Re: High Pressure Ridge Identification Algorithm?

Hi Alex,

My apologies, I didn't look closely enough at the local_min
documentation before I wrote that print statement. local_min returns the
index value of the latitudes and longitudes, not the latitudes and
longitudes themselves. So:
t = hgtnew&time
lat = hgtnew&lat
lon = hgtnew&lon
do i = 0,tend
    .....
    print("For t = "+i+", local_min exists at "+lat(nmin_at_yi)+"N, \
        "+lon(nmin_at_xi)+"E, with a value of "+nmin_at_minval)
end do

As far as writing out the local_mins to a text file, that can get a bit
tricky. Each time through local_min can return a different number of
local minima. This is a problem, since in order to write all the data at
once (using write_matrix) we will have to store the local_min data at
every timestep in one data array.

Thus, you will have to ID the max number of local minima returned from
local_min, and set up an array based on that number * the total number
of timesteps. So (untested):

t = x&time
nmax = 0
do h = 0,dimsizes(t)-1
    nmin = local_min(hgtnew(h,{60:90},:), True, 0.0)
    nmax = max((/nmin,nmax/))
end do
print("Maximum number of minima = "+nmax)

nmax = nmax*dimsizes(t) ; set up array at max possible size

ncol = 4
data = new ( (/nmax,ncol/) , typeof(x) )
cntr = 0
do i = 0,tend
    nmin = local_min(hgtnew(i,{60:90},:), True, 0.0)
    data(cntr:cntr+nmin-1,0) = t(i)
    data(cntr:cntr+nmin-1,1) = nmin_at_minval
    data(cntr:cntr+nmin-1,2) = nmin_at_xi
    data(cntr:cntr+nmin-1,3) = nmin_at_yi
    cntr = cntr+nmin
end do

opt = True
opt_at_fout = "alexg.txt"
write_matrix( data(:cntr-1,:), "f10.1,3f10.2", opt)

Adam

Alex Gardner wrote:
> Thanks Adam,
>
> I have 2 more questions:
>
> When I constrain my 2-D field to between 60 and 90 deg N ("(i,{60:90},:)" I
> get output for between 1-5 deg N. Any clue why this might be? I'm using NCEP
> netCDFs
>
> Instead of printing the output to the screen, is there a way to have each
> loop's output added to a text file?
>
> Cheers,
> Alex
>
> My Scrip now looks like this:
>
> begin
> ;
> tend = 24
> ;
> a = addfile("$NCARG_ROOT/netcdf/hgt.2002.nc","r")
> hgt=a->hgt(:,5,:,:)
>
> hgtnew = (hgt*hgt_at_scale_factor) + hgt_at_add_offset
> copy_VarMeta(hgt,hgtnew)
>
> do i = 0,tend,1
>
> x = hgtnew(i,{60:90},:)
> nmin = local_min(x, True, 0.0)
>
> print("For t = "+i+", local_min exists at "+nmin_at_yi+"N, \ "+nmin_at_xi+"E, with
> a value of "+nmin_at_minval)
>
>
> -----Original Message-----
> From: Adam Phillips [mailto:asphilli_at_cgd.ucar.edu]
> Sent: Wednesday, May 03, 2006 2:03 PM
> To: Alex Gardner
> Cc: ncl-talk_at_ucar.edu
> Subject: Re: High Pressure Ridge Identification Algorithm?
>
> Hi Alex,
>
> local_min / local_max only accept 2 dimensional data:
>
> http://www.ncl.ucar.edu/Document/Functions/Built-in/local_min.shtml
>
> As you have 3D data, you need to put in a do loop for the time dimension:
>
> t = x&time
> do gg = 0,dimsizes(t)-1
> nmin = local_min(x(gg,:,:), True, delta)
> print("For t = "+t(gg)+", local_min exists at "+nmin_at_yi+"N, \
> "+nmin_at_xi+"E, with a value of "+nmin_at_minval)
> end do
>
> Adam
>
> Alex Gardner wrote:
>
>>Thanks again Dennis, you've been a great help.
>>
>>I've tired running the scrip but I keep getting the error message:
>>
>>"Fatal: Number of dimensions in parameter (0) of (local_min) is (3), (2)
>>dimensions were expected"
>>
>>The scrip I have so for is:
>>
>>a = addfile("$NCARG_ROOT/netcdf/hgt.2002.nc","r")
>>x=a->hgt(:,5,{60:},:)
>>
>>t = x&time
>>ntim = dimsizes(t)
>>
>> delta = 0. ; user specified
>> nmin = local_min(x, True, delta)
>>
>> ncol = 4
>> data = new ( (/ntim,ncol/) , typeof(x) )
>> data(:,0) = t
>> data(:,1) = nmin_at_minval
>> data(:,2) = nmin_at_xi
>> data(:,3) = nmin_at_yi
>>
>> opt = True
>> opt_at_fout = "alexg.txt"
>> write_matrix( data, "f10.1,3f10.2", opt)
>>
>>end
>>
>>The problem has to do with the time component of the field "hgt".
>
> Local_min
>
>>doesn't seem to work when all time steps are extracted.
>>
>>Do you know how can I solve this issue?
>>
>>Alex
>>
>>-----Original Message-----
>>From: Dennis Shea [mailto:shea_at_cgd.ucar.edu]
>>Sent: Tuesday, May 02, 2006 2:46 PM
>>To: ncl-talk_at_ucar.edu; alexg_at_ualberta.ca
>>Subject: RE: [ncl-talk] High Pressure Ridge Identification Algorithm?
>>
>>
>>
>>>Thank you very much for the help Dennis,
>>>
>>>The min and max functions will do what I want. What I eventually want to
>
> do
>
>>>is output the min value, lat and long for each month of a 2 year monthly
>>>mean netCDF file for the region North of 60 deg North. Ideally I would
>
> like
>
>>>to run a script that could output a single txt file with the year, month,
>>>value, lat, and long for each month of the 24 months the netCDF. Does
>>
>>anyone
>>
>>
>>>have a script where they have used the locmin or locmax functions with a
>>>netcdf file and output the results as a txt file?
>>>
>>>Any help would be greatly appreciated.
>>
>>
>>quick/dirty/ and untested
>>
>> f = addfile (... , "r")
>> ; read data nort of 60
>> x = f->X(:,{60:},:) ; x(time,lat,lon) => (24,:,:)
>> t = x&time
>> ntim = dimsizes(t)
>>
>> delta = ... ; user specified
>> nmin = local_min(x, True, delta)
>> ; ????
>> ncol = 4
>> data = new ( (/ntim,ncol/) , typeof(x) )
>> data(:,0) = t
>> data(:,1) = nmin_at_minval
>> data(:,2) = nmin_at_xi
>> data(:,3) = nmin_at_yi
>>
>> opt = True
>> opt_at_fout = "alexg.txt"
>> write_matrix( data, "f10.1,3f10.2", opt)
>>
>>
>>
>>>
>>>
>>>-----Original Message-----
>>>From: Dennis Shea [mailto:shea_at_cgd.ucar.edu]
>>>Sent: Monday, May 01, 2006 3:35 PM
>>>To: ncl-talk_at_ucar.edu; alexg_at_ualberta.ca
>>>Subject: Re: High Pressure Ridge Identification Algorithm?
>>>
>>>
>>>
>>>>I know it's a long shot, but does anyone know if there is an atmospheric
>>>>ridge detection (from netCDF files) function that has been developed for
>>>>NCL? I'm looking to develop statistics on ridging event over a particular
>>>>region and was hoping that maybe someone has done similar work.
>>>>_______________________________________________
>>>
>>>I know of no ready made software.
>>>
>>>---
>>>Having said that ...
>>>
>>>What do you mean by "atmospheric ridge detection"?
>>>
>>>Are you talking about atmospheric "blocks" ?
>>>ie: blocking highs , blocking patterns
>>>
>>>---
>>>There are functions that can automate detection
>>>of local min and max of 2D fileds
>>>
>>>
>>>http://www.ncl.ucar.edu/Document/Functions/Built-in/local_max.shtml
>>>
>>>An example is at:
>>>http://www.ncl.ucar.edu/Applications/minmax.shtml
>>>
>>>good luck
>>>
>>
>>
>>
>>_______________________________________________
>>ncl-talk mailing list
>>ncl-talk_at_ucar.edu
>>http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>
>

-- 
--------------------------------------------------------------
Adam Phillips			             asphilli_at_ucar.edu
National Center for Atmospheric Research   tel: (303) 497-1726
ESSL/CGD/CAS                               fax: (303) 497-1333
P.O. Box 3000				
Boulder, CO 80307-3000	  http://www.cgd.ucar.edu/cas/asphilli
_______________________________________________
ncl-talk mailing list
ncl-talk_at_ucar.edu
http://mailman.ucar.edu/mailman/listinfo/ncl-talk