I don’t understand your position here. Only a strict, over-the-top Scientism insists that only that which is empirically detectable is real. It sounds like you are elevating Science to some kind of ultimate authority which it is not. And just because something is empirically undetectable at a given time in human history does NOT mean that something is “unreal.” (Was the Higgs boson “unreal” until it was empirically confirmed a few years ago? That view seems similar to the bizarre but popular question about whether a tree falling in the forest produces sound even if nobody is there to hear it. Of course there are sounds produced when a tree falls in the forest. They are real, regardless as to whether some human empirically detects the air molecule vibrations.)
And that discovery came through an empirical process: the mechanic checks the gasoline tank and the fuel filter, where the insoluble sugar accumulated harmlessly (though if clogs the fuel filter it may starve the engine of fuel until the filter is cleaned or replaced.)
As a has-been linguist, I must mention that there is a kind of semantic ambiguity at work in this argument. (I don’t want to call it an equivocation fallacy because that might give the impression that there was an intention to confuse the issue, and I certainly don’t believe that that is the case here.)
When we speak of natural (as in “natural causes”), we may mean:
(1) That which exists in or is caused by nature—and thereby not made or caused by humans.
(2) That which is opposite of supernatural, something explainable through the laws of physics as understood in the matter-energy world.
Your example of a mechanic discovering sugar clogging a fuel filter doesn’t qualify as natural in the aforementioned Definition #1. Nevertheless, the cause in the example is entirely natural as understood in Definition #2. No supernatural processes were involved. The mechanic empirically determined that an insoluble contaminant was added to the fuel line, and everybody has heard of the prank where humans sometimes pour sugar or sand into gasoline tanks in order to cause trouble. Nothing supernatural is suspected. The value of Methodological Naturalism is not somehow compromised in this scenario. MN in no way failed to provide an explanation.
By the way, suppose a chimpanzee is trained to pour sugar in gas tanks. In that case both Definition #1 and Definition #2 would label the chimpanzee a “natural cause.”
[And before a professional mechanic corrects me, I certainly concede that in the vast majority of cases, the sugar would merely accumulate on the bottom of the gasoline tank and most of it would probably not reach the fuel filter at all—depending on a lot of variables which are obviously beyond the context of this thread.]