Good effort.
Looking at your plot there could be more dynamics at play here, but hard to say as the two tests are not apples to apples.
In the shorter “200m” test were you travelling clockwise or anticlockwise? If clockwise, the LoRa disconnection could be coincident with the group of trees you went behind. E.g. red dotted line here:
Your “2km” test is only a line drawn clear of these tree obstructions so the circumstances of your actual path are not clear. Did you follow the same track as the first test and can you post it?
For example it’s possible it may not be a range issue per-se but something else like an inability to subsequently re-establish the connection after it drops. E.g. similar to this recent post re UHF: Corrections via UHF - Reach RS3 / RTK / PPK configuration - Emlid Community Forum
I’m not suggesting it is, but the underlying subsystems are very complex and the more consistent and detailed our tests are the better it will help Emlid to narrow down & fix.