About the spatial distribution of COMBINE_SA_ACON file

lotso · July 31, 2023, 8:41am

I used two versions of the ISAM module CMAQv5.4 and CMAQv5.3.2, respectively.The parameter settings are the same for both versions, but the spatial distribution of some labeled areas in the combine file is quite different, while others differ less.The following two images are the spatial distribution of different areas in the comebine file. For example, the distribution of O3_SH in 5.4 is quite different from that in 5.3.2, while the distribution of O3_CZ in 5.4 and 5.3.2 is similar. （All images are from screenshots of Verdi）
I wonder why this is, is it a problem with some parameter settings? How do I go about solving this problem?Thank you all for your help！

Ben_Murphy · July 31, 2023, 4:51pm

Hello,

It looks like the color scales are very different for the maps of O3_SH between the model versions. You may want to check that the emissions for the Shanghai region are being read properly. Are the bulk concentrations similar between the two model versions?

-Ben

hogrefe.christian · July 31, 2023, 5:14pm

Which ISAM_O3_WEIGHTS option are you setting in your CMAQv5.4 runs? Unless you are setting it to “1”, some differences in ISAM spatial patterns relative to the CMAQv5.3.3 approach are expected, quite aside from the effects of other science updates in CMAQv5.4 on bulk concentrations (and therefore ISAM ). For further details on the ISAM_O3_WEIGHTS option in CMAQv5.4, please also see the recent paper by Shu et al.

That said, I would expect the type of differences shown for O3_CZ in your second row of plots, but not the differences shown for O3_SH the first row of plots, especially given the different color scale noted by @Ben_Murphy in addition to the very different spatial patterns you pointed out. Since your plots show results after an additional step of post-processing (running COMBINE), have you confirmed that the same type of inconsistency exists for O3_SH in the raw SA_ACONC files? Is O3_SH the only tag for which you see very different spatial patterns?

lotso · August 17, 2023, 8:26am

Thanks for your reply.I also checked the results further.

ISAM_O3_WEIGHTS option is the default of 5.

2.I checked the ACONC file on the same day (2022.6.6) and adjusted the color maps to be consistent, but there was still a big difference in spatial distribution.

3.I also checked the SA_ACONC file on the same day (2022.6.6). The following GIF is the spatial distribution of the O3_SH on this day. (The concentration results of v532 and v54 are quite different, so this color map are not the same.) In contrast, it feels that the results of v532 are more reasonable. At least it began to spread outward from Shanghai.
test0606_O3_SH
test0606_O3_SH_v54

4.The O3_SH is not the only tag with differernt spatial distribution, and the spatial distribution of many tags is also strange.

Ben_Murphy · August 17, 2023, 1:47pm

Hello,

Thank you for posting these additional results. They are very helpful.

The differences in the bulk Ozone concentration between the v5.3.2 and v5.4 ACONC files is notable. It would be a good idea to focus on why those are so different first, and then move on to ISAM predictions.

Could you please send your RunScripts and an example CTM_LOG file from each simulation so that we can confirm there are no issues reading emissions in the v5.4 simulation? After that, we can look at the ISAM control file to see if there are issues there.

Best wishes,
Ben

lotso · August 20, 2023, 1:55pm

Hello

When I ran ISAM for v5.4, the ICON and BCON files were both generated as a result of v5.3.2. Will this have a big impact on the results? From the perspective of time series, v5.4 is also significantly low. Here ate the script and CTM_LOG files I ran, please see where the problem is.

Thanks a lot!

hogrefe.christian · August 24, 2023, 8:38pm

Thank you. Could you please also post the run script and log file (the CTM_LOG_000 one is sufficient) for your CMAQv5.3 simulation?

lotso · August 29, 2023, 6:53am

I have uploaded the script and LOG files a few days ago, is it that I did not upload successfully, can not view?
run_cctm_cn04_2206_O3_SA.csh (40.7 KB)
CTM_LOG_000.v54_ISAM_intel_cn04_20220525.txt (603.6 KB)

hogrefe.christian · August 29, 2023, 11:38am

I see the uploaded run script and log file for your CMAQv5.4 simulation, but not the corresponding files for your CMAQv5.3.2 simulation.

For us to help diagnose the issue further, could you please upload the files from both the CMAQv5.4 and CMAQv5.3.2 simulations?

lotso · August 29, 2023, 2:07pm

Here are the script and CTM_LOG_000 file for CMAQv5.3.2.
run_cctm_ISAM_O3_202206_CZ_REGION_0.5NOX_5VOC.csh (36.8 KB)
CTM_LOG_000.v532_intel_cn04_20220525.txt (925.8 KB)

hogrefe.christian · August 29, 2023, 2:26pm

Thank you. Could you please also upload the two isam_control files, i.e. /data5/projects/Changzhou/CMAQ/CCTM/script/ISAM/O3/REGION/isam_control_cn04_CZ.txt from run_cctm_ISAM_O3_202206_CZ_REGION_0.5NOX_5VOC.csh|attachment and ${WORKDIR}/isam_control_O3_REGION.txt from run_cctm_cn04_2206_O3_SA.csh

lotso · August 30, 2023, 2:52am

CMAQv5.3.2
isam_control_cn04_CZ_v532.txt (3.5 KB)
CMAQv5.4
isam_control_O3_REGION_v54.txt (4.1 KB)

hogrefe.christian · August 30, 2023, 7:44pm

Thank you for posting this additional information

I’ll admit that I’m a bit stumped here. The only difference I see in the DESID mapping portion of the log files is that in your v5.3.2 your emitted species XYL was mapped to the mechanism species XYLMN while in your v5.4 runs it was dropped altogether. Regardless of whether this is intentional behavior or not, I wouldn’t expect this to cause the type of differences you’re seeing, so I’ll have to defer to others to see if they have any additional suggestions on what might be happening.

Ryan · August 31, 2023, 5:35am

I would like to provide several suggestions as I also run some CMAQ cases over China with a number of CMAQ versions.

Firstly, consider to generate icon and bcon files using csh files under CMAQv5.4 to avoid any minor issues.
Secondly, I would question on biogenic emission used in these CMAQ runs. Are you using offline BVOC emissions calculated from MEGANv2.1 in both simulations? I would expect considerable ozone underestimation in CMAQ5.4 case if you are using the inline module of MEGAN3 in CMAQv5.4.
Last suggestion is that you can activate the IPR in both cases to see if any major changes in critical physical/chemical processes. Also, checking emission files and chemical mechanism used in modeling is warranted.

I concur with Christian that there is not sufficient details shown in LOG files, thus additional investigations is needed.
Hope it helps.

Bruteleenandrane · September 2, 2023, 1:13am

Thank you for posting these additional results. They are very helpful.

jinefune · September 11, 2023, 6:53pm

Thank you for posting these additional results. They are very helpful.

jinefune · November 15, 2023, 9:01am

Thanks a lot!..

Topic		Replies	Views
Different base PM2.5 concentration when applied Regional file in CMAQv5.3.2-ISAM CMAQ-ISAM	8	1020	February 18, 2021
CMAQv5.3 ISAM - difference with reference data CMAQ-ISAM	3	779	November 19, 2019
CMAQ ISAM ozone source apportionment weird situation CMAQ-ISAM	12	846	January 20, 2021
New ISAM module in CMAQv5.3.1 CMAQ-ISAM admin , smoke	1	469	November 10, 2020
Contrasting Source Apportionment Approaches with Zero Out Testing CMAQ-ISAM research-topic	1	267	November 14, 2023

About the spatial distribution of COMBINE_SA_ACON file

Related Topics