Re: [ferret_users] highlighting the region where most of the model agrees in sign

[saurabh rathore <rohitsrb2020@xxxxxxxxx>]

Dear Russ and Ryo,

Thank you for your suggestions. The method Suggested by Russ worked well.

Cheers, Saurabh

On Tue, Jul 17, 2018 at 6:03 PM saurabh rathore <rohitsrb2020@xxxxxxxxx> wrote:

Thank you Russ and Ryo,

I will revert you on getting the results in couple of days.

cheers , saurabh

On Tue, Jul 17, 2018 at 5:57 PM, Russ Fiedler <russell.fiedler@xxxxxxxx> wrote:

Hi,

I think it might be easier to assign +/-1 to regions depending on the sign of the trend. If you take the sum along the ensemble axis and get absolute values >=7 then you know that you must have had 9 or more ensembles with the same sign. i.e. >80%

let tsign = if trend ge 0 then 1 else (-1)
let agree_80 = if abs(tsign[m=1:11@sum]) ge 7 then 1


Russ

On 17/07/18 17:36, Ryo Furue wrote:

Hi Saurabh,

On Tue, Jul 17, 2018 at 1:29 PM saurabh rathore <rohitsrb2020@xxxxxxxxx> wrote:

I want to highlight those regions where linear trend of 80% models agree in sign. 

Ferret includes "@" operators to count the number of "events" along an axis.  So, I guess you want first to create a mask that indicates whether the trend has the same sign as the reference model and then count the number of "yes"s along the ensemble dimension using one of those operators.

Cheers,

Ryo

--

REGARDS

Saurabh Rathore

Research Scholar (PhD.)

Centre For Oceans, Rivers, Atmosphere & Land Science Technology

Indian Institute Of Technology, Kharagpur

contact :- 91- 8345984434

--

REGARDS

Saurabh Rathore

Research Scholar (PhD.)

Centre For Oceans, Rivers, Atmosphere & Land Science Technology

Indian Institute Of Technology, Kharagpur

contact :- 91- 8345984434