How to run WACCM with data Land model (CLM)?

[kzkaizhang@gmail_com]

Dear, I am running WACCM (FW compset) in CESM 1.2.0. But it is coupled with CLM which cannot be changed. Now I want to change the surface forcings like surface temperature to drive the model, that is, prescribe the CLM, just like prescribe the SST. Is there anyone who can help me on how to configure the data Land Model? And how to deal with it? Many thanks! I really need to know how to achieve it!

 

[mmills]

Kai,You may have better luck posting your question in a forum for CLM, or the data models. I don't know of anyone who has run WACCM with a data land model. You should ask the land model group what effects this would have on WACCM. 

 

[kzkaizhang@gmail_com]

Dear Mike,THanks very much for your help!I studied the tech note of CESM again and I found a solution - creat a new compset in /scripts/ccsm_utils/Case.template/config_compsets.xml by adding a sentence :2000_CAM4%WCCM_DLND%CPLHIST_CICE%PRES_DOCN%DOM_RTM_SGLC_SWAVThen I can use data land model in WACCM now. Hope this will help some others who meet the same situation.

 

[mmills]

Kai,I have spoken with David Lawrence, leads the CLM group at NCAR. He told me he is not at all familiar with the data land model, and isn't sure whether there really is one. There may just be a stub there.  He also said that he doesn't know any way to artificially modify the temperature in CLM. So it may not be possible to do what you wish to do.I've just learned that NCAR does not currently have resources to support a community liaison for the land model, so it may be some time before someone responds to your post in the land model forum.

 

[santos]

I would add three things to what Mike says:1. This is really more of a CAM question than a CLM question; if you are running data land, that means that you aren't running CLM at all.2. I believe that the current data land model was designed to feed data to CISM (the land ice model), not CAM. Therefore, I doubt that it will do the right thing here.3. For purposes of CAM testing, we view CLM as a dependency of CAM; if you wanted to replace it, you would not only have to specify surface temperatures, but many other quantities that feed into specific physics parameterizations, such as dust emissions and radiation. This is not a trivial set of changes, which is why we almost always use CLM with CAM, rather than a data land model.The only way that CAM can run without CLM is in certain minimalist test cases, such as "aquaplanet" cases (i.e. there is no land and a data ocean covers the whole globe), and ideal or adiabatic physics (where almost all physics and chemistry are turned off). None of these are really suitable for a scientific study using WACCM.

 

[kzkaizhang@gmail_com]

Dear Santos,Thanks very much for your information!I got some notes from CESM user guide. I can create my own compset for my own use. When I creat a new compset: 2000_CAM4%WCCM_DLND%SCPL_CICE%PRES_DOCN%DOM_RTM_SGLC_SWAV, that is, WACCM with data land model (DATA_MODE=CPLHIST,DATA_SNO_MODE=NULL), it comes out error when "cesm_setup":CAM writing dry deposition namelist to drv_flds_in Writing ocean component namelist to ./docn_in CAM writing namelist to atm_in ERROR(build-namelist::new): Required input variable yearfirst was not foundERROR: dlnd.buildnml.csh failedERROR from preview namelist - EXITINGBut if I create the new compset : 2000_CAM4%WCCM_SLND%SCPL_CICE%PRES_DOCN%DOM_RTM_SGLC_SWAV (DATA_MODE=NULL,DATA_SNO_MODE=CPLHIST). It successfully runs. How does it happen? And in http://www.cesm.ucar.edu/models/cesm1.2/cesm/doc/usersguide/x2382.html, it says that we can create our own compsets, why I cannot use WACCM with data land model in my case? Thanks very much for your help again!

 

[santos]

Hmm. As I said, the data land model was not created for use with CAM, but for CISM. In this case, because CISM is not on, it fails.I have no idea why the stub land model is working; it should basically be inert, so perhaps CAM is just getting fed all 0s, or constant data. It can't provide all the data necessary to get realistic output.Setting up a fully prescribed land model for use with CAM is a somewhat complex question, and not something we can provide support for. If you want to develop a method for doing this, I wish you luck, but CESM just does not have the capability that you want at this time.

 

[kauff]

In CESM, the land component is responsible for providing the atm/lnd fluxes. The choices for a land component are CLM and DLND (also a stub land, but that is just a "do nothing" stub componet).  If an active atm component is being used, it will be necessary to have atm/lnd fluxes.Computing the atm/lnd fluxes is a large part of what CLM does.  DLND currently does not have the functionality to compute atm/lnd fluxes.  While DLND can't compute the fluxes, hypothetically the DATM could read atm/lnd fluxes from an input data stream, and thus provide them to the coupled system -- but if it did so the atm/lnd fluxes would be totally inconsistent with the current state of the atmosphere, and thus the configuration wouldn't make any sense, scientifically (the simulation would most likely blow up).  This is why in CESM, if you are using and active atm component (CAM, WACCM), you must also use CLM.-Brian Kauffman, NCAR/CSEG 

 

[Yincheng]

In CESM, the land component is responsible for providing the atm/lnd fluxes. The choices for a land component are CLM and DLND (also a stub land, but that is just a "do nothing" stub componet). If an active atm component is being used, it will be necessary to have atm/lnd fluxes.Computing the atm/lnd fluxes is a large part of what CLM does. DLND currently does not have the functionality to compute atm/lnd fluxes. While DLND can't compute the fluxes, hypothetically the DATM could read atm/lnd fluxes from an input data stream, and thus provide them to the coupled system -- but if it did so the atm/lnd fluxes would be totally inconsistent with the current state of the atmosphere, and thus the configuration wouldn't make any sense, scientifically (the simulation would most likely blow up). This is why in CESM, if you are using and active atm component (CAM, WACCM), you must also use CLM.-Brian Kauffman, NCAR/CSEG

I meet the same problem with original poster that I want to run CESM 1.2.2 with DLAND mod for the purpose of fixed surface temperature, I have a question that why many simulations which prescribed the SST would not blow up except prescribed the surface temperature??

 

[peverley]

Moving this thread to the land forum in the hope that they can answer your question.

 

[slevis]

I will let others comment in this thread or in this related thread whether the capability of using dlnd to prescribe land temperature now exists in the CESM.

Regardless, it should go without saying that advanced model users may make modifications to the code to accomplish any task that the CESM does not already perform.