Re: How to do polar plots with logarithmic axis in radial coordinate?

[Paul van Delst <pvandelst(at)ncep.noaa.gov>]

Charlie Zender wrote:
> 
> Craig Markwardt wrote:
> 
> > Could you simply take the ALOG10() logarithm of the data before
> > plotting it?  Easier to re-label the axis than re-invent the world...
> >
> 
> This would cause the radial coordinate to be negative-valued which
> would have unpleasant results. It's possible someone could get
> this method to work but I tried without success.

I did the following:

IDL> r=10.0^(findgen(100)/20.0)
IDL> theta=findgen(100)/5.
IDL> !p.charsize=2.5
IDL> plot, alog10(r), theta, xsty=4,ysty=4,/polar
IDL> axis,0,0,xax=0,xtickformat='expticks_log'
% Compiled module: EXPTICKS_LOG.
IDL> axis,0,0,yax=0,ytickformat='expticks_log'

where the expticks_log.pro is:

FUNCTION expticks_log, axis, index, value
  tickmark = '10!E' + $
             STRTRIM(STRING(value,FORMAT='(f5.1)'),2) + $
             '!N'
  RETURN, tickmark
END

The !P.CHARSIZE was just so I could see the exponents. Is this what you are looking for?
Keep in mind that the format string for the expticks_log.pro won't work for exponents less
than 0.0 - you'll have to get smart about checking for the range and then setting the
format string based on that (rather than the f5.1 above, e.g.:

  CASE 1 OF

;   -- Exponent is less than zero ->
;   -- fractional ticklabel
    ( exponent LT 0 ): format = '( f' + $
                                STRTRIM(ABS(exponent)+2,2) + $
                                '.' + $
                                STRTRIM(ABS(exponent),2) + $
                                ' )'

;   -- Exponent is greater than or = to zero ->
;   -- whole number ticklabel
    ( exponent GE 0 ): format = '( i' + $
                                STRTRIM(ABS(exponent)+1,2) + $
                                ' )'

  ENDCASE

and then use the format string in the return value, e.g.

  RETURN, '10!E' + STRING( value, FORMAT = format ) + '!N'

or something similar depending on your tastes/needs.


Hope some of this is helpful, although I have to admit, the fact that IDL doesn't have a
stock polar plotting routine that produces a circular graph with the radial and concentric
circle tickmark axes is a bit ridiculous. Farting about with /POLAR and AXIS and whatnot
is sort of like using OG to plot, x, y - and in the end you still end up with
Cartesian-like axes.

Maybe there is a polar plotting routine out there in IDL software land somewhere. It is
sorely needed.

paulv

-- 
Paul van Delst           A little learning is a dangerous thing;
CIMSS @ NOAA/NCEP        Drink deep, or taste not the Pierian spring;
Ph: (301)763-8000 x7274  There shallow draughts intoxicate the brain,
Fax:(301)763-8545        And drinking largely sobers us again.
pvandelst@ncep.noaa.gov                   Alexander Pope.