Hm, yes, I can reproduce those step changes along the equator, although curiously not at lat=0, lon=0. Did you notice them anywhere other than for tiles having an edge on the equator? Do they only occur over water, or also over land tiles?
I didn't check in tileedit yet, but I would be surprised if it shows up there. All ocean tiles should have an elevation of 0, and the Earth elevation data have a vertical resolution of 1m, so a step of <1m can't be represented in the data. For now I don't have a strong suspicion, but maybe something is going wrong in the geometric transformation for the tile (in which case it should also occur over land).
I didn't check in tileedit yet, but I would be surprised if it shows up there. All ocean tiles should have an elevation of 0, and the Earth elevation data have a vertical resolution of 1m, so a step of <1m can't be represented in the data. For now I don't have a strong suspicion, but maybe something is going wrong in the geometric transformation for the tile (in which case it should also occur over land).