct the first run as described above and obtain original boundary files, and then update or re-calculate all species concentrations in the boundary files using sensitivity coefficients according to Eq.(1) as below. The results should be the the sensitivity boundary for down scaling finer domain. Am I right? Thank you for review and reply.

image

You are correct that BCON versions prior to CMAQv5.3 would likely have challenges creating boundary conditions from a CCTM_ASENS file specified as CTM_CONC_1 because these earlier versions used mechanism-specific GC, AE, NR, and TR namelist files to control the species in the output file and none of the species in your CCTM_ASENS file would be contained in the standard namelist files distributed with CMAQ.

However, starting with CMAQv5.3, BCON now simply writes whatever species it finds in the input file to the output file, and you no longer specify a mechanism or any namelist files in the run_bcon.csh run script. Therefore, you should be able to use BCON from CMAQv5.3 to implement the approach described by Sergey as long as the CCTM_ASENS file you specify as input to BCON contains the same number of layers and time steps as your CCTM_CONC concentration file. Please see Chapter 4 of the User Guide and the BCON README file for further information on running BCON in CMAQv5.3

Christian,
Thank you for your reply. I think you may be right for CMAQv5.3/BCON having the ability to process CCTM_ASENS file, but the point is we are discussing CMAQ/HDDM for nested domains. I do not think CMAQv5.3 with HDDM is available now. I am really looking forward to the new version of CMAQ with HDDM. We have to go back to CMAQv5.2 if we want to use HDDM. Am I right?
Thanks.
Fliu

Hi Fliu,

you are correct about using CMAQv5.2 for DDM as it is not yet implemented in CMAQv5.3. What I am suggesting is to use BCON from CMAQv5.3 instead of BCON from CMAQv5.2 to process your CMAQv5.2 DDM output files for nested domains.

Christian

Christian,
It should work using BCON from CMAQv5.3 to process output from CMAQv5.2, however CMAQv5.2/HDDM does not work with the outputs from BCON of CMAQv5.3 because of chemistry inconsistency. You may take time to try. Thank you.
FLiu

Christian,
In addition, the purpose for us to use HDDM boundaries is to reflect the modification of boundary due to emission reduction of some key ozone precursors like NOx and VOC in coarse domains. Straightforward processing boundaries using BCON could not realize that.
Thank you very much for your time to reply and check.
Fliu

Fliu,

Christian is suggesting that you use the utility distributed with CMAQv5.3 (BCON) to process output generated by CMAQv5.2. Then, use that BCON output to run the nested domain using CMAQv5.2.

Alternatively, there is an iopai utility you can look up that generates bcon files from ioapi files. Its called bcwndw.

https://www.cmascenter.org/ioapi/documentation/all_versions/html/BCWNDW.html

Sergey and Christian,
Thank you both very much for helps. Let me take some time to take your suggestions a step further. I will continue to keep you updated.
BTW, when can we have CMAQv5.3.x with HDDM?
Thanks.
Feng

Feng,

We are finishing the ISAM release first, which is very close to being done. The CMAQv5.3 DDM is complete for gas-phase species. PM may take a bit of time, because we are exploring alternate thermodynamic solutions. Either case, I can’t give you an estimate since I haven’t had a chance to work on it for a while now.

Sergey

It should work using BCON from CMAQv5.3 to process output from CMAQv5.2, however CMAQv5.2/HDDM does not work with the outputs from BCON of CMAQv5.3 because of chemistry inconsistency. You may take time to try. Thank you.

Can you please elaborate on what is not working? BCON from CMAQv5.3 does not take any mechanism information into account when generating BCs for a fine grid in regrid mode, all variables from the coarse grid input file are passed through to the fine grid BC file so there isn’t any chemistry inconsistency.

If BCON doesn’t work for you (although I think it should, but I may be missing something), you can instead use bcwndw as suggested by Sergey.