JGCRI / hector

The Hector Simple Climate Model
http://jgcri.github.io/hector/
GNU General Public License v3.0
107 stars 45 forks source link

Update Aerosols #724

Closed kdorheim closed 5 months ago

kdorheim commented 5 months ago

Related to #700, update the aerosol scalers

kdorheim commented 5 months ago

Implement aerosol forcing parameters to account for the new analysis by Zelinka et al. (2023)

Hector 3.0 is calibrated to forcing values in the AR6, which were informed by analysis of ESM model results as documented in Smith et al. (2021). We adjust aerosol forcing parameters in this version of the model from the values documented in Smith et al. (2021) to account for the new analysis by Zelinka et al. (2023). Zelinka et al. find, due to a coding error, a higher absorbing forcing, which is largely countered by a larger cloud indirect forcing. We therefore scale the magnitude of BC, other cooling aerosols, and F_ACI by the ratio of average CMIP6 model values from Zelinka et al. (2023, Table 2) and Smith et al. (2020) (Table 6). These changes largely offset each other, but result in a slightly higher net negative forcing for the current day and also result in slightly different total aerosol forcing pathways over time.


[Note two different references. Smith et al. 2021, and Smith et al. 2020. No relation :)

The resulting changes are to the following 5 aerosol constants:

SI Eq 12 -ρ𝐵𝐶 (new) = ρ𝐵𝐶(old) * ( (0.37+0.07) / (0.28 + 0.07) )**

SI Eq 13 through SI Eq 15 (for all the reflecting forcings) multiply by ( 0.65/0.63 )

SI Eq 16 -ρaci (new) = ρaci (old) * ( 0.88/ 0.81 )

The result will be a about a 5% larger net negative forcing, and a larger positive ARI from BC.

** Note that for BC we have assumed that the bulk of long-wave forcing is also due to BC absorption. Note that, while the long-wave forcing in Zelinka et al. (2023) and Smith et al. (2020) differ, this difference is not due to a coding change, but due to a different methodology in Zelinka et al. regarding how land surface warming is treated (Zelinka pers. communication). Because this is a definitional difference, we assume the LW forcing is the same between the two studies and allocate this to BC.

Smith et al. (2021) Smith, C. J., Kramer, R. J., Myhre, G., Alterskjær, K., Collins, W., Sima, A., Boucher, O., Dufresne, J.-L., Nabat, P., Michou, M., Yukimoto, S., Cole, J., Paynter, D., Shiogama, H., O'Connor, F. M., Robertson, E., Wiltshire, A., Andrews, T., Hannay, C., Miller, R., Nazarenko, L., Kirkevåg, A., Olivié, D., Fiedler, S., Lewinschal, A., Mackallah, C., Dix, M., Pincus, R., and Forster, P. M.: Effective radiative forcing and adjustments in CMIP6 models, Atmos. Chem. Phys., 20, 9591–9618, https://doi.org/10.5194/acp-20-9591-2020, 2020.

Zelinka et al. (2023) Zelinka, M. D., Smith, C. J., Qin, Y., and Taylor, K. E.: Comparison of methods to estimate aerosol effective radiative forcings in climate models, Atmos. Chem. Phys., 23, 8879–8898, https://doi.org/10.5194/acp-23-8879-2023, 2023.

kdorheim commented 5 months ago

Minor changes in global temp image

Changes are more noticeable for the specific aerosols image

github-actions[bot] commented 5 months ago

Differences in Hector outputs

Hello, this is leeyabot!

The current pull request's outputs do not differ from 3.1.1 (aab43a8).

codecov-commenter commented 5 months ago

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

:exclamation: No coverage uploaded for pull request base (dev@dbb7154). Click here to learn what that means.

:exclamation: Current head d06537d differs from pull request most recent head 5e6426d. Consider uploading reports for the commit 5e6426d to get more accurate results

Additional details and impacted files ```diff @@ Coverage Diff @@ ## dev #724 +/- ## ====================================== Coverage ? 80.15% ====================================== Files ? 61 Lines ? 5992 Branches ? 0 ====================================== Hits ? 4803 Misses ? 1189 Partials ? 0 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

ssmithClimate commented 5 months ago

^ Good point on SSPs, worth testing. I'd say just compare new/old, but if the total forcing is not that much difference could just be a comment in text - to keep the extra work down. (The nature of this change is that there's offsetting positive and negative forcing impacts, so its only going to be a pretty specific circumstance where this makes a large difference.).

github-actions[bot] commented 5 months ago

Differences in Hector outputs

Hello, this is leeyabot!

The current pull request's outputs do not differ from 3.1.1 (aab43a8).

kdorheim commented 5 months ago

Across the SPPs the total changes on RF total/temp are really small!

Total RF image MSE for total RF scenario MSE
hector_ssp119.ini 0.0001736
hector_ssp126.ini 0.0001539
hector_ssp245.ini 0.0002284
hector_ssp370.ini 0.0001664
hector_ssp434.ini 0.0001986
hector_ssp460.ini 0.0002071
hector_ssp534-over.ini 0.0002214
hector_ssp585.ini 0.0001971

Global Temp image

MSE Global Temp scenario MSE
hector_ssp119.ini 7.15e-05
hector_ssp126.ini 6.13e-05
hector_ssp245.ini 9.69e-05
hector_ssp370.ini 5.66e-05
hector_ssp434.ini 7.46e-05
hector_ssp460.ini 7.83e-05
hector_ssp534-over.ini 8.39e-05
hector_ssp585.ini 7.36e-05

RF BC image

MSE RF BC

scenario MSE
hector_ssp119.ini 0.0007222
hector_ssp126.ini 0.0007518
hector_ssp245.ini 0.0010348
hector_ssp370.ini 0.0030172
hector_ssp434.ini 0.0013467
hector_ssp460.ini 0.0016318
hector_ssp534-over.ini 0.0010274
hector_ssp585.ini 0.0012057
bpbond commented 5 months ago

Is MSE mean square error? If so I'm a little confused by that, but anyway, trust your judgment. Thanks!