How do you handle range maps when species records are patchy?

I'm working on a species distribution study where the occurrence records are very uneven across the region, and it's hard to tell whether gaps reflect real absence or just missing sampling.

I've tried a couple of approaches but none feel fully convincing.

How do you decide when a range map is reliable enough to publish or share with managers?