Crash of B1850 after 260 years run

[phou@mtu_edu]

Dear CESM experts,I'm still very confused why my B1850 and B20TR run will have a very sharp surface temperature decrease in the first 10 years, which make simulation result can't match the observation. So i'm trying to run B1850 much longer to see if it's a spin-up problem. But my B1850 run crashed after 260 years. I can't figure out why it crashed, the error likes this:forrtl: error (78): process killed (SIGTERM)I also attached the log file below here. Hope there's someone can help.By the way, I'm really eager to know how can I get rid of that big surface temperature decrease in the beginning of B compset. That decrease plot line even appeared in my nightmares... Shall I use B20TRC5 instead? Thank you very much!!Best regards,Pei

 

[hannay]

How do you initialize the model? Do you use a startup run, an hybrid ? How much is the temperature decrease ? What is the top of the atmosphere imbalance ?
Depending on your starting type, the model can take from several years to several decades to equilibrate. (If you start from Levitus, it will be probably even longer).If you have an imbalance at the top of the model, the temperature can warm or cool. IT is not recommended to use B20TRC5 instead. B20TRC5 shoudl be started from an stable  B1850 simulation.

 

[phou@mtu_edu]

Thank you so much, Hannay.I'm using CESM1_2_0, ./create_newcase -case xxx -compset B20TR -res T31_gx3v7 -mach XXX command to build the case, and only modify STOP_N option in env_run.xml. So it was a startup run. Both my B1850 and B20TR runs have a decrease nearly 2 degree in the first 10 years, also there's a big jump up at the very beginning of the run.(attachment) All of these stranges' increase and decrease are very different from F20TR run, which is much more like the observation result. Unfortunately, F20TR is not the case we want, we simulate the past, because we want to make sure it will work well in a BRCP compset, which has the similar problem now as B20TR. I'm not quite sure about what is " the top of the atmosphere imbalance". Does that mean the top level's temperature? If so, the top level's temperature has a decrease of 2.6 degree in the first simulated year, and then generally decrease 3.5 degree in the next 150 years, if I run a B20TR case from 1850 to 2005.After the case reach equilibrate, it is actually lower than the observation results. For example, if I want to use BRCP case to predict 2010's surface temperature. The case is not equilibrium yet after 5 years run. And even it finally get equilibrate, it will not be a correct number as I expected. How can I fix this problem? Thank you again!! I'm really appreciated for all the replies.Best regards,Pei 

 

[wanlingfeng_123@163_com]

i  have occur similar. so ，i wanna to konw，what the solution about it. i was used FW, and i have change solar data ,but have spinup have tens year.i don't konw what happened.

 

[hannay]

Hi,It is not enough to report: forrtl: error (78): process killed (SIGTERM)Thsi error could occur from many reasons.We need atl least log files/ 

 

[wanlingfeng_123@163_com]

i problem is the output have became sipnup after 20 years run. i used FW,f19_g16,start at 1955,then i have change the solar data.with startup to submited run for 30year,but after 20 years run , i see the 10 hpa T and 500 hpa T ,they are all have fluctuate in 1 degree range. all the run is success, but the resulte is outstand i suspition.