Temperature swings opposite from those expected by the diurnal radiation pattern (warming during the day, cooling at night) are typically due to air coming in from other locations. Synoptic (large-scale) fronts can bring such rapid shifts, but often tend to be more sharp (rather than gradual) and in one direction (as the front passes). However occasionally there can be interesting oscillations near more stationary fronts, such as this event from Oklahoma in February 2018:
More often, smaller nighttime fluctuations have to do with more mesoscale (local) features. Different regions cool differently, and the pressure differences caused by the resulting differences in density move the air around... two common causes of such shifts can be circulations around oceans or other large water bodies (probably not pertinent to you!) and terrain-induced circulations (probably much more pertinent!). The ones that are probably most impactful at night are downslope winds, which are caused by more rapid cooling on the mountain than the air around it, and downvalley winds, caused when enough air builds up from downsloping. This lecture from Stephan De Wekker gives some more details on some of these features and other terrain flows.
Often slight shifts in larger-scale wind direction and the varying temperatures over different degrees of mountains and valleys can lead to quite a complicated set of rises and falls in temperature through the night.
It's also not impossible that are some issues with your station, I was initially concerned with a few things:
- The winds in particular seemed odd -- due north the entire time is always suspicious, and less than 1 mph for an entire night and early morning, while certainly not impossible, makes it a bit more questionable as to whether significant temperature advection could occur.
- Plus it's also rare to see temperature and dewpoint changing almost 100% exactly in step over long periods of changes -- moisture tends to be slower to alter except around sharper fronts (and on nights when the temperature nudges lower while the air is saturated [$T=T_d$]).
But a look at observations across the region for the time period:
does show perhaps your data isn't impossible. The meteograms from Libby [KS59] and Missoula [KMSO] below do show very weak/calm winds and fairly similar temperature/dewpoints movement.
Still, there's no north winds (except when calm, which is a placeholder when there is no wind), and a slight bit more fluctuation between temperature and dewpoint, so it wouldn't hurt to check whether the station is properly observing.
I believe typically flows down a local mountain tend to be cooling at night, as the densest air (coldest) collects along the river. Other features driven by forced air movement can bring warm air to valleys, like foehns and large-scale downsloping, but those tend to bring drier air in.
So assuming it's a valid feature, it appears to be something complex. It can take years of observations to really understand the flows around complex topographic areas. But I would posit a guess that maybe a slight wind shift gradually pushed the pooled dense valley air away from you far enough to get you into slightly warmer air away from the river valley bottom. Or maybe winds become enough to mix down warmer air above. Differences in snowcover/melting effects or clouds/precipitation might also play a part in what you experienced. But in that part of the world, I'd expect you'll see many more peculiar local temperature fluctuations over your time observing!
