[gradsusr] question about calculation sequence

[LI Qi]

Dear GrADS community,


I would like to ask, if what I want is only one number, which is the correct sequence to calculate temporal statitics between forecast and observation.


For example:
the variable is hourly precipitation: var(t, y, x), 
aim: to calculate temporal correlation


method 1:
first, calculate the areal mean of var ---> get fcst_fldmean(t) and obs_fldmean(t)
then, calculate temporal correlation of fcst_fldmean(t) and obs_fldmean(t) ---> get one number


method 2:
first, calculate the temporal correlation of fcst var and obs var ---> get var_timcor(y,x)
then, calculate the areal mean of var_timcor(y,x) ---> get one number


Which is the correct calculation sequence?
Any hint would be much appreciated!


Best,
Qi
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://gradsusr.org/pipermail/gradsusr/attachments/20151127/2eea1397/attachment.html