The article implies that they assumed the beetles detected the IR of a fire, and then the researchers calculated the heat signature of the fire, and from that, how sensitive the sensors would have to be - IF in fact those beetles came from that far away, AND they used only IR to find the fire. It doesn't say that they know either of those assumptions are correct. They mention "stochastic resonance" - but the other IR 'noise' in their environment is not at all random, which would be required for that mechanism to improve their sensor accuracy. This is an awful lot of speculation, about an event that occurred in 1925. Colour me skeptical.
But it also says "before the fire could be extinguished" and it burned for 2 days. it's not clear that they arrived the first day. The beetles might have picked up on the smoke by then, or the light at night.