Anyone know of a greyscale map of Earth Topography?

[Malenfant]

I'm looking for a global map of the Earth which shows (or can be persuaded to show) land topography AND seafloor topography as an 8-bit greyscale image - i.e. with black being the deepest point and white being the highest point on the planet. I don't mind if it's in one piece or if it's in tiles that can be stitched together, so long as it's the same dataset. Resolution should be as high as possible (1 or 2 minute?) preferably, as long as vertical resolution is on the order of a metre or so. The sort of thing I can use to figure out the shape of coastlines if the sea level rises a bit.

Sorta like GLOBE, but with the seafloor on it too.

Any pointers would be appreciated!

[Malenfant]

Just an update: I found this:

http://www.ngdc.noaa.gov/mgg/image/2minrelief.html

But when I download and open the raw data as a raw image file in photoshop I get a weird greyscale palette where all the seafloor topo is clumped at the white end of the greyscale and all the land topo is clumped at the black end. Which is really annoying since this seems to be the only example of the sort of map I'm looking for...

Unless someone else can give me any pointers?

[t00fri]

Just an update: I found this:

http://www.ngdc.noaa.gov/mgg/image/2minrelief.html

But when I download and open the raw data as a raw image file in photoshop I get a weird greyscale palette where all the seafloor topo is clumped at the white end of the greyscale and all the land topo is clumped at the black end. Which is really annoying since this seems to be the only example of the sort of map I'm looking for...

Unless someone else can give me any pointers?

This is the standard NOAA topography data, but they are stored as 16bit integers (rather than 8bit)! That's why they look weired when displayed in 8bit PS routines. I have worked with these data a lot, but explaining in detail what you have to do is somewhat involved. The easiest and best quality option is to use Chris' normal map converter that works directly with 16 bit raw input data. The other method is to switch to lab mode in PS, use the 'curves' tool and filter the grayscale data with a special curve profile ( an 'S' shape clockwise rotated by 90 degrees and centered around 128 (medium gray) ). This gives the desired appearance. You then truncate to 8bit in PS.

Note that this requires a certain familiarity with PS and the respective image manipulation methods.

A third method is to employ a NOAA software tool that allows to display and save those data as JPG's . Forgot how the tool is called. I tried it once but never used it, of course, since it vasts a lot of image quality....

Bye Fridger

[cartrite]

Did you try to download the raw data? You may be able to use 3dem to open the .bin files.
Then save as USGS dem in ascII. Then use Wlibur.exe to convert to a greymap. I never did
it with these files but I did do this with other NOAA files when I was making a normal Map of earth. Those files had a missing Sea Floor though.

cartrite

EDIT: Sorry. It's early. I seen you did download the raw data.
If you want to try this the link for 3dem is:

http://www.visualizationsoftware.com/3dem.html

and the link for Wilbur is:

http://www.ridgenet.net/~jslayton/wilbur.html

[Malenfant]

This is the standard NOAA topography data, but they are stored as 16bit integers (rather than 8bit)! That's why they look weired when displayed in 8bit PS routines.

I should have mentioned - I did open it as a 16 bit image in Photoshop (opened as a 10800x5400 16-bit, 1 count IBM-PC greyscale image). Unless you mean that photoshop has trouble actually displaying those even in 16 bit mode.

The easiest and best quality option is to use Chris' normal map converter that works directly with 16 bit raw input data.

Is that available anywhere?

The other method is to switch to lab mode in PS, use the 'curves' tool and filter the grayscale data with a special curve profile ( an 'S' shape clockwise rotated by 90 degrees and centered around 128 (medium gray) ). This gives the desired appearance. You then truncate to 8bit in PS.

I tried doing that but it still didn't look right (I'm probably not getting the shape of the curve right). I suppose I'd have to invert the palette first so that the seas are dark and the land is bright?

A third method is to employ a NOAA software tool that allows to display and save those data as JPG's . Forgot how the tool is called. I tried it once but never used it, of course, since it vasts a lot of image quality....

I tried GEODAS Hydroplot and that was useless, it didn't even give an option to show greyscale contours. Was that what you were thinking of?

[Malenfant]

EDIT: Sorry. It's early. I seen you did download the raw data.
If you want to try this the link for 3dem is:

http://www.visualizationsoftware.com/3dem.html

Aha! That did it. I could save it as a BMP in 3DEM so I didn't need Wilbur.

Problem solved. Thanks!

[rra]

Malenfant,

I know that most of the popular programs hardly handle
16 bit gray (or 48 bir color) very well.
I did some exeriments with PaintShopPro and Photoshop a year ago with simple operations (scaling, convolution and such).
May be the latest versions do work well.

Also converting to gray will yield a depth resolution of only 256 levels
which is never enough to map the whole vertical range with 1 m. resolution.
Would be something like 60 meters or so I guess.

Ren?©

[t00fri]

This is the standard NOAA topography data, but they are stored as 16bit integers (rather than 8bit)! That's why they look weired when displayed in 8bit PS routines.

I should have mentioned - I did open it as a 16 bit image in Photoshop (opened as a 10800x5400 16-bit, 1 count IBM-PC greyscale image). Unless you mean that photoshop has trouble actually displaying those even in 16 bit mode.

That's what I meant. In PS 7.x only a /very/ limited number of routines are "16bit clean".

Also you might have to play with the "little endian/big endian" alternatives offered. PS calls it IBM-PC vs MAC ? I forgot the precise names they use. This does not refer to what YOU use in your PC but to the manner how the 16 bit numbers are stored in the file. That again depended on what machine (UNIX or PC) was mainly used for THEIR analysis. Typically things only work with one of the two options.

The easiest and best quality option is to use Chris' normal map converter that works directly with 16 bit raw input data.

Is that available anywhere?

Certainly: the source code (nm16.cpp) is part of the official Celestia sources (=>CVS) since long. The easiest compilation for Windows users is certainly via CYGWIN. The reason being that nm16 works with STDIN and STDOUT input/output such that nm16 can be "piped" together under UNIX/LINUX with other applications. CYGWIN allows to make use of this.

The trick about nm16 is that it does all compilations of the normal map in 16bit mode and truncates only at the very end to 8 bit.
That's why nm16 has very little noise.

The other method is to switch to lab mode in PS, use the 'curves' tool and filter the grayscale data with a special curve profile ( an 'S' shape clockwise rotated by 90 degrees and centered around 128 (medium gray) ). This gives the desired appearance. You then truncate to 8bit in PS.

I tried doing that but it still didn't look right (I'm probably not getting the shape of the curve right). I suppose I'd have to invert the palette first so that the seas are dark and the land is bright?

I don't recall precisely, but I saved the shape somewhere. If I find it, I send it to you.

Bye Fridger

[Malenfant]

WEll the point of this is to make a map of the earth with raised sealevels. Unfortunately the program I usually use to render this (POVRay) can only accept 8bit greyscales as heightfields, which means the vertical resolution (as rra said) is pretty low.

I can use 3DEM to show higher sealevels but it seems that the projection of the Topo2 heightmap doesn't quite match the earth maps I have.

Anyone got any tips on how I can do this? (Don?)

[Cormoran]

Malenfant,

I don't know if you've tried it or possess a copy, but Fractal Terrains Pro might be worth a look. I've used it to examine precisely the opposite issue to yourself, that of lowered sea-levels. Its not free, but I considered it worth every penny when I bought it.

I mention it because it handles map projection and reprojection quite well (at least from this layman's perspective) - Also it does icosahedral 'projection' output, which should appeal to the Traveller in you

Cormoran

[cartrite]

Malenfant,
I'm confused. You say that you used 3dem to open the file and saved it as a bmp. Didn't the bmp file come out as 1/10 'th scale? When I save file in image formats with 3dem they are a .10 of the original size. Thats why I save as USGS DEM. That way I get the full size.

cartrite

[Malenfant]

Malenfant,
I'm confused. You say that you used 3dem to open the file and saved it as a bmp. Didn't the bmp file come out as 1/10 'th scale? When I save file in image formats with 3dem they are a .10 of the original size. Thats why I save as USGS DEM. That way I get the full size.

cartrite

I could adjust the size up to 3000x6000 in 3DEM and then save it as a BMP (it's one of the function key options. F6 I think). That's more than enough for my purposes.

Do you know of a visible light earthmap that corresponds to the ETOPO2 data? The continent borders on the ETOPO2 map seem to be a bit out-of-sync with the earth maps that I currently have.

[cartrite]

Okay, Sorry, I'm UNconfused now. I forgot why I stoped trying to save image files with 3dem. I could only scale them to a little over 50%.
6000x3000 is .56 the size of the original. Did you also scale your visible light earthmap to .56?

cartrite

Edit: What is the original size of your visible earthmap?

Edit2: You mentioned above that you oppened as IMB-PC. As Fridger said above, that is little-endian. The site you found from NOAA has 2 files available. ETOPO2.dos.bin (little-endian) and ETOPO2.raw.bin (big-endian).
Download the right File? 3dem has an area to switch between the 2 when first opening the file. Map coordinates and other settings. It starts with little endian by default. I forget to change that alot. Just checking.

cartrite

[Malenfant]

My visible earthmap is from an entirely different (and now unknown) source - it's a completely different size to the ETOPO2 one (much smaller)... haven't tried finding a more uptodate one, maybe there's something on the motherlode that fits.

And yes, I downloaded the right version (dos.bin) and used the little-endian version. If you try opening it the wrong way then it looks obviously wrong .

[scaddenp]

Etopo2 data regarded as "projected' bitmap isnt going to match any
standard projection. Each cell is the height for a 2min square. It has
a number of errors especially when you match to coastline. You might
like to look also at
ftp://falcon.grdl.noaa.gov/pub/walter/G ... _blend.bi2
which blends the 1min Gebco data with the S&S map, fixing a numbers
that are in etopo. Be very careful of using this data in other contexts
though without making yourself aware of how it is derived.

The problem of projection is a bit of pain. We use Etopo2 and the gebcoSS
blend with GMT to map in any projection we like. You could use this tool
to do same but I hesitate to recommend as it has a big learning curve.
However if you want to overlay on say a mercator projection for the world
you will need software that is capable of doing the projection.

[cartrite]

The BMNG has a gebco_bathy file and a srtm_ramp2 file. I downloaded the small ones. With the Gimp and the gebco file, I painted 2 white squares at the bottom corners. Then I selected by color the land mass (black-000000 in hex). Then I inverted the selection and pasted as new. Then with the brightness and contrast tool, Darkened it twice at (-127) and Brightened the srtm file twice at +127. Then I combined the 2 files. I'm not sure if this is what your looking for but here is a 1024x512 image of the result. It should match earthmaps but there are problems along the coast. Mabey you can play with this with your image editor and get the results you want.

The link is: http://spike.arsc.edu/nasa/bmng/topography/

and here is a thumb:



If you like NOAA files They could probally be combined with the NOAA files and the files with NO ocean data. Good Luck

cartrite

[cartrite]

Malenfant,
If you are still working on this project, I noticed a capabilitly that 3dem has which may help. If you go into (color scale/modify scale) and set the terrain type as Lowlands or Ocean/Island, you can raise or lower the sealevel. If you are just raising the sealevel you could use the NOAA maps and set the sealevel to what ever and save as.

cartrite

[Malenfant]

Thanks, I figued out that 3DEM could show different sea levels so I just used that to show the whole ETOPO2 map and then rendered things in that.