Clutter data describes real world things on the planet’s surface that attenuate signals, for example trees, shrubs, buildings, bodies of water, etc, etc. There’s also different types of trees, some types of trees attenuate signals more than others, different types of buildings are the same.
Getting clutter data used to be crazy expensive, and done on a per country or even per region basis, until the European Space Agency dropped a global dataset free of charge for anyone to use, that covered the entire planet in a single source of data.
So we can use this inside Forsk Atoll for making our predictions.
First things first we’ll need to create an account with the ESA (This is not where they take astronaut applications unfortunately, it just gives you access to the datasets).
Then you can select the areas (tiles) you want to download after clicking the “Download” tab on the right.

We get a confirmation of the tiles we’re download and we’ll get a ZIP file containing the data.

We can load the whole ZIP file (Without needing to extract anything) into GlobalMapper which loads all the layers.

I found the _Map.tif
files the highest resolution, so I’m only exporting these.

Then we need to export the data to GeoTiff for use in Atoll (The specific GeoTiff format ESA produces them in is not compatible with Atoll hence the need to convert), so we export the layers as Raster / Image format.

Atoll requires square pixels, and we need them in meters, so we select “Calculate Spacing in Other Units”.

Then set the spacing to meters (I use 1m to match everything else, but the data is actually only 10m accurate, so you could set this to 10m).

You probably want to set the Export Bounds to just the areas you’re interested in, otherwise the data gets really big, really quickly and takes forever to crunch.
Now for the fancy part, we need to import it into Atoll.
When we import the data we import it as Raster data (Clutter Classes) with a pixel size of 1m.

Alas when we exported the data we’ve lost the positioning information, so while we’ve got the clutter data, it’s just there somewhere on the planet, which with the planet being the size it is, is probably not where you need it.
So I cheat, I start put putting the West and North values to match the values from a Cell Site I’ve already got on the map (I put one in the top left and bottom right corners of the map) and use that as the initial value.
Then – and stick with me, this is very technical – I mess with the values until the maps line up into the correct position. Increase X, decrease Y, dialing it it in until the clutter map lines up with the other maps I’ve got.


Right, now we’ve got the data but we don’t have any values.
Each color represents a clutter class, but we haven’t set any actual height or losses for that material.

To know what each colour means we need to RTFM – ESA WorldCover 2020 User Manual.
Which has a table:

Alas the Map Code does not match with the table in the manual, but the colours do, here’s what mine map to:

Which means when hovering over a layer of clutter I can see the type:

Next we need to populate the heights, indoor and outdoor losses for that given clutter. This is a little more tricky as it’s going to vary geography to geography, but there’s indicative loss numbers available online pretty easily.
Once you’ve got that plugged in you can run your predictions and off you go!