I've had far more rotary encoders fail than I've had capacitive screens fail, so even an argument of higher reliability is pretty moot. Most damage that would break the capacitive touch is going to damage the rest of the screen anyways.
Finally, if it's so I can change those settings while wearing gloves, wow I'm going to increase the complexity of the car and take up more space so I can change the settings on the secondary keys without taking off my gloves when it's really cold outside someday. So much stuff just so I can do that thing I rarely do anyways slightly easier for a few days of the year, assuming I'm changing those settings while also getting in and out of the car a lot so I wouldn't want to take off my gloves for a minute.
Just put the settings behind a touchscreen. It's fine.
That doesn't sound like a dial/knob. You'd give it a single big twist or scroll to get the cursor around the right spot first. Same as old-fashioned radios.
At least with an old-fashioned radio knob you got the feedback of if you were tuning into the station by hearing it. But moving a selector on a screen?
It's like you're arguing for the MacBook Wheel, as if a knob is the most optimal way to input arbitrary choices on a computer.
If I was making such an interface, that would be a dial or knob instead of buttons.