Erosion improvements

[JakBB]

Hello,
first of all I want to thank Stephen for the amazing new update.
The speed of WM is really quite incredible.

I’ve been experimenting with it for the last few days and am having troubles to get the erosion to look as I intend.

I will list a few improvements / issues I have with the current erosion device:

1 - COARSE DEPOSITION - maintaining detail in soil depositions. Right now the soil depositions look as flat as snow

in nature erosion depositions usually look distinctly rougher, as the terrain doesn’t erode into infinitely fine particles, but rocks tend to break off into differently sized chunks.
Right now you could emulate a rougher look in the depositions by displacing the areas with a fine noise, but if the erosion device would simulate this the results would vastly improve, as the erosion would interact with the noisy boulders and create organic flow patterns around the bigger boulders.
I don’t know how feasible this is. If there is any way to randomize the particle size, so as to make some particles from the erosion simulation bigger than others.
Also in the previous old erosion device I noticed that the relief of the pre eroded terrain always kind of stuck out from the deposition laying on top, which helped with maintaining details in deposition areas.

2 - GULLIES - in line with the 1st suggestion, right now I’m unable to create natural erosion gullies, which happen as a sort of cascading domino effect and create beautiful flow patterns on softer parts of the landscape.

Here an example of gullies on mars.
Also a really good post on erosion simulation in Houdini by Sam Krug https://www.artstation.com/artwork/BkwZJz

3 - CLIFF PRESERVATION - kinda considered as a holy grail by many, I still haven’t found a way to properly conserve the cliffs of my terrain while adding strong erosion details to the flatter parts of my terrain. The new Cliff Limit parameter is a good step forwards, but I’m still struggling. I will try to explain in the following image

Here I’m really digging the triangular shapes created on the flat plateaus, but I’m really bothered by the many small channels on the cliff faces. I would want the sediments from the flatter areas to fall off the cliffside, without doing much cutting into the rock.
These are the erosion settings from the screenshot.

This makes me wonder why I’m not able to change the bedrock strength all the way to 1.
I’m also wondering what the bedrock structure input on the erosion device is supposed to do, as currently it isn’t affecting the output in any way.

4 - SLIPPAGE CONTROL - I often use an erosion device to create more realistic erosion deposits compared to thermal weathering, usually by max combine the terrain before and after erosion

What I’m missing from the thermal weathering in this is the ability to control the angle of deposition.
That’s why I wonder if its possible to control the “slippage” of erosion deposition, kind of changing the gravity, or changing the weight / sliding for each particle.

That about sums it up. If I get other ideas, or if you have any, just reply here.

I really feel that the erosion algorithm is the core and workhorse of World Machine (and other terrain programs) and although it just recently got an amazing update I felt the need to address some of its shortcomings.

[Tomatiy]

I’m not user of erosion device, but may answer on some of points:

1 - This possible to do with Thermal Wethering device. Just play with Talus Size or scale, dont remember name of setting

2 - Also looks like Thermal Wethering device

3 - Thermal Wethering also helps to add sediments with saving of cliffs (but i may be wrong)

4 - o, lol, you already know about Thermal Wethering - sorry, maybe my answers gives no profit

[Stephen]

Hi there,

There are plenty of questions and ideas here, so I’m only going to hit a couple of them right now:

#1 I believe what you’re looking for is the Soil Diffusion parameter - adjust it down lower than 1.0 to make soil less diffusive (smooth). This also influences the creep of soil downslope, so if your feature scale is set very high it will have more dramatic global as well as local influence.

#4 Just to clarify - Thermal erosion allows you to control the angle of deposition. In particular, if you set the angle of repose to what you want and the mass balance to 1.0, bedrock will not be broken down - I believe achieving the “max” style operation that you’re looking for.

I don’t know if those solve your problems or not. I am currently updating the help center documentation for the new Erosion and Thermal Erosion devices to help explain all of this!

[Tomatiy]

okey, maybe i should ask chatgpt, but want to try ask here:

what “bedrock” means in terms of such software and general geology?

I do not know english, also know this word only from minecraft, so i’m confused to read this each time, eheheh

[WFab]

@Tomatiy Here ya go:

Bedrock

[Stephen]

I had an interesting thought occur tonight regarding your third question, and my answer to #4. This is not a final answer, but rather, a beginning of sorts…

Part A) So, roughly speaking, high gradient cliff areas SHOUD be profoundly eroded. However… material in nature often doesn’t erode isotropically - erosion occurs along weak planes in the rock, etc. As a result, it certainly doesn’t always have the strongly fluted/channeled character that often results from erosion in WM.

Part B) The mass balance parameter in Thermal erosion is a powerful tool to put your thumb on the scale, and cause either more depositional or erosive character.

…

→ A ‘mass balance’ parameter would thus be very useful for erosion for the same reasons its useful in Thermal Weathering.

The soil scour parameter is very close to this - if you set it to 1.0, the terrain will be eroded but no soil produced, causing extremely channeled character, as if a flood has carried away all the topsoil and talus. Set to 0.0, it is neutral. In my experiments, I found negative soil scour interesting because it simulates the volumetric expansion that occurs when bedrock is broken into talus and sediment - the same amount of mass takes up more space. Thus WM lets you set scour negative. The principle result in terms of look, however, is just “more soil”, which is useful but not profoundly interesting.

If negative values are instead interpreted as a mass balance parameter, things get interesting. Negative values now mean “produce soil, but don’t erode the bedrock by the proportional amount”. This is, in effect, what you are asking for in #3 above!

I conducted a little experiment to investigate the resulting character, and the results are quite promising:

Scour 0:

Scour -0.5:

You can see in this test scene, the resulting “positive mass balance” : the cliffs are not eroded severely while still producing depositional soil. The result is also, I think, quite pleasant… and is an interesting erosional character that was hard to capture before, and a good antidote to “my terrain looks too channeled”.

And of course, as a spatial parameter, you can vary it across the terrain to combine both styles…

I’m presenting this discovery in realtime here, essentially. There are some stability issues with negative scour in the simulation that would need to be overcome to be production quality, so it’s probably not reasonable to change for Hurricane Ridge’s release. But I DO think it very much worth pursuing further.

[JakBB]

Thanks for the replies.
It certainly helped to get some insight as to what the new erosion parameters can do, although I could have looked at the tooltips, I completely forgot about those.
The soil diffusion parameter is really quite wonderful.

Here are the results after playing around some more

I’m using a instance scatter for the scattered rocks to add more detail on top of the already detailed depositions.
I’m also using a slope mask to control erosion rate and bedrock strength, and this helps preserve cliff faces, although I really wonder why I’m not able to go below or above certain values. (erosion rate can’t go below 0.05 and bedrock strength above 0.95)

The negative scour pictures you posted look really neat, especially considering being able to control those spatially.
In connection to my 4th question in the original post, do you think it would be possible to control the slippage of the erosion sediments? In the screenshot with Scour -0.5 the taluses have a certain angle, but say I would want these sediments to form taluses which are less steep?
I know this function is already included in the thermal weathering device, but the taluses created by that device are more crude basic as the ones made by erosion

As for the generation of gullies / erosion channels I feel this is really missing in every terrain I’m creating.
I also realized that having this effect as a controllable erosion would allow for the creation of canyon like terrains, with deep erosion channels caused by flash floods or periods of heavy traumatic water erosion. Only when this happens in real life the underlying rock structure gets exposed, creating the usually blocky features of the grand canyon or similar. This wouldn’t be possible in WM and would have to be faked in post, because only the overlaying rock structure can be displayed by a heightmap.
Here a few more screenshots of gully erosion in the wild:

[Stephen]

One further answer:

The structural input works in concert with the structural magnitude parameter to simulate the natural variance in erosion alluded to earlier, where fracture planes and other macro-scale crystallize features end up dominating cliff faces.

If you do not supply any input, World Machine will default to basic noisiness. This is often good enough! Crank up the structural parameter and you’ll see smooth surfaces in the input become jagged and rough.

But if you supply the structural values, you can control what that emergent structure looks like. Good structure patterns to supply include gradients, voronoi cells, and more. Here is an example with a voronoi-cell style structure:

Or with the basic built-in:

Or with the structure magnitude set to zero (off):

[JakBB]

Really interesting. I will have to play around more with this structure system.

I’ve been able to produce some results with it, but I’m unsure if I understand how it works.

I guess that’s the tendency for this new erosion device. Really powerful features but hard to control if you don’t know intimately how each of the settings affect the outcome.
Maybe some example files should be included for the full release.
I might also experiment and build a few presets to share here on the forum which utilize all the parameters to its fullest, spatial and everything.

[Stephen]

Just a followup here. Build 4044 is now available, and it incorporates some of the ‘mass balance’ thoughts described above, although in different final form than I originally envisioned.

Rather than provide a mass balance parameter, I decided it made more sense in the context of the erosion device to remap that function to two existing parameters.

1. Bedrock strength previously governed the rate at which rock could be turned into sediment. It still does, except now it does not reduce the amount of sediment produced.

So for example, turning bedrock strength to 1.0 will preserve all existing highpoints of the terrain, while still producing eroded “stuff”. This is a violation of conservation of mass, but it’s very useful.

Soil scour is now fully responsible for what happens to that ‘stuff’, above. So setting bedrock strength and scour to the same value will produce the same results as before - but now you can vary them indepently as well. In practice this is extremely useful - you can use soil scour both positive and negative to tune the amount of relative deposition vs wear that occurs on the terrain.

Finally, the water power parameter has been expanded slighty and assumes now a dominant position among the parameterizations exposed. Give it a try.