Contact — Julia Hargreaves 2013/04/26 06:53 to contribute to this page, or leave a comment in the discussion box.
Estimating the model forcing for LGM_sens runs
A seed for discussion by Masa Yoshimori and Julia Hargreaves.
In order to investigate sensitivity of the climate system to different forcing (LGM vs. CO2 increase), we need to know the radiative forcing of each model rather accurately. As we consider the LGMGHG experiment, we would like to know the LGMGHG forcing as well. The forcing form individual models have not been provided before and often forcing from a particular model is used to interpret different model results.
One popular method to estimate radiative forcing is so-called Gregory method in which the forcing is applied instantaneously (e.g., CO2 is quadrupled abruptly), and the following 150 years of transient response is analyzed using the linear regression. This method is probably not useful to estimate the forcing that include large ice sheets.
Alternately, we can estimate the ice sheet and orbital forcing using approximate PRP method (APRP) of Taylor et al. (2006). Then, what we need to know is forcing of CO2 increase (2x or 4xCO2) and LGMGHG. The Gregory method requires 150 years of integrations to obtain these numbers, and this is perhaps costly for many groups for the purpose of just to get the forcing.
The alternate method is to use AGCM with fixed preindustrial SST. Although this includes the response of land surface temperature change by design, this is well-established method (aerosol forcing, including indirect effect, is estimated in this way for example). This likely requires only 10 to 15 years of integration without ocean and thus relatively inexpensive.
What we propose to start the discussion is
- To use the APRP method to estimate ice sheet and orbital forcing, separately (we need standard monthly output and land ice mask)
- To use AGCM with prescribed preindustrial SST to estimate 4xCO2 and LGMGHG forcing (the net radiative imbalance gives the effective forcing).
- We add LGMGHG forcing, ice sheet forcing from APRP, and orbital forcing from APRP to get the total LGM forcing.
[ PMIP3 Wiki Home ] - [ Help! ] - [ Wiki syntax ]
Discussion
Dear Masa and Jules,
I think that this is a good approach. I think that using the AGCM with prescribed preindustrial SST will give a more accurate estimate of the forcing than the Gregory method.
I think that Jonathan Gregory has recently experimented with calculating ice sheet forcing using SW APRP. I recall he had some issues with residuals. It might be worth contacting him about it.
Also, forgive my ignorance, but if you estimate the ice sheet forcing with SW APRP, how do you account for the effect on surface temperature of the increased surface elevation? Would running an AGCM with preindustrial SSTs and changes in orbital forcing and/or a prescribed ice sheet be a useful approach?
Regards,
Mark
Dear Mark,
Thank you very much for your comment, and I apologize for the very late response. I would like to reply to your comment with broader audience in mind.
If one first defines climate feedback as the surface-temperature-change-mediated radiation anomaly, then the forcing should be defined as the radiative anomaly at zero global mean surface temperature change. This is the basic idea of the Gregory method although it allows local temperature to respond (as long as the global mean is zero).
If one first defines the (effective) radiative forcing as radiative anomaly induced by atmospheric response to perturbation under fixed SST conditions (the Hansen method), then the forcing includes land surface temperature adjustments to the forcing.
As contribution of land surface temperature to the global mean temperature change under fixed SST conditions is relatively small, both methods are, roughly speaking, comparable for CO2 forcing although there are differences in details. The LGM is a bit different since the land surface temperature change is one of the dominant contributors to the global mean temperature response.
As you suggest, using the Hansen method for all forcing may be more consistent with effective radiative forcing calculation planned in RFMIP (Radiative Forcing Model Intercomparison Project) for various forcing agents. As long as the definition of forcing and feedback as well as the methodology to compute them are clarified, either way may be useful (We still have to check the residual issue of the APRP that you raised).
We probably need to start discussion on how the choice may or may not affect (conceptually and practically) on constraining of future climate change from the multimodel LGM experiments (and understanding of the outcome), which is the aim of this working group.
Regards,
Masa