case/Expt E0.1 Firex: Bias Scaling of CEDS 2019 SO2 Emissions

drnimbusrain commented 2 years ago

Case Details:

This is an experimental case run from July 1, 2019 -> August 31st, 2019. The model was restarted daily at 0z and forecasted out at 24 hours. The run uses the UFS-Aerosols branch feature/p7.1 with the updated inline AOD calculation as well as the fengsha sepd bug fix and new inputs. The purpose of this case is to evaluate the CEDS 2019 SO2 emissions, UFS-Aerosols AOD SO4 bias (against MERRA2), and applying global, gridded SO2 emissions scaling factors based the bias. Note: While this simple bias scaling methodology is applied here for only SO2 emissions, it could be applied to other CEDS emissions species in which speciated MERRA2 AOD analysis is available.

Model Configuration Details

Resolution: C96
Begin Date: 2019-07-01
End Date: 2019-08-31
Cold Start: True
FHOUT_GFS: 6
Forecast length (FHMAX_GFS): 24 hour cycling
na_init=1
LEVS: 65
CCPP suite: FV3_GFS_v16
FV3 output: every 6 hours

Aerosol Model Configuration

AERO_ExtData file: AERO_ExtData.rc
ANTHRO1 (surface emissions) : CEDS 2019
ANTHRO2 (elevated emissions) : CEDS 2019
SHIP : CEDS 2019
Biomass Burning: QFED where GBBEPx isn't available
Aviation: HTAP
NH3 Emissions: CEDS
Dust : Fengsha
- alpha: 2.2
- gamma: 1.0
- input file: FENGSHA_SOILGRIDS2017_GEFSv12_v1.0.nc
Sea Salt: 3
- scaling factor: 1
AOD output: 3 hours
Column Mass output: 3 hours
Surface Concentration output: every 3 hours

Measurements/Models available for comparison

drnimbusrain commented 2 years ago

Sulfate AOD bias against MERRA2 for August 2019:

Associated regression of sulfate AOD and SO2 emissions for August 2019: N = 450 pts Pearson (R) correlation = 0.7 Slope = 2.71e+09

Calculated fractional SO2 emissions change from regression analysis and sulfate AOD bias (against MERRA2) for August 2019:

drnimbusrain commented 2 years ago

@bbakernoaa Can you reprocess the daily CEDS 2019 emissions for July and August 2019 to account for this SO2 emissions bias scaling (above), so I can re-run this test? You would add the following lines to your HEMCO_Config file to scale only the July and August 2019 emissions:

ROOT : /groups/ESS/pcampbe8/emissions/hemco 9992 SO2_Bias_MASK_2019 $ROOT/MASKS/v2018-09/MERRA2_UFS_SO2_Bias_Scale_2019$MM.nc SO2_frac_emis 2019/1/1/0 C xy unitless 1 -180/-90/180/90

bbakernoaa commented 2 years ago

I can. Let's hold off on doing the simulation for the moment. We need to conserve some CPU allocation for the p7.1 runs. We have been running up against it and just got some more allotted to us this week.

I'll go ahead and process the emissions though

drnimbusrain commented 2 years ago

Sure Barry, I was going to ask about that. I have only been running at C96-L64, but will halt all emission test simulations until I hear from you all about the p7.1 runs. Thank you.

drnimbusrain commented 2 years ago

@zhanglikate This is the beginning of my CEDS 2019 SO2 emissions scaling tests, which we stopped because of limited resources on Hera. I see the underprediction in UFS-Aerosol Sulfate AOD compared to MERRA2, which agrees with what you are seeing.

However, per our discussion at the meeting, we need to be careful here for two reasons: 1) The CEDS 2019 emissions are definitely more representative of the August 2019 time period compared to CEDS 2014, and 2) The MERRA2 Sulfate AOD is a reanalysis dataset which may not be the best to use to scale the SO2 emissions themselves (but this is what we have right now).

It seems like there may be another issue in the model over East Asia, possibly too much wet scavenging? We have done a lot of analysis that shows the CEDS 2019 has decreased significantly from CEDS 2014 (~ 30% decrease globally), especially over East Asia. We could also look into the processed CEDS2019 emissions we are using, and make sure there are no errors.

Also, you can see from previous analysis in GEFS-Aerosol, using CEDS 2019 led to much better results compared to CEDS 2014 over East Asia, which reduced the overprediction in GEFS-Aerosol:
https://docs.google.com/presentation/d/1dJbpuJ-a6_UczUa_aFoDzfP_PUDhUH52V8uGQN5gYNs/edit#slide=id.gb491a0cb5c_0_76

zhanglikate commented 2 years ago

@drnimbusrain I think normally it is hard to validate the wet removal of sulfate if not correct over Eastern China. The large wet removal in GEFS-Aerosols is based on the GOCART default setting from NASA, which is different to GEFS-Aerosols. I have not yet touch that part now, but will be my next step to do some testing. Another concern about tuning the wet removal is because the model normally is using the same wet removal globally, while if tuning the wet removal for the eastern China, other areas may overpzedictetd. Also, the wet scavenging factor in UFS-Aerosols is smaller than the GEFS-Aerosols, which means we already have less wet scavenging over eastern China in UFS-Aerosols.

drnimbusrain commented 2 years ago

@zhanglikate OK, we should look at the CEDS 2019 emissions that @bbakernoaa processed for these UFS-Aerosol runs. I had processed the previous CEDS 2019 emissions for the GEFS-Aerosol runs, which showed largely improved performance vs. CEDS 2014. I will compare the August CEDS 2019 SO2 emissions between the two sets, and update you here soon.

zhanglikate commented 2 years ago

@drnimbusrain @gjfrost The comparison for using CEDS_2014 (EXP.5) and CEDS_2019 (EXP.6) in UFS-Aerosols at https://docs.google.com/presentation/d/1R5aXo5AxvCGmnMxhi0xlmDroMC0QS4VT/edit#slide=id.p1

You can see, using CEDS_2014, there is positive biases over Eastern China.

drnimbusrain commented 2 years ago

@zhanglikate , Right this makes sense. We saw the same thing in GEFS-Aerosol experiments. We went from an overprediction using CEDS 2014, to an underprediction using CEDS 2019. Again, this was based on comparison against reanalysis MERRA2 AOD sulfate, which should be taken with a grain of salt. The overall AOD performance (against AERONET) was better over East Asia when using CEDS 2019 compared to CEDS 2014: https://docs.google.com/presentation/d/1dJbpuJ-a6_UczUa_aFoDzfP_PUDhUH52V8uGQN5gYNs/edit#slide=id.gb491a0cb5c_0_76

Isn't this what you are also seeing?

I could also compare the CEDS 2019 emissions I generated for the GEFS-Aerosol experiments, and the CEDS 2019 emissions generated for these UFS-Aerosol runs from Barry. They should be very similar.

zhanglikate commented 2 years ago

@drnimbusrain We have not yet compare with AERONET in UFS-Aerosols using CEDS_2014 and CEDS_2019. @perthsb , right?

drnimbusrain commented 2 years ago

@zhanglikate Here are the relative (%) difference plots for CEDS 2019- CEDS 2014 emissions in August, and you can see the very large decrease in CEDS SO2 emissions over East Asia (~ 40-60%).

Global

East Asia

zhanglikate commented 2 years ago

Is it your version the same as what we used in the UFS-Aerosols?

Thanks.

Kate

On Tuesday, November 9, 2021, Patrick Campbell @.***> wrote:

@zhanglikate https://github.com/zhanglikate Here is the absolute and relative (%) difference plots for CEDS 2019- CEDS 2014 emissions, and you can see the very large decrease in SO2 emissions over East Asia (~ 40-60%).

Global [image: image] https://user-images.githubusercontent.com/26631222/141016522-5bdaaafc-1682-42df-9540-f0592b1b67d7.png

East Asia [image: image] https://user-images.githubusercontent.com/26631222/141016612-c6f6f6ba-da90-429c-908a-df41e4c6768b.png

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/noaa-oar-arl/UFS-Aerosol-Config/issues/24#issuecomment-964609612, or unsubscribe https://github.com/notifications/unsubscribe-auth/APJPDRDZTVLQMXJXRSS6SL3ULGPDTANCNFSM5DG47DCQ . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

bbakernoaa commented 2 years ago

Kate they are just for a different year. For GEFS we only used the CEDS 2019 data from the year 2019 (the target simulation year). But it is the same data used as in UFS-Aerosols.

On Tue, Nov 9, 2021 at 6:07 PM Kate.Zhang-NOAA @.***> wrote:

Is it your version the same as what we used in the UFS-Aerosols?

Thanks.

Kate

On Tuesday, November 9, 2021, Patrick Campbell @.***> wrote:

@zhanglikate https://github.com/zhanglikate Here is the absolute and relative (%) difference plots for CEDS 2019- CEDS 2014 emissions, and you can see the very large decrease in SO2 emissions over East Asia (~ 40-60%).

Global [image: image] < https://user-images.githubusercontent.com/26631222/141016522-5bdaaafc-1682-42df-9540-f0592b1b67d7.png

East Asia [image: image] < https://user-images.githubusercontent.com/26631222/141016612-c6f6f6ba-da90-429c-908a-df41e4c6768b.png

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub < https://github.com/noaa-oar-arl/UFS-Aerosol-Config/issues/24#issuecomment-964609612 , or unsubscribe < https://github.com/notifications/unsubscribe-auth/APJPDRDZTVLQMXJXRSS6SL3ULGPDTANCNFSM5DG47DCQ

. Triage notifications on the go with GitHub Mobile for iOS < https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675

or Android < https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub .

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/noaa-oar-arl/UFS-Aerosol-Config/issues/24#issuecomment-964626141, or unsubscribe https://github.com/notifications/unsubscribe-auth/AFIUVNYZMZOLLWIST26WLYLULGSTFANCNFSM5DG47DCQ .

zhanglikate commented 2 years ago

Barry, thanks for clarifying it.

Kate

On Tuesday, November 9, 2021, Barry Baker @.***> wrote:

Kate they are just for a different year. For GEFS we only used the CEDS 2019 data from the year 2019 (the target simulation year). But it is the same data used as in UFS-Aerosols.

On Tue, Nov 9, 2021 at 6:07 PM Kate.Zhang-NOAA @.***> wrote:

Is it your version the same as what we used in the UFS-Aerosols?

Thanks.

Kate

On Tuesday, November 9, 2021, Patrick Campbell @.***> wrote:

@zhanglikate https://github.com/zhanglikate Here is the absolute and relative (%) difference plots for CEDS 2019- CEDS 2014 emissions, and you can see the very large decrease in SO2 emissions over East Asia (~ 40-60%).

Global [image: image] < https://user-images.githubusercontent.com/26631222/141016522-5bdaaafc- 1682-42df-9540-f0592b1b67d7.png

East Asia [image: image] < https://user-images.githubusercontent.com/26631222/141016612-c6f6f6ba- da90-429c-908a-df41e4c6768b.png

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub < https://github.com/noaa-oar-arl/UFS-Aerosol-Config/issues/ 24#issuecomment-964609612 , or unsubscribe < https://github.com/notifications/unsubscribe-auth/ APJPDRDZTVLQMXJXRSS6SL3ULGPDTANCNFSM5DG47DCQ

. Triage notifications on the go with GitHub Mobile for iOS < https://apps.apple.com/app/apple-store/id1477376905?ct= notification-email&mt=8&pt=524675

or Android < https://play.google.com/store/apps/details?id=com.github. android&referrer=utmcampaign%3Dnotification-email%26utm medium%3Demail%26utm_source%3Dgithub .

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/noaa-oar-arl/UFS-Aerosol-Config/issues/ 24#issuecomment-964626141, or unsubscribe https://github.com/notifications/unsubscribe-auth/ AFIUVNYZMZOLLWIST26WLYLULGSTFANCNFSM5DG47DCQ .

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/noaa-oar-arl/UFS-Aerosol-Config/issues/24#issuecomment-964628056, or unsubscribe https://github.com/notifications/unsubscribe-auth/APJPDRDIHAJ54RFS37R7D6LULGS7RANCNFSM5DG47DCQ . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

drnimbusrain commented 2 years ago

@zhanglikate @bbakernoaa However, I also just did a quick comparison against the CAMS 2019 - 2014 SO2 emissions dataset, and it is actually quite different in the direction of emissions changes compared to CEDS 2019-CEDS 2014, particularly over East Asia:

CEDS 2019-2014

CAMS 2019-2014

This does bring up numerous questions as to why the CEDS vs. CAMS emissions trends are so different.

zhanglikate commented 2 years ago

That is interesting. Can you compare with MEIC over China: http://meicmodel.org/?lang=en ?

On Nov 9, 2021, at 4:14 PM, Patrick Campbell @.***> wrote:

@zhanglikate https://github.com/zhanglikate @bbakernoaa https://github.com/bbakernoaa However, I also just did a quick comparison against the CAMS 2019 - 2014 SO2 emissions dataset, and it is actually quite different in the direction of emissions changes compared to CEDS 2019-CEDS 2014, particularly over East Asia:

CEDS 2019-2014 https://user-images.githubusercontent.com/26631222/141019817-4cad8b32-7fa5-4303-a067-3c401c2537e1.png CAMS 2019-2014 https://user-images.githubusercontent.com/26631222/141020471-33a23435-90d4-4130-8b4c-0e0b802a83a4.png This does bring up numerous questions as to why the CEDS vs. CAMS emissions trends are so different.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/noaa-oar-arl/UFS-Aerosol-Config/issues/24#issuecomment-964629778, or unsubscribe https://github.com/notifications/unsubscribe-auth/APJPDRGKSYXVQFROS5F7UZLULGTLXANCNFSM5DG47DCQ. Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

drnimbusrain commented 2 years ago

@zhanglikate @bbakernoaa From what I know CAMS uses scenario data to extrapolate to recent years and perhaps uses more up to date emissions in Europe), but I’m not sure if they’ve documented their latest version.

There’s no reason to believe that global shipping SO2 decreased dramatically over this period (it would have in some ECA zones), and that’s not supported by the IMO assessment either (see graph copied below with IMO estimates out to 2018). Source: Fourth IMO GHG Study 2020 - Final report - MEPC 75-7-15.pdf

Emissions in China have definitely decreased over the period.

I’m surprised the India emissions are mainly decreasing in CEDS though, as the total 2019 SO2 emissions in India are slightly higher than 2014 in CEDS (but only by an annual average of 4%). Maybe this is driven by some local increases in SO2 emissions sources that influence the CEDS average, especially because this is only a 1-month comparison I show.

I suspect the CAMS estimates aren’t actually 2019 estimates, but projections.

drnimbusrain commented 2 years ago

CEDS2019-2014_SO2_Emissions

noaa-oar-arl / UFS-Aerosol-Config