[WIP] Atlas Maps of Valmere | From Real-World Height Data to Vector Map

[CredePendrel]

While I still haven't completed my world's custom height maps yet I wanted to go through the whole mapping process with some data I did have available. The colour scheme is the default one from ArcPro but will make a custom ramp later on. I did very little data clean up which is obvious as you zoom and see in the one image.

https://drive.google.com/drive/folde...H7?usp=sharing

[CredePendrel]

I started creating my own custom heightmap based on this tutorial and others scattered around the forums and finding it much harder than I thought. I figured it would be easy to recreate the workflow I used for the images in the previous post, which also uses real world data, but no. Some of the relief (generated with Contour-Polygons) looks ok but most of the contours themselves make so sense.

Heightmap in Wilbur


Contour Polygons


Some real ugly contour lines at the coast

[morne]

I started creating my own custom heightmap based on this tutorial and others scattered around the forums and finding it much harder than I thought. I figured it would be easy to recreate the workflow I used for the images in the previous post, which also uses real world data, but no. Some of the relief (generated with Contour-Polygons) looks ok but most of the contours themselves make so sense.

Heightmap in Wilbur


Contour Polygons


Some real ugly contour lines at the coast

The main issue here is adjusting for the coastline. You should be deciding what your "sea level" is in grayscale, and making sure that you are tapering off your coasts. It looks like you are doing that a little bit, but it's a very uniform and dramatic taper. If you look at real-world coastlines, it's more varied than that. If you do want certain parts of the coast to be steep/cliff-like, I've also found that doing targeted erosion in those areas can really help make it look more realistic and match the coastline a little bit better.

I'm working on a map right now where I did just that:



The above map is a mountainous region with fjords similar to Norway. When I first put DEM data in, the terrain didn't really match the fjords of the coastline. So I took it into an erosion tool and targeted erosion around the coastline to make it match up

You can create masks to erode in Wilbur. You don't have a scale in your maps, but the colors suggest you have pretty high elevations. Remember, 90% of Earth's topography is below 3000 meters (as a rough estimate. Parts of high elevation should be carefully placed.

One other thing that might help is taking your time and using smaller bits of data. In your example above, it looks like you've used entire chunks of data that represent your entire island/continent (Even if you didn't do this, the following advice is still useful). I would recommend being more selective and cutting, rotating, and blending data that conforms more closely to the plan you have for your terrain/coastline. Working out a rough tectonic layout (doesn't need to be detailed by any means) can be extremely helpful for determining where to place mountainous regions, instead of placing them haphazardly.

[CredePendrel]

Thank you 'morne' for your response.

I was definitely rushing too fast with the heightmap so I could get something into ArcPro and keep testing. I WAS really happy how the images in the google drive link turned out and wanted to see how easy it would be to replicate it...turns out not so easy lol. I understand your comment about the coastline and why that is contributing to the dense contours there, thank you. Using targeted erosion sounds cool. Is that a tutorial or documentation about creating masks to erode in Wilbur?

Near the end of the evening I was getting frustrated by the results of the contour-polygons and absolutely starting pasting in larger chunks of data. I was getting very sharp contours and weird square geometric patterns and was trying to find the root cause. I will go back and start with much smaller chunks of data and trying blending more.

This morning I went back to the Madagascar data I used in the Google Drive link and tried a different technique to modify the landform with some mild success. I loaded the HeightMap into ArcPro and then started adjusting the water levels bit by bit essentially flooding the landscape. This produced some interesting new shapes and made a new landmass based on Madagascar but different. This could work really well, especially after rotating the shapes a bit.




In the example map you showed above: are the rivers and lakes derived from spatial analysis?

[morne]

Thank you 'morne' for your response.

I was definitely rushing too fast with the heightmap so I could get something into ArcPro and keep testing. I WAS really happy how the images in the google drive link turned out and wanted to see how easy it would be to replicate it...turns out not so easy lol. I understand your comment about the coastline and why that is contributing to the dense contours there, thank you. Using targeted erosion sounds cool. Is that a tutorial or documentation about create masks to erode in Wilbur?

Near the end of the evening I was getting frustrated by the results of the contour-polygons and absolutely starting pasting in larger chunks of data. I was getting very sharp contours and weird square geometric patterns and was trying to find the root cause. I will go back and start with much smaller chunks of data and trying blending more.

This morning I went back to the Madagascar data I used in the Google Drive link and tried a different technique to modify the landform with some mild success. I loaded the HeightMap into ArcPro and then started adjusting the water levels bit by bit essentially flooding the landscape. This produced some interesting new shapes and made a new landmass based on Madagascar but different. This could work really well, especially after rotating the shapes a bit.




In the example map you showed above: are the rivers and lakes derived from spatial analysis?

No tutorial, but I would create a b/w mask and, in Wilbur, do Selection -> Load Selection. Then perform erosion steps as normal. I actually have started using Genbrush (https://mavichist.itch.io/genbrush) for targeted erosion and that works pretty well. Be aware, though, that I have to manually convert the raw data to tif to get my results out, since the current application only supports exporting 8-bit.

I've tried using the Arcgis Pro workflow for deriving rivers and lakes, but it ends up being difficult because of the necessity to perform a fill step. If there are too many basins that need to be filled you end up with a lot of straight rivers that pass through the filled basins which looks bad. Currently, I use World Machine to derive rivers and lakes, which for whatever reason their "find water" algorithms are more forgiving and gives better results. I take the raster output rivers from World Machine, then use the Raster -> Polyline in Arcgis Pro (Raster -> Polygon for lakes). Then I work on cleaning up errors and combining polylines to form "full" rivers. I export that into a vector program like Affinity, and create the tapered effect using variable stroke, adjusting all the rivers manually. I expand the stroke out and import back into ArcGIS Pro as polygons. It's a frustratingly manual process but it's better than hand-drawing rivers. The only thing this method lacks is any way to get river deltas.

[CredePendrel]

Wow I thought I was beginning to get the hang of it but the workflow gets more complex lol. I love World Machine and have Alpine Lakes but am not very good at it. I am much better at creating terrains and textures in Gaea but it lacks some key features WM does right now, like an actual river system. I haven't tried a heightmap produced in Gaea in ArcPro yet, but sounds like something I can try tonight.

Your process to get the rivers back into Pro sounds good albeit, fully manual like you said. I found this old blog post for ArcMap about tapering streams/rivers but it should still work in Pro.
https://www.esri.com/arcgis-blog/pro...m-to-your-map/


Some Random Gaea Maps
https://drive.google.com/file/d/1inZ...ew?usp=sharing
https://drive.google.com/file/d/14Jm...ew?usp=sharing
https://drive.google.com/file/d/1X1T...ew?usp=sharing

[cyanide_baby]

No tutorial, but I would create a b/w mask and, in Wilbur, do Selection -> Load Selection. Then perform erosion steps as normal. I actually have started using Genbrush (https://mavichist.itch.io/genbrush) for targeted erosion and that works pretty well. Be aware, though, that I have to manually convert the raw data to tif to get my results out, since the current application only supports exporting 8-bit.

I've tried using the Arcgis Pro workflow for deriving rivers and lakes, but it ends up being difficult because of the necessity to perform a fill step. If there are too many basins that need to be filled you end up with a lot of straight rivers that pass through the filled basins which looks bad. Currently, I use World Machine to derive rivers and lakes, which for whatever reason their "find water" algorithms are more forgiving and gives better results. I take the raster output rivers from World Machine, then use the Raster -> Polyline in Arcgis Pro (Raster -> Polygon for lakes). Then I work on cleaning up errors and combining polylines to form "full" rivers. I export that into a vector program like Affinity, and create the tapered effect using variable stroke, adjusting all the rivers manually. I expand the stroke out and import back into ArcGIS Pro as polygons. It's a frustratingly manual process but it's better than hand-drawing rivers. The only thing this method lacks is any way to get river deltas.

Hello,

Sorry for reviving this thread, but I'm very curious about your river tapering technique using Affinity Designer.
Tapering indeed works very well when dealing with a single, no-branch, river, but it can't seem to work when dealing with a multi branch river.

Any tips?

Note: my river network is generated from my fantasy DEM. I import the un-tapered vectorized river network in Affinity Designer.