Skypup, if you respond to this, keep it at an 8th grade level.  We're just simple cavemen around here.

I realize it has something to do with bare metal reflecting IR back at the lens (although that may be a crude or outright wrong explanation), but can some actually explain what's happening to cause bare metallic objects to appear so strangely when viewed with thermal?

All objects radiate thermal energy.  The amount of radiation detected by a thermal camera depends on several factors.  First and foremost is the emissivity of the surface of the object you are imaging.  Emissivity is defined as the ratio of the energy radiating from an object to the energy radiated from a perfect emitter (an object that reflects no energy).  Emissivity is a unit-less number that ranges from 0 (a perfect reflector) to 1 (a perfect emitter).  In simpler terms, emissivity is the opposite of reflectivity (how much energy an object's surface reflects).  For example, polished silver has an emissivity of around .03 whereas wood has an emissivity of .95.  the higher the emissivity the less it reflects.  Metal typically has a low emissivity so it reflects a lot of thermal energy.   Imagine a rooftop is a large mirror.  What object would typically be seen reflected in that mirror?  The sky, right?  If viewed with a thermal camera , the sky would appear as cold wrt the rest of the building. Emissivity is one of the most important parameters when trying to obtain the actual surface temperature of an object using a (calibrated) thermal camera.

Most metal roofs reflect the cold sky. Mine does. Like a mirror. Even though it is painted.  Like the bottom half of steel and aluminum wheels do. Spent brass in my gravel driveway (even the old, weathered, darkened and almost impossible to see, spent brass) shows up like black (cold) cigarette butts nestled in the rocks. Take your thermal to the range and see.
Finding almost any exposed metal object is pretty easy with thermal if you get the viewing angles right.

Different elements, molecules, and chemical compounds all have differing thermal emissivity coefficient values, which reflect their capability to transfer heat to the atmosphere.

Also, the texture of the surface itself also affects the transmission of midwave IR into the atmosphere from both animate and inanimate objects, smooth vs rough surfaces will vary due to the differing surface area transmitting the heat to the atomsphere.

Aluminum and water both have very high emissivity heat coefficients.

Try looking at a metal surface with a matte finish for fun, like a piece of sanded metal. Super shiny and reflects crisply like a mirror. Most metal surfaces have a surface finish, be it roughness or an oxide layer, that is thinner than the wavelengths that thermal images (~8-12 microns). Optically thin layers tend to be "transparent" in general. Only optically thick and thus non-transparent at visible wavelengths (roughly 20 times shorter).

Aluminum has very low emissivity. Don't confuse it with anodized aluminum that has a thick anodized layer. Emissivity there is set by the (thick) aluminum oxide layer itself.

Also, the reason different species of wood trees show up different with thermal is different tree species are more vascular than others and transpire hundreds if not thousands of gallons of water per day, ie they are big water pumps. The water holds the heat real well and adds to the emissivity of the wood.

In fact, water molecules transfer heat so efficiently that they are the definition of British Thermal Units (BTU) as 1 cc of water changed 1 degree Centrigrade takes 1 BTU of thermal energy to happen.

Same things go for different types of rocks giving off different thermal emissivities at night, I don't know how many times I have stalked out a rock outcropping at night, but it has been frequent.

Here is a pic of some hogs I shot at a waterhole, you can see their thermal emissivity reflection in the water itself.



The one with the crosshairs on his head dumped alot of hot blood into the cool water that you could see too....

This exactly
I use IR on electrical for a living

He might've conflated emissivity and reflectivity.

OP, most metals have low emissivity and high reflectivity. When using a thermal imager to check the cooling efficiency and effectiveness of track car mods, we either painted a section of metal items with black paint, or stuck a black piece of tape on it. Otherwise we'd just see the reflections of other items.

this

Yep.  When you IR metals with most IR stuff, you're just measuring the reflection of ambient surroundings.

Here is a couple of Hi-Res thermal photos I took, you can see the different reflections/emissivity from the various objects in the field, human skin, hot gas/metal, water glass, propane torch and canister, plastics, trees, etc.

These are true radiometric images.