Main menu

Navigation

Initial conditions for running CESM in forecast mode

7 posts / 0 new
Last post
jayant.pendhark...
Initial conditions for running CESM in forecast mode

I am a new user to cesm. I am using cesm1.2.2. I am interested in running cesm in the forecast mode, particularly, would like to investigate the El Nino and La Nina case of 1997-98. I already performed a simulation for the compset 'F_2000_CAM5' with SE core of 1 deg resolution. However, the results did not show the desired results. I have learnt that I need to check my initial conditions. Please suggest me where to find or how to create initial conditions for investigating the desired case. I want to have initial conditions for July 1997 and want to go for a higher resolution (1/4 deg). Also, please suggest what other inputs would be required.

Best,

JVP

olson

 

 You will need at least the analyses of the state variables

 T, U, V, Q, and Ps.  You will need to regrid these to the model

 vertical and horizontal grid.

  Era-Interim analyses exist for the period you are interested in.

  If you have access to NCAR HPSS, they can be found here:

 /FS/DSS/DS627.0/ei.oper.an.ml/199707

 

  I believe they are on the /glade directory as well.  For details, view:

 

http://rda.ucar.edu/datasets/ds627.0/

Jerry Olson

jayant.pendhark...

Thank you for the information. I am in the process of getting access to the data. Need help in identifying tools for regridding the data for CESM. Also, is it correct to start with the initial conditions of July 1997? What other data (initial / boundary conditions) will be required?

JVP

olson

Especially since you are looking at El Nino/La Nina, you would want to switch

to the FAMIPC5 compset which will automatically switch you to the monthly mean

SST analyses rather than climatology.  I don't think you need other datasets.  I believe

the appropriate datasets will be loaded under FAMIPC5 for the period you are interested

 in.  I cannot help you with the decision on the appropriateness of July 1997.  The period of choice

 is highly dependent upon the issue you are studying.

 

Jerry Olson

jayant.pendhark...

I am running the FAMIPC5 compset with SE dynamical core at 0.25 degree resolution. the 'env_mach_pes.xml' suggest to use 1024 TOTALPES. However, it is taking more than 24 hours to run one month simulation. Is it that slow or I overlooked something?

JVP

olson

 

 I haven't run ne120 yet but am about to myself.  But I believe 1024 is the default for

 1 deg.  Try increasing the processor count by a factor of 16.  But 1 month in 24 hours

 still sounds too slow to me even with only 1024.  There might still be something else you're overlooking.

Jerry Olson

olson

NCAR's NCL package has good tools for regridding

Jerry Olson

Log in or register to post comments

Who's new

  • rlove@...
  • katsumi@...
  • afox
  • shanru@...
  • yongxiao@...