@Kevin good question, and honestly a pretty punishing deployment environment — metal enclosure in direct sun can easily push junction temps well beyond the ambient 60°C you're seeing at the case.
On the TX power drift question: the STM32WB07 at the core of the MultiNav Pro+ has decent on-chip temp compensation, but like any RF front-end, the external PA and matching network components will have their own thermal coefficients that can cause some output variation at the extremes. In my experience the drift is usually more of a concern at the hot end than cold — at -20°C the silicon is actually happier in most respects, but startup behavior can be quirky if you're powering on from cold.
The sensors module pairing you're planning is actually a smart move here — if you're logging the onboard temp continuously you can correlate any link quality drops or RSSI changes against thermal events in post, which gives you real data on whether your specific enclosure and mounting situation is causing issues rather than guessing. That telemetry loop is genuinely useful for this kind of deployment characterization.
One thing worth thinking about: if that metal enclosure is sealed, you're also going to have condensation risk on the transition from -20°C nights back up to daytime temps, which can be a bigger killer than the heat itself. If you haven't already, worth factoring some conformal coating or at minimum desiccant into the build.
