how to cycle specific time periods to get 100 years simulation

[xiaoxiaokuishu]

Hi, all,
Recent, i modified the default land cover ( i mean both the fsurf and flandusetimeseries) during the 1990-2010, so i plan to make the initial data by myself.
But I have some questions about this issue,
(1) can i use the I compset to get the initial data, since F compset and B compset is time-consuming?
(2) I plan to cycle 1990-2010, 5 times to get 100years simulation, is there any way to set the namelists to get my target
(3) the reason of why cycle 1990-2010, as a apposed to simulate the 1850-1990 to get the initial data, is that outside the 1990-2010 time periods, i did not change the
landcover...so it is not meannningful to do such a simulation?

Hope to hear from you!
Best

 

[oleson]

(1) Sure. The caveat is that the I compset forcing data climate will be somewhat different from the coupled model climate, so there will be some adjustment needed when you use the initial file in the coupled run.
(2) For an I-compset spinup as you've specified. you could set:

./xmlchange RUN_STARDATE= 0001-01-01
./xmlchange DATM_CLMNCEP_YR_START=1990
./xmlchange DATM_CLMNCEP_YR_END=2010
./xmlchange DATM_CLMNCEP_YR_ALIGN=1

This aligns model year 1 with year 1990 forcing data.

(3) What you are doing seems reasonable, if I understand correctly.

 

[xiaoxiaokuishu]

（1）当然。需要注意的是，I compset 强迫数据气候与耦合模型气候有些不同，因此当您在耦合运行中使用初始文件时，需要进行一些调整。
(2) 对于您指定的 I-compset spinup。你可以设置：

./xmlchange RUN_STARDATE= 0001-01-01
./xmlchange DATM_CLMNCEP_YR_START=1990
./xmlchange DATM_CLMNCEP_YR_END=2010
./xmlchange DATM_CLMNCEP_YR_ALIGN=1

这使模型第 1 年与 1990 年的强迫数据保持一致。

(3) 如果我理解正确的话，你所做的似乎是合理的。

感谢您的回复！

对于（2）问题，我也不清楚如何提供100年的模拟结果，

./xmlchange RUN_STARDATE= 0001-01-01
./xmlchange DATM_CLMNCEP_YR_START=1990
./xmlchange DATM_CLMNCEP_YR_END=2010
./xmlchange DATM_CLMNCEP_YR_ALIGN=1

i did not see where can i set 5 (20years*5 cycle=100) as the cycle periods?

Is it just set the
NSUBMIT=5
NYEARS=20,

Best

 

[oleson]

I think you are probably confusing the datm settings with the total number of years you want the model to run. If you want to run 100 years and your system queue allows you to run the model for 100 years without resubmitting then set STOP_OPTION to nyears and STOP_N to 100. If it can't then you can use RESUBMIT to run your model for however many segments you need. For example, if your system queue will support running the model for 20 years, then set STOP_N to 20 and RESUBMIT to 4.
By the way, running the model 1990 through 2010 is 21 years, not 20 years.

 

[xiaoxiaokuishu]

I think you are probably confusing the datm settings with the total number of years you want the model to run. If you want to run 100 years and your system queue allows you to run the model for 100 years without resubmitting then set STOP_OPTION to nyears and STOP_N to 100. If it can't then you can use RESUBMIT to run your model for however many segments you need. For example, if your system queue will support running the model for 20 years, then set STOP_N to 20 and RESUBMIT to 4.
By the way, running the model 1990 through 2010 is 21 years, not 20 years.

Thanks! I clarify our discussion, i plan to cycle 1991-2010 simulation 5 times to get the 100 years result, here the cycle means cycle the datm for 5 times.

Best