I'm going to rebut some things that you mention. This rebuttal might seem subtle or pedantic, but that subtlety is hugely important.
most people using closed fists during a fight run significant risk of hand/knuckle damage
It depends on what you mean by most people. If you include untrained people in that group then yes, they do run a risk of injury - but that injury could be anywhere from slight (i.e. skinned knuckles) to major (i.e. broken wrist).
But if you are only talking about trained people then I would dispute the claim of significant risk. Two of the reasons why we train are to reduce risk to ourselves, and to reduce risk to our opponents through using the appropriate level of force.
it is very common for fists to land upon the opponent's limbs at awkward, mechanically unsound angles
Once again, if you are talking about untrained people then I would agree - your assertion is correct for undirected or uncontrolled strikes. But as martial artists we train to hit selectively and appropriately. I am very unlikely to hit an opponent in the head in such a way that I would injure myself because I will be aiming for specific targets; if I miss my shot and hit somewhere unexpected it is because my opponent moved unexpectedly, but even so my chances of incurring a significant injury are low.
One of the major impediments to many trained fighters in real-life combat threats/situations is the fear of causing serious damage to an opponent. With this in mind, is the risk of reducing range justified by a desire to reduce harm to an opponent?
Would I replace a punch with a palm strike simply to reduce force? Absolutely not - I would simply reduce the force of the punch.
Would I replace a punch with a palm strike because of the range? No I wouldn't - I use a technique appropriate for what I want to achieve. If I want to use a palm strike then I would simply close the range a little.
As for fear of causing damage, it reduces over time as a result of training. When you're new you absolutely do run the risk of using excessive force, but years of training and sparring reduce that fear and replace it with certainty.
Is it ever possible to reliably gauge an opponent's ability accurately enough to warrant a 'harm minimisation' approach?
If you are in a position where you need to inflict harm (i.e. self defense) then you should always adopt a harm minimisation approach, irrespective of the perceived ability of your opponent. The ideal outcome is to use minimum force necessary to be able to end the situation with minimal harm to yourself and your opponent - self defense related laws are often structured with that outcome in mind.
Is the increased range provided by the fist sufficient to outweigh the risk of damage when striking to the head?
Yes, because the risk of damage isn't as great as you might perceive, especially when you are using the appropriate strike with the appropriate force.