Work in progress: shaded-relief geological map
Posted 01 November 2010 - 07:05 AM
This is part of a shaded-relief geological map covering an area in the Pang Valley, which is in West Berkshire, ten-miles or so from where I live. The minimal grey base map was made using OS OpenData VectorMap District layers, the solid and drift geology layers are WMS layers from the British Geological Survey and the hill-shade tiff was made from an OS OpenData DEM. The contours were supplied as a DXF from the OS OpenData Landform PANORAMA data set (with no contour labels, or none that I can find!). The grid interval is 1000 metres and the contour vertical interval is 10 metres.
The geology is the story here, but of course geology and landform run together. However, I am concerned that the hill-shading might be affecting adversely the colours used for the geology, making identification confusing. I'm making a symbol set for the drift and this should help, there will also be a key for the solid and for the drift.
A poor thing but mine own, as they say. Comments on where I am now would be appreciated.
Posted 01 November 2010 - 08:09 AM
I like the style of your map.
My remarks: I think that the hill-shading is too strong and affects the clearness of your map. First I thought that the different colours represent elevation levels and I had to re-read your description to see that it's the geological categories. Then I wanted to ask: Does your map show the embankments/cuttings only in selected places (along the east-west divided road) or is the display of such symbols already complete? Are you planning to add bridge symbols? I think that if you make the hill-shading less strong and add (more) bridge/embankment/cutting symbols to the roads the map would be even more clear than it is now.
Posted 01 November 2010 - 08:46 AM
Posted 01 November 2010 - 10:25 AM
But you should somhow make the gological boundaries stand out more. I also don't understand why some of them are grey and some green. Is the colour somehow related to stratigraphic position, or does it have a tectonical meaning? In general, stratigraphical boundaries (where one layer of rock lies on top of another because it was deposited there) should be shown in one symbology, while tectonic boundaries (faults of any sort, basically where rocks broke apart and got shifted sometime after sedimentation) should be shown in another symbology. On Swiss Geological maps, stratigraphical boundaries are shown in black or dark grey (no matter whether you're looking at rocks or unconsolidated material like gravels, sands etc.), faults in red or in some cases dark blue (indicating the sense of movement).
I would use a filled symbol for buildings. Something like a medium-dark grey (60-70-%-ish) might work.
On the maps of the so-called Swiss Geological Atlas (1:25'000), they use a dark brown for all topographic features like buildings, roads, contours, even all the labels. this works really well, but then again, the topographic details are only there for orientation, the single main purpose of the maps is to show the geology, so topography really "steps in the background".
Hope this helps.
Posted 01 November 2010 - 10:42 AM
Posted 01 November 2010 - 03:34 PM
Johannes - The hill-shading is a difficult one because I am anxious to preserve the relief aspect of the map. I shall however remove the hatchure for the cuttings, because this doesn't add much to the map.
Dave - I'll definitely try a version without contours. The two geology layers are WMS PNGs, which will make strengthening the shape outlines an interesting exercise. It can done and there are artefacts (lines) in the layers that could do with getting rid of at the same time.
Kathi - As mentioned, I will try and sort out the problem with the shape outlines, at present the drift (superficial) is outlined in green and the solid (bedrock) in grey. The OS VectorMap District polygons in the 'settlement' layers (i.e. buildings) are pretty dreadful and I honestly doubt if it would add much to emphasise this by adding fill to them. Yes, if I keep the contours then they really ought to be labelled, or done away with altogether and replaced with some spot heights perhaps. Dark brown for the background mapping is certainly something that I'll try.
Jrat - thanks for the comments.
We are away for a week now, trying to find some sun, but I'll start on the revisions as soon as we return.
Posted 02 November 2010 - 01:52 AM
... at present the drift (superficial) is outlined in green and the solid (bedrock) in grey.
Ok, I see now. I was guessing something along those lines but it wasn't intuitively clear to me. I think the map would become more readable (at least to as a geologist) if the drift and bedrock were symbolized in clearly different colour schemes. Traditionally, drift is shown in mostly white with black or blue overprints (dots, triangles, dashes,... ) or very pale colours, while bedrock has more saturated colours.
You might also want to check out other geological maps of either the same area or of areas where rocks of the same age outcrop. There's a colour code for stratigraphic units that geologists have more or less agreed upon, which makes a geological map instantly much more readable. (Check the USGS website, for example.)
Posted 02 November 2010 - 11:26 PM
One suggestion: The shaded relief and the contour lines are showing two different levels of detail, which for me creates an unsettling dissonance.
The contours suit the level of detail in the geological data, but the relief shading seems too detailed. I'm not a geologist, but I'm thinking that at this scale the micro-lumpiness of the land surface doesn't add to the understanding of the larger landforms and the geology.
It would be interesting to see how it looked without the contour lines, but the geology seems to be correlated with them so it would be a shame to lose that relationship. With labels, the contours would allow you to compare the elevations of widely separated outcrops of the same material and see the slope of the beds (? not a geologist). With just shaded relief, you lose that. I would vote for both.
Are there any other DEMs that you could try? If you are into PhotoShop, it may be easier to add detail to an overly-smooth hillshade than to try to remove detail.
Posted 19 November 2010 - 07:28 AM
Attached is a JPG showing progress to date, its nominal scale is 1:25,000 . The consensus was that the contours ought to go, so I have removed them. Relief mapping seems to entail smoke, mirrors and other visual trickery and I thought that removing the contours might reduce the effect of the trick, but I hope this hasn't happened.
I've simplified the place-name labels and have also tried to put consistent borders on the geology shapes (this took some doing, the shapes were originally WMS layers delivered as PNGs). The cutting hatchure has gone and I have added labels and symbols to the geology shapes (with a key to follow).
These are still a couple of line artefacts on the map that I am having great problems removing.
Further comments would be greatly appreciated.
Posted 20 November 2010 - 03:03 AM
Nick my first observation is that areas of lighter brown are to vivid. Otherwise it is good-looking map and hillshading is fine considering the source resolution.
Posted 20 November 2010 - 10:31 AM
I'm going to dissent from the other opinions of the relief shading however. It is a little too strong, too much contrast. The light side is ok but your shadows are too dense and it's a little too high in the hierarchy over all. Could be softer. Not sure how your are creating the relief but you can fix in several ways:
1) adjust the whole image to be slightly lighter, but this may push the light side to pure white which can look solarized.
2) adjust the sun angle, bring the sun higher in the sky basically to reduce the contrast between both sides.
3) increase the ambient or scattered light value (if you have control over this). This keeps the sun angle and light side density but adds a more realistic reflected light to the shadows side reducing it's density.
GIS Reference and Instruction Specialist, Stanford Geospatial Center.
Posted 30 November 2010 - 07:21 AM
Contrary to others, I think the hillshade is fine as it is.
0 user(s) are reading this topic
0 members, 0 guests, 0 anonymous users