A Shady Topic – Hillshade layers

This post looks at hillshade layers – how they are used and created and how our brain interprets them.

I had an interesting mapping experience last winter.  While taking a break in the lodge at Mt Cheeseman, I was confronted, yes, confronted, by a strange phenomenon.  Now I take a fair bit of pride in my map reading and sense of roughly knowing where I am but this really threw me for a loop.  Perhaps it was someone’s sick idea of a joke, but what I saw was a 1:50,000 scale topographic map covering the Cheeseman area, pinned to the wall (wait for it), upside down.

At first glance I felt both disoriented (sorry, that’s a North American holdover – “disorientated” still sounds odd to me even after all these years) and slightly nauseous though it wasn’t immediately clear why.  When I focused in on the ski field, it got even worse but I slowly began to see why.  Does the image below look a bit strange to you?

Maybe it’s just me but when I look at this, what I know to be ridges look like river valleys and vice-versa.  The rivers look like they’re running uphill and flowing on the ridges.  And it appears like you have to drive downhill from a ridge to get to the Mt Olympus ski field.  If you’re seeing the same thing, you’re not alone.  It’s quite a common occurrence when maps get turned upside down.  Here’s that same area right-side up (well, with north at the top, to be more precise – who am I to say what the right way up is?):

Is that more like it?  Part of what’s going on here is the clash with what we expect to see on maps (like rivers running downhill) but this mostly has to do with what we do with the shadows.  Many people aren’t necessarily conscious of it but on this map (and most other maps you might see) there’s a background layer that simulates shadows and our brains are conditioned to use that to infer topography.  Look along the Craigieburn Range in the image above (especially if you’re familiar with the area).  The shadows help define where the ridges are and darker shadows imply steeper slopes.  And you probably did that without even thinking about it.  The fall of the shadows helps give us a sense of the topography in a far more intuitive way than contours ever could.  These layers are called (surprisingly enough) hillshade layers.  Let’s look at a few examples.

Here’s an area from Lyttelton Harbour.  I’ve added a basemap here to give it a bit of context:

And here’s the hillshade layer for the same extent:

Now I’ll combine these by layering the basemap on top of the hillshade and setting about 25% transparency to the basemap so we can see through to see the hillshade layer – does it make a difference?  Is it a better map?

Without a lot of extra overhead, we can now intuit the topography straightaway.  To create a hillshade layer, we need a digital elevation model (DEM) – a raster grid of elevations.  With a DEM as an input, the Hillshade tool first derives both slope and aspect from the elevations and then uses those values to calculate how much shade each grid cell gets as a result of the topography.  The tool window below shows the important parameters used:

As an input here I’m using a grid called “bpdem” (Banks Peninsula DEM – look for it in J:\Data\Digital_Elevation_Models) and saving the output as “HS”.  Two key parameters set where the sun is in the sky – Azimuth is the compass angle to the sun and Altitude is its angle above the horizon: the defaults are 315 and 45 degrees respectively which put the sun in the northwestern part of the sky halfway to overhead.  And therein lies a key point: you can set where the sun is in the sky with this tool – and this has a direct impact on where the shadows fall. It’s not quite Maui stopping the sun, but does give you a wee sense of power.  Ticking the “Model shadows” box does a slightly better job of creating realistic shadows.  (In the pre-digital days, these hillshades were hand drawn or airbrushed onto maps.)

So I’ll now rerun the the hillshade tool on the data shown above but I’ll put the sun in the opposite part if the sky – the southeast.  Here’s the hillshade:

And then with the basemap:

Does this give you a different sense from the previous map?  It’s a bit like what we saw at Cheeseman above and in a sense, it’s the same thing.  The hillshades on our national 1:50:000 scale topo maps are made with the sun in the northwest.  When we turn the map upside down, it’s as if the sun is coming in from the southeast – and it’s our brains that make it appear like high is low and vice-versa.

Another example just looking at hillshade layers now.  The first is the hillshade for Banks Peninsula based on the DEM (also in J:\Data\Digital_Elevation_Modles) with the sun in the northwest and 45 degrees above the horizon:

Looks okay, huh?  Here’s a hillshade with the sun in the southeast this time:

Do they look significantly different to you?  Here’s one more image that might make it a bit easier to see:

Those dots with light at the top (i.e. north) probably appear as bumps while those with light coming from the south appear as depressions.  I’ve yet to find a good explanation for this – it does appear to be something hard wired into our evolved brains.

It’s an interesting phenomenon, even more so when one considers that almost every map you may look at, regardless of what area it shows, has the sun in the northwest.  That’s right – northern hemisphere maps use shadows from directions that are physically impossible – here’s an example from the UK’s Ordnance Survey showing a portion of the Highlands of Scotland – where’s the sun?:

And here’s the northern tip of Iceland:

Weird, eh?  But this shading works on a subconscious level to help us better appreciate the topography without having to try too hard, even though it’s quite impossible.

Hillshades are invaluable mapping tools and I personally consider them to be one of the most effective cartographic techniques around; with the right data, they are very easy to create.  Have a go on your next map!

C