Re: [ferret_users] Finding correlation

[Paulo Henrique <paulohsm@xxxxxxxxx>]

Hi Savin,

Ferret comes with a go-file called 'variance' to calculate some simple statistics, including time-correlation.

To use this tool you need to define 'quietly' two variables, P and Q, 

let/quiet P = var1 ! var1==var1(t)
let/quiet Q = var2 ! var2==var2(t)

and then call variance:

go variance

Some new variables will be created, such as that named 'correl'. Now you can 'list correl' to see correlation of var2 relative to var1. It's important to note that both variables need to be in the same time-axis.

As you said, you have to correlate point-to-point a gridded data relative to a point data. I have not tested this case, but I think it should work.

More information about variance's variables names and statistical indices calculated can be collected with 'go/help variance' command.


Hope it helps,

Paulo Santiago

2008/6/30 Savin <schand@xxxxxxxxxxxxxxxxxxxx>:

Dear ferret users,

 

I have a time series data of say, Variable A to be correlated to each grid point of a time series map of say Variable B. Since there are many grids and each grid has a time series, is there any easy way out for carrying out this correlation?

 

Your help will be very much appreciated.

 

Thanks,

Savin