# Re: [ferret_users] How to do regression analysis by using two dataset

Dear Ansley
I am providing more information for understood the problem. My data looks like

currently SET data sets:
1> precip_anm_summer_1979_2006.nc

2> sst_summer_1979_2006.nc

yes? sh gr PRECIP_JJA[d=1]
GRID GEV1
name       axis              # pts   start                end
LON13_116 LONGITUDE          104mr   31.25E               71.25W
LAT25_60  LATITUDE            36 r   28.75S               58.75N
normal    Z
TYEARLY   TIME                28 r   01-JUL-1979 00:00    01-JUL-2006 00:00
normal    E
normal    F
yes? sh gr sst_JJA[d=2]
GRID GCP1
name       axis              # pts   start                end
LON1_18116_146 LONGITUDE     131mr   30E                  70W
LAT25_65  LATITUDE            41 r   40S                  40N
normal    Z
TYEARLY   TIME                28 r   01-JUL-1979 00:00    01-JUL-2006 00:00
normal    E
normal    F

let pre_nw = precip_jja[d=1,x=72:76@ave,y=14:20@ave]
let sst_new = sst_jja[d=2]
let p = pre_nw
let q =  sst_new
go regresst
fill slope

The unit of the precipitation is mm/day and SST is °C
after regresst the data. I want  the unit of slope like this °C per 5 mm/day

Thanking you

On Fri, Mar 3, 2017 at 6:58 AM, Ansley C. Manke wrote:

Hello,

What you have done looks correct, but without more information it is hard to tell what may be happening.  I assume the two variables have the same time and depth axes. If that was not the case, then Ferret would issue error messages about the grids not matching, and not return a result.

Have you considered the units of the two variables in thinking about how the result should look?  One way to use Ferret to look at the data and make sure the variables are what you expect, is to do a PLOT/VS plot, to see the cloud of data:

yes? plot/vs/symbol p, q

Ansley

On 3/2/2017 8:59 AM, Abhishek Savita wrote:
Hi Ferreters

Have anybody done the regression analysis between two data set for example sst[lon,lat,time] and precipitation[lon,lat,time] with 53 years.  I have  done like this by seeing some mail archives.

use  precipitation_summer_1979_2006.nc

let prep_new = precip_jja[d=1,x=72:76@ave,y=14:20@ave]  ! I have taken precipitation on a box average
let sst_new = sst_jja[d=2]   ! Here I have taken SST for hole Indian Ocean
let p = pre_new
let q =  sst_new
go regresst
fill slope

Because I have used same data and time period but I am not getting the same as I have seen in Literature

Am I doing Right ?

Any suggestion are highly appreciated

--
Regards
Abhishek Savita

Research Scholar (Earth System Science Technology)
Center For Oceans, Rivers, Atmosphere & Land Science Technology
Indian Institute of Technology, Kharagpur
+91-8609704619

--
Regards
Abhishek Savita

Research Scholar (Earth System Science Technology)
Center For Oceans, Rivers, Atmosphere & Land Science Technology
Indian Institute of Technology, Kharagpur
+91-8609704619