I wouldn't worry about it, so long as load temps aren't going up to like 90C it's not really important. And even then, it's only really a problem because the card will force itself to run slower and use less power. It's not usually possible for it to run up to temperatures that will damage it. It literally will not do that. You can't even make it do that intentionally.
I mean... could just be fan-stop. When a GPU goes idle, a lot of times the fans will stop until it hits a set point... which could be anywhere from 40-60C, depending on the make of card. This also means that on boot, the fans might not run since it never got hot enough. At that point, it could be in the 30s. But if you run it up to say... 60C, now the fans are going. But say they're set to stop as soon as it drops below 50. In that case, the card will take a very long time to get back down to what you saw before, because the fans aren't running. What are your fan speeds?
Or maybe at around that temperature the fans spin so slowly they don't bring the temperature down much. That could also make it so it takes a good bit of time for it to wind down after doing some work... maybe even flat-lining at a higher point, depending on room temperature and case conditions. Once it goes up, the fans may not spin fast enough to bring it down to that cool 30C, just because of how the curves are set. Either way, you'll almost never idle as low as you do right after booting up fresh, not running the machine. It can take hours for it to creep all the way back down once the card saturates. I've seen it happen many times with budget cards.
Fans... it would have to be fans. Shut the machine down when you go to bed and see how things look in the morning. If it goes back down to ~30C, then you know it's just the behavior of the card. Nothing to worry about... it just holds onto a little of the heat after you work it. Those temperatures are never gonna hurt it. It's not the best cooler, but honestly even if it always idled at ~45C it's good enough.
Please don't buy an aftermarket cooler for this. It's not a good idea to go and gut a card, possibly voiding the warranty, just because it doesn't idle as low as you think it should. Not to mention, there's always a chance you'll damage the card when you pull the cooler off. And that's not even considering that if we were having real temperature problems... like it was running 95C or more under load and slowing down, that's probably a faulty (or just really bad performing) card and you should return it for another, not start buying stuff to make it work like it should.
I really doubt that's the case, though.
Hippocratic oath, man... "Do first no harm." Meaning, before you go in potentially messing something up, try things that don't have that risk. You don't go right to brain surgery for a headache! Again... check out your fan speeds. If you don't have it, download MSI Afterburner and tweak them a bit. You may be able to get it down to where it was before. If there is automatic fan stop on the card, you can bypass it by applying a custom curve that makes them spin all of the time. I did that with my Strix 2060 because the fans make an annoying noise when they spin up. If I wasn't doing that, the fans wouldn't spin at all until around 50C. I mean ASUS seem to think that's fine, because that's how they run out of the box!
Idle may seem high, but it's not like other working cards don't idle in that range. My RX 580 would idle at around 45C, before that, it was a mining card, and now it gets pushed to 85C playing games in another room. Not even close to a problem.
