[gradsusr] 99% confidence level

Alan Robock robock at envsci.rutgers.edu
Thu Jan 26 09:24:42 EST 2017


Actually it is much better to shade or put dots on the regions that are 
NOT statistically significant.  That way you are covering information 
that is not important rather than covering the data you want people to see.

Alan

Alan Robock, Distinguished Professor
   Editor, Reviews of Geophysics
Department of Environmental Sciences             Phone: +1-848-932-5751
Rutgers University                                 Fax: +1-732-932-8644
14 College Farm Road                  E-mail: robock at envsci.rutgers.edu
New Brunswick, NJ 08901-8551  USA     http://envsci.rutgers.edu/~robockhttp://twitter.com/AlanRobock
Watch my 18 min TEDx talk at http://www.youtube.com/watch?v=qsrEk1oZ-54

On 1/26/2017 2:51 AM, Andrew Friedman wrote:
> Another option for plotting the significance areas (steps [2] and [3] below) is to use pattern filling: http://cola.gmu.edu/grads/gadoc/gradcomdsettile.html
> In this case, you can draw a tile over statistically significant regions directly in GrADS without needing to create shapefiles.
>
> -Andrew
>
>
>> On Jan 26, 2017, at 2:52 AM, Lyndon Mark Olaguera <olagueralyndonmark429 at gmail.com> wrote:
>>
>> Hello,
>>
>>
>> This website provides an example of doing a statistical significance test using grads:
>>
>> https://nelson.wisc.edu/ccr/resources/grads/significance-scripts.php
>>
>> Read the instructions carefully.
>>
>> As for the dots in the figure for areas which are statistically significant. You can do the following:
>>
>> [1] Mask out areas that are greater than the t statistic that you are using.
>> [2] Save the masked output as a point shapefile. You can also do this in grads. Everything is in this website:
>> http://cola.gmu.edu/grads/gadoc/shapefiles.html
>> [3] Plot the basemap (the anomaly from your figure) then draw the shapefile.
>>
>> Hope this helps.
>> Goodluck!
>>
>> Lyndon Mark P. Olaguera
>> PhD Student
>> Monsoon Climatology Laboratory
>> Department of Geography
>> Faculty of Urban Environmental Science
>> Minami-Osawa Campus
>> Tokyo Metropolitan University
>>
>>
>> On Thu, Jan 26, 2017 at 9:18 AM, mehwish ramzan <mehwish.ramzan at gmail.com> wrote:
>> Dear Andrew and GrADS users,
>>
>> I want to prepare a figure showing regions statistically significant at the 99% confidence level (using Student’s t-test) by subtracting model from observation.
>>
>> Please see attached file as a sample.The statistically significant areas are presented as black dots in the figure.
>>
>>
>> How can i obtain such figure using GrADS.
>>
>>
>> Please guide me.
>>
>> Thank you
>>
>> With Best Regards,
>>
>> Mehwish
>>
>>
>> PS: The figure attached as sample is just for reference purpose.
>> Figure reference: http://www.nature.com/articles/srep11847/figures/1. Retrieved on 25/01/2017
>>
>>
>> _______________________________________________
>> gradsusr mailing list
>> gradsusr at gradsusr.org
>> http://gradsusr.org/mailman/listinfo/gradsusr
>>
>>
>> _______________________________________________
>> gradsusr mailing list
>> gradsusr at gradsusr.org
>> http://gradsusr.org/mailman/listinfo/gradsusr
>
> _______________________________________________
> gradsusr mailing list
> gradsusr at gradsusr.org
> http://gradsusr.org/mailman/listinfo/gradsusr



More information about the gradsusr mailing list