如 何 利 用 'tcorr' 命 令对 CMIP5数 据 计 算 相 关 性

liuzf406

I am trying to use 'tcorr' command to calculate correlation between ENSO3.4 and 500-mb-hgt. All data are winter (djf) mean time series derived from CMIP5 data, and the script is below. I run this successfully on reanalyis data, but it doesn't work for CMIP5 data.

'reinit'
'sdfopen ts_Amon_IPSL-CM5A-LR_historical_r1-6i1p1_185001-200512.djf.mean.nc'
'sdfopen zg_Amon_IPSL-CM5A-LR_historical_r1-6i1p1_500mb_185001-200512.djf.mean.nc'
'set x 1'
'set y 1'
'set z 1'
'set t 1 155'
'define enso34=aave(ts.1,lon=-170,lon=-120,lat=-5,lat=5)'
'set lon -180 180'
'set lat -90 90'
'set z 1'
'set t 1'
'define r=tcorr(enso34,zg.2,t=1,t=155)'

'd r'

But this fails to work, and the error shows as fellow:
SDF: The double precision values of the variable lon_bnds will be automatically converted to single precision
SDF: The double precision values of the variable lat_bnds will be automatically converted to single precision
....
Data request warning: request beyond file limits
....

Any suggestion would be greatly appreciated.

沙颖凯

The error is about sdfopen, have you checked the version of both GrADS and CMIP5 datasets?

liuzf406

沙颖凯 发表于 2014-11-15 06:48
The error is about sdfopen, have you checked the version of both GrADS and CMIP5 datasets?

Thanks for your kind help!
The version of Grads I used is GrADS19, because it can open a 365-day calendars file using 'sdfopen'.