@James Jones great question, and honestly one that comes up a lot when people start planning real deployments rather than just benchtop testing.
The 50km figure is the drone-to-drone spec — both modules elevated, ground reflections eliminated, true open sky. That's the ceiling. For your agricultural deployment where the nodes are sitting on the ground or on low masts, you're looking at the ground-to-ground range which is closer to the 5km spec, and that's where terrain actually starts to matter.
The good news for your use case is that flat ag land is genuinely close to ideal for ground-level deployment — gentle undulation and the odd tree row will cost you some margin but shouldn't be catastrophic, especially if you can get the nodes even a meter or two off the ground on a pole or post. The honest answer is that a few rows of trees will hurt more than gradual elevation changes, because foliage at 2.4GHz is pretty lossy, but you're likely still looking at solid multi-kilometer links rather than total failure.
If you want to stress-test your node layout before committing, I'd suggest doing a quick walk-test with two units and the Connect app before finalizing positions — you'll see real RSSI feedback and can identify any dead zones that the terrain map wouldn't have predicted.
