[GRASS-dev] SRTM plus color table

[Michael Barton]

I agree with your synopsis. 

I started with the ETOPO2 ocean colors to see if I could improve them somewhat. 

I substituted your rule for the one I had for -100. Good on the continental shelf, but it still looked too dark in the deeps. Here is a little more modification. 

There are various ways of tweaking color tables for particular locales. I'm looking for something of a compromise that works reasonably well for most ocean depths. Probably not really possible, but I thought I'd play around with it some. I think this one is an improvement over what I did earlier.

-11000 0:0:0
-8000 0:0:50
-5000 10:10:100
-3000 30:30:150
-1000 70:70:200
-100 100:100:225
0 150:150:255

Just the ocean part.

Michael  
____________________
C. Michael Barton
Director, Center for Social Dynamics & Complexity 
Professor of Anthropology, School of Human Evolution & Social Change
Arizona State University

voice: 	480-965-6262 (SHESC), 480-727-9746 (CSDC)
fax:          480-965-7671 (SHESC),  480-727-0709 (CSDC)
www: http://www.public.asu.edu/~cmbarton, http://csdc.asu.edu





On Jan 27, 2013, at 9:13 PM, Hamish <hamish_b at yahoo.com>
 wrote:

> Michael wrote:
>> So, a little trial and error
>> experimentation produces this:
> ...
>> ...with the following color rules.
>> 
>> -11000 0:0:0
>> -8000 0:0:50
>> -5000 10:10:70
>> -3000 30:30:100
>> -1000 70:70:170
>> -100 100:100:200
>> 0 150:150:255
>> 0.1 57:151:105
>> 100 117:194:93
>> 200 230:230:128
>> 500 202:158:75
>> 1000 214:187:98
>> 2000 185:154:100
>> 3000 220:220:220
>> 5000 250:250:250
>> 8850 255:255:255
>> nv 255:255:255
>> default 255:255:255
> 
> I'll respond the long way, with some personal cosmetic
> observations and critique on ocean colors. YMM(& probably will)V.
> 
> my test data is the etopo1 global elevation dataset.
> See Global Datasets on the grass wiki for links.
> 
> r.colors etopo1 color=<>
> 
> etopo2 -- ocean colors are not too bad. land colors much too
> green then harsh transition to white.
> 
> 
> srtm -- only the shallow shelves are visible, most of the ocean
> is black. Georges Bank off Cape Cod and the North & Baltic Seas
> show up quite well (often < 100m). ~600m is already black, but
> for a global view 3000m should still be visible. I find srtm on
> land to be too brown, but in coastal areas it can look ok, e.g.:
>  http://adhoc.osgeo.osuosl.org/livedvd/docs/_images/grass-fractal.png
> 
> 
> elevation -- relative % based, only intended for land.
> 
> 
> terrain -- a mix of the green from etopo2 with the black oceans
> of srtm! worst of both worlds? elevation based.
> 
> 
> haxby -- specifically created by W.F. Haxby for highlighting
> terrain on the sea floor, but for segments of the sea -- not
> the full range. See also "Digital images of combined oceanic and
> continental data sets and their use in tectonic studies" in AGU's
> EOS from a couple months ago:
>  http://onlinelibrary.wiley.com/doi/10.1029/EO064i052p00995/abstract
> The haxby rules are relative % based scale, so you need to crop
> out the ocean first:
>  r.mapcalc "etopo1.sea = if(etopo1 <= 0, etopo1, null())"
>  r.mapcalc "etopo1.land = if(etopo1 >= 0, etopo1, null())"
> 
> (but even then the scale is too wide, just zoom in on one part
> of the ocean floor to use them)
> 
> 
> SRTMplus -- the shallow seas have a rather coral-reef look to
> them, structure of the deep sea seems a bit blurred out.
> 
> 
> proposed (above/dropbox'd): the moonlight effect of the sea
> floor is nice, although it's a little bit purplish. You can see
> a bit more structure in the Baltic, and still some structure on
> Georges Bank and the Hudson Canyon. Around the Bahamas the
> transitions are a bit harsh.  Like "srtm" I find the land to be
> too brown for my taste.
> 
> 
> One thing I often do with ocean floor plots is to put the land in
> greyscale, or as a greyscale shaded relief, with the blue->black
> ocean. That bypasses any high desert vs. ice cap coloring issues.
> (shaded relief works best in a non-LL projection)
> 
> d.erase 60:60:60
> r.colors etopo1.sea color=etopo2
> d.rast -o etopo1.sea
> d.rast -o etopo1.land.shade    # lowered resolution to 0:05 during creation else it looks to granular
> 
> 
> in summary, etopo2 is my favourite for the deep oceans, but may
> be too bright if you are strictly looking at the upper 100m of the
> continental shelves, in that case "srtm" may be better.
> 
> adding the following 100m rule to "etopo2" seems to look nice for
> adding detail in the shallow seas:  (e.g. east coast of the US
> and the North/Baltic Seas)
> 
> -100:100:100:225
> 
> (see attached screenshot; but you really need to zoom into a
> shallow sea like the Bay of Fundy or the Baltic to really see
> the difference)  It makes the Med look pretty nice too, but that
> already looked nice with the "etopo2" rules.
> 
> I propose to add the above rule to "etopo2".
> 
> 
> regards,
> Hamish
> 
> 
> -----
> fwiw,
> G> r.univar etopo1.sea -e percent=0,10,20,30,40,50,60,70,80,90,100
> minimum: -10803
> maximum: 0
> range: 10803
> mean: -3433.38
> standard deviation: 1716.47
> ...
> 0th percentile: -10803
> 10th percentile: -5286
> 20th percentile: -4902
> 30th percentile: -4532
> 40th percentile: -4226
> 50th percentile: -3890
> 60th percentile: -3509
> 70th percentile: -2958
> 80th percentile: -1863
> 90th percentile: -253
> 100th percentile: 0
> <etopo1_shaded_relief2.jpg>