[gradsusr] Standard deviation of an area

Thu Nov 20 12:39:26 EST 2014

Julio,

The result indeed looks fine, but I'm not sure about what it represents.
When you do:

set x 1
set y 1

It means (for my variable):

LON set to 0 0
LAT set to -90 -90

So the area is set to a point. I don't undersand the relation between these
coordinates and the LAT and LON limits you used to define the rectangle for
the "aave" function. So I'm not sure about what that plot represents. The
tloop-part is prefect.

Thanks once more,

On 20 November 2014 13:17, Julio Martinez <julioms24 at hotmail.com> wrote:

> David
>
> This is what I get when executing
>
> set y 1
> set x 1
> set t 1 last
> d tloop(sqrt(pow(aave(uwnd,lon=200,lon=300,lat=0,lat=30)-uwnd,2)))
>
> Im not sure if this is what you are expecting.
> Please let me know.
>
>
>
>
>
> Julio N. Martínez Sánchez
> M.C. en Oceanografía Física
> Consultor en Sitio
> Subgerencia de Pronóstico a Mediano y Largo Plazo
> Gerencia de Meteorología y Climatología
> Servicio Meteorológico Nacional
> Ext. 3156
>
>
> ------------------------------
> From: davidnielsen at id.uff.br
> Date: Thu, 20 Nov 2014 12:22:20 -0200
> To: sourabhbal at gmail.com
> CC: gradsusr at gradsusr.org
>
> Subject: Re: [gradsusr] Standard deviation of an area
>
> Sourabh,
>
> First, if you want it to run from t=1 to 21, you should fix the loop to
> "while (count < 22)", so that it will run until t=21.
>
> However, it will generate you a single number for each t: an average of
> the variable on the rectangle you defined, not a plot as you wish. I'm also
> curious to hear suggestions for that now.
>
> Good luck,
>
>
>
> On 20 November 2014 11:45, Sourabh Bal <sourabhbal at gmail.com> wrote:
>
> Hi David,
>
> Thanks for your suggestion. I tried with the following
> 'reinit'
> 'open SSt12.precip.lon-lat.allssws.ctl'
>
> 'x1=1000*24*60*60*(aprl+aprc)'
> while (count<1)
> 'set t' count
> 'define area=aave(x1,lon=120,lon=180,lat=-90,lat=90)
> 'd area'
> 'enable print figure1.zonalpre120-180E.mf'
> 'print'
> 'disable print'
> '!./gxeps -c -i figure1.zonalpre120-180E.mf -o
> SSt12.preci.120-180E.all.ssws.eps'
> count=count+1
> endwhile
>
>
> but does not yield my result.
>
> After running the script there is no error but the eps file produced is
> blank.
>
> Actually I want zonal average over longitude (120 to 180). The plot would
> be X as time ( t=1 to t =21) and Y as latitude (90N to 90S). More over I
> also want calculate statistical significance (t test values) over this plot
> ( may be in shaded).
>
>
> Please help me
>
> Sourabh
>
>
>
>
>
>
>
> On Thu, Nov 20, 2014 at 2:00 PM, David Nielsen <davidnielsen at id.uff.br>
> wrote:
>
> Julio,
>
> The way you suggested still returned a gridded result, not a number
> representig the area. Then I tried averaging the result over the same area
> (everything for each t-value):
>
>    define var1 = sqrt( pow( aave( sst, lat=..., lon=...)))
>    define var2 = aave( var1, lat=..., lon=...)
>
> Does it make any sense to you?
>
> Thanks again!
>
>
> Sourabh,
>
> So just checked, you can do:
>
> while (count < 22)
>    'set t' count
>    'define area = aave (variable-name, lon=120, lon=180, lat=-90, lat-90)
>    'd area'
>    count = count +1
> endwhile
>
> Works as fine as defining the area with x and y. Very small numerical
> difference though, don't know why.
>
> On 20 November 2014 10:23, David Nielsen <davidnielsen at id.uff.br> wrote:
>
> Sourabh,
>
> You can try:
>
> while (count < 22)
>    'set t 'count
>    'define average=ave("variable-name", x and y limits)'
>    'd average'
>    count=count+1
> endwhile
>
> You have to express lat and lon in terms of x and y in the "ave" formula.
> For that, open the file in Grads and:
>
> set lat -90 90
> q dim
>
> And it will tell you the equivalente x limits, for example.
>
> Not sure about this must, though. You can check it here:
> http://www.iges.org/grads/gadoc/gradfuncave.html
>
> Hope it helps!
>
> Julio, I will try your suggestion now! Thanks again!
>
>
> On 20 November 2014 09:39, Sourabh Bal <sourabhbal at gmail.com> wrote:
>
> Hello David and Julio,
>
> if I want to plot time-latitude cross section (Y and T varying), with the
> data being averaged over longitude (120 to 180), where shall I fix the
> longitude.
>
> set lat -90 90
> set lon ??????
> set t 1 21
> d tloop(ave(aprl,lon=120,lon=180))
>
>
> Sourabh
>
>
>
> On Thu, Nov 20, 2014 at 1:35 AM, Julio Martinez <julioms24 at hotmail.com>
> wrote:
>
> *David,*
>
> *Maybe you could try with*
>
> *d tloop(sqrt(pow(aave(uwnd,lon=200,lon=300,lat=0,lat=30),-uwnd,2)))*
>
>
>
>
> Julio N. Martínez Sánchez
> M.C. en Oceanografía Física
> Consultor en Sitio
> Subgerencia de Pronóstico a Mediano y Largo Plazo
> Gerencia de Meteorología y Climatología
> Servicio Meteorológico Nacional
> Ext. 3156
>
>
> ------------------------------
> From: davidnielsen at id.uff.br
> Date: Wed, 19 Nov 2014 21:13:14 -0200
> To: gradsusr at gradsusr.org
> Subject: Re: [gradsusr] Standard deviation of an area
>
>
> Hi, Julio,
>
> My problem is not with the loop in time. I am using a "While" sintax to
> execute the average along the series:
>
> while (count < "number of t-values +1")
>    'set t 'count
>    'define average=aave(sst,x=127.8,x=128.6,y=27.6,y=28.6)'
>    'd average'
>    average=subwrd(result,4)
>    rc = write ("c:\path\file.txt",average)
>    count=count+1
> endwhile
>
> This is working fine. What I need now, is a way to calculate the standard
> deviation of SST in the area of interest, also for each value of t, like
> something independent of time, for me to insert inside the loop.
>
> Everything Ive seen before is a standard deviation of the whole series
> (like below), and not for each time-value.
>
> 'define sdsst=sqrt(ave(pow(ave(sst,t=1,t=12)-sst,2),t=1,t=12))' ->
> calculates std of the series between t=1 and t=2. What I'd like to do is,
> for each t, an std of the area.
>
> Any suggestions?
>
> Thanks!
>
>
> On 19 November 2014 13:38, Julio Martinez <julioms24 at hotmail.com> wrote:
>
> *Hi David,*
>
>
> *I think you could try with tloop command...*
>
> Julio N. Martínez Sánchez
> M.C. en Oceanografía Física
> Consultor en Sitio
> Subgerencia de Pronóstico a Mediano y Largo Plazo
> Gerencia de Meteorología y Climatología
> Servicio Meteorológico Nacional
> Ext. 3156
>
>
> ------------------------------
> From: davidnielsen at id.uff.br
> Date: Wed, 19 Nov 2014 13:12:34 -0200
> To: gradsusr at gradsusr.org
> Subject: [gradsusr] Standard deviation of an area
>
>
> Dear all grads users,
>
> I would like to calculate the standard deviation of an area in Grads.
>
> I have a NetCDF wekly time series of SST (attached), and I am able to
> calculate the average of a certain area (defining LAT and LON) value using
> the "aave" formula, for each value of time, and generate a new series - ok.
> In addition to this seires, I would like to generate a series of standard
> deviations of this average, also for each time in the series. In other
> words, I need to to know, for a given area and time, what is the standard
> deviation of my variable?
>
> Please find attached an example of a NetCDF file I want to treat this way:
> 1. Calculate average of SST in a certain area (lat, lon) using aave, for
> each time value
> 2. Calculate standard deviation of SST in a certain area (lat, lon), for
> each time value
>
> Would anybody have any kind suggestions on how to address this #2 task?
>
> Many thanks in advance,
>
>
>
> --
> David Marcolino Nielsen
> Engenharia de Recursos Hídricos e do Meio Ambiente
> Universidade Federal Fluminense
> davidnielsen at id.uff.br
>
> _______________________________________________ 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
>
>
>
>
> --
> David Marcolino Nielsen
> Engenharia de Recursos Hídricos e do Meio Ambiente
> Universidade Federal Fluminense
> davidnielsen at id.uff.br
>
> _______________________________________________ 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
>
>
>
>
> --
> Dr. Sourabh Bal
> Post-Doc Fellow
> Institute of Meteorology, Freie Universitat, Berlin
>
>
>
>
> _______________________________________________
> gradsusr mailing list
> gradsusr at gradsusr.org
> http://gradsusr.org/mailman/listinfo/gradsusr
>
>
>
>
> --
> David Marcolino Nielsen
> Engenharia de Recursos Hídricos e do Meio Ambiente
> Universidade Federal Fluminense
> davidnielsen at id.uff.br
>
>
>
>
> --
> David Marcolino Nielsen
> Engenharia de Recursos Hídricos e do Meio Ambiente
> Universidade Federal Fluminense
> davidnielsen at id.uff.br
>
> _______________________________________________
> gradsusr mailing list
> gradsusr at gradsusr.org
> http://gradsusr.org/mailman/listinfo/gradsusr
>
>
>
>
> --
> Dr. Sourabh Bal
> Post-Doc Fellow
> Institute of Meteorology, Freie Universitat, Berlin
>
>
>
>
>
>
> --
> David Marcolino Nielsen
> Engenharia de Recursos Hídricos e do Meio Ambiente
> Universidade Federal Fluminense
> davidnielsen at id.uff.br
>
> _______________________________________________ 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
>
>

-- 
David Marcolino Nielsen
Engenharia de Recursos Hídricos e do Meio Ambiente
Universidade Federal Fluminense
davidnielsen at id.uff.br
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://gradsusr.org/pipermail/gradsusr/attachments/20141120/3a2325e8/attachment-0001.html 