EMI suppression is more an art than a science - it's very difficult to come up with a one-size-fits-all prescriptive approach. That said, one approach is to separate the emissions problem as to being radiated or conducted (or both).
For example, it appears that your furnace issue was a conducted problem as it came from the furnace via the cabling to the exposed thermostat. Not that it matters now, but it would have been interesting to see if placing a filter on the thermostat control cable at the furnace would have blocked the high-frequency signal from conducting to the thermostat.
OTOH, as I understand it your meter shows a strong reading in the vicinity of the engine but drops off away from the engine. This suggests you have a radiated issue since if the track was conducting the emissions like an antenna, the field strength would not be so localized.
I'm not familiar with the various engines you mentioned as to know what kind of "electronics" is inside. You can read a nice summary of the use of capacitors to suppress motor brush commutation noise/arcing as a source of radiated emissions here. A few 5 cent capacitors can do wonders. If the "electronics" is an old-school basic reversing unit, then I don't think that will be the source of any meaningful EMI. If the electronics has sound or other modern features typically implemented with digital processor circuitry, then there's a high frequency clock signal which can radiate. Then it could be a shielding issue which is do-able.
When you say the radiated emissions increase when the engine is pulling rolling stock, I assume you mean from the engine itself...likely from increased brush noise as more current is flowing thru the motor windings to pull the additional load. If you had lighted passenger cars, you will get arcing from intermittent wheel/roller contact with the track which I suppose is similar to brush/commutation noise. That would be a more interesting problem to solve since the arcing (the emissions) are occurring at the track so difficult to shield.
Same thing with the transformers. If the meter readings drop off with distance from the transformer irrespective of track, then it seems to me the dominant problem is radiated emissions from the transformer and its "electronics" if there is any. I'm pretty sure the CW-80 has a digital microprocessor circuit inside which has a clock circuit running the the MHz. I can't believe they would have designed it to limit RF emissions other than to "just" pass FCC (or similar) standards for consumer electronics. Obviously your personal requirements far exceed these standards. So like an engine, if a high frequency digital circuit is radiating from the transformer, some form of shielding (metal box) around the circuit can help. If it's a conducted problem (going out the cables like your furnace) then the low-pass filters will help. Messing with inductors/chokes can be tedious and you pretty much need to go to a specialty electronics components distributor. To that end, I dug out my old Radio Shack paper catalog as I recall they sold really easy to use clamp-on snap-on filters where you don't have to cut the cable. Depending on what frequencies you're trying to suppress, these can be a quick solution if you have a conducted problem (emissions from the input/output cables like your furnace).
