dear rneale:
Thanks for your help!
i find a problem that it has severe difference between the result of monthly average model output data runing with restart file and runing with initial file!
i do a experiment as thus:using the default setting of CAM3 model,with the climate SST file, integrate 6 months,then there is a restart file and a initial file in 0006-01-01. then ,i run the model with climate SST file with two files sepatately for one month.then ,i compare the result of monthly average ,there has a severe difference. then ,i think the result of restart run is more believable than initial run .
now, i think give a anomaly to the land suface temperature of east asia. based the result above,I think it is more believable to change the "ts" variable in restart file than in initial file. Is this right?
Thans!