Best settings to optimize bathymetry data

[y3nd]

Here are some produced tiles

The main issue is not the tile size but the extreme amount of polygons at low zoom levels.
I have tried many settings like polygon drops, coalesce etc.. without success in reducing the complexity without breaking the file completely. The source geojson is a 17 multipolygon file for each possible depth.

What would be the best tracks to reduce the complexity and display a lower amount of polygons at low zoom level ?

Thanks

[e-n-f]

If your polygons were pancaked instead of adjacent, I would suggest setting a tippecanoe:minzoom for every other contour level to make it drop out at lower zooms, but it looks like you need continuous coverage. Do you have the option of setting a larger tile size limit?

What we do at Felt is:

• set a 2.5 MB tile size limit with -M2500000
• set a lower tile resolution at low zooms with -D10
• use --no-simplification-of-shared-nodes to prevent introduction of shard gaps between adjacent rings
• use --coalesce-densest-as-needed to collapse small rings into larger ones when necessary

[laem]

Hi ! If only I had found this issue earlier : I've gone the same path, and got the same weight problems.

Bathymetry data makes MapLibre's new globe beautiful, but a lot heavier.

https://dev.cartes.app/#4.23/48.66/7.99

In case you've managed to lower the weight on low zoom, I'd be quite interested with the solution.

Current download size is ~2 Mo on high zoom, which is not unbearable but still quite heavy.